The Repair of Skeletal Muscle Requires Iron Recycling Through Macrophage Ferroportin Gianfranca Corna, Imma Caserta, Antonella Monno, Pietro Apostoli, Angelo A

The Repair of Skeletal Muscle Requires Iron Recycling through Macrophage Ferroportin Gianfranca Corna, Imma Caserta, Antonella Monno, Pietro Apostoli, Angelo A. Manfredi, Clara Camaschella and This information is current as Patrizia Rovere-Querini of September 25, 2021. J Immunol 2016; 197:1914-1925; Prepublished online 27 July 2016; doi: 10.4049/jimmunol.1501417

http://www.jimmunol.org/content/197/5/1914 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2016/07/26/jimmunol.150141 Material 7.DCSupplemental http://www.jimmunol.org/ References This article cites 53 articles, 16 of which you can access for free at: http://www.jimmunol.org/content/197/5/1914.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

by guest on September 25, 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 © 2016 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

The Repair of Skeletal Muscle Requires Iron Recycling through Macrophage Ferroportin

Gianfranca Corna,* Imma Caserta,*,† Antonella Monno,* Pietro Apostoli,‡ Angelo A. Manfredi,*,† Clara Camaschella,†,x and Patrizia Rovere-Querini*,†

Macrophages recruited at the site of sterile muscle damage play an essential role in the regeneration of the tissue. In this article, we report that the selective disruption of macrophage ferroportin (Fpn) results in iron accumulation within muscle-inﬁltrating macrophages and jeopardizes muscle healing, prompting fat accumulation. Macrophages isolated from the tissue at early time points after injury express ferritin H, CD163, and hemeoxygenase-1, indicating that they can uptake heme and store iron. At later time points they upregulate Fpn expression, thus acquiring the ability to release the metal. Transferrin-mediated iron uptake by regenerating myoﬁbers occurs independently of systemic iron homeostasis. The inhibition of macrophage iron export via the silencing of

Fpn results in regenerating muscles with smaller myoﬁbers and fat accumulation. These results highlight the existence of a local Downloaded from pathway of iron recycling that plays a nonredundant role in the myogenic differentiation of muscle precursors, limiting the adipose degeneration of the tissue. The Journal of Immunology, 2016, 197: 1914–1925.

esident macrophages are rare in healthy skeletal muscles. formation and assembly (1, 2). Accordingly, in mice carrying a Monocyte-derived macrophages are consistently detect- mutated Cebpb promoter, which selectively fail to express genes R able in the injured muscle (1) and, as demonstrated by associated with alternative activation, the clearance of remnants http://www.jimmunol.org/ depletion experiments using pharmacological and genetic tools is conserved, whereas myofiber regeneration is severely compro- (1–5), are required for muscle regeneration. mised (4). In agreement, IL-10, which is involved in macrophage- After damage, macrophages produce high levels of inflammatory alternative activation, has been recognized to have a role in muscle cytokines, such as TNF-a and IL-1b, and remove myofiber rem- regeneration (2, 6). Of importance, the pattern of macrophage ac- nants and apoptotic cells while sustaining the activation and the tivation controls the expression of iron-related genes and macro- proliferation of myogenic precursors (1). At later times during re- phage ability to store or release iron (7–9). Alternatively activated generation, alternatively activated macrophages predominate. They macrophages have a large intracellular labile iron pool but have secrete IL-10 and TGF-b, and favor myoblast fusion and neofiber limited storage ability, express the iron exporter ferroportin (Fpn), by guest on September 25, 2021 and spontaneously release the metal (7–9). Iron plays essential roles in many biological processes that *Division of Immunology, Transplantation and Infectious Diseases, San Raffaele reutilize the available molecules within a highly conservative Scientific Institute, 20132 Milan, Italy; †Vita-Salute San Raffaele University, 20132 Milan, Italy; ‡Department of Experimental and Applied Medicine, Section of Occu- metabolism (10). Splenic red pulp, bone marrow, and liver pational Health and Industrial Hygiene, University of Brescia, 25123 Brescia, Italy; macrophages phagocytose damaged and senescent RBCs, process and xDivision of Genetics and Cell Biology, San Raffaele Scientific Institute, 20132 Milan, Italy them to degrade hemoglobin-associated heme, and recycle the heme-associated iron (11–14) supplying the iron required for most ORCID: 0000-0003-1427-0143 (C.C.). of the daily needs (10, 13, 15, 16). Spleen and bone marrow Received for publication July 6, 2015. Accepted for publication June 26, 2016. macrophages accumulate intracellular heme and iron in hemolytic This work was supported by the Association Française contre les Myopathies (Grant 15440 to P.R.-Q.), the Italian Ministry of Health (Fondo per gli Investimenti della conditions, including autoimmune anemia and sickle cell disease Ricerca di Base-IDEAS to P.R.-Q.; Grant Ricerca Finalizzata 2011 to P.R.-Q. and (17). Of interest, heme recognition induces the expression of Fpn (18, A.A.M.), and the Italian Ministry of University and Research (Grant Progetto di Ricerca di Rilevanza Nazionale 2010 to A.A.M.). 19), thus delivering the metal to plasma, where it becomes available for tissue uptake and ensuing synthesis of novel heme groups G.C. designed experiments, performed experiments, analyzed the data, and wrote the manuscript; I.C. performed experiments; A.M. performed immunohistochemistry and and biogenesis of iron sulfur clusters in mitochondria (20, 21). immunofluorescence; P.A. was responsible for macrophage iron content measure- Iron homeostasis during skeletal muscle regeneration has not ment; C.C. discussed the results and wrote the manuscript; A.A.M. discussed the results and wrote the manuscript; and P.R.-Q. designed experiments and wrote the been extensively characterized. The tissue contains from 10 to 15% manuscript. of body iron, prevalently in the form of heme-iron bound myo- Address correspondence and reprint requests to Dr. Gianfranca Corna and Dr. Patrizia globin. Iron released during muscle fiber damage must be promptly Rovere-Querini, Ospedale San Raffaele, DIBIT-3A1, via Olgettina 58, 20132 Milano, removed to limit its oxidative effect; however, iron must become Italy. E-mail addresses: [email protected] (G.C.) and [email protected] (P.R.-Q.) available during skeletal muscle repair to allow the synthesis of The online version of this article contains supplemental material. myoglobin and prosthetic groups of the catalytic site of many Abbreviations used in this article: BMP, bone morphogenetic protein; CSA, cross- enzymes instrumental for fiber survival and function (22). Because sectional area; CTX, cardiotoxin; DFO, Desferal; DT, diphtheria toxin; FAP, fibroa- of their natural ability to uptake and release iron, macrophages are dipogenic precursor; Fpn, ferroportin; FtH, ferritin H; HIF1a, hypoxia inducible factor 1a; HO-1, hemeoxygenase-1; IF, immunofluorescence; IHC, immunohisto- strong candidates to accomplish these homeostatic functions: re- chemistry; KD, knock down; PB, peripheral blood; Q, quadriceps femoris; rm, moval of potentially dangerous free iron and transfer to regener- recombinant murine; Scr, scramble; shRNA, short hairpin RNA; TA, tibialis anterior; ating myofibers. Tf, transferrin; TfR1, Tf receptor I. In this study, we show that muscle-infiltrating macrophages after Copyright Ó 2016 by The American Association of Immunologists, Inc. 0022-1767/16/$30.00 acute sterile muscle damage undergo a time-dependent functional www.jimmunol.org/cgi/doi/10.4049/jimmunol.1501417 The Journal of Immunology 1915 modification of their ability to handle iron. By depletion of en- differentiation medium (IMDM supplemented with 2% horse serum, dogenous macrophages and adoptive transfer of cells characterized 100 U/ml penicillin, and 100 mg/ml streptomycin) to allow differentiation. by the selective inactivation of Fpn via short hairpin RNA (shRNA) In selected experiments differentiation was carried out in the presence of DFO. After 24 h of differentiation, cells were fixed with 4% PFA and interference, we show that Fpn expression by macrophages at the immunostained using an anti-sarcomeric myosin mAb (MF20; Develop- site of injury represents a requirement for appropriate activation of mental Studies Hybridoma Bank). Nuclei were stained with Hoechst 33342 myogenic precursors and eventual healing of the injured skeletal (Invitrogen). Samples were examined with a Nikon Eclipse 55i microscope muscle. (Nikon), and images were captured with a Digital Sight DS-5 M digital camera (Nikon). Fusion index was determined as the number of myosin- expressing myotubes with 2, 3–4, or $5 nuclei versus the total number of Materials and Methods myosin-expressing myotubes in the area analyzed. Muscle injury Myogenic precursors proliferation assay Two-month-old C57BL/6 female mice (from Charles River Laboratories) were anesthetized and injected with cardiotoxin (CTX; 10 mM; Naja After 48 h in proliferation medium, myogenic precursors were pulsed with mossambica mossambica; Sigma-Aldrich) in tibialis anterior (TA) and 10 mM BrdU for 2 h before fixation in 4% PFA and immunostained with rat quadriceps femoris (Q) muscles (50 ml for TA and 100 ml for Q muscles). anti-BrdU Abs 1:100 (R&D Systems). Nuclei were stained with Hoechst Animals were sacrificed at different time points after injury. All procedures 33342. Samples were examined as described earlier. were performed in the animal facility of the San Raffaele Scientific In- Tissue macrophage analysis stitute (Italy) in accordance with European Union guidelines and with the approval of the Institutional Animal Care and Use Committee of our In- Mononuclear cells retrieved from TA and Q muscles and processed as stitution. described earlier were analyzed by flow cytometry. Cells were incubated

+ with an allophycocyanin-conjugated anti-CD11b mAb (clone M1/70; BD Downloaded from Tissue digestion and retrieval of CD11b leukocytes Biosciences), a PE-conjugated anti-CD45.1 mAb (clone A20; BD Phar- Minced TA and Q muscles of each single mouse were collected and pooled. mingen), an FITC-conjugated anti-CD45.2 mAb (clone 104; BD Phar- Minced tissues underwent four cycles of enzymatic digestion at 37˚C for 10 mingen), an FITC-conjugated anti-CD80 mAb (clone 16-10A1; BD min in the presence of 3.5 mg/ml Dispase (Invitrogen) and 0.5 mg/ml Biosciences), a purified anti-CD163 polyclonal Ab (Santa Cruz Biotech- collagenase type V (Sigma-Aldrich). In selected experiments, macro- nology), followed by secondary staining with FITC-conjugated anti-rabbit phages were purified from muscle tissue of mice adoptively transferred Ab (BD Biosciences), an FITC-conjugated anti-CD86 mAb (clone 16-10A; BD Biosciences), an FITC-conjugated anti-IAb mAb (clone C3H.SW; BD with Fpn/scramble (Scr) or Fpn/knock down (KD) in vitro–derived cells. In b http://www.jimmunol.org/ theseexperiments,muscleswereretrieved3,5,and7dafterinjury. Biosciences), and an FITC-conjugated anti-H2k mAb (clone AF6-88.5; CD11b+ infiltrating cells were purified by magnetic bead sorting using BD Biosciences) and then analyzed on a FACSCanto flow cytometer (BD CD11b-conjugated magnetic beads (Miltenyi Biotec) as previously de- Biosciences). Analysis was performed with the FlowJo software (Tree scribed (23). When indicated, CD11b+ cells were also purified from the Star). peripheral blood (PB) of mice at the baseline by magnetic bead sorting. Transferrin saturation Cell sorting–mediated retrieval of fibroadipogenic and Blood was collected by retro-orbital bleeding at various time points. myogenic precursors Transferrin (Tf) saturation was calculated as a ratio of serum iron and total iron binding capacity levels, assessed using commercially available assays Fibroadipogenic precursor (FAP) and myogenic cells were isolated from (Randox Laboratories).

Pax7Zsgreen mice (kindly provided by E. Clementi, Milan, Italy). TA and Q by guest on September 25, 2021 muscles of 6- to 8-wk-old mice were pooled before digestion with 0.2% of Histochemistry, immunohistochemistry, and Collagenase II (Worthington Biochemical) and Dispase as described earlier. immunofluorescence Cells were stained with anti–CD31- PE/Cy7 (clone 390; eBioscience, San Diego, CA), anti–CD45-PE/Cy7 (clone 30-F11; eBioscience), anti- Serial sections (10 mm) of isopentane-frozen TA muscles were used for PDGFRa-allophycocyanin (CD140a, clone APA5; BioLegend), anti–LY- H&E staining and immunofluorescence (IF) or immunohistochemistry 2 6A/E SCA-1 allophycocyanin/Cy7 (clone B7; BD Biosciences), and (IHC). For IF, sections were blocked with BSA 5%, Triton 0.1% in PBS 7-aminoactinomycin D (Life Technologies) and purified by cell sorting (EuroClone) before incubation with a rabbit anti-laminin Ab (Sigma- 2 (MoFlow XDP; Beckman Coulter). FAP cells were identified as CD45 / Aldrich), anti–Tf receptor I (anti-TfR1; FITC anti-mouse CD71; Bio- 2 2 2 CD31 /SCA-1+/PDGFRa+/GFP ; myogenic cells were identified as CD45 / Legend), anti-developmental myosin (NCL MHCd, clone RNMy2/9D2; 2 2 2 CD31 /SCA-1 /PDGFRa /GFP+. Novocastra), anti-myosin H chain 2A (SC71; Developmental Studies Hybridoma Bank), anti-myosin H chain 2B (BF-F3; Developmental Culture and differentiation of FAP Studies Hybridoma Bank), anti-myosin H chain I (BAD7) followed by Freshly isolated FAP cells were plated at 1 3 104 cells/cm2 in multiwell Alexa Fluor 488–conjugated goat anti-rabbit Ab, Alexa Fluor 488– plates coated with growth factor reduced BD Matrigel Matrix (BD Bio- conjugated anti-rat Ab, Alexa Fluor 594–conjugated donkey anti-mouse Ab, sciences) and cultured for 6 d in growth medium (modified high-glucose Alexa Fluor 488–conjugated anti-mouse Ab, Alexa Fluor Cy3-conjugated DMEM supplemented with 20% FBS, 100 U/ml penicillin, and 100 mg/ml anti-mouse Ab, or Alexa Fluor 405–conjugated anti-mouse Ab (Invi- streptomycin) enriched with 5 ng/ml recombinant human basic fibroblast trogen). Specimens were counterstained with Hoechst 33342 (Molecular growth factor (Peprotech) as described by Cordani et al. (24). For adipo- Probes). To evaluate the extent of macrophage infiltration at days 3 and 5 genic differentiation, FAP cells were shifted to an adipogenic medium after injury, we stained tissue sections with rat anti-mouse CD68 mAb consisting of DMEM–Ham’s F-12 Nutrient Mixture, 10% FBS supple- (clone FA-11; AbD Serotec) and rat anti-mouse CD206 mAb (clone mented with 0.5 mM of phosphodiesterase inhibitor 3-isobutyl-1- MR5D3; AbD Serotec). Primary Abs were revealed using biotin- methylxanthine, 1 mM of dexamethasone, and 1 mg/ml insulin (all from conjugated anti-rat IgG (eBioscience) and R.T.U. HRP streptavidin (Vec- Sigma-Aldrich). After 3 d cells were switched to DMEM–Ham’s F-12 tor Laboratories), detected using Vector NovaRED substrate kit (Vector Nutrient Mixture with 10% FBS and 10 mg/ml insulin, which was Laboratories). Slides were counterstained with hematoxylin. Both IF and changed every 2 d. Desferal (DFO; 5 mM; Novartis) treatment started on IHC were examined with a Nikon Eclipse 55i microscope (Nikon). Images the third day of culture and continued during culture in adipogenic me- were captured with Digital Sight DS-5 M digital camera (Nikon) or with dium, replaced every other day. Oil Red O staining was carried out fol- an AxioImager M2m (Zeiss). Parallel slides without primary Abs were lowing manufacturer’s instructions (Bio-Optica). identically processed and used as negative controls. Macrophage quantification was performed on 20 and 10 independent images, respectively, Propagation and in vitro differentiation of myogenic derived from four different mice. Fiber cross-sectional area (CSA) analysis precursors was carried out on 250–500 fibers/muscle by using ImageJ software (http:// rsbweb.nih.gov/ij/). TA muscles retrieved 15 d after CTX injection were Myogenic cells were grown in proliferation medium (F10 with 20% cut transversally and sections were stained with the rabbit anti-laminin HyClone [Euroclone], 100 U/ml penicillin, 100 mg/ml streptomycin, and mAbs (Sigma-Aldrich) and Hoechst to evaluate the number of nuclei of 50 mg/ml gentamicin, plus 5 ng/ml recombinant human basic fibroblast regenerating myofibers. Oil Red O staining was carried out following growth factor [Peprotech]) for 5 d. DFO (5 mM) treatment started on day manufacturer’s instructions (Bio-Optica). Perls’ Prussian blue staining was three and was replaced every other day. After 5 d, cells were shifted to performed on sections incubated with 2% potassium ferrocyanide and 2% 1916 ROLE OF FERROPORTIN IN MUSCLE REGENERATION hydrochloric acid at a 1:1 ratio for 30 min. Nuclei were counterstained (exon 7), REV 59-ATGACTGAGATGGCGGAAAC-39 (exon 9); CD163 with Nuclear Fast Red (Vector Laboratories). (NM_001170395) FW 59-GGGCTCCGTCTGTGATTTT-39 (exon 7), REV 59-ACTACGCTGACATCCCTGCT-39 (exon 7); Tf (NM_133977) FW 59- Quantification of the iron muscle macrophage content GGCTGTGGTAAAGAAGGGAAC-39 (exon 5), REV 59-CCCGAGA- Iron was determined on macrophages retrieved from adoptively transferred AGAAACTGGACAC-39 (exon 6); PPARg (XM-006505743) FW 59-TT- mice (see later) at various times after injury by inductively coupled plasma GCTGAACGTGAAGCCCATCGAGG-39 (exon 13), REV 59-GTCCT- mass spectrometry using a Perkin Elmer ELAN DRC II instrument (Perkin TGTAGATCTCCTGGAGCAG-39 (exon 13); Adiponectin (NM_009605.4) Elmer Sciex, Woodbridge, ON, Canada) and total quant technique with FW 59-AGGGCTCAGGATGCTACTGTT-39 (exon 3), REV 59-CCAA- 9 b 9 external calibration, as described earlier (7). Two runs were performed per GAAGACCTGCATCTCC-3 (exons 3–4); TGF- (NM_011577) FW 5 - 9 9 each sample (two replicates each), with a dynamic reaction cell. Accuracy CGGCAGTGGCTGAACCAAGGA-3 (exon 3), REV 5 -CGTTTGGG- GCTGATCCCGTTGA-39 (exon 4); Hamp1 (AF_503444) FW 59-AAG- (around 87.5%) was determined in natural water, and bovine liver was used CAGGGCAGACATTGCGAT-39 (exons 2–3), REV 59-CAGGATGTGG- as standard reference materials (respectively, NIST 1640 and MS1577b; CTCTAGGCTATGT-39 (exon 3); Il10 (NM_010548) FW 59-ATTTGAAT- National Institute of Standard and Technology, Gaithersburg, MD). Vari- TCCCTGGGTGAGAAG-39 (exon 3), REV 59-TGTCTACAAGGCCATG- ation coefficients ranged among series from 6 to 8% and between series AATGAA-39 (exon 5); hypoxia inducible factor 1a yHIF1a) (XM_011243996) from 6 to 12%. The instrument was calibrated using standard solution, FW 59-TCAAGTCAGCAACGTGGAAG-39 (exon 5), REV 59-TATCGAGGC- concentration 10 mg/l (Multielement ICP-MS Calibration Standard 3, TGTGTCGAC-39 (exon 6); glut1 (NM_011400) FW 59-GCTGTGCTTATGG- Matrix per Volume: 5% HNO3/100 ml; Perkin Elmer). Limits of detection GCTTCT-39 (exon 6), REV 59-AGAGGCCACAAGTCTGCATT-39 (exon 7); were determined based on three SDs of the background signal: a value of CAIX (NM_139305) FW 59-GGAGTCCCTTGGGTTAGAGG-39 (exon 1), 0.005 was obtained. REV 59-GATGTCTACCGGGGACTGAA-39 (exon 4); RPL13a (NM_009438) In vitro macrophage generation and analysis FW 59-TCAAGGTTGTTCGGCTGAAG-39 (exon 6), REV 59-GCCCCAGG- TAAGCAAACTTT-39 (axon 7); SOD1 (NM_011434) FW 59-GGCTTCTCGT- Bone marrow precursors from C57BL/6 female mice were isolated and CTTGCTCTCT-39 (exon 1), REV 59-GTTCACCGCTTGCCTTCT-39 Downloaded from propagated for 7 d in anti-MEM (Life Technologies, Invitrogen) containing (exon1-2); and catalase (NM_009804) FW 59-CCCCCAACTATTACCC- 10% FBS (Lonza) in the presence of recombinant murine (rm)M-CSF (100 CAAC-39 (exon 10), REV 59-GTTCTCACACAGGCGTTTCC-39 (exon 11). ng/ml; Miltenyi Biotec) to generate macrophages as described previously The amplicon size was between 100 and 300 bp. The specificity of the primers (7). Cells were cultured for 2 additional d in the presence of rmM-CSF was assessed based on the presence of a single peak in the melting curve. (10 ng/ml) or rmIFN-g (50 ng/ml; Peprotech) to generate M1 cells. To Transcripts of target genes were quantified relative to the abundance of RPL13a obtain M2 macrophages (IL-4) or IL-10 macrophages, we cultured cells by the change-in-threshold method. for 4 additional d with rmIL-4 (10 ng/ml; Miltenyi Biotec) or rmIL-10 http://www.jimmunol.org/ (10 ng/ml; R&D Systems) and rmM-CSF (10 ng/ml). Macrophage polar- Western blot analysis ization was verified by flow cytometry after staining with fluorochrome- Lysates were prepared in Tris 10 mM at pH 8.0, NaCl 150 mM, Nonidet conjugated Abs. Cells were incubated with allophycocyanin-conjugated P-40 1%, NaDodSO4 (SDS) 0.1%, EDTA 10 mM, and protease inhibitors anti-CD11b Abs (BD Biosciences), FITC-conjugated anti-HMC class II (Sigma) and centrifuged at 16,000 3 g for 5 min at 4˚C. For Western blot (BD Biosciences), FITC-conjugated anti-CD86 (BD Biosciences), or PE- analyses, equal amounts of proteins were resolved by SDS-PAGE and conjugated anti-CD206 (AbD Serotec) or anti-CD163 Abs (Santa Cruz transferred onto Immobilon-P (Millipore). After Ponceau S staining, Biotechnology) followed by FITC-conjugated anti-rabbit Abs and relevant membranes were saturated in Tris-HCl 20 mM, pH 7.6, NaCl 150 mM controls at 4˚C for 20 min (final concentration, 5 mg/ml) in PBS containing (TBS) containing nonfat milk 5%, and Tween 20 0.1%. Ags were detected 10% FBS. Labeled cells were washed and analyzed using a FACSCanto using either rabbit polyclonal anti-FtH, kindly provided by S. Levi (Milan) flow cytometer and the FlowJo software (Tree Star). Results are expressed (25), or mouse monoclonal anti-TfR1 (Invitrogen), rabbit polyclonal anti– as relative fluorescence intensity, calculated dividing the mean fluores- HO-1 (H-105; Santa Cruz Biotechnology), rabbit anti-mouse Fpn IgG by guest on September 25, 2021 cence intensity obtained in the experimental sample by the one obtained (MTP11-A; Alpha Diagnostic International), rabbit anti-mouse CD163 with the isotype-matched control Ab. (M96; Santa Cruz Biotechnology), mouse monoclonal anti-GAPDH RNA extraction and quantitative real-time PCR (GAPDH-71.1; Sigma-Aldrich), rabbit anti-mouse monoclonal anti–a- tubulin (sc-8035; Santa Cruz Biotechnology), or mouse monoclonal TA and Q were retrieved immediately after necroscopy, pooled, minced, and anti–b-actin (clone AC15; Sigma-Aldrich) Abs. Primary Abs were flash frozen in liquid nitrogen after adding TRIzol (Life Technologies). revealed with HRP-conjugated secondary Abs (Amersham Biosciences) Samples were stored at 280˚C in TRIzol (Life Technologies) until ex- and a chemiluminescence kit (ECL; Amersham Biosciences). traction (maximum time of storage, 2 wk). RNA was extracted from muscle and liver using a homogenizer, whereas RNA from cells (FAPs, Lentiviral production + myogenic precursors, or CD11b cells) was obtained by pipetting before Lentiviral stocks were produced by polyethylenimine-assisted tran- chloroform extraction. For further details, see the manufacturer’s protocol. sient transfection of 293T packaging cells with pLKO.1 shFpn Nucleic acid concentration and quality were consistently assessed by (TRCN0000319650 or TRCN00000317773; Sigma-Aldrich), together with spectrophotometer (NanoDrop; Thermo Scientific). We obtained on aver- pCMV-dR8.74 and VSV-G/pMD 2.G plasmids. pLKO.1-PGK-puro-CMV age 1 mgRNA/mgmuscleand5mg/mg tissue in the case of the liver. In the scrambled shRNA control was used as comparison. Viral supernatants case of cells, we obtained about 10 mg RNA per million cells. RNA (1 mg) were collected after 30 h, concentrated by ultracentrifugation, and stored in was used for first-strand synthesis of cDNAs with the High-Capacity cDNA PBS at 280˚C. Reverse Transcription Kit (Applied Biosystems) using random primers (4 mM final concentration in 20 ml of reaction) and MultiScribe Reverse Tran- Bone marrow lentivirus infection scriptase according to the manufacturer’s instructions. Temperature and time of retrotranscription were 25˚C for 10 min, 37˚C for 120 min, 85˚C for 5 Total bone marrow samples were obtained from 2-mo-old C57BL/6 45.1 min, and 4˚C ‘. Transcribed products were analyzed by quantitative real- female mice and plated in the presence of rmM-CSF (100 ng/ml). The next time PCR with SYBR Green Master Mix (Applied Biosystems) and the day, the lentivirus was added with 8 mg/ml polybrene. Medium was appropriate primers on a Viia7 (Life Technologies). The real-time replaced the next day, and puromycin selection (3 mg/ml) started on day 3. PCR program was: HOLD stage: step 1 50˚C for 2 min, step 2 95˚C for Cells were maintained in rmM-CSF (100 ng/ml) and under puromycin 10 min; PCR stage: step 1 95˚C for 15 s, step 2 60˚C for 1 min (40 cycles); selection for an additional 4 d. MELT CURVE stage: step 1 95˚C for 15 s, step 2 60˚C for 1 min, step 3 Macrophage depletion and reconstitution (dissociation) 95˚C for 15 s. The final concentration of the primers in the assay was 1 mM. The sequences of the primers used are: Haptoglobin Two-mo-old CD45.2+ C57BL/6 female CD11bDTR mice (from Jackson (NM_017370) FW 59-GCTGGTGGAGATTGAGAAGG-39 (exon 5), REV Lab) were injected i.v. with 10 ng/g diphtheria toxin (DT; Sigma-Aldrich) 59-TTTGGAAGGCAGGCAGATAG-39 (exon 5); ferritin H (FtH; NM_010239) (day 21). The next day (day 0), CTX was injected (see earlier) and FW 59-TGCCTCCTACGTCTATCTGTC-39 (exon 1), REV 59-GTCATCACG- CD45.1+ (5 3 106, in a volume of 100 ml of saline) lentivirus-infected mac- GTCTGGTTTCTTT-39 (exons 2–3); hemeoxygenase-1 (HO-1) (NM_010442) rophages were i.v. injected. On day 1, both DT (10 ng/g) and CD45.1+ (5 3 FW 59-GACACCTGAGGTCAAGCACAG-39 (exon 4), REV 59-CCACTGC- 106) lentivirus-infected macrophages were injected. The CD45.2 - CD45.1 CACTGTTGCCAAC-39 (exon 6); Fpn (NM_016917) FW 59-CCAGTGTCC- mismatched allowed the actual tracing of donor-derived macrophages CCAACTACCAA-39 (exon 7), REV 59-GTCACCGTCAAATCAAAGGA-39 (CD45.1+) and the monitoring of the persistence of macrophages of re- (exon 8); TfR1 (NM_011638) FW 59-TGATTGTTAGAGCAGGGGAAA-39 cipient origin (CD45.2+). The Journal of Immunology 1917

Statistics At day 3 there is a consensual and significant upregulation of the Data are expressed as means 6 SEM. The normal distribution of each CD163, the receptor for haptoglobin–hemoglobin complexes, of continuous variable was assessed by use of the Kolmogorov–Smirnov test. the HO-1, which is involved in heme detoxification, and of the Statistical analysis was performed using Student t test or one-way ANOVA iron-storage protein, FtH (Fig. 1B). FtH protein, already present in followed by Tukey multiple comparison test when appropriate. Statistical healthy muscle, is induced starting from day 3 and remains ele- analysis were performed by Prism software. vated throughout the regeneration process. mRNA encoding for Fpn, the only known iron exporter, is induced at later time points, Results whereas that of TfR1 does not significantly change over time. The muscle response to injury encompasses modulation of iron TfR1 protein levels are virtually undetectable in healthy skeletal storage/export genes muscle; they substantially increase after injury, peak at day 5, and TA and Q muscles of 2-mo-old C57BL/6 mice were injected with abate at day 10 (Fig. 1B). TfR1 expression parallels the appear- CTX, with early disruption of most fibers and infiltration by in- ance of novel fibers (Fig. 1A) and is mostly restricted to small, flammatory cells. Infiltration peaks 3 d after and progressively centrally nucleated myofibers that express developmental myosin decreases (Fig. 1A). Centrally nucleated, regenerating myofibers H chain (Fig. 2A). Tf is transiently expressed in the muscle tissue are detectable at day 5 and the tissue eventually heals 15 d after with a peak at day 3 after injury to progressively decrease there- injury (Fig. 1A). The homeostatic response to injury comprises the after (Fig. 2B). In contrast, Tf saturation in the blood and hepcidin modulated expression of genes involved in heme and iron han- expression in the liver, a sensitive sensor of the serum iron level dling, as assessed by real-time PCR and Western blot analysis (21), do not change throughout muscle injury and regeneration,

(Fig. 1B). Haptoglobin is the earliest gene to be induced and is suggesting that the systemic iron homeostasis is not perturbed Downloaded from strongly and signiﬁcantly upregulated at day 1 after muscle injury. (Fig. 2C, 2D). http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 1. The expression of heme/iron genes is modulated during acute muscle injury. (A) H&E staining (upper panels) and IF for CD68 (lower panels, green color) in addition to Hoechst staining for nuclei (lower panels, blue color) on TA muscles left untreated or collected 1, 3, 5, and 15 d after CTX injection. Original magniﬁcation 320. Scale bars, 50 mm. (B) Time-course analysis of the mRNA expression of haptoglobin, CD163, FtH, HO-1, Fpn, and TfR1 (left) and of the protein levels of FtH and TfR1 (right) in pooled Q and TA muscles collected immediately before and 1, 3, 5, 7, 10, and 15 d after CTX injection in 2-mo-old mice. mRNA levels were normalized to RPL13a1. Bars indicate the mean 6 SEM. In Western blot, a-tubulin signal was used as internal protein loading control. Six different animals obtained from two independent experiments were analyzed for each time point. **p # 0.01, ***p # 0.001, signiﬁcantly different from time 0. 1918 ROLE OF FERROPORTIN IN MUSCLE REGENERATION Downloaded from http://www.jimmunol.org/

FIGURE 2. TfR1 is expressed by regenerating myoﬁbers. (A) IF for TfR1 (green color) and developmental myosin H chain (dMHC, red color) in

addition to Hoechst staining for nuclei (blue color) on TA muscles collected 3, 5, and 7 d after CTX injection. Boxed areas in the overlay images are by guest on September 25, 2021 enlarged in the right panels. Original magniﬁcation 320. Scale bars, 50 mm. (B) Time-course analysis of the Tf mRNA expression in Q and TA muscles from 2-mo-old mice immediately before and 1, 3, 5, 7, 10, and 15 d after CTX injection. mRNA levels were normalized to RPL13a1. Bars indicate mean 6 SEM. Six different animals from two independent experiments were analyzed for each time point. ***p # 0.001, signiﬁcantly different from time 0. (C) Time-course analysis of the percentage of Tf saturation measured on blood retrieved immediately before or 1, 3, 5, 7, 10, and 15 d after CTX injection. Bars indicate the mean 6 SEM. n = 6 animals for each time point. (D) Time-course analysis of the mRNA expression of hepcidin in livers collected from 2-mo- old mice immediately before and 1, 3, 5, 7, and 10 d after CTX injection. mRNA levels were normalized to RPL13a1. Bars indicate the mean 6 SEM. Four different animals were analyzed per each time point.

Skeletal muscle injury results in infiltration by inflammatory Fiber necrosis is expected to cause the release of muscle iron leukocytes. Early after injury, mostly neutrophils are present. In mainly in a form associated with the intracellular storage protein contrast, at later times, macrophages, that is, professional phago- myoglobin, which could be taken up by inflammatory macro- cytes, that are able to uptake and recycle iron, predominate (26). We phages. In fact, intact myoglobin is detectable in infiltrating investigated the expression of molecules involved in heme seques- macrophages from day 1 after injury (Fig. 3C, left panel). Con- tration and iron storage/export in macrophages retrieved by immu- versely, the myoglobin detectable in tissue slightly decreases in nomagnetic sorting from the injured and regenerating muscle in the first days after the injury to return to baseline levels after comparison with the PB monocytes. Muscle-infiltrating CD11b+ muscle regeneration (Fig. 3C, right panel). cells do not express TfR1. In contrast, they consistently upregulate The expression of molecules involved in iron handling also has CD163, HO-1, and FtH throughout the regeneration process. been evaluated in freshly purified myoblasts propagated in conditions that allow their expansion (proliferating cells) or favor their CD163 mRNA levels are relatively stable, suggesting that the re- differentiation into myotubes (differentiated cells). Both cell ceptor upregulation, which is known to occur physiologically in populations express high levels of TfR1 and FtH, but not of CD163, regenerating skeletal muscle because of the macrophage shift to- HO-1, and Fpn (Fig. 3D). ward an M2 profile, does not primarily occur at the mRNA level. Macrophages acquire the molecular machinery associated with Silencing of Fpn limits myofiber regeneration and leads to iron storage only upon recruitment at the injured sites, because PB deposition of fat + CD11b cells fail to express CD163, HO-1, or FtH (Fig. 3A, 3B). We thus established an in vivo model to selectively interfere with These changes strengthen the ability of infiltrating macrophages to the ability of macrophages to export iron by disrupting the ex- remove toxic iron released by damaged myofibers. In contrast, pression of Fpn. We stably silenced Fpn expression by infecting macrophage expression of the iron exporter Fpn increases at later bone marrow–derived macrophages with an ad hoc lentiviral vector time points, that is, when regenerating fibers need iron to sustain (Fig. 4A) and used these cells to rescue CD11bDTR transgenic their own regeneration (Fig. 3A, 3B). mice that underwent conditional endogenous macrophage ablation The Journal of Immunology 1919 Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 3. Changes of iron-related phenotype of muscle-inﬁltrating CD11b+ cells during muscle damage and regeneration. (A) Western blot analysis of TfR1, CD163, HO-1, FtH, and Fpn in CD11b+ cells retrieved from PB of mice at the baseline or from the skeletal muscles 1, 3, 5, 7, and 10 d after CTX injection in 2-mo-old mice. b-Actin was used as internal protein loading control. Results are representative of three independent experiments. (B) Time- course analysis of the mRNA expression of TfR1, CD163, HO-1, FtH, and Fpn CD11b+ cells retrieved from PB or from the skeletal muscles 1, 3, 5, 7, and 10 d after CTX injection. mRNA levels were normalized to RPL13a1. Bars indicate the mean 6 SEM. *p # 0.05, **p # 0.01, ***p # 0.001, signiﬁcantly different from PB. Results are from three independent experiments. (C, left panel) Western blot analysis of myoglobin in CD11b+ cells retrieved from the skeletal muscles of 2-mo-old mice 1, 3, 5, 7, and 10 d after CTX injection. Total muscle lysate (M) served as positive control. b-Actin was used as internal protein loading control. (C, right panel) Western blot analysis of myoglobin in pooled Q and TA muscles of 2-mo-old mice collected immediately before and 1, 3, 5, 7, 10, and 15 d after CTX injection. a-Tubulin was used as internal protein loading control. (D) Analysis of the mRNA expression of TfR1, CD163, HO-1, FtH, and Fpn in myoblasts retrieved from healthy skeletal muscles, either proliferating (prolif) or differentiated (differ) in vitro to form myotubes. mRNA levels were normalized to RPL13a1. Bars indicate the mean 6 SEM. Results are from three independent experiments. upon DT injection (Fig. 4B). Fpn silencing does not per se interfere macrophages, Fpn expression remains negligible at all analyzed time with macrophage ability to respond to cytokines (Supplemental Fig. 1) points (Fig. 4C). To trace silenced macrophages, we exploited the or with macrophage expression of other iron proteins, including TfR1 expression of the CD45.1 isoform of the pan-leukocyte marker CD45 and FtH (Supplemental Fig. 2A). After in vivo transfer of silenced by transduced macrophages injected into CD45.2+ DT-ablated mice 1920 ROLE OF FERROPORTIN IN MUSCLE REGENERATION Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 4. Lentiviral infection efficiently reduces the expression of Fpn. (A) Schematic representation of the protocol used to stably knock down the expression of Fpn. Total bone marrows were obtained from 2-mo-old C57BL/6 45.1 female mice and plated in the presence of rmM-CSF. The next day the cells were infected with a lentivirus containing either the control shRNA or the shRNA against Fpn. Puromycin selection started on day 3. Cells were maintained in rmM-CSF and under selection for an additional 4 d to obtain macrophages in which Fpn was silenced (Fpn/KD) and control macrophages (Fpn/Scr). (B) Schematic representation of the strategy used for in vivo transfer of transduced macrophages. To deplete endogenous macrophages, we injected CD45.2 CD11bDTR mice i.v. with DT 1 d before and 1 d after CTX damage. Fpn/KD and Fpn/Scr macrophages were delivered i.v. the same day as CTX and the day after. Adoptively transferred animals were sacrificed 3, 5, 7, 10, and 15 d after CTX injection. (C) Time-course analysis of the Fpn mRNA expression in Q and TA muscles from mice adoptively transferred with Fpn/Scr (empty bars) or Fpn/KD (filled bars) macrophages collected 3, 5, 7, and 10 d after CTX injury. mRNA levels were normalized to RPL13a1. Bars indicate the mean 6 SEM. Six different animals from two independent experiments were analyzed for each time point. (D) Quantification of CD68+ (upper panel) and CD206+ cells (lower panel) detected by IHC on TA muscles of mice adoptively transferred with Fpn/Scr or Fpn/KD macrophages collected 3 and 5 d after CTX injection. (E) Perls’ staining on TA muscles from mice adoptively transferred with Fpn/Scr or Fpn/KD macrophages collected 7 and 10 d after CTX injection. Original magnification 360. Scale bars, 20 mm. (F) Quantification of the iron content of CD11b+ cells retrieved after digestion of TA and Q muscle of mice adoptively transferred with Fpn/Scr or Fpn/KD macrophages. Bars indicate the mean 6 SEM. Four different animals were analyzed for each time point. **p # 0.01, ***p # 0.001, significantly different from mice adoptively transferred with Fpn/Scr macrophages. The Journal of Immunology 1921

FIGURE 5. Fpn deficiency in macrophages af- fects muscle regeneration. (A) IF staining for laminin (green color) in addition to Hoechst staining for nuclei (blue color) on TA sections of mice adoptively transferred with Fpn/Scr and Fpn/KD macrophages, collected 15 d after CTX injury. Original magnification 310. Scale bars, 200 mm. (B) Frequency histogram showing the distribution of myofiber CSA in mice adoptively transferred with Fpn/Scr (empty bars) and Fpn/KD (filled bars) C

macrophages collected 15 d after CTX injury. ( , Downloaded from left panel) Median CSA of myofibers of mice adoptively transferred with Fpn/Scr (empty bars) and Fpn/KD (filled bars) macrophages collected 15 d after CTX injury. Data from n = 9 mice/group. Bars indicate the mean 6 SEM. (C, right panel) Number of nuclei per myofiber of mice adoptively

transferred with Fpn/Scr (empty bars) and Fpn/KD http://www.jimmunol.org/ (filled bars) macrophages, 15 d after CTX injury. ***p # 0.001, **p # 0.01, significantly different from mice adoptively transferred with Fpn/Scr macrophages. (D) Western blot analysis of TfR1 in pooled Q and TA muscles of mice adoptively transferred with Fpn/Scr (left) or Fpn/KD (right) macrophages and collected 3, 5, 7, 10, and 15 d after CTX injection in 2-mo-old mice. GAPDH signal was used as internal protein loading control. (E)IFfor BrdU (red color) on myogenic precursor cells pro- by guest on September 25, 2021 liferating in culture treated or not with DFO. Nu- clei were counterstained with Hoechst (blue color). Original magnification 320. The percentage of BrdU+ nuclei of myogenic precursor cells treated or not with DFO is depicted in the right panel. *p # 0.05, significantly different from untreated myotubes. (F) IF for myosin H chain (green) on differentiated myotubes treated or not with DFO. Nuclei were counterstained with Hoechst (blue). Original magnification 320. The fusion index is depicted in the right panel. **p # 0.01, significantly different from untreated myotubes.

(Supplemental Fig. 2B). At day 3 after injury, .90% of muscle- damage, Fpn/Scr and Fpn/KD macrophages express comparable infiltrating leukocytes express the CD45.1 Ag, and as such are of levels of CD163. Upon Fpn silencing there is a substantial accu- donor origin (Supplemental Fig. 2B). mulation of iron within infiltrating macrophages of Fpn/KD, but Recruitment of CD68+ and CD206+ macrophages in the muscles not of Fpn/Scr adoptively transferred mice, confirming that Fpn is at days 3 and 5 after injury does not differ between Fpn/Scr and required for the ability of muscle macrophages to export the metal Fpn/KD transferred mice (Fig. 4D). Macrophages from Fpn/Scr (Fig. 4E, 4F). Iron accumulation within macrophages does not and Fpn/KD transferred mice express comparable levels of the apparently influence the extent of the oxidative stress, as detect- membrane molecules, CD80, CD86, H2Kb, and IAb, and of the able by the induction of key enzymes in the antioxidant response, cytokine IL-10 (Supplemental Fig. 3A, 3B). At day 3 and 5 after SOD1 and catalase (Supplemental Fig. 3B). 1922 ROLE OF FERROPORTIN IN MUSCLE REGENERATION

target genes, glut1 and caIX (Supplemental Fig. 4B). DFO-treated myoblasts proliferate significantly less than control myoblasts, as assessed by the incorporation of BrdU (Fig. 5E), and once induced to differentiate, yield small myotubes that rarely display more than two nuclei/cell (Fig. 5F). In contrast, myotube formation is unaf- fected by the addition of DFO immediately after the induction of differentiation (Supplemental Fig. 4A), indicating that iron is dis- pensable for cell fusion. These results suggest that the iron availability primarily limits the proliferation of myoblasts. The size of type IIb myofibers that express the myosin H chain 2B and that are characterized by glycolytic metabolism is significantly smaller in the absence of macrophage Fpn expression. In contrast, the size of myosin H chain expressing type IIa myofibers, with a mixed oxidative/glycolytic metabolism, does not differ between mice in which Fpn-silenced or control macrophages had been adoptively transferred (Fig. 6A, 6B). The relative ratio of type I, IIa, and IIb myofibers does not differ in the presence or the absence of macrophage Fpn (Fig. 6C). These results indicate a

selective requirement of iron for the regeneration of type IIb Downloaded from myoﬁbers. In the absence of macrophage Fpn, fat accumulation is prominent starting from day 7 after injury (Fig. 7A). mRNA expression of PPARg, a master regulator of adipogenesis, is upregulated 3 and 7 d after injury in the muscle tissue. Adiponectin, a marker of termi-

nally differentiated adipocytes, is also significantly upregulated at http://www.jimmunol.org/ day 10 of the regeneration process (Fig. 7B). In contrast, TGF-b expression is not affected (Fig. 7B). FAPs freshly purified from healthy muscle proliferate in vitro at a similar extent in the absence or in the presence of DFO (Fig. 7C) and differentiate efficiently toward adipocytes also if iron is restricted (Fig. 7D). DFO treatment of FAPs does not per se trigger the expression of HIF1a and its target genes, glut-1 and caIX (Supplemental Fig. 4B). by guest on September 25, 2021 Discussion The mechanisms by which macrophages sustain muscle healing FIGURE 6. Fiber type composition of muscle of mice adoptively transferred with Fpn/Scr or Fpn/KD macrophages. (A) IF for myosin H chain IIa have been only partially elucidated. In this study, we make the novel (green color), myosin H chain IIb (redcolor),andmyosinHchainI(blue observation that muscle-infiltrating macrophages after acute injury color) on Q muscles adoptively transferred with Fpn/Scr (left) or Fpn/KD rely on their professional ability to recycle iron to sustain muscle (right) macrophages and collected 15 d after CTX injection. Original mag- healing. This is a nonredundant event, because the targeting of a nification 320. Scale bars, 50 mm. (B) Mean CSA of type IIa (left panel) or single molecule expressed by macrophages, the sole cellular iron type IIb myofibers (right panel) from TA muscles of mice adoptively trans- exporter Fpn, is sufficient to jeopardize the regenerating process. Iron ferred with Fpn/Scr (empty bars) or Fpn/KD (filled bars) macrophages, col- recycling by spleen macrophages is critical in the homeostasis of 6 lected 15 d after CTX injury. Bars indicate the mean SEM. *p # 0.05, hemoglobin and red cell production, and prevents iron-restricted significantly different from mice adoptively transferred with Fpn/Scr mac- erythropoiesis and anemia (27). To the best of our knowledge, rophages. (C) Percentages of type I, type IIa, and type IIb fibers in TA sections our study provides the first demonstration that the macrophage iron of mice adoptively transferred with Fpn/Scr (left panel) or Fpn/KD (right panel) macrophages, collected 15 d after CTX injury. n.s., not significant. recycling machinery is essential also to support the homeostatic need of the skeletal muscle. However, although high erythropoiesis requests cause systemic iron changes, in our model iron muscle In the absence of macrophage Fpn, muscle regeneration is requirements are entirely satisfied on a local basis. compromised: the cross-sectional area of myofibers of mice During muscle damage and regeneration, we observe profound adoptively transferred with Fpn/KD macrophages is significantly changes in the muscle expression of heme and iron genes, which smaller than that of control mice (Fig. 5A–C [left graph], occur in the absence of changes in the systemic iron homeostasis, as Supplemental Fig. 2C). Regenerating myofibers of mice in which indicated by the stability of serum Tf saturation and hepcidin macrophages do not express Fpn also have significantly less nuclei mRNA levels in the liver, suggesting that all the events related to than control myofibers (Fig. 5C, right graph), indicating that iron release and uptake occur locally, by recycling the available macrophage-derived iron is required for satellite cell activation, iron. Muscle-infiltrating macrophages carry out this function: early proliferation, or fusion. In the absence of macrophage Fpn, after injury macrophages upregulate the expression of the mo- myofibers upregulate the expression of the TfR1 (Fig. 5D). This lecular machinery necessary for internalization and storage of the event indicates that they perceive a condition of iron deficiency. metal, thus preventing the noxious effects caused by its oxidizing To dissect in better detail the mechanism by which iron influences action (20, 21). Myoglobin, which is released because of the myogenic differentiation, we cultured freshly purified myoblasts in muscle necrosis, may represent a source of iron for macrophages the presence of reagents that limit iron availability (DFO). DFO and may possibly influence the expression of iron homeostatic treatment per se does not induce the expression of the HIF1a or of its genes. In fact, myoglobin is consistently contained in macro- The Journal of Immunology 1923

FIGURE 7. Fpn deficiency in macrophages results in muscle fat degeneration. (A) Oil Red O staining to detect fat on TA muscles of mice adoptively transferred with Fpn/Scr or Fpn/KD macrophages collected at 7, 10, and 15 d postinjury. Original magnification 34. Scale bars, 10 mm. Boxed areas in the overlay images are enlarged. (B) Time-course analysis of the PPARg, adiponectin, and TGF-b mRNA expression in Q and TA muscles of mice adoptively transferred with Fpn/Scr (empty bars) and Fpn/KD (filled bars) macro- Downloaded from phages, collected 3, 7, 10, and 15 d after CTX injury. mRNA levels were normalized to RPL13a1. Six different animals from two independent experiments were analyzed for each time point. Bars indicate the mean 6 SEM. ***p # 0.001, **p # 0.01, significantly different from mice adoptively http://www.jimmunol.org/ transferred with Fpn/Scr macrophages. (C) IF for BrdU (red color) on FAPs proliferating in culture treated or not with DFO. Nuclei were counterstained with Hoechst. Original magnification 320. The percentage of BrdU+ cells is depicted in right panel. (D) Red Oil staining of FAPs differentiated in the presence or absence of 5 mM DFO. Original magnification 320. The percentage by guest on September 25, 2021 of Oil Red+ cells is depicted in the right panel. n.s., not significantly different from untreated cells.

phages that infiltrate the damaged muscle, possibly reflecting ac- jeopardized. In this model, we do not impair other macrophage tive internalization of the iron-containing moiety released from crucial functions, and in particular, we do not interfere with their necrotic muscle fibers. At later times macrophages express high ability to scavenge the metal from the extracellular environment. levels of Fpn. In parallel, regenerating myofibers, which require Although defective clearance of iron might cause defective myo- iron for their metabolic reconstitution, express high levels of fiber regeneration, this has not been tested in our experimental TfR1, which binds and internalizes Tf-bound iron. setting. The process appears well coordinated: macrophages thus bridge Macrophages have been known for a long time to be associated the chronological gap between the time in which the muscle iron with skeletal muscle repair (28, 29), and data in various models storage protein myoglobin is released because of myofiber necrosis indicate that they are recruited to the injured skeletal muscle re- and the time in which the tissue requires iron for actual healing. gardless of the characteristics of the original noxa (30). In vivo They indeed play a dual function protecting the tissue against studies have unequivocally shown that macrophages actively oxidative damage and supplying the novel myofibers with an es- participate in the tissue repair process (1, 3, 4, 23, 31–38). A link sential nutrient. between iron release to regenerating myofibers and muscle healing We have established an in vivo system to selectively interfere has never been proposed. The interference with iron release de- with the latter event: by silencing the expression of macrophage termines a failure of wound healing in the case of human chronic Fpn, we prevented the export of iron, which as such remains sealed venous leg ulcers. In this model, the phenomenon appears to be within infiltrating macrophages. As a consequence, myofibers in related to a defect in the ability of macrophages to differentiate the regenerating tissue undergo a state of iron deficiency, as re- toward anti-inflammatory cells and not to the need for transfer of vealed by the upregulation of the TfR1, and tissue healing is the metal to precursor/stem cells (39). 1924 ROLE OF FERROPORTIN IN MUSCLE REGENERATION

Myogenic precursors are, in contrast, exquisitely sensitive to iron property is worth testing in other conditions of persistent damage of deprivation: in the absence of the metal they yield smaller myo- the skeletal muscle. tubes in vitro and smaller regenerated myofibers in vivo. The ability of myoblasts to proliferate appears to be the limiting step, although Acknowledgments they apparently do not depend on iron to fuse and type IIb fibers are The support of Annalisa Capobianco is gratefully acknowledged. We thank significantly more dependent on the metal. The balance between Enrica Gilberti for the measurement of macrophage iron content. satellite cell proliferation and differentiation is crucial to ensure muscle homeostasis. It depends on signaling pathways activated by growth factors, including hepatocyte, insulin, and epidermal Disclosures growth factors, which are finely regulated by iron availability (40) The authors have no financial conflicts of interest. and by bone morphogenetic protein (BMP) (41). Further studies are warranted to verify the molecular pathway by which iron References controls the fate of regenerating myofibers. The data indicate that 1. Arnold, L., A. Henry, F. Poron, Y. Baba-Amer, N. van Rooijen, A. Plonquet, the ability of macrophages to export iron via Fpn and to provide it R. K. Gherardi, and B. Chazaud. 2007. Inflammatory monocytes recruited after skeletal muscle injury switch into antiinflammatory macrophages to support to the regenerating myofibers is an essential nonredundant aspect myogenesis. J. Exp. Med. 204: 1057–1069. of their trophic actions. 2. Bosurgi, L., G. Corna, M. Vezzoli, T. Touvier, G. Cossu, A. A. Manfredi, Further work is also warranted to verify whether iron plays an S. Brunelli, and P. Rovere-Querini. 2012. Transplanted mesoangioblasts require macrophage IL-10 for survival in a mouse model of muscle injury. J. Immunol. instructive role in macrophage polarization in vivo and in particular 188: 6267–6277. in the transition between iron-storing and -releasing macrophages. 3. Warren, G. L., T. Hulderman, D. Mishra, X. Gao, L. Millecchia, L. O’Farrell, Downloaded from Interestingly, the heme moiety derived from senescent/damaged W. A. Kuziel, and P. P. Simeonova. 2005. Chemokine receptor CCR2 involve- ment in skeletal muscle regeneration. FASEB J. 19: 413–415. RBCs, which is known to directly induce the expression of Fpn 4. Ruffell, D., F. Mourkioti, A. Gambardella, P. Kirstetter, R. G. Lopez, (18, 19), has been recently demonstrated to influence the spleen- N. Rosenthal, and C. Nerlov. 2009. A CREB-C/EBPbeta cascade induces M2 specific development of macrophages (42). In particular, the heme macrophage-specific gene expression and promotes muscle injury repair. Proc. Natl. Acad. Sci. USA 106: 17475–17480. moiety derived from senescent or damaged erythrocytes triggers a 5. Sciorati, C., E. Rigamonti, A. A. Manfredi, and P. Rovere-Querini. 2016. Cell specific transcriptional program in monocytes necessary to drive death, clearance and immunity in the skeletal muscle. Cell Death Differ. 23: http://www.jimmunol.org/ the replenishment of red pulp and bone marrow macrophages 927–937. 6. Deng, B., M. Wehling-Henricks, S. A. Villalta, Y. Wang, and J. G. Tidball. 2012. during pathological hemolysis. IL-10 triggers changes in macrophage phenotype that promote muscle growth Besides favoring the myogenic differentiation of muscle-associated and regeneration. J. Immunol. 189: 3669–3680. 7. Corna, G., L. Campana, E. Pignatti, A. Castiglioni, E. Tagliafico, L. Bosurgi, precursors, the ability of infiltrating macrophages to export iron is A. Campanella, S. Brunelli, A. A. Manfredi, P. Apostoli, et al. 2010. Polarization crucial to avoid fat accumulation during the regeneration process. dictates iron handling by inflammatory and alternatively activated macrophages. Indeed, muscles of mice in which macrophages recruitment is Haematologica 95: 1814–1822. 8. Gaetano, C., L. Massimo, and M. Alberto. 2010. Control of iron homeostasis as a defective consistently undergo fat degeneration (3, 43). We observe key component of macrophage polarization. Haematologica 95: 1801–1803. that macrophage iron release significantly contributes to avoiding 9. Bories, G., S. Colin, J. Vanhoutte, B. Derudas, C. Copin, M. Fanchon, fat deposition. Indeed, upon macrophage Fpn targeting, fat degen- M. Daoudi, L. Belloy, S. Haulon, C. Zawadzki, et al. 2013. Liver X receptor by guest on September 25, 2021 activation stimulates iron export in human alternative macrophages. Circ. Res. eration occurs together with increased expression of adiponectin, a 113: 1196–1205. marker of adipocytes terminal differentiation (44). Also, the expres- 10. Pantopoulos, K., S. K. Porwal, A. Tartakoff, and L. Devireddy. 2012. Mecha- sion of PPARg, a master regulator of adipogenesis (45), was sig- nisms of mammalian iron homeostasis. Biochemistry 51: 5705–5724. 11. Kawane, K., H. Fukuyama, G. Kondoh, J. Takeda, Y. Ohsawa, Y. Uchiyama, and nificantly increased in the absence of macrophage Fpn. In contrast, S. Nagata. 2001. Requirement of DNase II for definitive erythropoiesis in the the antioxidant response, which is strictly associated to muscle injury mouse fetal liver. Science 292: 1546–1549. (23), is not apparently influenced by macrophage Fpn targeting. This 12. Knutson, M. D., M. Oukka, L. M. Koss, F. Aydemir, and M. Wessling-Resnick. 2005. Iron release from macrophages after erythrophagocytosis is up-regulated may occur because the iron overload is buffered by the upregulation by ferroportin 1 overexpression and down-regulated by hepcidin. Proc. Natl. of iron storage molecules such as ferritin, known to increase the Acad. Sci. USA 102: 1324–1328. 13. Ganz, T. 2012. Macrophages and systemic iron homeostasis. J. Innate Immun. 4: ability of macrophages to cope with oxidative stress (46, 47). 446–453. The actual cell population responsible for orchestrating adipo- 14. Beaumont, C., and C. Delaby. 2009. Recycling iron in normal and pathological genesis in the absence of macrophage iron needs to be identified. states. Semin. Hematol. 46: 328–338. 15. Andrews, N. C., and P. J. Schmidt. 2007. Iron homeostasis. Annu. Rev. Physiol. Various precursors including PDGFRa-expressing mesenchymal 69: 69–85. progenitors distinct from satellite cells, FAPs, and myoendothelial 16. Bratosin, D., J. Mazurier, J. P. Tissier, J. Estaquier, J. J. Huart, J. C. Ameisen, progenitors express adipogenic factors such as PPARg and yield D. Aminoff, and J. Montreuil. 1998. Cellular and molecular mechanisms of senescent erythrocyte phagocytosis by macrophages. A review. Biochimie 80: fat cells (48–50). Indeed, adipogenesis is actively restricted in 173–195. normal regenerating skeletal muscle in response to the BMP sig- 17. Guillaud, C., V. Loustau, and M. Michel. 2012. Hemolytic anemia in adults: naling (50), which also contributes to iron homeostasis. Interest- main causes and diagnostic procedures. Expert Rev. Hematol. 5: 229–241. 18. Delaby, C., N. Pilard, H. Puy, and F. Canonne-Hergaux. 2008. Sequential reg- ingly, iron deficiency restricts BMP signaling in the liver (51), and ulation of ferroportin expression after erythrophagocytosis in murine macro- a similar pathway could be activated in response to defective phages: early mRNA induction by haem, followed by iron-dependent protein macrophage iron transfer to myogenic precursors. expression. Biochem. J. 411: 123–131. 19. Marro, S., D. Chiabrando, E. Messana, J. Stolte, E. Turco, E. Tolosano, and Trigger adipogenesis might not be the only mechanism by which M. U. Muckenthaler. 2010. Heme controls ferroportin1 (FPN1) transcription PPARg influence muscle healing. Indeed, PPARs control the expres- involving Bach1, Nrf2 and a MARE/ARE sequence motif at position -7007 of sion of various genes involved in the maintenance of metabolic the FPN1 promoter. Haematologica 95: 1261–1268. 20. Camaschella, C. 2013. Iron and hepcidin: a story of recycling and balance. homeostasis and in inflammation (52). Mice with cardiomyocyte- Hematology Am. Soc. Hematol. Educ. Program 2013: 1–8. specific knockout of PPARg show cardiac hypertrophy (53). Data 21. Hentze, M. W., M. U. Muckenthaler, B. Galy, and C. Camaschella. 2010. Two to tango: regulation of mammalian iron metabolism. Cell 142: 24–38. obtained upon PPARg overexpression in cardiomyocytes suggest a 22. Beard, J. L. 2001. Iron biology in immune function, muscle metabolism and role in mitochondria biogenesis, which could be important for ef- neuronal functioning. J. Nutr. 131: 568S–579S; discussion 580S. fective healing of the injured skeletal muscle (54). 23. Vezzoli, M., P. Castellani, G. Corna, A. Castiglioni, L. Bosurgi, A. Monno, S. Brunelli, A. A. Manfredi, A. Rubartelli, and P. Rovere-Querini. 2011. High- In conclusion, our data demonstrate the essential role of iron mobility group box 1 release and redox regulation accompany regeneration and local recycling by macrophages for healing of skeletal muscle. This remodeling of skeletal muscle. Antioxid. Redox Signal. 15: 2161–2174. The Journal of Immunology 1925

24. Cordani, N., V. Pisa, L. Pozzi, C. Sciorati, and E. Clementi. 2014. Nitric oxide 39. Sindrilaru, A., T. Peters, S. Wieschalka, C. Baican, A. Baican, H. Peter, controls fat deposition in dystrophic skeletal muscle by regulating fibro- A. Hainzl, S. Schatz, Y. Qi, A. Schlecht, et al. 2011. An unrestrained proin- adipogenic precursor differentiation. Stem Cells 32: 874–885. flammatory M1 macrophage population induced by iron impairs wound healing 25. Santambrogio, P., A. Cozzi, S. Levi, E. Rovida, F. Magni, A. Albertini, and in humans and mice. J. Clin. Invest. 121: 985–997. P. Arosio. 2000. Functional and immunological analysis of recombinant mouse 40. Goodnough, J. B., E. Ramos, E. Nemeth, and T. Ganz. 2012. Inhibition of H- and L-ferritins from Escherichia coli. Protein Expr. Purif. 19: 212–218. hepcidin transcription by growth factors. Hepatology 56: 291–299. 26. Tidball, J. G., and S. A. Villalta. 2010. Regulatory interactions between muscle 41. Ono, Y., F. Calhabeu, J. E. Morgan, T. Katagiri, H. Amthor, and P. S. Zammit. and the immune system during muscle regeneration. Am. J. Physiol. Regul. 2011. BMP signalling permits population expansion by preventing premature Integr. Comp. Physiol. 298: R1173–R1187. myogenic differentiation in muscle satellite cells. Cell Death Differ. 18: 222– 27. Kohyama, M., W. Ise, B. T. Edelson, P. R. Wilker, K. Hildner, C. Mejia, 234. W. A. Frazier, T. L. Murphy, and K. M. Murphy. 2009. Role for Spi-C in the 42. Haldar, M., M. Kohyama, A. Y. So, W. Kc, X. Wu, C. G. Brisenõ, A. T. Satpathy, development of red pulp macrophages and splenic iron homeostasis. Nature 457: N. M. Kretzer, H. Arase, N. S. Rajasekaran, et al. 2014. Heme-mediated SPI-C 318–321. induction promotes monocyte differentiation into iron-recycling macrophages. 28. Robertson, T. A., M. A. Maley, M. D. Grounds, and J. M. Papadimitriou. 1993. Cell 156: 1223–1234. The role of macrophages in skeletal muscle regeneration with particular refer- 43. Sciorati, C., E. Clementi, A. A. Manfredi, and P. Rovere-Querini. 2015. Fat ence to chemotaxis. Exp. Cell Res. 207: 321–331. deposition and accumulation in the damaged and inflamed skeletal muscle: 29. McLennan, I. S. 1996. Degenerating and regenerating skeletal muscles contain cellular and molecular players. Cell. Mol. Life Sci. 72: 2135–2156. several subpopulations of macrophages with distinct spatial and temporal dis- 44. Yamauchi, T., and T. Kadowaki. 2013. Adiponectin receptor as a key player in tributions. J. Anat. 188: 17–28. healthy longevity and obesity-related diseases. Cell Metab. 17: 185–196. 30. Tidball, J. G. 2002. Interactions between muscle and the immune system during 45. Kawai, M., and C. J. Rosen. 2010. PPARg: a circadian transcription factor in modified musculoskeletal loading. Clin. Orthop. Relat. Res. 403(Suppl.): S100– adipogenesis and osteogenesis. Nat. Rev. Endocrinol. 6: 629–636. S109. 46. Garner, B., K. Roberg, and U. T. Brunk. 1998. Endogenous ferritin protects cells 31. St Pierre, B. A., and J. G. Tidball. 1994. Differential response of macrophage with iron-laden lysosomes against oxidative stress. Free Radic. Res. 29: 103– subpopulations to soleus muscle reloading after rat hindlimb suspension. J. Appl. 114. Physiol. 77: 290–297. 47. Orino, K., L. Lehman, Y. Tsuji, H. Ayaki, S. V. Torti, and F. M. Torti. 2001. 32. Chazaud, B., C. Sonnet, P. Lafuste, G. Bassez, A. C. Rimaniol, F. Poron, Ferritin and the response to oxidative stress. Biochem. J. 357: 241–247. Downloaded from F. J. Authier, P. A. Dreyfus, and R. K. Gherardi. 2003. Satellite cells attract 48. Joe, A. W., L. Yi, A. Natarajan, F. Le Grand, L. So, J. Wang, M. A. Rudnicki, and monocytes and use macrophages as a support to escape apoptosis and enhance F. M. Rossi. 2010. Muscle injury activates resident fibro/adipogenic progenitors muscle growth. J. Cell Biol. 163: 1133–1143. that facilitate myogenesis. Nat. Cell Biol. 12: 153–163. 33. Warren, G. L., L. O’Farrell, M. Summan, T. Hulderman, D. Mishra, M. I. Luster, 49. Uezumi, A., S. Fukada, N. Yamamoto, S. Takeda, and K. Tsuchida. 2010. W. A. Kuziel, and P. P. Simeonova. 2004. Role of CC chemokines in skeletal Mesenchymal progenitors distinct from satellite cells contribute to ectopic fat muscle functional restoration after injury. Am. J. Physiol. Cell Physiol. 286: cell formation in skeletal muscle. Nat. Cell Biol. 12: 143–152. C1031–C1036. 50. Huang, P., T. J. Schulz, A. Beauvais, Y. H. Tseng, and E. Gussoni. 2014. In-

34. Summan, M., G. L. Warren, R. R. Mercer, R. Chapman, T. Hulderman, N. Van tramuscular adipogenesis is inhibited by myo-endothelial progenitors with http://www.jimmunol.org/ Rooijen, and P. P. Simeonova. 2006. Macrophages and skeletal muscle regen- functioning Bmpr1a signalling. Nat. Commun. 5: 4063. eration: a clodronate-containing liposome depletion study. Am. J. Physiol. Regul. 51. Zumbrennen-Bullough, K. B., Q. Wu, A. B. Core, S. Canali, W. Chen, I. Theurl, Integr. Comp. Physiol. 290: R1488–R1495. D. Meynard, and J. L. Babitt. 2014. MicroRNA-130a is up-regulated in mouse 35. Tidball, J. G., and M. Wehling-Henricks. 2007. Macrophages promote muscle liver by iron deficiency and targets the bone morphogenetic protein (BMP) re- membrane repair and muscle fibre growth and regeneration during modified ceptor ALK2 to attenuate BMP signaling and hepcidin transcription. J. Biol. muscle loading in mice in vivo. J. Physiol. 578: 327–336. Chem. 289: 23796–23808. 36. Segawa, M., S. Fukada, Y. Yamamoto, H. Yahagi, M. Kanematsu, M. Sato, T. Ito, 52. Ahmadian, M., J. M. Suh, N. Hah, C. Liddle, A. R. Atkins, M. Downes, and A. Uezumi, S. Hayashi, Y. Miyagoe-Suzuki, et al. 2008. Suppression of mac- R. M. Evans. 2013. PPARg signaling and metabolism: the good, the bad and the rophage functions impairs skeletal muscle regeneration with severe fibrosis. Exp. future. Nat. Med. 19: 557–566. Cell Res. 314: 3232–3244. 53. Duan, S. Z., C. Y. Ivashchenko, M. W. Russell, D. S. Milstone, and 37. Lu, H., D. Huang, N. Saederup, I. F. Charo, R. M. Ransohoff, and L. Zhou. 2011. R. M. Mortensen. 2005. Cardiomyocyte-specific knockout and agonist of per-

Macrophages recruited via CCR2 produce insulin-like growth factor-1 to repair oxisome proliferator-activated receptor-gamma both induce cardiac hypertrophy by guest on September 25, 2021 acute skeletal muscle injury. FASEB J. 25: 358–369. in mice. Circ. Res. 97: 372–379. 38. Brigitte, M., C. Schilte, A. Plonquet, Y. Baba-Amer, A. Henri, C. Charlier, 54. Son, N. H., T. S. Park, H. Yamashita, M. Yokoyama, L. A. Huggins, K. Okajima, S. Tajbakhsh, M. Albert, R. K. Gherardi, and F. Chre´tien. 2010. Muscle resident S. Homma, M. J. Szabolcs, L. S. Huang, and I. J. Goldberg. 2007. Car- macrophages control the immune cell reaction in a mouse model of notexin- diomyocyte expression of PPARgamma leads to cardiac dysfunction in mice. induced myoinjury. Arthritis Rheum. 62: 268–279. J. Clin. Invest. 117: 2791–2801.