JOURNAL OF BONE AND MINERAL RESEARCHVolume 24, Number 3, 2009Published online on October 13, 2008; doi: 10.1359/JBMR.081006� 2009 American Society for Bone and Mineral Research

Activation of Renin–Angiotensin System Induces OsteoporosisIndependently of Hypertension

Yutaro Asaba,1,2 Masako Ito,3 Toshio Fumoto,1 Ken Watanabe,1 Ryoji Fukuhara,1 Sunao Takeshita,1

Yuji Nimura,2 Junji Ishida,4 Akiyoshi Fukamizu,4,5 and Kyoji Ikeda,1

ABSTRACT: Hypertension and osteoporosis are two major age-related disorders; however, the underlyingmolecular mechanism for this comorbidity is not known. The renin–angiotensin system (RAS) plays a centralrole in the control of blood pressure and has been an important target of antihypertensive drugs. Using achimeric RAS model of transgenic THM (Tsukuba hypertensive mouse) expressing both the human reninand human angiotensinogen genes, we showed in this study that activation of RAS induces high turnoverosteoporosis with accelerated bone resorption. Transgenic mice that express only the human renin gene werenormotensive and yet exhibited a low bone mass, suggesting that osteoporosis occurs independently of thedevelopment of hypertension per se. Ex vivo cultures showed that angiotensin II (AngII) acted on osteoblastsand not directly on osteoclast precursor cells and increased osteoclastogenesis-supporting cytokines,RANKL and vascular endothelial growth factor (VEGF), thereby stimulating the formation of osteoclasts.Knockdown of AT2 receptor inhibited the AngII activity, whereas silencing of the AT1 receptor paradox-ically enhanced it, suggesting a functional interaction between the two AngII receptors on the osteoblasticcell surface. Finally, treatment of THM mice with an ACE inhibitor, enalapril, improved osteoporosis andhypertension, whereas treatment with losartan, an angiotensin receptor blockers specific for AT1, resulted inexacerbation of the low bone mass phenotype. Thus, blocking the synthesis of AngII may be an effectivetreatment of osteoporosis and hypertension, especially for those afflicted with both conditions.J Bone Miner Res 2009;24:241–250. Published online on October 13, 2008; doi: 10.1359/JBMR.081006

Key words: osteoporosis, hypertension, renin–angiotensin system, bone resorption, osteoclast

INTRODUCTION

HYPERTENSION AND OSTEOPOROSIS are two major age-related disorders, which together account for signifi-

cant morbidity and mortality in the elderly by predisposingto cardiovascular diseases, fragility fractures, and theirsequelae, respectively.(1) Both hypertension and osteopo-rosis are multifactorial disorders, in which genetic andlifestyle factors contribute to the pathogenesis.(2,3) Epide-miological studies suggest a link between osteoporoticfractures and hypertension,(4) stroke,(5–7) or cardiovascularevents.(8,9) High blood pressure is associated with increasedbone loss at the femoral neck,(4) and low BMD has beenshown to be strongly associated with deaths from stroke.(5)

However, the cellular and molecular mechanisms under-lying the comorbidity of hypertension and osteoporosiswith aging are not known, and the two disorders have beentreated separately and additively by antihypertensive andanti-osteoporosis drugs, respectively.

The renin–angiotensin system (RAS) is an endocrinesystem that governs body fluid and electrolyte balance and

blood pressure.(10) In the classic endocrine RAS, angio-tensinogen produced in the liver is sequentially cleaved bypeptidases to form the biologically active octapeptideangiotensin II (AngII). Renin produced by the juxtaglo-merular apparatus of the kidney and secreted into thecirculation cleaves angiotensinogen to the inactive deca-peptide angiotensin I, which is cleaved by angiotensin-converting enzyme (ACE) to generate AngII. The initialreaction between the enzyme renin and the substrate an-giotensinogen is the rate-limiting step of the RAS, forwhich strict species specificity exists. We have previouslyestablished transgenic THM (Tsukuba hypertensive mouse),which by expressing both human renin and human angio-tensinogen genes and using endogenous murine ACE andAngII receptors, successfully reproduces a chimeric RAScascade with increased AngII production and hypertension.THM not only showed that RAS plays an important rolein the pathogenesis of hypertension(11) but has been widelyused for studying the role of RAS in various pathologicalconditions, such as pregnancy-associated hypertension.(12)

The RAS has been an important target of antihypertensivedrugs,(13) especially ACE inhibitors(14) and angiotensin re-ceptor blockers (ARBs).

1Department of Bone and Joint Disease, Research Institute, National Center for Geriatrics and Gerontology, Obu, Aichi, Japan;2Department of Surgery, Nagoya University Graduate School of Medicine, Nagoya, Japan; 3Department of Radiology, Nagasaki Uni-versity School of Medicine, Nagasaki, Japan; 4Life Sciences and Bioengineering, Graduate School of Life and Environmental Sciences,University of Tsukuba, Ibaraki, Japan; 5Aspect of Functional Genomic Biology, Center for Tsukuba Advanced Research Alliance(TARA), University of Tsukuba, Ibaraki, Japan.

The authors state that they have no conflicts of interest.

241

In an attempt to explore the mechanistic link betweenhypertension and osteoporosis, we studied the skeletalphenotype of THM mice, focusing our study on the mech-anism underlying the low bone mass phenotype, whether itis caused by osteoclast or osteoblast dysfunction, how RASimpinges on bone cell activities, whether it reflects systemicor local activation of RAS, and finally, from a therapeuticperspective, whether antihypertensive drugs that targetRAS are potentially an effective remedy to treat osteopo-rosis simultaneously.

MATERIALS AND METHODS

Reagents

AngII was purchased from the Peptide Institute (Osaka,Japan), losartan potassium and enalapril were from LTKLaboratories (St Paul, MN, USA), and propranolol hy-drochloride was from Sigma (St Louis, MO, USA). 1a,25-Dihydroxyvitamin D3 [1a,25(OH)2D3] was purchased fromNacalai Tesque (Kyoto, Japan). Mouse RANKL andmacrophage-colony stimulating factor (M-CSF) were pur-chased from R&D Systems (Minneapolis, MN, USA).

Generation of THM and animal experiments

THM mice were produced by mating doubly transgenicmice expressing human angiotensinogen gene (ANG/ANG)and those expressing the human renin gene (RN/RN),(11)

which were all on the C57BL/6J genetic background. Age-and sex-matched C57BL/6J wildtype mice were used forcontrols. THM were also produced by mating singletransgenic male mice expressing human angiotensinogen(ANG/+) and female transgenic mice expressing humanrenin gene (RN/+), because the mating of female ANG/+and male RN/+ did not give birth to healthy offspring,because of pregnancy-associated hypertension caused bythe combined action of maternal angiotensinogen andplacenta-derived renin.(12) In this case, nontransgenic lit-termates served as a control. The identification of fourdifferent genotypes (ANG/RN or THM, ANG/+, RN/+, +/+)was performed by Southern and PCR analyses of genomicDNA extracted from tail, using the following primers andprobes: 59-CTGCAGGCTTCTACTGCTC-39 and 59-GGGCCCCAGAACACAGTG-39 for angiotensinogen; 59-CACATCCACTCACTGTCCTTGTAC-39 and 59-GAGGGCAGGATGGTAATGCAGTC-39 for renin, a 200-bpApaI/PstI fragment that corresponds to a 59-untranslatedregion of the angiotensinogen gene, and a 431-bp RsaI/EcoRI fragment within the renin gene coding sequence ascDNA probes, respectively.

Mice were raised under standard laboratory conditionsat 24 ± 28C and 50–60% humidity and allowed free accessto tap water and commercial standard rodent chow (CE-2)containing 1.20% calcium, 1.08% phosphate, and 240IU/100 g vitamin D3 (Clea Japan). Enarapril, losartan, andpropranolol were dissolved in distilled water and admin-istered to mice in drinking water at the concentrations of5 mg/1.5 dl, 17 mg/dl, and 50 mg/dl for 4 wk. Blood pressurewas measured by noninvasive tail-cuff method as describedpreviously,(11) using Model BP-98A (Softron, Tokyo, Japan).

Urine was collected during the final 24 h, and blood sampleswere centrifuged to obtain the plasma and serum.

All experiments were performed in accordance withNCGG’s ethical guidelines for animal care, and the ex-perimental protocols were approved by the animal carecommittee of NCGG.

Blood and urine biochemistry

Serum and urinary calcium, phosphate, and creatinineconcentrations were determined by an autoanalyzer (Hitachi7170; Hitachi, Tokyo, Japan). Urinary deoxypyridinoline(DPD) was measured with a PYRILINKS-D assay kit(Metra Biosystems) and was corrected for creatinine. PlasmaAngII and serum osteocalcin (OC) and intact PTH concen-trations were measured using an RIA kit (MitsubishiChemical Medience, Tokyo, Japan), a mouse osteocalcinIRMA kit (Immutopics, San Clemente, CA, USA), and amouse intact PTH ELISA kit (Immutopics), respectively.

Bone and histological analyses

mCT scanning was performed on proximal tibias usingmCT-40 (SCANCO Medical) with a resolution of 12 mm,and microstructure parameters were calculated three di-mensionally as described previously.(15) Tibia were fixed in4% paraformaldehyde for bone histomorphometry. Eachsample was sectioned and stained for TRACP. Bone his-tomorphometry was performed on undecalcified sections,with calcein and tetracycline double labeling (administeredsubcutaneously at a 2-day interval), and histomorpho-metric parameters were measured at the Ito Bone ScienceInstitute (Niigata, Japan).

Cell culture and ex vivo osteoclastogenesis assay

Bone marrow cells were isolated from the tibias andfemurs of 8-wk-old male THM and nontransgenic controlmice and cultured in aMEM supplemented with 10% FBS,1% antibiotics (streptomycin and penicillin), and 10%CMG14–12 culture supernatant for 3 days, and adherentbone marrow macrophages (BMMs) were used as osteo-clast precursors, also as previously described.(16) BMMswere treated with M-CSF (50 ng/ml) and various con-centrations of RANKL for 3 days, fixed in 4% para-formaldehyde, and stained for TRACP. The number ofmultinucleate (at least three nuclei), TRACP+ cells wascounted as osteoclasts.

Osteoblasts were isolated from newborn mouse calvariaas described previously,(17) and co-cultures with bonemarrow cells were performed in the presence of 1028 M1a,25(OH)2D3, according to the established technique.(18)

The number of TRACP+ cells with more than three nucleiwas counted under light microscopy.

RNA interference using a retrovirus vector

The expression of AT1 and AT2 receptors in primaryosteoblastic cells was silenced by using specific hairpinsiRNAs and the pSilencer 5.1 Retro System (Ambion,Austin, TX, USA). Short hairpin RNA sequences targetingthe coding region of AT1a and AT2 were designed using

242 ASABA ET AL.

the Target Finder and Design Tool (Ambion). Selectedsequences did not show near-exact matches to any otherknown sequence on a BLAST search, confirming theirsequence specificity. Synthetic duplexed-deprotected siRNAswere purchased from Sigma and cloned into a pSilencer5.1-U6 vector. The 21-bp target sequences used for AT1aand AT2 were AATTCAAGATGACTGCCCCAG and AATGAGTCCGCCTTTAATTGC, respectively. The pSilencer5.1-U6 Retro Scrambled (Ambion) was used as thecontrol.

Retroviral vectors were transfected into GP2-293 pack-aging cells with pVSV-G (Clonthech Laboratories). Re-combinant retrovirus was infected into primary osteoblastsisolated from newborn mouse calvaria for 24 h, and after 3days of selection with puromycin (1.6 mg/ml), the cells wereused for osteoclastogenesis assays in co-cultures withBMMs and also for RNA extraction.

RNA extraction and RT-PCR

Total RNA was isolated from various cells and bone withTRIzol Reagent (Invitrogen, San Diego, CA, USA), andgene expression was assessed by RT-PCR at 948C, 30 s;558C for 1 min; and 728C for 1 min with the followingprimer sets; 59-GCATCATCTTTGTGGTGGG-39 and 59-GAAGAAAAGCACAATCGCC-39 for AT1 (35 cycles);59-AGTGCAAACTGGCATGGG-39 and 59-AAAACGCCTGGAATCTGA-39 for AT2 (35 cycles); 59-TGAAGACACACTACCTGACT-39 and 59-AAGATAGTCTGTAGGTACGCTT-39 for RANKL (27 cycles); 59-GACTTCATGCCAGATTGCC-39 and 59-GGTGGCTTTAGGGTACAGG-39 for M-CSF (22 cycles); 59-CCACTCTTATACGGACAGCT-39 and 59- TCTCGGCATTCACTTTGGTC-39 for OPG (28 cycles); 59-TCGGTGTCCTCTTGCTGTCC-39 and 59-TGGCGGAGTGTCTTTATGCTG-39 for SDF-1a (27 cycles); 59-CCTCACCATCATCCTCACTG-39 and 59-CCACTTCTTCTCTGGGTTGG-39 for CCL5 (27 cycles); 59-AGGACGCTAGCCTTCACTCCAA-39 and 59-GTGTGAAGGTTCAAGGCTGCAGA-39 for CCL8 (27 cycles); 59-GGGATCCACCATGAACTTTCTGCTCTCT-39 and 59-TTGCGGCCGCTCACCGCCTTGGCTTGTCACA-39 for VEGF (28cycles); 59-CCGATGAGTACTTGGACACCT-39 and 59-TTTCCAAGGAGGGTGCAGTT-39 for RANK; 59-ACTTTGTCAAGCTCATTTCC-39 and 59-TGCAGCGAACTTTATTGATG-39 for GAPDH (22 cycles); 59-CATCGTGGGCCGCTCTAGGCACCA-39 and 59-CGGTTGGCCTTAGGGTTCAGGGGG-39 for b-actin (26cycles).

For quantitative RT-PCR, total RNA was reverse tran-scribed using Superscript III (Invitrogen), and sampleswere analyzed using PowerSYBR Green PCR master mixand an ABI7300 real-time PCR system (Applied Biosys-tems). The sequences of the primers used for the real-timePCR were as follows: RANKL, 59-AGGATGAAACAAGCCTTTCAG-39 and 59-ACCATGAGCCTTCCATCATAG-39; GAPDH, 59-ATTGTCAGCAATGCATCCTG-39 and 59-ATGGACTGTGGTCATGAGCC-39. Theamount of target mRNA was corrected by that of GAPDHmRNA.

Statistical analysis

Data are expressed as the mean ± SD. Statistical analysiswas performed using unpaired Student’s t-test or ANOVAfollowed by Dunnett’s test or Student-Newman-Keuls test.Values were considered statistically significant at p < 0.05.

RESULTS

Activation of RAS induces osteopenia

THM mice were generated by mating doubly transgenicmice expressing the human angiotensinogen gene (ANG/ANG) and those expressing the human renin gene (RN/RN). THM (ANG/RN) mice, which produce human angi-otensinogen and human renin in the same individual, ex-hibit elevated serum AngII concentrations and hyperten-sion.(11) To examine how activation of RAS affects bonemetabolism in vivo, trabecular bone structure of THM micewas analyzed by mCT scanning of tibial metaphysis, usingsex- and age-matched C57BL/6J wildtype mice as controls.Representative mCT images of the 6-mo-old male THMmouse shown in Fig. 1A revealed marked osteopenia with asubstantial reduction in the 3D bone volume fraction (BV/TV), compared with an age- and sex-matched wildtypemouse. The decrease in bone volume in THM mice wasmore marked in males than in females (Fig. 1B).

To further confirm the low bone mass phenotype ofTHM mice, four different genotypes were generated bymating renin and angiotensinogen single-transgenic mice.As shown in Fig. 1C, THM mice exhibited markedly ele-vated plasma AngII concentrations and hypertension atthe age of 3 mo, and the bone volume fraction at theproximal tibia was significantly reduced compared withnontransgenic littermates, whereas single transgenic miceexpressing human angiotensinogen gene only (ANG/+) ex-hibited normal bone mass.

Interestingly and importantly, single transgenic micethat produce only human renin (RN/+) remained normo-tensive and yet exhibited significantly reduced bone mass,with marginally elevated AngII levels (Fig. 1C), suggestingthat the presence of hypertension is not a prerequisite forthe development of osteoporosis and that the activation ofRAS induces low bone mass phenotype independently ofhypertension.

Detailed microstructural analysis of the tibial metaphy-sis of THM mice showed that trabecular thickness de-creased significantly, whereas the structure model index(SMI) increased (Fig. 1D), suggesting that the trabeculaeof THM mice had become thinner and converted from aplate-like structure to a more fragile rod-like structure.

Osteopenia in THM mice is caused by highbone turnover

The cause(s) of osteopenia in THM mice was furtherexplored at the tissue and cell levels by histological andbiochemical analyses. TRACP staining of bone sectionsshowed an increased number of TRACP+ osteoclasts (Fig.2A), suggesting elevated bone resorption in THM. Detailed

RENIN–ANGIOTENSIN SYSTEM AND BONE METABOLISM 243

histomorphometric analysis at the proximal tibia showedthat histological indices of bone resorption, the numberof osteoclasts (N.Oc/BS), bone surface area covered byosteoclasts (Oc.S/BS), and eroded surface (ES/BS), wereall significantly elevated in THM compared with non-transgenic control mice (Fig. 2B).

Bone formation rate (BFR) was also significantly in-creased in THM mice (Fig. 2B), which together with theresults of accelerated bone resorption suggests that boneturnover is accelerated in THM mice. Whereas osteoblastsurface (Ob.S/BS) remained unchanged, the mineral ap-position rate (MAR) was significantly increased in THMmice (Fig. 2B), suggesting that the mineralizing function ofmature osteoblasts was increased in THM mice, whereasthe recruitment of new osteoblasts was not affected.

Biochemical analysis of urine and serum samples ofTHM mice indicated that both urinary excretion of DPD, amarker of bone resorption, and the serum osteocalcin (OC)concentration, a product of osteoblasts, were significantlyhigher in THM mice than in nontransgenic littermates (Fig.2C), which is consistent with a high bone turnover state.Serum calcium and PTH levels did not differ between the

THM and nontransgenic control mice (Fig. 2C and data notshown), suggesting that the activation of the RAS in THMmice affects bone remodeling locally but not as a secondaryconsequence of systemic alterations in calcium metabolism.

Osteoblasts as a target of AngII action in bone

As the first step to study the mechanism by which theactivated RAS leads to a high bone turnover state withelevated osteoclastic and osteoblastic activities, the ex-pression of AngII receptors, AT1 and AT2, was studied inosteoclast and osteoblast lineage cells. As shown in Fig.3A, both AT1 and AT2 receptors were expressed in pri-mary osteoblasts. The expression of AT1 mRNA seemedconstitutive, whereas that of AT2 increased along with thestage of osteoblast differentiation. In contrast, osteoclastlineage cells, specifically BMMs and pre-osteoclasts (pOCs),only weakly expressed the AT1 receptor, whereas mRNAfor AT2 receptor was not detected (Fig. 3A).

We studied whether hematopoietic precursor cells in thebone marrow of THM and nontransgenic control micediffered in their potential to differentiate into osteoclasts.

FIG. 1. Osteoporosis in THMmice expressing human reninand angiotensinogen genes. (A)Representative 3D images bymCT of trabecular bone at theproximal tibias of 6-mo-oldwildtype (WT) and THM malemice. BV/TV, 3D bone volumefraction per tissue volume inpercent. (B) Comparison of 3Dbone volume fraction (BV/TV)between 6-mo-old WT andTHM male and female mice.*p < 0.05, **p < 0.01 vs. WT(n = 8 each group). (C) Bloodpressure (BP), plasma angio-tensin II (AngII) concentra-tions, and 3D bone volumefraction (BV/TV) in four geno-types, nontransgenic controlmice (C), ANG/+ (single trans-genic mice expressing humanangiotensinogen gene), RN/+(single transgenic mice express-ing human renin gene), anddoubly transgenic THM malemice (3 mo old), which weregenerated by mating femaleANG/+ and male RN/+. **p <0.01 vs. control (n = 8 eachgroup). (D) Microstructureanalysis by mCT of trabecularbone at the proximal tibias of3-mo-old nontransgenic con-trol (C) and THM male mice.BV/TV, 3D bone volume frac-tion per tissue volume; Tb.N,trabecular number; Tb.Th, tra-becular thickness; Conn-Dens,connectivity density; SMI, struc-ture model index. *p < 0.05,**p < 0.01 (n = 6 each group).

244 ASABA ET AL.

Fig. 1 live 4/C

To this end, BMMs were isolated as osteoclast precursorcells from THM and nontransgenic littermates, and theirdifferentiation potential in response to RANKL andM-CSF was assessed in ex vivo cultures. As shown in Fig.3B, BMMs derived from THM and nontransgenic controlmice exhibited the same dose–response curve in terms ofthe number of TRACP+ multinucleated cells generated inresponse to increasing doses of RANKL (Fig. 3B) orM-CSF (data not shown), suggesting that there was nofunctional difference in the hematopoietic ‘‘seed cells’’ forosteoclasts between control and THM.

We also examined the direct response of BMMs to ex-ogenous AngII. As shown in Fig. 3C, AngII itself, between10210 and 1026 M, had no activity of inducing osteoclastdifferentiation or exerted any stimulatory effect on osteo-clastogenesis triggered by a lower (permissive) concentra-

tion of RANKL. In contrast, AngII stimulated the forma-tion of osteoclasts in the co-culture of calvaria-derivedprimary osteoblasts and BMMs in a dose-dependent man-ner between 1028 and 1026 M (Fig. 3D). Taken together,these results suggest that AngII stimulates osteoclasto-genesis by acting on osteoblastic cells (i.e., ‘‘the soil cells’’)and not through a direct action on hematopoietic ‘‘seed cells.’’

AngII increases RANKL and vascular endothelialgrowth factor in osteoblasts

Osteoblasts modulate osteoclast differentiation by pro-ducing both positive and negative regulators, most notablyRANKL and osteoprotegerin (OPG), respectively.(19)

To gain further insight into the mechanism by whichAngII, acting mainly on osteoblasts, stimulates osteoclast

FIG. 2. High bone turnoverin THM mice. (A) Bone sec-tions of THM mice exhibitmore TRACP+ osteoclasts thannontransgenic control mice. (B)Results of histomorphometryat the tibial metaphysis. N.Oc,number of osteoclasts; Oc.S,osteoclast surface; ES, erodedsurface; Ob.S, osteoblast sur-face; BFR, bone formation rate;MAR, mineral apposition rate.Data are normalized for thebone surface (BS). *p < 0.05,**p < 0.01 (n = 5 each group).(C) Urinary DPD excretion,corrected for creatinine (Cr),serum OC, and serum PTHconcentrations. *p < 0.05, **p <0.01 (n = 5 each group).

RENIN–ANGIOTENSIN SYSTEM AND BONE METABOLISM 245

Fig. 2 live 4/C

differentiation, the expression of osteoblast-derived reg-ulators of osteoclastogenesis was investigated in primaryosteoblasts. As shown in Fig. 4A, treatment with AngII,alone or in combination with 1a,25(OH)2D3, increasedthe expression of RANKL and vascular endothelialgrowth factor (VEGF) mRNA levels in osteoblasts,whereas the M-CSF, OPG, SDF-1, and CCL5 mRNAlevels remained unaltered. The increase in RANKLmRNA by AngII was confirmed by quantitative RT-PCR(Fig. 4B). Thus, it is suggested that RANKL, and perhapsalso VEGF, mediates the stimulation of osteoclastogenesis-supporting potential of osteoblastic cells by AngII.

To further substantiate this concept, RNA was extractedfrom the bone of THM and nontransgenic control mice,and gene expression was examined by RT-PCR. The re-sults indicate that RANKL expression in bone was sub-stantially increased in THM mice compared with controlmice, and VEGF expression also was increased modestly(Fig. 4C). It was evident that ACE mRNA was expressedin bone and was decreased in THM mice (Fig. 4C).

Functional interaction between AT1 and AT2in osteoblasts

To examine whether functional interaction exists be-tween AT1 and AT2 receptors and to determine the rela-tive contribution of the two receptors for transducing theosteoclastogenesis-supporting function of AngII in osteo-blasts, the expression of each of the receptors was knockeddown with siRNA in primary osteoblasts in culture. Asshown in Fig. 4D, the expression level of AT1 and AT2receptor was reduced by 90% and 98%, respectively, byspecific siRNA. In AT1-knockdown osteoblasts, the stim-ulatory effect of AngII on osteoclast formation wassomewhat enhanced (Figs. 4E and 4F). In AT2-knockdownosteoblasts, in contrast, the osteoclastogenic potential wasmarkedly attenuated (Figs. 4E and 4F).

Taken together, it is suggested that AT2 is the majortransducing receptor for AngII in osteoblasts. In view ofthe findings that the inhibition of AT1 alone resulted instimulation of osteoclastogenesis, it is conceivable that,under functional knockdown of the AT1 receptor, signal-ing through AT2 may be facilitated.

ACE inhibition improves but ARB exacerbatesosteoporosis of THM mice

Because the AngII produced through activation of theRAS stimulates the expression of osteoclastogenic cyto-kines in osteoblasts, thereby leading to a high turnoverosteoporosis, it is reasonable to hypothesize that blockadeof the RAS may ameliorate osteoporosis and hypertension.To test this hypothesis, THM mice were treated orally for4 wk with losartan, an ARB specific to AT1, and the effectson the low bone mass were determined by mCT scanningof proximal tibia. As expected, losartan was effective intreating hypertension. However, contrary to our expecta-tions, the administration of losartan resulted in a furtherdecrease in bone mass (Fig. 5A). Because osteoblastsexpress both AT1 and AT2 receptors (Fig. 3A), we

FIG. 3. Angiotensin II acts on osteoblasts that express both AT1and AT2 receptors. (A) Expression of AngII receptors, AT1 andAT2, was examined by RT-PCR using specific primer sets. b-actinwas used as the loading control. RNA was extracted from primaryosteoblasts (OB) from newborn mouse calvaria, BMMs, and pre-osteoclasts (pOC) and mature osteoclasts (mOC) generated fromBMMs after treatment with M-CSF and RANKL for 2 and 4 days,respectively. RNA from brain and liver was used as positive con-trols of AT2 and AT1 mRNA, respectively. (B) BMMs derivedfrom THM mice generated the same number of TRACP+ osteo-clasts as from the bone marrow of nontransgenic control mice inresponse to increasing doses of RANKL. Ex vivo culture wasperformed in the presence of M-CSF (50 ng/ml) and the indicatedconcentrations of RANKL, and the number of TRACP+ cells withmore than three nuclei was counted. (C) Ex vivo culture wasperformed in the presence of M-CSF (50 ng/ml), RANKL (25 ng/ml), and the indicated concentrations of AngII, and the number ofTRACP+ cells with more than three nuclei was counted. Note thatAngII itself (at 1026 M) had no capacity to induce osteoclast dif-ferentiation and that AngII (at 10210–1026 M) did not potentiateosteoclast formation triggered by RANKL. (D) Co-cultures ofmouse calvaria-derived osteoblasts and BMMs were performed inthe presence of 1028 M 1a,25(OH)2D3 and the indicated dosesof AngII, and TRACP+ cells with more than three nuclei werequantified. *p < 0.05, **p < 0.01 (n = 5 each group).

246 ASABA ET AL.

considered the possibility that blockade of AT1 alonesomehow activated signaling through AT2 receptor by thestill high levels of circulating AngII.

To investigate this concept, we next treated THM micewith enalapril, an ACE inhibitor, which inhibits the con-version of AngI to the active hormone AngII, therebyinhibiting signaling through both AT1 and AT2. As shownin Fig. 5B, enalapril corrected the low bone mass pheno-type and hypertension of THM mice. These results suggestthat, to combat the deleterious effects of RAS in bonetissue, treatments that inhibit the synthesis of AngII areneeded, because there is functional cross-talk betweenAT1 and AT2 receptors on the osteoblastic cell surface.Furthermore, the findings that the two antihypertensiveagents, losartan and enalapril, had opposite effects on bonemass are consistent with our contention that the RAS isinvolved in the regulation of bone metabolism indepen-dently of its effect on blood pressure.

Finally, to address the involvement of the sympatheticnervous system in the low bone mass phenotype of THM,the effects of a b-adrenergic antagonist, propranolol, wereexamined. As shown in Fig. 5C, treatment with propranololcaused a marked elevation of blood pressure in THM butnot in wildtype mice, and the bone volume tended to fur-ther decrease in THM, although not to the level of statis-tical significance.

DISCUSSION

We showed in this study that activation of RAS inducesnot only hypertension but also osteopenia with micro-structural deterioration, reminiscent of osteoporosis, sug-gesting that aberrant activation of RAS may contribute tothe co-occurrence of the hypertensive disorders and oste-oporosis, which is often seen with advancing age.

FIG. 4. Functional interaction between AT1 and AT2 receptors in osteoblasts. (A and B) Expression of molecules involved in theregulation of osteoclastogenesis by RT-PCR (A) and quantitative RT-PCR (B). RNA was extracted from calvaria-derived primaryosteoblasts cultured in the presence or absence of AngII at 1026 M and 1a,25(OH)2D3 at 1028 M, and mRNA levels were analyzed byusing specific primer sets. *p < 0.01, **p < 0.0001 (n = 3 each group). (C) Expression of the indicated regulators of osteoclast differ-entiation was examined by RT-PCR in the bone of nontransgenic control and THM mice (from three mice each). GAPDH was used as apositive control. (D) Knockdown of AT1 and AT2 mRNA by siRNA in primary osteoblasts derived from newborn mouse calvaria. (E)AT1- and AT2-knockdown osteoblasts were co-cultured with BMMs in the presence of 1028 M 1a,25(OH)2D3 and 1026 M AngII for 7days, and TRACP staining was performed. Representative dishes are shown. (F) AT1- and AT2-knockdown osteoblasts were co-culturedwith bone marrow cells in the presence of 1028 M 1a,25(OH)2D3, with (closed bars) or without (open bars) 1026 M AngII for 7 days, andthe number of TRACP+ cells with more than three nuclei was counted. **p < 0.01 (n = 5 each group).

RENIN–ANGIOTENSIN SYSTEM AND BONE METABOLISM 247

Fig. 4 live 4/C

Two lines of evidence, however, suggest that the pres-ence of hypertension per se is not the determinant of thelow bone mass phenotype. First, renin/+ single transgenicmice were normotensive and yet showed a significant re-duction in bone volume compared with nontransgeniccontrol mice. Although there is species specificity for thereaction between renin and angiotensinogen, human reninoverproduced in renin/+ transgenic mice is thought topossess a residual capacity to cleave the endogenous,mouse angiotensinogen, which is evidenced by a modestbut significant increase in circulating AngII concentrationsin renin/+ transgenic mice (Fig. 1C). It may follow that thesensitivity of bone tissue to this marginally elevated AngIIis higher than that of the pressor response. Alternatively, inlight of the current understanding that, in addition to theclassic, endocrine RAS, many tissues possess a local RASthat serves important physiological functions(20,21) and thefindings that ACE is in fact expressed in bone, as shown inthis study (Fig. 4C) and reported by others,(22,23) locallyproduced AngII may be sufficient to cause osteopenia.

Second, treatment of THM with an ARB, losartan, im-proved hypertension and yet exacerbated the osteopenicphenotype. Thus, it is plausible that local RAS contributesto the development of osteoporosis, independently of itssystemic effect on blood pressure.

Histomorphometric and biochemical analyses showedTHM to be in a state of high bone turnover, which is likelyto be responsible for the development of osteoporosis.Both the AT1 and AT2 receptors were expressed in oste-oblasts, whereas BMMs and pre-osteoclasts showed only aweak expression of AT1, but not AT2, on RT-PCR anal-ysis. There have been several reports on the effects ofAngII on bone cell function in vitro, including inhibition ofosteoblastic differentiation and mineralization,(24) stimu-lation of proliferation and collagen synthesis in osteo-blasts,(25) and stimulation of osteoclastic bone resorp-tion.(22) Hatton et al.(22) showed that, although AngII hadno effect on osteoclast formation or bone resorption byisolated osteoclasts, it stimulated bone resorption in co-cultures with bone cells. In agreement with this study, wedid not observe any effect of AngII on osteoclast precursorcells or mature osteoclasts. In fact, our results suggest thatAngII acts on osteoblasts to increase the expression ofRANKL and, by providing the pro-osteoclastogenic cyto-kine locally, stimulates the formation of osteoclasts indi-rectly. AngII also increased the expression of VEGF, whichwe have recently shown to stimulate osteoclastogenesis atleast in part through the VEGF receptor 1 (Flt-1) ex-pressed on hematopoietic cells.(26) Chemokines producedby osteoblasts, such as SDF-1 and CCL5, have been knownto regulate osteoclastogenesis(27,28); however, we did notfind any alteration in the expression of these chemokinesby AngII. Further studies are needed to examine the in-volvement of vascular endothelial cells in bone, in relationto VEGF and RAS, in the osteoporotic phenotype of THM.

Because AngII is known to stimulate the release ofnorepinephrine from sympathetic nerve endings,(29) andbecause sympathetic tone has been shown to be an im-portant regulator of bone metabolism,(30) we attempted toaddress the role of the sympathetic tone in the low bonemass phenotype of THM. Contrary to our expectations,treatment of THM with a b-adrenergic blocker, propran-olol, resulted in a further elevation of blood pressure butnot in wildtype mice. Although the reason is not clear, wesuspect that the blockade of b1 and b2 adrenergic recep-tors by propranolol caused relative stimulation of thea1 receptor in the blood vessels, leading to increased vas-cular resistance and further elevation of blood pressure,specifically in the setting of the elevated AngII of THM.Under these circumstances, the osteopenia of THM did notimprove, but rather, tended to worsen after treatment withpropranolol. Further studies are needed to clarify thefunctional interaction between the RAS and sympatheticnervous system in the regulation of bone remodeling.

AngII binds to two G protein–coupled receptors: AT1and AT2. Although the vasopressive and aldosterone-secreting actions of AngII are mainly mediated throughAT1, AT2 serves important physiological roles, and func-tional interactions exist between the two receptors.(31,32)

Knockdown experiments with siRNA are consistent with

FIG. 5. Angiotensin receptor blockade exacerbates and ACEinhibition ameliorates osteoporosis in THM mice. Two-month-oldmale THM mice were treated orally with losartan (A), an angio-tensin II receptor blocker specific for AT1, or enalapril (B), anangiotensin-converting enzyme blocker, for 4 wk, and arterialblood pressure and 3D bone volume fraction (BV/TV) at theproximal tibia were determined by the noninvasive tail-cuffmethod and mCT, respectively. Nontransgenic littermates servedas control (C). *p < 0.05, **p < 0.01 (n = 5 each group). (C) Theeffects of propranolol, a b-blocker, were examined in 5-mo-oldmale THM and age- and sex-matched wildtype (WT) mice. *p <0.05, **p < 0.01 (n = 3–5 each group).

248 ASABA ET AL.

the notion that the AngII action on osteoblasts in terms ofstimulating osteoclastogenesis is mainly mediated throughthe AT2 receptor, because suppression of AT2 expressionmarkedly inhibited osteoclast formation by AngII. On theother hand, knockdown of AT1 resulted in a further in-crease in osteoclast formation by AngII, suggesting thatAT1 may exert an inhibitory effect on AT2. The functionsof AT1 and AT2 are in many cases counter-regulatory toeach other, and these findings in osteoblasts are consistentwith this notion. Further studies are needed to dissect thesignaling pathways downstream of each receptor in osteo-blasts and their interaction with sex steroids, in view of thesubstantial difference in the bone phenotype between maleand female THM in this study (Fig. 1B) and of the knownsex difference in AT2 expression.(33)

The cross-talk between AT1 and AT2 receptors was alsoobserved in pharmacological experiments in vivo. It is in-triguing that treatment with losartan exacerbated osteo-porosis despite the amelioration of hypertension and thatinhibition of AngII synthesis with an ACE inhibitor wasneeded to correct the osteoporotic phenotype. It has re-cently been shown in myopathy model mice that losartanrestores muscle regeneration by blocking TGF-b signalingin skeletal muscle,(34) and the action of losartan of reducingbone volume may be related to the inhibition of TGF-bsignaling in bone.

It has been proposed that inhibition of RAS with ACEinhibitors or ARBs has beneficial effects beyond thoseresulting from lowering blood pressure alone, such asrenoprotective effects and cardiovascular outcomes.(35)

Epidemiological studies indicate that patients who haveundergone stroke have an increased risk of hip frac-ture.(6,7,36) Our results hint on the utility of RAS-basedmedicine in the management of elderly people in whomhypertension/cardiovascular events and osteoporosis/fra-gility fracture are the two most important comorbidities. Itis to be noted that a cross-sectional study of Chinese pop-ulation showed an association of ACE inhibitor use withhigher BMD.(37)

In conclusion, this study showed a common causalmechanism in the pathogenesis of hypertension and oste-oporosis. Given the high prevalence of the two disorders inthe aging population, RAS may provide a crucial target fortherapeutic intervention.

During the course of preparing this manuscript, we noticedan online publication reporting that AngII stimulatedRANKL expression in osteoblasts and thereby osteoclasto-genesis and that infusion of AngII in ovariectomized ratsincreased bone resorption and decreased BMD,(38) whichsupports our contention that AngII stimulates osteoclasto-genesis through RANKL. In this study, the effects of AngIIinfusion to increase bone resorption and decrease BMD wereshown only in the setting of ovariectomy in rats. In apparentcontrast to our findings that losartan caused an exacerbationof the low bone mass phenotype of THM, they showed thattreatment of ovariectomized spontaneously hypertensive rat(SHR) with olmesartan, another ARB, ameliorated the el-evated bone resorption and reduced BMD.(38) Furtherstudies are thus needed to clarify the pathophysiological rolesof the AT1 receptor in bone metabolism.

ACKNOWLEDGMENTS

The authors thank Kumi Tsutsumi and Mie Suzuki fortechnical assistance, members of our Department (NCGG)for stimulating discussion, and Dr Shinji Fukada (NationalHospital for Geriatric Medicine, NCGG) and Dr ToshiyukiArai (Department of Surgery, Nagoya University School ofMedicine) for encouragement and support throughout thestudy. This study was supported in part by a grant-in-aid forLongevity Science from the Ministry of Health, Labor andWelfare of Japan (to KI) and by a grant for the Promotion ofFundamental Studies in Health Sciences of the NationalInstitute of Biomedical Innovation (NIBIO) of Japan (MF-14 and 06-31 to KI). Pacific Edit reviewed the manuscriptbefore submission.

REFERENCES

1. NIH Consensus Development Panel on Osteoporosis Pre-vention, Diagnosis and Therapy 2001 Osteoporosis preven-tion, diagnosis, and therapy. JAMA 285:785–795.

2. Lifton RP 1996 Molecular genetics of human blood pressurevariation. Science 272:676–680.

3. Ralston SH, de Crombrugghe B 2006 Genetic regulation ofbone mass and susceptibility to osteoporosis. Genes Dev 20:2492–2506.

4. Cappuccio FP, Meilahn E, Zmuda JM, Cauley JA 1999 Highblood pressure and bone-mineral loss in elderly white women:A prospective study. Study of Osteoporotic Fractures Re-search Group. Lancet 354:971–975.

5. Browner WS, Seeley DG, Vogt TM, Cummings SR 1991 Non-trauma mortality in elderly women with low bone mineraldensity. Study of Osteoporotic Fractures Research Group.Lancet 338:355–358.

6. Kanis J, Oden A, Johnell O 2001 Acute and long-term increasein fracture risk after hospitalization for stroke. Stroke 32:702–706.

7. Dennis MS, Lo KM, McDowall M, West T 2002 Fracturesafter stroke: Frequency, types, and associations. Stroke 33:728–734.

8. von der Recke P, Hansen MA, Hassager C 1999 The associ-ation between low bone mass at the menopause and cardio-vascular mortality. Am J Med 106:273–278.

9. Tanko LB, Christiansen C, Cox DA, Geiger MJ, McNabb MA,Cummings SR 2005 Relationship between osteoporosis andcardiovascular disease in postmenopausal women. J BoneMiner Res 20:1912–1920.

10. Peach MJ 1977 Renin-angiotensin system: Biochemistry andmechanisms of action. Physiol Rev 57:313–370.

11. Fukamizu A, Sugimura K, Takimoto E, Sugiyama F, Seo MS,Takahashi S, Hatae T, Kajiwara N, Yagami K, Murakami K1993 Chimeric renin-angiotensin system demonstrates sus-tained increase in blood pressure of transgenic mice carryingboth human renin and human angiotensinogen genes. J BiolChem 268:11617–11621.

12. Takimoto E, Ishida J, Sugiyama F, Horiguchi H, Murakami K,Fukamizu A 1996 Hypertension induced in pregnant mice byplacental renin and maternal angiotensinogen. Science 274:995–998.

13. Dzau V 2005 The cardiovascular continuum and renin-angiotensin-aldosterone system blockade. J Hypertens Suppl23:S9–S17.

14. Ondetti MA, Rubin B, Cushman DW 1977 Design of specificinhibitors of angiotensin-converting enzyme: New class oforally active antihypertensive agents. Science 196:441–444.

15. Ito M, Ikeda K, Nishiguchi M, Shindo H, Uetani M, Hosoi T,Orimo H 2005 Multi-detector row CT imaging of vertebralmicrostructure for evaluation of fracture risk. J Bone MinerRes 20:1828–1836.

RENIN–ANGIOTENSIN SYSTEM AND BONE METABOLISM 249

16. Takeshita S, Kaji K, Kudo A 2000 Identification and charac-terization of the new osteoclast progenitor with macrophagephenotypes being able to differentiate into mature osteoclasts.J Bone Miner Res 15:1477–1488.

17. Hiramatsu K, Asaba Y, Takeshita S, Nimura Y, Tatsumi S,Katagiri N, Niida S, Nakajima T, Tanaka S, Ito M, Karsenty G,Ikeda K 2007 Overexpression of gamma-glutamyltransferasein transgenic mice accelerates bone resorption and causes os-teoporosis. Endocrinology 148:2708–2715.

18. Takahashi N, Akatsu T, Udagawa N, Sasaki T, Yamaguchi A,Moseley JM, Martin TJ, Suda T 1988 Osteoblastic cells areinvolved in osteoclast formation. Endocrinology 123:2600–2602.

19. Boyle WJ, Simonet WS, Lacey DL 2003 Osteoclast differen-tiation and activation. Nature 423:337–342.

20. Campbell DJ 1987 Circulating and tissue angiotensin systems.J Clin Invest 79:1–6.

21. Carey RM, de Siragy HM 2003 Newly recognized componentsof the renin-angiotensin system: Potential roles in cardiovas-cular and renal regulation. Endocr Rev 24:261–271.

22. Hatton R, Stimpel M, Chambers TJ 1997 Angiotensin IIis generated from angiotensin I by bone cells and stimulatesosteoclastic bone resorption in vitro. J Endocrinol 152:5–10.

23. Strawn WB, Richmond RS, Ann Tallant E, Gallagher PE,Ferrario CM 2004 Renin-angiotensin system expression in ratbone marrow haematopoietic and stromal cells. Br J Haematol126:120–126.

24. Hagiwara H, Hiruma Y, Inoue A, Yamaguchi A, Hirose S1998 Deceleration by angiotensin II of the differentiation andbone formation of rat calvarial osteoblastic cells. J Endocrinol156:543–550.

25. Lamparter S, Kling L, Schrader M, Ziegler R, Pfeilschifter J1998 Effects of angiotensin II on bone cells in vitro. J CellPhysiol 175:89–98.

26. Niida S, Kondo T, Hiratsuka S, Hayashi S, Amizuka N, NodaT, Ikeda K, Shibuya M 2005 VEGF receptor 1 signaling isessential for osteoclast development and bone marrow for-mation in colony-stimulating factor 1-deficient mice. Proc NatlAcad Sci USA 102:14016–14021.

27. Yu X, Huang Y, Collin-Osdoby P, Osdoby P 2003 Stromalcell-derived factor-1 (SDF-1) recruits osteoclast precursorsby inducing chemotaxis, matrix metalloproteinase-9 (MMP-9)activity, and collagen transmigration. J Bone Miner Res 18:1404–1418.

28. Winslow MM, Pan M, Starbuck M, Gallo EM, Deng L,Karsenty G, Crabtree GR 2006 Calcineurin/NFAT sig-naling in osteoblasts regulates bone mass. Dev Cell 10:771–782.

29. Brasch H, Sieroslawski L, Dominiak P 1993 Angiotensin IIincreases norepinephrine release from atria by acting on

angiotensin subtype 1 receptors. Hypertension 22:699–704.

30. Takeda S, Elefteriou F, Levasseur R, Liu X, Zhao L, ParkerKL, Armstrong D, Ducy P, Karsenty G 2002 Leptin regulatesbone formation via the sympathetic nervous system. Cell111:305–317.

31. Hunyady L, Catt KJ 2006 Pleiotropic AT1 receptor signalingpathways mediating physiological and pathogenic actions ofangiotensin II. Mol Endocrinol 20:953–970.

32. Nouet S, Nahmias C 2000 Signal transduction from the angi-otensin II AT2 receptor. Trends Endocrinol Metab 11:1–6.

33. Okumura M, Iwai M, Ide A, Mogi M, Ito M, Horiuchi M 2005Sex difference in vascular injury and the vasoprotective effectof valsartan are related to differential AT2 receptor expres-sion. Hypertension 46:577–583.

34. Cohn RD, van Erp C, Habashi JP, Soleimani AA, Klein EC,Lisi MT, Gamradt M, ap Rhys CM, Holm TM, Loeys BL,Ramirez F, Judge DP, Ward CW, Dietz HC 2007 AngiotensinII type 1 receptor blockade attenuates TGF-beta-inducedfailure of muscle regeneration in multiple myopathic states.Nat Med 13:204–210.

35. Casas JP, Chua W, Loukogeorgakis S, Vallance P, Smeeth L,Hingorani AD, MacAllister RJ 2005 Effect of inhibitors of therenin-angiotensin system and other antihypertensive drugs onrenal outcomes: Systematic review and meta-analysis. Lancet366:2026–2033.

36. Ramnemark A, Nyberg L, Borssen B, Olsson T, Gustafson Y1998 Fractures after stroke. Osteoporos Int 8:92–95.

37. Lynn H, Kwok T, Wong SY, Woo J, Leung PC 2006 Angio-tensin converting enzyme inhibitor use is associated withhigher bone mineral density in elderly Chinese. Bone 38:584–588.

38. Shimizu H, Nakagami H, Osako MK, Hanayama R, KunugizaY, Kizawa T, Tomita T, Yoshikawa H, Ogihara T, Morishita R2008 Angiotensin II accelerates osteoporosis by activatingosteoclasts. FASEB J 22:2465–2475.

Address reprint requests to:Kyoji Ikeda, MD

Department of Bone and Joint Disease

National Center for Geriatrics and Gerontology36-3 Gengo, Morioka, Obu

Aichi 474-8522, Japan

E-mail: kikeda@nils.go.jp

Received in original form February 18, 2008; revised form August19, 2008; accepted October 7, 2008.

250 ASABA ET AL.