Abstract Although osteoclasts play crucial roles in the skeletal system, the mechanisms that underlie oxidative stress during osteoclastogenesis remain unclear. The transcription factor Nrf2 and its suppressor, Keap1, function as central mediators of oxidative stress. To further elucidate the function of Nrf2/Keap1-mediated oxidative stress regulation in osteoclastogenesis, DNA microarray analysis was conducted in this study using wild-type (WT), Keap1 knockout (Keap1 KO), and Nrf2 knockout (Nrf2 KO) osteoclasts. Principal component analysis showed that 403 genes, including Nqo1, Il1f9, and Mmp12, were upregulated in Keap1 KO compared with WT osteoclasts, whereas 24 genes, including Snhg6, Ccdc109b, and Wfdc17, were upregulated in Nrf2 KO compared with WT osteoclasts. Moreover, 683 genes, including Car2, Calcr, and Pate4, were upregulated in Nrf2 KO cells compared to Keap1 KO cells. Functional analysis by Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway analysis showed upregulated genes in Nrf2 KO osteoclasts were mostly enriched in oxidative phosphorylation. Furthermore, GeneMANIA predicted the protein–protein interaction network of novel molecules such as Rufy4 from genes upregulated in Nrf2 KO osteoclasts. Understanding the complex interactions between these molecules may pave the way for developing promising therapeutic strategies against bone metabolic diseases caused by increased osteoclast differentiation under oxidative stress. Keywords: osteoclast, oxidative stress, Nrf2, Keap1, transcriptome, gene ontology, GeneMANIA 1. Introduction Osteoclasts are multinucleated cells responsible for bone resorption and play crucial roles in physiological bone remodeling, pathological osteoporosis, rheumatoid arthritis, and periodontal diseases [[26]1,[27]2,[28]3]. Physiologically, the receptor activator of nuclear factor-kappa B ligand (RANKL) promotes osteoclast differentiation. Nuclear factor of activated T cells c1 (NFATc1) is an important transcription factor that promotes osteoclastogenesis [[29]4]. Inflammatory cytokines such as interleukin 1 (IL-1), IL-6, and TNF-α [[30]5], as well as lipopolysaccharides (LPS) and oxidative stress [[31]6], promote osteoclast differentiation. In an in vitro culture system using bone marrow cells from adult mice, osteoclasts matured in monocyte-macrophage lineage-derived progenitor cells grown with macrophage colony-stimulating factor (M-CSF) when RANKL was added [[32]7]. During osteoclast differentiation, there is an increase in the expression of carbonic anhydrase 2, calcitonin receptor, vacuolar ATPase, cathepsin K, Ocstamp, Dcstamp [[33]3], Oscar [[34]8], Steap4 [[35]9], and Rab38 [[36]10]. Nuclear factor erythroid 2-related factor 2 (Nrf2) is a crucial transcription factor involved in antioxidant responses [[37]11]. Physiologically, Nrf2 binds to its inhibitor, Keap1, promoting its degradation in a proteasome-dependent manner, maintaining low intracellular levels. Under oxidative stress, Keap1 is released from Nrf2, and Nrf2 translocates to the nucleus to upregulate cytoprotective genes [[38]12,[39]13]. A previous study using Nrf2-specific siRNA probes and Keap1 and Nrf2 expression plasmids demonstrated that the Keap1/Nrf2 axis regulates RANKL-dependent osteoclast differentiation [[40]14]. Moreover, natural compounds with Nrf2-activating ability, such as epigallocatechin gallate [[41]15], fisetin [[42]16], resveratrol [[43]17], and sulforaphane [[44]15,[45]18], inhibit osteoclast differentiation. Studies using bone marrow cells from Nrf2 knockout (Nrf2 KO) mice have reported enhanced osteoclast differentiation [[46]19,[47]20]. Because Keap1 knockout (Keap1 KO) mice with Nrf2 hyperactivation are juvenile lethal [[48]13], we previously used splenocytes from newborn mice instead of bone marrow cells as osteoclast progenitor cells, cultured them with M-CSF, and stimulated the cells with RANKL to determine whether osteoclast differentiation occurred. We observed that osteoclast differentiation was markedly suppressed in Keap1 KO cells [[49]21], which is particularly relevant to the finding that many Nrf2-activating natural compounds suppress osteoclast differentiation [[50]22]. However, the mechanisms underlying this suppression are not fully understood. Microarray analysis is a robust technique for identifying the genes and pathways involved in biological processes in an unbiased manner. In this study, we used gene array technology to identify changes in gene expression during the differentiation of mouse splenocyte-derived macrophages into osteoclasts. We analyzed the transcriptome of RANKL-stimulated wild-type (WT), Nrf2 KO, and Keap1 KO cells to identify novel osteoclast regulators and elucidate the mechanisms underlying osteoclast differentiation via the Nrf2/Keap1 system. Here, we report all transcripts discovered in a microarray comparison of highly purified WT, Keap1 KO, and Nrf2 KO osteoclasts stimulated with RANKL. When compared with Keap1 KO cells, Nrf2 KO osteoclasts generally showed higher expression of osteoclast marker genes such as Oscar, Calcr, Mmp9, Acp5, Ctsk, and Car2. Apart from the above-mentioned osteoclastogenic genes, whose expression is upregulated by Nrf2 KO, the relationship of the following mitochondria-related genes, including Ndhfs, Cox, and Sdhc, as well as novel genes, with osteoclastogenesis is not yet clear. 2. Materials and Methods 2.1. Reagents RANKL was prepared according to a previously described method using an expression vector containing the hexahistidine-tagged human soluble RANKL extracellular domain, kindly provided by Dr. H. Amano (Tokyo Medical and Dental University, Tokyo, Japan) [[51]23]. In brief, Codon-Plus BL21 (DE3) RL-competent cells were used to express recombinant RANKL, which was then purified using Ni-NTA column chromatography. Contaminated LPS was removed by phase separation using Triton X-114, resulting in a final LPS concentration below the detection limit (1 pg/μg protein). In an osteoclastogenesis assay using mouse bone marrow cells, the purified recombinant RANKL protein demonstrated bioactivity comparable to human soluble RANKL (PeproTech EC, London, UK). 2.2. Mice Breeding pairs of Nrf2^−/− mice (Nrf2 KO) (RBRC01390) [[52]24] were obtained by RIKEN BRC through the National Bio-Resource Project of MEXT, Japan (Tsukuba, Japan). Keap1^+/− (RBRC01388) [[53]13] were provided by RIKEN BRC and were self-mated to generate Keap1 wild-type (WT) control mice and Keap1^−/− homozygous mice (Keap1 KO). The Animal Care and Use Committee of Nagasaki University Graduate School of Biomedical Sciences approved all animal experimental protocols (Approval Number: 210216169-2). All the experiments were conducted in accordance with the rules of animal experimentation at Nagasaki University, Japan. 2.3. Cell Culture Osteoclast precursors obtained from the spleen of neonatal mice were used for DNA microarray analysis as previously described [[54]21]. In brief, splenocytes were cultured in α-minimal essential medium supplemented with 10% fetal bovine serum, penicillin (100 U/mL), streptomycin (100 μg/mL), and amphotericin B (0.25 μg/mL) in the presence of macrophage colony-stimulating factor (M-CSF; 50 ng/mL) for 16 h at 37 °C under 5% CO[2]. Non-adherent cells were collected and cultured in M-CSF (50 ng/mL). After 72 h, the adherent splenic macrophages were cultured for three days with M-CSF (30 ng/mL) and RANKL (50 ng/mL) to produce osteoclasts. As previously described [[55]25], the cells were fixed with 4% paraformaldehyde on day three and stained for tartrate-resistant acid phosphatase (TRAP) activity to verify osteoclast formation. For bone marrow macrophage (BMM)-derived osteoclast formation, marrow cells from the femurs and tibias of 5-week-old male Nrf2 KO and WT mice were cultured using the method described above, in which osteoclasts were prepared from splenocytes. 2.4. Microarray Analysis Three days after the addition of RANKL, total RNA was extracted from WT, Nrf2 KO, and Keap1 KO cells using TRIzol Reagent (Thermo Fisher Scientific, Waltham, MA, USA) and subsequently purified using an RNeasy Mini Kit (QIAGEN, Tokyo, Japan) according to the manufacturer’s instructions. The RNA quality was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and a NanoDrop spectrophotometer (Thermo Fisher). The values of 260/280 were more than 2.00 for all samples. Total RNA (100 ng) was reverse transcribed using an Affymetrix GeneChip WT Plus Reagent Kit (Affymetrix, Santa Clara, CA, USA), labeled with biotin using a GeneChip WT Terminal Labeling Kit (Affymetrix), hybridized to a GeneChip Mouse Gene 2.0 ST Array (Affymetrix) using a GeneChip Hybridization, Wash, and Stain Kit (Affymetrix), and scanned according to the manufacturer’s instructions. Numerical data files of the three arrays of WT, Nrf2 KO, and Keap1 KO mice (CHP files obtained by the Robust Multichip Analysis algorithm on the Affymetrix Expression Console) were compared and analyzed using GeneSpring GX analysis software (version 7.3.1, Agilent Technologies, Palo Alto, CA, USA). For functional enrichment analysis, gene ontology (GO) analysis was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID; [56]https://david.ncifcrf.gov/, accessed on 24 October 2024), and pathway analysis was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG). Protein–protein interaction networks were predicted using GeneMANIA ([57]https://genemania.org/, accessed on 24 October 2024). Probes that had values of fold change (log) > 1 or (log) < −1 between the samples to be compared were considered statistically significant. 2.5. Quantitative Real-Time Polymerase Chain Reaction (RT-PCR) Analysis To verify the microarray data, Quantitative RT-PCR (qRT-PCR) analysis was performed using Quantstudio 3 (Thermo Fisher Scientific) according to the manufacturer’s instructions. Briefly, total RNA was extracted with TRIzol Reagent and reverse transcription was performed using oligo(dT)15 primer (Promega, Madison, WI, USA) and Revertra Ace (Toyobo, Osaka, Japan). Complementary DNA was amplified using Brilliant III Ultra-Fast SYBR Green QPCR Master Mix (Agilent Technologies). The primer sets used are listed in [58]Supplementary Table S1. All samples were assayed in triplicate, and the relative mRNA expression was normalized to that of the housekeeping gene β-actin. Data are presented as the mean ± standard deviation (SD) from three independent experiments. Statistical significance was determined using Student’s t-test, and differences were considered statistically significant at * p < 0.05 and ** p < 0.01. 3. Results 3.1. Identification of Differentially Expressed Genes We previously demonstrated that in the presence of RANKL, Keap1 KO splenic macrophages failed to differentiate into osteoclasts. However, compared to WT cells, Nrf2 KO splenic macrophages produced significantly higher osteoclasts ([59]Figure 1A). To elucidate the role of the Nrf2/Keap1 system during osteoclast differentiation, we performed a comparative analysis of gene expression in the following three combinations to identify the genes that fluctuate in relation to the system. 1. Keap1 KO vs. WT; 2. Nrf2 KO vs. WT; 3. Nrf2 KO vs. Keap1 KO. Figure 1. [60]Figure 1 [61]Open in a new tab Microarray analysis of WT, Nrf2 KO, and Keap1 KO cells. (A) Splenic macrophages from WT, Keap1 KO, and Nrf2 KO mice were cultured with 30 ng/mL M-CSF and 50 ng/mL RANKL for three days, followed by TRAP staining. Representative photographs showing red-colored osteoclasts. (a) WT, (b) Keap1 KO, and (c) Nrf2 KO mice. Scale bars: 100 μm. (B) Splenic macrophages from two mice each of WT, Keap1 KO, and Nrf2 KO were cultured with 30 ng/mL M-CSF and 50 ng/mL RANKL for three days, and RNA was collected from each cell for DNA microarray analysis (single microarray analysis for each cell). Graphs showing scatter plots of (a) Keap1 KO cells vs. WT osteoclasts, (b) Nrf2 KO osteoclasts vs. WT osteoclasts, and (c) Nrf2 KO osteoclasts vs. Keap1 KO cells. Green lines indicate log[2]2 or log[2]0.5. Principal component analysis (PCA) results showed that the gene expression patterns in Keap1 KO cells differed from those in WT and Nrf2 KO osteoclasts, whereas the expression patterns in Nrf2 KO osteoclasts were similar to those in WT osteoclasts. This may be because Keap1 KO cells did not form osteoclasts, unlike WT and Nrf2 KO cells ([62]Figure 1B). Among the 34,390 genes, 403 were upregulated by at least 2-fold in Keap1 KO cells than in WT osteoclasts ([63]Supplementary Table S2), and 553 were downregulated by less than 0.5-fold ([64]Supplementary Table S3). The top 20 genes, including NAD(P)H dehydrogenase, quinone1 (Nqo1), interleukin 1 family member 9 (Il1f9), and matrix metallopeptidase 12 (Mmp12), and bottom 20 downregulated genes, including carbonic anhydrase 2 (Car2), calcitonin receptor (Calcr), and osteoclast-associated receptor (Oscar), are listed in [65]Table 1 and [66]Table 2. We validated the mRNA expression of these genes by qRT-PCR ([67]Figure 2 and [68]Figure 3). Consistent with the microarray results, significant upregulation of Nqo1, Il1f9, Mmp12, Slc39a4, Fabp7, Cxcl14, Gsta3, Rnf128, Ly6g, Tanc2, and Gclm was confirmed in Keap1 KO compared to WT ([69]Figure 2), while Calcr, Scin, Ctsk, Pate4, Ocstamp, Ccr3, Tm4sf19, and Steap4 were significantly decreased ([70]Figure 3). Table 1. Top 20 upregulated genes in Keap1 KO/WT. Gene Symbol Description Fold Change (log) Fold Change Nqo1 NAD(P)H dehydrogenase, quinone1 4.539824 23.260722 Il1f9 Interleukin 1 family, member 9 3.949011 15.444399 A5 30064D06Rik RIKENcDNAA530064D06gene 3.487380 11.215177 Mmp12 Matrix metallopeptidase 12 3.240210 9.449317 Slc39a4 Solute carrierfamily39, member 4 3.164959 8.969074 Fabp7 Fatty acid binding protein7, brain 3.153433 8.897708 Cxcl14 Chemokine (C-X-C motif) ligand14 3.124515 8.721130 Gsta3 Glutathione S-transferase, alpha3 3.060010 8.339784 Prl2c2 Prolactinfamily2, subfamily c 2.981902 7.900270 Rnf128 Ring finger protein 128 2.961491 7.789289 Treml4 Triggering receptor expressed on myeloid cells-like 4 2.820159 7.062402 Ly6g Lymphocyte antigen 6 complex, locus G 2.7996016 6.962482 Car3 Carbonic anhydrase 3 2.7337945 6.652029 Tanc2 Tetra tri co peptide repeat 2.7162542 6.571643 Gclm Glutamate-cysteine ligase, modifier subunit 2.683442 6.423867 Cd36 CD36antigen 2.649393 6.274032 Ceacam10 Carcinoembryonic antigen-related cell adhesion molecule 10 2.6434088 6.248062 Gsta4 Glutathione S-transferase, alpha 4 2.447667 5.455330 Mfsd6 Major facilitator superfamily domain containing 6 2.433969 5.403780 Itga8 Integrin alpha 8 2.415440 5.334820 [71]Open in a new tab Table 2. Bottom 20 downregulated genes in Keap1 KO/WT. Gene Symbol Description Fold Change (log) Fold Change Car2 Carbonic anhydrase 2 −6.374898 0.012049 Calcr Calcitonin receptor −6.056974 0.015020 Oscar Osteoclast associated receptor −5.767575 0.018356 Scin Scinderin −5.017828 0.030866 Ctsk Cathepsin K −4.974964 0.031797 Akr1c18 Aldo-keto reductase family 1, member C18 −4.904193 0.033396 Pate4 Prostate and testis expressed 4 −4.870079 0.034195 Ocstamp Osteoclast stimulatory transmembrane protein −4.830698 0.035141 Ccr3 Chemokine (C-C motif) receptor 3 −4.745497 0.037279 Tm4sf19 Transmembrane 4L six family member 19 −4.693895 0.038636 Slc9b2 Solute carrier family 9, subfamily B −4.684105 0.038899 Akap6 A kinase (PRKA) anchor protein 6 −4.575950 0.041928 Steap4 STEAP family member 4 −4.502150 0.044128 Cd200 CD200 antigen −4.479137 0.044838 Dcstamp Dendrocyte expressed seven transmembrane protein −4.210025 0.054033 Trav9d-3 T cell receptor alpha variable 9D-3 −4.199091 0.054444 Atp6v0d2 ATPase, H+ transporting, lysosomal V0 subunit D2 −4.196109 0.054556 Myo1d Myosin I D −4.178478 0.055227 Adck3 Aar F domain containing kinase 3 −3.985752 0.063120 Rasgrp1 RAS guanyl releasing protein 1 −3.979324 0.063402 [72]Open in a new tab Figure 2. [73]Figure 2 [74]Open in a new tab Validation of microarray data. Upregulated genes in [75]Table 1 were confirmed by qRT-PCR. The relative mRNA levels of Nqo1, Il1f9, Mmp12, Slc39a4, Fabp7, Cxcl14, Gsta3, Rnf128, Ly6g, Tanc2, and Gclm in Keap1 KO were confirmed. Data are presented as the mean ± SD from three independent experiments (* p < 0.05 and ** p < 0.01). Figure 3. [76]Figure 3 [77]Open in a new tab Validation of microarray data. Downregulated genes in [78]Table 2 were confirmed by qRT-PCR. The relative mRNA levels of Calcr, Scin, Ctsk, Pate4, Ocstamp, Ccr3, Tm4sf19, and Steap4 in Keap1 KO were confirmed. Data are presented as the mean ± SD from three independent experiments (* p < 0.05 and ** p < 0.01). Among the 34,390 genes, 24 genes were upregulated by at least 2-fold ([79]Supplementary Table S4) and 112 genes were downregulated in Nrf2 KO osteoclasts compared with WT osteoclasts ([80]Supplementary Table S5). The top 20 upregulated genes, including small nuclear RNA host gene 6 (Snhg6), coiled-coil domain containing 109 B (Ccdc109b), and WAP four-disulfide core domain 17 (Wfdc17), and the bottom 20 downregulated genes, including cathepsin E (Ctse), interferon activated gene 202 B (Ifi202b), malic enzyme 1, NADP(+)-dependent, and cytosolic (Me1), are listed in [81]Table 3 and [82]Table 4. We validated the mRNA expression of these genes by qRT-PCR ([83]Figure 4 and [84]Figure 5). The qRT-PCR results showed a significant upregulation of Snhg6 and Ccdc109b genes in splenic macrophage-derived Nrf2 KO osteoclasts ([85]Figure 4A). Since Nrf2 KO mice grow to adulthood, osteoclasts prepared from BMMs of 5-week-old mice similarly showed a significant increase in Snhg6, Ppbp, Wfdc17, and Ctsk, consistent with the microarray results ([86]Figure 4B). Moreover, a significant decrease in Ctse, Ifi202b, Me1, Cbr3, Thy1, Lrrc32, Rnf128, Cxcl14, Slc7a11, and Nqo1 were confirmed in splenic macrophage-derived Nrf2 KO osteoclasts ([87]Figure 5). Table 3. Top 20 upregulated genes in Nrf2 KO/WT. Gene Symbol Description Fold Change (log) Fold Change Snhg6 Small nucleolar RNA host gene 6 1.870724 3.657160 Ccdc109b Coiled-coil domain containing 109 B 1.842320 3.585861 Wfdc17 WAP four-disulfide core domain 17 1.541663 2.911298 Mir1945 Micro RNA 1945 1.474677 2.779214 Ear1|Ear-ps9 Eosinophil-associated, ribonuclease A family, member1 1.378982 2.600848 1700049L16Rik Hematological and neurological expressed 1-like pseudogene 1.346564 2.543057 H2-Q6|H2-Q8 Histocompatibility 2, Q region locus 6 1.3387494 2.529320 Fam136b-ps Family with sequence similarity136, member B, pseudogene 1.301304 2.464516 H2-Ab1 Histocompatibility 2, classII antigen A, beta1 1.263337 2.400504 Igh-VJ558 Immunoglobulin heavy chain (J558 family) 1.254574 2.385967 Pf4 Pro-platelet basic protein 1.177933 2.262523 Luzp4 Leucine zipper protein 4 1.128823 2.186803 Ppbp Pro-platelet basic protein 1.121270 2.175384 Mir3093 microRNA3093 1.114681 2.165472 Trbc1|Tcrb-J T cell receptor beta, constant region 1 1.100635 2.144490 Syt10 Synaptotagmin X 1.081524 2.116271 Tcrg-V4 T cell receptor gamma, variable4 1.050665 2.104195 Saa3 Serum amyloid A3 1.050665 2.071484 Klrb1 Killer cell lectin-like receptor subfamily B member 1 1.031235 2.043773 Serpinb6b Serine (or cysteine) peptidase inhibitor, clade B, member 6b 1.021454 2.029964 [88]Open in a new tab Table 4. Bottom 20 downregulated genes in Nrf2 KO/WT. Gene Symbol Description Fold Change (log) Fold Change Ctse Cathepsin E −3.9513383 0.064644 Ifi202b Interferon activated gene 202B −3.260101 0.104379 Me1 Malic enzyme 1, NADP (+)-dependent, cytosolic −3.171412 0.110997 Cbr3 Carbonyl reductase 3 −2.686977 0.155289 Bgn Biglycan −2.376685 0.192551 Thy1 Thymus cell antigen 1, theta −2.071330 0.237940 Ppp2r1b Protein phosphatase 2, regulatory subunit A −2.040244 0.243123 Lrrc32 Leucine rich repeat containing 32 −2.038478 0.243420 Rnf128 Ring finger protein 128 −2.035712 0.243888 Ednrb Endothelin receptor type B −1.992971 0.251221 Cxcl14 Chemokine (C-X-C motif) ligand 14 −1.991431 0.251489 Loxl1 Lysyl oxidase-like 1 −1.920795 0.264109 Slc7a11 Solute carrier family 7 −1.908276 0.266411 Serpine1 Serine (or cysteine) peptidase inhibitor −1.894347 0.268995 Cnn1 Calponin 1 −1.877003 0.272249 Ddah1 Dimethyl arginine dimethyl amino hydrolase 1 −1.857229 0.276006 Des Desmin −1.813092 0.284580 Pydc3 Pyrin domain containing protein 3 −1.807523 0.285681 Ctgf Connective tissue growth factor −1.790487 0.289074 [89]Open in a new tab Figure 4. [90]Figure 4 [91]Open in a new tab Validation of microarray data. Upregulated genes in [92]Table 3 were confirmed by qRT-PCR. (A) Significant upregulation of Snhg6 and ccdc109b in Nrf2 KO osteoclasts derived from splenocyte were confirmed. Ppbp gene expression tended to increase. (B) Significant upregulation of Snhg6, Wfdc17, Ppbp, and Ctsk in Nrf2 KO osteoclasts derived from BMMs were confirmed. Data are presented as the mean ± SD from three independent experiments (* p < 0.05 and ** p < 0.01). Figure 5. [93]Figure 5 [94]Open in a new tab Validation of microarray data. Downregulated genes in [95]Table 4 were confirmed by qRT-PCR. Significant downregulation of Ctse, Ifi202b, Me1, Cbr3, Thy1, Lrrc32, Rnf128, Cxcl14, Slc7a11, and Nqo1 in Nrf2 KO osteoclasts were confirmed. Data are presented as the mean ± SD from three independent experiments (* p < 0.05 and ** p < 0.01). Among the 34,390 genes, 683 were upregulated by at least 2-fold in Nrf2 KO osteoclasts compared to Keap1 KO cells ([96]Supplementary Table S6), and 644 were downregulated in Nrf2 KO osteoclasts compared with Keap1 KO cells ([97]Supplementary Table S7). The top 20 upregulated genes, including carbonic anhydrase 2 (Car2), calcitonin receptor (Calcr), and prostate and testis expressed 4 (Pate4); and the bottom 20 downregulated genes, including NAD(P)H dehydrogenase, quinone 1 (Nqo1), cathepsin E (Ctse), and chemokine (C-X-C motif) ligand 14 (Cxcl14), are listed in [98]Table 5 and [99]Table 6. We validated the mRNA expression of these genes by qRT-PCR ([100]Figure 6 and [101]Figure 7). Consistent with the microarray results, significant upregulation of Calcr, Pate4, Oscar, Scin, Akr1c18, Ctsk, Steap4, Adck3, Tm4sf19, Atp6v0d2, and Ccr3 in Nrf2 KO osteoclasts were confirmed ([102]Figure 6), while significant downregulation of Nqo1, Ctse, Cxcl14, Rnf128, Me1, Mmp12, Slc39a4, Gclm, Slc7a11, Cbr3, and Fabp7 in Nrf2 KO osteoclasts were confirmed ([103]Figure 7). Table 5. The increased top 20 genes in Nrf2 KO/Keap1 KO. Gene Symbol Description Fold Change (log) Fold Change Car2 Carbonic anhydrase 2 6.611182 97.760632 Calcr Calcitonin receptor 5.738293 53.382416 Pate4 Prostate and testis expressed 4 5.416601 42.712931 Oscar Osteoclast associated receptor 5.297467 39.327512 Scin Scinderin 5.254921 38.184653 Akr1c18 Aldo-keto reductase family 1, member C18 5.221919 37.321084 Ctsk Cathepsin K 5.070246 33.596662 Ocstamp Osteoclast stimulatory transmembrane protein 4.729692 26.532560 Slc9b2 Solute carrier family 9, subfamily B 4.721196 26.376764 Cd200 CD200 antigen 4.569846 23.749842 Akap6 A kinase (PRKA) anchor protein 6 4.497392 22.586545 Steap4 STEAP family member 4 4.302776 19.736254 Adck3 aarF domain containing kinase 3 4.216301 18.588011 Dcstamp Dentrocyte expressed seven transmembrane protein 4.199798 18.376600 Tm4sf19 Transmembrane 4L six family member 19 4.109665 17.263643 Atp6v0d2 ATPase, H+ transporting, lysosomal V0 subunit D2 4.063965 16.725356 Rasgrp1 RAS guanyl releasing protein 1 3.950285 15.458039 Ccr 3 Chemokine (C-Cmotif) receptor 3 3.859836 14.518659 Trav9d-3 T cell receptor alpha variable 9D-3 3.780108 13.738075 [104]Open in a new tab Table 6. Bottom 20 downregulated genes in Nrf2 KO/Keap1 KO. Gene Symbol Description Fold Change (log) Fold Change Nqo1 NAD(P)H dehydrogenase, quinone 1 −6.131591 0.014263 Ctse Cathepsin E −5.380067 0.024013 Cxcl14 Chemokine (C-X-C motif) ligand 14 −5.115946 0.028837 Rnf128 Ring finger protein 128 −4.997204 0.031311 Me1 Malic enzyme 1, NADP (+)-dependent,cytosolic −4.568336 0.042150 Mmp12 Matrix metallo peptidase 12 −4.524167 0.043460 Slc39a4 Solute carrier family 39(zinc transporter) −4.320142 0.050062 Ednrb Endothelin receptor type B −4.040073 0.060788 Gclm Glutamate-cysteine ligase, modifier subunit −3.970209 0.063804 Tanc2 Tetra tri copeptide repeat −3.957151 0.064384 Slc7a11 Solute carrier family 7, member 11 −1.991431 0.251489 Cbr3 Carbonyl reductase 3 −3.728080 0.075463 Gatm Glycine amidino transferase −3.394802 0.095074 Fabp7 Fatty acid binding protein 7, brain −3.334516 0.099131 LOC100038947 Signal-regulatory protein beta1-like −3.313690 0.100573 Nlrp1c-ps NLR family, pyrin domain containing 1C, pseudogene −3.270247 0.103647 Lrrc32 Leucine rich repeat containing 32 −3.263585 0.104127 Gsta3 Glutathione S-transferase, alpha 3 −3.252826 0.104906 Gsta4 Glutathione S-transferase, alpha 4 −3.232622 0.106386 [105]Open in a new tab Figure 6. [106]Figure 6 [107]Open in a new tab Validation of microarray data. Upregulated genes in [108]Table 5 were confirmed by qRT-PCR. Significant upregulation of Calcr, Pate4, Oscar, Scin, Akr1c18, Ctsk, Steap4, Adck3, Tm4sf19, Atp6v0d2, and Ccr3 in Nrf2 KO osteoclasts were confirmed. Data are presented as the mean ± SD from three independent experiments (** p < 0.01). Figure 7. [109]Figure 7 [110]Open in a new tab Validation of microarray data. Downregulated genes in [111]Table 6 were confirmed by qRT-PCR. Significant downregulation of Nqo1, Ctse, Cxcl14, Rnf128, Me1, Mmp12, Slc39a4, Gclm, Slc7a11, Cbr3, and Fabp7 in Nrf2 KO osteoclasts were confirmed. Data are presented as the mean ± SD from three independent experiments (* p < 0.05 and ** p < 0.01). 3.2. GO Analysis of Nrf2/Keap1-Mediated Osteoclastogenesis GO analysis, using the DAVID showed that upregulated genes identified in Keap1 KO cells/WT osteoclasts were associated with the terms “response to stress”, “response to external stimulus”, “inflammatory response”, “response to bacterium”, and “defense response”, whereas downregulated genes were associated with the terms “mitochondrion”, “cytoplasmic part”, “cytoplasm”, “mitochondrial part”, and “mitochondrial inner membrane” ([112]Figure 8A). Figure 8. [113]Figure 8 [114]Figure 8 [115]Figure 8 [116]Open in a new tab GO enrichment analysis. (A) Up- or downregulated genes were analyzed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) for GO enrichment analysis in Keap1 KO cells compared with WT osteoclasts. (B) Up- or downregulated genes were analyzed using DAVID for GO enrichment analysis in Nrf2 KO osteoclasts compared with WT osteoclasts. (C) Up- or downregulated genes were analyzed using DAVID for GO enrichment analysis in Nrf2 KO osteoclasts compared with Keap1 KO cells. GO analysis showed that upregulated genes identified in Nrf2 KO osteoclasts/WT osteoclasts were associated with the terms “negative regulation of megakaryocyte differentiation”, “regulation of megakaryocyte differentiation”, “DNA replication-independent nucleosome organization”, “DNA replication-independent nucleosome assembly”, and “DNA replication-dependent nucleosome organization”, whereas downregulated genes were associated with the terms “extracellular matrix”, “proteinaceous extracellular matrix”, “tissue development”, “response to stress”, and “extracellular space” ([117]Figure 8B). GO analysis showed that upregulated genes identified in Nrf2 KO osteoclasts/Keap1 KO cells were associated the with terms “mitochondrial protein complex”, “cytoplasm”, “mitochondrial membrane part”, “respiratory chain”, “inner mitochondrial membrane protein complex”, “organelle envelope”, “mitochondria respiratory chain”, and “oxidoreductase complex”, whereas downregulated genes were associated with the terms “response to stress”, “regulation of multicellular organismal process”, “single-multicellular organism process”, “binding”, and “system development” ([118]Figure 8C). 3.3. KEGG Pathway Analysis of Nrf2/Keap1-Mediated Osteoclastogenesis The effect of the Nrf2/Keap1 system on osteoclastogenesis was also confirmed by KEGG analysis using DAVID for functional annotation of differentially expressed genes. Among the 683 upregulated genes with more than a 2-fold change in Nrf2 KO osteoclasts compared with Keap1 KO cells, 54 were associated with the oxidative phosphorylation pathway. Genes surrounded by red lines are significantly upregulated genes such as NADH:ubiquinone oxidoreductase core subunit (Ndhfs), cytochrome C oxidase (Cox), and succinate dehydrogenase complex (Sdhc) ([119]Figure 9A). These genes also coincided with those whose expression was reduced in Keap1 KO cells relative to WT osteoclasts. Figure 9. [120]Figure 9 [121]Figure 9 [122]Figure 9 [123]Open in a new tab KEGG pathway enrichment analysis. Compared to Keap1 KO cells, Nrf2 KO osteoclasts exhibited a marked increase in expression of genes (surrounded by red lines) involved in oxidative phosphorylation (A) and osteoclast differentiation (B), whereas marked decreased in expression of genes (surrounded by blue lines) involved in focal adhesion (C) and ECM–receptor interaction (D). Among the 683 upregulated genes with more than a 2-fold change in Nrf2 KO osteoclasts than in Keap1 KO cells, 13 were associated with the osteoclast differentiation pathway. Red asterisks indicate significantly upregulated genes, such as cathepsin K (Ctsk), calcitonin receptor (Calcr), and nuclear factor of activated T cells, cytoplasmic, calcineurin-dependent 1 (Nfatc1) ([124]Figure 9B). These genes were also consistent with those that were downregulated in Keap1 KO cells relative to WT osteoclasts. Moreover, 27 other genes, including integrin alpha V (Itgav), integrin beta3 (Itgb3), ATPase (Atp6v), and transporter 1 ATP-binding cassette (Tap1), whose expression is upregulated, are involved in the phagosome pathway. Interestingly, among the 644 downregulated genes with less than a 0.5-fold change in Nrf2 KO osteoclasts compared with Keap1 KO cells, 24 genes were associated with the focal adhesion pathway. Blue asterisks indicate significantly decreased expression of genes such as caveolin-1 (Cav1), integrin alpha1 (Itga1), protein kinase C, beta (Prkcb), calpain2 (Capn2), pavin alpha (Parva), and dedicator of cytokinesis1 (Dock1) ([125]Figure 9C). Among 644 downregulated genes with less than a 0.5-fold change in Nrf2 KO osteoclasts compared with Keap1 KO cells, 16 genes were associated with the ECM-receptor interaction pathway. Genes surrounded by blue lines are significantly decreased expression of genes such as collagen, type1, alpha1 (Col1a1), laminin B1 (Lamb1), fibronectin1 (Fn1), tenascin (Tnc), and nephronectin (Npnt) ([126]Figure 9D). 3.4. Protein–Protein Interaction Network Interactions between proteins encoded by the upregulated genes in Nrf2 KO osteoclasts relative to Keap1 KO cells were predicted using GeneMANIA, the generated protein association network, in which proteins are presented in the form of nodes, and different colored lines between nodes represent specific meanings. Each color code within a node reflects its function. Of the 683 genes whose expression was upregulated in Nrf2 KO osteoclasts, the nodes colored for mitochondria-related functions, such as the mitochondrial protein complex, mitochondrial inner membrane, respiratory chain complex, mitochondrial matrix, oxidative phosphorylation, mitochondrial transport, and mitochondrial gene expression, are shown in [127]Supplementary Figure S1. Similarly, nodes colored for osteoclast-related functions, such as bone remodeling, response to tumor necrosis factor, bone resorption, macrophage migration, proton-transporting V-type ATPase complex, regulation of bone resorption, and regulation of osteoclast differentiation, are shown in [128]Supplementary Figure S2. Interactions between proteins encoded by the top 40 upregulated genes in Nrf2 KO osteoclasts relative to Keap1 KO cells were predicted using GeneMANIA, and the following four genes: mucolipin 3 (Mcoln3); T-cell acute lymphocytic leukemia 2 (Tal2); potassium-inward centrally rectifying channel, subfamily J, member 2 (Kcnj2); and plexin D1 (Plxnd1) were revealed. ([129]Figure 10A). These four genes were not upregulated in the Nrf2 KO cells in the present study. Among the top 40 genes, Acp5, Adcy3, Atp6v0d2, Ctsk, Dcstamp, and Ocstamp were predicted to be coexpressed with Mcoln3. Moreover, Dcstamp, Jdp2, Mst1r, Serpind1, and Tnfrsf9 were predicted to be co-expressed with Tal2. Acp5, Adcy3, Car2, Ccr3, Jdp2, Ocstamp, and Oscar were predicted to be coexpressed with Kcnj2. Furthermore, Acp5, Adcy3, Akap6, Car2, Ccr3, Ctsk, Cyp2s1, Dcstamp, Slc9b2, Jdp2, and Ocstamp were predicted to be coexpressed with Plxnd1 ([130]Figure 10A). Moreover, some molecules were predicted to interact with the top 40 genes, such as matrix metallopeptidase 9 (Mmp9), solute carrier family 37 (Slc37a2), and G protein-coupled receptor 137 B (Gpr137b), which were significantly upregulated in Nrf2 KO osteoclasts ([131]Supplementary Table S6). Figure 10. [132]Figure 10 [133]Figure 10 [134]Open in a new tab Protein–protein interaction network analysis by GeneMANIA. (A) Predicted network of proteins that interact with proteins encoded by the top 40 upregulated genes in Nrf2 KO osteoclast against Keap1 KO cells. (B) Predicted network of proteins interacting with proteins encoded by genes upregulated in Nrf2 KO osteoclast compared with WT osteoclasts. Among the top 40 genes, Rufy4 was not predicted to interact with the other genes. Therefore, we reanalyzed the interaction between Rufy4 and all other 682 genes whose expression was upregulated in Nrf2 KO and predicted the interaction with arginine vasopressin-induced 1 (Avpi1), mitochondrial ribosomal protein L47 (Mrpl47), pleckstrin homology domain-containing family F (with FYVE domain) member 1 (Plekhf1), pleckstrin homology domain-containing family N member 1 (Plekhn1), and WD repeat and FYVE domain-containing 1 (Wdfy1). However, the functions of these genes in osteoclasts have not yet been reported. As shown in [135]Supplementary Table S4, no genes were upregulated more than 4-fold in Nrf2 KO osteoclasts relative to WT osteoclasts, likely because osteoclasts form in both Nrf2 KO and WT cells ([136]Figure 1A). Nrf2 KO cells exhibit enhanced osteoclastogenesis and bone resorption compared to WT cells [[137]21]. The genes that were upregulated in the current microarray suggested that they contributed to enhanced osteoclastogenesis in Nrf2 KO conditions. GeneMANIA predicted interactions between the 24 genes whose expression was upregulated more than 2-fold in Nrf2 KO cells and predicted that WAP four-disulfide core domain 17 (Wfdc17), histocompatibility 2, class II antigen A, beta1 (H2-Ab1), and platelet factor 4 (Pf4) interact with mitochondrial calcium uniporter dominant negative beta subunit (Mcub), triggering receptor expressed on myeloid cells-like 1 (Treml1), S100 calcium protein A9 (S100a9), and chemokine (C-C motif) receptor 2 (Ccr2), which are not included in the 24 genes ([138]Figure 10B). These molecules are involved in humoral immune response, myeloid leukocyte migration, cell chemotaxis, blood coagulation, and wound healing, suggesting that these functions might effectively induce osteoclastogenesis. Among the 24 genes, leucine zipper protein 4 (Luzp4) showed no interaction with the other genes ([139]Figure 10B). 4. Discussion In this study, we used WT, Nrf2 KO, and Keap1 KO neonatal mouse spleen cells as osteoclast progenitor cells and performed DNA microarray analysis of the RNA obtained from the cells three days after RANKL addition. The results obtained by PCA are shown in [140]Supplementary Tables S2–S7. Using these data, we performed GO enrichment analysis, KEGG pathway analysis, and constructed a protein–protein interaction network using GeneMANIA. Although various factors have been reported to regulate osteoclast differentiation, this study focused on Nrf2 and Keap1, which are essential for oxidative stress regulation, and comprehensively investigated genes whose expression was upregulated during RANKL-induced osteoclast differentiation. In this study, Nrf2 KO cells with accelerated osteoclast differentiation compared with Keap1 KO cells showed significant upregulation of mitochondria-related genes, including cytochrome C oxidase (Cox), mitochondrial ribosomal protein 12 (Mrpl12), and succinate dehydrogenase complex assembly factor 1 (Sdhaf1), as well as osteoclast differentiation marker genes, including carbonic anhydrase2 (Car2), integrin beta3 (Itgb3), osteoclast-associated receptor (Oscar), and dendrocyte-expressed seven transmembrane protein (Dcstamp). These results suggest that mitochondrial biogenesis was induced in Nrf2 KO osteoclasts. Previous studies have shown that cells and mitochondria isolated from Nrf2 KO mice have low respiration and ATP levels, whereas Keap1 knockout and knockdown mice have increased respiration and ATP levels [[141]26,[142]27]. Furthermore, mitochondrial fatty acid oxidation was decreased in mitochondria isolated from Nrf2-deficient mice, indicating that Nrf2 positively regulates mitochondrial biogenesis [[143]28]. However, these results are inconsistent with our findings of elevated expression of mitochondria-related genes in Nrf2 KO osteoclasts. The increased expression of peroxisome proliferative-activated receptor gamma coactivator 1-beta (Ppargc1b) [[144]29] in Nrf2 KO osteoclasts compared to that in Keap1 KO cells in the current study suggests that ROS, increased by Nrf2 KO, induced the phosphorylation of CREB and increased Ppargc1b, possibly resulting in increased mitochondrial biogenesis. Alternatively, because osteoclasts are mitochondria-rich cells, the comparison of Nrf2 KO, which promotes osteoclast formation, with Keap1 KO, in which only a few osteoclasts are formed, may have resulted in the higher expression of mitochondria-related genes, reflecting prominent osteoclast characteristics. Microarray analysis has led to the identification of genes with unknown functions in osteoclasts. Our microarray results showed that the expression of Rufy4, a novel autophagy regulator, was higher in Nrf2 KO osteoclasts. We also found that Rufy4 expression was upregulated during RANKL-induced osteoclast differentiation and that bone resorption was enhanced in osteoclasts overexpressing Rufy4 [[145]30]. Kim et al. reported that Rufy4 promotes extracellular secretion of cathepsin K and increases bone resorption [[146]31]. Moreover, qRT-PCR results showed that the gene expression of Snhg6, Ccdc109b, Wfdc17, and Ppbp was upregulated in Nrf2 KO osteoclasts compared with WT osteoclasts, but the roles of these genes in osteoclasts have not been clarified. However, the present GeneMANIA analysis predicted Mcub, Treml1, S100a9, and Ccr2 as genes that interact with Wfdc17 and Ppbp through protein–protein interactions. MCUb is the inhibitory subunit of the MCU complex. MCU is a mitochondrial inner membrane protein that transports cytosolic calcium into the mitochondrial matrix and reduces intracellular oxidative phosphorylation levels in Mcu-KO mice [[147]32]. The MCU inhibitor ruthenium red has been reported to inhibit RANKL-induced osteoclast differentiation and bone resorption [[148]33]. Treml1 has been reported to be a negative regulator of osteoclastogenesis [[149]34]. S100a9 repressed RANK expression and inhibited monocyte differentiation into osteoclasts. Ccr2 and its ligand Ccl2 have been reported to promote osteoclastogenesis [[150]35]. Wfdc17 and Ppbp have been suggested to function in osteoclasts via these genes. Bone metabolism is closely related to oxidative stress, and the Nrf2/Keap1 signaling axis regulates osteoclastogenesis. Therefore, bioactive natural compounds are considered to have potential Nrf2-activating effects and to reduce the risk of oxidative stress-related osteoporosis [[151]22]. Recent studies have reported that plant phenolic compounds such as 4-methylcatechol [[152]36], oroxylin A [[153]37], notopterol [[154]38], and tussilagone [[155]39] have inhibitory effects on osteoporosis by acting on the Nrf2/Keap1 signaling axis and reducing oxidative stress levels. It is unclear whether these natural compounds are involved in the expression of the genes identified in this study; however, further studies are expected in the future. The lack of Nrf2 can induce osteoclastogenesis; however, the role of Nrf2 in osteoblast differentiation is more complex. Our previous study revealed that during osteoblast differentiation using primary osteoblasts isolated from newborn mice calvaria cultured with ascorbic acid, β-glycerophosphate, and dexamethasone, expression levels of osteogenic essential genes such as Runx2, Sp7, and Alpl were significantly upregulated in Nrf2 KO cells. In contrast, the expression of these genes was suppressed by approximately 50% in Keap1 KO cells compared to WT cells. In co-culture experiments with primary osteoblasts from newborn mice and splenic macrophages, Keap1 KO macrophages show completely abolished osteoclastogenesis; however, Keap1 KO osteoblasts could support osteoclastogenesis. This is probably due to the fact that the expression of the osteogenic essential genes is not completely suppressed but is expressed even in half levels of the WT cells [[156]21]. Consistent with our results, Park et al. reported that the gene expression of ALP, Runx2, and Sp7 was significantly upregulated in Nrf2 KO osteoblasts [[157]40]. In other reports, Nrf2 KO mice showed lower bone mass and decreased bone formation rate [[158]41], and Nrf2-deficient osteoblasts exhibited impaired differentiation and mineralization [[159]42]. The role of Nrf2 in osteoblast differentiation and osteogenesis remains controversial. It may depend on the age of the mice and the level of oxidative stress in each experimental system. Loss of Keap1 promotes hyperactivation of Nrf2, resulting in juvenile lethality. Therefore, it was not possible to observe hard tissues of adult individuals of global Keap1 KO mice. Yoshida et al. used viable Keap1 KO mice by esophageal Nrf2 deletion in Keap1 KO mice and revealed impaired differentiation of both osteoclasts and osteoblasts [[160]43]. These studies suggest that the Nrf2/Keap1 signaling axis plays an important role in osteoblasts as well as osteoclastogenesis. 5. Conclusions The results of the current DNA microarray transcriptome analysis revealed elevated expression of novel genes, such as Rufy4, Avpi1, Mrpl47, Wfdc17, and H2-Ab1, in Nrf2 KO osteoclasts. It is expected that the data from this study will be further analyzed by researchers for their own purposes, resulting in the elucidation of a new regulatory mechanism of osteoclast differentiation and the development of therapeutic agents for diseases caused by excessive bone resorption by osteoclasts, such as osteoporosis, rheumatoid arthritis, and periodontal diseases. Acknowledgments