**4. Pathological and physiological relevance of redox signal-regulated TRPA1 activation**

TRPA1 is abundantly expressed in a portion of primary sensory neurons and mediates noxious pain sensation evoked by ROS to avoid deleterious DNA damage. In an inflammatory milieu, phagocytes release large amounts of ROS that promote the production of endogenous lipid compounds, including the RCS 4-HNE and 15d-PGJ2 . These compounds activate TRPA1 to cause inflammatory pain [36, 47].

TRPA1 is readily modified by electrophilic compounds. An analysis of the redox potentials of reactive disulfides revealed that TRPA1 has the lowest redox potential threshold among TRPA1, TRPV1, TRPV2, TRPV3, TRPV4, and TRPC5, indicating its high sensitivity to electrophilic compounds [48]. Thus, TRPA1 could act as a hyperoxia sensor in vagal and sensory nerves (**Figure 3**). Heterologously expressed TRPA1 was activated through covalent modification of cysteines at high oxygen conditions in excess of atmospheric oxygen concentration (20% O2 ) (hyperoxia), and the activation was diminished by point mutation of TRPA1 cysteine residues (Cys633Ser and Cys856Ser) (**Figure 1B**). In addition, hypoxic conditions also activated TRPA1, which could be suppressed by overexpression of recombinant prolyl hydroxylase (PHD). The oxygen-sensitive PHD catalyzes proline hydroxylation of TRPA1 under normoxic conditions that in turn inhibits channel function. A PHD inhibitor activated TRPA1 even under normoxic conditions, suggesting the involvement of proline hydroxylation in TRPA1 activation. Under hypoxic conditions, basal enzymatic activity of PHD is inhibited and TRPA1 is activated upon relief of proline hydroxylation-mediated inhibition. A proline residue located in the TRPA1 N-terminus (Pro394) is suggested to be involved in proline hydroxylationmediated TRPA1 inhibition (**Figure 1B**). Indeed, vagal nerves isolated from wild type, but not TRPA1-knockout (KO) mice, showed [Ca2+] **i** increases respond to both hyperoxic and hypoxic conditions. These data suggest that TRPA1 could function as an O2 sensor to control respiratory, cardiac and vascular functions through cysteine oxidation by ROS and proline modification by the oxygen-sensitive enzyme PHD. Proline hydroxylation of TRPA1 is also considered to participate in cold sensitivity of this channel, although this possibility remains a subject of debate (**Figure 3**). Temperature sensitivity of mammalian TRPA1 varies among species [49] in

**Figure 3.** TRPA1 regulation by Cys-oxidation (Cys-Ox) and Pro-hydroxylation at different oxygen concentrations. PHD, proline hydroxylase.

**Figure 2.** (A) Scheme showing the activation cascade of TRPM2 by reactive oxygen species (ROS). (B) H2

O2

points indicated in the upper column. (C) H2

208 Redox - Principles and Advanced Applications

The temperature-fura2 ratio relationship is plotted for H<sup>2</sup>

temperature thresholds for each group (Mean ± SEM) are shown.

heat-evoked single channel opening observed in inside-out single channel recordings. (a,b) Magnified traces at the time


O2

O2



O2


that human TRPA1 and monkey TRPA1 are not activated by cold stimulus whereas rat TRPA1 and mouse TRPA1 are. Mutation of Gly878 in S5 of rat TRPA1 to Val that is present in human TRPA1 abolished cold sensitivity. A proline hydroxylation-deficient mutant of human TRPA1 (Pro394Ala) was activated by cold stimulation in the presence of low concentrations of H2 O2 (0.1 μM), but wild-type TRPA1 showed no activation [50]. PHD inhibitors could result in activation of wild-type TRPA1 by cold temperature, which is attenuated by mitochondrial ROS scavenging. H2 O2 -evoked responses were significantly larger for Pro394Ala TRPA1 than for wild type, and a PHD inhibitor increased the response of wild-type TRPA1. In contrast, mouse wild-type TRPA1 was activated by cold temperature and effects of the PHD inhibitor and ROS scavengers were also observed in a mouse TRPA1 clone, confirming that the inhibitory effect of proline hydroxylation was conserved between human and mouse. Several chemotherapeutic agents can cause cold allodynia [51–53]. For instance, the drug oxaliplatin potentiates H2 O2 induced responses in mouse DRG neurons possibly by inhibiting PHD [50].

TRPA1 is also reported to play a physiological role in artery vasodilation, although its activity varies among different types of vascular beds. In cerebral arteries, TRPA1 expressed in the endothelium mediates "endothelium-dependent" vasodilation. ROS such as superoxide anions (O2 − ) and hydrogen peroxide (H2 O2 ) are known to cause vasodilation, leading to increases in cerebral microcirculation [54]. TRPA1 mRNA expression was observed in the endothelium of cerebral arteries, but not in peripheral vascular beds in renal, coronary and mesenteric arteries [55]. TRPA1 protein in the endothelium of cerebral arteries preferentially colocalizes with NOX2, a ROS-generating enzyme, in fenestration of internal elastic lamina where plasma membranes of endothelium and smooth muscle cell are in close contact. ROS generated by NADPH-induced NOX activity led to cerebral artery vasodilation following TRPA1 activation. This vasodilation could be abolished in a variety of ways, including by a NOX inhibitor, catalase-mediated degradation of H2 O2 , deferoxamine, which inhibits the Fenton reaction that generates hydroxyl radical (OH<sup>−</sup> ), the TRPA1 inhibitor HC-030031 and TRAM34, a blocker of the intermediate conductance Ca2+-sensitive K channel (IK). Moreover, vasodilation could be mimicked by application of 4-hydroxy-nonenal (4-HNE), a product of lipid peroxidation. These results suggest that ROS-derived lipid peroxidation products activate TRPA1, leading to cytosolic Ca2+ elevation, which in turn activates intermediate conductance Ca2+-sensitive K channels (IK) and membrane hyperpolarization in the endothelium. This change in membrane potential is propagated through gap junctions to smooth muscle cells to promote additional vasodilation. On the other hand, in peripheral arteries, TRPA1 expressed in primary sensory neurons is reportedly involved in "neurogenic" vasodilation. TRPA1 expressed in sensory neurons likely mediates vasodilation in peripheral arteries because TRPA1 is not expressed in the endothelium of peripheral arteries [56]. Topical application of cinnamaldehyde, a TRPA1 agonist, onto mouse ears caused vasodilation in wild-type mice, but not in TRPA1KO mice. TRPA1 agonist-induced vasodilation could be attenuated by a CGRP antagonist, a nonselective NOS or a neuronal NOS (nNOS) inhibitor, suggesting the possible involvement of CGRP and NO release from sensory neurons. In addition, TRPA1 agonist-induced vasodilation is mediated by formation of superoxide and peroxynitrite.

Taken together, TRPA1 expressed in the endothelium of central arteries is involved in endothelial-dependent vasodilation, whereas TRPA1 expressed in vagal and primary sensory neurons functions in neurogenic vasodilation in peripheral arteries.
