Cutting Edge: Parathyroid Hormone Facilitates Efferocytosis in Bone Marrow via Proresolving Mediators D1 and Resolvin D2 This information is current as of September 30, 2021. Laurie K. McCauley, Jesmond Dalli, Amy J. Koh, Nan Chiang and Charles N. Serhan J Immunol 2014; 193:26-29; Prepublished online 2 June 2014; doi: 10.4049/jimmunol.1301945 Downloaded from http://www.jimmunol.org/content/193/1/26

Supplementary http://www.jimmunol.org/content/suppl/2014/05/31/jimmunol.130194 http://www.jimmunol.org/ Material 5.DCSupplemental References This article cites 20 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/193/1/26.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 30, 2021

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Th eJournal of Cutting Edge Immunology

Cutting Edge: Parathyroid Hormone Facilitates Macrophage Efferocytosis in Bone Marrow via Proresolving Mediators Resolvin D1 and Resolvin D2 Laurie K. McCauley,*,† Jesmond Dalli,‡ Amy J. Koh,* Nan Chiang,‡ and Charles N. Serhan‡ Bone marrow stimulate skeletal wound fates, but none has detailed the aftermath of apoptotic osteo- repair and osteoblastic bone formation by poorly de- blasts. Inefficient apoptotic clearance leads to a local inflam- fined mechanisms. Specialized proresolving mediators matory environment (1). Hence, there are likely mechanisms of inflammation drive macrophage efferocytosis (phago- for the efficient removal of apoptotic osteoblasts, and a likely

of apoptotic cells) and resolution, but little is facilitator is the resident macrophage. The term “osteomac” has Downloaded from known regarding this process in the bone marrow. In been attributed to macrophages credited with stimulating bone this study, metabololipidomic profiling via liquid chro- formation through mechanisms that link to their phagocytic matography mass spectrometry revealed higher levels capacity (2). Macrophages facilitate apoptotic cell clearance in of specialized proresolving mediators in the bone mar- a process termed efferocytosis, which converts the dead and row relative to the spleen. The endocrine and bone an- dying cell toxic environment to an anti-inflammatory environ- abolic agent parathyroid hormone increased specialized ment. http://www.jimmunol.org/ With the identification of resolution-phase mediators, recent proresolving mediator levels, including (Rvs), efforts have focused on the resolution phase, an active pro- in bone marrow. Human and murine primary macro- cess orchestrated by special mediators that direct cellular and phages efferocytosed apoptotic osteoblasts in vitro, and biochemical pathways to facilitate return to homeostasis (3). RvD1 and RvD2 (10 pM–10 nM) enhanced this pro- Such specialized proresolving mediators (SPMs) of inflam- cess. These findings support a unique profile of special- mation include (LXs), resolvins (Rvs), protectins, and ized lipid mediators in bone marrow that contribute to (4). Among other actions, these mediators are potent a feedback system for resolution of inflammation and stimulators of macrophage efferocytosis, but little is known maintenance of skeletal homeostasis. The Journal about their role in bone remodeling. The purpose of this by guest on September 30, 2021 of Immunology, 2014, 193: 26–29. investigation was to determine the LM profiles of two lym- phoid organs—bone marrow and spleen—assess their mod- ulation by parathyroid hormone (PTH), and determine their primary function of bone is to house marrow, the ability to facilitate macrophage efferocytosis of apoptotic major hematopoietic organ in mammals. Bone as an organ is composed of various cell types, including osteoblasts. The implications identify a new feedback process A in bone that has potential impact in normal bone remodeling, those traditionally considered “bone” cells (osteoblasts, oste- as well as wound healing in bone. oclasts, osteocytes), hematopoietic cells (hematopoietic stem cells, progenitor cells, megakaryocytes), immune cells (lym- Materials and Methods phocytes, macrophages, dendritic cells), and stromal cells In vivo models and lipidomics (mesenchymal stem cells). Osteoblasts are “repair” cells respon- Male C57BL/6 mice (4–5 wk; The Jackson Laboratory, Bar Harbor, ME) sible for development, secretion, and extracellular matrix min- were injected once with recombinant human PTH 1-34 (50 mg/kg; Bachem, eralization. Mesenchymal stem cells are recruited to bone sur- Torrance, CA) or vehicle (saline) 2 h prior to sacrifice. Spleens and long bones faces where they differentiate into osteoblasts. Once committed were frozen in liquid nitrogen and then placed in ice-cold methanol con- to their lineage, osteoblasts are destined to embed into the taining deuterated internal standards (d8-5S-hydroxyeicosatetraenoic acid, d4-leukotriene B4,d4-PGE2,andd5-LXA4;500pgeach)andhomogenizedus- mineralized matrix as osteocytes, become flattened lining cells, ing a PTFE dounce (Kimble Chase). Proteins were precipitated (4˚C), solid- or apoptose. Studies have evaluated the trajectory toward these phase extracted using Biotage RapidTrace+ (5), and analyzed using liquid

*School of Dentistry, University of Michigan, Ann Arbor, MI 48109; †Department of The online version of this article contains supplemental material. Pathology, Medical School, University of Michigan, Ann Arbor, MI 48109; and ‡Center Abbreviations used in this article: AA, ; DHA, ; for Experimental Therapeutics and Reperfusion Injury, Department of Anesthesiology, EPA, ; hFOB, human fetal osteoblast; LC-MS-MS, liquid chro- Perioperative and Pain Medicine, Brigham and Women’s Hospital, Harvard Institutes of matography coupled with tandem mass spectrometry; LM, lipid mediator; LX, ; Medicine, Boston, MA 02115 PTH, parathyroid hormone; Rv, resolving; SPM, specialized proresolving mediator. Received for publication August 26, 2013. Accepted for publication May 6, 2014. Ó This work was supported by National Institutes of Health Grants R01DK53904 (to L.K.M.) Copyright 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 and P01GM095467 (to C.N.S.). Address correspondence and reprint requests to Dr. Laurie K. McCauley, University of Michigan, 1011 North University Avenue, Ann Arbor, MI 48109-1078. E-mail address: [email protected]

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1301945 The Journal of Immunology 27 Downloaded from http://www.jimmunol.org/ by guest on September 30, 2021

FIGURE 1. Proresolving LM levels in hematopoietic tissues. LM levels were assessed following solid-phase extraction by LC-MS-MS–based LM metab- ololipidomoics. (A) Three-dimensional score plot of bone marrow (green) and spleen (purple) endogenous LMs. (B) Three-dimensional loading plot displaying the mediators that correlate with bone marrow (green) and spleen (blue) (n = 5 mice/group). (C and D) Mice were administered vehicle or PTH (50 mg/kg), and bone marrow LM levels were assessed 2 h later. (C) Three-dimensional score plot of mouse bone marrow LM profiles for vehicle-treated (purple) and PTH-treated (green) mice. (D) Three-dimensional loading plot displaying the mediators that correlate with vehicle-treated (purple) and PTH-treated (green) mice (n = 8–9 mice/group). Ellipses in (A) and (C) denote 95% confidence region. chromatography-UV-tandem mass spectrometry, QTrap 5500 (AB SCIEX, reaction monitoring with signature ion fragments for each molecule was used, Framingham, MA) equipped with an Agilent HP1100 binary pump (Santa with six diagnostic ions used for identification, and quantification was ach- Clara, CA). An Agilent Eclipse Plus C18 column (100 mm 3 4.6 mm 3 1.8 ieved using calibration curves (5). Principal component analysis was per- mm) maintained at 50˚C was used with a gradient of methanol/water/acetic formed using SIMCA 13.0.3 software (Umetrics, Umea, Sweden) following acid of 55:45:0.01 (v/v/v) to 100:0:0.01 at 0.4 ml/min flow rate. Multiple mean centering and unit variance scaling of LMs.

FIGURE 2. Upregulation of bone marrow proresolving LMs by PTH. Mice were administered vehicle (CT) or PTH (50 mg/kg); femurs and tibias were collected 2 h later, and LM levels were assessed following solid-phase extraction by LC-MS-MS–based LM metabololipidomoics. Bioactive LM families significantly increased by PTH. For complete LM list, see Supplemental Fig. 1. Results are shown as mean 6 SEM (n = 8–9 mice/group). *p , 0.05, **p , 0.01 versus CT mice. 28 CUTTING EDGE: PTH AND BONE MARROW RESOLVINS

FIGURE 3. Murine macrophage efferocytosis. (A) Four microscopic still images from video with macrophages (Green BIODIPY label) phagocytosing apo- ptotic osteoblasts (white arrow; CellTracker Orange label) taken over 24 h. Original magnification 363. (B) Images from coincubation of apoptotic osteoblasts (Red CMTPX label with blue DAPI nuclear stain) and bone marrow macrophages (Green BIODIPY label) at various states of macrophage ingestion of apoptotic osteoblasts from left (early) to right (late). Original magnification 3315. Supplemental materials include video for (A) (Supplemental Video 1) and three-di- mensional video format for (B) (Supplemental Videos 2–5).

Human mononuclear cell isolation, culture, and efferocytosis assays proresolving mediator levels in bone marrow from unchallenged PBMCs were isolated from human venous blood by density gradient cen- mice compared with spleens. RvD2, RvE1, and LXB4, among trifugation (6). Blood was obtained from healthy human volunteers under others, were significantly elevated in bone marrow. Principal a Partners Human Research Committee–approved protocol. Cells were cul- component analysis demonstrated that LM profiles obtained Downloaded from tured in RPMI 1640 with 10% FBS, Pen/Strep, and GM-CSF (10 ng/ml; with mouse bone marrow clustered separately from those ob- R&D Systems) 7 d prior to efferocytosis assays. Macrophages differentiated from PBMCs were plated overnight (96-well plates; 50,000/well) and then tained with mouse spleens (Fig. 1A) (e.g., with incubated with 1 pM–100 nM of RvD1 or RvD2 15 min prior to the addi- with spleen) and RvD2 with bone marrow (Fig. 1B). These tion of apoptotic human osteoblasts (5:1 osteoblasts/macrophages). Human results expand the recent identification of LM biosynthesis in fetal osteoblasts (hFOBs; American Type Culture Collection, Manassas, VA), cultured in DMEM F12 media (10% FBS 1% Pen/Strep, glutamine; Invi- murine spleen (7) and bone marrow (8) with the identification m trogen, Life Technologies, Grand Island, NY), were labeled with CFSE (3 M; of novel SPMs that, with the present results, include RvD2. http://www.jimmunol.org/ Invitrogen) 15 min prior to the induction of via UV radiation. The spleen and bone marrow serve as stem and leukocyte cell re- Cells were exposed to UV light for 30 min, returned to 37˚C for 2 h, har- servoirs; an important difference is that the marrow is enclosed vested, and enumerated via trypan blue dye exclusion. More than 90% of cells were positive for trypan blue, reflecting cell death. hFOBs were resuspended by a skeletal boundary and, hence, hematopoietic and leukocytic in PBS and incubated with macrophage cultures for 4 h, followed by incu- cells are juxtaposed with osteoblasts, osteoclasts, and a mineralized bation with trypan blue to quench extracellular fluorescence and exclude cell- extracellular matrix. That PGs were increased in bone marrow associated nonphagocytosed cells (i.e., not internalized), prior to rinsing and scanning at 487–517 nM on a microplate reader (SpectraMax M3; Molecular versus spleen was not surprising because years of work have shown Devices, Sunnyvale, CA). Results are expressed as the fold increase in CFSE the diverse actions of PGs in bone (9). Unique to this study were above control, as a reflection of cell engulfment. findings that members of the DHA bioactive metabolome by guest on September 30, 2021 Murine macrophage efferocytosis (i.e., the Rvs) were higher in marrow. This suggests that bone is poised to facilitate resolution of inflammation. Indeed, injurious Marrow was flushed from 4-wk-old C57BL/6 mice, expanded with 30 ng/ml M-CSF (eBioscience) in a-MEM media (10% FBS Pen/Strep, glutamine) for and infectious insults to the marrow are more deleterious than 5–7 d to enrich for macrophages, plated at 10,000 cells/well in eight-well to the spleen; hence, protective mechanisms for resolving in- chambers, attached overnight, stained for 15 min with 2 mM Green BODIPY flammation in a timely manner would be essential for survival. in PBS, returned to complete medium, and incubated for 1 h to finalize dye cross-linking. MC3T3-E1 cells (previously stained as above with Red In the current study, PTH administration led to a rapid (2-h) CMTPX; Invitrogen) were induced to undergo apoptosis for 48 h with 50 and selective increase in proresolving LM levels in marrow, mM etoposide (Sigma, St. Louis, MO), and 40,000 cells/well were coincu- including RvD1 and RvD2, giving characteristic LM profiles, bated with macrophages for 3 h. Cells were washed, fixed with ice-cold methanol, washed, and overlaid with VECTASHIELD (Vector Laborato- ries, Burlingame, CA) to visualize DAPI nuclei. Z-stack images were per- formed on a Leica Inverted SP5X Confocal Microscope System using Leica Application Suite software (Leica, Wetzlar, Germany) and compiled into a three-dimensional video format using Imaris (Bitplane Scientific Software, South Windsor, CT). Videography was performed using the DeltaVision RT Live Cell Imaging System with SoftWoRx 3.5.1 software (Applied Precision, Issaquah, WA). Results and Discussion LM profiles from spleen and bone were obtained and com- pared using liquid chromatography coupled with tandem mass spectrometry (LC-MS-MS)–based LM metabololipido- mics. Specific mediators from the arachidonic acid (AA), docosahexaenoic acid (DHA), and eicosapentaenoic acid (EPA) bioactive metabolomes were identified, including RvD1 and RvD2 from DHA, RvE1 and RvE3 from EPA, and LXA4 FIGURE 4. Rvs increase macrophage efferocytosis. Macrophage-enriched and LXB from the AA bioactive metabolomes (Fig. 1). PBMCs were incubated with different concentrations of RvD1 (s) or RvD2 4 n These mediators were identified with matching retention times, ( ) 15 min prior to the addition of CFSE-labeled apoptotic hFOBs (4 h), rinsed, and scanned at excitation 493 nm, emission 525 nm. Results are tandem mass spectrometry fragmentation patterns, and least expressed as the CFSE levels above control, as a reflection of cell engulfment, six characteristic and diagnostic ions, as illustrated for RvD1 for RvD1 or RvD2. Experiments were performed in quadruplicate for each donor, and RvD2 (Supplemental Fig. 1). Quantification of individ- and dose and data are shown as mean 6 SEM from six or seven donors ual LMs demonstrated elevated AA-, EPA-, and DHA-derived performed at different times. *RvD1 or **RvD2 p , 0.05 versus 1-pM dose. The Journal of Immunology 29 as demonstrated by principal component analysis (Figs. 1C, these results suggest a new feedback system for the resolution 1D, 2, Supplemental Fig. 1), but no significant difference in of inflammation and restoration of skeletal homeostasis. the spleen (data not shown). This regulation included D-series and E-series Rvs. The D-series Rvs facilitate macrophage Acknowledgments efferocytosis of apoptotic neutrophils and reduce the resolu- We thank M. Shinohara, M. Kibi, and H. Arnardottir for cell procurement. tion interval postinflammatory challenge (5). Of note, statis- tically significant differences were not observed for LM levels Disclosures 5 d after PTH administration (data not shown). C.N.S. is an inventor on patents (Rvs) assigned to Brigham and Women’s One of the prime actions of SPMs is to facilitate macrophage Hospital and licensed to Resolvyx Pharmaceuticals. C.N.S. is a scientific founder efferocytosis (1, 3). Efferocytosis studies most commonly in- of Resolvyx Pharmaceuticals and owns equity in the company. C.N.S.’s interests volve macrophage clearance of apoptotic neutrophils and re- were reviewed and are managed by the Brigham and Women’s Hospital and sult in macrophage production of anti-inflammatory factors, Partners HealthCare in accordance with their conflict of interest policies. The such as TGFb (10). If dying cells are not cleared, their in- other authors have no financial conflicts of interest. tracellular contents are expelled, creating an unfavorable en- vironment (11). In the bone marrow, the fate of osteoblasts References that undergo apoptosis is unclear. Experiments performed in 1. Korns, D., S. C. Frasch, R. Fernandez-Boyanapalli, P. M. Henson, and D. L. Bratton. 2011. Modulation of macrophage efferocytosis in inflammation. which labeled osteoblasts were incubated with macrophages Front. Immunol. 2: 57. identified that macrophages rapidly efferocytose apoptotic 2. Cho, S. W., F. N. Soki, A. J. Koh, M. R. Eber, P. Entezami, S. I. Park, N. van Downloaded from Rooijen, and L. K. McCauley. 2014. Osteal macrophages support physiologic osteoblasts (Fig. 3, Supplemental Videos 1–5). skeletal remodeling and anabolic actions of parathyroid hormone in bone. Proc. Studies to determine whether increased bone marrow SPMs Natl. Acad. Sci. USA 111: 1545–1550. 3. Serhan, C. N., and J. Savill. 2005. Resolution of inflammation: the beginning facilitate macrophage efferocytosis revealed that RvD1 and programs the end. Nat. Immunol. 6: 1191–1197. RvD2 increased uptake of apoptotic osteoblasts (Fig. 4). Of 4. Serhan, C. N., and N. Chiang. 2013. Resolution phase lipid mediators of inflam- note, RvD1 (100 pM–10 nM) and RvD2 (100 pM–100 nM) mation: agonists of resolution. Curr. Opin. Pharmacol. 13: 632–640. 5. Dalli, J., and C. N. Serhan. 2012. Specific lipid mediator signatures of human increased efferocytosis of osteoblasts (30–60%) to a similar phagocytes: microparticles stimulate macrophage efferocytosis and pro-resolving http://www.jimmunol.org/ extent as efferocytosis of apoptotic polymorphonuclear neu- mediators. Blood 120: e60–e72. ∼ 6. Krishnamoorthy, S., A. Recchiuti, N. Chiang, S. Yacoubian, C. H. Lee, R. Yang, trophils ( 30%) (12). These experiments demonstrate the abil- N. A. Petasis, and C. N. Serhan. 2010. Resolvin D1 binds human phagocytes with ity of macrophages at physiologically relevant concentrations, evidence for proresolving receptors. Proc. Natl. Acad. Sci. USA 107: 1660–1665. 7. Ramon, S., F. Gao, C. N. Serhan, and R. P. Phipps. 2012. Specialized proresolving based on our LC-MS-MS quantification, to engulf apoptotic mediators enhance human B cell differentiation to antibody-secreting cells. J. osteoblasts, a process facilitated by Rvs produced in the local Immunol. 189: 1036–1042. environment. Other studies identified that RvD1 reduced 8. Poulsen, R. C., K. H. Gotlinger, C. N. Serhan, and M. C. Kruger. 2008. Identi- fication of inflammatory and proresolving lipid mediators in bone marrow and their macrophage-derived TNF-a, promoting the resolution of in- lipidomic profiles with ovariectomy and omega-3 intake. Am. J. Hematol. 83: 437– 445.

flammation, and that the process of efferocytosis involves by guest on September 30, 2021 high 9. Blackwell, K. A., L. G. Raisz, and C. C. Pilbeam. 2010. Prostaglandins in bone: bad subsets of macrophages reprogrammed from CD11b to cop, good cop? Trends Endocrinol. Metab. 21: 294–301. CD11blow (13, 14). 10. Xiao, Y. Q., C. G. Freire-de-Lima, W. P. Schiemann, D. L. Bratton, R. W. Vandivier, and P. M. Henson. 2008. Transcriptional and translational reg- The process of bone remodeling has been likened to in- ulation of TGF-beta production in response to apoptotic cells. J. Immunol. 181: flammation, where the inflammatory stage of “injury” corre- 3575–3585. 11. Nagata, S., R. Hanayama, and K. Kawane. 2010. Autoimmunity and the clearance lates with pressure and microfracture. The “reaction” stage of dead cells. Cell 140: 619–630. correlates with osteoclastic bone resorption, and the “repair” 12. Dalli, J., J. W. Winkler, R. A. Colas, H. Arnardottir, C. Y. Cheng, N. Chiang, stage correlates with osteoblastic bone formation (15). PTH is N. A. Petasis, and C. N. Serhan. 2013. Resolvin D3 and aspirin-triggered resolvin D3 are potent immunoresolvents. Chem. Biol. 20: 188–201. a potent inducer of bone remodeling and is well known to 13. Schif-Zuck, S., N. Gross, S. Assi, R. Rostoker, C. N. Serhan, and A. Ariel. 2011. increase bone formation through mechanisms that are still Saturated-efferocytosis generates pro-resolving CD11b low macrophages: modula- tion by resolvins and glucocorticoids. Eur. J. Immunol. 41: 366–379. not entirely clear. PTH drives robust increases in bone accrual 14. Lee, H. N., J. K. Kundu, Y. N. Cha, and Y. J. Surh. 2013. Resolvin D1 stimulates during wound healing (16, 17); hence, our interest was in efferocytosis through p50/p50-mediated suppression of tumor necrosis factor-a expression. J. Cell Sci. 126: 4037–4047. determining whether PTH was associated with LMs respon- 15. Martin, T. J., and G. A. Rodan. 2008. Intercellular communication during bone siblefortheresolutionofinflammation.PTHstimulatesIL-6 remodeling. In Osteoporosis, 3rd Ed., R. Marcus, D. Feldman, D. A. Nelson, and and CCL2 production from osteoblasts, which recruits and C. J. Rosen, eds. Academic Press, San Diego, CA, p. 547–560. 16. Aspenberg, P., H. K. Genant, T. Johansson, A. J. Nino, K. See, K. Krohn, activates myeloid cells (18, 19). During wound healing, mac- P. A. Garcı´a-Herna´ndez, C. P. Recknor, T. A. Einhorn, G. P. Dalsky, et al. 2010. rophages phagocytose debris and apoptotic cells. PTH has Teriparatide for acceleration of fracture repair in humans: a prospective, random- ized, double-blind study of 102 postmenopausal women with distal radial fractures. robust anabolic actions during traumatic injuries associated J. Bone Miner. Res. 25: 404–414. with increased apoptosis, such as bone marrow ablation (20). 17. Bashutski, J. D., R. M. Eber, J. S. Kinney, E. Benavides, S. Maitra, T. M. Braun, W. V. Giannobile, and L. K. McCauley. 2010. Teriparatide and osseous regener- Hence, PTH could act in the marrow to recruit macrophages ation in the oral cavity. N. Engl. J. Med. 363: 2396–2405. and promote the release of SPMs that facilitate efferocytosis. 18. Cho, S. W., F. Q. Pirih, A. J. Koh, M. Michalski, M. R. Eber, K. Ritchie, B. Sinder, S. Oh, S. A. Al-Dujaili, J. Lee, et al. 2013. The soluble interleukin-6 receptor is Such a proresolving circuit would be enhanced when PTH a mediator of hematopoietic and skeletal actions of parathyroid hormone. J. Biol. is administered during osseous wound healing where it ex- Chem. 288: 6814–6825. pedites skeletal homeostasis in favor of anabolism. In summary, 19. Li, X., L. Qin, M. Bergenstock, L. M. Bevelock, D. V. Novack, and N. C. Partridge. 2007. Parathyroid hormone stimulates osteoblastic expression of MCP-1 to recruit thebonemarrowhasauniqueprofileofspecializedLMs, and increase the fusion of pre/osteoclasts. J. Biol. Chem. 282: 33098–33106. and PTH increases select SPMs. Bone marrow macrophages 20. Zhang, Q., J. Carlson, H. Z. Ke, J. Li, M. Kim, K. Murphy, N. Mehta, J. Gilligan, and A. Vignery. 2010. Dramatic increase in cortical thickness induced by femoral efferocytose apoptotic osteoblasts, which is facilitated by marrow ablation followed by a 3-month treatment with PTH in rats. J. Bone Miner. RvD1 and RvD2. Taken together with reports that macro- Res. 25: 1350–1359. b b 21. Tang, Y., X. Wu, W. Lei, L. Pang, C. Wan, Z. Shi, L. Zhao, T. R. Nagy, X. Peng, phages produce TGF during efferocytosis and that TGF J. Hu, et al. 2009. TGF-beta1-induced migration of bone mesenchymal stem cells recruits mesenchymal stem cells to renew osteoblasts (21), couples bone resorption with formation. Nat. Med. 15: 757–765.