Why a soft touch can hurt

The neurophysiological literature on nociceptive neurons with ascending projections from the spinal cord has been riven for almost 40 years by a dogmatic controversy: many investigators professed that pain is subserved by so-called ‘wide dynamic range’ (WDR) neurons that respond to both low- and high-threshold cutaneous stimuli, whereas others preferred cells with nociceptive-specific (NS) response properties (reviewed in: Craig, 2003). WDR proponents claimed that such cells fulfilled the once popular (but now discredited: Inui et al. 2006) ‘gate control theory’ and that they could easily explain the allodynia commonly described by neuropathic pain patients (that is, pain elicited by low-threshold contact, as in sunburn). On the other hand, proponents of NS neurons claimed that such cells could serve as ‘labelled lines’ that required no central integration to signal pain.

Attention has been focused in recent years on lamina I, which is the main output layer of the superficial dorsal horn, where the small-diameter afferents that innervate all tissues of the body terminate; lamina I contains an abundance of NS neurons (as well as neurons specific for cool, warm, itch, and so on), it contains the main concentration of spinal substance P-responsive neurons, and in primates it is a major source of the spinothalamic tract, which is classically associated with pain and temperature sensation. In rodents, which have a very diminutive spinothalamic projection, the major lamina I projection target is the brainstem parabrachial nucleus, consistent with the concept that the fundamental role of the small-diameter afferent/lamina I pathway is homeostasis (Craig, 2003). Molecular physiological evidence now indicates that rat lamina I spinoparabrachial neurons are normally mostly NS cells with few WDR cells, but that these proportions can be dramatically reversed by nerve injury, and most tellingly, that NS neurons can become WDR neurons following local glial activation or chloride channel disruption (Keller et al. 2007); this evidence provides a cogent role for NS-neurons-that-become-WDR-neurons in allodynia and hyperalgesia. Yet, such unmasked low-threshold input was presumed to be conveyed by large-diameter mechanoreceptors via an unknown polysynaptic route, because those fibres terminate in the deep dorsal horn, not the superficial dorsal horn.

Low-threshold C-fibre tactile (CT or ‘slow brush’) afferents have been known for many years to terminate in the superficial dorsal horn, but across dozens of studies, no recordings of lamina I projection neurons responsive to ‘slow brush’ had been reported. Nevertheless, recent functional imaging work revealed that the human lamina I projection terminus in insular cortex is activated not only by pain and temperature, but also by CT fibres (Björnsdotter et al. 2009). This evidence suggested strongly that lamina I projection neurons responsive to ‘slow brush’ must exist, consistent with a homeostatic role for CT fibres, which are thought to signal emotional safety (and promote bonding) in primates, rather than danger.

The quantitative evidence provided now by Andrew in this issue of The Journal of Physiology (Andrew, 2010) convincingly demonstrates that rats indeed have lamina I spinoparabrachial neurons responsive to CT fibres, and by extension, that primates must have lamina I spinothalamic neurons sensitive to CT fibres. The surprise is that the neurons Andrew identified were not ‘labelled lines’ for ‘slow brush’, like NS and cool, warm and itch lamina I cells, rather they were all WDR cells!

This finding has several potential ramifications. First, it suggests that the lamina I WDR neurons described in previous reports may in fact have responded to low-threshold brushing due to CT fibres, rather than large-diameter mechanoreceptors. That would be consistent with the often reported fact that few, if any, lamina I neurons respond to electrical stimulation of Aβ-fibres; it would provide a second way in which WDR cells can be misinterpreted (in addition to the confusion caused by the activation of polymodal nociceptive cells by forceps at room temperature), and it would differentiate WDR lamina I neurons from the WDR neurons in lamina V, which receive Aβ input and subserve skeletal motor control, because they receive small-diameter afferent input and subserve homeostatic emotion, like other lamina I cells. Second, it suggests that the homeostatic safety signal ascribed to CT fibres requires comparison with NS activity in order to guide behaviour. That implies that the hedonic valence of intimate conspecific contact can easily reverse under inappropriate circumstances, like the gentle touch that irritates instead of soothes. Third, and most significant clinically, this finding suggests that allodynia may be due to unmasked low-threshold activation of NS cells by CT fibres, not by mechanoreceptors, which in turn implies that a selective pharmaceutical target might exist for the alleviation of allodynic pain. Indeed, new molecular physiological and behavioural evidence has serendipitously just appeared in Nature (Seal et al. 2009) which confirms the critical role of CT fibres, and thus the neurons that Andrew identified, in allodynic pain.

The absence of CT-specific lamina I neurons could of course be due to the anaesthesia or to sensitization by prior stimulation or the surgical trauma inherent in the acute experiments that Andrew performed, although earlier reports of ‘WDR’ lamina I trigeminothalamic neurons in awake monkeys suggest otherwise (see Craig, 2003). Perhaps recordings obtained from lamina I or parabrachial neurons while rats are grooming will provide evidence that confirms or denies the lack of specificity of CT-responsive lamina I neurons and the inherent potential shift in hedonic valuation that is implied. In the meantime, as we all learn eventually, it is important to keep an eye on someone who touches us softly.