The present findings provide several novel insights into the pathology and potential treatment of neuropathic pain. We have demonstrated that PNI induces a major reorganization of translation regulation signaling, translation machinery and RNA-binding proteins in the injured PNS. These changes are directly linked to increased eIF4F complex formation and augmented nascent protein synthesis. We have identified AMPK activation as a novel avenue for the potential treatment of neuropathic pain in humans. Metformin and A769662 both alleviated neuropathic allodynia, inhibited translation regulation pathways associated with PNI and decreased sensory neuronal excitability. No clinical trials have assessed the efficacy of metformin for neuropathic pain. These studies provide a compelling preclinical rationale for the clinical assessment of metformin for neuropathic pain in humans and for the future development of more efficacious AMPK activators for the treatment of chronic pain disorders.
Translation control in the axonal compartment of neurons contributes to development of the peripheral and central nervous systems , is involved in axonal regeneration following injury [26, 27] and contributes to pain plasticity [7, 8]. Likewise, the repertoire of mRNAs localized to the axonal compartment is developmentally regulated and shows plasticity upon injury [26–30]. Interestingly, a recent non-biased approach to mRNA profiling of the axonal compartment of DRG neurons revealed a developmental shift toward localization of mRNAs involved in immune regulation and nociception in the adult DRG axon . In support of this mRNA profile, we have recently demonstrated that NGF and/or IL-6-induced allodynia is dependent on local, axonal translation from existing pools of axonally localized mRNAs . The present findings enhance our understanding of plasticity of translation control after PNI. While changes in mRNA localization have been observed after pre-conditioning peripheral nerve lesions , we found that PNI induces profound changes in activity of kinases associated with translation control (e.g. mTOR and ERK), phosphorylation of their downstream targets and in overall levels of proteins involved in RNA processing and transport (e.g. Mov10, FMRP and rck/p54). This reorganization results in increased eIF4F complex formation and nascent protein synthesis in the injured sciatic nerve. Moreover, these changes are directly correlated to normalization of neuropathic allodynia as metformin reversed changes in protein synthesis in injured sciatic nerves and resolved PNI-induced allodynia.
Protein synthesis is an energy intensive process and, for this reason, an intricate system to control energy consumption has evolved in the form of the ubiquitous kinase, AMPK . Activated AMPK blocks protein synthesis by inhibiting components of the mTOR and ERK signaling pathways . Concurrent inhibition of multiple signaling pathways inherently prevents signaling crosstalk, which is commonly observed with the inhibition of a single kinase involved in these signaling cascades (e.g. inhibition of mTORC1 with rapamycin ). Receptor tyrosine kinases are associated with IRS, which mediates activation of the mTOR and ERK pathways. Inhibition of mTORC1 with rapamycin removes negative feedback onto IRS, meditated by phosphorylation of IRS (S1101) by rS6K . Thus, inhibiting mTORC1 and, in turn, rS6K, with rapamycin releases disinhibition of IRS signaling resulting in activation of ERK and mTORC2/AKT pathways . ERK activation in the PNS is a well known mechanism for increasing the excitability of nociceptors . For these reasons we focused on AMPK as a therapeutic target for neuropathic pain as AMPK suppresses IRS signaling by phosphorylation on Serine 794 . Activation of AMPK with distinct pharmacological tools failed to promote ERK or AKT activation and, in the case of A769662 and AICAR, led to inhibition of these kinases. These findings highlight advantages of targeting AMPK for the treatment of neuropathic pain.
We have also demonstrated that activation of AMPK in mouse sensory neurons leads to decreased excitability. We used ramp current-evoked spiking to assess the influence of AMPK activation on sensory neuron excitability. Our findings are consistent with a potential modulation of the voltage-gated sodium channel Nav1.7 by AMPK activators. Human genetic studies clearly demonstrate an important role for Nav1.7 in inherited pain conditions and a growing body of evidence suggests an important role for Nav1.7 in acquired chronic pain states . In humans, Nav1.7 expression is increased in painful neuromas  and dental pulp [35, 36]. Moreover, inhibition of Nav1.7 decreases sensory neuron excitability [37, 38]. Genetic deletion of Nav1.7 in mice leads to marked decreases in acute and inflammatory pain . Finally, pharmacological inhibition of Nav1.7 with several distinct classes of inhibitors leads to a reduction in allodynia in preclinical neuropathic pain models [40–43]. Hence, human clinical findings and pharmacological inhibition of Nav1.7 creates a compelling rationale for targeting Nav1.7 in neuropathic pain disorders. The present findings indicate that the AMPK signaling axis regulates sensory neuronal activity by decreasing action potential firing induced by ramp current injection and increasing the latency to the first action potential, both of which are consistent with modulation of Nav1.7 . We hypothesize that this AMPK-mediated modulation of sensory neuron excitability may be linked to inhibition of ERK because ERK phosphorylates Nav1.7 altering channel gating properties toward a hyperexcitable state and leads to decreased neuronal hyperexcitability . While further work will be needed to directly test the effect of AMPK activators on ERK-mediated Nav1.7 phosphorylation and Nav1.7 current kinetics, the present findings demonstrate a role for AMPK modulation in sensory neuronal excitability.
The results presented here are consistent with a peripheral action for AMPK activators in the alleviation of SNI- and SNL-induced allodynia; however, we cannot exclude a potential central mechanism of action. Several recent studies have demonstrated an important role for dorsal horn mTOR in preclinical pain models [9, 45–49] and AMPK activators influence the mTOR pathway in central neurons [50, 51]. Moreover, metformin crosses the blood brain barrier . We favor a peripheral mechanism of action for several reasons. AMPK activators had a clear effect on mTOR and, in some cases, ERK signaling, in cultured sensory neurons. These compounds also negatively influenced the excitability of these neurons, consistent with the alleviation of neuropathic pain. Moreover, in vivo treatment led to a reversal of PNI-induced enhanced nascent protein synthesis, consistent with a direct action of AMPK activators on the injured PNS. Finally, inhibition of translation regulation signaling (e.g. with mTORC1 inhibitors) in the CNS is thought to play a critical role in the initiation but not maintenance of plasticity . To this end, we have recently shown that mTOR inhibiton in the dorsal horn is incapable of reversing an established preclinical pain state  and, in the setting of neuropathic pain, other investigators have concluded that even centrally applied mTOR inhibitors act via inhibition of DRG neuron excitability . Nevertheless, we cannot rule out a potential CNS site of action for AMPK activators and, due to a key role of translation regulation in learning and memory  we also cannot rule out a possible effect of AMPK activators on these processes. Additional pharmacokinetic/pharmacodynamic studies will be required to resolve this question with certainty.
The present findings have important implications for AMPK-based drug discovery for the treatment of pain. Metformin activates AMPK via LKB1 stimulation  and inhibition of AMP deaminase . AICAR results in 5-aminoimidazole-4-carboxamide ribonucleoside accumulation in cells mimicking AMP binding to AMPK . On the other hand, A769662 is a direct, positive allosteric modulator of AMPK  which requires the β1 subunit of the kinase heteromer . Unlike metformin and AICAR which activate AMPK by inducing the phosphorylation T172 on the alpha subunit, A769662 induced AMPK activation does not require this post-translational modification . While both metformin and A769662 led to a full reversal of neuropathic allodynia, A769662 was more potent in vivo and had a more profound effect on sensory neuron excitability with a near complete blockade of ramp current evoked spiking. Moreover, A769662 (and AICAR) led to robust ERK inhibition in sensory neurons in culture whereas metformin had no impact on ERK activity. While further experimentation will be needed to gain insight into the exact mechanisms through which different modes of AMPK activation achieve differential signaling endpoints, these results point to a potential pharmacological advantage for positive allosteric modulators of AMPK for the treatment of chronic pain. Furthermore, the efficacy of A769662 in mouse models of neuropathic pain suggests that targeting the β1-subunit of AMPK may be a viable drug development target for the pain pathway.
In conclusion, we have demonstrated a novel pathway for the potential treatment of neuropathic pain, AMPK activation. Pharmacological AMPK activation negatively regulates aberrant translation control after PNI, resolves neuropathic allodynia and decreases sensory neuron excitability. Due to the clinical availability and safety of metformin, these preclinical findings have the potential to lead to rapid translation into the clinic.