Endoplasmic reticulum–mitochondria interplay in chronic pain: The calcium connection

Abstract

Chronic pain is a debilitating condition that affects roughly a third to a half of the world’s population. Despite its substantial effect on society, treatment for chronic pain is modest, at best, notwithstanding its side effects. Hence, novel therapeutics are direly needed. Emerging evidence suggests that calcium plays an integral role in mediating neuronal plasticity that underlies sensitization observed in chronic pain states. The endoplasmic reticulum and the mitochondria are the largest calcium repositories in a cell. Here, we review how stressors, like accumulation of misfolded proteins and oxidative stress, influence endoplasmic reticulum and mitochondria function and contribute to chronic pain. We further examine the shuttling of calcium across the mitochondrial-associated membrane as a mechanism of cross-talk between the endoplasmic reticulum and the mitochondria. In addition, we discuss how endoplasmic reticulum stress, mitochondrial impairment, and calcium dyshomeostasis are implicated in various models of neuropathic pain. We propose a novel framework of endoplasmic reticulum–mitochondria signaling in mediating pain hypersensitivity. These observations require further investigation in order to develop novel therapies for chronic pain.

Introduction

Pain is an adaptive response to noxious stimuli that cause tissue injury or have the potential to cause an injury.¹ It is the most common symptom across a wide variety of medical conditions ranging from autoimmune disorders like rheumatoid arthritis and multiple sclerosis, to traumatic nerve injury^2,3 and is a major reason for seeking consultation with a physician.^4,5 When pain loses its adaptive value and becomes persistent, even after the initial injury has healed, it becomes a disease in its own right.¹

Studying pain in a laboratory setting is particularly difficult due to the subjective nature of pain perception. To overcome this obstacle, researchers carefully control the extent of the injury or disease and closely monitor nocifensive behaviors in animals. Models of neuropathic pain, such as spinal nerve ligation (SNL) and spared nerve injury (SNI), typically induce traumatic damage to the nerve resulting in evoked and spontaneous pain behaviours.^6,7 Other disease-oriented pain models, like diabetic neuropathy and chemotherapy-induced pain, affect the molecular integrity of nerves and typically induce a more global, complex injury response involving inflammation, immune activation, and gliosis as well as changes in the neurovasculature.^8,9 The underlying pathology responsible for pain hypersensitivity across these conditions remains enigmatic. It is clear, however, that pain is a common feature of many medical conditions, regardless of their aetiologies. Whether these pain states share common or divergent mechanisms is being actively investigated.

Chronic neuropathic pain results from a lesion or disease of the somatosensory nervous system comprising of peripheral sensory neurons as well as the spinal cord and the brain.¹⁰ Enhanced excitability, or sensitization, throughout this pain axis is thought to mediate neuropathic pain. In the central nervous system (CNS), changes in the response properties of complex neuronal and glial networks, such as disinhibition, contribute to persistent evoked and spontaneous pain.^11,12 In the periphery, however, specialized cells, known as nociceptors, are responsible for transducing painful stimuli. These cells typically have a high stimulation threshold allowing them to fire action potentials when exposed to noxious heat, chemicals, and mechanical stimulation.^13,14 In concert with other sensory neurons, nociceptors inform the CNS about the nature, location, and intensity of the painful stimulus.¹⁵ Enhancement of the response properties of sensory neurons, known as peripheral sensitization, can arise and is response to neuropathy caused by tissue injury and inflammation or lesions to the nerve fibers.^16,17 Sensitization leads to hyperalgesia (increased sensitivity to painful stimulus) and allodynia (pain from a previously non-noxious stimulus).¹⁸

Major efforts have been made to understand the mechanisms that underlie the sensitization of nociceptors.^15,18,19 Some molecules that might be held responsible include reactive oxygen species (ROS), adenosine triphosphate (ATP), and calcium (Ca²⁺).²⁰ Besides providing energy for many cellular processes, ATP acts as a nociceptive transmitter via the purinergic receptors, sensitizing nociceptors to further stimulation.²¹ ROS such as superoxide (SO) are chemically reactive molecules that are produced as a by-product of the chemical processes in the mitochondria, lysosomes, and the endoplasmic reticulum (ER). High levels of ROS during times of stress lead to oxidation of nucleic acids, lipids, and proteins in a process known as oxidative stress.²² SO and its descendant peroxynitrite have been linked to alterations in protein kinases, glutamatergic signaling, and ion channel modulation in the development of hypersensitivity.²³ Ca²⁺ is a multifunctional ion that is involved in neuronal plasticity by regulating neurotransmitter release, action potential propagation, and changes in gene expression.²⁴ The role of plasmalemmal Ca²⁺ signaling through voltage-gated Ca²⁺ channels and transient receptor potential (TRP) channels is well-documented in pain pathophysiology.²⁵ Ca²⁺-signaling also results in increased production of both ROS and ATP by mitochondria, and thus may serve as a master regulator between these molecules and the development of nociceptor sensitization. However, the precise mechanisms that regulate intracellular Ca²⁺ (iCa²⁺) stores to mediate pain remain to be fully elucidated.

In this regard, the ER and mitochondria are two organelles that act together as an important Ca²⁺ signaling hub and as such, they may represent a fundamental regulator of iCa²⁺, excitability and neuronal plasticity.²⁶ Indeed, emerging evidence is beginning to shed light on the critical role these two organelles play in nociceptor sensitization and chronic pain. Given their important position within the cell for Ca²⁺ regulation, this review will discuss the complex relationship between the ER and mitochondria and explore how dysregulation within and between these organelles can impact on nociceptor sensitization and the implications this has for chronic pain.

The endoplasmic reticulum

The ER is perhaps the largest tubular organelle in the cell.²⁷ In neurons, the ER is continuous throughout the cell, extending from the dendritic arbors to the cell soma, across the axon and into the presynaptic terminals.²⁸ The ER has classically been subdivided into the rough ER (containing ribosomes), the smooth ER (without ribosomes) and the nuclear envelope.²⁷ The ER performs a variety of functions. Most notably, a host of luminal and transmembrane chaperones and foldases help synthesize and mature proteins in the ER.²⁹ In addition, the ER is also important for lipid synthesis. As such, the smooth ER synthesizes phospholipids, glycolipids, and cholesterol which are used for various functions such as synthesizing steroid hormones from cholesterol.³⁰ The ER also contains many detoxifying enzymes that metabolize lipid soluble molecules by converting them into polar, water-soluble compounds that are then secreted from the cell and the body.³¹ In addition, the ER is the cell’s largest dynamic repository of Ca²⁺.²⁶ Many of the ER’s chaperones (e.g., BiP, calreticulin, calnexin) are low-affinity, high-capacity Ca²⁺ binding proteins (CBPs).²⁶ Their function as chaperones also heavily depends on the luminal concentration of Ca²⁺.³²

Since the ER extends throughout the cell, it also serves as a highway for RNAs, proteins, lipids, and ions.²⁷ Changes in the intraluminal environment of the ER, such as increased protein synthesis and altered Ca²⁺ homeostasis, may lead to ER stress which is initially protective but detrimental in the long-term.³³

The unfolded protein response

Stressful conditions such as amino acid deprivation, viral infection, heme deficiency, and accumulation of unfolded proteins lead to the induction of a larger translational regulation pathway known as the integrated stress response (ISR). The ISR is mediated by the stimulation of four kinases, general control nonderepressible 2 kinase (GCN2), protein kinase R (PKR), heme-regulated eIF2α kinase (HRI), and protein kinase RNA-like ER kinase (PERK), respectively.³⁴ These kinases converge onto the phosphorylation of eukaryotic initiation factor 2 α (eIF2α) regulating its translational capacity and allowing the cell to aptly react to the stressor.³⁴

Protein synthesis, folding, and quality control are canonical functions of the ER. Upon increased protein demand, the ER responds to an accumulation of unfolded or misfolded proteins by initiating a series of cellular processes collectively known as the unfolded protein response (UPR) (Figure 1).³⁴ The goal of the UPR is to reduce translation of proteins, increase the expression of chaperones, and degrade misfolded proteins. This response is mediated by three downstream transducers: PERK, inositol-requiring enzyme 1α (IRElα), and activating transcription factor 6 (ATF6).³⁴ Each of these ER resident proteins contains a single transmembrane domain and a single luminal domain.³⁵ Under stressed conditions, binding immunoglobulin protein (BiP; also known as heat shock protein A 5, HSPA5, and glucose response protein 78, GRP78) dissociates from ATF6, PERK, and IRE1α activating their respective arm of the UPR.³⁴ BiP, then, actively folds unfolded proteins using its allosteric ATPase activity and prevents the formation of protein aggregates by binding and restraining unfolded proteins.³⁶

Figure 1.

The unfolded protein response of the ER. The dissociation of BiP from PERK and IRE1 induces oligomerization and autophosphorylation of these kinases, initiating the unfolded protein response. Activated IRE1 acts as an endoribonuclease and splices XBP1 mRNA (XBP1s) generating a potent transcription factor that enhances the production of ER chaperones and genes associated with ER-associated degradation. PERK activation enhances phosphorylation of eIF2α as part of the integrated stress response. This results in the suppression of global translation, and expression of stress-induced genes like ATF4, CHOP, and GADD34. In concert with protein phosphatase 1, GADD34 functions to reverse eIF2α phosphorylation and resume global translation. Many of ER’s chaperones, including BiP, require Ca²⁺ for proper functioning. SERCA actively pumps Ca²⁺ into the ER while the activation of IP3Rs and RyRs releases Ca²⁺ into the cytosol. Stressors, like inflammation, may modulate the activity of ER Ca²⁺ transporters and thus influence ER functioning.

Once BiP is sequestered from PERK’s luminal domain, PERK autophosphorylates and oligomerizes as either a dimer or a transient tetramer.³⁷ The extent of oligomerization of PERK may represent the degree of unfolded proteins in the ER since the tetramer configuration is a more potent kinase than the dimer.³⁴ Upon activation, PERK phosphorylates eukaryotic initiation factor 2 α (eIF2α) reducing the general translation of proteins while enhancing translation of select mRNAs. In particular, PERK-eIF2α signaling increases translation of those transcripts that encode an upstream open reading frame (uORF).³⁸ eIF2α phosphorylation enhances the expression of activating transcription factor 4 (ATF4) which in turn increases the expression of the proapoptotic transcription factor C/EBP homologous protein (CHOP).³⁹ ER stress and the UPR have been linked to several neuropathic pain states, including peripheral nerve injury (PNI), diabetic neuropathy, chemotherapy-induce neuropathic pain. Furthermore, inhibiting certain arms of the ER stress pathway attenuates pain hypersensitivity.^40,41 Mechanisms supporting these observations will be discussed below.

Mitochondria

Mitochondria are unique organelles thought to have originated from bacteria that were endocytosed by an early eukaryote. They have their own DNA (mtDNA) that encodes for proteins which allow for efficient aerobic energy production.⁴² They are composed of both an outer mitochondrial membrane (OMM) and inner mitochondrial membrane (IMM). The IMM encloses the matrix, and has many folds, referred to as cristae. Cristae increase the surface area available to harbor the five transmembrane proteins involved in energy production by oxidative phosphorylation (Complex I–V).⁴³ These transmembrane proteins pump hydrogen ions into the intermembrane space, creating a potential difference across the IMM (ΔΨm) that is used by the protein ATP synthase (Complex V) to make ATP. The ΔΨm also provides an electrochemical driving force for the movement of ions such as Ca²⁺ across the IMM.⁴⁴ An important byproduct of mitochondria metabolism is the generation of ROS, which are important signaling molecules, but under pathological conditions can lead to oxidative stress and exacerbated damage.⁴⁵

Mitochondria form dynamic and integrated tubular networks, regulated by proteins such as MFN1 and MFN2, which promote mitochondrial fusion, and by Fis1 and Drp1 which regulate fission.^46,47 These networks not only allow for the coordination of ATP production spatially and temporally but also play a role in regulating mitochondria–cell communication involved in metabolism, the cell cycle, stress responses, and cell death cascades.⁴⁶ Although the mechanism of how mitochondrial networks could mediate these processes is still under investigation, there are numerous hypotheses. For example, mitochondrial networks may control energy production through Ca²⁺ signaling. In a fused mitochondrial network, Ca²⁺ can diffuse throughout the network thus allowing small changes in Ca²⁺ to have a greater affect and reach more sites of energy production.^48,49

Mitochondrial-associated membrane

Mitochondria interact with the ER through a specialized domain called the mitochondria-associated ER membrane (MAM). The MAM is composed of contact sites tethered by protein complexes, which facilitate ER–mitochondria crosstalk and juxtaposition (as reviewed by Veeresh et al.⁵⁰ and Vance⁵¹). The ER–mitochondria encounter structures (ERMES) complex regulates phospholipid transfer as well as mitochondrial dynamics and mitophagy.⁵² Phosphofurin acidic sorting cluster 2 (PACS2) regulates communication between the ER and mitochondria, ER homeostasis, and apoptosis.⁵³ Presenilin 2 (PS2) in the presence of mitochondrial fusion protein mitofusin (MFN2) promotes ER–mitochondrial tethering.⁵⁴ The IP3R-S1R-GRP75-VDAC complex is crucial for Ca²⁺ transfer from the ER to mitochondria. The inositol triphosphate receptor (IP3) receptor in the ER is stabilized by the sigma1 receptor (S1R) to allow Ca²⁺ efflux. This complex is tethered by the chaperone glucose response protein 75 (GRP75) to the voltage-dependent anion channel (VDAC) to facilitate mitochondrial uptake of Ca²⁺.^50,55 In addition, vesicle-associated membrane protein-associated protein B (VAPB) in the ER associates with protein tyrosine phosphatase-interacting protein 51 (PTPIP51) in the mitochondria to help facilitate Ca²⁺ transfer.⁵⁶

In neurons, ER–mitochondrial interactions through the MAM can significantly influence many key processes relevant to nociceptor function such as mitochondrial bioenergetics, Ca²⁺ homeostasis, and action potential generation.^55,57,58 For example, Ca²⁺ regulates activity-dependent signaling and neuronal excitability, and in nociceptors this regulation is important to prevent aberrant signaling and pain.⁵⁹ Thus, Ca²⁺ modulation at the MAM can play a role in pain control through nociceptor excitability.

ER stress results in various functional and structural changes at the MAM.⁶⁰ ER stress can lead to increased MAM contacts and strengthened Ca²⁺ transfer to mitochondria, which boosts energy production.⁶¹ Interestingly, some proteins known to be involved in the UPR, IRE1, and PERK perform noncanonical functions at the MAM.^62,63 IRE1 localized at the MAM regulates IP3 receptors and maintains mitochondrial bioenergetics.⁶² Whereas PERK at the MAM regulates juxtaposition between the ER and mitochondria and promotes apoptosis by conveying ROS signals and maintaining high levels of CHOP.⁶³

Sigma1 receptors (S1Rs) are also important in the development of neuropathic pain. In models of neuropathic pain, hypersensitivity can be reduced using S1R knockouts and pharmacological S1R antagonists.^64–66 In addition to their role at the MAM, S1Rs can translocate to both the plasma membrane and nucleus to modulate voltage-gated channel activity.⁶⁷ These diverse functions present multiple pathways through which S1Rs are be involved in neuropathic pain. The S1R antagonist MR309 reduced chemotherapy-induced neuropathic pain in clinical trials, demonstrating its promise as a therapeutic target in pain.⁶⁸

ER Ca²⁺ regulation and the impact on pain

Voltage gated Ca²⁺ channels, TRP channels, glutamatergic channels, and purinergic receptors on the plasma membrane regulate the influx of extracellular Ca²⁺ in sensory neurons and are heavily involved in pain pathogenesis following disease, injury, and inflammation (for an extensive review, see Bourinet et al.²⁵). Plasma membrane Ca²⁺-ATPase (PMCA) and Na⁺/Ca²⁺ exchanger (NCX) are primarily responsible for extruding cytosolic Ca²⁺.⁶⁹ The bulk of the intracellular Ca²⁺ is stored in the ER.²⁷ Since Ca²⁺ is a multifunctional ion involved in a variety cellular processes, the ER can directly impact gene expression, bioenergetics, ion channel physiology, and enzyme function.²⁷ With the presence of Ca²⁺ transporters on the ER membrane, the ER can be excited to release or sequester Ca²⁺, tightly regulating the concentration of available cytosolic Ca²⁺. In the ER, Ca²⁺ is stored as bound to CBPs, many of which are chaperones, like calnexin, calreticulin, and BiP.^26,27 By and large, each CBP binds to many Ca²⁺ molecules in a low-affinity, high-capacity manner such that the opening of a Ca²⁺ exporters allows for the immediate dissociation of Ca²⁺ from the CBPs.²⁶ The ER also maintains a concentration of free Ca²⁺ unbound to CBPs. Free Ca²⁺ determines the driving force of Ca²⁺ release, while Ca²⁺ bound to CBPs determines the total amount of Ca²⁺ that can be released.^70,71 Ca²⁺ efflux from the ER is mediated by Ca²⁺-sensitive ryanodine receptors (RyRs) and inositol triphosphate receptors (IP3Rs), while the influx of Ca²⁺ into the ER is mediated by sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) pumps in concert with store-operated Ca²⁺ entry (SOCE) (Figure 1).

Ryanodine receptors and pain

RyRs are named after their antagonist, ryanodine, which was originally used as an insecticide. RyRs are one of the largest membrane proteins consisting of four-identical subunits arranged in a four-leaf clover-shaped structure.^72,73 There are three variants of RyRs (RyR1, RyR2, and RyR3) all of which can be activated by free Ca²⁺ in the cytosol in a process known as Ca²⁺-induced Ca²⁺ release or CICR.²⁷ RyR1 is the most sensitive to cytosolic Ca²⁺ followed by RyR2 and then RyR3^74,75 RyR3s were classically known as the “brain” isoform; however, all three isoforms can be detected in the nervous system spanning across different cell types and being expressed in virtually every cellular compartment.^26,76,77 RyRs are not exclusively selective for Ca²⁺. Instead, they also counter-conduct other cations such as Mg²⁺, Li⁺, Na⁺, and K⁺.⁷⁸ This allows the channel to release Ca²⁺ without largely affecting the membrane potential.⁷⁸ RyRs can associate with other proteins to further enhance a cellular Ca²⁺ response.⁷⁹Voltage-gated calcium channels and Ca²⁺-sensitive large-conductance K⁺ (BK) channels are functionally coupled to RyRs.^80,81 As a result, cytosolic Ca²⁺ increase due to neuronal activity can enhance Ca²⁺ release from the ER which in turn readily affects action potential dynamics via Ca²⁺ sensitive channels.⁸¹

A growing number of diseases and conditions are linked to impaired RyR function such as genetic RyR mutations in malignant hyperthermia, central core disease, and catecholaminergic polymorphic ventricular tachycardia.⁸² Aberrant RyR function has also been implicated in neurodegenerative disorders like Alzheimer’s disease.⁸³ Recently, pain and hyperalgesic priming have been linked to RyR sensitization in rats. A single dose of a priming stimulus, like ryanodine, induces long-lasting changes in the nervous system such that a second pain-inducing stimulus, like prostaglandin E2, provokes a much more profound algesic response than would be observed without the primer. In female mice, a low dose (1 pg) of ryanodine is sufficient to induce priming compared to a much higher dose of 100 ng in males.⁸⁴ This heightened sensitivity to ryanodine in females is mediated by estrogen receptor α (ERα) such that antagonizing the ERα using antisense oligonucleotides abolishes the priming phenotype. Activating ERα using an agonist, propyl pyrazole triol, promotes priming in the absence of ryanodine.⁸⁴

IP3Rs and pain

IP3Rs are another class of ER Ca²⁺ release channel. IP3Rs respond to inositol triphosphate (IP3) which is catalyzed by phospholipase C (PLC) typically after the activation of plasmalemmal G-protein coupled receptors (GPCRs) such as metabotropic glutamate receptors (Figure 3).²⁶ Like the RyRs, there are three types of IP3Rs: IP3R1, IP3R2, and IP3R3. The channel may form as homo- or hetero-tetramers.²⁶ IP3Rs can be further modulated by (de)phosphorylation by various kinases, like protein kinase A, B, and C, and phosphatases like calcineurin.⁸⁵ Further splice variations in the IP3R genes allow for even more functional diversity of the channel despite 60%–80% shared homology among IP3R isotypes.²⁶ IP3Rs open when both IP3 and Ca²⁺ bind to the receptor.⁸⁶ It is believed that IP3 primes the receptor to cytosolic Ca²⁺ which allows for the opening of the channel in a Ca²⁺-dependent manner.⁸⁶ IP3R1s are maximally activated around 300–400 nM of Ca²⁺ beyond which they lose their sensitivity to intracellular Ca²⁺.⁸⁷ In contrast, IP3R2s and IP3R3s are not inhibited by high Ca²⁺ concentration.⁸⁸ Activation of the IP3Rs may lead to widespread Ca²⁺ waves or be confined to microdomains depending on the sensitivity of IP3Rs to IP3, the density of IP3Rs, degradation of IP3, and the cell’s buffering capacity for IP3.²⁷ Since IP3Rs are co-activated by intracellular Ca²⁺, RyR-mediated CICR enhances Ca²⁺ release from IP3Rs (and vice versa) possibly contributing to widespread Ca²⁺ waves in the ER.^27,86

Figure 3.

Mitochondria both regulate, and are regulated by Ca²⁺. Ca²⁺ enters the outer mitochondrial membrane through the VDAC, then subsequently is transported into the mitochondial matrix via the MCU. Ca²⁺ homeostasis within the mitochondrial matrix is maintained by extrusion via the NCLX, which exchanges three Na⁺ for one Ca²⁺. However, mitochondria can also buffer excess Ca²⁺ by precipitating it as calcium-phosphate. Ca²⁺ can also regulate various mitochondrial processes such as fission and fusion, ATP production, ROS production, and mPTP opening. Reducing ATP and increasing ROS can cause widespread dysfunctions within the cell. Opening of the mPTP depolarizes the mitochondrial membrane potential as it allows Ca²⁺ efflux, and it can also trigger apoptosis by release of cytochrome C.

The coupling of IP3 production with stimulation of GPCRs allows the cell to regulate intracellular Ca²⁺ release in response to the external environment (Figure 3). As such, the binding of bradykinin, an inflammatory mediator, to its B2 receptor induces IP3 production, enhances Ca²⁺ release from the ER, activates Ca²⁺-sensitive calmodulin, and contributes to neuronal sensitization.^89,90 The link between bradykinin and inflammatory pain has been well-documented.^90,91Fractalkine-induced inflammation and subsequent hyperalgesia was recently shown to be mediated by IP3-induced Ca²⁺ transients in microglia.⁹² Like ryanodine, a low dose (1–3 pg) of IP3 induces hyperalgesic priming in female, but not male, rats in a ERα dependent manner.⁹³ Interestingly, both ryanodine-induced and IP3-induced hyperalgesic priming can be blocked by inhibiting either IP3Rs or RyRs, respectively.⁹³ This observation in females implies that Ca²⁺ release from the ER intricately links RyRs and IP3Rs, possibly in a CICR manner that is sex specific.

SERCA pumps and pain

Cytosolic Ca²⁺ concentrations are tightly regulated by SERCA pumps. These channels belong to a class of P-type Ca²⁺ ATPases that respond to increases in intracellular Ca²⁺ levels by actively sequestering the ion from the cytosol into the ER (Figure 1).²⁶ Like other ER Ca²⁺ transporters, there are three paralogs of SERCA: SERCA1, SERCA2, and SERCA3. SERCA2 is further differentiated by its spliced isoforms, SERCA2a and SERCA2b. Of particular interest is SERCA2b which is ubiquitously expressed in nervous tissue.²⁷ SERCA2b activity is modulated by Ca²⁺-sensitive calreticulin and ER resident protein 57 (ERp57), both of which also function as chaperones.^94,95 When Ca²⁺ decreases in the ER lumen, calreticulin binds to SERCA2b increasing its activity, while ERp57 functions as a thiol oxidoreductase, catalyzing disulfide bridges on the SERCA channel and consequently enhancing Ca²⁺ uptake. Hence, luminal Ca²⁺ concentration inversely determines the rate of Ca²⁺ uptake by SERCA such that channel activity peaks at low luminal Ca²⁺ concentrations and wanes at higher concentrations.²⁷ With a drastic increase in cytosolic Ca²⁺ (e.g., during an action potential), other transporters with a greater capacity for Ca²⁺, such as mitochondrial calcium uniporter (MCU) and Na⁺/Ca²⁺ exchanger (NCX), can further support Ca²⁺ homeostasis.²⁶

A loss of SERCA function has been linked to neuropathic pain after PNI. Reduced SERCA function in response to PNI leads to a slower recovery of Ca²⁺ transients in small diameter nociceptors.⁹⁶ The inability of the ER to buffer increases in cytosolic Ca²⁺ after injury leads to Ca²⁺-signaling-induced hyperexcitability of these afferents and neuropathic pain.^20,97 This is further exemplified in axons of sensory neurons obtained from diabetic rats where a loss of SERCA activity contributes to diminished Ca²⁺ levels in intracellular stores.⁹⁸Thrombospondin-4, an extracellular matrix glycoprotein commonly elevated after nerve injury, was shown to suppress SERCA activity, enhance PMCA activity, drain ER Ca²⁺, and stimulate store-operated Ca²⁺-entry⁹⁹ (Figure 3). The net result of this is a loss of intracellular Ca²⁺ levels. Similar observations have been made following axotomy.¹⁰⁰

Store-operated Ca²⁺ entry and pain

During a Ca²⁺ transient, Ca²⁺ enters the cell through the plasma membrane and is also released from the ER. To replenish this loss of ER Ca²⁺ pool, store operated Ca²⁺ channels (SOCCs), particularly the Orai channels, pump extracellular Ca²⁺ into the cytosol and into the ER via the SERCA channels.^101,102 The function of these SOCCs is dependent on the intraluminal ER Ca²⁺ concentration.¹⁰³ Stromal interaction molecule (STIM) proteins were identified as Ca²⁺ sensors in the ER.¹⁰⁴ There are two homologs of STIM proteins, STIM1 and STIM2 in vertebrates that are expressed in a variety of tissues and organs including nervous tissue.¹⁰⁵ These single-pass transmembrane ER proteins are localized diffusely throughout the ER under resting conditions.¹⁰⁶ Upon emptying of the Ca²⁺ store, STIM proteins redistribute and cluster on the ER membrane nearest to the plasma membrane where their CRAC activation domain (CAD; also known as STIM-Orai activation region or SOAR) binds and activates Orai channels on the plasma membrane.^104,107 Orai channels also redistribute and cluster to ER–plasma membrane junction sites creating microdomains of store mediated Ca²⁺ entry (SOCE).^106,108 There are three human homologs of Orai channels, Orai1-3, each of which share considerable homology (∼62% overall identity) with each other but have very little overlap with other ion channels.¹⁰⁷ Orai channels are found ubiquitously in various tissues.^109–111 Of note, high Orai1 expression in the brain implicates its role in the nervous system.¹¹² Various mechanisms, such as glycosylation and pH sensing, and proteins, including golli proteins and septins, can further regulate SOCE.¹⁰³ Through the formation of ER:plasma membrane microdomains, the ER—an intracellular Ca²⁺ store—becomes functionally linked to the extracellular supply of Ca²⁺. An extensive review of SOCE and its regulatory mechanisms can be found elsewhere.^103,107

The role of SOCE in mediating disease and injury has only recently come to light primarily because the Orai channels were only identified and cloned at the turn of the century.^113–115 Since then, SOCCs have been implicated in various disease states including autoimmunity,¹¹⁶ neurodegeneration,¹¹⁷ arthritis,¹¹⁸ and tumorigenesis.¹¹⁹ Only recently has the role of SOCCs been established in pain pathophysiology. It was noticed that intraplantar, intraperitoneal, and intrathecal administration of YM-58483, an inhibitor of SOCE, prevented both acute and chronic mechanical and thermal hypersensitivity in a dose-dependent manner following a SNI and CFA-administration as well as reducing both phases of the formalin response.¹²⁰ Similarly, YM-58483 administration in a model of rheumatoid arthritis was found to reduce cytokine production in the periphery and ERK and CaMKIIα expression in the spinal cord, resulting in attenuated pain hypersensitivity.¹¹⁸ Although these reports demonstrate the analgesic effects of inhibiting SOCE, the molecular underpinning of this phenotype remained unknown until recently. In a series of experiments employing Orai1 knock-out animals, Dou et al.¹²¹ showed that Orai1 contributed to neuronal hyperexcitability in the spinal cord and pain behaviors following formalin and carrageenan administration. In the periphery, knockdown of STIM1, STIM2, Orai1, and Orai3 were found to diminish SOCE in DRG neurons.¹²² Thapsigargin, a SERCA blocker that depletes ER Ca²⁺ and thereby induces SOCE, promoted hyperexcitability of sensory neurons in vitro. This phenotype was abolished with Orai1 and Orai3 double knockdown.¹²² These observations clearly implicate SOCE in regulating pain pathophysiology. The sequence of events that lead to the imbalance of intracellular Ca²⁺, however, remains unclear. Is the ER Ca²⁺ dysregulated first and that is followed by increased SOCE or is the SOCE impaired that then induces ER Ca²⁺ release? The answer is probably dependent on the disease or injury. In either case, a vicious cycle of Ca²⁺ dyshomeostasis becomes detrimental to the cell.

Mitochondrial Ca²⁺ regulation and the impact on pain

Mitochondria were not considered to be major Ca²⁺ regulators under physiological conditions until fluorescent Ca²⁺ dyes made it possible to measure Ca²⁺ in organelles of live cells.^123,124 Although they store less Ca²⁺ than the ER, mitochondria are nonetheless important buffers of intracellular Ca²⁺.^125,126 Ca²⁺ influx into the mitochondrial matrix occurs through the MCU located on the IMM, which has high Ca²⁺ selectivity, but low affinity.¹²⁷ The MAM overcomes this barrier by creating microdomains of high Ca²⁺ at the IP3R-S1R-GRP75-VDAC complex (Figure 2).^50,128 In this complex, VDACs on the OMM of mitochondria are tethered by the chaperone GRP75 to IP3Rs in the ER. Thus, when IP3Rs open, and are stabilized by S1Rs, there can be efficient transport of Ca²⁺ into the mitochondrial matrix by the MCU.^50,127,128 The VAPB-PTPIP51 complex also facilitates Ca²⁺ uptake by tethering the MAM.^56,129

Figure 2.

The ER and the mitochondria are connected by Ca²⁺. Extracellular Ca²⁺ commonly enters sensory neurons via voltage-gated Ca²⁺ channels, store-operated Ca²⁺ channels (ORAI1/3), and transient receptor potential channels upon stimulation. This increase in cytosolic Ca²⁺ is rapidly neutralized by the SERCA pump and stored in the ER. Various CBPs including chaperones, like BiP, buffer ER Ca²⁺ levels. The ER and mitochondria are connected by the MAM, which facilitates Ca²⁺ transfer from the ER through IP3 receptors, to the mitochondria through the VDAC in a complex tethered by the chaperone GRP75. This exchange is regulated by the MAM complex VAPB-PTPIP51. Ca²⁺ subsequently enters the mitochondrial matrix through the MCU and is extruded by the NCLX. Ca²⁺ can also exit the mitochondria through the mPTP under pathological conditions. Ca²⁺ dyshomeostasis along this axis can not only disrupt ER and mitochondrial function but also affect iCa²⁺ and cell excitability, contributing to pain hypersensitivity.

Ca²⁺ regulates various mitochondrial processes (Figure 3) including ATP and ROS production, fission and fusion, opening of the mitochondrial permeability transition pore (mPTP), and mitochondria-mediated apoptosis.^24,130 For example, Ca²⁺ can stimulate ATP production via activation of Krebs cycle dehydrogenases.¹³¹ ROS and Ca²⁺ have a complex relationship. Ca²⁺ signaling can stimulate of ROS-generating enzymes and upregulate ROS production during mitochondrial respiration.¹³² In addition, ROS can influence Ca²⁺-dependent processes. Oxidation of BK channels leads to a loss of Ca²⁺-sensing ability and thus a loss of activation.¹³³ Since BK channels prevent repetitive firing, loss of their function can lead to nociceptor hyperexcitability and pain.¹³⁴

Normally, mitochondrial Ca²⁺ concentration is maintained steadily by a balance between influx through MCUs and efflux through the mitochondrial sodium-calcium exchanger (NCLX).^126,135,136 However, mitochondria can also buffer Ca²⁺ elevations in the cytosol by uptake and precipitation as Ca²⁺-phosphate.^24,137 In DRG neurons, both the MCU and NCLX are critical to this process.¹³⁸ Knocking down the NCLX can reduce both MCU Ca²⁺ uptake, and TRPV1 receptor-mediated Ca²⁺ currents in nociceptors. Thus, mitochondrial Ca²⁺ buffering is important for regulating intracellular Ca²⁺ homeostasis. This capacity to buffer intracellular Ca²⁺ can significantly influence Ca²⁺ transients in nociceptors, the overall level of nociceptor excitability, and ultimately, pain.¹³⁹

Pathophysiology impacting the ER and mitochondria

ER stress in PNI

Along with various other homeostatic processes, ER stress has also been implicated in the pathophysiology of PNIs (Table 1). Damage to peripheral nerves induces a robust UPR in axons and soma of sensory neurons in the DRG (reviewed by Valenzuela et al.¹⁴⁰) Interestingly, ER chaperones, like BiP, and transcription factors, like CHOP, are elevated at the site of injury following sciatic nerve crush and are retrogradely transported back to the soma as markers of injury.¹⁴¹ It is unlikely that UPR by-products from injured axons are the only source of ER stress markers observed in the DRG considering that ER stress markers are elevated in the DRG as early as one day after injury.¹⁴² Regardless, axonal ER stress is proposed to be beneficial and essential for nerve regeneration and locomotor recovery possibly by addressing the high demand of protein synthesis required for axonal outgrowth. In this respect, the upregulation of ERp57, an ER resident disulfide isomerase, was found to improve axonal regeneration and locomotion following sciatic nerve injury.¹⁴³ Furthermore, regeneration and recovery after sciatic nerve crush depended on the activation of the more protective IRE1α-XBP1 axis as opposed to the more detrimental PERK-ATF4 pathway.⁴¹ Moreover, axotomy-induced Ca²⁺ release from the ER promotes IRE1 and PERK phosphorylation.¹⁴⁴ Inhibiting PERK and IRE1 signaling using GSK2606414 and 4µ8c, respectively, collapses the growth cone, fragments axonal ER, and impairs axonal regeneration. Interestingly, axotomy-induced PERK activation does not increase CHOP levels in the axon, although increased phosphorylation of eIF2α is observed.¹⁴⁴ Combined with increased splicing of XBP1 and BiP expression in the axons following axotomy, this observation further suggests that enhancing the ER folding capacity and reducing ER load after a PNI enhances the regenerative capacity of the nerve. However, which cytoprotective proteins are specifically folded following nerve injury remains to be investigated and will require high-throughput translatomic and proteomic screening.

Table 1.

Overview of ER pathologies.

Disease	Reference	Findings
Peripheral nerve injury pain	Castillo et al.143	Overexpressing ERp57 improved nerve regeneration and locomotion after sciatic nerve crush.
	Ge et al.148	Inhibiting PDI, a chaperone, in SDH after SNL reduced ER stress and mechanical pain.
	Liu et al.147	Intrathecal 4-PBA administration alleviated mechanical and thermal hypersensitivity in SNL rats.
	Melemedjian et al.145	SNL induced phosphorylation of eIF2α in the sciatic nerve
	Ohtake et al.144	Nerve regeneration following axotomy requires Ca2+-dependent activation of IRE1 and PERK signaling.
	Onate et al.41	Nerve regeneration and motor recovery was associated with IRE1-XBP1 signaling, but not PERK-ATF4 pathway, in a nerve crush model.
	Yamaguchi et al.142	ER stress markers are elevated following SNL. Salubrinal, an eIF2α phosphatase inhibitor, reduced pain hypersensitivity.
	Ying et al.141	Mild ER stress enhanced nerve regeneration after crush injury. UPR by-products are retrogradely transported from the nerve to the DRG.
	Zhang et al.146	SNL induced BiP and ATF6 in SDH neurons and PERK signaling in astrocytes. ATF6 knockdown in the SDH reduced BiP levels and pain behaviours in SNL rats.
Diabetic neuropathy	Barragan-Iglesias et al.165	MGO triggers ISR in IB4+ sensory neurons and produces mechanical hypersensitivity. Treatment with ISRIB resolves eIF2α phosphorylation and pain in MGO administered and diabetic mice. MGO enhanced Ca2+ responses of sensory neurons.
	Griggs et al.163,164	MGO sensitizes lamina II SDH neurons of db/db mice via the TRPA1-AC1-Epac pathway. Spinal neurons demonstrate enhanced Ca2+ responses upon stimulation.
	Inceoglu et al.160	STZ-induced diabetes robustly generates ER stress in the sciatic nerve and glabrous skin. 4-PBA treatment attenuated mechanical hypersensitivity in a time- and dose-dependent manner.
	Lupachyk et al.159	4-PBA and TMAO attenuated diabetes-induced ER stress, nerve dysfunction, fiber loss, and oxidative stress in the sciatic nerve and the spinal cord. CHOP−/− mice were less susceptible to diabetic peripheral neuropathy.
	Zherebitskaya et al.98	Hypoalgesia associated with chronic STZ-induced diabetes is a result of reduced ER Ca2+ loading.
Chemotherapy-induced neuropathic pain	Andoh et al.40,185	Ptx treatment promoted CHOP expression in the sciatic nerve, enhanced excitability of SDH neurons, and produced tactile hypersensitivity.
	Megat et al.182	Ptx enhances excitability of sensory neurons
	Resham and Sharma184	Wnt signaling inhibitors normalized sciatic nerve function, reduced axonal BiP expression, and attenuated pain hypersensitivity in Ptx treated rats.
	Tanimukai et al.186,187	In vitro, Ptx induced CHOP and caspase 3 expression in SK-N-SH neuroblastoma cells. Inducing sigma 1 receptor and BiP reduced ER stress and was cytoprotective.

Only recently has ER stress been linked to pain in PNI models. Melemedjian et al.¹⁴⁵ was one of the first studies to correlate increases in eIF2α phosphorylation in injured nerves of rodents following SNL; however, no mechanistic role of the ISR in inducing mechanical pain was investigated. Another study found elevated BiP, XBP1, p-eIF2α, and ATF6 expression in the dorsal horn of the spinal cord two weeks after SNL surgery which correlated with microglial activation and mechanical hypersensitivity in these animals.¹⁴⁶ Furthermore, knockdown of ATF6 using an intrathecal siRNA reduced neuronal BiP expression and partially reversed mechanical hypersensitivity. Intrathecal treatment of chemical chaperones, 4-PBA and tauroursodeoxycholic acid (TUDCA), was also found to partially reverse nociceptive behaviours after SNL.^147,148 In the periphery, Yamaguchi et al.¹⁴² found that BiP, XBP1, and CHOP expression was elevated in the sensory neurons as early as one day post-SNL only to recover within a week. Interestingly, treatment with salubrinal, an eIF2α phosphatase inhibitor, reduced mechanical, thermal, and cold hypersensitivity¹⁴² suggesting that induction of the ISR may ameliorate pain possibly by reducing global translation, enhancing ER folding capacity, and thereby dampening ER load.

As an oxidoreductase and a chaperone, protein disulfide isomerase (PDI) reduces disulfide bonds and produces H2O2 as a by-product. Along with other chaperones, SNL induces PDI expression in the spinal cord. Inhibiting PDI using bacitracin, an antibiotic, reduces ER stress, mitochondrial SOD levels, and mechanical hypersensitivity associated with nerve injury.¹⁴⁸ This suggests that SNL-induced pain may be a result of ER–mitochondria cross-talk.

Mitochondria in PNI

Mitochondrial dysfunction is evident in both peripheral and central tissues in PNI models (Table 2). Mitochondria after PNI exhibit fragmentation, depolarized membrane potentials, reduced ATP synthesis, and oxidative stress.¹⁴⁹ Mitochondrial dysfunctions can trigger apoptosis and axonal degeneration in the periphery and spinal cord, leading to pain.^149–152 However, neuropathic pain in PNI is not exclusively caused by degeneration. Both peripheral and central sensitization contribute to pain maintenance.¹⁵³ Elevated Ca²⁺ signaling underlies nociceptor hyperexcitability in other neuropathic pain conditions.⁹⁷ Interestingly, mitochondrial dysfunction observed in PNI models may also contribute to iCa²⁺ dyshomeostasis, but the relationship is likely complicated. Mitochondria both regulate and are regulated by Ca²⁺,²⁴ so, while Ca²⁺ elevations can overwhelm mitochondrial buffering and lead to dysfunction, mitochondrial dysfunction can also be a starting point for pathological iCa²⁺ elevation. This complex association presents a significant target for future study to probe into how mitochondrial dysfunction might contribute to PNI-induced neuropathic pain.

Table 2.

Overview of mitochondrial pathologies.

Disease	Study	Finding
Peripheral nerve injury	Bae et al.197	Mitochondrial superoxide in the spinal dorsal horn of mice after PNI contributes to sensitization.
	Englezou et al.152	DRG neurons exposed to H2O2 in vitro display mitochondrial fission and eventual mitochondria-mediated apoptosis, which may give insight into how peripheral neurons are lost after PNI.
	Graham et al.198	PNI in mice results in increased mitochondrial fission and mitophagic protein expression in whole muscle tissue.
	Kim et al.151	Inhibiting mPTP opening in PNI with cyclosporin A attenuates allodynia and spinal cord cytochrome c release.
	Wang et al.150	Pro-BDNF induces mitochondria-mediated apoptosis of neurons and satellite glia in vitro
Diabetic neuropathy	Hall et al.199	DRG neurons from diabetic rats showed increased voltage-dependent Ca2+ currents.
	Kostyuk et al.175	Prolonged cytosolic Ca transients in cells from diabetic mice are dependent on mitochondrial dysfunction, and may contribute to neuropathic pain.
	Russell et al.200	Hyperglycemia induces ROS production and mitochondrial-mediated apoptosis in vitro.
	Srinivasan et al.168	DRG neurons from diabetic rats in vitro have increased apoptosis due to mitochondrial dysfunction.
	Vincent et al.170	An imbalance in mitochondrial biogenesis and fission results from hyperglycemia in DRG neurons in vitro and may contribute to diabetic neuropathy via neuronal injury.
Chemotherapy-induced neuropathic pain	Duggett et al.188	Rats treated with paclitaxel display energy dysfunction in the DRG that correlates with pain.
	Flatters and Bennett et al.176	In rats, paclitaxel-induced mitochondrial abnormalities may be responsible for pain, as their timelines correlate.
	Jia et al.189	Paclitaxel treatment in rats induces mitochondrial damage and production of reaction oxygen species that may activate the NLRP3 inflammasome and contribute to pain.
	Kidd et al.193	Paclitaxel-induced intracellular Ca2+ elevations can be inhibited with cyclosporin A, suggesting that paclitaxel may cause pain by inducing mPTP-opening and aberrant Ca release.
	Xiao and Bennett194	Spontaneous discharge of sensory neurons in rats treated with paclitaxel is likely a factor in paclitaxel-induced peripheral neuropathy.

ER stress in diabetic neuropathy

Diabetes mellitus (DM) is a metabolic disorder characterized by abnormally high blood glucose levels. The cause of DM is linked to either reduced insulin production (type 1) or a resistance to insulin (type 2).^154,155 Peripheral neuropathy and subsequent neuropathic pain is a common phenomenon in DM (Table 1).¹⁵⁶ Initially, patients with DM and preclinical rodent models of DM typically present with mechanical and thermal hypersensitivity.¹⁵⁷ Later disease stages are associated with hyposensitivity as a result of sensory fiber retractions from the skin.¹⁵⁸ Pain in DM follows a “glove and stocking” pattern in which distal limbs are affected first, followed by more proximal areas of the body.¹⁵⁶ This section will highlight mechanisms involved in neuropathic pain associated with DM, rather than loss of sensation, that have been linked to impaired ER function, Ca²⁺ dyshomeostasis, and mitochondrial damage.

Several studies have directly investigated the role of ER stress in mediating diabetes-induced pain (Table 1). ER stress in the streptozotocin (STZ) model of type 1 DM is characterized by increased phosphorylation of PERK, eIF2α, and IRE1α, as well as the cleavage of ATF6 in the sciatic nerve and the glabrous skin of the paw.^159,160 Other markers of ER stress such as BiP and spliced XBP1 are also known to be upregulated with DM.¹⁶⁰ The spinal cord is also known to increase CHOP and BiP expression in diabetic rats.¹⁵⁹ Alleviating UPR/ER stress using chemical chaperones, 4-phenylbutyric acid (4-PBA) and trimethylamine oxide (TMAO), reverses tactile allodynia and heat hypoalgesia in rats.^159,160 Denervation associated with late-stage DM is also reversed in rodents administered with 4-PBA and TMAO.¹⁵⁹STZ-induced DM in transgenic CHOP-deficient mice have reduced nerve fiber loss from the glabrous skin as compared to STZ-induced DM in wild-type mice.¹⁵⁹CHOP-knockout mice experience minimal thermal hypoalgesia as compared to wild-types in response to STZ-induced diabetes. However, CHOP^−/− does not confer protection against mechanical and tactile hypersensitivity.¹⁵⁹ Collectively, these studies suggest that diabetes-induced ER stress contributes to neuropathic pain.

Methylglyoxal (MGO), a common metabolite present in high concentrations (up to 5 µM) in the plasma of diabetic patients, is known to induce heat, mechanical, and affective pain in DM by activating peripheral and spinal TRPA1 channels.^161–166 Recently, the involvement of the UPR in MGO-induced pain hypersensitivity was postulated.¹⁶⁵ Intraplantar injection of MGO induces rapid and transient mechanical hypersensitivity in rats and mice that is associated with increased phosphorylation of PERK and eIF2α, particularly in the IB4+ nociceptors.¹⁶⁵ Preventing the phosphorylation of eIF2α using the ISR inhibitor, ISRIB, attenuated mechanical pain in both MGO-induced neuropathy and STZ-induced DM.¹⁶⁵ Systemic 4-PBA treatment was also able to ameliorate MGO-induced mechanical hypersensitivity.¹⁶⁵ Hence, MGO-associated ER stress and ISR in IB4+ nonpeptidergic nociceptors, may contribute to mechanical hypersensitivity observed in type 1 DM. Furthermore, acute MGO treatment in vitro enhances KCl-induced cytosolic Ca²⁺ transients in dissociated mouse DRG neurons.¹⁶⁵ These studies propose that rises in MGO in diabetes can enhance intracellular Ca²⁺ signaling in painful diabetic neuropathy.

In contrast, axons of sensory neurons from hypoalgesic type 1 STZ-induced diabetic rats have dampened Ca²⁺ responses to KCl and caffeine stimulation.⁹⁸ This suggests that Ca²⁺ stores in the ER are likely depleted in DM as a result of reduced SERCA-mediated Ca²⁺ uptake.⁹⁸ The discrepancy between these studies may stem predominantly from the differences in their experimental paradigms (acute vs. chronic), disease model (type 1 vs. type 2 DM), and the presence of hyperalgesia. Perhaps, increased Ca²⁺ signaling contributes to the initial hyperalgesia in DM-induced neuropathy, while at later stages, diabetic neuropathy is characterized by reduced intracellular and intra-ER Ca²⁺. Both, the increase in and the loss of intracellular Ca²⁺ contribute to detrimental processes within the cell including excitotoxicity,⁶⁹ ER stress,²⁶ mitochondrial dysfunction (see below), and cell death.⁶⁹

Mitochondria in diabetic neuropathy

Numerous studies have revealed mitochondrial dysfunction in models of DM (Table 2). This includes increased fission, ΔΨm depolarization, reduced ATP production, disrupted calcium homeostasis, and oxidative stress.^97,167–170 While mitochondria dysfunction plays a role in diabetic neuropathy, both the pathways that lead to this dysfunction, and the mechanism by which it induces neuropathy are contested. The two hallmark factors of type 1 DM and its relevant models are a lack of insulin, and resultant hyperglycemia. Whether only one or both of these factors is responsible for mitochondrial dysfunction in DM is questioned in the literature.^171–173 Several competing theories have been put forward to explain how mitochondrial dysfunction may be implicated in diabetic neuropathy. Hyperglycemia may increase mitochondrial fission, ROS production, and mitochondrial DNA damage, leading to neuropathy through apoptosis.^171,174 Although these mechanisms would explain sensory fiber retraction, this model does not link mitochondrial dysfunction to nociceptor hyperexcitability. Alternatively, impaired mitochondrial ATP production and reduced mitochondrial membrane potential could arise in response to reduced insulin receptor stimulation. This would significantly effect Ca²⁺ homeostasis and trigger ER stress, leading to changes in transcriptional and translational programs of the cell.¹⁷² Considering that increased intracellular Ca²⁺ has been linked to nociceptor hyper-excitability and neuropathic pain in diabetic neuropathy, impaired ER–mitochondrial Ca²⁺ homeostasis in diabetic neuropathy is a promising potential mechanism to explain the aberrant nociceptor activity and pain associated with the disease.^97,175

ER stress in chemotherapy-induced peripheral neuropathy

Neuropathy associated with chemotherapy agents is the single largest reason for reducing and halting treatment of cancer.¹⁷⁶ Approximately 68% of patients receiving chemotherapy will develop chemotherapy-induced peripheral neuropathy (CIPN) in their first month of treatment.¹⁷⁷ CIPN affects the nerves in a “glove and stocking” pattern where the longest axons are affected first, similar to DM-induced neuropathy.¹⁷⁸ Unlike DM-induced neuropathy, however, CIPN symptoms appear suddenly, affect multiple areas simultaneously, and progresses rapidly.¹⁷⁸ As a result of the CIPN, patients can experience acute and chronic sensory, motor, and autonomic dysfunction.¹⁷⁹ Sensory impairments typically involve mechanical, thermal, and cold pain hypersensitivity as well as paresthesias and dysesthesias.¹⁷⁹ Chronic use of chemotherapeutic agents may lead to permanent damage of the nerves, resulting in a loss of sensation.¹⁷⁸ Platinum derivatives (e.g., oxaliplatin), vinca alkaloids (e.g., vincristine), and taxanes (e.g., paclitaxel, Ptx) are commonly used chemotherapy agents known to cause CIPN and the associated pain phenotype.^178,180

Treatment with Ptx and oxaliplatin usually accompanies an acute pain syndrome three to four days after the start of the treatment as well as chronic neuropathic pain even after treatment has stopped.¹⁸¹ It is well established that chemotherapeutics contribute to neuronal injury and inflammation, especially in the peripheral nervous system (Table 1).¹⁸¹ Sensory neurons of the DRG from mice treated with paclitaxel fire spontaneously and become hyperexcitable, both of which are evidence of painful neuropathy.¹⁸² The protective and detrimental effects of ER stress and the ISR have previously been described in various cancer models (reviewed by Nam and Jeon¹⁸³) However, the effects of ER dysfunction in the context of pain in CIPN are not well understood. A recent report found that intraperitoneal Ptx (2 mg/kg) administration reduced sensory nerve conduction velocity and increased pain behaviors in rats.¹⁸⁴ Resham and Sharma¹⁸⁴ further observed that heat, cold, and mechanical hypersensitivity was correlated with increased expression of BiP in the sciatic nerves of Ptx treated animals.¹⁸⁴ Another study found that a single intraperitoneal injection of Ptx (5 mg/kg) was sufficient to induce CHOP expression in the sciatic nerve, increase the excitability of spinal dorsal horn neurons, and lead to mechanical allodynia in C57BL/6 NCr mice.^40,185 In vitro, Ptx treatment of SK-N-SH neuroblastoma cells increased the phosphorylation of eIF2α and induced the expression of CHOP as well as the pro-apoptotic caspase 3^186,187 Furthermore, pretreatment with fluvoxamine, a Sigma 1 receptor inducer, and BiP-inducer X (BIX) attenuated ER stress and protected the SK-N-SH cells from paclitaxel-associated cytotoxicity.¹⁸⁷ However, whether these drugs can mitigate pain in CIPN remains to be investigated.

Mitochondria in chemotherapy-induced peripheral neuropathy

Mitochondrial dysfunction is a well-characterized feature of chemotherapy-induced peripheral neuropathy (Table 2). The drug paclitaxel, for example, affects both mitochondrial structure and function in sensory neurons. Some mitochondrial changes linked to Ptx include swelling, vacuolation, decreased respiration, reduced ATP production, loss of ΔΨm, mitochondrial permeability transition pore (mPTP) formation, ROS production, and Ca²⁺ release.^{176,188–190} These dysfunctions can be accounted for by the theory that Ptx interacts with the mitochondrial membrane to induce mPTP-opening.¹⁹¹ Opening of the mPTP allows for aberrant passage of molecules into and out of the mitochondria, which can alter oxidative phosphorylation and ATP production, as well as trigger cell death cascades.¹⁹² In addition, in a nerve-injury model, treatment with the mPTP blocker cyclosporin A not only prevented mitochondrial dysfunction, but also reduced pain.^151,193 Taken together, this evidence suggests that Ptx-induced mPTP-opening and subsequent Ca²⁺ release may underlie nociceptor hyperexcitability and neuropathic pain.^176,194

Conclusions and future directions

There is a new appreciation for the roles of the ER and mitochondria in mediating chronic pain. What is emerging from these studies is a recognition of the complex interplay between these two organelles and the impact that disturbances in this interplay may have on Ca²⁺ signaling and ultimately, nociceptor excitability. Although the examples discussed here generally involve some form of axonal injury or trauma, emerging research also suggests that chronic pain associated with inflammatory diseases such as Multiple Sclerosis and rheumatoid arthritis may also be linked to ER and mitochondria dysfunction.^195,196 In all of these pathologies, it is clear that pain is a major overlapping symptom. A common thread linking pain to these conditions is stress at the ER and the ensuing mitochondrial dysfunction and Ca²⁺ dyshomeostasis. This has opened an exciting new avenue for research that is leading toward a more comprehensive understanding of the underlying mechanisms driving chronic, neuropathic pain in these states. A better understanding of how disturbed ER–mitochondrial interactions can lead to nociceptor hyperexcitability may provide insight into novel, therapeutic targets that can be manipulated to reverse the changes that have sensitized the system and thus restore the homeostatic state.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: MSY was supported by the Alexander Graham Bell Canada Graduate Scholarship from NSERC. This work was supported by a grant from the Canadian Institutes of Health Research (CIHR): FRN-162434.

ORCID iD

Muhammad Saad Yousuf https://orcid.org/0000-0003-0829-8182