Regulation of osteoclastogenesis by lipoteichoic acid and adiponectin

국문 초록 (Abstract)

목적
골조직은 불필요한 골을 파괴시키는 파골세포와 새로운 골을 생성하는 조골세포의 균형에 의해 골 항상성이 유지된다. 골 흡수를 담당하는 파골세포는 선천성 면역세포인 대식세포로부터 분화하며, 파골세포의 비정상적 분화는 다양한 골질환의 발병과 밀접하게 연관되기 때문에 파골세포 분화를 조절하는 것은 골질환의 예방과 치료에 매우 중요하다. 파골세포의 분화는 골 미세환경에 존재하는 세포와 내인성 인자들 뿐만 아니라 미생물 또는 미생물 유래 인자들에 의해 영향을 받는다. 특히, 감염성 골질환의 주요 원인균인 그람양성세균은 세포 표면에 리포테이코산을 발현하는데 현재까지 파골세포 분화 조절에 리포테이코산이 미치는 영향에 대해서는 전혀 보고된 바 없다. 또한 대표적 골질환인 골다공증은 골량의 감소와 함께 골조직 내 지방세포가 증가되는 특성을 가지는데, 지방세포에서 주로 분비되는 아디포넥틴은 면역 및 대사과정에 중요한 역할을 한다고 보고되었으나 골 항상성과 파골세포의 분화 조절에 미치는 영향에 대해서는 아직 명확하지 않다. 따라서 본 연구에서는 파골세포의 분화 조절에 리포테이코산과 아디포넥틴이 미치는 영향에 대해 알아보고 그 작용기전을 규명하고자 하였다.

실험방법
리포테이코산이 골 항상성에 관여하는지 확인하기 위해, 야생형 황색포도상구균 또는 리포테이코산-결여 균을 쥐의 복강 내 주입하거나 황색포도상구균에서 분리한 리포테이코산을 지질펩타이드와 단독 또는 복합으로 쥐의 두개골 위에 이식하였다. 파골세포의 분화는 기질세포를 제거한 골수세포를macrophage colony-stimulating factor (M-CSF)와 배양하여 얻은 대식세포를 이용해 M-CSF와 receptor activator of NF-κB ligand (RANKL)과 추가 배양하여 유도하였으며, 이 때 황색포도상구균 또는 리포테이코산을 함께 처리하였다. 또한 후기 파골전구세포인 RANKL이 전처리 된 대식세포를 이용하여M-CSF만 존재하는 상태에서 리포테이코산을 황색포도상구균, 지질펩타이드, 내독소, 종양괴사인자-알파, 또는 RANKL과 단독 또는 복합 처리하였다.
파골세포 분화 조절에 아디포넥틴의 역할을 규명하기 위해, 야생형 쥐와 아디포넥틴-결여 쥐를 이용하였다. 두개골에서 조골전구세포를 분리한 후 조골세포 또는 지방세포로 분화를 유도하였으며, 골수 유래 대식세포를 이용하여 파골세포로 분화를 유도하였다. 아디포넥틴 결핍 시 파골세포 분화능의 차이가 대식세포 아형과 관련되는지 확인하기 위해, granulocyte/macrophage colony-stimulating factor에 의해 분화된 M1 대식세포와 M-CSF에 의해 분화된 M2 대식세포를 사용하였다. 골 미세환경을 대변하는 시험관 내 방법인 조골전구세포-대식세포 공배양을 통해 파골세포 분화를 유도하였다. 생체 내 골 소실 정도를 확인하기 위해, RANKL을 쥐의 복강 내 주입 또는 두개골 위에 이식하였다.
미세단층촬영기법을 이용해 대퇴골 내 소주골 또는 두개골의 골량을 분석하였다. 헤마톡실린/에오신 또는 tartrate-resistant acid phosphatase (TRAP)이 염색된 조직절편에서 골흡수, 골생성, 지방생성 관련 인자들을 측정하였다. 파골세포 분화는 TRAP 염색 시 세 개 이상의 핵을 갖는 TRAP-양성 다핵세포의 수를 측정하여 확인하였다. 파골세포 활성은 칼슘인산-부착 판에서 세포배양 후 생성된 흡수 면적을 측정하여 평가하였다. 신호전달인자의 인산화와 c-Fos와 NFATc1의 발현 정도를 Western blotting으로 확인하였고, AP-1과 NFATc1의 DNA-결합 활성은 electrophoric mobility shift assay를 통해 확인하였다. 조골세포 분화는 alkaline phosphatase 또는 Alizarin Red S 염색으로, 지방세포 분화는 Oil Red O 염색을 통해 확인하였다. RANKL과 osteoprotegerin (OPG)의 발현은 세포 배양액, 혈청, 골수 내 세포외액을 사용해 효소면역측정법으로 확인하였다.

결과
리포테이코산-결여 황색포도상구균은 야생형 황색포도상구균에 비해 더 강한 골 소실을 유발했고, 후기 파골전구세포로부터 파골세포의 분화를 강하게 유도하였다. 하지만 리포테이코산 단독은 생체 내/외 모두에서 골 소실 또는 파골세포 분화에 영향을 주지 않았다. 대신 리포테이코산은 RANKL에 의한 대식세포의 파골세포 분화 억제하였으며, 이 억제기전은 톨유사수용체-2 (toll-like receptor 2, TLR2)를 경유하여 MyD88의존/비의존적 경로를 통해 일어났다. 또한 리포테이코산은 황색포도상구균, 지질펩타이드, 내독소, 종양괴사인자-알파에 의해 유도되는 후기 파골전구세포로부터의 파골세포 분화 역시 강하게 억제하였다. 이 억제기전은 TLR2를 경유하지 않으며 파골세포 분화와 활성 과정 중 액틴 결합반응에 필수적인 액틴-젤솔린의 해리를 억제하였다. 따라서 리포테이코산은 파골세포 분화를 직접적으로 억제하는 핵심 인자임을 확인하였다.
아디포넥틴-결여 쥐는 대퇴골 또는 두개골의 골량이 정상 쥐에 비해 감소되어 있었고, 대퇴골 내 골흡수의 증가, 골생성의 감소, 골 내 지방 증가되어 있었다. 생체 내 현상과 동일하게, 아디포넥틴-결여 쥐에서 분리한 조골전구세포는 정상세포와 비교 시 조골세포의 분화는 억제된 반면 지방세포의 분화는 향상되었다. 하지만 파골세포의 분화를 확인한 결과 아디포넥틴-결여 대식세포가 정상세포에 비해 TRAP-양성 다핵세포의 생성, 골 흡수능, AP-1과 NFATc1의 DNA-결합 활성이 현저히 억제되었다. 또한 저해된 파골세포 분화능을 가진 아디포넥틴 결여 쥐의 대식세포는 M1 대식세포의 특성을 가지고 있었는데, M1 대식세포 분화의 핵심 인자인 인터페론 조절인자 5의 발현 증가가 RANKL 신호전달을 억제하기 때문인 것을 확인하였다. 다른 한편으로 골 미세환경을 대변하는 조골전구세포와 대식세포의 공배양 시, 파골세포 분화가 아디포넥틴-결여 조골전구세포와 공배양한 경우 증가함을 확인하였다. 특히, 파골세포 활성지표인 RANKL/OPG 비율의 증가가 아디포넥틴-결여 조골세포의 배양액과 아디포넥틴-결여 쥐의 골수 내 세포외액에서 동일하게 관찰되었다. 또한 생체 내 골소실의 정도가 아디포넥틴-결여 쥐에서 정상 쥐와 비교 시 증가하였다. 이를 통해 아디포넥틴이 파골세포와 조골세포의 분화를 모두 조절하는 인자임을 확인하였다.

결론
이상의 연구결과들로부터 다음과 같은 결론을 얻을 수 있었다. 리포테이코산은 직접적으로 파골세포 분화에 영향을 주지 않으나, 파골세포 활성화 인자들에 의해 유도되는 파골세포의 분화를 억제하는 것을 알 수 있었다. 아디포넥틴은 직접적으로 파골세포 분화를 향상시키나 조골세포로부터 분비되는 RANKL/OPG 비율을 억제하여 궁극적으로 파골세포 분화를 억제하며, 복합적으로 조골세포의 분화는 활성화시켜 골량을 증가시킨다. 결론적으로 리포테이코산과 아디포넥틴은 파골세포의 분화를 억제함으로써 궁극적으로 골대사를 조절하는데 기여한다.

다국어 초록 (Multilingual Abstract)

Objectives
Bone homeostasis is maintained by a balance between bone-resolving osteoclasts and bone-forming osteoblasts. Osteoclasts are derived from monocyte/macrophage lineages of hematopoietic stem cells, whereas osteoblasts are derived from mesenchymal stem cells which are likely to differentiate into adipocytes. Because the differentiation of osteoclasts and osteoblasts is important for maintaining appropriate bone metabolism, disruption of the differentiation can result in bone diseases such as bacteria-induced bone destruction and osteoporosis. Bacterial infection triggers inflammatory bone destruction resulting from excessive osteoclast differentiation, and microbe-associated molecular patterns (MAMPs) expressed on bacteria are known to induce inflammatory responses. Lipoteichoic acid (LTA) is considered as a major virulence factor expressed on the cell wall of Gram-positive bacteria, but little is known about its effect on the osteoclastogenesis. On the other hand, osteoporosis is known to be a representative bone disease accompanying reduced bone mass with increased bone marrow adiposity. Adiponectin is the member of adipokines secreted primarily from adipocytes and plays an important role in the regulation of metabolic processes, but its role in osteoclastogenesis is still uncertain. In the present study, the roles of LTA and adiponectin in the regulation of osteoclastogenesis and the action mechanisms were investigated.

Methods
Ethanol-inactivated wild-type and LTA-deficient Staphylococcus aureus were prepared and LTA was purified from S. aureus. Mice were intraperitoneally administered with ethanol-inactivated bacteria. Mouse calvarial bone was implanted with collagen sheet soaked in LTA and/or a synthetic peptide Pam2CSK4 mimicking Gram-positive bacterial lipoproteins. Bone marrow-derived macrophages (BMMs) were prepared from stroma-free bone marrow cells (BMs) by incubating with macrophage colony-stimulating factor (M-CSF), and the BMMs were further differentiated into osteoclasts by incubating with M-CSF and receptor activator of nuclear factor κB ligand (RANKL) in the presence or absence of the activated bacteria or LTA. Committed osteoclast precursors were prepared from BMMs by incubating with RANKL and M-CSF. The committed osteoclast precursors were stimulated with LTA and/or osteoclast-inducing factors, such as bacteria, Pam2CSK4, lipopolysaccharide (LPS), TNF-α, and RANKL, in the presence of M-CSF.
To investigate the role of adiponectin in osteoclastogenesis, wild-type and adiponectin-deficient mice were used. Osteoblast precursors were isolated from the calvaria of one-day-old mice and incubated with ascorbic acid and β-glycerophosphate to induce osteoblast differentiation. For induction of adipocyte differentiation, osteoblast precursors were incubated with dexamethasone, insulin, and 3-isobutyl-1-methylxanthine. BMMs were prepared from stroma-free BMs by incubating with granulocyte/macrophage colony-stimulating factor (GM-CSF) for M1 macrophages or M-CSF for M2 macrophages. To examine osteoclast differentiation, the BMMs were incubated with M-CSF and RANKL or co-cultured with osteoblast precursors. RANKL-induced bone destruction was made by intraperitoneal administration or by implantation on mouse calvarial bone.
Bone parameters of the femur and the calvaria were analyzed by micro-computed tomography. Osteoclast differentiation and bone resorption capacity were evaluated by enumerating tartrate-resistant acid phosphatase (TRAP)-positive multinucleated cells (MNCs) and by measuring the areas of resorption pits. Expression of cell surface proteins, such as TLR2, CD14, c-Fms, RANK, and TREM2, was analyzed by flow cytometry. Osteoblast differentiation was evaluated by staining with alkaline phosphatase and Alizarin Red S. Adipocyte differentiation was evaluated by staining with Oil Red O. Expression of RANKL, OPG, IL-6, TNF-α, and IL-10 in the culture supernatants, the bone marrow extracellular fluids, or the serum was measured by enzyme-linked immunosorbent assay. Intracellular signaling pathways were determined by Western blotting, electrophoretic mobility shift assay, or immunoprecipitation.

Results
LTA-deficient S. aureus augmented massive bone destruction compared with wild-type S. aureus. Consistently, the LTA-deficient S. aureus increased the number of TRAP-positive MNCs. LTA alone affected neither differentiation of committed osteoclast precursors nor bone destruction. However, LTA inhibited differentiation of BMMs into TRAP-positive MNCs in response to RANKL. In contrast, LTA augmented phagocytic and inflammatory potential of the BMMs. The inhibitory effect of LTA on RANKL-induced osteoclastogenesis was mediated by attenuated activation of ERK, JNK, and AP-1 through toll-like receptor 2 (TLR2) with partial involvement of MyD88. In addition to BMMs, LTA also inhibited differentiation of committed osteoclast precursors into TRAP-positive MNCs triggered by S. aureus, Pam2CSK4, LPS, TNF-α, or RANKL. The inhibitory effect of LTA was not altered in the osteoclasts derived from mice lacking TLR2. LTA suppressed activation of both MAPKs and actin polymerization. Taken together, these results suggest that LTA inhibits not only M-CSF/RANKL-mediated osteoclastogenesis from macrophages through TLR2-dependent pathways but also osteoclastogenesis from committed osteoclast precursors triggered by TLR ligands, pro-inflammatory cytokines, and RANKL through TLR2-independent pathways.
Adiponectin-deficient mice exhibited a decrease of bone mass but, instead, an increase of bone marrow adiposity. Bone histomorphometric analysis also demonstrated that the bone of adiponectin-deficient mice showed an increase of bone resorption, a decrease of bone formation, and an increase of adiposity formation. Concordant with the in vivo results, adiponectin-deficient osteoblast precursors were scarcely differentiated into osteoblasts but, instead, efficiently differentiated into adipocytes. However, unlike the in vivo results, adiponectin-deficient BMMs were hardly differentiated into TRAP-positive MNCs. Both formation of bone resorption pits and DNA-binding activities of AP-1 and NFATc1 were reduced in the adiponectin-deficient BMMs. The weak osteoclastogenic potential of the adiponectin-deficient BMMs was associated with properties of M1 macrophages. M1 macrophages were barely differentiated into osteoclasts by attenuating activation of NFATc1 conferred by the increased expression of IRF5. However, when adiponectin-deficient BMMs were co-cultured with adiponectin-deficient osteoblast precursors, the number of TRAP-positive MNCs was increased. In addition, a RANKL/OPG ratio was up-regulated under the same condition. The up-regulated RANKL/OPG ratio was also observed in bone marrow extracellular fluid of adiponectin-deficient mice. Indeed, RANKL-induced bone destruction was facilitated in the adiponectin-deficient mice. These results suggest that adiponectin is an important factor inhibiting osteoclastogenesis by down-regulation of RANKL/OPG ratio, consequently contributing to an increase of bone mass.

Conclusion
Controlling the osteoclastogenesis is important for an efficient treatment or prevention of bone diseases. Therefore, it is important to identify the molecular targets controlling osteoclastogenesis and to characterize their regulatory mechanisms for osteoclastogenesis. The present study demonstrated the inhibitory roles of LTA and adiponectin in osteoclastogenesis. Although LTA alone affected neither osteoclast differentiation from committed osteoclast precursors nor bone mass, LTA inhibited not only M-CSF/RANKL-induced osteoclast differentiation from macrophages but also TLR ligands, TNF-α, and RANKL-induced osteoclastogenesis from committed osteoclast precursors, consequently resulting in suppression of bone resorption. Adiponectin deficiency enhanced osteoclast differentiation by down-regulation of RANKL/OPG ratio and also attenuated osteoblast differentiation by skewed differential potential of osteoblast precursors toward adipocytes. Conclusively, LTA and adiponectin might be important factors capable of inhibiting osteoclastogenesis and could be potential target molecules controlling bone diseases accompanying excessive generation of osteoclasts.

목차 (Table of Contents)

• Abstract
• Contents
• List of figures
• List of tables
• Abbreviations
• Chapter I. Introduction 1
• 1. Bone and osteoclastogenesis 1
• 1.1. Bone remodeling 1
• 1.2. Bone diseases 3
• 1.3. Osteoclastogenesis 4
• 1.3.1. Macrophages 6
• 1.3.2. Committed osteoclast precursors 8
• 1.3.3. Regulation of osteoclastogenesis 8
• 2. Microbes and lipoteichoic acid (LTA) 9
• 2.1. Microbial regulation of osteoclastogenesis 9
• 2.2. Effect of MAMPs on osteoclastogenesis 9
• 2.3. LTA 12
• 3. Adipokines and adiponectin 13
• 3.1. Adipokines 13
• 3.2. Adiponectin 15
• 4. Aims of this study 15
• Chapter II. Materials and Methods 16
• 1. Reagents and chemicals 16
• 2. Experimental animals 18
• 3. Preparation of ethanol-inactivated S. aureus 18
• 4. Preparation of LTA 18
• 5. Cells 19
• 5.1. RANKL/M-CSF-induced osteoclastogenesis of macrophages 19
• 5.2. Osteoclast differentiation of committed osteoclast precursors 20
• 5.3. Isolation of calvarial osteoblast precursors 20
• 5.4. Osteoclast differentiation by co-culture of osteoblast precursors with osteoclast precursors 20
• 5.5. Differentiation of calvarial cells into osteoblasts or adipocytes 21
• 6. Cell staining 21
• 6.1. TRAP staining 21
• 6.2. Alkaline phosphatase (ALP) staining 21
• 6.3. Alizarin Red S staining 21
• 6.4. Oil Red O staining 22
• 7. In vitro bone resorption assay 22
• 8. Micro-computed tomography (micro-CT) analysis 22
• 9. Bone histomorphometric analysis 23
• 10. In vivo bone resorption assay 23
• 10.1. Trabecular bone resorption assay 23
• 10.2. Calvarial bone resorption assay 23
• 11. Measurement of cell viability 24
• 12. Flow cytometric analysis 24
• 13. Endocytic activity assay 25
• 14. Western blotting 25
• 15. Immunoprecipitation 26
• 16. Electrophoretic mobility shift assay (EMSA) 26
• 17. Immunofluorescence 27
• 18. Enzyme-linked immunosorbent assay (ELISA) 28
• 19. Reverse transcription-polymerase chain reaction (RT-PCR) 28
• 20. Measurement of intracellular calcium and oscillation 29
• 21. IRF5 overexpression by transfection 30
• 22. IRF5 knock-down by transfection with small interfering RNA (siRNA) 30
• 23. Statistical analysis 30
• Chapter III. Results 31
• 1. Role of LTA in osteoclastogenesis 31
• 1.1. LTA-deficient S. aureus induces massive bone resorption and excessive osteoclastogenesis compared with wild-type S. aureus 31
• 1.2. S. aureus LTA affects neither osteoclastogenesis nor bone mass 34
• 1.3. S. aureus LTA inhibits M-CSF/RANKL-induced osteoclastogenesis from macrophages 36
• 1.3.1. S. aureus LTA inhibits M-CSF/RANKL-induced osteoclast differentiation from macrophages 36
• 1.3.2. S. aureus LTA enhances both endocytic capacity and TNF-α production in macrophages 39
• 1.3.3. Inhibitory effect of S. aureus LTA on osteoclastogenesis is mediated through TLR2 but partial involvement of MyD88 41
• 1.3.4. S. aureus LTA inhibits both M-CSF and RANKL signaling and reduces DNA-binding activity of AP-1 44
• 1.4. S. aureus LTA inhibits osteoclast differentiation triggered from committed osteoclast precursors 46
• 1.4.1. S. aureus LTA inhibits S. aureus- and Pam2CSK4-induced osteoclast differentiation from committed osteoclast precursors 46
• 1.4.2. S. aureus LTA inhibits Pam2CSK4-induced osteoclast activation and bone resorption 49
• 1.4.3. S. aureus LTA inhibits differentiation of committed osteoclast precursors into osteoclasts in response to LPS, TNF-α, or RANKL 51
• 1.4.4. TLR2 signaling is not necessary for S. aureus LTA-mediated inhibition of osteoclast differentiation 53
• 1.4.5. S. aureus LTA inhibits Pam2CSK4-induced MAPK activation and DNA-binding activity of NFATc1 55
• 1.4.6. S. aureus LTA inhibits dissociation of gelsolin-actin complex induced by Pam2CSK4 or RANKL 57
• 2. Role of adiponectin in osteoclastogenesis 59
• 2.1. Adiponectin-deficient mice exhibit decreased bone mass but increased bone marrow adiposity 59
• 2.2. Adiponectin-deficient osteoblast precursors preferentially differentiate into adipocytes rather than osteoblasts 62
• 2.3. Adiponectin-deficient macrophages have weak osteoclastogenic potential 65
• 2.3.1. Adiponectin-deficient BMMs scarcely differentiate into osteoclasts 65
• 2.3.2. The percentages of osteoclast progenitors in whole BMs are similar between wild-type and adiponectin-deficient mice 67
• 2.3.3. Adiponectin-deficient BMMs show reduced activation of AP-1 and NFATc1 by decreasing activation of p38 kinase and CREB in response to RANKL 69
• 2.3.4. Weak osteoclastogenic potential of adiponectin-deficient macrophages is associated with properties of M1 macrophages 71
• 2.3.5. M1 macrophages have weak osteoclastogenic potential 75
• 2.3.5.1. M1 macrophages scarcely differentiate into osteoclasts 75
• 2.3.5.2. M1 macrophages have reduced bone resorptive capacity 77
• 2.3.5.3. Efficient osteoclastogenic potential of M2 macrophages is not associated with expression of osteoclastogenesis-associated receptors 79
• 2.3.5.4. M2 macrophages remarkably induces activation of c-Fos and NFATc1 through up-regulation of calcium oscillation followed by CREB activation 81
• 2.3.5.5. Osteoclastogenic potential of macrophage is reversible by converting macrophage subtypes 85
• 2.3.5.6. Expression of IRF5 modulates osteoclast differentiation 87
• 2.4. Up-regulated RANKL/OPG ratio enhances osteoclast differentiation in adiponectin deficiency 90
• 2.4.1. Adiponectin-deficient osteoblast precursors have up-regulated RANKL/OPG ratio 90
• 2.4.2. Osteoclast differentiation is enhanced by osteoblast from adiponectin-deficient mice through up-regulation of RANKL/OPG ratio 92
• 2.4.3. RANKL/OPG ratio is increased in the bone marrow of adiponectin-deficient mice 95
• 2.5. Adiponectin deficiency facilitates RANKL-induced bone resorption 97
• Chapter IV. Discussion 99
• Chapter V. References 108
• 국문초록