Origin, identity, and function of terminal Schwann cells

Abstract

The highly specialized non-myelinating glial cells present at somatic peripheral nerve endings, known collectively as terminal Schwann cells (TSCs), play critical roles in the development, function and repair of their motor and sensory axon terminals and innervating tissue. Over the past decades, research efforts across various vertebrate species have revealed that while TSCs are a diverse group of cells, they share a number of features among them. In this review, we summarize the state-of-knowledge about each TSC type and explore the opportunities that TSCs provide to treat conditions that afflict peripheral axon terminals.

Terminal Schwann cells are an integral part of end organ structures

Axon terminals of the peripheral nervous system (PNS) are present in every organ of the vertebrate body. These axonal terminals are points of contact at which bidirectional information is relayed between peripheral tissues and the central nervous system (CNS). Within the somatic division of the PNS, axon terminals transmit motor information to skeletal muscles at the neuromuscular junction (NMJ) or receive various types of sensory information from hair follicles, Meissner corpuscles (MCs), Pacinian corpuscles (PCs), muscle spindles, Golgi tendon organs (GTOs), Merkel cell-neurite complexes, and free nerve endings (FNEs). Each of these peripheral targets of innervation, to which we will refer collectively as “end organ structures”, is highly specialized yet shares a common feature of containing a unique population of peripheral glia. These glial cells were first reported in 1878[1] and later dubbed “teloglia”[2,3]. Nomenclature for these cells further evolved when they began to be referred to as “terminal Schwann cells”[3–5], based on observations that they appear to originate from the same progenitors as Schwann cells (SCs) along the nerve[5–9]. Although other common names for these cells persist, we will use terminal Schwann cells (TSCs) as a unifying designation of any non-myelinating Schwann cell that is associated with an axon terminal, regardless of end organ structure, in this review.

The last few decades have provided insights into how TSCs influence the development and physiology of adult end organ structures. We have learned that their association with axon terminals differs between each unique end organ structure. At the NMJ, TSCs cover the entire axon terminal[10]; at MCs and PCs, they elaborate lamellae around the axon terminal[11–13]; they wrap around the low-threshold mechanoreceptor (LTMR) axon terminals associated with hair follicles[14]; and, finally, they extend fine processes around nociceptive FNEs in skeletal muscle, connective tissue, and the epidermis [15–18]. Despite their different anatomical locations and cellular features, evidence continues to indicate that TSCs share many functional features with each other and with CNS glia, particularly astrocytes and microglia[19]. These include active participation in synaptic development[20], facilitating and maintaining neural transmission[21,22], and repairing damages following injury [23,24]. They perform these functions by sensing neurotransmission[21,25], releasing cytokines[26], influencing synaptic competition[27], and removing supernumerary axons[28]. Additionally, there is growing evidence that TSCs may undergo a myriad of changes that cause them to lose their homeostatic functions in diseases and aging[29–31]. Despite advances in understanding the molecular makeup of TSCs, including single cell and spatial transcriptomics (Table 1), the mechanisms that specify their identity and function remain largely unknown.

Table 1.

Genes highly expressed in TSCs as compared to myelinating Schwann cells.

Gene	End organ structure	Method*	Reference
Adgrg6	NMJ	RNA seq, IHC	[35,127]
Agrn	NMJ	RNA seq, WB	[35,128]
Ajap1	NMJ	RNA seq, FISH	[35,38]
Apod	NMJ	RNA seq, FISH	[35,38,129]
Aqp1	FNE	IHC	[16]
Asic2	MC, PC	IHC	[79,80]
Bche	FNE, MC, NMJ, PC	RNA-seq, IHC, enzymatic activity	[35,75,130,131]
Calb1	MC, PC	IHC	[132]
Calb2	MC, PC	IHC	[132]
Chrm5	NMJ	IHC	[64]
Cspg4 (aka NG2)	NMJ	RNA seq, FISH, IHC	[26,32,35,38]
Egfr	MC, PC	IHC	[132]
Gfap	PC	IHC	[133]
Grin2a (aka NMDA2A)	Hair Follicle	IHC	_[59]_
Grm1	PC	IHC	[83]
Kcnj10 (aka Kir4.1)	NMJ	RNA seq, IHC	[35,52]
L1cam	FNE	IHC	[134]
Mcam (aka CD146)	FNE	IHC	[134]
Ncam1	FNE, MC, NMJ, PC	RNA seq, IHC	[35,54,134,135]
Nes	Hair Follicle	IHC	[59,136]
Ngfr	FNE, MC,PC	IHC	[53,132,134]
Ntrk1	MC, PC	IHC	[132]
P2ry1	NMJ	RNA seq, IHC	[35,64]
Pvalb	MC, PC	IHC	[132]
Sema3a	NMJ	RNA seq, IHC	[35,109]
Slc17a1	PC	IHC	[83]
Trpv4	MC	IHC	[78]
Ush2a	Hair Follicle, MC	IHC, functional analysis	[84]
Vim	FNE, MC, PC	IHC	[54,132,134,137]

^*RNA seq = RNA sequencing, IHC = immunohistochemistry, WB = western blot, FISH = fluorescence in situ hybridization, NMJ = neuromuscular junction, MC = Meissner corpuscle, PC = Pacinian corpuscle, FNE = free nerve ending.

Recently discovered genetic markers that are specific for TSCs (Table 2) make it possible to interrogate TSCs to the exclusion of myelinating SCs. These advances have opened new possibilities in TSC research. This review aims to summarize progress made about the origin and function of TSCs with a particular focus on the better understood end organ structures, including NMJs, MCs, PCs, FNEs, and hair follicles. We will synthesize data generated across vertebrate species, including frogs, birds, rodents, cats, and humans, although differences between human TSCs and other vertebrates have been noted[32,33]. We will document similarities and differences among TSCs and point out unresolved questions about TSCs and their roles at end organ structures. We will also highlight how maladaptive changes in TSCs could have negative consequences for the function and health of the end organ structure and the neuron that innervates it. Additionally, we will discuss how current understanding of TSCs is providing clues to treat acute and chronic conditions that impair the function of neurons and peripheral tissues.

Table 2.

Transgenic mouse lines validated for TSC expression.

SC population expression	Mouse line	Identifier	Reference
TSCs	NG2-DsRed	IMSR_JAX:008241	[26,35]
Non-Myelinating SCs	Kir4.1-CreER	N/A	[52]
All SCs	Dhh-Cre	IMSR_JAX:012929	[127]
All SCs	Plp1-CreER	MGI:2663093	[49]
All SCs	Plp1-CreER	IMSR_JAX:005975	[13,16,36,40,63]
All SCs	Plp1-GFP	IMSR_JAX:033357	[22]
All SCs	S100b-GFP	IMSR_JAX:005621	[95,100,101]
All SCs	S100b-rtTa	N/A	[138]
All SCs	Sox10-CreER	MGI:5910373	[16,17,49]
All SCs	Wnt1-Cre	IMSR_JAX:009107	[49,59,65]

Morphological diversity of TSCs

TSCs are present at end organ structures in both the motor (Figure 1) and sensory divisions of the somatic PNS (Figure 2). Two end organ structure types are found in the motor PNS, the NMJ (Figure 1B) and the muscle spindle. The NMJ consists of the axon terminal of an α-motor neuron, an extrafusal muscle fiber and approximately 3 TSCs, also referred to as perisynaptic Schwann cells, in young adult mice[26,34–36]. The NMJ is a cholinergic synapse in mammals responsible for initiating voluntary movements by causing muscle contraction[34]. At the NMJ, TSCs completely cover the arborized axon terminal and separate the synaptic cleft from the surrounding cellular and molecular milieu of skeletal muscles[10,34]. Muscle spindles consist of axon terminals from γ-motor neurons and proprioceptive sensory neurons as well as intrafusal muscle fibers[37]. Muscle spindles modulate voluntary movements and their γ-motor axon terminals are associated with TSCs[37,38].

Figure 1. TSC morphology at skeletal muscle end organ structures.

A) Innervation pattern of skeletal muscle, including neuromuscular junctions (dark blue), free nerve endings (orange), Golgi tendon organs (purple), and muscle spindles (yellow and light blue). B) The NMJ lies on an extrafusal muscle fiber and its axon terminal (blue) is entirely capped by TSC processes (green). In cross section, the axon terminal lies in the synaptic gutter (gray) and is capped by the TSC. mSCs = myelinating SCs.

Figure 2. TSC morphology at cutaneous end organ structures.

A) Peripheral nerve end organ structures in the skin, including Meissner corpuscles (MCs, blue), Merkel cell-neurite complexes (MNCs, maroon), free nerve endings (FNEs, orange), Pacinian corpuscles (PCs, purple), and hair follicles (yellow). B) MC innervated by two axon terminals. The TSCs produce lamellae in the form of a coin stack around the axon terminals. C) The complete coverage of FNE axon terminals in the skin by TSC processes transitions to partial coverage after crossing the dermal/epidermal border. D) In the PC, the axon terminal region is covered in TSC lamellae in the form of an onion bulb. E) Hair follicle innervated by lanceolate axon terminals running perpendicular to the follicle shaft and circumferential axon terminals wrapping around the follicle shaft (gray). In B-E, axon terminals are depicted in dark blue and TSC processes are depicted in green. mSCs = myelinating SCs.

The end organ structures of the sensory PNS are far more diverse than those of the motor system and, for the most part, do not utilize chemical synapses to transmit sensory information from peripheral tissues to the CNS. The largest group of end organ structures in the sensory PNS are in the skin and fall under the category of mechanoreceptors and nociceptors that are innervated by axon terminals from Aβ, Aδ and C sensory neurons[22,33]. These end organ structures include PCs, MCs, hair follicles, and FNEs[15,16,22], covering the sensory modalities of touch, vibration, skin stretch, pain, temperature, and proprioception. PCs (Figure 2D) are the largest of these end organ structures with axon terminals at the center surrounded by multiple thin layers of TSCs that together form an onion bulb-like structure. Each PC contains upwards of 70 TSCs and functions to transmit forces associated with vibration in the skin to the CNS[11,13,39]. MCs (Figure 2B) are comprised of one or two axon terminals that intermingle with around 4 TSCs that form stacked layers and occupy the bulk of the end organ structure[13]. MCs detect fine touch against the skin. Because of their layered morphology, TSCs of both the PC and MC have been classically referred to as lamellar cells[22]. Hair follicles (Figure 2E) contain axon terminals that are either parallel with the hair shaft, known as lanceolate nerve endings, or encircle it, known as circumferential nerve endings[13,40]. The nerves innervating hair follicles sense hair movement. These axon terminals intermingle with one another and are flanked by multiple TSCs that maintain distinct territories within the hair follicle structure[13,14]. However, individual TSCs simultaneously associate with multiple axon terminals[13,40]. FNEs (Figure 2C), which sense pain among other modalities, are classically called “free” because they do not terminate in some larger structure such as a corpuscle. Nevertheless, morphological studies indicate that the axon terminals of FNEs are associated with TSC processes throughout their length[15,16].

Far less is known about the TSCs associated with proprioceptive sensory end organ structures in skeletal muscle, which include muscle spindles and Golgi tendon organs (GTOs). GTOs are present near the interface of myofibers and tendons, known as the myotendinous junction, and relay information about the amount of force that is being placed on this junction as muscles are stretched[41]. The branched axon terminals in this organ are surrounded by as many as 15 TSCs[41]. Sensory axons in muscle spindles wrap around intrafusal muscle fibers and detect muscle stretch[42]. To date, TSCs have not been found associated with these types of proprioceptive sensory axon terminals[37].

Origin, differentiation, maturation, and expansion of TSCs

Origin of TSCs:

It is generally agreed that TSCs and other SCs found along the length of peripheral nerves are derived from neural crest cells (NCCs) (Figure 3). Studies across vertebrate species, including frogs, birds, and rodents, indicate that during embryonic development, NCCs exit the neural tube, attach to nascent axons and differentiate into multipotent Schwann cell precursors (SCPs)[43]. SCPs migrate along growing axons as they approach their peripheral tissues. During this time, they continuously proliferate, leaving a trail of multipotent cells along the way, which give rise to nerve SCs. It has been presumed that TSCs come from those same SCPs because TSC precursors have consistently been described in ultrastructural studies as having similar morphologies to SCPs[5,6,8,9,44–47]. However, direct experimental evidence of the origin of TSCs has been sparse with the notable exception of a lineage tracing study that showed TSCs of the avian PC analogue do in fact have the same origin as nerve SCs[48]. Lineage tracing via single cell transcriptomics in mice has also provided evidence that NMJ TSCs likely arise from SCPs[49]. These experiments showed that SCP-derived cells induce expression of TSC-specific genes as they differentiate into NMJ TSCs. Moreover, mature NMJ TSCs retain expression of numerous SCP and immature SC markers[35,49]. It is also worth noting that TSCs in different end organ structures, or possibly within the same end organ structure, may not universally share the same progenitor. This was highlighted in a study showing that a subpopulation of TSCs in FNEs and hair follicles arise from Prss56⁺ boundary cap cells[50].

Figure 3. Developmental lineage of TSCs.

All TSCs originate from the neural crest. A subset of TSCs at hair follicles and FNEs are derived from a transient pool of boundary cap cells. TSCs at other end organ structures descend from the same Schwann cell precursors as SCs in the nerve, although it is unknown whether developing TSCs go through an immature Schwann cell state as do nerve SCs. Abbreviations: FNE: free nerve ending, PC: Pacinian corpuscle, MC: Meissner corpuscle, GTO: Golgi tendon organ, NMJ: neuromuscular junction.

TSC differentiation and maturation:

While the factors that drive precursor cells to differentiate into TSCs remain largely unknown, there is a close relationship between TSCs acquiring their distinctive cellular and molecular characteristics and the maturation of the end organ structure. For instance, as the mouse NMJ matures, TSCs form cellular processes that envelop the axon terminal arbor[28], organize into tiled territories[51], begin to express neuron-glial antigen 2 (NG2), a cell surface proteoglycan[35], and continue to express Kir4.1, an inward-rectifying K+ channel, unlike myelinating SCs[52]. A similar relationship has been found at PCs and MCs. In rodents and humans, as both of these end organ structures mature, TSCs express high levels of vimentin[53,54]. These relationships and the dependence of TSCs on axonal contact during development[41,55,56] suggest that end organ structure specific factors contribute to the differentiation and subsequent maturation of TSCs.

TSCs expansion:

The number of TSCs increases as end organ structures mature and grow in size during development. The number of TSCs increases from 1 to about 3 per NMJ in maturing rodents[57]. This increase in TSCs correlates with the transformation of the NMJ from a simple and small plaque-like structure into a large pretzel-like synapse[57]. There is a similar relationship in PCs and MCs. At maturing PCs in rodents and humans, TSCs greatly expand in number and alter their shape to form the thin lamellae of the onion bulb structure that surrounds the axon terminal ending[45,46,54]. At maturing murine MCs, TSCs increase from 1 to around 4 per end organ structure and elaborate lamellae in which the axon terminals are interspersed[13,47].

Studies have identified local factors important for the expansion of TSCs. Axon-derived NRG1 type-III has been shown to significantly impact the number of TSCs per NMJ during development and in adult mice[58]. In human PCs, evidence suggests that NCAM-1 may play a role in increasing the number of TSCs during development, as it is localized to axon-TSC junctions and upregulated as PCs mature[54]. In hair follicles of mice, glutamatergic signaling between TSCs and axon nerve endings has been shown to impact TSC proliferation[59]. TSCs express the glutamate receptor, NMDA. Here, disruption of glutamatergic signaling via conditional knockout of vesicular glutamate transporter 2 (VGLUT2) in sensory axons reduced the number of TSCs at the developing hair follicle. Along the same lines, administration of an NMDA antagonist stunted the formation of TSC processes that intermingle with axon terminals at hair follicles.

TSCs at maturing end organ structures

While TSCs do not appear to be required for axon terminals to reach and connect with end organ structures[60–63], there is evidence indicating important roles for TSCs in the maturation of the axon terminal and end organ structure. This has been best characterized at the rodent NMJ where TSC precursors have been shown to stabilize initial contacts between the axon and myofiber by inhibiting muscle derived thrombins[60,61]. TSCs are necessary for NMJ maturation to proceed past initial contact, as selective ablation of their precursors in-utero causes axons to degenerate shortly after they contact the myofiber[60–63]. Furthermore, those nascent NMJs are initially innervated by multiple motor axons[34], and TSCs participate in the elimination of all but one motor axon. In the context of NMJ development, this process is often referred to as synaptic elimination. TSCs participate in synaptic elimination by detecting differences in synaptic efficacy of competing axonal inputs using the P2RY1 receptor and calcium-mediated signaling[64,65]. Inhibition of P2YRY1 receptors renders TSCs unable to sense synaptic transmission and delays synaptic elimination[27]. TSCs also jockey with motor axons for contact with the muscle fiber surface and phagocytose axons that lose this competition[28]. Through these mechanisms, NMJ TSCs sense the relative strengths of competing synaptic inputs and influence the removal of weaker ones. Elsewhere in the PNS, it remains unknown whether and how TSCs participate in the maturation of their end organ structures.

TSCs at healthy adult end organ structures

TSCs insulate the site of contact between the axon terminal and end organ structure:

In each end organ structure, TSCs separate the axon terminal and its partner end organ structure region from the surrounding peripheral tissue. TSCs work together within a given end organ structure to create a physical barrier using finger-like cellular processes that tightly wrap around the perimeter of the axon terminal. This leaves the leading portion of the axon terminal exposed to communicate with the end organ structure. In the case of the NMJ, the exposed part of the axon terminal directly abuts the myofiber, and is the site of synaptic transmission[34]. In FNEs [15,16], MCs, PCs, and hair follicles [13], protrusions of the axon terminals beyond the TSC covering have been proposed to be involved in the conversion of external stimuli to afferent neuronal signal.

Research efforts in rodents and humans are revealing the identity of molecules with potential roles in the formation and function of TSCs processes. NMJ TSCs have been found to express high levels of transcripts for molecules involved in phagocytosis, cell adhesion, and extracellular matrix formation[35]. These include Ncam1, Ajap1, Megf10, members of the dystroglycan complex as well as connexin, nidogen, laminin and collagen subtypes. NCAM1 has also been found in TSCs of MCs and PCs[54]. In addition to anchoring TSC processes to the site of contact between the axon terminal and end organ structure, these molecules may also be involved in TSCs transducing signals among each other and to axon terminals. In support of this model, adherens junctions have been shown to connect TSCs and axon terminals and to possibly play roles in sensory signal transduction in hair follicles, MCs, and PCs[13].

There are a number of outstanding questions about the barrier TSC processes form. For instance, it is unknown which molecules are able to pass the TSC barrier. If the permeability of the TSC barrier is stringent, how are nutrients and waste carried in and out of the end organ structure? One possibility is that TSCs perform these functions by forming a barrier that is reminiscent in some ways of the blood brain barrier (BBB), and thereby connecting axon terminals and abutting end organ structure region with the vasculature. The BBB, which is the physical barrier that strictly regulates the influx of molecules from the blood stream into the CNS[66], is paralleled in the PNS by the blood-nerve barrier (BNB) that ensconces large peripheral nerves, such as the sciatic nerve[67]. Like astrocytes in the CNS, Schwann cells regulate the permeability of capillaries at the BNB[67]. Whether the BNB continues to extend along axons that branch out from these larger nerve structures as they fan out into peripheral tissues and whether TSCs influence its permeability to blood-derived molecules is largely unknown. Addressing these questions may be needed to fully understand the function of TSCs in the maturation and later in the maintenance of axon terminals and end organ structures.

TSCs detect and modulate neurotransmission:

TSCs across various end organ structures have been consistently shown to be involved in, and indeed critical for, proper neuronal signaling. NMJ TSCs in mice and frogs detect synaptic activity through muscarinic acetylcholine receptors[64,68], adenosine receptors[69], and purinergic receptors[70], which trigger a release of intracellular stores of Ca²⁺ and initiates G protein mediated signaling within the TSCs[70–74] (Figure 4A). NMJ TSCs also express butyrylcholinesterase (BChE), which fine tunes their modulation of synaptic plasticity in response to acetylcholine spillover [75]. Furthermore, differential nerve stimulation patterns result in different TSC Ca²⁺ wave patterns, and when that Ca²⁺ activity was optically mimicked in TSCs, the synaptic plasticity seen with nerve stimulation was replicated[76]. Despite decades of evidence that TSCs respond to synaptic activity in a way that can alter the function of the synapse itself, the precise mechanisms TSCs use to influence synaptic plasticity at NMJs remain to be fully elucidated.

Figure 4. TSCs actively participate in the physiology of their axon terminals.

A) At the neuromuscular junction, axon terminals release vesicles containing transmitters such as acetylcholine and ATP. These interact with muscarinic acetylcholine receptors (AChRs) and purinergic receptors on TSCs, which cause an increase in intracellular Ca²⁺, leading to synaptic function modulation by the TSC. Butyrylcholinesterase (BChE), expressed by the TSC, also plays a role in modulating synaptic plasticity. B) At the Pacinian corpuscle, TSCs and the axon terminal express ion channels involved in mechanosensitivity (ASIC2 and Piezo2, respectively). Membrane deformation via physical force depolarizes the axon terminal. This causes the axon to release glutamate onto the TSC, inducing the TSC to release GABA onto the axon, which in turn shuts down neuronal transmission.

TSCs of sensory end organ structures have been far less studied compared to those at NMJs. It is clear, however, that TSCs at sensory end organ structures facilitate the transduction of forces associated with stretch and vibration into neural signaling. Current evidence in mice, ducks, and humans suggests TSCs express mechanoreceptors, as do other PNS glia[77]. TSCs may modulate afferent signaling through mechanosensitivity, vesicular release of molecules onto the axon terminal, and/or by forming anchors with the axon terminal membrane. TSCs in the MC express receptors involved in mechanosensitivity, TRPV4[78] and ASIC2[79,80], which is also expressed by TSCs in PCs. Depolarization of TSCs via these mechanosensitive receptors may be important for sensory signaling[16,81], as evidenced by the depolarization of FNE TSCs being sufficient to induce a pain response in mice[16,82].

Evidence in mice and cats also suggests that TSCs influence the depolarization of their axon terminals by releasing select molecules and transferring mechanical force to axon terminals. TSCs have long been known to contain vesicles[11,39,40], and PC TSCs express components required for glutamatergic and GABAergic signaling[83]. Experiments with PCs utilizing glutamate and GABA antagonists suggest a model where mechanical excitation of the axon terminal causes it to release glutamate, which in turn causes TSCs to emit GABA onto the axon, inhibiting the release of glutamate and shutting down transmission shortly after response onset[25] (Figure 4B). TSCs may also influence neuronal excitation by transferring mechanical force from the surrounding tissue to the axon terminal. Mouse MC and hair follicle TSCs express the transmembrane protein Ush2A, which is necessary for proper sensation in these end organ structures, and may cause axon terminals to be mechanically affected when the TSC membrane is deformed by an external stimulus[84]. In the PC, MC and in hair follicles, adherens junctions between TSCs and the axon terminal are thought to be responsible for mechanical stretching of the terminal membrane in response to force. It has been proposed that this TSC-mediated deformation of the axonal membrane is necessary for activation of Piezo2 mechanoreceptors, which are exclusively located on axon terminals within these end organ structures[13]. These studies suggest that TSCs are conduits for transducing mechanical forces into action potentials by providing structural support or possibly through mechanoreceptor-mediated signaling. At the same time, TSCs may have a critically important role in fine-tuning the depolarization responses of sensory axons to mechanical stimuli through the release of GABA and glutamate, or possibly other signaling molecules.

TSCs maintain the function and structure of adult axon terminals:

Given that TSCs create a physical barrier to protect the end organ structure from the external environment and are important for proper neuronal signaling, it stands to reason that they may be necessary for the integrity and health of axon terminals. To examine whether that is the case, a number of studies have ablated TSCs at various end organ structures. TSC ablation at frog and mouse NMJs does not immediately impact electrophysiological properties, suggesting that TSCs are not required for synaptic transmission per se[62,85]. However, after 3–7 days without TSCs, neurotransmitter release and compound muscle action potential amplitude are decreased[62,63,85]. Likewise, when FNE TSCs are ablated, pain responses in mice are altered[17]. When TSCs are stripped away from PCs in cats, the axon terminal’s response to stimuli shifts from a rapidly adapting response to a slowly adapting response[86]. These effects may be, at least in part, due to full or partial degeneration of axon terminals in the absence of TSCs or off-target ablation of the SCs that myelinate these axons. While adult murine motor and circumferential LTMR axon terminals seem to be more resilient to TSC ablation[36,40], lanceolate LTMR and FNEs degenerate without their TSCs[17,40]. Altogether, published data support roles for TSCs in the function and stability of adult axon terminals and end organ structures.

TSCs in injury, disease, and aging

Injuries to peripheral nerves:

In adult rodents, TSCs repair end organ structures following axonal degeneration by clearing axonal debris, recruiting axons back to the end organ structure, and assisting with rebuilding the end organ structure upon reinnervation. With the exception of axon terminals at FNEs[17] and circumferential LTMR endings[40], where TSCs die shortly after axon retraction, TSCs remain at end organ structures following axonal degeneration in adults[24,40,87,88]. Upon denervation, TSC enter a reactive state characterized by a number of morphological and molecular changes. At the NMJ[24] and the PC[23], TSCs engulf components of the degenerating axon terminal. NMJ TSCs upregulate markers of phagocytosis (galectin-3[30]), migration (galectin-1[89]), and astrocytic reactivity (GFAP[90]), as well as other repair markers (p75-NTR[91], GAP-43[92], Nestin[93–95], and Zipro1[96]). The lamellae of MC TSCs shrink in size, with reduced expression of BChE after axotomy[88].

Once the debris from the former axon terminal has been cleared, NMJ TSCs have been posited to attract motor axons back to denervated NMJs by extending cellular processes far away from their original synaptic territory[91]. This includes extension of processes towards healthy neighboring NMJs, where they provide a path for the axon terminal of the innervated NMJ to sprout towards the denervated NMJ, in a process known as compensatory reinnervation[95,97–101]. TSC repair responses have not been as extensively studied in other end organ structures. While elimination of redundant inputs occurs in the PC after nerve injury, the role of TSCs in this process is not known[102]. TSCs do actively phagocytose debris from injured axons in the PC, suggesting that they may participate in the repair of this end organ structure following denervation. Re-innervation of the MC causes TSCs to regain healthy phenotypes, though the extent to which TSCs facilitate MC reinnervation is unknown[103].

Disease:

There is evidence that TSCs impact the health of motor and sensory end organ structures afflicted by disease. The NMJ is an early site of motor neuron degeneration in amyotrophic lateral sclerosis (ALS), and work in rodent ALS models suggests that TSCs acquire properties that contribute to deterioration of this synapse[31,104]. TSCs at the NMJ show altered sensitivity to synaptic transmission[105], migration away from the NMJ[106], aberrant insertion of processes into the synaptic cleft[107], impaired phagocytosis of degenerating axonal debris[108], and increased expression of harmful factors such as complement proteins[32] and Sema3A[109]. Many of these deleterious changes in TSCs have been observed before NMJ degeneration is evident. Thus, TSCs appear to acquire unique characteristics in ALS that may precipitate NMJ degeneration. Mouse models of spinal muscular atrophy (SMA) show that NMJ TSCs are impacted during the early stages of NMJ degeneration. This includes subtle changes in morphology, reduced S100B expression, increased presence of TSC sprouts associated with motor axons, and fewer TSCs in SMA mice[110–113]. In the mdx mouse model of Duchenne muscular dystrophy, NMJ TSCs form extensive sprouts that are seldomly associated with motor axons, exhibit reduced coverage of motor axon terminals, and show an increased number of TSCs per NMJ[114,115]. In the early stages of peripheral neuropathy associated with diabetes, the nerve endings at FNEs degenerate and TSCs are lost[116,117]. These findings indicate that TSCs at other end organ structures are at a minimum indirectly impacted by diseases.

Aging:

The ability of TSCs to repair end organ structures becomes particularly important as these structures begin to degrade during the course of aging. In the case of the NMJ, degeneration has been linked to sarcopenia and loss of motor function[104,118] and evidence suggests that TSCs may be involved in this process in both a beneficial and harmful manner. TSCs at aged NMJs in mice upregulate inflammatory and phagocytic genes, expand in numbers, and form processes, often associated with an axonal sprout, that contact neighboring NMJs[26]. They are thought to slow age-related muscle dysfunction by promoting compensatory reinnervation of denervated muscle fibers by nearby healthy motor axons[26,119,120]. However, there is also evidence that age-related changes in TSCs may contribute to NMJ damage. In aging humans and rodents, TSCs not only fail to tightly cap the synaptic cleft but aberrantly insert their processes into it[26,121–123]. Age-related deficits are not limited to the motor PNS. There is evidence that some sensory axon terminals at muscle spindles degenerate in aged mice[124] and humans[42], and given that proprioception is necessary for proper motor function, this degeneration may contribute to age-related loss of motor function[124]. Aged humans lose the ability to sense touch, impacting their ability to grasp objects[125], and this has been linked to age-related loss of PCs and MCs[126]. The roles of TSCs in preserving the health of aging end organ structures have not yet been fully established. Future research should address the fate of TSCs and their contribution to the health of end organ structures during aging.

Concluding remarks and future perspectives

Much progress has been made to understand the developmental origins of TSCs as well as their roles in the maturation, function, health, and repair of associated axon terminals and end organ structures. TSCs are derived from NCCs and the majority appear to emanate directly from SCPs. They are also dependent on axonal contact to survive during development and are actively involved in the maturation of their end organ structures. Research over the last decades has revealed that TSCs play important roles in the proper function of adult end organ structures and their repair upon damage. Additionally, it is becoming apparent that TSCs lose their homeostatic functions in diseases and aging, and thus likely contribute to end organ structure degeneration. Despite these advances, many questions remain (see Outstanding Questions). Addressing these and other questions about the biology of TSCs could pave the path for developing therapies to treat a variety of PNS conditions. It may also provide leads for treating CNS conditions, given the many functional parallels between TSCs and CNS glia, such as their roles in synaptic pruning, axon guidance and modulation of neuronal activity.

Outstanding Questions.

• The presence of TSCs at the axon terminals is conserved across various end organs, yet TSCs at different end organs display different morphologies and molecular functions. What are the TSC-intrinsic molecules that specify their identity and various functions?

• Current evidence suggests that TSCs are derived from the same progenitors as SCs along the nerve. What factors, if any, do axon terminals and other end organ structures produce to promote TSC differentiation and maturation?

• TSCs have been found to be altered in aging and various diseases. Which of the cellular and molecular changes to TSCs in these conditions are beneficial or detrimental to the sustained physiology of their associated axon terminal?

• Do TSCs acquire different functional phenotypes and lose their homeostatic functions under chronic stress conditions, such as in disease and aging, as has been found for astrocytes and microglia in the CNS?

• What is the consequence of TSCs’ losing the appropriate coverage of their axon terminal on the peripheral tissue and the innervating neuron residing either in the PNS or CNS?

• Do TSCs influence the permeability of an end-organ level blood-nerve barrier? If so, is this structure damaged by disease and aging?

Highlights Box.

• Terminal Schwann cells are highly specialized, non-myelinating glia, associated with axon terminals of the somatic peripheral nervous system.

• Terminal Schwann cells mediate communication between the nervous system and peripheral tissue. They play key roles in the development, stability, and repair of axon terminals and peripheral tissues.

• Accruing evidence points to maladaptive changes in terminal Schwann cells contributing to degeneration of peripheral tissues and innervating neurons in diseases and during aging.

• Terminal Schwann cells contain a unique molecular fingerprint compared to myelinating Schwann cells.