To our knowledge, this is the first work that aimed to analyze the implications of the BDNF and the Tropomyosin Kinease B (Trk-B) receptor in the effects of PRP therapy in OA animals. Previous works reporting the presence of BDNF in other sites rather than neuronal tissue and implicated in functional activities of platelets have been reported by Chacón-Fernández et al.[24], Fujimura et al. [46], and Le Blanc et al. [47].
Evolutionarily, the platelet system was conceived in mammals to control hemostasis, ensure the mitigation of vascular damage, and avoid bleeding and tissue injury [48]. Such property allowed an effective regulation of the blood leakage associated with injuries of diverse nature, triggered from the most primitive animal experience to more particular contexts such as maintenance of placental and embryo development [48, 49].
In such reparative and physiological role, products such as ATP and calcium ions act as main agonists in platelet degranulation [12, 50].We observed that using 10% of CaCl2 could stimulate platelet degranulation and BDNF release. Similar results were obtained by Le Blanc et al. (2020), that suggest a plasmatic predominance in BDNF concentrations coming from activated platelets through a series of agonists such as ADP, collagen, calcium and arachidonic acid.
Previous findings suggest that BDNF is stored in cytosol and at α-granules in platelets [24, 47]. These α-granules are structures that are present platelets plasmatic membrane and oscillations in intracellular calcium stores stimulate their exocytosis and content release [12]. Although it seems reasonable that, as previously mentioned, different sets of molecules implicated in tissue injury signaling can trigger BDNF release in platelets, the studies regarding mechanisms involved on the docking and release of this neurotrophin in platelets will be of quite interest to elucidate if classical exocytosis protein complexes that led to granule release in platelets, such as interactions between VAMP-8 and SNARE proteins, are also involved in BDNF release from platelet granules [12, 51].
Perhaps one of the most crucial cellular interactions for tissue repair where platelets play a pivotal role is the crosstalk with macrophages [52–54]. Macrophages are versatile cells and have an essential role in maintaining joint tissue homeostasis [55–57]. Beyond signaling molecules capable of initiating and drive a local inflammatory process such as IL-8, CCL2, and epinephrine, platelets may release other molecules that act on macrophage polarization and thus interfere with the dynamic plasticity of these cells, triggering the M1 (IL-1β, TNF- α) or M2 phenotype (TGF-β, Fibrinogen) of macrophages present at the injured site [52, 58, 59].
In OA, the importance of macrophage polarization takes place as a tool to improve tissue healing in compartments with poor vascularization, such as diarthrodial joints [56] . In releasing growth factors such as TGF-β or IGF-1, or anti-inflammatory cytokines and peptides such as IL-10, IL-4, and TIMPs, macrophages contribute for the shifting to an anabolic state in OA joints, that ultimately leads to a mitigation in tissue injury associated with catabolism and oxidative stress, especially in chondrocytes [60]. Importantly, the macrophage polarization triggered by platelet-rich plasma is one of the main therapeutic properties of PRP and further relief of OA long-term pain [58, 61, 62].
Among several growth factors present in the platelet compartment, BDNF is certainly stored at high concentrations. Of note, studies had shown that platelets compared with brain homogenates, have about 100 to 1,000 times fold of BDNF [24–26] . Thus, three relevant evidences were equal important in this work to propose a modulation effect of macrophage polarization induced by PRP depending on a BDNF-Trk-B receptor interaction: the significant bioavailability of BDNF in platelets allied to the immunomodulatory potential from this neurotrophin [63–65] and the Trk-B receptor expression in macrophages [66, 67].
In our hands, BMDM cells that were exposed to LPS/IFN-γ exhibited classical M1-like phenotype, confirmed by the significant increase in Nos-2 mRNA in comparison with controls (unstimulated BMDM cells). Given that in our platelet degranulation assays we showed significant amounts of BDNF release by activated platelets, we decided to incorporate the supernatant collected from activated samples of LP-PRP (SLPPRP) in our polarization assays to investigate the outcomes of BDNF in M1-macrophage inflammatory phenotype induced by LPS/IFN-γ. Notably, our results suggest that when cells previously triggered for M1-like phenotype were treated with SLPPRP for 2 hours, a decrease in Nos-2 mRNA expression was seen. Moreover, these effects of SLPPRP on M1 phenotype were significantly reversed in the presence of ANA-12, indicating a role of Trk-B receptor in the suppression of M1-phenotype induced by SLPPRP.
Although it seems reasonable that the effects of SLPPRP would be preventive in our assays due the short time window between the cell stimuli and treatments, evidence shows that 1 hour after the LPS there is already a substantial increase in iNOS expression and message in macrophage cells that lasts at least for 24 hours [68]. In this sense, it seems reasonable that SLPPRP would be acting in these M1-like cells to reprogram or counteract an inflammatory state, which Trk-B receptor seems to be involved.
Despite significant modulation was seen by SLPPRP in M1-phenotype, at least after a 2-hour incubation period, no changes in Arg-1 expression were detected, indicating no changes for an M2-like phenotype could be triggered by SLPPRP. Indeed, the in-vitro time-coursefor M2-phenotype triggered by Arg-1 modulation induced by PRP and its therapeutic derivatives such as LP-PRP and PRGF seems to occur in a timely manner, with significant levels of Arg-1 mRNA being seen between 6 and 24h after incubation [69–71]. In this sense, the failure of SLPPRP in induced an M2-like phenotype in our experiments could be attributed to the earlier cells harvesting in the third hour after LPS/IFN-γ stimuli, which would reflect undetectable changes in Arg-1 message in comparison with our unstimulated control cells.
The interaction between BDNF and Trk-B receptor in neurons led to the activation of signaling cascades mainly ruled by PI3K, MAPK, PLCγ, and GTPases enzymes that ultimately lead to cell growth & dendrite branching, cell survival and synaptic plasticity [72, 73]. Of interest, BDNF signaling through PI3K leads to further phosphorylation of AKT kinase inhibiting the intrinsic and extrinsic pathways and further neuronal apoptosis. It is also suggested that blockade of Trk-B receptor by selective antagonists such as ANA-12 increases expression of the p38 protein, IL-6, TNF-α, IL-1β, and iNOS, indicating that BDNF has an anti-inflammatory role depending on TrkB activation [74].
In this sense, the interactions between BDNF and AKT signaling could be implicated in macrophage polarization. It is known that the isoforms AKT1 and AKT2 directly regulate macrophage polarization-modulating iNOS expression [75, 76]. The absence of AKT1 in macrophages promotes hyperresponsiveness to LPS and increased iNOS gene expression, exacerbating M1 inflammatory response mediated by increase in IL-1β, TNF-α, and IL-6 cytokines while the opposite is observed in the absence of AKT2, that leads to an exacerbation of anti-inflammatory response by upregulating Arg-1[76].
After sets of in-vitro experiments suggesting the LP-PRP role in macrophage polarization associated with the TrK-B activation, we investigated the implications of the Trk-B receptor in the repair activities induced by LP-PRP in OA animals. This receptor is expressed in synoviocytes, chondrocytes, osteoclasts, and osteoblasts [22, 77–80]. In addition, although not directly linked with OA pathogenesis, BDNF might have potential therapeutic activities through the anti-inflammatory, anti-apoptotic, and antioxidant actions and modulation of autophagy in the affected OA tissue damage, that leads to the joint healing process [73, 81].
The OA model triggered by MIA insult comprehends a chronic-degenerative pathology that led to tissue injury and release of pro-inflammatory molecules into the synovial fluid, such as ADP, Ca2+, collagen peptides, and aggrecan [82–84]. Thus, once these molecules are major agonists of platelet degranulation, the local intra-articular injection of LP-PRP preparations without prior activation was conducted to ensure more bioavailability of BDNF in the knee joint compartment.
As in humans, MIA-OA animals display similar pathological changes seen resulting from oxidative stress in joint cells, which leads to ongoing tissue catabolism, inflammation, cartilage breakdown, nerve injury, and chronic pain [82, 83].
These OA-like alterations triggered by MIA are divided into two phases: during the acute phase (inflammatory phase), previous studies shown an increase in inflammatory cytokines such as IL-1β, TNF-α, IL-6, IL-8, and chemokines such as CCL-2 are observed [82, 85–90]. Additionally, an enhancement on cell recruitment occurs in the knee joint, resulting on an increasing number of inflammatory cells such as macrophages, dendritic cells, neutrophils, and mast cells [75, 90–93]. Particularly, a shifting to M1 profile is noticed in macrophages, driven by the massive released of tissue degradation products and cytokines (as previously mentioned) that ultimately contribute to boost the late phase of MIA-OA model looping the pro-inflammatory context through their released pro-inflammatory products and stablishing an ongoing tissue injury that sets the degenerative process [82, 87, 90, 94]. At the late phase (chronic phase), a decrease in the expression of enzymes linked with cell survival against oxidative damage, such as SOD-2 and Catalase, and increased expression of matrix metalloproteinases (MMP), such as MMP3 and 9 are observed [82, 86, 95, 96]. As this catabolic state goes by, the cell-death tied to a hyperactivity of MMPs lead to cartilage breakdown and further sensitization of sensory fibers due to nerve injury of primary afferent terminals in subchondral bone. In addition, this sensitization of joint nociceptors is also driven by inflammatory products that could be present on synovial fluid such as IL-1β, TNF-α, IL-6 and NGF [55, 82, 90].
In our studies, an interesting decrease in knee joint injury followed by functional recovery of both joint-flexion mechanical thresholds and gait was observed after LP-PRP treatment. These mitigation on OA-like injuries triggered by MIA are dependent on Trk-B receptor, since the effects of LP-PRP, both at behavioral and wound-healing outcomes (joint and nerve injury) were significantly reversed in the presence of the selective Trk-B antagonist ANA-12.
Despite limited studies investigating the mechanisms behind the healing properties of PRP in OA models, studies conducted Chiou et al. [32], Woodell-May et al. [97] and Aniss et al. [98] support the use of PRP as a therapeutic intervention to promote joint repair in OA. These works indicated an essential role of the PRP for in-vitro cellular differentiation of chondrocytes via regulation of cyclins B1, D1, and E2, decrease in apoptotic signaling induced by caspase-3 and p53 expression, and alleviates ROS-induced oxidative damage triggered by IL-1β treatment in chondrocyte cells. Of note, chondrocyte differentiation is a key activity to boost tissue matrix peptides synthesis and further mitigate joint catabolism triggered in OA. Interestingly, PI3K and AKT signaling that are activated by BDNF-Trk-B interaction, play an important role in chondrocyte terminal differentiation in murine embryos [99]. In this sense, studies showing a relationship between BDNF and chondrogenic signaling with impact in OA pathology will be of interest. Particularly, an interaction of BDNF and Runx-2, a transcriptional factor known to regulate chondrogenesis and mitigate OA cartilage injury downregulating matrix metalloproteinases as MMP-13 and Adamts-5, is described in a non-OA animal model [100].
In-vivo studies demonstrated that the PRP treatment decreases the expression of MMPs 1, 3, and 13 [32, 97, 98]. Of interest, some of these repair activities mentioned for PRP, are seen also for BDNF as well, such as the decrease in ROS-induced oxidative damage towards modulation of the SOD-2 enzyme [101–103] and modulation of the apoptosis proteins as caspase-3 [104, 105]. Complementary, some studies have shown that BDNF can activate FOXO3 protein signaling [106], which integrates the FOXO proteins complex that is involved in cell defense against oxidative stress and modulation of autophagy genes such as Sesn3 and Bnip3 that prevent chondrocyte apoptosis and cartilage aging[107].
In summary, it seems reasonable to conclude that the interaction between BDNF-TrkB arising after LP-PRP injection and further platelet degranulation, has as one of main therapeutic axis the attenuation of oxidative stress triggered in MIA-OA model [82, 95, 96]. Complementary, the anti-inflammatory profile of PRP and also seen for BDNF, such as a decrease in inflammatory IL-1β/TNF-α and enhance of the anti-inflammatory IL-10/IL-4 cytokine axis [63, 65]could be contributing to a suppression on macrophage-M1 profile, and thus, a decrease in further ongoing-tissue injury, as our therapeutic protocol with LP-PRP was started at day 7 after MIA insult, which comprehends the transitioning period between the acute phase and chronic phase of MIA-OA model [82, 87].
In this sense, this downward on tissue inflammatory and catabolic state clearly provided by LP-PRP and BDNF-TrkB interaction reflects the findings in our tissue pathology assays (joint and DRG) and corroborates with our behavior results: MIA-OA outcomes are attenuated in the presence of LP-PRP, where its therapeutic activities are reversed in the presence of Trk-B antagonist. The importance of these BDNF-TrkB biological activities for joint repair are highlighted at the functional (gait and pain behavior) outcomes obtained in presence and absence of the antagonist, with some worsening seen particularly on joint pain and inflammation, as the increase in pain phenotype in groups where TrkB receptor blockade was made by intra-articular injection.
Joint nociceptive neurons typically evoke action potentials in response to a range of stimulus that would breakthrough the tissue physiological working range[108–110]. Thus, when a static or dynamic pressure stimulus exceeds the joint baseline resistancesuch asintense pressure, sharp rotation or twisting, the result is an avoidance of harmful motion, withdrawal of and preservation of the affected limb.
Interestingly, although some murine models shown an overlap between antalgic gait and pain behavior such as seen in diabetic neuropathy [111], a study have shown no reversal of antalgic gait parameters in presence of analgesics, especially in CFA and spared nerve injury pain models [112] . Particularly for MIA-OA models, there is still no concise data, where some studies report none to mild changes in gait after MIA [113] and some seen markable joint dysfunction regarding spatiotemporal parameters at the acute and late phase of the model [85, 114–116]. In these mentioned studies, it is a common understanding that joint dysfunction triggered by MIA is followed by changes in print-area, print length and print width parameters[85, 114–116].
In our experiments, the LP-PRP therapeutic protocol was able to increase the max contact area as well as print-length, and print-width gait parameters at day 14 and 21st after MIA injection. These improvements in gait were also followed in joint mechanical thresholds and reversed in the presence of Trk-B antagonist ANA-12. Interestingly, rats in the ANA-12 and ANA-12+LP-PRP groups displayed significantly decreased joint-flexion mechanical thresholds in comparison with both OA controls treated with antagonist vehicle (ANA-12 vehicle group). However, despite this worsening in joint-flexion mechanical thresholds, this event was not replicated in the gait parameters analyzed between these 3 groups.
It is important to note that although there is some significant degree of interdependence between antalgic gait and mechanical thresholds linked with joint injury (as joint flexion thresholds or secondary tactile allodynia), these parameters might not follow strictly covariance over time [85, 113, 115].
In this sense, joint nociceptive afferents have prominent response to noxious flexion or twisting stimuli [108–110, 117]. Although it seems plausible that the nerve injury tied to the persistent inflammatory phenotype driven by increased TNF-α in ANA-12 and ANA-12 + LP-PRP groups would imply an increase in gait behavior related to joint noxious dysfunction, the animal ability in promote adjustment in joint motion would bias our conclusions showing no changes related to spatiotemporal parameters, at least at the endpoint measurements. Or the antalgic gait seen in MIA-OA model could be related to a significant impairment in joint mechanobiology properties, that limit joint movement and further proper animal ambulation. This impairment would be more depreciated in the model early phase due to an inflammatory input linked with swelling and joint sensitization. Such gait alterations would been resolving as the inflammatory burst decreases and acquire a latent characteristic, and the pain phenotype could be implicated with the ectopic activity of primary afferents which the background relies on nerve injury triggered in the model, where nociception and pain behavior would be achieved through induction of joint flexion [117, 118].
In addition, this nerve injury seen mechanistically by increasing on ATF-3 expression in dorsal root ganglia neurons implicated in joint nociceptive response can be associated with a persistent plasticity of these cells triggered by massive release of inflammatory peptides in the early phase of MIA or due to a mechanical injury of peripheral terminals present at joint capsule and subchondral bone [119, 120]. Of note, chemical (inflammatory cytokines and peptides, algogens) or mechanical (chronic constriction, axotomy) injury are primarily mechanisms involved in ATF-3 increase on the DRG [121–123].
Concluding, the improvements in joint performance by LP-PRP could be attributed to a shift in inflammation profile driven by this therapeutic after injection. This alternate in inflammatory and degenerative profile would be likely associated with a downregulation activity in macrophage M1-profile, mitigation in tissue catabolism and preservation in joint integrity that ultimately led to a decrease in inflammation and nerve injury of nociceptive afferents. These biological activities seem to involve an interaction between BDNF and Trk-B receptors with implications in functional aspects as well, once both changes in histopathological, molecular, and pain phenotype evoked by LP-PRP were reversed in presence of Trk-B antagonist.