Received: 13 January 2020 | Accepted: 2 April 2020
DOI: 10.1002/pros.23983
OR I G I N A L AR T I C L E
Canine prostatic cancer cell line (LuMa) with osteoblastic
bone metastasis
Said M. Elshafae PhD1,2,3 | Wessel P. Dirksen PhD1,4 |
Aylin Alasonyalilar‐Demirer PhD1,5 | Justin Breitbach DVM1 | Shiyu Yuan BS6 |
Noriko Kantake PhD6 | Wachiraphan Supsavhad DVM, PhD7 | Bardes B. Hassan PhD1,8 |
Zayed Attia MS1,9 | Lucas B. Alstadt BS1 | Thomas J. Rosol DVM, PhD, MBA6
1Department of Veterinary Biosciences,
College of Veterinary Medicine, The Ohio Abstract
State University, Columbus, Ohio
2 Background: Osteoblastic bone metastasis represents the most common complica-Department of Pathology, Faculty of
Veterinary Medicine, Benha University, Benha, tion in men with prostate cancer (PCa). During progression and bone metastasis,
Egypt
PCa cells acquire properties similar to bone cells in a phenomenon called osteomi-
3Department of Neuroscience and
Pharmacology, Carver College of Medicine, micry, which promotes their ability to metastasize, proliferate, and survive in the
University of Iowa, Iowa City, Iowa bone microenvironment. The mechanism of osteomimicry resulting in osteoblastic
4Department of Veterinary Clinical Sciences, bone metastasis is unclear.
College of Veterinary Medicine, The Ohio
State University, Columbus, Ohio Methods: We developed and characterized a novel canine prostatic cancer cell line
5Department of Pathology, Faculty of (LuMa) that will be useful to investigate the relationship between osteoblastic bone
Veterinary Medicine, Bursa Uludag University,
Bursa, Turkey metastasis and osteomimicry in PCa. The LuMa cell line was established from a primary
6Department of Biomedical Sciences, Heritage prostate carcinoma of a 13‐year old mixed breed castrated male dog. Cell proliferation
College of Osteopathic Medicine, Ohio and gene expression of LuMa were measured and compared to three other canine
University, Athens, Ohio
7 prostatic cancer cell lines (Probasco, Ace‐1, and Leo) in vitro. The effect of LuMa cells onDepartment of Pathology, Faculty of
Veterinary Medicine, Kasetsart University, calvaria and murine preosteoblastic (MC3T3‐E1) cells was measured by quantitative
Bangkok, Thailand
reverse‐transcription polymerase chain reaction and alkaline phosphatase assay. LuMa
8Department of Pathology, Faculty of
Veterinary Medicine, Cairo University, Giza, cells were transduced with luciferase for monitoring in vivo tumor growth and metastasis
Egypt using different inoculation routes (subcutaneous, intratibial [IT], and intracardiac [IC]).
9Department of Animal Medicine and Xenograft tumors and metastases were evaluated using radiography and histopathology.
Infectious Diseases, Faculty of Veterinary
Medicine, Sadat City University, Sadat City, Results: After left ventricular injection, LuMa cells metastasized to bone, brain, and
Egypt adrenal glands. IT injections induced tumors with intramedullary new bone forma-
Correspondence tion. LuMa cells had the highest messenger RNA levels of osteomimicry genes
Thomas Rosol, DVM, PhD, Department of (RUNX2, RANKL, and Osteopontin [OPN]), CD44, E‐cadherin, and MYOF compared to
Biomedical Sciences, Heritage College of
Osteopathic Medicine, 1 Ohio University, 225 Ace‐1, Probasco, and Leo cells. LuMa cells induced growth in calvaria defects and
Irvine Hall, Athens, OH 45701. modulated gene expression in MC3T3‐E1 cells.
Email: rosolt@ohio.edu
Conclusions: LuMa is a novel canine PCa cell line with osteomimicry and stemness
properties. LuMa cells induced osteoblastic bone formation in vitro and in vivo.
LuMa PCa cells will serve as an excellent model for studying the mechanisms of
osteomimicry and osteoblastic bone and brain metastasis in prostate cancer.
K E YWORD S
bone, canine, dog, metastasis, osteoblast, prostate cancer, tumor
698 | © 2020 Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/pros The Prostate. 2020;80:698–714.
ELSHAFAE ET AL. | 699
1 | INTRODUCTION characterized cell lines were developed.21,23,24,26 Few cancer and tu-
mor cell lines have been developed that recapitulate the osteoblastic
Prostate cancer (PCa) is the most common malignant tumor and the nature of PCa. These include Ace‐1, Probasco, LuCAP‐23.1, and MDA‐
second greatest cause of cancer‐related death in men after lung cancer PCa 2a & b. 27‐33 In this study, we have established and characterized a
in Western countries. One of the most frequent complications in PCa novel canine prostatic cell line (LuMa) that has unique osteomimicry
patients is bone metastasis, which is most often osteoblastic (bone‐ properties, stem cell and invasive characteristics, and induces osteo-
inducing). Some tumor cells that undergo skeletal metastasis gain fea- blastic bone metastases in nude mice. These properties demonstrate
tures that are usually restricted to bone cells, especially osteoblasts, and that LuMa cells will be a valuable model in prostate cancer research.
this process has been termed osteomimicry.1‐3 Acquiring osteoblastic
properties by cancer cells promotes their survival and growth.4 In ad-
dition, some neoplastic cells at their primary site have acquired osteo- 2 | MATERIALS AND METHODS
mimicry and this promotes their metastasis to bone. The timing and
mechanisms of osteomimicry acquisition are not well understood. 2.1 | Cell lines, calvaria, and frozen tissues
Previous studies have shown that several factors including RANKL,
insulin‐like growth factor 1 (IGF‐1), bone morphogenetic protein 2 Frozen samples of dog normal prostates (N = 2), benign prostate hy-
(BMP‐2), and transforming growth factor beta (TGFβ) were involved in perplasia (N = 2), and primary prostatic cancers (OC, CB, Probasco, and
acquiring an osteomimicry phenotype in prostate cancer.5,6 These factors LuMa) were obtained from The Ohio State University (OSU) College of
activate intracellular signaling molecules such as Wnt, nuclear factor κB Veterinary Medicine Biospecimen Repository. Calvaria were removed
(NF‐κB), and Twist that consequently promote the osteogenic program in from 5 to 12‐day old nude mouse pups (NCr‐nu/nu) supplied by the
PCa cells.7‐11 RANKL‐RANK is an important signaling pathway that ac- OSU Comprehensive Cancer Center (OSUCCC) Target Validation
tivates several transcription factors that have a role in multiple processes, Shared Resource (TVSR). Four canine prostate cancer cell lines were
including osteomimicry (Sox2, HIF1α, and Sox9), epithelial‐to‐ used in this study including Ace‐1,27 Probasco,33 Leo,34 and LuMa
mesenchymal transition (EMT and Twist1), neuroendocrine differentia- (available from ABM Inc, Vancouver, Canada).
tion (FoxA2, Sox9, and HIF1α) and stem cell properties (Nanog and
Sox2).12 Some prostate and breast cancer cells mimic the osteoblast
phenotype by expressing bone matrix factors such as, OPN,13,14 osteo- 2.2 | Establishment and validation of canine
calcin (OC)15 and bone sialoproteins.13,16‐18 In addition, cancer cells can prostate carcinoma cell line (LuMa)
express alkaline phosphatase (ALP)18 and Runt‐related transcription
factor 2 (Runx2) proteins, which are important osteoblastic markers.19 A sample of fresh cancer tissue (about a 1 cm3 antemortem biopsy)
Runx2 signaling is important in the regulation of bone home- was removed from a primary prostate carcinoma of a 13‐year old
ostasis and development. Knockout of RUNX2 led to early prenatal mixed breed male castrated dog the day before euthanasia and sub-
mortality in mice due to the lack of osteoblast differentiation.20 sequent necropsy. The tumor tissue was washed 3× with sterile
Runx2 plays a role in mammary tumorigenesis by increasing pro- Dulbecco phosphate‐buffered saline (DPBS) and minced into small
liferation and inhibition of apoptosis in mammary acini.21 Runx2 also pieces (≤1mm3) with a sterile scalpel. The pieces were cultivated in
directly induces genes associated with angiogenesis, invasiveness and Dulbecco's modified Eagle medium (DMEM)/F12 medium supple-
metastasis including OPN, vascular endothelial growth factor (VEGF), mented with 20% fetal bovine serum (FBS) (only for the first 2 weeks)
and matrix metallopeptidase 9 (MMP‐9), and promotes EMT of primary or 10% FBS (in the remaining weeks) and 1% penicillin/streptomycin
cancers.21‐23 Upregulation of RUNX2 was accompanied by the in- (100 unit/mL penicillin and 100 µg/mL streptomycin) in a 75 cm2 tissue
crease in Gleason score and metastasis that occurred in prostate culture flask. The medium was replaced every other day in the first
cancer.22,24 Furthermore, Runx2 enhanced growth and migration and 2 weeks and every 3 days afterward. Differential trypsinization with
promoted osteolytic activity of breast cancer. The messenger RNA 0.25% trypsin/ethylenediamine‐tetra acetic acid (EDTA) was used for
(mRNA) expression level of RUNX2 was shown to be greater in the 10 to 30 seconds before medium change to remove stromal cell con-
metastatic human prostatic cell line, PC3, compared to less‐ tamination. The epithelial cells were allowed to become 70% to 90%
metastatic cells lines (LNCaP, C4‐2B, and RWPE).25 confluent before passaging. LuMa cells (1 × 106) passage 78 were
Metastatic cancer cells interact with osteoclasts and osteoblasts submitted to IDEXX BioResearch (Columbia, MO) to perform short
in bone to stimulate their activity and disturb the normal balance of tandem repeat (STR) DNA profiling and multiplex PCR analysis for
bone remodeling. The imbalance may promote an osteosclerotic/os- detection of any cross contamination or misidentification.
teoblastic response, as in PCa, or an osteolytic process, as in multiple
myeloma and breast cancer. The interactions between PCa cells and
bone cells and the PCa vicious cycle are not completely understood. To 2.3 | Lentiviral luciferase transduction
improve our understanding of the mechanism of skeletal metastasis in
prostate cancer and the interaction between PCa and the bone mi- LuMa cells (100 000) were cultivated with DMEM/F12, 10% FBS,
croenvironment, animal models of skeletal metastasis using well and 1% penicillin/streptomycin (culture medium) in a six‐well plate.
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
700 | ELSHAFAE ET AL.
Once LuMa cells were approximately 90% to 95% confluent, 2.6 | Induction of osteoblast differentiation
the medium was replaced with 100 µL of luciferase‐containing
virus (VC 2192 ConcpLuc [VSV‐G]),33 1.5 mL of DMEM/F12 Murine preosteoblast cells, MC3T3‐E1 (ATCC, Manassas, Virginia)
culture medium, and 1.6 µL of polybrene stock (8 µg/mL). The plate (200 000 cells), were cultivated in six (25 cm2) flasks and cultured for
was centrifuged at 2700 rpm for 1 hour at 30°C and incubated at 2 days in α‐minimum Eagle medium (α‐MEM) supplemented with 10%
37°C and 5% CO2 for 48 hours after which the virus‐containing FBS and 1% penicillin/streptomycin. On day 3, the media were re-
medium was removed and replaced with culture medium. placed in three flasks with α‐MEM medium preincubated with LuMa
After 48 hours the cells were evaluated for bioluminescence cells (LuMa condition medium [CM]) for 48 hours. Photographs of
using an IVIS 100 (Caliper Life Sciences, Hopkinton, MA) and MC3T3‐E1 cells were taken at day 5 of the experiment using a Nikon
photon signal intensity was quantified using Living Image 300 Diaphot inverted microscope supplied with a Tucsen camera. To
software version 2.50 (Caliper Life Sciences). Luciferase transfected confirm the differentiation of MC3T3‐E1 cells, ALP activity assay
LuMa cells were trypsinized and cultivated in T75 tissue culture (Anaspec, Fremont, CA) and qRT‐PCR of osteogenic‐related genes
flasks. (OPN and OC) were performed for the MC3T3‐E1 cells.
2.4 | In vitro growth rate 2.7 | Coculture of mouse calvaria with LuMa and
ACE‐1 cells
LuMa cells (passage 15) were seeded in six‐well culture plates in
triplicate. LuMa cells were trypsinized and harvested at days 1, 3, To evaluate the induction of new bone in the presence of LuMa
and 5. The cell number, size, and viability were counted using an cells, an in vitro bone formation model was used. Calvaria of 5‐ to
automated cell counter (Nexcelom Bioscience, Lawrence, MA) and 12‐day old pups were aseptically removed and dissected from
trypan‐blue dye exclusion to differentiate between live and dead surrounding soft tissues. Two bone disks (1.5 mm) were punched
cells. The doubling time was calculated using the following formula: and removed from the parietal bone on both sides in each calvaria
doubling time = duration (hours) × log (2)/[log (final concentration) − using biopsy punches to create artificial defects. Calvaria were
log (initial concentration)]. washed twice in DPBS and basal medium consisting of BGJb cul-
ture medium (Gibco, Invitrogen, Carlsbad, CA) supplemented with
0.1% bovine serum albumin (BSA) fraction V (Sigma‐Aldrich Corp,
2.5 | RNA extraction and quantitative reverse‐ St Louis, Missouri) and 100 μg/mL Normocin (Invivogen, San Diego,
transcription polymerase chain reaction (qRT‐PCR) CA). The calvaria with defects were divided into three groups and
incubated for 12 days in BGJb culture medium ± LuMa cells or Ace‐
RNA was extracted from canine prostatic cancer cell lines (LuMa, 1 cells (10 000 cells seeded on the first day) in six‐well plates.
Probasco, Ace‐1, and Leo), benign prostatic hyperplasia (BPH), BGJb medium was changed every other day in both control
normal canine prostate tissues, murine MC3T3 E1 cells ± LuMa (BGJb + calvaria only) and treatment groups (LuMa or Ace‐
conditioned media and LuMa cells ± mouse calvaria using the 1 + BGJb + calvaria) until the end of the experiment. The calvarial
QuickGene RNA Extraction Kit (AutoGen, Holliston, MA, Cat. No. FK‐ defects were imaged at day 12 using a dissecting microscope
RC‐S2). Reverse‐transcription (RT) of total RNA was performed using (Nikon SMZ‐U) and Nikon Diaphot 300 inverted‐phase contrast
the Superscript II First Strand cDNA synthesis kit (Invitrogen). microscope supplied with a Tucsen camera.
Quantitative RT polymerase chain reaction (PCR) was performed for
the reference gene, glyceraldehyde 3‐phosphate dehydrogenase
(GAPDH) for LuMa cells, and ubiquitin C (UBC) for MC3T3‐E1 cells, 2.8 | Tartrate‐resistant acid phosphatase and ALP
as well as for the osteomimicry, stemness, and prostate cancer staining
progression and metastasis genes including myoferlin (MYOF),
runt‐related transcription factor 2 (RUNX2), snail homolog 1 (SNAIL), Calvaria were rinsed with deionized water for 10minutes and in-
RANKL, SPP1 (osteopontin, OPN), E‐cadherin, TWIST, VIMENTIN, an- cubated with 0.2M citrate buffer for 5minutes at room temperature.
drogen receptor (AR), calcium‐sensing receptor (CaR), folate The calvaria were incubated with 50mM sodium L‐tartrate dehydrate
hydrolase (FOLH1 or PSMA), TGFβ, CD44, and CD133 (PROMININ) (Sigma‐Aldrich) in deionized water for 1 hour and transferred to
using canine‐specific primers (Table 1). qRT‐PCR was completed for acetate buffer containing 0.5 mg/mL naphthol AS‐MX phosphate
the following genes in MC3T3‐E1 cells; OPG (osteoprotegerin), disodium salt (Sigma‐Aldrich) and 1.1 mg/mL Fast Red TR salt 1,
RANKL, OPN, OC and RUNX2. For testing primer specificity, all qRT‐ 5‐naphthalenedisulfonate salt (Sigma‐Aldrich). Calvaria were kept at
PCR products were verified by electrophoresis on a 2% agarose gel 37°C for 2 hours until the red color product was developed and then
and stained with ethidium bromide to confirm a single amplification rinsed using deionized water.
product of the expected size. Sequences were verified by a BLAST ALP staining of calvaria was performed using the Vector blue
search using the NCBI website. substrate kit (Burlingame, CA, Cat. No. SK‐5300). Calvaria were
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ELSHAFAE ET AL. | 701
TABLE 1 Primers used for qRT‐PCR in LuMa and mouse MC3T3‐E1 cells
Gene Forward primers Reverse primers
Canine primers
TWIST GGCAGGGCCGGAGACCTAGATG TCCACGGGCCTGTCTCGCTT
SNAIL GTCTGTGGCACCTGCGGGAAG GAAGGTTGGAGCGGTCGGCA
RUNX2 TGCCTCTGGCCTTCCACTCTCAG TGCATTCGTGGGTTGGAGAAGCG
CaR TCTCCACGGCTGTGGCAAACC GAGGAGTCTGCTGGAGGAGGCAT
E‐cadherin GCTGCTGACCTGCAAGGCGA GGCCGGGGTATCGGGGACAT
GAPDH CCCACTCTTCCACCTTCGAC AGCCAAATTCATTGTCATACCAGG
TGFβ GGCAGAGTTGCGCCTGCTGA CCGGTTGCTGAGGTAGCGCC
MYOF TGCCCCCGAAAGGCTGGGAAT ACTCCGTGTGCCCTGCGTCT
RANKL TCCGAGCCGCTGTACAAAA AGTATGAGTCTTGCCCCTCCT
VIMENTIN GAGGACATCATGCGGCTGCGG CGCTCAAGGTCAAGACGTGCC
Osteopontin GGTTCATATGATGGCCGAGGT CAGAGGTGCCTCTCACTGTC
OPG ATGCCCAGATGGGTTCTTCTC AGAATGCCTCCTCACACAAGG
CD44 CACCTCCCAGTACGACACG CATCGTCAGTGGGGTTGCT
CD133 CATTCACCGCAATTTGCCCA ATGAGGGTCAGCAAACAGCA
Mouse primers
OPG AGCTGCTGAAGCTGTGGAA TCGAGTGGCCGAGAT
UBC CGTCGAGCCCAGTGTTACCACCAAGAAGG CCCCCATCACACCCAAGAACAAGCACAAG
RUNX2 CGACAGTCCCAACTTCCTGT TACCTCTCCGAGGGCTACAA
RANKL GCTGGCTACCACTGGAACTC TGTGCACACCGTATCCTTGT
Osteocalcin CTCACAGATGCCAAGCCC CCAAGGTAGCGCCGGAGTCT
Osteopontin GCTTGGCTTATGGACTGAGG CGCTCTTCATGTGAGAGGTG
incubated with substrate working solution for 30 minutes in the 2.10 | Confocal laser scanning visualization
dark, washed in 200 mM Tris‐HCl for 5 minutes and rinsed in
deionized water. Calvaria were imaged using a dissecting micro- Runx2 and ALDH1 protein expression levels were evaluated in both
scope with camera after Tartrate‐resistant acid phosphatase Ace‐1 and LuMa cells using confocal microscopy. A total of 100 000
(TRAP) and ALP staining. Immunofluorescence of ALP in osteo- Ace‐1 and LuMa cells were seeded on cover slips in 24‐well plates
blasts and newly formed bone using the Vector AP kit was de- and incubated for 24 hours at 37°C and 5% CO2. On the next day, the
tected and imaged using a Nikon Diaphot 300 fluorescent cells were washed twice with PBS and fixed with 4% paraformalde-
microscope. hyde for 30minutes followed by PBS washing (3×) and permeabili-
zation using methanol for 30 seconds. Ace‐1 and LuMa cells were
washed twice with PBS and incubated with 5% heat‐inactivated goat
2.9 | Flow cytometry serum for 2 hours. The cells were incubated with Runx2 (Cat No
ab23981; 1:100 dilution; Abcam) or ALDH1A1 (Cat No ab23375;
Expression of Runx2 was measured in four prostatic cancer cell 1:100 dilution; Abcam) primary antibodies for 2 hours at room
lines (Ace‐1, Probasco, Leo, and LuMa) using flow cytometry. The temperature followed by PBS washing (3×). Secondary anti‐rabbit
cultured PCa cells were washed twice with DPBS, trypsinized, fluorescent antibody (Cat No FI‐1000; 1:1000 dilution) was added to
centrifuged and counted using a Cellometer Auto T4 (Nexcelom the cells for 1 hour at room temperature. The cells were washed 3×
Bioscience). One million cells were incubated with anti‐Runx2 pri- with PBS and incubated with 1 µg/mL DAPI (Cat No 62248; Life
mary antibody (Cat No ab23981, 1:100 dilution; Abcam, Cambridge, Technologies, Grand Island, NY) for 10minutes at room temperature.
MA) for 2 hours at room temperature. The cells were incubated Images were captured using laser scanning confocal fluorescence
with fluorescein goat antirabbit immunoglobulin G antibody (FI‐ microscope with a 60× objective (Olympus Fluoview FV10i).
1000, 1:100 dilution; Vector Laboratories, Burlingame, CA) for
30 minutes at 4°C, washed three times with FACS buffer (PBS, 1%
BSA), and centrifuged at 1500 rpm for 5 minutes. Cells were ana- 2.11 | ALP activity assay
lyzed by flow cytometry (Accuri C6; BD Biosciences, San Jose, CA).
The data were analyzed with Accuri C6 Flow software (BD The LuMa cells (before and after 5 days of incubation with calvaria)
Biosciences). and MC3T3‐E1 cells (before and after 3 days of incubation with
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
702 | ELSHAFAE ET AL.
LuMa CM) were lysed using 1× assay buffer and Triton X‐100 (0.2%), (1 mm to the left and lateral from the sternum). Once a jet of blood
agitated for 5minutes and centrifuged at 2500 g for 10minutes at was present in the hub of the needle, the cell suspension was then
4°C. The standards, LuMa cells, and MC3T3‐E1 cells were loaded in a slowly injected for 30 seconds.33 Bioluminescent imaging was per-
96‐well plate. All samples were incubated with ALP substrate formed directly postinjection (7‐10minutes) to ensure successful IC
solution (p‐nitrophenylphosphate [pNPP]) at room temperature for injection and weekly to monitor for metastases.
1 hour, the stop solution was added and the absorbance was read at
405 nm (Perkin Elmer, CA). The ALP activity (u, unit) of cell lysates
was measured based on the ALP standard curve. 2.13 | Bioluminescent imaging
D‐Luciferin (0.1 mL of 40mg/mL, dissolved in DBPS) was injected
2.12 | In vivo experiments intraperitoneally in each mouse before anesthesia using a 1mL tu-
berculin syringe (Caliper Life Sciences). Mice were anesthetized with
All animal experiments were approved by the Institutional Animal Care 3% isoflurane and maintained with 2% isoflurane during biolumi-
and Use Committee (IACUC) of The Ohio State University Institutional nescent imaging. The IVIS 100 (Caliper Life Sciences, Hopkinton, MA)
Laboratory Animal Care and Use Committee. Athymic 5‐7‐week‐old was used to detect the bioluminescence and the photon signal in-
male nude mice (NCr‐nu/nu) were purchased from the OSU Compre- tensity (total photons/sec) was measured for each region of interest
hensive Cancer Center (OSUCCC) Target Validation Shared Resource using Living Image software version 2.50 (Caliper Life Sciences).
(TVSR). Mice were maintained according to the NIH standards estab- Imaging was performed every 1minute until the peak signal was
lished in the “Guidelines for the Care and Use of Laboratory Animals.” obtained (10‐15minutes postinjection). The IVIS 100 was set to a
1‐minute exposure with medium binning.
2.12.1 | Subcutaneous, intratibial, and intracardiac
injection of LuMa cells into nude mice 2.14 | Radiography
Subcutaneous Radiography was performed on the right hind leg in IT and both
Two million luciferase/YFP‐transduced LuMa cells (LuMa‐Luc) sus- the right and left hind legs for IC experiments. Formalin‐fixed
pended in 0.25mL of sterile DPBS were injected subcutaneously in four legs were soaked in water for 8 hours and then placed centrally
mice above the right shoulder using a 25‐guage needle. Tumor volume on a Faxitron laboratory radiography system LX‐60 (Faxitron
was measured twice weekly using a digital caliper. Three dimensions X‐ray Corp, Wheeling, IL) imaging platform and high resolution
were measured, and tumor volume was calculated using the formula; DICOM radiograph images were taken at 25 KeV with 5‐second
length ×width × height × 1/2. Tumor growth rate in SQ xenografts was exposures.
measured weekly using bioluminescent imaging. The doubling time for
the tumor volume was calculated using the formula; doubling time =
duration (days) × log (2)/log (final volume/initial volume). 2.15 | Histopathology
Intratibial The animals were euthanized after 21 days in the IC experiment and
Nude mice (N = 5) were maintained under isoflurane anesthesia 30 days in both the SQ and IT experiments. Tumors, tissue specimens
(2.5%) in supine position during intratibial (IT) injection. The right leg and bones were collected at necropsy. Tumor and tissue specimens
was held and the knee joint was bent so the femur made a 90° angle were fixed in 10% neutral‐buffered formalin at room temperature for
with the tibia. LuMa‐Luc cells (50 000 in 10 µL) were loaded in a 72 hours, embedded in paraffin, cut in 4 µm sections, and stained
Hamilton syringe with 27‐guage needle. The needle was placed with hematoxylin and eosin (H&E). Bones were decalcified with mild
through the patellar ligament and the articular cartilage of the tibia Decalcifier (formaldehyde, methanol and formic acid; Leica Biosys-
into the metaphyseal marrow space of the tibia. IT tumor growth was tems, Buffalo Grove, IL) at room temperature for 6 hours. Histolo-
monitored weekly by bioluminescent imaging.33 gical images of the slides were taken using an Olympus BX51
microscope equipped with a Nikon digital camera and analyzed using
Intracardiac ImageScope software (version 11.2; Leica Biosystems, Buffalo
Nude mice (N = 6) were anaesthetized using 3% isoflurane initially Grove, IL).
and 2.5% during the intracardiac (IC) procedure. The mice were
placed on their back and their front and hind legs were restrained to
the procedure table with surgical tape. A tuberculin syringe with 2.16 | Immunohistochemistry
0.1 mL of DPBS containing 100 000 LuMa‐Luc cells and 0.1 mL of air
was prepared. The syringe with a 27‐gauge needle was introduced Paraffin‐embedded tissue sections (4‐μm thick) were preheated at
into the left ventricle of the heart through the third intercostal space 60°C for 1 hour, dewaxed, rehydrated and incubated with antigen
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ELSHAFAE ET AL. | 703
retrieval solution (Dako, Carpinteria, CA) for 50minutes in a steamer tumor implantation, and the average tumor volume was 2.1 ± 1.2 cm3.
(Black & Decker HS1000, Newark, DE) followed by slow cooling. To SQ xenografts were easily dissectible from the SQ tissue. The his-
eliminate endogenous peroxidase activity, slides were treated with tological characteristics of the SQ tumors were similar to the primary
3% hydrogen peroxide (H2O2), washed, and coated with serum free prostatic carcinoma and contained foci of necrosis (Figure 1C,D).
protein block (Dako). Tissue sections were incubated with primary
antibodies to RUNX2, E‐cadherin, cytokeratin AE1/AE3, PTHrP, and
Twist overnight at 4°C (Table 2). Secondary biotinylated antibodies 3.2 | In vitro and in vivo growth and culture
were applied on the second day for 30minutes, followed by in- characteristics
cubation with diaminobenzidine (DAB; Dako) for 5 minutes and
counterstained with hematoxylin. Sections were dehydrated through The LuMa cells formed a cobblestone growth pattern in culture. The
graded alcohols and xylene and coverslipped. DAB brown color average diameter of LuMa cells in cell culture was 13.5 to 15.3 mi-
staining was recorded as a positive reaction. Positivity and intensity crons. LuMa cells grew in a monolayer sheet consisting of polygonal
of staining was evaluated as weak, moderate, or strong using Aperio cells and few moderate to large oval cells (22‐28 microns) with a
digital scans, ImageScope software (version 11.2; Leica Biosystems) large vesicular nucleus and prominent nucleoli (Figure 2A). The
and the algorithm positive pixel count analysis (version 9). doubling time of LuMa cells in culture was 30.8 ± 1.4 hours in stan-
dard culture conditions (Figure 2B). In vivo, the doubling time of the
tumor volume in LuMa SQ xenografts was approximately 8.6 ± 2.6
2.17 | Statistical analysis days (Figure 2C).
All data were displayed as mean ± standard deviation. One‐way ana-
lysis of variance was used to analyze normalized gene expression data 3.3 | STR DNA profiling and multiplex PCR analysis
and tumor volume and bioluminescence in mice followed by Sidak's of LuMa cells
multiple comparisons test using GraphPad Prism version 6.03 (La Jolla,
CA). Data with P ≤ .05 were considered statistically significant. LuMa cell line was confirmed to be of canine origin with no mam-
malian interspecies contamination. A genetic profile for LuMa was
generated using a panel of 14 STR markers for genotyping compared
3 | RESULTS to Ace‐1 cells (Supporting Information Table S).
3.1 | Histopathology of the primary carcinoma and
subcutaneous xenograft 3.4 | Immunohistochemistry, confocal microscopy,
and flow cytometry
The primary prostate carcinoma had a papillary to cribriform pattern
(Figure 1A,B). Foci of transitional cell differentiation with vacuolated Immunohistochemistry revealed strong staining of Runx2 in ;the cy-
cells were present. Desmoplasia was pronounced throughout the tumor toplasm of LuMa cells in both SQ and bone xenografts (Figure 3A,B).
parenchyma with multifocal areas of necrosis and squamous metaplasia. Primary LuMa tumors were strongly positive for epithelial markers
Chronic submucosal edema and urothelial hyperplasia were reported in (cytokeratin AE1/AE3 and E‐cadherin; Figure 3C,D). LuMa primary
the bladder with mild mesothelial hypertrophy and hyperplasia. No cancer cells had intense cytoplasmic and nuclear staining for PTHrP
neoplastic changes were observed in the bladder epithelium. Multifocal and moderate cytoplasmic staining for Twist (Figure 3E,F).
metastatic well‐differentiated carcinomas with central necrosis and Flow cytometric analysis revealed that few Ace‐1, Probasco, and
mineralization were present in the lungs of the dog. Leo cells were positive for Runx2 (0.2 ± 0.05%, 0.16 ± 0.05%, and
LuMa subcutaneous (SQ) xenografts were grossly visible in nude 2.8 ± 1.5%, respectively; Figure 4A‐C) while many LuMa cells
mice 1 week after implantation. Mice were euthanized 1 month after expressed Runx2 (12.4 ± 0.43%; Figure 4D). LuMa and Leo cells
TABLE 2 Antibodies used for immunohistochemistry of LuMa primary tumor and mouse xenografts
Antibody Species Clone Company Dilution
CK AE1/AE3 Mouse M3515 Dako Corporation, Carpinteria, CA 1:100
Twist Rabbit H81 Santa Cruz Biotechnology, Santa Cruz, CA 1:150
PTHrP Goat N‐19 Santa Cruz Biotechnology, Santa Cruz, CA 1:200
Runx2 Rabbit ab23981 Abcam, Cambridge, MA 1:100
E‐cadherin Mouse Clone 36 BD Biosciences, San Jose, CA 1:200
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
704 | ELSHAFAE ET AL.
F IGURE 1 Histopathology of LuMa
primary prostatic carcinoma (A,B) and the
subcutaneous xenografts (C,D) in nude
mice (H&E stain). LuMa neoplastic cells in
the primary tumor had a cribriform
pattern (A,B) (200× and 400×,
respectively). (C,D) LuMa cells in SQ
xenograft tumors had a cribriform to
papillary pattern with transitional cell‐like
differentiation (200×) [Color figure can be
viewed at wileyonlinelibrary.com]
F IGURE 2 In vitro and in vivo growth characteristics of LuMa cells. A, Phase contrast microscopy of LuMa cells in culture with a cobblestone
pattern. B, In vitro growth curve of LuMa cells in vitro. Data are mean ± SD of three replicates for each time point. C, Graph shows mean LuMa
tumor volume in SQ xenografts. Data presented as a mean ± standard deviation from five mice
had greater Runx2 (46 ± 1.6‐fold and 11 ± 0.9‐fold, respectively) CDH1 (E‐cadherin) (P = .049) compared to normal prostate tissue, BPH,
compared to the Ace‐1 cells. and canine prostatic cancer cell lines (Ace‐1, Probasco, and Leo)
The differences between Ace‐1 and LuMa cells in Runx2 and (Figure 5A,B). There was little to moderate mRNA expression of me-
ALDHA1 expression and cellular distribution were also examined by senchymal markers (VIMENTIN and SNAIL) in LuMa cells compared to
confocal microscopy. Both Ace‐1 and LuMa cells were positively Ace‐1 cells (Figure 5C). There was weak to no expression of AR, TWIST,
stained with Runx2 and ALDHA1 antibodies. Runx2 staining was CDH2 (N‐cadherin), GRPR, and IGF‐1 mRNA in LuMa cells (data not
strong and diffuse in the cytoplasm of LuMa cells compared to Ace‐1 shown). LuMa cells expressed significantly less FOLH1 mRNA than Leo
cells which had weak cytoplasmic staining (Figure 4E,F). Intense (P = .0044) and Probasco (P = .0429) cells and a similar level of CD133
ALDH1A staining was observed mainly around the nuclei of LuMa mRNA compared to Probasco cells (Figure 5D).
cells compared with Ace‐1 cells (Figure 4G,H).
3.6 | Effect of LuMa on bone formation in mouse
3.5 | qRT‐PCR of LuMa cells in vitro calvaria defects
LuMa cells had significantly greater mRNA expression for RUNX2 LuMa cells improved the healing response of calvarial defects when
(P = .037), MYOF (P = .029), CD44 (P = .0002), OPN (P = .0001) and compared to Ace‐1 and control calvaria. The calvarial defects in the
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ELSHAFAE ET AL. | 705
F IGURE 3 IHC staining of Runx2 in LuMa
subcutaneous (A) and intratibial (B) xenograft
tumors (200×). IHC staining of cytokeratin AE1/
AE3 (C) (200×), E‐cadherin (D) (400×), PTHrP (E)
(200×) and Twist (F) (200×) in LuMa primary
prostatic carcinoma. LuMa cells were strongly
positive for Runx2, cytokeratin, E‐cadherin,
PTHrP and moderately positive for Twist. Many
LuMa cells had nuclear staining for PTHrP. IHC,
immunohistochemistry [Color figure can be
viewed at wileyonlinelibrary.com]
control group (without any cancer cells) had no stromal cell growth 3.7 | Effect of LuMa on MC3T3‐E1 osteoblast
and an eroded border with no evidence of osteoblast cell prolifera- differentiation
tion (Figure 6A,B). Calvaria incubated with Ace‐1 cells had mild to
moderate stromal growth in the defects with eroded margins There were morphological differences in MC3T3‐E1 cells after
(Figure 6C,D). The calvarial bone defects developed growth of os- incubation with LuMa CM. MC3T3‐E1 cells formed a monolayer
teoblasts and stromal cells into the defects with bone mineralization of ovoid to pyriform cells with clear cytoplasm (Figure 7A).
in culture with LuMa cells (Figure 6E‐H). After incubation with LuMa CM, MC3T3‐E1 cells had many small va-
ALP and TRAP staining were performed to identify osteoblasts cuoles in the cytoplasm close to the cell border (Figure 7B). To de-
and osteoclasts in the reparative tissue. The majority of the cells termine whether LuMa CM induced MC3T3‐E1 cell differentiation,
were ALP‐positive blue (using an inverted microscope) or green ALP activity and the expression of osteogenic‐related genes mRNA
(using a fluorescent microscope) and were interpreted as osteoblasts were measured in MC3T3‐E1 cells. Cotreatment of MC3T3‐E1
(Figure 6I‐K). There were few TRAP‐positive osteoclasts at the bor- cells with LuMa CM increased ALP activity in the MC3T3‐E1
der of the defects (data not shown). cells compared to untreated cells (P = .011) (Figure 7C). In addition,
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
706 | ELSHAFAE ET AL.
F IGURE 4 Flow cytometric and confocal
microscopic analysis of Runx2 and ALDH1 in
canine prostate cancer cell lines. A‐D, Flow
cytometric analysis of Ace‐1 (A), Probasco (B),
Leo (C), and LuMa (D). The number in the lower
right corner indicates the percentage of Runx2‐
positive cells in each cell line. E‐H, Confocal
microscopic images of Runx2 (E‐F) and ALDH1
(G‐H) expression level in Ace‐1 (E, G) and LuMa
(F, H) cell lines [Color figure can be viewed at
wileyonlinelibrary.com]
LuMa CM upregulated the mRNA expression level of RUNX2 (P = .0062) after incubation with calvaria for 5 days (Figure 7E).
(2.0‐fold) (P = .0017) in MC3T3 E1 cells compared to untreated Calvaria increased the expression of CaR (5.6‐fold) (P = .00028) and
cells (Figure 7D). There was a significant increase in OPG mRNA TGFβ (1.7‐fold) (P = .024) mRNA in LuMa cells. In addition, there was
(2.3‐fold) (P = .00019) in the treated MC3T3‐E1 cells and no change in a trend for RUNX2 expression to be increased (Figure 7F). Calvaria
RANKL mRNA (P = .85) (Figure 7D). This resulted in an increased the downregulated the mRNA expression of OPG (0.4‐fold) (P = .0001)
OPG/RANKL ratio in LuMa CM‐treated MC3T3‐E1 cells (P = .0134). and OPN (0.14‐fold) (P = .0046) in LuMa cells (Figure 7F).
3.8 | Effect of bone (calvaria) and bone CM on LuMa 3.9 | IT injection of LuMa cells in nude mice
ALP activity and gene expression
The tibias of nude mice with LuMa tumors had radio‐opaque areas in
LuMa cells had greater ALP activity (P = .0059) compared to Ace‐1 the epiphysis, metaphysis, and diaphysis of the marrow cavity
cells. However, there was a marked decrease in ALP in LuMa cells (Figure 8A). Histopathological examination of the tibias showed
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ELSHAFAE ET AL. | 707
F IGURE 5 A‐B, The mRNA expression of RUNX2, MYOF, CD44, Osteopontin, and E‐cadherin in normal dog prostate (n = 2), BPH (n = 2) and
prostate cancer cell lines (Ace‐1, Probasco, Leo, and LuMa). C‐D, The expression of mesenchymal markers (VIMENTIN and SNAIL), FOLH1
(PSMA) and CD133 (PROMININ) in normal prostate (n = 2) and BPH (n = 2) tissues, and four canine prostatic cancer cell lines. The graphs
represent the relative mRNA expression in Ace‐1, Probasco, Leo, and LuMa cells in comparison to normal prostate gland. Significant differences
are indicated as *P ≤ .05, **P ≤ .01, ***P ≤ .001, and ****P ≤ .0001 different from normal prostate. BPH, benign prostatic hyperplasia
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
708 | ELSHAFAE ET AL.
F IGURE 6 Effect of LuMa cells on mouse calvarial defects. Dissecting (A‐H), phase contrast (I) and fluorescent (J, K) microscopic images showing the
effect of LuMa cells on growth and mineralization of calvarial defects in vitro. A‐B, The calvarial defect incubated with only BGJb showed regular and
irregular perimeters. C‐D, The calvarial defect incubated with Ace‐1 cells was partially (15‐20%) filled by osteoblasts and mineralized matrix with
irregular perimeters. E‐H, The defect was partially to almost closed with osteoblast and stromal cells in calvaria incubated with LuMa cells. I‐K,
The defects were almost completely closed by osteoblast and stromal cells in LuMa‐incubated calvaria; the osteoblasts stained positive for alkaline
phosphatase (blue, light microscopy) (I) or (green, fluorescent microscopy) (J‐K) [Color figure can be viewed at wileyonlinelibrary.com]
compact sheets of neoplastic LuMa cells and large multifocal areas of moderate osteoblast cell proliferation, new bone formation, and
new intramedullary woven bone lined by hypertrophic cuboidal os- displacement of bone marrow cells (Figure 9C,D). LuMa metastatic
teoblasts adjacent to neoplastic cells. LuMa cells induced new bone tumors occupied most of the vertebral medullary bone and sur-
formation from the endosteal and trabecular surfaces (Figure 8B‐D). rounded by hypertrophic and hyperplastic osteoblasts (Figure 9E).
LuMa metastases were also in the adrenal glands, brain, and man-
dibular alveolar bone (Figure 9F‐H).
3.10 | IC injections and bioluminescent imaging of
LuMa in nude mice
4 | DISCUSSION
The metastasis of LuMa cells after left ventricular intracardiac in-
jection was monitored by bioluminescence imaging. Successful IC Canine prostate cancer is considered an excellent model of human
injections were confirmed using bioluminescence at 7 minutes, which PCa for several reasons. Prostate carcinoma occurs spontaneously in
demonstrated distribution of tumor cells throughout the body older dogs, is usually androgen‐independent, and commonly metas-
(Figure 9A). The bioluminescence diminished after 1 to 2 days and tasizes to bone, lung, and regional lymph nodes.35 Several canine PCa
was negative until day 18. At 3 weeks, four of six nude mice had cell lines have been developed that recapitulate the late stages of
bioluminescence in their long bones (tibia, humerus, and femur), prostate cancer progression and bone metastasis and are useful to
head, vertebrae, and adrenal glands (Figure 9A). Radiographic images investigate the interaction that occurs between PCa cells and bone
demonstrated mild medullary sclerosis in the metaphysis of long microenvironment.36,37 Few human PCa tumor and cell lines can be
bones (tibia and humerus) (Figure 9B). The locations of LuMa me- used to model osteoblastic bone metastasis, these include LuCaP
tastases (confirmed by microscopic examination) were reported 23.1, LAPC‐9, VCaP, and LNCaP C4‐2B.29,34,38,39 MDA‐PCa 2b is a
(Table 3). Histological examination showed that LuMa metastases in human prostatic cell line that forms osteoblastic bone metastasis
the tibias, humeri, and femurs were characterized by mild to after about 2 to 3 months of intraosseous injection.32 In this study,
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ELSHAFAE ET AL. | 709
F IGURE 7 A‐B, Effect of LuMa CM on MC3T3‐E1 cells. Phase‐contrast photomicrographs of control (untreated) (A) and LuMa CM‐treated
(B) MC3T3‐E1 cells. B, MC3T3‐E1 cells had cytoplasmic vacuoles (arrows) after LuMa CM treatment. C‐D, Effect of LuMa CM on MC3T3‐E1
cells, alkaline phosphatase activity and gene expression. C, Concentration of alkaline phosphatase in control MC3T3‐E1 cells and cells treated
with LuMa CM. D, qRT‐PCR of mRNA expression of bone‐related genes (RUNX2, OPG, Osteocalcin, Osteopontin, and RANKL) in MC3T3‐E1
cells with and without treatment with LuMa CM. E‐F, Effect of bone (calvaria) on LuMa ALP activity and gene expression. E, Concentration of
alkaline phosphatase in Ace‐1 and LuMa cells (before and after incubation with calvaria). F, qRT‐PCR of expression of RUNX2, OPG, TGFβ,
Osteopontin and CaR mRNA in LuMa cells ± calvaria. Significant differences are indicated as *P ≤ .05, **P ≤ .01, ***P ≤ .001, and ****P ≤ .0001.
CM, condition media; OPG, osteoprotegrin; qRT‐PCR, quantitative reverse‐transcription polymerase chain reaction; TGFβ, transforming growth
factor β [Color figure can be viewed at wileyonlinelibrary.com]
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
710 | ELSHAFAE ET AL.
F IGURE 8 LuMa xenografts in the tibia
of nude mice. A, Radiography of intratibial
LuMa tumors with radiopaque
intramedullary osteoblastic tumors
(arrows). Histopathology of LuMa tibial
tumor (B‐D) showing intramedullary new
bone formation emanating from endosteal
and trabecular bone lined with
hypertrophic osteoblasts (200×) [Color
figure can be viewed at
wileyonlinelibrary.com]
we showed that LuMa cells formed osteoblastic bone metastasis dogs.41,42 DPC‐1 cells showed metastatic potential in vivo indicated
after IT or IC injection. by the presence of lung and iliac lymph node micrometastases in
Seven canine prostate cancer cell lines (Ace‐1, Probasco, Leo, immunosuppressed dogs after orthotopic injection. DPC‐1 cells in-
DPC‐1, CPA‐1, CHP‐1, and CT‐1258) have been described. The Ace‐1 duced mixed osteoblastic and osteolytic bone metastases similar to
cell line has been used extensively in prostatic cancer research. Ace‐1 Ace‐1 cells.42 The CPA‐1 cell line produces well‐differentiated
cells have been easily transfected with human or dog genes such as prostatic tumors with acinar pattern in athymic mice similar to the
PTHrP, GRPr, and DKK1 for studying different signaling pathways in primary tumor with no evidence of metastasis.43 It has been de-
prostate cancer. Orthotopic injection of Ace‐1 cells in dogs resulted monstrated that CT‐1258 cell line is tumorigenic in NOD‐SCID mice
in prostatic tumors with metastases to lungs and lymph nodes.39 following SQ or intraperitoneal implantation without any further
Ace‐1 cells grow in vitro, metastasize to long bones after intracardiac metastases.44
injection and induced mixed osteoblastic/osteolytic metastases after LuMa cells had increased osteomimicry properties compared to
intraosseous injection in nude mice.27 The Leo cell line has been Probasco, Ace‐1, and Leo cells as evidenced by the expression of
useful to investigate brain metastasis and osteolytic bone metastases bone‐associated markers (RUNX2, ALP, and OPN) and by their ability
in prostate cancer.34 The LuMa and Probasco33 cell lines induce to induce calcified bone matrix in a calvarial defect in vitro compared
prominent osteoblastic bone metastases in nude mice following IT to Ace‐1 cells. Upregulation of Runx2 expression in carcinomas was
injection. Probasco induces new woven bone that emanates from the found to be correlated with enhanced tumor progression, skeletal
periosteal surface in the form of a radio‐opaque starburst pattern metastasis45,46 and a poor prognosis.26 Some previous studies de-
radiographically and histologically, while the endosteal new bone was monstrated that knockin of RUNX2 was accompanied with an in-
characterized by sclerosis of bone marrow cavity with loss of bone crease in the expression of osteogenic genes including OPN, OC,
trabeculae.33 In contrast, LuMa cells mainly induced new woven bone RANKL, and MMP‐9, which promoted skeletal metastases of solid
formation from the endosteum and within the marrow cavity with tumors.47‐51 Some functional studies have shown that over-
disruption of the trabeculae. CHP‐1 is a useful cell line to study expression of OPN in PCa cell lines increased cancer cell invasion and
androgen receptor (AR) signaling in both AR‐dependent and ‐ enhanced their ability to enter the circulation in mouse models.52
independent PCa.40 CHP‐1 cells have not been extensively Based on these data, we propose that Runx2 and OPN signaling plays
characterized in terms of their metastasis sites and bone metastasis an important role in osteomimicry, tumorigenesis, and bone metas-
phenotype. The DPC‐1 cell line has been utilized for imaging of PCa tasis of LuMa cells.
in both mice and dogs by targeting PSMA. DPC‐1 cells are highly LuMa cells induced the differentiation of murine MC3T3‐E1 cells to
tumorigenic in nude mice and orthotopically in immunodeficient express osteoblast genes (OPN and OC) and proteins (ALP). ALP activity,
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ELSHAFAE ET AL. | 711
F IGURE 9 Bioluminescence and
histopathology of LuMa IC xenografts. A,
Bioluminescent images of nude mice after
7minutes (1), 15minutes (2), 18 days (3)
and 21 days (4, 5). B, Radiographic image
of intratibial LuMa tumors 21 days after
intracardiac injection showing a
radiopaque area in epiphysis and
metaphysis of the tibia (arrowhead). C‐H,
Histopathological images showing LuMa
metastases (arrowheads) in the tibia (C)
(100×), humerus with inset (D) (100×),
cervical vertebra (E) (200×), adrenal gland
(F) (200×), brain (G) (400×) and alveolar
bone of a tooth (H) (200×) [Color figure
can be viewed at wileyonlinelibrary.com]
TABLE 3 LuMa metastases in nude mice at 21 d following left production of bone sialoprotein, OC, and mineralization are parameters
53
ventricular injection that are typically used to identify osteoblastic differentiation.
LuMa cells may have inhibited osteoclastic bone resorption in
Mouse # Metastasis sites
the osteosclerotic bone metastases by increasing the production of
Mouse 1 Tibia, femur, vertebrae and adrenal gland
OPG and decreasing RANKL. In vitro, LuMa cells increased the
Mouse 2 Tibia, mandible and brain (undetected by OPG/RANKL ratio in MC3T3‐E1 cells suggesting that LuMa de-
bioluminescence)
creased osteoclastogenesis by its regulation of OPG/RANKL gene
Mouse 3 Tibia, femur, vertebrae, brain and adrenal gland expression with subsequent disruption of RANK/RANKL signaling in
Mouse 4 Null osteoclasts. The RANK/RANKL/OPG triad is the master regulator of
Mouse 5 Tibia, femur, vertebrae, brain and adrenal gland osteoclast function in vivo. Binding of RANKL produced from os-
teoblasts to its receptors (RANK) on osteoclasts promotes osteo-
Mouse 6 Tibia, femur, vertebrae, mandible and adrenal gland
clast differentiation and its resorptive activity.54‐57 OPG is a decoy
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
712 | ELSHAFAE ET AL.
receptor for RANKL and inhibits the ability of RANKL to induce ACKNOWLEDGMENTS
osteoclast function.55,58 The authors thank Dr. Prosper Boyaka, Department of Veterinary Bios-
LuMa cells will also be useful to investigate cancer stem cell (CSC) ciences, The Ohio State University, Columbus, Ohio for his help with
markers in prostate cancer. They expressed high levels of ALDH1 analyzing the flow cytometry data and Shady Estfanous, Department of
protein and CD133 and CD44 mRNA. It has been shown that CD44, Microbial Infection and Immunity, The Ohio State University, Columbus,
α2β1 and CD133‐expressing PCa cells exhibited more (3.7‐fold) self‐ Ohio for helping with the confocal microscope. The Ohio State College of
renewal capability than CD133 negative cells. PCa cells with high Veterinary Medicine Biospecimen Repository was supported by the fol-
expression of ALDH1 exhibited CSC properties and a positive corre- lowing grants: UL1TR001070 from the National Center for Advancing
lation with Gleason stage and poor survival in PCa patients.59,60 Translational Sciences and P30CA016058 from the National Cancer In-
Previous studies showed that CD44 has a role in cancer stem cell stitute to The Ohio State University. The research was supported by the
formation.61 The expression of CD44 receptor is correlated with Ministry of Higher Education and Scientific Research, Egypt.
RANKL expression in a CD44 knockout mouse model.24 One study
showed that RANKL and MMP9 expression were partially mediated by CONFLICT OF INTERESTS
CD44 and RUNX2‐dependent signaling.62 Consistent with these stu- The authors declare that there are no conflicts of interest.
dies, LuMa cells had high levels of CD44, RANKL, and RUNX2, thus
supporting the potential crosstalk between these factors. ORCID
MYOF has an important role in the promotion of tumor invasion Aylin Alasonyalilar‐Demirer http://orcid.org/0000-0001-6637-3254
and metastasis.63 Knockdown of MYOF in breast cancer markedly Thomas J. Rosol http://orcid.org/0000-0003-3737-1190
decreased the expression of MMP1 while its depletion suppressed
breast cancer cell motility and the phosphorylation of receptor tyr- REFERENCES
osine kinases, FGFR2, IGF‐IR, JAK2, TXK, and VEGFR2.64‐66 LuMa 1. Koeneman KS, Yeung F, Chung LW. Osteomimetic properties of
cells had higher mRNA expression of MYOF compared to normal prostate cancer cells: a hypothesis supporting the predilection
‐ of prostate cancer metastasis and growth in the bone environment.prostate, BPH and other canine PCa cell lines (Ace 1, Probasco, and
Prostate. 1999;39(4):246‐261.
Leo). It would be useful to investigate phosphorylation events of 2. Lecrone V, Li W, Devoll RE, Logothetis C, Farach‐Carson MC. Calcium
these pathways in LuMa cells. signals in prostate cancer cells: specific activation by bone‐matrix
LuMa cells express a low level of androgen receptors (AR) and proteins. Cell Calcium. 2000;27(1):35‐42.
3. Thomas R, True LD, Bassuk JA, Lange PH, Vessella RL. Differential
this may be a limitation for the model. LuMa cells were established
expression of osteonectin/SPARC during human prostate cancer
from a castrated dog similar to other canine PCa cell lines (Ace‐1, progression. Clin Cancer Res. 2000;6(3):1140‐1149.
Leo, Probasco, DPC‐1, and CHP‐1)41 and this may explain the lack 4. Towler DA. Angiogenesis and marrow stromal cell fates: roles in bone
of AR expression in LuMa and other canine PCa cell lines, since strength. Osteoporos Int. 2003;14(Suppl 5):S46‐S50. discussion S50‐43.
most canine prostate cancers are androgen‐independent. Although 5. Graham TR, Agrawal KC, Abdel‐Mageed AB. Independent and co-
operative roles of tumor necrosis factor‐alpha, nuclear factor‐kappaB,
most human prostate cancer patients respond to androgen depri- and bone morphogenetic protein‐2 in regulation of metastasis and
vation therapy initially, men with metastatic prostate cancer often osteomimicry of prostate cancer cells and differentiation and miner-
develop aggressive androgen‐independent prostate cancer (AIPC) alization of MC3T3‐E1 osteoblast‐like cells. Cancer Sci. 2010;101(1):
after therapy.67 AR signaling remains active in most AIPC patients, 103‐111.
68,69 6. Kavitha CV, Deep G, Gangar SC, Jain AK, Agarwal C, Agarwal R.which is in contrast to dogs with advanced PCa. Despite this
Silibinin inhibits prostate cancer cells‐ and RANKL‐induced osteo-
difference, canine PCa shares similar characteristics (clinical pre- clastogenesis by targeting NFATc1, NF‐kappaB, and AP‐1 activation
sentation, pathogenesis, and bone metastasis phenotype) with ad- in RAW264.7 cells. Mol Carcinog. 2014;53(3):169‐180.
vanced human androgen refractory prostate cancer making the dog 7. Chu C‐Y. The Role of RankL in Prostate Cancer Progression and Bone
Metastasis [PhD thesis]; Georgia State University. 2011.
cancer cell lines valuable models for studying androgen‐
8. Cox RF, Jenkinson A, Pohl K, O'Brien FJ, Morgan MP. Osteomimicry
independent PCa.69,70 of mammary adenocarcinoma cells in vitro; increased expression of
bone matrix proteins and proliferation within a 3D collagen en-
vironment. PLOS One. 2012;7(7):e41679.
5 | CONCLUSIONS 9. Glait‐Santar C, Benayahu D. Regulation of SVEP1 gene expression by
17beta‐estradiol and TNFalpha in pre‐osteoblastic and mammary
adenocarcinoma cells. J Steroid Biochem Mol Biol. 2012;130(1‐2):36‐44.
LuMa is a novel canine prostate carcinoma cell line with osteomi- 10. Hu P, Chu GCY, Zhu G, et al. Multiplexed quantum dot labeling of
micry and stemness properties that preferentially metastasizes to activated c‐Met signaling in castration‐resistant human prostate
cancer. PLOS One. 2011;6(12):e28670.
bone and induces osteoblastic bone metastasis. LuMa cells induced
11. Yuen HF, Kwok WK, Chan KK, et al. TWIST modulates prostate
differentiation of MC3T3‐E1 cells and stimulated bone formation in cancer cell‐mediated bone cell activity and is upregulated by osteo-
vitro. The LuMa cell line provides a clinically relevant and unique genic induction. Carcinogenesis. 2008;29(8):1509‐1518.
model for understanding the molecular mechanisms of osteomimicry, 12. Chu GCY, Zhau HE, Wang R, et al. RANK‐ and c‐Met‐mediated signal
osteoblastic bone metastasis, and stemness in the pathogenesis of network promotes prostate cancer metastatic colonization. Endocr
Relat Cancer. 2014;21(2):311‐326.
prostate cancer.
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ELSHAFAE ET AL. | 713
13. Brown LF, Papadopoulos‐Sergiou A, Berse B, et al. Osteopontin ex- 32. Yang J, Fizazi K, Peleg S, et al. Prostate cancer cells induce osteoblast
pression and distribution in human carcinomas. Am J Pathol. 1994; differentiation through a Cbfa1‐dependent pathway. Cancer Res.
145(3):610‐623. 2001;61(14):5652‐5659.
14. Carlinfante G, Vassiliou D, Svensson O, Wendel M, Heinegard D, 33. Simmons JK, Dirksen WP, Hildreth BE, et al. Canine prostate cancer
Andersson G. Differential expression of osteopontin and bone sialo- cell line (Probasco) produces osteoblastic metastases in vivo. Prostate.
protein in bone metastasis of breast and prostate carcinoma. Clin Exp 2014;74(13):1251‐1265.
Metastasis. 2003;20(5):437‐444. 34. Thudi NK, Shu ST, Martin CK, et al. Development of a brain metastatic
15. Yeung F, Law WK, Yeh CH, et al. Regulation of human osteocalcin canine prostate cancer cell line. Prostate. 2011;71(12):1251‐1263.
promoter in hormone‐independent human prostate cancer cells. J Biol 35. Bell FW, Klausner JS, Hayden DW, Feeney DA, Johnston SD. Clinical
Chem. 2002;277(4):2468‐2476. and pathologic features of prostatic adenocarcinoma in sexually intact
16. Ibrahim T, Leong I, Sanchez‐Sweatman O, et al. Expression of bone and castrated dogs: 31 cases (1970‐1987). J Am Vet Med Assoc. 1991;
sialoprotein and osteopontin in breast cancer bone metastases. Clin 199(11):1623‐1630.
Exp Metastasis. 2000;18(3):253‐260. 36. Rosol TJ, Tannehill‐Gregg SH, Corn S, Schneider A, McCauley LK. Animal
17. Waltregny D, Bellahcene A, de Leval X, Florkin B, Weidle U, models of bone metastasis. Cancer Treat Res. 2004;118:47‐81.
Castronovo V. Increased expression of bone sialoprotein in bone 37. Rosol TJ, Tannehill‐Gregg SH, LeRoy BE, Mandl S, Contag CH. Animal
metastases compared with visceral metastases in human breast and models of bone metastasis. Cancer. 2003;97(3 Suppl):748‐757.
prostate cancers. J Bone Miner Res. 2000;15(5):834‐843. 38. Zhang J, Dai J, Qi Y, et al. Osteoprotegerin inhibits prostate cancer‐
18. Waltregny D, Bellahcene A, Castronovo V, et al. Prognostic value of induced osteoclastogenesis and prevents prostate tumor growth in
bone sialoprotein expression in clinically localized human prostate the bone. J Clin Invest. 2001;107(10):1235‐1244.
cancer. J Natl Cancer Inst. 1998;90(13):1000‐1008. 39. Keller JM, Schade GR, Ives K, et al. A novel canine model for prostate
19. Li Y, Cozzi PJ. Targeting uPA/uPAR in prostate cancer. Cancer Treat cancer. Prostate. 2013;73(9):952‐959.
Rev. 2007;33(6):521‐527. 40. Azakami D, Nakahira R, Kato Y, et al. The canine prostate cancer cell
20. Komori T, Yagi H, Nomura S, et al. Targeted disruption of Cbfa1 re- line CHP‐1 shows over‐expression of the co‐chaperone small
sults in a complete lack of bone formation owing to maturational glutamine‐rich tetratricopeptide repeat‐containing protein alpha. Vet
arrest of osteoblasts. Cell. 1997;89(5):755‐764. Comp Oncol. 2017;15(2):557‐562.
21. Pratap J, Imbalzano KM, Underwood JM, et al. Ectopic runx2 ex- 41. Anidjar M, Villette JM, Devauchelle P, et al. In vivo model mimicking
pression in mammary epithelial cells disrupts formation of normal natural history of dog prostate cancer using DPC‐1, a new canine
acini structure: implications for breast cancer progression. Cancer Res. prostate carcinoma cell line. Prostate. 2001;46(1):2‐10.
2009;69(17):6807‐6814. 42. Anidjar M, Scarlata E, Cury FL, et al. Refining the orthotopic dog
22. Chua CW, Chiu YT, Yuen HF, et al. Suppression of androgen‐ prostate cancer (DPC)‐1 model to better bridge the gap between
independent prostate cancer cell aggressiveness by FTY720: vali- rodents and men. Prostate. 2012;72(7):752‐761.
dating Runx2 as a potential antimetastatic drug screening platform. 43. Eaton CL, Pierrepoint CG. Growth of a spontaneous canine prostatic
Clin Cancer Res. 2009;15(13):4322‐4335. adenocarcinoma in vivo and in vitro: isolation and characterization of
23. Zelzer E, Glotzer DJ, Hartmann C, et al. Tissue specific regulation of a neoplastic prostatic epithelial cell line, CPA 1. Prostate. 1988;12(2):
VEGF expression during bone development requires Cbfa1/Runx2. 129‐143.
Mech Dev. 2001;106(1‐2):97‐106. 44. Fork MA, Escobar HM, Soller JT, et al. Establishing an in vivo model of
24. Gupta A, Cao W, Chellaiah MA. Integrin alphavbeta3 and CD44 canine prostate carcinoma using the new cell line CT1258. BMC
pathways in metastatic prostate cancer cells support osteoclasto- Cancer. 2008;8(240):240.
genesis via a Runx2/Smad 5/receptor activator of NF‐kappaB ligand 45. Javed A, Barnes GL, Pratap J, et al. Impaired intranuclear trafficking of
signaling axis. Mol Cancer. 2012;11:66. Runx2 (AML3/CBFA1) transcription factors in breast cancer cells inhibits
25. Ge C, Zhao G, Li Y, et al. Role of Runx2 phosphorylation in prostate osteolysis in vivo. Proc Natl Acad Sci USA. 2005;102(5):1454‐1459.
cancer and association with metastatic disease. Oncogene. 2016;35(3): 46. Pratap J, Javed A, Languino LR, et al. The Runx2 osteogenic tran-
366‐376. scription factor regulates matrix metalloproteinase 9 in bone meta-
26. Akech J, Wixted JJ, Bedard K, et al. Runx2 association with pro- static cancer cells and controls cell invasion. Mol Cell Biol. 2005;
gression of prostate cancer in patients: mechanisms mediating bone 25(19):8581‐8591.
osteolysis and osteoblastic metastatic lesions. Oncogene. 2010;29(6): 47. John A, Tuszynski G. The role of matrix metalloproteinases in tumor
811‐821. angiogenesis and tumor metastasis. Pathol Oncol Res. 2001;7(1):14‐23.
27. LeRoy BE, Thudi NK, Nadella MVP, et al. New bone formation and 48. Purushothaman A, Chen L, Yang Y, Sanderson RD. Heparanase sti-
osteolysis by a metastatic, highly invasive canine prostate carcinoma mulation of protease expression implicates it as a master regulator of
xenograft. Prostate. 2006;66(11):1213‐1222. the aggressive tumor phenotype in myeloma. J Biol Chem. 2008;
28. Al Nakouzi N, Bawa O, Le Pape A, et al. The IGR‐CaP1 xenograft 283(47):32628‐32636.
model recapitulates mixed osteolytic/blastic bone lesions observed in 49. Roodman GD. Genes associate with abnormal bone cell activity in
metastatic prostate cancer. Neoplasia. 2012;14(5):376‐387. bone metastasis. Cancer Metastasis Rev. 2012;31(3‐4):569‐578.
29. Corey E, Quinn JE, Bladou F, et al. Establishment and characterization 50. Shevde LA, Das S, Clark DW, Samant RS. Osteopontin: an effector and
of osseous prostate cancer models: intra‐tibial injection of human an effect of tumor metastasis. Curr Mol Med. 2010;10(1):71‐81.
prostate cancer cells. Prostate. 2002;52(1):20‐33. 51. Weber GF. The metastasis gene osteopontin: a candidate target for
30. Raheem O, Kulidjian AA, Wu C, et al. A novel patient‐derived intra‐ cancer therapy. Biochim Biophys Acta. 2001;1552(2):61‐85.
femoral xenograft model of bone metastatic prostate cancer that 52. Bellahcene A, Castronovo V, Ogbureke KU, Fisher LW, Fedarko NS.
recapitulates mixed osteolytic and osteoblastic lesions. J Transl Med. Small integrin‐binding ligand N‐linked glycoproteins (SIBLINGs): mul-
2011;9:185. tifunctional proteins in cancer. Nat Rev Cancer. 2008;8(3):212‐226.
31. Thalmann GN, Anezinis PE, Chang SM, et al. Androgen‐independent 53. Barrere F, van Blitterswijk CA, de Groot K. Bone regeneration: mo-
cancer progression and bone metastasis in the LNCaP model of hu- lecular and cellular interactions with calcium phosphate ceramics. Int
man prostate cancer. Cancer Res. 1994;54(10):2577‐2581. J Nanomedicine. 2006;1(3):317‐332.
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
714 | ELSHAFAE ET AL.
54. Lacey DL, Timms E, Tan HL, et al. Osteoprotegerin ligand is a cytokine 65. Li R, Ackerman WE, Mihai C, et al. Myoferlin depletion in breast
that regulates osteoclast differentiation and activation. Cell. 1998; cancer cells promotes mesenchymal to epithelial shape change and
93(2):165‐176. stalls invasion. PLOS One. 2012;7(6):e39766.
55. Roodman GD. Mechanisms of bone metastasis. Discov Med. 2004; 66. Rowe RG, Weiss SJ. Breaching the basement membrane: who, when
4(22):144‐148. and how? Trends Cell Biol. 2008;18(11):560‐574.
56. Nakagawa N, Kinosaki M, Yamaguchi K, et al. RANK is the essential 67. Buchanan G, Ricciardelli C, Harris JM, et al. Control of androgen re-
signaling receptor for osteoclast differentiation factor in osteoclas- ceptor signaling in prostate cancer by the cochaperone small gluta-
togenesis. Biochem Biophys Res Commun. 1998;253(2):395‐400. mine rich tetratricopeptide repeat containing protein alpha. Cancer
57. Kong YY, Feige U, Sarosi I, et al. Activated T cells regulate bone loss Res. 2007;67(20):10087‐10096.
and joint destruction in adjuvant arthritis through osteoprotegerin 68. Imamura Y, Sadar MD. Androgen receptor targeted therapies in
ligand. Nature. 1999;402(6759):304‐309. castration‐resistant prostate cancer: bench to clinic. Int J Urol. 2016;
58. Simonet WS, Lacey DL, Dunstan CR, et al. Osteoprotegerin: a novel 23(8):654‐665.
secreted protein involved in the regulation of bone density. Cell. 1997; 69. Simmons JK, Elshafae SM, Keller ET, McCauley LK, Rosol TJ. Review
89(2):309‐319. of animal models of prostate cancer bone metastasis. Vet Sci. 2014;
59. Vassalli G. Aldehyde dehydrogenases: not just markers, but functional 1(1):16‐39.
regulators of stem cells. Stem Cells Int. 2019;3904645. 70. Leroy BE, Northrup N. Prostate cancer in dogs: comparative and
60. Li T, Su Y, Mei Y, et al. ALDH1A1 is a marker for malignant prostate clinical aspects. Vet J. 2009;180(2):149‐162.
stem cells and predictor of prostate cancer patients' outcome. Lab
Invest. 2010;90(2):234‐244.
SUPPORTING INFORMATION
61. Chanmee T, Ontong P, Kimata K, Itano N. Key roles of hyaluronan and
its CD44 receptor in the stemness and survival of cancer stem cells. Additional supporting information may be found online in the Sup-
Front Oncol. 2015;5:180. porting Information section.
62. Teplyuk NM, Galindo M, Teplyuk VI, et al. Runx2 regulates G protein‐
coupled signaling pathways to control growth of osteoblast progeni-
tors. J Biol Chem. 2008;283(41):27585‐27597.
63. Eisenberg MC, Kim Y, Li R, Ackerman WE, Kniss DA, Friedman A. How to cite this article: Elshafae SM, Dirksen WP,
Mechanistic modeling of the effects of myoferlin on tumor cell inva- Alasonyalilar‐Demirer A, et al. Canine prostatic cancer cell
sion. Proc Natl Acad Sci USA. 2011;108(50):20078‐20083.
line (LuMa) with osteoblastic bone metastasis. The Prostate.
64. Deftos LJ, Barken I, Burton DW, Hoffman RM, Geller J. Direct
evidence that PTHrP expression promotes prostate cancer pro- 2020;80:698–714. https://doi.org/10.1002/pros.23983
gression in bone. Biochem Biophys Res Commun. 2005;327(2):
468‐472.
 10970045, 2020, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pros.23983 by Bursa Uludag University, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License