We assessed the impact of two hours of acute neck muscle inflammation on the properties of SDH neurons in the upper cervical spinal cord. Our c-Fos experiments showed carrageenan-induced muscle inflammation produced neuronal activation in both ipsilateral and contralateral laminae I – II, and to a lesser extent in lamina V of the C1 and C2 spinal segments. These findings confirmed that neurons in laminae I - II should be the principal target in subsequent electrophysiological recordings from acutely prepared spinal cord slices.
Assessment of c-Fos expression, the protein product of the proto-oncogene c-fos, is a well-established method for determining neuronal activation after noxious peripheral stimulation [19, 20]. In our study, c-Fos expression occurred in neurons located predominantly in laminae I - II. This matches the terminal field distribution of nociceptive afferent fibres , and the location of c-Fos expressing neurons after application of noxious peripheral stimuli . Some c-Fos labelling was also observed in deeper layers (lateral lamina V) of the dorsal horn. These superficial and deep locations closely match those noted in other rodent injury models involving deep tissues, such as muscle. For example in rat, chronic constriction injury of the sciatic nerve , adjuvant-induced arthritis , and nerve growth factor injection into semispinal neck muscles  resulted in c-Fos labelling in laminae I-II and V-VII. In contrast, capsaicin injection into the trapezius and splenius muscles of cats does not label many neurons in lamina II . Thus, some differences exist in the extent to which neurons in laminae II are involved in processing noxious input from muscle across species.
No labelling was observed in control or sham (saline injection) animals in our c-Fos experiments. These data suggest neither handling nor the manipulation of muscle tissue associated with the injection protocol resulted in significant activation of dorsal horn interneurons. Similar results have been observed for limb muscles when c-Fos expression has been compared with that in vehicle-injected animals . Our unilateral carrageenan injections resulted in c-Fos labelling in both ipsilateral and contralateral superficial laminae (I-II) with some labelling in deeper laminae. Bilateral c-Fos expression has been observed after capsaicin injection into other neck muscles (trapezius and splenius capitis) in cats . This contrasts with the data for limb muscles where inflammation produces neuronal activation that is confined to the ipsilateral side of the spinal cord. Thus, the major pattern of c-Fos activation we described following carrageenan injection into RCM, fits with activation of muscle nociceptive pathways in axial musculature. Thus, recording from neurons in ipsilateral laminae I-II should detect inflammation-related changes in the spinal dorsal horn.
In our electrophysiological analysis of SDH neurons in spinal cord slices prepared two hours after RCM inflammation we observed two major changes: i) AMPA-mediated excitatory drive decreased and ii) the amplitude and kinetics of the rapid, but not slow IA type potassium current were altered. Assuming acute neck muscle inflammation enhanced nociceptive signaling (cf. our c-Fos results above), it is surprising that we found excitatory drive to SDH neurons was reduced (as assessed by sEPSC amplitude and charge). It is noteworthy, however, that other studies have also described reduced excitatory drive to SDH neurons in pain models, specifically in GAD67 positive inhibitory interneurons . In these experiments the frequency of spontaneous excitatory input, rather than amplitude, was reduced in neuropathic animals. The authors proposed that reduced excitatory drive to inhibitory interneurons, when placed in the context of nociceptive signaling, would reduce inhibitory drive in the SDH and contribute to hyperalgesia. While our data does not allow us to identify recorded neurons as excitatory or inhibitory, similar plasticity in the form of reduced sEPSCs onto inhibitory interneurons would contribute to enhanced nociceptive signaling in acute neck muscle inflammation.
Regardless of the identity of neurons that undergo reduced excitatory drive in our model, numerous studies have confirmed that inflammation alters the expression of AMPA type glutamate receptors. For example, expression of calcium permeable AMPA receptors containing GluR1 subunits has been shown to increase, whereas expression of calcium impermeable GluR2 containing subunits was reduced [27–31]. Importantly, all these studies assessed receptor expression 24 hours after peripheral inflammation. In contrast, our data showing sEPSC kinetics were unaltered in carrageenan-injected recordings imply that AMPA receptor expression is unchanged at the two-hour time point used following axial muscle inflammation. Phosphorylation of GluR1 and GluR2 has, however, been demonstrated at time points commensurate with those in our study . This would enhance excitatory drive in SDH neurons. Importantly, most GluR1/2 subunit expression and phosphorylation studies have used biomolecular techniques in spinal cord homogenates where synaptic and extrasynaptic receptors and the precise laminae location of neurons are not known. Thus our experiments, which assessed synaptic receptor function in specific laminae, suggest levels of GluR plasticity vary according to time after inflammatory insult.
Several electrophysiological studies have examined the properties of primary afferent synapses in the SDH after peripheral inflammation, and report that primary afferent synaptic function was generally enhanced [33, 34]. Again this contrasts with the reduced excitatory drive that we observed. However, our sEPSCs recordings will have included excitatory currents arising from local interneurons, descending systems, as well as primary afferents. Relevant to this point, previous work has shown that ablation of afferent input by dorsal rhizotomy, or selective removal of peptidergic afferents with neonatal capsacin treatment, does not reduce mEPSC frequency recorded from SDH neurons in spinal slices [35, 36]. This suggests that the majority of sEPSCs recorded in an isolated spinal slice such as that in our experiments, come from local interneurons and not primary afferent terminals. Thus, any change to primary afferent synapses may have gone undetected in our experiments.
When sEPSCs have been recorded in the SDH in peripheral inflammation models it appears their properties are not altered within the first 24 hours after inflammation [37, 38]. Interestingly, the Li et al. study showed that sEPSC frequency was enhanced when inflammatory insults were delivered to neonates but not in animals older than P14. The most likely explanation for the reduced sEPSC amplitudes we observed comes from work that has assessed the contribution of the initial primary afferent barrage and its time course during an inflammatory insult. For example, peripheral nerve block with lignocaine during muscle inflammation prevents the development of plasticity (or central sensitisation) in spinal neurons, whereas nerve block outside the first two hours of inflammation does not protect against central sensitisation . This finding suggests primary afferent input within the first two hours of inflammation plays a crucial role in establishing central sensitisation. In our experiments, the barrage of primary afferent input would be expected to be the principle driver of central sensitisation, which is reflected as the expression of c-Fos in certain neuronal populations. The fact, however, that we did not observe an enhancement of sEPSC frequency or amplitude may partly be explained by the in vitro slice preparation we used. Specifically, previous work has shown that the removal of populations of primary afferent inputs either by dorsal rhizotomy or neonatal capsacin does not affect the properties of sEPSCs recorded in slices [35, 38]. This implies that most sEPSCs recorded in slices originate from local interneurons rather than primary afferents. Further support of our interpretation comes from work in an in vitro hemisected spinal cord preparation where inflammation-induced central sensitisation was only detected six hours after inflammation in the absence of peripheral input . These data, when combined with our finding that excitatory drive is decreased two hours after acute muscle inflammation, suggests a complex sequence of events occurs over the onset, establishment and maintenance of inflammatory pain and associated central sensitisation.
The reduced excitatory drive we observed two hours after acute neck muscle activation was accompanied by a reduction in the amplitude and time course of the IAr - type potassium current. IAr currents are expressed widely in the CNS and regulate AP discharge and neuronal output. In the dorsal horn, IAr currents are thought to play an important role in reducing neuronal excitability. For example, IAr currents are preferentially associated with the delayed firing pattern of AP discharge (Figure 4A). Likewise, SDH neurons expressing delayed firing and IAr currents exhibit reduced AP discharge compared to tonic firing and initial bursting neurons when activated by current protocols they are likely to receive in vivo (synaptic vs. square step current injection) . Furthermore, a number of studies have suggested IAr expression is a feature of excitatory interneurons [41, 42]. This suggests that diminished IAr would substantially increase excitability and nociceptive signaling in the dorsal horn. Thus, any intervention that reduces IAr current in SDH neurons leads to enhanced excitability.
Comparable modulation of IAr in the dorsal horn has been reported after peripheral inflammation. For example, enhanced excitability in dorsal horn neurons is observed as early as one hour after carrageenan injection into the hindpaw of young (P7-12) mice . A detailed analysis of IAr potassium currents in these animals identified a shift in the current’s steady state inactivation. The authors suggested this would act like “a relaxation of the brake on excitability at physiologically meaningful potentials”. Interestingly, we have previously demonstrated the converse (i.e., enhanced braking) in spastic mice with glycine receptor mutations . In this case, modulation of IAr was seen as a compensatory mechanism to enhance neuronal inhibition and maintain stable sensory processing. Together, this work suggests IAr can be strongly modulated to maintain an excitability set-point in the SDH. Importantly, a signalling pathway involving metabotropic glutamate receptor activation and extracellular signal-regulated kinase dependent phosphorylation of Kv4.2 channels has been shown to regulate IAr currents in SDH neurons [44, 45]. Furthermore, this work demonstrated that inflammation of peripheral tissues can activate this pathway and reduce IAr. This provides a potential mechanism for our observations during acute neck muscle inflammation.
Given IAr currents are diminished in SDH neurons from carrageenan-injected animals one might expect this to impact on neurons with the delayed- and reluctant firing AP discharge pattern as the two have been associated in a number of previous studies [18, 46, 47]. For example, diminished IAr would either reduce the delay before AP discharge or reduce the proportion of neurons exhibiting delayed and reluctant firing and convert them to more excitable forms of discharge such as tonic firing or initial bursting. There was, however, no difference in the incidence of AP discharge patterns across the three conditions: control, sham and carrageenan-injected. This finding may be explained, at least partly, by the biophysical properties of IAr currents, which are inactivated at depolarised membrane potentials [16, 47]. The membrane potential of neurons in our recordings was approximately −60 mV across all three conditions. At this potential much of IAr (~80 - 90%) is inactivated and therefore not able to reduce or delay AP discharge. Under in vivo conditions, however, when membrane potential is fluctuating due to the combination of inhibitory and excitatory synaptic input, IAr currents could have a greater impact on neuronal output. Thus, under in vivo conditions the reduction in IAr we observed could enhance excitation and disrupt normal sensory processing.
Ultimately, our findings must be placed in the context of the whole animal. Fortunately, relevant data are available on the time course of pain related behaviours following carrageenan injections. Most of this work has involved carrageenan injection into the plantar surface of the hindpaw [48–52] (rather than muscle). Carrageenan causes rapid edema, which peaks 3–4 hours post injection and largely resolves by 24–72 hours . Behaviourally, carrageenan’s capacity to induce thermal and mechanical hyperalgesia is highly reproducible. These indices of altered pain sensation have been detected as early as 15 minutes after carrageenan injection , though typically studies report effects to be fully developed at 1–4 hours . Following onset, some variability exists in duration of the carrageenan-induced mechanical and thermal hyperalgeisa. This variability might reflect the carrageenan concentration (typically 1-3% is injected) as well as the species studied. However all work suggests mechanical and thermal hyperalgeisa persists for at least 24 hours . The few studies that have injected carrageenan into muscle also support a rapid onset for mechanical and thermal hyperalgesia, which is fully developed by two hours and can last as long as 4–8 weeks [54, 55]. Thus our study has described a number of central changes to both the synaptic and intrinsic membrane properties of SDH neurons that equate to a relatively early time point in the carrageenan-induced inflammatory pain model.
In summary, the present study suggests two hours of acute neck muscle inflammation represents a transition point between the involvements of peripheral and/or central sensitisation. As our study was undertaken at one time point post inflammation it is difficult to predict where each of the alterations we observed lies in terms of the train of events that leads to central sensitisation. The surprising result that sEPSC amplitude and charge are reduced, suggest some excitatory inputs are depressed in the hours following inflammation. This may reduce local excitatory drive in the face of the initial excitatory barrage arising from the periphery. Alternatively, if the neurons that experience a reduction in excitatory drive were inhibitory, an associated reduction in the activity of this population would reduce inhibition in the SDH and contribute to enhanced nociceptive signaling. It will be important in future studies to assess whether our observation persists after peripheral drive abates following the resolution of inflammation in the periphery. The diminished IAr-current after two hours of acute inflammation is, however, consistent with inflammation-induced hyperexcitability in the spinal cord dorsal horn. Furthermore, there is a clear signalling pathway that links hyperexcitation (and neuronal depolarisation), with ERK kinase-dependent down regulation of A-currents . Thus, the early barrage of afferent input from inflamed muscle could underpin this observation. Our findings when combined with the existing literature on altered spinal cord processing, albeit at longer times after acute insult, suggest a complex sequence of events occurs in the dorsal horn following neck muscle inflammation. A greater understanding of the order and duration of these changes may help to uncover new strategies for blocking the transition from peripheral to central sensitisation.