Bradykinin is produced within the interstitium of most tissues and plays an important role in normal and pathological conditions, being considered an important inflammatory pain mediator (reviewed in ) that is associated with chronic musculoskeletal pain syndromes [41, 42]. Recent findings have shown that in the central nervous system bradykinin triggers astrocyte-neuron signaling via glutamate release followed by NMDA receptors activation . In peripheral tissues, bradykinin receptors have already been described in sensory neurons of dorsal root ganglia (DRG), where it exerts direct and indirect effects via neuronal excitation and threshold modulation by the release of excitatory signaling molecules, respectively . To the best of our knowledge, this is the first report demonstrating that fibroblasts isolated from the human subcutaneous connective tissue respond to bradykinin by releasing ATP into the extracellular medium through the activation of B2 receptors, which are constitutively expressed in most of the tissues exhibiting bradykinin sensitivity. Although our experiments were conducted in non-stressful conditions, the involvement of inducible B1 receptors mediating bradykinin effects in human subcutaneous fibroblasts cannot be ruled out . Bradykinin-induced ATP release from these cells was demonstrated by two distinct experimental approaches: the luciferin-luciferase bioluminescence assay and ATP-binding quinacrine dye destaining by time-lapse confocal microscopy. Our data also suggest that bradykinin-induced ATP release from human subcutaneous fibroblasts requires Ca2+ recruitment from thapsigargin-sensitive internal stores and occurs independently of extracellular Ca2+ by electrodiffusional translocation of the nucleotide via hemichannels containing connexins (most probably Cx43) and Panx1.
In several cell types [24, 25], including human fibroblasts of the foreskin , bradykinin elicits an abrupt rise in [Ca2+]i to a peak that depends on Ca2+ recruitment from internal stores, which then declines to a plateau that is sustained by Ca2+ entry from the extracellular compartment. The long lasting elevation of [Ca2+]i seen after the initial [Ca2+]i peak probably results from a receptor- or second messenger-operated Ca2+ current activated by bradykinin, rather than via a capacitative pathway . Here, we demonstrate for the first time that ADP-sensitive P2Y12 purinoceptors may be, at least in part, responsible for the Ca2+ influx that is characteristic of the plateau phase of elevated [Ca2+]i caused by bradykinin in human subcutaneous fibroblasts. This conclusion is supported by results showing that: (1) human subcutaneous fibroblasts exhibit P2Y12 receptor immunoreactivity, (2) bradykinin induces the release of ATP that is subsequently hydrolyzed into ADP by membrane-bound ectonucleotidases, (3) ADP has a tendency to accumulate transiently in the cultures, given that we found that the extracellular catabolism of ADP into AMP is the rate-limiting step of the ectonucleotidase cascade, and, finally, (4) adenine nucleotides-induced [Ca2+]i rises in human subcutaneous fibroblasts are dependent on extracellular Ca2+.
The pathways underlying bradykinin-induced long lasting Ca2+ influx resulting from the cooperation with ADP-sensitive P2Y12 purinoceptors need to be further investigated. Interestingly, bradykinin causes a rapid translocation of all protein kinase C isoforms from the cytosol to the plasma membrane within the time frame (30 s) of ATP appearance in the cultures , which may favor the synergism between B2 and P2Y12 receptors activation. The involvement of ADP-sensitive P2Y12 receptors in neuropathic pain, has been already reported; authors suggested that this could be due to an unclear mechanism involving non-neuronal cells, such as spinal microglia [45, 46] and trigeminal ganglion satellite cells . The close proximity between P2Y12 receptor-expressing fibroblasts of the subcutaneous tissue and sensory nerve fibers exhibiting numerous ligand-gated receptors (including ionotropic P2X and metabotropic P2Y purinoceptors) implies that adenine nucleotides released from stimulated fibroblasts may alter acute and chronic pain perception. During acute tissue injury, excessive ATP release from damaged fibroblasts, keratinocytes, blood vessels and inflammatory cells may cause pain by activating excitatory purinoceptors on nociceptive sensory nerve endings [46, 48–50]. Lower levels of ATP released from intact cells in response to mild mechanical and thermal stimulation may participate in normal tactile sensation and also contribute to the spontaneous pain and tactile hypersensitivity that occurs under chronic painful conditions involving the subcutaneous connective tissue when nerve endings become sensitized [48, 51]. We, now, postulated a new type of interaction involving an autocrine action of ATP, via P2Y12 receptors, triggered by the activation of bradykinin B2 receptors in human subcutaneous fibroblasts. Moreover, increased sensitivity to extracellular ATP has been described in fibroblasts from patients affected with systemic sclerosis .
Little is known about the mechanisms upstream the nucleotide release from human fibroblasts despite the importance of connective tissue ATP signaling in the pathogenesis of acute and chronic inflammatory pain [18, 19]. Multiple nucleotide-releasing pathways have been identified in intact cells, which represent a critical component for the initiation of the purinergic signaling cascade (reviewed by ). Experiments designed to manipulate exocytosis of vesicles/organelles containing compartmentalized ATP suggest that it might not represent an important pathway for releasing ATP from human fibroblasts stimulated with bradykinin. Several non-vesicular ATP release mechanisms have been proposed, yet many remain controversial and are complicated by the non-specificity of available inhibitors. Hemichannels possessing connexin and Panx1 subunits represent an important mechanism for the cellular release of ATP. The opening of hemichannels occurs in response to many physiological and pathological situations, including volume regulation, proliferation, calcium wave propagation by extracellular messengers and cell death during metabolic inhibition (reviewed in ). Using immunofluorescence confocal microscopy and Western blot analysis, we demonstrated that fibroblasts of the human subcutaneous tissue exhibit strong anti-Panx1 immunoreactivity in addition to Cx43 that is characteristic from fibroblasts of other tissue origins [27, 28]. Moreover, functional data using non-selective connexin inhibitors targeting Cx43 hemichannels (e.g. 2-octanol, CBX) strongly depressed the plateau phase of bradykinin-induced [Ca2+]i response. Because connexin hemichannels are activated by moderate (< 500 nM) [Ca2+]i elevations, these channels may open in response to bradykinin during the initial [Ca2+]i rise  and contribute to ATP release and to the subsequent purinoceptor-mediated signaling. Closing of connexin containing hemichannels, which unlike Panx1 hemichannels seal when the spike amplitude rises above 500 nM, contributes to shape bradykinin-induced [Ca2+]i oscillations as demonstrated by the partial reduction of the initial [Ca2+]i rise of bradykinin in the presence of 2-octanol and CBX.
Considering the relatively high potency of CBX (300 μM) and the fact that this compound also blocks Panx1 containing hemichannels , we tested the effect of the selective Panx1 mimetic inhibitory peptide, 10Panx, which also depressed the plateau phase of the bradykinin-induced [Ca2+]i response in parallel to the decline of ATP release from fibroblasts loaded with quinacrine. Like the effects obtained with apyrase and with the selective P2Y12 antagonist, AR-C 66096, inhibition of hemichannels containing Cx43 and Panx1 were more effective in depressing the plateau phase rather than the peak of the bradykinin response. Taken together, data suggest that Cx43 and Panx1 containing hemichannels have a predominant role on ATP release from human subcutaneous fibroblasts stimulated with bradykinin, thereby instigating the regenerative propagation of intracellular Ca2+ signals. Our findings agree with the observation that mechanically stimulated cardiac fibroblasts release ATP in a CBX-sensitive manner, an effect that the authors attributed to Cx43 hemichannels not excluding a possible involvement of Panx1 containing hemichannels .
Regardless of whether channel-mediated efflux or vesicle exocytosis comprises the predominant ATP release mechanism, most studies (but not all, see e.g.) have identified elevation of cytosolic Ca2+ as an important regulator of nucleotide export in different cell model systems. The molecular mechanism by which bradykinin releases ATP through the opening of Cx43 and Panx1 hemichannels in human subcutaneous fibroblasts may be the generation of inositol trisphosphate (IP3) by phospholipase C and the downstream [Ca2+]i mobilization from internal stores [57, 58]. Our data, showing that intracellular Ca2+ depletion with thapsigargin impaired quinacrine dye destaining induced by bradykinin is in favor with the hypothesis that Ca2+ mobilization is necessary for ATP release in these cells. Further experiments are required to discard the ability of bradykinin, like certain Gq-coupled receptors, to additionally stimulate Rho-GTPase acting to strongly potentiate a Ca2+-activated ATP release pathway [59, 60]. Seminario-Vidal and col. (2009) , demonstrated that Ca2+- and RhoA/Rho kinase-dependent ATP release from thrombin-stimulated A549 lung epithelial cells occurs via connexin or pannexin hemichannels, a pathway that seems to be not competent for ATP release in human astrocytoma cells . Given the actions exerted by Rho/Rho kinase on cytoskeleton components (e.g. regulating myosin-light chain phosphorylation and actin polymerization), those authors speculated that Rho-promoted membrane-cytoskeletal rearrangements facilitate the insertion of hemichannel subunits within the plasma membrane.
Bradykinin can increase glutamate release from mouse astrocytes through volume-sensitive outwardly rectifying (VSOR) anion channels without cell swelling via a mechanism that is regulated by high intracellular Ca2+ in nanodomains . Although we did not test directly whether this pathway plays a role in bradykinin-induced [Ca2+]i-dependent ATP release from human subcutaneous fibroblasts, it appears that VSOR currents would exhibit a slow activation kinetics requiring 15-20 min after bradykinin application to reach a sustained plateau . This activation pattern is entirely different from the rapid (within 30 s) ATP releasing response to bradykinin observed in human fibroblasts (see Figure 3), thus indicating that slow activating but prolonged VSOR currents play a minor, if some, role in the release of ATP in our experimental time frame. In our hands mefloquine (MFQ, 3 μM, Figure 4Bi), which also blocks non-selectively anion channels, failed to modify bradykinin-induced [Ca2+]i signals. Moreover, modulation of VSOR channel permeability through the activation of protein kinase C with the phorbol ester (12-myristate 13-acetate, 1-10 μM, n = 3) did not mimic the effect of bradykinin.
Involvement of soluble signaling mediators, such as ATP and/or its metabolites, may also explain heterogeneity of individual [Ca2+]i responses to bradykinin within a cell population; confocal microscopy studies showed that some cells displayed no plateau phase whereas others were not noticeably affected by bradykinin removal and continued to respond for a few minutes (see Figure 1C). This is consistent with the generation of concentration gradients by released ATP and metabolites formation (namely ADP) via ectonucleotidases, which enables differential targeting of subtype-specific P2 purinoceptors and, thus, cell-to-cell communication depending on proximity. Therefore, we may speculate that bradykinin-stimulated fibroblasts trigger a “purinergic wave” mediated by released ATP and metabolites formation that can affect sensory afferent nerve endings localized in the vicinity, representing the first insights of a fibroblast-neuron communication unproved so far.
Recently, it has been reported that mechanical stimulation of human epidermal keratinocytes induces propagating Ca2+ waves depending on non-vesicular release of ATP through connexin hemichannels . In view of the potential contribution of the cutaneous release of ATP to acute and chronic pain syndromes, this and other groups demonstrated that human epidermal keratinocytes co-cultured with neurons of the dorsal root ganglia interplay through the release of ATP following keratinocytes-born [Ca2+]i waves . Likewise, subcutaneous inflammation or injection of ATP causes pain sensation through the activation of P2X3 receptors expressed in sensory nerve endings, which may become sensitized in both animal models and human patients [63, 64]. Knocking down or selectively antagonizing P2X3 receptor activity results in reduced responses to ATP, as well as reduced thermal and mechanical hyperalgesia in inflammatory and neuropathic pain rat models [65, 66]. P2Y purinoceptors, especially P2Y1 and P2Y2, expressed in primary sensory endings have also been implicated in chronic pain states . Authors from the latter study agree that cutaneous ATP release does not appear to contribute to pain sensation in the absence of tissue injury. However, under chronic painful conditions, such as inflammation and nerve injury, nerve endings may become sensitized and a normally innocuous level of subcutaneous ATP may now be sufficient to reach the firing threshold of nociceptors . Despite direct modulation of nociceptors threshold by ATP released from different cell types may play a key role to the association between subcutaneous connective tissue injury and musculoskeletal pain, there are alternative mechanisms that should also be considered in this context. These include interference with muscle tension due to fusimotor effects induced by group III and IV afferents activation, which by projecting to γ-motoneurons amplify muscle spindles activity and, thereby, increases muscle tone, generating metabolites (e.g. bradykinin) that enter in a positive feedback loop (reviewed by [41, 42, 68]).
In addition, it has been proposed that mechanical deformation of the skin by needles and application of heat or electrical current leads to release of large amounts of ATP from keratinocytes, fibroblasts and other cells in skin . Impulses generated by P2 purinoceptors in sensory fibers in the skin connect with interneurons that may negatively modulate neural pathways to the pain centers in the cortex. This is the basis for the novel hypothesis for the involvement of purinergic signaling in acupuncture . Furthermore, purines are known to cause intracellular Ca2+-dependent transient changes in cultured human fibroblast cytoarchitecture, which share similarities with the increase in cross-sectional area of fibroblasts in response to acupuncture . Although evidence has been presented of the role of adenosine in acupuncture-mediated anti-nociception by demonstrating that the local concentrations of the nucleoside increase in human subjects  and that adenosine is implicated in the proliferation of fibroblasts and remodeling of the skin, liver and lung (reviewed in , we failed to demonstrate any contribution of the nucleoside to bradykinin-induced [Ca2+]i signals in cultured human subcutaneous fibroblasts. Controversy still exists regarding adenosine participation in wound healing and scarring. For instance, the adenosine A2A receptor promotes skin fibrosis and scarring  and increases collagen production in human dermal fibroblasts  probably by activating the Gs/cyclicAMP pathway , yet it remains controversial how A2A receptors increase collagen production since cyclicAMP has been found to decrease the synthesis of collagen and DNA by fibroblasts . Thus, further studies are required to test long-term effects of extracellular adenosine, which may originate from the hydrolysis of ATP released from fibroblasts stimulated either mechanically (e.g. acupuncture) or by inflammatory mediators, such as bradykinin. Our findings demonstrating that adenosine accumulates as an end product of the catabolism of released ATP in the vicinity of fibroblasts within the subcutaneous connective tissue may be of clinical relevance, given the role of the nucleoside on dermal fibrosis (via A2A receptors) and its anti-nociceptive properties (via A1 receptors) on free nerve endings and sensory afferents (reviewed in [37, 75]).