Peripheral glia diversity

Abstract

Recent years have seen an evolving appreciation for the role of glial cells in the nervous system. As we move away from the typical neurocentric view of neuroscience, the complexity and variability of central nervous system glia is emerging, far beyond the three main subtypes: astrocytes, oligodendrocytes, and microglia. Yet the diversity of the glia found in the peripheral nervous system remains rarely discussed. In this review, we discuss the developmental origin, morphology, and function of the different populations of glia found in the peripheral nervous system, including: myelinating Schwann cells, Remak Schwann cells, repair Schwann cells, satellite glia, boundary cap‐derived glia, perineurial glia, terminal Schwann cells, glia found in the skin, olfactory ensheathing cells, and enteric glia. The morphological and functional heterogeneity of glia found in the periphery reflects the diverse roles the nervous system performs throughout the body. Further, it highlights a complexity that should be appreciated and considered when it comes to a complete understanding of the peripheral nervous system in health and disease.

Glia found in the peripheral nervous system are anatomically, morphologically, and functionally diverse. They include: myelinating Schwann cells, Remak Schwann cells, repair Schwann cells, satellite glia, boundary cap‐derived glia, perineurial glia, terminal Schwann cells, glia found in the skin, olfactory ensheathing cells, and enteric glia. The morphological and functional heterogeneity of glia found in the periphery reflects the diverse roles the nervous system performs throughout the body.

1. INTRODUCTION

The term glia, a word derived from the ancient Greek work for glue, was coined by Rudolf Virchow in 1856 who described the “connective tissue” of the brain and spinal cord that embedded the parts of the nervous system (Virchow, 1856). Less than a century later, it was discovered that the “connective tissue” Virchow had described was actually multiple populations of nonneuronal cells, including astrocytes (Von Lenhossék, 1893), Schwann cells (Ranvier, 1871), microglia (del Rio Hortega, 1920), and oligodendrocytes (del Río Hortega, 1921). It is now appreciated that glia play much more than a mere supportive role in both the central and peripheral nervous systems. The interdependent relationship between glia and neurons, and perhaps underappreciated, glia and glia, is necessary for proper nervous system development and function (see reviews by Allen & Lyons, 2018; Freeman, 2006; Nave & Trapp, 2008). Functional roles for glia found in the central nervous system (CNS) and peripheral nervous system (PNS) include provision of trophic and metabolic support (Dai et al., 2003; Tachikawa et al., 2004; Wilkins et al., 2003), participation in signaling (Christopherson et al., 2005; Liao et al., 1995; Mauch et al., 2001; Peng et al., 2003), and response to injury (see reviews by Buffo et al., 2010; Fitch & Silver, 2008; Jessen & Mirsky, 2016, and original research by, e.g., Szalay et al., 2016). Moreover, in higher organisms, glial cells are essential for neuronal and therefore the organism's, survival (Booth et al., 2000; Riethmacher et al., 1997; Woldeyesus et al., 1999). An appreciation for the complexity of CNS glia is beginning to emerge; however, the PNS has typically been considered the CNS' simpler counterpart. In this review, we discuss the wide range of glia subtypes found in the mammalian PNS, both from a morphological and functional stand point.

2. MYELINATING SCHWANN CELLS

As neural crest cells migrate and proliferate alongside growing axons in the early embryonic nerve, they differentiate into Schwann cell precursors (SCPs) (Jessen et al., 1994). Despite what their name might suggest, SCPs remain incredibly multipotent cells, in a manner similar to radial cells of the CNS (reviewed by Pinto & Gotz, 2007), retaining the ability to generate numerous cell types including peripheral glia (Dong et al., 1995; Jessen et al., 1994; Riethmacher et al., 1997; Woodhoo et al., 2009), parasympathetic and enteric neurons (Dyachuk et al., 2014; Espinosa‐Medina et al., 2014; Uesaka et al., 2015), cells of the sympathoadrenal system such as chromaffin cells and sympathetic neurons (Furlan et al., 2017; Kameneva et al., 2021), and possibly endoneurial fibroblasts (Joseph et al., 2004). In late embryonic and perinatal nerves, continued axonal signals, including neuregulin 1 type III (NRG1‐III) (Dong et al., 1995) and the notch ligand jagged 1 (Woodhoo et al., 2009), determine SCP differentiation into elongated immature Schwann cells. In the early postnatal phase, immature Schwann cells orchestrate a phenomenon called radial sorting, in which they select out large diameter axons with which to form 1:1 relationships (each Schwann cell associates with one large diameter axon) (Webster et al., 1973, and reviewed by Feltri et al., 2016). As the immature Schwann cell and the axon begin to form a symbiotic relationship (reviewed by Taveggia, 2016, and Wilson et al., 2020), the immature Schwann cell differentiates into a pro‐myelinating Schwann cell, surrounded by its own basal lamina (Webster et al., 1973). Ultimately, continued signals from axons and the extracellular matrix guide the formation of mature myelinating Schwann cells, the most well‐known subtype of peripheral glia. The most critical fate‐determining factor appears to be the dose of axonal NRG1‐III, with axon diameter correlating to NRG1‐III expression and high doses of NRG1‐III promoting Schwann cell myelination (Garratt et al., 2000; Michailov et al., 2004; Taveggia et al., 2005). Extracellular matrix signals are also crucial for the formation of myelinating Schwann cells. Signals from collagens (IV and XV) and laminins (211 and 411) through receptors such as the G‐protein‐coupled receptor GPR126 (ADGRG6) (Monk et al., 2011; Petersen et al., 2015), dystroglycan 1 (Saito et al., 2003), and the integrins (α6β1 and α7β1) (Feltri et al., 2002; Pellegatta et al., 2013) cause a rise in intracellular cyclic adenosine monophosphate (cAMP) (Monk et al., 2009; Mogha et al., 2013) and activate components of various signaling pathways (including small Rho nucleotide guanosine triphosphates, Rho GTPases, integrin‐linked kinase, ILK, and focal adhesion kinase, FAK) which promote Schwann cell proliferation and cytoskeletal reorganization necessary for radial sorting and myelination (Benninger et al., 2007; Grove et al., 2007; Nodari et al., 2007; Pereira et al., 2009). Furthermore, growing evidence for the role of mechanical forces and extracellular rigidity in the proper development of myelinating Schwann cells is also emerging (Poitelon et al., 2016; Urbanski et al., 2016, and see review by Belin et al., 2017). Myelinating Schwann cells possess a remarkable morphology, exemplified by the fact that they can expand their surface area approximately 2000 fold in order to enwrap axons in the main trunk of peripheral nerves with the multilayered myelin sheath which can be up to 20 mm² in surface area (Figure 1) (see review by Kidd et al., 2013). Prior to the evolutionary emergence of myelinating glia, action potential speed was directly proportional to axon diameter. Concentric wrapping of axons with layers of myelin sheath membrane allows the generation of specialized subdomains: nodes of Ranvier, paranodes, juxtaparanodes, and internodes, which together allow the propagation of action potentials via saltatory conduction (see review by Carroll, 2017), likely an adaptation to the relatively long peripheral nerves found in mammals. In recent years, a revised appreciation for Schwann cells has emerged, and more than providing myelin sheath to axons, it is becoming evident that these Schwann cells also have an important role to play in providing trophic support essential for neuronal health (Beirowski et al., 2014; Funfschilling et al., 2012; Nave, 2010; Nunes et al., 2021; Viader et al., 2011; Viader et al., 2013) and following neuronal injury (Babetto et al., 2020). The role of Schwann cells in neuronal health and disease is further exemplified by the fact that mutations in Schwann cell proteins can be sufficient to cause axonal neuropathy. For example, mutations in connexin32 (GJB1) can cause X‐Linked Charcot‐Marie‐Tooth disease type 1 with axonal pathology (Hattori et al., 2003; Vavlitou et al., 2010), and while mutations in myelin protein zero (MPZ) most commonly cause demyelinating Charcot‐Marie‐Tooth disease type 1, they can also cause the predominantly axonal Charcot‐Marie‐Tooth disease type 2 (Chapon et al., 1999; De Jonghe et al., 1999; Hattori et al., 2003; Shy et al., 2004).

FIGURE 1.

Glia found in the peripheral nervous system are anatomically, morphologically, and functionally diverse. Depiction of the diverse types of glial cells found within the peripheral nervous system. (a) Myelinating Schwann cells wrap large diameter axons with concentric layers of myelin sheath, generating myelinating internodes with gaps in the myelin sheath (nodes of Ranvier) which allow for saltatory conduction. Each myelinating Schwann cell associates with one axon. (b) In contrast, Remak Schwann cells associate with numerous smaller axons. (c) Following peripheral nerve injury, myelinating and Remak Schwann cells are capable of transdifferentiation into repair Schwann cells, which, among other functions, elongate to form bands of Büngner in order to guide axonal regrowth. (d) Satellite glia are found in sensory and autonomic ganglia, are flat, and wrap around the neuronal cell body. (e) Boundary cap cells are a transient population found close to the PNS/CNS boundary. (f) Perineurial glia cells compartmentalize axons and their associated Schwann cells into nerve fascicles. (g) Terminal Schwann cells associate with nerve terminals at the neuromuscular junction. (h) Numerous Schwann cells are found in the skin, including cutaneous Schwann cells which extend into the epidermal layer. (i) Olfactory ensheathing cells ensheath axons bundled together. (j) Enteric glia functionally and morphologically vary depending on their location within the gut wall. Peripheral glia cells which are largely considered to be derived from the neural crest are colored blue. Perineurial glia cells, which are at least in part of CNS origin, are shaded green. Neuronal cells are depicted in yellow

3. REMAK SCHWANN CELLS

In the trunk of the peripheral nerve, the alternative fate of immature Schwann cells is that of the nonmyelinating Remak Schwann cell, so named for Robert Remak, who first described them in the 1800s (Remak, 1838, and reviewed by Griffin & Thompson, 2008). Despite being lesser studied than their myelinating counterpart, Remak Schwann cells, in fact, occupy a far greater percentage of the Schwann cells found within the normal human sural nerve (Ochoa & Mair, 1969). Following completion of radial sorting (around postnatal day 15 in rodents), immature Schwann cells will extend processes to encompass the remaining axons, those which are typically smaller than 1 μm in diameter (Webster et al., 1973, and see review by Monk et al., 2015). These cells will then differentiate into mature Remak Schwann cells, engulfing numerous axons clustered in bundles (Figure 1). Thus, it should be noted that NRG1‐III is also required for proper formation of Remak bundles, and mouse models in which NRG1‐III is deleted show abnormally numerous axons within bundles and perturbed bundle structure by way of improper separation of neighboring axons by Remak Schwann cell cytoplasm (Fricker et al., 2009; Taveggia et al., 2005). Remak Schwann cells share many protein markers with immature Schwann cells including neural cell adhesion molecule 1 (NCAM1), L1 cell adhesion molecule (L1CAM), glial fibrillary acidic protein (GFAP), nerve growth factor receptor (NGFR, also known as p75 neurotrophin receptor), and growth associated protein 43 (GAP43) (Curtis et al., 1992; Jessen et al., 1990, and reviewed by Mirsky et al., 2008). As it stands, there are no tools available to exclusively explore the role of Remak Schwann cells in vivo and so less is known about the signals governing Remak Schwann cell fate, compared to myelinating Schwann cells (see review by Harty & Monk, 2017). However, there is evidence for the involvement of fibroblast growth factor (Furusho et al., 2009), γ‐aminobutyric acid (GABA) (Faroni et al., 2014; Procacci et al., 2013), insulin (Hackett et al., 2019), and lysophosphatidic acid signaling (Anliker et al., 2013), as well as signaling via the low‐density lipoprotein receptor‐related protein‐1 (LRP1) (Orita et al., 2013), since disruption to each of these components perturbed the organization of unsorted axons. The phospholipase neuropathy target esterase (NTE, also called patatin‐like phospholipase domain containing 6, PNPLA6) was also shown to be required for proper axonal ensheathment in Remak bundles, but not for myelination (McFerrin et al., 2017). Remak Schwann cells display qualities reminiscent of the wrapping glia of drosophila melanogaster, raising the possibility that these two types of glia share a common evolutionary heritage. In drosophila, aberrant differentiation of wrapping glia, or indeed complete absence of wrapping glia, impacts axonal diameter, in turn resulting in disturbance of conductance velocity (Kottmeier et al., 2020). It is feasible that Remak Schwann cells in mammals play a role in determining the speed of axonal signals in unmyelinated fibers in a similar manner. As aforementioned, Schwann cells also provide peripheral axons with metabolic support. This seems to be particularly important for axons within Remak bundles, since disruption of Schwann cell metabolism results in abnormal Remak bundle structure (Funfschilling et al., 2012; Pooya et al., 2014) and degeneration of Remak axons, which is more pronounced than that of the larger axons supported by myelinated Schwann cells (Beirowski et al., 2014; Viader et al., 2011; Viader et al., 2013).

4. REPAIR SCHWANN CELLS

One of the unique aspects of the PNS is its ability to regenerate and remyelinate. This is due in a large part to the plasticity of Schwann cells and their ability to switch between differentiation states (reviewed by Jessen & Mirsky, 2016). After injury, disruption of axonal contact triggers Remak and myelinating Schwann cells to re‐enter the cell cycle and transdifferentiate into a phenotypically unique repair cell. This transition is now understood to be transcriptionally controlled by JUN (Arthur‐Farraj et al., 2012), STAT3 (Benito et al., 2017), and H3K27 trimethylation (Arthur‐Farraj et al., 2017; Ma et al., 2016). Upregulation of these proteins occurs quite quickly after axonal disruption (Lee et al., 2009; Sheu et al., 2000) and increases expression of genes implicated in regulation and trophic support of neurons like Bdnf, Gdnf, Shh, and Gap34, while decreasing myelin and axonal‐glial interface protein expression (Arthur‐Farraj et al., 2012; Parkinson et al., 2008, and reviewed by Jessen & Arthur‐Farraj, 2019). The initial role of a repair Schwann cell after injury is to assist in the degeneration process. Disruption of axonal contact and NRG1‐III signaling is replaced by neuregulin 1 type I (NRG1‐I) produced by denervated Schwann cells (Stassart et al., 2013). Similar to NRG1‐III, para‐and autocrine derived NRG1‐I signals through the ERBB2 receptor on the Schwann cell surface and activates the ERK1/2 kinase pathway. Unlike in development where a sharp rise in ERK1/2 activity leads to increase in myelin protein expression, NRG1‐I activation of ERK1/2 induces an inflammatory response that is important for nerve repair (Napoli et al., 2012). Cytokines produced by Schwann cells, like monocyte chemoattractant protein‐1 (MCP‐1/CCL2), recruit macrophages to digest myelin and axonal debris (Fischer et al., 2008a; Fischer et al., 2008b, reviewed by Martini et al., 2008; Napoli et al., 2012). Schwann cells are also capable of digesting their own myelin in a process called “myelinophagy” (Gomez‐Sanchez et al., 2015; Jang et al., 2016, and reviewed by Park et al., 2019), acting cooperatively with macrophages and other phagocytic cells (reviewed by Hirata & Kawabuchi, 2002; Martini et al., 2008). Once denervated, repair Schwann cells elongate two to three times their previous cell length and form long branching process, lining up in columns or “regeneration tracks” known as Büngner bands (Figure 1) (Gomez‐Sanchez et al., 2017), a process that is also regulated by JUN (Arthur‐Farraj et al., 2012). Büngner bands serve as a growth tube expressing critical extracellular matrix proteins on the basal lamina needed for regenerating axons (Fugleholm et al., 1994; Longo et al., 1984). Growing neurite contact with the Schwann cell triggers a decrease in JUN expression causing shortening of the cell as it redifferentiates into a myelinating cell (Gomez‐Sanchez et al., 2017; Parkinson et al., 2008). As in development, remyelination is regulated by signals from the axon; however, neuronal NRG1 does not appear as critical in remyelination as it is in the development of myelin, as although recovery after injury is delayed in mice lacking neuronal NRG1, they do eventually reach levels of remyelination on par with controls (Fricker et al., 2013). Despite the plasticity of Schwann cells, remyelination does not generate the myelin sheaths created during developmental myelination. It has been well documented that the thickness of myelin sheaths of remyelinated axons is reduced compared to sheath thickness in naïve nerves, and functional recovery is often incomplete (Schröder, 1972, and reviewed by Höke, 2006). Despite this flaw, the PNS remains unique compared to the CNS in its ability to repair and regain function after injury.

5. DIVERSITY OF SCHWANN CELLS IN THE PERIPHERAL NERVE TRUNK

While morphologically Schwann cells in the trunk of peripheral nerves are neatly divided into either myelinating or Remak, as detailed above, it is plausible that functionally, Schwann cells may be more diverse. Typically, Remak bundles contain sensory C fibers and autonomic axons, whereas Schwann cell myelination occurs on A and B fibers, encompassing primarily motor axons but also some large sensory axons (see review by Griffin & Thompson, 2008). The existence of numerous neuronal subtypes poses the question: do there exist further subtypes of Schwann cell within the myelinating and Remak divisions? One controversial hypothesis that argues for the existence of motor and sensory Schwann cell diversity comes from the fact that after peripheral nerve injury, preferential motor reinnervation is reported, which could be the result of repair Schwann cells retaining a molecular identity more compatible with motor axons (Brushart, 1993; Brushart et al., 2013; Hoke et al., 2006; Madison et al., 2009, and reviewed by Bolívar et al., 2020). At least at postnatal day 1, single‐cell transcriptomics reported four different types of Schwann cell in the rat sciatic nerve (although this possibly reflects various developmental stages) (Zhang et al., 2020) and Schwann cells harvested from the motor and the sensory branches of the rat femoral nerve have unique gene expression profiles (Jesuraj et al., 2012). Moreover, a recent study revealed that Schwann cells purified from motor and sensory nerve roots display differences on a proteomic level (Shen et al., 2020). The human natural killer‐1 (HNK‐1) epitope is found on numerous cell adhesion molecules including some of those found in the PNS such as myelin associated glycoprotein (MAG), myelin protein zero (MPZ), NCAM1, and L1CAM. In the mouse femoral nerve, motor axon‐associated myelinating Schwann cells were reported to be HNK‐1 positive, while sensory axon‐associated myelinating Schwann cells were not (Martini et al., 1994; Saito et al., 2005). Although Schwann cells express proteins that enable them to respond to and support neuronal activity (reviewed by Samara et al., 2013), it remains entirely speculative whether Schwann cells within the main trunk of the peripheral nerve also exhibit direct functional excitatory activity. Indeed, recent evidence has shown that Schwann cells found in the skin do play a role in the transmission of mechanotransduction (Abdo et al., 2019), which is discussed in greater detail below in our section discussing Schwann cells of the skin. If Schwann cells within nerves also contribute to functional performance or perception, the type of neuron they associate with (such as motor, nociceptive, mechanoreceptive, chemoreceptive, and so forth) could thus necessitate functionally different Schwann cell types.

6. SATELLITE GLIA

Satellite glia are found in sensory ganglia (including dorsal root ganglia, trigeminal ganglia, and nodose ganglia) and the sympathetic and parasympathetic ganglia of the autonomic nervous system (reviewed by Hanani & Spray, 2020). Morphologically, they are distinct and unlike most other peripheral glia, they are flat and wrap around the neuronal cell body, with the complete neuronal‐satellite glia unit encompassed by a single basal lamina (Figure 1) (see reviews by Hanani & Spray, 2020; Pannese, 2010). In most cases, neurons are wrapped with several satellite glial cells (the exact number being relative to the size of the neuronal cell body and dependent on the species) (Ledda et al., 2004; Pannese, 1960), forming a complete envelope with a narrow 20 nm gap between the neuronal soma and the satellite glia (see reviews by Huang et al., 2013, and Hanani & Spray, 2020). Occasionally, one satellite glial cell encompasses multiple neuronal cell bodies to form a cluster (Huang et al., 2013; Hanani & Spray, 2020). Mature satellite glia appear postnatally (George et al., 2018) and are derived from neural crest cells; however, some evidence demonstrates that a subset of satellite glial cells are derived from boundary cap cells (see section below) (Maro et al., 2004). Satellite glia have been proposed to be a population of Schwann cells arrested from complete differentiation due to contact with the DRG neuronal soma (George et al., 2018). That being said, RNA sequencing studies suggest satellite glia share many molecular similarities with astrocytes, including expression of Fabp7, Pparα, and Aldoc transcripts (Avraham et al., 2020; Jager et al., 2020) and, similarly to their CNS counterpart, can be transcriptionally subdivided depending on different stages of maturation or activation (van Weperen et al., 2021). Indeed, in a similar manner to astrocyte activation, peripheral injury can increase GFAP expression in satellite glial cells (Liu et al., 2012; Woodham et al., 1989; Zhang et al., 2009). Satellite glia interact closely with the neuronal cells they surround and with their neighboring satellite glia. The most well‐known function of satellite glia in sensory ganglia is their contribution to pain and their role in the development of chronic pain (Huang et al., 2013; Hanani and Spray, 2020). In injury or inflammatory states, activation of satellite glia causes release of pro‐inflammatory cytokines which can disrupt normal neuronal firing causing neuropathic pain (Afroz et al., 2019; Dubový et al., 2010; Souza et al., 2013). Furthermore, satellite glia are thought to play a role in the depolarization of neurons in sensory ganglia through gap junctions which are upregulated after injury (Kim et al., 2016; Komiya et al., 2018). Blocking these gap junctions prevents activation of the dorsal root ganglia neurons and could provide a target to treat chronic neuropathic pain. Finally, satellite glia display plasticity when cultured in vitro (Belzer et al., 2010) and can give rise to cells expressing markers of neuronal cells (Arora et al., 2007; Li et al., 2007), Schwann cells (Fex Svenningsen et al., 2004; Li et al., 2007), oligodendrocytes (Fex Svenningsen et al., 2004), and astrocytes (Fex Svenningsen et al., 2004), and in vivo can be triggered to form oligodendrocyte‐like cells, through elevated expression of Sox10 (Weider et al., 2015). Thus, it is possible that satellite glia represent a potential pool of progenitor cells. Notably, most of the studies exploring satellite glia utilize those found in sensory ganglia, so there may exist differences in the satellite glia residing in the autonomic ganglia.

7. BOUNDARY CAP‐DERIVED GLIA

Boundary cap cells are a transient population of multipotent cells that originate from the neural crest and cluster in the spinal nerve roots, where dorsal (sensory) and ventral (motor) roots enter and exit the spinal cord (see review by Radomska & Topilko, 2017). Whether the boundary cap cell population in the ventral (motor) root also originates from the neural crest remains controversial (Radomska & Topilko, 2017). Boundary cap cells appear around embryonic day 10.5 in mice, potentially playing a role in the guidance of sensory and motor axons in the dorsal root entry zones and motor exit points, respectively, but disappear soon after birth (Golding & Cohen, 1997; Radomska & Topilko, 2017). Boundary cap cells give rise to the majority of the Schwann cells found in the spinal nerve roots, as well as a subpopulation of endoneurial fibroblasts and some satellite glia and sensory neurons within the dorsal root ganglia (Maro et al., 2004, and discussed in review by Jessen & Mirsky, 2005). Remarkably, a subpopulation of boundary cap derivatives migrates along the spinal nerves to the skin, where they form mostly nonmyelinating Schwann cells and terminal glia associated with lanceolate sensory endings and free nerve endings (Gresset et al., 2015). A population of boundary cap‐derived cells in the dermis remains multipotent; when transplanted into dorsal root ganglia, they can give rise to sensory neurons and when transplanted into sciatic nerve, differentiate into peripheral fibroblasts and Schwann cells (Gresset et al., 2015), thus representing something of a stem cell‐like population. The CNS and PNS come into close contact at motor exit point transition zones, and boundary cap cells have also been shown to play a gatekeeping role, preventing the motor neuron cell body from escaping the CNS (Bron et al., 2007; Garrett et al., 2016; Vermeren et al., 2003; see review by Radomska & Topilko, 2017). Boundary cap cells express specific markers, including Egr2 (Krox20, a transcription factor widely considered a master regulator of myelination) and Prss56 (previously named L20) (Coulpier et al., 2009). Notably Egr2 is expressed in boundary cap cells (despite no production of myelin) prior to the onset of Egr2 expression in myelinating Schwann cells (Topilko et al., 1994). Although boundary cap cells have not been identified in zebrafish, glia with some boundary cap cell qualities, motor exit point (MEP) glia, have been described. Intriguingly, these centrally derived cells seem to be a hybrid CNS and PNS glial cell, marked by Nkx2.2a, responsible for myelinating axons in the spinal cord exit zones (Fontenas and Kucenas, 2021). An attractive hypothesis is that a subpopulation of boundary cap cells might represent the mammalian equivalent of MEP glia (discussed further in review by Fontenas & Kucenas, 2018).

8. PERINEURIAL GLIA

Perineurial glia represent one of the most recently identified populations of peripheral glia. Perineurial glia, which were historically considered to be fibroblasts due to their shape, are thin, with their own basal lamina, and connected via tight junctions, contributing to the multilayered perineurium that encases axons and Schwann cells into compartmentalized nerve fascicles (Figure 1) (Shanthaveerappa & Bourne, 1962; see review by Kucenas, 2015). Unlike most other peripheral glia, perineural glia do not develop from the neural crest (Joseph et al., 2004). Evidence from zebrafish (Kucenas et al., 2008) and mice (Clark et al., 2014) suggests that at least a subset of these glia is in part CNS‐derived and, in zebrafish, rely on NOTCH signaling for appropriate migration out of the CNS (Binari et al., 2013). Yippee like 3 (ypel3) was recently identified to be required for the timely exit of perineurial glial precursors from the CNS in zebrafish (Blanco‐Sanchez et al., 2020). Physically, perineurial cells contribute to the blood‐nerve barrier, a semipermeable barrier in place to control the exchange of molecules between the blood and nervous tissue (Parmantier et al., 1999; Sharghi‐Namini et al., 2006). Evidence, from zebrafish and mice, has suggested that beyond their blood‐nerve barrier role, perineurial cells are also involved in nerve development (Binari et al., 2013; Clark et al., 2014; Kucenas et al., 2008; Parmantier et al., 1999) and response to injury (Bajestan et al., 2006; Lewis & Kucenas, 2014; Sharghi‐Namini et al., 2006). Signaling between myelinating Schwann cells and perineurial cells, via desert hedgehog (DHH) and its receptor patched (PTCH), is essential for the complete formation of the perineurium (Parmantier et al., 1999; Sharghi‐Namini et al., 2006) and in particular for perineurial cell expression of the key gap junction protein connexin 43 (also known as gap junction α‐1 protein, GJA1) (Parmantier et al., 1999). Interestingly, human patients with mutations in the DHH gene exhibit abnormal peripheral nerve morphology presenting as the formation of multiple mini fascicles (Sato et al., 2017; Umehara et al., 2000). Conditional ablation of perineurial cells in mice causes disorganized nerve fascicles and reduction in myelination along motor nerves (Clark et al., 2014), while zebrafish models show that perineurial glia respond rapidly after injury, phagocytosing debris, recruiting macrophages to the site, and forming a bridge across the injury gap (Lewis and Kucenas, 2014). In a similar manner to boundary cap cells, it has been shown that perineurial glia also play a role in preventing the escape of motor neuron bodies from the spinal cord (Clark et al., 2014; Kucenas et al., 2008). Perturbing proper formation of either Schwann cells or perineurial cells demonstrates that the Schwann cell–perineurial cell interaction is necessary for the development of both types of glia (Binari et al., 2013; Blanco‐Sanchez et al., 2020; Clark et al., 2014; Kucenas et al., 2008; Morris et al., 2017) and the proper response to injury (Lewis and Kucenas, 2014), highlighting the importance of interglia communication.

9. TERMINAL SCHWANN CELLS

The glial cells that surround neuromuscular junctions are named terminal or perisynaptic Schwann cells (Figure 1) and have been shown to have a unique role both in maintenance of the synapse and in reinnervation after injury. Terminal Schwann cells are nonmyelinating Schwann cells derived from neural crest cells that migrate along with the growing axon during development as SCPs. In mouse neuromuscular junctions, terminal Schwann cells align directly opposite the acetylcholine receptors on the motor endplate, while in human junctions, the terminal Schwann cells, which are notably smaller than mouse terminal Schwann cells, seldom completely cover all of the acetylcholine receptors (Alhindi et al., 2021). Castro et al. recently demonstrated that specific labeling of terminal Schwann cells can be achieved with the markers S100 calcium‐binding protein B (S100B) and neuron‐glia antigen 2 (NG2/CSPG4) (Castro et al., 2020). By crossing two transgenic mice, one of which labels S100 positive cells in GFP and the other which labels NG2 positive cells in red, they generated a new transgenic mouse that labels terminal Schwann cells in yellow. Further studies using this mouse will allow for the investigation of the unique genetic profile of terminal Schwann cells compared to other Schwann cell types. Very initial formation of nerve‐muscle contact does not require terminal Schwann cells as multiple mouse models lacking Schwann cells still demonstrate early neuromuscular junctions and acetylcholine receptor (AChR) clustering on the muscle (Lin et al., 2000; Morris et al., 1999; Riethmacher et al., 1997). However, these models clearly demonstrate the importance for terminal Schwann cells in the maintenance of junctions as most of these early junctions were rather disorganized and the nerve terminals quickly denervated from the muscle body after initial contact. As junctions mature and grow, they gain more terminal Schwann cells through migration of cells to the periphery and division of already present Schwann cells (Love and Thompson, 1998). In healthy junctions, terminal Schwann cells play a significant role in regulating synaptic activities. Early in development, they are responsible for phagocytosing the retracting terminal axons during synapse elimination and remodeling (Lee et al., 2016; Smith et al., 2013). Once a neuromuscular junction has formed, terminal Schwann cells continue to play an active role in responding to and regulating synaptic function. Receptors for acetylcholine and other neurotransmitters are expressed on the membranes of terminal Schwann cells, and neuromuscular synaptic activity has been shown to increase intracellular Ca²⁺ through activation of G protein‐coupled receptors (GPCRs) within the Schwann cells (Heredia et al., 2018; Jahromi et al., 1992) ultimately modulating neurotransmitter release (Castonguay et al., 2001; Robitaille, 1998, and reviewed by Auld & Robitaille, 2003). Terminal Schwann cells also play an active role in reinnervation after injury. Following denervation and loss of axonal contact, terminal Schwann cells phagocytose disconnected axon terminals in preparation for junction remodeling. After axonal degeneration, terminal Schwann cells spread out and increase their processes which guides growing axons and facilitates reinnervation of the muscle (Son & Thompson, 1995). Local ablation of terminal Schwann cells in a mouse model prevented the remodeling of neuromuscular junctions and limited reinnervation of the muscle after injury (Hastings et al., 2020). In disease, antibodies targeted to terminal Schwann cells have been implicated in a number of immune‐mediated peripheral neuropathies (Halstead et al., 2004; Halstead et al., 2005) while abnormal terminal Schwann cell function has been observed in animal models of certain neurological disorders, such as amyotrophic lateral sclerosis (Carrasco et al., 2016; Harrison & Rafuse, 2020) and spinal muscular atrophy (Murray et al., 2013; Voigt et al., 2010) and is thought to play a role in the degeneration and denervation that is characteristic of these disorders (reviewed by Santosa et al., 2018).

10. SCHWANN CELLS OF THE SKIN

The subepidermal nerve plexus makes up part of the largest sensory organ in the body, the skin. Large bundles of axons are arranged in horizontal layers in the superficial dermis extending thin nerve fibers into the epidermis, the majority of which remain unmyelinated. The association of terminal Schwann cells with myelinated and unmyelinated nerve endings has been observed for a number of years and early work using electron microscopy identified Schwann cell basal lamina surrounding nerve endings within the dermis (Cauna & Ross, 1960; Cauna, 1973), with more recent work in both mouse and human skin showing the extension of Schwann cell cytoplasm into the epidermis (Abdo et al., 2019; Rinwa et al., 2021) (Figure 1). Many of the Schwann cells within the dermis were found to be associated with low threshold mechanoreceptors found in the glaborous skin (Meissner, Pacinian, & Ruffini corpuscles) which transmit information about touch, pressure, vibration, and cutaneous force to the CNS. Schwann cells within the Meissner and Pacinian corpuscles form sheet‐like laminar formations around the inner axon, while axon processes extend between Schwann cells within Ruffini corpuscles (Cauna & Mannan, 1958; Cauna & Ross, 1960; Hachisuka et al., 1984; Halata et al., 1985, and reviewed by Cobo et al., 2021). Schwann cells are also found in hairy skin, ensheathing mechanosensors known as lanceolate endings arranged in a palisade around the hair follicle, and Merkel's disks which send thin nonmyelinated nerve fibers towards the hair follicle or into the epidermis (Halata et al., 2003; Li & Ginty, 2014; Takahashi‐Iwanaga, 2000; Yamamoto, 1966). The density of these Schwann cells within the dermis varies upon location of the human body, with the highest density on the trunk and about half those numbers in the distal extremities (Reinisch & Tschachler, 2012). Reduction of the number of Schwann cells in the dermis has been seen in pathological states like diabetic neuropathy, where loss of Schwann cells could be both a direct consequence of the disease or secondary to axonal degeneration (Reinisch et al., 2008). Terminal Schwann cells within the dermis can also be the targets of infectious agents (reviewed by Bray et al., 2020) such as herpes simplex virus (HSV) (Worrell & Cockerell, 1997) and mycobacterium leprae (Imaeda & Imaeda, 1986; Rambukkana et al., 1998, and reviewed by Rambukkana, 2000), and mutations affecting these cells can trigger the formation of neurofibromas in patients with neurofibromatosis (Ortonne et al., 2018; Zhu et al., 2002). Recent work has shown that Schwann cells also extend up into the epidermis and that this may contribute to a functional role for Schwann cells in nociception (Abdo et al., 2019; Rinwa et al., 2021) (Figure 1). Schwann cells in the skin are neural crest‐derived glia, sharing a number of markers with nonmyelinating Schwann cells of the peripheral nerve including S100, vimentin, NGFR, NCAM, L1CAM, and CD146 (Reinisch & Tschachler, 2012). Using an optogenetic approach where channelrhodopsin‐2 was exclusively expressed in peripheral glia but not axons, Abdo et al. discovered that optogenetic stimulation could elicit a pain response and increase in nociceptor firing rates in the paws of mice (Abdo et al., 2019). Interestingly, inhibition of nociceptive Schwann cells did not reduce the withdrawal response to hot or cold stimulus but did increase the threshold of activation of the response to mechano‐stimuli. Equally, electrophysiological studies of cultured cutaneous Schwann cells showed rapid response to changes in mechanical force, suggesting that nociceptive Schwann cells have the greatest contribution to the sensation of mechanical pain (Abdo et al., 2019). The transmembrane protein usherin (USH2A) is found in Schwann cells within the dermis and mutations in this protein in humans can cause Usher's syndrome which results in deafness and touch impairment (Schwaller et al., 2021). Genetic ablation of USH2A in mice reduces mechanoreceptor and vibration perception (Schwaller et al., 2021), providing further evidence for functional roles of Schwann cells in the skin. The previous understanding of nociception was that it was thought to be an axonally driven process. However, this discovery of nociceptive Schwann cells that form a glia‐neural network and are capable of responding in conjunction with nerve endings to mechanical stimuli from the environment changes this understanding and provides a potential new cellular target for pain management.

11. OLFACTORY ENSHEATHING CELLS

Very early microscopy work identified spindle‐shaped cells associated with and ensheathing olfactory nerve fibers (De Lorenzo, 1957; Gasser & Palade, 1956) that appeared similar to Schwann cells. Later morphological studies were able to distinguish these cells from Schwann cells and found that they shared traits with glial cells of the CNS, and they were given their unique name of olfactory ensheathing cells (OECs) (reviewed by Barnett, 2004). Unlike Schwann cells, OECs do not ensheath individual axons but rather bundles of axons packed tightly together (Figure 1) (Barber & Lindsay, 1982). More recent analysis however determined that OECs share transcriptional homology with Schwann cells, including similar expression of NGFR (Ulrich et al., 2014). OECs are also derived from neural crest cells that migrate to the periphery closely associated with growing axons during development (Barraud et al., 2010; Chuah & Au, 1991) and share the unique plasticity of Schwann cells. The olfactory nervous system in general is unique in its ability for constant renewal and regeneration. Throughout the human lifespan olfactory neurogenesis remains continually active and in times of infection or injury, this process occurs even more rapidly (Bayer, 1983; Cancalon, 1987). The plasticity of this system is in part attributed to the OECs that are crucial for survival and growth of olfactory neurons (Doucette, 1990; Kafitz & Greer, 1998; Kafitz & Greer, 1999). In fact, successful isolation of OECs from the olfactory bulb (Barnett et al., 1993; Barnett & Roskams, 2002; Barnett & Roskams, 2008) has led to the discussion of using OECs in therapeutics for peripheral and CNS injuries (reviewed by Barnett & Riddell, 2007; Franklin & Barnett, 2000; Fairless & Barnett, 2005; Nazareth et al., 2021). Although they do not produce myelin normally, researchers discovered that OECs were able to myelinate axons after they were transplanted into a demyelinated spinal cord. In multiple rodent models, OECs were successfully transplanted into an injured spinal cord (García‐Alías et al., 2004; Imaizumi et al., 2000; Kato et al., 2000; Pearse et al., 2007; Smith et al., 2002) or peripheral nerve injury site (Guérout et al., 2011; Radtke et al., 2009), improving regeneration and remyelination of axons and restoring conduction across the injured site. This benefit is believed to be mostly due to the secreted neurotrophic factors from the OECs that aid in neuronal survival and the ability of the OECs to produce myelin, in a similar manner to Schwann cells (Franklin et al., 1996; Smith et al., 2001). Furthermore, unlike Schwann cells which cause reactive gliosis when cultured with astrocytes or transplanted into the CNS, OECs have a unique ability to seamlessly integrate with astrocytes making them a more desirable treatment than transplanted Schwann cells after CNS injury (Lakatos et al., 2000; Lakatos et al., 2003; Santos‐Silva et al., 2007).

12. ENTERIC GLIA

Within the gastrointestinal (GI) system resides a division of the PNS that controls GI motility, exocrine and endocrine secretions, circulation, and regulation of the gut microbiome, called the enteric nervous system (Cooke, 1994; Goyal & Hirano, 1996; see reviews by Kunze & Furness, 1999; Sharon et al., 2016). In the middle of the 1700s the idea that the gut has a “brain” of its own started to develop as intestines removed from the body were observed to have preserved peristaltic function (Temkin, 1936). In the late 1800s German physicians discovered neurons within the gut wall following the observation of evoked polarization after stimulation and characterized the first enteric glial cells in sections of gut wall (Dogiel, 1899; Lüderitz, 1890; Lüderitz, 1891). Within the gut wall neuronal cell bodies are gathered into ganglia extending their axons to form two major plexuses, the myenteric or Auerbach and submucosal or Meissner plexuses which reside between the gut mucosa and the smooth or longitudinal muscle (Figure 1). Surrounding these neurons are populations of transcriptionally unique (Howard et al., 2021) nonmyelinating glial cells derived from a common pool of neural crest derivatives (Figure 1) but phenotypically influenced by the microenvironment of their location within the gut wall (Le Douarin & Teillet, 1973; Yntema & Hammond, 1954). Enteric glia express proteins characteristic of other glia, including proteolipid protein 1 (PLP1), S100B, SOX10, and GFAP and share morphological similarities with astrocytes (Boesmans et al., 2015; Jessen & Mirsky, 1983; Rao et al., 2015). Hanani was the first to classify enteric glial populations into four subgroups based on morphological differences (Hanani, 1994). Others have separated populations based on their location within the gut wall: mucosal, intraganglionic, and intramuscular enteric glia, reflecting differences in function of each group (reviewed by Gulbransen & Sharkey, 2012). Enteric glia within the mucosal layer assist in barrier formation and repair by directing the differentiation and maturation of epithelial cells (Bach‐Ngohou et al., 2010; Savidge et al., 2007; Van Landeghem et al., 2011, and reviewed by Yu & Li, 2014). They also appear to play a role in inflammation as glial cells in the mucosa can act as antigen‐presenting cells, further propagating immune cell activation and break down of the intestinal epithelial barrier (Geboes et al., 1992). Deeper in the gut wall, glia are involved in trophic support of neurons and modulating synaptic activity. They play a key role regulating neurotransmitter levels as they are capable of clearing neurotransmitters from the synapse to terminate activity and have also been shown to respond to ATP released from enteric neurons suggesting that they play an active role in neurotransmission and activity (Braun et al., 2004; Gulbransen & Sharkey, 2009; Lavoie et al., 2011). Recent evidence has demonstrated a number of roles for enteric glia in gastrointestinal disorders. Glial modulation of neurotransmitters and activation of macrophages has been shown to play a role in activation and sensitization of visceral nociceptors, contributing to the pain that is associated with gastrointestinal disorders (reviewed by Grubišić et al., 2020; Morales‐Soto & Gulbransen, 2019). In mouse models, targeted disruption of enteric glial cells reduces intestinal motility (Aube et al., 2006; Nasser et al., 2006), while histopathological intestine samples from patients with slow transit constipation were found to have significantly reduced numbers of enteric glial cells (Bassotti et al., 2006a; Bassotti et al., 2006b). Loss of enteric glia and disruption of the enteric nervous system is also a hallmark characteristic of necrotizing enterocolitis in infants (Sigge et al., 1998; Wedel et al., 1998). Furthermore, genetic deletion or CD8 T cell‐mediated destruction of enteric glia in mice causes severe bowel inflammation and hemorrhage within the small bowel suggesting a role for enteric glia in inflammatory bowel disorders like Crohn's disease and ulcerative colitis (Bush et al., 1998; Cornet et al., 2001). That being said, human data have been less clear, as patients with Crohn's disease have decreased numbers of enteric glia while patients with ulcerative colitis or certain types of infectious colitis actually have more glial cells and gliosis throughout the affected bowel (see review by Neunlist et al., 2014).

13. CONCLUSION

Unlike the CNS, populations of glia in the PNS remain less explored and less defined. There may remain further subtypes of glia in the periphery yet to be defined, and indeed, a very recent study described spleen glia, which may play a role in neuroimmune crosstalk (Lucas et al., 2021). Lineage tracing and increasing sophistication of research tools may continue to uncover novel subtypes of peripheral glia or specific cell markers to further characterize different populations. Meanwhile, glia and neurons are not the only cells that are found in the PNS (Chen et al., 2021; Gerber et al., 2021). Other cell types in the PNS, including pericytes, endoneurial fibroblasts, and macrophages, have their roles in development and disease and may even be classified as unique subsets of glia. Unlike microglia in the CNS, macrophages are in limited numbers in the healthy peripheral nerve and have little known importance in nerve development but do play a major role in injury and disease. The population of fibroblasts found in the peripheral nerve, considered largely supportive, is thought to have differentiated from SCPs and may transpire to play a far more functional role in the PNS (Joseph et al., 2004; see also review by Jessen & Mirsky, 2005). Fibroblasts express extracellular matrix molecules such as collagen and trophic factors like neuregulin that can promote Schwann cell proliferation, assembly of the basal lamina, and ensheathment of neurites in culture (Dreesmann et al., 2009; Obremski et al., 1993). Moreover, there is evidence showing that nerve fibroblasts play a role in recovery following peripheral nerve injury (Parrinello et al., 2010) and signaling to macrophages in disease states (Groh et al., 2012). Targeted ablation of peripheral nerve fibroblasts would allow for further understanding of the interdependent role between fibroblasts and Schwann cells in peripheral nerve development, maintenance, and response to injury.

Within the PNS resides multiple unique populations of glial cells almost all derived from the neural crest. Despite the heterogeneity, these glial cells are all essential for the proper development and function of the PNS and in particular the function of their corresponding neurons. Moreover, glial cells contribute to the pathogenesis of neurodevelopmental and neurodegenerative disease within the PNS and thus represent important populations for potential therapeutic target.

CONFLICT OF INTEREST

The authors declare no conflict of interests.

AUTHOR CONTRIBUTIONS

CBR, MLF, and ERW contributed to the interpretation of literature, concepts discussed, and design of this manuscript; CBR and ERW wrote the article; CBR, MLF, and ERW revised the manuscript; ERW created the figure.

Reed, C.B. , Feltri M.L.&Wilson E.R. (2022) Peripheral glia diversity. Journal of Anatomy, 241(5), 1219–1234. 10.1111/joa.13484