The Hippo pathway: Horizons for innovative treatments of peripheral nerve diseases

Abstract

Initially identified in Drosophila, the Hippo signaling pathway regulates how cells respond to their environment by controlling proliferation, migration and differentiation. Many recent studies have focused on characterizing Hippo pathway function and regulation in mammalian cells. Here, we present a brief overview of the major components of the Hippo pathway, as well as their regulation and function. We comprehensively review the studies that have contributed to our understanding of the Hippo pathway in the function of the peripheral nervous system and in peripheral nerve diseases. Finally, we discuss innovative approaches that aim to modulate Hippo pathway components in diseases of the peripheral nervous system.

1 ∣. COMPONENTS, REGULATORS, AND FUNCTION OF THE HIPPO PATHWAY

1.1 ∣. Canonical Hippo pathway components

In mammalian cells, the core components of the Hippo pathway are Serine/Threonine kinases; mammalian sterile 20-like (MST1 and MST2) kinases, large tumor suppressor (LATS1 and LATS2) kinases, the adaptor proteins Salvador homolog 1 (SAV1) and Mps one binder kinase activator proteins (MOB1A and MOB1B) (Figure 1). These proteins were grouped into one signaling pathway - the Hippo pathway - named after observing that mutants of their Drosophila orthologues had a “hippopotamus” appearance, caused by the overgrowth of tissues due to hypercellularity.^1-5 The main function of the Hippo pathway is to negatively regulate the activity of the transcription coactivator Yes-associated protein 1 (YAP, encoded by YAP1) and its paralog, transcriptional coactivator with PDZ-binding motif (TAZ, encoded by WWTR1) (Figure 1). YAP and TAZ have many similarities in their structures, functions and regulations, yet they are not completely redundant (^6,7; and reviewed in⁸). In this review, we use the term YAP/TAZ when it has been shown that YAP and TAZ share the mentioned regulations or functions. However, we use YAP or TAZ when data is available only for either YAP or TAZ. The central axis of the Hippo pathway is a phosphorylation cascade from MST1/MST2-LATS1/LATS2 to YAP/TAZ. Mechanistically, MST1/MST2 form heterotetramers with SAV1 to promote MST1/MST2 activation by trans-autophosphorylation (^9,10). Phosphorylated MST1/MST2-SAV1 heterotetramers then phosphorylate MOB1A/MOB1B and the hydrophobic motif of LATS1/LATS2 kinases.¹¹ Phosphorylation on MOB1A/MOB1B promotes their interaction with LATS1/LATS2 and the autophosphorylation of LATS1/LATS2 on their activation loop. Both phosphorylation of the hydrophobic LATS1/LATS2 motif by MST1/MST2 and autophosphorylation of LATS kinases activation loop are critical for their activation (reviewed in¹²). Activated LATS1/LATS2 directly phosphorylate YAP and TAZ at multiple sites. Phosphorylated YAP/TAZ bind to 14-3-3 proteins and are then sequestered in the cytoplasm, resulting in their inhibition. Additional phosphorylation of YAP/TAZ by casein kinase 1 leads to YAP/TAZ ubiquitination and degradation by the proteasome.^13,14 In addition, YAP can also be degraded by autophagy.¹⁵ Because TAZ has a rapid turnover, with a half-life of less than 2 hours, it has been suggested that protein degradation, rather than cytoplasmic sequestration, may be the main route for TAZ inhibition.¹³ In contrast, YAP is a relatively stable protein, thus it may be inhibited primarily by cytoplasmic sequestration. Interestingly, it was also shown that YAP may have a cytoplasmic function, namely to promote cell migration by favoring activation of CDC42,.^16,17 To sum up, once activated, the Hippo pathway inhibits YAP/TAZ nuclear localization through their phosphorylation (reviewed in¹⁸).

FIGURE 1.

Core components of the Hippo pathway. The Hippo pathway control the cytoplasmic-nuclear shuttling of YAP/TAZ. A, When the Hippo pathway is active (ON), LATS1/LATS2 kinases phosphorylate YAP/TAZ, causing either their cytoplasmic retention or degradation. NF2 contributes to the Hippo pathway activation. B, When the Hippo pathway if inactive (OFF), YAP and TAZ are not phosphorylated and shuttle into the nucleus to regulate gene expression. Mechanical stimuli and Actin dynamics can regulate LATS1/LATS2 kinases

On the other hand, when the Hippo pathway is inactive/off, YAP/TAZ are not phosphorylated and are instead translocated into the nucleus, where they promote gene transcription¹⁴ (Figure 1). YAP and TAZ are transcriptional co-regulators, but do not contain DNA-binding domains and TEAD family transcription factors are the primary binding partners of YAP/TAZ.¹⁹ TEADs mediate the main transcriptional output of the Hippo pathway in mammalian cells.²⁰ As detailed below, YAP/TAZ/TEAD-mediated gene expression regulates many cellular processes, spanning form proliferation to migration and to cell differentiation. In the absence of nuclear YAP/TAZ, TEAD transcription factors function as repressors by binding to VGLL4 (transcription cofactor vestigial-like protein 4), repressing target gene expression.^21,22

1.2 ∣. Upstream signals regulating the canonical Hippo pathway

The central role of the Hippo pathway is to influence the localization and protein stability of the effectors YAP/TAZ. A multitude of extracellular, intracellular and non-biological stimuli can activate the Hippo pathway and regulate YAP/TAZ transcriptional activity. Here we briefly present the most understood mechanisms that regulate the Hippo pathway. For a comprehensive review detailing all signals and regulators of the Hippo pathway, see.²³

Cell polarity is essential for a wide range of cellular processes, such as synaptic communication between neurons or myelin formation (reviewed in^24-26). The Crumbs complex is a key regulator of cell polarity and contains CRB proteins (CRBs), which are necessary to organize apicobasal polarity and the associated cytoplasmic proteins PALS1 and PATJ.²⁷ CRBs act as scaffold proteins for multiple components of the Hippo pathway such as LATSs/YAP/TAZ, and play a role in YAP/TAZ cytoplasmic localization.^28-30 Consistent with the key role of cell polarity proteins in promoting activation of the Hippo pathway, components of tight junctions or adherens junction complexes regulate Hippo signaling.³¹ For example, loss of E-cadherins, α- or β-Catenin can result in increased YAP transcriptional activity.^32,33 CRB proteins also recruit to the plasma membrane other known Hippo regulators such as Angiomotins (AMOTs) or Neurofibromin 2 (NF2), also called Merlin or Schwannomin.^29,34-37 Angiomotins were first reported to promote the activation of the Hippo pathway³⁸ through the sequestration of YAP/TAZ (see Section 1.3). However, a further report has also suggested that Angiomotins may serve as direct YAP activators by preventing its phosphorylation.³⁹ NF2 is a tumor suppressor gene, whose loss of function causes Neurofibromatosis type 2, a disorder characterized by the development of tumors in the peripheral nervous system (PNS) (see Section 3.1).⁴⁰ NF2 is a positive regulator of the Hippo pathway through several mechanisms. NF2 binds and recruits LATS proteins to the cell membrane where they get activated by the MST1/MST2/SAV1 complex.⁴¹ In addition, NF2 also binds and inhibits the ubiquitination and degradation of LATS proteins.⁴²

Cells can sense stiffness of the extracellular matrix (ECM) through focal adhesion complexes that link the ECM to the actin cytoskeleton, and consequently change cell behavior such as migration and spreading (reviewed in⁴³). Physical forces such as stress or strain that physiologically impact cell density, stiffness of the extracellular environment and cell geometry are now known to regulate the localization and activation of YAP/TAZ through mechanisms that are both dependent or independent (see Section 1.3 below) of the Hippo pathway.^44-46 In response to ECM stiffness or cell spreading, YAP/TAZ transcriptional activity is regulated, making them essential effectors of mechanical stimuli into the cell.⁴⁴ Integrin signaling has been revealed to inhibit LATS1/LATS2 activity and thus facilitates YAP nuclear translocation and transcriptional activation through the β1-integrin-FAK-Src-PI3K-PDK1 pathway.⁴⁷ The Src-Rac1-PAK pathway also inhibits the phosphorylation of YAP by LATS1/LATS2 downstream of β1-integrin, through phosphorylation of NF2.⁴⁷

In addition to local and mechanical signals (eg, cell polarity, cell-cell contact and ECM stiffness), paracrine or endocrine signals can also regulate the Hippo pathway. For example, ligands of G-protein-coupled receptors such as sphingosine 1-phosphate and lysophosphatidic acid have been shown to inhibit LATS1/LATS2 through G-proteins.⁴⁸ Similarly, epidermal growth factor signaling inactivates the Hippo pathway through the epidermal growth factor receptor (EGFR) and both the PI3K-PDK1 and the Ras-Raf-MAPK signaling pathways.^49,50 Activation of the EGFR-RAS-MAPK pathway increases the phosphorylation of the Ajuba family proteins, promoting their interaction with LATS proteins, and thus preventing LATS-dependent YAP phosphorylation.^51,52

In addition to the MAPK signaling pathway, the Hippo pathway interacts with many other signaling pathways, such as the Wnt/β-catenin and Notch pathways. Most studies have focused on the crosstalk between the Hippo signaling and the Wnt/β-catenin signaling pathway due to their critical roles in tumorigenesis (see Section 3).⁵³ When Wnt signaling is activated, β-catenin induces the transcriptional up-regulation of YAP. Yet, (i) when the Hippo pathway is activated, YAP inhibits Wnt signaling by sequestering β-catenin in the cytoplasm; (ii) when the Hippo pathway is inactive, nuclear YAP up-regulates Wnt/β-catenin signaling.^54,55 In addition, the Notch receptor and ligand Jagged1 are direct target genes of YAP, and the Notch intracellular domain enhances the transcriptional activity of YAP/TAZ.^56,57

Finally, the Hippo pathway is also regulated by dephosphorylation events, such as by striatin-interacting phosphatase and kinase (STRIPAK) complexes. The phosphatase component of STRIPAK complexes is protein phosphatase 2A (PP2A). PP2A interacts with and dephosphorylates MST1/MST2 kinases and the functionally related kinase MAP4K4, which leads to the dephosphorylation of YAP/TAZ.^10,58-62 The interaction of STRIPAK complexes with MAP4K4, and possibly with MST1/MST2, is regulated in part by RhoA activity-dependent NF2 association with STRIPAK components.⁵⁸

1.3 ∣. Hippo pathway-independent regulation of YAP/TAZ

While the cytoplasmic retention of phosphorylated YAP/TAZ is a consistent outcome of Hippo pathway activation, in several studies the cytoplasmic-nuclear shuttling of YAP was shown to be independent from the Hippo pathway.^44,63,64 Thus, it is currently proposed that in addition to canonical Hippo signaling (i), kinases other than LATS1/LATS2 may also regulate YAP/TAZ transcriptional activity (reviewed in⁶⁵), and (ii) YAP/TAZ cytoplasmic-nuclear shuttling may also be regulated independently of their phosphorylation.^66,67

Among Hippo- independent YAP/TAZ regulation, the role of several kinases or binding proteins have been elucidated. This is the case for several receptor tyrosine kinases (GFRF, RET, MERTK, PYK2), phosphofructokinase (PFK1), AMP-activated protein kinase (AMPK), the pre-mRNA splicing factor 4 kinase (PRP4K) and Nemo-like kinase (NLK), which have been shown to directly phosphorylate a subset of LATS sites on YAP/TAZ and either phosphorylate YAP/TAZ or promote their exclusion from the nucleus.^68-71 For example, AMPK is a key cellular metabolic sensor that is activated by a high AMP to ATP ratio. PFK1 is an enzyme regulating the first step of glycolysis. Their regulation of YAP/TAZ indicates that YAP/TAZ can be regulated by metabolic pathways.^72,73 In addition to phosphorylation, YAP/TAZ are also regulated by a variety of molecules that can sequester YAP/TAZ. Cell-cycle kinase (CDK1), α-catenin, Claudin18 and protein tyrosine phosphatase non-receptor type 14 (PTPN14) directly bind to YAP to sequester it in the cytoplasm or at the plasma membrane.^23,37,74,75

Finally, several recent studies have shown that mechanical cues can induce regulation of YAP/TAZ also independently of their phosphorylation by the Hippo pathway.^44,45,76,77 Physical forces lead to the reorganization of the actin cytoskeleton, and to activation of Rho GTPase.^76,78 Knockdown of actin capping protein (CAPZB) causes an increase in F-actin, and an increase in nuclear YAP/TAZ coupled with an increase in expression of YAP/TAZ target genes.⁴⁴ Reverse regulations also exist, as YAP promotes the expression of the Rho guanine exchange factor ARHGEF17 that activates the RhoA-actomyosin axis. Thus, YAP and RhoA signals are likely to be mutually regulated by feedback loops.^79,80 Spectrin, a cytoskeletal protein associated to actin filaments, also plays an important role in the activation of YAP.^81-83 While the mechanical regulation of YAP/TAZ independently of the Hippo pathway was controversial for some time, recent studies have started to clarify its mechanisms. Actomyosin fibers connecting focal adhesions to the LINC (Linker of the Nucleoskeleton and Cytoskeleton) complex of the nuclear envelope, may regulate YAP import in the nucleus by opening nuclear pores.^66,67 Interestingly, in this model it was also suggested that NF2 forms a complex with YAP/TAZ to mediate their nuclear export.⁸⁴

2 ∣. THE HIPPO SIGNALING PATHWAY IN PERIPHERAL NERVE DEVELOPMENT

YAP, TAZ and TEADs regulates gene expression that leads cells to respond to their environment in an exquisitely cell and context-specific manner. The PNS is formed from the soma and axons of sensory and autonomic neurons, and from the axons of motor neurons which exit the central nervous system (CNS) at motor exit points. These axons are ensheathed by glial cells and surrounded by blood vessels, endoneurial fibroblasts, perineurial and epineurial cells. Many of these cell types derive from the neural crest. Here we will briefly review the developmental steps from the neural crest to PNS components in which the Hippo pathway is known to be involved (Figure 2).

FIGURE 2.

Non-comprehensive, schematic representation of peripheral nerve development (green background) and pathology (blue background). The PNS is formed from the soma and axons of sensory and autonomic neurons, and from the axons of motor neurons. These axons are ensheathed by glial cells and surrounded by blood vessels, endoneurial fibroblasts, perineurial and epineurial cells. Many of these cell types derive from the neural crest. Schwann cells are the major type of ensheathing glia. Yet, the PNS also includes region-specific ensheathing glia: PNS ganglia are associated with satellite cells and axons at the sensory terminals are surrounded by nociceptive glial cells. Magenta notes indicate the steps where an involvement of the Hippo pathway or YAP/TAZ have been described. Ax, axon; BL, basal lamina; CM, compact myelin

2.1 ∣. Neural crest cells

Neural crest cells originate after closure of the neural tube and expand and migrate extensively in the whole organism. Depending on the anterior/posterior region from which they delaminate, neural crest cells are classified into cranial, cardiac, vagal, trunk or sacral. The region of origin and path of migration greatly influence the fate of neural crest cells, which includes PNS components, melanocytes, chromaffin organs, smooth muscle cells, cardiac cells, facial bone and cartilage (reviewed in⁸⁵). YAP promotes human neural crest cell fate and migration.⁸⁶ In addition, the Hippo pathway regulates migration and differentiation of neural crest cells into smooth muscle cells, craniofacial and dental structures and melanocytes.^87-91 Most of the neural crest-derived PNS cells, including sensory and autonomic neurons, all Schwann cells, satellite cells and endoneurial fibroblasts, originate from trunk neural crest cells. The Hippo pathway is active in trunk neural crest cells.^92,93

To our knowledge, there are no published reports on the Hippo pathway in terminal Schwann cells at the neuromuscular junction. Thus, only the studied roles of Hippo pathway components in Schwann cell, sensory neuron precursors and fibroblasts are reviewed below.

2.2 ∣. Dorsal root ganglia neurons and satellite cells

After delamination from the neural tube, some neural crest cells migrate ventrally between the neural tube and the somites, where they coalesce to form dorsal root ganglia (DRG) neurons and glia.^94,95 DRG neurons send axons centrally to form spinal roots and synapse with sensory neurons in the spinal cord, and peripherally to sensory terminal. Satellite glia tightly surround the soma of DRG neurons, probably providing metabolic and electrical support, similar to astrocytes in the CNS (reviewed in⁹⁶). DRG neurons and satellite glia derive from a common neural crest precursor, in separate waves spanning embryonic day (E) E11.5 to E13.5 in the mouse, that give rise to different subsets of DRG neurons and to satellite cells.^97-99 Elegant work involving expression in vivo and deletion or hyperactivation of members of the Hippo pathway revealed that YAP is expressed in immature DRG neurons, glia precursors, and satellite cells, but not in mature DRG neurons. Inhibition of YAP, mediated by NF2 positive regulation of the Hippo pathway (see Section 1.2), prevents over-expansion of the precursor cell population and increases the number of glial cells while deleting NF2 causes neuron depletion.⁹³ This suggests that YAP may favor glia differentiation. Inhibition of the differentiation of various neuronal types by YAP transcriptional activity is also suggested by several studies showing that softer matrices and tissues, which inhibit YAP dephosphorylation, favor neuronal differentiation and axonal growth.¹⁰⁰

2.3 ∣. Schwann cells

Trunk neural crest precursors that migrate ventrally give rise to DRG neurons and Schwann cells.¹⁰¹ Schwann cells migrate to the periphery along neuronal projections, while progressively changing their spatial association with axons and differentiating into Schwann cells precursors, immature Schwann cells, promyelinating Schwann cells and finally into mature myelinating Schwann cells or Remak (non-myelin forming) Schwann cells (Figure 2). It is important to mention that some Schwann cells, especially those myelinating dorsal and ventral nerve roots, derive from a different population of neural crest derivatives, boundary cap cells, that transiently localize and preserve the boundary between the CNS and PNS at sensory entry points and motor exit points,^102,103 and reviewed in.¹⁰⁴ Boundary cap cells are in part molecularly distinct from Schwann cells, and share many characteristics with a similar population of cells, motor exit point cells, characterized in the zebrafish¹⁰⁵ and reviewed in.¹⁰⁶

Schwann cells precursors (E12.5-E15.5 in the mouse) surround large bundles of mixed caliber axons.¹⁰⁷ Immature Schwann cells (E16.5-early postnatal) start penetrating nerve bundles and recognize axons destined to be myelinated (usually larger than 1 μm), based on the amount of Neuregulin 1 type III on the surface of axons.^108,109 By progressive proliferation to match the number of axons and segregation, immature Schwann cells reach a 1:1 relationship with larger axons start to myelinate. Collectively, this process is called radial sorting of axons by Schwann cells, and spans from perinatal times to post-natal day (P) 10 in the mouse. Finally, when all axons larger than 1 μm have been isolated, small axons remain ensheathed in smaller bundles that differentiate into Remak bundles, around P15 in the mouse, surrounded by non-myelin forming Schwann cells (Figure 2). Radial sorting is highly dependent on signals from basal lamina components laminin 211, 411 and collagen XV,^110-112 which are transmitted through the laminin receptors integrins α6β1 and α7β1, dystroglycan and GPR126.^113-115 It has been shown that one way by which laminin signals promote radial sorting is by regulating the cytoskeleton through FAK, ILK, CDC42, Pinch, and RAC1, allowing the formation of lamellipodia-like processes from Schwann cells that interact with, recognize and sort axons.^116-120 Other mechanisms involve regulation of cAMP levels and PKA cross-talk with the Neuregulin pathway.^121,122 Detailed reviews of Schwann cell development can be found elsewhere.^123-126

Recent work from several laboratories has highlighted how inhibition of the Hippo pathway, with consequent activation of YAP/TAZ, is essential for multiple steps of normal Schwann cell development, including radial sorting, proliferation and subsequent myelination.^92,127,128 YAP/TAZ are expressed throughout the Schwann cell lineage, but TAZ expression is higher between P3 and P15, the period corresponding to active radial sorting and myelination.^127,128 Nuclear localization of YAP/TAZ, denoting that they are active, is detected in vivo in mice at all times examined, from E16.5 to adulthood (P60), suggesting that YAP/TAZ may be required throughout the life of a Schwann cell.^{28,92,127,128} Deletion of both alleles of YAP/TAZ specifically in Schwann cells using Schwann-cell specific Cres (Mpz-Cre)^92,128 or Dhh-Cre,¹²⁷ which recombine Schwann cell precursors at E13.5 and E12.5, respectively, profoundly impairs radial sorting of axons, preventing immature Schwann cells from reaching the promyelinating stage (1:1 relationship with axons) and therefore preventing myelination. Interestingly, homozygous deletion of only YAP causes virtually no radial sorting problems, while deletion of TAZ causes moderate radial sorting arrest, suggesting that TAZ is more important for radial sorting than YAP. However, YAP/TAZ are functionally redundant, as deletion of one of two YAP alleles in the context of TAZ deletion aggravates the radial sorting phenotype.^92,127,128

As for other cell types and as introduced in part 1 in this review, the inhibition of the Hippo pathway in Schwann cells to promote radial sorting of axons is probably caused by multiple mechanical and chemical stimuli. Peripheral nerves traverse every tissue of the body and thus must possess mechanical resistance and elasticity. These are provided by abundant extracellular matrix, including collagen fibers, that renders nerves among the stiffer (after bones) and most elastic tissues in the body.¹²⁹ Throughout all their developmental steps, Schwann cells are subjected to mechanical stimulation due to various forces, including axonal stretching and elongation and basal lamina polymerization and stiffening. It thus likely that some of these mechanical stimuli promote Schwann cell development through activation of YAP/TAZ. Like other cells, Schwann cells are mechanosensitive^128,130 and activate YAP/TAZ in response to increased matrix stiffness^128,130 and axonal stretching during nerve elongation,²⁸ but also in response to laminin 111 and 211. Laminins likely inhibit the Hippo pathway and activate YAP/TAZ through mechanoreceptors, such as integrin α6β1, which in turn regulate the actin cytoskeleton via RAC1 activation.^{119,128,131,132} Downstream of YAP/TAZ, radial sorting is promoted via the transcriptional activation of several genes required for this process, including genes encoding for regulators of proliferation (Gα proteins, ErbB2/3 receptors), SOX10, ZEB2, laminin γ1 and the laminin receptors integrin α6β1 and dystroglycan. In several of these cases, TAZ and TEAD1 bind directly to enhancers of these genes.^127,133 It is possible that other Laminin receptors such as GPR126/cAMP are also involved in regulating the Hippo pathway in Schwann cells.

YAP/TAZ also have a role in myelination, after radial sorting is complete, however some findings suggest that the regulation of myelination by the Hippo pathway and YAP/TAZ is complex and subject to multiple levels of cross-regulation. Studies in mice clearly indicate that YAP/TAZ promote myelination in concert with SOX10 by targeting enhancers of the master myelin-promoting transcription factor EGR2/KROX20 and of genes coding for myelin proteins and lipid synthesis regulators.^{28,92,127,128,133,134,200} However, the absence of NF2 causes transient hypomyelination, but whether this is due to the lack of Hippo-mediated repression of YAP/TAZ is unknown.¹³⁵ In addition, deletion of TEAD4 downstream of HDAC3 causes remarkable hypermyelination in mice, increases PI3K and ERK signaling and promotes upregulation of myelin-related genes,¹³⁶ suggesting that TEAD1 and TEAD4 may have opposite functions in myelination. This, and the fact that TAZ seems more important during development than YAP,¹²⁸ suggest that specific components of the Hippo pathway may have specific effects. Interestingly, YAP and TAZ act differently in the nucleus, as only TAZ forms liquid-liquid interphase biomolecular condensates that are associated with super enhancers, while YAP does not, indicating potentially different nuclear functions (reviewed in¹³⁷). It is also possible that YAP or TAZ interact with TEAD4 to inhibit precocious myelination during radial sorting, in a way similar to the inhibitory effect of laminin on Neuregulin 1 type III signaling, to ensure the timely and correct amount of myelin formation.¹²¹ More transcriptomic and epigenetic experiments on single gene mutants or after activation of single components will be required to dissect these possibilities. Furthermore, a better understanding of the intersection between specific components of the Hippo pathway and other promyelinating pathways such as those activated by Neuregulins is also of interest. As mentioned, NF2 is a particularly important tumor suppressor in Schwann cells, and it is a known activator of the Hippo pathway at different levels. NF2 also cross-talks with integrins and antagonizes RAC1.^131,138 However, how NF2 and RAC1, and possibly NF2 and CRB3, regulate each other is only partially understood. Finally, CRB3-FRMD6/willin, upstream regulators of the Hippo pathway, localize at the nodes of Ranvier and regulate the elongation of myelin segments (internodal length) during development through YAP.²⁸

2.4 ∣. Endoneurial fibroblasts

Endoneurial fibroblasts are neural crest derived, and play crucial roles during nerve regeneration, by communicating with Schwann cells via EphB signaling to coordinate segregation and collective migration of both cell types.¹³⁹ Interestingly, FRMD6/willin, an upstream activator of the Hippo pathway, is highly expressed in endoneurial fibroblasts, where it appears to inhibit YAP/TAZ activation to promote collective migration necessary for regeneration.¹⁴⁰ This is remarkably similar to the role of NF2 in Schwann cells after injury.¹³⁵ Thus, is possible that the specificity of activation of the Hippo pathway in Schwann cells and fibroblasts after injury may be achieved using cell-specific receptors.

3 ∣. THE HIPPO PATHWAY IN PNS DISEASES

As the Hippo pathway controls so many PNS cells and their function during nerve development, we would expect that components of this pathway contribute also to peripheral nerve pathology. So far, the most prominent roles for the Hippo pathway are reported in tumorigenesis, nerve injury and repair and in neuropathic pain, which will be reviewed below (Figure 2). In contrast, the role of the Hippo pathway in acquired or inherited neuropathies are largely unexplored, so we will only speculate briefly about this topic.

3.1 ∣. Tumorigenesis

Early indications of a link between a defective Hippo pathway and tumorigenesis were suggested by the overgrowth phenotypes of Drosophila Hippo pathway mutants, observations of elevated tumorigenesis in LATS kinases mutant mice, and an association between the NF2 tumor suppressor and Hippo signaling.¹⁴¹ Subsequent observations have shown that elevated nuclear YAP/TAZ levels were associated with a wide variety of tumors, including in the PNS.^134,142

Tumors of peripheral nerve, schwannomas and neurofibromas, are benign in the vast majority of clinically symptomatic cases (reviewed in ¹⁴³), with lesser percentages made up of other benign tumors and of malignant tumors such as malignant peripheral nerve sheath tumors (MPNSTs), lymphoma and metastases.¹⁴⁴ MPNSTs are frequent in the context of neurofibromatosis. For patients with neurofibromatosis type 2 (NF-2), the clinical presentation is characterized by the development of multiple schwannomas and meningiomas.^42,145,146 Because the NF2 gene was shown to function upstream of the Hippo pathway, there are numerous evidences that strongly suggest that YAP/TAZ are involved in schwannoma tumorigenesis and NF-2 pathophysiology.^135,147-150 A very recent study confirmed this hypothesis by demonstrating that Hippo signaling and YAP/TAZ are required for schwannoma formation.¹⁵¹ For patients with neurofibromatosis type 1 (NF-1), the incidence of malignancy is significantly greater. About 10% of patients with NF-1 will develop a MPNST during their lifetimes, and nearly 50% of patients with MPNST have NF-1¹⁵² Several studies have demonstrated evidence for alterations of the Hippo pathway in MPNSTs. Analysis from MPNST samples shows an increase in the copy number of Hippo pathway effectors (ie, TAZ, CTGF and BIRC5) and loss of gene loci of YAP/TAZ activity inhibitors (ie, LATS2 and AMOTL2).¹³⁴

Transcriptome sequencing of human MPNST samples also revealed elevated YAP-activated gene expression in MPNST when compared to normal nerves and NF1-associated benign tumors.^153,154 Furthermore, the YAP signature is present in MPNST samples regardless of their NF-1 genetic status, suggesting that activation of the YAP/TAZ is common to both genetic and sporadic MPNSTs. Importantly, ablation of both LATS kinases in Schwann cells in genetic mouse models abolished the regulation of YAP and TAZ, which remained active in the nucleus, causing the development of aggressive nerve-associated tumors, similar to human MPNSTs in their pathological and invasive characteristics.¹³⁴ Together, these studies validate the role of the Hippo pathway in MPNST tumorigenesis, likely in combination with other oncogenic pathways, such as EGFR-RAS, WNT, GPCR or PI3K signaling, and with changes in the tumor microenvironment, such as inflammation and increased ECM stiffness.

3.2 ∣. Traumatic injury

YAP/TAZ proteins play key roles in wound repair and regeneration. This is due to their roles in promoting cell differentiation and organ growth, and to their activation by loss of tissue integrity, as it happens in damaged tissues. Indeed, activation of YAP/TAZ proteins occurs both in response to changes in cell-cell contact, cell density, and cell stretching that can accompany wounding. Following traumatic injury, the peripheral nerve starts a series of events distal to the site of the injury (Wallerian degeneration reviewed in¹⁵⁵), which includes the degeneration of axons, infiltration of macrophages, and reprogramming of Schwann cells to transdifferentiate and assume a repair Schwann cell phenotype, which facilitates axonal regeneration through formation of tubular guidance structures (Bands of Büngner) and functional repair of the peripheral nerve.¹⁵⁶ Recent data indicated that both YAP and TAZ regulate peripheral nerve regeneration, but in different contexts. The Hippo pathway activator NF2 is crucial for the generation of repair Schwann cells, as nerves deficient for NF2 display a severe impairment of Bunger band formation, axon regrowth and remyelination. Consistently, NF2-null nerves present with activation of YAP following nerve injury in Schwann cells.¹³⁵ Despite the fact that YAP in Schwann cell itself is not necessary for axonal regrowth or remyelination following injury, further removal of YAP in of NF2-Schwann cells null crushed nerves leads to normal axonal regrowth and functional recovery after injury.¹³⁵ Thus, Schwann cell YAP may have a role in axonal regrowth following traumatic nerve injury. Ablation of TAZ alone does not affect axonal regrowth, rather, it delays remyelination after nerve injury. This phenotype is amplified in the injured nerve of animals lacking both YAP/TAZ, which present a failure in remyelination.^157,158 This failure to remyelinate is due to YAP/TAZ being necessary in repair Schwann cells for their redifferentiation into myelinating Schwann cells.¹⁵⁸ Thus, YAP/TAZ are both regulating developmental and post-traumatic myelin formation, yet while absence of TAZ cannot be compensated by YAP, YAP absence seems to be compensated by the expression of TAZ in Schwann cells.

3.3 ∣. Neuropathic pain

Neuropathic pain may be due to genetic mutations in ion channels or may be a consequence of lesions or inflammation of the nervous system, and accounts for about 10% of chronic pain cases.¹⁵⁹ The transmission of nociception in unmyelinated C fibers and small thinly myelinated Aδ fibers depends on mechanosensitive ion channels at sensory terminals such as Piezo ion channels, transient receptor potential (TRP) channels and sodium channels (eg, Nav1.7-1.9) (reviewed in¹⁶⁰). While a conventional role for YAP/TAZ directly at the synaptic terminal of neurons is unlikely due to the distance between the neuronal cell bodies and synaptic terminals, a functional role for the Hippo pathway was suggested by upregulation of FRMD6 in sensory neurons cell bodies in the dorsal root ganglion after injury,¹⁶¹ and further in the spinal dorsal horn of the CNS.¹⁶² Interestingly, peripheral glia associated with these afferent fiber's terminals may also act as sensors. Activation of skin-resident glial cells at the sensory terminals are both necessary and sufficient for mechanical pain sensation.¹⁶³ These newly identified nociceptive glial cells are sensitive to mechanical stimulation, a known regulator of YAP/TAZ.¹⁶³ In addition, Piezo channels can modulate activation of YAP/TAZ.¹⁶⁴ These studies suggest new mechanisms to target the Hippo pathway and specific glial cells involved in pain perception (see Section 4.2).

3.4 ∣. Speculation on inherited neuropathies

Peripheral neuropathies are common, affecting up to 8% of people over 55 years of age (reviewed in¹⁶⁵). Neuropathies can be acquired, such as diabetic neuropathies, or inherited, such as Charcot-Marie-Tooth. Despite different pathogenic mechanisms, neuropathies eventually converge on a common phenotype characterized by loss of nerve fibers and increased ECM. It is likely that these changes dysregulate the mechanosensitive Hippo pathway, potentially increasing Schwann cell proliferation and impeding recovery. Even though studies on the potential role of the Hippo pathway in acquired and inherited peripheral neuropathies are currently lacking, one could make the argument that dysregulation of the Hippo pathway is a potential contributor to inherited neuropathies due to altered expression of Peripheral Myelin Protein 22 (PMP22) in Charcot Marie Tooth 1A (CMT1A) and Hereditary Neuropathy with Liability to Pressure Palsy (HNPP).^166-170 As the disease name suggests, in HNPP innocuous mechanical pressure causes conduction block and demyelination (reviewed in¹⁷¹). While normal myelinated fibers can withstand remarkable compression without loss of integrity and quickly reverse to normal shape, PMP22 null nerves have less elasticity.¹²⁹ The possible multiple roles of PMP22 in myelinated fiber elasticity and resistance, at cellular junctions and in the cell-cycle make it an ideal candidate to be intimately linked to the Hippo pathway, which is also regulated by mechanical signals and junctional proteins and tied to proliferation. Interestingly, the YAP/TAZ co-transcription factor TEAD1 directly regulates PMP22 gene expression.¹³³ Thus, a putative dysregulation of the Hippo pathway may further contribute to dysregulated PMP22 expression caused by genetic copy number variation of PMP22.

4 ∣. INNOVATIVE THERAPIES BASED ON THE HIPPO PATHWAY

4.1 ∣. Drug based therapies

Targeting of the Hippo pathway with drug-based therapies is of great significance due to its roles in cancer biology and tissue regeneration. Interestingly, these fields usually have opposite goals, with cancer research pursuing drugs to reduce YAP/TAZ-TEAD transcriptional activity and regenerative medicine pursuing drugs to increase YAP/TAZ-TEAD transcriptional activity. To date, most studies have investigated small molecules or peptides capable of activating the Hippo pathway for cancer treatment and these drugs have been extensively reviewed.^172-174 Due to the considerable amount and variety of these small molecules and peptides, Pobbati et al. have classified Hippo-modifying drugs into three groups (see Figure 3). Group I drugs target the upstream regulators of YAP/TAZ, enhancing LATS1/LATS2-dependent phosphorylation of YAP/TAZ, thus reducing YAP/TAZ transcriptional activity (see Table 1). This group includes many compounds that are already U.S. Food and Drug Administration (FDA)-approved (eg, pazopanib, dasatinib, dobutamine, losartan) and therefore hold potential for drug repurposing. However, these compounds are not specific to the Hippo pathway and dysregulate broad signaling downstream of cell surface receptors and intracellular kinases. Group II drugs directly modify the YAP/TAZ-TEAD interaction, by targeting TEAD or YAP (see Table 1). Thus, these drugs are more directed towards the Hippo/YAP/TAZ pathway. Nevertheless, verteporfin, a commonly used disruptor of the YAP/TAZ-TEAD interaction, also does not solely target YAP and may target other proteins.^184,185 Finally, group III drugs (eg, celecoxib) target the genes regulated by YAP/TAZ-TEAD (see Figure 3). Using these drugs can be challenging as they target only one of the many proteins regulated by YAP/TAZ. Combinatory use of group I, II and III drugs could also be considered.

FIGURE 3.

Schematic representing the different approaches to pharmacologically target the Hippo pathway. Group I drugs enhance phosphorylation of YAP/TAZ. Group II drugs directly modify the YAP/TAZ-TEAD interaction. Group III drugs target the genes regulated by YAP/TAZ-TEAD

TABLE 1.

Examples of FDA-approved drugs targeting the Hippo pathway

Group	Drug Name	Mechanism of Action	Effect on Hippo pathway	Reference
I	dasatinib	Src family tyrosin kinase inhibitor	activates LATS1/2	(175,32,176,177)
I	dobutamine	β1-adrenergic receptor agonist	Increase YAP/TAZ phosphorylation	178
I	pazopanib	c-KIT, FGF PDGF, VEGF, receptors inhibitor	LATS1/2 activation	(179,180,181)
I	losartan	Angiotensin II receptor inhibitor	Activates LATS1/1	182
II	verteporfin	YAP ligand	inhibits YAP/TAZ-TEAD interaction	183
III	celecoxib	COX-2 inhibitor	-	148

The therapeutic potential for Hippo pathway modulation in the context of PNS diseases is of interest for neuropathies such as HNPP, CMT1A, and MPNSTs (see Section 3). Upstream Hippo pathway activation (Group I drugs) or direct inhibition of the YAP/TAZ-TEAD interaction (Group II drugs) may alleviate the oncogenic effects of high YAP transcriptional activity in MPNSTs. In addition, celecoxib, a cyclooxygenase-2 (COX-2) inhibitor that has been FDA approved since 1998, is a particularly exciting example of a widely available Group III Hippo-modifying drug with potential to treat Schwann cell-derived tumors. Treatment in vitro with celecoxib has been shown to reduce cell proliferation of NF2-null schwannoma tumor cells in a dose-dependent manner.¹⁴⁸ Moreover, when these cells were implanted into mouse sciatic nerves as an in vivo schwannoma tumor model, tumorigenesis was markedly reduced following 10 days of once daily oral celecoxib administration. Further studies will be needed to clarify the therapeutic potential and side effect profile of the use of Hippo-modifying drugs in PNS diseases.

4.2 ∣. Mechanical based therapies

Peripheral nerves are often damaged by compression, stretch, and avulsion. There are many proposed methods of speeding nerve regeneration, including physical methods, such as electrical stimulation,¹⁸⁶ magnetic field stimulation,¹⁸⁷ laser stimulation¹⁸⁸ and ultrasound therapy.¹⁸⁹ It is possible that the Hippo pathway and YAP/TAZ are stimulated by the application of physical stimuli. First, in vitro studies have shown that application of low-intensity ultrasound can regulate the Hippo pathway and YAP/TAZ.^190-192 Second, in vitro studies in Schwann cells have shown that ultrasound regulates signaling pathways upstream of the Hippo pathway.^193,194 Finally, extensive studies provide evidence that low-intensity ultrasound activates and promotes Schwann cell proliferation, a major outcome of the Hippo pathway (reviewed in¹⁹⁵).

Similarly, application of an electromagnetic field to cultured Schwann cells promotes their proliferation and migration, at least in part through the Hippo pathway. Colciago et al., have shown that the expression of genes known to be upstream or downstream mediators of the Hippo pathway was altered in Schwann cells, leading to dysregulation of YAP.¹⁹⁶ While further studies are needed to support this hypothesis, the physical control of the Hippo pathway and YAP/TAZ opens potential new avenues for the treatment of peripheral neuropathies. However, because of the implication of the Hippo pathway in tumorigenesis, ectopic activation of the Hippo pathway and YAP/TAZ in physiological and pathological contexts should be carefully controlled.^197,198

5 ∣. CONCLUSIONS AND OPEN QUESTIONS

The Hippo pathway and YAP/TAZ rapidly came to the forefront of PNS biology and pathology only in the last few years. In light of the multitude of stimuli and signals regulating the Hippo pathway, and of the major role that it plays in many developmental and pathological systems, one of the challenges now faced by the scientific community is to understand how the signals that converge on the Hippo pathway, YAP and TAZ, regulate each other in feed-back loops (reviewed in¹⁹⁹), how integration with other major cellular signaling pathways occurs, the extent of cell-specificity and context-specificity and the different and overlapping roles of single components of the Hippo pathway. This specific understanding will allow more targeted and specific modulation of YAP and TAZ activity in peripheral nerve disease.