CREB is a major transcription factor that mediates activity-dependent gene transcription across a wide variety of cell types
. CREB is a crucial component of transcriptional enhancers that regulate functions of the developing and mature brain, including in neuronal survival, synaptogenesis, synaptic plasticity and drug addiction
[22, 24–26]. CREB has also been demonstrated to be crucial for the development of mechanical and thermal hypersensitivity in a variety of preclinical pain models
[27–31]. However, post-translational mechanisms of CREB regulation (e.g. phosphorylation) or CREB DNA targets have been the focus of those studies
[27, 32–36]. Our current results add a new layer of evidence to the role of CREB in nociceptive plasticity and uncover a new mechanism for the development of chronic pain – local translation and retrograde transport of a transcription factor.
Our findings support a model wherein IL-6-mediated mechanical sensitization is mediated by the local translation of CREB, its retrograde transport and subsequent induction of BDNF expression in the DRG. Evidence for this model is three pronged. 1) IL-6 treatment leads to nascent synthesis of CREB and the de novo presence of augmented nascently synthesized CREB in the distal sciatic nerve. 2) Disruption of axonal transport prevents the development of mechanical hypersensitivity and hyperalgesic priming and prevents increased BDNF expression in the DRG. 3) Employment of i.t. delivered, CRE decoy oligonucleotides demonstrates that the IL-6-dependent change in BDNF expression and nociceptive plasticity are selectively dependent on CREB within the DRG. Hence, these findings demonstrate a novel mechanism linking locally generated signals in the form of the nascent synthesis of a transcription factor within the peripheral axons of sensory neurons to remote molecular effects within the cell bodies that manifest as altered mechanical hypersensitivity and hyperalgesic priming.
BDNF is well recognized as an important player in pain sensitization
[14, 37–39]. Previous studies have demonstrated that nociceptor-specific knockout of BDNF leads to a loss of pain sensitization across inflammatory pain models
. Moreover, hindpaw inflammation, hindpaw treatment with NGF
 or DRG neuron treatment in vitro with NGF
 increases BDNF mRNA expression. BDNF mRNA is heavily alternatively spliced with 8 possible 5’ untranslated region (UTR) exons spliced to exon 9 which contains the coding sequence. Inflammation increases expression of several exons but the effect is most robust for exon 1
, which is known to respond to CREB-dependent promoters
. Here we observed an increase in BDNF protein in DRG after hindpaw IL-6 treatment that was dependent on retrograde transport via the sciatic nerve and that was blocked by i.t. treatment with CRE decoy oligonucleotides. A possible alternative interpretation is that i.t. injection of CRE decoys attenuates CREB signaling in the dorsal horn of the spinal cord. Our finding that CRE decoy treatment fails to alleviate i.t. BDNF-induced mechanical hypersensitivity argues against a dorsal horn mediated effect pointing to a DRG-dependent mechanism. We have previously demonstrated a key role for BDNF in the initiation and maintenance of IL-6-induced hyperalgesic priming
[14, 42]. We have also shown a critical role for local translation in the initiation but not maintenance of IL-6-induced hyperalgesic priming
. Several other studies have demonstrated similar links between translation control in DRG neurons and hyperalgesic priming
[43–45]. Our present findings present a possible unification between these two mechanisms in the form of locally translated CREB at the time of IL-6 injection linking the initiation and maintenance of hyperalgesic priming to CRE-dependent changes in BDNF expression in the DRG. Future studies will be needed to evaluate the time course of altered BDNF expression in hyperalgesic priming and whether BDNF dependency during the maintenance phase requires nociceptor-derived BDNF or if it originates from other cells, such as microglia
Our results are consistent with a previous study linking axonal translation of CREB in embryonic DRG neurons to NGF-induced survival
. This study demonstrated the presence of CREB mRNA in DRG axons and induction of local CREB translation upon NGF exposure. This nascently synthesized CREB then undergoes retrograde transport and influences CRE-dependent transcription in sensory neuron nuclei. Interestingly, this retrograde transport of axonally synthesized CREB was required to account for nuclear accumulation of phosphorylated CREB despite the uninterrupted presence of retrograde signaling endosomes. This suggests that local translation, not retrograde endosomal signaling, constitutes the necessary machinery for CRE-dependent changes in transcription via axonal signaling. In development this locally synthesized, retrograde signal is important for NGF-dependent survival of sensory neurons
. In adulthood our results suggest that this same mechanism is used to link local insult to changes in nociceptive sensitivity potentially contributing to the development of a chronic pain state. Although axonal CREB transcripts have been observed in DRG neurons
, similar studies in sympathetic neurons have failed to detect axonal CREB transcripts
. At this point it is not clear if axonal mRNA sorting for CREB transcripts is an exclusive feature of sensory axons or if sympathetic axons are an exception. CREB is a critical mediator of neuronal plasticity in Aplysia californica
[1, 50], an organism where axon crush creates hyperexcitability in a manner reminiscent of observations in neuropathic pain models and patients
[1, 4, 8]. Interestingly, this effect is mediated by local translation
 and involves retrograde transport of an unidentified signaling protein
. Hence, the mechanism we describe here has striking similarity to sensory neuron plasticity in this phylogenetically distinct organism and suggests an evolutionarily conserved mechanism of nociceptive plasticity.
The time course of effects noted in this study deserves comment in relation to measurements of axonal transport. We estimate the distance between the axonal terminals in the hindpaw of mice used in this study to the DRG cell body at ~ 45 mm. At the upper end of estimations for retrograde transport it would take ~ 2.5 hrs for axonally synthesized CREB to reach the DRG nucleus
[51, 52]. A previous study in embryonic DRG neurons in culture observed retrograde transport rates of ~ 8-9 mm/hr for axonally synthesized CREB labeled with the photoactivatable florescent probe Dendra
. We observed robustly increased nascently synthesized CREB (AHA-labelled) in the distal sciatic nerve 2 hrs following intraplantar injection of IL-6. Moreover, we noted an increase in BDNF protein in the DRG, which was reversed by either blockade of retrograde transport or by CRE DNA decoy treatment
[18–20], within 3 hrs of IL-6 treatment. These effects are within the time frame permissible from previous measures of retrograde transport
[51, 52] but are slightly slower than observations made with photactivatable CREB in embryonic DRG neurons in vitro
. A very recent study indicated that prior injury can increase anterograde and retrograde transport rates in DRG neurons in vivo and in vitro
. While this study used prior nerve crush as the conditioning stimulus, it is interesting to speculate that inflammatory stimuli, such as IL-6, may be able to induce plasticity in axonal transport rates.
There are several caveats to our work that should be addressed. First, the i.t. approach to the CRE DNA decoy experiment can be mediated either by an effect on DRG neurons or spinal cord neurons. CREB has been linked to pain plasticity at both of these locations. As mentioned above, BNDF given i.t. at the same time as the CRE DNA decoy treatment failed to modulate BDNF-induced pain plasticity. This is despite the previous demonstration that BDNF induces CREB phosphorylation in dorsal horn neurons
. Hence, we favor an interpretation of our findings wherein CRE DNA decoy treatment is disrupting CREB-mediated actions in DRG neurons that are linked to changes in BDNF expression. We favor a role for retrograde transport of CREB in this effect as blocking retrograde transport attenuated the induction of BDNF in the DRG, however, it is formally possible that this effect is mediated by retrograde signaling endosomes that stimulate CREB activity in the DRG. Ultimately, specific knockdown of axonal CREB mRNA will be needed to provide definitive proof for this model. Second, the compounds used here to disrupt axonal transport have influences on other cells. We cannot completely exclude an inhibitory effect on local inflammation, for instance, within the nerve in response to IL-6 treatment, however, such an effect seems unlikely given that IL-6 treatment itself does not cause any noticeable swelling or flare at the site of injection. We can exclude a potential systemic effect of these compounds based on experiments where they were injected into the contralateral leg. It is also not likely that these transport inhibitors interfere with nerve conduction as ciliobrevin D had no effect on capsaicin-induced flinching while this was completely blocked by lidocaine. Finally, our interpretation of protein localization of CREB and AHA-incorporated CREB in the distal sciatic nerve after IL-6 injection is consistent with retrograde transport, however, we cannot completely rule out a possible contribution of other cells. Having said this, the sciatic nerve was taken 2 cm from the site of AHA injection and it is unlikely that sufficient AHA would travel this distance and incorporate into CREB in sufficient levels for detection using the methods employed here. Moreover, even if this contributed to our observations, such a local effect is completely incompatible with our findings linking retrograde transport and CREB-mediated effects in the DRG to behavioral responses induced by IL-6. Therefore, despite these caveats, we propose that the current evidence favors our model and excludes other alternatives.
We have demonstrated a novel mechanism of nociceptive plasticity linking local translation to changes in gene expression in the distal DRG. Translation control in DRG neurons has emerged as an important mechanism of pain plasticity
[5, 7, 54] but, to our knowledge, no studies have definitively linked a specific locally translated protein to functional changes in the nociceptive system. CREB, an evolutionarily conserved gene controlling neural plasticity
[50, 55], emerges from our work as a first candidate in this category.