Snail/Slug-YAP/TAZ complexes cooperatively regulate mesenchymal stem cell function and bone formation

ABSTRACT

Snail and Slug are zinc-finger transcription factors that play key roles in directing the epithelial-mesenchymal transition (EMT) programs associated with normal development as well as disease progression. More recent work suggests that these EMT-associated transcription factors also modulate the function of both embryonic and adult stem cells. Interestingly, YAP and TAZ, the co-transcriptional effectors of the Hippo pathway, likewise play an important role in stem cell self-renewal and lineage commitment. While direct intersections between the Snail/Slug and Hippo pathways have not been described previously, we recently described an unexpected cooperative interaction between Snail/Slug and YAP/TAZ that controls the self-renewal and differentiation properties of bone marrow-derived mesenchymal stem cells (MSCs), a cell population critical to bone development. Additional studies revealed that both Snail and Slug are able to form binary complexes with either YAP or TAZ that, together, control YAP/TAZ transcriptional activity and function throughout mouse development. Given the more recent observations that MSC-like cell populations are found in association throughout the vasculature where they participate in tissue regeneration, fibrosis and cancer, the Snail/Slug-YAP/TAZ axis is well-positioned to regulate global stem cell function in health and disease.

KEYWORDS:

• Hippo
• mesenchymal stem cells
• RUNX2
• skeletal stem cells
• Slug
• Snail
• TAZ
• TEAD
• YAP

Mesenchymal stem/stromal cells (MSCs) are a specialized population of progenitor cells that are positioned throughout host tissues during embryogenesis and postnatal life.^1,2 Though the full multipotent range of these cells remains the subject of ongoing debate, bone marrow-derived MSCs (also referred to as skeletal stem cells) of either mesodermal or neural crest origin are able to differentiate into osteoblasts, chondrocytes or adipocytes as well as haematopoietic stem cell-supportive stroma.^2-4 Loss of stem cell function not only leads to developmental defects, but also contributes to postnatal osteoporosis with associated defects in fracture repair, haematopoietic potential and tissue homeostasis.^1,2 Despite their importance, however, the molecular mechanisms by which MSC self-renewal and function are maintained in the in vivo setting remain unclear.

The zinc-finger transcriptional repressors, Snail and Slug (alternatively termed, Snail1 and Snail2, respectively), are best known for their respective roles in orchestrating the epithelial-mesenchymal transition (EMT) programs normally associated with embryogenesis, early development and a growing list of pathologic states, ranging from cancer to fibrosis.^5-12 More recently, however, Snail and Slug functions have been expanded to include the control of fate determination in stem cell populations where expression of either transcription factor has been proposed to selectively modulate self-renewal, pluripotency, proliferative activity and senescence.^9,13-22 Nevertheless, the ability of endogenous Snail or Slug to regulate MSC function in the in vivo setting has, until recently, only been the subject of speculation.

In recent work, we discovered that Snail and Slug are co-expressed in bone marrow MSCs during both pre- and post- natal states in vivo.²³ Though effects of Snail/Slug on stem cell function are most frequently assigned to embryonic or epithelial cell populations,^5-7 these data demonstrated that MSCs – despite their mesenchymal origin – co-express both transcription factors, reinforcing the fact that these transcription factors assume functional roles outside of classic EMT-related programs.^24,25 However, following in vivo targeting of either MSC-derived Snail or Slug alone, we noted only subtle defects in bone development.²³ These results contrasted with earlier reports proposing roles for Snail alone in regulating osteoblastogenesis.^25-27 Likewise, whereas individual roles for either Snail or Slug have been linked to the control of chondrogenesis,^28,29 we were unable to document frank defects in cartilage development, a finding similar to that first reported by Gridley and colleagues.^23,30 Despite these negative results, spurred by recent reports documenting the compensatory regulation of Snail and Slug during early cartilage formation as well as craniofacial development,^30-32 we elected to target both MSC transcription factors in tandem. Remarkably, under these conditions, MSC self-renewal and differentiation were impaired strikingly with attendant defects in osteoprogenitor function that completely undermined bone formation in vivo as well as terminal osteoblast differentiation in vitro.²³ These results should not, however, be construed to suggest that Snail and Slug necessarily compensate for one another under all circumstances. For example, the tissue-specific deletion of Snail alone undermines developmental programs associated with vascular development,³³ intestinal homeostasis,¹⁹ pancreatic acinar cell maintenance²² and neural maturation.³⁴ Likewise, Slug alone appears to regulate adipogenesis via a Snail-independent process while Snail can, in and of itself, modulate insulin-induced lipolysis.^35,36 Indeed, Snail and Slug function may even exert opposite effects in selected events, such as stem cell reprogramming.³⁷ Nevertheless, similar to defects observed in chondrocytes or neural crest stem cells,^30-32 only the dual loss of Snail and Slug unmasked the key roles played by these transcription factors in developing bone.²³

From a functional perspective, the ability of Snail and Slug to cooperatively regulate MSC proliferation as well as differentiation was unexpected, leaving us to conjecture as to how those transcription factors might control bone formation. We were, however, intrigued by the fact that the observed effects of Snail/Slug targeting on MSC self-renewal and lineage commitment dovetailed with those previously ascribed to the transcriptional co-activators, Yes-associated protein (YAP) and transcriptional co-activator with PDZ-binding motif (TAZ).^38-43 These molecules, through their direct binding interactions with the TEAD family of transcription factors, serve as key mediators of the Hippo pathway that regulates organ size, cell proliferation and cell fate determination.^38-43 Seminal work in the field had also established the fact that YAP and TAZ play critical roles in stem cell self-renewal, maintenance and expansion.^38-44 Indeed, with specific regard to bone marrow-derived MSCs, YAP/TAZ not only control stem cell fate, but also serve as downstream mediators of cell-extracellular matrix interactions.^45,46 Though the possibility that Snail and Slug coordinate MSC behavior by working in tandem with the YAP and TAZ axis had not been explored previously, we further noted that YAP and TAZ, like Snail/Slug, bind to 14.3.3, translocate from the cytosolic to nuclear compartment, undergo phosphorylation-dependent β-TRCP-mediated ubiquitination and proteasomal degradation, interface with the canonical Wnt pathway as well as TGFβ signaling, and are responsive to mechanical inputs.^6-8,38-43 Furthermore, like Snail and Slug, several recent reports have established compensatory roles for YAP and TAZ during development.^47-53 Indeed, these multiple, but potentially coincidental, similarities proved informative as we found that Snail and Slug control both the transcriptional and post-transcriptional regulation of YAP/TAZ in mouse as well as human MSCs.²³ That is, in the combined absence of both Snail and Slug, YAP/TAZ protein, but not mRNA, levels decreased markedly.²³ Likewise, Snail/Slug co-deletion inhibited expression of multiple YAP/TAZ-TEAD targets, including Ctgf, Ankrd1, Axl, Dkk1 and Cyr61 in vitro as well as in vivo.²³ Interestingly, effects on gene expression did not reflect global changes in MSC transcriptional program as Runx2 expression (both at the mRNA and protein levels), a master regulator of bone and cartilage development,^54-56 remained unaffected. Indeed, changes in stem cell-associated transcription factors, e.g., Nanog, Oct4 or Sox2, are not detected in Snail/Slug double-null MSCs relative to wild-type controls (Fig. 1). Hence, despite the combined absence of Snail/Slug, MSCs commit to the osteoprogenitor lineage in normal fashion, but are unable to complete the terminal steps of the osteoblastogenic program in vitro or in vivo.²³

Figure 1. Relative expression of stem cell markers in MSCs isolated from Snail^f/f/Slug^f/f mice and transduced with adenoviral- GFP or Cre expression vectors.

Under normal conditions, YAP/TAZ stability is regulated, in large part, by the Hippo cascade-associated kinases, MST1 and MST2 as well as the Lats kinases, LATS1 and LATS2; the latter kinases mediating YAP/TAZ phosphorylation, thereby controlling YAP/TAZ subcellular localization and proteosomal degradation.^38-43 However, Snail/Slug co-deletion did not affect MST or LATS activity in isolated MSCs.²³ Nevertheless, in the absence of Snail/Slug, the levels of phosphorylated YAP/TAZ were markedly increased with concomitant decreases in YAP/TAZ protein half-lives and steady-state levels.²³ As such, next considered the possibility that Snail/Slug might modulate YAP/TAZ function via direct binding interactions given recent precedents wherein i) the tyrosine phosphatase, SHP2, was reported to bind and modulate YAP/TAZ-TEAD-dependent transcriptional activity ⁵⁷ ii) and the transcription factor, TXB5, was found to form function-modulating complexes with YAP and β-catenin.⁵⁸ Indeed, we found that biomolecular complexes were readily detected between Snail and either YAP or TAZ, or alternatively, between Slug and YAP or TAZ, both in vitro and in vivo.²³ Consistent with the fact that binding interactions between YAP/TAZ and nuclear transcription factors are mediated by their respective WW domains,^38,39 the formation of Snail/Slug-YAP/TAZ complexes in MSCs was likewise WW domain-dependent.²³ In turn, Snail/Slug binding interactions with the YAP/TAZ WW domains were dependent on their extended, N-terminal SNAG domains that also serve to recruit chromatin modifying enzymes critical for Snail/Slug-dependent transcriptional regulation.^5,6 WW domains classically interact with protein partners via embedded PPxY motifs,^38,39,59 but the extended SNAG domains of neither Snail nor Slug include this sequence.⁵ Though additional molecules might be integrated into Snail/Slug-YAP/TAZ complexes that serve to bridge these proteins together, we find that bacterially-expressed Snail/Slug and YAP/TAZ are able to form binary complexes directly (Fig. 2). Interestingly, the extended SNAG domains of Snail and Slug contain a conserved PxY motif that potentially participates in YAP/TAZ binding,⁵ but further studies are needed to identify the critical sequences involved in WW-Snail/Slug domain interactions and whether these interactions might apply to other members of the Snail Family.⁶⁰

Figure 2. Snail/Slug-YAP/TAZ binding interactions. GST-YAP, GST-TAZ or GST alone were bound to glutathione beads and incubated with recombinant Flag-tagged Snail or Slug proteins. Aliquots of Flag-tagged proteins and each bound fraction were subjected to SDS-PAGE and immunoblotted with the anti- Flag and GST antibodies.

Given the identification of Snail/Slug-YAP/TAZ binding interactions, their potential functional significance was underlined following the detection of these complexes in the promoter regions of a series of YAP/TAZ-TEAD-targeted genes.²³ Likewise, confirming and extending earlier studies describing the ability of Runx2 to bind YAP or TAZ directly,^61,62 Runx2-YAP/TAZ-Snail/Slug complexes were likewise identified in the promoter region of multiple genes known to regulate osteoblast differentiation and function.²³ A requirement for Snail/Slug-YAP/TAZ complex formation in MSC differentiation programs was further reinforced by the ability of a dominant-negative Snail-deletion mutant that efficiently blocks binding interactions between endogenous Snail/Slug and YAP/TAZ, to completely inhibit osteoblastogenesis in a fashion that recapitulated the phenotype of Snail/Slug-null or YAP/TAZ hemizygote MSCs.²³ Hence, in the absence of Snail and Slug, YAP/TAZ alone is unable to support bone development while partial, but combined, deficiencies of YAP and TAZ undermine the ability of Snail and Slug to drive osteoblastogenesis.²³ Taken together, these data describe a model where, Snail, Slug, YAP and TAZ play critical roles in cooperatively regulating the transcriptional events downstream of Runx2 activation that play required roles in osteoblast differentiation and bone formation (Fig. 3).²³ Interestingly, whole-genome ChIP studies have identified both TEAD and Runx family members as YAP-binding transcription factors.^62-65 However, at least in the case of YAP/TEAD complexes, only a handful of target genes were identified whose expression was regulated by YAP-dependent binding in specific promoter regions.^62-64 Instead, transcriptional control was found to be exerted by YAP/TEAD complex-mediated action on distal enhancers that contacted target promoters through chromatin looping.^62-65 Even with this information in hand, the transcriptional targets regulated by these types of interactions in MSCs remains to be determined as DNA-binding sites for YAP-TEAD complexes, to date analyzed most frequently in cancer cells, are distinct when comparing transformed to non-transformed cell lines.⁶³

Figure 3. Schematic model of Snail/Slug-YAP/TAZ complexes in regulating MSC function.

Until recently, bone marrow MSCs were often times treated as a homogenous population.^1,2 However, it is now appreciated that multiple and distinct MSC populations participate in early bone formation as opposed to postnatal skeletal homeostasis and fracture repair.^1-4,66-68 As such, these findings raise the question as to whether the functional roles defined for Snail/Slug-YAP/TAZ complexes are confined to early MSC development or extend into the adult setting.^66-68 We found, however, that MSCs recovered from adult mice continue to rely on Snail/Slug-YAP/TAZ activity to maintain self-renewal properties as well as undergo osteoblastogenesis.²³ Nevertheless, Snail/Slug-YAP/TAZ-dependent MSC function is unlikely to be maintained throughout the lifespan of normal mice. Indeed, we find that the well-known decrease in the size and function of MSC reserve pools that occurs during aging ⁶⁹ coincides with marked losses in Snail, Slug, and YAP/TAZ expression levels (Figs. 4 and 5). Given complementary studies demonstrating important roles for Snail/Slug as well as YAP in chondrogenesis,^30,70 the identification of the factors that regulate MSC Snail/Slug and YAP/TAZ levels could have important ramifications for improving overall skeletal health in aging populations. Likewise, we presume that it will prove informative to identify the mediators that control the Snail/Slug-YAP/TAZ axis and the ability of MSCs to re-engage this pathway as a function of age following bone or cartilage damage.

Figure 4. Expression of Snail and Slug in young versus aged mice. (A) LacZ expression in 10 day-old vs. 6 month-old parietal bone. Scale bar: 50 µm. (B) LacZ expression in 2 day-old vs. 12 month-old mice femur. (C) LacZ expression in 10 day-old vs. 6 month-old mice femur. Scale bar: 100 µm.

Figure 5. Expression of YAP and TAZ in young vs. aged mice. Immunostaining of YAP/TAZ in the femurs of 2 day-old vs. 8 month-old mice are shown. Scale bar: 100 µm.

Studies performed to date have focused on the regulation of the Snail/Slug-YAP/TAZ axis in bone marrow-derived MSCs, but pericyte-like MSCs are now known to be associated with the vasculature throughout the body where they have likewise been linked to events ranging from tissue repair and cancer to fibrosis.^71-76 As Snail/Slug as well as YAP/TAZ can participate in all of these events,^7,42,43 we presume that the functional roles for the Snail/Slug-YAP/TAZ axis outlined herein may also extend to MSC function outside the confines of the bony skeleton, or even the vasculature.⁷⁷ Further, as cancer cells themselves can display stem cell-like characteristics,^8,9,13,16-18 and are known to express Snail/Slug as well as YAP/TAZ, functional roles for the Snail/Slug-YAP/TAZ axis in neoplastic processes bear further consideration. Finally, from the perspective of YAP/TAZ function, limiting such analyses to interactions with Snail/Slug alone is likely premature as complementary studies have recently established similar binding interactions between YAP and ZEB1, an additional, but structurally distinct, member of the family of EMT-associated zinc-finger transcription factors.^6,8,78 In this case, the interaction of ZEB1 and YAP switches ZEB1 from a transcriptional repressor to a transcriptional co-activator that promotes stemness, proliferation, metastatic activity and chemoresistance in breast carcinoma cells.⁷⁸ As such, and given the widespread expression of Snail/Slug as well as YAP/TAZ in both epithelial as well as mesenchymal cell populations in vivo, the interacting transcriptional networks we have outlined herein will likely prove pertinent to the regulation of stem cell-associated programs in a wide variety of both normal and neoplastic states.

Methods

Mice

Snail^flox/flox and Snail^+/LacZ mice were generated as described previously and maintained on a C57BL/6 genetic background.^15,24,33 YAP^f/f/TAZ^f/f mice were previously developed and provided by E. Olsen.⁴⁷ Slug^+/LacZ mice were obtained from T. Gridley and maintained on a C57BL/6 background.⁷⁹ All mouse work was performed with IACUC approval.

GST-pull down assay

Flag-Snail and Flag-Slug expression plasmids ²³ were transfected into HEK-293 cells, and cell lysates were harvested 2 d after transfection. The Flag-Snail and Flag-Slug fusion proteins were isolated and purified with Anti-Flag M2 affinity resin (Sigma-Aldrich), and the purified proteins eluted with a Flag peptide (Sigma-Aldrich). GST-YAP and GST-TAZ (Addgene) fusion proteins were produced in bacteria and purified by glutathione-agarose affinity chromatography. Protein-protein interactions were determined using a GST-Pull Down assay kit according to the manufacturer's instructions (Thermo Scientific).

RT-PCR

Bone marrow MSCs have been isolated and cultured as described previously.²³ Total RNA was harvested from the cultured cells and RT-PCR performed according to the manufacturer's instructions (Power SYBR Green, Applied Biosystems). Primers for Nanog, forward: AGGGTCTGCTACTGAGATGCTCTG and reverse: CAACCACTGGTTTTTCTGCCACCG. Oct4, Forward: TGGAGACTTTGCAGCCTGAG and reverse: AAGCGACAGATGGTGGTCTG. Sox2, Forward: GAAGGGGAGAGATTTTCAAAG and the reverse: CCAGATCTATACATGGTCCGATTCCC.

LacZ staining and immunostaining

Mouse bone was harvested and the LacZ staining and immunostaining performed as described.²³

Disclosure of potential conflicts of Interest

The authors declare no conflict of interest.

Funding

This work is supported by the National Institutes of Health under grant number AR065524 (SJW) and the Breast Cancer Research Foundation (SJW).