ARTICLE DOI: 10.1038/s41467-018-04047-5 OPEN Oxidation-speciﬁc epitopes restrain bone formation 1 2 2 2 1 Elena Ambrogini , Xuchu Que , Shuling Wang , Fumihiro Yamaguchi , Robert S. Weinstein , 3 1 2 1 Sotirios Tsimikas , Stavros C. Manolagas , Joseph L. Witztum & Robert L. Jilka Atherosclerosis and osteoporosis are epidemiologically linked and oxidation speciﬁc epitopes (OSEs), such as phosphocholine (PC) of oxidized phospholipids (PC-OxPL) and malondialdehyde (MDA), are pathogenic in both. The proatherogenic effects of OSEs are opposed by innate immune antibodies. Here we show that high-fat diet (HFD)-induced bone loss is attenuated in mice expressing a single chain variable region fragment of the IgM E06 (E06-scFv) that neutralizes PC-OxPL, by increasing osteoblast number and stimulating bone formation. Similarly, HFD-induced bone loss is attenuated in mice expressing IK17-scFv, which neutralizes MDA. Notably, E06-scFv also increases bone mass in mice fed a normal diet. Moreover, the levels of anti-PC IgM decrease in aged mice. We conclude that OSEs, whether produced chronically or increased by HFD, restrain bone formation, and that diminished defense against OSEs may contribute to age-related bone loss. Anti-OSEs, therefore, may represent a novel therapeutic approach against osteoporosis and atherosclerosis simultaneously. Division of Endocrinology and Metabolism, Center for Osteoporosis and Metabolic Bone Diseases, University of Arkansas for Medical Sciences and Central Arkansas Veterans Healthcare System, 4301W. Markham, Little Rock, AR 72205, USA. Division of Endocrinology and Metabolism, University of California San Diego, La Jolla, CA 92093-0682, USA. Department of Medicine, Cardiololgy, University of California San Diego, 9500 GilmanDrive, La Jolla, CA 92093-0682, USA. Correspondence and requests for materials should be addressed to E.A. (email: email@example.com) NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications 1 | | | 1234567890():,; ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 pidemiologic studies have established a strong correlation activate additional macrophages and lymphocytes in a pathologic between atherosclerosis and osteoporosis, suggesting that a cascade that eventually leads to the development of atheroma, 15–17 Ecommon pathogenic mechanism may underlie both con- vascular inﬂammation and calciﬁcation . Consistent with 1–4 ditions . In a meta-analysis of twenty-ﬁve studies involving their anti-OSE properties, administration of either E06 IgM or more than 10,000 subjects, atherosclerotic lesions were IK17-scFv—a single chain variant of the antigen binding domain signiﬁcantly increased in patients with osteopenia and of the human anti-MDA antibody (IK17)—prevents the osteoporosis independently of age, sex, body mass index, and formation of foam cells and atheroma in murine models of 5 15,22–24 other cardiovascular risk factors . Consistent with the human atherosclerosis . Moreover, genetic maneuvers that increase data, mouse models of hyperlipidemia/atherosclerosis exhibit overall levels of anti-OSE antibodies in mice attenuate 25,26 decreased bone mass, primarily due to a reduction in osteoblast atherosclerosis . 6–10 number and bone formation . In bone, SRs and TLRs are present on osteoblasts, osteoclasts, Oxidation of polyunsaturated fatty acids (PUFAs) by reactive and bone marrow macrophages, raising the possibility that oxygen species (ROS) is a feature of many physiologic and physiologic or pathologic generation of OSEs could affect bone 27,28 pathologic processes, including apoptosis, cellular senescence, and homeostasis . Previous studies have demonstrated in murine inﬂammation , as well as the bone loss that accompanies models of atherosclerosis that OxLDL stimulates a 9,10 10 atherosclerosis in rodents and humans . Lipid peroxidation pro-inﬂammatory response of bone and loss of bone mass . generates highly reactive degradation products such as Additionally, OxLDL attenuates osteoblast generation and 9,29–33 malondialdehyde (MDA), 4-hydroxynonenal (4-HNE), and promotes osteoblast apoptosis in vitro . Because of this oxidized phospholipids (OxPL), such as oxidized phosphati- evidence, we investigated the role of OSEs in bone homeostasis. dylcholine. These moieties react with amino groups on proteins To do this, we used mice overexpressing the antigen recognition and other lipids to form adducts that are collectively known as portion of E06 (E06-scFv), or IK17-scFv, which block the effects oxidation-speciﬁc epitopes (OSEs) . OSEs are members of a of PC-OxPL or MDA, respectively. We found that the E06-scFv larger group of proinﬂammatory and immunogenic molecules— transgene increases cancellous bone mass and attenuates the loss known collectively as damage-associated molecular patterns of cortical bone mass caused by high fat diet (HFD) in the LDL- (DAMPs)—that are produced by excessive ROS, necrosis, ische- receptor knockout (LDLR-KO) model of atherosclerosis. An 11–13 mia reperfusion, or chemically induced tissue injury . increase in cancellous, as well as cortical, bone mass is observed DAMPs often share structural homology with moieties present on also in HFD-fed LDLR-KO mice expressing the IK17-scFv microbes, which are collectively known as pathogen-associated transgene. Unexpectedly, the bone anabolic properties of E06- molecular patterns (PAMPs). PAMPs and DAMPs bind to scFv on cancellous bone is present not only in the setting of evolutionary conserved pattern recognition receptors (PRRs) . enforced atherogenesis, but also in C57BL/6 J mice maintained Binding of PAMPs to PRRs activates mechanisms that kill on a normal diet. Lastly, consistent with evidence in humans that 34–37 invading microbes. Binding of DAMPS to PRRs sequesters them B-1 cells decline with age and low levels of anti-PC IgM and also mounts defenses that attempt to prevent cell damage. antibodies are associated with increased incidence of cardiovas- The PRRs are either cell bound, such as the large family of cular diseases , we show that sera from 22- to 26-month-old scavenger receptors (SRs) and toll-like receptors (TLRs), or female C57BL/6 J mice have dramatically decreased levels of anti- soluble, such as the natural antibodies (NAbs) produced by B-1 PC IgM as compared to 6- or 7-month-old mice. These ﬁndings 14–18 lymphocytes . NAbs are predominantly of the IgM class and indicate that OSEs, produced chronically and in response to HFD, comprise about 80% of the total serum IgM in uninfected mice . restrain bone formation in mice; and that the age-related bone Their antigen binding sites are generated by rearrangement of loss might be due in part to diminished defense against OSEs. germline-encoded variable region genes in the complete absence Anti-OSEs, therefore, may represent a novel therapeutic approach of foreign antigen exposure. NAbs constitute the ﬁrst line of to the prevention and treatment of osteoporosis and athero- defense against microbial pathogens. In a similar manner, NAbs sclerosis simultaneously. may also maintain homeostasis against OSEs, which are ubiqui- tous and increased in inﬂammatory states . In addition, because some OSEs share molecular signatures with PAMPs, there may Results have been selective pressure from both OSEs and PAMPs for the E06 prevents the effects of OxLDL on osteoblastic cells. 11 9,29–33 preservation of NAbs that bind to both . Consisent with earlier reports , addition of OxLDL to The IgM NAb E06 recognizes the phosphocholine (PC) moiety cultured osteoblastic cells attenuated their proliferation and of oxidized phosphatidylcholine present on the membrane of differentiation, and stimulated apoptosis (Fig. 1a–c). All these apoptotic cells and oxidized low-density lipoproteins (OxLDLs) . effects were prevented by the monoclonal murine E06 IgM, but The PC moiety is normally inaccessible to PRRs, but it is exposed not by IgM obtained from normal mice (Fig. 1a–c). E06 IgM by following oxidation of the sn2 PUFA of phosphatidylcholine and itself had no effect on any of these parameters in the absence of thereby becomes a conformational neo-epitope (PC-OxPL) . OxLDL. Importantly, E06 prevents the binding of OxLDL by SRs and We made transgenic C57BL/6 J mice expressing a single-chain TLRs on macrophages and thereby reduces the proinﬂammatory variable region fragment (scFv) of the E06 IgM. The transgene activity of OxPLs . Similar to E06, both the murine NAb E014 (E06-scFv) consists of a fusion protein of the heavy and light- and the human IK17, recognize MDA and prevent the proin- chain variable domains covalently joined by a ﬂexible peptide 11,21 ﬂammatory effect of MDA on macrophages . linker (Fig. 2a). The transgene is expressed under the control of Emerging evidence indicates that when OSEs persist and/or are the Apo-E promoter to provide for a high level of secretion of the excessive, the protective effect of the innate immune system protein by the liver and macrophages. E06-scFv transgenic mice against them is overwhelmed. Under such circumstances, the had normal body weight and did not display any morphologic or inﬂammatory response initiated by TLR and SR activation reproductive abnormalities . The His-tagged E06-scFv protein eventually leads to tissue damage and disease . In athero- was detected by ELISA in the serum of E06-scFv transgenic mice, sclerosis, OSEs in OxLDL bind to their cognate receptors present but not in WT mice (Fig. 2b). The E06-scFv is present in plasma in endothelial cells and macrophages and stimulate the of mice at a concentration of 20–30 μg/ml, and the level of production of inﬂammatory cytokines, which in turn recruit and endogenous IgM E06 was not affected by the expression of the 2 NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 ARTICLE Vehicle OxLDL a b Proliferation Alkaline phosphatase activity NS 10,000 P < 0.001 NS P < 0.001 P < 0.001 P = 0.007 0 0 No Ab E06 IgM No Ab E06 IgM Caspase 3 activity P < 0.001 P < 0.009 NS No Ab E06 IgM Fig. 1 E06 prevents the effects of OxLDL on osteoblast apoptosis, differentiation and proliferation. Osteoprogenitor cells from bone marrow were pre- treated for 1 h without (no Ab), or with E06 (2 µg/ml), or with murine IgM ('IgM', 2 µg/ml), in the absence ('veh') or presence of OxLDL (50 µg/ml) for a 2 days prior to measurement of BrDU incorporation, or b 2 days prior to measurement of alkaline phosphatase activity, or c 12 h prior to measurement of Caspase 3 activity. All measures were performed in triplicate cultures. Essentially identical results were obtained in a separate experiment. Data were analyzed by ANOVA; the p values reﬂect comparison to the respective vehicle. Data shown represent the mean (±SD). AFU: Arbitrary ﬂuorescence units, RLU: Relative luminescence units, NS: not signiﬁcant ATG TAA Myc EO6 heavy chain His EO6 light chain Linker lgκ (G4S)3 Spe I Poly(A) Sac II EO6 -scFv ApoE 3’UTR LE6 ApoE promoter Liver element Exon1 Exon2 b c His-tag ELISA Caspase 3 activity C57BL/6J E06-scFv Vehicle OxLDL P < 0.001 P < 0.001 P < 0.001 5 P < 0.001 3 NS NS NS No Ab E06 1:4 1:8 1:4 1:8 Serum from Serum from E06-scFv mice C57BL/6J mice Fig. 2 Generation of E06-scFv transgenic mice. a E06-scFv transgene structure. The transgene consists of a fusion protein of the heavy- and light-chain variable domains, which constitute the antigen recognizing portion of the natural antibody E06. The domains are covalently linked by a ﬂexible peptide linker. The transgene is expressed under the control of the ApoE promoter. Reproduced from Que et al. with permission. b Quantiﬁcation of E06-scFv in the serum of C57BL/6 J and E06-scFv 5-month-old male mice as determined by ELISA. C57BL/6 J n = 5, E06-scFv n = 6. Data were analyzed by Rank Sum test. c E06-scFv prevents the pro-apoptotic effect of OxLDL. Osteoblastic cells obtained from neonatal calvaria were pre-treated for 1 h without (no Ab), or with E06 (2 µg/ml), or with serum from E06-scFv or C57BL/6 J mice at the indicated dilutions (1:4, 1:8). Serum from 4 mice (per genotype) was pooled for this analysis. Caspase 3 activity was then measured 6 h after addition of OxLDL (50 µg/ml). All measures were performed in triplicate cultures. Essentially identical results were obtained in a separate experiment. Data were analyzed by ANOVA; the p values reﬂect comparison to the respective vehicle. Data shown represent the mean (±SD). AFU: Arbitrary ﬂuorescence units NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications 3 | | | BrdU (RLU × 10 ) RLU (x10 ) per 100 ms –1 –1 AFU min μg –1 –1 AFU min μg –1 –1 μm min μg protein ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 E06-scFv antibody . Similar to the effect of the E06 IgM, serum thickness (Fig. 3b) in LDLR-KO mice. This effect was due to from E06-scFv transgenic mice prevented OxLDL-induced reduced periosteal apposition, as evidenced by a decrease in the osteoblast apoptosis while serum from WT mice did not (Fig. 2c). total area without a change in the medullary area. HFD-fed The presence of the transgene did not change plasma lipid or LDLR-KO;E06-scFv mice had similar cortical thickness and lipoprotein levels in response to the various diets . cortical bone area as LDLR-KO mice on normal diet (Fig. 3b). However, the E06-ScFv transgene preserved cortical thickness in the HFD-fed LDLR-KO;E06-ScFv mice by decreasing medullary E06-scFv attenuates cortical bone loss in HFD-fed mice. Two- area. Thus, E06-scFv increased endosteal bone apposition, but did month-old LDLR-KO male mice were placed on a normal diet not prevent the negative effects of HFD on periosteal expansion (ND), or on a HFD containing 0.5% cholesterol and 21% milk fat. (Fig. 3b). Endosteal osteoblast number and bone formation rate A cohort of LDLR-KO;E06-scFv mice was placed on the HFD at were reduced in HFD-fed LDLR-KO mice as compared to the ND the same age. At 6.5 months, the LDLR-KO mice on HFD had controls, but these effects were prevented in LDLR-KO;E06-scFv lower femoral bone mineral density (BMD) compared to the mice (Fig. 3c). Neither the HFD nor the E06-scFv transgene LDLR-KO mice fed ND, as determined by dual-energy X-ray affected osteoclast number. absorptiometry (DXA) (Fig. 3a). Strikingly, LDLR-KO;E06-scFv mice were protected from the adverse effect of HFD on femoral BMD. At 7.5 months of age, the mice were euthanized and both E06-scFv increases cancellous bone in HFD-fed mice. Femoral femoral cortical and cancellous compartments were analyzed by metaphyseal cancellous bone mass, measured by micro-CT, was micro-CT and histology. The HFD reduced cortical diaphyseal unaffected by the HFD; albeit, there being a modest increase in LDLR-KO,HFD LDLR-KO;E06-scFv,HFD LDLR-KO,ND a b Femoral BMD Cortical thickness Bone area Total area Medullary area P = 0.015 P = 0.017 P < 0.001 P < 0.001 P < 0.001 P = 0.003 P = 0.049 NS NS P < 0.05 0.08 0.25 1.5 2.4 1.0 0.07 1.2 2.0 0.20 0.06 0.9 0.8 1.6 0.05 0.15 0 0 0 0 0 Ob .N/bone perimeter Oc.N/bone perimeter BFR/BS P = 0.015 P = 0.021 NS NS P = 0.05 P = 0.006 30 0.6 20 0.4 10 0.2 0 0 0.0 BV/TV (%) Tb number Tb thickness Tb separation P = 0.019 NS P < 0.001 P = 0.001 NS NS P = 0.007 P = 0.006 0.4 0.08 0.04 0.2 0 0 0 0 Ob .N/bone perimeter Oc.N/bone perimeter BFR/BS P = 0.032 NS P < 0.001 P = 0.036 NS NS 0.3 30 6 0.2 20 4 10 2 0.1 0 0 0.0 Fig. 3 E06-scFv attenuates the HFD-induced loss of cortical bone and increases cancellous bone. a Femoral BMD in 6.5-month-old male mice (LDLR-KO, ND, n = 10; LDLR-KO, HFD, n = 11; LDLR-KO;E06-scFv, HFD, n = 10). b Femoral cortical bone architecture in 7.5-month-old male mice (cortical thickness: LDLR-KO, ND, n = 10; LDLR-KO, HFD, n= 11; LDLR-KO;E06-scFv, HFD, n = 10; area measurements: LDLR-KO, ND, n = 10; LDLR-KO, HFD, n = 8; LDLR-KO; E06-scFv, HFD, n = 10). c Histomorphometry of the endocortical femoral surface (osteoblast and osteoclast numbers: LDLR-KO, ND, n = 10; LDLR-KO, HFD, n = 8; LDLR-KO;E06-scFv, HFD, n = 9; BFR/BS: LDLR-KO, ND, n = 9; LDLR-KO, HFD, n = 9; LDLR-KO;E06-scFv, HFD, n= 7). d Femoral cancellous bone architecture at 7.5 months (LDLR-KO, ND, n = 10; LDLR-KO, HFD, n = 11; LDLR-KO;E06-scFv, HFD, n = 10). e Histomorphometry of the cancellous femoral surface (osteoblast and osteoclast numbers: LDLR-KO, ND, n = 10; LDLR-KO, HFD, n = 8; LDLR-KO;E06-scFv, HFD, n = 10; BFR/BS: LDLR-KO, ND, n = 10; LDLR-KO, HFD, n = 9; LDLR-KO;E06-scFv, HFD, n= 9). Data analyzed by ANOVA except for medullary area which were analyzed by ANOVA on Ranks. Data shown represent the mean (±SD). Ob.N osteoblast number, Oc.N osteoclast number, BFR bone formation rate, BS bone surface, NS not signiﬁcant 4 NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications | | | –2 mg cm Per mm Per mm mm Per mm Per mm Per mm 2 –1 –1 2 –1 –1 mm μm μm day μm μm day mm mm NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 ARTICLE ATG TAA His IK17 heavy chain Myc IK17 light chain Linker Igκ Spe I Poly(A) Sac II LE6 IK17 -scFv ApoE 3 ′UTR ApoE promoter Liver element Exon2 Exon1 b c 180,000 BSA MDA-BSA 160,000 MAA-BSA 25,000 MAA-LDL 140,000 MDA-LDL 20,000 Cu-OxLDL 120,000 PC-BSA 15,000 LDL 100,000 C57BL/6 10,000 IK17-scFv 80,000 BSA 60,000 40,000 0.00 0.05 0.10 0.15 0.20 20,000 Dilution of plasma (reciprocal of dilution) C57BL/6 IK17scFv-Tg d e Vertebrae BV/TV Tb number Tb thickness Tb separation 0.08 0.30 P = 0.005 P = 0.008 40 NS 0.25 0.06 P = 0.004 0.20 30 0.04 0.15 0.02 0 0 0 0 LDLR-KO LDLR-KO;IK17-scFv fg Femur BV/TV Tb number Tb thickness Tb separation 0.08 0.30 NS 0.25 0.06 P = 0.007 P = 0.007 P = 0.014 0.20 0.04 0.15 0.02 0 0 0 0 Cortical thickness LDLR-KO 0.25 P = 0.016 LDLR-KO;IK17-scFv LDLR-KO; LDLR-KO IK17-scFv 0.20 Fig. 4 IK17-scFv increases cancellous bone and cortical thickness in HFD-fed mice. a IK17-scFv transgene structure. The transgene consists of a mutated single chain variant of the antigen binding domain of the human anti-MDA antibody (IK17). The transgene contains a fusion protein of the heavy- and light- chain variable domains, covalently linked by a ﬂexible peptide linker. The transgene is expressed under the control of the ApoE promoter. b Binding properties of plasma from wild-type C57BL/6 and IK17-scFv-Tg mice (n = 6 per group) to indicated antigens. Data shown represent the mean (±SEM). BSA: bovine serum albumin, MDA-BSA: malondialdehyde-modiﬁed BSA, MAA-BSA: malondihyldehyde-acetaldehyde-modiﬁed BSA, MAA-LDL: malondihyldehyde-acetaldehyde-modiﬁed LDL, MDA-LDL: malondialdehyde-modiﬁed LDL, Cu-OxLDL: copper-oxidized LDL, PC: phosphocholine. c Inhibition of binding of biotinylated MDA-modiﬁed LDL (Bt-MDA-LDL) to J774 macrophages by IK17-scFv plasma. d–h Micro-CT determination of (d, e) vertebral, and (f–h) femoral bone architecture in 6-month-old male mice fed a HFD diet. d Lateral images of vertebral cancellous bone. e Quantiﬁcation of vertebral cancellous bone architecture. f Longitudinal images of distal femur, with diaphysis at the top. Boxes indicate area of quantiﬁcation of cancellous bone. g Quantiﬁcation of cancellous bone architecture. h Quantiﬁcation of diaphyseal cortical thickness. n = 5 per group. e–g data were analyzed by Student’s t-test except for vertebral trabecular number which were analyzed by Rank Sum test. Data in h were analyzed by Rank Sum test. Data shown represent the mean (±SD). BV/TV: bone volume/total volume, Tb: trabecular, NS: not signiﬁcant NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications 5 | | | RLU/100 ms mm Bt-MDA-LDL Bound (RLU/100 ms) Per mm Per mm mm mm mm mm ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 E06-scFv C57BL/6J a b Vertebrae BV/TV Tb number Tb thickness Tb separation 0.10 P = 0.0002 P = 0.00002 40 6 0.25 NS 0.06 0.15 P = 0.000009 0 0 0 C57BL/6J E06-scFv c d Femur BV/TV Tb number Tb thickness Tb separation 0.10 P = 0.015 P = 0.004 6 0.25 NS 0.06 0.15 P = 0.005 0 0 0 0 Diaphysis Cortical thickness Medullary area Total area 0.30 2.0 1.2 NS NS NS C57BL/6J E06-scFv 0.8 0.20 1.6 0 0 0 Metaphysis Cortical thickness Total area Medullary area 0.30 1.6 2.4 NS NS NS 0.20 0.8 1.6 0 0 0 Fig. 5 Overexpression of E06-scFv increases cancellous bone in mice fed a normal diet. Micro-CT determination of (a, b) vertebral, and (c–f) femoral bone architecture in 5-month-old male mice fed a normal diet. a Lateral images of vertebral cancellous bone. b Quantiﬁcation of vertebral cancellous bone architecture. c Longitudinal images of femur with proximal end at the top. Boxes indicate area of quantiﬁcation of cancellous bone. d Quantiﬁcation of cancellous bone architecture. e Quantiﬁcation of diaphyseal cortical architecture. f Quantiﬁcation of metaphyseal cortical architecture. C57BL/6 J n = 5, E06-scFv n= 6. Data analyzed by two-tailed Student’s t-test. Data shown represent the mean (±SD). BV/TV: bone volume /total volume, Tb: trabecular, NS: not signiﬁcant trabecular number and decrease in trabecular separation in the variable domains and linker (Fig. 4a). The IK17-scFv is present in HFD-fed LDLR-KO mice (Fig. 3d). HFD-fed LDLR-KO;E06-scFv plasma of mice at an average concentration of 40 μg/ml (range mice exhibited an increase in femoral cancellous bone mass as 16.2 to 64 μg/ml). IK17-scFv mice gained weight equally and did compared to LDLR-KO fed either ND or HFD, due to a sub- not exhibit any developmental defects. The binding speciﬁcity of stantial increase in trabecular number and decrease in trabecular the IK17-scFv transgene was determined using ELISA (Fig. 4b). separation. (Fig. 3d). HFD caused a decline in the number of Plasma from IK17-scFv transgenic mice bound to MDA or MAA cancellous osteoblasts, as well as a reduction in osteoclasts (malonaldehyde-acealdehyde) epitopes but not to PC epitopes or (Fig. 3e). In contrast, the number of osteoblasts was signiﬁcantly non-oxidized LDL. Plasma from IK17-scFv transgenic mice also higher, whereas osteoclast number was unchanged, in HFD-fed prevented the binding of MDA-containing LDL to macrophages LDLR-KO;E06-scFv mice as compared to HFD-fed LDLR-KO compared to plasma from C57BL/6 mice (Fig. 4c). controls (Fig. 3e). Bone formation rate (BFR) mirrored the LDLR-KO;IK17-scFv male mice and LDLR-KO controls were changes in osteoblast number in the two groups, but did not fed a HFD containing 60% fat from lard for 4 months, beginning reach statistical signiﬁcance. at 2 months of age. HFD-fed LDLR-KO;IK17-scFv transgenic mice exhibited a signiﬁcant increase in cancellous bone mass both IK17-scFv increases bone mass in HFD-fed mice. To investigate in the vertebra (Fig. 4d, e) and in the femur (Fig. 4f, g) as whether the bone effects of the E06-scFv transgene were a compared to HFD-fed LDLR-KO controls. The increased property of anti-OSEs in general, we next examined mice over- cancellous bone mass was due to increased trabecular number expressing IK17-scFv, which speciﬁcally binds to the OSE and reduced trabecular separation, in both the vertebra and the MDA . The transgene construct is analogous to the one used for femur. Moreover, LDLR-KO;IK17-scFv transgenic mice had the E06-scFv, except for different heavy chain and light-chain increased cortical thickness (Fig. 4h). 6 NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications | | | mm % mm Per mm mm mm Per mm mm mm mm mm mm mm NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 ARTICLE Table 1 E06-scFv increases cancellous bone mass by increasing bone formation rate and decreasing osteoclasts C57BL/6 J E06-scFv Vertebral BV/TV (%) 22.5 ± 2.2 29.6 ± 4.2 (P= 0.002) Femoral metaphysis BV/TV (%) 4.5 ± 1.6 14.2 ± 3.7 (P= 0.003) Femoral diaphyseal Cortical Thickness (mm) 0.20 ± 0.01 0.20 ± 0.02 (NS) Vert. BFR/BS (µm /µm/d) 0.05 ± 0.02 0.10 ± 0.04 (P = 0.05) Vert. N.Oc/B.Pm (#/mm) 8.63 ± 1.52 4.19 ± 1.62 (P = 0.03) Quantiﬁcation of cancellous and cortical bone mass as determined by micro-CT and histomorphometric analysis of cancellous bone in 11-month-old mice fed a normal diet. n = 4, 2 females and 2 males per genotype. Data shown represent the mean (±SD). Data were analyzed by two-tailed Student’s t-test BV/TV Bone volume/total volume, BFR/BS bone formation rate/bone surface, N.Oc osteoclast number, B. Pm bone perimeter, NS not signiﬁcant Anti-PC IgM a bc 6-month-old 6-month-old 7-month-old 26-month-old 22-month-old 25-month-old 120,000 100,000 20,000 100,000 16,000 80,000 80,000 P = 0.03 12,000 60,000 60,000 8000 40,000 40,000 P < 0.001 P < 0.001 4000 20,000 20,000 0 0 0 Dilution 1: 100 Dilution 1: 100 Dilution 1: 100 Fig. 6 Aging is associated with reduced anti-PC IgM levels. Quantiﬁcation of anti-PC IgM levels by ELISA in serum of a 6- (n = 10) and 26- (n = 9) month- old female C57BL/6 J mice; b 6- (n = 9) and 22- (n = 8) month-old female C57BL/6 J mice, and c 7- (n = 10) and 25- (n = 10) month-old female C57BL/6 J mice. Data analyzed by Rank Sum test. Data shown represent the mean (±SD). RFU: Relative luminescence units E06-scFv increases cancellous bone in mice fed a normal diet. Discussion Having observed an anabolic effect of two different anti-OSE The evidence reported herein reveals that anti-OSE antibodies antibodies in the context of atherosclerosis induced by HFD, we recognizing the PC and MDA moieties counter the adverse examined whether the overexpression of one of them (E06-scFv) effects of these oxidized lipids in mice. Unexpectedly, the E06- had an effect on bone of mice maintained on a normal diet. At scFv also promoted bone anabolism not only in the context of 5 months of age, E06-scFv transgenic mice exhibited a signiﬁcant hyperlipidemia, but also in mice maintained on a normal diet. increase in cancellous bone mass in vertebrae and femora, as This ﬁnding strongly suggests that OSEs generated throughtout determined by micro-CT (Fig. 5a–d). This increase was due to life chronically restrain bone formation. Speciﬁcally, OSEs increased trabecular number and reduced trabecular separation, negatively affect osteoblasts, and may increase osteoclasts, thereby while trabecular thickness was unchanged. Notably, the cancel- causing a previously unappreciated effect on skeletal homeostasis lous bone of the femur extended into the diaphyseal region in regardless of diet. E06-scFv mice—well-beyond the usual metaphyseal location of OSEs are ubiquitous, and the innate responses that they trigger this bone compartment in wild type (WT) mice (Fig. 5c). The have important physiological, as well as pathophysiological con- E06-scFv transgene had no effect on femoral cortical bone as sequences . OSE generation occurs during oxidation of LDL and determined by measurements of cortical thickness, total area or it is also a well-recognized feature of physiological cell apoptosis, medullary area at the diaphysis (Fig. 5e), or distal metaphysis a process that continuously occurs in highly regenerating tissues, 11,43,44 (Fig. 5f). The same ﬁndings were reproduced in a second such as bone marrow and bone . Apoptotic bodies and experiment employing a cohort of 11-month-old male and female microparticles shed by apoptotic cells contain abundant OSEs 11,43–45 E06-scFv mice fed a normal diet (Table 1). Histomorphometric identical to those formed during lipid peroxidation . Such analysis of vertebral cancellous bone in these mice showed OSEs mark oxidatively modiﬁed endogenous molecules, which increased bone formation rate and a decrease in osteoclast can be then recognized and removed by the innate immune number (Table 1). system . Indeed, these OSEs represent a class of many 'eat-me' signals recognized by macrophages during efferocytosis . The only known activity of the anti-OSE scFv antibody fragments is their ability to bind to their target antigens, thereby leading to Aging is associated with reduced anti-PC IgM levels. Age- sequestration or masking of exposed PC or MDA neo-epitopes. related loss of cancellous bone in mice is due to an insufﬁcient In turn, this could either inhibit the uptake of these OSEs, and/or number of osteoblasts and it is associated with increased lipid 39–41 block their proinﬂammatory properties . The bone anabolic peroxidation . In view of evidence that B-1 cells decline with 34–37 effect of these antibody fragments is likely unrelated to increased age in humans , we measured the level of anti-PC IgM clearance of apoptotic cells because this requires the Fc portion of antibodies in aged female C57BL/6 J mice, in which we had a full antibody. Instead, their anabolic efﬁcacy is probably due to previously documented age-dependent decline in cortical and sequestration of OSEs present on OxLDL or apoptotic cells and cancellous bone . We found that, compared to 6- or 7-month- the inhibition of the proinﬂammatory and anti-osteogenic effects old mice, sera from 22 or 26-month-old mice had much lower of such OSEs. levels of anti-PC IgM (Fig. 6a–c). NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications 7 | | | RLU per 100 ms ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 antibodies is insufﬁcient to maintain maximal cancellous bone SR mass under normal conditions. The clinical relevance of Oxidation-specific anti-OSE-speciﬁc IgM has recently been elucidated by the evi- epitopes Osteoblast dence that low levels of anti-PC IgM antibodies are associated with increased incidence of cardiovascular diseases . In humans TLR 34–37 Natural IgM B-1 cells decline with age , raising the possibility that declining NAb levels against OSEs contribute to the pathogenesis SR of age-related diseases. The decline of anti-PC IgM levels in older Inflammatory mice may indicate that such decline contributes to the patho- Oxidation-specific cytokines epitopes genesis of involutional osteoporosis. Although further work will be required to establish a cause-effect association, we submit that TLR increasing the titer of antibodies to OSEs may attenuate and Macrophage perhaps reverse age-dependent bone loss. This could be achieved Fig. 7 Model for the anabolic effect of anti-OSE antibodies on bone. OSEs by immunization strategies to increase endogenous titers of OSE may directly inhibit the differentiation and survival of osteoblasts via antibodies (either against PC and/or MDA), or by passive scavenger receptor (SR) or toll-like receptors (TLR). OSEs may stimulate 23,24,57 immunization with exogenously generated antibodies . the production of anti-osteogenic cytokines via activation of these Such novel therapeutic approaches could be used for the treat- receptors on macrophages. Natural IgM that recognize OSEs may exert an ment of osteoporosis and atherosclerosis simultaneuosly. anabolic effect by preventing their binding to SR and TLR of osteoblastic cells and /or macrophages Methods Animals. All animal procedures were approved by the Institutional Animal Care and Use Committees of the University of Arkansas for Medical Sciences, and by the The histomorphometric evidence presented in this paper UCSD. The transgenic E06-scFv founder line was generated in the C57BL/6 J shows that in both LDLR-KO;E06-scFv mice fed a HFD or E06- background and backcrossed to each other to generate homozygous E06-scFv lines scFv mice fed a normal diet, osteoblast number and bone for- expressing high titers of the E06-scFv in plasma. These were crossed into LDLR- mation rate in cancellous and cortical bone was increased. This KO mice on C57BL/6 J background present in the UCSD vivarium. The transgenic IK17-scFv founder line was similarly generated and backcrossed into the LDLR-KO ﬁnding clearly indicates that the E06-scFv transgene affects background. Transgenic mice were maintained as homozygous breeding pairs. osteoblast by blocking the anti-osteogenic effects of OSEs. Contemporaneous age- and sex-matched C57BL/6 J or LDLR-KO mice were used Scavenger receptor class B type I (SR-B1) and CD36 are as controls. Offspring of E06 transgenic mice were screened both for plasma E06- 47,48 11 receptors for OxLDL and the two major SRs for PC-OxPL . scFv titer by ELISA (see below), and for the transgene by PCR genotyping of tail DNA using KAPA Mouse Genotyping Kit with the upstream primer sequence 5′- Both these receptors are expressed on macrophages, as well as TAC AAT TGA GCT GGC TAG CCA CCA TGG AG-3′ and the downstream osteoblasts. In osteoblasts, they have been implicated in the reverse primer sequence 5′-GCT GTA CCA AGC CTC CTC CAG ACT CCA CCA uptake of OxLDL, cholesteryl ester, and estradiol . The role of G-3′ to yield a 540-bp product. Offspring of IK17 transgenic mice were screened these SRs in bone homeostasis has been studied only in mice that for the transgene by PCR genotyping of the tail DNA using KAPA Mouse lack SRs globally, but the data from those studies are conﬂicting. Genotyping Kit with the upstream primer sequence 5′-AGC GAT TAG TGG TAC TGG TCG TAG C -3′ and the downstream reverse primer sequence 5′-GTC GAC Indeed, it has been reported that although mice with global 50 GGC GCT ATT CAG ATC CTC-3′ to yield a 295-bp product corresponding to the deletion of CD36 have low bone mass , mice with global deletion nucleotide sequence. LDLR-KO mice were genotyped using the following primers of SR-B1 maintained on a normal diet have high bone LDLR- 5′-AAT CCA TCT TGT TCA ATG GCC GAT C -3′ and LDLR 5′-CCA 27,51,52 mass , as well as striking architectural and histomorpho- TAT GCA TCC CCA GTC TT -3′ to yield a 350-bp product in KO animals. The transgenic mice are morphologically normal and healthy. metric similarities with E06-scFv transgenic mice. Be that as it Animals were group housed under speciﬁc pathogen-free conditions, may, the global nature of the deletion makes it difﬁcult to maintained with a constant temperature of 23 °C, a 12:12 h light–dark cycle, and understand the role of SR-B1 speciﬁcally in osteoblasts and does had access to food and water ad libitum. The animals were matched for age, body not rule out the possibility that the phenotype is due to effects of weight, and total cholesterol prior to feeding the ND (TD22/5 Rodent Diet from Harlan Laboratories that contains 5.5% fat, 0.003% cholesterol and has 17% of kcal OSEs in other cell types. In addition to the direct effects on from fat) or HFD diet. The HFD used in experiment of Fig. 3 contains 21.2% fat, osteoblasts, the OSEs may act indirectly via the stimulation of 0.5% cholesterol and has 41.8% of kcal from fat (TD00457 from Harlan anti-osteogenic cytokine (e.g., TNFα and IL1-β) production by Laboratories). Body weight of mice fed ND vs. this HDF was indistinguishable at macrophages (Fig. 7). The dissection of direct vs. indirect the time of euthanasia: LDLR-KO mice on ND (n = 10), 28.9 ± 2.2 g; LDLR-KO mice on HFD (n = 11), 30.7 ± 4.9 g; LDLR-KO;E06-scFv mice on HFD (n = 10), mechanisms will require further studies. 31.8 ± 4.4 g; NS by Kruskal–Wallis ANOVA on Ranks. The HFD diet used in We have shown previously that HFD had no effect on endosteal experiment in Fig. 4 (TD12492 from Research Diets) contains 34.9% fat, 24.5% lard osteoclast number, but caused a signiﬁcant decrease of osteoclasts and has 60% of kcal from fat. Mice were injected intraperitoneally with tetracycline in cancellous bone . The latter effect is in agreement with studies hydrochloride (30 µg per gram body weight) 6 days and 2 days before sacriﬁce to showing that RANKL-induced differentiation of the RAW264.6 permit measurement of bone formation rate. Animals were euthanized by CO2 inhalation. Genotype was re-checked after euthanasia. pre-osteoclastic cell line is inhibited by OxLDL . However, others have reported conﬂicting results. Namely, OxLDL increased basal Bone imaging. BMD of the left femur was determined by dual-energy X-ray levels of RANKL in MG-63 human osteosarcoma cells , and 55,56 absorptiometry (DXA) of sedated mice (2% isoﬂurane) using a PIXIimus densit- hyperlipidemia promoted osteoclastogenesis . In the present ometer (GE Lunar) as described . The mean coefﬁcient of variation of BMD of a work, the E06-scFv transgene did not change osteoclast number in proprietary phantom (performed prior to each use) during the conduct of these the LRLR-KO;E06-scFv transgenic mice fed a HFD. Albeit, a studies was 0.40%. For micro-CT determination of skeletal archietecture, the ﬁfth lumbar vertebra and the left femur were dissected, cleaned of soft tissues, placed in decrease in osteoclasts was seen in the older E06-scFv transgenic ﬁxative solution, and either transferred to 100% ethanol (experiment in Fig. 3)or mice. Thus, the effects of anti-OSEs on osteoclastogenesis remain wrapped with saline and stored at −20 °C until analysis (experiment in Fig. 4 and presently unclear and future studies will be needed to clarify Fig. 5). Bones were loaded into 12.3 mm diameter scanning tubes and imaged whether the OSEs exert direct or indirect effects on osteoclasto- (micro-CT40, Scanco Medical, Basserdorf, Switzerland). Scans were integrated into 3D voxel images (1024 × 1024 pixel matrices for each individual planar stack) and a genesis or whether anti-OSEs prevent those effects. Gaussian ﬁlter (sigma = 0.8, support = 1) was used to reduce signal noise. A In addition to a pathogenic role of OSEs in the setting of threshold of 200 was applied to all scans, at medium resolution (E = 55 kVp, hyperlipidemia, our data suggest for the ﬁrst time that anti-OSE I = 145 µA, integration time = 200 ms). For cancellous bone determinations, the antibodies have the potential to protect the skeleton from adverse distal femur was scanned from the boundary between the growth plate and effects of OSEs; and that the titer of endogenous anti-OSE metaphysis towards the diaphysis to obtain 151 slices (12 µm/slice). The entire 8 NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 ARTICLE For the binding proﬁle of plasma from IK17-scFv mice or control C57BL/6 vertebral body was scanned with a transverse orientation from the rostral growth mice the microplates were coated with the various antigens at 5 μg/ml (50 μl per plate to the caudal growth excluding any bone outside the vertebral body plate to well) in PBS overnight at 4 °C and then serum was added as above. The bound obtain 233 slices. All cancellous measurements were made by drawing contours IK17-scFv was detected with anti-His antibody as described above. every 10 to 20 slices and using voxel counting for bone volume per tissue volume and sphere ﬁlling distance transformation indices, without pre-assumptions about the bone shape as a rod or plate for cancellous microarchitecture. In both vertebra Macrophage binding assay. To demonstrate the speciﬁcity of IK17-scFv binding and femur, the analysis was performed on contours of the cross sectional acquired to biotinylated MDA-LDL and ability to inibhit their binding to J774 macrophages, images drawn to exclude the primary spongiosa and cortex. Cortical thickness, total a chemoluminescent binding assay was performed as described . Brieﬂy, area, and medullary area were measured at the femoral mid-diaphysis or distal biotinylated MDA-LDL (5 μg/ml) was incubated in the absence or presence of metaphysis. The mean coefﬁcient of variation of the micro-CT phantom (per- IK17-scFv transgenic plasma or controls at various diluitions overnight at 4 °C. The formed weekly) during the conduct of these studies was 1.23%. supernatatans were then added to macrophages plated in 96-well microplates and the binding of biotinylated MDA-LDL was detected by AP-labeledNeutrAvidin and Histomorphometry. Bones were ﬁxed in 10% Millonig’s formalin (Leica Micro- chemoluminescent ELISA. systems Inc.), dehydrated in 100% ethanol, and then embedded in methyl methacrylate (Sigma-Aldrich) as described . The analysis was done on 5 µm thick Statistics. Data are shown as mean ±SD, or SEM where indicated; or as box plots longitudinal sections. Measurements were made on both of the endosteal surfaces which represent the 25th and 75th percentile of data points. Whiskers represent the of a longitudinal section in a blinded fashion using Osteomeasure version XP 3.1 10th and 90th percentile of data points, dots represent data points outside the 10th (Osteometrics Inc.). For measurement of static indices of osteoclast and osteoblast and 90th percentile, and the line in the box denotes the median value. Numerical number, the sections were stained for tartrate-resistant acid phosphatase with data presented in Table 1 are shown as the mean and SD. Statistical analysis was toluidine blue counterstaining. Osteoblasts were identiﬁed as teams of cells (≥2) performed using Sigma Plot (version 12.5). Group mean values were compared, by overlying osteoid. The determination of ﬂuorescent tetracycline labeling was done Student’s two-tailed t-test, or by ANOVA as appropriate. If the normality or equal using ﬂuroresce microscopy on unstained sections. Bone fomation rate (BFR) is variance assumptions for standard parametric analysis methods were not met, the deﬁned as the distance between the double labels divided by the interval between data were analyzed by the Mann-–Whitney Rank Sum test or the Kruskal–Wallis the ﬂuorocrome administrations and multiplied by the sum of the double labeled One Way ANOVA on Ranks test. When ANOVA indicated a signiﬁcant effect, perimeter and one-half of the single-labeled perimeter. If only a single label was pairwise multiple comparisons were performed and the P values adjusted using the present, data were treated as a missing value for statistical purposes. Histomor- Holm-Sidak method. P values <0.05 were considered statistically signiﬁcant. Based phometric data are reported using the nomenclature recommended by the on our previous studies, the sample size is adequate to detect changes with American Society for Bone and Mineral Research . sufﬁcient power . For in vitro experiments replicates were of sufﬁcient size to provide conﬁdence in measurements. Each ﬁgure legend includes the number of mice used in the experiments. There were no data excluded. In Fig. 3 poor section Culture of osteoblastic cells. Bone marrow cells were obtained by ﬂushing the quality precluded histological analysis of some samples and the numbers of sam- femoral diaphysis from 3- to 4-month-old C57BL/6 J mice with a-MEM medium ples analyzed are speciﬁed in the ﬁgure legend. All data were collected by personnel (Invitrogen, Carlsbad, CA, USA). Calvaria cells were isolated from neonatal pups as blinded to the identity of the samples. described before by sequential digestion with collagenase type 2 (Worthington, CLS-2, lot 47E17554B). Bone marrow stromal and calvaria cells were maintained in a-MEM containing 10% preselected FBS (Sigma-Aldrich), 1% penicillin/strepto- Data availability. The datasets generated during and/or analysed during the mycin/glutamine in presence of 1 mM ascorbate-2-phosphate up to 70% current study are available from the corresponding author upon reasonable request. conﬂuence. Oxidized LDL was obtained from Alfa Aesar, E06 IgM from AVANTI (Cat Number 330001 S lot number E06-19 and E06-17), mouse IgM from Santa Received: 22 November 2017 Accepted: 26 March 2018 Cruz Biotechnology (Catalog number sc-3881). Apoptotic cells were quantiﬁed by measuring caspase 3 activity in cell lysates as described previously . Brieﬂy, lysates (100 μg protein) were incubated with 50 μM DEVD-AFC in 50 mM HEPES (pH 7.4), 100 mM NaCl, 0.1% CHAPS, 10 mM DTT, 1 mM EDTA, and 10% glycerol, in the absence or presence of the irreversible inhibitor DEVD-CHO for 60 min. The released ﬂuorescent AFC was measured in a microplate ﬂuorescence reader FL500 (Bio Tek Instruments, Winooski, Vermont, USA) with excitation/emission wave- References lengths of 340/542 nm. Protein concentration in the lysate was measured using a 1. Farhat, G. N. et al. Volumetric BMD and vascular calciﬁcation in middle-aged Bio-Rad detergent–compatible kit (Bio-Rad). women: the Study of Women’s Health Across the Nation. J. Bone Miner. Res. For the alkaline phosphatase (ALP) activity measurement the cells were lysed in 21, 1839–1846 (2006). 100 mM glycine, 1 mM MgCl2, and 1% Triton X-100 at pH 10 using a buffer 2. Kiel, D. P. et al. Bone loss and the progression of abdominal aortic calciﬁcation containing 2-amino-2-methylpropanol and p-nitrophenylphosphate (Sigma- over a 25 year period: the Framingham Heart Study. Calcif. Tissue Int. 68, Aldrich Inc). ALP activity was normalized to total protein concentration measured 271–276 (2001). as described above. Proliferation was measured by BrdU incorporation with a kit 3. Schulz, E., Arfai, K., Liu, X., Sayre, J. & Gilsanz, V. Aortic calciﬁcation and the from Roche Diagnostics following the manufacturer instructions. For all assays risk of osteoporosis and fractures. J. Clin. Endocrinol. Metab. 89, 4246–4253 triplicate cultures were analyzed. (2004). 4. Sennerby, U. et al. Cardiovascular diseases and risk of hip fracture. JAMA 302, ELISA. For measurement of E06 single-chain antibody (E06-scFv) in serum or 1666–1673 (2009). plasma, 96-well round-bottomed microplates were coated with phosphocholine- 5. Ye, C. et al. Decreased bone mineral density is an independent predictor for KLH (Santa Cruz Biotechnology) at 5 μg/ml (50 μl per well) in PBS overnight at the development of atherosclerosis: a systematic review and meta-analysis. 4 °C. The plates were then washed with PBS 3 times, blocked with 1% BSA in PBS PLoS One. 11, e0154740 (2016). (75 μl per well) for 45 min at room temperature, and washed three times with PBS. 6. Hirasawa, H. et al. ApoE gene deﬁciency enhances the reduction of bone E06-scFv transgenic mouse serum or serum from C57BL/6 J controls (25 μl diluted formation induced by a high-fat diet through the stimulation of p53-mediated 1:20 with 1% BSA-PBS) was added to the wells, and incubated for 90 min at room apoptosis in osteoblastic cells. J. Bone Miner. Res. 22, 1020–1030 (2007). temperature. Bound E06-scFv was detected with anti-His tag antibody conjugated 7. Pirih, F. et al. Adverse effects of hyperlipidemia on bone regeneration and with alkaline phosphatase (Sigma-Aldrich Inc, Catalog number A5588, Clone HIS- strength. J. Bone Miner. Res. 27, 309–318 (2012). 1, Lot number 3085M-4836V), in Tris buffered saline (TBS) buffer containing 1% 8. Soares, E. A., Nakagaki, W. R., Garcia, J. A. & Camilli, J. A. Effect of BSA (1:2000 dilution, 50 μl per well), followed by washing with PBS, a rinse with hyperlipidemia on femoral biomechanics and morphology in low-density deionized water to remove phosphates, and the addition of 25 μl of 50% (1:1 in di lipoprotein receptor gene knockout mice. J. Bone Miner. Metab. 30, 419–425 H2O) LumiPhos 530 (Lumigen, Southﬁeld, MI) as luminescent substrate. Che- (2012). miluminescence was measured as relative light units (RLU) over 100 ms using a 9. Tintut, Y. & Demer, L. L. Effects of bioactive lipids and lipoproteins on bone. plate Luminometer (DYNEX Technologies). To measure the absolute levels of the Trends Endocrinol. Metab. 25,53–59 (2014). E06-scFv or IK17-scFv in plasma, a standard curve was prepared using puriﬁed 10. Liu, Y. et al. Skeletal inﬂammation and attenuation of Wnt signaling, Wnt E06-scFv or IK17-scFv, respectively. The puriﬁed E06-scFv and IK-17scFv were ligand expression, and bone formation in atherosclerotic ApoE-null mice. Am. obtained from culture supernatants of transfected HEK293 cells and isolated with J. Physiol. Endocrinol. Metab. 310, E762–E773 (2016). Ni-NTA agarose beads (Qioagen) according to manifacture’s protocol. 11. Binder, C. J., Papac-Milicevic, N. & Witztum, J. L. Innate sensing of oxidation- To quantify serum levels of IgM that bind PC, 25 µL of diluted serum (in 1% speciﬁc epitopes in health and disease. Nat. Rev. Immunol. 16, 485–497 BSA-PBS) was added to microplates coated with PC-KLH. Bound anti-PC IgM was (2016). quantiﬁed by chemiluminescence with anti-µ antibody conjugated with alkaline phosphatase at 1:10.000 dilution (Sigma-Aldrich Inc, Catalog number A6988, Lot 12. Kapetanovic, R., Bokil, N. J. & Sweet, M. J. Innate immune perturbations, accumulating DAMPs and inﬂammasome dysregulation: a ticking time bomb number #SLBC1772V), as above. All determinations were done in triplicate. in ageing. Ageing Res. Rev. 24,40–53 (2015). NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications 9 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 13. Feldman, N., Rotter-Maskowitz, A. & Okun, E. DAMPs as mediators of sterile 43. Huber, J. et al. Oxidized membrane vesicles and blebs from apoptotic cells inﬂammation in aging-related pathologies. Ageing Res. Rev. 24,29–39 (2015). contain biologically active oxidized phospholipids that induce monocyte- 14. Miller, Y. I. et al. Oxidation-speciﬁc epitopes are danger-associated molecular endothelial interactions. Arterioscler. Thromb. Vasc. Biol. 22, 101–107 (2002). patterns recognized by pattern recognition receptors of innate immunity. Circ. 44. Chang, M. K. et al. Apoptotic cells with oxidation-speciﬁc epitopes are Res. 108, 235–248 (2011). immunogenic and proinﬂammatory. J. Exp. Med. 200, 1359–1370 (2004). 15. Witztum, J. L. & Lichtman, A. H. The inﬂuence of innate and adaptive 45. Tsiantoulas, D. et al. Circulating microparticles carry oxidation-speciﬁc immune responses on atherosclerosis. Annu. Rev. Pathol. 9,73–102 (2014). epitopes and are recognized by natural IgM antibodies. J. Lipid Res. 56, 16. Chou, M. Y. et al. Oxidation-speciﬁc epitopes are important targets of innate 440–448 (2015). immunity. J. Intern. Med. 263, 479–488 (2008). 46. Chang, M. K. et al. Monoclonal antibodies against oxidized low-density 17. Leibundgut, G., Witztum, J. L. & Tsimikas, S. Oxidation-speciﬁc epitopes and lipoprotein bind to apoptotic cells and inhibit their phagocytosis by elicited immunological responses: translational biotheranostic implications for macrophages: evidence that oxidation-speciﬁc epitopes mediate macrophage atherosclerosis. Curr. Opin. Pharmacol. 13, 168–179 (2013). recognition. Proc. Natl Acad. Sci. USA 96, 6353–6358 (1999). 18. Canton, J., Neculai, D. & Grinstein, S. Scavenger receptors in homeostasis and 47. Kunjathoor, V. V. et al. Scavenger receptors class A-I/II and CD36 are the immunity. Nat. Rev. Immunol. 13, 621–634 (2013). principal receptors responsible for the uptake of modiﬁed low density 19. Baumgarth, N., Tung, J. W. & Herzenberg, L. A. Inherent speciﬁcities in lipoprotein leading to lipid loading in macrophages. J. Biol. Chem. 277, natural antibodies: a key to immune defense against pathogen invasion. 49982–49988 (2002). Springe. Semin. Immunopathol. 26, 347–362 (2005). 48. Gillotte-Taylor, K., Boullier, A., Witztum, J. L., Steinberg, D. & Quehenberger, 20. Binder, C. J. et al. Innate and acquired immunity in atherogenesis. Nat. Med. O. Scavenger receptor class B type I as a receptor for oxidized low density 8, 1218–1226 (2002). lipoprotein. J. Lipid Res. 42, 1474–1482 (2001). 21. Shaw, P. X. et al. Human-derived anti-oxidized LDL autoantibody blocks 49. Brodeur, M. R., Brissette, L., Falstrault, L., Luangrath, V. & Moreau, R. uptake of oxidized LDL by macrophages and localizes to atherosclerotic Scavenger receptor of class B expressed by osteoblastic cells are implicated in lesions in vivo. Arterioscler. Thromb. Vasc. Biol. 21, 1333–1339 (2001). the uptake of cholesteryl ester and estradiol from LDL and HDL3. J. Bone 22. Faria-Neto, J. R. et al. Passive immunization with monoclonal IgM antibodies Miner. Res. 23, 326–337 (2008). against phosphorylcholine reduces accelerated vein graft atherosclerosis in 50. Kevorkova, O. et al. Low-bone-mass phenotype of deﬁcient mice for the apolipoprotein E-null mice. Atherosclerosis 189,83–90 (2006). cluster of differentiation 36 (CD36). PLoS One. 8, e77701 (2013). 23. Binder, C. J. et al. Pneumococcal vaccination decreases atherosclerotic lesion 51. Martineau, C., Kevorkova, O., Brissette, L. & Moreau, R. Scavenger receptor formation: molecular mimicry between Streptococcus pneumoniae and class B, type I (Scarb1) deﬁciency promotes osteoblastogenesis but stunts oxidized LDL. Nat. Med. 9, 736–743 (2003). terminal osteocyte differentiation. Physiol. Rep. 2, e12117 (2014). 24. Tsimikas, S. et al. Human oxidation-speciﬁc antibodies reduce foam cell 52. Martineau, C., Martin-Falstrault, L., Brissette, L. & Moreau, R. Gender- and formation and atherosclerosis progression. J. Am. Coll. Cardiol. 58, 1715–1727 region-speciﬁc alterations in bone metabolism in Scarb1-null female mice. (2011). J. Endocrinol. 222, 277–288 (2014). 25. Gruber, S. et al. Sialic acid-binding immunoglobulin-like lectin g promotes 53. Mazière, C. et al. Oxidized low density lipoprotein decreases Rankl-induced atherosclerosis and liver inﬂammation by suppressing the protective functions differentiation of osteoclasts by inhibition of Rankl signaling. J. Cell. Physiol. of B-1 cells. Cell Rep. 14, 2348–2361 (2016). 221, 572–578 (2009). 26. Rosenfeld, S. M. et al. B-1b cells secrete atheroprotective igm and attenuate 54. Maziere, C., Salle, V., Gomila, C. & Maziere, J. C. Oxidized low density atherosclerosis. Circ. Res. 117, e28–e39 (2015). lipoprotein enhanced RANKL expression in human osteoblast-like cells. 27. Martineau, C., Martin-Falstrault, L., Brissette, L. & Moreau, R. The Involvement of ERK, NFkappaB and NFAT. Biochim. Biophys. Acta 1832, atherogenic Scarb1 null mouse model shows a high bone mass phenotype. 1756–1764 (2013). Am. J. Physiol. Endocrinol. Metab. 306, E48–E57 (2014). 55. Graham, L. S. et al. Oxidized lipids enhance RANKL production by T 28. Takemura, K. et al. Class A scavenger receptor promotes osteoclast lymphocytes: implications for lipid-induced bone loss. Clin. Immunol. 133, differentiation via the enhanced expression of receptor activator of NF-kB 265–275 (2009). (RANK). Biochem. Biophys. Res. Commun. 391, 1675–1680 (2010). 56. Tintut, Y., Morony, S. & Demer, L. L. Hyperlipidemia promotes osteoclastic 29. Parhami, F. et al. Atherogenic diet and minimally oxidized low density potential of bone marrow cells ex vivo. Arterioscler. Thromb. Vasc. Biol. 24, lipoprotein inhibit osteogenic and promote adipogenic differentiation of e6–e10 (2004). marrow stromal cells. J. Bone Miner. Res. 14, 2067–2078 (1999). 57. Gonen, A. et al. Atheroprotective immunization with malondialdehyde- 30. Hamel, P., Abed, E., Brissette, L. & Moreau, R. Characterization of oxidized modiﬁed LDL is hapten speciﬁc and dependent on advanced MDA adducts: low-density lipoprotein-induced hormesis-like effects in osteoblastic cells. Am. implications for development of an atheroprotective vaccine. J. Lipid Res. 55, J. Physiol. 294, C1021–C1033 (2008). 2137–2155 (2014). 31. Brodeur, M. R., Brissette, L., Falstrault, L., Ouellet, P. & Moreau, R. Inﬂuence 58. O’Brien, C. A. et al. IL-6 is not required for parathyroid hormone stimulation of oxidized low-density lipoproteins (LDL) on the viability of osteoblastic of RANKL expression, osteoclast formation, and bone loss in mice. Am. J. cells. Free Radic. Biol. Med. 44, 506–517 (2008). Physiol. Endocrinol. Metab. 289, E784–E793 (2005). 32. Klein, B. Y. et al. Cell death in cultured human Saos2 osteoblasts exposed to 59. Weinstein, R. S., Jilka, R. L., Parﬁtt, A. M. & Manolagas, S. C. The effects of low-density lipoprotein. J. Cell. Biochem. 90,42–58 (2003). androgen deﬁciency on murine bone remodeling and bone mineral density are 33. Parhami, F. et al. Lipid oxidation products have opposite effects on calcifying mediated via cells of the osteoblastic lineage. Endocrinology 138, 4013–4021 vascular cell and bone cell differentiation. A possible explanation for the (1997). paradox of arterial calciﬁcation in osteoporotic patients. Arterioscler. Thromb. 60. Dempster, D. W. et al. Standardized nomenclature, symbols, and units for Vasc. Biol. 17, 680–687 (1997). bone histomorphometry: A 2012 update of the report of the ASBMR 34. Grifﬁn, D. O., Holodick, N. E. & Rothstein, T. L. Human B1 cells in umbilical Histomorphometry Nomenclature Committee. J. Bone Miner. Res. 28,2–17 cord and adult peripheral blood express the novel phenotype CD20+CD27 (2013). +CD43+CD70-. J. Exp. Med. 208,67–80 (2011). 61. Jilka, R. L. Parathyroid hormone-stimulated development of osteoclasts in 35. Holodick, N. E., Vizconde, T., Hopkins, T. J. & Rothstein, T. L. Age-related cultures of cells from neonatal murine calvaria. Bone 7,29–40 (1986). decline in natural igm function: diversiﬁcation and selection of the B-1a Cell 62. Plotkin, L. I. et al. Prevention of osteocyte and osteoblast apoptosis by Pool with Age. J. Immunol. 196, 4348–4357 (2016). bisphosphonates and calcitonin. J. Clin. Invest. 104, 1363–1374 (1999). 36. Holodick, N. E. & Rothstein, T. L. B cells in the aging immune system: time to 63. Montano, E. N. et al. Development and application of a nonradioactive consider B-1 cells. Ann. N. Y. Acad. Sci. 1362, 176–187 (2015). binding assay of oxidized low-density lipoprotein to macrophage scavenger 37. Rothstein, T. L. Natural Antibodies as Rheostats for Susceptibility to Chronic receptors. J. Lipid Res. 54, 3206–3214 (2013). Diseases in the Aged. Front. Immunol. 7, 127 (2016). 38. Que, X. et al. Oxidized Phospholipids are proinﬂammatory and proatherogenic in hypercholesterolemic mice. Nature, https://doi.org/10.1038/ Acknowledgements s41586-018-0198-8 (2018). We thank Michela Palmieri, Han Li, Annick DeLoose, Stu Berryhill, and Kanan Vyas for 39. Manolagas, S. C. From estrogen-centric to aging and oxidative stress: a revised technical assistance, and Jeff Thostenson for help with the statistical analysis. This work perspective of the pathogenesis of osteoporosis. Endocr. Rev. 31,266–300 (2010). was supported by the Biomedical Laboratory Research and Development Service of the 40. Lips, P., Courpron, P. & Meunier, P. J. Mean wall thickness of trabecular bone Veterans Administration Ofﬁce of Research and Development (I01 BX000514 to R.L.J.), packets in the human iliac crest: changes with age. Calcif. Tissue Res. 26,13–17 the National Institutes of Health (P01 AG-13918 to S.C.M.), the University of Arkansas (1978). for Medical Sciences Tobacco Funds and Translational Research Institute (239 G1- 41. Parﬁtt, A. M. in Bone: A Treatise. The Osteoblast and Osteocyte Vol. 1 50893-01; 1UL1 RR-029884 to E.A.) and the University of Arkansas for Medical Sciences (ed. Hall, B. K.) 351–429 (Telford Press and CRC Press, Boca Raton, 1990). Barton Endowment funding (271 G1-51451-99 to E.A.); NIH HL P01-088093 (X.Q., T.S., 42. Piemontese, M. et al. Old age causes de novo intracortical bone remodeling J.L.W.) HL 119828 (J.L.W.), HL R35-135737 and HL P01 136275 (J.L.W.). The work was and porosity in mice. JCI Insight 2, 93771 (2017). also supported by NIH P20GM125503. 10 NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04047-5 ARTICLE Author contributions Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in E.A., S.W., X.Q., F.Y., S.T., J.L.W., and R.L.J. developed the concept, designed the published maps and institutional afﬁliations. experiments and analyzed the data. E.A., S.W., F.Y., and X.Q. perfomed the analyses. R.S. W. provided advice. E.A., R.L.J., S.M., and J.L.W. wrote the manuscript with the con- tribuitions from all authors, who commented on it at all stages. E.A. created the ﬁgures. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, Additional information adaptation, distribution and reproduction in any medium or format, as long as you give Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- appropriate credit to the original author(s) and the source, provide a link to the Creative 018-04047-5. Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless Competing interests: X.Q., S.T., and J.L.W. are co-inventors and receive royalties from indicated otherwise in a credit line to the material. If material is not included in the patents owned by the University of California San Diego on oxidation-speciﬁc antibodies. article’s Creative Commons license and your intended use is not permitted by statutory S.T. currently has a dual appointment at UCSD and as an employee of Ionis regulation or exceeds the permitted use, you will need to obtain permission directly from Pharmaceuticals. J.L.W. is a consultant to Ionis Pharmaceuticals. S.C.M. is a co-founder the copyright holder. To view a copy of this license, visit http://creativecommons.org/ and owns equity of Radius Health Inc. The remaining authors declare no competing licenses/by/4.0/. interests. Reprints and permission information is available online at http://npg.nature.com/ © The Author(s) 2018 reprintsandpermissions/ NATURE COMMUNICATIONS (2018) 9:2193 DOI: 10.1038/s41467-018-04047-5 www.nature.com/naturecommunications 11 | | |
Nature Communications – Springer Journals
Published: Jun 6, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera