Peripheral and spinal circuits involved in mechanical allodynia

Sensory information from nociceptors and touch receptors, such as Aβ-fiber mechanoreceptors and C-fiber low-threshold mechanoreceptors (C-LTMRs), is normally relayed and processed by separate neural circuits in the spinal cord.19 After nerve injury or inflammation, however, touch-related information is also relayed to nociceptive circuits in the superficial dorsal horn, which results in touch-evoked pain.16 Here, we depict the spinal circuits that facilitate this modality crosstalk and form the cellular basis of mechanical allodynia. A role of Aβ fibers in allodynia was originally proposed by human studies showing that compression block of Aβ fibers abolishes touch-evoked pain.3,7 Subsequent studies confirmed that after nerve injury, neurokinin-1 receptor (NK1R) expressing projection neurons in lamina I receive excitatory input from Aβ fibers via a preexisting polysynaptic connection that includes somatostatin (SOM) expressing interneurons, which also receive input from nociceptors.2,6,17 This connection is normally inhibited by dynorphin-/GAD67-expressing GABAergic interneurons in lamina II,4,6 the activity of which is controlled by polysynaptic input from Aβ fibers and by mono- and poly-synaptic inputs from Aδ- and C-fibers nociceptors. Another circuit links Aβ-fibers with lamina I NK1R− projection neurons through PKCγ+/SOM+ interneurons, central cells and SOM+ interneurons in outer lamina II. Information flow through this connection is controlled by feed-forward inhibition mediated by glycinergic and dynorphin-expressing interneurons.6,13 The Aβ-fiber subtypes that provide input to these circuits are unknown. Studies in humans14 and mice with impaired glutamate release from central synapses of C-LTMRs18 suggested that these afferents, which normally signal pleasant touch,10 might also signal mechanical allodynia. This hypothesis was, however, challenged by others who could not find differences in allodynia in mice lacking C-LTMRs11 and who were unable to induce tactile allodynia in patients lacking Aβ fibers.9 A possible explanation for this controversy is that C-LTMRs co-release the protein TAFA4, which counterbalances excitatory actions of glutamate.5 Thus, C-LTMR–specific loss of glutamatergic neurotransmission only,18 would result in a different phenotype than does the complete loss of C-LTMRs.11 C-fiber low-threshold mechanoreceptors project to lamina IIi and provide polysynaptic input to lamina I spinoparabrachial neurons,1 most of which express NK1R. They do not project to PKCγ+ neurons in the same lamina, which receive input from Aβ fibers.15 Putative C-LTMRs, as identified by means of conduction velocity and fiber diameter, also provide excitatory drive onto GABAergic islet cells in lamina II, which regulate information flow from C-fiber nociceptors via central and vertical cells to NK1R+ projection neurons.12 Existence of this C-LTMR–driven pain-inhibiting connection is consistent with recent findings suggesting an analgesic effect of C-LTMR signaling,5 which is reduced during tactile allodynia.9 Whether there is a direct crosstalk between C-LTMR– and Aβ-fiber–processing circuits or whether these 2 pathways only converge at the level of lamina I projection neurons is still unclear. However, given the evidence for a role of both fiber types in mechanical allodynia and considering the columnar organization of the spinal projections of Aβ fibers and C-LTMRs from the same skin area,8 a functional connection between these circuits seems highly likely.