Touch and Pain

Capsaicin, the molecule that makes chili peppers burn, carries no heat. Yet when the cloned rat capsaicin receptor is put into cells that normally lack it, both capsaicin and heat above about 43 °C open the same ion channel.

So the burn is a reading taken by a sensory neuron. This lesson follows such readings from skin to brain, and then asks how the nervous system turns them up and down.

One distinction first. Nociception is the neural encoding of a damaging or potentially damaging stimulus. Pain is an experience. The notes to the revised 2020 definition of the International Association for the Study of Pain state that pain cannot be inferred solely from activity in sensory neurons (Raja and colleagues, 2020). A firing nociceptor or a withdrawing mouse paw is evidence about nociception. It is not a report of pain.

Four receptors, three fibre sizes

Touch and pain from the body start in neurons whose cell bodies sit in the dorsal root ganglia beside the spinal cord (a ganglion is a cluster of neuron cell bodies outside the brain and cord). These neurons arise from the neural crest, and many nociceptors depend on nerve growth factor while they develop (Trk and p75). The face uses trigeminal pathways, left aside here.

In the hairless (glabrous) skin of the human hand, four kinds of low-threshold mechanoreceptor afferent are recognised. Each has a receptive field, the patch of skin where a stimulus changes its firing, and an adaptation rate, how fast its firing falls during a held stimulus. Both came up in sensory coding. Think of four reporters: two describe what is there, two what is changing.

• Merkel cell afferents (SAI) adapt slowly and have small receptive fields. Single human SAI afferents can resolve spatial detail down to about 0.5 mm.
• Meissner corpuscle afferents adapt rapidly and also have small fields. They respond best to low-frequency vibration, the range of flutter and slip.
• Pacinian corpuscle afferents adapt faster still and have large fields. They respond best to high-frequency vibration, with a peak near 250 Hz in reviews of fingertip data. These are bands of best sensitivity, not cutoffs.
• SAII afferents adapt slowly, have large fields and are particularly sensitive to skin stretch. They are conventionally assigned to Ruffini endings, but in the review by Abraira and Ginty (2013) that pairing has only been postulated, without direct evidence.

Axons are also sorted by diameter and myelin, which set conduction velocity, as the action potential lesson explains. Large myelinated Aβ fibres carry most of these touch signals. Thinly myelinated Aδ fibres, often summarised at roughly 5 to 30 m/s, and unmyelinated C fibres, at roughly 0.5 to 2 m/s, carry most nociceptive and thermal signals. Most, not all: some C fibres are low-threshold touch afferents (C-LTMRs), not nociceptors. Speed does not tell you what a fibre reports.

Two pathways, two crossings

Touch and pain enter through the same dorsal root and then part company. Ipsilateral means on the same side as a reference point, and contralateral means on the opposite side. A tract is a bundle of axons inside the brain or spinal cord, and a nucleus is a cluster of neuron cell bodies inside them. In the spinal cord, dorsal means toward the back.

The dorsal column-medial lemniscus pathway carries fine touch, vibration and position sense from the body. The first-order axon enters the cord and ascends, without synapsing, in the dorsal column on the same side. It synapses in the gracile or cuneate nucleus of the caudal medulla. Axons from those nuclei cross the midline as internal arcuate fibres, form the medial lemniscus, and ascend to the thalamus, which relays to primary somatosensory cortex. This pathway crosses in the medulla.

The spinothalamic pathway, the main part of the anterolateral system, carries pain and temperature. The first-order axon synapses in the dorsal horn near its entry. The second-order axon crosses through the anterior white commissure, within a segment or two, and ascends in the contralateral anterolateral cord to the thalamus. This pathway crosses in the spinal cord. Both are major routes, not the whole of touch or pain processing, and where they cross lets you locate a lesion.

Worked problem, part (a). A clean cut through the right half of the spinal cord at the tenth thoracic segment (a hemisection at T10). What is lost, and on which side?

1. Dorsal columns. Fibres from the right side of the body below T10 ascend in the right dorsal column and have not crossed yet. Fine touch, vibration and position sense are lost on the right, below T10.
2. Spinothalamic tract. The right anterolateral cord at T10 carries axons that have already crossed from the left. Pain and temperature are lost on the left. Fibres entering just below the cut cross a segment or two higher, and some cross above it and escape, so the left-sided loss begins a little lower, from about T11 to T12 down.
3. Corticospinal tract. The lateral corticospinal tract crossed in the caudal medulla, at the pyramidal decussation (see the basal ganglia lesson), so in the cord it serves its own side. Weakness is on the right, below the lesion.

Touch and strength are lost on one side and pain on the other: the Brown-Séquard pattern.

Worked problem, part (b). Now speed. Take an illustrative path of 1.00 m from foot to cord, and assume constant velocity and conduction alone.

t = L / v

• Aβ at 50 m/s: 1.00 m ÷ 50 m/s = 0.020 s = 20 ms.
• Aδ at 10 m/s: 1.00 m ÷ 10 m/s = 0.100 s = 100 ms.
• C at 1.0 m/s: 1.00 m ÷ 1.0 m/s = 1.0 s = 1,000 ms.
• Aδ to C difference: 1,000 ms − 100 ms = 900 ms. Check: 1.00 m × (1/1.0 − 1/10) s/m = 0.900 s.

The velocities are illustrative, with the Aδ and C values inside the summarised ranges. This is conduction only, leaving out transduction, synapses and central processing, so it is not a measured interval between a sharp first pain and a later burning one.

Channels that feel heat, cold and damage

Nociceptors end as free nerve endings, so transduction happens in the terminal membrane. The first molecular handle came from expression cloning. Caterina, Julius and colleagues (1997) expressed pools of cDNA from rat dorsal root ganglia in HEK293 cells and screened for capsaicin-evoked calcium influx. One cDNA, named VR1 and now called TRPV1, made the cells respond. It encodes a nonselective cation channel, and in heterologous cells it opened with heat above about 43 °C. In Xenopus oocytes expressing it, capsaicin had an EC50 of 711.9 nM. Both numbers belong to the cloned rat channel in a foreign cell, not to human perception. How TRP and Piezo channels gate belongs to ion channels as receptors; here the question is what each one reports.

• TRPV1 responds to capsaicin, noxious heat and protons. Tominaga and colleagues (1998) found that protons lowered the temperature threshold of cloned VR1, and at pH 5.9 or below it could be activated at room temperature.
• TRPM8 is the menthol receptor, cloned from rat trigeminal neurons as CMR1 and activated by temperatures in the cool-to-cold range (McKemy and colleagues, 2002).
• TRPA1 is activated by thiol-reactive irritants of diverse structure through covalent modification of cysteines in its N-terminus (Hinman and colleagues, 2006, heterologous cells).
• Piezo2 is a mechanically activated channel needed for gentle touch in mice (Ranade and colleagues, 2014). Merkel cells use it too (Woo and colleagues, 2014).

Whether the animal needs a channel is a separate question. Caterina and colleagues (2000) found that mice lacking VR1 kept normal responses to noxious mechanical stimuli, so TRPV1 is not “the pain receptor”. Szczot and colleagues (2018) found that Piezo2-knockout mouse sensory neurons still responded robustly to noxious pinch, while people with loss-of-function PIEZO2 variants failed to develop painful touch after the tested skin inflammation. Piezo2 carries gentle touch, including touch that hurts after injury; something else also senses pinch, although in mice Piezo2 contributes to part of the response to painful mechanical stimuli (Murthy and colleagues, 2018).

Turning pain up and down: sensitization, gates and descending control

The gain on pain changes at three levels.

At the ending. Injured tissue releases mediators such as bradykinin and protons. In cultured small sensory neurons from neonatal rats, bradykinin increased the heat-activated inward current, and activators of protein kinase C mimicked it (Cesare and McNaughton, 1996), the pathway covered in PLC, IP3, DAG and PKC. Two terms name the results. Hyperalgesia is greater pain from a normally painful stimulus. Allodynia is pain from a normally innocuous one, such as a light touch.

In the dorsal horn. Nociceptor input can trigger a prolonged but reversible increase in the excitability and synaptic efficacy of central nociceptive neurons, called central sensitization (reviewed by Woolf, 2011). It can involve NMDA receptor-mediated rises in postsynaptic calcium, a form of the plasticity in synaptic plasticity. Convergence matters too. In isolated thoracic spinal cord from 10- to 14-day-old rats, somatic and visceral C fibres converged monosynaptically onto individual lamina I neurons (Luz and colleagues, 2015). Such convergence is a substrate for referred pain, pain felt away from its source, though not shown to be the whole explanation.

Melzack and Wall’s gate control theory (1965) proposed that substantia gelatinosa cells inhibit afferent terminals, with large-fibre activity increasing and small-fibre activity decreasing that inhibition. Touch input would act like a volume knob on nociceptive transmission. That wiring is not settled. Duan and colleagues (2014) tested one form of gating in mice by genetically ablating six spinal neuron lineages: dynorphin-lineage inhibitory neurons were required to gate mechanical pain, and loss of somatostatin-lineage neurons markedly impaired mechanical withdrawal. That supports spinal gating by particular populations, not the 1965 diagram as drawn.

From the brainstem. The periaqueductal gray (PAG) of the midbrain supplies top-down input to the rostral ventromedial medulla (RVM), which projects to the dorsal horn. In rodent recordings, functionally defined RVM OFF cells suppress and ON cells facilitate pain-related processing. Control runs both ways: in mice, activating GABAergic RVM projections to the dorsal horn facilitated mechanical nociception, because they inhibited enkephalin- and GABA-releasing interneurons (François and colleagues, 2017). Mu-opioid receptor agonists act at several of these sites (transmitter systems). How the functional ON and OFF classes map onto molecularly defined cells remains open (De Preter and Heinricher, 2024).

Design problem

A cDNA from rat sensory neurons makes HEK cells respond to capsaicin and to heat above about 43 °C. Does this show how mammals sense noxious heat? Design the test.