Basal lamina: A novel pH regulator at the neuromuscular junction

Abstract

Proton concentration can change within the cleft during synaptic activity due to vesicular release and Ca²⁺ extrusion from cellular compartments. These changes within the synaptic cleft can impact neural activity by proton-dependent modulation of ion channel function. The pH transient differs in magnitude and direction between synapses, requiring different synapse types to be measured to generate a complete understanding of this mechanism and its impacts on physiology. With a focus on the mouse neuromuscular junction (NMJ), the recently published “Postsynaptic Calcium Extrusion at the Mouse Neuromuscular Junction Alkalinizes the Synaptic Cleft” measured synaptic cleft pH at a cholinergic synapse and found a biphasic pH transient. The study demonstrated that the changes in proton concentration found were due to postsynaptic signaling when measuring pH at the muscle membrane, despite the expectation of a presynaptic contribution. This result suggests a diffusional barrier within the NMJ isolates pH transients to presynaptic versus postsynaptic compartments. Generating a Donnan equilibrium that impacts protons, evidence suggests the basal lamina may be a key regulator of pH at the NMJ. Exploring synaptic pH, proton regulating factors, and downstream pH transient effects at presynaptic versus postsynaptic membranes may lead to new insight for a variety of diseases.

Introduction

Changes in proton concentrations across different synaptic clefts have been documented and shown to occur through vesicle release, acid/base transporters, or Ca²⁺ extrusion during neurotransmission. ¹ These proton transients act on several voltage- and/or ligand-gated ion channels that are integral to neurotransmitter release, impacting the excitability of neurons.^1–4 Understanding the directionality of these transients and how they regulate neural communication provides an additional layer of synaptic plasticity, which is currently underappreciated. Understanding how these transients become dysregulated could provide insight into multiple disease states that impact the nervous system. Research into the role of synaptic pH is hindered by difficulties in generalizing changes in pH that occur during neurotransmission and identifying the underlying mechanisms at the synaptic cleft. The directionality, magnitude, and effect of pH changes at the cleft depend on a variety of factors, including the synapse being measured. For example, canonical examples, such as ribbon synapse, have been shown to acidify during transmission; however, cleft pH alkalizes at the calyx of Held and hippocampal glutamatergic synapses.^2–4 Thus, increased efforts to define these transients at a variety of synapses are required for a more holistic understanding of synaptic pH dynamics. Toward this goal, we recently sought to characterize transient changes in pH, specifically at the neuromuscular junction (NMJ) during transmission. ⁵ Physiological pH transients that caused both acidification and alkalization of the cleft were identified; however, the results left open questions regarding how it is regulated at the NMJ. Below, we speculate on possible mechanisms for the postsynaptic acidification we observed at the NMJ endplate in that study.

A summary of activity-dependent pH transients at the NMJ

Previous studies have had contradictory findings when measuring pH at the NMJ. Drosophila melanogaster glutamatergic NMJs were found to have large cleft alkalization in vivo and ex vivo at the synaptic cleft when measured either pre- or postsynaptically. ⁴ At mammalian cholinergic NMJs, motor neuron terminals showed intracellular acidification followed by long alkalization, suggesting that reciprocal transients should exist extracellularly in the synaptic cleft. ⁶ However, the cleft only shows slight acidification under strenuous stimulation when presynaptic activity was isolated, and measurements were made from the postsynaptic muscle membrane at the endplate. ⁷ While some differences between vertebrate and invertebrate NMJs could be expected due to evolution and different transmitter types (acetylcholine versus glutamate, respectively), the large discrepancy between the intracellular and cleft pH of the mammalian NMJ was surprising.

Measuring activity-dependent synaptic events in the NMJ, using either electrophysiology or optical approaches, is difficult due to muscle movement during contraction. Most ways of managing contractions involve inhibiting early targets in the excitation-contraction cascade in muscle pharmacologically, namely by blocking nicotinic acetylcholine receptors to completely incapacitate the muscle. Specifically, these inhibitors block cytosolic Ca²⁺ influx in muscle, both from the opening of voltage-gated calcium channels on the membrane and intracellular release from stores. Consequently, they affect proton movement via the plasma membrane Calcium ATPase.^3,4,6 The recent work from Durbin et al. bypassed this problem to assess the link between muscle Ca²⁺ and cleft pH using the myosin inhibitor 3-(N-butylethanimidoyl)-4-hydroxy-2H-chromen-2-one (BHC) to stop muscle contraction but allowing the large intracellular release of Ca²⁺ from the sarcoplasmic reticulum during activity. ⁵ Similarly, the muscle-specific voltage-gated sodium channel blocker μ-conotoxin was used in parallel experiments, allowing some small exogenous Ca²⁺ influx but blocking intracellular release.

Using these inhibitors and postsynaptic expression of the ratiometric fluorescent probe pHusion-Ex in the fast-twitch levator auris longus muscle, the innervating nerve was stimulated, and pH transients in the synaptic cleft were measured. Transients were activity-dependent, biphasic, proportional to stimulation duration, and required postsynaptic intracellular release of calcium. For example, a short alkalization transient was seen when the preparation was stimulated for 5 s in BHC, and an additional acidification of the cleft was seen when stimulated for 20 s in BHC. These transients were found to be modulated specifically by intracellular calcium release from the muscle. The exact mechanism for acidification remains unknown despite testing for a role of lactate transport through monocarboxylate transporters and a role of sodium-proton exchange by pharmacological blockade. Rather than the movement of protons, it thus stands to reason that the movement of a basic ion, like HCO₃−, into the cell could be responsible for this change. Regardless, the mechanism behind the alkalization transient was solved. PMCA was localized to the NMJ, and a specific PMCA inhibitor, caloxin 1b1, significantly decreased the alkalization found, suggesting that PMCA substantially contributes to the alkalization that occurs at the synapse. In separate experiments, μ-conotoxin treatment showed only a slight alkalization when stimulating for 20 s, confirming that PMCA activity and alkalization increase proportionally to intracellular muscle Ca²⁺ levels. ⁵ These results ultimately demonstrated the importance of muscle in the generation of the cleft pH transient and how muscle contributes to a possible retrograde signal that can impact neurotransmission. Additionally, the identified transients seemed to extend past the synaptic cleft into the surrounding muscle; however, the magnitude, dynamics, and spread of interstitial transients were not characterized because the ROIs used for analysis were restricted to the endplate. ⁵ This may indicate possible diffusion of pH signals throughout the muscle in response to neurotransmission, but more extensive investigation is required to understand this phenomenon. However, it is still ambiguous whether these transients are generalizable to all NMJs, or specific to fast-twitch muscle, due to differences in Ca²⁺ extrusion or general proton handling.

The NMJ as a divided synapse

Interestingly, other effects were seen that suggest the existence of more regulatory elements that were not considered originally. For example, there were only marginal presynaptic pH changes when muscle depolarization was blocked,^5,7 illustrating a disagreement between extracellular pH differences predicted by intracellular pH transients in the motor neuron terminal and what was observed on the endplate membrane of the muscle. ⁶ Further, the contributions of the muscle measured at the cleft, although reciprocal of what was found in motor neuron terminals, were 2–6 times larger than what would be predicted to be in the cleft based on presynaptic data.^5,6 Given that the intracellular alkalization at motor neuron terminals was found to be the result of prolonged association of vesicular ATPase on the cellular membrane, it is likely those protons causing the assumed corresponding extracellular acidification are deposited into the cleft after exocytosis. ⁶ There was also difficulty in the experimental disruption of pH buffering of the transients. Increasing the buffering capacity using a high concentration of HEPES only decreased the time to recovery and not the amplitude of the transient at this synapse. ⁵ These results indicate the presence of some sort of obstruction slowing the diffusion of protons in the NMJ synaptic cleft.

If the idea of an obstruction was extended to include the data showing presynaptic signaling, a diffusion barrier within the cleft of the NMJ could stand as a reason for the lack of impact of the presynaptic transients when measuring pH postsynaptically. This diffusion barrier dividing the synapse would have to restrict the flow of protons yet allow for the fast diffusion of neurotransmitters to the muscle across the synaptic space. Given that protons will be much smaller than acetylcholine, a physical restriction of space would not be possible, so an electrochemical force would need to be present to specifically trap these small positive ions released from the presynaptic terminal.

The idea of a diffusional barrier trapping ions between tissues is not a new one. Diffusional barriers restricting the movement of ions have been found surrounding neurons within the central nervous system. A specialized portion of the extracellular matrix, called a perineuronal net (PNN), forms a mesh surrounding fast-acting neurons, allowing neurons to spike faster and manage reactive oxygen species because it traps cations and repels anions near the cell surface. ⁸ Cations such as Zn²⁺, Fe²⁺, Ca²⁺, Na⁺, and K⁺ are retained due to immobile chondroitin sulfate proteoglycans in the PNN that are negatively charged, establishing a Donnan equilibrium. This diffusion barrier likely affects how protons and bases move around neurons beyond the synaptic contacts. Perineuronal nets were suggested to sensitize dorsal root ganglion cultures to low pH. ⁹ Thus, PNNs can greatly affect the excitability and function of the cells that have them.

The existence of diffusion barriers impacting proton concentration provides proof of concept, and evidence supports the existence of one at the NMJ. The basal lamina is a specialized extracellular matrix that surrounds the muscle. This basement membrane is found at the NMJ, separating the presynaptic motor neuron and the postsynaptic endplate and providing structure and support to the NMJ as a whole.^10,11 Like PNNs, the basal lamina also contains many negatively charged proteoglycans, such as agrin, that are localized to the NMJ. Agrin, in particular, has also been implicated in the generation of lipid rafts at the endplate, allowing for possibly additional negatively charged glycolipids and proteins to aggregate in the NMJ and contribute to a Donnan equilibrium. ¹² And pores at the NMJ that are ∼25–30 nm in diameter that would allow acetylcholine to pass through.^10,11 We propose that the basal lamina is a good candidate for a proton-specific diffusion barrier at the NMJ and could establish its own Donnan equilibrium.

Modeling the role of the basal Lamina in regulating pH

Assuming the existence of a diffusional barrier at the NMJ with effects like PNNs, there are a few feasible ways that diffusion of protons could work at the NMJ to cause the synapsic acidification identified under physiological conditions. We present several hypotheses (Figure 1):

Protons released from the presynaptic terminal move toward the muscle. Given the draw of the negatively charged proteoglycans, protons from both presynaptic and postsynaptic sources could aggregate at the muscle surface. The presynaptic signal could have been small enough that it could not be resolved from noise; however, data of similar size to the expected pH change due to presynaptic effects (change in pH ∼0.1–0.2 units) were resolved. ⁶ While a formal possibility, this hypothesis is poorly supported by data and we think unlikely.

The presynaptic and postsynaptic sides of the NMJ are proton islands. Due to either the relatively large (∼100 nm) space of the synapse, the existence of a basal lamina, or both, localized proton transients do not affect pH gradients on the other side of the synapse. ¹¹ This would suggest that the pH transient serves as a self-modulating feedback loop, affecting the motor neuron and the muscle independently.

The release of protons is primarily a retrograde signal. Due to the size of the cleft and the size of the presynaptic transients, pH changes generated by presynaptic motor neurons diffuse rapidly to insignificant levels in the extracellular space before reaching probes anchored on the muscle endplate membrane. Alternatively, the eerily similar time course of both presynaptic and postsynaptic changes in pH and the fact that postsynaptic transients are at least double the size of the corresponding presynaptic one indicates the basal lamina acts as a one-way valve.^5,6 Although most protons become trapped near the surface of the muscle, a sufficiently large postsynaptic release of protons could overwhelm the Donnan equilibrium and diffuse toward the motor neuron because the concentration of protons will always be lower in that direction during transmission. The full amount of postsynaptic acid would not make it to the neuron in this case, but the presynaptic transient would likely become more acidic due to contributions from postsynaptic signaling.

Figure 1.

Graphical models of basal lamina function in proton signaling. Each panel displays a different hypothesis for how the basal lamina acts on presynaptic and postsynaptic protons that enter the synaptic cleft of the neuromuscular junctions. The nerve (gray) is positioned above the muscle (red), and the basal lamina is represented by the dashed line between them. The arrows indicate the number of protons (scaled by line weight) and indicate the direction in which protons are flowing. The dashed arrow indicates a possible retrograde proton transient that crosses the basal lamina. Presynaptic pH is shown to be influenced by vesicular release and temporary association of vATPase to the membrane. Although the cause of postsynaptic acidification is unknown, sodium/hydrogen exchangers (NHE) and bicarbonate transporters are suggested mechanisms for increasing postsynaptic proton concentrations (created with BioRender.com).

Speculated targets and effects of pH transients at the NMJ

In any case, these proton signals and organization of the basal lamina are likely having some downstream effects at the NMJ, both pre- and postsynaptically. In terms of presynaptic targets, it is likely that voltage-gated calcium channels will be affected by the drop in pH facilitated by the brief association of vATPase to the membrane, vesicular release, and possibly an acidic transient from the muscle.^2,6 Along with vATPase, vesicular acetylcholine transporters may also associate with the membrane and cause nonvesicular acetylcholinesterase (AChE) release after vesicular fusion and before endocytosis during this expected drop in cleft pH on the presynaptic side of the synapse. Further, AChE has been found to be pH sensitive, and it may have reduced activity as protons aggregate near the basal lamina. ¹³ Being a tripartite synapse, the NMJ still has an unknown quantity we have not mentioned: perisynaptic Schwann cells. Perisynaptic Schwann cells remain unstudied in regards to cleft pH, and it is unknown what role they may play in contributing to or buffering pH at the NMJ.

Disease relevance: Muscle function and acidosis

Regardless of how the basal lamina influences pH at the NMJ, identifying how pH is regulated and how pH changes in the synaptic cleft allows us to predict the effects of these pH transients on NMJ function. It is still relatively unclear what downstream effects these transients have on the motor neuron and muscle. Recently published work suggests that reduced pH increases endplate potentials due to an increase in SV release via presynaptic homeostatic potentiation. ¹⁴ This potentiation occurs when there is a block of postsynaptic nicotinic acetylcholine receptors, but it could also be a compensatory action for the acid-dependent inhibition of voltage-gated Ca²⁺ channels on the motor neuron terminal membrane.

An important open question remains regarding what happens when synaptic pH is dysregulated. A drop in extracellular pH is increasingly acknowledged in the pathogenesis of diseases within the nervous system. Several neurodegenerative diseases are associated with progressive pH changes, such as ALS, where acidosis slowly develops and reducing proton buffering power has been shown to negatively affect survival in mice. ¹⁵ Inhibiting matrix metalloproteases, which are proteins that break down the basal lamina, was shown to improve survival in an ALS model, suggesting the basal lamina plays an important role in ALS progression as well. ¹⁶ Even psychological disorders like schizophrenia are incorporating dysregulated pH into their models and hypotheses. ¹⁷ Measurements of synaptic pH transients and understanding their underlying regulatory factors and mechanisms may prove useful in providing rational targets for these various mental disorders and disease states. More extensive testing is required to validate the synaptic pH transients, proton regulatory factors, and pH effects across other model systems.