Sensory nerves directly promote osteoclastogenesis by secreting Cyp40

afferent sensors found and directly alter physiology 1 . sensory nerves been in and the direct evidence We treated mice with resiniferatoxin (RTX) or complete Freund’s adjuvant (CFA) to induce sensory hypersensitivity. Bone histomorphometry including 48 micro-ct, three-point bending assay, von kossa staining, calcein double 49 labeling, toluidine blue staining, and trap staining were performed to monitor bone quality and bone cells. Multiple virus vectors were applied to trace 51 signals between sensory nerves and osteoclasts. Sensory neurons (SN) and 52 osteoclasts were cocultured to study the effects and mechanisms of the sensory nerves on osteoclasts in vitro. Isobaric tag for relative and absolute 54 quantitation (iTRAQ) was used to identify secreted proteins in the sensory nerve. direct, and quick response of sensory nerves to the changes in bone. Targeting the Cyp40 could therefore be a strategy to promote bone repair at the early stage of bone injury.

This study revealed a novel mechanism of sensory nerves on osteoclasts: the 68 direct promotion of osteoclastogenesis by the Cyp40. This mechanism may 69 represent a direct, and quick response of sensory nerves to the changes in 70 bone. Targeting the Cyp40 could therefore be a strategy to promote bone 71 repair at the early stage of bone injury.
Background 78 In the tissues that are closely contacted with the external environment, such 79 as skin, lung and gut, sensory nerves detect damaging stimuli, and can 80 regulate the ensuing immune response by releasing neuropeptides seceretion 6, 81 7,8 . However, other than the perception of pain, the role of sensory endings in 82 the deep tissue such as the bone is not well understood. 83 Bones are innervated by a prolific network of neurochannels, around 77% of 84 which consisted of sensory endings 9 . In bone, sensory endings closely 85 contact with osteoblasts 10 , or osteoclasts 11 . Multiple studies have found 86 sensory nerves have a direct effect on osteoblasts and bone formation 12,13,14 . 87 A novel direct mechanism by which sensory nerves regulates HSC 88 mobilization in bone marrow was also found recently 9,15 163 and RANKL (100 ng/mL, Peprotech) used, accompanied by treatments (co-164 culture, sensory nerve homogenate, recombinant proteins) for another 3 days. 165 On day 6, cells were fixed and stained for tartrate-resistant acid phosphatase 166 (Trap) using the leukocyte acid phosphatase kit (Sigma). Trap+ osteoclasts 167 with more than three nuclei were quantified using ImageJ software.   (Fig 1b). However, trabecular 314 thickness (Tb.Th), trabecular bone surface/bone volume (Tb.BS/BV), and all 315 key parameters in cortical bone were not significantly altered (Fig.S1a). 316 Three-Point Bending Test also showed the decrease of bone quality (Fig 1c). 317 Serum biochemical analyses showed a dramatic reduction of serum inorganic 318 phosphorus (P, Fig. 1d). Since the serum P abnormal were always induced by 319 renal dysfunction, we tested serum biochemical makers (Fig.S1c) related to 320 renal function, and found all serum renal markers were not changed in RTX 321 mice. Our results above have hinted the efferent influence of sensory nerve 322 on bone homeostasis. 323 Von Kossa staining showed decreased osteoid in RTX treat mice (Fig 1e). 324 Calcine double labeling confirmed the reduced bone formation and mineral 325 apposition rate (Fig 1f). Toluidine blue staining showed decreased osteoblast 326 function and bone formation in RTX mice (Fig 1g). Accordingly, the serum 327 level of procollagen type N-terminalpropeptide (P1NP , Fig 1h) was 328 significantly decreased. All these were consistent with our published before, 329 that is sensory nerve exert direct efferent regulation on bone marrow stem 330 cells (BMSC) differentiation 14 . Then we found the serum levels of collagen 331 type I cross-linked C-telopeptide (CTX), an osteoclast bone resorption 332 marker, were significantly elevated in RTX treated mice compared with 333 control littermates (Fig 1h). 334 In order to further dissect the mechanism of sensory nerve action, we purified 335 sensory neuron (SN) from the dorsal root ganglion (DRG, Fig. S1.d). The SN 336 were isolated co-cultured with osteoclasts such that direct cell contacts were 337 avoided ( Fig. 1.i). The presence of SN significantly increased the number of osteoclasts (Fig 1j), osteoclast resorption activity (Fig 1 k), and resorption 339 related genes (Fig 1 l) in the bone cells from PNS and from sensory nerve (Fig 2g, and Fig 2c-f). 378 Our results showed that the PHP.S-EGFP signal also enters various cells in 379 the bone (Fig 2h). But we did not find significant difference in the distribution of signals from sensory nerves and peripheral nerves into 381 osteoclasts (Fig 2i). This suggests that the regulation of bone by sensory 382 nerves is a common process. 383 384 Cyp40 is crucial in sensory nerve elevating osteoclastogenesis 385 To search for key factors through which sensory nerve elevates 386 osteoclastogenesis, we screened the neuropeptides in sensory nerve axons. 387 Since the functional neuropeptides synthesized in the dorsal root ganglion are 388 transported peripherally and stored in the axon, before being released to 389 target cells 26 . Isobaric tag for relative and absolute quantitation (iTRAQ) was 390 used to identify proteins in saphenous nerve axoplasm (Fig 3a). The  Both of Cyp40 and Mif increase osteoclastic differentiation (Fig 3 f, g) and 411 resorption activity of osteoclasts (Fig 3 h, i). To further study the roles of the 412 two factors in osteoclastogenesis promotion by sensory nerve, we 413 downregulated Cyp40 (shCyp40) and Mif (shmif) in SN (Fig S3 a, b). 414 Osteoclasts were then cultured with the modified SN. Downregulation of 415 Cyp40 significantly attenuated the ability of the SN to promote 416 osteoclastogenesis. However, downregulation of Mif did not significantly 417 affect the ability of SN to promote osteoclastogenesis (Fig. 3 j-m). Thus, we 418 concluded that Cyp40 is crucial in the ability of SN to promote 419 osteoclastogenesis in vitro. 420 Further, we screened the instant alteration of secrete proteome after sensory 421 hypersensitivity. We applied CFA to induce sensory hypersensitivity ( Fig  S4.a), and found 272 secreted proteins changed 48 hours after CFA injection 423 (Fig S4.b). KEGG enrichment of differentially proteins showed top 15 424 pathways. Among them, proteins associated with tight junction showed the 425 most prominent difference (Fig S4.c). Notably, Cyp40 were downregulated 426 by CFA (Fig S4.d).

428
Cyp40 is a neuropeptide, enters osteoclasts by transmembrane, 429 downregulates Ras/c-Raf/p-Erk to promote osteoclastogenesis 430 Back in vivo, we found Cyp40 level was upregulated in the RTX treat mice 431 (Fig 4a), including serum Cyp40 level (Fig 4b); the recombinant Cyp40 432 dramatic reduced osteoblastic differentiation (Fig S3 d); these results 433 confirmed the crucial role of Cyp40 in the sensory nerve efferent functions. 434 Then we constructed and transfected an EGFP-tagged Cyp40 vector into SN 435 to verify if Cyp40 is released from SN into osteoclasts (Fig S3 a, b).  culturing the modified SN with osteoclasts led to the detection of EGFP-437 tagged Cyp40 in osteoclasts (Fig 4c), indicating Cyp40 was secreted by SN 438 and taken up by osteoclasts. ELISA results show a dose-response curve of 439 Cyp40 in response to increasing numbers of SN (Fig 4d). These results 440 therefore prove that Cyp40 is a secreted factor from SN. 441 The transportation of Cyp40 has not yet been studied 32 . Then we traced the 442 transportation of Cyp40 between cells. In the co-culture medium, Cyp40 was 443 found outside exosomes (Fig 4e), indicating Cyp40 was not transported 444 between SN and osteoclasts via exosomes. IEM that targets Cyp40 shows the 445 factor entering osteoclasts through the cell membrane (Fig 4f, II, and Fig 4g; 446 red arrow), instead wrapped by membrane (Fig 4f, II, and Fig 4g; green 447 arrow). Since Cyp40 is abundant in neurons, we detected its traffic also in the 448 brain. Cyp40 is transported between neurons in the brain also through the cell 449 membrane (Fig 4g, red arrow). As shown in the figures, Cyp40 was crossing 450 the cell membrane (g 4f, II, and Fig 4g, red arrow). After entering 451 osteoclasts, Cyp40 became widely distributed in the cells. It was present at 452 low levels in the nucleus (Fig 4f, III) and higher levels throughout the 453 cytoplasm (Fig 4f, I), including the ruffled border responsible for bone 454 resorption (Fig 4d, IV). Thus, Cyp40 secreted from SN enters osteoclasts via 455 a non-exosomal mechanism to promote osteoclastogenesis. 456 Ras/c-Raf/p-Erk signaling has been implicated in osteoclast survival, 457 proliferation, apoptosis, formation, polarity, and differentiation 33 . Thus, we 458 studied the effects of SN on ERK signaling in osteoclasts. We found that SN in the RTX treat mice, when Cyp40 was upregulated in DRG from the same mice (Fig 4i). These results indicates that Cyp40 plays a crucial role in the 465 ability of sensory nerves to negatively regulate Ras/c-Raf/p-Erk in osteoclasts 466 (Fig 4h). P-ERK can inhibit osteoclastogenesis by downregulating 467 NFATc1 34, 35 . NFATc1 is a key transcription regulator in osteoclasts 36 . We 468 found an increase in NFATc1 mRNA expression level and simultaneous 469 increases in osteoclastogenesis when p-Erk was downregulated in osteoclasts 470 (Fig 4j). Moreover, these changes in NFATc1 expression were dependent on 471 Cyp40 (Fig 4j). Overall, these results suggest that the sensory nerves promote found a decrease of AhR in RTX treat mice (Fig 4i). Then, we found an 482 interaction between AhR in osteoclasts and Cyp40 from SN (Fig 5a). 483 Furthermore, western blotting showed that co-culture with SN reduced the 484 expression of AhR in osteoclasts; and the AhR downregulation was 485 attenuated by knocking down Cyp40 in SN (Fig 5b). Above results indicate 486 Cyp40 binds to and downregulates AhR in osteoclasts. 487 Then, we found that AhR knock down in osteoclast dramatic decreased 488 osteoclastogenesis, and significantly attenuated sensory efferent promotion 489 on osteoclastogenesis (Fig 5c). Moreover, Ras/c-Raf/p-Erk level was also 490 decreased after AhR knock down in osteoclasts (Fig 5d).  (Fig 5g). Collectively, these findings suggest that 498 Cyp40 binds, and down-regulates AhR, to decrease Ras/c-Raf/p-Erk level, 499 and finally promote osteoclastogenesis.  47,48,49,50,51,52,53

855
(f) Maximal loading of femur by three-point bending assay.

856
(g) ELISA analysis of serum CTX and P1NP levels.