Injury-induced activation of the endocannabinoid system promotes axon regeneration

Summary Regeneration after a peripheral nerve injury still remains a challenge, due to the limited regenerative potential of axons after injury. While the endocannabinoid system (ECS) has been widely studied for its neuroprotective and analgesic effects, its role in axonal regeneration and during the conditioning lesion remains unexplored. In this study, we observed that a peripheral nerve injury induces axonal regeneration through an increase in the endocannabinoid tone. We also enhanced the regenerative capacity of dorsal root ganglia (DRG) neurons through the inhibition of endocannabinoid degradative enzyme MAGL or a CB1R agonist. Our results suggest that the ECS, via CB1R and PI3K-pAkt pathway activation, plays an important role in promoting the intrinsic regenerative capacity of sensory neurons after injury.


INTRODUCTION
Peripheral nerve regeneration is limited after injury, mainly due to the large distance the axons have to grow in the limbs. 1,2 However, the regenerative potential of sensory dorsal root ganglia (DRG) neurons after a peripheral nerve injury or a spinal cord injury (SCI), is enhanced by a prior injury of their peripheral axon, a remarkable ability called the conditioning lesion. 3,4 This conditioning lesion paradigm has been thoroughly used as a model to identify multiple regenerative signaling pathways and molecules affecting the intrinsic ability of neurons to regenerate after injury. 5 Interestingly, new pathways are discovered every year using this paradigm. [6][7][8][9] While the endocannabinoid system (ECS) has been widely studied for its neuroprotective and analgesic effects, 10 especially in the context of neuropathic pain, 11 little is known about its putative role in axonal regeneration, and during the conditioning lesion.
The ECS is an important neuromodulatory system highly expressed in the nervous system. It is composed by cannabinoid receptors, the most important being cannabinoid receptor type 1 (CB1R), highly expressed in neurons, and cannabinoid receptor type 2 (CB2R), predominantly expressed in immune cells. Exogenous and endogenous ligands (endocannabinoids) act on this system by binding to these receptors. The most abundant endocannabinoids in the nervous system are 2-arachidonoylglycerol (2-AG) and arachidonoylethanolamine (AEA or anandamide). Endocannabinoid levels are regulated in an activity-dependent manner by specific synthesis and degradation enzymes. N-acylphosphatidylethanolamine phospholipase D (NAPE-PLD) is the main responsible for the synthesis of AEA, and diacylglycerol lipase (DAGL) is the responsible for the synthesis of 2-AG. 10 In contrast, monoacylglycerol lipase (MAGL) is predominantly responsible for the degradation of 2-AG, 12 whereas fatty acid amide hydrolase (FAAH) is responsible for the degradation of AEA. 13 After nerve injury in mouse models of neuropathic pain, convincing evidence exists regarding increases in local endocannabinoid levels. 14,15 This increase is rapidly counteracted by swift cellular uptake and subsequent degradation. 16 In fact, increasing the endocannabinoid levels by inhibiting their degrading enzymes has been previously used as a therapeutic strategy to increase their analgesic effects. 17 This approach has profited from a local activation of cannabinoid receptors at sites with high endocannabinoid turnover, rather than global activation of CB1R, which can result in side effects. 16 neuronal CB1R activation. Additionally, the CB1R agonist ACEA also enhances neurite outgrowth in cultured DRG neurons, via activation of the PI3K signaling pathway. The administration of this agonist also improves axon growth in vivo in mice after sciatic nerve crush. Interestingly, we found that increased neurite outgrowth induced by a conditioning sciatic nerve lesion was blocked by rimonabant administration, a CB1R antagonist, suggesting that CB1R is involved in the conditioning injury-dependent axonal regeneration. Altogether, we have shown the importance of the ECS in promoting the intrinsic regenerative capacity of sensory neurons after injury, opening new venues for the treatment of axonal injuries.

Sciatic nerve-conditioning injury promotes axonal outgrowth through 2-AG synthesis
Since the ECS is known to respond upon damage to DRG neurons, we first sought to test if injury-induced endocannabinoid synthesis would support axonal regeneration, including after sciatic ''regenerative'' conditioning lesion. We measured axon outgrowth in cultured DRG, after a conditioning sciatic lesion combined with LEI-401, a high affinity NAPE-PLD inhibitor, 18 DO34, a selective DAGL inhibitor, or vehicle. We found that neurite outgrowth induced by a conditioning lesion was partially blocked by administration of DO34, immediately before crush ( Figure 1A). Contrarily, LEI-401 did not affect injury-induced DRG outgrowth. These results indicate that the synthesis of 2-AG is necessary for the conditioning effect.
Next, we examined whether 2-AG from DRG neurons is sufficient for DRG outgrowth. We cultured DRG cells in vitro and treated them with the MAGL inhibitor JZL184. We observed that the administration of 1 mM JZL184 was sufficient to enhance neurite outgrowth ( Figure 1B), although when administered at a higher concentration (3 mM) did not, likely reflecting unspecific effects at this dose. In line with the results of the blockade of the synthesis of AEA through NAPE-PLD inhibition, the inhibition of AEA hydrolysis via URB597, a selective inhibitor of FAAH, did not produce any changes in DRG in vitro neurite outgrowth ( Figure S1). iScience Article To further understand whether the sciatic nerve injury may alter the expression of 2-AG synthesis and degradation enzymes or their cannabinoid receptors, we measured the expression of Cnr1, the gene encoding for CB1R, that was significantly reduced 24h after sciatic nerve injury. Cnr2 expression, the gene encoding for CB2R was instead significantly increased. Neither Magl nor Dagl expression were altered 24h after sciatic nerve injury ( Figure S2).
Together, these data show that the ECS responds in the DRG after a regenerative nerve lesion and that 2-AG is necessary for axonal outgrowth and regeneration of sensory axons.

CB1R activation promotes axonal regeneration after sciatic nerve injury
In order to test the effects of CB1R activation after sciatic nerve injury, we measured axon outgrowth in ex vivo cultured DRG, 24h after injury combined with rimonabant, a selective CB1R antagonist, or vehicle. Results indicate that neurite outgrowth induced by a conditioning lesion is blocked by administration of rimonabant immediately before crush ( Figure 2A).
To test the specificity of the effects induced by JZL184 administration, we combined the administration of the MAGL inhibitor with rimonabant. Interestingly, we found that rimonabant co-administration blocked the JZL184-induced neurite outgrowth ( Figure 2B), indicating the CB1R specificity of the effects.
We then sought to phenocopy the regenerative capacity of DRG neurons after sciatic nerve injury, by activating CB1R with ACEA (a highly selective CB1R agonist). We first measured the in vitro DRG neurite outgrowth after ACEA administration and found a significant increase in outgrowth compared to vehicle ( Figure 3A). We then asked whether exogenous CB1R activation with ACEA would induce nerve regeneration in vivo in the sciatic nerve crush model. We found that ACEA delivered to the sciatic nerve at the time of the crush enhanced the regenerative capacity of injured axons ( Figure 3B). These data indicate that CB1R activation is sufficient to promote growth in DRG neurons.

CB1R activation promotes axonal growth through PI3K-Akt pathway activation
In order to identify the underlying molecular mechanisms responsible for the CB1R-induced axonal growth, we targeted the two main known downstream actuators of CB1R activation, the MEK pathway and the PI3K pathway, both previously known to be involved in axon growth. 19,20 PD032591, an inhibitor of MEK1/2, did not prevent the ACEA-induced outgrowth. Conversely, pharmacological inhibition of PI3K with wortmannin, a selective inhibitor of PI3K, blocked the neurite outgrowth induced by ACEA ( Figure 4A), as evidenced by the significant difference versus the Veh-ACEA group and the lack of difference versus the Veh-Veh group. Additionally, ACEA induced an increase in the phosphorylation of Akt 2h after its administration ( Figure 4B), but not ERK, proving the role of the PI3K-Akt pathway in the ACEA-CB1R induced neurite growth.
To gain further insight into the downstream molecular pathways supporting CB1R-dependant axonal regeneration, we checked the expression of well-established regeneration associated genes (RAGs), documented to be upregulated after conditioning injury, on DRG neurons 24 h after administration of ACEA ( Figure S3), however, we did not observe any change in the expression of any of the analyzed RAGs.
As CB1R activation did not seem to trigger a relevant shift in the expression of RAGs, we tested whether local actin cytoskeleton remodeling was the responsible for the CB1R-dependant induced axonal growth.
To this aim, we measured the in vitro DRG neurite outgrowth combining the CB1R agonist ACEA with different pharmacological inhibitors of the actin cytoskeleton dynamics: Wiskostatin, an N-WASP inhibitor and C3T, a Rho GTPase inhibitor that stabilizes pre-formed actin filaments and inhibits their disassembly, inhibiting in turn the progression of the growth cone. However, we did not find any of the mentioned inhibitors to affect ACEA-induced outgrowth ( Figure S4). In our study, we identified the ECS as an intrinsic modulator of regeneration. Specifically, we found that this system promotes axon regeneration by activating the CB1R-PI3K-Akt pathway following sciatic nerve injury in mice. The ECS plays an important modulatory part in neurotransmission, affecting different cellular and tissue processes in both homeostatic and pathological settings. [22][23][24][25] Although we did not observe any change in the expression of the main enzymes responsible for the synthesis and degradation of 2-AG, there seems to be a regulation of their activity, leading to increased 2-AG synthesis, as evidenced by the partial inhibition of the injury-induced neurite growth, when inhibiting DAGL. Additionally, we induced axonal growth in vitro by blocking the 2-AG degrading enzyme MAGL. However, the molecular mechanism underlying how a nerve injury induces 2-AG synthesis or DAGL activation remains to be elucidated.
There is strong evidence that the ECS modulates neuronal signaling after injury, a process that has been extensively studied for its analgesic effect during neuropathic pain. 11,26 However, in terms of axonal iScience Article growth, most of the available studies have focused on the role of this system during development of CNS neurons, such as retinal pathways, in which both the activation of CB1R 27 and CB2R 28 have been shown to negatively affect neurite growth and induce growth cone collapse during the formation of these pathways. Similarly, CB1R agonists have also shown to cause growth cone collapse in hippocampal GABAergic neurons. 29 In contrast, but in line with our results, other studies have shown that CB1R antagonists inhibit axonal growth responses 30 and that pyramidal neurons rely on ECS signaling to initiate the elongation and fasciculation of their long-range axons during development. 31 The authors of these studies propose that this divergence may be caused by different targeted growth, i.e., cell autonomous guidance versus target-derived cues. Nonetheless, to this date no studies have evaluated the role of the ECS in the regeneration of peripheral sensory axons after injury. Our results revealed that CB1R activation induces growth in proprioceptive neurons and CB1R antagonism blocks the ECS and injury-dependant growth on these neurons.
We also found that nerve injury reduced the expression of CB1R mRNA in the DRG, in accordance with unbiased RNA-Seq data showing a significant reduction in CB1R expression after nerve injury. 32 Two recent studies have shown different injury-dependant epigenetic mechanisms that control both the downregulation of CB1R expression, 33 as well as the upregulation of CB2R expression. 34 A previous study showed that dorsal root ligation injury did not alter the mRNA level of CB1Rs in the injured DRG, however they found a reduction in the percentage of CB1R/IB4 and CB1R/CGRP double-labeled DRG neurons. 35 This indicates that the apparent overall CB1R expression decrease might be specific to IB4 + and CGRP + nociceptors, and the expression on large-diameter proprioceptors might be unaffected.
We also found that the synthetic specific CB1R agonist ACEA promotes axon growth both in vitro and in vivo. At the same time, pharmacological inhibition of the PI3K-Akt but not the MEK-ERK pathway, the two main downstream mediators of CB1R effects, inhibited ACEA-dependent outgrowth. Additionally, activating CB1R with ACEA leads to the activation of the PI3K-Akt pathway 2h after treatment in cultured DRG neurons, as evidenced by an increased Akt phosphorylation. These results are in line with other studies defining PI3K-Akt as the underlying mechanism of the effects induced by CB1R. [36][37][38] The activation of this pathway in turn promotes the initiation of the regenerative response in DRG neurons as previously reported via other stimuli. 20,39,40 For instance, after nerve injury, Akt is phosphorylated and has been shown to modulate multiple cellular processes affecting axonal growth, including transcriptional and epigenetic alterations, resulting in increased expression of several RAGs, 8,20 increased retrograde transport 9,40 as well as stimulation of cytoskeleton dynamics. 41,42 This CB1R rapid activation of pAkt suggests that endocannabinoids may be part of the early response cascade following a sciatic nerve crush (SNC) and is a key component of the axon regeneration program. Strikingly, the activation of CB1R with ACEA did not affect the expression of the main RAGs known to be upregulated after nerve injury, nor seemed to induce changes in the local actin cytoskeleton.
The present data demonstrate the importance of the ECS in axon growth after injury, while the use of CB1R agonists may represent a novel therapy for nerve injuries. Overall, the present study reveals that the activation of the ECS is associated with axon regeneration and impacts axon regeneration by modulating the activation of PI3K-Akt pathway after nerve injury. This provides deeper insight into the molecular events involved in intrinsic axon regeneration control. Besides the effect of the ECS on PNS nerve regeneration, it would also be interesting to further examine whether the ECS also participates in the axon regeneration process following a CNS injury such as an optic nerve crush or a spinal cord injury.

Limitations of the study
Currently, there is a lack of techniques to measure endocannabinoid levels in DRG neurons on a single-cell level, which limits our understanding of the cell-specific changes that occur. If new methods become available in the future, these would open the possibility to define how a nerve injury changes the endocannabinoid levels. Additionally, while the mechanisms underlying PI3K-pAkt activation have been extensively studied, the methodology used in this study did not allow for a comprehensive evaluation of gene expression changes or metabolic reprogramming after its activation, and all outcomes analyzed did not show significant changes. Despite this, the PI3K-pAkt pathway has been well established as a key player in promoting axonal regeneration.

DECLARATION OF INTERESTS
The authors declare no competing interests.

INCLUSION AND DIVERSITY
We support inclusive, diverse, and equitable conduct of research.

Primary cultures
Primary DRG neurons were obtained from freshly dissected DRG from adult (10-12 weeks old) male and female mice. Dissociated cells were plated on poly-D-lysine / laminin coated culture multiwell plates and kept at 37 C in a 5% CO 2 atmosphere.

Sciatic nerve regeneration
Briefly, the biceps femoris and the gluteus superficialis were separated by blunt dissection, and sciatic nerve was exposed, immobilized and crushed using fine forceps 2x10s orthogonally. The required compounds were injected locally into the sciatic nerve 5mm rostrally to the crush site. 24h later sciatic nerves were dissected and fixed in 4% paraformaldehyde in PBS at 4 C for 2h, or else, the DRGs were collected and processed for RNA extraction. Whole nerves were immunostained for SCG10/Stathmin-2 (Rabbit, Novus) a marker for regenerating axons. 43 Number of regenerating axons caudal to the axotomy and their distance from the lesion epicenter were analyzed for both nerves per animal with a confocal fluorescence Zeiss microscope (LSC800).
When necessary, compounds were added 2h after plating while for combinatorial experiments, cells were treated with pharmacological inhibitors at 1,5h after plating and ACEA, JZL184 or the corresponding ll OPEN ACCESS iScience 26, 106814, June 16, 2023