Wnt-YAP interactions in the neural fate of human pluripotent stem cells and the implications for neural organoid formation

ABSTRACT

Human pluripotent stem cells (hPSCs) have shown the ability to self-organize into different types of neural organoids (e.g., whole brain organoids, cortical spheroids, midbrain organoids etc.) recently. The extrinsic and intrinsic signaling elicited by Wnt pathway, Hippo/Yes-associated protein (YAP) pathway, and extracellular microenvironment plays a critical role in brain tissue morphogenesis. This article highlights recent advances in neural tissue patterning from hPSCs, in particular the role of Wnt pathway and YAP activity in this process. Understanding the Wnt-YAP interactions should provide us the guidance to predict and modulate brain-like tissue structure through the regulation of extracellular microenvironment of hPSCs.

KEYWORDS:

• neural organoid
• pluripotent stem cell
• three-dimensional
• Wnt
• YAP

ABBREVIATIONS

3-D	=	three-dimensional
DKK	=	Dickkopf
EBs	=	embryoid bodies
ECM	=	extracellular matrix
ESCs	=	embryonic stem cells
FGF	=	fibroblast growth factor
GSK 3	=	glycogen synthase kinase 3
hESCs	=	human embryonic stem cells
hiPSCs	=	human induced pluripotent stem cells
hPSCs	=	human pluripotent stem cells
IWP2 (or 4)	=	inhibitor of Wnt-production 2 (or 4)
IWR-1	=	inhibitor of Wnt-response 1
LRP6	=	low-density lipoprotein receptor-related protein 6
MSCs	=	mesenchymal stem cells
NPCs	=	neural progenitor cells
PSCs	=	pluripotent stem cells
RA	=	retinoic acid
ROCK	=	Rho-associated protein kinase
SHH	=	sonic hedgehog
TAZ	=	transcriptional coactivator with PDZ-binding motif
YAP	=	Yes-associated protein

INTRODUCTION

Human pluripotent stem cells (hPSCs), including human embryonic stem cells (hESCs) and human induced pluripotent stem cells (hiPSCs), have great potential in drug screening, disease modeling, and potentially in cell therapy.^1-3 Recently, it is found that hPSCs can self-assemble into three-dimensional (3-D) organoids, in particular the “mini-brains”, which have the organized structure to mimic human tissue morphogenesis.^4-6 Midbrain organoids have also been derived from hPSCs, containing long-lived dopaminergic neurons which otherwise have short-term life span in vitro.⁷ These mini-organoids provide a novel platform as physiologically relevant human brain tissue models for drug screening and disease modeling.^5,8 Wnt signaling is one of the major pathways that serve as extrinsic signaling to regulate stem cells for neural tissue development and regeneration.⁹ Wnt signaling can switch rostral and caudal brain tissue identity in human neural tissue morphogenesis in a concentration-dependent manner.¹⁰ In the meanwhile, Hippo/Yes-associated protein (YAP) signaling has been identified as another important pathway that plays a critical role in organ growth control as well as stem cell propagation and differentiation.¹¹ The interactions of Wnt signaling and YAP expression attract extensive interests in regenerative medicine recently,^12,13 but how Wnt-YAP interactions affect neural tissue morphogenesis from hPSCs has been unexplored. Here, the biological significance of Wnt signaling and YAP expression in neural organoid formation from hPSCs is discussed.

THE ROLE OF CANONICAL WNT/β-CATENIN SIGNALING IN HPSC FATE DECISIONS

Wnt signaling has been found to play an important role in regulating hPSC self-renewal and differentiation through intercellular interactions (Table 1 and Fig. 1).¹⁴ At the undifferentiated state, hPSCs express high level of β-catenin at cell membrane due to the formation of cadherin-catenin complex with little nuclear localization. Increased E-cadherin expression can sequester β-catenin at the adherent junctions, which results in less β-catenin available for translocation to the nucleus and the down-regulation of Wnt signaling.¹⁵ The loss of E-cadherin expression enables β-catenin translocation into nucleus to activate Wnt signaling. During differentiation, the influence of Wnt/β-catenin has the context-dependent effects on hPSC fate decisions.

Figure 1. Illustration of the interactions of extracellular microenvironment (soluble factors, matrices etc.) with Wnt/β-catenin signaling through biochemical regulation. The effects of Wnt activation (e.g., Wnt 3A or CHIR99021) and inhibition (e.g., DKK1 or IWP4) on pluripotent stem cell (PSC) lineage commitment are shown. GSK-3β: glycogen synthase kinase-3β; TCF: T-cell factor; LEF: lymphoid enhancer factor. FGF, fibroblast growth factor; HGF: Hepatocyte growth factor; IGF: Insulin-like growth factor; TGF, transforming growth factor.

TABLE 1. Summary of the role of Wnt signaling in pluripotent stem cell fate decisions.

Applications	The role of Wnt signaling	Characterization	Ref.
Pluripotency and reprogramming	Activation of Wnt signaling with Wnt3a or GSK inhibitor BIO maintains self-renewal of hESCs	Expression of pluripotent markers in a short-term culture.	Sato et al, 2004 [49].
	Wnt activation (Wnt3a) promotes the reprograming of the fibroblasts into iPSCs	Addition of Wnt3a can reprogram fibroblasts in the absence of c-Myc. About 20-fold more drug resistant colonies were formed in the presence of Wnt3a.	Marson et al, 2008 [50]
	Wnt/β-catenin promotes mouse Episc (epiblast stem cells) and hESC self-renewal when stabilized β-catenin is retained in the cytoplasm.	Using Wnt activator CHIR or Wnt inhibitor IWR-1 to determine how β-catenin localization affects the pluripotency of hESCs. β-catenin translocation into the nucleus induces differentiation.	Kim et al, 2013[51]
Mesoderm and cardiovascular differentiation	Temporal modulation of Wnt signaling promotes mesoderm and cardiac differentiation of PSCs	CHIR activation of Wnt signaling followed by the Wnt inhibition with IWP4 results in the efficient generation of high purity of cardiomyocytes (>90%).	Lian et al, 2012 [17]
	Efficient differentiation of hPSCs to endothelial progenitors by Wnt activation.	CHIR activation of Wnt signaling without the following Wnt inhibition results in the efficient generation of endothelial progenitors. Additional magnetic purification step is required.	Lian et al, 2014 [27]
Ectoderm and neural differentiation (early induction)	Inhibition of Wnt signals promotes anterior character and enhances neurectodermal differentiation in embryoid bodies (EB).	Using Wnt3a for activation and DKK1 for inhibition. Characterized by EB structure (polarity) and the expression of Pax6, Otx2, Sox1 ectoderm genes.	ten Berge et al, 2008 [16]
	Wnt^low hPSCs are prone to neuroectodermal induction.	Wnt^high cells differentiated into endodermal and cardiac cells (FOXA2^+), while Wnt^low cells differentiated into neuroectodermal cells (PAX6^+ and Otx2^+)	Blauwkamp et al, 2012 [9]
Neural tissue patterning (from rostral to caudal identity)	Wnt inhibition (using DKK1 or XAV-939) in combination with SHH signaling promotes ventral patterning of telencephalic neural precursors	In combination with the dual SMAD inhibitors, XAV-939 treatment increases FOXG1 and Nkx2.1 expression (with forebrain/midbrain identity).	Nicoleau et al, 2013 [18]
	Regulate the heterogeneity of hPSC-derived NPCs and the subsequent neural differentiation along rostral and caudal identity.	With high endogenous Wnt signaling, NPCs adopt a posterior neural fate by expressing PITX2, FGF8, IRX3, HOXB4, the markers for hindbrain/spinal cord, low Wnt signaling enriches forebrain neurons.	Moya et al, 2014 [19]
	Activation of Wnt signaling promotes motor neuron differentiation from hPSCs	In the presence of Wnt activator CHIR, motor neuron progenitors (>95% Oligo2^+) were generated in 12 days. Spinal motor neurons (74% ISL1^+HB9^+, use CHIR 3 μM and RA 1 μM) were generated in 14 days.	Du et al, 2015 [20] Maury et al, 2015 [22]

The role of Wnt signaling in early stage neural differentiation from hPSCs

During early stage ectoderm differentiation from PSCs, inhibition of Wnt signals promotes anterior character and enhances neurectodermal differentiation, while local activation of Wnt signaling posteriorizes embryoid body (EB) and promotes mesendodermal differentiation.¹⁶ Consistently, efficient cardiac differentiation from hPSCs requires temporal modulation of Wnt signaling, i.e., early activation of canonical Wnt signaling followed by the inhibition of Wnt activity.¹⁷ Characterization of Wnt^high and Wnt^low populations in hESCs reveals that Wnt^high cells mainly differentiate into endodermal and cardiac cells (FOXA2⁺), while Wnt^low cells differentiate into neuroectodermal cells (PAX6⁺ and Otx2⁺).⁹ For neural differentiation of hPSCs using LDN193189 and SB431542 (dual SMAD inhibitors), Wnt inhibition (e.g.,, using XAV-939) in combination with sonic hedgehog (SHH) signaling promotes the telencephalic specification and ventral patterning of telencephalic neural precursors.¹⁸

The role of Wnt signaling in neural tissue patterning of hPSCs

Wnt signaling was found to regulate the heterogeneity of hPSC-derived neural progenitor cells (NPCs) and the subsequent neural differentiation along rostral and caudal positional identity. In the presence of high endogenous Wnt signaling, NPCs adopt a posterior neural fate by expressing PITX2, FGF8, IRX3, and HOXB4, the caudal neural markers for hindbrain and spinal cord.¹⁹ In contrast, in the absence of endogenous Wnt signaling, NPCs are enriched for rostral forebrain and anterior-specific markers (e.g., FOXG1, DLX2, LHX2 and LHX8) [19]. The intermediate level of Wnt signaling leads to the enrichment of midbrain neurons. Consistently, exogenous activation of Wnt signaling with CHIR99021 increases the expression of posterior markers such as HOXB4, while inhibitor of Wnt-production 2 (IWP2) treatment increases the expression of anterior markers such as FOXG1 [19]. Therefore, Wnt signaling plays a critical role in the regional patterning and positional identity of hPSC-derived NPCs.

Due to the neural patterning effect, Wnt signaling can efficiently promote motor neuron differentiation from hPSCs, in combination with caudalization factor retinoic acid (RA) and ventralization factor SHH.^20,21 Wnt signaling may elevate the threshold level of SHH signaling to enrich motor neural progenitors. Using dual SMAD inhibition in the presence of Wnt activator CHIR99021, high purity of motor neural progenitors (>95% Oligo2⁺) can be generated in 12 days.²⁰ Consistently, high purity of spinal motor neurons (74% ISL1⁺HB9⁺ cells, by activating Wnt signaling with CHIR99021 at 3 μM and RA at 1 μM) or cranial motor neurons (52% ISL1⁺PHOX2B⁺ cells, by lowering Wnt and RA concentrations with CHIR99021 at 1 μM and RA at 0.01 μM) were generated in 14 days by early exposure to Wnt signaling.²² These studies indicate that Wnt activation induces a caudal fate of neural progenitors that more easily differentiate into motor neurons.

While the mentioned studies demonstrated the effects of Wnt signaling on generating specific types of neural cell populations with high purity at cellular levels, the influence of Wnt signaling on spatial patterning of different neural populations from hPSCs in 3-D culture at tissue/mini-organ levels is still unexplored and needs further investigation.

Wnt activators and inhibitors

To achieve effective Wnt activation or inhibition, various Wnt activators and inhibitors have been developed to modulate Wnt signaling, which may act through different mechanisms (Table 2).

TABLE 2. Wnt antagonists and agonists.

Category	Molecules	Action mechanism	References
Wnt antagonists	Dickkopf (DKK) family, e.g., DKK1	Inducing internalization and degradation of low-density lipoprotein receptor-related protein 6 (LRP6) through transmembrane Kremen proteins	MacDonald et al, 2009 [23]
	WISE/SOST family	The LRP5/6 ligands/antagonists, disrupt Wnt-induced Frizzled-LRP6 complex	MacDonald et al, 2009 [23]
	Inhibitors of Wnt-production, e.g., IWP2, IWP4	Prevents palmitoylation of Wnt proteins by Porcupine	Chen et al, 2009 [24]
	XAV939	A Tankyrase inhibitor that inhibits Wnt/β-catenin signaling.	Davidson et al, 2012 [14]; Nicoleau et al, 2013 [18]
	Inhibitors of Wnt-response, e.g., IWR-1	An antagonist which inhibits Wnt/β-catenin through the stabilization of Axin	Chen et al, 2009 [24]
Wnt agonists	Wnt ligands (canonical), e.g., Wnt3a	Bind to Frizzled receptor or its coreceptor LRP6 to activate Wnt pathway	Azzolin et al, 2014 [12].
	Wnt ligands (non-canonical), e.g., Wnt5a	Wnt activation independent of β-catenin	Park et al, 2015 [13]
	GSK inhibitors, such as CHIR99021	Activate Wnt through GSK-3β inhibition	Lian et al, 2012 [27]; Maury et al, 2015 [22]
	Norrin	A specific ligand for Frizzled 4	MacDonald et al, 2009 [23]
	R-spondin	Exhibit synergy with Wnt, Frizzled receptor, and LRP6; a ligand for Kremen and antagonizes DKK/Kremen interactions	Astudillo et al, 2014 [29]

1. Wnt antagonists (inhibitors). Dickkopf (DKK) family and the WISE/SOST family are the well-known Wnt antagonists. DKK1 inhibits Wnt signaling by inducing internalization and degradation of low-density lipoprotein receptor-related protein 6 (LRP6) through transmembrane Kremen proteins.²³ Similarly, WISE/SOST also belongs to the LRP5/6 ligands/antagonists that can inhibit Wnt signaling. Inhibitor of Wnt-production 4 (IWP4) prevents palmitoylation of Wnt proteins by Porcupine, thereby blocking Wnt protein secretion and activity.²⁴ Addition of IWP4 in the slow rotary culture was shown to reduce the expression of cardiac genes MLC-2v and αMHC and inhibit cardiac differentiation of PSCs.²⁵ Similarly, IWP2-treated hESCs have high neuroectodermal differentiation potential and express higher levels of Pax6 and Otx2 but lower level of Gata4.⁹ XAV939 is a Tankyrase inhibitor that inhibits Wnt/β-catenin signaling. XAV939 can block the effect of Wnt3a on the induction of mesoderm and primitive streak.¹⁴ Inhibitor of Wnt response (IWR)-1, acting similarly to XAV939, is an antagonist which inhibits Wnt/β-catenin through the stabilization of Axin.

2. Wnt agonists (activators): Wnt proteins such as Wnt3a promote Wnt/β-catenin signaling by acting as the ligands to Wnt receptors. Inhibition of glycogen synthase kinase 3 (GSK3) (e.g.,, using GSK3 inhibitor CHIR99021) leads to nuclear accumulation of β-catenin, which associates with T-cell factor/lymphoid enhancer-binding factor and activates Wnt-targeted gene transcription.¹⁷ CHIR99021 in combination with PD0325901 (a potent mitogen-activated protein kinase inhibitor) supports PSC long-term self-renewal and can significantly improve the reprogramming efficiency of mouse embryonic fibroblasts.²⁶ Activation of Wnt using CHIR99021 also promotes mesoderm differentiation that can lead to the high purity of cardiac lineage and endothelial lineage cells.²⁷

WNT INTERACTIONS WITH EXTRACELLULAR MATRIX AND THE EFFECT OF CULTURE SYSTEMS

The extracellular control of Wnt signaling depends on the extracellular storage of Wnt proteins and the secretion of Wnt antagonists such as frizzled-related proteins and DKK1.²⁸ In this regard, proteoglycans and glycosaminoglycan chain components may affect Wnt/β-catenin pathway through the interactions with Wnt ligands or inhibitors. Extracellular matrix (ECM) components, such as biglycan, glypican, and heparan sulfate, are able to modulate the activation or the inhibition of Wnt/β-catenin signaling by binding Wnt to the frizzled receptor and its coreceptor LRP6.²⁸ It has been reported that biglycan directly interacts with Wnt3a at C-terminus, which enhances β-catenin/T cell factor mediated transcriptional activity.²⁸ Syndecan-4 was shown to inhibit Wnt/β-catenin signaling through regulation of LRP6 and R-spondin 3.²⁹ The influence of acellular ECMs on β-catenin expression and the response to Wnt activator and inhibitor were demonstrated recently.³⁰ The ECMs derived from PSC-NPCs reduce the expression of β-catenin compared to the ECMs derived from undifferentiated PSCs and spontaneously differentiated EBs. ECM rigidity also affects Wnt signaling through the Rho GTPase signaling (Fig. 2).³¹ Rho GTPase is a key regulator of intracellular contractility and, thus, allows cells to sense matrix stiffness and respond to mechanical cues. Increased matrix rigidity up-regulates Rho and Rac, which lead to the synthesis of phosphoinositide lipid (PIP2) and the phosphorylation of LRP5/6 coreceptors, and thus indirectly activates canonical Wnt signaling (Fig. 2). Matrix rigidity was also found to influence Wnt signaling by the downregulation of DKK1 protein, which results in elevated membrane type 1 matrix metalloproteinase expression.³² Collectively, these studies demonstrate the important role of ECMs in modulating Wnt signaling.

Figure 2. Illustration of the interactions of extracellular microenvironment (matrix rigidity, geometry etc.) with Wnt/β-catenin signaling through biophysical regulation. Rigid matrix increases the RhoA and Rac which activate phosphatidylinositol 4-phosphate 5-kinase (PIP5K) and phosphatidylinositol 4-kinase (PI4K). These kinases synthesize the phosphoinositide lipid (PIP2) which enhances Wnt signaling. Rigidity also results in elevated membrane type 1 matrix metalloproteinase (MT1-MMP) expression, a transcriptional target of β-catenin/T-cell factor (TCF), and thereby promotes Wnt signaling. Dvl: Dishevelled.

Changes in the spatial localization and phosphorylation state of β-catenin expression can be regulated by the PSC aggregation process through modulating culture systems, e.g., the speed of rotary culture (25–55 rpm) or 3-D microwell culture, which affect the subsequent cardiomyocyte differentiation.^15,25 Slower rotary speed (25 rpm) was shown to increase the nuclear accumulation of β-catenin at early time point and the expression of cardiac genes MESP-1 and MEF-2C.²⁵ 3-D microwell culture was shown to promote the membrane localization of β-catenin and thus downregulate Wnt signaling. However, the EBs formed from 3-D microwell culture contain more canonical Wnt signaling activity at early stage of differentiation and promote cardiac differentiation of hESCs.¹⁵ These studies indicate that modulation of culture systems of PSC aggregates affects Wnt signaling.

THE ROLE OF HIPPO/YAP SIGNALING IN STEM CELL FATE DECISIONS

Hippo/YAP signaling has been found to play a critical role in the effects of substrate mechanics on stem cell propagation and differentiation (Table 3). It also affects cell density and organ size control through the interactions with Wnt signaling.³³ Hippo pathway promotes cytoplasmic retention of YAP and inhibits nuclear localization of YAP. The downstream YAP activity is the effector in the transcriptional regulation of cellular behaviors of Hippo pathway.

TABLE 3. Summary of the role of YAP in stem cell fate decisions.

Function	Culture systems	Characterization	Ref.
Cell density and YAP (organ size control)	Heart General organs	High cell density induces YAP phosphorylation and cytoplasmic translocation, reduce proliferation and restrain organ size.	Zhao et al, 2011 [11] Heallen et al, 2011 [45]
Stress fibers and YAP	NIH 3T3 cells, MTD-1A cells. YAP regulates cell density-dependent control of proliferation.	Disruption of stress fibers using cytochalasin D reduces nuclear YAP. Inhibiting myosin II contractility using Blebbistatin or the inhibition of ROCK signaling using Y27632 to decrease the actomyosin contractility also reduce nuclear YAP	Wada et al, 2011 [41]
YAP/TAZ in hPSC self-renewal and reprogramming	hESC self-renewal YAP mediated by Rho signaling	Stiff surface based on glycosaminoglycan-binding hydrogels promoted the nuclear localization of co-activator YAP, which supports the self-renewal of hPSCs	Musah et al, 2012; [35]. Ohgushi et al, 2015 [40]
	Reprogramming	Activation of YAP can increase reprogramming efficiency and prevents differentiation of PSCs	Ramos et al, 2012 [36]
YAP/TAZ can control the switch between cardiac and endothelial lineages	The proliferation and differentiation of Sca-1^+ adult cardiac progenitor cells on substrates with different stiffness	YAP silencing (or soft substrates) correlated to the down-regulation of cardiomyocyte genes and the up-regulation of endothelial-specific genes as well as matrix metalloproteinase genes.	Mosqueira et al, 2014 [39]
YAP cytoplasmic retention to promote neural differentiation	Polydimethylsiloxane micropost arrays to control the stiffness of substrates. hPSC motor neural differentiation.	Inhibition of Hippo/YAP signaling on soft surface accelerates motor neuron differentiation (PAX6^+ cells at day 6: 60–70% vs. 20–30% for 5 kPa vs. 1200 kPa; 33% Tuj^+ and 18% HB9^+ at day 23 vs. 10% Tuj^+ and 3% HB9^+ cells).	Sun et al, 2014 [37]
	hPSC neural differentiation on the soft substrate.	Neuronal specification of hPSCs was enhanced in the absence of neurogenic factors on soft substrate (0.7 kPa vs. 3–10 kPa) that inhibits nuclear YAP localization.	Musah et al, 2014 [38]
YAP for mechanical memory behavior	Poly(ethylene glycol) hydrogels with different elastic modulus (stiffness) (2–10 kPa), MSC osteogenic differentiation and adipocyte differentiation.	Mechanical dosing of MSC on stiff surface bias the differentiation into osteogenic lineage. YAP/TAZ acts as intracellular rheostat, storing the information from past physical environments.	Yang et al, 2015 [42]
Bio-directional regulation of YAP and Wnt	Wnt activation (Wnt3a, GSK3 inhibitor) leads to the nuclear YAP localization. β-catenin dependent. Tested HEK293, MSCs, and ESCs.	In the absence of Wnt signaling, YAP/TAZ is sequestered to the destruction complex in the cytoplasm, which consists of a central protein named Axin, GSK3, β-catenin, as well as YAP/TAZ. This destruction complex is used to degrade β-catenin.	Azzolin et al, 2014 [12]
	Wnt activation (by Wnt5a) leads to nuclear YAP localization. β-catenin independent. Tested osteogenic and adipocyte differentiation. HEK293A cells, 3T3-L1 cells.	Activation of YAP/TAZ inhibits Wnt/β-catenin signaling. YAP/TAZ depletion abolishes Wnt4-induced enhancement of osteogenic differentiation. YAP/TAZ are mediators of alternative Wnt signaling.	Park et al, 2015 [13]

Substrate stiffness regulates Hippo/YAP signaling in stem cell differentiation

Using human mesenchymal stem cells (MSCs), stiff ECM (40 kPa) was found to result in cell spreading and active nuclear expression of YAP/transcriptional coactivator with PDZ-binding motif (TAZ) activity, which lead to osteoblast differentiation.³⁴ In contrast, soft substrates (0.7 kPa) inhibit YAP/TAZ and promote adipocyte differentiation. Similarly, stiff surface based on glycosaminoglycan-binding hydrogels was found to promote nuclear localization of YAP, which leads to the self-renewal of hPSCs,³⁵ while compliant surface inhibits nuclear YAP and cannot support the self-renewal of hPSCs. Moreover, activation of YAP can increase reprogramming efficiency and prevent differentiation of PSCs.³⁶

The inhibition of nuclear YAP through the biophysical properties of substrates was found to promote neuronal differentiation from hPSCs. Using polydimethylsiloxane micropost arrays, inhibition of Hippo/YAP signaling (i.e., cytoplasmic YAP expression) on soft surface was observed to accelerate motor neuron differentiation in the presence of soluble neurogenic factors (e.g., higher PAX6⁺ cells for 5 kPa vs. 1200 kPa: 60–70% vs. 20–30%).³⁷ Similarly, neuronal specification (i.e., GABAergic interneurons) of hPSCs was enhanced in the absence of neurogenic factors on soft substrates (0.7 kPa vs. 3–10 kPa) that inhibited the nuclear localization of YAP.³⁸ While these studies were performed in 2-D cultures, inhibition of nuclear YAP for neural differentiation of hPSCs based on 3-D cultures remains to be revealed.

YAP/TAZ also plays a critical role in the proliferation and differentiation of Sca-1⁺ adult cardiovascular progenitor cells on substrates with different stiffness.³⁹ Increased stiffness leads to the shuttling of YAP from cytoplasm to nucleus. Nanostructure of the matrix created by thermo-responsive crosslinked polycaprolactone polymers induces cytoplasmic YAP and removing nanopattern on the surface leads to the shift in YAP/TAZ nuclear expression.³⁹ More importantly, YAP silencing correlates to the down-regulation of cardiomyocyte genes and the up-regulation of endothelial-specific genes.³⁹ Consistently, soft matrix leads to cytoplasmic expression of YAP and the ability of the cells to form branching endothelial morphology. Therefore, YAP/TAZ can control the switch between cardiac and endothelial lineages for potential vascularization of neural organoids.

Rho and actin cytoskeleton (stress fibers) are required for the nuclear localization of YAP/TAZ and the mechano-transduction role of YAP/TAZ is found to be related to the interactions between Hippo signaling and the Rho GTPase signaling.⁴⁰ Stress fibers reduce YAP phosphorylation and promote nuclear YAP accumulation,⁴¹ so disruption of stress fibers using cytochalasin D reduces nuclear YAP. Inhibiting myosin II contractility using Blebbistatin or the inhibition of Rho-associated protein kinase (ROCK) signaling using Y27632 to decrease the actomyosin contractility also reduces nuclear YAP.^34,39 Therefore, nuclear YAP/TAZ transduces the signals exerted by ECM stiffness and cell morphology to regulate the downstream transcription factors through Rho GTPase activity and the stress fibers.

YAP has been observed to play a central role in “mechanical memory behavior” of stem cells,⁴² which means that stem cells possess mechanical memory and the past physical environments influence the cell fate. Human MSCs exposed to stiff surface longer (7 days vs. 3 days) have higher tendency to differentiate into osteogenic lineage even when the cells are cultured on soft substrates, appearing to “remember” the stiff environment. In this process, YAP/TAZ acts as the intracellular mechanical rheostat that stores the past information. A threshold mechanical dose can lead to the constitutive expression of nuclear YAP even after the mechanical dose is removed. Given the role of YAP regulation in neural differentiation of hPSCs, it is likely that the mechanical memory behaviors would exist in neural differentiation.

The role of Hippo/YAP pathway in cell density and organ size control

Hippo/YAP pathway plays a critical role in controlling cell density and organ size. At low cell density (cells spread), accumulated nuclear YAP promotes cell proliferation. While at high cell density (cells round up),⁴¹ strong Hippo signaling inhibits nuclear YAP (through YAP phosphorylation and cytoplasmic translocation) and suppresses cell proliferation.⁴¹ Changes in cell morphology due to different cell density were found to affect YAP localization through stress fibers. In organ growth, when the organ reaches a certain size (i.e., high cell density), the activated Hippo pathway suppresses cell proliferation. The dysregulation of this process leads to abnormal cell growth and cancer. For example, in multiple human cancers, the elevated YAP protein expression and nuclear localization have been observed.¹¹

Cell-cell contact also partially triggers Hippo pathway. It was found that the tight-junction protein complex (consists of PATJ, PALS1/MPDZ and Lin7 proteins) inhibits YAP/TAZ by promoting YAP/TAZ localization to the tight junctions and the YAP/TAZ phosphorylation.¹¹ Disruption of cell-cell junctions in epithelium was found to result in nuclear localization of YAP/TAZ and abnormal growth.⁴³ Therefore, maintaining cell-cell contacts is important for normal Hippo function and organ size control.

RELATIONSHIP OF HIPPO/YAP PATHWAY AND WNT/β-CATENIN PATHWAY

YAP localization regulates Wnt signaling

Hippo pathway promotes cytoplasmic localization of YAP/TAZ and inhibits Wnt signaling through two distinct mechanisms.⁴⁴ In cell nuclei, YAP and β-catenin can form YAP-β-catenin complex and the activation of YAP due to inhibition of Hippo pathway activates β-catenin-regulated genes (i.e., nuclear YAP promotes Wnt). In cell cytoplasm, cytoplasmic YAP can sequester β-catenin and thus promotes the retention of cytoplasmic β-catenin, which inhibits Wnt signaling (i.e., cytoplasmic YAP inhibits Wnt). α-catenin inhibits YAP activity by forming the trimeric complex of α-catenin, 14-3-3, and YAP, which sequesters YAP in cytoplasm and inhibits YAP nuclear localization.³³

The cross-interactions of Hippo pathway and Wnt pathway were found to restrain the proliferation of cardiomyocytes and control heat size.⁴⁵ Wnt/β-catenin signaling is required for the upregulation of cardiac cell growth. Since nuclear YAP forms complex with β-catenin and promotes β-catenin activity (e.g., upregulating SOX2 and Snail2 genes), Hippo pathway promotes the cytoplasmic retention of YAP and thus inhibits Wnt/β-catenin signaling. Consequently, blocking Hippo pathway was found to generate the oversized heart due to the promoted cell growth.⁴⁵

Wnt activation regulates YAP localization

It was recently found that Wnt activation leads to the nuclear YAP localization.¹² In the absence of Wnt signaling, YAP/TAZ is sequestered to the destruction complex in the cytoplasm, which consists of a central protein named Axin, GSK3, β-catenin, as well as YAP/TAZ. When associated with destruction complex, YAP/TAZ is transcriptionally inactive. This destruction complex is used to degrade β-catenin and also serves as cellular “sink” for YAP/TAZ. The destruction complex interacts with Wnt-activated Fz-LRP6 receptors through Axin/LRP6 interactions. With the activation of Wnt by Wnt3a, progressive dissociation of YAP/TAZ from Axin is observed along with the increased association of Axin with LRP6. In the meanwhile, β-catenin is released from the complex and translocated to the nuclei.

The biological relevance of Wnt activation to regulate YAP localization was demonstrated in the growth of crypt organoids.¹² YAP/TAZ was shown as the downstream effector of Wnt activation. Combined deletion of YAP/TAZ blocks the growth of crypt, indicating that YAP and TAZ are required for Wnt-induced response. The biological relevance of Wnt activation to regulate YAP localization is also demonstrated in maintaining undifferentiated mouse ESCs. Cytoplasmic YAP/TAZ inhibits β-catenin which leads to the loss of ESC self-renewal ability and activates differentiation. Loss of YAP/TAZ has the similar effect compared to GSK inhibition (i.e., add CHIR99021) and compensates for the absence of GSK inhibitor. To date, the biological relevance of Wnt activation to regulate YAP localization in neural organoid formation from hPSCs has not been well studied.

An “alternative Wnt-YAP/TAZ signaling axis” was proposed recently, which is independent of canonical Wnt/β-catenin signaling.¹³ Alternative Wnt ligands Wnt5a/b and other Wnt inhibitors including DKK1, bone morphogenetic protein 4, and insulin like growth factor binding protein 4 were found to be involved in the “alternative Wnt-YAP/TAZ signaling axis.” Treatment on the cells with Wnt5a/b (ligands for non-canonical Wnt signaling) results in the nuclear accumulation of YAP but has no effect on β-catenin. YAP-TAZ was suggested to induce the secretion of Wnt/β-catenin inhibitors (e.g., sFRP1) to inhibit Wnt/β-catenin signaling. In this mechanism, Wnt5a was shown as both an upstream activator and a downstream targeted gene of YAP/TAZ, forming a positive feedback loop between Wnt pathway and YAP/TAZ expression.

Biophysical perturbation of YAP to affect Wnt pathway

Given the bi-directional regulation between Wnt and YAP,⁴⁶ it is reasonable to hypothesize that biophysical perturbation of YAP protein localization should affect Wnt pathway, and thus neural tissue patterning of hPSCs. Biophysical perturbation of YAP is usually achieved using different small molecules.⁴⁷ For example, cytochalasin D and latrunculin A are actomyosin cytoskeletal molecules that disrupt F-actin and then inhibit YAP nuclear localization. Blebbistatin is a non-muscle myosin II inhibitor that inhibits YAP nuclear localization. Lysophosphatidic acid, sphingosine-1-phosphate and thrombin activate Rho and actin and thus stimulate YAP nuclear localization. Treating the cells with these small molecules should affect Wnt signaling like Wnt modulators. Biophysical perturbation of YAP can also be achieved by modulating the biophysical microenvironment of the cells,³⁷ such as the stiffness of 3-D substrates, to impact neural organoid formation.

CONCLUSIONS AND PERSPECTIVES

Wnt signaling affects neural tissue patterning of hPSCs by switching rostral and caudal brain tissue identity. Wnt-on condition induces nuclear YAP localization and the caudalization of neural cells. On the other hand, biophysical perturbation of YAP may affect Wnt pathway and thus neural tissue patterning of hPSCs. However, the biological relevance of Wnt-YAP interactions in neural organoid formation from hPSCs remains to be revealed.

There are several approaches to modulate Wnt pathway and Hippo/YAP pathway in order to generate different types of neural organoids from hPSCs. The most straightforward approach is to use soluble Wnt activators or inhibitors, or YAP modulators, which can be presented to the stem cell aggregates at appropriate time windows and locations to generate the 3-D neural organoids from different brain regions (such as forebrain, hippocampus, midbrain, hindbrain etc.). In particular, using the disease-relevance hPSC lines to generate neural organoids should be helpful to understand neurological disease pathology during brain tissue development.⁴⁸ Since the spatial presentation of the pathway modulators is important, the molecules can be tethered with ECMs or loaded with mciro/nanoparticles to create the in vivo-like gradient of the pathway modulators.

Potential problems and challenges still remain to fully re-create brain-like organoids in vitro. For example, the impact of Wnt signaling is context-dependent. The signaling network of Wnt-YAP with other signaling pathways (such as SHH, RA, fibroblast growth factors) may be required. Moreover, just like any de novo generated organoids, current neural organoids still lack the vascularization structure and fully mimicking specific brain region needs to consider the contributions of many non-neural cell types such as brain microvascular endothelial cells, the supportive brain stromal cells, etc. By better understanding of intracellular signaling pathways and their interactions with extracellular microenvironment, it should be possible in future to increase the complexity and function of hPSC-derived neural organoids.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

No potential conflicts of interest were disclosed.

Funding

We acknowledge grant support from the National Science Foundation (grant No. 1342192) and Florida State University start up fund.