Mitochondrial modulation-induced activation of vagal sensory neuronal subsets by antimycin A, but not CCCP or rotenone, correlates with mitochondrial superoxide production

Inflammation causes nociceptive sensory neuron activation, evoking debilitating symptoms and reflexes. Inflammatory signaling pathways are capable of modulating mitochondrial function, resulting in reactive oxygen species (ROS) production, mitochondrial depolarization and calcium release. Previously we showed that mitochondrial modulation with antimycin A, a complex III inhibitor, selectively stimulated nociceptive bronchopulmonary C-fibers via the activation of transient receptor potential (TRP) ankyrin 1 (A1) and vanilloid 1 (V1) cation channels. TRPA1 is ROS-sensitive, but there is little evidence that TRPV1 is activated by ROS. Here, we used dual imaging of dissociated vagal neurons to investigate the correlation of mitochondrial superoxide production (mitoSOX) or mitochondrial depolarization (JC-1) with cytosolic calcium (Fura-2AM), following mitochondrial modulation by antimycin A, rotenone (complex I inhibitor) and carbonyl cyanide m-chlorophenyl hydrazone (CCCP, mitochondrial uncoupling agent). Mitochondrial modulation by all agents selectively increased cytosolic calcium in a subset of TRPA1/TRPV1-expressing (A1/V1+) neurons. There was a significant correlation between antimycin A-induced calcium responses and mitochondrial superoxide in wild-type ‘responding’ A1/V1+ neurons, which was eliminated in TRPA1-/- neurons, but not TRPV1-/- neurons. Nevertheless, antimycin A-induced superoxide production did not always increase calcium in A1/V1+ neurons, suggesting a critical role of an unknown factor. CCCP caused both superoxide production and mitochondrial depolarization but neither correlated with calcium fluxes in A1/V1+ neurons. Rotenone-induced calcium responses in ‘responding’ A1/V1+ neurons correlated with mitochondrial depolarization but not superoxide production. Our data are consistent with the hypothesis that mitochondrial dysfunction causes calcium fluxes in a subset of A1/V1+ neurons via ROS-dependent and ROS-independent mechanisms.


Introduction
The activation of nociceptive sensory nerves by noxious stimuli evokes defensive reflexes and pain, which serves to protect the body from further damage. Inflammation, infection, physical damage and exogenous irritants are capable of significant nociceptor activation. As such, excessive activation of this sensory subpopulation contributes to a wide-range of disease states including chronic pain, irritable bowel disease, colitis, arthritis and asthma [1,2]. Nociceptive sensory nerves are selectively activated by noxious stimuli due to their expression of a variety of receptors, including the allyl isothiocyanate (AITC)-sensitive transient receptor potential (TRP) ankyrin 1 (A1) channel and the capsaicin-sensitive TRP vanilloid 1 (TRPV1) channel [3,4]. Both TRPA1 and TRPV1 are nonspecific cation channels whose activation causes neuronal depolarization and action potential discharge. TRPV1 is expressed on the majority of nociceptive neurons [5][6][7], although it should be noted that some specific subsets of nociceptive Aδ-and C-fiber nociceptors lack TRPV1 [8][9][10]. TRPA1 is expressed on approximately 66% of TRPV1-expressing neurons and <10% of neurons lacking TRPV1 expression [5][6][7]11,12]. TRPA1 is activated by a wide variety of electrophilic compounds, including reactive oxygen species (ROS) and environmental pollutants [13][14][15][16]. TRPV1 is activated directly by capsaicin, extracellular protons and heat, and indirectly downstream of second messenger signaling [3,17].
One potential way in which inflammation may cause nociceptor activation is by inducing the dysfunction of mitochondria, which are densely packed within sensory nerve terminals [18,19]. Mitochondrial dysfunction encompasses events including morphology changes, ROS production, mitochondrial depolarization and calcium release. While catastrophic mitochondrial dysfunction induces apoptosis, mitochondrial dysfunction also initiates signaling events involved in cellular stress responses. Chronic inflammation induces changes in mitochondrial morphology associated with dysfunction [20], and there are reports of an increased rate of mitochondrial DNA mutations and altered activity of ROS scavengers such as glutathione and superoxide dismutase in inflammatory states [21][22][23]. Inflammatory signaling has been shown to directly inhibit complexes I, II, and IV of the mitochondrial electron transport chain (mETC), resulting in ROS production and mitochondrial depolarization [24][25][26][27].
We have previously demonstrated that mitochondrial dysfunction induced by antimycin A, an mETC complex III inhibitor, activates vagal sensory neurons and bronchopulmonary Cfibers via the gating of TRPA1 and TRPV1 [28]. Inhibition of complex III by antimycin A induces ROS production and mitochondrial depolarization [29,30], but the role of these events in mitochondrial dysfunction-induced neuronal activation has not been established. Here we investigated the influence of mitochondrial dysfunction on the activation of vagal sensory neurons using three mitochondrial agents: antimycin A, carbonyl cyanide m-chlorophenyl hydrazone (CCCP, mitochondrial uncoupling agent), and rotenone (mETC complex I inhibitor). Using live cell co-imaging, we simultaneously monitored cytosolic calcium (as a surrogate for neuronal activation) and either mitochondrial superoxide production or changes in mitochondrial polarization. We found that the three agents activated (i.e. caused significant calcium fluxes) only a subset of A1/V1+ neurons, yet there was no difference in mitochondrial superoxide production or mitochondrial depolarization between 'responding' and 'non-responding' A1/V1+ neurons. Nevertheless, we observed a significant correlation between antimycin Ainduced calcium responses and mitochondrial superoxide production in wild-type 'responding' A1/V1+ neurons, which was eliminated in TRPA1 -/neurons, but not TRPV1 -/neurons. CCCP-induced calcium responses did not correlate with either superoxide production or mitochondrial depolarization. Rotenone-induced calcium responses correlated with mitochondrial depolarization but not superoxide production. Our data are consistent with the hypothesis that mitochondrial dysfunction causes the activation of a subset of A1/V1+ neurons via ROS-dependent and ROS-independent mechanisms.

Mouse models
Female TRPV1 -/mice were mated with male TRPV1 -/mice. Female TRPA1 -/mice were mated with male TRPA1 +/mice. Genotype of the offspring was confirmed using polymerase chain reaction. Wild type C57BL/6J mice were purchased from The Jackson Laboratory. All experiments were performed with approval from the University of South Florida Institutional Animal Care and Use Committee (AAALAC #000434).

Neuronal dissociation
Male 6-12-week-old mice were euthanized by CO 2 asphyxiation followed by exsanguination. Vagal ganglia were isolated followed by an enzymatic and mechanic dissociation then individual neurons were plated onto poly-D-lysine and laminin coated coverslips as previously described [28]. Neurons were incubated at 37˚C in antibiotic-free L-15 media supplemented with 10% fetal bovine serum and used within 24 hours.

Imaging analysis
Images were analyzed using Nikon Elements (Nikon, Melville, NY). Each individual neuron was analyzed separately using a region of interest (ROI) that encompassed the entire intracellular region. If required, ROIs were tracked over time to account for any cellular movement. For each imaging method (Fura-2AM, MitoSOX, and JC-1), ROIs with an unstable baseline, noisy baseline or insufficient loading were eliminated from the analysis. For Fura-2AM imaging, cells which failed to exhibit an increase in [Ca 2+ ] i to both AITC & capsaicin and KCl challenges (> 30% the ionomycin maximal response) were eliminated. Additionally, neurons that possessed an initial JC-1 red/green (525nm/610nm) ratio under 0.5 were eliminated as this indicated mitochondrial depolarization prior to the start of the experiment. For statistical analyses, we analyzed the maximal Fura-2AM, JC-1 and MitoSOX responses during the three-minute drug treatment (see below for calculations).
Relative changes in [Ca 2+ ] i were determined ratiometrically (R) using Fura-2 fluorescence: 340nm/380nm. As such the impact of variations in Fura-2AM loading are negated. The average calcium response to the combined AITC and capsaicin was used to classify A1/V1+ vs. A1-V1neurons: A1/V1+ neurons had an R AITC/Caps > (R bl + 2 Ã SD bl ), where R AITC/Caps is the average 340/380 ratio during AITC/capsaicin treatment, R bl is the baseline 340/380 ratio prior to treatment, and SD bl is the standard deviation of R bl . In the time series analyses, responses are presented as the change of the 340/380 ratio (ΔR = R 1 -R 0 ). In the bar graphs, scatterplots and statistical analyses, the response to mitochondrial modulation was determined using the maximal Δ340/380 (ΔR max = Rmax-R 0 ) for each neuron. An A1/V1+ neuron was defined as 'responding' to a particular mitochondrial agent if the [Ca 2+ ] i ΔR max was greater than 3 Ã SD of the ΔR max of A1-V1-neurons: ΔR max(A1/V1+) > ΔR max(A1-V1-) + 3 Ã SD (A1-V1-) where ΔR max(A1/V1+) is the maximal response to mitochondrial modulation in an individual A1/V1+ neuron, ΔR max(A1-V1-) is the average maximal response in A1-V1-neurons, and SD (A1-V1-) is the standard deviation of ΔR max(A1-V1-) . We have evaluated the effect of Fura-2AM loading on the calcium responses to mitochondrial modulation in these experiments and determined that there is no correlation between Fura-2 intensity at 380nm at baseline, an indicator of Fura-2AM loading, and the magnitude of ΔR max (data not shown). This suggests that the effects of calcium buffering by Fura-2 are negligible in the dissociated vagal neurons used in these studies.
Mitochondrial superoxide production was analyzed using MitoSOX fluorescence (F, excitation at 535nm). Given that the reaction of superoxide with MitoSOX red is irreversible and that baseline physiologic ROS production by mitochondria will vary between individual neurons, we have chosen to correct for the slope in the time series studies using F 1(adjusted) = F 1 -(S 0 Ã t), where F 1 is the raw fluorescence at a given timepoint (in arbitrary fluorescent units), S 0 is the slope of the baseline (arbitrary fluorescent units/s) and t is time of the F 1 measurement (in seconds). As MitoSOX loading varies between individual neurons, we have normalized the slope corrected data: F 1(adjusted) /F 0(adjusted) , where F 1(adjusted) is the slope corrected fluorescence at a given timepoint and F 0(adjusted) is the average of the slope corrected baseline. In order to determine the percentage of neurons that exhibited meaningful mitochondrial superoxide production compared to vehicle, we calculated the percentage of neurons whose individual maximal F 1(adjusted) /F 0(adjusted) was greater than the average maximal F 1(adjusted) /F 0(adjusted) for vehicle treatment + 3 Ã standard deviation of the maximal F 1(adjusted) /F 0(adjusted) for vehicle treatment.
Mitochondrial polarization was determined using JC-1 fluorescence at both the green (525nm, which measures the monomeric form of JC-1) and red (610nm, which measures the aggregated form of JC-1) emission wavelengths. JC-1 accumulates in negatively charged mitochondria and forms 'red' aggregates. Upon mitochondrial depolarization, JC-1 disperses into the cytosol as the 'green' monomeric form. As mentioned above, the 525nm/610nm ratio was used to identify neurons with depolarized mitochondria at baseline. However, we chose to determine the effect of mitochondrial modulation on mitochondrial polarization using only the 'green' monomeric form (F, 525nm fluorescence), as JC-1 fluorescence of 'red' aggregates is susceptible to quenching, prone to oxidation and requires~90 minutes to reach equilibrium (versus~15 minutes for monomers) [34]. Data were normalized (F 1 /F 0 ) where F 1 is the emission fluorescence at 610nm at a given time point and F 0 is the baseline emission fluorescence at 610nm. This normalization process allowed for the calculation of the relative change of mitochondrial membrane potential independent of baseline polarization and JC-1 loading. In order to determine the percentage of neurons that exhibited meaningful mitochondrial depolarization compared to vehicle, we calculated the percentage of neurons whose individual maximal F 1 /F 0 was greater than the average maximal F 1 /F 0 for vehicle treatment + 3 Ã standard deviation of the maximal F 1 /F 0 for vehicle treatment.

Statistical analyses
Statistical analyses were performed using R studio using maximal responses of the normalized responses observed during the three-minute drug treatment. To determine statistical differences in calcium response, mitochondrial polarization, and ROS production between mitochondrial agents and specific neuronal populations we performed two-way ANOVAs followed by post hoc comparison using the Bonferroni correction. To evaluate the correlation between cytosolic calcium and ROS production/mitochondrial depolarization, we performed a onetailed Pearson correlation. A chi square test for homogeneity was used to evaluate the percentage of responding neurons for each inhibitor. A p value of 0.05 was taken as the threshold for significance.
Nevertheless the increase in [Ca 2+ ] i in A1/V1+ neurons in response to antimycin A, CCCP and rotenone was significantly greater than that evoked by vehicle (p<0.05) (Fig 1I). This data is consistent with our previous studies demonstrating antimycin A evoked calcium transients that were dependent on extracellular calcium and TRPA1 and TRPV1 [28]. Only CCCP evoked an increase in [Ca 2+ ] i in A1-V1-neurons that was greater than vehicle (p<0.05)( Fig  1B, 1D and 1I). MitoSOX measurements in the dual imaging studies showed that antimycin A and CCCP evoked significant mitochondrial superoxide production compared to vehicle in A1/V1+ neurons, A1-V1-neurons and the total neuronal population (p<0.05) (Fig 1E, 1F, 1H and 1J) (p<0.05). Despite the differences in calcium responses between A1/V1+ and A1-V1-populations, there was no differences in the magnitude of mitochondrial superoxide production in A1/V1+ and A1-V1-neurons evoked by either antimycin A or CCCP (p>0.05) (Fig 1E, 1F and 1J). Rotenone failed to evoke significant mitochondrial superoxide production in either A1/V1+, A1-V1-or total populations compared to vehicle (Fig 1G, 1H and 1J). In general, rotenone had a minor inhibitory effect on the accumulation of MitoSOX fluorescence, which was more pronounced in the A1/V1+ population compared to the A1-V1-population (p< 0.05) (Fig 1G and 1J).
Scatterplot analysis of the calcium and MitoSOX responses (Fig 2) showed that mitochondrial superoxide production in response to antimycin A and CCCP was highly variable in all neuronal populations (i.e. 'responding' A1/V1+, 'non-responding' A1/V1+, and A1-V1groups). Overall, only 40% and 23% of neurons exhibited an increase in mitochondrial superoxide production following treatment with antimycin A and CCCP, respectively (compared to 1.6% for vehicle, see methods for calculation). Antimycin A-evoked calcium responses correlated with mitochondrial superoxide production in the responding A1/V1+ population and in the total A1/V1+ population (p<0.05) (Fig 2A, Table 1). However, there was no correlation between calcium and MitoSOX responses for CCCP in any neuronal population (p>0.05) ( Fig 2B, Table 1). Rotenone failed to evoke mitochondrial superoxide production in any neuronal population (0%) and as such there was no correlation between calcium and MitoSOX responses (p>0.05) (Fig 2C, Table 1).

Discussion
Nociceptive activation is a hallmark of inflammation in multiple organs, resulting in debilitating symptoms (e.g. pain) and reflexes (e.g. vasodilation, edema, bronchospasm, etc.) [36][37][38]. While inflammatory signaling pathways involving TNFα, ceramides and neurotrophins inhibit complexes of the mETC and induce mitochondrial dysfunction (ROS production, depolarization, and calcium release) [24][25][26][27], the mechanisms linking these factors to C-fiber activity are largely unknown. In a previous study, we demonstrated that inhibition of mETC complex III with antimycin A activated bronchopulmonary C-fibers via the activation of TRPA1 and TRPV1 [28]-Ca 2+ -permeable cation channels selectively expressed on many nociceptive sensory neurons [7,28,39]. However, the influence of specific aspects of mitochondrial function on TRPA1 and TRPV1 activation remains unknown. Here, we investigated the correlation of  Table 2.
https://doi.org/10.1371/journal.pone.0197106.g006 Correlation of mitochondrial dysfunction and sensory neuron activation mitochondrial depolarization and superoxide production on calcium fluxes in dissociated vagal neurons following mitochondrial modulation. As inflammatory signaling inhibits the mETC at multiple sites, we performed these experiments using three mitochondrial agents: antimycin A (complex III inhibitor), rotenone (complex I inhibitor), and CCCP (mitochondrial uncoupling agent). Here, we found that all three agents caused robust increases in [Ca 2+ ] i in the vagal A1/V1+ neuronal population. These data are consistent with our previous study that antimycin A evoked a selective increase in [Ca 2+ ] i in A1/V1+ vagal neurons that was abolished by the removal of extracellular Ca 2+ or by the genetic and pharmacological blockade of the nociceptive-specific ion channels TRPA1 and TRPV1 [28]. Given that TRPA1/TRPV1 were required for antimycin A-induced bronchopulmonary C-fiber activation [28], it is likely that the mitochondrial agent-evoked increases in [Ca 2+ ] i in A1/V1+ neurons observed in this study represents preferential activation of A1/V1+ neurons. However, not all A1/V1+ neurons 'responded' to mitochondrial modulation: approximately 60%, 35% and 35% of A1/V1+ neurons exhibited an increase in [Ca 2+ ] i to antimycin A, CCCP and rotenone, respectively. In general, antimycin A and CCCP evoked greater increases in [Ca 2+ ] i in 'responding' A1/V1+ neurons than rotenone. There was no correlation between responses in individual A1/V1+ neurons and their proximity to other neurons (data not shown), suggesting that calcium responses in each neuron was independent of the effects of mitochondrial modulation in other neurons.
Mitochondrial modulation can evoke ROS production, most notably at complexes I and III, when electrons leak and incompletely oxidize O 2 resulting in O 2 •-(superoxide) formation [40][41][42][43]. Superoxide is membrane impermeable and highly reactive, thus is capable of damaging DNA, proteins and lipids. In cells, superoxide dismutase rapidly converts superoxide to H 2 O 2 which is membrane permeable but less reactive [44]. Here, we found that antimycin A increased mitochondrial superoxide production acutely in both A1/V1+ and A1-V1-neurons. Furthermore, there was no difference in the mitochondrial superoxide production evoked by antimycin A between the 'responding' A1/V1+ and 'non-responding' neurons. Nevertheless, there was a significant correlation between antimycin A-evoked mitochondrial superoxide production and the magnitude of calcium responses in 'responding' A1/V1+ neurons. Our interpretation of these data is that mitochondrial superoxide produced by antimycin A can lead to A1/V1+ neuronal activation, but this is dependent on another unknown factor that is not present in all A1/V1+ neurons. It should be noted that only 40% of all neurons exhibited increased superoxide production in response to antimycin A, raising the possibility that some neurons are insensitive to this mitochondrial modulating agent. However, this seems unlikely given the observation that 97% of neurons exhibited antimycin A-induced mitochondrial depolarization. Both superoxide and H 2 O 2 are potent activators of TRPA1 [15,16]. Here, we found a significant correlation between antimycin A-evoked mitochondrial superoxide production and the magnitude of calcium responses in 'responding' trpv1 -/-A1+ neurons, whose responses are presumed to be mediated by ROS-sensitive TRPA1 alone [28]. Nevertheless, some trpv1 -/-A1+ neurons failed to be activated despite substantial superoxide production, thus again suggesting that some other unknown factor is required for mitochondrial ROS-mediated TRPA1 activation. Separate studies confirmed that many non-responding trpv1 -/-A1+ neurons were activated by exogenously-applied H 2 O 2 . Importantly, there was no correlation between antimycin A-evoked mitochondrial superoxide production and the magnitude of calcium responses in 'responding' trpa1 -/-V1+ neurons, whose responses are likely mediated by TRPV1 [28]. There is little evidence to suggest that TRPV1 is a major target of ROS and oxidative stress: TRPV1 is not activated by physiological concentrations of H 2 O 2 [16,35] or most electrophilic products of lipid peroxidation [45], although it is activated by high concentrations of 4-oxononenal [11]. Instead, TRPV1 is activated by PIP 2 hydrolysis, GPCR signaling and arachidonic acid metabolites [45][46][47][48]. Our data further suggests that TRPV1-mediated activation by antimycin A is not downstream of superoxide production.
Like antimycin A, CCCP evoked significant mitochondrial superoxide production in all neuronal populations, consistent with previous reports in isolated rat brain mitochondria [49], although other studies indicate that uncoupling agents such as CCCP do not increase ROS production [50,51]. However, there was no correlation between CCCP-induced mitochondrial superoxide production and the magnitude of calcium responses in 'responding' A1/V1+ neurons. Approximately 80% of neurons exhibited mitochondrial depolarization to CCCP, but there were no significant differences between the mitochondrial depolarization in 'responding' A1/V1+ and 'non-responding' neurons and there was no correlation between mitochondrial depolarization and the magnitude of calcium responses in 'responding' A1/V1+ neurons. We therefore conclude that factors other than mitochondrial superoxide production determine the CCCP-induced activation of A1/V1+ neurons and that mitochondrial depolarization alone is unable to account for these responses.
Rotenone failed to cause mitochondrial superoxide production in any neuronal population. Indeed, rotenone decreased baseline mitochondrial ROS production in A1/V1+ neurons. Previous studies have shown that rotenone can increase or decrease ROS production depending on metabolism of complex I versus complex II substrates [42,52]. Despite the lack of superoxide production, rotenone evoked an increase in [Ca 2+ ] i in approximately 35% of A1/V1+ neurons in this study. Although rotenone caused mitochondrial depolarization in approximately 38% of all vagal neurons, there were no significant differences between the mitochondrial depolarization in 'responding' A1/V1+ and 'non-responding' neurons. Nevertheless, there was a significant correlation between rotenone-evoked mitochondrial depolarization and the magnitude of calcium responses in 'responding' A1/V1+ neurons. This suggests that mitochondrial depolarization alone can lead to A1/V1+ neuronal activation, but again this is dependent on another unknown factor that is not present in all A1/V1+ neurons.
Our data indicates that mitochondrial modulators (in particular antimycin A and CCCP) evoke mitochondrial superoxide production in a limited percentage of neurons despite the evident mitochondrial depolarization in the majority of neurons. The present study used quantitative measurements of mitochondrial superoxide using MitoSOX, which is neither ratiometric nor reversible, making it possibly susceptible, despite our efforts to normalize the data, to variations in loading and baseline redox state. MitoSOX is targeted to the mitochondrial matrix, thus it is possible that some of the superoxide produced from complex I and III may evade MitoSOX detection following its release into the intermembrane space [41,53,54]. ROS produced within sensory neurons has 3 potential fates: reacting with signaling pathways, with endogenous antioxidant systems (e.g. glutathione) or with the ROS-sensitive dye. As such, the quantification of ROS production may be susceptible to significant error on a cellby-cell basis. Nevertheless, even a qualitative analysis suggests that (A) mitochondrial modulation does not evoke significant calcium responses in all A1/V1+ neurons and (B) significant ROS production does not necessarily evoke calcium responses in A1/V1+ neurons. This suggests that factors in addition to superoxide determine the TRPA1-and TRPV1-mediated calcium responses in A1/V1+ neurons. The identity of these factors is presently unknown, but it is possible that these either involve the activity of antioxidant systems or the spatial/temporal arrangement of mitochondria, TRP channels and signaling molecules. Our rotenone studies suggest that mitochondrial depolarization is able to contribute, but experiments with CCCP (and antimycin A) indicate that mitochondrial depolarization is not guaranteed to result in significant calcium responses in A1/V1+ neurons. How mitochondrial depolarization could contribute to A1/V1+ neuronal activation is not yet known. Mitochondrial depolarization causes calcium efflux from the matrix, and calcium is a regulator of multiple pathways such as phospholipases and protein kinases [55,56] that can alter TRP channel activity [45,[57][58][59].
Unlike antimycin A and rotenone, CCCP evoked a significant increase in [Ca 2+ ] i in A1-V1neurons, although these were much smaller than those in A1/V1+ neurons. As such, the threshold for determining the calcium 'responsiveness' of a particular neurons is greater for CCCP treatments. A1-V1-neurons do not express TRPA1 or TRPV1. Thus, the source of the CCCP-evoked calcium responses in A1-V1-is presently unknown. It is possible that calcium release from the mitochondrial matrix after robust mitochondrial depolarization contributes to these calcium transients [60], and this is consistent with (A) the significant calcium responses with CCCP in the nominal absence of extracellular calcium and (B) the significant correlation between CCCP-evoked mitochondrial depolarization and the magnitude of calcium responses in A1-V1-neurons (slope 0.035, p<0.05, data not shown).
The present study has investigated the correlation between mitochondrial ROS/depolarization and calcium transients evoked by mitochondrial modulators in vagal neurons. Our previous study identified the critical role of TRPA1 and TRPV1 in antimycin A-induced calcium responses and bronchopulmonary C-fiber activation [28]. Although the specific role of TRPA1 and TRPV1 has not been extensively investigated in the present study, our data is consistent with TRPA1 and TRPV1 mediating the mitochondrial modulator-evoked calcium responses. It is important to note, however, that TRP channel sensitivity itself is not static and that multiple intracellular pathways, including oxidative stress, can promote either sensitization or desensitization of TRPA1 and TRPV1 channels [3,[61][62][63][64]. Furthermore, TRPA1 and TRPV1 have been reported to functionally modulate each other's sensitivity [65,66]. Although not addressed here, it is possible that mitochondrial modulation may impact TRP function beyond channel activation.
In summary, we have shown that mitochondrial modulation by antimycin A-induced TRPA1-mediated calcium responses in vagal A1/V1+ neurons correlates with mitochondrial superoxide production, but that other factors including mitochondrial depolarization may contribute to A1/V1+ neuronal calcium responses via TRPV1 and downstream of complex I inhibition and mitochondrial uncoupling. It should be emphasized that the conclusions here are based upon correlations, which do not necessarily imply causation. Thus, we note that these studies do not show a conclusive role of superoxide production in the A1/V1+ neuronal calcium responses by mitochondrial modulators. Furthermore, we have focused our present study on dissociated vagal neurons and it is possible that mechanisms linking mitochondrial superoxide production and depolarization with TRP channel activation may be different within sensory terminals in vivo. More studies are needed to fully understand the events linking mitochondrial dysfunction and nociceptor activation.