Redox-dependent activation of protein kinase G1α contributes to transient receptor potential cation channel subfamily V member 1-mediated acute nociceptive pain behavior

ABSTRACT

Background

Sensory neurons relay the pain signals to the brain via the nociceptive system. Notably, reactive oxygen species (ROS) serve as signaling molecules in the somatosensory system; however, their contribution to sensing noxious stimuli remains poorly understood.

Methods

Herein, the role of protein kinase G (PKG)1α, which is highly expressed in sensory neurons and serves as a ROS target, was investigated in sensory neurons in the processing of acute nociceptive pain. Cys42Ser PKG1α-knock-in (PKG1α-KI) mice, devoid of redox-dependent PKG1α activation, were subjected to behavioral testing, ROS detection assays, gene expression experiments, and imaging analyses.

Results

Interestingly, PKG1α-KI mice showed reduced behavioral responses to noxious heat and the transient receptor potential cation channel subfamily V member 1 (TRPV1) agonist capsaicin. Moreover, capsaicin-induced sensory neuron stimulation upregulated ROS production and redox-dependent PKG1α activation. Calcium imaging results and patch-clamp recordings revealed that capsaicin-induced calcium flux and neuronal excitability was reduced in sensory neurons of PKG1α-KI mice.

Conclusion

Altogether, the findings of this study show the effects of redox-dependent PKG1α activation on capsaicin/TRPV1-mediated signaling in sensory neurons during acute nociceptive pain.

GRAPHICAL ABSTRACT

Introduction

Sensory neurons are crucial for detecting acute nociceptive stimuli and initiating important defense mechanisms against bodily damage and injury. Reportedly, reactive oxygen species (ROS) are critical signaling mediators in the processing of pain [1–3]. For instance, antioxidants or ROS scavengers could reduce pain behavior in different animal models of pain [4–6]. Additionally, deleting or inhibiting ROS-producing enzymes, such as nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) oxidases, can alleviate acute nociceptive or chronic pain behavior [7–13], suggesting a specific contribution of ROS to pain processing. However, despite the reported involvement of different ROS signaling pathways with chronic pain, the mechanisms underlying ROS actions in acute nociceptive settings remain elusive.

Protein kinase G (PKG)1α, an important ROS target in sensory neurons, can be activated by a redox-dependent disulfide bridge formation between the two cysteine 42 residues in the PKG1α homodimer, independently of cyclic guanosine monophosphate (cGMP) [14,15]. PKG1α has been implicated in chronic pain processing, as reflected by its upregulation following tissue damage and nerve injury and by considerably reduced responses of global and sensory neuron-specific PKG1-deficient mice in models of chronic pain [16–19]. Moreover, redox-dependent PKG1α activation is critically involved in signaling processes regarding peripheral nerve injury. Redox-insensitive Cys42Ser PKG1α-knock-in (PKG1α-KI) mice have demonstrated reduced neuropathic pain behavior after sciatic nerve injury [20] and altered regenerative capacity of sensory neurons [21]. However, unlike chronic pain, the potential contribution of ROS induced PKG1α activation in the sensing of acute noxious stimuli remains unelucidated.

This study aimed to explore the involvement of redox-dependent PKG1α activation in acute nociceptive signaling. Herein, the behavior of PKG1α-KI mice in multiple pain models was investigated. Additionally, cultured sensory neurons and dorsal root ganglia (DRGs) were comprehensively analyzed to identify ROS-dependent interactions of PKG1α and transient receptor potential cation channel subfamily V member 1 (TRPV1), an essential pain sensor.

Material and methods

Animals

The ‘redox-dead’ PKG1α-KI mice model (Prkg1tm1.1Arte) was generated as previously described [22]. Heterozygous mice on a C57BL/6N background were bred to obtain wildtype (WT) and PKG1α-KI mice for this study. In addition, some tissue expression studies and assays were performed with PKG1α-KI mice obtained from homozygous breeding and C57BL/6N mice (Charles River) or other WT animals of different breeding strains of our animal facility. Animals were housed under a 12/12 h light/dark cycle with free access to water and food. All experiments were performed per the International Guiding Principles for Biomedical Research Involving Animals (EC Directive 86/609/EEC) and Ethical Issues of the International Association for the Study of Pain and approved by the local Ethics Committee for Animal Research (Regierungspräsidium Darmstadt, Germany; V54-19 c20/15-F95/49, V54-19c20/15- FR/1012, and V54-19c18-FR/2025). Experiments were conducted between 8 am and 6 pm in experimental rooms at 22 ± 2°C and 55 ± 10% humidity.

Behavioral experiments

Behavioral experiments were performed by an observer blinded to the genotype of littermate male and female mice (age: 10–30 weeks).

Hot plate test

The hot plate test was performed as described previously [23]. Measurements at different temperatures (48, 50, or 52°C with cutoff times of 60, 40, and 20 s, respectively) were performed because the extent of activation of different heat receptors and the processing of thermal stimuli depend on stimulus intensity [24,25].

Thermal place preference (TPP) test

For the TPP test, two metallic plates caged with acrylic glass cylinders were connected by a bridge (Thermal Place Preference, Ugo Basile); baseline plate, 30°C and testing plate, 30°C (control) or 45°C. Mice were placed on the baseline plate, and their movement in the setup was tracked for 20 min using a video tracking system (VideoMot; TSE Systems). Their movements were analyzed using the TSE VideoMot software [26].

Cold plantar test

Implementation of the cold plantar test has been described in earlier studies [27,28]. The latency time was defined as the first lifting of the paw and measured at least five times with alternating paws for each mouse (cutoff time: 30 s and at least 5 min in between applications of one paw).

Tape response test

The tape response assay was performed as described previously [27]. The time until the first reaction, including biting and scratching of the tape or ‘wet-dog-shaking’, and the total number of reactions were measured for 5 min [29,30].

Von Frey tests

Mice were acclimatized in boxes with a metal grid floor for at least 60 min. Calibrated von Frey filaments (0.70–39.2 mN; 0.07–4.0 g; #37450-275, Ugo Basile) were used to stimulate the plantar side of the hind paws. The number of reactions (including paw withdrawal and flinching) up to 3 s after 10 applications were counted for all filaments [31,32]. Measuring the mechanical sensitivity threshold using von Frey filaments has been described before [33]. Six critical data points were measured after reaching the first turning point in filament strength [31], and with these the 50% threshold was calculated using an online algorithm (https://bioapps.shinyapps.io/von_frey_app/) [34].

Capsaicin, allyl isothiocyanate (AITC), and bradykinin tests

Mice were acclimated to an acrylic glass cylinder before the experiment and capsaicin (#M2028, Sigma-Aldrich; 820 μM (5 µg in 20 µL 2% dimethyl sulfoxide [DMSO] in phosphate-buffered saline [PBS;#14190144, Thermo Fisher])), AITC (#377430, Sigma-Aldrich; 500 μM (10 nmol in 20 µL 0.05% DMSO in PBS)), or bradykinin (#J63131, Thermo Fisher; 400 μM (4 nmol in 10 µL PBS)) was injected subcutaneously into the plantar side of one hind paw. After injection, mice were videotaped for 5 min (capsaicin, bradykinin) or 20 min (AITC) to record the time spent licking the injected paw. A vehicle control (20 µL of 2% DMSO in PBS) was injected into the contralateral paw at least 7 d after the compound injection.

In experiments with the ROS scavenger N-tert-butyl-α-phenylnitrone (PBN; #B7263, Sigma-Aldrich) [35], PBN (50 mg/kg in PBS) was intraperitoneally (i.p.) administered to mice 30 min prior the intraplantar (i.pl.) capsaicin injection. Behavioral observations were performed as described above.

Euthanasia and tissue preparation

Naïve mice were killed by CO₂ inhalation and DRGs (T1–T13, L1–L6) were excised to generate primary DRG neuron cultures as previously described [36].

Dorsal root ganglia (DRG) neuron cultures

For calcium imaging and ROS detection assays, DRGs were cultured in Neurobasal (NB) medium (#21103049, Thermo Fisher) supplemented with 2% B27 (#17504044, Thermo Fisher), 0.5 mM GlutaMax (#35050061, Thermo Fisher), and 1% penicillin/streptomycin (#15140122, Thermo Fisher) at 37°C and under 5% CO₂ overnight. For ROS detection assays, NB medium without phenol-red (#12348017, Thermo Fisher) was used to avoid background noise. For patch-clamp recordings, penicillin, streptomycin, or trypsin were not used to ensure membrane integrity.

Cellular ROS detection assays

Chloromethyl-2’,7'-dichlorodihydrofluorescein diacetate (CM-H₂DCFDA) assay

The cellular ROS indicator CM-H₂DCFDA (#C6827, Sigma-Aldrich) was used to measure intracellular ROS production in DRG neurons as previously described, with slight modification [37]. CM-H₂DCFDA was dissolved in 100% DMSO, and DRG neurons, cultured for 24 h and acclimatized to RT for 10 min in PBS, were stained for 30 min using 5 µM CM-H₂DCFDA in PBS. The cells were transferred into NB medium to allow recovery for 30 min. ROS production was measured in a live cell imaging setup at 490/520 nm (excitation/emission) in Ringer solution (145 mM NaCl (#S9888, Sigma-Aldrich), 1.22 mM CaCl₂·2H₂O (#5239.1, Carl Roth), 1 mM MgCl₂·6H₂O (#3532.1, Carl Roth), 5 mM KCl (#104936, Merck Millipore), 10 mM Glucose (#X997.1, Carl Roth), 10 mM HEPES (#6763, Carl Roth) in water; pH adjusted to 7.4 with NaOH (##35256-1L, VWR International)). Images were taken once every minute at 10 ms exposure time over 40 min in the dark. A perfusion system was used to perform a baseline measurement with Ringer for the first 10 min, followed by stimulation with H₂O₂ (#H1009, Sigma-Aldrich; 50 µM in Ringer; positive control), capsaicin (1 µM in Ringer), AITC (2 mM in Ringer), or Ringer (negative control) starting at minute 11. The measurements and analysis were performed using the NIS-Elements (Nikon) software package and fluorescence intensities were calculated as follows:

$$\mathrm{Fluorescence}\mathrm{increase}\left(\text{\%}\right)=\left(\left[\mathrm{value}\mathrm{minute}X-\mathrm{value}\mathrm{minute}11\right]/\mathrm{value}\min 11\right)\times 100$$

Amplex Red assay

Relative H₂O₂ levels were measured through Amplex Red/horse radish peroxidase (HRP) assay (#A22188, Invitrogen/Life Technologies, Darmstadt, Germany) per the manufacturer’s instructions. Capsaicin samples and H₂O₂ standard controls were incubated with 100 µM Amplex Red and 0.2 U/mL HRP for 60 min at 20°C, followed by transfer into a 96-well plate (µCLEAR, BLACK; Greiner Bio One, Kremsmünster, Austria). Fluorescence intensities were acquired at 540/595 nm (excitation/emission) using a microplate reader (Tecan Group Ltd., Männedorf, Switzerland).

Quantitative real-time reverse transcription polymerase chain reaction (RT–PCR)

DRGs (L1–L6) were used to extract total RNA using the innuPREP Micro RNA Kit (#845-KS-2030010, IST Innuscreen, Germany), quantified at 260 nm using a NanoDrop 2000c spectrophotometer (Thermo Scientific), and reverse transcribed into complementary DNA using the First Strand cDNA Synthesis Kit (#K1621, Thermo Fisher) with random primers per the manufacturer’s protocols. Quantitative real-time RT–PCR was performed on a CFX96 Touch Real-Time PCR Detection System (#1845097, Bio-Rad, Germany) equipped with a C1000 Touch Thermal Cycler (#1841100, Bio-Rad, Germany) using the iTaq Universal SYBR Green SuperMix (#1725120, Bio-Rad) and primers (Biomers, Germany). Specific primers were used to analyze expression levels of transient receptor potential cation channel subfamily V member 1 (Trpv1: forward: 5’-ACTCTTACCACACAGCAGCC-3’; reverse: 5’-GCCCAATTTGCAACCAGCTA-3’), transient receptor potential ankyrin 1 (Trpa1: forward: 5’-GGAAATACCCCACTGCATTGT-3’; reverse: 5’-CAGCTATGTGAAGGGGTGACA-3’), transient receptor potential cation channel subfamily M member 3 (Trpm3: forward: 5’-GGTGTGGCTTCAGGAGTACT-3’; reverse: 5’-CCAGATACTTGTTCACGCCG-3’) and Glycerinaldehyd-3-phosphat-dehydrogenase (Gapdh: forward: 5'-CAATGTGTCCGTCGTGGATCT-3’; reverse: 5'-GTCCTCAGTGTAGCCCAAGATG-3’) . Reactions were carried out in duplicates under following conditions: incubation at 95°C for 5 min, followed by 40 cycles of 30 s at 95°C and 60 s at 60°C, with a fluorescence readout after each cycle. The relative expression of target genes was calculated using the 2^−ΔΔCt method.

In situ hybridization (ISH)

For ISH, DRGs (L4–L5) were dissected, post-fixed, sectioned and stored at −70°C as previously described [38]. ISH was performed using a QuantiGene ViewRNA Tissue Assay (Thermo Fisher, Frankfurt, Germany) per the manufacturer’s instructions, including extensive washing steps [38]. The following specific probes were used: mouse messenger RNA of protein kinase G1 (Prkg1; #VB1-19757, Thermo Fisher, NCBI #NM_001013833.3), Trpv1 (#VB6-16610, Thermo Fisher, NCBI #NM_001001445.2), Trpa1 (#VB6-18246, Thermo Fisher, NCBI #NM_177781.4), Trpm3 (#VB6-18246-VT, Thermo Fisher, NCBI #NM_177781.4), and scramble control for type 1 (#VF1-17155, Thermo Fisher) and type 6 (#VF6-18580, Thermo Fisher).

Calcium imaging

Calcium imaging of DRG neurons cultured for 24 h was performed as previously described [39]. Neurons were loaded with Fura-2-AM-ester (5 µM; #50033, Biotium) in HBSS (#14175095, Thermo Fisher), transferred to the perfusion chamber, and continuously superfused with a physiological Ringer solution. Images taken every 2 s at two wavelengths (340 and 380 nm) were processed using the NIS-Elements software (Nikon). Baseline measurements were recorded using Ringer solution at a flow rate of 1–2 mL/min for 3 min. Cells were exposed to capsaicin (100 nM) or AITC (200 µM) dissolved in Ringer solution by bath perfusion for 30 s at RT. Next, cells were washed for 5 min with Ringer solution and finally stimulated for 30 s with 75 mM KCl (#104936, Merck Millipore) to identify viable neurons. A calcium response was defined when the 340/380 nm ratio, normalized to baseline, exceeded 20% of the baseline level. Acquired images are presented as the 340/380 nm ratio.

Patch-clamp assay

Whole-cell patch-clamp experiments of DRG neurons cultured for 24 h were performed. Neurons cultured on coverslips were transferred into a bath chamber (#RC-26G, Warner Instruments) containing a physiological extracellular solution (140 mM NaCl (#S9888, Sigma-Aldrich), 5 mM KCl (#104936, Merck Millipore), 2 mM CaCl₂ (#CN93.1, Carl Roth), 2 mM MgCl₂ (#M8266, Sigma-Aldrich), and 10 mM HEPES (#6763, Carl Roth) in water; pH adjusted to 7.4 with NaOH (#35256-1L, VWR International)). Cells were stained with 10 µg/mL Alexa Fluor 488-conjugated isolectin B4 (IB4; #I21411, Thermo Fisher) for 5 min at RT. For whole-cell patch-clamp recordings, smaller IB4-negative neurons were patched with borosilicate glass microelectrodes containing an intracellular pipette solution (140 mM KCl (#104936, Merck Millipore), 2 mM MgCl₂ (#M8266, Sigma-Aldrich), 5 mM EGTA (#3054.1, Carl Roth), and 10 mM HEPES (#6763, Carl Roth) in water; pH adjusted to 7.4 with KOH (#1.09108.1000, VWR International)) and were pulled using a micropipette puller (#P-97, Sutter Instruments). Patch-clamp recordings were performed using an EPC 9 amplifier with the Patchmaster software (HEKA Electronics) and analyzed using the Fitmaster software (version 2 × 91, HEKA Electronics). The protocols were as follows: 10 ms current injections starting at 0 pA with 20 pA increments up to 280 pA for evoked action potentials (APs); 1,000 ms current injections from 0 pA to 350 pA with 50 pA increments to measure AP-firing. Patch-clamp recordings with vehicle or capsaicin stimulation were performed using a timed protocol to ensure the comparability of measured cells. First, baseline vehicle measurements were performed by stimulating cells for 30 s with extracellular solution, starting 5 s before measuring evoked APs, followed by baseline AP-firing measurements after 45 s. Subsequently, these cells were flushed with 100 nM capsaicin using the Trio-245-L/SmartSquirt microinjection system (Science-Products, Hofheim) for 30 s. Evoked AP recordings were performed 5 s after starting the capsaicin stimulation, followed by AP-firing recordings after 45 s (Figure S3, A).

Western blotting

Redox-dependent PKG1α dimerization was analyzed using non-reducing Western blots as described before [20,23]. Briefly, DRGs (L1–L6) were rapidly dissected, stored in ice-cold PBS incubated in Accutase (500 µL; #A1110501, Thermo Fisher) for 30 min at RT, washed in PBS, and stimulated for 30 min at RT with 1 mM capsaicin or vehicle. Next, samples were further homogenized with Phosphosafe buffer (#71296, Merck Millipore) mixed in a 7:1 ratio with a protease inhibitor cocktail (Complete Mini; #11836170001, Sigma-Aldrich; 1 tablet dissolved in 1 ml water) and supplemented with 10% maleimide (#129585, Merck Millipore). Proteins were separated by SDS-polyacrylamide gel electrophoresis and blotted onto a 0.2 µm nitrocellulose membrane. The following antibody was used: rabbit anti-PKG1 (1:1000; #KAP-PK005-D, Enzo/Stressgene) dissolved in blocking buffer containing 0.1% Tween-20. After incubation with secondary antibodies for 1 h at RT, proteins were detected using an Odyssey Infrared Imaging System (LI-COR Bioscience), and band densities were quantified using the Image studio lite software (LI-COR Bioscience).

For reducing western blots samples were homogenized in Phosphosafe buffer combined with a protease inhibitor cocktail. The samples were subjected to western blotting and detection as described above. The following primary antibodies were used: rabbit anti-phospho-TRPV1 (1:1000; #PA5-64860, Invitrogen) or mouse anti-α-tubulin (1:500; #3144508, Merck Millipore).

PKG activity assay

PKG1 activity in DRGs in response to capsaicin was measured using the CycLex® cyclic GMP-dependent protein kinase assay kit (#CY-1161, MBL Internal Corporation) per the manufacturer’s instructions. DRGs (L1–L6) were dissected, pooled in ice-cold PBS, incubated with Accutase (500 µL; #A6964, Sigma-Aldrich) for 30 min and washed in PBS. Following the addition of samples to 100 µL of extraction buffer (20 mM Tris (#4855.2, Carl Roth), 150 mM NaCl, protease inhibitor cocktail, pH adjusted to 7.4), extracted protein samples were centrifuged at 9,500 g at 4°C for 10 min. Samples (10 µL) were stimulated with capsaicin (1 mM) or vehicle in 2.5 mM ATP containing kinase buffer provided by the kit for 30 min at RT. After washing, samples were incubated for 60 min with the antibody solution, and the reaction was stopped using the stop solution. Finally, photometric densities were analyzed at 450 and 540 nm per the manufacturer`s instructions.

Statistical analyses

The GraphPad Prism 9 software was used for statistical analyses. Herein, statistical power calculations were not conducted before the study; the sample sizes were based on similar studies [40,41]. Outliers were detected using the Robust regression and Outlier removal method with a 1% coefficient Q and were excluded from the statistical analysis, if applicable. For each data set, Anderson–Darlin, D’Agostino–Pearson, Kolmogorov–Smirnov, or Shapiro–Wilk tests were used to assess the normal distribution of data within groups and to select the appropriate parametric or non-parametric test. For normally distributed data, differences between two groups in behavioral tests were analyzed using the two-tailed unpaired Student’s t test. Statistical differences in behavior experiments with multiple time points were assessed using the two-way repeated-measure analysis of variance (ANOVA) with Bonferroni post-hoc test and multiple comparisons. In the ROS detection assay, an ordinary one-way ANOVA with Bonferroni post-hoc test and multiple comparisons was used to compare endpoint values of more than two groups. For patch-clamp recordings, paired t-tests were used to compare before–after values. Non-normally distributed values in calcium imaging experiments were analyzed using the Mann–Whitney test. For all tests, a P-value of <0.05 was considered to be statistically significant. Results for normally distributed data are expressed as the mean ± standard error of the mean, whereas those for non-normally distributed data are expressed as a median with interquartile range.

Results

Redox-dependent PKG1α activation modulates nociceptive signaling

Acute nociceptive behavior of redox-insensitive PKG1α-KI mice was investigated after inducing noxious stimuli via mice hind paws. Interestingly, after administration of the transient receptor potential vanilloid 1 (TRPV1) agonist capsaicin, the paw licking behavior was significantly reduced in PKG1α-KI mice compared to WT littermates (Figure 1A). In contrast, TRPA1 agonist AITC and the B2 receptor agonist bradykinin evoked a paw licking behavior in PKG1α-KI mice comparable to that in WT mice (Figure 1B and C). Administration of vehicle (20 µL 2% DMSO in PBS) did not induce any paw licking behavior (Supplemental Figure 1A), in line with other publications [42,43]. These data suggest that redox-dependent PKG1α activation contributes to nociceptive processing after the activation of TRPV1, but not TRPA1 or the B2 receptor.

Figure 1.

Reduced responses to capsaicin and noxious heat in Cys42Ser protein kinase G (PKG)1α-knock-in (PKG1α-KI) mice. (A) After intraplantar injection of capsaicin (820 μM in 20 µL 2% dimethyl sulfoxide [DMSO] in phosphate-buffered saline [PBS]),: paw licking was assessed for 5 min. Pain behavior was markedly reduced in PKG1α-KI mice compared with that in littermate wild-type (WT) mice (WT: mean = 33.5, SEM = 4.6, SD = 13.1; PKG1α-KI: mean = 12.4, SEM = 3.7, SD = 10.5; n = 8). (B and C) No difference in paw licking time was observed in the first 20 or 5 min between genotypes after injection of (B) allyl isothiocyanate (500 µM in 20 µL 0.05% DMSO in PBS) (WT: mean = 46.16, SEM = 9.9, SD = 22.1; PKG1α-KI: mean = 43.7, SEM = 8.7, SD = 21.2; n = 5–6) or (C) bradykinin (400 µM in 10 µL PBS) (WT: mean = 99.8, SEM = 19.1, SD = 42.8; PKG1α-KI: mean = 120.2, SEM = 15.6, SD = 38.1; n = 5–6), respectively. (D–F) In the hot plate test, PKG1α-KI mice showed markedly increased latency times at 50°C (E) and 52°C (F), but not at 48°C (D), compared to WT mice (48°C: WT: mean = 33.8, SEM = 3.0, SD = 11.0; PKG1α-KI: mean = 44.8, SEM = 7.7, SD = 23.1; 50 °C: WT: mean = 27.3, SEM = 1.4, SD = 5.7; PKG1α-KI: mean = 35.7, SEM = 2.0, SD = 8.1; 52°C: WT: mean = 15.6, SEM = 0.8, SD = 3.1; PKG1α-KI: mean = 20.3, SEM = 0.9, SD = 3.4; n = 9–16). (G) Results of the thermal place preference test in an observation period of 20-min revealed a significant increase in time (%) PKG1α-KI mice stayed on the 45°C testing plate, when compared to WT littermates (45°C: WT: mean = 9.1, SEM = 2.0, SD = 4.8; PKG1α-KI: mean = 17.6, SEM = 2.1, SD = 5.2; 30 °C: WT: mean = 82.7, SEM = 4.9, SD = 11.9; PKG1α-KI: mean = 74.8, SEM = 2.8, SD = 6.7; bridge: WT: mean = 8.1, SEM = 3.3, SD = 8.1; PKG1α-KI: mean = 7.6, SEM = 1.9, SD = 4.6; n = 6–7). (H) In the cold plantar test, latency times were similar in both genotypes (WT: mean = 11.4, SEM = 0.4, SD = 1.4; PKG1α-KI: mean = 11.8, SEM = 0.7, SD = 2.2; n = 11). (I) The tape response test did not show significant differences between PKG1α-KI and WT littermates when assessed for 5 min (WT: mean = 27.8, SEM = 1.8, SD = 6.0; PKG1α-KI: mean = 27.0, SEM = 2.9, SD = 9.7; n = 10–11). (J) Both genotypes demonstrated similar mechanical thresholds in the Von Frey up and down assay (WT: mean = 1.3, SEM = 0.2, SD = 0.7; PKG1α-KI: mean = 1.1, SEM = 0.2, SD = 0.5; n = 11). (K) After stimulation with Von Frey filaments of different forces (10 applications per filament), no significant differences in paw withdrawal were detected between PKG1α-KI and WT mice (0.04 g: WT: mean = 0.4, SEM = 0.1, SD = 0.4; PKG1α-KI: mean = 0.3, SEM = 0.1, SD = 0.4; 0.07 g: WT: mean = 0.9, SEM = 0.1, SD = 0.4; PKG1α-KI: mean = 0.5, SEM = 0.1, SD = 0.4; 0.16 g: WT: mean = 1.8, SEM = 0.3, SD = 1.0; PKG1α-KI: mean = 1.5, SEM = 0.3, SD = 0.8; 0.4 g: WT: mean = 4.2, SEM = 0.5, SD = 1.8; PKG1α-KI: mean = 3.9, SEM = 0.6, SD = 2.0; 0.6 g: WT: mean = 4.5, SEM = 0.6, SD = 2.1; PKG1α-KI: mean = 4.2, SEM = 0.5, SD = 1.8; 1.0 g: WT: mean = 4.6, SEM = 0.7, SD = 2.2; PKG1α-KI: mean = 5.0, SEM = 0.7, SD = 2.2; 1.4 g: WT: mean = 6.1, SEM = 0.6, SD = 2.0; PKG1α-KI: mean = 5.3, SEM = 0.5, SD = 1.5; 2.0 g: WT: mean = 6.7, SEM = 0.6, SD = 2.1; PKG1α-KI: mean = 6.4, SEM = 0.6, SD = 1.8; 4.0 g: WT: mean = 7.6, SEM = 0.6, SD = 1.9; PKG1α-KI: mean = 6.6, SEM = 0.5, SD = 1.7; n = 11). Data are shown as mean ± standard error of the mean. Statistical differences were assessed using a student ‘s t-test (A–F, H–J) or a two-way analysis of variance (G, K). * P < 0.05, ** P < 0.01, *** P < 0.001.

TRPV1 is an essential heat sensor [44]; hence, the responses to noxious heat stimuli were investigated using the hot plate test. Interestingly, PKG1α-KI mice showed increased latencies compared to WT littermates at 48°C, 50°C, or 52°C (Figures 1D–F); notable significant differences were detected between genotypes at 50°C and 52°C. These results indicate that redox-dependent PKG1α activation affects the detection of noxious heat at distinct temperatures. Additionally, in the TPP test PKG1α-KI mice spent considerably more time at 45°C compared to WT littermates (Figure 1G), further confirming that heat detection is impaired in the absence of oxidant-mediated PKG1α activation. In contrast, the responses of PKG1α-KI mice to cold stimuli were similar to those of WT mice (Figure 1H). Moreover, the perception of mechanical stimuli was normal in PKG1α-KI mice, as revealed by tape response and von Frey filament assays (Figures 1I–K). Altogether, these data suggest that redox-dependent PKG1α activation specifically contributes to capsaicin- or noxious heat-induced pain-related behavior.

Capsaicin induces a redox-dependent PKG1α activation

The DCFDA/H2DCFDA assay was performed to detect cellular ROS production. Notably, primary DRG neuron cultures stimulated with capsaicin (1 µM) [37] showed increased ROS production compared with vehicle-stimulated neurons (Figures 2A–C). H₂O₂ (50 µM; positive control) stimulation also increased ROS content; however, AITC (2 mM) stimulation did not increase the ROS production in sensory neurons. In control experiments, capsaicin's role as an oxidant was assessed through an Amplex Red assay in a cell-free setting; however, no ROS was detected (fold-increase ROS production after the addition of capsaicin compared to vehicle: 0%), suggesting that capsaicin itself does not serve as an oxidant. Altogether, these data indicate that capsaicin-stimulated sensory neurons display enhanced intracellular ROS production.

Figure 2.

Reactive oxygen species (ROS) production in capsaicin-stimulated sensory neurons. (A–C) Live cell imaging using chloromethyl-2’,7'-dichlorodihydrofluorescein diacetate to detect ROS production in cultured sensory neurons from wild-type (WT) mice. (A) Representative fluorescence images 40 min after the start of the cell imaging experiment. (B) The percentage increase in fluorescence denoting altered ROS production after stimulation with H₂O₂ (50 µM; positive control; n = 1351 cells, 4 mice) and capsaicin (1 µM; n = 510 cells, 6 mice); stimulation with allyl isothiocyanate (AITC; 2 mM; n = 288 cells, 6 mice) did not change the fluorescence increase compared to the vehicle control (n = 206 cells, 6 mice). (C) The percentage increase in fluorescence at 40 min was significantly higher in cells stimulated with H₂O₂ or capsaicin compared to vehicle control; no difference was observed in AITC-stimulated cells. (D) The capsaicin-induced paw licking behavior was similar in Cys42Ser protein kinase G1α-knock-in and littermate WT mice after pretreatment with the ROS-scavenger n-tert-butyl-α-phenylnitrone (50 mg/kg intraperitoneal injection) 30 min before the intraplantar injection of capsaicin (820 µM) (WT: mean = 19.02, SEM = 3.70, SD = 10.46; PKG1α-KI: mean = 12.12, SEM = 3.92, SD = 1.48; n = 7–8). Data are presented as mean ± standard error of the mean. Statistical analysis was performed using a one-way analysis of variance (C) or Student ‘s t-test (D). **** P < 0.0001.

To evaluate the role of ROS in capsaicin-induced nociceptive behavior, paw licking was assessed after PBN pretreatment (50 mg/kg i.p.) 30 min before capsaicin injection. Interestingly, both WT and PKG1α-KI mice showed similar paw licking responses (Figure 2D) compared with the experiment without PBN pretreatment (Figure 1A), suggesting that ROS scavengers ameliorate the capsaicin-induced nociceptive processing. These data further support the finding that sensory neurons produce ROS in response to capsaicin administration in vivo, potentially affecting redox-dependent PKG1α-mediated nociceptive processing.

Furthermore, the role of capsaicin-mediated dimerization and subsequent activation of PKG1 in DRG after ex vivo stimulation was investigated. Interestingly, western blots under non-reducing conditions revealed that the proportion of the PKG1 dimer was enhanced after capsaicin stimulation, whereas that of the PKG1 monomer decreased (Figures 3A and B), indicating that capsaicin stimulates PKG1 dimerization. In control experiments, no dimer formation was detected in DRGs isolated from PKG1α-KI mice after stimulation with capsaicin (Figure 3C). Consistent with these results, PKG1 activity in DRG lysates was enhanced in response to ex vivo capsaicin stimulation (Figure 3D). In contrast, the activity of isolated PKG enzyme was not affected by capsaicin (fold change: 0.54 ± 0.17, P > 0.05), suggesting that capsaicin-induced PKG1α activation in sensory neurons depends on intracellular signaling. These results show that capsaicin mediates PKG1α activation in sensory neurons in a redox-dependent manner.

Figure 3.

Protein kinase G (PKG)1α dimerization in capsaicin-stimulated dorsal root ganglions (DRGs). (A and B) Western blots of lumbar DRGs stimulated ex vivo with capsaicin (1 mM); PKG1α dimerization is significantly increased compared to the vehicle control (n = 6–7). (C) No PKG1α dimer formation was observed upon ex vivo stimulation of DRGs of Cys42Ser PKG1α-knock-in mice with capsaicin (1 mM). (D) In a PKG1 activity assay, DRGs of wild-type mice showed an increase in relative PKG1 activity after stimulation with capsaicin (1 mM) compared to vehicle (n = 12–13). Data are presented as mean ± standard error of the mean. For statistical analysis a two-way analysis of variance (B) or a Student’s t-test (D) was used. * P < 0.05, ** P < 0.01.

Control experiments revealed that messenger RNA (mRNA) levels of Trpv1, Tpra1, and Tprm3 (Supplemental Figure 1B–D) in lumbar DRGs of naïve PKG1α-KI and WT mice were comparable between genotypes. Moreover, western blotting revealed similar phospho-TRPV1 levels in DRG or spinal cord lysates of naïve PKG1α-KI and WT mice (Supplemental Figure 1E) and in DRG lysates of PKG1α-KI and WT mice stimulated with capsaicin (Supplemental Figure 1 F). The distribution of PKG1 mRNA in DRG neurons of WT mice was determined using in situ hybridization. Our results confirmed that PKG1 mRNA was expressed in most DRG neurons (Supplemental Figure 2A), which is consistent with previous studies [17,45]. Consequently, double-staining in situ hybridization showed that PKG1 mRNA was co-expressed with mRNA of Trpv1, Trpa1, and Trpm3 in DRG neurons (Supplemental Figure 2B–D). Altogether, these data suggest that redox-dependent PKG1α activation specifically contributes to TRPV1-dependent signaling in sensory neurons.

Neuronal activity of DRG neurons in response to capsaicin is reduced in PKG1α-KI mice

The function of redox-dependent PKG1α activation in capsaicin-mediated signaling was further assessed through calcium imaging experiments in capsaicin-stimulated cultured DRG neurons of PKG1α-KI and WT mice. Interestingly, consistent with the reduced paw licking phenotype (Figure 1A), capsaicin (100 nM) stimulation markedly reduced peak values and peak areas of calcium traces in sensory neurons from PKG1α-KI mice compared with those in WT mice (Figures 4A–C). However, the percentage of capsaicin-reactive cells was not altered between genotypes (Figure 4D). In contrast, AITC (200 µM) stimulation evoked similar peak values and peak areas in calcium traces of PKG1α-KI and WT mice (Figure 4E–G), and the percentage of AITC-reactive cells was comparable between genotypes (Figure 4H). These data suggest that redox-dependent PKG1α activation specifically enhances capsaicin-mediated calcium increase in sensory neurons.

Figure 4.

Capsaicin-induced calcium influx is decreased in dorsal root ganglion (DRG) neurons of Cys42Ser protein kinase G1α-knock-in (PKG1α-KI) mice. Graphs showing the results of calcium imaging experiments of cultured DRG neurons from PKG1α-KI mice and littermate wild-type (WT) mice upon stimulation with (A–D) capsaicin (100 nM) or (E–H) allyl isothiocyanate (AITC; 200 µM). (A and B) The peak area (A) and peak value (B) of capsaicin-induced calcium traces were significantly decreased in DRG neurons of PKG1α-KI mice compared to WT mice (n = 236–287 neurons from 7–9 mice per group). (C) Representative traces of intracellular calcium during the calcium imaging experiment with capsaicin stimulation. (D) No difference in the percentage of capsaicin-responsive cells was observed between PKG1α-KI and WT neurons. (E and F) Upon stimulation with AITC, DRG neurons of WT and PKG1α-KI mice revealed similar peak areas (E) and peak values (F) of the calcium traces (n = 380–451 neuros from 6–8 mice per group). (G) Representative images of calcium traces in AITC-stimulated live cells. (H) AITC-stimulated WT and PKG1α-KI neurons showed similar percentages of responsive cells. Data of calcium traces are presented as box-and-whisker plots with the box representing the 25th to 75th percentile, the median (50th percentile) as the central line, and the whiskers ranging from minimum to maximum values (A, B, E, and F). Percentages of responsive cells are shown as mean ± standard error of the mean (D, H). For statistical analysis, non-parametric medians were analyzed using the Mann–Whitney test (a, b, e, f), whereas responsive cells were compared using a Student‘s t-test (d, h). * P < 0.05, *** P < 0.001.

Finally, the role of capsaicin-mediated redox-dependent PKG1α activation on neuronal excitability on a cellular level was investigated. Whole-cell patch-clamp recordings of small IB4-negative DRG neurons from PKG1α-KI and WT mice were performed to address TRPV1-expressing peptidergic C-fibers [46]. For each cell, evoked APs and AP-firing (generation of ≥ two APs) were measured in the presence of vehicle (extracellular solution incubated for 30 s) or capsaicin (100 nM; incubated for 30 s; Supplemental Figure 3A). In the evoked AP recordings, capsaicin stimulation increased the resting membrane potential (RMP) in WT and PKG1α-KI neurons to a similar extent (Figure 5A) and did not significantly alter the rheobase (the minimum impulse required to evoke an AP) in both genotypes (Supplemental Figure 3B). Representative traces of WT and PKG1α-KI neurons are depicted in Figure 5B. Furthermore, the characteristics of evoked APs, specifically the amplitude size, AP duration at 90% depolarization, amplitude rise time (time from 20–80% amplitude increase), and time to reach the maximum amplitude (Supplemental Figures 3C–F) were similar between genotypes and not altered after capsaicin stimulation. These data suggest that the characteristics of evoked APs in sensory neurons of WT and PKG1α-KI mice are comparable in the presence of vehicle or capsaicin.

Figure 5.

Capsaicin-induced hyperexcitability of dorsal root ganglion (DRG) neurons involves redox-dependent activation of protein kinase G (PKG)1α. Graphs show the results of whole-cell patch-clamp action potential (AP)-evoked (A and B) experiments and AP-firing (C-G) in cultured DRG neurons from Cys42Ser PKG1α-knock-in (PKG1α-KI) and wild-type (WT) mice. (A) Experiments in which only a single AP was evoked, showed a significant increase in resting membrane potential (RMP) for both genotypes after capsaicin stimulation. (B) Representative traces of experiments in which a single AP was evoked by 10 ms currents (0-to-280 pA in 20 pA steps). (C) Number of APs that were observed when injecting increasing currents from 0 pA to 350 pA in 50 pA steps for 1 s. Capsaicin stimulation increased the number of APs at lower currents in WT and PKG1α-KI neurons, which then decreased at higher currents. The capsaicin-induced increase in AP-firing was significantly diminished in PKG1α-KI neurons compared with that in WT neurons (n = 9–10 neurons from 4 mice). (D) Representative traces of AP-firing experiments at 100 pA currents for WT and PKG1α-KI neurons, during vehicle incubation and capsaicin stimulation. (E) A significant increase in AP-firing detected after injecting 100 pA currents in capsaicin-stimulated WT neurons but not PKG1α-KI neurons. (F) The rheobase, which is the minimum impulse required to evoke an AP, decreased significantly with capsaicin stimulation in cells of both genotypes. (G) Capsaicin stimulation significantly increased the RMP of firing cells in both genotypes. Data are presented as mean ± standard error of the mean (C) or before–after plots that show the mean as bar charts (A, E–G). For statistical analysis, a one-way analysis of variance (C) or paired and unpaired t-tests were used (A, E–G). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

In the AP-firing recordings in the presence of vehicle, the number of APs increased in both genotypes with increasing current injections to a similar extent (0–350 pA, 1-s duration; Figure 5C). However, capsaicin stimulation in WT mice significantly increased the number of APs, whereas a significant increase was not observed in PKG1α-KI mice. Representative traces and results from 100 pA current injections are presented in Figures 5D and E, respectively. Capsaicin stimulation led to a notable decrease in rheobase (Figure 5F) and a marked increase in RMP (Figure 5G), which were similar for both WT and PKG1α-KI neurons. These data suggest that capsaicin-induced hyperexcitability of sensory neurons partly depends on the redox-dependent PKG1α activation.

Discussion

This study aimed to investigate the role of redox-dependent PKG1α activation in the processing of acute painful stimuli in the somatosensory system. Notably, the behavioral responses to capsaicin or noxious heat were reduced in PKG1α-KI mice, suggesting that redox-dependent PKG1α activation affects distinct nociceptive pathways. The pronociceptive function of ROS in acute pain was further supported by the fact that the capsaicin-induced nociceptive behavior was attenuated by pretreatment with the ROS scavenger PBN. Moreover, upregulated ROS levels were found in primary DRG neuron cultures stimulated with capsaicin, which is consistent with previous studies [37]. Ex vivo stimulation of DRGs with capsaicin enhanced the dimer formation of PKG1α, indicating redox-dependent activation of PKG1α. There are various potential ROS sources in DRG neurons, including NADPH oxidases (NOX), mitochondria, xanthine oxidase, and superoxide dismutase [2,3,47,48]. Interestingly, increased ROS production as a consequence of TRPV1 activation has been proposed in earlier studies. For example, stimulation of TRPV1 in a cardiac myoblast cell line resulted in increased calcium levels and mitochondrial dysfunction leading to enhanced ROS production [49]. In line with this, rat thymocytes showed dissipation of mitochondrial transmembrane potential in response to TRPV1 activation [50], pointing to a contribution of mitochondrial ROS production. Recent studies further demonstrated a contribution of TRPV1-induced mitochondrial dysfunction in neuropathic pain [51], suggesting that mitochondria-induced ROS production might act downstream of TRPV1. Potential other ROS sources, which are induced by activation of TRPV1, are NADPH oxidases. For instance, in cultured organ of Corti transformed cells, TRPV1-dependent ROS production seems to rely on the NADPH oxidase isoform NOX3 [52], while in microglia cells TRPV1-dependent ROS production depends on NOX2 [53]. In addition, NOX4 has also been described to act downstream of TRPV1 in renal tubular cells [54]. Thus, there are several potential sources of ROS acting downstream of TRPV1. In contrast, other studies suggest that enhanced ROS production modulates TRPV1 expression and activity, thereby acting upstream of TRPV1 [55–57]. Moreover, the induction of antioxidative pathways in response to capsaicin-induced TRPV1 activation has been proposed [58]. Therefore, further studies are needed to elucidate the question which ROS sources are activated after capsaicin-mediated TRPV1 stimulation and how this contributes to the activation of PKG1α.

Several studies have reported the functional role of PKG1α in the nociceptive system [19,59]. For instance, immunostaining and single-cell RNA sequencing studies have shown that PKG1 is expressed in most sensory neurons and in populations of dorsal horn neurons of the spinal cord [17,20,47,60,61]. Additionally, reduced acute nociceptive, inflammatory, and neuropathic pain behavior in global and sensory neuron-specific PKG1α-deficient mice or after administration of PKG1 inhibitors has been reported [16–18,62,63]. Redox-dependent PKG1α activation has been shown to contribute to the processing of neuropathic pain, but it is dispensable for inflammatory processing [20]. Herein, the responses to capsaicin and noxious heat were attenuated in PKG1α-KI mice, whereas the responses to AITC, bradykinin, or mechanical stimuli were normal. This suggests that redox-dependent PKG1α activation is involved in only distinct nociceptive pathways. This may be attributed to the fact that ROS are produced only in specific conditions in somatosensory neurons. Thus, subcellular colocalization of PKG1α and ROS sources or cGMP producers might influence the kinase activation. Moreover, colocalization of PKG1α with downstream targets most likely affects which proteins are phosphorylated by the kinase in dependence of its mode of activation. So far identified downstream targets of PKG1α in the nociceptive system include the myosin light chain phosphatase subunit MYPT1, large conductance calcium-activated potassium channels, inositol 1,4,5-trisphosphate receptor-associated cGMP-kinase substrate 1(IRAG)/inositol 1,4,5-triphosphate receptor (IP3R), sarco/endoplasmatic reticulium calcium ATPase or cysteine-rich protein 4 [19]. Nevertheless, it remains to be elucidated whether these targets are modulated by PKG1α in a cGMP-dependent and/or redox-dependent manner. Of note, redox-dependent activation of PKG1α in cardiomyocytes reduces phosphorylation of tuberous sclerosis complex 2 (TSC2), thereby enhancing mammalian target of rapamycin complex 1 (mTORC1) activity, while the cGMP activated enzyme promotes TSC2 phosphorylation and consequently supresses mTORC1 activity [64]. Moreover, in the heart selective phosphorylation of downstream targets such as phospholamban, depending on whether the kinase is activated by cGMP or oxidants, has been shown through phosphoproteomics [65]. In sensory neurons redox-dependent activation of PKG1α leads to the phosphorylation of cofilin, which is required for peripheral nerve regeneration [21]. Importantly, specific phosphorylation of PKG1α targets might differ between cell types, depending on its mode of activation. Thus, which targets are selectively phosphorylated in sensory neurons depending on whether the kinase is activated by cGMP or oxidants needs to be elucidated in further studies.

Consistent with the present study, previous studies have revealed that PKG1α affects capsaicin-induced pain behavior. For example, secondary hyperalgesia induced by intraplantar capsaicin administration was shown to be reduced in sensory neuron-specific PKG1-deficient mice [17]. In line with these behavioral effects, we observed that the capsaicin-induced calcium transients in sensory neurons and the hyperexcitability of sensory neurons were reduced in PKG1α-KI mice, as demonstrated by calcium imaging experiments and whole-cell patch-clamp recordings. Specifically, the reduction of capsaicin-induced calcium influx might contribute to the decreased excitability of sensory neurons, as described earlier [66,67]. For instance, studies pointed to a direct link between the activity of PKG and L-type calcium channels, which might contribute to the results presented here [68]. Furthermore, there are several lines of evidence that elevated pain behavior is a direct consequence of increased neuronal AP-firing [69–71]. As our patch-clamp recordings revealed decreased AP-firing in sensory neurons from PKG1α-KI mice in presence of capsaicin, it seems likely that the reduced excitability underlies the attenuated nociceptive behavior after intraplantar capsaicin injection. Nevertheless, identification of specific phosphorylation targets of redox-activated PKG1α is required to better understand the associated signaling pathways.

Importantly, our findings contradict the results of previous studies on sensory neuron-specific PKG1 knockout mice [17], in which capsaicin stimulation did not affect calcium responses. Thus, capsaicin-induced nociceptive behavior might generally involve PKG1α, independent of its activation; however, downstream signaling pathways might differ. These observations further support that redox- and cGMP-dependent activation of the kinase can modulate different intracellular pathways. To date, only a few studies have reported TRPV1–PKG1α interaction in sensory neurons. For example, PKG activity was induced in response to capsaicin and inhibiting PKG1α reduced capsaicin-induced pain-related behavior [72]. Moreover, it has been shown that the NO-cGMP/PKG pathway can alter TRPV1 activation [73]. Importantly, some studies suggest that PKG1 might act upstream of TRPV1 in a model of osteoarthritic pain and cardiac hypertrophy and heart failure [72,74]. However, as (i) we detected enhanced ROS production in sensory neurons, (ii) increased PKG1 dimer formation and activity in response to capsaicin, and (iii) no alterations in TRPV1 phosphorylation after stimulation with capsaicin, our data suggest that redox-dependent activation of PKG1 acts downstream of capsaicin-induced TRPV1 activation. We cannot exclude the possibility that alterations in capsaicin-induced calcium transients in PKG1α-KI mice might be attributed to signaling cascades that lead to the release of calcium from intracellular storages are modulated in the redox-deficient PKG1α mice. Of note, phosphorylation of IRAG by PKG1α and its interaction with IP3 at the membrane of the endoplasmic reticulum promotes calcium release leading to the regulation of smooth muscle contraction and platelet activation [75,76]. A similar mechanism might contribute to the altered calcium dynamics observed here. In line with electrophysiological studies, capsaicin-associated neuronal hyperexcitability has been reduced after administering the PKG1 inhibitor KT5823 [77], supporting the finding that TRPV1- and PKG1-dependent pathways interact in sensory neurons.

Altogether, the findings of this study show that redox-dependent PKG1α activation affects capsaicin/TRPV1-mediated signaling in sensory neurons. Nevertheless, further studies are warranted to elucidate the molecular mechanisms upstream and downstream of redox-activated PKG1α.

CRediT statement

Conceptualization: AS, WKG

Data curation: TB, KM, NB, EW, MK

Formal analysis: TB, WKG

Funding acquisition: WKG

Investigation: TB, KM, NB, EW, MK

Methodology: TB, KM, PE, WKG

Project administration: AS, WKG

Resources: PE

Supervision: AS, WKG

Visualization: TB, WKG

Writing – original draft: TB, AS, WKG

Writing – review & editing: PE, AS, WKG

Funding Statement

This work was supported by the Else Kröner-Fresenius-Stiftung under grant 2021_EKEA.08.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sets of this study are available from the corresponding author upon reasonable request.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/13510002.2025.2549954.