Sensory Neuron-Specific Deletion of Tropomyosin Receptor Kinase A (TrkA) in Mice Abolishes Osteoarthritis (OA) Pain via NGF/TrkA Intervention of Peripheral Sensitization

Tropomyosin receptor kinase A (TrkA/NTRK1) is a high-affinity receptor for nerve growth factor (NGF), a potent pain mediator. NGF/TrkA signaling elevates synovial sensory neuronal distributions in the joints and causes osteoarthritis (OA) pain. We investigated the mechanisms of pain transmission as to whether peripheral sensory neurons are linked to the cellular plasticity in the dorsal root ganglia (DRG) and are critical for OA hyperalgesia. Sensory neuron-specific deletion of TrkA was achieved by tamoxifen injection in 4-week-old TrkAfl/fl;NaV1.8CreERT2 (Ntrk1 fl/fl;Scn10aCreERT2) mice. OA was induced by partial medial meniscectomy (PMM) in 12-week-old mice, and OA-pain-related behavior was analyzed for 12 weeks followed by comprehensive histopathological examinations. OA-associated joint pain was markedly improved without cartilage protection in sensory-neuron-specific conditional TrkA knock-out (cKO) mice. Alleviated hyperalgesia was associated with suppression of the NGF/TrkA pathway and reduced angiogenesis in fibroblast-like synovial cells. Elevated pain transmitters in the DRG of OA-induced mice were significantly diminished in sensory-neuron-specific TrkA cKO and global TrkA cKO mice. Spinal glial activity and brain-derived neurotropic factor (BDNF) were significantly increased in OA-induced mice but were substantially eliminated by sensory-neuron-specific deletion. Our results suggest that augmentation of NGF/TrkA signaling in the joint synovium and the peripheral sensory neurons facilitate pro-nociception and centralized pain sensitization.


Introduction
Osteoarthritis (OA) is characterized by progressive, degenerative joint damage involving articular cartilage deformation of joint structure and severe pain. OA is a global disease and is a leading cause of disability that affects over 300 million worldwide [1]. Despite the availability of pharmacological and non-pharmacological options, persistent pain associated with OA pain is a debilitating condition that impairs the quality of life and causes heavy burdens on the healthcare system [2,3]. In the United Sates, one out of five individuals suffer chronic pain, and chronic pain affects more people than cardiovascular diseases (CVDs), cancer, and diabetes combined [4][5][6]. Therefore, understanding the mechanisms of pain transmission is the foremost important step for identification of molecular

Cartilage-Specific Deletion of TrkA Did Not Improve OA-Associated Hypersensitivity to Pain or Knee Joint Pathology with OA Progression in Mice
OA was induced by PMM in 12-week-old TrkA fl/fl ;Aggrecan CreERT2 -positive mice (n = 13) and their littermate controls, TrkA fl/fl ;Aggrecan CreERT2 -negative mice (n = 6), followed by harvesting of knee joints at 12 weeks after PMM surgery. Development of mechanical allodynia by von Frey filament testing in the ipsilateral hind paw was compared with TrkA fl/fl ;Aggrecan CreERT2 -negative mice and TrkA fl/fl ;Aggrecan CreERT2 -positive mice every week until 12 weeks after PMM surgery. Mice with a cartilage-specific deletion of TrkA did not exhibit significant differences in pain sensation ( Figure 2A) and other pain-related behavior such as ambulation test measured by horizontal photo beam crossings ( Figure 2B) and spontaneous rearing activity measured by vertical photo beam crossings ( Figure 2C). Both ambulation and rearing activities were reduced between pre-PMM and post-PMM mice

Global Conditional Deletion of TrkA Resulted in Reduced Expression of TrkA and NGF in the DRG Sensory Neurons in Mice
OA was induced by PMM in 12-week-old TrkA fl/fl ;Rosa CreERT2 -positive mice (n = 5) and their litter mate controls, TrkA fl/fl ;Rosa CreERT2 -negative mice (n = 11). Immunofluorescence staining of TrkA and NGF in the lumbar L3-L5 DRG isolated at 12 weeks after PMM surgery were monitored. The expression levels of TrkA were significantly (p < 0.00001) reduced in TrkA fl/fl ;Rosa CreERT2 -positive mice in the lumbar DRG ( Figure 1D) over 95%, as shown by quantification of TrkA expression per area ( Figure 1E). Deletion of TrkA significantly (p < 0.013) reduced the expression of NGF ( Figure 1F) by over 80%, as shown in the quantification of NGF expression per area ( Figure 1G). These results are consistent with the concept that OA-induced hyperalgesia is associated with increased NGF/TrkA signaling.

Cartilage-Specific Deletion of TrkA Did Not Improve OA-Associated Hypersensitivity to Pain or Knee Joint Pathology with OA Progression in Mice
OA was induced by PMM in 12-week-old TrkA fl/fl ;Aggrecan CreERT2 -positive mice (n = 13) and their littermate controls, TrkA fl/fl ;Aggrecan CreERT2 -negative mice (n = 6), followed by harvesting of knee joints at 12 weeks after PMM surgery. Development of mechanical allodynia by von Frey filament testing in the ipsilateral hind paw was compared with TrkA fl/fl ;Aggrecan CreERT2 -negative mice and TrkA fl/fl ;Aggrecan CreERT2 -positive mice every week until 12 weeks after PMM surgery. Mice with a cartilage-specific deletion of TrkA did not exhibit significant differences in pain sensation ( Figure 2A) and other pain-related behavior such as ambulation test measured by horizontal photo beam crossings ( Figure 2B) and spontaneous rearing activity measured by vertical photo beam crossings ( Figure 2C). Both ambulation and rearing activities were reduced between pre-PMM and post-PMM mice ( Figure 2B,C). In hot plate tests ( Figure S2A), both TrkA fl/fl ;Aggrecan CreERT2 -positive and -negative control group mice showed significant (p < 0.001) sensitivity to heat after 8 weeks after PMM, but there was no differences between cKO mice or negative control mice. In the acetone tests, TrkA fl/fl ;Aggrecan CreERT2 -negative mice displayed significant cold sensitivity in 12 weeks after PMM (p < 0.05), while TrkA fl/fl ;Aggrecan CreERT2 -positive mice displayed no significant sensitivity compared with pre-PMM or cKO mice ( Figure S2B). negative control group mice showed significant (p < 0.001) sensitivity to heat after 8 weeks after PMM, but there was no differences between cKO mice or negative control mice. In the acetone tests, TrkA fl/fl ;Aggrecan CreERT2 -negative mice displayed significant cold sensitivity in 12 weeks after PMM (p < 0.05), while TrkA fl/fl ;Aggrecan CreERT2 -positive mice displayed no significant sensitivity compared with pre-PMM or cKO mice ( Figure S2B).
Cartilage-specific deletion of TrkA also did not alter proteoglycan depletion, as established by histological assessment with Safranin-O fast green staining ( Figure 2D) nor the severity of articular cartilage degradation as determined using the Osteoarthritis Research Society International (OARSI) scoring system ( Figure 2E). These results suggest that loss of NGF/TrkA signaling in cartilage does not affect catabolic pathways that perturb the cartilage extracellular matrix, nor does it affect the progression of OA hyperalgesia and cartilage deterioration. Cartilage-specific deletion of TrkA in mice did not improve either OA-associated hypersensitivity to pain or knee joint pathology with OA progression. OA was induced by PMM in 12week-old TrkA fl/fl ;Aggrecan CreERT2 -positive (n = 13) and TrkA fl/fl ;Aggrecan CreERT2 -negative control mice (n = 6), followed by harvesting knee joints at 12 weeks after PMM. Each knee shown is representative of a group of mice. (A) Development of mechanical allodynia (von Frey filament testing) in the ipsilateral hind paw, comparing TrkA fl/fl ;Aggrecan CreERT2 -negative and TrkA fl/fl ;Aggrecan CreERT2 -positive mice following PMM. There were no significant differences. (B) Ambulation (horizontal photo beam crossings) and (C) spontaneous rearing activity (vertical photo beam crossings) were not different in TrkA-deleted mice compared with the control after PMM, while both activities were reduced between pre-and advanced OA. (D) Histological assessment for proteoglycan depletion by Safranin-O fast green staining (×10). (E) Severity of articular cartilage degradation was graded using the Osteoarthritis Research Society International (OARSI) scoring system. Values are mean ± SEM. * p < 0.05, ns: not significant. p = 0.644 (sham), p = 0.376 (PMM).

Figure 2.
Cartilage-specific deletion of TrkA in mice did not improve either OA-associated hypersensitivity to pain or knee joint pathology with OA progression. OA was induced by PMM in 12-week-old TrkA fl/fl ;Aggrecan CreERT2 -positive (n = 13) and TrkA fl/fl ;Aggrecan CreERT2 -negative control mice (n = 6), followed by harvesting knee joints at 12 weeks after PMM. Each knee shown is representative of a group of mice. (A) Development of mechanical allodynia (von Frey filament testing) in the ipsilateral hind paw, comparing TrkA fl/fl ;Aggrecan CreERT2 -negative and TrkA fl/fl ;Aggrecan CreERT2 -positive mice following PMM. There were no significant differences. (B) Ambulation (horizontal photo beam crossings) and (C) spontaneous rearing activity (vertical photo beam crossings) were not different in TrkA-deleted mice compared with the control after PMM, while both activities were reduced between pre-and advanced OA. (D) Histological assessment for proteoglycan depletion by Safranin-O fast green staining (×10). (E) Severity of articular cartilage degradation was graded using the Osteoarthritis Research Society International (OARSI) scoring system. Values are mean ± SEM. * p < 0.05, ns: not significant. p = 0.644 (sham), p = 0.376 (PMM).
Cartilage-specific deletion of TrkA also did not alter proteoglycan depletion, as established by histological assessment with Safranin-O fast green staining ( Figure 2D) nor the severity of articular cartilage degradation as determined using the Osteoarthritis Research Society International (OARSI) scoring system ( Figure 2E). These results suggest that loss of NGF/TrkA signaling in cartilage does not affect catabolic pathways that perturb the cartilage extracellular matrix, nor does it affect the progression of OA hyperalgesia and cartilage deterioration.
2.4. Intra-Thecal Injection of Anti-NGF Antibody Showed No Effect on Joint Pain, While Intra-Articular Injection of Anti-NGF Antibody Improved OA-Induced Chronic Pain Next, we tested anti-NGF antibody therapy to relieve pain and examined whether the action of NGF in pain transmission was peripheral. When an anti-NGF antibody was injected intra-thecally (IT) in C57Bl/6 mice at 13 weeks after PMM surgery, there was no improvement in pain reduction from chronic osteoarthritis ( Figure 3A). However, when the anti-NGF antibody was delivered locally by intra-articular (IA) injection, the OA pain was significantly reduced ( Figure 3B). IA injection of anti-NGF antibody was given twice a week to OA-induced PKCδ null mice by DMM, and significant pain reduction is shown in Figure 6H of Kc et al. [11]. These results indicate that pain is transmitted via peripheral sensory neurons and that peripheral inhibition of NGF/TrkA signaling can be effective. Hence, local inhibition of pain in OA can significantly reduce side effects, which may result from the whole-body application. Next, we tested anti-NGF antibody therapy to relieve pain and examined whether the action of NGF in pain transmission was peripheral. When an anti-NGF antibody was injected intra-thecally (IT) in C57Bl/6 mice at 13 weeks after PMM surgery, there was no improvement in pain reduction from chronic osteoarthritis ( Figure 3A). However, when the anti-NGF antibody was delivered locally by intra-articular (IA) injection, the OA pain was significantly reduced ( Figure 3B). IA injection of anti-NGF antibody was given twice a week to OA-induced PKC null mice by DMM, and significant pain reduction is shown in Figure 6H of Kc et al. [11]. These results indicate that pain is transmitted via peripheral sensory neurons and that peripheral inhibition of NGF/TrkA signaling can be effective. Hence, local inhibition of pain in OA can significantly reduce side effects, which may result from the whole-body application. -NGF antibody in C57BL/6 mice at 13 weeks after PMM did not alleviate hyperalgesia from chronic osteoarthritis, indicating that pain transmission is not a central event but may occur through peripheral sensory neurons. OA was induced by partial medial meniscectomy (PMM) in 12-week-old wild-type C57Bl/6 mice (n = 7), followed by pain test with von Frey filament. At 13 weeks after PMM, 10 g of anti-NGF antibody in 5 L (2 g/L) was injected intrathecally, and pain was assessed at 1 h, 4 h, 8 h, and 24 h after IT injection. (B) Von Frey filament testing in the ipsilateral hindpaw of PKCδ null mice (n = 10) receiving intaarticular (IA) injection of anti-NGF-2.5S antibody (30 μg in 5 μL saline) twice a week until 8 weeks after DMM surgery. Mice injected with anti-NGF antibody showed significantly reduced pain from the third week of anti-NGF IA injection. Black arrows indicate the IA injection. Adapted from Kc et al., Figure  6H [11]. * p < 0.05, ** p < 0.01, ns: not significant. p = 0.35. To understand the function of NGF/TrkA signaling in peripheral sensory neurons, we selectively and conditionally deleted TrkA using Cre recombination under control of the promoter for voltage-gated sodium channel (NaV1.8/SCN10A). TrkA fl/fl ;NaV1.8 CreERT2 mice were generated by crossing TrkA fl/fl mice with NaV1.8 CreERT2 , and conditional ablation  To understand the function of NGF/TrkA signaling in peripheral sensory neurons, we selectively and conditionally deleted TrkA using Cre recombination under control of the promoter for voltage-gated sodium channel (Na V 1.8/SCN10A). TrkA fl/fl ;Na V 1.8 CreERT2 mice were generated by crossing TrkA fl/fl mice with Na V 1.8 CreERT2 , and conditional ablation was induced by systemic administration of 2 mg tamoxifen for 5 consecutive days starting at 4 weeks of age. OA was induced by PMM in 12-week-old TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice and their corresponding littermate controls (TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice), followed by pain testing every week and harvesting knee joints, DRG, and spinal cords (SCs) at 12 weeks after PMM surgery.

Sensory Neuron-Specific
Development of mechanical allodynia was measured by von Frey filament testing in the ipsilateral hind paw every week, and the pain response was compared in TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice following PMM or sham surgery. Sensory-neuron-specific deletion of TrkA in mice (TrkA fl/fl ;Na V 1.8 CreERT2positive) abolished OA-associated hypersensitivity ( Figure 4A). This effect is comparable to global deletion of TrkA in mice (TrkA fl/fl ;Rosa CreERT2 -positive), as shown in Figure 1A. However, proteoglycan content as visualized by Safranin-O fast green staining (×20) showed that the joint continued to deteriorate upon OA-induced damage in mice with sensoryneuron-specific deletion of TrkA. Hence, loss of TrkA did not significantly improve cartilage structure protection compared to mice in the TrkA fl/fl ;Na V 1.8 CreERT2 -negative control group ( Figure 4B). Severity of articular cartilage degradation was graded using the Osteoarthritis Research Society International (OARSI) scoring system to quantitate the results of cartilage protection. The OARSI scores further confirmed no significant difference in cartilage protection against OA-induced damage with sensory neuron-specific ablation of TrkA ( Figure 4C).
was induced by systemic administration of 2 mg tamoxifen for 5 consecutive days starting at 4 weeks of age. OA was induced by PMM in 12-week-old TrkA fl/fl ;NaV1.8 CreERT2 -positive mice and their corresponding littermate controls (TrkA fl/fl ;NaV1.8 CreERT2 -negative mice), followed by pain testing every week and harvesting knee joints, DRG, and spinal cords (SCs) at 12 weeks after PMM surgery.
Development of mechanical allodynia was measured by von Frey filament testing in the ipsilateral hind paw every week, and the pain response was compared in TrkA fl/fl ;NaV1.8 CreERT2 -negative and TrkA fl/fl ;NaV1.8 CreERT2 -positive mice following PMM or sham surgery. Sensory-neuron-specific deletion of TrkA in mice (TrkA fl/fl ;NaV1.8 CreERT2 -positive) abolished OA-associated hypersensitivity ( Figure 4A). This effect is comparable to global deletion of TrkA in mice (TrkA fl/fl ;Rosa CreERT2 -positive), as shown in Figure 1A. However, proteoglycan content as visualized by Safranin-O fast green staining (×20) showed that the joint continued to deteriorate upon OA-induced damage in mice with sensoryneuron-specific deletion of TrkA. Hence, loss of TrkA did not significantly improve cartilage structure protection compared to mice in the TrkA fl/fl ;NaV1.8 CreERT2 -negative control group ( Figure 4B). Severity of articular cartilage degradation was graded using the Osteoarthritis Research Society International (OARSI) scoring system to quantitate the results of cartilage protection. The OARSI scores further confirmed no significant difference in cartilage protection against OA-induced damage with sensory neuron-specific ablation of TrkA ( Figure 4C).   Immunofluorescence staining for TrkA in the DRG of TrkA-deleted mice and control mice ( Figure 4D), as well as the quantitative analysis of TrkA expression in the DRG ( Figure 4E) indicate that TrkA protein expression in DRG is significantly reduced (over 90%, p < 0.000001), reflecting the efficiency of sensory neuron-specific deletion of TrkA in TrkA fl/fl ;Na V 1.8 CreERT2positive mice. TrkA intensity measured in ImageJ software from the National Institutes of Health (NIH imagej.nih.gov/ij/download.html: accessed on 8 July, 2020) significantly increased sixfold from a baseline of 2.96 to 17.71 upon OA induction (p < 0.0002 in sham vs. TrkA fl/fl ;Na V 1.8 CreERT2 -negative PMM mice). However, TrkA intensity decreased to 1.82 upon sensory-neuron-specific loss of TrkA (p < 0.00001 in TrkA fl/fl ;Na V 1.8 CreERT2 -negative PMM vs. TrkA fl/fl ;Na V 1.8 CreERT2 -positive PMM) ( Figure 4E). These immunofluorescence results of DRG establish that TrkA is effectively deleted by tamoxifen-induced Cre activation in sensory neurons of TrkA fl/fl ;Na V 1.8 CreERT2 mice ( Figure 4D,E).

Sensory
Neuron-Specific Deletion of TrkA in Mice (TrkA fl/fl ;Na V 1.8 CreERT2 -Positive) Reduced NGF Significantly NGF expression was significantly reduced in TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice on the basis of immunofluorescence staining ( Figure 4F) and quantitative analysis of NGF protein in DRG ( Figure 4G). NGF intensity increased dramatically (69-fold) from a baseline of 1.01 to 69.30 in TrkA fl/fl ;Na V 1.8 CreERT2 -negative OA mice (p < 0.00001 in sham vs. TrkA fl/fl ;Na V 1.8 CreERT2 -negative PMM). However, levels of NGF expression return to close to the control level 1.92 (p < 0.00001 in TrkA fl/fl ;Na V 1.8 CreERT2 -negative PMM vs. TrkA fl/fl ;Na V 1.8 CreERT2 -positive PMM) ( Figure 4F,G). These results clearly indicate that OAinduced hyperalgesia is associated with increased NGF/TrkA axis signaling in DRG.

Decreased Density of Peripheral Nerve Fibers in OA Synovium Correlated with Decreased Joint Pain Sensation
We next investigated whether sensory nerve sprouting in the knee joints of mice may contribute to hyperalgesia as a principal pathological feature of OA. Peripheral nerve innervation increased in the ipsilateral knee joint synovium at 12 weeks after PMM surgery, as shown by immunofluorescence staining with anti-protein gene product 9.5 (PGP 9.5) antibody ( Figures 5A and 6A). The results show that induction of OA by PMM surgery significantly increased the density of PGP9.5-positive nerve fibers in synovial/capsular regions of TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2 -negative mice compared to sham control mice. PGP9.5-positive nerve fiber density was calculated as the signal intensity per nerve fiber surface area, as reflected by the number of nuclei stained with 4 ,6-diamidino-2-phenylindole (DAPI) blue. PGP9.5-positive structures (green) in the knee joint synovium were significantly reduced upon loss of TrkA in both TrkA fl/fl ;Na V 1.8 CreERT2 -positive and TrkA fl/fl ;Rosa CreERT2 -positive mice compared with both negative control TrkA fl/fl ;Na V 1.8 CreERT2negative and TrkA fl/fl ;Rosa CreERT2 -negative mice ( Figures 5A and 6A), and this was confirmed in the quantification of the intensity of PGP 9.5 ( Figures 5B and 6B). PGP9.5 in shamoperated control mice were 0.69 and 0.74, and they are significantly increased 6-fold and 13-fold in intensity to 4.30 and 9.47 in TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2negative PMM induced OA mice, respectively (p < 0.0018 and p < 0.0016 in sham vs. TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2 -negative PMM mice, respectively). However, PGP9.5 expression levels in OA induced mice were completely reduced to 0.43 and 0.63 in TrkA fl/fl ;Na V 1. itive OA mice, respectively (p < 0.0008 and p < 0.0014 in TrkA fl/fl ;NaV1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2 -negative PMM vs. TrkA fl/fl ;NaV1.8 CreERT2 -positive and TrkA fl/fl ;Rosa CreERT2positive PMM, respectively) ( Figures 5A,B and 6A,B).
These results indicate that the degree of OA pain correlated with increased peripheral nerve fibers in knee joints of OA mice and deletion of TrkA in the peripheral sensory neurons. Hence, inactivation of NGF/TrkA signaling can effectively block OA-associated pain transmission.    These results indicate that the degree of OA pain correlated with increased peripheral nerve fibers in knee joints of OA mice and deletion of TrkA in the peripheral sensory neurons. Hence, inactivation of NGF/TrkA signaling can effectively block OA-associated pain transmission.

OA-Induced Angiogenesis in Joint Synovium Was Mitigated by SensoryNeuron-Specific Deletion of TrkA Mice
The endothelium marker CD31, which is a potent angiogenesis factor ( Figure 5C,D and Figure 6C,D), was monitored to understand OA-mediated angiogenesis after loss of TrkA in sensory neurons. Immunofluorescence staining for CD31 in the synovial vasculature of the ipsilateral knee joint of 12 weeks after PMM surgery showed no obvious aggressive angiogenesis in TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice and TrkA fl/fl ;Rosa CreERT2 -positive mice compared to TrkA fl/fl ;Na V 1.8 CreERT2 negative and TrkA fl/fl ;Rosa CreERT2 negative PMM mice. CD31 expression increased strongly (18-fold) in the ipsilateral knee joint of both TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice and TrkA fl/fl ;Rosa CreERT2 -negative mice at 12 weeks after PMM surgery, as shown by immunofluorescence staining with anti-CD31 antibody ( Figures 5C and 6C). However, the fluorescence intensity of CD31 (green) in knee joints was significantly reduced upon loss of TrkA in both of TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice and TrkA fl/fl ;Rosa CreERT2 -positive mice comparing both of TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2 -negative mice ( Figures 5D and 6D). CD31 intensities in sham-operated control mice were 0.73 and 1.32 and were significantly increased to 13.24 upon induction of OA in TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice (p < 0.0002) and to 24.26 in TrkA fl/fl ;Rosa CreERT2negative PMM mice (p < 0.0004), respectively. This elevation of CD31 signals was severely reduced to intensity to 2.66 in TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice (p < 0.0007) and 4.32 in TrkA fl/fl ;Rosa CreERT2 -positive PMM mice (p < 0.0015) compared to their respective negative control PMM mice ( Figure 5C,D, and Figure 6C,D).
Consistent with these findings, VEGF expression was strongly increased (14-fold and 20-fold, respectively) in the ipsilateral knee joint of TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2 -negative mice upon induction of OA at 12 weeks after PMM surgery, as measured by immunofluorescence signals for anti-VEGF antibody (Figures 5E and 6E). However, VEGF expression in the knee joint synovium was significantly reduced upon loss of TrkA in two different genetic mouse models comparing both TrkA fl/fl ;Na V 1.8 CreERT2positive and TrkA fl/fl ;Rosa CreERT2 -positive mice with both TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice and TrkA fl/fl ;Rosa CreERT2 -negative mice (Figures 5F and 6F). As expected, VEGF intensities were significantly increased by OA induction, comparing 0.53 and 0.90 in shamoperated control mice to 7.42 with TrkA fl/fl ;Na V 1.8 CreERT2 -negative PMM mice (p < 0.015) and 18.18 with TrkA fl/fl ;Rosa CreERT2 -negative PMM mice (p < 0.013), respectively. However, this increase in VEGF levels was strongly mitigated (85% and 97%) by loss of TrkA to intensity 1.59 (p < 0.025) in TrkA fl/fl ;Na V 1.8 CreERT2 -positive OA mice and 1.36 (p < 0.015) in TrkA fl/fl ;Rosa CreERT2 -positive PMM mice ( Figure 5E,F and Figure 6E,F).
Collectively, our results indicate that NGF/TrkA signaling in sensory nerves is important for angiogenesis and that inhibition of this pathway may alleviate both OA-enhanced angiogenesis and pain sensation.

SensoryNeuron-Specific Deletion of TrkA Mitigated Pro-Inflammatory Cytokines
We next measured tumor necrosis factor-α (TNF-α/TNF) and interleukin-1β (IL-1β/IL1B) expression, the most common pro-inflammatory cytokines in OA pathology. To identify the tissue source of inflammatory cytokine production, we examined expression levels of TNF-α and IL-1β. Protein levels of TNF-α ( Figure 5G,H and Figure 6G,H) and IL-1β ( Figure 5I,J and Figure 6I,J) significantly increased in cartilage and synovium at 12 weeks after PMM (comparing TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2negative mice with sham control mice). However, the OA-induced elevation of these cytokines was impeded substantially upon TrkA loss in TrkA fl/fl ;Na V 1.8 CreERT2 -positive and TrkA fl/fl ;Rosa CreERT2 -positive mice.
TNF-α expression was over 16-fold increased in the ipsilateral knee joint of both TrkA fl/fl ;Na V 1.8 CreERT2 -negative and TrkA fl/fl ;Rosa CreERT2 -negative mice at 12 weeks after PMM surgery, as shown in immunofluorescence staining ( Figures 5G and 6G). However, TNF-α expression (green fluorescence) was 60% reduced in knee joints of TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice, while TrkA fl/fl ;Rosa CreERT2 -positive mice exhibited 93% reduction compared to their control PMM mice ( Figures 5G and 6G), and this was confirmed in the quantification of the density of TNF-α ( Figures 5H and 6H). TNF-α intensity in shamoperated control mice was 0.22, and it was greatly increased to an intensity of 24.14 in TrkA fl/fl ;Na V 1.8 CreERT2 -negative OA mice (p < 0.0001) and the intensity increased from 1.6 to 25.8 in TrkA fl/fl ;Rosa CreERT2 -negative PMM mice (p < 0.0005). However, expression of TNF-α in OA-induced mice was reduced to 9.84 in TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice (p < 0.0018) and 3.21 in TrkA fl/fl ;Rosa CreERT2 -positive mice (p < 0.001) (Figures 5H and 6H).
IL-1β expression was increased 2.7-fold in the ipsilateral knee joint of TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice and increased 34.7-fold in TrkA fl/fl ;Rosa CreERT2 -negative mice at 12 weeks after PMM surgery, as shown by immunofluorescence staining with anti-IL-1β antibody ( Figures 5I and 6I). However, green fluorescence indicating IL-1β expression in knee joints was reduced 70% in TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice, while TrkA fl/fl ;Rosa CreERT2 -positive mice showed 74% reduction in fluorescence intensity ( Figures  5J and 6J). IL-1β intensity in sham-operated control mice was 3.96, and it was significantly increased to an intensity of 10.73 in TrkA fl/fl ;Na V 1.8 CreERT2 -negative OA mice (p < 0.00017) and increased from 0.53 to 18.5 in TrkA fl/fl ;Rosa CreERT2 -negative mice (p < 0.028). The expression decreased to an intensity of 5.97 in TrkA fl/fl ;Na V 1.8 CreERT2 -positive PMM mice (p < 0.029) and 5.29 in TrkA fl/fl ;Rosa CreERT2 -positive OA-induced mice (p = 0.073) (Figures 5J and 6J).
Although TrkA-deleted OA mice showed substantial reductions in the expression of pro-inflammatory cytokines, this inhibition was only partial because pro-inflammatory cytokine activities still remained in the knee joints of PMM-induced TrkA cKO mice and facilitated the destruction of joint integrity. This result may explain why loss of TrkA greatly relieves pain transmission (see Figure 4A) but does not provide noticeable protection in joint pathology shown (see Figure 4B,C).

Cartilage-Degrading Enzyme MMP-13 Remained Activated upon SensoryNeuron-Specific Deletion of TrkA
The expression of cartilage-degrading enzyme MMP-13/MMP13, a critical target in OA progression, was not eliminated in OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice or TrkA fl/fl ;Rosa CreERT2 -positive mice (Figures 5K and 6K) at all. MMP-13 expression was increased 16-fold in the ipsilateral knee joint of TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice and increased 22-fold in TrkA fl/fl ;Rosa CreERT2 -negative mice at 12 weeks after PMM surgery, compared to the sham control, as shown by immunofluorescence staining (Figures 5K and 6K). However, loss of TrkA did not reduce MMP-13 level in OA-induced TrkA deleted mice (TrkA fl/fl ;Na V 1.8 CreERT2 -positive and TrkA fl/fl ;Rosa CreERT2positive mice compared with their negative control mice), as confirmed in the intensity quantification ( Figures 5L and 6L). MMP-13 level in sham-operated mice was 1.12, and it was increased 16-fold (intensity 17.2) in OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice (p < 0.001) and increased from 1.09 to 23.5 in TrkA fl/fl ;Rosa CreERT2 -negative mice (p < 0.001). However, expression of MMP-13 in OA-induced TrkA cKO mice remained high in both sensory-neuron-specific deletion (17.0, p = 0.91) and in global deletion (23.5, p = 0.99) of TrkA (Figures 5L and 6L).
Immunofluorescence of MMP-13 showed NGF/TrkA is not a critical mediator in joint destruction in OA. Deletion of TrkA, and thus inhibition of NGF signaling specifically at the sodium channel sensory neuron (Na V 1.8), fails to eliminate MMP-13, a cartilagedegrading enzyme in PMM-induced OA mice. This finding indicates that TrkA/NGF signaling impacts pain transmission but not the joint pathology, at least in part, because it has no effect on the MMP-13 activity in mediating cartilage destruction.

SensoryNeuron-Specific and Global Deletion of TrkA Reduced SP and CGRP Significantly in the DRG of OA-Induced Mice
Even though the degree of cartilage degeneration was similar in both TrkA fl/fl ;Na V 1.8 CreERT2positive and TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice ( Figure 4B,C), there was significant pain relief in TrkA fl/fl ;Na V 1.8 CreERT2 -positive PMM mice ( Figure 4A). Therefore, we compared the expression levels of pain transmitter substance P (SP) and calcitonin-gene-related peptide (CGRP) in DRG.
Immunofluorescence staining for substance P (SP: green) co-stained with the neuronal marker NeuN (red) in DRG ( Figure 7A) and their quantitative analysis of SP expression in the NeuN-positive cells in DRG ( Figure 7C) exhibited significant reduction in both TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice and TrkA fl/fl ;Rosa CreERT2 -positive OA mice (p < 0.003). SP intensity in sham control mice was 0.43 and was significantly increased to 78 in TrkA fl/fl ;Na V 1.8 CreERT2 -negative PMM mice (p < 0.0003) and 37.25 in TrkA fl/fl ;Rosa CreERT2negative OA mice (p < 0.00005). However, loss of TrkA resulted in 99% reduction of SP expression to control level 0.80 in TrkA fl/fl ;Na V 1.8 CreERT2 -positive OA mice (p < 0.00030) and 77% reduction to 8.65 in TrkA fl/fl ;Rosa CreERT2 -positive OA mice (p < 0.004) ( Figure 7A,C).
This indicates that the deletion of TrkA in sensory neurons significantly reduces pain molecules SP and CGRP in DRG. SP and CGRP are also known as angiogenesis neuropeptides and it is suggested that the NGF/TrkA signaling in sensory nerve has significant effects on angiogenesis for alleviating pain sensation and OA-induced hyperalgesia.

SensoryNeuron-Specific Deletion of TrkA in Mice (TrkA fl/fl ;NaV1.8 CreERT2 -Positive) Reduced Glial Cell Activity, Inflammatory Cytokines, and CGRP Significantly in Spinal Cords
In order to determine whether the mechanism involved in the improvement of hyperalgesia by peripheral sensory neuronal inhibition of NGF/TrkA signaling was based on retrograde centralized pain transportation, we examined glial cell activity, inflammatory cytokines, and pain molecules in the spinal cords of OA-induced mice. The microglial cell marker IBA-1 in OA-induced TrkA fl/fl ;NaV1.8 CreERT2 -negative mice was increased sixfold in spinal cords (p < 0.0004) and fourfold in TrkA fl/fl ;Rosa CreERT2 -negative OA mice (p < 0.0031) but was reduced by 83% in OA-induced TrkA fl/fl ;NaV1.8 CreERT2 -positive mice (p < 0.0029) and 100% in TrkA fl/fl ;Rosa CreERT2 -positive OA mice (p < 0.0009). Quantitative analyses of IBA-1 was calculated as the intensity of IBA-1(green) area ( Figure 8A,B). This indicated the increased microglial cell activity due to OA induction was greatly reduced in TrkA cKO mice in that NGF and pain molecule retrograde transportation was diminished by deletion of TrkA in sensory neurons.  In order to determine whether the mechanism involved in the improvement of hyperalgesia by peripheral sensory neuronal inhibition of NGF/TrkA signaling was based on retrograde centralized pain transportation, we examined glial cell activity, inflammatory cytokines, and pain molecules in the spinal cords of OA-induced mice. The microglial cell marker IBA-1 in OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice was increased sixfold in spinal cords (p < 0.0004) and fourfold in TrkA fl/fl ;Rosa CreERT2 -negative OA mice (p < 0.0031) but was reduced by 83% in OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice (p < 0.0029) and 100% in TrkA fl/fl ;Rosa CreERT2 -positive OA mice (p < 0.0009). Quantitative analyses of IBA-1 was calculated as the intensity of IBA-1(green) area ( Figure 8A,B). This indicated the increased microglial cell activity due to OA induction was greatly reduced in TrkA cKO mice in that NGF and pain molecule retrograde transportation was diminished by deletion of TrkA in sensory neurons.

8C,D).
This finding indicated that astroglial cell activation due to OA induction was abolished in TrkA fl/fl ;NaV1.8 CreERT2 -positive mice and drastically reduced in TrkA fl/fl ;Rosa CreERT2positive OA mice. This change in astroglial activation was due to deletion of TrkA in sensory neurons, which impedes retrograde transportation of augmented NGF levels upon OA induction. Hence, astroglial activity to produce inflammatory cytokines and pain molecules is significantly abolished. Interestingly, complete reduction of astroglial cell activity is achieved in sensory-neuron-specific deletion in that NGF and pain molecule retrograde transportation is solely accountable to deletion of TrkA in sensory neurons.  The astroglial cell marker GFAP in OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice was increased 16-fold and in TrkA fl/fl ;Rosa CreERT2 -positive OA mice in spinal cords compared to sham control mice, which increased 8-fold (p < 0.0001 and p < 0.0025, respectively). However, loss of TrkA was reduced completely in OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice (p < 0.00008) and by 93% in TrkA fl/fl ;Rosa CreERT2 -positive OA mice (p < 0.003) ( Figure 8C,D).
This finding indicated that astroglial cell activation due to OA induction was abolished in TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice and drastically reduced in TrkA fl/fl ;Rosa CreERT2positive OA mice. This change in astroglial activation was due to deletion of TrkA in sensory neurons, which impedes retrograde transportation of augmented NGF levels upon OA induction. Hence, astroglial activity to produce inflammatory cytokines and pain molecules is significantly abolished. Interestingly, complete reduction of astroglial cell activity is achieved in sensory-neuron-specific deletion in that NGF and pain molecule retrograde transportation is solely accountable to deletion of TrkA in sensory neurons.
Even though sensory-neuron-specific deletion of TrkA exhibited partial reduction of pro-inflammatory cytokines in the knee joints of OA-induced TrkA fl/fl ;Na V 1.8 CreERT2positive mice, expression of TNF-α and IL-1β was reduced by over 90% in spinal cords. This indicated that glial cell activity due to retrograde pain transportation was significantly reduced by the sensory neuron-specific deletion of TrkA.
Loss of TrkA completely inhibited CGRP expression in the SC of OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice and TrkA fl/fl ;Rosa CreERT2 -positive mice ( Figure 8I,J) compared to their negative control PMM mice. CGRP intensity in sham-operated mice 2.0 had a 6.5-fold increase to 13.2 in OA-induced TrkA fl/fl ;Na V 1.8 CreERT2 -negative mice (p < 0.017) and a 7-fold increase to 14.1 in OA-induced TrkA fl/fl ;Rosa CreERT2 -negative mice (p < 0.003). However, expression of CGRP in PMM-induced mice was reduced to 0.86 in TrkA fl/fl ;Na V 1.8 CreERT2positive PMM mice (p < 0.012) and to 2.73 in TrkA fl/fl ;Rosa CreERT2 -positive OA mice (p < 0.004), indicating complete inhibition of glial activity ( Figure 8I,J) was achieved in sensory-neuronspecific TrkA-deleted mice. Finally, we investigated whether increased peripheral sensory NGF due to OA stimulated brain derived neurotrophic factor (BDNF), a key pain mediator in the central nervous system (CNS), and whether BDNF is diminished upon sensory-neuron-specific deletion of TrkA, hence inhibiting the NGF/TrkA signaling pathway.
Immunofluorescence staining of BDNF in TrkA fl/fl ;Na V 1.8 CreERT2 -positive mice and TrkA fl/fl ;Rosa CreERT2 -positive mice following PMM surgery ( Figure 8K,L) showed complete inhibition of BDNF expression in SCs compared to control PMM mice. BDNF intensity in sham control mice was 0.79, and it was increased 13-fold to 10.13 in TrkA fl/fl ;Na V 1.8 CreERT2negative OA mice (p < 0.019) and increased 14-fold to 11.29 in TrkA fl/fl ;Rosa CreERT2 -negative OA mice (p < 0.00003). However, expression of BDNF was reduced to 0.41 in TrkA fl/fl ;Na V 1.8 CreERT2 -positive PMM mice (p < 0.016) and to 2.67 in TrkA fl/fl ;Rosa CreERT2positive OA mice (p < 0.0004). This showed a complete inhibition of the central pain mediator BDNF ( Figure 8K,L) was accomplished by sensory-neuron-specific TrkA deletion, thus disabling the NGF/TrkA signaling pathway.

Discussion
OA is a disease of the whole joint, and OA pain is associated with the presence of bone marrow lesions and synovitis [26]. Our previous studies have shown that ablation of the PKCδ/Prkcd gene prevents cartilage destruction and yet exacerbates OA-associated pain [11]. Pain perception during progression of OA may not strictly correlate with the degree of cartilage damage because neurons and blood vessels are absent in mature articular cartilage.
Here, we present evidence that the NGF/TrkA pathway is critical for transducing pain and changes in nerve processing, as well as the fact that both peripheral and central sensitization may contribute to pain in OA. Augmentation of NGF/TrkA signaling in the joint synovium and the peripheral level of sensory neurons are the major determinants that facilitate pro-nociception and the transition of OA from an asymptomatic joint degenerative process to a painful disease. However, inhibiting the NGF/TrkA pathway to desensitize OA-associated hyperalgesia does not improve cartilage preservation, consistent with the concept that OA pain may develop independently of joint destruction.
Many growth factors and cytokines, including NGF, bFGF, VEGF, and IL-1, are regulated by autocrine/paracrine feedback mechanisms. We observed that blocking VEGFR1/Flt1 or VEGF2/Flk1 significantly downregulated VEGF expression (Ma et al., unpublished data). We previously observed that augmentation of NGF caused increased expression of TrkA, which is the cognate receptor for NGF [11]. In our current study, increased NGF transportation to DRG was impaired due to sensory-neuron-specific deletion of NGF receptor TrkA. Therefore, NGF was downregulated after TrkA was deleted in the sensory neurons.
In knee joints, NGF is produced by fibroblast-like synoviocytes and/or chondrocytes under OA conditions. Yet, cartilage-specific deletion of TrkA (as the cognate NGF receptor) fails to protect from OA-associated hyperalgesia. This finding suggests that NGF signaling in cartilage may not be catabolic or that TrkA deletion in cartilage is not sufficient to block the progression of OA. In contrast, deletion of TrkA in sensory neurons modulated the activity of NGF, angiogenesis, and pro-inflammatory cytokines in the knee joint of mice upon OA induction after PMM surgery. More importantly, mice experienced remarkable pain relief during early to advanced stages of OA in our study. The protection from OAinduced hyperalgesia upon sensory-neuron-specific deletion of TrkA is directly supported by our observations that loss of TrkA decreased peripheral nerve fibers in knee joints and reduced angiogenesis.
Expression of pro-inflammatory cytokines was partially reduced in the synovium but was completely abolished in the spinal cord. This result indicates that these proinflammatory cytokines in the synovium play important roles in joint integrity. The com-plete ablation of cytokine production observed in the spinal cord was the result of the inactivation of glial cells, which are able to produce these inflammatory factors when provoked. NGF may upregulate IL-1β through TrkA and NF-κB/NFKB1-dependent caspase-1/CASP1 activation in human monocytes [21]. Pro-inflammatory cytokines TNF-α and IL-1β were upregulated in cartilage and synovium in advanced human OA and upon surgical induction of OA in mice [11]. Ligand binding of NGF to TrkA activates transcriptional and posttranscriptional pathways that mediate IL-1β release [27], while IL-1β induces chondrocyte inflammation and osteoarthritis via the NF-κB signaling pathway [28]. Sensory neuron-specific deletion of TrkA in OA-induced mice did not fully deactivate pro-inflammatory cytokine production, in agreement with the idea that the levels of TNF-α and IL-1β correlate with joint pathology but not with the level of pain.
The matrix metalloprotease MMP-13/MMP13 is a major collagen-degrading enzyme responsible for cartilage destruction in OA, and its inhibition has been considered a strategy to block OA development [29]. Our study showed that MMP-13 was still fully active during OA progression in mice upon sensory-neuron-specific TrkA deletion or global TrkA deletion. Therefore, we propose that MMP-13 is a critical factor for joint pathology protection, but its expression is not completely dependent on the NGF/TrkA pathway.
Deletion of TrkA in sensory neurons significantly reduced expression of pain molecules such as SP and CGRP in DRG and spinal cords that also act as angiogenic neuropeptides. NGF/TrkA signaling in sensory neurons stimulated angiogenesis factors such as VEGF and CD31, and, as a consequence, contributed to angiogenesis. This result suggests that NGF/TrkA signaling in sensory nerves contributes significantly to angiogenesis while alleviating pain sensation and OA-induced hyperalgesia.
Activated glial cell activity in our OA mouse model may have involved retrograde centralized pain transportation as a key mechanism that improves hyperalgesia upon peripheral sensory neuronal inhibition of NGF/TrkA signaling. Both microglial and astroglial cell activities were modulated, indicating that hyperalgesia in both early and advanced OA can be treated by TrkA inactivation. OA induces expression of the brain-derived pain mediator BDNF by peripheral NGF, and this induction of BDNF was completely abolished upon sensory-neuron-specific deletion of TrkA, which inactivates the NGF/TrkA signaling pathway.
Our findings strongly support an emerging concept for the etiology of OA in that the severity of joint pain is due to pathological changes in the synovium as well as cellular and molecular plasticity in the sensory neurons of the innervating DRG. Our results also support the mechanistic interpretation that NGF/TrkA signaling is one of the key determinants for the painful symptomatic transition of OA. Peripheral inhibition of NGF/TrkA signaling can effectively reduce the pain mechanism that operates via retrograde centralized pain sensitization. Hence, sensory-neuron-specific deletion of TrkA is as efficient as a global deletion and is sufficient for alleviating pain. Therefore, disease intervention would benefit from local inactivation of TrkA to eliminate the risk and safety issues of systemic treatments (which are modulated in mouse OA by global deletion of TrkA). Beyond localized inhibition of TrkA, other interventions will be required for cartilage repair and protection in advanced OA.
Clinical studies with analgesics, including non-steroidal anti-inflammatory drugs (NSAIDs), opioid agonists, and calcitonin, have been evaluated for therapeutic benefits to treat chronic pain [30]. Our finding that intervention in NGF/TrkA signaling suppresses OA hyperalgesia has important clinical ramifications and could be developed further as a superb and safe therapy against chronic pain.
One study limitation is that our findings on NGF/TrkA signaling in nociceptive responses associated with OA were based on a post-traumatic surgical mouse model for OA that may not strictly correlate with the age-related degenerative human OA. However, our results can be further tested for human applications using human-derived sensoryneuron-like stem cells. Another limitation is that targeting NGF/TrkA signaling is effective only for the alleviation of nociception but not for preserving joint pathology. Therefore, future studies may require the intercalation of OA disease-modifying drugs that inhibit nociception and preserve joint integrity simultaneously. A new experimental model could include chondrocytic three-dimensional cultures with primary cells from human patients and mouse models in which NGF/TrkA signaling is modified using nanoparticles with controlled release. Immunofluorescence microscopy and Western blots are both semiquantitative because immunofluorescence and chemiluminescence each have a limited dynamic range. We were not able to present additional assessments of NGF/TrkA in Western blotting due to the size limitation of DRG, where the NGF receptor TrkA is selectively expressed in nociceptive DRG neurons [31]. However, fluorescence signals are exquisitely sensitive and provide definitive information of the local presence of proteins in situ within histological slides. In our study, we used well-characterized antibodies that have been extensively used in the field to detect TrkA and NGF in mouse sensory neurons in the DRG in situ. Our study did not present cortex/hippocampus-related data since Na V 1.8 is typically selectively expressed at high levels in sensory ganglion neurons but not within the CNS. Moreover, no loss of neuronal cells nor any differences in motor function or other abnormal behavior in tamoxifen-inducible Cre-ERT2 recombinase deletion has been reported [32]. The behavioral response to pain assessments between global deletion of TrkA and sensory-neuron-specific deletion of TrkA (Na V 1.8 expressing neurons) are similar.
Our findings can be summarized as follows: (1) Conditional whole-body deletion of TrkA in mice abolished OA pain. (2) Sensory neuron-specific deletion of TrkA alleviated OA pain as effectively as whole-body deletion of TrkA in mice. (3) Cartilage-specific deletion of TrkA in mice led to development of joint pain and damage in OA, showing no improvement on OA progression. (4) Peripheral sensory neuron-specific deletion of TrkA modulated OA-induced sensory neuron plasticity and abolished OA pain by interfering with the NGF/TrkA signaling pathway of retrograde pain transmission. (5) Sensory neuronal TrkA deletion did not downregulate the cartilage-degrading enzyme, hence leading to cartilage degeneration progresses. (6) Peripheral inhibition of NGF/TrkA signaling can effectively reduce pain, but other interventions are needed for cartilage repair and protection. In conclusion, this study provides compelling evidence for the role of NGF/TrkA in OArelated pain perception and supports the idea that this pathway is a viable target for novel disease-modifying pharmacotherapies.
Tamoxifen-inducible conditional deletion of TrkA was achieved globally with TrkA fl/fl ;Rosa CreERT2 mice, cartilage-specific deletion with TrkA fl/fl ;Aggrecan CreERT2 mice, and sensory neuron-specific deletion with TrkA fl/fl ;Nav1.8 CreERT2 mice by intraperitoneal injection of tamoxifen (2 mg/day) for 5 consecutive days at 4 weeks of age. The dose was optimized and ensured for the safety of growing young mice bone health [33]. Our Nav1.8-CreERT2 constructs were specific for sensory neurons, and the Na V 1.8 driver does not support Cre recombination in the CNS [32,34,35]. Tamoxifen-independent recombination was not a major concern in our system, since less than 1% of neurons showed Cre expression in the absence of tamoxifen treatment [32].
Mice were housed under standard laboratory conditions (in a temperature-controlled (21 ± 1 • C) room with a normal 12 h light/12 h dark cycle). Animal studies were performed according to guidelines in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animal procedures were approved by the Jessie Brown Veterans Affairs Medical Center's Institutional Animal Care and Use Committee (IACUC,