Local translation in primary afferents and its contribution to pain

1. Introduction

Chronic pain is a complex but common problem. According to the Global Burden of Disease 2016, pain and pain-related diseases are the leading cause of disability and disease worldwide [64]. In 2019, approximately 20.4% (50 million) of individuals in the United States alone suffer from some sort of persistent pain [204] and is estimated that this results in $635 billion in lost productivity and healthcare expenditures [29,166]. A primary cause of this burden is the relative dearth of consistently effective therapeutic interventions for pain devoid of deleterious side effects. While the opioid crisis has highlighted the devastating consequences of some of these side effects, a systematic review of non-opioid therapeutic clinical trials found only modest effect sizes and a high number needed to treat for the majority of these alternative approaches [55].

More effective therapeutic strategies will ultimately depend on a better understanding of the mechanisms that underlie persistent pain and the transition from acute to chronic pain. In this context, the primary afferent has been a focus of research for a number of compelling reasons, not the least of which is evidence that blocking afferent activity is sufficient to eliminate ongoing pain and hypersensitivity for the vast majority of pain patients, including neuropathic pain [78], visceral pain [138], osteoarthritis [167,182], and painful temporomandibular joint (TMJ) disorder [109]. Although changes throughout the central nervous system (CNS) clearly contribute to the manifestation of acute and chronic pain, the majority of these changes appear to depend on aberrant afferent activity [72,74,101] underscoring the importance of primary afferents in the development of chronic pain. An advantage of targeting primary afferents as opposed to the CNS is that peripherally restricted interventions may limit side effect liability associated with CNS-penetrant drugs. However, while effective in the short term, blocking peripheral nerves is not a viable long term therapeutic strategy given the importance of normal afferent activity to survival. A growing body of evidence suggests that the specific mechanisms underlying aberrant afferent activity depends on a number variables including the type of injury [198], the site of injury [108], previous history [132], time after injury [73,172], sex [13,20,54], age [199], and genetics [41,160,209].

The list of potential mechanisms contributing to the initiation and maintenance of aberrant afferent activity observed in the presence of chronic pain continues to grow. These include a variety of mediators released from cells around peripheral and central terminals of nociceptive afferents, transducers responsible for converting mechanical, thermal, and chemical stimuli into neural activity, and the channels that control neural excitability [69]. Thus, much of the focus of research into the identification of mechanisms underlying chronic pain has been on afferent terminals and changes in the patterns of gene expression detected in the afferent cell body [84,107,108,153,183,186]. However, there is a growing body of evidence to suggest that local translation in the axons of these nociceptors may be effectors of some of these changes [140,148]. As a result, local translation in primary afferents may represent a viable target for alternative therapeutic strategies.

2. Methods

The purpose of this review is to summarize the results of research into the interaction between local and in particular axonal translational and pain. A primary literature search was conducted on PubMed using the search query: ((“pain”[Title/Abstract] OR “injury”[Title/Abstract] AND “axonal translation”[Title/Abstract] OR “local translation”[Title/Abstract”]. Titles and abstracts were used to screen for relevant studies. The results were then filtered manually to include studies that met at least two of the following criteria in terms of focus or purpose: 1) translation mechanisms, 2) translation in the peripheral nervous system at site(s) distant from the cell body, 3) axonal translation in nerve injury, 4) local/axonal translation in the maintenance of pain. Studies identified using this methodology were also evaluated to determine whether they directly or indirectly investigated axonal protein synthesis. In addition, articles familiar to the authors were reviewed for relevance and included. These sources were then organized into the four sections of this review: Axonal Protein Synthesis, Initiation of Local Translation, Second Messenger Systems, and Pain Effector Molecules in Local Translation. A secondary search was conducted using each of these sections as the key words to ensure a thorough investigation of each topic. References from each primary or review source were assessed to identify additional studies missed in original searches. Reviews were omitted when possible. Sources were confined to the last 15 years, unless used for primary/historical context.

3. Axonal Protein Synthesis

Despite initial skepticism, it is now well established that most axons contain a diverse population of mRNAs for local translation that occurs within the axonal compartment [38,76,164,181]. The capacity for axonal translation appears to be one that is evolutionarily conserved as mRNA encoding protein synthesis machinery is detected in axons of species ranging from invertebrates, such as Aplysia, to vertebrate, such as rats [97,134], highlighting the importance of this biological process. Genome-wide microarray profiling of RNA isolated from a variety of in vitro culture systems, including but not limited to Xenopus retinal ganglion cells, rat dorsal root ganglia (DRG), and rat cortical neurons, confirmed that these neuronal compartments contain transcripts encoding protein machinery such as ribosomal proteins, translation initiation factors, and translation elongation factors [76,169,208]. Importantly, these mRNAs have been shown to be axonally translated [97] and provide the necessary protein synthetic machinery for the local translation of the wide repertoire of mRNAs detected in these subcellular compartments [6,76,162,169,208].

Local translation of mRNA provides several advantages over protein trafficking from the cell body. First, translationally silent mRNAs can be stored locally in ribonucleoprotein (RNP granules) structures that can undergo translation in response to local stimulation, thus providing tight spatio-temporal control of local protein pools [4,58,157,175]. Consistent with this notion, axonal translation of localized mRNAs occurs in response to neuronal activity and extrinsic stimuli [21,106,111,188,201]. For example, growth cone turning in response to guidance cues, such as brain-derived neurotrophic factor (BDNF) and netrin-1, requires the localization of β-actin mRNA to distal axons and its local translation ([14,158,187,196]. Second, since axonal trafficking is an ATP-dependent process [103], local translation is more energetically efficient than the trafficking of protein from the cell body to axons. This is especially true for neurons with long axons that provide innervation to the distal appendages some of which can span a meter in length in humans. RNAs can also reach the distal portion of long axons much more quickly than many proteins and are trafficked at rates of approximately 50–400 mm/day (fast axonal transport) while a large population of proteins are transported at rates of approximately 0.2–10 mm/day (slow axonal transport) [154] and this disparity is likely explained by a reliance on different motor proteins [75]. As the sciatic nerve can exceed 1 meter in humans, these differences in transport rate and transport related-ATP consumption are of particular significance in the peripheral nervous system as compared to the CNS, where the average length of cortical axons is approximately 8 mm [65]. One can appreciate the significant temporal advantage that local translation provides when considering the example of the hind limb axons of rodents. Assuming the hind limb axons of mice may be up to 40 mm in length and those of rats may be up to 80 mm, fast axonal transport of mRNA from the DRG to these distal axons may take somewhere between 144 minutes (approximately 2.5 hours) and 19 hours in the mouse and between 288 minutes (almost 5 hours in the rat) and 1.6 days. The trafficking of many proteins from the soma to these sites is estimated to be at least 5-fold slower. Therefore, the ability to quickly respond to extrinsic stimuli is greatly enhanced by the trafficking of mRNA and the translation of localized mRNA.

Finally, the 3’ untranslated region (3’UTR) of mRNA often contains information lost in its protein form such as localization signals, translation signals, and microRNA binding sites, which function in RNA silencing [3,7,18,44,111,176]. Thus, 3’UTRs provide another layer of translation regulation. Intriguingly, neuron-enriched transcripts tend to have longer 3’UTRs and transcripts localized to neuronal processes, such as axons and dendrites, have longer 3’UTRs than those localized to the cell body [133,176]. Several alternative, elongated 3’UTR isoforms have been found to drive axonal localization and these are also enriched in microRNA binding sites [79,82,176]. Together these data suggest that the additional level of control over localization and translation of mRNA provided by 3’UTRs is particularly important to the axon function. It is also important to note that while only a small percentage of total mRNA is locally translated (approximately 5% in sensory axons), local translation appears to be critical for axon outgrowth and guidance during development [15], and maintenance and regeneration in adults [45,174,200]. Inhibition of local protein synthesis results in reduced axonal growth and loss of axonal viability [81,97]. Thus, local translation of axonal localized mRNA supports the proper function of these subcellular compartments.

4. Initiation of Local Translation

4.1. Nerve Growth Factor (NGF)

Nerve Growth Factor (NGF) is critical for the survival of subpopulations of primary afferent neurons during development, and the maintenance of the phenotype of subpopulations of afferents in the adult. However, NGF is also elevated in response to tissue injury where it will sensitize this subpopulation of afferents and therefore contribute to pain associated with injury. Indeed, NGF injected into healthy human skin can cause localized pain and hyperalgesia that develops within minutes but can last for hours to days [145]. The spatially restricted nature of this hyperalgesia suggests a peripheral rather than central mechanism of this sensitization.

NGF binds the TrkA receptor, which is found in peptidergic C fibers and Aδ fibers, sensory neurons that express the neuropeptide calcitonin gene-related peptide (CGRP) [9]. TrkA receptor activation results in the initiation of numerous different intracellular signaling cascades including at least two implicated in local translation [8,59,96]. Neuronal survival signals elicited by NGF induced TrkA receptor activation are dependent on axonally synthesized protein [35]. These same proteins have been implicated in the modulation and maintenance of pain, and therefore are thought to contribute to some of the longer lasting effects of NGF [99,113,195,205], which may be generated via local translation. Supporting the supposition that local protein translation mediates NGF- induced mechanical hypersensitivity is the finding that co-injection of NGF with a general translation inhibitor attenuates the mechanical hypersensitivity observed with intraplantar injections of NGF alone [125].

In addition to the role of NGF in promoting the local translation of downstream signaling molecules, local translation of TrkA, may also contribute to persistent pain. For example, in a rat model of trinitrobenzene sulfonic acid (TNBS)-induced colitis, there is an increase in the retrograde transport of the TrkA protein from the colon and commensurate increase in TrkA protein in the colonic afferent cell bodies. In contrast, there is an increase in the anterograde transport of TkrA mRNA in colonic afferent axons [150]. This suggests that local translation of TrkA contributes to persistent effects of elevated NGF observed at sites of inflammation.

4.2. Other signaling molecules

Several other trophic factors have been implicated in the initiation and/or maintenance of chronic pain in addition to NGF. These include Brain Derived Tropic Factor (BDNF) and Neurotrophin-3 (NT3). A role for BDNF in pain associated with tissue injury has been most extensively characterized in the context of traumatic nerve injury, where it has been shown to mediate the decrease in inhibition observed in the spinal cord dorsal horn [32,39]. However, there is evidence for BDNF signaling in the periphery that contributes to inflammatory hypersensitivity [203]. TrkB is the high affinity receptor for BDNF, and similar to NGF, BDNF-induced activation of TrkB results in the initiation of signaling cascades shown to underlie local translation [112,168]. While the most compelling evidence in support of BDNF-induced local translation comes from CNS neurons, the fact that BDNF contributes to the sensitization of nociceptive afferents suggests that local translation in peripheral axons may contribute to the persistent pro-nociceptive effects associated with BDNF.

The high affinity receptor for NT3 is TrkC, which is primarily expressed in non-nociceptive sensory neurons. In contrast to NGF and BDNF, NT3 appears to attenuate nociceptive signaling. For example, in the chronic constriction injury (CCI) model of trigeminal nerve injury, NT3 induces a significant and long-lasting decrease in injury-induced heat hyperalgesia [46]. In the CCI model of sciatic nerve injury, NT3 significantly reduced Nav1.8 and Nav1.9 mRNA levels and attenuated neuronal upregulation of Nav1.8 and Nav1.9 protein at the site of injury [189]. Additional experiments are needed to confirm whether suppression of local translation contributes to the anti-nociceptive effects of NT3. However, NT3 signaling has been shown to redistribute RNA granules to neurites in cortical neurons [104] and stimulates the phosphorylation of translational initiation machinery in oligodendrocytes [27] suggesting a link between NT3 signaling and the regulation of local translation.

In addition to trophic factors, several inflammatory mediators likely contribute to the initiation and/or maintenance of chronic pain via local translation. IL-6 is one of the most extensively studied of these pro-inflammatory molecules in the context of local translation, and the data suggests these signaling cascades parallel those activated by NGF [127] Indeed, there is evidence of the retrograde trafficking of locally translated proteins in the sciatic nerve within two hours after an injection of IL-6 into the mouse hindpaw [127].

Indirect evidence in support of a role for local translation contributing to the actions of a variety of other signaling molecules comes from the characterization of a phenomenon referred to as hyperalgesic priming, that is suggested to underlie the transition from acute to chronic pain. Hyperalgesic priming involves an initiating stimulus which appears to change nociceptive afferents such that the duration of the response to a subsequent challenge vastly outlasts the duration of a same challenge applied to naïve tissue [2,93,143]. For example, an intradermal injection of the inflammatory mediator prostaglandin E2 (PGE2) results in hypersensitivity that lasts for 45 to 60 minutes in naïve animals, but 24 to 36 hours in a primed animal [2,143]. While there is evidence that changes in CNS circuitry may contribute to the primed state [95], the bulk of evidence points to a long last plasticity in subpopulations of nociceptive afferents where the specific subpopulation of afferents involved depends on the nature of the priming stimulus. Intriguingly, local protein synthesis has been implicated in the establishment of hyperalgesic priming [50,51] and local administration of protein synthesis inhibitors attenuates priming-induced hyperalgesia [51]. Of note, hyperalgesic priming may involve changes at the central as well as peripheral terminals, both of which appear to be dependent on local protein synthesis [52]. Importantly, evidence that local injection of protein synthesis inhibitors can reverse hyperalgesic priming [50] suggests local protein synthesis is not only required for the initiation of the primed state, but for its maintenance. Thus, local translation may play a critical role in the transition from acute to chronic pain. Notably, there is evidence of sex differences in the mechanisms underlying hyperalgesic priming [93], which raises the possibility that the population of locally translated proteins generated differs between sexes and, consequently, the need to consider sex when developing therapeutic interventions targeting this potential mechanism of chronic pain.

4.3. Support cells and local translation

A growing body of evidence suggests that Schwann cells may support axonal protein synthesis. Studies suggests Schwann cells have the capacity to transfer ribosomes and mRNA to the sciatic nerve after injury [22,33], and both appear to be transferred via exosomal transport [22,33,34,119,151]. Additionally, Schwann cells may provide energy substrates to support ATP production in distal axons [90,164]. Intriguingly, sensory neuron axons associate with Schwann cell-secreted exosomes in compartmentalized cultures and Schwann cell-derived exosomes promote axonal regeneration both in vitro and in vivo [117]. Highlighting the potential relevance of these exosomes to local translation is the finding that in compartmentalized cell cultures, Schwann cell derived exosome promote axonal regeneration when added to the axonal compartment but not when added to the cell body compartment [117]. Although the link between Schwann cells and local translation is still tenuous, given the presence of the machinery needed for local translation in peripheral axons, and evidence that mRNA is packaged in exosomes, it is at least reasonable to suggest that other cells may help guide the rapid response to injury by serving as a reservoir of mRNA.

5. Second Messenger Systems in Local Translation

5.1. ERK and PI3K

Of the many signaling cascades initiated by Trk- and/or cytokine receptor activation, at least two have been implicated in local translation. One of these is phosphoinositide 3-Kinase (PI3K), a lipid kinase responsible for the phosphorylation of a number of downstream molecules involved in protein synthesis, including AKT [25,36,80,102,116]. The PI3K/AKT pathway is essential to the development and maintenance of chronic pain (for full review, see [25]). Spontaneous nociceptive behaviors induced by capsaicin and NGF injections are attenuated after intradermal injections of PI3K inhibitors wortmannin or LY294002 [193,207]. In the distal axon, this signaling pathway promotes retrograde transport of BDNF and NGF [12,67], whereas inhibition of PI3K has been shown to attenuate NGF- and IL-6-induced local translation [48,121].

The second signaling cascade implicated in local translation involves the activation of extracellular signal-regulated kinase (ERK). ERK was first implicated in nociceptive processing in the spinal cord dorsal horn, where it was demonstrated that noxious, but not non-noxious stimuli resulted in an intensity and NMDA receptor-dependent increase in the phosphorylation of ERK (pERK). More importantly, inhibition of pERK attenuated the second phase of the formalin response, which is thought to involve spinal neuron sensitization [91]. Subsequent studies showed that pERK is detectable in nociceptive afferents, where it contributes to nociceptor sensitization [37]. The rapid sensitization of nociceptive afferents was initially shown to be due to the phosphorylation of nociceptive signaling molecules such as TRPV1 [207]. More recently, there is evidence to suggest ERK-dependent nociceptor sensitization may also be due to local translation. ERK phosphorylates several down-stream targets implicated in local translation [120,147]. Similar to PI3K, inhibition of ERK blocks local translation as well as NGF- [8,121] and IL-6- [42,43,122] induced hypersensitivity and hyperalgesic priming.

5.2. mTOR signaling

mTOR is a kinase downstream of PI3K activation. While it is involved in a variety of cellular functions including autophagy, metabolism, and transcription, its role in mRNA translation is well established [57,114,184]. For example, mTOR, as well as its downstream targets 4EBP1, S6K, and S6, are extensively expressed in a subset of myelinated primary afferent sensory fibers in the glabrous and adjacent hairy skin of the rat hindpaw [92]. 4EBP1, S6K, and S6 are involved in protein synthesis through their association with eukaryotic initiation factors and ribosomes, and constitute critical components of the local translation machinery [92]. This machinery is not only present, but functional since downstream targets for these kinases were largely phosphorylated in distal axons [66,92] and phosphorylation is blocked by intraplantar injection of rapamycin, an inhibitor of mTOR and translation [92]. The presence of translational machinery in myelinated sensory fibers does not necessarily imply that local protein translation may have a role in nociception, as it is also possible that local protein synthesis maintains the health and viability of these axons [81]. However, further investigation into this population of A-fibers demonstrated that the inhibition of protein synthesis in vivo increased mechanical thresholds and reduced mechanical hypersensitivity. In two rodent models of pain, capsaicin-induced secondary hyperalgesia and spared nerve injury (SNI)-induced neuropathic pain, rapamycin treatment increased mechanical thresholds in the von Frey filament and pinprick tests [66,92]. Furthermore, intrathecal administration of rapamycin blocked mTORC1 activity in dorsal roots but not dorsal root ganglion (DRG) suggesting the effect of translation on the modulation of pain behavior occurred at the level of axons rather than the soma [66]. Nevertheless, it is important to note that while markers of mTORC1 activation were measured, resulting axonal translation was not directly demonstrated in this study.

As alluded to above, data demonstrating altered nociception in FMR1 knockout (KO) mice supports the supposition that mTOR-mediated alterations in pain behavior are a result of the modulation of axonal translation [149]. FMR1 encodes Fragile X mental retardation protein (FMRP), an RNA-binding protein (RBP) that has an important role in the negative regulation of local translation at synapses. Large repeat expansions in the FMR1 gene, which encodes the FMRP protein, leads to transcriptional silencing of the gene and is a genetic cause of Fragile X syndrome (FXS), an intellectual disability that is also associated with number of co-morbid neuropsychiatric illnesses [63,190]. Although the role of mTOR signaling in FXS is not completely understood, a number of studies have identified mechanisms through which FMRP regulates mTOR signaling and vice versa [190]. Intriguingly, FMR1 KO mice show decreased responses to ongoing nociception as evidenced by a reduction in pain-related behaviors in the second phase of the formalin test [149]. Intrathecal injections of the mTOR inhibitor rapamycin decreased second phase pain-related behaviors in wild-type (WT) mice in a dose-dependent manner without affecting the behavior of FMR1 KO mice in response to formalin [149]. These results suggest that the mTOR pathway and localized protein synthesis may mediate formalin-induced nociception. This supposition is supported by evidence that formalin-induced neuronal hyperexcitability and second phase nociceptive behavior are mediated by rapid protein synthesis at the spinal level: these responses are attenuated by rapamycin [5], which may be due to inhibition of mRNA translation secondary to the inhibition of mTOR signaling [10]. Additional evidence supporting the importance of axonal protein synthesis in mediating these effects includes the finding that FMR1 KO mice exhibit a 3-week delay in the development spared nerve injury (SNI)-induced mechanical hyperalgesia [149]. This suggests that proteins involved in mediating SNI-induced mechanical hyperalgesia are axonally synthesized in WT mice but need to be trafficked from the cell body into nociceptive axons in FMR1 KO mice [149], based on estimates of transport times for mRNA and proteins summarized above. Together, these data support the idea that translational regulation via mTOR and FMRP are an important feature of nociceptive plasticity.

5.3. Eukaryotic Initiation Factor

Eukaryotic Initiation Factor Complex (eIF4F) is a downstream target of mTOR signaling. It is comprised of eIF4E, which binds to the 5’ cap protein, eIF4A, which unwinds the 5’ UTR of mRNAs to expose its translation initiation sites, and eIF4G, a scaffolding protein that binds eIF4A and eIF4E [1,130]. When assembled, this complex promotes eukaryotic translation initiation and protein synthesis in primary sensory neurons and their axons [125,165]. mTORC1 indirectly affects the assembly of this complex via the phosphorylation of the 4E-binding proteins (4E-BP). These inhibitory proteins actively suppress the eIF4F complex assembly by tonically binding eIF4E which prevents its interaction with eIF4G. However, activation of mTORC1 phosphorylates 4E-BPs, causing dissociation from eIF4E and thus enabling complex assembly and translational initiation.

There is an abundance of evidence supporting the role of the eIF4F complex in regulating persistent pain states [100]. Prevention of eIF4E phosphorylation at Serine 209 abolished an increase in excitability of DRG neurons, while also mitigating mechanical and thermal hypersensitivity, affective pain expression, and hyperalgesic priming induced by IL-6, NGF, and protease-activated receptor 2 (PAR2) [56,135]. This phosphorylation is dependent on MNK1/2 signaling, as inhibition of MNK1/2 recapitulates these behavioral and electrophysiological effects [135]. Suppression of MNK1/2 signaling was also shown to alleviate mechanical hypersensitivity in the setting of inflammatory pain [125,135]. Most relevantly, local inhibition of eIF4E phosphorylation prevents NGF- and IL-6-induced hypersensitivity [135]. Together, these data suggest MNKK1/2 and eIF4E underlie local translation driven by trophic factors and cytokines which subsequently contributes to persistent pain states and hyperalgesic priming.

6. Pain Effector Molecules in Local Translation

6.1. GAP43 and nerve regeneration

Growth-associated protein 43 (GAP-43) is a membrane protein that plays a major role in axonal pathfinding and branching for the developing nervous system. In the adult peripheral nervous system, GAP-43 is constitutively downregulated, but is upregulated following nerve injury to promote axonal sprouting [16,68,83,191]. Furthermore, intrathecal injection of IL-6 and NGF upregulates the expression of GAP-43 through the PI3K/AKT/mTOR signaling pathways [194]. Importantly, GAP-43 protein can be axonally synthesized and its local translation promotes axonal elongation and regeneration after injury [44,45]. While it is clear that GAP-43 plays a critical role in nerve regeneration following injury, there is conflicting evidence exists surrounding its role in nociceptive signaling where the interpretation of the results associated with manipulations of GAP-43 on nociceptive behavior have been confounded by the potential impact of these manipulations on glial cells as well as regenerating axons.

More compelling evidence exists for a pain-related role for GAP-43 in humans. GAP-43 is detected in cutaneous NGF+ and CGRP+ nerve fibers [49,179]. In patients with diabetes, GAP-43 expression levels have not only been shown to increase [19,61,62,197], but there appears to be a correlation between the level of GAP-43 expression and pain, and this correlation is independent of fiber loss [19,61,62]. Similar results were reported in patients with chronic pancreatitis [40].

6.2. CREB

As discussed above, both NGF and IL-6 are important signaling molecules in the initiation of local translation. One interesting point of convergence between IL-6 and NGF-mediated protein translation is their effect on axonal translation of cyclic AMP response element binding protein (CREB), a transcription factor implicated in the modulation and maintenance of pain [113,195,205]. It was previously established that CREB mRNA is present is distal axons of sensory neurons and that NGF stimulates its local translation [35]. Importantly, when distal axons were severed from their respective cell bodies, in vitro, the presence of axonal CREB protein was dependent on both NGF and translation demonstrating that CREB mRNA is locally translated in NGF-dependent manner. Moreover, this axonally synthesized CREB was retrogradely trafficked to the soma and was also found to be present in its phosphorylated form in the nucleus [35]. Retrogradely trafficked CREB was also found to be colocalized with TrkA-signaling endosomes following NGF application [35], which is significant because TrkA-containing signaling endosomes mediate the activation of a kinase that is required for phosphorylation of CREB in response to axonal application of neurotrophins [185]. Intriguingly, axonal CREB was able to induce cyclic response element (CRE)-dependent transcription [35] raising the possibility that axonally translated CREB could induce transcription of pain-related genes in response to noxious stimuli. Supporting this possibility, more recent studies demonstrate that local translation and retrograde transport of CREB contributes to IL-6-induced nociceptor plasticity. One study demonstrated that intraplantar injection of IL-6 in mice resulted in the detection of CREB protein in the sciatic nerve within 2 hours. Click chemistry was used to confirm that the increased CREB protein detected in the sciatic nerve was nascently synthesized at the site of injection. Significantly, disruption of retrograde transport of CREB protein with microtubule disrupting agents prevented IL-6 mediated mechanical sensitivity and hyperalgesic priming and inhibited IL-6 mediated BDNF upregulation in the DRG [127], which was previously shown to be an important component of nociceptor sensitization [126]. The use of a cAMP response element (CRE) consensus sequence as an oligonucleotide decoy to disrupt CREB’s action as a transcription factor, prevented IL-6 mediated mechanical hypersensitivity and hyperalgesic priming as well as IL-6 mediated increases in BDNF expression [126,127]. Together, these data demonstrate the intimate link between CREB’s local translation and retrograde trafficking, its role as a transcription factor, and IL-6-induced plasticity. It is tempting to speculate that local translation of CREB, or other yet to be identified proteins, may effect transcriptional changes in DRGs that are important for the maintenance of persistent pain, especially since retrograde signaling effects transcriptional changes in the soma [131]. Support for this comes from research in regenerating sensory neurons. After nerve crush injury, an axonally localized Importin β1 mRNA isoform, a classical nuclear import receptor, is upregulated and locally translated. Importantly, this locally translated Importin β1 associates with the motor protein dynein to coordinate the trafficking of an injury-signaling complex containing transcription factors that are axonally translated and retrogradely trafficked to the nucleus [17,77] Conditional Cre/lox mediated knockout of the axonally localizing element of Importin β1 mRNA attenuates the transcriptional response to nerve injury detected at the cell body and delays recovery indicating that local translation of this transcript and its retrograde signaling, and presumably nuclear translocation, is critical for functional plasticity of the sensory neuron after injury [144].

6.3. CGRP

In addition to the local translation and retrograde transport of transcription factors such as CREB, there is also evidence that pain-related proteins are locally synthesized. Calcitonin gene-related peptide (CGRP) is a protein widely expressed in central and peripheral nociceptive pathways where it plays a role in orchestrating the response to tissue injury and infection [26,88,110], and is involved in migraine pathophysiology, as well as somatic, visceral, neuropathic, and inflammatory pain [159]. CGRP is also hypothesized to play a key role in peripheral sensitization [87]. Interestingly, CGRP mRNA and protein levels increase in regenerating sensory axons after peripheral nerve injury both in cultured DRG and intact nerves [174]. αCGRP mRNA is also concentrated at growth cones and its local translation may play a role in nerve regeneration [174]. Indeed, CGRP was found to be locally translated in peripheral sensory nerve axons and knockdown of CGRP mRNA by local administration of siRNA reduced the capacity for nerve regeneration [174]. The local translation of CGRP may also have relevance to nociceptive signaling [141,142,174].

6.4. Voltage-gated sodium channels (VGSCs)

Elevated protein and mRNA levels of several ion channels, including Nav1.8, have been detected in peripheral sensory axons after injury [70,156,171]. Increased Nav1.8 resurgent currents are thought to contribute neuronal hyperexcitability and hypersensitivity associated with peripheral neuropathies [192] and Nav1.8 expression is suggested to be necessary for the expression of spontaneous, inflammatory, and neuropathic pain after injury [70,94,155]. Thus, the increased axonal levels of Nav1.8 mRNA and subsequent locally synthesized protein is highly relevant to multiple types of persistent pain. As soma levels of Nav1.8 are reduced in some models of neuropathic pain [70,163], one interpretation of these findings is that pre-formed Nav1.8 protein is redistributed to axons following injury rather than being locally translated [70,139]. However, the upregulation of axonal Nav1.8 mRNA following nerve injury has been replicated by several investigators [28,82,171], and suggests that increased axonal Nav1.8 protein is due to local translation. Moreover, an alternative, elongated Nav1.8 3’UTR has been recently identified in sensory axons following sciatic nerve entrapment [82]. As mRNAs with extended 3’UTR have been shown to preferentially localize to axons [79,176] and distinct 3’UTR isoforms can differentially mediate activity dependent translation of mRNA in neurons [111], it is possible that this alternative Nav1.8 mRNA isoform is preferentially localized and axonally translated in states of hyperexcitability like those seen in neuropathic pain.

While Nav1.8 mRNA and protein is consistently found to be upregulated in the axon following peripheral nerve injury [82,156,171], it is highly likely that other VGSCs are axonally translated following injury and contribute to the initiation and/or maintenance of persistent pain states. For example, following infraorbital nerve entrapment (IoNE), an animal model of trigeminal neuralgia, there is a significant increase in Nav1.3, Nav1.7, and Nav1.8 mRNA in the infraorbital nerve of the IoNE group as compared to the sham group [137]. Conversely, in the trigeminal ganglion (TG) of IoNE animals, there was a reduction in Nav1.1, Nav1.6, and Nav1.8 mRNA. That the pattern of changes local Nav subunit translation may be injury specific is suggested by our recent observation that Nav1.1, but not Nav1.3, Nav1.7, or Nav1.8 is increased in the infraorbital nerve following chronic constriction injury [146]. Together, these observations suggest that there is a redistribution of specific Nav subtype mRNA from TG to peripheral axons following injury. It is important to note that axons contain the necessary secretory machinery, such as protein components of the signal recognition particle, endoplasmic reticulum, and Golgi apparatus, to traffic axonally translated VGSCs to the plasma membrane of axons [71,118,128,129,180]. Consistent with these data are the observations that nerve entrapment [137] and constriction injury [146] result in changes in the sensitivity of action potential propagation in the infraorbital nerve to Nav subtype selective blockers. Importantly, local pharmacological block of Nav1.1 at the site of the constriction injury of the infraorbital nerve attenuated injury-induced mechanical hypersensitivity, suggesting that these axonally translated proteins are not only functional but may serve as therapeutic targets for the treatment of pain.

6.5. MMP-9

Matrix metalloproteinase-9 (MMP-9) has been implicated in pain and inflammation [177]. Although it has traditionally been thought to be secreted in its pro-, or inactive form, and subsequently operate in the extracellular space to cleave proteins such as growth factors, cell surface receptors, and cell adhesion molecules, a growing body of literature supports its role in the intracellular space as well [30,177,178]. Of note, MMP-9 has emerged as an important molecule in the control of synaptic plasticity. In the CNS, MMP-9 mRNA is trafficked into dendrites and locally translated in an activity dependent manner [47,89,105,210]. Although there is no direct evidence for local translation of MMP-9 in the peripheral nervous system, MMP-9 mRNA has found to be upregulated in the sciatic nerve following multiple models of sciatic nerve injury [24,82], suggesting it could be axonally translated and locally secreted following injury, where it could contribute to pain via the cleavage of several targets including myelin basic protein [24,115]. It is likely that in addition to the cleavage of proteins directly involved in the control of nociceptor excitability, MMP-9 contributes to injury-induced pain via indirect actions. For example, there is evidence that MMP-9 induced cleavage of cytokines, such as TNF-α and IL1-β, as well as upregulation of chemokines are important components of MMP-9 mediated neuropathic pain [98,206]. Further supporting this point is the finding that intrathecal injections of an MMP-9 targeting antibody not only decreases mechanical allodynia associated with chemotherapy induced neuropathy but downregulates inflammatory mediators such as IL-6 and TNF-α in DRG neurons [173] suggesting that these cytokines may be the pain effector molecules.

7. Conclusions and Future Directions

A general model of the steps associated with local translation in primary afferents is illustrated in Figure 1. As highlighted throughout this article, while the phenomenon of local translation is well-established, much of the evidence in support of its role in chronic pain is indirect. A variety of tools are now available, however, to visualize the trafficking of mRNA and translation machinery, as well as local changes in protein levels. Thus, in the immediate future, many of the steps involved in local translation can and should be directly demonstrated in the context of tissue injury and pain.

Figure 1:

Summary of steps associated with local translation in primary afferent axons. The process is initiated (Initiation Process) via neural activity (not shown) and the activation of receptors for cytokines and neurotrophins. There is also evidence that the process may be started by the transfer of mRNA and translation machinery via exosomes released from local cells such as Schwann cells. Receptor activation initiates second messenger cascades (Second Messenger Signaling) ultimately resulting in the translation of mRNA. The proteins locally translated (Effector Molecules) may contribute to nociceptive signaling (CGRP), the facilitation of action potential propagation (voltage-gated sodium channels, VGSCs), and proteins trafficked back to the cell body that influence gene expression (CREB).

In the context of future directions and potential therapeutic implications of local translation for the treatment of pain, there are critical questions to be answered with respect to each step of the process outlined in Figure 1. For example, while there is abundant evidence that neural activity is sufficient to trigger local translation [47,53,111,175], less is known about the link between injury-induced patterns of activity and the patterns of local proteins translated, let alone the site(s) throughout the neuron where local translation occurs. Similarly, while we highlighted several of the mechanisms implicated in the initiation of local translation, this is likely to involve more than a focal increase in NGF or IL6. That said, while focal increases in pro-nociceptive mediators may explain why local translation occurs at one point in a neuron, but not another, little is known about the mechanisms controlling site(s) of local translation in response to injury. Recent evidence that organelles called ribosome associated vesicles (RAVs), a structure that may prove to be critical for local translation, are transported along microtubules [23], suggests a mechanism for local translation at the site of nerve injury. That is, if the nerve injury is associated with microtubule disruption, RAVs will necessarily cluster at the site of damage.

The trafficking and release of mRNA in RNA binding proteins (RPBs) would provide another mechanism influence site(s) of local translation. It remains to be determined which RBPs transport pain-related mRNAs to axons let alone whether they differ from those that regulate the transport of mRNAs that are critical in the regeneration of axons following injury. For example, the RBP zipcode-binding protein 1 (ZBP1) transports both GAP-43 and β-actin mRNA to axons to support axonal outgrowth [44,45] and the RBP nucleolin forms transports mTOR mRNA to sensory axons where it is locally translated following nerve injury [170]. However, whether these RBPs are also implicated in the axonal localization of pain-related mRNAs remains to be demonstrated. Additionally, determining whether RBPs coordinate and transport pain-related mRNAs in the same RNA granule, akin to how the RBP splicing factor proline and glutamine rich (SFPQ) orchestrates and transports an RNA regulon that includes mRNAs whose local translation is critical for the viability of sensory axons [31,60], could be of clinical significance as it could provide an opportunity to target the axonal translation of multiple pain-related mRNAs at once. If a common RNA binding motif is identified in multiple mRNAs that are locally translated in the context of chronic pain, therapeutics such as antisense oligonucleotides (ASOs) could be developed to prevent the trafficking of these mRNAs to distal axons. Supporting this approach is the finding that competitively inhibiting the RBP poly(A) binding protein (PABP), a protein important in RNA stability, reduces axonal protein synthesis in vitro and injection of the competitive oligonucleotide into the hindpaw of mice attenuates pain in multiple models of inflammatory pain including NGF- and IL6- induced hyperalgesic priming [11].

If the emergence of ectopic activity generation observed at sites proximal to that of nerve injury [101,172] not only contribute to neuropathic pain but are due to local translation, it will be critical to identify mechanisms controlling the site(s) or local translation to target this mechanism therapeutically. However, if the site of local translation contributing to pain associated with injury is known, targeted application of inhibitors of steps critical for local translation may provide pain relief with minimal side effects. Indeed peripheral administration of translational inhibitors has been shown to reduce pain related symptoms in animal models of chronic pain [50,123] and metformin, an AMPK activator, which acts upstream of mTOR, corrects abnormal mRNA translation and reverses enhanced neuronal excitability and behavioral hypersensitivity in several models of chronic pain [85,86,202]. This finding is particularly significant as metformin is already FDA approved for the treatment of type 2 diabetes and could potentially be repurposed for the treatment of chronic pain or alleviate pain associated with diabetic peripheral neuropathy.

Based on a strategy analogous to that proposed for the therapeutic efficacy of metformin, eFT508 is another molecule that should provide pain relief via the block of local translation. eFT508 is an inhibitor that selectively blocks MNK-mediated phosphorylation of the translation initiation factor eIF4E [152]. As phosphorylation of 4E-BP by mTOR relieves inhibition of eIF4E, eFT508 can inhibit translation of molecules downstream of mTOR activation. The MNK-eIF4E signaling axis has already been established as an important player in the development of chronic pain [135,136] and systemic administration of eFT508 has been shown to alleviate neuropathic pain in rodent models [124,161]. Excitingly, eFT508 is currently in phase II clinical trials for the treatment of solid tumors. If approved, this drug could potentially serve as another treatment strategy to target the transition from acute to chronic pain.

Thus, while there is still much to be learned about local translation, together the available data suggests that local translation is not only important for nerve regeneration but plays a critical role in chronic pain. That this process is dynamically regulated suggests it may contribute to the initiation of the chronic pain state and/or the transition from acute to chronic pain. Furthermore, evidence suggests that targeting points of convergence such as mTOR and eIF4 may be viable treatment strategies. As more is learned about local translation, it should be possible to develop even more specific therapeutic strategies. However, as is becoming clear for all known therapeutic targets, the relative contribution of any specific mechanism is likely to be context dependent, and will vary as a function of a variety of factors including time, site of injury, and nature of the injury. For example, the subsets of mRNAs translated in response to inflammation are likely to differ from those translated in response to nerve injury. As such, targeted treatment approaches are likely to be necessary to adequately target chronic pain in patients with osteoarthritis and post-surgical pain versus those with traumatic nerve injury. Nonetheless, the emerging evidence strongly suggests that inhibition of axonal translation has the capacity to be both a personalized and effective therapeutic strategy for the treatment of chronic pain.