Formalin-induced behavioural hypersensitivity and neuronal hyperexcitability are mediated by rapid protein synthesis at the spinal level

We used in vivo electrophysiology (see methods) to study the effect of rapamycin on neuronal responses from naive rats in order to determine the importance of rapamycin-sensitive pathways under physiological conditions. When rapamycin was administered onto the exposed spinal cord (250 nM or 11.43 ng in 50 μl), there was a significant reduction in nociceptive-specific C-fibre-mediated transmission onto WDR neurones when compared to DMSO control (Figure 1A), therefore indicating that rapamycin-sensitive pathways mediate stimulus-evoked responses of nocicpetors. We therefore expected that wind-up, which is a potentiated response mediated by nociceptive C-fibre activity and a measure of neuronal hyperexcitability, would also be significantly inhibited. However, this was not case, although in all cases, wind up was inhibited to some extent. Furthermore, in some cases, there were clearly strong inhibitory effects when compared to DMSO (Figure 1B). Although this data my appear conflicting, one must bear in mind that measuring the number of action potentials attributable to C-fibres after a train of stimuli does not directly correlate with wind up since wind up involves the added feature of non linear recruitment of initially silent NMDA receptors [13] which likely requires a higher degree of inhibition achievable, assumingly by a higher dose of rapamycin.

Rapamycin also significantly inhibited neuronal responses to von Frey filaments (8 – 60 g) when compared to DMSO (Figure 1C). Previous studies have shown that in naive rats, the 50% behavioural mechanical withdrawal threshold varies from around 11 to 19 g [14, 15] so it is apparent from these results that rapamycin-sensitive pathways are important in mediating neuronal responses to innocuous as well as noxious mechanical stimuli under physiological conditions. Rapamycin-sensitive pathways appear to be more important in mechanically-evoked responses since thermally-evoked responses were weakly altered by rapamycin (Figure 1D). Therefore, although rapamycin-sensitive pathways are important for stimulus-evoked neuronal responses under physiological conditions, this appears to only be true for specific sensory modalities.

We used in vivo electrophysiology to study spinal neuronal hyperexcitability induced by formalin injection into the hind paw. When rapamycin was administered onto the exposed spinal cord 3 min prior to formalin injection into the hind paw, there was a significant reduction in neuronal activity from 40 – 60 min when compared to DMSO administration. Area under the curve (AUC) analysis confirmed that overall, rapamycin elicited a significant reduction in the second phase of the formalin test (Figure 2A). There were no significant effects of rapamycin on the first phase of the formalin test. To confirm that these effects were indeed due to inhibition of mRNA translation, we also examined the effects of the general mRNA translation inhibitor anisomycin, on the formalin test in the same manner as that for rapamycin. There was a strong trend for inhibition of neuronal hyperexcitability from 10 – 30 min compared to DMSO administration and just like rapamycin, AUC analysis confirmed that overall, anisomycin elicited a significant reduction in the second phase of the formalin test (Figure 2B). A time of 3 min incubation with rapamycin or anisomycin was chosen due to direct evidence from studies on the importance of rapamycin-sensitive pathways in hippocampal LTP pointing towards a time-restricted role for the involvement of rapamycin-sensitive pathways at LTP induction [7]. Spinal neurones selected for 25% DMSO and rapamycin treatment or 10% DMSO and anisomycin treatment prior to formalin being injected into the hind paw comprised equivalent populations for all measures (Tables 1 and 2) i.e. there was no selection bias for neurones for either treatment.

For behavioural studies, we first administered rapamycin 5 min prior to injecting formalin into the hind paw. We found that unlike the results produced with in vivo electrophysiology, there was no significant effect of rapamycin on formalin-induced behavioural hypersensitivity (data not shown). We assume this to be due to the differences in the experimental conditions since the in vivo electrophysiology set up involves applying the drug directly to the exposed spinal cord (dura removed) whereas the behavioural studies involve injecting the drug onto the surface of the cord (dura in tact). In addition, we cannot rule out the possibility that the rats had completely recovered from anaesthesia within 5 min even though they appeared to be fully alert. When rapamycin was spinally administered 20 min prior to formalin injection into the hind paw, there was a significant reduction in the total behaviour for both the first phase at 5 min and also the second phase at 20, 25 and 30 min when compared to DMSO. This was confirmed with AUC analysis (Figure 3A). The effects of rapamycin were found to be more selective for licking and biting as there was a significant reduction in the length of this behaviour in the first phase at 5 min and also in the second phase at 30 min. Again, this was confirmed with AUC analysis (Figure 3B). Rapamycin was however ineffective in attenuating lifting and flinching behaviour (Figure 3C).

For all studies, male Sprague Dawley rats (250 – 280 g) were used. These were supplied by the Biological Services Unit (BSU, University College London, UK). All procedures were carried out in accordance the UK Animals (Scientific Procedures) Act, 1986 and were in agreement with the IASP guidelines [37].

In vivo electrophysiology studies were carried out according to a well established protocol [38]. Rats were initially anaesthetised in an induction box with 4% isofluorane in a mixture of nitrous oxide (66% v/v) and oxygen (33% v/v). Once the rats had lost consciousness and were completely areflexic, the trachea was exposed and isolated and a cannula was inserted into the trachea and fastened with 3-0 silk threads. This was used to maintain anaesthesia throughout the recording period. At this stage, the isofluorane was reduced to 2.5% v/v (areflexia was maintained). Rats were then secured in a stereotaxic frame and a rectal probe attached to a heating blanket was used to maintain a core temperature of 37°C.

An incision was made through the skin along the length of vertebrae and the skin was then separated from the underlying muscle. Muscle, connective tissue and vertebrae were specifically removed from lumbar vertebral segments L1 – L3 of the spinal cord. Muscle and connective tissue from surrounding areas were kept intact and this created a well in the exposed spinal cord area into which, drug solutions could be added. Clamps were used to stabilise and straighten the cord. The dura mater was also removed to aid drug penetration. When the set up was complete, the isofluorane was reduced to 1.8% v/v, a level sufficient for anaesthesia, whilst maintaining areflexia. 11.43 ng rapamycin (sirolimus, LC laboratories) dissolved in a 50 μl saline/dimethyl sulphoxide (DMSO, Sigma) mix comprising 25% v/v DMSO (250 nM rapamycin); 62.35 μg anisomycin (Sigma) dissolved in a in a 50 μl saline/DMSO mix comprising 10% v/v DMSO (4.7 mM anisomycin); 25% v/v DMSO and 10% v/v DMSO were applied directly onto the exposed spinal cord.

Recordings were obtained with an AC recording system (NeuroLog system, Digitimer). An electrode (parylene insulated tungsten microelectrode, 125 μm diameter, 2 MΩ, A-M systems Inc.) inserted into a head stage attached to a 3-axis manipulator was manually lowered into the exposed cord (L4 – L5) to a depth of 500 – 1000 μM. This is an area occupied by WDR neurones that are important in pain processing. An oscilloscope was used to isolate single neurones and a number of stimuli were applied to the receptive field. Mechanical stimuli (von Frey filaments) were applied to the most sensitive part of the receptive field for 10 s. This was also the case for thermal stimuli, where increasing heat was applied using a jet of water from a 60 ml syringe attached to a needle.

Before formalin was administered to the hind paw, a neurone was selected and characterised. Electrical stimuli were delivered by inserting two stimulating electrodes intradermally into the most sensitive part of the receptive field of the hind paw. Firstly, Aβ- and C-fibre thresholds were determined depending on their latencies to respond to stimuli (Aβ-fibres = <20 ms post-stimulus; C-fibres = 90 – 300 ms post-stimulus). The stimulator was then set to three times C-fibre threshold and a train of 16 stimuli (0.5 Hz, 2 ms pulse width) was delivered to the receptive field to determine the number of action potentials attributable to Aβ-fibres (0 – 20 ms); Aδ-fibres (20 – 90 ms); C-fibres (90 – 300 ms) and post-discharge (300 – 800 ms) which is attributable to the wind up elicited by repeated stimuli of nociceptive C-fibres. The input (non-potentiated response) and the wind up (potentiated response) were calculated as follows: C-fibre Input = action potentials (90 – 800 ms) evoked by the first pulse at three times C-fibre threshold multiplied by the total number of pulses (16). This represents the theoretical baseline in the absence of wind up. Wind up = total action potentials (90 – 800 ms) after the 16-train stimulus at three times C-fibre threshold minus the input. This represents the excess activity above the theoretical baseline due to wind up. For characterising thermally-evoked responses prior to hind paw formalin injection, increasing heat was applied using a water jet directed at the receptive field for 10 s. When determining the effect of rapamycin on baseline neuronal responses, only stable cells where 3 consecutive stimulus-evoked responses that were within 10% of the previous result for the same test were selected for further pharmacological study. A 'test' comprising electrical, mechanical and thermal stimuli was carried out every 20 min. Maximum changes (positive or negative) from control in neuronal activity were used for data analysis (all raw values).

To monitor spontaneous neuronal activity as a result of formalin-induced inflammation, 50 μl of a 5% v/v formalin solution made from 40% v/v formaldehyde solution (BDH Chemicals Ltd) was intradermally injected into the hind paw ipsilateral to the WDR neurone which had already been characterised, using a 0.5 ml insulin syringe (BD Micro-Fine™). A WDR neurone was selected on one side of the cord for treatment with spinally administered (intrathecal or i.t.) vehicle prior to formalin injection into the corresponding hind paw. Only after a biphasic control response was achieved was a neurone then selected on the opposite side for treatment with the drug prior to formalin injection into the corresponding hind paw. Neuronal activity was separated into bins of 10 min.

Before each behavioural study, each rat was allowed to acclimatise for 30 min in individual open top clear Plexiglass chambers (length, width, height = 25 × 25 × 25 cm). In order to determine the effect of drugs at the spinal level on pain-like behaviour, rats were first lightly anaesthetised on 2% v/v isofluorane in a mixture of nitrous oxide (50% v/v) and oxygen (50% v/v) after which they were disinfected and lightly shaved across their backs. A 0.5 ml insulin syringe (BD Micro-Fine™) was used to inject a 20 μl i.t. dose of drug solution (250 μM or 11.43 μg in a 50 μl saline/DMSO mix comprising 25% v/v DMSO) through the skin, into the L5 – L6 vertebral interspace after which, the rats were allowed to recover prior to formalin injection into the hind paw. A 20 μl volume has been shown to produce uniform coverage of the spinal cord which is restricted to the sacral and cauda equina levels and extends up to thoracic T13 – lumbar L1 [39]. Behavioural studies used a higher dose of drug (250 μM) compared with electrophysiological studies (250 nM) since 250 μM rapamycin has been shown to be effective in attenuating capsaicin- and nerve-injury-induced behavioural hypersensitivity when injected locally into the hind paw [6]. After the rats had recovered, they were restrained and 5% v/v formalin solution was then administered to the left hind paw. The rats were then placed back into their chambers and observed for 1 hr. The following behaviours were measured: 1) licking and biting and 2) lifting and flinching [40]. Behavioural data were separated into bins of 5 min. The drug regimen was blinded until the analysis was complete.

After all studies, rats were overdosed on a rising concentration of CO₂, after which, death was ensured by cervical dislocation of the neck.