Osteocytes regulate anabolic and catabolic responses to PTH 1 Parathyroid Hormone

30 Pages·2013·1.21 MB·English

JBC Papers in Press. Published on June 2, 2013 as Manuscript M112.441360
The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M112.441360
Osteocytes regulate anabolic and catabolic responses to PTH
Parathyroid Hormone (PTH)/PTH-related Peptide Type 1 Receptor (PPR) Signaling in Osteocytes
Regulates Anabolic and Catabolic Skeletal Responses to PTH*

Vaibhav Saini1, Dean J. Marengi1 Kevin J. Barry1, Keertik S. Fulzele1, Erica Heiden1, Xiaolong
Liu1, Christopher Dedic1, Akira Maeda1, Sutada Lotinun2, Roland Baron1,2, and Paola
Divieti Pajevic1

1Endocrine Unit, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA,
2Harvard School of Dental Medicine, Boston MA, USA

*Running title: Osteocytes regulate anabolic and catabolic responses to PTH

To whom correspondence should be addressed: Paola Divieti Pajevic, Harvard Medical School,
Massachusetts General Hospital, Endocrine Unit, Thier 1101, Boston, MA, USA, Tel.: [phone];
Fax: [phone]; E-mail: [email protected]

Keywords: Osteocyte; PTH; anabolic; catabolic

D
Background: Osteocytes, the most abundant cells catabolic PTH regimen, Ocy-PPRKO animals ow
n
in adult bone, express PPR. demonstrated blunted skeletal responses. PTH lo
a
d
Results: Mice with constitutive PPR deletion in failed to suppress SOST/Sclerostin or induce ed
osteocytes demonstrate blunted anabolic and RANKL expression in Ocy-PPRKO animals fro
m
catabolic bone responses, and inability to recruit compared to controls. In vitro osteoclastogenesis h
ttp
osteoblasts and osteoclasts upon PTH was significantly impaired in Ocy-PPRKO upon ://w
administration. PTH administration, indicating that osteocytes w
w
Conclusion: PPR in osteocytes is needed for a full control osteoclast formation through a PPR- .jb
c
skeletal response to PTH administration mediated mechanism. Taken together, these data .o
rg
Significance: PPR signaling in osteocytes is indicate that PPR signaling in osteocytes is b/
y
necessary for PTH-driven anabolic effects during required for bone remodeling and receptor g
u
e
osteoporosis therapy. signaling in osteocytes is needed for anabolic and st o
catabolic skeletal responses. n A
p
SPuarmatmhyarroyid hormone (PTH) is the only FDA IntroduPctairoanth yroid hormone (PTH) is an 84 ril 5, 2
0
approved anabolic agent to treat osteoporosis; amino acid single-chain polypeptide synthesized 1
9
however cellular targets of PTH action in bone and secreted by the parathyroid glands in a
remain controversial. PTH modulates bone calcium-regulated manner. PTH maintains serum
turnover by binding to the PTH/PTH- related calcium homeostasis, controls renal phosphate
Peptide (PTHrP) type 1 receptor (PPR), a G- resorption, and vitamin D 1α-hydroxylation (1) by
protein-coupled-receptor highly expressed in bone activating the PTH/PTH-related Peptide (PTHrP)
and kidneys. Osteocytes, the most abundant cells type 1 receptor (PPR), a G-protein coupled
in adult bone, also express PPR. However, receptor capable of activating multiple G protein-
physiological relevance of PPR signaling in coupled pathways, including those signaling
osteocytes remains to be elucidated. Towards this through cAMP/Protein kinase A (PKA),
goal, we generated mice with PPR deletion in phospholipase C (PLC)/Protein kinase C (PKC),
osteocytes (Ocy-PPRKO). Skeletal analysis of and non-PLC-dependent PKC and Ca++ (1). The
i
these mice revealed a significant increase in bone first 34 amino-acids of PTH are necessary and
mineral density, and trabecular and cortical bone sufficient to fully activate the receptor (1), and the
parameters. Osteoblast activities were reduced in amino-terminal region received FDA approval in
these animals, as demonstrated by decrease 2002 as the first anabolic agent to treat
Col1α1 mRNA and RANKL expression. osteoporosis, a health disorder estimated to affect
Importantly, when subjected to anabolic or 200 million people worldwide (2)(3). In a 2002
1
Copyright 2013 by The American Society for Biochemistry and Molecular Biology, Inc.
Osteocytes regulate anabolic and catabolic responses to PTH
prevalence report by the National Osteoporosis showed an increase in bone mineral density and
Foundation, it was estimated that over 10 million osteopetrosis (19, 20), suggesting a role of
people suffered from osteoporosis in the US, and osteocyte-derived RANKL in bone remodeling.
~80% of these were women. According to the We previously generated transgenic mice
World Health Organization, osteoporotic fractures in which the PPR was conditionally ablated from
are a major cause of morbidity and disability in osteocytes post-natally upon tamoxifen injection
older people, and hip fractures could lead to (10Kb DMP1-Cre-Ert2). These mice demonstrated
premature death (3). Therefore, understanding the mild osteopenia by 4-6 weeks of age,
cellular targets of PTH action in bone is central for characterized by a reduction in trabecular bone,
the development of future osteoporosis accompanied by a tonic elevation of SOST and
therapeutics. Sclerostin, and a lack of PTH-induced SOST and
The PPR is abundantly expressed in bone Sclerostin suppression (4). To further investigate
and kidney, and its expression has also been the physiological role of PPR in osteocytes, we
reported in a variety of other tissues where it likely have generated mice with constitutive PPR
reflects the local paracrine role of PTHrP (4). In ablation in osteocytes by using the constitutive
bone, PPR is expressed in cells of the osteoblast 10Kb DMP1 promoter to drive Cre recombinase
lineage, including osteocytes. Osteocytes are expression in PPR-floxed mice. Osteocyte-PPR
terminally differentiated osteoblasts and comprise KO mice (Ocy-PPRKO) showed a significant
D
90-95% of all cells in adult bone (5). They survive increase in bone mineral density (BMD), and ow
n
for decades in their mineralized microenvironment trabecular and cortical bone parameters at 12 lo
a
d
and have the longest lifespan among all skeletal weeks of age, indicating that PPR on osteocytes is ed
cells. Recent evidence suggests direct or indirect required for normal bone remodeling. Interestingly fro
m
interactions of osteocytes with organs, such as Ocy-PPRKO displayed normal serum calcium, h
ttp
kidneys, muscles, heart, and bone, through various phosphate, and PTH, suggesting that under ://w
molecules, such as FGF-23 and sclerostin (5, 6). physiological conditions PPR signaling in w
w
Moreover, osteocytes might play important roles osteocytes is not needed to maintain normal .jb
c
in diseases, such as hypophosphatemic rickets, mineral homeostasis. When subjected to .o
rg
osteopenia, and osteopetrosis (5, 6). Emerging intermittent or continuous PTH administration, b/
y
literature also supports crucial roles of osteocytes Ocy-PPRKO generated blunted anabolic and g
u
e
in normal physiological processes, such as catabolic skeletal responses, indicating that PPR st o
lactation and bone modeling and remodeling (4, 7, signaling in osteocytes is necessary for full n A
8). It has been previously proposed that sOkceyle-PtaPl RKreOsp onfasielse.d Utpoo ni ncPrTeaHs e adCmoilnlaigsterna-t1ioαn1, pril 5, 2
0
osteocytes may mediate, in part, the anabolic (Col1α1) mRNA expression in osteoblasts or 1
9
effects of PTH by suppressing Sclerostin suppress SOST/Sclerostin expression in
expression through a Mef2C-mediated pathway osteocytes. Moreover, reduced RANKL
(9–12). Sclerostin, encoded by the SOST gene, is expression in osteocytes was observed in Ocy-
secreted by osteocytes and inhibits osteoblast PPRKO, and PTH failed to induce a catabolic
function and bone formation by binding to the response in these animals. All together, these data
low density lipoprotein receptor-5 and 6 and highlight the necessity of PPR signaling in
suppressing the Wnt signaling pathway (13–16). osteocytes for proper bone remodeling and full
Moreover, PPR on osteoblasts is known to skeletal responses to PTH.
regulate RANKL expression, and thereby control
osteoclast formation and bone resorption (17, 18). Experimental procedures
Importantly, activities of both osteoblasts and Mice
osteoclasts are needed for bone remodeling, and To generate Ocy-PPRKO animals, mice in
anabolic and catabolic bone responses. which 10Kb DMP1 promoter drives the expression
Recently osteocytes have been shown to of Cre-recombinase (10Kb DMP1-Cre, kindly
modulate osteoclast function(s) through a provided by Dr. J. Feng) were mated with mice in
RANKL-mediated mechanism (19, 20). Mice which the E1 exon of the PPR gene is flanked by
lacking RANKL specifically in osteocytes lox-P sites (control, kindly provided by Dr. T.
2
Osteocytes regulate anabolic and catabolic responses to PTH
Kobayashi). The genotypes of the mice were antigen retrieval was performed with Tris-EDTA
determined by PCR analysis of the genomic DNA buffer (10mM Tris base, 1mM EDTA, 0.05%
extracted from tail biopsies. For the 10Kb Dmp1- Tween-20, pH 9.0) in a boiling water bath at 95°C
Cre transgene, the forward Cre primer (5’ for 12 min. Slides were cooled to RT and blocked
CGCGGTCTGGCAGTAAAAACTATC-3’) and with TNB (TSA ™ Biotin Tyramide Kit, Perkin
the reverse Cre primer (5’- Elmer, Waltham MA, USA) for 30 min, RT, and
CCCACCGTCAGTACGTGAGATATC-3’) were incubated with anti-RANKL antibody (N-19,
used to generate a PCR product of approximately Santa Cruz Biotechnology), 4°C, ON in Can Get
400bp. For the floxed PPR allele, the P1 primer Signal immunostain Solution A (Toyobo, Japan).
(5’-ATG AGG TCT GAG GTA CAT GGC TCT Slides were washed, and incubated with Biotin-
GA -3’) and the P2 primer (5’-CCT GCT GAC SP-conjugated AffiniPure Rabbit Anti-Goat IgG
CTC TCT GAA AGA ATG T -3’) were used, (H+L), (Jackson ImmunoResearch) for 30 min at
which recognized the sequence spanning the RT. Afterwards, the slides were washed and
3’lox-P site, as previously reported (4). Wild-type incubated for 30 min with streptavidin (SA)
and mutant alleles give ~ 210 bp and 290 bp conjugated horseradish peroxidase (HRP) and
products, respectively. tyramide following the manufacturer’s protocol
Institutional Animal Care and Use (TSA™ Biotin Tyramide Kit), developed with a
Committee, Subcommittee on Research Animal 3,3'-diaminobenzidine substrate-chromogen
D
Care, at Massachusetts General Hospital approved system (Vector laboratories, Burlingame, CA, ow
n
all animal protocols. USA), and counterstained with methyl green. lo
a
d
Allele specific DNA recombination was Brightfield microscopy was performed and images ed
performed on DNA isolated from osteocyte- were acquired with a 40x objective (Nikon Eclipse fro
m
enriched long bones (generated by sequential E800). For quantification, ≥8 fields were imaged h
ttp
collagenase and EDTA digestions as previously from ≥4 tibiae per group (control vehicle, control ://w
described in (4)), calvaria, kidney, liver, lung, PTH, Ocy-PPRKO vehicle, and Ocy-PPRKO w
w
heart, skeletal muscle, and spleen. Multiplex PCR PTH). Subsequently, ImageJ (21) was used to .jb
c
analysis was performed using allele specific quantify the RANKL-stained osteocyte area and .o
rg
primers, P1, P2 and P3 (5’ACA TGG CCA TGC normalized to total number of osteocytes per field. b/
y
CTG GGT CTG AGA 3’) and following the Statistical analysis was performed using 2-way g
u
e
manufacturer protocol (Quiagen™ Multiplex PCR ANOVA and Tukey's HSD test. st o
kit, Quiagen Valencia, CA USA) Similar staining protocol was followed for n A
Histology Panetriiboostdiyn (IAHFC2 9o5n5 ,t iRbi&aeD eSxycsetpetm tsh)e waanst id-Pileurtieods tiinn pril 5, 2
0
Tibiae and vertebrae were fixed in 10% Can Get Signal immunostain Solution B (Toyobo, 1
9
formalin/PBS solution at 4°C, ON, decalcified in Japan).
20% EDTA pH 8.0 for 7-15 days, processed, IHC for sclerostin expression was
embedded in paraffin, and sectioned. Sections performed on deparaffinized vertebrae. Antigen
were stained with hematoxylin and eosin, or used retrieval was performed using proteinase K for 15
for immunohistochemistry. Undecalcified femurs, min at RT. The subsequent steps were similar to
fixed in 70% ethanol, were embedded in those described for RANKL IHC above, except no
methylmethacrylate (Aldrich Chemical Co, secondary antibody was used because we used the
Milwaukee, WI, USA), sectioned with a diamond- biotinylated anti-mouse sclerostin antibody (1:50
embedded wire saw, and stained by the Von Kossa dilution; R&D Systems, Inc., Minneapolis, MN,
method. USA) which was diluted in tris-NaCl blocking
buffer. Brightfield microscopy was performed and
Immunohistochemistry (IHC) images were acquired with a 20x objective (Nikon
Immunohistochemical RANKL expression Eclipse E800). For quantification, ≥12 fields were
detection was performed as described in (19) with imaged from ≥5 vertebrae per group (control
minor modifications. On deparaffinized tibiae, vehicle, control PTH, Ocy-PPRKO vehicle, and
endogenous peroxidase activity was inhibited by Ocy-PPRKO PTH), sclerostin positive osteocytes
3% H O treatment for 10 min. Subsequently, were counted and normalized to total number of
2 2
3
Osteocytes regulate anabolic and catabolic responses to PTH
osteocytes per field. 2-way ANOVA was analysis, and left tibia was used for histology.
performed and the significance of p values was Right femurs and right tibiae were used to isolate
determined based on the Q scores obtained using RNA.
Tukey's HSD test.
Histomorphometric Analysis
In Situ Hybridization Mice were intraperitoneally injected with
In situ hybridization was carried out as previously 20 mg/kg calcein and 40 mg/kg demeclocycline on
described (22). The antisense probe for col11 has days 9 and 2 before sacrifice, respectively. The
been reported (22). distal femur metaphyses were removed and fixed
in 70% ethanol. They were then dehydrated,
Serology infiltrated, and embedded in methyl methacrylate.
Serum was isolated from the blood Undecalcified 4-μm-thick sections were cut using
collected by retro-orbital bleeding. Serum mouse a microtome (Leica RM 2255, Heidelberg,
TRAP5b (Mouse TRAP Assay), PINP (Rat/Mouse Germany) and mounted unstained for dynamic
PINP EIA), CTX (RatLaps EIA) and RANKL measurements. Consecutive sections were
(Quantikine ELISA Kit, R&D systems) were toluidine blue stained for static measurements.
measured using ELISA (Immunodiagnostic Histomorphometric parameters were measured
Systems Limited, UK). Serum total calcium was using an Osteomeasure image analysis system
D
measured by calcium liquicolor arsenazo kit (OsteoMetrics, Atlanta, GA) coupled to a ow
n
(Stanbio Laboratory, USA) and inorganic photomicroscope and personal computer and the lo
a
d
phosphate was quantified by phospho liquid-UV results were expressed according to standardized ed
kit (Stanbio Laboratory, USA) as per the nomenclature (23). A sampling site of 2 mm2 was fro
m
manufacturer’s instructions. Serum PTH was established in the secondary spongiosa at 400 µm h
ttp
measured using mouse intact PTH ELISA kit below the growth plate. Trabecular bone volume ://w
(Immutopics, USA) according to the was assessed as a percentage of total tissue volume w
w
manufacturer’s protocol. (BV/TV, %). Tb.Th (μm), Tb.N (1/mm), and .jb
c
Tb.Sp (μm) were calculated. Mineral apposition .o
rg
Anabolic and Catabolic PTH treatment rate (MAR, µm/day) was derived from the mean b/
y
Human PTH(1–34) (MGH Peptide Core distance between fluorescent labels divided by the g
u
e
Facility) was dissolved in 0.1% TFA, aliquoted, labeling interval. The cancellous bone surface st o
stored frozen at –80°C, and subsequently diluted lined by osteoblasts and osteoclasts were n A
thoe atth-ien aacptipvraotperdi amteo ucsoen sceernutmra,t i0o.n1 Nin H vCelh ainclde 0(.92%% msuerafascuere d. O(Osbte.Sob/BlaSs,t%s )w, ere ebxopnree ssed ppeerri mbeotneer pril 5, 2
0
sodium chloride) immediately before (N.Ob/B.Pm, /mm) and tissue area (N.Ob/T.Ar, 1
9
administration. For anabolic treatment, control and /mm2). Osteoclasts were expressed per bone
Ocy-PPRKO female mice, randomly subdivided surface (Oc.S/BS,%), bone perimeter (N.Oc/B.Pm,
into sham or treatment groups, were injected with /mm) and tissue area (N.Oc/T.Ar, /mm2). The data
100μL of either vehicle or PTH peptide. We in Table 2 comparing control vs. Ocy-PPRKO
treated mice with 80μg/kg PTH(1–34) anabolic were analyzed using an unpaired t-test. The data in
dose sc 5 days per week for 4 consecutive weeks. Table 5 comparing catabolic response among 4
For catabolic treatment, control and Ocy-PPRKO groups (Control vehicle, Control PTH, Ocy-
male mice, randomly subdivided into sham or PPRKO vehicle, and Ocy-PPRKO PTH), were
treatment groups, were sc implanted with Alzet analyzed using ANOVA with Tukey/Kramer.
Micro-Osmotic Pump Model 1002 (Durect
corporation, USA) to deliver 100μL of either RNA extraction and purification
vehicle or PTH peptide (100μg/kg/day) for 2 RNA was extracted from whole bones (femur and
consecutive weeks. Upon sacrifice, left femurs and tibia) and from osteocyte-enriched calvaria.
left tibiae were cleaned of soft tissue, fixed in 10% Calvariae from control and Ocy-PPRKO mice
phosphate-buffered formalin (pH 7.2) for 48h and were subjected to sequential collagenase and
stored in 70% ethanol until further use. Left femur EDTA digestions to remove endosteal and
was used for microCT and histomorphometry periosteal osteoblasts and bone marrow cells as
4
Osteocytes regulate anabolic and catabolic responses to PTH
detailed in (4). For basal mRNA expression (50mM sodium acetate/0.05% sodium azide, pH
analysis, osteocyte-enriched calvariae were frozen 6.2) for measurement of cAMP by a specific RIA,
in liquid nitrogen. RNA was extracted by as previously described (38). Bones were washed
homogenization in Trizol (Invitrogen, USA) as per twice with 0.5mL acetone and once with 0.5mL
the manufacturer’s instructions. RNA quality and ether and were air-dried and weighed. The results
quantity was ascertained by UV were normalized for the bone weight, and the data
spectrophotometry (NanoDrop 8000, Thermo were expressed as picomole of cAMP produced
Fisher, USA). per mg of dry bone. Each experiment was repeated
at least three times.
Quantitative RT-PCR
Reverse transcription was performed on Total Body Weight and dual-X-ray
0.5-1 µg of DNAse-treated total RNA and absorptiometry (DXA)
oligoDT primers using Omniscript (Invitrogen, Total body weight in g was recorded at the
USA) or QuantiTect (Qiagen, USA) according to ages mentioned in the results. Mouse BMD was
the manufacturer’s instructions. Quantitative PCR measured by DXA using a Lunar PIXImus II
for PPR, SOST, RANKL, OPG, and β-actin was densitometer (GE Medical System Luna, Madison,
performed using the QuantiTect SYBR Green WI). In brief, animal were briefly anesthetized by
PCR Kit (Qiagen, USA) on StepOnePlus (Applied sc. administration of tri-bromoethanol, placed on a
D
Biosystems, USA) as described previously (4). tray and total body mineral density was acquired ow
n
Primer sequences are available upon request. The using manufactures protocols. Post-acquisition lo
a
d
data were normalized to beta actin, and statistical skeletal analysis was performed by excluding the ed
significance was determined using a t-test in Excel mouse head from the final BMD (g/cm2) fro
m
and p≤0.05 was considered significant. determination. For total body weight and BMD, h
ttp
statistical significance was determined using a t- ://w
Cyclic AMP measurement test and p≤0.05 was considered significant. w
w
Tibiae, femurs, and calvariae were .jb
c
isolated from 8-12-week old mice. Briefly, tibiae MicroCT analysis .o
rg
were dissected and cleaned of adherent tissues. Bone morphology and microarchitecture were b/
y
Distal and proximal epiphyses were removed and analyzed using a desktop high-resolution µCT g
u
e
the bone marrow was flushed out using 2-3mL of (µCT40, Scanco Medical, Brüttisellen, st o
α-MEM supplemented with 0.1% bovine serum Switzerland), as described previously (24). n A
arelbmuaminiinn, gB dSiaAp haynsdia l2 e5nmriMch eHd erpeegsio np Ho f 7th.4e. b oTnhees Bmriidesfhlya,f t,t haen dd isLta5l vfeermteobrraall mboedtayp hwyseirce, sfceamnonreadl pril 5, 2
0
was cut into three pieces and sequentially digested using an X-ray energy of 70keV, integration time 1
9
as described above for RNA isolation. Each piece of 200ms, and a 12µm isotropic voxel size.
was then placed in ice-cold cAMP-assay buffer Longitudinal sections were scanned using an X-
(Dulbecco Modified Essential Medium containing ray energy of 70keV, a current of 114µA,
10mM HEPES, 0.1% heat-inactivated BSA and integration time of 200ms, and a 18µm isotropic
1mM isobutylmethylxantine). Bone pieces were voxel size; subsequently the half of total femur
then incubated in cAMP-assay buffer with the length was measured, and distal to mid-diaphysis
appropriate treatment at 37°C for 15 min. The in the femurs were used for measuring trabecular
three pieces of each tibia were incubated with parameters. For the trabecular bone region BV/TV
vehicle alone (assay buffer), 100nM human PTH (%), Tb.N (/mm), Tb.Sp (mm), and Tb.Sp (mm);
(1-34) or 0.1µM forskolin . At the end of the and for the cortical bone region Cort.Th (mm),
incubation, the reaction was terminated by quickly Cort.A (mm2), MA (mm2), Cort.Por (%), Cort.
removing the bones and placing them in 0.3 mL of Den (mgHA/ccm), and pMOI (mm4) were
cold 90% 2-propanol in 0.5M HCl. Bones were assessed.
then incubated for 16–18 h at 4°C. Propanol
extraction was repeated, and the combined extracts Ex Vivo Osteoclastogenesis
were evaporated by vacuum centrifugation. The We followed the protocol described in
dried extracts were redissolved in acetate buffer (19) with modifications. Briefly, non-adherent
5
Osteocytes regulate anabolic and catabolic responses to PTH
spleen cells isolated from ~12-week-old control p=0.25, NS) confirming that receptor expression
mice were cultured with 10 ng/ml mouse M-CSF was not altered in osteoblasts. In addition,
(R&D system) for 3 days and were used as osteocalcin mRNA expression was also similar
osteoclast precursors. Osteocyte-enriched long (control = 100±22.1 %, Ocy-PPRKO = 54.2±13.0
bones were isolated from ~10-week-old Ocy- %, expression normalized for -actin; p=0.14, NS)
PPRKO and control, weighed, cut into small confirming the osteoblastic nature of these cells.
pieces using aseptic technique, added to 24-well Functional ablation of the receptor was ascertained
plates in 500µl complete medium (αMEM with L- by assessing cAMP accumulation in response to
Glutamine, 10% FBS, and Pen-Strep) containing PTH in osteocyte-enriched long bone explants
0.1µM dexamethasone, and incubated overnight at (OEBE). OEBE were treated with 100nM
37ᵒC in a humidified atmosphere with 5% CO . hPTH(1-34) or 10µM forskolin. Ocy-PPRKO
2
The next day, 20,000 osteoclast precursors were bone explants had a statistically significant
added to each well of the 24-well plates and reduction (86.5% reduction) in cyclic AMP
cultured with 100nM hPTH(1-34), 10nM 1,25- accumulation in response to PTH compared to
dihydroxyvitamin D3, and/or 100ng/mL controls (Ocy-PPRKO = 15.04.7 picomol/mg
murineOPG/Fc chimera (R&D Systems) in the bone vs. controls = 110.820.5 picomol/mg bone;
presence of dexamethasone and M-CSF for 13 p<0.001) whereas both Ocy-PPRKO and control
days with fresh ingredients added every 3-4 days. mice demonstrated a robust cyclic AMP response D
Next, these were stained for TRAP, and to forskolin (Ocy-PPRKO = 94.258.3 ow
n
multinucleate cells with 2 or more nuclei were lo
picomol/mg bone vs. controls = 66.810.6 a
d
smciocrreods caosp eT. RTAhPe -pnousmitbiveer ocef llTs RuAsiPn-gp oasnit iivnev ecretellds pdiecmoomnoslt/rmatgin g thbaotn Oe;c y-PpP=R0K.3O59 m) ice (rFeitga.i n in1taCc)t ed from
per well were normalized to the weight of the bone adenylate cyclase responsiveness, but lack cAMP http
pieces added to that well. Statistical significance accumulation following PTH administration due to ://w
was determined using a t-test in Microsoft Excel w
reduced PPR expression in osteocytes (Fig. 1C). w
with p≤0.05 as significant. Photographic evidence Moreover, primary calvarial osteoblasts from Ocy- .jb
c
was recorded with a Nikon Eclipse E800 .o
PPRKO and control mice showed robust cAMP rg
microscope at a 10x objective. production when treated with either PTH or by/
g
forskolin (Fig. 1D), demonstrating intact PPR u
e
Results signaling in osteoblasts. st o
Generation of mice lacking PPR expression in n A
osteocytTeso generate mice lacking PPR expression Oancdy -cPoPrtRicKaOl b monicee displayed increased trabecular pril 5, 2
0
primarily in osteocytes (Ocy-PPRKO), mice with 1
PTH is the most important hormonal 9
10kb-DMP1-Cre were mated with mice in which
regulator of calcium and phosphate homeostasis. It
exon E1 of PPR was flanked by Lox-P sites (Fig.
exerts its effects by binding PPR expressed on its
1A). Receptor ablation was demonstrated by
target organs, namely bone and kidney. To
genomic recombination as assessed by PCR
investigate if receptor deletion in osteocytes
analysis of genomic DNA. Presence of the ~600
altered mineral homeostasis, we measured serum
bp band, expected with successful Cre-based
calcium and phosphate concentration. Both serum
recombination, confirmed PPR gene deletion in
calcium (control = 7.24±0.35 mg/dL, Ocy-PPRKO
calvaria and the osteocytes from long bones, but
= 6.88±0.62, mg/dL; n=5, NS) and phosphate
not in other tissues (Fig. 1B). PPR gene knockout
(control = 9.3±0.89 mg/dL, Ocy-PPRKO =
resulted in significantly reduced PPR mRNA
8.4±0.41, mg/dL, n=5, NS) were indistinguishable
expression in osteocyte-enriched long bones from
between Ocy-PPRKO and controls, suggesting
Ocy-PPRKO as compared with littermate controls,
that other PPR-expressing cells, likely osteoblasts
as we have recently reported (25). In contrast, PPR
or kidney cells, regulate serum calcium and
mRNA expression in primary calvarial osteoblasts
phosphate concentrations. Moreover, intact
derived from control and Ocy-PPRKO was
circulating PTH was also similar between Ocy-
unchanged (control = 100±6.9 %, Ocy-PPRKO =
PPRKO and control (control = 111.5±25.1 pg/mL,
112.1±6.9 %, expression normalized for -actin;
6
Osteocytes regulate anabolic and catabolic responses to PTH
Ocy-PPRKO = 104.6±25.2, pg/mL; NS), inertia (pMOI), in Ocy-PPRKO as compared with
indicating that the skeletal phenotype observed in control (Table 2). Cortical density (Cort. Dens).
Ocy-PPRKO mice is not a consequence of an medullary area (MA) and cortical porosity
altered parathyroid state. (Cort.Por) were comparable in Ocy-PPRKO and
We then analyzed the bone phenotype of control. (Table 2).
Ocy-PPRKO mice. Considering that bone weight, To investigate whether the increase in
fat tissue, and lean tissue contribute towards total BMD was driven by reduced bone resorption or
body weight, we measured it at different ages. increased bone formation, or a combination of
Total body weight was comparable between Ocy- both, we measured, in these animals, serum
PPRKO and control at 4 and 8 weeks of age (Fig. markers of bone formation and bone resorption.
2A). However, significantly increased body PINP (control = 232.8±26.8 ng/mL, Ocy-PPRKO
weight was recorded at 12 weeks for Ocy-PPRKO = 185.1±11.5, ng/mL; 12-week-old females, n≥4,
as compared with control (Fig. 2A). Increased NS), CTX (control = 9.5±1.9 ng/mL, Ocy-PPRKO
body weight was not present in control (PPRfl/fl =7 .0±0.3, ng/mL; 12-week-old females, n≥4, NS),
mice) and DMP1-Cre expressing mice (Fig. 2A), and TRAP5b (control = 2.35±0.31 U/L, Ocy-
indicating that the higher weight observed in Ocy- PPRKO = 2.17±0.18, U/L; 12-week-old females,
PPRKO animals was associated with the loss of n≥4, NS) were unchanged, as compared to
PPR expression in osteocytes. As assessed by controls.
D
DXA, fat tissue (control = 839.5±37.6 g/Kg, Ocy- Osteoblasts and osteoclasts number and ow
n
PPRKO = 845.3±27.4, g/Kg; n≥6, NS) and lean function were reduced, although not significantly, lo
a
d
tissue (control = 160.6±35.9 g/Kg, Ocy-PPRKO = in Ocy-PPRKO animals compared to controls ed
154.1±25.7, g/Kg; n≥6, NS) weights were similar (Table 5, vehicle treated animals), as demonstrated fro
m
between 12-week-old Ocy-PPRKO and control. by a decrease in Ob.S/BS, N.Ob./T.Ar, h
ttp
BMD was similar between Ocy-PPRKO and N.Ob./B.Pm, osteoid surface (OS/BS) and osteoid ://w
control mice at 8 weeks in both females (Fig. 2B) volume (OV/TV), Oc.S/BS, N.Oc/T.Ar, w
w
and males (Fig. 2C). However, at 12-15 weeks a N.Oc/B.Pm, and erosion surface (ES/BS). These .jb
c
statistically significant increase in BMD was findings were associated with a significant .o
rg
observed in both sexes and it persisted at 20 weeks increase in both SOST mRNA expression (Fig. b/
y
(Fig. 2B, C). DMP1-Cre only mice at 12 weeks of 2G) and in the number of Sclerostin-positive g
u
e
age were indistinguishable from controls (Fig. osteocytes (Fig. 3E, top row, and 3F) in bones st o
2B). from these animals, suggesting that the n A
increaseHd &tEra bsetaciunlianrg (bFoinge. 2iDn ) oOfc yti-bPiaPeR KshOow eads sduripvper ethssei orned oufc ethde o Wstenotb/la-sctast ennuimn bpeart hawnday asc tmiviigthyt. pril 5, 2
0
compared with control, starting at 8 weeks. Reduced osteoblast function was also 1
9
Similarly, Von Kossa staining of femurs showed demonstrated by a dramatic reduction in Col1α1
increased trabecular bone at 14 weeks (Fig. 2E). mRNA expression, as shown by in situ
To further assess the skeletal phenotype of hybridization in both trabecular and cortical bone
these animals, we performed microCT analysis of (Fig. 3C), in Ocy-PPRKO animals compared to
both vertebrae (L5) and distal femurs in 5-6 week controls. Interestingly, comparable number of
old Ocy-PPRKO and controls. At this age, there Sclerostin- expressing osteocytes was observed in
were no statistically significant differences the cortical region of the vertebrae from Ocy-
between Ocy-PPRKO and control mice, as shown PPRKO and control, suggesting that cellular
in Table 1. MicroCT analysis at 12 weeks, on the responses to PTH are site specific (Fig. 3G).
other hand, showed a significant increase in To examine cells of the osteoclast lineage,
trabecular bone volume fraction (BV/TV), we analyzed RANKL and OPG expression. There
trabecular number (Tb.N), and thickness (Tb.Th), was a significant decrease in both OPG (Fig. 2H)
and decrease in separation (Tb.Sp) in the vertebral and RANKL (Fig. 2I) mRNA expression in
body (L5) of Ocy-PPRKO as compared with osteocyte-enriched calvariae from Ocy-PPRKO
control (Table 2). Analysis of femur midshafts compared with control whereas the ratio
showed a significant increase in cortical thickness RANKL/OPG remained unchanged. A significant
(Cort.Th), area (Cort.A), and polar moment of 85% decrease in RANKL protein expression was
7
Osteocytes regulate anabolic and catabolic responses to PTH
also observed in osteocytes from Ocy-PPRKO as osteoclast activity, remained unchanged in both
compared with control (Fig. 5E, top row, and 5F). groups (Fig. 3B). In Ocy-PPRKO, intermittent
To investigate whether the decrease in PTH failed to stimulate Collα1 mRNA expression
osteoclast activity was associated with the (Fig. 3C). Periostin, a regulator of
presence of unresorbed calcified cartilage, as SOST/Sclerostin expression (26), was unchanged
recently reported for chondrocyte and osteoblast- upon intermittent PTH administration both in
specific RANKL knockout mice (20), we control and Ocy-PPRKO mice (Fig. 3D).
performed safranin O staining on tibiae from Ocy- SOST/Sclerostin expression has been reported to
PPRKO and control. Growth plates in Ocy- return to baseline by 24 hours after PTH treatment
PPRKO were indistinguishable from controls (Fig. (27), which limited out ability to analyze
2F), indicating that PPR activation in osteocytes is Sclerostin expression in intermittent PTH-treated
not required for proper bone growth and cartilage bone samples because, as per our anabolic
resorption. Moreover, the presence of a normal protocol, these were harvested 24 hours after the
growth plate demonstrates normal PPR expression last PTH injection. Therefore, to study Sclerostin
and signaling in other bone cells, such as regulation, we treated control and Ocy-PPRKO
chondrocytes and osteoblasts. with vehicle or 50nmol/Kg hPTH(1-34) for 1.5
Altogether, these findings indicate that in hours. While PTH induced a decrease in Sclerostin
the absence of PPR signaling in osteocytes there is expression in the osteocytes from control, it failed
D
an age-dependent increases in BMD, and in to suppress Sclerostin expression in Ocy-PPRKO ow
n
trabecular and cortical bone parameters, that is (Fig. 3E, F). Interestingly, similar sclerostin lo
a
d
associated with a decrease in osteoblast and expression was observed in the cortical bone of ed
osteoclast functions but not number. vertebrae from both vehicle- and PTH-treated fro
m
Ocy-PPRKO and control (Fig. 3G), suggesting a h
ttp
Blunted skeletal response to anabolic and differential PPR-mediated regulation of sclerostin ://w
catabolic PTH administration in Ocy-PPRKO expression in trabecular and cortical sites. w
w
Ocy-PPRKO and control were then Next, we examined the ability of Ocy- .jb
c
subjected to intermittent or continuous PTH PPRKO to respond to continuous PTH .o
rg
treatment. As expected, intermittent hPTH(1-34) administration. Mice were subjected to 2 weeks of b/
y
at a dose of 80µg/Kg/day, induced a significant continuous administration of 100µg/Kg/day of g
u
e
increase in vertebral BV/TV and Tb.N, and hPTH(1-34). MicroCT (Fig. 4C and Table 4) and st o
decrease in Tb.Sp (Table 3) in control. Analysis of Von Kossa staining (Fig. 4D) showed a catabolic n A
mCoidrts.hTahf t afnedm uCrso rst.hAo,w eadn da psMignOiIf iciann tP iTnHcr-etaresaet eind rOecsyp-oPnPseR KupOo.n P TPTHH t retraetamtmenetn itn dinu cceodn tar osli,g nbiufti cnanott pril 5, 2
0
controls compared to vehicle treated animals decrease in Tb.N, and increase in Tb.Sp in control 1
9
(Table 3). In contrast, these trabecular and cortical mice, as shown by microCT analysis (Table 4). In
parameters remained unchanged between vehicle- contrast, these parameters remained unchanged
and PTH-treated Ocy-PPRKO (Table 3). No between vehicle- and PTH-treated Ocy-PPRKO
significant changes were observed in the other (Table 4), demonstrating that mice lacking PPR
bone parameters between vehicle- and PTH- expression in osteocytes are resistant to continuous
treated control or Ocy-PPRKO mice (Table 3). infusion of PTH. No significant changes were
These data demonstrate that osteocyte-mediated observed in other parameters between vehicle- and
signals are required for anabolic skeletal responses PTH-treated control or Ocy-PPRKO mice (Table
to PTH. 4).
To evaluate the role of PPR signaling in Histomorphometric analysis demonstrated
osteocytes in modulating osteoblast and osteoclast a significant increase in both osteoblasts and
activities during anabolic PTH treatment, PINP osteoclasts number and function, upon PTH
and CTX serum levels were measured. treatment in controls, as demonstrated by a
Intermittent administration of PTH significantly significant increase in Ob.S/BS, N.Ob./T.Ar,
increased circulating PINP, a marker of osteoblast N.Ob./B.Pm, osteoid surface (OS/BS), osteoid
activity, in control whereas had no effect in Ocy- volume (OV/TV), Oc.S/BS, N.Oc/T.Ar,
PPRKO (Fig. 3A). Serum CTX, a marker of N.Oc/B.Pm, and erosion surface (ES/BS) (Table
8
Osteocytes regulate anabolic and catabolic responses to PTH
5). None of these parameters were significantly significant difference in genotypes, but not the
changed in Ocy-PPRKO animals upon PTH PTH treatments or genotype and PTH treatment
administration (Table 5), indicating that deletion interaction. Importantly, Tukey HSD identified a
of PPR in osteocytes blunts the skeletal response significantly higher RANKL expression in both
to continuous PTH. Interestingly, PTH did vehicle- and PTH- treated control osteocytes as
significantly decrease Tb.N in Ocy-PPRKO mice compared with both vehicle- and PTH-treated
and not in controls; however this change was not Ocy-PPRKO osteocytes (Fig. 4E and F). To
associated with a decrease (or change) in any other further assess the role of RANKL in these animals,
bone parameter (Table 5). Moreover, we measured its level in the serum. There was no
histomorphometry further confirmed the increased difference in circulating RANKL between control
BV/TV, Tb.N and decreased in Tb.Sp in Ocy- and Ocy-PPRKO animals and PTH treatment had
PPRKO compared to controls (Table 5). no effect (control-veh = 129.7.8±10.9 pg/mL,
Serum marker of osteoblast activity, control-PTH = 101.78 ±18.6 pg/mL, p=0.19, NS;
PINP, was significantly elevated in PTH-treated Ocy-PPRKO-veh =98.8±7.8, pg/mL, Ocy-
control, but remained unchanged in Ocy-PPRKO, PPRKO-PTH =107.4±20.5, pg/mL; p=0.65, NS)
further confirming the lack of PTH-responsiveness suggesting that circulating RANKL might not
in Ocy-PPRKO (Fig. 4A). Interestingly, serum reflect the local effects of this cytokine. RANKL
levels of CTX, a marker of bone resorption, were levels in bone marrow supernatant of femurs and
D
significantly increased in Ocy-PPRKO (p<0.05) tibias of Ocy-PPRKO and controls were below the ow
n
whereas in controls the increase (167%) did not limit of detection of the assay used. Moreover, lo
a
d
reach significance (p=0.0508) (Fig.4B). These RANKL mRNA expression in RNA isolated from ed
results suggest that PTH-induced CTX elevation is intact bones (including both bone marrow and fro
m
independent of PPR signaling in osteocytes and osteoblasts) showed no changes in vehicle- or h
ttp
most likely reflect a osteoblast-mediated, PTH PTH-treated control and Ocy-PPRKO (Relative % ://w
dependent mechanism(s). RANKL mRNA expression: control-veh = w
w
To investigate if the lack of response to 100.0±12.0%, control-PTH = 83.0 ±2.9%, .jb
c
continuous PTH was due to an osteoclast defect, p=0.198, NS; Ocy-PPRKO-veh =53.4±5.2%, Ocy- .o
rg
we analyzed the osteoclast-forming ability of PPRKO-PTH =86.2±22.7%; p=0.270, NS). These b/
y
osteocytes from Ocy-PPRKO and control in vitro. findings are in agreement with previous studies g
u
e
OEBE from Ocy-PPRKO and control were co- (20) and suggest that the contribution of RANKL st o
cultured with osteoclast precursors derived from produced by osteocytes in small relative to the n A
wmiulldt inutcylpeea te ospstleeoenclsa. sts U(MpoNnC ) PwTeHre obtrseeartvmede nitn, tcoatna lb Re AdeNteKcLte dp raosd ulaccekd ,o yf ePt TitHs bcaiotalobgoilcica l aecftifoecntss. pril 5, 2
0
both cultures (Fig. 5A). However, the number of In the absence of PPR signaling in osteocytes, 1
9
MNC formed in Ocy-PPRKO cultures was continuous PTH administration fails to increase
significantly reduced (Fig. 5B; Control MNC/mg both osteoblasts and osteoclasts and induce a
OEBE = 4.6±0.2, Ocy-PPRKO MNC/mg OEBE = skeletal effect.
1.8±0.2, p=0.001). Importantly, OPG completely
abrogated the osteoclastogenic ability of both Ocy- Discussion
PPRKO and control OEBE (Fig. 5A,B), It has been known for over a decade that
suggesting a RANKL-dependent mechanism. The PTH can exert its anabolic or catabolic effect on
number of multinucleated osteoclasts formed upon bone via activation of the PPR widely expressed
1,25-dihydroxyvitamin D treatment was similar in on cells of the osteoblast lineage; however, the
both Ocy-PPRKO and control cultures (Fig. 5), exact cellular target of its action on bone has
demonstrating reduced osteoclast formation in remained elusive. Several mechanisms have been
Ocy-PPRKO because of ablated PPR signaling. proposed and they comprise recruitment and
To further investigate the reduced proliferation of osteoprogenitors cells, activation
osteoclast activity in response to PTH observed by of bone lining cells (28, 29), inhibition of
histomorphometry (Table 5), we analyzed osteoblast and osteocyte apoptosis (30), and
RANKL expression in osteocytes upon catabolic suppression of SOST/Sclerostin expression (27).
PTH administration. ANOVA showed a Osteocytes, the most abundant cells in bone, have
9
Osteocytes regulate anabolic and catabolic responses to PTH
recently emerged as important modulator of bone in osteocytes does not significantly alter the
modeling and remodeling (5). In this regard, the number of osteoblasts and osteoclast but reduces
molecular roles of sclerostin and RANKL, two their function with a net increase in BMD, as
predominantly osteocytic skeletal factors, have demonstrated by both DXA, microCT and
been reported (5). Interestingly, both molecules histomorphometric analysis. On the other hand,
are known targets of PTH actions with when Ocy-PPRKO animals were subjected to a
SOST/Sclerostin being suppressed and RANKL skeletal perturbation (such as intermittent or
being increased upon receptor activation. continuous PTH administration), they failed to
Herein, using a murine model of PPR properly increase both osteoblast and osteoclast
ablation in osteocytes, we provide, for the first number and activitiess as shown by
time, evidences that receptor signaling in these histomorphometric analysis (Table 5), expression
cells is important to control both osteoblast and of Col1α1(Fig. 3C) and Sclerostin (Fig. 3E).
osteoclast functions via SOST/Sclerostin and Interestingly, continuous administration of PTH
RANKL-mediated mechanisms. Importantly, our failed to induce trabecular bone loss or increase
studies demonstrated that PPR signaling in osteoclast number in Ocy-PPRKO animals despite
osteocytes is required for generating full skeletal a significant increase in serum CTX. We can
responses to anabolic and catabolic PTH speculate that CTX might derive from skeletal
administration. During PTH anabolic regimen, sources other than the one analyzed or from the
D
Sclerostin suppression is needed to allow a full modest increase in osteoclast numbers detected by ow
n
hormonal effect and, in the absence of PPR histomorphometry (Table 5). lo
a
d
signaling in osteocytes, PTH has little, if any, We previously reported that mice with ed
skeletal effects. inducible PPR deletion (postnatally, Ocy-PPRcKO) fro
m
It has been previously reported that mice (4) have mild osteopenia and homeostatic defects h
ttp
expressing a constitutively active PPR in whereas constitutive ablation of the receptor, as ://w
osteocytes (DMP1-CaPTHR1) have increased achieved in Ocy-PPRKO induce an increase in w
w
trabecular and cortical bone (8, 31), suggesting BMD. We can speculate that the difference .jb
that receptor signaling in osteocytes drives bone between Ocy-PPRcKO and Ocy-PPRKO could be c.o
rg
formation. Interestingly, mice lacking receptor ascribed mostly to the timing of receptor ablation b/
y
expression in osteocytes, the Ocy-PPRKO mice, (postnatally vs. embryonic), different promoter g
u
e
also displayed increased trabecular and cortical activities or tamoxifen effects independent of PPR st o
bone. The apparently similar phenotype, i.e., activation. Moreover, in the conditional PPR n A
ipnacrraemaseete rs,i n obtsrearbveecdu lari n amndu rinceo rtmicoadl elsb onoef kdnepoecnkd-oeuntt othne t hsee vaegrei toyf omf icteh ea t pwhheincoht ytpaem omxiigfehnt pril 5, 2
0
constitutive expression (DMP1-caPTHR1) or treatment was started, the dose of tamoxifen 1
9
deletion (Ocy-PPRKO) of PPR in osteocytes is administered, and the duration of receptor
driven by two distinct mechanisms. The increased ablation. To investigate whether a transient
trabecular and cortical bone in DMP1-caPTHR1 is osteopenia was present in the Ocy-PPRKO as
a result of increased osteoblast and osteoclast well, we performed microCT analysis on L5
activities, increased bone formation rate, and high vertebrae and femurs of mice at ~5 weeks of age
bone turnover. In contrast, the increased bone in (similar age as for the inducible animals). As
the Ocy-PPRKO is the net result of reduced reported in Table 1, there were no detectable
osteoblast and osteoclast function, and low bone skeletal differences between controls and Ocy-
turnover state. Bone formation rate in Ocy- PPRKO animals, suggesting that timing of
PPRKO mice is not changed, as is the mineral receptor ablation, promoter activity, and tamoxifen
apposition rate (data not shown). effects might drive the osteopenic phenotype in
Histomorphometric analysis showed a reduced, the inducible model.
although not significant, number of both Increased osteoblast function in DMP1-
osteoblasts and osteoclast, and, upon PTH caPTHR1 was associated with decreased
administration, this cellular defect became evident Sclerostin expression (31). Similarly, decreased
(see Table 5). We can hypothesize that, under osteoblast function in Ocy-PPRKO was associated
normal conditions, the lack of receptor expression with increased Sclerostin expression, supporting
10