Inhibitions and Down-Regulation of Motor Protein Eg5 Expression in Primary Sensory Neurons Reveal a Novel Therapeutic Target for Pathological Pain

The motor protein Eg5, known as kif11 or kinesin-5, interacts with adjacent microtubules in the mitotic spindle and plays essential roles in cell division, yet the function of Eg5 in mature postmitotic neurons remains largely unknown. In this study, we investigated the contribution and molecular mechanism of Eg5 in pathological pain. Pharmacological inhibition of Eg5 and a specific shRNA-expressing viral vector reversed complete Freund’s adjuvant (CFA)-induced pain and abrogated vanilloid receptor subtype 1 (VR1) expression in dorsal root ganglion (DRG) neurons. In the dorsal root, Eg5 inhibition promoted VR1 axonal transport and decreased VR1 expression. In the spinal cord, Eg5 inhibition suppressed VR1 expression in axon terminals and impaired synapse formation in superficial laminae I/II. Finally, we showed that Eg5 is necessary for PI3K/Akt signalling-mediated VR1 membrane trafficking and pathological pain. The present study provides compelling evidence of a noncanonical function of Eg5 in primary sensory neurons. These results suggest that Eg5 may be a potential therapeutic target for intractable pain. Supplementary Information The online version contains supplementary material available at 10.1007/s13311-022-01263-2.


Introduction
Eg5, known as kif11 or kinesin-5, is an evolutionarily conserved kinesin that is essential for centrosome movement and bipolar spindle formation during cell division [1][2][3][4]. The bipolar structure of Eg5 enables the motor protein to crosslink and slide antiparallel microtubules apart, generated an outward pushing force, and promotes centrosome separation and bipolar spindle formation [5,6]. The biological functions of Eg5 are modulated by phosphorylation [3,7,8] and acetylation [9] at different amino acid residues of Eg5. Eg5 inhibitors abolish spindle assembly, block mitotic progression and show anticancer effects [10,11]. Beyond mitotic cells, Eg5 was detectable in postmitotic mature neurons, acting as a brake on microtubule movements and promoting the regeneration of injured axons [12,13]. However, the function of Eg5 in mature neurons has not been fully explored.
The noncanonical function of Eg5 in cytokinesis and postanaphase was reported in plants [14,15] but not in mammals. Accumulating evidence has shown the contributions of kinesin proteins (KIF17, KIF13B and KIF5B) to the membrane trafficking of pain-related molecules (NMDA receptors, VR1 and Na V 1.8) [16][17][18], suggesting that kinesins might be neoteric and efficacious targets for pathological pain [19]. The molecular mechanisms of Eg5 in pain modulation remain incompletely understood.
Here, we discovered a noncanonical function of the motor protein Eg5 in pathological pain. Pharmacological and shRNA virus results show that Eg5 modulates VR1 expression in dorsal root ganglion (DRG) neuronal soma, nerve fibres and central terminals as well as synapse formation in the dorsal horn. We further found that Eg5 contributes to pathological pain and VR1 membrane trafficking through the PI3K/Akt signalling pathway. Our results lay the groundwork for targeting Eg5 in mature postmitotic neurons in pathological pain modulation. Na Wei, Yang Yu and Yan Yang contributed equally.

Experimental Animals
The behavioural experiments were performed on male Sprague-Dawley albino rats (purchased from Laboratory Animal Center of FMMU) weighing 150-200 g (age > 4 postnatal weeks). The animals had access to water and food ad libitum and were maintained at room temperature (22-26 °C) with a light/dark cycle of 12 h. The number of animals used and their suffering were minimized. All animal experiments were performed in accordance with the ARRIVE guidelines and approved by the Institutional Animal Care and Use Committee of FMMU.

Inflammatory Pain Model and Behaviour
Complete Freund's adjuvant (CFA, 100 μl, 1:1 dissolved in 0.9% sterile saline) was administered to the plantar surface of the rat left hindpaw to induce inflammatory pain one day after CFA injection [20]. For assessment of heat hypersensitivity, rats were placed in a plastic chamber on the surface of a 2-mm-thick glass plate, and the sensitivity to heat stimuli was detected by an RTY-3 radiant heat stimulator (Xi'an Bobang Technologies of Chemical Industry Co., Ltd., China). The heat stimuli were applied to both the injection site and the corresponding area of the contralateral paw, and the latency was determined as the duration from the beginning of heat stimuli to the occurrence of a marked withdrawal reflex. Five stimuli were repeated for each site, and the latter three values were averaged as the mean paw withdrawal thermal latency (PWTL, s).

Virus Intraganglionic Injection
AAV virus intraganglionic injection was performed as previously described [21]. The rats were anaesthetized with sodium pentobarbital (50 mg/kg, i.p.), and a laminectomy was performed from the L4 to L5 vertebrae to expose the DRGs for intraganglionic injection (2 μl, 1 × 10 13 viral genomes (vg)/mL, Biowit Technologies, China) into the left L4-5 DRGs. The injection needle was withdrawn 10 s after the end of the injection. Rats were allowed to recover for 3-4 weeks prior to behaviour or staining experiments.

Vesicle Transport and Axonal Branching
Live cell imaging in AAV9 virus-transfected DRG neurons was performed as described previously [12]. For analysis of fluorescence-positive vesicle transport, 360 time-lapse images were taken at 40 ms exposure at 5 s intervals for each axon in a live cell imaging system (Cell^R MT20, Olympus, Japan). The velocity of fluorescence-positive vesicle transport was analysed using distance/time (μm/min). For the live cell imaging of stepwise axon and branch growth, image acquisition was performed at 15-s intervals for 1 h using a heated stage apparatus to maintain the temperature at 37 °C.

Membrane Loading with FM Dye and Photobleaching
FM dye membrane loading and photobleaching methods were modified from a previous report [22]. For loading of the membrane, DRG neurons were bathed in a solution containing 5 μM SynaptoGreen C4 (FM1-43) (Biotium, USA, 70,022) for 90 s. Regions of interest (ROIs) for measuring recovery after photobleaching were 15 μm × 15 μm in size and contained approximately a quarter of the membrane region of one DRG neuron. The standard fluorescence recovery after photobleaching (FRAP) protocol involved six prebleach images at 2-s intervals (60% laser intensity, 488 nm), one bleach iteration (100% intensity, 405 nm, 20 s) and 43 postbleach images at 1.6 s intervals (60% intensity, 488 nm). A recovery curve was obtained simultaneously by measuring the average fluorescence intensity. Fluorescence during recovery was corrected for the bleaching produced by the 60% intensity acquisition illumination, which was measured in control experiments using the same imaging protocol but omitting the intense bleaching stimulus.

Electron Microscopy
Electron microscopy was performed as previously described [23]. Rats were anaesthetized with sodium pentobarbital (50 mg/kg, i.p.) and perfused with 2.5% glutaraldehyde and 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Tissues were dissected, cut into 2 mm × 2 mm segments and postfixed in 2.5% glutaraldehyde (SPI Supplies, USA, 02607-BA). Segments were washed in 0.1 M phosphate buffer, fixed using 1% osmium tetroxide (Ted Pella, Inc., USA, 18456), followed by dehydration with an increasing concentration gradient of ethanol and propylene oxide. The osmicated tissue blocks were embedded in Epon-812 (SPI Supplies, USA, 02659-AB). Semithin sections (1 μm) were cut by an Ultratome (EM UC7, Leica, Germany) and collected into slides. To wash the epoxy resin, we prepared a stock solution with 5 g NaOH in 50 ml of alcohol. Slides were incubated in working solution with dimethylbenzene 3:1 (v/v) for 5 min at room temperature and washed in graded ethanol and distilled water. Sections were stained in 1% toluidine blue (Sinopharm Chemical Reagent Co., Ltd., China) for 5 min, dehydrated with graded ethanol, vitrified in dimethylbenzene and then mounted under coverslips with neutral resin. Images were taken under an optical microscope (Olympus, Japan). Ultrathin sections (50 nm) were cut by an Ultratome (EM UC7, Leica, Germany) and collected by copper grids (200 mesh). The ultrathin sections were stained with uranyl acetate (Electron Microscopy China, China, GZ02625-3) and lead citrate (Electron Microscopy China, China, GZ10701-2). Images were taken using an electron microscope (TH7700, Hitachi, Japan).

Immunohistochemistry Staining
Immunohistochemistry staining was performed as previously described [20,24]. The rats were anaesthetized with sodium pentobarbital (50 mg/kg, i.p.). The lumbar spinal cord and L4-6 DRGs were dissected, postfixed for 8 h and cryoprotected in 20% sucrose in PBS overnight at 4 °C. Transverse frozen Sects. (20-μm thick) were cut on a CM1900 freezing microtome (Leica, Germany), incubated for 4 h in 0.05% Triton X-100 and 10% goat serum in phosphate buffered saline (PBS) at room temperature, and incubated with primary antibodies at 4 °C overnight with agitation. After three washes with PBS, the sections were incubated with secondary antibodies for 2 h at room temperature. The following primary antibodies were used:

Statistical Analysis
For behavioural experiments, n refers to the number of animals. For immunohistochemistry imaging data, the number of animals used is indicated in the legend. For photobleaching imaging experiments, n refers to the number of cells responding to any stimulus. For vesicle transport experiments, n refers to the number of recorded fluorescence-positive vesicles. For electron microscopy experiments, n refers to the synapse number of regions of interest. No sample size calculation was performed; however, our samples are similar to those used in the field. Statistical analyses were performed using GraphPad Prism Software. All data are expressed as the mean ± standard error (SEM) unless otherwise stipulated. Descriptions of the tests used and n are located in the figure legends. Significance was defined as follows: ns not significant, *P < 0.05, **P < 0.01 and ***P < 0.001.

Eg5 Expression Profile in DRG Neurons and the Analgesic Effects of an Eg5 Inhibitor
To reveal the expression profile of Eg5, we performed immunohistochemistry of Eg5 and phospho-EG5 (pEg5, phosphorylated at threonine 926, which is the key site for the Eg5 and microtubule interaction [3]) and found that 89% of neurons were pEg5-positive in the DRG (Fig. 1A). Triple immunofluorescence labelling showed that pEg5 colocalized with the small-sized neural marker IB4 or the large-sized neural marker NF-200 (Fig. 1B). These data prompted us to examine the roles of neuronal Eg5 in pain modulation. The Eg5-specific inhibitor monastrol (1 mM, 10 μl) was given via i.t. administration twice daily for three consecutive days. Our unpublished data showed that i.t. monastrol had no effects on motor function. The behavioural test showed that monastrol reversed CFA-induced heat hyperalgesia (9.7 ± 0.8 s versus 4.6 ± 0.5 s, P < 0.001) but had no effects on the control group (Fig. 1C).
Morphological and pharmacological results showed that Eg5 was widely expressed in primary sensory neurons and that Eg5 blockade attenuated VR1 expression and inflammatory pain behaviours. Our data indicated a previously unknown role of the motor protein Eg5 in sensory neurons for pathological pain development.

Eg5 shRNA Inhibits Inflammatory Pain and VR1 Expression in DRG Neurons
To fully investigate the mechanism of Eg5 in pathological pain, we generated adeno-associated virus 9 (AAV9) expressing Eg5 short hairpin-structured RNA (Eg5 shRNA) and the fluorescence protein ZsGreen (Fig. 2A). Genotype PCR results confirmed the knockdown effects of Eg5 shRNA virus (Fig. 2B, C). Immunohistochemistry results showed similar distributions of ZsGreen in DRG neurons of different sizes after scramble and shRNA virus intraganglionic injections (n = 4 mice and 17-20 sections for each group, Fig. 2D, E). Based on the similar transfection efficiency of scramble and shRNA, we could compare the effects of these viruses in the following studies. Notably, we found that the shRNA virus was efficiently transfected in vivo into DRG neurons (Fig. S1A) and in vitro into cultured neurons (Fig. S1B, C). These data supported the high efficiency of shRNA virus in neuronal transfection in the DRG.
By vesicle transport assays, we found that Eg5 shRNA accelerated the average velocity of anterograde transport of the ZsGreen fluorescence protein in DRG neuronal axons (3.9 ± 0.4 versus 27 ± 3.1 μm/min, P < 0.001, Fig. 4F, G and H). Then, we generated an AAV9 virus containing the VR1-mCherry fusion protein (Fig. S2) and confirmed the colocalization of VR1 and the reporter protein mCherry in the soma and axon of DRG neurons ( Fig. 4I and S2D). Using the VR1-mCherry virus, we directly observed VR1 transport in axons (Fig. 4J). Monastrol application increased the average velocity of VR1 anterograde transport in axons of cultured DRG neurons (63.8 ± 5.2 versus 93.5 ± 8.6 μm/min, P < 0.01, Fig. 4K, Movies S1-S2). Therefore, Eg5 inhibition accelerated VR1 anterograde transport in the axons of DRG neurons and impaired VR1 expression in the dorsal root of DRG neurons.
Collectively, our data showed that the motor protein Eg5 regulated VR1 expression in DRG neuronal soma, nerve fibres and central terminals as well as synapse formation in superficial lamina I/II of the dorsal horn (Fig. 6F). Eg5 inhibition generated analgesia at multiple sites in the sensory nervous system, indicating that Eg5 might be a novel molecular target for pathological pain treatment.

Discussion
In the present study, we first reported a previously unknown noncanonical function of the motor protein Eg5 in pathological pain. In vivo and in vitro data showed that Eg5 inhibition abrogated PI3K/Akt signalling-mediated VR1 plasma membrane trafficking and reversed pathological pain. Targeting Eg5-dependent membrane trafficking of pain regulators will be a novel strategy for pathological pain treatment.
The VR1 ion channel is essential for the thermal sensation of neurons, and its plasma membrane expression largely dominates the whole-cell response in pain [25,26]. The process of VR1 membrane trafficking can be modulated by cyclin-dependent kinase 5 (Cdk5), synaptotagmin and synaptic vesicle membrane protein VAMP1 [17,[34][35][36]. Phosphoinositide 3-kinase (PI3K) interacts directly with the N-terminal region of VR1 and facilitates its trafficking to the plasma membrane [29,32]. Activation of the PI3K/ Akt signalling pathway contributes to pathological pain [30,31]. In the present study, we further showed that Eg5 was upstream of PI3K/Akt signalling-mediated VR1 plasma trafficking. Pharmacological inhibition and shRNA knockdown impaired VR1 trafficking and reversed inflammatory pain behaviour.
The contribution of VR1 axonal transport in nerve fibres to pain was reported two decades ago [37,38]. However, the molecules mediating the process remain unclear. Previous data showed that Eg5 inhibition enhanced the transport frequency of short microtubules but not rhodamine-dextranlabelled vesicles and FM-labelled mitochondria [12]. Here, we found that Eg5 inhibition could speed up the anterograde transport of ZsGreen fluorescence protein in axons. Furthermore, by using the VR1-mCherry virus, we revealed that Eg5 inhibition accelerated the average velocity of VR1 anterograde transport and abrogated VR1 expression in the axons of DRG neurons. Our results indicate that Eg5 is the key molecule modulating VR1 axonal transport in sensory nerve fibres.
In the present study, we found that local and systemic applications of monastrol and shRNA virus suppressed synapse formation in the superficial laminae of the spinal cord. Eg5 inhibition-induced synapse impairment was unexpected. One interesting finding was that monastrol caused dendritic spine loss in cultured mouse hippocampal neurons [39]. Here, we showed that the loss of synapsin, the specific protein maintaining synapse organization [28], was responsible for synapse impairment in the spinal cord. FM1-43 photobleaching tests showed that Eg5 inhibition abrogated fluorescence recovery, suggesting the roles of Eg5 in maintaining membrane fluidity. Membrane fluidity contributes to the formation and function of synapses [40]. Therefore, membrane fluidity perturbation induced by Eg5 inhibition contributed to synapse impairment. Additionally, synapse maturation requires a high amount of energy [41] and is involved in a series of steps from neurite formation, outgrowth and branching in the early stage to synapse formation, maturation and plasticity in the late stage [42]. Our results suggested that, under Eg5 inhibition, the increased number of input fibre bundles and axonal branching might not efficiently support synapse formation. Our study provide some reasonable explanations for the roles of Eg5 in synapse formation in the spinal cord.
The origins of synapse connections in the spinal cord are tremendously complex, containing nerve fibre terminals from primary DRG neurons, local excitatory and inhibitory interneurons and descending GABAergic or serotonergic neurons from the brain rostroventral medulla [43]. We found that Eg5 shRNA virus intraganglionic injection decreased the number of synapses in superficial laminae, indicating that the decreased synapses were arising from primary DRG neurons.
Spinal cord superficial laminae received afferent projections of thinly myelinated Aδ, non-myelinated C-fibres and even heavily myelinated Aβ fibres coming from DRG neurons [44][45][46][47][48][49]. By the ZsGreen + virus transfection approach, we found that large-sized DRG neurons projected axonal fibres crossing the deep laminae and finally terminated in superficial laminae. Therefore, the synapse originating from Aβ, Aδ or C fibre that is the target of Eg5 still needs to be clarified. Additionally, Eg5 is widely expressed in small-, mediumand large-sized DRG neurons, transmitting diverse somatosensory information coming from proprioceptors, mechanoreceptors, thermoceptors and chemoreceptors [46][47][48]. It will be of interest to determine the potential functions of motor protein Eg5 in other types of somatic sensory, such as itch and touch.

Conclusions
In the present study, our data show that Eg5 regulates VR1 expression in DRG neuronal soma, nerve fibres and central terminals as well as synapse formation in the superficial lamina of the spinal cord. We further identify a novel molecular mechanism by which Eg5 contributes to pathological pain and VR1 membrane trafficking through the PI3K/Akt signalling pathway. These findings lay the groundwork for targeting Eg5 in mature postmitotic neurons in pathological pain modulation. Eg5 inhibition might be an efficient strategy to alleviate intractable pain.