THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 32, Issue of August 12, pp.

29011–29016, 2005
© 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

A Novel Pathophysiological Mechanism for Osteoporosis Suggested

by an in Vivo Gene Expression Study of Circulating Monocytes*□
S

Received for publication, February 1, 2005, and in revised form, May 3, 2005
Published, JBC Papers in Press, June 17, 2005, DOI 10.1074/jbc.M501164200

Yao-Zhong Liu‡§¶, Volodymyr Dvornyk‡¶储, Yan Lu‡, Hui Shen‡§, Joan M. Lappe‡,
Robert R. Recker‡, and Hong-Wen Deng‡§**‡‡§§
From the ‡Osteoporosis Research Center, Creighton University Medical Center and the §Department of Biomedical
Sciences, Creighton University, Omaha, Nebraska 68131, **The Key Laboratory of Biomedical Information and
Engineering, Ministry of Education, and Institute of Molecular Genetics, School of Life Science and Technology, Xi’an
Jiaotong University, Xi’an 710049, China, the ‡‡Laboratory of Molecular and Statistical Genetics, College of Life Sciences,
Hunan Normal University, Changsha, Hunan 410081, China, and the 储Department of Biological Sciences,
Kent State University, Kent, Ohio 44242

Bone mineral density (BMD) is a major risk factor for Osteoporosis is most frequently characterized by low bone
osteoporosis. Circulating monocytes may serve as early mineral density (BMD).1 The well accepted pathophysiological
progenitors of osteoclasts and produce a wide variety of mechanisms for low bone mass include prolongation of the life
factors important to bone metabolism. However, little is span of osteoclasts (1) and early apoptosis of osteoblasts (2) and

Downloaded from http://www.jbc.org/ by guest on November 22, 2018

known about the roles of circulating monocytes in rela-
tion to the pathophysiology of osteoporosis. Using the
Affymetrix HG-U133A GeneChip® array, we performed a
comparative gene expression study of circulating mono-
cytes in subjects with high and low BMD. We identified
in total 66 differentially expressed genes including some
novel as well as some already known to be relevant to
bone metabolism. Three genes potentially contributing
to bone metabolism, CCR3 (chemokine receptor 3), HDC
(histidine decarboxylase, i.e. the histamine synthesis en-
zyme), and GCR (glucocorticoid receptor), were con-
firmed by quantitative real-time reverse transcriptase-
PCR as up-regulated in subjects with lower BMD. In
addition, significant negative correlation was ob-
served between expression levels of the genes and
BMD Z-scores. These three genes and/or their products
mediate monocyte chemotaxis, histamine production,
and/or sensitivity to glucocorticoids. Our results sug-
gest a novel pathophysiological mechanism for osteo-
porosis that is characterized by increased recruitment
of circulating monocyte into bone, enhanced monocyte
differentiation into osteoclasts, as well as osteoclast
stimulation via monocyte functional changes. This is
the first in vivo microarray study of osteoporosis in
humans. The results may contribute to identification
of new genes and their functions for osteoporosis and
suggest genetic markers to discern individuals at
higher risk to osteoporosis with an aim for preventive
intervention and treatment.
* This work was supported by grants from National Institutes of
Health (Grants K01 AR02170-01, R01 AR45349-01, and R01 GM60402-
01A1) and an LB595 grant from the State of Nebraska. The study also
benefited from grants from National Science Foundation of China, Huo
Ying Dong Education Foundation, HuNan Normal University, Xi’an
Jiaotong University, and the Ministry of Education of China. The costs
of publication of this article were defrayed in part by the payment of
page charges. This article must therefore be hereby marked “advertise-
ment” in accordance with 18 U.S.C. Section 1734 solely to indicate this
fact.
□S The on-line version of this article (available at http://www.jbc.org)
contains supplemental exclusion criteria and a supplemental table.
¶ Both authors contributed equally to the work.
§§ To whom correspondence should be addressed: Osteoporosis Re-
search Center, Creighton University Medical Center, 601 N. 30th St.,
Ste. 6787, Omaha, NE 68131. Tel: 402-280-5911; Fax: 402-280-5034;
E-mail: deng@creighton.edu.
osteocytes (3, 4). Recently, the imbalance between osteoblasto-
genesis and adipogenesis of bone marrow mesenchymal stem
cells was suggested as a new mechanism for osteoporosis (5).
However, the above mechanisms were discovered mainly
through in vitro studies or in vivo studies using animals. Their
comprehensiveness and exhaustiveness is unclear, and their
direct relevance to osteoporosis in humans is yet to
be established.
Circulating monocytes may serve as progenitors of oste-
oclasts (6 –9). Monocytes isolated from peripheral blood can
differentiate in vitro into the multinuclear cells that express
typical osteoclast markers and are able to resorb bone (6, 8). All
osteoclasts in peripheral skeleton (7, 10) and a considerable
amount of osteoclasts in the central skeleton (11) come from
circulating monocytes. Blood monocytes also produce a wide
variety of factors involved in bone metabolism, such as inter-
leukin-1, tumor necrosis factor-␣, interleukin-6, platelet-de-
rived growth factor, transforming growth factor-␤, and
1,25(OH)2D3 (12). Circulating mononuclear cells of osteoporotic
patients, when induced in vitro into osteoclasts, demonstrated
increased levels of bone resorption activity as compared with
normal controls (13).
For both peripheral and central skeleton, the access route of
circulating monocytes into the bone microenvironment is
through the structure called sinusoid (14) that occupies the
center of a basic multicellular unit, the functional unit of bone
remodeling. Recruitment of circulating monocytes into bone is
modulated mainly by chemotaxis (15), whereby monocyte che-
moattractants from bone and their receptors on monocytes,
such as chemokine receptors, interact to initiate and maintain
a directed migration of monocytes toward bone tissue.
Given the importance of the circulating monocytes in bone
metabolism, currently, surprisingly little is known about the in
vivo roles of circulating monocytes in the pathophysiology of
osteoporosis. Microarrays are a powerful tool to provide a com-
prehensive picture of cell function as they can assay expression
of tens of thousands of genes simultaneously; however, they
have not been attempted for in vivo studies in humans to
investigate the molecular mechanisms underlying osteoporo-
1
The abbreviations used are: BMD, bone mineral density; DEG, differ-
entially expressed genes; FDR, false discovery rate; CCR3, chemokine
receptor 3; HDC, histidine decarboxylase; GCR, glucocorticoid receptor.

This paper is available on line at http://www.jbc.org 29011

29012 Microarray Study of Osteoporosis
TABLE I
Characteristics of study subjects
ZBMD represents the BMD Z score.
Premenopausal Postmenopausal
n Age (S.D.) L1–4 ZBMD (S.D.) n Age (S.D.) L1–4 ZBMD (S.D.)
High BMD 5 49.4 (3.21) 2.71 (0.901) 5 52.6 (1.82) 2.22 (0.928)
Low BMD 4 50.5 (2.89) ⫺1.11 (0.429) 5 51.4 (1.52) ⫺1.69 (0.159)

sis. Here, with an in vivo approach directly in humans, we

explore other potential mechanisms for osteoporosis, which are
mediated by or reflected in circulating monocytes.
We hypothesized that gene expression profiles of circulating
monocytes differ between high and low BMD women. These
differentially expressed genes (DEGs) are potentially impor-
tant for the pathogenesis of osteoporosis. To identify these
genes, we applied the Affymetrix HG-U133A GeneChip® array
to analyze the gene expression profiles of circulating monocytes
in subjects with extremely discordant BMD values.
MATERIALS AND METHODS
Subjects

Downloaded from http://www.jbc.org/ by guest on November 22, 2018

The study was approved by the Creighton University Institutional
Review Board. All the study subjects signed informed consent docu-
ments before entering the project. We established strict criteria to
exclude non-genetic factors that might cause BMD variation. Additional
criteria were implemented to exclude diseases/conditions that may po-
tentially lead to gene expression changes of circulating monocytes. The
detailed criteria are included in the Supplemental Data.
The recruited women were all Caucasians and aged between 47 and
55 with an average age of ⬃51. The narrow age window may minimize
the potential age effects on monocyte gene expression (16, 17). Among
the 19 subjects, 10 had high spine BMD, with an average L1–L4 Z-score
of ⫹2.47, and 9 had low spine BMD, with an average L1–L4 Z-score of
⫺1.43. The Z-score is defined as the number of standard deviations of a
BMD measurement above (i.e. a positive Z-score) or below (i.e. a nega-
tive Z-score) the age-, gender-, and ethnicity-matched mean BMD. The
average Z-score for the high BMD subjects places them in the top 0.68%
of the normal population distribution, whereas the Z-score for the low
BMD subjects places them in the bottom 7.6% of the normal population
distribution. Among the 10 high BMD subjects, five were premeno-
pausal, and five were postmenopausal. Among the nine low BMD sub-
jects, four were premenopausal, and five were postmenopausal. In our
study, menopause was defined as cessation of menstruation for at least
one year. The recruitment scheme for menopausal status ensures that
the DEGs found between the 10 high and 9 low BMD subjects are only
related to BMD. Detailed characteristics of the subjects are available in
Table I.
We used a method recommended by Affymetrix to evaluate our
sample size (e.g. the number of replicates per condition) (18). According
to the method, coefficient of variation (CV) is calculated for certain
percentiles (e.g. the 25th, 50th, and 75th) of signal intensities for a data
set containing 1, 2, 3, . . ., and n replicates. Comparison is made
between the calculated CVs of certain percentiles for different numbers
of replicates. The optimal sample size is achieved when the value of CV
is stabilized, that is, does not change appreciably from n replicates to n
⫹ 1 replicates. Under such a situation, it is unlikely that additional
replicates will improve significantly the accuracy of the estimation for
the standard deviations of the sample, which is ultimately used to
determine statistical significance in parametric statistical tests. FIG. 1. Comparison of CVs of microarray signal intensities for
Based on the calculations with the raw microarray data in our study, different numbers of replicates.
we observed that the CV became stabilized at seven replicates for the
25th and the 50th percentile of signal intensities and at nine replicates
for the 75th percentile of signal intensities (Fig. 1). Therefore, our Experimental Procedures
sample size is sufficient and decent to provide adequate statistical Monocyte Isolation—Monocyte isolation from 30 ml of whole blood
power for detecting differential gene expression. was performed with a monocyte-negative isolation kit (Dynal Biotech
Inc., Lake Success, NY) following manufacturer’s recommendation. The
BMD Measurement kit contains a highly optimized antibody mix, blocking reagent, and
BMD (g/cm2) for the lumbar spine (L1–L4) was measured with a Depletion Dynabeads® to deplete T cells, B cells, and natural killer cells
Hologic 4500 dual energy x-ray absorptiometer scanner (Hologic Corp., from mononuclear cells, leaving monocytes untouched and free of the
Waltham, MA). The machine was calibrated daily. The coefficient of surface-bound antibody and beads. Our flow cytometry results showed
variation of the dual energy x-ray absorptiometer measurements for an average monocyte purity of 86% ⫾ 3% in the samples.
BMD was 0.9%. None of our subjects had any known conditions that Total RNA Extraction—Total RNA from monocytes was extracted
might artificially increase BMD values, such as osteophytes and using Qiagen RNeasy Mini kit (Qiagen, Inc., Valencia, CA).
facet sclerosis. Preparation of cRNA and Gene Chip Hybridization—Preparation of
 Microarray Study of Osteoporosis
cRNA, hybridization, and scanning of the HG-U133A GeneChip® was
performed according to the manufacturer’s protocol (Affymetrix, Santa
Clara, CA).
Real-time RT-PCR—Real-time RT-PCR was performed in two steps,
i.e. the first step of reverse transcription for synthesis of cDNA from
total RNA and the secondary step of real-time quantitative PCR.
Reverse transcription reactions were performed in a 30-␮l reaction
volume, containing 3 ␮l of 10⫻ PCR Buffer II (100 mM Tris-HCl, pH 8.3,
500 mM KCl, 0.01% w/v gelatin), 6 ␮l of 2.5 mM MgCl2, 8 ␮l of dNTPs,
1 ␮l of murine leukemia virus reverse transcriptase, 1 ␮l of RNase
inhibitor, 1 ␮l of oligo(dT), 0.5 ␮g of total RNA, and water to 10 ␮l. All
of the above reagents were supplied by Applied Biosystems (Foster City,
CA). Reaction conditions were as follows, 10 min at 20 °C, 60 min at
42 °C, 5 min at 95 °C.
Real-time quantitative PCR was performed in a 50-␮l reaction vol-
ume using standard protocols on an Applied Biosystems 7000 sequence
detection system. Briefly, 5 ␮l of cDNA was mixed with 25 ␮l of TaqMan
universal PCR master mix (2⫻), 2.5 ␮l of Assays-on-DemandTM gene
expression assay mix (contains forward and reverse primers and la-
beled probe), 2.5 ␮l of human GAPDH (20⫻), and 15 ␮l of water. The
thermocycling conditions were as follows: 2 min at 50 °C, 10 min at
95 °C, 50 cycles of 15 s at 95 °C plus 1 min at 60 °C. The thermal
denaturation protocol was run at the end of the PCR to determine the
copy number of products that were present in the reaction. All reactions
were run in triplicates and included “no template” controls for
each gene.
The TaqMan gene expression assays all have amplification efficien-
cies very close to 1 (19). Thus, it was not necessary for us to perform
validation experiments to test the equality of the amplification efficien-
cies between the target genes and the reference gene (GAPDH).
Data Analyses
We used a method by Golub et al. (20) to preselect a subset of genes
from the total of around 14,500 genes in the array for further analyses.
The preselection process minimizes the multiple-testing problem by
excluding from further analysis the genes with relatively constant
expression across the samples. It was based on the expression values
generated with the MAS 5.0 software (Affymetrix). We selected a subset
of 8,623 probe sets from a total of 22,283 for around 14,500 genes.
Subsequent data analyses were performed on these selected probe sets
using Bioconductor packages (21). The Bioconductor Affy package (22)
was used to process the probe-level data (raw data) and transform the
data into expression data using the robust multiarray average algo-
rithm (23).
Based on the expression data generated with the robust multiarray
average algorithm, the Bioconductor Multtest package was used to
identify DEGs between high versus low BMD subjects. To account for
multiple testing, the Benjamini and Hochberg stepwise procedure (24),
which controls for the false discovery rate, was implemented to gener-
ate adjusted p values.
RESULTS
DEGs Suggested by Microarray Analyses—All the raw mi-
croarray data are available.2 We also submitted the data to the
NCBI Gene Expression Omnibus data repository under acces-
sion number GSE2208. Based on the analysis with the MAS 5.0
software (Affymetrix), on average, 39.69% of a total of 22,283
probe sets in the array were called “present” for our samples.
Of the preselected 8,623 probe sets, we identified 67 (corre-
sponding to 66 genes) differentially expressed between the high
and low BMD subjects (Benjamini and Hochberg stepwise pro-
cedure-adjusted p ⬍ 0.05). These genes are listed in the order
of the adjusted p values in Table 1 of the Supplemental Data.
Forty-six genes are up-regulated and 20 are down-regulated in
the low BMD subjects as compared with the high BMD ones.
Since the majority of the genes listed in the table have not been
confirmed with real-time RT-PCR, the significance of their
differential expression (i.e. the p values) should be interpreted
with caution.
DEGs Validated through Real-time RT-PCR—The number of
monocytes we obtained from the 30 ml of blood of each subject
2
Y.-Z. Liu, V. Dvornyk, Y. Lu, H. Shen, J. M. Lappe, R. R. Recker,
and H.-W. Deng, personal communication.
29013
was limited, and the total RNA was scarce. Therefore, our
criteria for selection of genes for further real-time RT-PCR
validation were stringent and comprehensive. They were based
not only on the statistical significance of the differential ex-
pression of a gene as detected by microarray (i.e. the Benjamini
and Hochberg stepwise procedure-adjusted p ⬍ 0.05) but also
on its potential functional relevance to bone metabolism. Ac-
cordingly, we selected six genes for real-time RT-PCR valida-
tion, which are CCR3 (chemokine (C-C motif) receptor 3), HDC
(histidine decarboxylase), GCR (glucocorticoid receptor),
RAMP3 (receptor (calcitonin) activity-modifying protein 3),
SCYE1 (small inducible cytokine subfamily E, member 1), and
FCER1A (Fc fragment of IgE, high affinity I, receptor for
␣-polypeptide). These genes and/or their products, through
modulation of monocytes, are known to participate in bone
monocyte recruitment (CCR3, HDC) (15, 25), osteoclastogen-
esis (GCR, HDC, CCR3) (26 –31), osteoclast regulation (HDC,
RAMP3) (32, 33), and/or production of proinflammatory cyto-
kines important to osteoporosis (SCYE1, FCER1A) (34, 35).
Based on the relative gene expression (i.e. 2⫺⌬CT, where
⌬CT ⫽ (CT Target Gene⫺CT GAPDH)) for each sample obtained
Downloaded from http://www.jbc.org/ by guest on November 22, 2018
through real-time RT-PCR, the Student’s t test was employed
to compare expression for each gene between the high and low
BMD subjects. As our subjects were also stratified by meno-
pausal status, comparison of gene expression between high and
low BMD subjects was also performed among premenopausal
as well as postmenopausal women. Since the major aim of this
study was focused on identifying genes related to BMD varia-
tion, factorial analysis for identifying differential gene expres-
sion due to menopause and/or interaction between BMD and
menopause was not performed here and was pursued else-
where (36). The t tests confirmed the differential expression for
three genes. CCR3 and HDC were up-regulated, by 3- (p ⫽
0.0002) and 4-fold (p ⫽ 0.022), respectively, in the low versus
the high BMD subjects, and GCR was 2-fold up-regulated in
the premenopausal low versus high BMD subjects (p ⫽ 0.027).
We also observed significant negative correlation between
the expression levels of the three genes as quantified by real-
time PCR and BMD Z-scores. For the HDC, CCR3, and GCR
genes, the correlation coefficients were ⫺0.458 (p ⫽ 0.048),
⫺0.687 (p ⫽ 0.001), and ⫺0.816 (p ⫽ 0.007), respectively.
Although there was significant difference in expression be-
tween the high and the low BMD groups for the three genes,
there is no clear threshold level for their expression, which
might be associated with the low BMD subjects. Therefore, the
expression levels may represent continuous variables, which
negatively correlate with BMD.
The reason for only three out of the six candidate genes being
confirmed by real-time RT-PCR might be related to the abso-
lute expression levels of the genes. Genes with moderate ex-
pression levels tend to have higher correlation between expres-
sion quantification results obtained by microarray and RT-PCR
(37). Indeed, in our study, among the six selected genes, the
three confirmed ones had expression levels within the middle
range (i.e. from 800 to 1,300 according to microarray quantifi-
cation using the MAS 5.0 software), whereas the other three
had more extreme expression levels (i.e. less than 300 or more
than 3,000). Statistically, this may also be due to problems
associated with multiple testing and statistical power for rep-
lication (see detailed elaboration under “Discussion”).
We chose GAPDH as the endogenous control gene for the
real-time RT-PCR because its expression levels in microarrays
were very consistent and stable across all the samples. In the
HG-U133A array, there are three independent probe sets,
among a total of 22,283, for the GAPDH gene. The three probe
sets for the GAPDH gene were ranked by their CVs of expres-
29014 Microarray Study of Osteoporosis

FIG. 2. Potential effects on bone metabolism caused by the up-regulation of the CCR3, HDC, and GCR gene in circulating monocytes.
Mono indicates monocytes; MNC indicates mononuclear cells; the drawing at the upper right indicates osteoclasts; the drawing at the lower right
indicates the kidney; the plus sign indicates stimulate; the large boldface arrow indicates collective results from the functional changes of monocytes,
i.e. more monocytes enter into the bone tissue, more monocytes differentiates into osteoclasts, and osteoclasts are up-regulated in both function and
number; the up arrow indicates an increase in expression or production; and the down arrow indicates a decrease in production. IL, interleukin;

Downloaded from http://www.jbc.org/ by guest on November 22, 2018

GM-CSF, granulocyte-macrophage colony-stimulating factor; TNF, tumor necrosis factor.

sion at the upper 0.2th, 0.3th, and 1.7th percentile, respec-
tively, higher than all other well known housekeeping genes,
including the actin and the 18S gene. Therefore, the GAPDH
gene has the most stable expression as compared with other
well known endogenous control genes and is also ranked among
the most consistently expressed genes on the array.
DISCUSSION
It has already been suggested that functional changes of
osteoclasts (1), osteoblasts (2), osteocytes (3, 4), and mesenchy-
mal stem cells (5) underlie the pathogenesis of osteoporosis.
Although circulating monocytes are an important cell type for
bone metabolism (6 –13), their role in the pathophysiology of
osteoporosis has not been systematically explored in vivo. In
this study, we scanned expression of ⬃14,500 genes of circu-
lating monocytes from human subjects using the microarray
approach and identified, among the others, three important
genes related to BMD variation. The potential functions of the
three genes allowed us to propose a model (Fig. 2) suggesting a
potential and novel monocyte-mediated pathophysiological
mechanism for osteoporosis.
Chemotaxis is a major mechanism whereby circulating
monocytes migrate into the bone microenvironment (15). Given
that CCR3 ligand, RANTES (Chemokine, CC motif, ligand 5),
was found in both osteoblasts and osteoclasts (38 – 40), CCR3
may play an important role in the recruitment of monocytes
into bone. In addition, in dendritic cells that are very similar to
monocytes in terms of phenotype and origin, the CCR3 gene
mediates in vitro chemotactic response (41), a process essen-
tially the same as that by which monocytes migrate into the
bone microenvironment. The CCR3 gene was also found to be
up-regulated during the receptor activator of NF-␬B ligand-
induced osteoclastogenesis (28), which suggests its potential
role in promoting monocyte differentiation into osteoclasts.
Thus, the up-regulation of CCR3 in circulating monocytes of
low BMD women as discovered in the present study may facil-
itate both the access of monocytes to the bone microenviron-
ment and their subsequent differentiation into osteoclasts.
These two concurrent processes may lead to increased bone
resorption.
HDC is the only enzyme for synthesis of histamine in the
human body (42). With the HDC gene up-regulated, monocyte
histamine production may increase in subjects with lower
BMD. A high level of circulating histamine was associated with
osteoporosis and abnormal bone histomorphometric indices of
resorption (43– 45). Histamine receptor antagonists were found
to have treatment effects on osteoporosis (46, 47). HDC knock-
out mice (suffering from histamine deficiency) demonstrated
increased bone formation and decreased bone loss induced by
ovariectomy (42). Histamine may also have effects on the kid-
ney by decreasing synthesis of calcitriol and leading to conse-
quent secondary hyperparathyroidism (42). Locally, histamine
can stimulate osteoclast activity (33) and increase osteoclast
number (33), production of granulocyte-macrophage colony-
stimulating factor and interleukin-6 by mononuclear cells (29,
30), and expression of tumor necrosis factor-␣ by monocytes
(31), which are important cytokines promoting osteoclastogen-
esis (48, 49). Histamine may also stimulate monocytes to pro-
duce MCP-1 (monocyte chemoattractant protein-1) and CCR2
(chemokine receptor-2) (25), two important factors mediating
chemotaxis of monocytes (50 –52), thereby increasing monocyte
recruitment into bone.
Higher expression of the GCR gene in subjects of lower BMD
may increase sensitivity of their monocytes to glucocorticoids.
Glucocorticoid treatment is known to promote secondary osteo-
porosis (53). Glucocorticoids may serve to prime monocytes/
macrophages toward differentiation into osteoclasts (26, 27). A
polymorphism of the GCR gene has been associated with a
higher sensitivity to glucocorticoids and lower BMD (54).
As illustrated in Fig. 2, the three genes and/or their products
may have similar functions in bone remodeling. For example,
both CCR3 and HDC may increase bone monocyte recruitment,
whereas both HDC and GCR may promote osteoclastogenesis.
In such a way, the genes may act synergistically, thereby
resulting in a higher rate of monocyte recruitment into bone, an
increased number of monocytes committed to differentiation
into osteoclasts, and, subsequently, lower BMD.
It has already been suggested that there is no increase in the
number of circulating osteoclast precursors in osteoporotic ver-
sus normal subjects (13). Thus, an increased accessibility of
these precursors (i.e. monocytes) to the bone tissue, as sug-
gested in our study, may represent an important and novel
mechanism for osteoporosis. This hypothesis is supported by
studies on chemokines and chemokine receptors, the factors
responsible for monocyte recruitment, which have shown their
 Microarray Study of Osteoporosis
important roles in BMD regulation and bone remodeling
(52, 55).
Our study for the first time suggested monocyte production
of histamine as a pivotal factor underlying osteoporosis. As
shown in Fig. 2, histamine affects bone metabolism at multiple
levels, locally in osteoclasts (33), mononuclear cells (29, 30),
and monocytes (25, 31) and systemically in the kidney (42).
Although the primary cell type for histamine production is
mast cells, they are normally not found in the circulation and
thus are probably clinically unimportant for osteoporosis, ex-
cept in hyperplastic conditions, such as mastocytosis (43– 45).
In monocytes, the expression of the HDC gene and production
of histamine has already been observed in several studies (56 –
58). Histamine produced by monocytes was also implicated in
the pathogenesis of chronic inflammatory diseases, such as
atherosclerosis (59). Recently, in a study showing the effects of
histamine on trabecular bone loss in ovariectomized rats, the
phagocyte/monocyte lineage was suggested as a cellular source
of histamine causing the bone loss in the animals (46). Our
findings, together with the above, suggested histamine produc-
tion from circulating monocytes as an important factor modu-
lating bone loss and proposed histamine receptor antagonists
as a potential drug for treatment of osteoporosis.
According to our real-time RT-PCR results, there was no
significant difference in the GCR gene expression between the
low and the high BMD subjects in the postmenopausal females,
although such difference was detected in the premenopausal
females. This may suggest an interaction between menopause
and BMD for the GCR gene expression. Indeed, by fitting the
real-time RT-PCR data of the gene into a general linear model,
a significant interaction (p ⫽ 0.016) between BMD and meno-
pause was detected (data not shown). Consistent with this, the
GCR gene was down-regulated due to menopause (p ⫽ 0.03)
only in low BMD females.
In our study, only three out of six selected genes suggested by
microarrays were confirmed with real-time RT-PCR. Although
this may reflect a high false-positive discovery rate associated
with microarray technology, statistically it may also be attrib-
uted to the multiple-testing procedure in the process of mi-
croarray gene identification. Generally, due to multiple-testing
procedures in a microarray experiment, the power required to
suggest a gene differentially expressed by microarrays is con-
siderably higher than that required to confirm the differential
expression of that gene in a subsequent experiment. If, for
example, for a list of 10 genes, the power to detect each one of
them individually is 10%, then the power required to suggest,
in a microarray experiment, any one of them (i.e. at least one of
them) is increased to about 65% (i.e. 1-(1–10%)10). However, for
such a gene suggested by microarrays, the power decreases to
essentially 10% again when it comes to confirm it in a subse-
quent focused replication experiment, either in an independent
sample, or as in our study, in the same sample but on a
different platform (e.g. RT-PCR, Northern blotting, etc.). This
drastic drop of power makes the confirmation of the differential
expression of a gene suggested by microarrays difficult and
may contribute to a moderate confirmation rate in our study.
Although our data suggested a significant correlation be-
tween the expression levels of the three genes and BMD Z-
scores, the difference in expression levels of monocyte HDC,
GCR, and CCR3 gene among women, whose BMD Z-scores
span from normal to severely osteoporotic, might be moderate.
In normal conditions, circulating monocytes are mature cells of
finite differentiation stage, functionally stable and homogene-
ous in phenotype until entering peripheral tissues (60). There-
fore, we expected that, unless pathologic conditions occur, such
as immune diseases and inflammation, their expression change
29015
should be generally small in magnitude. Although in our study
the low BMD subjects were not severely osteoporotic, most of
the high BMD subjects had very high Z-scores, which made the
difference in Z-score between the two groups (high versus low
BMD) as large as 3.90. Even with the two groups so extremely
segregated in terms of BMD, the gene expression fold change
for the three genes between the two groups was only from 2
to 4.
Minor functional differences of monocytes, as detected by
microarrays, might magnify exponentially in the bone micro-
environment, where many specific bone-related factors may
interact or synergize. For example, in the bone tissue, an in-
crease in monocyte histamine production may locally induce
MCP-1 and CCR2 production (25), leading to monocyte accu-
mulation. As more monocytes accumulate, local histamine lev-
els may increase, which in turn, may induce production of more
MCP-1 and CCR2. With such cycles, a tiny increase in mono-
cyte histamine production may be enhanced locally in the bone
to physiologically significant levels necessary to initiate a
higher rate of monocyte recruitment and osteoclastogenesis
and, consequently, bone resorption as illustrated in Fig. 2.
The present study reports data about differential expression
Downloaded from http://www.jbc.org/ by guest on November 22, 2018
of genes at the mRNA level only. In future functional genomics
studies of circulating monocytes, the results may be validated
also at the protein level. Another interesting direction is to
genetically manipulate the expression of the three genes in
freshly isolated monocytes using such approaches as RNA in-
terference and viral overexpression and look at the ability of
the cells to attach to bone, undergo osteoclastogenesis, or
resorb bone ex vivo. This may further test the functions of the
genes as suggested by the present study.
The present study for the first time provides evidence for the
functional difference of circulating monocytes between low and
high BMD women. The model proposed here (Fig. 2) suggests
how the three genes, CCR3, HDC, and GCR, may potentially
contribute to increased bone resorption, making them a new
group of functionally related skeletal genes in monocytic cell
lineages. To the best of our knowledge, this is the first in vivo
microarray study in humans to suggest another potential
pathophysiological mechanism underlying osteoporosis.
Acknowledgments— We thank JoAnn Wilde and Theresa Conway for
subject recruitment, Bill Snyder for assistance in real-time RT-PCR
experiments, and Peng Xiao for help in preparing the manuscript.
REFERENCES
1. Boyce, B. F., Hughes, D. E., Wright, K. R., Xing, L., and Dai, A. (1999) Lab.
Investig. 79, 83–94
2. Manolagas, S. C., Kousteni, S., and Jilka, R. L. (2002) Recent Prog. Horm. Res.
57, 385– 409
3. Tomkinson, A., Gevers, E. F., Wit, J. M., Reeve, J., and Noble, B. S. (1998)
J. Bone Miner. Res. 13, 1243–1250
4. Tomkinson, A., Reeve, J., Shaw, R. W., and Noble, B. S. (1997) J. Clin.
Endocrinol. Metab. 82, 3128 –3135
5. Nuttall, M. E., and Gimble, J. M. (2004) Curr. Opin. Pharmacol. 4, 290 –294
6. Udagawa, N., Takahashi, N., Akatsu, T., Tanaka, H., Sasaki, T., Nishihara, T.,
Koga, T., Martin, T. J., and Suda, T. (1990) Proc. Natl. Acad. Sci. U. S. A.
87, 7260 –7264
7. Zambonin, Z. A., Teti, A., and Primavera, M. V. (1984) J. Cell Sci. 66, 335–342
8. Quinn, J. M., Neale, S., Fujikawa, Y., McGee, J. O., and Athanasou, N. A.
(1998) Calcif. Tissue Int. 62, 527–531
9. Fujikawa, Y., Quinn, J. M., Sabokbar, A., McGee, J. O., and Athanasou, N. A.
(1996) Endocrinology 137, 4058 – 4060
10. Horton, M. A., Spragg, J. H., Bodary, S. C., and Helfrich, M. H. (1994) Bone
(N. Y.) 15, 639 – 646
11. Parfitt, A. M. (1994) J. Cell. Biochem. 55, 273–286
12. Nathan, C. F. (1987) J. Clin. Investig. 79, 319 –326
13. Jevon, M., Hirayama, T., Brown, M. A., Wass, J. A., Sabokbar, A., Ostelere, S.,
and Athenasou, N. A. (2003) Scand. J. Rheumatol. 32, 95–100
14. Parfitt, A. M. (1998) Bone (N. Y.) 23, 491– 494
15. Graves, D. T., Jiang, Y., and Valente, A. J. (1999) Front Biosci. 4, D571-D580
16. Perkins, S. L., Gibbons, R., Kling, S., and Kahn, A. J. (1994) Bone (N. Y.) 15,
65–72
17. Cohen-Solal, M. E., Boitte, F., Bernard-Poenaru, O., Denne, M. A., Graulet,
A. M., Brazier, M., and De Vernejoul, M. C. (1998) J. Clin. Endocrinol.
Metab. 83, 1687–1690
29016 Microarray Study of Osteoporosis
18. Affymetrix, Inc. (2002) GeneChip Expression Analysis Technical Manual, Af-
fymetrix, Inc., Santa Clara, CA
19. Applied Biosystems (2004) Applied Biosystems Application Note, Amplification
Efficiency of TaqMan® Gene Expression Assays, Application Note
127AP05– 02, Applied Biosystems, Foster City, CA
20. Golub, T. R., Slonim, D. K., Tamayo, P., Huard, C., Gaasenbeek, M., Mesirov,
J. P., Coller, H., Loh, M. L., Downing, J. R., Caligiuri, M. A., Bloomfield,
C. D., and Lander, E. S. (1999) Science 286, 531–537
21. Gentleman, R. C., Carey, V. J., Bates, D. M., Bolstad, B., Dettling, M., Dudoit,
S., Ellis, B., Gautier, L., Ge, Y., Gentry, J., Hornik, K., Hothorn, T., Huber,
W., Iacus, S., Irizarry, R., Leisch, F., Li, C., Maechler, M., Rossini, A. J.,
Sawitzki, G., Smith, C., Smyth, G., Tierney, L., Yang, J. Y., and Zhang, J.
(2004) Genome Biol. 5, R80
22. Gautier, L., Cope, L., Bolstad, B. M., and Irizarry, R. A. (2004) Bioinformatics
(Oxf.) 20, 307–315
23. Irizarry, R. A., Hobbs, B., Collin, F., Beazer-Barclay, Y. D., Antonellis, K. J.,
Scherf, U., and Speed, T. P. (2003) Biostatistics 4, 249 –264
24. Benjamini, Y., and Hochberg, Y. (1995) J. R. Stat. Soc. B 57, 289 –300
25. Kimura, S., Wang, K. Y., Tanimoto, A., Murata, Y., Nakashima, Y., and
Sasaguri, Y. (2004) Pathol. Int. 54, 465– 474
26. Hirayama, T., Sabokbar, A., and Athanasou, N. A. (2002) J. Endocrinol. 175,
155–163
27. Takuma, A., Kaneda, T., Sato, T., Ninomiya, S., Kumegawa, M., and Hakeda,
Y. (2003) J. Biol. Chem. 278, 44667– 44674
28. Lean, J. M., Murphy, C., Fuller, K., and Chambers, T. J. (2002) J. Cell.
Biochem. 87, 386 –393
29. Falus, A. (1993) Immunology 78, 193–196
30. Mor, S., Nagler, A., Barak, V., Handzel, Z. T., Geller-Bernstein, C., and
Fabian, I. (1995) J. Leukocyte Biol. 58, 445– 450
31. Wang, K. Y., Arima, N., Higuchi, S., Shimajiri, S., Tanimoto, A., Murata, Y.,
Hamada, T., and Sasaguri, Y. (2000) FEBS Lett. 473, 345–348
32. McLatchie, L. M., Fraser, N. J., Main, M. J., Wise, A., Brown, J., Thompson,
N., Solari, R., Lee, M. G., and Foord, S. M. (1998) Nature 393, 333–339
33. Dobigny, C., and Saffar, J. L. (1997) J. Cell Physiol. 173, 10 –18
34. Ko, Y. G., Park, H., Kim, T., Lee, J. W., Park, S. G., Seol, W., Kim, J. E., Lee,
W. H., Kim, S. H., Park, J. E., and Kim, S. (2001) J. Biol. Chem. 276,
23028 –23033
35. Kraft, S., Novak, N., Katoh, N., Bieber, T., and Rupec, R. A. (2002) J. Investig.
Dermatol. 118, 830 – 837
36. Dvornyk, V., Liu, Y. Z., Lu, Y., Shen, H., Wilde, J., Conway, T., Recker, R. R.,
and Deng, H. W. (2004) 54th Annual Meeting of American Society of Human
Genetics, Toronto, Canada, October 26 –30, 2004, p. 307 (Abstr. 1650), The
University of Chicago Press, Chicago, IL
37. Etienne, W., Meyer, M. H., Peppers, J., and Meyer, R. A., Jr. (2004) BioTech-
niques 36, 618 – 6
38. Murdoch, C., and Finn, A. (2000) Blood 95, 3032–3043
39. Lisignoli, G., Toneguzzi, S., Grassi, F., Piacentini, A., Tschon, M., Cristino, S.,
Gualtieri, G., and Facchini, A. (2002) Cytokine 20, 231–238
40. Day, C. J., Kim, M. S., Stephens, S. R., Simcock, W. E., Aitken, C. J., Nicholson,
G. C., and Morrison, N. A. (2004) J. Cell. Biochem. 91, 303–315
41. Beaulieu, S., Robbiani, D. F., Du, X., Rodrigues, E., Ignatius, R., Wei, Y.,
Ponath, P., Young, J. W., Pope, M., Steinman, R. M., and Mojsov, S. (2002)
J. Immunol. 169, 2925–2936
42. Fitzpatrick, L. A., Buzas, E., Gagne, T. J., Nagy, A., Horvath, C., Ferencz, V.,
Mester, A., Kari, B., Ruan, M., Falus, A., and Barsony, J. (2003) Proc. Natl.
Acad. Sci. U. S. A. 100, 6027– 6032
43. Graves, L., III, Stechschulte, D. J., Morris, D. C., and Lukert, B. P. (1990)
J. Bone Miner. Res. 5, 1113–1119
44. de Gennes, C., Kuntz, D., and De Vernejoul, M. C. (1992) Clin. Orthop. 279,
281–291
45. Delsignore, J. L., Dvoretsky, P. M., Hicks, D. G., O’Keefe, R. J., and Rosier,
R. N. (1996) Iowa Orthop. J. 16, 126 –134
46. Lesclous, P., Guez, D., Baroukh, B., Vignery, A., and Saffar, J. L. (2004) Bone
(N. Y.) 34, 91–99
47. Tyan, M. L. (1993) J. Intern. Med. 234, 143–148
48. Hodge, J. M., Kirkland, M. A., Aitken, C. J., Waugh, C. M., Myers, D. E., Lopez,
C. M., Adams, B. E., and Nicholson, G. C. (2004) J. Bone Miner. Res. 19,
190 –199
49. Blair, H. C., and Athanasou, N. A. (2004) Histol. Histopathol. 19, 189 –199
50. Luster, A. D. (1998) N. Engl. J. Med. 338, 436 – 445
51. Rahimi, P., Wang, C. Y., Stashenko, P., Lee, S. K., Lorenzo, J. A., and Graves,
D. T. (1995) Endocrinology 136, 2752–2759
52. Volejnikova, S., Laskari, M., Marks, S. C., Jr., and Graves, D. T. (1997) Am. J.
Pathol. 150, 1711–1721
53. Campbell, I. A., Douglas, J. G., Francis, R. M., Prescott, R. J., and Reid, D. M.
Downloaded from http://www.jbc.org/ by guest on November 22, 2018
(2004) Thorax 59, 761–768
54. Huizenga, N. A., Koper, J. W., De Lange, P., Pols, H. A., Stolk, R. P., Burger,
H., Grobbee, D. E., Brinkmann, A. O., De Jong, F. H., and Lamberts, S. W.
(1998) J. Clin. Endocrinol. Metab. 83, 144 –151
55. Yamada, Y., Ando, F., Niino, N., and Shimokata, H. (2002) Genomics 80, 8 –12
56. Higuchi, S., Tanimoto, A., Arima, N., Xu, H., Murata, Y., Hamada, T., Mak-
ishima, K., and Sasaguri, Y. (2001) FEBS Lett. 505, 217–222
57. Zwadlo-Klarwasser, G., Vogts, M., Hamann, W., Belke, K., Baron, J., and
Schmutzler, W. (1998) Inflamm. Res. 47, 434 – 439
58. Laszlo, V., Rothe, G., Hegyesi, H., Szeberenyi, J. B., Orso, E., Schmitz, G., and
Falus, A. (2001) Inflamm. Res. 50, 428 – 434
59. Sasaguri, Y., and Tanimoto, A. (2004) J. Atheroscler. Thromb. 11, 122–130
60. Bloom, W., and Fawcett, D. W. (1975) in A Text Book of Histology (Bloom, W.,
and Fawcett, D. W., eds) p. 153, W. B. Saunders Co., Philadelphia, PA
 A Novel Pathophysiological Mechanism for Osteoporosis Suggested by an in Vivo
Gene Expression Study of Circulating Monocytes
Yao-Zhong Liu, Volodymyr Dvornyk, Yan Lu, Hui Shen, Joan M. Lappe, Robert R.
Recker and Hong-Wen Deng
J. Biol. Chem. 2005, 280:29011-29016.
doi: 10.1074/jbc.M501164200 originally published online June 17, 2005

Access the most updated version of this article at doi: 10.1074/jbc.M501164200

Alerts:
• When this article is cited
• When a correction for this article is posted

Downloaded from http://www.jbc.org/ by guest on November 22, 2018

Click here to choose from all of JBC's e-mail alerts

Supplemental material:
http://www.jbc.org/content/suppl/2005/06/28/M501164200.DC1

This article cites 56 references, 10 of which can be accessed free at

http://www.jbc.org/content/280/32/29011.full.html#ref-list-1