Distinct brainstem to spinal cord noradrenergic pathways inversely regulate spinal neuronal activity

Abstract Brainstem to spinal cord noradrenergic pathways include a locus coeruleus origin projection and diffuse noxious inhibitory controls. While both pathways are traditionally viewed as exerting an inhibitory effect on spinal neuronal activity, the locus coeruleus was previously shown to have a facilitatory influence on thermal nocioception according to the subpopulation of coerulean neurons activated. Coupled with knowledge of its functional modular organisation and the fact that diffuse noxious inhibitory controls are not expressed in varied animal models of chronicity, we hypothesized a regulatory role for the locus coeruleus on non-coerulean, discrete noradrenergic cell group(s). We implemented locus coeruleus targeting strategies by microinjecting canine adenovirus encoding for channelrhodopsin-2 under a noradrenaline-specific promoter in the spinal cord (retrogradely labelling a coeruleospinal module) or the locus coeruleus itself (labelling the entire coerulean module). Coeruleospinal module optoactivation abolished diffuse noxious inhibitory controls (two-way ANOVA, P < 0.0001), which were still expressed following locus coeruleus neuronal ablation. We propose that the cerulean system interacts with, but does not directly govern, diffuse noxious inhibitory controls. This mechanism may underlie the role of the locus coeruleus as a ‘chronic pain generator’. Pinpointing the functionality of discrete top-down pathways is crucial for understanding sensorimotor modulation in health and disease.

In situ hybridisation data from the Allen Brain Atlas supports this expression pattern.
Additionally, adrenoceptors are present on descending fibre terminals where they function as autoreceptors (see datasets in Allen Brain Atlas, www.proteinatlas.org, and review 2 ), as well as on primary afferents (mouse DRG neurons express only one type of adrenoceptor: Adra2c 3 ).
Furthermore, a significant expression of Adra1d is reported in the Allen Brain Atlas's in situ hybridisation data (cell type unspecific, thus likely non-neuronal), while others' RNAseq data 4 suggests a significant expression of Adra1a on Hes1+ astrocytes in the dorsal lumbar dorsal horn, which role was tied to coerulean control of mechanical hypersensitivity. Activation of DNIC (also phasic) in naïve animals inhibits WDR neurons by α2-ARs-mediated mechanism. These receptors may be located either directly on the WDR neurons or may have an indirect impact (for example via presynaptic inhibition of sensory afferents). Note the body of evidence regarding protein distribution of adrenoceptors in the cord [5][6][7][8] , which is consistent with the RNAseq data (A, B above) despite known low specificity (i.e. knock-out unvalidated) of adrenoceptors antisera 9 . From this we infer that the majority of α2-ARs immunoreactivity in the cord has peripheral sensory neuron origin, as confirmed by the significant signal reduction in animals with ablated sensory neurons (either by dorsal rhizotomy or perinatal capsaicin treatment) 8 . Abbreviations: markers of glutamatergic neurons: CCK-Cholecystokinin, NTS-Neurotensin, NKB-Neurokinin B, GRP-Gastrin-releasing peptide, SP-substance P, Slc17a6-Vesicular glutamate transporter 2 (VGluT2). Markers for GABAergic neurons: Gal-Galanin, Dyn-Dynorphin, NPY-Neuropeptide Y, Calret-Calretinin, Parvalb-Parvalbumin, ENK-enkephalin, Slc32a1-Vesicular inhibitory amino acid transporter (VGAT).

DSP4 injections
For ablation of the coerulean noradrenergic fibres across the neuroaxis, a selective neurotoxin N-(2-chloroethyl)-N-ethyl-2-bromobenzylamine (DSP4) hydrochloride (Sigma, Dorset, UK) was used 12,13 . Six rats weighting 60-80 g were injected intraperitoneally with freshly prepared 50 mg/kg DSP4 in saline. Additional six control rats received vehicle injection. 15% lethality rate from the DSP4 toxin is expected 14 , therefore, to minimise eventual animals suffering injections were given early on the day and rats were kept for up to 4 hours in the recovery incubator (set at 32°C) and carefully monitored for up 8 hours to minimise lethality. 14-16 days after the toxin injection terminal electrophysiological characterisation of lumbar deep dorsal horn WDR neurons was performed followed by transcardial perfusion with cold saline followed by 4% PFA and tissue collection for histological analysis.

Virus injections The LC:LC module
To transduce catecholaminergic coerulean neurons, ipsilateral LC stereotaxic injections were made (Kopf Instruments, UK) analogously to described in detail earlier 15
In brief, male Sprague-Dawley rats weighting 60-80 g were used for injections. Following the induction of anaesthesia (5% isoflurane in 1 l/min oxygen) rats were placed in a stereotaxic frame (without fixing the head) and maintained with 1.

Spinal cord in vivo electrophysiology
In vivo electrophysiology was performed on animals weighing 240-300 g as previously described 19 . Briefly, after the induction of anaesthesia, a tracheotomy was performed, and the rat was maintained with 1.5% of isoflurane in a gaseous mix of N2O (66%) and O2 (33%). Core body temperature was monitored and maintained at 37°C by a heating blanket unit with differential rectal probe system. Electrocardiogram (ECG) was monitored by two intradermal needles inserted in front limbs with signal amplified by the Neurolog system consisting of AC preamplifier (Neurolog NL104, gain x200), through filters (NL125, bandwidth 300 Hz to 5 KHz) and a second-stage amplifier (Neurolog NL106, variable gain 600 to 800) to an analogueto-digital converter (Power 1401 625kHz, CED). Craniotomy was performed to gain stereotaxic access to the ipsilateral LC for either optic fibre or micropipette insertion as described in following sections. A laminectomy was performed to expose the L3-L5 segments of the spinal cord, the cord was clamped to minimise movement, dura was carefully removed with the aid of surgical microscope, and the recording area was secured by saline-filled well made in solidified 2% low melting point agarose (made in saline also). Using a parylene-

In vivo spinal pharmacology with electrophysiological monitoring
After collection of predrug baseline control data as outlined above, atipamezole (a α2-AR (as compared to mean pre-drug baseline) were judged by the maximal change in recorded action potential rate. All data plotted represents the time point of peak change based on these criteria.
Atipamezole is a selective α2-adrenoceptor antagonist (α2/α1 ratio is >8300) 21  Sigma, UK) and the pipette was re-inserted in the same brain region and 500 nl of the Lucifer Yellow Solution was injected therein. 10 minutes were allowed for the dye diffusion after which the animal was sacrificed by the anaesthetic overdose and immediate transcardial perfusion followed by brain extraction for anatomical verification post-mortem.

Optogenetics Light stimulation of coerulean neurons during spinal WDR recordings
The 450 nm laser (Doric Lenses, Quebec, Canada) was externally TTL-triggered by the neurolog system (NeuroLog system, Digitimer, UK) to deliver defined light pulses (20 ms pulse width at 5 Hz). In brief, the laser light was coupled to a multimode 200 um patch cord The power density was adjusted for each preparation. After desired power was achieved the cannula was slowly inserted in the ventral LC ipsilateral to the recorded spinal WDR neurons.
Spinal WDR neurons were characterised by three stable baseline responses followed by three optically modulated responses. For combined optogenetics and spinal pharmacology, after collecting three stable baseline and three stable optoactivation responses (averaged, if stable), a drug (either 100 µg atipamezole or 20 µg prazosin) was applied topically on the exposed spinal cord surface, right above the recording site. To assess simultaneous action of the drug and the LC optoactivation, light pulses were delivered 30 s before and throughout each series of tests (approximately 5 minutes per series) and minimally 5 minutes of the recovery time was allowed between the tests. Pharmacology was monitored every 10 minutes for 30-40 minutes (each test with optoactivation) and the 60 minutes time point was to test neuron returning to the baseline (no optoactivation). At the end of every experiment, animals were sacrificed by the overdose of isoflurane and transcardially perfused with cold saline followed by 4% paraformaldehyde for anatomical evaluation.

LC neuron recording and optoactivation
A simultaneous recording and optical stimulation of the transduced LC neurons were made using microoptrodes as described earlier with minor modifications 27 . LC neurons were identified as described before 15

Passive Tissue Clearing (PACT)
A passive CLARITY tissue clearing technique (PACT) (described in detailed 28

Quantification and Statistical Analysis
All data plotted in represent mean ± SEM. Typically, up to 4 WDR neurons were characterised per preparation (n), and data were collected from at least 6 rats per group (N). Single Additional references: