Bone mass and structure in adults are maintained by constant bone remodeling with balanced bone formation by osteoblasts and bone resorption by osteoclasts (Zaidi, 2007; Siddiqui and Partridge, 2016). Under pathological conditions, however, excessive bone resorption caused by either increased number or exaggerated activities of osteoclasts leads to bone loss which is a hallmark of many metabolic bone diseases such as osteoporosis, rheumatoid arthritis, Paget’s disease of bone, periodontal disease, and tumor bone metastasis (Mbalaviele et al., 2017; Novack and Teitelbaum, 2008). Osteoclasts are multinucleated cells formed by fusion of mononuclear precursors that are differentiated from the monocyte/macrophage lineage of hematopoietic cells. This process is triggered by two indispensable cytokines M-CSF (macrophage colony-stimulating factor) and RANKL (receptor activator of NF-κB ligand) (Walker, 1975; Xiao et al., 2017; Boyle et al., 2003; Nakashima et al., 2012). Upon attachment to bone, osteoclasts organize their actin cytoskeleton to form actin-rings that seal the resorptive microenvironment (Väänänen et al., 2000; Teitelbaum, 2011). The dissolution of bone minerals and digestion of organic bone matrix mainly type I collagen are executed by hydrochloric acid and acidic hydrolase cathepsin K, respectively (Zhao, 2012; Fujiwara et al., 2016). Both osteoclast formation and bone-resorbing activity of mature osteoclasts demand high energy. However, the pathways and molecular mechanisms regulating osteoclast energy metabolism in osteoclasts remain largely unknown.

Both glycolysis and mitochondrial oxidative phosphorylation (OXPHOS) increase during osteoclast differentiation (Kubatzky et al., 2018; Arnett and Orriss, 2018; Kim et al., 2007; Indo et al., 2013; Li et al., 2020a). Osteoclasts contain numerous mitochondria (Ch’uan, 1931; Holtrop and King, 1977). The mitochondrial respiratory complex I and the key mitochondria transcriptional regulator PGC-1β (peroxisome proliferator-activated receptor γ coactivator 1β) are crucial for osteoclast differentiation (Jin et al., 2014; Ishii et al., 2009). Mitochondrial ROS (reactive oxygen species) stimulates osteoclast differentiation (Srinivasan et al., 2010; Kim et al., 2013; Kim et al., 2017). On the contrary, it has been reported that decreased mitochondrial biogenesis and activity by PGC-1β deletion in osteoclast lineage cells disrupt osteoclast cytoskeletal organization and function but not osteoclastogenesis (Zhang et al., 2018). Loss of G-protein Gα13 in osteoclast progenitors promotes mitochondrial respiration and cytoskeleton organization but has little effects on osteoclastogenesis in cultures and in mice (Nakano et al., 2019). Heretofore, the mechanisms by which energy metabolism regulates osteoclast differentiation and function remain unclear.

Iron is a nutritional element that plays a fundamental role in mitochondrial metabolism and the biosynthesis of heme and Fe-S clusters which are critical components of the mitochondrial respiratory complexes (Figure 1—figure supplement 1A and Xu et al., 2013). Both systemic and cellular iron homeostasis play a pivotal role in bone remodeling (Balogh et al., 2018). Osteoporosis and pathological bone fractures are commonly associated with the hereditary iron-overload disease hemochromatosis and the acquired iron-overload conditions in treatment of thalassemia and sickle cell disease (Almeida and Roberts, 2005; Fung et al., 2008; Dede et al., 2016; Jandl et al., 2020). Moreover, increased bone resorption and/or decreased bone formation have been observed in genetic mouse models of iron overload diseases and in mice fed with excessive iron (Tsay et al., 2010; Guggenbuhl et al., 2011; Xiao et al., 2015). In contrast to iron overload, much less is known about the influence of iron deficiency on bone cells and bone homeostasis. Severe iron deficiency impairs both bone resorption and bone formation and causes osteopenia in rats (Katsumata et al., 2009; Diaz-Castro et al., 2012), whereas iron chelation inhibits osteoclastogenesis and bone resorption in vitro and in vivo and attenuates estrogen deficiency induced bone loss in mice (Ishii et al., 2009; Guo et al., 2015).

There are two transferrin (Tf) receptors (Tfrs) in mammalian cells (Kawabata, 2019). Transferrin receptor 1 (Tfr1), encoded by Tfrc gene, is ubiquitously expressed with high affinity to Fe³⁺-loaded holo-Tf. Germline deletion of Tfrc in mice leads to early embryonic lethality due to severe defects in erythroid and neuronal development (Levy et al., 1999). A missense mutation in TFRC causes combined immunodeficiency in human (Jabara et al., 2016). The tissue-specific deletion of Tfrc gene in neurons, skeletal muscles, cardiomyocytes, hematopoietic stem cells, and adipocytes in mice disrupts cellular iron homeostasis and causes severe metabolic defects in these tissues (Matak et al., 2016; Barrientos et al., 2015; Xu et al., 2015; Wang et al., 2020; Li et al., 2020b). In contrast to Tfr1, Tfr2 has lower affinity to holo-Tf and is predominantly expressed in liver and erythroid precursor cells (Kawabata, 2019). Mutations of TFR2 in human and germline or hepatocyte-specific deletion of Tfr2 in mice cause type 3 hemochromatosis (Camaschella et al., 2000; Fleming et al., 2002). More recently, Tfr2 has been reported to regulate osteoblastic bone formation and bone mass in mice, independent of its function in iron homeostasis (Rauner et al., 2019). Nevertheless, the cell autonomous functions of Tfr1 and Tfr2 in osteoclast lineage cells remain unknown.

To elucidate the role of Tfr1 and Tfr1-mediated iron uptake in osteoclast lineage cells, we generated Tfrc conditional knockout mice in myeloid osteoclast precursors and differentiated osteoclasts by crossing Tfrc-flox mice with Lyz2-Cre and Ctsk-Cre mice, respectively, and used them for comprehensive skeletal phenotyping in vivo and mechanistic studies in vitro.

Loss of Tfr1 in osteoclast lineage cells results in >50% decrease in the total intracellular iron content in mature osteoclasts, whereas the iron levels in monocytes and pre-osteoclasts are not affected by Tfr1-deficiency (Figure 1). Since the heme transporter Hcp1 is upregulated during osteoclast differentiation and the fact that hemin but not FAC (absorbed through NTBI) is able to partially rescue the Tfr1-deficient osteoclast phenotypes (Figure 5), the residual cellular iron in Tfr1-deficient osteoclasts could be obtained through uptake of heme. Given that the key molecules involved in Tf-dependent and heme-dependent iron uptake pathway are significantly upregulated during osteoclast differentiation and that the expressions of NTBI iron transporter Zip14 and ferritin transporter (Scara-5 and Tim-2) are either decreased in osteoclasts or undetectable in osteoclast lineage cells (Figure 1A and data not shown), these results indicate that mature osteoclasts acquire extracellular iron largely through uptake of Tf and heme.

Although iron chelation by desferrioxamine has been shown to inhibit osteoclasts and prevent bone loss in ovariectomized mice (Ishii et al., 2009) and in a mouse model of Alzheimer’s disease (Guo et al.), the intrinsic functions of Tfr1 and Tfr1-mediated iron uptake in osteoclast lineage cells in vivo have not been investigated using genetically modified mouse models. To fill in this knowledge gap, we have generated two lines of Tfrc conditional knockout mice in which the Tfr1 expression is disrupted in Lyz2-Cre expressing myeloid osteoclast precursors and Ctsk-Cre expressing differentiated osteoclasts, respectively. The μCT analysis of control and Tfr1^ΔlysM mice at three different ages ranging from the pre-pubertal 3-week-old to the post-pubertal 10-week-old and the adult 6-month-old has unveiled a significant increase in trabecular bone mass in post-pubertal and adult female, but not male, Tfr1^ΔlysM mice compared to their age- and gender-matched control mice (Figure 2 and Figure 2—figure supplement 4). Although the significant increase of trabecular bone mass was observed in both genders of 10-week-old Tfr1^ΔCtsk mice, this phenotype is more pronounced in female knockout mice (Figure 2—figure supplement 6). These results indicate that the accrual of trabecular bone mass in Tfr1-deficient mice is gender-dependent. The exact mechanisms underlying the gender difference in the trabecular bone phenotype of mice upon the Tfr1-deficiency in osteoclasts are unknown and need further investigation since sex as a biological variable in mice and humans has been increasingly appreciated (Bhargava et al., 2021). A similar finding has been recently reported in Slc2a1 (encoding the glucose transporter protein type 1, Glut1) myeloid conditional knockout mice (Li et al., 2020a). While Glut1-deficient BMM cultured from both male and female knockout mice have similar degree of blunted osteoclast differentiation in vitro, deletion of Slc2a1 in Lyz2-Cre expressing myeloid cells in mice inhibits osteoclastogenesis and leads to increased trabecular bone mass in female but not male mice.

Estrogen has been reported to regulate systemic and cellular iron homeostasis by inhibiting the expression of hepcidin, a liver-derived iron regulating hormone that binds to and induces degradation of iron exporter Fpn (Hou et al., 2012; Yang et al., 2012; Ikeda et al., 2012); however, the levels of serum hepcidin and total iron are similar in control and Tfr1^ΔlysM male and female mice (Figure 2—figure supplement 3) excluding this possible mechanism. On the other hand, our in vitro supplemental experiment has demonstrated that 17-β estradiol inhibits the mRNA expression of Tfrc in mature osteoclasts and synergistically with Tfr1-deficiency attenuates the mRNA expression of mt-Co1 (encoding the mitochondrial cytochrome c oxidase subunit I, mt-Co1), whereas dihydrotestosterone has no effects on Tfr1, Fpn, and mt-Co1 expression in osteoclast lineage cells (Figure 2—figure supplement 8). This finding is consistent with our recent published work that estrogen attenuates mitochondrial OXPHOS and ATP production in osteoclasts (Kim et al., 2020). The tissue- and bone cell type-specific sex differences in mitochondria and the underlying mechanisms in skeletons in both mice and humans need further investigation as recently reported in adipose tissue (Chella Krishnan et al., 2021).

It is well established that estrogen-mediated mechanisms play a key role in mediating bone accretion during pubertal growth period. During this developmental period, estrogen is known to exert an anabolic effect on osteoblast besides its well-established role in inhibiting osteoclast functions. However, the loss of estrogen as seen during menopause or due to ovariectomy in adults leads to bone loss primarily due to increased bone resorption. The increased trabecular bone volume of Tfr1^ΔLysM mice is evident at the post-pubertal age of 10 weeks but is not different at the pre-pubertal age of 3 weeks (Figure 2), thus suggesting that the skeletal phenotype gets manifested during pubertal and post-pubertal growth periods. While loss of Tfr1 does not affect estrogen deficiency-induced bone loss, the trabecular bone mass of Tfr1^ΔLysM mice is persistently higher than control mice after OVX. The issue of whether the skeletal phenotype in 10-week-old Tfr1^ΔLysM mice is dependent on pubertal increase in estrogen remains to be evaluated by performing OVX at pre-pubertal age of 3 weeks in future.

Both Lyz2-Cre and Ctsk-Cre mice are commonly used to conditionally delete the floxed genes at different stages of osteoclast lineage cells (Dallas et al., 2018). While the Lyz2-Cre deletes genes in monocytes of osteoclast precursor cells, it also targets other myeloid cells such as neutrophils and dendritic cells (Abram et al., 2014). Although Nfatc1-deletion by Mx1-Cre attenuates osteoclastogenesis and bone resorption in mice, Nfatc1^ΔlysM mice displayed no osteoclast and bone phenotypes (Aliprantis et al., 2008), suggesting that the deletion efficiency of Lyz2-Cre is weaker than Mx1-Cre and is allele-dependent (Dallas et al., 2018 and Clausen et al., 1999). The Ctsk-Cre, especially the Ctsk-Cre knock-in line, has been reported to delete genes at the late stages of differentiated osteoclasts (Nakamura et al., 2007). It should be noticed that the Ctsk-Cre has been reported to cause gene-disruption in mesenchymal progenitor cells including periosteal cells and osteoblasts (Yang et al., 2013 and Ruiz et al., 2016). Thus, the differences of skeletal phenotypes between Tfr1^ΔlysM and Tfr1^ΔCtsk mice might be caused by the in vivo efficacy and the specificity of targeting cells of these two Cre mouse lines.

The high bone mass observed in Tfr1^ΔlysM and Tfr1^ΔCtsk mice is more dramatic in the appendicular skeleton (femurs and tibias) (Figure 2 and Figure 2—figure supplement 6) than the axial skeleton (vertebras) (Figure 2—figure supplements 5 and 7). The trabecular bone is more vulnerable to Tfr1-deficiency than the cortical bone. The exact explanation(s) for distinct effects of Tfr1-deficiency on different sites and compartments of bone are not available. Our qPCR result demonstrated that deletion efficiency of Tfr1 gene, Tfrc, in vertebra is less than femurs (Figure 2—figure supplement 9). It is also likely that the different physical properties, remodeling rate, microenvironment, and the heterogeneous osteoclast subtypes in the distinct sites and compartments of bone may contribute to these site- and compartment-specific phenotypes which have been reported in other genetically modified mouse models (Cruz Grecco Teixeira et al., 2016 and Sato et al., 2020).

In this study, we have demonstrated that energy metabolism regulated by Tf11-mediated iron uptake is specifically indispensable for osteoclast activation and function but has little influence on osteoclast differentiation. In contrast, Ishii et al., 2009 have reported that in vitro knockdown of Tfr1 expression in BMM by short-hairpin RNA (shRNA) inhibits osteoclast differentiation. This discrepant finding might be caused by the off-target effects of Tfr1 shRNA on expression of genes essential for osteoclast differentiation or by the distinct effects of short-term Tfr1 downregulation by shRNA vs long-term genetic deletion of Tfr1 on osteoclast lineage cells. In support of our finding, it has been recently reported that deletion of mitochondria coactivator PGC-1β in myeloid osteoclast precursor cells by Lyz2-Cre diminishes mitochondrial biogenesis and function in osteoclasts, leading to cytoskeletal disorganization and bone resorption retardation with normal osteoclasts differentiation (Zhang et al., 2018). In this study, the PGC-1β-stimulated osteoclast cytoskeleton activation is identified to be mediated by GIT1 (G protein coupled receptor kinase 2 interacting protein 1) which has been reported to regulates osteoclast function and bone mass (Menon et al., 2010 and Jones and Katan, 2007). Another report by Nakano et al. has shown that deletion of G-protein Gα₁₃ in myeloid cells augments mitochondrial biogenesis and metabolism in osteoclasts. The gain-of-function of mitochondrial pathway induced by Gα₁₃-deletion in osteoclasts promotes actin cytoskeleton dynamic and bone resorption via activation of cytoskeleton regulators c-Src, Pyk2, and RhoA-Rock2. Again, Gα₁₃-deletion in osteoclast lineage cells has minimum effects on osteoclastogenesis in vivo and in vitro (Nakano et al., 2019). Intriguingly, we have unveiled that Tfr1-mediated iron uptake regulates osteoclast cytoskeleton probably via a distinct mechanism. The level of GIT1, Rac1, Cdc42, and Rock2 remains unchanged in Tfr1-null osteoclasts (Supplementary file 5). However, the activation of c-Src-Rac1-WRC axis by M-CSF is attenuated and mostly downregulated in the absence of Tfr1, indicating that Tfr1 modulates osteoclast cytoskeleton through promoting the stability of WRC complex. In supporting of this premise, overexpression of Hem1 (also known as Nckap1l) partially rescues the cytoskeletal defects in Tfr1-deficient osteoclasts (Figure 9).

In summary, we have provided evidence demonstrating that Tfr1-mediated iron uptake is a major iron acquisition pathway in osteoclast lineage cells that differentially regulates trabecular bone remodeling in perpendicular and axial bones of female and male mice. The increased intracellular iron facilitated by Tfr1 is specifically required for osteoclast mitochondrial energy metabolism and cytoskeletal organization.

The animal work follows the ARRIVE guidelines. All the in vivo and in vitro experiments were performed and analyzed in double-blinded manner.

All data generated or analysed during this study are included in the manuscript and supporting file.

1. 1. Abram CL
   2. Roberge GL
   3. Hu Y
   4. Lowell CA

   (2014) Comparative analysis of the efficiency and specificity of myeloid-Cre deleting strains using ROSA-EYFP reporter mice

   Journal of Immunological Methods 408:89–100.

   https://doi.org/10.1016/j.jim.2014.05.009

   • PubMed
   • Google Scholar
2. 1. Aliprantis AO
   2. Ueki Y
   3. Sulyanto R
   4. Park A
   5. Sigrist KS
   6. Sharma SM
   7. Ostrowski MC
   8. Olsen BR
   9. Glimcher LH

   (2008) NFATc1 in mice represses osteoprotegerin during osteoclastogenesis and dissociates systemic osteopenia from inflammation in cherubism

   The Journal of Clinical Investigation 118:3775–3789.

   https://doi.org/10.1172/JCI35711

   • PubMed
   • Google Scholar
3. 1. Almeida A
   2. Roberts I

   (2005) Bone involvement in sickle cell disease

   British Journal of Haematology 129:482–490.

   https://doi.org/10.1111/j.1365-2141.2005.05476.x

   • PubMed
   • Google Scholar
4. 1. Arnett TR
   2. Orriss IR

   (2018) Metabolic properties of the osteoclast

   Bone 115:25–30.

   https://doi.org/10.1016/j.bone.2017.12.021

   • PubMed
   • Google Scholar
5. 1. Balogh E
   2. Paragh G
   3. Jeney V

   (2018) Influence of iron on bone homeostasis

   Pharmaceuticals 11:E107.

   https://doi.org/10.3390/ph11040107

   • PubMed
   • Google Scholar
6. 1. Barrientos T
   2. Laothamatas I
   3. Koves TR
   4. Soderblom EJ
   5. Bryan M
   6. Moseley MA
   7. Muoio DM
   8. Andrews NC

   (2015) Metabolic catastrophe in mice lacking transferrin receptor in muscle

   EBioMedicine 2:1705–1717.

   https://doi.org/10.1016/j.ebiom.2015.09.041

   • PubMed
   • Google Scholar
7. 1. Bhargava A
   2. Arnold AP
   3. Bangasser DA
   4. Denton KM
   5. Gupta A
   6. Hilliard Krause LM
   7. Mayer EA
   8. McCarthy M
   9. Miller WL
   10. Raznahan A
   11. Verma R

   (2021) Considering sex as a biological variable in basic and clinical studies

   An Endocrine Society Scientific Statement Endocr Rev 42:219–258.

   https://doi.org/10.1210/endrev/bnaa034

   • Google Scholar
8. 1. Blangy A
   2. Bompard G
   3. Guerit D
   4. Marie P
   5. Maurin J
   6. Morel A
   7. Vives V

   (2020) The osteoclast cytoskeleton - current understanding and therapeutic perspectives for osteoporosis

   Journal of Cell Science 133:jcs244798.

   https://doi.org/10.1242/jcs.244798

   • PubMed
   • Google Scholar
9. 1. Boyle WJ
   2. Simonet WS
   3. Lacey DL

   (2003) Osteoclast differentiation and activation

   Nature 423:337–342.

   https://doi.org/10.1038/nature01658

   • PubMed
   • Google Scholar
10. 1. Camaschella C
    2. Roetto A
    3. Calì A
    4. De Gobbi M
    5. Garozzo G
    6. Carella M
    7. Majorano N
    8. Totaro A
    9. Gasparini P

    (2000) The gene TFR2 is mutated in a new type of haemochromatosis mapping to 7q22

    Nature Genetics 25:14–15.

    https://doi.org/10.1038/75534

    • PubMed
    • Google Scholar
11. 1. Chella Krishnan K
    2. Vergnes L
    3. Acín-Pérez R
    4. Stiles L
    5. Shum M
    6. Ma L
    7. Mouisel E
    8. Pan C
    9. Moore TM
    10. Péterfy M
    11. Romanoski CE
    12. Reue K
    13. Björkegren JLM
    14. Laakso M
    15. Liesa M
    16. Lusis AJ

    (2021) Sex-specific genetic regulation of adipose mitochondria and metabolic syndrome by Ndufv2

    Nature Metabolism 3:1552–1568.

    https://doi.org/10.1038/s42255-021-00481-w

    • PubMed
    • Google Scholar
12. 1. Chen AC
    2. Donovan A
    3. Ned-Sykes R
    4. Andrews NC

    (2015) Noncanonical role of transferrin receptor 1 is essential for intestinal homeostasis

    PNAS 112:11714–11719.

    https://doi.org/10.1073/pnas.1511701112

    • PubMed
    • Google Scholar
13. 1. Ch’uan CH

    (1931) Mitochondria in osteoclasts

    The Anatomical Record 49:397–401.

    https://doi.org/10.1002/ar.1090490408

    • Google Scholar
14. 1. Clausen BE
    2. Burkhardt C
    3. Reith W
    4. Renkawitz R
    5. Förster I

    (1999) Conditional gene targeting in macrophages and granulocytes using LysMcre mice

    Transgenic Research 8:265–277.

    https://doi.org/10.1023/A:1008942828960

    • Google Scholar
15. 1. Cruz Grecco Teixeira MB
    2. Martins GM
    3. Miranda-Rodrigues M
    4. De Araújo IF
    5. Oliveira R
    6. Brum PC
    7. Azevedo Gouveia CH

    (2016) lack of α2C-adrenoceptor results in contrasting phenotypes of long bones and vertebra and prevents the thyrotoxicosis-induced osteopenia

    PLOS ONE 11:e0146795.

    https://doi.org/10.1371/journal.pone.0146795

    • PubMed
    • Google Scholar
16. 1. Dallas SL
    2. Xie Y
    3. Shiflett LA
    4. Ueki Y

    (2018) Mouse Cre models for the study of bone diseases

    Current Osteoporosis Reports 16:466–477.

    https://doi.org/10.1007/s11914-018-0455-7

    • PubMed
    • Google Scholar
17. 1. Dede AD
    2. Trovas G
    3. Chronopoulos E
    4. Triantafyllopoulos IK
    5. Dontas I
    6. Papaioannou N
    7. Tournis S

    (2016) Thalassemia-associated osteoporosis: a systematic review on treatment and brief overview of the disease

    Osteoporosis International 27:3409–3425.

    https://doi.org/10.1007/s00198-016-3719-z

    • PubMed
    • Google Scholar
18. 1. Diaz-Castro J
    2. López-Frías MR
    3. Campos MS
    4. López-Frías M
    5. Alférez MJM
    6. Nestares T
    7. Ojeda ML
    8. López-Aliaga I

    (2012) Severe nutritional iron-deficiency anaemia has a negative effect on some bone turnover biomarkers in rats

    European Journal of Nutrition 51:241–247.

    https://doi.org/10.1007/s00394-011-0212-5

    • PubMed
    • Google Scholar
19. 1. Donovan A
    2. Roy CN
    3. Andrews NC

    (2006) The ins and outs of iron homeostasis

    Physiology 21:115–123.

    https://doi.org/10.1152/physiol.00052.2005

    • PubMed
    • Google Scholar
20. 1. Fleming RE
    2. Ahmann JR
    3. Migas MC
    4. Waheed A
    5. Koeffler HP
    6. Kawabata H
    7. Britton RS
    8. Bacon BR
    9. Sly WS

    (2002) Targeted mutagenesis of the murine transferrin receptor-2 gene produces hemochromatosis

    PNAS 99:10653–10658.

    https://doi.org/10.1073/pnas.162360699

    • PubMed
    • Google Scholar
21. 1. Fujiwara T
    2. Ye S
    3. Castro-Gomes T
    4. Winchell CG
    5. Andrews NW
    6. Voth DE
    7. Varughese KI
    8. Mackintosh SG
    9. Feng Y
    10. Pavlos N
    11. Nakamura T
    12. Manolagas SC
    13. Zhao H

    (2016) PLEKHM1/DEF8/RAB7 complex regulates lysosome positioning and bone homeostasis

    JCI Insight 1:e86330.

    https://doi.org/10.1172/jci.insight.86330

    • PubMed
    • Google Scholar
22. 1. Fung EB
    2. Harmatz PR
    3. Milet M
    4. Coates TD
    5. Thompson AA
    6. Ranalli M
    7. Mignaca R
    8. Scher C
    9. Giardina P
    10. Robertson S
    11. Neumayr L
    12. Vichinsky EP
    13. Multi-Center Iron Overload Study Group

    (2008) Fracture prevalence and relationship to endocrinopathy in iron overloaded patients with sickle cell disease and thalassemia

    Bone 43:162–168.

    https://doi.org/10.1016/j.bone.2008.03.003

    • PubMed
    • Google Scholar
23. 1. Gammella E
    2. Buratti P
    3. Cairo G
    4. Recalcati S

    (2017) The transferrin receptor: the cellular iron gate

    Metallomics : Integrated Biometal Science 9:1367–1375.

    https://doi.org/10.1039/c7mt00143f

    • PubMed
    • Google Scholar
24. 1. Gu Z
    2. Wang H
    3. Xia J
    4. Yang Y
    5. Jin Z
    6. Xu H
    7. Shi J
    8. De Domenico I
    9. Tricot G
    10. Zhan F

    (2015) Decreased ferroportin promotes myeloma cell growth and osteoclast differentiation

    Cancer Research 75:2211–2221.

    https://doi.org/10.1158/0008-5472.CAN-14-3804

    • PubMed
    • Google Scholar
25. 1. Guggenbuhl P
    2. Fergelot P
    3. Doyard M
    4. Libouban H
    5. Roth MP
    6. Gallois Y
    7. Chalès G
    8. Loréal O
    9. Chappard D

    (2011) Bone status in a mouse model of genetic hemochromatosis

    Osteoporosis International 22:2313–2319.

    https://doi.org/10.1007/s00198-010-1456-2

    • PubMed
    • Google Scholar
26. 1. Guo JP
    2. Pan JX
    3. Xiong L
    4. Xia WF
    5. Cui S
    6. Xiong WC

    (2015) Iron chelation inhibits osteoclastic differentiation in vitro and in Tg2576 mouse model of alzheimer’s disease

    PLOS ONE 10:e0139395.

    https://doi.org/10.1371/journal.pone.0139395

    • PubMed
    • Google Scholar
27. 1. Hentze MW
    2. Muckenthaler MU
    3. Galy B
    4. Camaschella C

    (2010) Two to tango: regulation of Mammalian iron metabolism

    Cell 142:24–38.

    https://doi.org/10.1016/j.cell.2010.06.028

    • PubMed
    • Google Scholar
28. 1. Holtrop ME
    2. King GJ

    (1977) The ultrastructure of the osteoclast and its functional implications

    Clinical Orthopaedics and Related Research 123:177–196.

    https://doi.org/10.1097/00003086-197703000-00062

    • PubMed
    • Google Scholar
29. 1. Hou Y
    2. Zhang S
    3. Wang L
    4. Li J
    5. Qu G
    6. He J
    7. Rong H
    8. Ji H
    9. Liu S

    (2012) Estrogen regulates iron homeostasis through governing hepatic hepcidin expression via an estrogen response element

    Gene 511:398–403.

    https://doi.org/10.1016/j.gene.2012.09.060

    • PubMed
    • Google Scholar
30. 1. Ikeda Y
    2. Tajima S
    3. Izawa-Ishizawa Y
    4. Kihira Y
    5. Ishizawa K
    6. Tomita S
    7. Tsuchiya K
    8. Tamaki T

    (2012) Estrogen regulates hepcidin expression via GPR30-BMP6-dependent signaling in hepatocytes

    PLOS ONE 7:e40465.

    https://doi.org/10.1371/journal.pone.0040465

    • PubMed
    • Google Scholar
31. 1. Indo Y
    2. Takeshita S
    3. Ishii KA
    4. Hoshii T
    5. Aburatani H
    6. Hirao A
    7. Ikeda K

    (2013) Metabolic regulation of osteoclast differentiation and function

    Journal of Bone and Mineral Research 28:2392–2399.

    https://doi.org/10.1002/jbmr.1976

    • PubMed
    • Google Scholar
32. 1. Ishii K
    2. Fumoto T
    3. Iwai K
    4. Takeshita S
    5. Ito M
    6. Shimohata N
    7. Aburatani H
    8. Taketani S
    9. Lelliott CJ
    10. Vidal-Puig A
    11. Ikeda K

    (2009) Coordination of PGC-1beta and iron uptake in mitochondrial biogenesis and osteoclast activation

    Nature Medicine 15:259–266.

    https://doi.org/10.1038/nm.1910

    • PubMed
    • Google Scholar
33. 1. Jabara HH
    2. Boyden SE
    3. Chou J
    4. Ramesh N
    5. Massaad MJ
    6. Benson H
    7. Bainter W
    8. Fraulino D
    9. Rahimov F
    10. Sieff C
    11. Liu ZJ
    12. Alshemmari SH
    13. Al-Ramadi BK
    14. Al-Dhekri H
    15. Arnaout R
    16. Abu-Shukair M
    17. Vatsayan A
    18. Silver E
    19. Ahuja S
    20. Davies EG
    21. Sola-Visner M
    22. Ohsumi TK
    23. Andrews NC
    24. Notarangelo LD
    25. Fleming MD
    26. Al-Herz W
    27. Kunkel LM
    28. Geha RS

    (2016) A missense mutation in TFRC, encoding transferrin receptor 1, causes combined immunodeficiency

    Nature Genetics 48:74–78.

    https://doi.org/10.1038/ng.3465

    • PubMed
    • Google Scholar
34. 1. Jandl NM
    2. Rolvien T
    3. Schmidt T
    4. Mussawy H
    5. Nielsen P
    6. Oheim R
    7. Amling M
    8. Barvencik F

    (2020) Impaired bone microarchitecture in patients with hereditary hemochromatosis and skeletal complications

    Calcified Tissue International 106:465–475.

    https://doi.org/10.1007/s00223-020-00658-7

    • PubMed
    • Google Scholar
35. 1. Jin Z
    2. Wei W
    3. Yang M
    4. Du Y
    5. Wan Y

    (2014) Mitochondrial complex I activity suppresses inflammation and enhances bone resorption by shifting macrophage-osteoclast polarization

    Cell Metabolism 20:483–498.

    https://doi.org/10.1016/j.cmet.2014.07.011

    • PubMed
    • Google Scholar
36. 1. Jones NP
    2. Katan M

    (2007) Role of phospholipase Cgamma1 in cell spreading requires association with a beta-Pix/GIT1-containing complex, leading to activation of Cdc42 and Rac1

    Molecular and Cellular Biology 27:5790–5805.

    https://doi.org/10.1128/MCB.00778-07

    • PubMed
    • Google Scholar
37. 1. Katsumata S
    2. Katsumata-Tsuboi R
    3. Uehara M
    4. Suzuki K

    (2009) Severe iron deficiency decreases both bone formation and bone resorption in rats

    The Journal of Nutrition 139:238–243.

    https://doi.org/10.3945/jn.108.093757

    • PubMed
    • Google Scholar
38. 1. Kawabata H

    (2019) Transferrin and transferrin receptors update

    Free Radical Biology & Medicine 133:46–54.

    https://doi.org/10.1016/j.freeradbiomed.2018.06.037

    • PubMed
    • Google Scholar
39. 1. Kim JM
    2. Jeong D
    3. Kang HK
    4. Jung SY
    5. Kang SS
    6. Min BM

    (2007) Osteoclast precursors display dynamic metabolic shifts toward accelerated glucose metabolism at an early stage of RANKL-stimulated osteoclast differentiation

    Cellular Physiology and Biochemistry 20:935–946.

    https://doi.org/10.1159/000110454

    • PubMed
    • Google Scholar
40. 1. Kim H
    2. Kim T
    3. Jeong BC
    4. Cho IT
    5. Han D
    6. Takegahara N
    7. Negishi-Koga T
    8. Takayanagi H
    9. Lee JH
    10. Sul JY
    11. Prasad V
    12. Lee SH
    13. Choi Y

    (2013) Tmem64 modulates calcium signaling during RANKL-mediated osteoclast differentiation

    Cell Metabolism 17:249–260.

    https://doi.org/10.1016/j.cmet.2013.01.002

    • PubMed
    • Google Scholar
41. 1. Kim H
    2. Lee YD
    3. Kim HJ
    4. Lee ZH
    5. Kim HH

    (2017) SOD2 and Sirt3 control osteoclastogenesis by regulating mitochondrial ROS

    Journal of Bone and Mineral Research 32:397–406.

    https://doi.org/10.1002/jbmr.2974

    • PubMed
    • Google Scholar
42. 1. Kim H-N
    2. Ponte F
    3. Nookaew I
    4. Ucer Ozgurel S
    5. Marques-Carvalho A
    6. Iyer S
    7. Warren A
    8. Aykin-Burns N
    9. Krager K
    10. Sardao VA
    11. Han L
    12. de Cabo R
    13. Zhao H
    14. Jilka RL
    15. Manolagas SC
    16. Almeida M

    (2020) Estrogens decrease osteoclast number by attenuating mitochondria oxidative phosphorylation and ATP production in early osteoclast precursors

    Scientific Reports 10:11933.

    https://doi.org/10.1038/s41598-020-68890-7

    • PubMed
    • Google Scholar
43. 1. Kubatzky KF
    2. Uhle F
    3. Eigenbrod T

    (2018) From macrophage to osteoclast - How metabolism determines function and activity

    Cytokine 112:102–115.

    https://doi.org/10.1016/j.cyto.2018.06.013

    • PubMed
    • Google Scholar
44. 1. Levy JE
    2. Jin O
    3. Fujiwara Y
    4. Kuo F
    5. Andrews NC

    (1999) Transferrin receptor is necessary for development of erythrocytes and the nervous system

    Nature Genetics 21:396–399.

    https://doi.org/10.1038/7727

    • PubMed
    • Google Scholar
45. 1. Li B
    2. Lee WC
    3. Song C
    4. Ye L
    5. Abel ED
    6. Long F

    (2020a) Both aerobic glycolysis and mitochondrial respiration are required for osteoclast differentiation

    The FASEB Journal 34:11058–11067.

    https://doi.org/10.1096/fj.202000771R

    • Google Scholar
46. 1. Li J
    2. Pan X
    3. Pan G
    4. Song Z
    5. He Y
    6. Zhang S
    7. Ye X
    8. Yang X
    9. Xie E
    10. Wang X
    11. Mai X
    12. Yin X
    13. Tang B
    14. Shu X
    15. Chen P
    16. Dai X
    17. Tian Y
    18. Yao L
    19. Han M
    20. Xu G
    21. Zhang H
    22. Sun J
    23. Chen H
    24. Wang F
    25. Min J
    26. Xie L

    (2020b) Transferrin receptor 1 regulates thermogenic capacity and cell fate in brown/beige adipocytes

    Advanced Science 7:1903366.

    https://doi.org/10.1002/advs.201903366

    • PubMed
    • Google Scholar
47. 1. Matak P
    2. Matak A
    3. Moustafa S
    4. Aryal DK
    5. Benner EJ
    6. Wetsel W
    7. Andrews NC

    (2016) Disrupted iron homeostasis causes dopaminergic neurodegeneration in mice

    PNAS 113:3428–3435.

    https://doi.org/10.1073/pnas.1519473113

    • PubMed
    • Google Scholar
48. 1. Mbalaviele G
    2. Novack DV
    3. Schett G
    4. Teitelbaum SL

    (2017) Inflammatory osteolysis: a conspiracy against bone

    The Journal of Clinical Investigation 127:2030–2039.

    https://doi.org/10.1172/JCI93356

    • PubMed
    • Google Scholar
49. 1. Menon P
    2. Yin G
    3. Smolock EM
    4. Zuscik MJ
    5. Yan C
    6. Berk BC

    (2010) GPCR kinase 2 interacting protein 1 (GIT1) regulates osteoclast function and bone mass

    Journal of Cellular Physiology 225:777–785.

    https://doi.org/10.1002/jcp.22282

    • PubMed
    • Google Scholar
50. 1. Nakamura T
    2. Imai Y
    3. Matsumoto T
    4. Sato S
    5. Takeuchi K
    6. Igarashi K
    7. Harada Y
    8. Azuma Y
    9. Krust A
    10. Yamamoto Y
    11. Nishina H
    12. Takeda S
    13. Takayanagi H
    14. Metzger D
    15. Kanno J
    16. Takaoka K
    17. Martin TJ
    18. Chambon P
    19. Kato S

    (2007) Estrogen prevents bone loss via estrogen receptor alpha and induction of Fas ligand in osteoclasts

    Cell 130:811–823.

    https://doi.org/10.1016/j.cell.2007.07.025

    • PubMed
    • Google Scholar
51. 1. Nakano S
    2. Inoue K
    3. Xu C
    4. Deng Z
    5. Syrovatkina V
    6. Vitone G
    7. Zhao L
    8. Huang XY
    9. Zhao B

    (2019) G-protein Gα13 functions as a cytoskeletal and mitochondrial regulator to restrain osteoclast function

    Scientific Reports 9:e40974-z.

    https://doi.org/10.1038/s41598-019-40974-z

    • PubMed
    • Google Scholar
52. 1. Nakashima T
    2. Hayashi M
    3. Takayanagi H

    (2012) New insights into osteoclastogenic signaling mechanisms

    Trends in Endocrinology and Metabolism 23:582–590.

    https://doi.org/10.1016/j.tem.2012.05.005

    • PubMed
    • Google Scholar
53. 1. Novack DV
    2. Teitelbaum SL

    (2008) The osteoclast: Friend or foe?

    Annual Review of Pathology 3:457–484.

    https://doi.org/10.1146/annurev.pathmechdis.3.121806.151431

    • PubMed
    • Google Scholar
54. 1. Pantopoulos K
    2. Porwal SK
    3. Tartakoff A
    4. Devireddy L

    (2012) Mechanisms of mammalian iron homeostasis

    Biochemistry 51:5705–5724.

    https://doi.org/10.1021/bi300752r

    • PubMed
    • Google Scholar
55. 1. Rauner M
    2. Baschant U
    3. Roetto A
    4. Pellegrino RM
    5. Rother S
    6. Salbach-Hirsch J
    7. Weidner H
    8. Hintze V
    9. Campbell G
    10. Petzold A
    11. Lemaitre R
    12. Henry I
    13. Bellido T
    14. Theurl I
    15. Altamura S
    16. Colucci S
    17. Muckenthaler MU
    18. Schett G
    19. Komla-Ebri DSK
    20. Bassett JHD
    21. Williams GR
    22. Platzbecker U
    23. Hofbauer LC

    (2019) Transferrin receptor 2 controls bone mass and pathological bone formation via BMP and Wnt signalling

    Nature Metabolism 1:111–124.

    https://doi.org/10.1038/s42255-018-0005-8

    • PubMed
    • Google Scholar
56. 1. Rishi G
    2. Secondes ES
    3. Wallace DF
    4. Subramaniam VN

    (2016) Normal systemic iron homeostasis in mice with macrophage-specific deletion of transferrin receptor 2

    American Journal of Physiology. Gastrointestinal and Liver Physiology 310:G171–G180.

    https://doi.org/10.1152/ajpgi.00291.2015

    • PubMed
    • Google Scholar
57. 1. Rottner K
    2. Stradal TEB
    3. Chen B

    (2021) WAVE regulatory complex

    Current Biology 31:R512–R517.

    https://doi.org/10.1016/j.cub.2021.01.086

    • PubMed
    • Google Scholar
58. 1. Ruiz P
    2. Martin-Millan M
    3. Gonzalez-Martin MC
    4. Almeida M
    5. González-Macias J
    6. Ros MA

    (2016) CathepsinKCre mediated deletion of βcatenin results in dramatic loss of bone mass by targeting both osteoclasts and osteoblastic cells

    Scientific Reports 6:36201.

    https://doi.org/10.1038/srep36201

    • PubMed
    • Google Scholar
59. 1. Sato T
    2. Iwata T
    3. Usui M
    4. Kokabu S
    5. Sugamori Y
    6. Takaku Y
    7. Kobayashi T
    8. Ito K
    9. Matsumoto M
    10. Takeda S
    11. Xu R
    12. Chida D

    (2020) Bone phenotype in melanocortin 2 receptor-deficient mice

    Bone Reports 13:e100713.

    https://doi.org/10.1016/j.bonr.2020.100713

    • PubMed
    • Google Scholar
60. 1. Senyilmaz D
    2. Virtue S
    3. Xu X
    4. Tan CY
    5. Griffin JL
    6. Miller AK
    7. Vidal-Puig A
    8. Teleman AA

    (2015) Regulation of mitochondrial morphology and function by stearoylation of TFR1

    Nature 525:124–128.

    https://doi.org/10.1038/nature14601

    • PubMed
    • Google Scholar
61. 1. Siddiqui JA
    2. Partridge NC

    (2016) Physiological bone remodeling: Systemic regulation and growth factor involvement

    Physiology 31:233–245.

    https://doi.org/10.1152/physiol.00061.2014

    • PubMed
    • Google Scholar
62. 1. Srinivasan S
    2. Koenigstein A
    3. Joseph J
    4. Sun L
    5. Kalyanaraman B
    6. Zaidi M
    7. Avadhani NG

    (2010) Role of mitochondrial reactive oxygen species in osteoclast differentiation

    Annals of the New York Academy of Sciences 1192:245–252.

    https://doi.org/10.1111/j.1749-6632.2009.05377.x

    • PubMed
    • Google Scholar
63. 1. Teitelbaum SL

    (2011) The osteoclast and its unique cytoskeleton

    Annals of the New York Academy of Sciences 1240:14–17.

    https://doi.org/10.1111/j.1749-6632.2011.06283.x

    • PubMed
    • Google Scholar
64. 1. Tsay J
    2. Yang Z
    3. Ross FP
    4. Cunningham-Rundles S
    5. Lin H
    6. Coleman R
    7. Mayer-Kuckuk P
    8. Doty SB
    9. Grady RW
    10. Giardina PJ
    11. Boskey AL
    12. Vogiatzi MG

    (2010) Bone loss caused by iron overload in a murine model: importance of oxidative stress

    Blood 116:2582–2589.

    https://doi.org/10.1182/blood-2009-12-260083

    • PubMed
    • Google Scholar
65. 1. Väänänen HK
    2. Zhao H
    3. Mulari M
    4. Halleen JM

    (2000) The cell biology of osteoclast function

    Journal of Cell Science 113:377–381.

    https://doi.org/10.1242/jcs.113.3.377

    • PubMed
    • Google Scholar
66. 1. Walker DG

    (1975) Control of bone resorption by hematopoietic tissue. The induction and reversal of congenital osteopetrosis in mice through use of bone marrow and splenic transplants

    The Journal of Experimental Medicine 142:651–663.

    https://doi.org/10.1084/jem.142.3.651

    • PubMed
    • Google Scholar
67. 1. Wang L
    2. Fang B
    3. Fujiwara T
    4. Krager K
    5. Gorantla A
    6. Li C
    7. Feng JQ
    8. Jennings ML
    9. Zhou J
    10. Aykin-Burns N
    11. Zhao H

    (2018) Deletion of ferroportin in murine myeloid cells increases iron accumulation and stimulates osteoclastogenesis in vitro and in vivo

    The Journal of Biological Chemistry 293:9248–9264.

    https://doi.org/10.1074/jbc.RA117.000834

    • PubMed
    • Google Scholar
68. 1. Wang S
    2. He X
    3. Wu Q
    4. Jiang L
    5. Chen L
    6. Yu Y
    7. Zhang P
    8. Huang X
    9. Wang J
    10. Ju Z
    11. Min J
    12. Wang F

    (2020) Transferrin receptor 1-mediated iron uptake plays an essential role in hematopoiesis

    Haematologica 105:2071–2082.

    https://doi.org/10.3324/haematol.2019.224899

    • PubMed
    • Google Scholar
69. 1. Xiao W
    2. Beibei F
    3. Guangsi S
    4. Yu J
    5. Wen Z
    6. Xi H
    7. Youjia X

    (2015) Iron overload increases osteoclastogenesis and aggravates the effects of ovariectomy on bone mass

    The Journal of Endocrinology 226:121–134.

    https://doi.org/10.1530/JOE-14-0657

    • PubMed
    • Google Scholar
70. 1. Xiao Y
    2. Palomero J
    3. Grabowska J
    4. Wang L
    5. de Rink I
    6. van Helvert L
    7. Borst J

    (2017) Macrophages and osteoclasts stem from a bipotent progenitor downstream of a macrophage/osteoclast/dendritic cell progenitor

    Blood Advances 1:1993–2006.

    https://doi.org/10.1182/bloodadvances.2017008540

    • PubMed
    • Google Scholar
71. 1. Xu W
    2. Barrientos T
    3. Andrews NC

    (2013) Iron and copper in mitochondrial diseases

    Cell Metabolism 17:319–328.

    https://doi.org/10.1016/j.cmet.2013.02.004

    • PubMed
    • Google Scholar
72. 1. Xu W
    2. Barrientos T
    3. Mao L
    4. Rockman HA
    5. Sauve AA
    6. Andrews NC

    (2015) Lethal cardiomyopathy in mice lacking transferrin receptor in the heart

    Cell Reports 13:533–545.

    https://doi.org/10.1016/j.celrep.2015.09.023

    • PubMed
    • Google Scholar
73. 1. Yang Q
    2. Jian J
    3. Katz S
    4. Abramson SB
    5. Huang X

    (2012) 17β-Estradiol inhibits iron hormone hepcidin through an estrogen responsive element half-site

    Endocrinology 153:3170–3178.

    https://doi.org/10.1210/en.2011-2045

    • PubMed
    • Google Scholar
74. 1. Yang W
    2. Wang J
    3. Moore DC
    4. Liang H
    5. Dooner M
    6. Wu Q
    7. Terek R
    8. Chen Q
    9. Ehrlich MG
    10. Quesenberry PJ
    11. Neel BG

    (2013) Ptpn11 deletion in a novel progenitor causes metachondromatosis by inducing hedgehog signalling

    Nature 499:491–495.

    https://doi.org/10.1038/nature12396

    • PubMed
    • Google Scholar
75. 1. Zaidi M

    (2007) Skeletal remodeling in health and disease

    Nature Medicine 13:791–801.

    https://doi.org/10.1038/nm1593

    • PubMed
    • Google Scholar
76. 1. Zhang Y
    2. Rohatgi N
    3. Veis DJ
    4. Schilling J
    5. Teitelbaum SL
    6. Zou W

    (2018) PGC1β organizes the osteoclast cytoskeleton by mitochondrial biogenesis and activation

    Journal of Bone and Mineral Research 33:1114–1125.

    https://doi.org/10.1002/jbmr.3398

    • Google Scholar
77. 1. Zhao H

    (2012) Membrane trafficking in osteoblasts and osteoclasts: new avenues for understanding and treating skeletal diseases

    Traffic 13:1307–1314.

    https://doi.org/10.1111/j.1600-0854.2012.01395.x

    • PubMed
    • Google Scholar
78. 1. Zhou J
    2. Ye S
    3. Fujiwara T
    4. Manolagas SC
    5. Zhao H

    (2013) Steap4 plays a critical role in osteoclastogenesis in vitro by regulating cellular iron/reactive oxygen species (ROS) levels and cAMP response element-binding protein (CREB) activation

    The Journal of Biological Chemistry 288:30064–30074.

    https://doi.org/10.1074/jbc.M113.478750

    • PubMed
    • Google Scholar

Decision letter after peer review:

Thank you for submitting your article "Transferrin receptor 1-mediated iron uptake regulates bone mass in mice via osteoclast mitochondria and cytoskeleton" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen Mone Zaidi as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Nathan J Pavlos (Reviewer #2).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

The reviewers found significant merit in the work. However, the evaluation identified sufficient weaknesses that prevent publication of this work in the present form. The main shortcomings of the presented work include the relatively underdeveloped in vitro data which does not convincingly support the conclusions; the concern that the mouse models result in decreased Ctsk expression which may along at least in part explain the phenotype; the potential of evaluation in an ovarectomized mice and / or mice at different ages to extend the translational potential and underlying mechanism of these findings, respectively; and some degree of evaluation of systemic iron metabolism and erythropoiesis. Specifically, the in vitro data can be extended to more fully explore osteoclastogenesis and in vivo data by experimentally addressing the observed gender disparity and specificity of effect in the trabecular region of long bones.

Reviewer #1 (Recommendations for the authors):

Some additional questions and clarifications are delineated below.

1. It is unclear why only or mostly female osteoclast-selective TFR1-null mice exhibit a more robust phenotype, with increased bone mass in long bones. The authors discuss this point in the Discussion section but the proposed experiment, to evaluate whether hepcidin expression is altered in the liver or hepcidin concentration changes in the serum and IHC using anti-FPN1 antibody in osteoclasts or at least cellular iron concentration would be useful to corroborate the hypothesis of estrogen-mediated changes in the hepcidin:ferroportin axis. Do osteoclasts even express FPN1 to make such a hypothesis viable? There is some data on testosterone and hepcidin regulation (Latour Hepatology 2014) that may be worth mentioning or discussing in the setting of the authors hypothesis about differential gender-related results. Finally, are the differences between male and female mice significant in Figure 2? Are these differences in the expected range for male and female mice?

2. It is similarly unclear why only long bones but not vertebrae reveal these differences. Can the authors speculate and discuss this point in the Discussion section?

3. The authors suggest that they have effectively excluded an effect from TFR2 on osteoclast function. However, this is inaccurate and needs to be communicated differently. At best, the authors have confirmed what has long been understood that TFR2 serves as an iron sensor rather than importer of iron. This should be restated in all mention of TFR2 in the manuscript. Furthermore, Rauner et al. published an important manuscript looking at the role of TFR2 in bone, demonstrating that TFR2 competes with BMPR for binding BMP2 and TFR2:BMP binding is enhanced in the presence of Fe-TF, together suggesting that presence of iron may favor bone resorption in the presence of TFR2. Something akin to this would benefit the readership and should be included in the manuscript as it pertains to the data on TFR1 here.

4. Figure 1, Y axis on mRNA expression panels is unclear what the control is. Typically, expectation is of change from baseline (1) in control. Minor: panel for Flvcr2 is missing from legend.

5. Figure 1 B, what mechanism leads to some cellular iron in TFR1 fl/LysM Cre OC cells? This may be the reason that the results on differentiation are not seen in the current work relative to Ishii et al. 2009. Please explain and include in the Discussion.

6. Figure 3 A-E is redundant and can be moved to the Supplement. In general, the data on LysM-cre mice is distracting and does not provide substantive contributing information to the main point. Consider moving all LysM-cre data to the supplement to better focus the manuscript.

7. Figure 3G, the difference in female mice is driven by 3 mice with elevated values; are these 3 mice in any way different from the others? Are they outliers?

8. Figure 3 states that there is no impact on osteoclast number and bone resorption in mice but while BFR is unchanged in Figure 3J, serum P1NP is increased. How do the authors reconcile this confusing result?

9. Given the number of figures, it would be worthwhile to consolidate and reorganize to move some additional elements into the supplementary figures. For example, there is no reason in the mind of the reviewer to include both male and female samples in vitro. Especially since there are no visible differences, please move one gender to the supplement.

10. Figure 6 title suggests that there are no differences in markers of differentiation, e.g. Nfatc1 and Ctsk, but both of these appear to be decreased in f/f;c/c samples, both female but especially male (especially since tubulin appears greater in those samples). Please provide the quantification here.

11. Figue 12B and 12C do not include control cells with and without Hem1 overexpression. This is important since the mechanism may be completely different and yet additive to yield the observed results. Please edit.

12. The Introduction is too long and unfocused. Please edit.

13. The first half of the Discussion is a restating of the results and needs to be re-written to further extend the interpretation of the results without restating them.

Reviewer #2 (Recommendations for the authors):

This is a well-written study that examines the role of transferrin receptor 1 in osteoclast function and bone remodelling. The importance of iron homeostasis in bone turnover has gained momentum in recent years, most notably the role of transferrin receptor 2 in osteoblasts (Rauner et al., Nat Med 2019), however the physiological significance of iron and the mechanism(s) responsible for its uptake in osteoclasts remains poorly understood. As such, the present study detailing the physiological role of Tfr1 in osteoclasts would likely be of considerable interest to the field, complementing existing studies in osteoblasts, but also reconciling the physiological contribution of cellular iron handling on the other side of the bone remodelling equation. The strengths of this study lie the multiple osteoclast-specific mouse lines generated to interrogate the role of Tfr1 in both osteoclast-lineage cells and mature osteoclasts along with the suite of complementary cell biological, biochemical and proteomic analyses used to dissect its contribution in osteoclasts. Nonetheless, there remain some weaknesses, particularly concerning the expressivity of the bone phenotype and mechanistic link between Tfr1 and the cytoskeleton that should be considered to strengthen the overall resolve of the conclusions drawn.

I have the following comments:

1. The studies conducted on LysoM-Cre mice appear to be well-executed and carefully controlled. The skeletal phenotype is however complex, being restricted to trabeculae of the long bones of female mice, but not the axial skeleton (e.g. vertebrae) at 10-weeks of age. The authors rationalise that the unexpected sexual dimorphism is associated with an intersection between estrogen and iron export, which is reasonable, but they do not offer experimental evidence to support this. Further, the skeletal phenotyping is limited to a single time-point i.e. 10-weeks of age. As such, it would be beneficial if the authors could provide data in older animals to show whether the bone phenotype becomes more pronounced or resolves with age. It would also be useful for the reader if the authors could offer some explanation for the observed site-specific differences in the Discussion e.g. potential differences in osteoclast heterogeneity etc.

2. Although widely employed as a Cre-transgenic mouse line to conditionally delete target genes in mature osteoclasts there are some caveats that should be considered to avoid over-interpretation of data arising from the cathespin K (Ctsk)-Cre lines. Foremost, this cre-line is known to result in the loss one allele of Ctsk owing to partial replacement the Ctsk locus with Cre in the targeting cassette (Dallas et a., Genetics 2018). Thus, endogenous Ctsk expression is reduced. Considering that deficiency and mutations in Ctsk lead to high bone mass phenotype (Saftig et al., 1997) together with findings that Cstk K-Cre can also lead to unexpected germline deletion of genes (e.g. Winkeler et al., PLos 2012), any data arising from these mice need to be interpreted with a degree of caution. The authors have rightly attempted to control for this using littermates mice on heterozygous KI (Ctsk-Cre/+), which is appropriate although these mice are on a mice mixed background C57BL6/129 which can complicate skeletal phenotypes. Overall, these underlying complexities may account, in part, for the discrepancies observed between the LysoM and Ctsk-cre males and females, with Ctsk-Cre leading to a more pronounced phenotype. As a minimum, these limitations need to be acknowledged and discussed. A minor oversight- for Figure 4B is that the mice sexes are unlabelled in the reconstructed uCT scans.

3. The in vitro data concerning the potential effects of Tfr1-deficiency on osteoclast different and mitochondrial metabolism are extensive and overall support the conclusion drawn. On the other hand the mechanism linking Tfr1 regulation of the osteoclast cytoskeleton remains largely descriptive and thus is comparatively underdeveloped. While the quantitative proteomics data goes some way to lend support to the notion that loss of Tfr1 leads to a global dysregulation of downstream Tfr1-associated protein pathways including those involved with cytoskeletal signalling/organization, the extent and diversity of proteins affected in osteoclasts >2000 proteins is remarkable. Can the authors please clarify how they arrived at the 1.2-fold threshold for these data. Does the 1.2-fold represent an arbitrary cut-off or does it reflect >1.2 Log2Fold, p<0.05? To validate these proteomic data the authors hone in on hem1 (a component of the Wave regulatory complex, WRC) and show that it partly restores the cytoskeletal spreading defect in Tfr1-deficient osteoclasts. This finding is interesting but would be bolstered by confirmation of a disruption of additional components of the WRC or downstream components such as Arf2/3 which has an established role in OC cytoskeletal organization in osteoclasts (Hurst et al., JBMR 2009). Considering the well-established roles of B3-integrin, c-SRC and Pyk2 in osteoclasts, it would be valuable to corroborate/or discount their direct involvement in these spreading and cytoskeletal defects by independently assessing their expression levels immunoanalytically in Tfr1-deficient osteoclasts.

Reviewer #3 (Recommendations for the authors):

1. The authors state that Tfr1 does not inhibit osteoclast differentiation, but just spreading and resorption. I do not agree with this statement. By "spreading", do the authors mean "fusion"? I believe in any case that this is part of the differentiation process to a fully mature osteoclast. Also, even though it is stated that osteoclasts have been counted as TRAP-positive and with more than 3 nuclei, I know from experience that it is tricky to identify the nuclei just using TRAP staining, especially in such small cells as presented in Suppl. Figure 9A. I believe it may be more accurate to assess the number of osteoclasts in this case with so many small ones using IF and DAPI staining for the nuclei. Using that, the authors could also count how many osteoclasts have e.g. 3-10 or 10-50 nuclei (giving insights into the fusion process). Finally, I believe more markers for osteoclasts could be used to address this point of fusion/spreading, such as DC-STAMP, SWAP70, or CTSK. In Suppl. Figure 9, e.g., it looks like CtsK expression is reduced in the Tfr1;Lysm-cre mice, although the WB is a bit over-exposed, so it is hard to tell… quantification would be helpful.

2. It appears that the appendicular skeleton is more affected than the axial skeleton. Did the authors check the expression of Tfr1 in e.g. femoral vs. vertebral bone using immunohistochemistry? Is there a difference in expression pattern that may explain the altered susceptibility of the sites to Tfr1 deletion? Similarly, it would be interesting to check Tfr1 in trabecular vs. cortical bone – as those also seem to be differentially affected.

3. Along those lines, is the expression of Tfr1 in osteoclasts (bone) lower in young animals compared to adult ones? Could that explain the more severe effects of iron deficiency in adult mice compared to the 3-week-old ones?

4. The increase in BV/TV in female 10-week-old mice is quite strong (especially in the Ctsk-cre mice). Did the authors detect extramedullary hematopoiesis? Are those mice iron deficient also in the liver? How is their systemic iron levels?

5. The lack of osteoclast function during Tfr1-deficiency would be beneficial in a more clinical context in e.g. ameliorating bone loss via ovariectomy. I think performing such an experiment (with the Tfr1;Ctsk-cre mice) would add more clinical benefit to the paper.

6. How do the authors explain that Tfr1 deficiency in osteoclast precursors has less effects than using the Ctsk-cre mice? I would assume that using the Lysm-cre, all osteoclasts – also the later stages – are affected, whereas in the Ctsk-cre, only the mature ones are. Thus, I am not sure why the Lysm-cre don´t also show a stronger effect.

7. Figure 4: to me, it actually looks like there are fewer osteoclasts in the cKO mice… the number of mice is not that high for histology, which tends to have high standard deviations. Did the authors also check the BFR in those mice? Was it also a bit increased, like in the Lysm-cre mice?

8. A WB should be shown validating the reduction of OXPHOS complex I-V expression. Similarly, it would be good to show a verification of other key proteins involved in cytoskeletal reorganization that were identified in the proteomic studies.

9. It is not clear to me how Mitotracker/Mitosox MFI and mitochondrial membrane potential were quantified? Why are there so many dots? How many cells and/or mitochondria were quantified? Was the MFI of each individual cell measured? Please add more details.

10. The link between Tfr1 and cytoskeleton organization is not quite clear. Is that connection e.g. iron-dependent and/or dependent on non-canonical functions of Tfr1? If iron-dependent, how does iron affect the regulation of WRC/Arp2/3 complex?

11. I personally think that Suppl. Figure 1 should be included into the main figures as it shows the first mechanistic part. I would rather suggest to potentially combine Figure 1+3 and/or 2+4 (or put parts of them into the supplement).

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Transferrin receptor 1-mediated iron uptake regulates bone mass in mice via osteoclast mitochondria and cytoskeleton" for further consideration by eLife. Your revised article has been evaluated by Mone Zaidi (Senior Editor) and a Reviewing Editor.

While the authors have added a substantive amount of work to significantly improve this manuscript, there are some remaining issues that need to be addressed, still a few places that need further explanation, a few pieces of data that do not evidently communicate what the authors claim:

1) While Ctsk mRNA expression is increased in Tfr1f/f Cre+ mice, protein concentration is decreased. This is unclear and presumptively, the focus is on the protein concentration. It is unclear what this means in total but the authors state that this indicates that Tfr1 deletion in osteoclasts has minimum effect on osteoclast differentiation. This needs to be clarified further in the results and Discussion sections.

2) Tfr1 loss appears to increase the number of osteoclasts with fewer nuclei and decreased fusion. The authors state that this impacts osteoclast function (decreased) rather than number or differentiation. This is fine as they show good evidence of decreased osteoclast functional endpoints. However, later in the manuscript, osteoblast data is also alluded to but not presented, noting only increased serum P1NP and Runx2 expression (Figure 3 – suppl Figure 2) and a comment suggesting that the phenotype in Tfr1f/f;Lyz2-Cre mice therefore maybe the consequence of altered osteoblast function. The response to Reviewer 1, point #8 is thus confusing and requires further clarification. Perhaps the authors are stating that the altered osteoclast function leads to an altered ability of the osteoclasts to suppress osteoblasts? This should be made clear in the discussion at least as speculation if the data cannot be shown fully. Response to Reviewer 2 point #4 is confusing with circular logic along the same lines.

3) Figure 6D needs a quantification. The authors state that Tfr1 loss results in decreased energy metabolism but there is no obvious evidence of this in osteoclasts in this figure (maybe BMM are bit more decreased in Tfr1-/- cells). Therefore, it is erroneous to state that is is restored by hemin as there is no evidence of a defect or visually appreciable differences relative to control in any of the wells. Please clarify.

4) While OVX Tfr1f/f;LysMCre mice exhibit a persistent difference from OVX control, the authors state that the phenotype observed in non-OVX mice is estrogen dependent (page 9, line 227). This is conceptually inaccurate in my understanding as if it were estrogen dependent, the difference would be abrogated after OVX which it is not. It in fact supports the opposite conclusion, that the phenotype is NOT estrogen dependent. Please clarify in discussion.

5) Figure 9C and 9D demonstrate the effect of Hem1 overexpression. The authors note "rescue" but this is only partial and a more appropriate conclusion would be that other mechanisms of iron delivery can reverse, at least partially, the effect of iron restriction induced by Tfr1 loss.

6) Unclear if there is a difference between 10uM hemin in control vs. Tfr1-/- cells in Figure 5A, 5C, and 5D. A preserved difference would suggest that there is incomplete rescue even at the highest hemin dose because osteoclasts may not have every option for iron import available in other cells. This is important to expand in the Discussion section.

7) Figure 9B requires quantification to corroborate the assertions made by the authors.

Essential revisions:

The reviewers found significant merit in the work. However, the evaluation identified sufficient weaknesses that prevent publication of this work in the present form. The main shortcomings of the presented work include:

1) The relatively underdeveloped in vitro data which does not convincingly support the conclusions.

To further support our conclusion that loss of Tfr1 in osteoclast lineage cells has little effects on differentiation but attenuates cytoskeleton organization and inhibits activation/function of mature osteoclasts, we performed the following extended in vitro experiments. The new data are included in new Figure 4, Figure 6, and Figure 9.

i. qPCR detection of mRNA expression of osteoclast marker genes, including Acp5 (encoding TRAP), Ctsk (encoding Cathepsin K), Nfatc1 (encoding osteoclast master transcription factor Nfatc1), and Dcstamp (encoding osteoclast fusion regulator DC-Stamp), in control and Tfr1^ΔlysM bone marrow monocytes, pre-osteoclasts, and mature osteoclasts. As shown in Figure 4B, the mRNA levels of Acp5, NFATc1, and dcstamp in Tfr1^ΔlysM osteoclast precursor and mature cells were similar to those of respective control cells. The mRNA expression of Ctsk in Tfr1^ΔlysM pre-osteoclasts and mature osteoclasts was slightly higher than that in control cells. These results indicate that deletion of Tfr1 in osteoclast myeloid precursors has minimum effects on osteoclast differentiation as indicated in Figure 4A.

ii. We repeated western blotting for Nfatc1 and Cathepsin K and quantified the specific bands of these osteoclast markers by densitometry using NIH Image J software. As demonstrated in the right panels of Figure 4C and Figure 4—figure supplement 1B, deletion of Tfr1 had no effects on the protein level of Nfatc1 in osteoclast precursor and mature cells, whereas the level of Cathepsin K in Tfr1^ΔlysM pre- and mature-osteoclasts was slightly decreased compared to control and Tfr1-f/f;c/+ osteoclasts.

iii. We counted the number of osteoclasts with 3-10, 10-30, and >30 nuclei, respectively. As shown in Figure 4D, the number of osteoclasts with 3-10 nuclei in Tfr1^ΔlysM culture was higher than control and there was no difference in the number of osteoclasts with 10-30 nuclei between control and Tfr1^ΔlysM cultures. Only the number of osteoclasts with > 30 nuclei in Tfr1^ΔlysM culture was lower than control although the mRNA expression of osteoclast fusion gene Dcstamp was similar in control and Tfr1^ΔlysM osteoclasts. The exact mechanism(s) of decreased fusion in the late stage of osteoclastogenesis in Tfr1^ΔlysM culture need further investigation in future.

iv. To consolidate the finding from quantitative proteomic results that cellular iron regulates the components of mitochondrial respiration chain, we did immuno-blotting to detect the changes of mitochondrial respiratory complexes in control and Tfr1^ΔlysM monocytes and mature osteoclasts treated with either vehicle or 10μM hemin using a cocktail of antibodies recognizing the representative components of mitochondrial respiratory complex I-V. As shown in Figure 6D, there was an obvious decrease in the mitochondrial complex II and III in Tfr1^ΔlysM monocytes and hemin stimulated the mitochondrial complex I and II in control and Tfr1^ΔlysM osteoclasts, indicating that Tfr1-mediated iron uptake plays an important role in mitochondrial oxidative phosphorylation pathway.

v. To identify the mechanisms by which Tfr1 regulates osteoclast WRC complex and cytoskeleton, we examined the activation (phosphorylation) of protein tyrosine kinase c-Src which plays a pivotal role in osteoclast cytoskeleton organization downstream of integrin and M-CSF as well as RANKL signaling pathways in control and Tfr1^ΔlysM osteoclasts. We also assayed the activation of small GTPases Rac1 and Arf1 that are upstream stimulators of WAVE-regulating complex (WRC) in responses to M-CSF and RANKL in control and Tfr1^ΔlysM osteoclasts using GST-pulldown kits from Thermo-Fisher Scientific Inc. As shown in Figure 9B, we found that the phosphorylation of c-Src and the GTP-bound (active form) of Rac1, but not GTP-bound Arf1, stimulated by M-CSF were declined in Tfr1^ΔlysM osteoclasts. In contrast, Tfr1-deficiency had little effects on RANKL-activated c-Src, Rac1, and Arf1 in osteoclasts.

vi. We repeated the Hem1 overexpression experiment with proper controls suggested by the reviewers. As demonstrated in Figure 9C and 9D, overexpression of HA-tagged Hem1 enhanced the number of activated osteoclasts in control culture and rescued the cytoskeletal defect in Tfr1^ΔlysM osteoclasts.

2) The concern that the mouse models result in decreased Ctsk expression which may along at least in part explain the phenotype;

As shown in Figure 4B and 4C, although the mRNA expression of Ctsk in Tfr1^ΔlysM pre-osteoclasts and mature osteoclasts detected by qPCR was slightly higher than that in control cells, the quantification of western blots demonstrated that the levels of Ctsk protein in Tfr1^ΔlysM pre- and mature-osteoclasts were slightly decreased compared to control cells. This decreased Ctsk expression might contribute attenuated bone resorption in Tfr1^ΔlysM osteoclasts. It should be pointed out that Ctsk heterozygous mice had no defects in osteoclast function (Gowen, et al., JBMR 1999). The phenotypes of osteoclasts in vitro and skeletal homeostasis in Tfr1^ΔlysM mice are mainly resulted from the decreased energy metabolism and defect in cytoskeleton organization in Tfr1^ΔlysM osteoclasts.

3) The potential of evaluation in an ovarectomized mice and / or mice at different ages to extend the translational potential and underlying mechanism of these findings, respectively.

In addition to 10 weeks old post-pubertal mice, we have examined the skeletal phenotypes of 3-weeks-old pre-pubertal and 6 months old sham and ovariectomized control and Tfr1-flox;LysM-Cre mice by μCT. These data are presented in Figure 2 and Figure 2—figure supplement 4-5. The results are incorporated in the third paragraph on page 8-9. The possible mechanisms are discussed in the second paragraphs on pages 20-21.

4) Some degree of evaluation of systemic iron metabolism and erythropoiesis.

We have measured the serum levels of total iron and iron-regulating hormone hepcidin. In addition, we have counted the number of red blood cells in peripheral blood from both male and female control and Tfr1-flox;LysM-Cre mice. These data are presented in the Figure 2—figure supplement 3. The results have been added in the second paragraph on page 8, which demonstrate that loss of Tfr1 in myeloid lineage cells has no effects on systemic iron homeostasis and erythropoiesis.

5) Specifically, the in vitro data can be extended to more fully explore osteoclastogenesis and in vivo data by experimentally addressing the observed gender disparity and specificity of effect in the trabecular region of long bones.

See specific point-by point responses below to individual reviewer’s critiques related to this issue.

Reviewer #1 (Recommendations for the authors):

Some additional questions and clarifications are delineated below.

1. It is unclear why only or mostly female osteoclast-selective TFR1-null mice exhibit a more robust phenotype, with increased bone mass in long bones. The authors discuss this point in the Discussion section but the proposed experiment, to evaluate whether hepcidin expression is altered in the liver or hepcidin concentration changes in the serum and IHC using anti-FPN1 antibody in osteoclasts or at least cellular iron concentration would be useful to corroborate the hypothesis of estrogen-mediated changes in the hepcidin:ferroportin axis. Do osteoclasts even express FPN1 to make such a hypothesis viable? There is some data on testosterone and hepcidin regulation (Latour Hepatology 2014) that may be worth mentioning or discussing in the setting of the authors hypothesis about differential gender-related results. Finally, are the differences between male and female mice significant in Figure 2? Are these differences in the expected range for male and female mice?

The serum levels of iron and hepcidin in 10-week-old male and female control and Tfrc-flox; Lyz2-Cre mice are similar (see Figure 2—figure supplement 3). It is well known that the trabecular bone mass of male mice is higher than their age-matched female mice in the adults. Since Tfr1-deficiency results in significant increase in trabecular bone mass solely in female but not male mice, there is no difference of bone mass between female Tfrc-flox;Lyz2-Cre mice and male Tfrc-flox;Lyz2-Cre as well as male control mice. This phenomenon has been added on page 9.

2. It is similarly unclear why only long bones but not vertebrae reveal these differences. Can the authors speculate and discuss this point in the Discussion section?

Mouse genetic studies have clearly established that the genetic regulation of peak bone mass development is skeletal site dependent. Thus, it is not surprising that Tfr1 deficiency exerts different effects on the skeletal phenotype of long bones versus vertebra. These differences are discussed in the context of what is known in the literature on osteoclast heterogeneity and genetic regulation of bone development at these two sites on pages 22-23. Our qPCR result demonstrated that deletion efficiency of Tfr1 gene, Tfrc, in vertebra is less potent than femurs (Figure 2—figure supplement 9).

3. The authors suggest that they have effectively excluded an effect from TFR2 on osteoclast function. However, this is inaccurate and needs to be communicated differently. At best, the authors have confirmed what has long been understood that TFR2 serves as an iron sensor rather than importer of iron. This should be restated in all mention of TFR2 in the manuscript. Furthermore, Rauner et al. published an important manuscript looking at the role of TFR2 in bone, demonstrating that TFR2 competes with BMPR for binding BMP2 and TFR2:BMP binding is enhanced in the presence of Fe-TF, together suggesting that presence of iron may favor bone resorption in the presence of TFR2. Something akin to this would benefit the readership and should be included in the manuscript as it pertains to the data on TFR1 here.

We agree. Although the results in Figure 1B indicate that Tfr1-mediated iron uptake is a major route for iron acquisition in osteoclasts, we did not have direct evidence regarding the effects of Tfr2 on osteoclast iron uptake and function. The statement of “The role of Tfr2 in osteoclast iron uptake and function needs to be further elucidated in the future” has been added in the second paragraph on page 7.

4. Figure 1, Y axis on mRNA expression panels is unclear what the control is. Typically, expectation is of change from baseline (1) in control. Minor: panel for Flvcr2 is missing from legend.

The information has been added in Figure 1A and its legend.

5. Figure 1 B, what mechanism leads to some cellular iron in TFR1 fl/LysM Cre OC cells? This may be the reason that the results on differentiation are not seen in the current work relative to Ishii et al. 2009. Please explain and include in the Discussion.

The residual cellular iron in Tfr1-deficient osteoclasts may be obtained through uptake of heme because the heme transporter Hcp1 is up-regulated during osteoclast differentiation and the fact that hemin but not ferric ammonium citrate (FAC, which is absorbed through NTBI) is able to rescue the Tfr1-deficient osteoclast phenotypes (Figure 5). While Ishii et al. reported that knockdown of Tfr1 expression in bone marrow monocytes by short-hairpin RNA (shRNA) inhibits osteoclast differentiation and transferrin promotes osteoclastogenesis in vitro, the treatment of mice with iron chelator has no effects on osteoclast number in vivo. Iron chelator inhibits iron effect totally and not specifically to Tfr1. The use of mice with conditional deletion of Tfr1 gene specifically in osteoclasts to investigate the role of Tfr1 mediated iron transport function in osteoclasts has not been done previously. The information has been added on page 23.

6. Figure 3 A-E is redundant and can be moved to the Supplement. In general, the data on LysM-cre mice is distracting and does not provide substantive contributing information to the main point. Consider moving all LysM-cre data to the supplement to better focus the manuscript.

Please see our response to General Comment #1, both the Lyz2-Cre and Ctsk-Cre models have advantages and disadvantages. Since Lyz2-Cre is relatively more specific than Ctsk-Cre in gene-deletion in osteoclast lineage cells (see the reference in Yang W et al., Nature, 2013), we moved Tfrc-flox;Ctsk-Cre results to the Figure 2—figure supplement 6-7 and Figure 3—figure supplement 1/3.

7. Figure 3G, the difference in female mice is driven by 3 mice with elevated values; are these 3 mice in any way different from the others? Are they outliers?

Since the other histological parameters of these three mice (See the Figure 3 source data file) are not quite different from other mice, we think that the data in Figure 3G reflects the high variation in in vivo data. We decide to keep the data of these three mice.

8. Figure 3 states that there is no impact on osteoclast number and bone resorption in mice but while BFR is unchanged in Figure 3J, serum P1NP is increased. How do the authors reconcile this confusing result?

The calculation of BFR in Osteomeasure software is normalized to bone surface (BS). Since there is a significant increase in trabecular bone mass (also in BS) in Tfrc-flox;Lyz2-Cre female mice, the total level of P1NP (procollagen type 1 N-terminal propeptide, a specific marker of type 1 collagen deposition during bone formation) is up in these mice. This is supported by increased mRNA level of osteoblast transcription factor Runx2 in Tfrc-flox;Lyz2-Cre female mice (Figure 3—figure supplement 2). See also the reply in #14 of Reviewer 3.

9. Given the number of figures, it would be worthwhile to consolidate and reorganize to move some additional elements into the supplementary figures. For example, there is no reason in the mind of the reviewer to include both male and female samples in vitro. Especially since there are no visible differences, please move one gender to the supplement.

We moved all the male in vitro and in vivo results to the Supplemental Figures. The male osteoclasts have similar phenotypes to female cells in vitro. Thus, the differences of in vivo phenotypes between male and females are resulted from the distinct effects of sex hormones on osteoclasts in vivo.

10. Figure 6 title suggests that there are no differences in markers of differentiation, e.g. Nfatc1 and Ctsk, but both of these appear to be decreased in f/f;c/c samples, both female but especially male (especially since tubulin appears greater in those samples). Please provide the quantification here.

During the revision of this manuscript, we examined the mRNA expression of osteoclast marker genes by qPCR in control and Tfr1^ΔlysM osteoclast lineage cells. As shown in Figure 4B, the mRNA levels of Acp5, Nfatc1, and dcstamp in Tfr1^ΔlysM osteoclast precursor and mature cells were similar to those of respective control cells. The mRNA expression of Ctsk in Tfr1^ΔlysM pre-osteoclasts and mature osteoclasts was even higher than that in control cells. We also quantified the western blots of Nfatc1 and Ctsk. Deletion of Tfr1 in both male and female mice had no effects on the protein level of Nfatc1 in osteoclast precursor and mature cells, whereas the level of Ctsk in Tfr1^ΔlysM pre- and mature-osteoclasts was slightly decreased compared to control and Tfr1-f/f;c/+ osteoclasts (the right panels in Figure 4C and Figure 4—figure supplement 1B).

11. Figure 12B and 12C do not include control cells with and without Hem1 overexpression. This is important since the mechanism may be completely different and yet additive to yield the observed results. Please edit.

Please see our response to General Comment #4.

12. The Introduction is too long and unfocused. Please edit.

The Introduction has been shortened.

13. The first half of the Discussion is a restating of the results and needs to be re-written to further extend the interpretation of the results without restating them.

We have revised the discussion to reduce redundancy with the Results section.

Reviewer #2 (Recommendations for the authors):

This is a well-written study that examines the role of transferrin receptor 1 in osteoclast function and bone remodelling. The importance of iron homeostasis in bone turnover has gained momentum in recent years, most notably the role of transferrin receptor 2 in osteoblasts (Rauner et al., Nat Med 2019), however the physiological significance of iron and the mechanism(s) responsible for its uptake in osteoclasts remains poorly understood. As such, the present study detailing the physiological role of Tfr1 in osteoclasts would likely be of considerable interest to the field, complementing existing studies in osteoblasts, but also reconciling the physiological contribution of cellular iron handling on the other side of the bone remodelling equation. The strengths of this study lie the multiple osteoclast-specific mouse lines generated to interrogate the role of Tfr1 in both osteoclast-lineage cells and mature osteoclasts along with the suite of complementary cell biological, biochemical and proteomic analyses used to dissect its contribution in osteoclasts. Nonetheless, there remain some weaknesses, particularly concerning the expressivity of the bone phenotype and mechanistic link between Tfr1 and the cytoskeleton that should be considered to strengthen the overall resolve of the conclusions drawn.

I have the following comments:

1. The studies conducted on LysoM-Cre mice appear to be well-executed and carefully controlled. The skeletal phenotype is however complex, being restricted to trabeculae of the long bones of female mice, but not the axial skeleton (e.g. vertebrae) at 10-weeks of age. The authors rationalise that the unexpected sexual dimorphism is associated with an intersection between estrogen and iron export, which is reasonable, but they do not offer experimental evidence to support this. Further, the skeletal phenotyping is limited to a single time-point i.e. 10-weeks of age. As such, it would be beneficial if the authors could provide data in older animals to show whether the bone phenotype becomes more pronounced or resolves with age. It would also be useful for the reader if the authors could offer some explanation for the observed site-specific differences in the Discussion e.g. potential differences in osteoclast heterogeneity etc.

See the reply to #3 in Essential Revisions and the results in new Figure 2 and supplemental Figures 5-6. The explanations for the different phenotypes in trabecular verse cortical bone compartments and in long bone versus vertebra are discussed in the third paragraph on page 22.

2. Although widely employed as a Cre-transgenic mouse line to conditionally delete target genes in mature osteoclasts there are some caveats that should be considered to avoid over-interpretation of data arising from the cathespin K (Ctsk)-Cre lines. Foremost, this cre-line is known to result in the loss one allele of Ctsk owing to partial replacement the Ctsk locus with Cre in the targeting cassette (Dallas et a., Genetics 2018). Thus, endogenous Ctsk expression is reduced. Considering that deficiency and mutations in Ctsk lead to high bone mass phenotype (Saftig et al., 1997) together with findings that Cstk K-Cre can also lead to unexpected germline deletion of genes (e.g. Winkeler et al., PLos 2012), any data arising from these mice need to be interpreted with a degree of caution. The authors have rightly attempted to control for this using littermates mice on heterozygous KI (Ctsk-Cre/+), which is appropriate although these mice are on a mice mixed background C57BL6/129 which can complicate skeletal phenotypes. Overall, these underlying complexities may account, in part, for the discrepancies observed between the LysoM and Ctsk-cre males and females, with Ctsk-Cre leading to a more pronounced phenotype. As a minimum, these limitations need to be acknowledged and discussed. A minor oversight- for Figure 4B is that the mice sexes are unlabelled in the reconstructed uCT scans.

The advantages and disadvantages of Lyz2(LysM)-Cre and Ctsk-Cre models and the possible explanations for the different phenotypes between the Lyz2-Cre and Ctsk-Cre males and females have been added in and discussed in the second paragraph on page 22. See also the reply to the #1 of General Comments of Reviewer 1.

3. The in vitro data concerning the potential effects of Tfr1-deficiency on osteoclast different and mitochondrial metabolism are extensive and overall support the conclusion drawn. On the other hand the mechanism linking Tfr1 regulation of the osteoclast cytoskeleton remains largely descriptive and thus is comparatively underdeveloped. While the quantitative proteomics data goes some way to lend support to the notion that loss of Tfr1 leads to a global dysregulation of downstream Tfr1-associated protein pathways including those involved with cytoskeletal signalling/organization, the extent and diversity of proteins affected in osteoclasts >2000 proteins is remarkable. Can the authors please clarify how they arrived at the 1.2-fold threshold for these data. Does the 1.2-fold represent an arbitrary cut-off or does it reflect >1.2 Log2Fold, p<0.05? To validate these proteomic data the authors hone in on hem1 (a component of the Wave regulatory complex, WRC) and show that it partly restores the cytoskeletal spreading defect in Tfr1-deficient osteoclasts. This finding is interesting but would be bolstered by confirmation of a disruption of additional components of the WRC or downstream components such as Arf2/3 which has an established role in OC cytoskeletal organization in osteoclasts (Hurst et al., JBMR 2009). Considering the well-established roles of B3-integrin, c-SRC and Pyk2 in osteoclasts, it would be valuable to corroborate/or discount their direct involvement in these spreading and cytoskeletal defects by independently assessing their expression levels immunoanalytically in Tfr1-deficient osteoclasts.

The advantages and disadvantages of Lyz2(LysM)-Cre and Ctsk-Cre models and the possible explanations for the different phenotypes between the Lyz2-Cre and Ctsk-Cre males and females have been added in and discussed in the second paragraph on page 22. See also the reply to the #1 of General Comments of Reviewer 1.

Reviewer #3 (Recommendations for the authors):

1. The authors state that Tfr1 does not inhibit osteoclast differentiation, but just spreading and resorption. I do not agree with this statement. By "spreading", do the authors mean "fusion"? I believe in any case that this is part of the differentiation process to a fully mature osteoclast. Also, even though it is stated that osteoclasts have been counted as TRAP-positive and with more than 3 nuclei, I know from experience that it is tricky to identify the nuclei just using TRAP staining, especially in such small cells as presented in Suppl. Figure 9A. I believe it may be more accurate to assess the number of osteoclasts in this case with so many small ones using IF and DAPI staining for the nuclei. Using that, the authors could also count how many osteoclasts have e.g. 3-10 or 10-50 nuclei (giving insights into the fusion process). Finally, I believe more markers for osteoclasts could be used to address this point of fusion/spreading, such as DC-STAMP, SWAP70, or CTSK. In Suppl. Figure 9, e.g., it looks like CtsK expression is reduced in the Tfr1;Lysm-cre mice, although the WB is a bit over-exposed, so it is hard to tell… quantification would be helpful.

The word “spreading” is often used to describe the cellular process in which the morphology of multinucleated osteoclasts changes from the irregular-shaped status characterized by scattered distribution of podosomes to a round-shaped status featured by the peripherally localized podosome-belt when osteoclasts are cultured on glass coverslips or plastic culture dishes. Thus, the spreading process in osteoclasts is regulated by cytoskeleton organization. The defects in osteoclast spreading are associated with decreased osteoclast activation and bone resorption. The reviewer’s point is correct that the “fusion” is a part of osteoclast differentiation process where mononuclear osteoclast precursor cells are fused to form multinucleated osteoclasts. This process is regulated by osteoclast fusion genes including DC-Stamp induced by RANKL and Nfatc1.

In response to the reviewer suggestion, we have now quantitatively measured the mRNA expression of osteoclast marker genes in control and Tfr1^ΔlysM osteoclast precursor and mature cells. As shown in Figure 4B, the mRNA levels of Acp5, Nfatc1, and Dcstamp in Tfr1^ΔlysM osteoclast precursor and mature cells were similar to those of respective control cells. The mRNA expression of Ctsk in Tfr1^ΔlysM pre-osteoclasts and mature osteoclasts was slightly higher than that in control cells. These results indicate that deletion of Tfr1 in osteoclast myeloid precursors has minimum effects on osteoclast differentiation as we observed in Figure 4A (see also the reply in #i of Essential Revisions #1).

Per the reviewer’s suggestion, we counted the number of osteoclasts with 3-10, 10-30, and >30 nuclei, respectively. As shown in Figure 4D, the number of osteoclasts with 3-10 nuclei in Tfr1^ΔlysM culture was higher than control and there was no difference in the number of osteoclasts with 10-30 nuclei between control and Tfr1^ΔlysM cultures. Only the number of osteoclasts with > 30 nuclei in Tfr1^ΔlysM culture was lower than control although the mRNA expression of osteoclast fusion gene Dcstamp was similar in control and Tfr1^ΔlysM osteoclasts. The exact mechanism(s) of decreased fusion in the late stage of osteoclastogenesis in Tfr1^ΔlysM culture need further investigation in future.

The suppl Figure 9 was in our previous but not in this submission. Instead, we repeated western blotting for Nfatc1 and Ctsk and quantified the specific bands of these osteoclast markers by densitometry using NIH Image J software. As demonstrated in the right panels of Figure 4C and Figure 4—figure supplement 1B, deletion of Tfr1 had no effects on the protein level of Nfatc1 in male and female osteoclast precursor and mature cells, whereas the level of Ctsk in Tfr1^ΔlysM pre- and mature-osteoclasts was slightly decreased compared to control and Tfr1-f/f;c/+ osteoclasts in both gender. This slight decrease in Ctsk protein level in Tfr1^ΔlysM osteoclasts might be caused by the fewer number of large osteoclasts with > 30 nuclei in Tfr1^ΔlysM culture than in control.

2. It appears that the appendicular skeleton is more affected than the axial skeleton. Did the authors check the expression of Tfr1 in e.g. femoral vs. vertebral bone using immunohistochemistry? Is there a difference in expression pattern that may explain the altered susceptibility of the sites to Tfr1 deletion? Similarly, it would be interesting to check Tfr1 in trabecular vs. cortical bone – as those also seem to be differentially affected.

As discussed on page 22, the exact explanation(s) for distinct effects of Tfr1-deficiency on the appendicular vs axial bones and the trabecular vs cortical bones are not available. Unfortunately, the anti-Tfr1 antibody we used in Figure 4A does not work for immunohistochemistry. By qPCR, we demonstrated that the deletion efficiency of Tfr1 gene, Tfrc, in vertebra is less than femurs (Figure 2—figure supplement 9).

3. Along those lines, is the expression of Tfr1 in osteoclasts (bone) lower in young animals compared to adult ones? Could that explain the more severe effects of iron deficiency in adult mice compared to the 3-week-old ones?

As mentioned above, we could not answer this question because the anti-Tfr1 antibody does not work for immunohistochemistry on EDTA-decalcified, paraffin-embedded bone tissue sections.

4. The increase in BV/TV in female 10-week-old mice is quite strong (especially in the Ctsk-cre mice). Did the authors detect extramedullary hematopoiesis? Are those mice iron deficient also in the liver? How is their systemic iron levels?

See the reply for #4 in Essential Revisions. Loss of Tfr1 in osteoclasts has no effects on systemic iron homeostasis and erythropoiesis as detected by serum iron and peripheral blood RBC count. There is no splenomegaly in Tfr1^ΔLysM and Tfr1^ΔCtsk mice (data not shown). Serum hepcidin, a liver-derived iron-regulating hormone, levels are similar between control and Tfr1-deficient mice, thus suggesting that there is no iron deficiency in liver. However, we have not measured liver iron level in mice with disruption of Tfr1 in osteoclasts.

5. The lack of osteoclast function during Tfr1-deficiency would be beneficial in a more clinical context in e.g. ameliorating bone loss via ovariectomy. I think performing such an experiment (with the Tfr1;Ctsk-cre mice) would add more clinical benefit to the paper.

When the lab moved to California in 2018, we discontinued the Tfrc-flox;Ctsk-Cre strain. We performed OVX on 4.5-monthold control and Tfr1^ΔLysM female mice. After 6-week post OVX, the mice were sacrificed, and trabecular bone mass and cortical bone thickness were measured by μCT. As shown in Figure 2, estrogen deficiency resulted in decreased trabecular bone in both female control and Tfr1^ΔlysM mice, indicating that Tfr1 in osteoclasts does not influence bone loss in OVX mice. This finding is discussed on page 20-21.

6. How do the authors explain that Tfr1 deficiency in osteoclast precursors has less effects than using the Ctsk-cre mice? I would assume that using the Lysm-cre, all osteoclasts – also the later stages – are affected, whereas in the Ctsk-cre, only the mature ones are. Thus, I am not sure why the Lysm-cre don´t also show a stronger effect.

The distinct effects of Tfr1-deficiency by Lyz2-Cre and Ctsk-Cre may be due to the different in vivo efficiency of these two Cre lines and may also be caused by the off-target effects of the Cre drivers as mentioned in the reply for the #1 critic in General Comments of Reviewer 1.

7. Figure 4: to me, it actually looks like there are fewer osteoclasts in the cKO mice… the number of mice is not that high for histology, which tends to have high standard deviations. Did the authors also check the BFR in those mice? Was it also a bit increased, like in the Lysm-cre mice?

The number of osteoclasts in control and Tfr1^ΔCtsk mice is the same. The staining in cKO mice was faint by unknown reason. We only did histology in decalcified, paraffin-embedded bone sections for TRAP staining. We did not do tetracycline-labeling in these mice.

8. A WB should be shown validating the reduction of OXPHOS complex I-V expression. Similarly, it would be good to show a verification of other key proteins involved in cytoskeletal reorganization that were identified in the proteomic studies.

See Figure 6B and the reply in #iv of Essential Revisions above. We had difficulties in finding antibodies with high specificity and affinity against the components of WRC complex by western blotting. Instead, we found that the M-CSF stimulated activation of c-Src and Rac1, which are upstream regulators of WRC complex, were decreased in Tfr1^ΔlysM osteoclasts (Figure 9B). See the reply in #v of Essential Revisions above and in #10 below.

9. It is not clear to me how Mitotracker/Mitosox MFI and mitochondrial membrane potential were quantified? Why are there so many dots? How many cells and/or mitochondria were quantified? Was the MFI of each individual cell measured? Please add more details.

Please see “Mitochondrial mass, ROS production, and membrane potential measurements” in the Materials and methods section for detailed methodology. The following information has been added (See grey highlights on page 32).

The epifluorescent images of 10-15 randomly chosen areas per group (control or knockout) were taken in a blinded fashion. Each area had between 7-15 cells. The images were analyzed with Image J software also in a blinded fashion. The fluorescent intensities of cells were individually measured, and the results were presented as mean fluorescent intensity (MFI) per cell in arbitrary units.

10. The link between Tfr1 and cytoskeleton organization is not quite clear. Is that connection e.g. iron-dependent and/or dependent on non-canonical functions of Tfr1? If iron-dependent, how does iron affect the regulation of WRC/Arp2/3 complex?

As shown in Figure 5C and 5D, 10μM hemin (ferric chloride heme), which provides iron through the alternative heme uptake in the absence of Tfr1, rescued the cytoskeleton organization defects (podosome-belt formation in osteoclasts cultured on glass coverslips and the actin-ring formation in osteoclasts cultured on bovine cortical bone slices, see also Figure 4D and 4E) in Tfr1^ΔlysM osteoclasts. These results indicate that Tfr1-mediated cellular iron homeostasis regulates osteoclast cytoskeleton organization although we don’t have evidence whether the non-canonical functions of Tfr1 participates this process at the moment. To identify the mechanisms by which Tfr1-mediated iron uptake regulates WRC/Arp2/3 complex, we turned to its upstream activators, the small GTPases Rac1/Arf1 and the protein tyrosine kinase c-Src which plays a central role in activating Rac1 and osteoclast cytoskeleton stimulated by the integrins, M-CSF as well as RANKL. As shown in Figure 9B, we found that the phosphorylation (activation) of c-Src and the GTP-bound (active form) of Rac1 but not GTP-bound Arf1 stimulated by M-CSF were declined in Tfr1^ΔlysM osteoclasts. In contrast, Tfr1-deficiency had little effects on RANKL-activated c-Src, Rac1, and Arf1 in osteoclasts.

11. I personally think that Suppl. Figure 1 should be included into the main figures as it shows the first mechanistic part. I would rather suggest to potentially combine Figure 1+3 and/or 2+4 (or put parts of them into the supplement).

The Supplemental Figure 1 has been changed to Figure 1. The previous Figure 1 and Figure 2 are now Figure 2 and Figure 3, respectively. We moved previous Figure 3 and Figure 4 into the supplemental Figures. They are now Figure 2—figure supplement 6 and Figure 3—figure supplement 3, respectively.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

While the authors have added a substantive amount of work to significantly improve this manuscript, there are some remaining issues that need to be addressed, still a few places that need further explanation, a few pieces of data that do not evidently communicate what the authors claim:

1) While Ctsk mRNA expression is increased in Tfr1f/f Cre+ mice, protein concentration is decreased. This is unclear and presumptively, the focus is on the protein concentration. It is unclear what this means in total but the authors state that this indicates that Tfr1 deletion in osteoclasts has minimum effect on osteoclast differentiation. This needs to be clarified further in the results and Discussion sections.

We apologize for lack of clarity on the Ctsk mRNA expression and protein data. Cathepsin k is a well-accepted marker of osteoclast activity. Thus, our data on the reduced CTSK protein levels in Tfr1^ΔLysM mice suggest reduced osteoclast activity. The result and conclusion have been revised on page 12 and page 13.

2) Tfr1 loss appears to increase the number of osteoclasts with fewer nuclei and decreased fusion. The authors state that this impacts osteoclast function (decreased) rather than number or differentiation. This is fine as they show good evidence of decreased osteoclast functional endpoints. However, later in the manuscript, osteoblast data is also alluded to but not presented, noting only increased serum P1NP and Runx2 expression (Figure 3 – suppl Figure 2) and a comment suggesting that the phenotype in Tfr1f/f;Lyz2-Cre mice therefore maybe the consequence of altered osteoblast function. The response to Reviewer 1, point #8 is thus confusing and requires further clarification. Perhaps the authors are stating that the altered osteoclast function leads to an altered ability of the osteoclasts to suppress osteoblasts? This should be made clear in the discussion at least as speculation if the data cannot be shown fully. Response to Reviewer 2 point #4 is confusing with circular logic along the same lines.

We apologize for not clearly stating the osteoclast-osteoblast connection to the observed skeletal phenotype. Loss of Tfr1 expression in osteoclasts lead to decreased osteoclast function as shown by reduced CTSK protein levels and osteoclast functional data. Based on what is known about osteoclast-mediated coupling of bone formation, one would predict that the reduced osteoclast activity in Tfr1f/f Cre+ mice should reduce bone formation. By contrast, the increased serum PINP levels and Runx2 expression in these mice suggest increased bone formation. Thus, we don’t have an explanation at present if the increased bone formation is a direct consequence of inhibition of osteoclast activity or by some other mechanisms. Further studies are needed to address this issue.

3) Figure 6D needs a quantification. The authors state that Tfr1 loss results in decreased energy metabolism but there is no obvious evidence of this in osteoclasts in this figure (maybe BMM are bit more decreased in Tfr1-/- cells). Therefore, it is erroneous to state that is is restored by hemin as there is no evidence of a defect or visually appreciable differences relative to control in any of the wells. Please clarify.

Quantitative data of Figure 6D is added as Figure 6—figure supplement 1. The results are revised based on this quantification. Although Figure 6D provides complementary evidence for proteomic data (shown in Figure 6B and 6C) by simultaneously demonstrating the changes of mitochondrial respiratory complexes in vehicle and hemin treated control and Tfr1^ΔLysM bone marrow monocytes and osteoclasts, it should point out that the antibody cocktail used in this immunoblotting contains antibodies against only one component of each of five mitochondrial respiratory complexes as indicated in Figure 6D. Thus, the result of Figure 6D could not reflect complete changes of mitochondria in Tfr1-deficient osteoclast lineage cells. This limitation of the assay has been added on page 16. We agree that Figure 6 does not provide strong evidence to support the conclusion that loss of Tfr1 in osteoclasts attenuates energy metabolism. This conclusion is removed to Figure 8.

4) While OVX Tfr1f/f;LysMCre mice exhibit a persistent difference from OVX control, the authors state that the phenotype observed in non-OVX mice is estrogen dependent (page 9, line 227). This is conceptually inaccurate in my understanding as if it were estrogen dependent, the difference would be abrogated after OVX which it is not. It in fact supports the opposite conclusion, that the phenotype is NOT estrogen dependent. Please clarify in discussion.

We agree that the OVX data in Figure 2 could not fully support the conclusion that the high trabecular bone mass phenotype in Tfr1^ΔLysM mice is estrogen dependent though showing obvious gender-dependent difference. We revised the results and conclusion in the contest of what we know about estrogen regulation of bone modeling and remodeling during development and in adulthood in the Results and Discussion sections on page 9 and page 23, respectively. It is now well established that estrogen-mediated mechanisms play a key role in mediating bone accretion during pubertal growth period. During this developmental period, estrogen is known to exert an anabolic effect on osteoblast besides its well-established role in inhibiting osteoclast functions. However, the loss of estrogen as seen during menopause or due to ovariectomy in adults lead to bone loss primarily due to increased bone resorption. As shown in Figure 2, trabecular bone volume of Tfr1^ΔLysM mice was increased at the post-pubertal age of 10 weeks but was not different at the prepubertal age of 3 weeks, thus suggesting that the skeletal phenotype gets manifested during pubertal and post-pubertal growth periods. To determine if loss of Tfr1 in osteoclasts affects OVX-induced bone loss, control and Tfr1^ΔLysM mice at 4.5-month of age were ovariectomized to mimic human postmenopausal osteoporosis and skeletal phenotype was determined at 6 months of age. While OVX induced bone loss in Tfr1^ΔLysM mice, the trabecular bone mass of Tfr1^ΔLysM mice was significantly higher than control mice after OVX. While these data suggest that loss of Tfr1 does not affect estrogen deficiency-induced bone loss, the issue of whether the skeletal phenotype in 10-week-old Tfr1^ΔLysM mice is dependent on pubertal increase in estrogen remains to be evaluated by performing OVX at prepubertal age of 3 weeks in future.

5) Figure 9C and 9D demonstrate the effect of Hem1 overexpression. The authors note "rescue" but this is only partial and a more appropriate conclusion would be that other mechanisms of iron delivery can reverse, at least partially, the effect of iron restriction induced by Tfr1 loss.

We agree. We have revised the conclusion to indicate that the rescue is partial and suggest that the other component(s) of the WRC complex or distinct pathways might also be involved in the Tfr1-mediated regulation of osteoclast cytoskeleton in the first paragraph on page 20.

6) Unclear if there is a difference between 10uM hemin in control vs. Tfr1-/- cells in Figure 5A, 5C, and 5D. A preserved difference would suggest that there is incomplete rescue even at the highest hemin dose because osteoclasts may not have every option for iron import available in other cells. This is important to expand in the Discussion section.

We agree. We have added the statistical analysis results of 10uM hemin-treated control vs. Tfr1^ΔLysM cultures in Figure 5A, 5C, and 5D. We have revised the conclusion to indicate that the rescue is partial and there are limited alternative iron acquisition pathways in osteoclasts as in other cells in the first paragraph on page 15.

7) Figure 9B requires quantification to corroborate the assertions made by the authors.

Quantitative data is added as Figure 9—figure supplement 1, and the results and conclusion are revised based on the quantitation in the first and second paragraphs on page 19.

Author details

1. Bhaba K Das

   Southern California Institute for Research and Education, Long Beach, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing - original draft

   Contributed equally with

   Lei Wang and Toshifumi Fujiwara

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0256-5489

2. Lei Wang

   1. Department of Orthopedics, The Third People’s Hospital of Hefei, Third Clinical College, Anhui Medical University, Hefei, China
   2. Center for Osteoporosis and Metabolic Bone Diseases, Division of Endocrinology, Department of Internal Medicine, University of Arkansas for Medical Sciences, Little Rock, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing - original draft

   Contributed equally with

   Bhaba K Das and Toshifumi Fujiwara

   Competing interests

   No competing interests declared

3. Toshifumi Fujiwara

   1. Center for Osteoporosis and Metabolic Bone Diseases, Division of Endocrinology, Department of Internal Medicine, University of Arkansas for Medical Sciences, Little Rock, United States
   2. Department of Orthopedic Surgery, Kyushu University Hospital, Fukuoka, Japan

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing - original draft

   Contributed equally with

   Bhaba K Das and Lei Wang

   Competing interests

   No competing interests declared

4. Jian Zhou

   Department of Orthopedics, First Affiliated Hospital, Anhui Medical University, Hefei, China

   Contribution

   Data curation, Formal analysis, Investigation, Validation

   Competing interests

   No competing interests declared

5. Nukhet Aykin-Burns

   Division of Radiation Health, Department of Pharmaceutical Sciences, University of Arkansas for Medical Sciences, Little Rock, United States

   Contribution

   Data curation, Formal analysis, Investigation, Methodology, Visualization

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-8574-4102

6. Kimberly J Krager

   Division of Radiation Health, Department of Pharmaceutical Sciences, University of Arkansas for Medical Sciences, Little Rock, United States

   Contribution

   Data curation, Formal analysis, Investigation

   Competing interests

   No competing interests declared

7. Renny Lan

   Department of Pediatrics, University of Arkansas for Medical Sciences, Little Rock, United States

   Contribution

   Data curation, Formal analysis, Investigation, Methodology, Visualization

   Competing interests

   No competing interests declared

8. Samuel G Mackintosh

   Department of Biochemistry and Molecular Biology, University of Arkansas for Medical Sciences, Little Rock, United States

   Contribution

   Data curation, Formal analysis, Investigation, Methodology, Validation

   Competing interests

   No competing interests declared

9. Ricky Edmondson

   Department of Biochemistry and Molecular Biology, University of Arkansas for Medical Sciences, Little Rock, United States

   Contribution

   Data curation, Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

10. Michael L Jennings

    Department of Physiology and Cell Biology, University of Arkansas for Medical Sciences, Little Rock, United States

    Contribution

    Data curation, Formal analysis, Investigation, Methodology, Writing - review and editing

    Competing interests

    No competing interests declared

11. Xiaofang Wang

    Department of Biomedical Sciences, Texas A&M University, Dallas, United States

    Contribution

    Data curation, Formal analysis, Investigation, Visualization

    Competing interests

    No competing interests declared

12. Jian Q Feng

    Department of Biomedical Sciences, Texas A&M University, Dallas, United States

    Contribution

    Data curation, Formal analysis, Investigation, Validation, Writing - review and editing

    Competing interests

    No competing interests declared

13. Tomasa Barrientos

    Department of Orthopedics, Duke University, Durham, United States

    Contribution

    Data curation, Formal analysis, Investigation, Validation

    Competing interests

    No competing interests declared

14. Jyoti Gogoi

    Southern California Institute for Research and Education, Long Beach, United States

    Contribution

    Data curation, Formal analysis, Investigation

    Competing interests

    No competing interests declared

15. Aarthi Kannan

    1. Southern California Institute for Research and Education, Long Beach, United States
    2. Division of Dermatology, Department of medicine, Long Beach VA Healthcare System, Long Beach, United States

    Contribution

    Data curation, Formal analysis, Investigation, Project administration

    Competing interests

    No competing interests declared

16. Ling Gao

    1. Southern California Institute for Research and Education, Long Beach, United States
    2. Division of Dermatology, Department of medicine, Long Beach VA Healthcare System, Long Beach, United States

    Contribution

    Conceptualization, Data curation, Formal analysis, Investigation, Validation, Writing - review and editing

    Competing interests

    No competing interests declared

17. Weirong Xing

    Musculoskeletal Disease Center, VA Loma Linda Healthcare System, Loma Linda, United States

    Contribution

    Data curation, Formal analysis, Investigation, Methodology, Validation, Writing - review and editing

    Competing interests

    No competing interests declared

18. Subburaman Mohan

    Musculoskeletal Disease Center, VA Loma Linda Healthcare System, Loma Linda, United States

    Contribution

    Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing - original draft, Writing - review and editing

    For correspondence

    Subburaman.Mohan@va.gov

    Competing interests

    Reviewing editor, eLife

    "This ORCID iD identifies the author of this article:" 0000-0003-0063-986X

19. Haibo Zhao

    1. Southern California Institute for Research and Education, Long Beach, United States
    2. Center for Osteoporosis and Metabolic Bone Diseases, Division of Endocrinology, Department of Internal Medicine, University of Arkansas for Medical Sciences, Little Rock, United States
    3. Department of Physiology and Cell Biology, University of Arkansas for Medical Sciences, Little Rock, United States

    Contribution

    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Visualization, Writing - original draft, Writing - review and editing

    For correspondence

    Haibo.zhao@va.gov

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0003-0836-7555

• 1,442

  Page views

• 492

  Downloads

• 8

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.