An in vivo genome wide gene expression study of circulating monocytes suggested GBP1, STAT1 and CXCL10 as novel risk genes for the differentiation of peak bone mass

Bone 44 (2009) 1010-1014

Contents lists available at ScienceDirect

Bone j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / b o n e

An in vivo genome wide gene expression study of circulating monocytes suggested GBP1, STAT1 and CXCL10 as novel risk genes for the differentiation of peak bone mass Shu-Feng Lei a, Shan Wu a, Li-Ming Li a, Fei-Yan Deng a, Su-Mei Xiao a, Cheng Jiang a, Yuan Chen a, Hui Jiang a, Fang Yang a, Li-Jun Tan a, Xiao Sun a, Xue-Zhen Zhu a, Man-Yuan Liu a, Yao-Zhong Liu b, Xiang-Ding Chen a, Hong-Wen Deng a,b,⁎ a Laboratory of Molecular and Statistical Genetics, The Key Laboratory of Protein Chemistry and Developmental Biology of Ministry of Education, College of Life Sciences, Hunan Normal University, Changsha, Hunan 410081, P.R. China b Department of Orthopedic Surgery, School of Medicine, University of Missouri-Kansas City, 2411 Holmes Street, Kansas City, MO 64108, USA

a r t i c l e

i n f o

Article history: Received 5 October 2007 Revised 7 May 2008 Accepted 9 May 2008 Available online 28 May 2008 Edited by: Bjorn Olsen Keywords: Circulating monocyte Gene expression Peak bone mass DNA microarray

a b s t r a c t Peak bone mass (PBM) is an important determinant of osteoporosis. Circulating monocytes serve as early progenitors of osteoclasts and produce important molecules for bone metabolism. To search for genes functionally important for PBM variation, we performed a whole genome gene differential expression study of circulating monocytes in human premenopausal subjects with extremely low (N = 12) vs. high (N = 14) PBM. We used Affymetrix HG-U133 plus2.0 GeneChip® arrays. We identified 70 differential expression probe sets (p < 0.01) corresponding to 49 unique genes. After false discovery rate adjustment, three genes [STAT1, signal transducer and activator of transcription 1; GBP1, guanylate binding protein 1; CXCL10, Chemokine (CX-C motif) ligand 10] expressed significantly differentially (p < 0.05). The RT-PCR results independently confirmed the significantly differential expression of GBP1 gene, and the differential expression trend of STAT1. Functional analyses suggested that the three genes are associated with the osteoclastogenic processes of proliferation, migration, differentiation, migration, chemotaxis, adhesion. Therefore, we may tentatively hypothesize that the three genes may potentially contribute to differential osteoclastogenesis, which may in the end lead to differential PBM. Our results indicate that the GBP1, STAT1 and CXCL10 may be novel risk genes for the differentiation of PBM at the monocyte stage. © 2008 Published by Elsevier Inc.

Introduction Osteoporosis is a serious health problem of excessive skeletal fragility leading to low trauma fractures among the elderly [1]. Low bone mass in the elderly is determined by peak bone mass (PBM) in young adults and subsequent bone loss with aging later in life [2]. Attainment and maintenance of high PBM in young and middle aged adults is of primary importance in protecting against late-life osteoporosis. Peripheral blood monocytes can serve as early precursors of osteoclasts [3–6]. In vitro, multinuclear mature functional osteoclasts can be derived from circulating monocytes when placed in a suitable microenvironment [3,6]. Blood monocytes also produce a wide variety of factors involved in bone metabolism, such as interleukin-1, tumor necrosis factor-α (TNF-α), interleukin-6, platelet-derived growth factor, transforming growth factor-β, and 1,25(OH)2D3 [7]. All osteoclasts in peripheral skeleton [4,8] and a considerable amount of osteoclasts in the central skeleton [9] come from circulating monocytes. Recently,

⁎ Corresponding author. Laboratory of Molecular and Statistical Genetics, College of Life Sciences, Hunan Normal University, Changsha, Hunan 410081, P.R. China. Fax: +86 731 8872791. E-mail address: [email protected] (H.-W. Deng). 8756-3282/$ – see front matter © 2008 Published by Elsevier Inc. doi:10.1016/j.bone.2008.05.016

Liu et al. [10] studied peripheral blood monocytes for differential expression genes (DEGs) between elderly women with extremely low vs. high bone mineral density (BMD). They suggested a novel pathophysiological mechanism for osteoporosis that is characterized by increased recruitment of circulating monocyte into bone, and enhanced monocyte differentiation into osteoclasts. The course of differentiation from circulating monocytes to osteoclasts contains a series of osteoclastogenic processes including monocytes’ motility, adhesion, transendothelial migration, prodifferentiation, proliferation, chemotaxis, adhesion, activation, and maturation etc. Therefore, we hypothesize that the DEGs at the precursors of osteoclasts functionally involved in the osteoclastogenic processes may contribute to differential osteoclastogenesis, which may in the end lead to differential PBM. Microarray technology is a high-throughput and powerful tool in identifying and comparing the patterns of gene expression. The technology has substantially improved in recent years and shown reliable and repeatable results across various labs and platforms [11,12]. Here, we analyzed the gene expression profile of circulating monocytes (via the Affymetrix HG-U133 plus2.0 GeneChip® array) in human subjects with extremely low vs. high PBM to identify DEGs functionally potentially relevant to the differentiation of osteoclastogenesis.

S.-F. Lei et al. / Bone 44 (2009) 1010-1014

1011

Materials and methods Subjects The study was approved by Hunan Normal University, ChangSha, China. All the recruited volunteers signed informed consent form before entering this project. All the study subjects belong to Chinese Han ethnic group. We first recruited 878 healthy Chinese premenopausal females aged 20-45 y with an average of 27.3 y when PBM is attainted and maintained [13,14]. Then, we distributed the total sample according to the hip Z-score of PBM. From the bottom 100 and top 100 subjects of the PBM phenotypic distribution, we selected 12 subjects (Mean Z-score ± SD = -1.72 ± 0.60) and 14 (Mean Z-score ± SD = 1.57 ± 0.57) with extremely low and high PBM for further DNA microarray experiments. We adopted strict exclusion criteria detailed by Liu et al. [10], to minimize any known potential confounding effects on the variation of bone phenotype. Briefly, patients with chronic diseases/conditions that may potentially affect bone mass were excluded. These diseases/conditions included chronic disorders involving vital organs (heart, lung, liver, kidney, brain), serious metabolic diseases (diabetes, hypo- or hyperparathyroidism, hyperthyroidism), other skeletal diseases (Paget's disease, osteogenesis imperfecta, rheumatoid arthritis), chronic use of drugs affecting bone metabolism (corticosteroid therapy, anticonvulsant drugs), and malnutrition conditions (chronic diarrhea, chronic ulcerative colitis). BMD measurement

Fig. 1. The phenotypic distribution of 14 and 12 subjects with extremely high vs. low PBM.

BMD (g/cm2) for the lumbar spine (L1-4) and total hip (femoral neck, trochanter, and intertrochanter region) was measured with a Hologic 4500W dual energy X-ray absorptiometer (DEXA) scanner (Hologic Corporation, Waltham, Massachusetts, USA). The machine was calibrated daily. The coefficient of variation of BMD values from the DEXA measurements, obtained from 7 individuals repeatedly measured five times, of the DEXA measurements was 0.80% at the hip.

represented by analysis of over 47,000 transcripts and variants. Then, microarrays were washed (Affymetrix Fluidics Station 450), stained with phycoerythrin-streptavidin, and scanned using an Affymetrix scanner (Gene array Scanner 3000).

Experiment procedures Monocyte isolation A monocyte negative isolation kit (Dynal Biotech Inc., Lake Success, NY) was used to isolate circulating monocytes from 50 ml whole blood following the procedures recommended by the manufacturer. The kit can deplete T cells, B cells, and natural killer cells from mononuclear cells, leaving only monocytes untouched and native and free of the surface-bound antibody and beads. This is of particular importance as binding antibody-coagulated beads to the cell surface may activate the cells and thus change their gene expression profiles. The purity of the isolated monocyte sample was examined by BD-FACScalibur flow cytometry (BD Biosciences, San Jose, CA USA) with fluorescence labeled antibodies PE-CD14 and FITC-CD45. The purity of the isolated monocyte ranged from 70%-90% in our samples. Total RNA extraction and microarray assays Total RNA from monocytes was extracted using Qiagen kit (Qiagen, Inc., Valencia, CA) following the procedures recommended by the manufacturer. Experimental procedures for microarray assays were performed according to the manufacturer's protocol (Affymetrix, Santa Clara, CA). Briefly, RNA was converted to double-stranded cDNA. In vitro transcription was performed to produce biotin-labeled cRNA (BioArray HighYield RNA Transcription Labeling Kit; Enzo Diagnostics). Biotinylated cRNA was cleaned, fragmented, and hybridized (Affymetrix Genechip Hybridization Oven 640) to Affymetrix HG-U133 plus2.0 GeneChips, containing about 38,500 human genes as

Table 1 Basic characteristics of the studied subjects p value

Trait

All

Top 100 subjects

Bottom 100 subjects

Low PBM group for microarrays

High PBM group for microarrays

Number Weight (kg) Height (cm) Age (years) PBM (g/cm2) Hip Spine

878 50.7 ± 6.2

100 47.7 ± 5.5

100 54.4 ± 6.6

12 51.5 ± 7.3

14 55.8 ± 5.7

0.10

158.2 ± 5.1

156.5 ± 4.9

159.7 ± 5.2

158.9 ± 4.4

158.9 ± 5.3

0.98

27.3 ± 4.8

27.1 ± 4.6

27.5 ± 5.0

25.3 ± 3.1

28.7 ± 4.7

0.045

0.87 ± 0.10 0.94 ± 0.10

0.71 ± 0.04 0.85 ± 0.08

1.05 ± 0.05 1.06 ± 0.09

0.70 ± 0.06 0.85 ± 0.07

1.03 ± 0.05 1.04 ± 0.09

0.0001 0.0001

Notes: 1. Values are presented as means ± standard deviation. 2. Top 100 and bottom 100 subjects means the top 100 and bottom 100 subjects of the PBM phenotypic distribution (by hip Z-score). 3. p value is the statistic level of t-test in the low vs. high hip PBM groups for microarray experiments.

Real-time RT-PCR We used two-step real-time RT-PCR to confirm the selected DEGs, i.e., reverse transcription for synthesis of cDNA from total RNA followed by real-time quantitative PCR. RT reactions were performed in a 30 μL reaction volume, containing 3 μL 10× PCR Buffer II, 6.6 μL 25 mM MgCl2, 6 μL dNTPs, 0.75 μL MULV reverse transcriptase, 0.6 μL RNase inhibitor, 1.5 μL Oligo d(T), 0.6 μg total RNA and water to 30 μL. All the above reagents were supplied by Applied Biosystems (Foster City, CA). Reaction conditions were as follows, 10 min at 25 °C, 30 min at 48 °C, 5 min at 95 °C. Real-time quantitative PCR was performed in 20 μL reaction volume using standard protocols on an Applied Biosystem's 7000 Sequence Detection System. Briefly, 2 μL cDNA was mixed with 2 μL 10× TaqMan Buffer A, 4.4 μL 25 mM MgCl2, 0.4 μL 10 mM dATP, 0.4 μL 10 mM dCTP, 0.4 μL 10 mM dGTP, 0.4 μL 20 mM dUTP, 0.2 μL AmpErase UNG, 1 μL Assays-on-DemandTM Gene Expression Assay Mix (contains forward and reverse primers and labeled probe), 0.4 μL GAPDH Probe, 0.4 μL GAPDH Forward Primer, 0.4 μL GAPDH Reverse Primer, 0.1 μL AmpliTaq Gold DNA Polymerase (5.0 U/μL) and water to 20 μL. The thermocycling conditions are as follows: 2 min at 50 °C, 10 min at 95 °C, 40 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 presented in the reaction. All reactions were run in triplicates and included “no template” controls for each gene. As the TaqMan Gene Expression Assays all have amplification efficiencies very close to one, we did not perform validation experiments to test the equality of the amplification efficiencies between the target genes and the reference GAPDH gene. Data analyses GCOS 1.2 (GeneChip Operating Software) was used to process the probe-level raw data. We used the RMA (Robust Multiarray Average) algorithm [15] molded in R package to transform the probe-level raw data into gene expression data. RMA can give most reproducible results and show the highest correlation coefficients with RT-PCR data among currently available algorithms [16]. Based on the expression data generated with the RMA algorithm, Matlab software package was used to perform student t-test to compare the expression signals in subject groups with extremely low vs. high PBM to identify DEGs. To account for multiple testing, a false discovery rate (FDR) method, Benjamini and Hochberg stepwise procedure [17], was used to generate adjusted p values. To better understand the differential expression profile of circulating monocyte between subjects with extremely low vs. high PBM, gene ontological analyses of the DEGs were performed by Onto-Express [18], available at http://vortex.cs.wayne.edu/ ontoexpress/. Pathway-Express [19] was performed to identify potentially interesting pathways for these DEGs. Real-time RT-PCR data were generated with the ABI Prism 7900 sequence detection system software (Applied Biosystems). The cycle number at which the reaction crossed a predetermined cycle threshold (CT) was identified for each gene, and the expression of each target gene relative to GAPDH gene was determined using the equation 2-ΔCT, where ΔCT = (CTTarget Gene-CTGAPDH). Based on the relative gene expression, we performed student's t-test to validate the DEGs between the discordant PBM groups.

Results Table 1 lists the basic characteristics of the studied sample. We can find significant difference of hip PBM between the low and high hip PBM groups for microarray experiments. Fig. 1 intuitively shows that

1012

S.-F. Lei et al. / Bone 44 (2009) 1010-1014

the distribution of 12 (Mean Z-score ± SD = -1.72 ± 0.60) and 14 (Mean Z-score ± SD = 1.57 ± 0.57) selected subjects with extremely low and high PBM, respectively. The average difference of hip Z-score is about 3.29 in the extremely high vs. low PBM groups. Empirically, this large gap in PBM phenotype between the two groups may enhance chance in finding genes related to the differetiation of PBM. Affymetrix HG-U133 plus2.0 GeneChips, containing about 38,500 human genes for analysis of over 47,000 transcripts and variants, were used to disclose the expression profile of circulating monocytes in Chinese young adults. The target signal value in our arrays is 500. The percentage of the probe sets called “Present” relative to the total of 54675 probe sets ranged from 25.6%~33.6% with an average 31.4%. Using RMA transformed expression values, we identified 70 differential expression probe sets (p < 0.01) corresponding to 49 unique genes (Table 2 and Appendix 1). Among these DEGs, compared with the group with extremely high PBM, 37 genes were up-regulated and 12 were down-regulated in the extremely low PBM group. According to the gene ontology (GO) biological process classification, we found 17 major categories for the 49 DEGs including cellular metabolism, response to stimulus, localization, cell communication, cell proliferation etc, which may be potentially relevant to the differentiation of monocyte to osteoclast. Our identified DEGs belong to eleven interesting known pathways. Some of these pathways, e.g., cytokine-cytokine receptor interaction, toll-like receptor signaling pathway, Jak-STAT signaling pathway etc., are significantly involved in bone metabolism and osteoclast differentiation [20–22]. After FDR adjustment for multiple testing, there remain three significantly DEGs [STAT1, signal transducer and activator of transcription 1; GBP1, guanylate binding protein 1; CXCL10, chemokine (C-X-C motif) ligand 10] (FDR adjusted p < 0.05). The fold changes for their expressions in the low vs. high PBM group range from 1.39-1.67. The three genes are closely associated with bone metabolism and the processes of osteoclastogeneis (see details in Discussion) (Fig. 2). Due to limited amount of total RNA of each sample, we selected STAT1 and GBP1 for further RT-PCR validation. The selection of the two genes was based on their statistical significance of differential expression detected by DNA microarray analyses, and/or on their prior known function relevance to bone metabolism. STAT1 is a primary mediator of interferon signaling pathways that are involved in osteoclast differentiation [21–23]. Although prior knowledge of GBP1 has not shown directly functional relevance to bone metabolism, GBP1 is a STAT1-dependent protein, and functional studies have shown a strong relationship of GBP1 and STAT1 genes [24,25] (see details in Discussion). Moreover, three probe sets of 202269_x_at, 231577_s_at, 202270_at in GBP1 gene showed very strong and consistent significantly differential expression (p = 0.0000, 0.0004, and 0.0013, respectively). Using the relative gene expression values for each sample (i.e., 2-ΔCT), student's t-test indicated that GBP1 gene had significantly (p = 0.032) higher expression (up-regulated by 2.7-fold) in the low vs. high PBM groups by RT-PCR, which confirmed the result

Table 2 Information of three significantly DEGs Gene name

Chemokine (C-X-C motif) ligand 10

Signal transducer and activator of transcription 1

Guanylate binding protein 1

Gene symbol Probe set Location RAM p value FDR p value Fold L/H

CXCL10 204533_at 4q21 <10- 4 0.004 1.39

STAT1 209969_s_at 2q32.2 <10- 4 0.0049 1.67

GBP1 202269_x_at 1p22.2 <10- 4 0.0267 1.56

Notes: 1. Fold L/H is the ratio of mean expression value (RMA transform) of gene in the low vs. high PBM groups. 2. p value is the statistical level of t-test using the RMA transform expression value.

Fig. 2. Potential effects of STAT1, CXCL10 and GBP1 on osteoclastogenesis. Notes: monocyte; multinucleate cell; mature osteoclast; active osteoclast.

of this gene in DNA microarray study. We also confirmed the expression trend of the STAT1 gene in the low vs. high PBM groups. The STAT1 gene is up-regulated by 1.7 fold in the low vs. high PBM group, respectively. Discussion In this study, we systemically investigated the role of circulating monocytes in the differentiation of PBM in vivo in Chinese premenopausal female samples. After screening the expression of about 38,500 human genes in extremely low vs. high PBM groups, our results revealed interesting differential expression profiling. Three tentatively confirmed genes (STAT1, GBP1 and CXCL10) may have potential function relevant with the differentiation of PBM at the monocyte stage. Previous knowledge reported the functional relevance of STAT1, CXCL10, and GBP1 genes to bone metabolism. STAT1 may serve as a primary mediator of interferon (IFN) signaling pathways involving osteoclast differentiation [20,26]. Its interaction effects with c-fos and p21 genes enhanced the proliferation of lymphocytes in rheumatoid arthritis patients [27]. Through the p38 MAPK pathway, RANKL stimulates the serine phosphorylation of STAT1 resulting in the migration and adhesion of osteoclast precursors [28]. No direct evidence supports that GBP1 gene is involved in bone metabolism or osteoclast differentiation. However, GBP1 is a STAT1-dependent protein, induced by IFNs [29]. Functional studies have addressed the strong relationship of GBP1 and STAT1 genes [24,25]. The sumoylation-defective STAT1 mutant displayed increased induction of GBP1 and transporters associated with antigen presentation 1 (TAP1) transcription [24]. The mutation in STAT1 gene dramatically reduced the inducibility of the GBP1 and TAP1 genes by IFN [25]. The CXCL10 gene is also an IFNinducible protein. It is expressed in human osteoclasts with changing expression levels during osteoclast differentiation [30]. Ingenuity Pathways Analysis (IPA) (Ingenuity® Systems, www.ingenuity.com) suggested that the three genes are associated with some biological processes of proliferation, migration, differentiation, migration, chemotaxis, adhesion [26,30–32]. During the course of differentiation from circulating moncytes to osteoclasts, there are a series of biological processes including monocytes' motility, adhesion, transendothelial migration, differentiation, proliferation, chemotaxis, adhesion, activation, and maturation etc. Therefore, we may tentatively hypothesize that the three DEGs functionally involved in the above biological processes may contribute to differential osteoclastogenesis, which may in the end lead to differential PBM (Fig. 2). Future functional studies are needed to provide in-depth evidence to support the above hypothesis.

S.-F. Lei et al. / Bone 44 (2009) 1010-1014

Our sampling strategy conferred advantage of detecting interesting genes. First, all the studied subjects are in their PBM period when bone mass is fairly stable due to a balance between bone formation by osteoblasts and bone resorption by osteoclasts. This homeostasis is much buffered against in- and/or ex-trinsic confounding factors and perturbations. Thus, the difference of PBM may most likely derive from the intrinsic expressing and functional differentiation of the involved genes. The PBM period may provide us with a special time window to cross-sectionally explore the “sustaining-effective” factors contributing to variation of PBM. Second, the average difference of hip Z-score is about 3.29 in the extremely high vs. low PBM groups. This extreme sampling scheme may greatly enhance the statistical power in gene mapping and fine mapping [33,34] and may also have high chance to detect genes potentially affecting the differentiation of PBM. Previous studies addressed the importance of osteoclasts [35], osteoblasts [36], osteocytes [37,38], and mesenchymal stem cells [39] on the pathogenesis of osteoporosis, but few studies focused on circulating monocyte – precursors of osteoclasts (except Liu et al.) [10]. In particular, few earlier studies have been conducted like this study in vivo in humans for bone related cells. In the present study we selected circulating monocytes as our target cells because circulating monocytes serve as progenitors of osteoclasts [3–6], and secrete osteoclastogenic cytokines, such as IL-1, IL-6, and TNF-α [40–42]. Moreover, circulating monocytes are relatively easy to isolate for in vivo studies. Using circulating monocytes, previous studies have suggested a novel pathophysiological mechanism for osteoporosis through an in vivo gene expression study in Caucasians [10]. We systemically screened a total of 54,675 probe sets for each sample. Although in our sample the phenotypic difference of average hip PBM Z-score in the two extreme PBM groups is greater than 3.0, we only identified 70 differential expression probe sets corresponding to 49 different genes with p < 0.01. This may implicate that during the PBM period, gene expression in circulating monocyte may be under fine control for homeostasis maintenance. Minor functional differences of circulating monocytes, as detected by DNA microarray, may amplify exponentially the quantitative changes on phenotype level (i.e., large PBM difference), via a series of complex regulation during the differentiation from circulating monocytes to osteoclasts. Such functional difference may be further amplified at the phenotypic level after menopause, when monocytes are functionally activated due to the largely deficient inhibitory effects of estrogen on monocytes [43]. Only a few of DEGs such as the fumarate hydratase gene and lymphoid-restricted membrane protein gene overlapped between this study and a previous study [10] using circulating monocyte as target cell. However, most of the genes identified are different between the two studies. Ethnic difference and different sampling scheme may be responsible for the low overlap rate of DEGs. The present study was on Chinese, while Liu et al. [10] studied Caucasians. There may be ethnic genetic differences in pathogenesis of osteoporosis [44–46]. This implicates that the genes underling bone mass variation may not be totally the same across different ethnic populations. Alternatively and in particular, our samples are in their PBM period and in premenopausal period, but Liu et al. [10] recruited Caucasian women aged between 47 and 55 with an average age of ~ 51, which included both pre-menopausal and in post-menopausal women. The global gene expression in monocytes in different physiological periods may be different. This is plausible particularly given that the estrogen levels of the subjects in PBM and those approaching or just after menopause ages are drastically different and the functional relevance (and thus gene expression) of monocytes is strongly under the influence of estrogen age of ~51 [47], which included both pre-menopausal and in post-menopausal women. The global gene expression in monocytes in different physiological periods may be different. This is plausible particularly given that the estrogen levels of the subjects in PBM and those approaching or just after menopause ages are drastically

1013

different and the functions (and thus gene expression) of monocytes is strongly under the influence of estrogen. In conclusion, the present study represents our first effort on DNA microarray study of circulating monocytes in vivo via extreme sampling scheme in Chinese to search for genes affecting the differentiation of PBM. Three genes (STAT1, CXCL10 and GBP1) may have potential importance in the differentiation of PBM and deserves further specific functional studies. Acknowledgments The study was partially supported by grants from Natural Science Foundation of China (30600364, 30470534, and 30230210). HWD was partially supported by grants from NIH (R01AR050496, R21AG027110, K01 AR 02170) and the Dickson/Missouri endowment. Appendix A. Information for other 46 DEGs with p < 0.01

Probe set

Gene symbol

200986_at

SERPING1

205241_at 214511_x_at 203773_x_at 210592_s_at 201649_at 223502_s_at 201480_s_at 216950_s_at 238327_at 208436_s_at 202748_at 227609_at 205992_s_at 224983_at 222670_s_at 1558924_s_at 217502_at 224701_at 208012_x_at 218085_at 204224_s_at 202687_s_at 200748_s_at 213236_at 202626_s_at 224173_s_at 204994_at 219938_s_at 226929_at 230966_at 217933_s_at 202446_s_at 228055_at 204415_at

Gene name

p L/ value H

Serine (or cysteine) proteinase inhibitor, clade G (C1 inhibitor), member 1 SCO2 SCO cytochrome oxidase deficient homolog 2 (yeast) LOC440607 Fc-gamma receptor I B2 BLVRA Biliverdin reductase A SAT Spermidine/spermine N1-acetyltransferase UBE2L6 Ubiquitin-conjugating enzyme E2L 6 TNFSF13B Tumor necrosis factor (ligand) superfamily, member 13b SUPT5H Suppressor of Ty 5 homolog (S. cerevisiae) FCGR1A Fc fragment of IgG, high affinity Ia, receptor for (CD64) ECGF1 Endothelial cell growth factor 1 (platelet-derived) IRF7 Interferon regulatory factor 7 GBP2 Guanylate binding protein 2, interferon-inducible EPSTI1 Epithelial stromal interaction 1 (breast) IL15 Interleukin 15 SCARB2 Scavenger receptor class B, member 2 MAFB v-maf musculoaponeurotic fibrosarcoma oncogene homolog B (avian) RSN Restin (Reed-Steinberg cell-expressed intermediate filament-associated protein) IFIT2 Interferon-induced protein with tetratricopeptide repeats 2 PARP14 Poly (ADP-ribose) polymerase family, member 14 SP110 SP110 nuclear body protein SNF7DC2 SNF7 domain containing 2 GCH1 GTP cyclohydrolase 1 (dopa-responsive dystonia) TNFSF10 Tumor necrosis factor (ligand) superfamily, member 10 FTH1 Ferritin, heavy polypeptide 1 SASH1 SAM and SH3 domain containing 1 LYN v-yes-1 Yamaguchi sarcoma viral related oncogene homolog MRPL30 Mitochondrial ribosomal protein L30 MX2 Myxovirus (influenza virus) resistance 2 (mouse) PSTPIP2 Praline-serine-threonine phosphatase interacting protein 2 MTHFR 5,10-methylenetetrahydrofolate reductase (NADPH) IL4I1 Interleukin 4 induced 1 LAP3 Leucine aminopeptidase 3 PLSCR1 Phospholipids scramblase 1 NAPSB Napsin B aspartic peptidase pseudogene G1P3 Interferon, alpha-inducible protein (clone IFI-6-16)

0.0001 ↑ 0.0002 ↑ 0.0002 0.0006 0.0006 0.0008 0.0015

↑ ↑ ↑ ↑ ↑

0.0018 0.0019

↓ ↑

0.0021

↓

0.0022 ↑ 0.0022 ↑ 0.0028 0.0029 0.0031 0.0032

↑ ↑ ↑ ↑

0.0033 ↑ 0.0036 ↓ 0.0036 ↑ 0.0036 ↑ 0.0043 ↓ 0.0044 ↑ 0.0045 ↑ 0.0052 ↓ 0.0054 ↓ 0.0054 ↑ 0.0054 ↓ 0.0056 ↑ 0.0059 ↑ 0.006

↓

0.0067 0.0067 0.0068 0.0069 0.007

↑ ↑ ↑ ↓ ↑

(continued on on next page) (continued

1014

S.-F. Lei et al. / Bone 44 (2009) 1010-1014

(continued) Appendix A (continued) Probe set

Gene symbol

Gene name

p L/ value H

225344_at 214453_s_at 206133_at 208751_at

NCOA7 IFI44 HSXIAPAF1 NAPA

0.0071 0.0072 0.0078 0.008

218543_s_at 215001_s_at

ZC3HDC1 GLUL

200629_at 210980_s_at

WARS ASAH1

201669_s_at

MARCKS

Nuclear receptor coactivator 7 Interferon-induced protein 44 XIAP associated factor-1 N-ethylmaleimide-sensitive factor attachment protein, alpha Zinc finger CCCH type domain containing 1 Glutamate-ammonia ligase (glutamine synthase) Tryptophanyl-tRNA synthetase N-acylsphingosine amidohydrolase (acid ceramidase) 1 Myristoylated alanine-rich protein kinase C substrate Membrane-associated protein 17 Extracellular matrix protein 1

1553589_a_at MAP17 209365_s_at ECM1

↑ ↑ ↑ ↓

0.0081 ↑ 0.0083 ↑ 0.0083 ↑ 0.0086 ↓ 0.0089 ↑ 0.0098 ↑ 0.0098 ↓

Notes: 1. L/H is the ratio of mean expression value (RMA transform) of gene in the low vs. high PBM groups. 2. p value is the statistical level of t-test

References [1] Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359(9319):1761–7. [2] Teegarden D, Proulx WR, Martin BR, et al. Peak bone mass in young women. J Bone Miner Res 1995;10(5):711–5. [3] Udagawa N, Takahashi N, Akatsu T, et al. Origin of osteoclasts: mature monocytes and macrophages are capable of differentiating into osteoclasts under a suitable microenvironment prepared by bone marrow-derived stromal cells. Proc Natl Acad Sci U S A 1990;87(18):7260–4. [4] Zambonin ZA, Teti A, Primavera MV. Monocytes from circulating blood fuse in vitro with purified osteoclasts in primary culture. J Cell Sci 1984;66:335–42. [5] Fujikawa Y, Quinn JM, Sabokbar A, McGee JO, Athanasou NA. The human osteoclast precursor circulates in the monocyte fraction. Endocrinology 1996;137(9):4058–60. [6] Quinn JM, Neale S, Fujikawa Y, McGee JO, Athanasou NA. Human osteoclast formation from blood monocytes, peritoneal macrophages, and bone marrow cells. Calcif Tissue Int 1998;62(6):527–31. [7] Nathan CF. Secretory products of macrophages. J Clin Invest 1987;79(2):319–26. [8] Horton MA, Spragg JH, Bodary SC, Helfrich MH. Recognition of cryptic sites in human and mouse laminins by rat osteoclasts is mediated by beta 3 and beta 1 integrins. Bone 1994;15(6):639–46. [9] Parfitt AM. Osteonal and hemi-osteonal remodeling: the spatial and temporal framework for signal traffic in adult human bone. J Cell Biochem 1994;55(3):273–86. [10] Liu YZ, Dvornyk V, Lu Y, et al. A novel pathophysiological mechanism for osteoporosis suggested by an in vivo gene expression study of circulating monocytes. J Biol Chem 2005;280(32):29011–6. [11] Shi L, Reid LH, Jones WD, et al. The MicroArray Quality Control (MAQC) project shows inter- and intraplatform reproducibility of gene expression measurements. Nat Biotechnol 2006;24(9):1151–61. [12] Canales RD, Luo Y, Willey JC, et al. Evaluation of DNA microarray results with quantitative gene expression platforms. Nat Biotechnol 2006;24(9):1115–22. [13] Yao WJ, Wu CH, Wang ST, et al. Differential changes in regional bone mineral density in healthy Chinese: age-related and sex-dependent. Calcif Tissue Int 2001;68(6):330–6. [14] Qin M, Yu W, Meng X, Xing X, Xu L. [Normal spinal changes of bone mineral density in 445 individuals: assessment by quantitative computed tomography]. Zhongguo Yi Xue Ke Xue Yuan Xue Bao 1996;18(6):439–43. [15] Irizarry RA, Hobbs B, Collin F, et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003;4(2):249–64. [16] Millenaar FF, Okyere J, May ST, et al. How to decide? Different methods of calculating gene expression from short oligonucleotide array data will give different results. BMC Bioinformatics 2006;7:137. [17] Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B 1995;57:289–300. [18] Khatri P, Draghici S, Ostermeier GC, Krawetz SA. Profiling gene expression using onto-express. Genomics 2002;79(2):266–70. [19] Khatri P, Sellamuthu S, Malhotra P, et al. Recent additions and improvements to the Onto-Tools. Nucleic Acids Res 2005;33(Web Server issue):W762–5. [20] Hayashi T, Kaneda T, Toyama Y, Kumegawa M, Hakeda Y. Regulation of receptor activator of NF-kappa B ligand-induced osteoclastogenesis by endogenous interferon-beta (INF-beta) and suppressors of cytokine signaling (SOCS). The possible counteracting role of SOCSs-in IFN-beta-inhibited osteoclast formation. J Biol Chem 2002;277(31):27880–6.

[21] Hayashi S, Yamada T, Tsuneto M, et al. Distinct osteoclast precursors in the bone marrow and extramedullary organs characterized by responsiveness to Toll-like receptor ligands and TNF-alpha. J Immunol 2003;171(10):5130–9. [22] Fox SW, Haque SJ, Lovibond AC, Chambers TJ. The possible role of TGF-betainduced suppressors of cytokine signaling expression in osteoclast/macrophage lineage commitment in vitro. J Immunol 2003;170(7):3679–87. [23] Lovibond AC, Haque SJ, Chambers TJ, Fox SW. TGF-beta-induced SOCS3 expression augments TNF-alpha-induced osteoclast formation. Biochem Biophys Res Commun 2003;309(4):762–7. [24] Ungureanu D, Vanhatupa S, Gronholm J, Palvimo JJ, Silvennoinen O. SUMO-1 conjugation selectively modulates STAT1-mediated gene responses. Blood 2005;106(1):224–6. [25] Kovarik P, Mangold M, Ramsauer K, et al. Specificity of signaling by STAT1 depends on SH2 and C-terminal domains that regulate Ser727 phosphorylation, differentially affecting specific target gene expression. EMBO J 2001;20 (1-2):91–100. [26] Huang W, O'Keefe RJ, Schwarz EM. Exposure to receptor-activator of NFkappaB ligand renders pre-osteoclasts resistant to IFN-gamma by inducing terminal differentiation. Arthritis Res Ther 2003;5(1):R49–59. [27] Hikasa M, Yamamoto E, Kawasaki H, et al. p21waf1/cip1 is down-regulated in conjunction with up-regulation of c-Fos in the lymphocytes of rheumatoid arthritis patients. Biochem Biophys Res Commun 2003;304(1):143–7. [28] Kwak HB, Lee SW, Jin HM, et al. Monokine induced by interferon-gamma is induced by receptor activator of nuclear factor kappa B ligand and is involved in osteoclast adhesion and migration. Blood 2005;105(7):2963–9. [29] Cheng YS, Patterson CE, Staeheli P. Interferon-induced guanylate-binding proteins lack an N(T)KXD consensus motif and bind GMP in addition to GDP and GTP. Mol Cell Biol 1991;11(9):4717–25. [30] Grassi F, Piacentini A, Cristino S, et al. Human osteoclasts express different CXC chemokines depending on cell culture substrate: molecular and immunocytochemical evidence of high levels of CXCL10 and CXCL12. Histochem Cell Biol 2003;120(5):391–400. [31] Nguyen AN, Stebbins EG, Henson M, et al. Normalizing the bone marrow microenvironment with p38 inhibitor reduces multiple myeloma cell proliferation and adhesion and suppresses osteoclast formation. Exp Cell Res 2006;312 (10):1909–23. [32] Hillyer P, Mordelet E, Flynn G, Male D. Chemokines, chemokine receptors and adhesion molecules on different human endothelia: discriminating the tissuespecific functions that affect leucocyte migration. Clin Exp Immunol 2003;134 (3):431–41. [33] Risch N, Zhang H. Extreme discordant sib pairs for mapping quantitative trait loci in humans. Science 1995;268(5217):1584–9. [34] Deng HW, Li J. The effects of selected sampling on the transmission disequilibrium test of a quantitative trait locus. Genet Res 2002;79(2):161–74. [35] Boyce BF, Hughes DE, Wright KR, Xing L, Dai A. Recent advances in bone biology provide insight into the pathogenesis of bone diseases. Lab Invest 1999;79 (2):83–94. [36] Manolagas SC, Kousteni S, Jilka RL. Sex steroids and bone. Recent Prog Horm Res 2002;57:385–409. [37] Tomkinson A, Gevers EF, Wit JM, Reeve J, Noble BS. The role of estrogen in the control of rat osteocyte apoptosis. J Bone Miner Res 1998;13(8):1243–50. [38] Tomkinson A, Reeve J, Shaw RW, Noble BS. The death of osteocytes via apoptosis accompanies estrogen withdrawal in human bone. J Clin Endocrinol Metab 1997;82(9):3128–35. [39] Nuttall ME, Gimble JM. Controlling the balance between osteoblastogenesis and adipogenesis and the consequent therapeutic implications. Curr Opin Pharmacol 2004;4(3):290–4. [40] Cohen-Solal ME, Graulet AM, Denne MA, et al. Peripheral monocyte culture supernatants of menopausal women can induce bone resorption: involvement of cytokines. J Clin Endocrinol Metab 1993;77(6):1648–53. [41] Eghbali-Fatourechi G, Khosla S, Sanyal A, et al. Role of RANK ligand in mediating increased bone resorption in early postmenopausal women. J Clin Invest 2003;111 (8):1221–30. [42] Pacifici R. Estrogen, cytokines, and pathogenesis of postmenopausal osteoporosis. J Bone Miner Res 1996;11(8):1043–51. [43] Sorensen MG, Henriksen K, Dziegiel MH, Tanko LB, Karsdal MA. Estrogen directly attenuates human osteoclastogenesis, but has no effect on resorption by mature osteoclasts. DNA Cell Biol 2006;25(8):475–83. [44] Dvornyk V, Liu PY, Long JR, et al. Contribution of genotype and ethnicity to bone mineral density variation in Caucasians and Chinese: a test for five candidate genes for bone mass. Chin Med J (Engl) 2005;118(15):1235–44. [45] Dvornyk V, Liu XH, Shen H, et al. Differentiation of Caucasians and Chinese at bone mass candidate genes: implication for ethnic difference of bone mass. Ann Hum Genet 2003;67(Pt 3):216–27. [46] Lei SF, Chen Y, Xiong DH, Li LM, Deng HW. Ethnic difference in osteoporosis-related phenotypes and its potential underlying genetic determination. J Musculoskelet Neuronal Interact 2006;6(1):36–46. [47] Ben Hur H, Mor G, Insler V, et al. Menopause is associated with a significant increase in blood monocyte number and a relative decrease in the expression of estrogen receptors in human peripheral monocytes. Am J Reprod Immunol 1995;34(6): 363–9.