A labeled-line for cold from the periphery to the parabrachial nucleus

Spinal projection neurons are a major pathway through which somatic stimuli are conveyed to the brain. However, the manner in which this information is coded is poorly understood. Here, we report the identification of a modality-selective spinoparabrachial (SPB) neuron subtype with unique properties. Specifically, we find that cold-selective SPB neurons are differentiated by selective afferent input, reduced neuropeptide sensitivity, distinct physiological properties, small soma size, and low basal drive. In addition, optogenetic experiments reveal that cold-selective SPB neurons are distinctive with respect to their connectivity, with little to no input from either Pdyn or Nos1 inhibitory interneurons. Together, these data define a neural substrate supporting a labeled-line for cold from the periphery to the brain.


INTRODUCTION 45
Cold, along with heat, pain, itch, and some aspects of touch are conveyed from the spinal 46 cord to the brain via the anterolateral tract (Braz et al., 2014;Todd, 2010). This pathway 47 is made up of neurons that arise from many distinct laminae of the spinal cord and project 48 to numerous regions of the brain, ultimately giving rise to autonomic, affective and 49 discriminative aspects of somatosensation. Although numerous groups have recorded 50 from neurons that contribute to the anterolateral tract, the number of distinct spinal output 51 subtypes remains unclear, and how these parallel channels of information give rise to 52 discrete aspects of somatosensation is not known. 53 To better understand somatosensory coding, it is critical to identify the different 54 channels of output from the spinal cord. In this regard, one of the key challenges is that 55 most spinal output neurons respond to several types of stimuli. For instance, the majority 56 of temperature-responsive spinal output neurons also respond to mechanical input 57 Similarly, all spinal output neurons that appear to be tuned for itch also respond to the 59 ( Figures 2D, 2E and 2F) or heat ( Figures 2G, 2H and 2I). These data suggest that cold-117 selective SPB neurons receive input from only one modality, cold. 118

Cold-selective neurons are unresponsive or only weakly responsive to Substance P. 119
It has previously been shown that the majority of SPB neurons express Tacr1 (also 120 known as the Neurokinin 1 Receptor), which is the receptor for Substance P (also known 121 as Tac1) (Cameron et al., 2015). We therefore assessed whether cold-selective SPB 122 neurons contained functional Tacr1 to determine whether they belong to this category. 123 Sixteen retrogradely-labeled SPB neurons were studied in voltage clamp mode, and 124 neurons were considered to express functional Tacr1 if an inward current was induced by 125 bath-applied Substance P (2 µM). Interestingly, cold-selective SPB neurons showed little 126 to no inward current in response to Substance P ( Figures 3A and 3C). In contrast, the 127 majority of other lamina I SBP neurons (i.e., those that were not cold-selective) showed 128 strong inward current to Substance P ( Figures 3B and 3C) neurons with respect to resting membrane potential ( Figure 4A). However, we found that cold-selective neurons had significantly lower membrane capacitance ( Figure 4B) and 141 higher membrane resistance ( Figure 4C). These findings raised the possibility that cold-142 selective neurons are smaller in size than the others. To address this idea directly, we 143 measured the soma area of SPB neurons that were filled with Alex 488 during recording. 144 In agreement with capacitance and resistance measurements, the soma size of cold-145 selective SPB neurons was significantly smaller than that of other lamina I SPB neurons 146 ( Figures 4D and 4E). 147

Cold-selective SPB neurons are distinct with respect to excitatory input 148
The finding that cold-selective SPB neurons have significantly higher membrane  These findings reinforce the idea that cold-selective SPB neurons are part of a distinct 159 neural circuit that is quiet with respect to ongoing activity. 160

Cold-selective SPB neurons belong to distinct spinal circuits. 161
To further characterize this cold-selective circuit, we examined whether SPB neurons are This finding suggested that, as general rule, SPB neurons receive direct inhibitory input 181 from Pdyn neurons but not Nos1 neurons. 182 To ensure that this apparent absence of synaptic input from Nos1 neurons onto 183 SPB neurons was not simply a technical artifact, we next recorded random lamina I 184 neurons ( Figure 6F). We found that 6 of 15 random lamina I neurons showed IPSCs in 185 response to optogenetic stimulation of Nos-CreER neurons ( Figures 6F and 6G). 186 Moreover, the IPSC amplitude in random neurons was much stronger (median = 158.7 187 pA, n = 6) than that observed in the singe SPB neuron that received input (20.8 pA, n=1).
Thus, Nos1 neurons provide functional inhibitory input to some lamina I neurons, but not 189 SPB neurons as a general class. 190 Next, we analyzed the Pdyn input in more detail ( Figure 6I). We found that Our study provides evidence that cold is mediated by a subset of SPB neurons that have 202 many distinctive properties including little to no response to Substance P, low 203 capacitance, high resistance, small soma size, low basal drive, and a lack of inhibitory 204 input from either Pdyn or Nos1 inhibitory neurons. Thus, these cells represent a distinct 205 output channel through which cold information is conveyed from the periphery to the 206 brain (Figure 7). 207 The idea that cold is conveyed by a distinct output channel is not without 208 response to Substance P is consistent with the possibility that they may be pyramidal in 243 shape. In future studies, it will be important to perform detailed morphological analyses 244 and comprehensive neurochemical profiling to tests these predictions.

Stereotaxic injection of DiI
Four-to six-week-old mice were anesthetized with isoflurane and placed in a stereotaxic 285 apparatus. A small hole was made in the skull bone with a dental drill. A glass pipette 286 was used to inject 100 nl of FAST DiI oil (2.5 mg/ml; Invitrogen, Carlsbad, CA) into the 287 left lateral parabrachial area (relative to lambda: anteroposterior −0.5 mm; lateral 1.3 288 mm; dorsoventral −2.4 mm). The head wound was closed with stitches. After recovery 289 from the anesthesia, the animals fed and drank normally. DiI was injected at least five 290 days prior to electrophysiological recordings. 291

Semi-intact somatosensory preparation 292
Semi-intact somatosensory preparation was made as previously described with small 293 modification (Hachisuka et al., 2016). Briefly, young adult mice (5-9 weeks old) were 294 deeply anesthetized and perfused transcardially through the left ventricle with oxygenated 295 (95% O2 and 5% CO2) sucrose-based artificial cerebrospinal fluid (ACSF) (in mM; 234 296 sucrose, 2.5 KCl, 0.5 CaCl2, 10 MgSO4, 1.25 NaH2PO4, 26 NaHCO3, 11 Glucose) at 297 room temperature. Immediately following perfusion, the skin was incised along the 298 dorsal midline and the spinal cord was quickly exposed via dorsal laminectomy. The right 299 hindlimb and spinal cord (~C2 -S6) were excised, transferred into Sylgard-lined 300 dissection/recording dish, and submerged in the same sucrose-based ACSF, which 301 circulated at 50 ml/min to facilitate superfusion of the cord. Next, the skin innervated by 302 the saphenous nerve and the femoral cutaneous nerve was dissected free of surrounding 303 tissue. L2 and L3 DRG were left on the spine. Dural and pial membranes were carefully 304 removed and spinal cord was pinned onto the Sylgard chamber with the right dorsal horn 305 facing upward. Following dissection, the chamber was transferred to the rig. Then the 306 preparation was perfused with normal ACSF solution (in mM; 117 NaCl, 3.6 KCl, 2.5 307 5% CO2 at 31 °C. Tissue was rinsed with ACSF for at least 30 min to wash out sucrose. 309 Thereafter, recordings were performed for up to 6 h post-dissection. 310

Patch clamp recording from dorsal horn neurons 311
Neurons were visualized using a fixed stage upright microscope (BX51WI Olympus 312 microscope, Tokyo, Japan) equipped with a 40x water immersion objective, a CCD 313 camera (ORCA-ER Hamamatsu Photonics, Hamamatsu City, Japan) and monitor screen. The data were low-pass filtered at 2 kHz and digitized at 10 kHz with an A/D converter 326 (Digidata 1322A, Molecular Devices) and stored using a data acquisition program 327 (Clampex version 10, Molecular Devices). The liquid junction potential was not 328 corrected. 329

Natural stimulation to the skin 330
To search for a cell's receptive field, a firm brush or a 4 g von Frey filament was applied 331 systematically over the skin. If no response to mechanical stimulation was observed, then 332 hot (50 °C) or cold (0 °C) saline was applied in pseudorandom order across the skin. 333 Once a receptive field was located, stimuli were reapplied directly to the receptive field 334 for 1 s. For mechanosensitive neurons, a variety of mechanical stimuli were applied 335 (small firm paintbrush and/or von Frey filaments (1, 2 and 4 g), but these data were 336 pooled for in this study for simplicity. Thermal stimulation was applied using 1 ml of hot 337