Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2024 Aug 25;43(1):84–93. doi: 10.1002/jor.25963

DNA methylation of bone morphogenetic protein 7 in leukocytes as a possible biomarker for hand osteoarthritis: A pilot study

Takashi Kuroiwa 1, Yoshiki Tsuboi 2, Takehiro Michikawa 3, Kaori Tajima 1, Yuki Uraya 1, Atsushi Maeda 1, Kanae Shizu 1, Katsuji Suzuki 1, Koji Suzuki 2, Yusuke Kawano 1, Nobuyuki Fujita 1,✉
PMCID: PMC11615413  PMID: 39182186

Abstract

Hand osteoarthritis (HOA), characterized by an earlier onset age and reduced susceptibility to mechanical stress compared with knee and hip osteoarthritis, is considered a suitable disease for identifying predictive biomarkers of osteoarthritis. In particular, DNA methylation variants, expected to contribute to HOA susceptibility, hold potential as osteoarthritis biomarkers. In this study, leukocyte DNA methylation patterns were analyzed in blood samples from patients with HOA, aiming to identify disease‐specific biomarkers for osteoarthritis. Using DNA methylation microarrays, we analyzed samples from three subjects with HOA and three age‐ and gender‐matched healthy individuals. For validation, pyrosequencing analysis was conducted using samples from 16 to 9 subjects with and without HOA, respectively. From 735,026 probes in the DNA methylation array, the Top 100 CpG sites associated with HOA, based on low adjusted P‐values, including those targeting bone morphogenetic protein 7 (BMP7), SBF2‐AS1, PLOD2, ICOS, and CSF1R were identified. Validation analysis revealed significantly higher methylation levels in the BMP7‐related site in the HOA group compared with the control group, even after adjusting for age, gender, and body mass index (p = 0.037). In contrast, no significant difference was observed in the other selected CpG sites between the HOA and control groups. This study highlights the significantly increased frequency of methylation at the specific BMP7 site in leukocytes of patients with HOA, suggesting its potential as a biomarker for HOA. Measurement of methylation levels at the CpG sites identified in this study offers a potential approach to prevent future osteoarthritis progression, providing valuable insights into disease management.

Keywords: biomarkers, BMP7, DNA methylation microarray, hand osteoarthritis, leukocyte

1. INTRODUCTION

Hand osteoarthritis (HOA) involves arthritis in the distal interphalangeal joints, proximal interphalangeal joints, and the thumb's carpometacarpal joint, referred to as Heberden's nodes, Bouchard's nodes, and carpometacarpal joint arthrosis, respectively. According to radiological assessments, HOA prevalence ranges from 21% in the United States population to 92% in the Japanese population, surpassing than that of hip or knee osteoarthritis, whereas the prevalence of symptomatic HOA is estimated at approximately 8%. 1 Onset age is lower for HOA than for knee and hip osteoarthritis, and within osteoarthritis clusters, HOA may predict the onset of knee and hip osteoarthritis. 2 Despite the necessity for establishing predictive osteoarthritis biomarkers for early detection, such biomarkers have not been adopted clinically, whereas biomarkers for osteoporosis and rheumatoid arthritis have been clinically applied. 3 , 4 Given its earlier onset and reduced susceptibility to mechanical stress factors, such as body weight and physique, HOA may offer a more suitable focus for predictive osteoarthritis biomarker discovery compared with knee and hip osteoarthritis.

DNA methylation is an epigenetic phenomenon altering gene activity without changing genetic information. Methylation predominantly occurs at CpG dinucleotides, playing a crucial role in suppressing gene expression. DNA methylation patterns, varying based on developmental process, cell type, and environmental factors, are vital for cellular differentiation and physiological maintenance. 5 Additionally, DNA methylation is associated with disease onset, with changes in DNA methylation holding potential as disease‐specific biomarkers in cancer, neurological diseases, and cardiovascular diseases. 5 Methylation pattern variations in certain genetic sites are valuable for early detection and risk assessment of these diseases. 5 In osteoarthritis, DNA methylation variants contributing to its susceptibility are considered potential diagnosis and prognosis biomarkers, 6 although studies on these aspects remain limited. 7 Therefore, the objective of this study was to comprehensively analyze leukocyte DNA methylation in blood samples from patients, aiming to identify disease‐specific methylation patterns in HOA. We hypothesized that, through this analysis, distinct osteoarthritis‐specific methylation patterns would emerge.

2. MATERIALS AND METHODS

2.1. Patient selection

The flow chart in Figure 1 illustrates sample selection for our comprehensive analysis. Initially, we enrolled 61 subjects (17 male and 44 female patients; average age: 64.1 ± 10.1 years) who sought consultation for hand and finger joint pain at our department. These participants were subjected to blood tests as well as full‐length lower limb, spine, and hand radiographs, and their medical histories were reviewed. Following the American College of Rheumatology criteria, 49 subjects were diagnosed with HOA. 8 Subsequently, 20 women with HOA presenting with Heberden's nodes were selected from these subjects. Cases with a history of other diseases, including cancer and autoimmune diseases, previous orthopedic surgeries, or spine or lower limb osteoarthritis in radiographs were excluded. Consequently, three remaining subjects were designated for comprehensive analysis. Additionally, three age‐ (±6 years) and gender‐matched healthy individuals with no history of cancer or autoimmune diseases nor hand, lower limb, or spine osteoarthritis in radiographs were selected as controls.

Figure 1.

Figure 1

Open in a new tab

Flow chart depicting the sample selection process for comprehensive analysis.

The flow chart in Figure 2 illustrates sample selection for validation analysis. Out of the 61 patients who consulted to our department with hand and finger joint pain, 20 subjects remained after excluding those with knee, hip, or spine osteoarthritis or other diseases. From these subjects, 16 individuals were diagnosed with HOA, whereas 4 were not, with the 16 subjects being designated for validation analysis. Among these subjects, there were four patients who had only carpometacarpal arthropathy. An additional five healthy individuals were enrolled, totaling nine subjects used as controls in the validation analysis. In the validation analysis, three individuals from each group used in the comprehensive analysis were also included.

Figure 2.

Figure 2

Open in a new tab

Flow chart depicting the sample selection process for validation analysis.

2.2. DNA extraction

Patient samples were obtained from our institution's bioresource room. Peripheral blood, collected under fasting conditions in tubes containing ethylenediaminetetraacetic acid 2Na, was processed to obtain the buffy coat using centrifugation. DNA extracted from the buffy coat underwent bisulfite conversion using the EpiTect Fast DNA Bisulfite Kit (QIAGEN).

2.3. DNA methylation microarray

DNA methylation was measured using Illumina Infinium MethylationEPIC (Illumina). Probes were filtered using ChAMP (version 2.24.0) R packages, with criteria including low quality (detected p > 0.01), low bead count (<5%), non‐cg probes, probes with a CpG near a SNP, probes aligning to multiple locations, and probes from X and Y chromosomes. The filtered beta matrix was normalized using the beta mixture quantile dilation method. Singular value decomposition analysis was performed, and a heatmap illustrating the effect of factors on the original data set was generated using ChAMP (version 2.24.0) R packages. Differentially methylated probes were identified using a method implementing the limma package, and p values obtained using a linear model were adjusted via the Benjamini and Hochberg method. Differentially methylated regions were detected using the Bumphunter method with its default parameters.

2.4. Pyrosequencing validation

PCR was performed using bisulfite‐treated DNA (10 ng/μL) with TaKaRa EpiTaq HS (for bisulfate‐treated DNA; Takara). Using PCR products, quantitative CpG methylation analysis was conducted on a PyroMark Q24 Advanced (QIAGEN) system using PyroMark Q24 Advanced Reagents (QIAGEN) following the manufacturer's instructions. This procedure was described in detail in a previous study. 9 Five CpG sites (bone morphogenetic protein 7 [BMP7] cg20955836, SBF2 antisense RNA 1 [SBF2‐AS1] cg20103036, procollagen‐lysine, 2‐oxoglutarate 5‐dioxygenase 2 [PLOD2] cg24790788, inducible T cell costimulator [ICOS] cg18561976, colony stimulating factor 1 receptor [CSF1R] cg01875467) were measured using specific primers (BMP7: forward, 5′‐ATGAGGGAGAGATATTAAAGTGATAG‐3′; reverse, 5′‐CTTTTAAAATACAACCCTCTCAAACTAC‐3′; sequencing, 5′‐TTTAGGTGGTGAGGGG‐3′. SBF2‐AS1: forward, 5′‐TGTTATTGATAAAGTGTTGGTTGATATATG ‐3′; reverse, 5′‐CACAAATCCCTCAATAAACTAAACCTATA ‐3′; sequencing, 5′‐AGTGTTGGTTGATATATGT ‐3′. PLOD2: forward, 5′‐TTGTGTTAAGAGAGTTTTAGATTTTTT‐3′; reverse, 5′‐CTCATTCCAACAAATTAAACCCTTAAC‐3′; sequencing, 5′‐AGTAATGAGGTTTGTAAATTT ‐3′. ICOS: forward, 5′‐ AGAAGTAGGGTGGTTTTGAAAAATATAG‐3′; reverse, 5′‐ AAAAAAAAAATACCAAAAACCTAACTTCA‐3′; sequencing, 5′‐ ATTTATTGTTAGTTTTGAATATTGA ‐3′. CSF1R: forward, 5′‐ TTTGGAATTTGGGTTTTAGTAGTTGTTTG‐3′; reverse, 5′‐ CTCCACCATATACTTTAACTTTAACTATCA‐3′; sequencing, 5′‐ TTAGTAGTTGTTTGTTATAGAG ‐3′). CpG methylation level results were analyzed using PyroMark Q24 Advanced software (QIAGEN).

2.5. Statistical analyses

PCR analysis values were compared using t tests. Univariable differences in methylation levels between HOA and control groups were assessed, followed by analysis of covariance (ANCOVA) adjusting for age, sex, and body mass index (BMI) as potential confounding factors. STATA 16 software (Stata Corporation) was used for ANCOVA, with the significance level set at p < 0.05.

3. RESULTS

Table 1 shows sample characteristics for DNA methylation microarray and Pyrosequencing validation. In both analyses, no significant differences in age and gender were observed between the HOA group and the control group, though a significant difference in BMI was noted. Figure 3 presents representative hand radiographs of patients with HOA and control subjects used for DNA methylation microarray analysis. In this microarray, 735,026 probes were used. Figure 4 shows Manhattan plot of the microarray. We identified the Top 100 CpG sites associated with HOA based on low adjusted p values (Table 2). Among these sites, the target genes that were repeatedly reported to be associated with osteoarthritis in the past were cg24790788 of PLOD2, cg18561976 of ICOS, cg01875467 of CSF1R, cg20955836 of BMP7, and cg05439665 of MCF2L. 10 , 11 , 12 , 13 , 14 Therefore, in the validation analysis, we focused on these five candidate genes. Additionally, cg20103036 of SBF2‐AS1, one of Top five CpG sites, was also included in this analysis. For MCF2L, we were unable to create matching primers. Validation analysis results are shown in Figure 5. Notably, BMP7 DNA methylation levels were significantly higher in patients with HOA than in control subjects (25.3% ± 7.6% and 16.5% ± 4.2%, respectively; p = 0.004). Importantly, these results remained consistent even after adjusting for age, gender, and BMI (p = 0.037) (Table 3). Conversely, no significant difference was observed in DNA methylation levels of SBF2‐AS1 (57.8% ± 2.7% and 56.0% ± 6.0%, respectively; p = 0.397), PLOD2 (32.4% ± 7.5% and 33.8% ± 9.6%, respectively; p = 0.709), ICOS(28.7% ± 4.4% and 27.7% ± 8.1%, respectively; p = 0.734), and CSF1R (81.5% ± 4.9% and 77.7% ± 5.1%, respectively; p = 0.087) between the HOA and control groups (Figure 5) (Table 3).

Table 1.

Sample characteristics for each analysis.

Comprehensive analysis Validation analysis
Control (n = 3) HOA (n = 3) p Value Control (n = 9) HOA (n = 16) p Value
Male/female 0/3 0/3 1 2/7 2/14 0.946
Age (years) 53.7 ± 2.1 49.0 ± 4.4 0.170 60.9 ± 12.2 65.3 ± 10.8 0.364
BMI (kg/m2) 20.3 ± 1.3 25.9 ± 1.9 0.012 20.9 ± 2.2 23.9 ± 2.2 0.004
Open in a new tab

Abbreviations: BMI, body mass index; HOA, hand osteoarthritis.

Figure 3.

Figure 3

Open in a new tab

Representative radiographs of the hands of patients with HOA (right panel) and control subjects (left panel). Yellow arrows indicate Heberden's nodes.

Figure 4.

Figure 4

Open in a new tab

Manhattan plot of the microarray.

Table 2.

Top 100 CpG sites associated with HOA with low adjusted p Values.

CpG p Value Adjusted p Value Control average HOA average deltaBeta Target gene
1 cg19375418 3.43801E‐07 0.190729504 0.948858425 0.70373243 −0.243111117
2 cg05267394 9.71997E‐07 0.190729504 0.236278217 0.388143409 0.152001086 ASIP
3 cg16440058 9.86111E‐07 0.190729504 0.226923338 0.396595836 0.16981009 ASIP
4 cg15974673 1.03795E‐06 0.190729504 0.088616445 0.273826348 0.185281651 EVI5L
5 cg20103036 1.74717E‐06 0.256843145 0.779340282 0.665014794 −0.111543119 SBF2‐AS1
6 cg13883027 2.42328E‐06 0.2968617 0.745973578 0.107225327 −0.636078745
7 cg22184818 3.07955E‐06 0.323364469 0.790220995 0.696229161 −0.091547348 TTC29
8 cg07790718 4.75431E‐06 0.367113397 0.708238738 0.608591621 −0.097165584 MNT
9 cg09343458 4.92412E‐06 0.367113397 0.662518562 0.55234623 −0.107451806 MNT
10 cg20382790 5.30746E‐06 0.367113397 0.920912809 0.575681833 −0.342949128
11 cg15641364 5.51577E‐06 0.367113397 0.07500366 0.219215395 0.144258255 TAGLN2
12 cg02220129 6.20459E‐06 0.367113397 0.723156087 0.829437899 0.109437316
13 cg14221852 6.52305E‐06 0.367113397 0.64802679 0.749281031 0.104422021 KIAA0226
14 cg27256213 6.99239E‐06 0.367113397 0.687881628 0.588033253 −0.096747253 VSTM2B
15 cg04035553 7.53378E‐06 0.368476117 0.339182772 0.557934286 0.22143624
16 cg09328330 8.04137E‐06 0.368476117 0.162223119 0.061028938 −0.101180404 LOC728323
17 cg19650697 9.33975E‐06 0.368476117 0.284351826 0.429689213 0.148888958 RBM42
18 cg21756147 9.4814E‐06 0.368476117 0.412019628 0.547563582 0.138752154 DLX3
19 cg26003873 9.82019E‐06 0.368476117 0.115895908 0.032826506 −0.080332016
20 cg24790788 1.09351E‐05 0.368476117 0.408212004 0.535680993 0.130897466 PLOD2
21 cg25843713 1.20142E‐05 0.368476117 0.848562655 0.934006293 0.087947126 CCBE1
22 cg18561976 1.23899E‐05 0.368476117 0.462670267 0.296353272 −0.163125608 ICOS
23 cg04091927 1.24373E‐05 0.368476117 0.615579664 0.723290819 0.1114096
24 cg15528852 1.43409E‐05 0.368476117 0.83019154 0.748027004 −0.079628466
25 cg13689756 1.46255E‐05 0.368476117 0.389935271 0.495471541 0.110006752 NSMAF
26 cg16866321 1.50046E‐05 0.368476117 0.532905507 0.364280134 −0.166035777
27 cg14637885 1.5094E‐05 0.368476117 0.521161305 0.329888098 −0.18791353
28 cg02760300 1.76219E‐05 0.368476117 0.93405041 0.861108903 −0.072914614 ODZ4
29 cg10179547 1.95771E‐05 0.368476117 0.805200758 0.677649335 −0.125734457 PKIG
30 cg16183741 2.06492E‐05 0.368476117 0.837323893 0.926709296 0.091777836 TBCD
31 cg00792091 2.143E‐05 0.368476117 0.722009482 0.62767114 −0.091348819
32 cg18648427 2.15281E‐05 0.368476117 0.813857608 0.734773475 −0.076966459 LINC01207
33 cg15391397 2.216E‐05 0.368476117 0.697757272 0.591877539 −0.103278433 LOC100507002
34 cg09208311 2.23372E‐05 0.368476117 0.482158164 0.554257908 0.075450063 KIRREL3
35 cg01485599 2.32933E‐05 0.368476117 0.938070612 0.845391674 −0.090180942 ARHGEF28
36 cg12841860 2.33115E‐05 0.368476117 0.811564038 0.87220098 0.063686544 NHP2L1
37 cg20118822 2.39915E‐05 0.368476117 0.826700182 0.908661832 0.084529313 SYNPO2
38 cg22933195 2.40948E‐05 0.368476117 0.804578747 0.678459411 −0.123711721
39 cg05057634 2.40968E‐05 0.368476117 0.09959544 0.181514312 0.085325937 MGAT5B
40 cg22859267 2.49128E‐05 0.368476117 0.840193811 0.593108213 −0.244814841 LOC643339
41 cg22891191 2.49826E‐05 0.368476117 0.529118716 0.408727174 −0.117101324 PTPRE
42 cg16868253 2.49862E‐05 0.368476117 0.774604731 0.683765326 −0.089101359 BMPER
43 cg08136221 2.7823E‐05 0.368476117 0.813115703 0.698083001 −0.112479924 GEFT
44 cg22866430 2.78334E‐05 0.368476117 0.838825709 0.906660389 0.070541851 SNORA59A
45 cg15018066 2.79262E‐05 0.368476117 0.789026383 0.701054591 −0.085916274
46 cg08178168 2.80121E‐05 0.368476117 0.548392899 0.451968427 −0.096463795
47 cg16321846 2.80216E‐05 0.368476117 0.547257863 0.671203554 0.127782271 SDK1
48 cg14313918 3.05535E‐05 0.368476117 0.730831611 0.578032864 −0.150160999 LMLN
49 cg03356492 3.12126E‐05 0.368476117 0.450744912 0.627827376 0.180984657 BRUNOL4
50 cg15426035 3.14004E‐05 0.368476117 0.265966357 0.403448068 0.141160412 GAPDHS
51 cg07822980 3.22088E‐05 0.368476117 0.801910768 0.673167845 −0.125690986
52 cg22243466 3.22177E‐05 0.368476117 0.82795042 0.733495152 −0.092183211 ANK2
53 cg02322492 3.24221E‐05 0.368476117 0.694472013 0.619574122 −0.071981604 TMEM72‐AS1
54 cg05468156 3.28342E‐05 0.368476117 0.727390428 0.546029789 −0.178139893
55 cg25695450 3.29866E‐05 0.368476117 0.19190609 0.11860012 −0.070002873
56 cg19645298 3.3399E‐05 0.368476117 0.232636744 0.383567694 0.15390643 C8orf47
57 cg11594299 3.34087E‐05 0.368476117 0.73517337 0.802910492 0.070777651 RADIL
58 cg18647039 3.41299E‐05 0.368476117 0.513238012 0.599097218 0.088797348 DNAJC18
59 cg01306563 3.45869E‐05 0.368476117 0.431874288 0.32488497 −0.103850094 FOXP2
60 cg01875467 3.62019E‐05 0.368476117 0.686620263 0.571306675 −0.11246392 CSF1R
61 cg26504421 3.68988E‐05 0.368476117 0.665641783 0.447331642 −0.215519013 SDK2
62 cg13593427 3.74308E‐05 0.368476117 0.889457618 0.807011758 −0.080200766
63 cg20955836 3.76782E‐05 0.368476117 0.267108458 0.501257056 0.237324889 BMP7
64 cg12224030 3.83257E‐05 0.368476117 0.103055783 0.182898818 0.083805919 DLX4
65 cg09764150 3.83927E‐05 0.368476117 0.275458798 0.202402711 −0.069265437 ESR2
66 cg04311403 3.87919E‐05 0.368476117 0.638968138 0.550153913 −0.085797091 EDARADD
67 cg06349780 3.89105E‐05 0.368476117 0.598430835 0.485593361 −0.109640396
68 cg20118157 3.9059E‐05 0.368476117 0.903205361 0.813209002 −0.088005202 ATP13A5
69 cg26271127 3.91101E‐05 0.368476117 0.050268592 0.164147005 0.117359993 FUZ
70 cg09236831 3.91767E‐05 0.368476117 0.53937583 0.635787678 0.099391591 CXCR5
71 cg19536401 3.95726E‐05 0.368476117 0.366131802 0.174005239 −0.18866391
72 cg16208049 3.95735E‐05 0.368476117 0.431345913 0.526654512 0.098499068
73 cg19218509 3.97309E‐05 0.368476117 0.476699033 0.57665747 0.099983256 ASIP
74 cg10740245 4.11818E‐05 0.368476117 0.507112106 0.436186817 −0.068328228
75 cg14683916 4.16062E‐05 0.368476117 0.781602406 0.686441417 −0.092696682 ZNF717
76 cg01289141 4.22999E‐05 0.368476117 0.91358133 0.842180695 −0.069320812 NCR2
77 cg09736490 4.23957E‐05 0.368476117 0.158320147 0.373462035 0.220039238
78 cg16494747 4.25066E‐05 0.368476117 0.66214923 0.578972674 −0.080783998 CDH13
79 cg19286989 4.28394E‐05 0.368476117 0.905815642 0.818465147 −0.087133237 SDK1
80 cg18757016 4.34418E‐05 0.368476117 0.601098819 0.748160856 0.150450028
81 cg26408858 4.37717E‐05 0.368476117 0.567297065 0.480884867 −0.083567327
82 cg14337614 4.41929E‐05 0.368476117 0.418421189 0.305198263 −0.109713782
83 cg00105512 4.42582E‐05 0.368476117 0.543875681 0.442423933 −0.098937338 SYCE1L
84 cg12845650 4.45126E‐05 0.368476117 0.828743645 0.658699772 −0.167506962
85 cg20811856 4.4571E‐05 0.368476117 0.174108626 0.052135829 −0.121929256 TMEM51
86 cg05984439 4.51202E‐05 0.368476117 0.94105597 0.800296382 −0.138620096
87 cg05245822 4.53773E‐05 0.368476117 0.681246048 0.58511038 −0.093221162
88 cg07638857 4.53914E‐05 0.368476117 0.24424127 0.177075599 −0.06350555 ANKS1B
89 cg20683108 4.59107E‐05 0.368476117 0.73807376 0.647220816 −0.088298918
90 cg14028115 4.59182E‐05 0.368476117 0.720438567 0.561269115 −0.156542365 NOSTRIN
91 cg10144336 4.59398E‐05 0.368476117 0.295695471 0.362050823 0.070439915
92 cg23357813 4.64669E‐05 0.368476117 0.697611149 0.784054907 0.090225861
93 cg13242768 4.73602E‐05 0.368476117 0.375829712 0.205941498 −0.166729265 LOC101928851
94 cg20356323 4.90372E‐05 0.368476117 0.920650149 0.765389258 −0.152963039
95 cg14472052 4.91658E‐05 0.368476117 0.791471207 0.706957882 −0.082131223 NRP2
96 cg04148237 4.93272E‐05 0.368476117 0.952072447 0.68451775 −0.265727955
97 cg18375068 4.96461E‐05 0.368476117 0.636201303 0.404317209 −0.228846942
98 cg05439665 5.00723E‐05 0.368476117 0.666877151 0.578127776 −0.088776663 MCF2L
99 cg02869243 5.18312E‐05 0.368476117 0.689760405 0.60279294 −0.083649107 SERPINA12
100 cg09560062 5.26765E‐05 0.368476117 0.91185504 0.733141371 −0.176781244
Open in a new tab

Abbreviation: HOA, hand osteoarthritis.

Figure 5.

Figure 5

Open in a new tab

Pyrosequencing validation for two specific CpG sites. P represents the unadjusted values.

Table 3.

Least square means of DNA methylation between HOA and Control.

Control HOA p Valuea
LSmeansa SE LSmeana SE
BMP7 16.8 2.7 25.1 1.9 0.037
SBF2‐AS1 58.8 2.0 55.4 1.4 0.232
PLOD2 33.0 3.6 33.5 2.5 0.924
ICOS 25.8 2.6 29.4 1.8 0.319
CSF1R 79.8 2.0 78.7 1.4 0.701
Open in a new tab
a

Adjusted for age, gender, and BMI.

Abbreviation: HOA, hand osteoarthritis.

4. DISCUSSION

In this study, we compared the DNA methylation of leukocytes between subjects with and without HOA using both DNA methylation microarray and pyrosequencing analyses. Notably, we observed a significant increase in the methylation of BMP7 in patients with HOA.

Although several biomarkers for osteoarthritis have been reported, including cartilage oligomeric matrix protein, collagen type II alpha 1 (col2‐1), and col2‐1 NO2, C‐terminal cross‐linked telopeptide of type II collagen, CXC motif chemokine ligand 12/CXC chemokine receptor 2, and matrix metalloproteinases (MMPs), 15 , 16 , 17 , 18 , 19 , 20 , 21 in serum and synovial fluid, none have been established as clinically useful biomarkers. In the context of HOA, serum interleukin‐1 levels have been linked to hand function and radiological joint damage. 22 Additionally, several studies have explored the potential roles of DNA methylation in specific genes associated with osteoarthritis. 23 Notably, 12 methylation sites from various genes, including Meis homeobox 1, GABAA receptor γ3, retinoid X receptor α, and engrailed homeobox 1, have been previously identified as candidate biomarkers of methylation in the cartilage of patients with osteoarthritis. 24 A previous pilot study indicated that peripheral blood DNA methylation models could serve as potential biomarkers for knee osteoarthritis progression. 7 However, to the best of our knowledge, no previous studies have specifically focused on DNA methylation as a biomarker for HOA. The present study is the first to identify significantly increased methylation at specific DNA sites within leukocytes of patients with HOA. Previous epidemiological studies have associated DNA methylation in leukocytes with various diseases, including cancers, schizophrenia, metabolic syndrome, and myelodysplastic syndrome. 25 In addition, peripheral blood DNA methylation models have been also3reported to have the potential as predictors of knee OA. 7 Nevertheless, given the influence of diverse environmental factors on DNA methylation, 26 thorough evaluation is necessary to establish relevant biomarkers. In our comprehensive analysis, we used leukocytes from patients with so‐called “pure HOA” confirmed to have no history of other diseases and no osteoarthritis in the knee, hip, or spine. Despite the limited sample size, this focused approach enhances the specificity of our findings to HOA. In our validation analysis, a subset of methylation array results was verified, but the sample size remained limited. Thus, further evaluation with a larger sample size is crucial for future investigations.

If the association between HOA and knee/hip osteoarthritis is valid, 2 the identified methylation sites in this study may serve as predictive factors for osteoarthritis in general. Identifying middle‐aged individuals with methylation in these sites through blood samples could enable the provision of lifestyle guidance and education on preventive methods targeting knee and hip osteoarthritis. 27 This, in turn, could potentially contribute to the prevention of these diseases.

The observed significantly higher methylation rate of BMP7 (also known as osteogenic protein‐1) in the leukocyte DNA of patients with HOA aligns with BMP7's known role in mesenchymal cell differentiation into osteoblasts and chondrocytes. 13 BMP7's anabolic properties in cartilage homeostasis maintenance through MMP13 expression reduction and stimulation of proteoglycan synthesis have been established in fundamental research. 28 , 29 , 30 , 31 In a Phase 1 clinical trial, BMP7 was administered via intra‐articular injection to patients with severe knee osteoarthritis, with no reported adverse events. 32 Moreover, the severity of knee osteoarthritis has been correlated with the concentration of BMP7 in both serum and synovial fluid. 33 Therefore, based on our current findings, it is hypothesized that in HOA patients, an increase in DNA methylation levels of BMP7 may lead to systemic decreased expression of BMP7, thus reducing its protective function against OA. Although our study does not directly indicate that BMP7 could be a potential treatment or biomarker for human osteoarthritis, it supports its potential role in osteoarthritis, particularly in HOA.

This study has several limitations, including a modest sample size and potential differences in background characteristics between cases and controls, despite matching subjects and statistically adjusting for specific confounding factors. In particular, there were significant differences between the two groups in terms of BMI, which has been reported to be associated with HOA. 34 In these analyses, we could not completely eliminate the influence of BMI. Second, only a limited number of HOA‐specific methylation sites identified using the methylation microarray were confirmed through validation analysis. Third, although this study identified only a single methylation pattern difference, it is possible that a combination of methylation differences in several genes may be involved in HOA. Fourth, although the American College of Rheumatology criteria were used for the diagnosis of HOA in this study, HOA is heterogeneous disease, with different characteristics and symptoms varying among patients. Ideally, selection should have been based on these criteria along with a more comprehensive set of perspectives. For these reasons, future research with larger cohorts is necessary for robust validation based on our findings. Nevertheless, this study is the first to identify potential biomarkers of DNA methylation for HOA. Moreover, our results may provide insights into not only HOA but also the broader mechanisms of osteoarthritis, including epigenetic aspects.

In conclusion, this study revealed a significantly higher frequency of methylation at a specific BMP7‐related site in the leukocytes of patients with HOA. The site's DNA methylation may emerge as a potential biomarker for HOA and, potentially, for osteoarthritis as a whole. We propose that measuring methylation levels at the CpG sites identified in this study could serve as a valuable approach to prevent future progression of osteoarthritis.

AUTHOR CONTRIBUTIONS

Takashi Kuroiwa conducted experiments, performed the statistical analysis, interpret the data, and drafted the manuscripts. Yoshiki Tsuboi, Takehiro Michikawa, and Koji Suzuki conducted the experiments, performed the statistical analysis, and revised the manuscript. Kaori Tajima, Yuki Uraya, Atsushi Maeda, Kanae Shizu, and Yusuke Kawano conducted the experiments and revised the manuscript. Katsuji Suzuki secured funding, helped interpret the data, and revised the manuscript. Nobuyuki Fujita conceived the study, participated in its design and coordination, drafted the manuscript. All authors made the final approval of the article.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENT

This study received approval from our institution's ethics committee, and all patients provided written informed consent for use of their clinical data in accordance with the hospital's ethics guidelines. All methods used in the study adhered to the guidelines of the Declaration of Helsinki.

Kuroiwa T, Tsuboi Y, Michikawa T, et al. DNA methylation of bone morphogenetic protein 7 in leukocytes as a possible biomarker for hand osteoarthritis: a pilot study. J Orthop Res. 2025;43:84‐93. 10.1002/jor.25963

REFERENCES

  • 1. Wan J, Qian X, He Z, Zhu Z, Cheng P, Chen A. Epidemiological trends of hand osteoarthritis from 1990 to 2019: estimates from the 2019 global burden of disease study. Front Med. 2022;9:922321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Prieto‐Alhambra D, Judge A, Javaid MK, Cooper C, Diez‐Perez A, Arden NK. Incidence and risk factors for clinically diagnosed knee, hip and hand osteoarthritis: influences of age, gender and osteoarthritis affecting other joints. Ann Rheum Dis. 2014;73:1659‐1664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Taylor P, Gartemann J, Hsieh J, Creeden J. A systematic review of serum biomarkers anti‐cyclic citrullinated peptide and rheumatoid factor as tests for rheumatoid arthritis. Autoimmune Dis. 2011;2011:815038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Diem SJ, Joffe H, Larson JC, et al. Effects of escitalopram on markers of bone turnover: a randomized clinical trial. J Clin Endocrinol Metab. 2014;99:E1732‐E1737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Yousefi PD, Suderman M, Langdon R, Whitehurst O, Davey Smith G, Relton CL. DNA methylation‐based predictors of health: applications and statistical considerations. Nat Rev Genet. 2022;23:369‐383. [DOI] [PubMed] [Google Scholar]
  • 6. Waheed A, Rai MF. Osteoarthritis year in review 2023: genetics, genomics, and epigenetics. Osteoarthritis Cartilage. 2024;32:128‐137. [DOI] [PubMed] [Google Scholar]
  • 7. Dunn CM, Nevitt MC, Lynch JA, Jeffries MA. A pilot study of peripheral blood DNA methylation models as predictors of knee osteoarthritis radiographic progression: data from the Osteoarthritis Initiative (OAI). Sci Rep. 2019;9:16880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Altman R, Alarcon G, Appelrouth D, et al. The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hand. Arthritis Rheum. 1990;33:1601‐1610. [DOI] [PubMed] [Google Scholar]
  • 9. Tsuboi Y, Yamada H, Munetsuna E, et al. Relationship between long interspersed nuclear element‐1 DNA methylation in leukocytes and dyslipidemia in the Japanese general population. J Atheroscler Thromb. 2018;25:1231‐1239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Tsezou A. Osteoarthritis year in review 2014: genetics and genomics. Osteoarthritis Cartilage. 2014;22:2017‐2024. [DOI] [PubMed] [Google Scholar]
  • 11. Revell PA, Matharu GS, Mittal S, Pynsent PB, Buckley CD, Revell MP. Increased expression of inducible co‐stimulator on CD4+ T‐cells in the peripheral blood and synovial fluid of patients with failed hip arthroplasties. Bone Joint Res. 2016;5:52‐60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Zeng J, Jiang X, Jiang M, Cao Y, Jiang Y. Bioinformatics analysis of hub genes as osteoarthritis prognostic biomarkers. Sci Rep. 2023;13:22894. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Chubinskaya S, Hurtig M, Rueger DC. OP‐1/BMP‐7 in cartilage repair. Int Orthop. 2007;31:773‐781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Rodriguez‐Fontenla C, Gonzalez A. Genética de la artrosis. Reumatología Clínica. 2015;11:33‐40. [DOI] [PubMed] [Google Scholar]
  • 15. Tseng S, Reddi AH, Di Cesare PE. Cartilage oligomeric matrix protein (COMP): a biomarker of arthritis. Biomark Insights. 2009;4:BMI.S645. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Kapoor M, Mahomed N. Osteoarthritis: Pathogenesis, Diagnosis, Available Treatments, Drug Safety, Regenerative and Precision Medicine. ADIS; 2015:1‐278. [Google Scholar]
  • 17. Ruiz‐Romero C, Blanco FJ. Proteomics role in the search for improved diagnosis, prognosis and treatment of osteoarthritis. Osteoarthritis Cartilage. 2010;18:500‐509. [DOI] [PubMed] [Google Scholar]
  • 18. Xu Q, Sun X, Shang X, Jiang H. Association of CXCL12 levels in synovial fluid with the radiographic severity of knee osteoarthritis. J Investig Med. 2012;60:898‐901. [DOI] [PubMed] [Google Scholar]
  • 19. Sherwood J, Bertrand J, Nalesso G, et al. A homeostatic function of CXCR2 signalling in articular cartilage. Ann Rheum Dis. 2015;74:2207‐2215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Garnero P. Cross sectional evaluation of biochemical markers of bone, cartilage, and synovial tissue metabolism in patients with knee osteoarthritis: relations with disease activity and joint damage. Ann Rheum Dis. 2001;60:619‐626. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Mehana ESE, Khafaga AF, El‐Blehi SS. The role of matrix metalloproteinases in osteoarthritis pathogenesis: an updated review. Life Sci. 2019;234:116786. [DOI] [PubMed] [Google Scholar]
  • 22. Roux CH, Foltz V, Maheu E, et al. MRI and serum biomarkers correlate with radiographic features in painful hand osteoarthritis. Clin Exp Rheumatol. 2016;34:991‐998. [PubMed] [Google Scholar]
  • 23. Reynard LN. Analysis of genetics and DNA methylation in osteoarthritis: what have we learnt about the disease? Semin Cell Dev Biol. 2017;62:57‐66. [DOI] [PubMed] [Google Scholar]
  • 24. Wu Z, Shou L, Wang J, Huang T, Xu X. The methylation pattern for knee and hip osteoarthritis. Front Cell Dev Biol. 2020;8:602024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Terry MB, Delgado‐Cruzata L, Vin‐Raviv N, Wu HC, Santella RM. DNA methylation in white blood cells: association with risk factors in epidemiologic studies. Epigenetics. 2011;6:828‐837. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Ryan J, Wrigglesworth J, Loong J, Fransquet PD, Woods RL. A systematic review and meta‐analysis of environmental, lifestyle, and health factors associated with DNA methylation age. J Gerontol A. 2020;75:481‐494. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Wellsandt E, Golightly Y. Exercise in the management of knee and hip osteoarthritis. Curr Opin Rheumatol. 2018;30:151‐159. [DOI] [PubMed] [Google Scholar]
  • 28. Stöve J, Schneider‐Wald B, Scharf HP, Schwarz ML. Bone morphogenetic protein 7 (bmp‐7) stimulates proteoglycan synthesis in human osteoarthritic chondrocytes in vitro. Biomed Pharmacother. 2006;60:639‐643. [DOI] [PubMed] [Google Scholar]
  • 29. Badlani N, Inoue A, Healey R, Coutts R, Amiel D. The protective effect of OP‐1 on articular cartilage in the development of osteoarthritis. Osteoarthritis Cartilage. 2008;16:600‐606. [DOI] [PubMed] [Google Scholar]
  • 30. Im HJ, Pacione C, Chubinskaya S, Van Wijnen AJ, Sun Y, Loeser RF. Inhibitory effects of insulin‐like growth factor‐1 and osteogenic protein‐1 on fibronectin fragment‐ and interleukin‐1β‐stimulated matrix metalloproteinase‐13 expression in human chondrocytes. J Biol Chem. 2003;278:25386‐25394. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Ripmeester EGJ, Caron MMJ, van den Akker GGH, et al. BMP7 reduces the fibrocartilage chondrocyte phenotype. Sci Rep. 2021;11:19663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Hunter DJ, Pike MC, Jonas BL, Kissin E, Krop J, McAlindon T. Phase 1 safety and tolerability study of BMP‐7 in symptomatic knee osteoarthritis. BMC Musculoskelet Disord. 2010;11:232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Honsawek S, Chayanupatkul M, Tanavalee A, et al. Relationship of plasma and synovial fluid BMP‐7 with disease severity in knee osteoarthritis patients: a pilot study. Int Orthop. 2009;33:1171‐1175. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Shah K, Cai H, Lane JCE, et al. Prognostic factors for finger interphalangeal joint osteoarthritis: a systematic review. Rheumatology. 2021;60:1080‐1090. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Orthopaedic Research are provided here courtesy of Wiley

RESOURCES