ADAM10Contributes To Long Bone Growth Defect And Correlates With Osteosarcoma Progression

The chondro-osseus junction (COJ) is the site of dynamic interactions between severaldifferent cell types that drive bone growth during endochondral ossification andosteosarcoma tumorigenesis and progression. The most prominent cell types at the growthplate are the hypertrophic chondrocytes on one side of the COJ, and invading osteoclastsand endothelial cells as well as osteoblasts on the other side. A considerable amount ofinformation is available on the regulation of chondrocytes, osteoclasts and osteoblasts at theCOJ during endochondral ossification, but only little is known about the role of endothelialcells in this process. We have previously identified a defect in long bone growth in micelacking the cell surface metalloprotease ADAM10(a disintegrin and metalloprotease10) inendothelial cells, raising questions about the underlying cause of this growth defect,additionally, whether ADAM10is involved in osteosarcoma is still unknown. ADAM10is aprincipal regulator of Notch signaling, and mice lacking ADAM10in endothelial cells(ADAM10△EC mice) display a characteristic increase in vascular branching in thedeveloping retinal vasculature that is a hallmark for defects in Notch signaling. MostADAM10△EC mice survive for many months, providing a unique opportunity to study howthe lack of ADAM10in endothelial cells affects the growth of long bones. On the otherhand, recent studies have shown that Notch signaling contribute to osteosarcomatumorigenesis and invasion, providing clues for the involvement of ADAM10as a majorsheddase of Notch in osteosarcoma progression through angiogenesis. This study was toexamine the long bones of ADAM10△EC mice at different times of development comparedto controls, with an emphasis on identifying possible abnormalities in the appearance anddistribution of endothelial cells, chondrocytes and osteoclasts at the COJ. In addition,ADAM10on the progression of human osteosarcoma and its angiogenesis were alsoinvestigated. Part1. ADAM10in mouse long bone growth defectMaterials and methods1. Establishment of ADAM10△EC mice. ADAM10△EC mice were generated bymating ADAM10lox/loxmice with transgenic mice expressing the endothelial-specificTie2-Cre transgene, to yield endothelial cell depleted ADAM10mice (ADAM10△EC,genotype as ADAM10lox/loxTie2-Cre+/-) and normal control mice (Control, genotype asADAM10lox/loxTie2-Cre-/-). To identify these mice, DNA was extracted from the tail andPCR was done to examine the amplification products of loxP and Tie2in order todistinguish whether it was from ADAM10△EC or control mice.2. Dynamic alterations of long bones growth defect in ADAM10△EC mice. FaxitronX-ray scanner was used to monitor the dynamic alterations of long bone growth defect bycomparing ADAM10△EC mice and control mice, trabecular bone density was alsocalculated. The overall alterations of bone and cartilage of ADAM10△EC mice wereexamined by Alizarin Red&Alcian Blue whole-mount staining. The alteration of thegrowth plate and its cartilage of femurs and tibiae from tissue sections of mouse knee jointwas examined by H&E staining and Safranin O&Fast Green cartilage staining, and theeffect of the chondrocyte growth by ADAM10deletion was also calculated.3. Vascular development at COJ affected by ADAM10deletion in endothelial cells.Mouse endothelial cell specific marker-endomucin antibody was used to detect the vascualrmorphological changes at the COJ region after deletion of ADAM10in endothelial cells byimmunofluorescent staining of the knee joint sections. In addition, methacrylate resin wasused as a casting material to establish the3D replica of the femoral vascular structure at theCOJ.4. The effect of ADAM10depletion to osteoclast at COJ and the relationship betweenthe alteration of osteoclast and the abnormalities of the vasculature at the COJ. TRAPstaining was performed to examine the alterations of the osteoclast number at the COJ ofthe femur after deletion of ADAM10in endothelial cells, and endomucin-TRAP doublestaining was performed to elucidate the correlation between the vascular abnormality andthe osteoclast alteration.5. The effect of ADAM10deletion in endothelial cells to osteoclastogenesis. Bonemarrow cells from the femurs and tibiae of both ADAM10△EC mice and control mice were isolated and cultured in vitro by stimulating with M-CSF to differentiate intomacrophages, which were continued stimulated with/without RANKL to differentiate intoosteoclast, as examined by TRAP staining to obtain the numerical and morphologicalalterations.6. RANKL/OPG signaling in ADAM10△EC mice. RNA was extracted in the COJregion from the forelimbs and hindlimbs of ADAM10△EC and control mice. Real timequantitative RT-PCR was performed to detect the changes of the RANKL/OPG ratio toaddress the effect of RANKL/OPG pathway on abnormal osteoclastogenesis in ADAM10△EC mice.Results1. Deletion of ADAM10in mouse endothelial cells caused multiple long bone growthdefect. Faxitron data analysis revealed that ADAM10△EC femurs were slightly shorter thanthose of littermate controls at P7, and that this growth defect persisted at all time points, butwas most severe at6months of age. Measurements of the length of the tibiae by faxitronanalysis showed that there was a significant growth defect at all stages from P7to P42inADAM10△EC mice compared to littermate controls.The length of the humerus, ulna andradius were slightly, but significantly shorter in ADAM10△EC mice compared to controllittermates starting at P14and at all later stages until P42. The length of the metacarpals ofADAM10△EC mice was comparable to that of controls at all stages of development. Theincreased density of the trabecular bone in ADAM10△EC femurs was also visible onfaxitron images at P28and P42. The tibia growth plates also displayed an increase in thedensity of trabecular bone at P21and later stages of development.2. Deletion of ADAM10led to structural abnormalities of long bone growth plate. Thehistopathological analysis of the growth plate at P7and P14did not uncover any evidentabnormalities in ADAM10△EC femurs at these early time points compared to controls. AtP21, the posterior part of the ADAM10△EC distal femur growth plate had developed adiscontinuity in three out of four samples. By P28, the discontinuity of the posterior aspectof the distal femur growth plate had progressed to a point where the entire growth plate wasbisected in ADAM10△EC mice, whereas the anterior aspects of the growth plate hadirregular enlargements of the zone of hypertrophic chondrocytes compared to controls. Inaddition, ADAM10△EC femurs showed an increase in the density of trabecular bone under the growth plate at P28, which was more pronounced towards the cortical aspects of thebone, and resulted in narrowing of the central bone marrow cavity compared to wild typecontrols. By P42, several large discontinuities spanning the entire zone of chondrocytes haddeveloped, and there was an overall increase in the density of trabecular bone. When asimilar analysis of tibiae was performed by H&E staining at different stages ofdevelopment, the growth plate in ADAM10△EC mice appeared normal at P7, but had anincreased size of the central part of the zone of hypertrophic chondrocytes starting at P14and persists until P42that was monitored, which caused subchondral bone deficiencytherein. However, unlike at the distal femur growth plate, there were no discontinuities ofthe proximal tibial growth plate. In addition, whole-mount Alizarin Red&Alcian Bluestaining revealed an early closure of the epiphyseal line of the femur of ADAM10△ECmice at P42. Safranin O&Fast Green cartilage staining also showed that the cartilagesynthesis in the proliferating zone of the femoral and tibial growth plate of ADAM10△ECmice markedly decreased starting at P21.3. Loss of ADAM10in mouse endothelial cells led to vascular developmentabnormality at COJ. The endomucin staining at the COJ at P14was stronger in samplesfrom ADAM10△EC mice compared to controls, and that there was an increased density ofendomucin-positive cells at the COJ. A quantification of the vessel density at the COJ atP14confirmed the increase in ADAM10△EC femurs at P14compared to controls. Bloodvessel invasion was evident at the discontinuity site at P21, and this blood vessel invasionwas found both at the superior and inferior of the bisected part of the growth plate. In tibia,bulb-like vessels were found at the COJ, and loss of blood vessels was underneath theenlarged hypertrophic growth plate, while compared with ADAM10△EC mice, the bloodvessel at the COJ of control mice shows a regular and even distribution pattern and theendomucin staining was less intensed. Examination of the vascular casts of the femur undera stereo microscope showed a large central artery with increased branching towards thegrowth plates that ended in regularly distributed fine vascular loop structures in controlmice. By comparison, the vascular casts from ADAM10△EC mice showed that the largecentral artery branched out in vascular loops that were less parallel and less well organizedand had numerous small bulb-like expansions.4. Long bone growth defect of ADAM10△EC mice was caused by the decrease of the osteoclast number at COJ. We found similar numbers and distribution of TRAP stainedcells in the growth plates of ADAM10△EC femurs compared to controls at P7and P14.However, at P21, the distal femoral growth plate in ADAM10△EC mice had a significantreduction in the number of TRAP-stained cells at the COJ compared to controls, and thisdifference persisted at P28. Similar observations were made in an analysis of the proximalgrowth plate of the tibia, where the number of TRAP+cells was comparable at P7and P14in ADAM10△EC mice and controls, but was significantly reduced in ADAM10△EC micefrom P21onward.5. The decrease of osteoclast number at COJ caused by ADAM10deletion inendothelial cells colsely related with disorganized vascular development at COJ.Co-staining of these samples for TRAP followed by merging of the images revealed a tightassociation of TRAP+cells with endothelial cells at the COJ at P14in ADAM10△EC miceand controls. At P28, there was a similar increase in the endomucin-positive cells at theCOJ in femurs from ADAM10△EC mice, but very few TRAP+cells could be detected atthe COJ.6. Loss of ADAM10in endothelial cells led to a decrease of osteoclastogenesis in vitro,while the RANKL/OPG pathway was not affected. When we cultured osteoclast precursorsfrom ADAM10△EC mice, we found that the osteoclastogenesis was slightly delayedcompared to the control mice, resulting in a small but significant reduction in the number ofTRAP+multinucleated osteoclasts. However, there was no significant difference in theRANKL/OPG ratio as determined by qPCR in samples extracted from the zone ofhypertrophic and terminally differentiation/COJ zones in the hindlimbs and forelimbs ofADAM10△EC or control mice.Part2. ADAM10in osteosarcoma progressionMaterials and methods1. Human osteosarcoma samples. Main samples were purchased osteosarcoma tissuechip slides, each includes40cases of human osteosarcoma tissue spots treated with PFAand embedded in paraffin. The pathological stages were from IA to IIB and the histologicaldiagnosis include ostoeblastic osteosarcoma, chondroblastic osteosarcoma, fibroblasticosteosarcoma, and giant cell rich osteosarcoma. Some samples were from the Departmentof Pathology of our Hospital. 2. H&E staining. To address the origin, histological morphology, and pathologicaltypes of osteosarcoma tissues.3. Immunofluorescent staining. Osteosarcoma tissue sections were stained withanti-CD31, anti-cytoplasmic ADAM10, anti-activated Notch1(NICD, Notch1intracellulardomain) antibodies，and DAPI (to stain the neuclei), the expression density of CD31andNICD and the ADAM10+tumor cells were calculated and correlated with tumor stagingand histological types. In addition, CD31/ADAM10, CD31/NICD doubleimmunofluorescent stainings were done to investigate the correlation of ADAM10/Notch1with osteosarcoma vascular endothelial cells.4. TRAP staining. To investigate the expression of osteoclast in osteosarcoma tissuesand address whether osteoclast-like multinucleated giant cell is an authentic osteoclast.Results1. ADAM10expression positively correlated with osteosarcoma progression. Apolarized and condensed expression pattern of cytoplasmic ADAM10was observed in allcases and stages of osteosarcoma. ADM10+tumor cells increase dramatically asosteosarcoma advances from IA to IIB, the average ADAM10+tumor cell numbers of stageIIA and IIB were significantly higher than that of stage IA. In addition, the number ofADAM10+tumor cells was significantly larger in osteoblastic osteosarcoma than that inchondroblastic and fibroblastic osteosarcoma.2. A cluster of angiogenic tumor cells were identified in stage IA/IB osteosarcoma,which might form vascular lumens through “cytoplasmic invagination”. H&E stainingidentified a cluster of basophilic round tumor cells, its positivity for CD31staining suggestsits angiogenic property, some of these cells underwent different stages of “cytoplasmicinvagination”, which might be a mechanism utilized by osteosarcoma to form vascularlumen. Additionally, cytoplasmic ADAM10was homogeneously expressed in theseangiogenic tumor cells.3. Osteoclast-like multinucleated giant cell is not an osteoclast but an angiogenictumor cell, where ADAM10/Notch1signaling was activated. TRAP/DAPI double stainingshows that multinucleated giant cell was negative for TRAP staining in giant cell-richosteosarcoma, while osteoblastic, chondroblastic, and fibroblastic osteosarcoma all hadTRAP+tumor cells respectively. CD31was expressed at intermediate level in the cytoplasm of ginat cells, ADAM10and activated Notch1were also expressed in giant cellsin a similar pattern as CD31.4. The vasculature of partial osteosarcoma had endothelium deficiency. CD31stainingrevealed red blood cells in some osteosarcoma from stage IA to IIB osteosarcomavasculature were not surrounded by endothelium, which was stained positive for CD31anddisproved the reason as CD31was not expressed in endothelium of osteosarcoma but ratheras endothelium deficiency, which might be related with multinucleated giant cells.Conclusions1. The role of bone vessels at the growth plate in regulating the function of osteoclastsonly manifests itself in the large long bones such as the femur, humerus, tibia, ulna andradius at the later stages of their growth.2. ADAM10in endothelial cells is required for the proper development and function ofthe specialized vasculature in the bone during the process of endochondral ossification.3. The reduced number of osteoclasts at the COJ in joints of long bones of ADAM10△EC mice are likely the consequence of alterations in Notch-dependent differentiation ofendothelial cells in the absence of ADAM10, instead of defects in release of a soluble factorfrom endothelial cells by ADAM10.4. The expression of ADAM10in tumor cells positively correlates with osteosarcomaprogression.5. ADAM10participated in the cytoplasmic invagination of angiogenic tumor cells ofstage IA/IB osteosarcoma, which might contribute to vascular lumen formation.6. ADAM10/Notch1signaling was activated in multinucleated giant cell, which wasnot osteoclast, but rather was angiogenic and might be related with endothelium deficiencyin partial osteosarcoma vasculature.