作为一个图书馆,NLM 提供科学文献的访问。纳入 NLM 数据库并不意味着认可或同意, NLM 或国立卫生研究院的内容。
Learn more: PMC Disclaimer | PMC Copyright Notice
了解更多: 免责声明 | PMC版权声明
关节炎风湿病。 11.4【2023最新】13.3Q1医学-1区,风湿病学-1区 2016年7月; 68(7):1648-1659。
Mesenchymal Stem Cell Alterations in Bone Marrow Lesions in Patients With Hip Osteoarthritis
髋骨关节炎患者骨髓病变中间充质干细胞的改变
T. 马克·坎贝尔、 1莎拉·M·丘奇曼、 2亚历杭德罗·戈麦斯、 2丹尼斯·麦格纳格尔、 2菲利普·G·科纳汉、 2弗雷德里克·庞切尔、 2和埃琳娜·琼斯 2
Sarah M. Churchman
2University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
Alejandro Gomez
2University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
Dennis McGonagle
2University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
Philip G. Conaghan
2University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
Frederique Ponchel
2University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
Elena Jones
2University of Leeds and NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds, UK
作者信息文章注释版权和许可信息PMC 免责声明
Associated Data 相关数据
- Supplementary Materials 补充材料
- ART-68-1648-s001.docx (37K)GUID: DD850F66-5F81-4338-8581-5A088C0E7557ART-68-1648-s002.tif (3.7M)GUID: DB48F591-2E1C-4096-9DA4-18174ADC419CART-68-1648-s003.tif (2.9M)GUID: 13B56ED9-1AD0-4B2D-BC17-F329E0305D95
Abstract 抽象的
Objective 客观的
In patients with osteoarthritis (OA), bone marrow lesions (BMLs) are intimately linked to disease progression. We hypothesized that aberrant multipotential stromal cell (also known as mesenchymal stem cell [MSC]) responses within bone tissue contributes to BML pathophysiology.
在骨关节炎 (OA) 患者中,骨髓病变 (BML) 与疾病进展密切相关。我们假设骨组织内异常的多能基质细胞(也称为间充质干细胞 [MSC])反应有助于BML病理生理学。
The aim of this study was to investigate BML and non‐BML native subchondral bone MSCs for numeric, topographic, in vitro functional, and gene expression differences.
本研究的目的是研究BML和非BML天然软骨下骨 MSC 的数量、形态、体外功能和基因表达差异。
Methods 方法
Ex vivo 3T magnetic resonance imaging (MRI) of the femoral heads of 20 patients with hip OA was performed.
对 20 名髋关节 OA 患者的股骨头进行离体 3T 磁共振成像 (MRI)。
MRI‐determined BML and non‐BML regions were excised and enzymatically treated to extract cells and quantify MSCs using flow cytometry and colony‐forming unit–fibroblast (CFU‐F) assay. Immunohistochemical analysis was performed to determine in vivo CD271+ MSC distribution.
切除 MRI 确定的BML和非BML区域,并进行酶处理以提取细胞,并使用流式细胞术和集落形成单位成纤维细胞 (CFU-F) 测定对 MSC 进行定量。进行免疫组织化学分析以确定体内 CD271+ MSC 分布。
Culture‐expanded CD271+ cells were analyzed for tripotentiality and gene expression.
分析培养扩增的 CD271+ 细胞的三潜能和基因表达。
Results 结果
BML regions were associated with greater trabecular bone area and cartilage damage compared with non‐BML regions. The proportion of CD45−CD271+ MSCs was higher in BML regions compared with non‐BML regions (median difference 5.6‐fold; P < 0.001); the CFU‐F assay showed a similar trend (median difference 4.3‐fold; P = 0.013). Immunohistochemistry revealed CD271+ cell accumulation in bone adjacent to cartilage defects and areas of osteochondral angiogenesis. BML MSCs had lower proliferation and mineralization capacities in vitro and altered expression of TNFSF11/RANKL and CXCR4/stromal cell–derived factor 1 receptor. OA MSCs showed up‐regulated transcripts for CXCR1 and CCR6 compared with MSCs derived from healthy or osteoporotic bone.
与非 BML 区域相比,BML 区域与更大的骨小梁面积和软骨损伤相关。与非 BML 区域相比,BML 区域 CD45−CD271+ MSC 的比例较高(中位差异 5.6 倍; P < 0.001); CFU-F 检测显示出类似的趋势(中位差异 4.3 倍; P = 0.013)。免疫组织化学显示 CD271+ 细胞在邻近软骨缺损和骨软骨血管生成区域的骨骼中积聚。 BML MSC 在体外具有较低的增殖和矿化能力,并且TNFSF11 /RANKL 和CXCR4 /基质细胞衍生因子 1 受体的表达发生改变。与来自健康或骨质疏松骨的 MSC 相比,OA MSC 的CXCR1和CCR6转录本上调。
Conclusion 结论
This study is the first to show numeric and topographic alterations in native MSCs in the diseased bone of patients with hip OA. Given the associated functional perturbation of MSCs, these data suggest that subchondral bone MSC manipulation may be an OA treatment target.
这项研究首次显示了髋关节 OA 患者病变骨中天然 MSC 的数量和形态变化。鉴于 MSC 的相关功能扰动,这些数据表明软骨下骨 MSC 操作可能是 OA 治疗目标。
Osteoarthritis (OA) is the most common form of arthritis and a major cause of chronic pain and disability 1. As the population ages, the projected number of older adults with OA is expected to increase substantially in the next 2 decades 1. The pathophysiology of OA is complex, symptomatic treatment is often ineffective, and no licensed structure‐modifying OA drugs are currently available.
骨关节炎 (OA) 是最常见的关节炎形式,也是慢性疼痛和残疾的主要原因1 。随着人口老龄化,预计患有 OA 的老年人数量将在未来 20 年大幅增加1 。 OA 的病理生理学复杂,对症治疗往往无效,目前尚无获得许可的结构修饰 OA 药物。
Established OA involves pathology in multiple tissues, but subchondral bone plays an important role in pathogenesis and symptomatology 2, 3.
已确定的 OA 涉及多种组织的病理学,但软骨下骨在发病机制和症状学中起着重要作用2 , 3 。
With the use of magnetic resonance imaging (MRI), subchondral bone pathology, including bone marrow lesions (BMLs) 3, 4, 5, 6, can be visualized on fluid‐sensitive MRI sequences. Such BMLs are associated with overlying cartilage pathology, pain, and progression of structural abnormalities over time 5, 7. Histologically, BMLs represent mesenchymal tissue abnormalities including bone marrow fibrosis, necrosis, swollen/dying adipocytes, and alterations in trabecular bone structure 7.
通过使用磁共振成像 (MRI),软骨下骨病理学,包括骨髓病变 (BML) 3 , 4 , 5 , 6 ,可以在流体敏感的 MRI 序列上可视化。此类 BML 与上覆软骨病理、疼痛以及结构异常随时间的进展有关5 , 7 。在组织学上,BML 代表间充质组织异常,包括骨髓纤维化、坏死、脂肪细胞肿胀/死亡以及骨小梁结构的改变7 。
Multipotential stromal cells (also known as mesenchymal stem cells [MSCs]) are nonhematopoietic, clonogenic, multipotential cells that are present in numerous tissues 8. They have attracted great interest because of their regenerative and immunoregulatory properties, as well as their increased use in cell‐based therapies 9. However, MSC behavior in vivo and any potential contribution to the development of OA remain poorly understood 10.
多能基质细胞(也称为间充质干细胞 [MSC])是非造血、克隆形成、多能细胞,存在于多种组织中8 。由于它们的再生和免疫调节特性,以及它们在细胞疗法中越来越多的使用,它们引起了极大的兴趣9 。然而,间充质干细胞在体内的行为以及对 OA 发展的任何潜在贡献仍然知之甚少10 。
MSCs are abundant in trabecular bone, where they have been observed in both perivascular and bone‐lining locations 11, 12, 13. Bone‐resident MSCs are important for bone repair and remodeling by virtue of being precursors of osteoblasts, which not only form new bone but also control osteoclast activation by producing RANKL and osteoprotegerin 14. Within BMLs, perturbations in bone MSC function may lead to abnormal bone remodeling, which could affect the overlying cartilage 3.
间充质干细胞在骨小梁中含量丰富,在血管周围和骨衬位置均观察到它们11 , 12 , 13 。骨驻留间充质干细胞作为成骨细胞的前体,对于骨修复和重塑非常重要,它们不仅形成新骨,还通过产生 RANKL 和骨保护素来控制破骨细胞活化14 。在 BML 中,骨 MSC 功能的扰动可能会导致骨重塑异常,从而影响上层软骨3 。
We therefore hypothesized that subchondral bone MSCs contribute to BML pathophysiology and compared the numbers, topography, in vitro differentiation capacities, and gene expression profiles of MSCs extracted from paired BML and non‐BML regions from the same OA‐affected femoral head.
因此,我们假设软骨下骨 MSC 有助于BML病理生理学,并比较了从同一受 OA 影响的股骨头的配对BML和非BML区域提取的 MSC 的数量、形态、体外分化能力和基因表达谱。
PATIENTS AND METHODS 患者和方法
Patients and cells 患者与细胞
Patients with primary hip OA who were scheduled to undergo total hip arthroplasty (THA) were recruited from the orthopedic unit at Chapel Allerton Hospital, Leeds. All patients met the American College of Rheumatology criteria for the classification of hip OA 15. Exclusion criteria included a history of inflammatory arthritis, previous hip surgery, metastatic cancer, or disorders affecting bone. All patients gave written informed consent, and the study was approved by the National Research Ethics Committee Yorkshire and Humberside.
计划接受全髋关节置换术 (THA) 的原发性髋关节 OA 患者是从利兹 Chapel Allerton 医院骨科部门招募的。所有患者均符合美国风湿病学会对髋关节 OA 15的分类标准。排除标准包括炎症性关节炎病史、既往髋关节手术、转移性癌症或影响骨骼的疾病。所有患者均签署了书面知情同意书,该研究得到了约克郡和亨伯赛德国家研究伦理委员会的批准。
As a control, trabecular bone was harvested from the iliac crest in 9 age‐matched patients with pelvic fracture who were otherwise healthy (median age 57 years, range 39–84 years).
作为对照,从 9 名年龄匹配的骨盆骨折患者的髂嵴中采集了骨小梁,这些患者在其他方面都很健康(中位年龄 57 岁,范围 39-84 岁)。
Femoral heads were also collected from 5 patients with osteoporosis (OP) who had a femoral neck fracture (median age 83 years, range 74–92 years).
还收集了 5 名股骨颈骨折的骨质疏松症 (OP) 患者的股骨头(中位年龄 83 岁,范围 74-92 岁)。
Femoral heads from patients with OA or OP were collected immediately after removal during THA and placed in phosphate buffered saline (PBS) (Invitrogen).
OA 或 OP 患者的股骨头在 THA 期间取出后立即收集,并置于磷酸盐缓冲盐水 (PBS) (Invitrogen) 中。
Twenty‐one femoral heads from patients with OA (median age 65.3 years, range 48–83 years) were subjected to MRI, and femoral heads from 7 patients (median age 71 years, range 43–78 years) were used as controls in real‐time quantitative polymerase chain reaction (qPCR) and flow cytometry validation experiments.
对 21 名 OA 患者(中位年龄 65.3 岁,范围 48-83 岁)的股骨头进行 MRI,并将 7 名患者(中位年龄 71 岁,范围 43-78 岁)的股骨头用作真实对照。 ‐时间定量聚合酶链反应(qPCR)和流式细胞术验证实验。
For MRI, explants were mounted on a nonmetal bracket and clamp with orienting cut‐outs (Figure (Figure1A)1A) and secured in a PBS‐containing polyethylene jar (Figure (Figure1B).1B). Cod liver oil pills were fixed externally to provide further position markers. These orienting markers could be visualized on MRI and during manual processing. Samples were kept at room temperature, and MRI was performed within 4 hours of sample collection.
对于 MRI,外植体安装在带有定向切口的非金属支架和夹具上(图(图 1A) 1 A),并固定在含有 PBS 的聚乙烯罐中(图(图 1B) 。1 B)。将鱼肝油丸固定在外部以提供进一步的位置标记。这些定向标记可以在 MRI 上和手动处理过程中可视化。样品保存在室温下,并在样品采集后 4 小时内进行 MRI 检查。
MR image acquisition and analysis
MR 图像采集和分析
MR images were obtained with a Verio 3.0T MRI system (Siemens). MRI was performed using fat‐suppressed, fast spin‐echo, proton density–weighted sequences. T1‐weighted sequences were obtained in 1 plane to identify BMLs (Figure (Figure1C).1C). To ensure accurate cutting measurements in 3 dimensions for separation of BML from non‐BML regions, MRI sequences were obtained in 3 planes (axial, coronal, and sagittal) (Figure (Figure1C).1C). BMLs were defined as areas of increased signal intensity adjacent to subcortical bone on proton density–weighted sequences that had low intensity on T1‐weighted sequences (Figure (Figure1C).1C). For additional information, including MRI sequence settings, see Supplementary Table 1 (available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract).
MR 图像是用 Verio 3.0T MRI 系统(西门子)获得的。 MRI 使用脂肪抑制、快速自旋回波、质子密度加权序列进行。在 1 个平面中获得 T1 加权序列以识别 BML(图(图 1C) 。1 C)。为了确保在 3 个维度上进行准确的切割测量,以将BML与非BML区域分离,在 3 个平面(轴向、冠状和矢状)中获得了 MRI 序列(图(图 1C)。1 C)。 BML 被定义为质子密度加权序列上与皮质下骨相邻的信号强度增加的区域,而 T1 加权序列上的信号强度较低(图(图1C) 。1 C)。有关其他信息,包括 MRI 序列设置,请参阅补充表 1(可在Arthritis & Rheumatology网站http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract上获取)。
MR images were examined using Siemens syngo fastView software for DICOM images. Identification of BML and non‐BML regions was first performed by a physiatrist with experience in musculoskeletal imaging (TMC).
使用西门子 syngo fastView 软件检查 DICOM 图像的 MR 图像。 BML和非BML区域的识别首先由具有肌肉骨骼成像 (TMC) 经验的理疗师进行。
Interpretation was repeated independently by a radiologist with expertise in musculoskeletal imaging (RH), with blinding with regard to all patient data and prior interpretation. Consensus regarding BML and non‐BML regions was then reached by discussion.
由具有肌肉骨骼成像(RH)专业知识的放射科医生独立重复解释,并对所有患者数据和先前的解释进行盲法。然后通过讨论就BML和非BML区域达成共识。
Because the patients recruited into the study had severe OA, we acknowledge that non‐BML regions are unlikely to represent “normal bone”; however, the most normal‐appearing bone was selected.
由于参与研究的患者患有严重的 OA,我们承认非BML区域不太可能代表“正常骨骼”;然而,选择了外观最正常的骨头。
Excision of specimens from BML and non‐BML regions, and downstream tissue processing
从 BML 和非 BML 区域切除标本,以及下游组织处理
MRI measurements were performed in all 3 proton density–weighted sequence planes (Figure (Figure1C).1C). Bone cutting measurements were planned using the measuring tool of the image visualization software. Measurements and sample orientation were accomplished with reference to sample anatomic landmarks (e.g., outer cortex apex, ligamentum teres, osteophytes).
在所有 3 个质子密度加权序列平面上进行 MRI 测量(图(图 1C) 。1 C)。使用图像可视化软件的测量工具计划骨切割测量。测量和样品定向是参考样品解剖标志(例如,外皮质顶点、圆韧带、骨赘)来完成的。
A sterile marker was used to mark orienting lines on the sample corresponding to the anteroposterior and mediolateral landmarks on the bracket, so that orientation was maintained.
使用无菌标记在样品上标记对应于支架上的前后和中间侧界标的定向线,从而保持定向。
All bone processing was performed in a tissue culture hood, using sterile technique.
所有骨处理均在组织培养罩中使用无菌技术进行。
After the femoral heads were removed from the bracket, they were placed in a vice, and the remainder of the femoral neck was used to hold the sample to prevent damage to the subchondral regions (Figure (Figure1D).1D). BML and non‐BML portions of bone were removed from the femoral head, referencing the 3‐dimensional cut plan (Figure (Figure1D).1D). Because BMLs have an ill‐defined border, a 2–3‐mm gap was left between BML and non‐BML regions to ensure that BML bone was well separated from non‐BML bone.
将股骨头从支架上取下后,将其放置在虎钳中,并使用股骨颈的其余部分来固定样本,以防止损坏软骨下区域(图(图1D)。1 D)。参照 3 维切割计划,从股骨头中去除骨的BML和非BML部分(图(图 1D) 。1 D)。由于 BML 的边界不明确,因此BML和非BML区域之间留有 2-3 毫米的间隙,以确保BML骨与非BML骨良好分离。
Excised pieces of BML and non‐BML regions were each divided into 2 portions: one for histology and one for cell extraction following enzymatic release 11. Samples prepared for collagenase treatment were minced using a rongeur (Figure (Figure1E)1E) and then placed in low‐glucose Dulbecco's modified Eagle's medium (DMEM; Life Technologies) with 20% fetal calf serum (FCS) (Sigma) and animal origin–free collagenase (3,000 units/gm bone) (Worthington Biochemical Corporation) for 4 hours at 37°C 11. After completion of the collagenase treatment, a fraction of the cells (average 1.5 × 105) was obtained for flow cytometry, and the rest of the cells were frozen in FCS supplemented with 10% dimethylsulfoxide (Sigma) for later use. Samples for histologic analysis were placed in 10% formalin (Sigma) before processing.
BML 和非 BML 区域的切除片段均分为 2 部分:一部分用于组织学,另一部分用于酶释放后的细胞提取11 。使用咬骨钳将准备用于胶原酶处理的样品切碎(图(图1E) 1 E),然后放入含有 20% 胎牛血清 (FCS) (Sigma) 和动物的低葡萄糖 Dulbecco 改良 Eagle 培养基(DMEM;Life Technologies)中无来源胶原酶(3,000 单位/克骨)(沃辛顿生化公司)37°C 4 小时11 。胶原酶处理完成后,获得一部分细胞(平均1.5×10 5 )用于流式细胞术,其余细胞冷冻在补充有10%二甲亚砜(Sigma)的FCS中备用。用于组织学分析的样品在处理前置于 10% 福尔马林 (Sigma) 中。
Histologic and immunohistochemical analysis
组织学和免疫组织化学分析
For each patient, separate BML and non‐BML femoral head pieces were decalcified using 12.5% EDTA (Sigma) in deionized water for 3–4 months and then mounted on paraffin blocks.
对于每位患者,使用去离子水中的 12.5% EDTA (Sigma) 将单独的BML和非BML股骨头脱钙 3-4 个月,然后安装在石蜡块上。
Decalcified tissue specimens were stained with hematoxylin and eosin or Safranin O, using standard protocols.
使用标准方案,用苏木精和伊红或番红 O 对脱钙组织样本进行染色。
Digital image analysis was performed to evaluate relative trabecular bone area and cartilage damage. For each sample, the whole tissue area was scanned using a Nikon E1000 microscope under brightfield mode and a multispectral Nuance camera (PerkinElmer).
进行数字图像分析以评估相对骨小梁面积和软骨损伤。对于每个样本,使用 Nikon E1000 显微镜在明场模式下和多光谱 Nuance 相机 (PerkinElmer) 扫描整个组织区域。
Using the femoral head cartilage surface to orient the tissue, the overall section was then separated into cartilage (including the cartilage–bone interface) and bone. The presence of cysts in the bone area (Figure (Figure2A)2A) was taken into consideration, and tissue that was immediately adjacent was not used for quantitative bone area measurements.
利用股骨头软骨表面对组织进行定向,然后将整个切片分为软骨(包括软骨-骨界面)和骨。考虑到骨区域中存在囊肿(图(图2A) 2 A),并且紧邻的组织不用于定量骨区域测量。
Depending on the size of the section, at least 5 (nonoverlapping) images were captured for the bone area and 2–5 images were captured for the cartilage area.
根据切片的大小,骨骼区域至少捕获 5 个(不重叠)图像,软骨区域捕获 2-5 个图像。
Nuance version 3.0.1.2 software (Caliper Life Sciences) was used for digital image analysis.
Nuance 版本 3.0.1.2 软件(Caliper Life Sciences)用于数字图像分析。
Educating the software to recognize trabecular bone was done by manually determining small representative areas of bone and repeating the process until >95% accuracy in identifying bone was achieved automatically.
通过手动确定骨骼的小代表性区域并重复该过程,直到自动实现识别骨骼的 >95% 准确度,来训练软件识别骨小梁。
For each tissue section, the full area was measured, and the trabecular bone area was calculated as a percentage (mean ± SD of a minimum of 5 images).
对于每个组织切片,测量整个面积,并将骨小梁面积计算为百分比(至少 5 个图像的平均值±SD)。
For cartilage assessment, 10 positions were spaced out over the length of available cartilage in each image and repeated over 2–5 images (depending on the size of the tissue).
对于软骨评估,在每张图像中可用软骨的长度上间隔开 10 个位置,并在 2-5 个图像上重复(取决于组织的大小)。
Cartilage appearance was classified as “less damaged” or “more damaged” based on superficial zone smoothness, clefts, fibrillation, and presence of sclerotic bone or reparative tissue within denuded surface 16 (Figures (Figures2D2D and E).
根据浅表区域的光滑度、裂缝、纤维颤动以及裸露表面内是否存在硬化骨或修复组织,软骨外观被分为“受损程度较低”或“受损程度较高” (图(图2D 、2D和E))。
Immunohistochemistry for CD271 staining was performed as optimized by Tormin et al 12. Mouse anti‐human CD271 monoclonal antibody (Abcam) was used at a dilution of 1:50. (For a complete list of reagents, see Supplementary Table 2, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract).
CD271 染色的免疫组织化学按照 Tormin 等人的优化进行12 。小鼠抗人 CD271 单克隆抗体 (Abcam) 以 1:50 的稀释度使用。 (有关试剂的完整列表,请参阅补充表 2,可在关节炎和风湿病学网站http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract上获取)。
Flow cytometry 流式细胞术
For MSC enumeration, flow cytometry was performed on freshly enzymatically treated samples using a BD LSRII flow cytometer (BD Biosciences). Depending on the cellularity of collagenase‐treated samples, ∼1–2 × 105 cells were resuspended in 50 μl of fluorescence‐activated cell sorting (FACS) buffer (PBS plus 0.5% bovine serum albumin) and incubated in a 10% Fc receptor–blocking reagent solution (Miltenyi Biotec) before fluorophore‐conjugated antibodies were added.
对于 MSC 计数,使用 BD LSRII 流式细胞仪(BD Biosciences)对新鲜酶处理的样品进行流式细胞术。根据胶原酶处理样品的细胞结构,将 ~1–2 × 10 5 个细胞重悬于 50 μl 荧光激活细胞分选 (FACS) 缓冲液(PBS 加 0.5% 牛血清白蛋白)中,并在 10% Fc 受体中孵育– 添加荧光团缀合抗体之前的封闭试剂溶液(Miltenyi Biotec)。
Staining was performed for CD90, CD73, CD45, and CD271 (for additional information, see Supplementary Table 3, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract), and dead cells were excluded using live cell marker calcein violet and dead cell marker aqua‐fluorescent reactive dye (Invitrogen) 17. The proportion of MSCs gated as CD45−CD271+ cells 11, 12 was calculated relative to total live cells (Figure (Figure3A).3A). MSC extended phenotype was investigated using CD73 and CD90 markers. The percentages of lymphocytes 18 were similarly calculated relative to the total number of live cells.
对 CD90、CD73、CD45 和 CD271 进行染色(更多信息,请参见补充表 3,可在关节炎和风湿病学网站http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract上获取) ),并使用活细胞标记物钙黄绿素紫和死细胞标记物水荧光活性染料(Invitrogen)排除死细胞17 。相对于总活细胞计算门控为 CD45−CD271+ 细胞11 、 12的 MSC 比例(图(图 3A) 。3 A)。使用 CD73 和 CD90 标记物研究 MSC 扩展表型。类似地计算淋巴细胞18相对于活细胞总数的百分比。
The MSC identity of BML‐ and non‐BML–derived adherent cultures was investigated following staining with a standard panel of antibodies defining the phenotype of cultured MSCs 19. Passage 3 cultures grown from BML‐ and non‐BML CD271–selected cells were trypsinized and resuspended at 107 cells/ml in FACS buffer.
使用定义培养的 MSC 表型的标准抗体组染色后,研究了BML和非BML来源的贴壁培养物的 MSC 身份19 。将BML和非BML CD271 选择的细胞生长的第 3 代培养物进行胰蛋白酶处理,并以 10 7 个细胞/ml 的浓度重悬于 FACS 缓冲液中。
Antibody combinations included the following: phycoerythrin–Cy7 (PE–Cy7)–conjugated CD45, PerCP‐conjugated CD34, allophycocyanin (APC)–conjugated CD271, APC‐H7–conjugated CD14, and fluorescein isothiocyanate (FITC)–conjugated CD19 and PE‐labeled antibodies, including PE‐conjugated CD73 (ecto‐5′‐nucleotidase), PE‐conjugated CD105 (endoglin), and PE‐conjugated CD90 (Thy‐1) (for additional information, see Supplementary Table 3).
抗体组合包括以下内容:藻红蛋白-Cy7 (PE-Cy7)-缀合CD45、PerCP-缀合CD34、别藻蓝蛋白(APC)-缀合CD271、APC-H7-缀合CD14和异硫氰酸荧光素(FITC)-缀合CD19和PE-标记抗体,包括 PE 缀合的 CD73(外切-5′-核苷酸酶)、PE 缀合的 CD105(内皮糖蛋白)和 PE 缀合的 CD90 (Thy-1)(更多信息,请参见补充表 3)。
All antibodies were used at the concentrations recommended by the manufacturers, with matched isotype controls. Dead/dying cells (normally <5% of total cells) were excluded from the analysis using 10 μl/ml DAPI (Sigma).
所有抗体均按制造商推荐的浓度使用,并具有匹配的同种型对照。使用 10 μl/ml DAPI (Sigma) 将死亡/垂死细胞(通常占总细胞的 <5%)排除在分析之外。
All flow cytometry data were analyzed using Diva version 6.2 software (BD Biosciences).
所有流式细胞术数据均使用 Diva 6.2 版软件(BD Biosciences)进行分析。
Colony‐forming unit–fibroblast (CFU‐F) assay and MSC expansion from BML and non‐BML collagenase digests
集落形成单位-成纤维细胞 (CFU-F) 测定以及 BML 和非 BML 胶原酶消化物中的 MSC 扩增
A CFU‐F assay was performed as described previously 11, with a minor modification using methylene blue, before scoring was performed in a blinded manner. MSC expansion was used to measure MSC proliferation rates and to produce a sufficient number of cells for trilineage differentiation assays and gene expression analysis.
如前所述进行 CFU-F 测定11 ,并使用亚甲基蓝进行细微修改,然后以盲法进行评分。 MSC 扩增用于测量 MSC 增殖率并产生足够数量的细胞用于三系分化测定和基因表达分析。
MSCs were expanded in StemMACS MSC Expansion Media after preenrichment using CD271 MACSelect MicroBeads (both from Miltenyi Biotec), and culture population doublings were calculated as previously described 20. MSCs from the bone of healthy controls and patients with OP were expanded similarly following their enzymatic release from bone 11.
使用 CD271 MACSelect MicroBeads(均来自 Miltenyi Biotec)预富集后,MSC 在 StemMACS MSC 扩增培养基中扩增,并按先前所述计算培养物群体倍增20 。来自健康对照者和 OP 患者骨骼的 MSC 在从骨骼中释放酶后也进行了类似的扩增11 。
Trilineage differentiation
三系分化
Passage 2/3 MSCs (n = 5 matched donor–derived cultures for BML and non‐BML bone tissue digests) were induced toward osteogenesis, chondrogenesis, and adipogenesis, using standard protocols 21. For osteogenesis and chondrogenesis, we used StemMACS OsteoDiff and ChondroDiff medium, respectively (Miltenyi Biotec); adipogenic cultures were grown in DMEM with 10% FCS, antibiotics, 10% horse serum (StemCell Technologies), 0.5 mM isobutylmethylxanthine, 60 μM indomethacin, and 0.5 μM hydrocortisone (all from Sigma).
使用标准方案21诱导第 2/3 代 MSC(n = 5 匹配的供体来源的 BML 和非 BML 骨组织消化培养物)向成骨、软骨生成和脂肪生成方向诱导。对于成骨和软骨形成,我们分别使用 StemMACS OsteoDiff 和 ChondroDiff 培养基(Miltenyi Biotec);脂肪形成培养物在含有 10% FCS、抗生素、10% 马血清 (StemCell Technologies)、0.5 m M异丁基甲基黄嘌呤、60 μM吲哚美辛和 0.5 μM氢化可的松(均来自 Sigma)的 DMEM 中生长。
Differentiation assessment was performed as previously described 21. Briefly, alkaline phosphatase activity was visualized on day 14 postinduction. Calcium deposits were stained using alizarin red on day 21, and total calcium produced by cultures was measured using a Calcium Detection Kit (Sentinel Diagnostics).
如前所述进行分化评估21 。简而言之,在诱导后第14天观察碱性磷酸酶活性。第 21 天使用茜素红对钙沉积物进行染色,并使用钙检测试剂盒(Sentinel Diagnostics)测量培养物产生的总钙。
Biochemical assessment of the glycosaminoglycans (GAGs) was performed on 3 of 4 chondrogenic pellets grown for 21 days. The remaining pellet was used for histologic analysis; 4‐μm sections were cut using a Leica CM1950 cryostat, fixed, and stained with toluidine blue.
对生长 21 天的 4 个软骨形成颗粒中的 3 个进行糖胺聚糖 (GAG) 生化评估。剩余的沉淀用于组织学分析;使用 Leica CM1950 低温恒温器切割 4 μm 切片,固定并用甲苯胺蓝染色。
Adipogenic cultures were stained with oil red O on day 21 postinduction.
诱导后第21天用油红O对脂肪形成培养物进行染色。
Real‐time qPCR 实时定量PCR
To investigate MSC molecular profiles, paired CD271 bead–selected BML and non‐BML passage 2 cultures were analyzed for their relative expression of genes involved in MSC tripotentiality, collagen metabolism, chemotaxis, angiogenesis, and control of osteoclast activation.
为了研究 MSC 分子谱,分析了配对 CD271 珠选择的BML和非BML第 2 代培养物中涉及 MSC 三潜能、胶原代谢、趋化性、血管生成和破骨细胞活化控制的基因的相对表达。
Other selected genes included those previously described as being associated with OA 3, 22 (see also Supplementary Table 4, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract). MSCs derived from trabecular bone from age‐matched controls as well as MSCs derived from the femoral heads of patients with OP were included as controls.
其他选定的基因包括先前描述为与 OA 3 , 22相关的基因(另见补充表 4,可在关节炎和风湿病学网站http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract上找到) )。源自年龄匹配对照的小梁骨的 MSC 以及源自 OP 患者股骨头的 MSC 作为对照。
Reverse transcription and qPCR were performed using a custom Format 48 TaqMan low‐density array (Life Technologies) (see Supplementary Table 4), as previously described 20. Mean fold changes were calculated and were considered further if the change was ≥2‐fold. Selected transcripts were validated on additional samples using individual TaqMan assays, matching those included in the TaqMan low‐density array.
使用定制的 Format 48 TaqMan 低密度阵列(Life Technologies)进行逆转录和 qPCR(参见补充表 4),如前所述20 。计算平均倍数变化,如果变化≥2倍则进一步考虑。使用单独的 TaqMan 检测在其他样品上验证选定的转录本,与 TaqMan 低密度阵列中包含的转录本相匹配。
Tests were performed in triplicate on 5 ng complementary DNA per well.
每孔 5 ng 互补 DNA 进行三次重复测试。
Statistical analysis 统计分析
Differences between paired BML and non‐BML samples for trabecular bone area, flow cytometric measurements of MSC and lymphocyte proportions, CFU‐F assays, MSC growth, gene expression, and differentiation data were compared using the 2‐sample paired sign test. Gene expression differences using control MSC cultures (3 groups: OA, healthy control, and OP MSCs) were tested by Kruskal‐Wallis test with Dunn's multiple comparison tests.
使用 2 样本配对符号检验比较配对BML和非BML样本之间的骨小梁面积、MSC 和淋巴细胞比例的流式细胞术测量、CFU-F 测定、MSC 生长、基因表达和分化数据之间的差异。使用对照 MSC 培养物(3 组:OA、健康对照和 OP MSC)的基因表达差异通过 Kruskal-Wallis 检验和 Dunn 多重比较检验进行测试。
The chi‐square test was used to establish associations between cartilage appearance (less damaged versus more damaged) and BML/non‐BML images. P values less than or equal to 0.05 were considered significant. All tests were performed using IBM SPSS Statistics 21.
卡方检验用于建立软骨外观(受损较少与受损较多)和BML /非BML图像之间的关联。 P值小于或等于 0.05 被认为是显着的。所有测试均使用 IBM SPSS Statistics 21 执行。
RESULTS 结果
Subject recruitment 受试者招募
Twenty‐one patients were considered for recruitment into the ex vivo MRI study, but 1 of the patients showed no detectable BMLs on MRI and was excluded. Half of the patients were women, and all were white.
考虑招募 21 名患者参加离体 MRI 研究,但其中 1 名患者在 MRI 上未显示出可检测到的 BML,因此被排除在外。一半的患者是女性,而且都是白人。
The median age was 65.3 years (range 48–83 years), and the median body mass index was 28.4 kg/m2 (range 20.7–42.8). Half of the patients were receiving nonsteroidal antiinflammatory drugs, and none were receiving bisphosphonates. Twenty‐five percent of the patients were smokers, and 85% had Kellgren/Lawrence grade 4 hip OA 23.
中位年龄为 65.3 岁(范围 48-83 岁),中位体重指数为 28.4 kg/m 2 (范围 20.7-42.8)。一半患者正在接受非甾体抗炎药治疗,没有人接受双磷酸盐治疗。 25% 的患者是吸烟者,85% 患有 Kellgren/Lawrence 4 级髋关节 OA 23 。
Histologic features segregating BML from non‐BML regions
将 BML 与非 BML 区域分开的组织学特征
To confirm the accuracy of our BML sampling, we performed histologic assessment of bone and cartilage in paired BML and non‐BML samples from 14 randomly selected patients. BMLs have been characterized by a high bone volume fraction 24. For trabecular area analysis, 320 images of bone were acquired: 198 images for BML samples (average of 9.3 images per sample [range 5–14]) and 122 images for non‐BML samples (average of 6.6 images per sample [range 5–9]).
为了确认BML采样的准确性,我们对来自 14 名随机选择的患者的配对BML和非BML样本进行了骨和软骨的组织学评估。 BML 的特点是骨体积分数高24 。对于小梁区域分析,采集了 320 个骨骼图像:198 个BML样本图像(每个样本平均 9.3 个图像[范围 5-14])和 122 个非BML样本图像(每个样本平均 6.6 个图像[范围 5-14])。 9])。
The average trabecular bone area per sample (n = 14 pairs) was widely distributed between the patients; however, a clearly higher trabecular area was observed in BML compared with paired non‐BML samples (P = 0.001) (Figures (Figures2C,2C, F, and G).
每个样本(n = 14 对)的平均骨小梁面积在患者之间分布广泛;然而,与配对的非BML样本相比,在BML中观察到明显更高的小梁面积( P = 0.001)(图(图 2C、 2 C、F 和 G)。
BMLs were also associated with overlying cartilage defects 25, 26. Cartilage assessment was performed in the group of samples from 14 donors.
BML 还与上覆软骨缺损相关25 , 26 。对来自 14 名捐赠者的样本组进行了软骨评估。
Following decalcification, 9 of 14 paired tissue samples had sufficient amounts/quality of cartilage to enable paired analysis, and 74 images were acquired for the cartilage region: 43 BML (average of 5 images per sample [range 2–9]) and 31 non‐BML (average of 3 images per sample [range 2–14]).
脱钙后,14 个配对组织样本中的 9 个具有足够数量/质量的软骨以进行配对分析,并获取了软骨区域的 74 个图像:43 个BML (每个样本平均 5 个图像 [范围 2-9])和 31 个非‐ BML (每个样本 3 个图像的平均值 [范围 2-14])。
Images showing more damaged cartilage (assessed by comparing the relative smoothness of cartilage, presence of clefts, fibrillation, and sclerotic bone or reparative tissue within denuded surface [16]) were closely associated with BML samples (24 of 43 images) compared with non‐BML samples (5 of 31 images; P = 0.01), where most of the cartilage surface was considered to be less damaged (Figures (Figures2D2D and E). Taken together, these findings were consistent with the expected histologic features of BML and non‐BML regions, confirming the accuracy of the excision method.
与非 BML 样本(43 幅图像中的 24 幅)相比,显示更多受损软骨的图像(通过比较裸露表面内软骨的相对光滑度、裂缝、纤维颤动和硬化骨或修复组织进行评估[16])与BML样本密切相关(43 幅图像中的 24 幅)。 BML样本(31 个图像中的 5 个; P = 0.01),其中大部分软骨表面被认为损伤较小(图(图2D 2 D 和 E)。总而言之,这些发现与BML的预期组织学特征一致和非BML区域,证实了切除方法的准确性。
MSC enumeration by flow cytometry and CFU‐F assay
通过流式细胞术和 CFU-F 测定进行 MSC 计数
Representative flow cytometry plots for CD45−CD271+ MSC populations are shown in Figure Figure3A.3A. Despite broad donor‐to‐donor variation, a greater proportion of CD45−CD271+ cells as a percentage of total live cells was observed in BML compared with non‐BML tissue digests (median difference 5.6‐fold; P < 0.001) (Figure (Figure3B).3B). The expression of 2 additional markers (CD73 and CD90) was also assessed on CD45−CD271+ cells to confirm their MSC identity. In both BML and non‐BML regions, >85% of CD45−CD271+ cells were CD73+ (for BML, mean 87.5%; for non‐BML, mean 89.5% [P = not significant]). The CD45−CD271+CD90+ cell subpopulation, a recently described phenotype of the most clonogenic MSCs 27, was also higher in BML regions (median 1.7‐fold; P = 0.041) (Figure (Figure3B).3B). In contrast to the observed differences in MSC numbers, no difference in the percentages of lymphocytes (gated as CD45brightSSClow cells [18]) in BML and non‐BML digests was observed (P = 0.824) (Figure (Figure3C).3C). To further validate this difference in MSC frequency, a CFU‐F assay was performed in 14 pairs of tissue digests, and the same trend was observed (median difference 4.3‐fold; P = 0.013) (Figure (Figure33D).
CD45−CD271+ MSC 群体的代表性流式细胞术图如图3A 所示。 3 A. 尽管供体之间存在广泛差异,但与非 BML 组织消化物相比,在 BML 中观察到 CD45−CD271+ 细胞占总活细胞的比例更高(中位差异 5.6 倍; P < 0.001)(图(图3B ) 。还评估了 CD45−CD271+ 细胞上 2 个附加标记(CD73 和 CD90)的表达,以确认其 MSC 身份。在 BML 和非 BML 区域,>85% 的 CD45−CD271+ 细胞为 CD73+(对于 BML,平均为 87.5%;对于非 BML,平均为 89.5% [ P = 不显着])。 CD45−CD271+CD90+ 细胞亚群是最近描述的最具克隆性的 MSC 27的表型,在 BML 区域也较高(中位数 1.7 倍; P = 0.041)(图(图 3B) 。3 B)。与观察到的 MSC 数量差异相反,BML 和非 BML 消化物中的淋巴细胞百分比(门控为 CD45亮SSC低细胞 [18])没有观察到差异( P = 0.824)(图(图 3C)。 3C )。为了进一步验证 MSC 频率的这种差异,在 14 对组织消化物中进行了 CFU-F 测定,并观察到相同的趋势(中位数差异 4.3 倍; P = 0.013)(图(图3 D)。
Topography of CD271+ MSCs in excised BML and non‐BML specimens
切除的 BML 和非 BML 标本中 CD271+ MSC 的形貌
We next performed immunohistochemical analysis to investigate the localization of CD271+ cells within the excised samples. In both BML and non‐BML tissues, CD271 staining was distributed as expected in a perivascular and reticular pattern in marrow cavities (Figure (Figure4B)4B) 12, 13. Additionally, CD271 positivity was clearly detectable in bone lining locations (Figure (Figure44C).
接下来,我们进行了免疫组织化学分析,以研究切除样本中 CD271+ 细胞的定位。在 BML 和非 BML 组织中,CD271 染色按预期分布在骨髓腔中的血管周围和网状图案中(图(图 4B) 4 B) 12 , 13 。此外,在骨衬位置可以清楚地检测到 CD271 阳性(图(图 4 4 C)。
A highly heterogeneous distribution of CD271 positivity (due to the large intersubject heterogeneity of bone pathology and cartilage OA architectural changes) did not allow reliable quantification of CD271+ cells using digital imaging analysis to directly compare excised non‐BML and BML specimens.
CD271 阳性的高度异质性分布(由于骨病理学和软骨 OA 结构变化的受试者间异质性较大)不允许使用数字成像分析对 CD271+ 细胞进行可靠的定量,以直接比较切除的非BML和BML标本。
However, in the BML samples, accumulation of CD271 staining was particularly evident in the regions adjacent to subchondral bone cysts (Figure (Figure4D)4D) and cartilage damage (Figures (Figures4D–F)4D–F) at osteochondral junctions where overlying cartilage loss was more pronounced. In addition to perivascular staining (Figure (Figure4F),4F), there was substantial staining of fibrous stromal tissue extending toward and up to the cement line from the subchondral bone (Figures E and F), suggesting that MSCs had accumulated at regions of cartilage damage.
然而,在BML样本中,CD271 染色的积累在邻近软骨下骨囊肿的区域(图(图 4D) 4 D)和骨软骨连接处的软骨损伤(图(图 4D–F) 4 D–F)尤其明显。其中上覆软骨损失更为明显。除了血管周围染色(图(图4F), 4 F)外,纤维基质组织也有大量染色,从软骨下骨延伸到骨水泥线(图E和F),表明间充质干细胞在这些区域积累软骨损伤。
Stained stromal tissue was often seen invading more damaged cartilage “from below” in BMLs (see Supplementary Figures 1A and B, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract); however, there was no positive CD271+ staining within the cartilage itself (Figure (Figure44G).
在 BML 中,经常可以看到染色的基质组织“从下方”侵入更多受损的软骨(参见补充图 1A 和 B,可在关节炎和风湿病学网站http://onlinelibrary.wiley.com/doi/10.1002/art.39622上找到) /抽象的);然而,软骨本身内没有阳性 CD271+ 染色(图(图 4 4 G)。
In vitro growth and differentiation capacities of CD271+ MSC–derived cultures
CD271+ MSC 衍生培养物的体外生长和分化能力
CD271+ bead–selected cells from BML and non‐BML digests possessed the standard MSC phenotype following culture expansion (Figure (Figure5A).5A). To test whether BML‐resident CD271+ MSCs had altered functional capacities, the growth kinetics of CD271+ cell cultures were examined. BML cultures had slightly longer population doubling times compared with non‐BML cultures (P = 0.049 for all passages combined; n = 5 donors) (Figure (Figure55B).
来自 BML 和非 BML 消化物的 CD271+ 珠子选择的细胞在培养扩增后具有标准 MSC 表型(图(图 5A) 。5 A)。为了测试 BML 驻留 CD271+ MSC 是否改变了功能能力,检查了 CD271+ 细胞培养物的生长动力学。与非 BML 培养物相比,BML 培养物的群体倍增时间稍长(所有传代组合的P = 0.049;n = 5 个供体)(图(图5 B)。
Regarding their differentiation capabilities, paired BML and non‐BML cultures showed similar levels of adipogenesis (Figure (Figure5C).5C). No differences were observed in alkaline phosphatase staining on day 14 postinduction of osteogenesis (Figure (Figure5D);5D); however, on day 21 postinduction, BML cultures produced lower amounts of calcium (P = 0.043) (Figure (Figure5D).5D). No obvious trends were observed in chondrogenesis assays, assessed either qualitatively (chondrogenic pellet staining with toluidine blue) or quantitatively (GAG assay) (Figure (Figure55E).
关于它们的分化能力,配对的BML和非BML培养物显示出相似的脂肪生成水平(图(图 5C) 。5 C)。成骨诱导后第 14 天,碱性磷酸酶染色未观察到差异(图(图 5D); 5 D);然而,在诱导后第 21 天, BML培养物产生的钙量较低( P = 0.043)(图(图 5D) 。5 D)。在软骨形成测定中没有观察到明显的趋势,无论是定性评估(用甲苯胺蓝进行软骨形成颗粒染色)还是定量评估(GAG 测定)(图(图5 E)。
Comparative gene expression signatures of MSCs from BML and non‐BML digests
BML和非BML消化物中 MSC 的基因表达特征比较
The expression of 46 genes involved in MSC function, collagen metabolism, chemotaxis, angiogenesis, and control of osteoclast activation was measured using qPCR in CD271+ cell–derived cultured MSCs from OA patients (n = 7 BML/non‐BML pairs).
使用 qPCR 测量来自 OA 患者的 CD271+ 细胞来源的培养 MSC 中涉及 MSC 功能、胶原代谢、趋化性、血管生成和破骨细胞活化控制的 46 个基因的表达(n = 7 BML /非BML对)。
The 2 differentially expressed bone‐related genes between BML and non‐BML cultures (CXCR4 and TNFSF11) (Figure (Figure6A)6A) were subsequently validated by individual TaqMan assays and flow cytometry. Consistent with TaqMan low‐density array data, expression of the receptor for stromal cell–derived factor 1 (SDF‐1), CXCR4, was lower in BML MSCs, using qPCR (Figure (Figure6B).6B). CXCR4 surface protein was present only in a small percentage 28 of cells in all 5 paired BML/non‐BML MSC cultures (see Supplementary Figure 2A, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract). The mean fluorescence intensity of the surface protein RANKL (encoded by TNFSF11) was lower in 4 of 5 cultures of BML MSCs compared with non‐BML MSCs, as demonstrated using flow cytometry (Figure (Figure6B).6B). The remaining genes were not differentially expressed between BML and non‐BML digests.
BML和非BML培养物之间的 2 个差异表达的骨相关基因( CXCR4和TNFSF11 )(图(图 6A) 6 A)随后通过单独的 TaqMan 测定和流式细胞术进行了验证。与 TaqMan 低密度阵列数据一致,使用 qPCR 时, BML MSC 中基质细胞衍生因子 1 (SDF-1) 受体CXCR4的表达较低(图(图 6B) 。6 B)。 CXCR4 表面蛋白仅存在于所有 5 配对BML /非BML MSC 培养物中的一小部分细胞中(参见补充图 2A,可在关节炎和风湿病学网站http://onlinelibrary.wiley.com/doi上找到) /10.1002/art.39622/摘要)。与非BML MSC 相比,5 种BML MSC 培养物中的 4 种表面蛋白 RANKL(由TNFSF11编码)的平均荧光强度较低,如流式细胞术所示(图(图 6B) 。6 B)。其余基因在BML和非BML消化物之间没有差异表达。
Comparative gene expression signatures of MSCs from OA, OP, and healthy bone
来自 OA、OP 和健康骨的 MSC 的基因表达特征比较
To determine whether differential MSC gene expression in femoral heads was a feature of hip OA, we compared all OA MSC transcripts (averaging BML and non‐BML TaqMan low‐density array data for the genes that were not differentially expressed) with healthy control MSCs (n = 5) and OP MSCs from femoral heads (n = 3).
为了确定股骨头中差异性 MSC 基因表达是否是髋关节 OA 的一个特征,我们将所有 OA MSC 转录本(未差异表达的基因的BML和非BML TaqMan 低密度阵列数据的平均值)与健康对照 MSC 进行了比较( n = 5) 和来自股骨头的 OP MSC (n = 3)。
When OA MSCs were compared with healthy control MSCs, significantly different levels were observed for 8 transcripts. Transcript levels for 5 of these transcripts were higher in OA MSCs: CXCR1/interleukin‐8 (IL‐8) receptor α‐chain and CCR6/macrophage inflammatory protein 1α (MIP‐1α) receptor (mostly below detection in healthy control MSCs), GDF5/growth differentiation factor 5 (8‐fold), MMP1/matrix metalloproteinase 1 (23‐fold), and TGFBR2/transforming growth factor β receptor 2 (2‐fold). The levels of another 3 transcripts were lower: ACAN/aggrecan (2‐fold), NTRK1/high‐affinity nerve growth factor receptor (10‐fold), and NGFR/low‐affinity nerve growth factor receptor (2‐fold). We next used qPCR (Figure (Figure6C)6C) and flow cytometry (see Supplementary Figure 2B, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract) to validate some of these putative OA‐specific genes, using additional cultures of OA MSCs (n = 7), and compared expression with that in healthy control MSCs (n = 9).
当 OA MSC 与健康对照 MSC 进行比较时,观察到 8 个转录物的水平存在显着差异。其中 5 个转录物的转录水平在 OA MSC 中较高: CXCR1 /白细胞介素 8 (IL-8) 受体 α 链和CCR6 /巨噬细胞炎症蛋白 1α (MIP-1α) 受体(大部分低于健康对照 MSC 中的检测值), GDF5 /生长分化因子5(8倍)、 MMP1 /基质金属蛋白酶1(23倍)和TGFBR2 /转化生长因子β受体2(2倍)。另外 3 个转录物的水平较低: ACAN /聚集蛋白聚糖(2 倍)、 NTRK1 /高亲和力神经生长因子受体(10 倍)和NGFR /低亲和力神经生长因子受体(2 倍)。接下来我们使用了 qPCR(图(图 6C) 6 C)和流式细胞术(参见补充图 2B,可在关节炎和风湿病学网站http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract上找到)使用额外的 OA MSC 培养物 (n = 7) 来验证其中一些假定的 OA 特异性基因,并将其表达与健康对照 MSC (n = 9) 中的表达进行比较。
We were mindful that these differences could be attributable to OA MSCs and healthy control MSCs having been derived from anatomically different bones (femoral head versus iliac crest 29); therefore, we analyzed OP MSCs (n = 5), which were also derived from femoral head bone. This analysis confirmed that expression of CXCR1, CCR6, and GDF5 was significantly different across the 3 groups (Figure (Figure6C),6C), and that CXCR1 and CCR6 were OA‐specific (as shown by significantly higher expression in OA MSCs compared with both healthy control MSCs and OP MSCs).
我们注意到,这些差异可能归因于 OA MSC 和健康对照 MSC 源自解剖学上不同的骨骼(股骨头与髂嵴29 );因此,我们分析了同样源自股骨头的 OP MSC (n = 5)。该分析证实, CXCR1 、 CCR6和GDF5的表达在 3 组中存在显着差异(图(图 6C)、 6 C),并且CXCR1和CCR6是 OA 特异性的(如与与健康对照 MSC 和 OP MSC 一起)。
CXCR1 surface protein expression was also higher in OA MSCs compared with healthy control MSCs and OP MSCs, and shown using flow cytometry (see Supplementary Figure 2B, available on the Arthritis & Rheumatology web site at http://onlinelibrary.wiley.com/doi/10.1002/art.39622/abstract).
与健康对照 MSC 和 OP MSC 相比,OA MSC 中的 CXCR1 表面蛋白表达也较高,并使用流式细胞术显示(参见补充图 2B,可在关节炎和风湿病学网站http://onlinelibrary.wiley.com/doi上获取) /10.1002/art.39622/摘要)。
Finally, we investigated whether the expression of CXCR4 and TNFSF11 was OA‐specific (Figure (Figure6D).6D). Using additional OA, healthy control and OP MSC cultures, we observed that CXCR4 expression was significantly different across the 3 groups and higher in OA MSCs compared with healthy control MSCs (Figure (Figure6D).6D). TNFSF11/RANKL expression was variable across the 3 groups, as demonstrated by the results of qPCR (Figure (Figure6D)6D) and flow cytometry, respectively (Supplementary Figure 2B).
最后,我们研究了CXCR4和TNFSF11的表达是否具有 OA 特异性(图(图 6D) 。6 D)。使用额外的 OA、健康对照和 OP MSC 培养物,我们观察到CXCR4表达在 3 组之间存在显着差异,并且与健康对照 MSC 相比,OA MSC 中的 CXCR4 表达更高(图(图 6D) 。6 D)。 TNFSF11 /RANKL 表达在 3 组中存在差异,分别由 qPCR(图(图 6D) 6 D)和流式细胞术结果证明(补充图 2B)。
DISCUSSION 讨论
Relatively little is known about the role of endogenous MSCs in OA‐related bone pathophysiology 10. Given that MSCs are thought to be master regulators of joint and bone homeostasis 30, we investigated whether they might be involved in OA BMLs, which is known to be associated with both pain and structural changes 31. To our knowledge, this study is the first to evaluate native subchondral bone MSCs in human OA in relation to tissue damage.
关于内源性 MSC 在 OA 相关骨病理生理学中的作用知之甚少10 。鉴于 MSC 被认为是关节和骨稳态的主要调节者30 ,我们研究了它们是否可能参与 OA BML,众所周知,OA BML 与疼痛和结构变化相关31 。据我们所知,这项研究首次评估了人类 OA 中天然软骨下骨 MSC 与组织损伤的关系。
Our findings show numeric, topographic, gene expression, and functional perturbations in MSCs from patients with hip OA, especially from areas of cartilage loss in BMLs.
我们的研究结果显示了髋关节 OA 患者的 MSC 中的数量、形态、基因表达和功能扰动,尤其是来自 BML 软骨损失区域的 MSC。
Previous studies investigated OA MSCs from anatomic sites remote from damaged areas, i.e., from iliac crest bone marrow 32 and from femoral canal bone marrow 33, both after culture expansion. Our previous analysis of CD45−CD271+ cells sorted from whole OA femoral heads did not reveal any significant signs of premature aging or gross osteogenic abnormality compared with control bone 11. In the current study, we carefully excised BML and non‐BML areas of a femoral head, as segregated using MRI, and were able to detect subtle differences in MSC features within the same affected joint but in relation to the amount of tissue damage.
先前的研究调查了来自远离受损区域的解剖部位的OA MSC,即来自髂嵴骨髓32和股管骨髓33 ,两者均经过培养扩增。我们之前对从整个 OA 股骨头分选的 CD45−CD271+ 细胞进行的分析,与对照骨相比,没有发现任何明显的过早衰老或总体成骨异常的迹象11 。在当前的研究中,我们仔细切除了股骨头的BML和非BML区域,使用 MRI 进行分离,并且能够检测到同一受影响关节内 MSC 特征的细微差异,但与组织损伤的数量有关。
The histologic features of excised BML and non‐BML regions in our study were consistent with anticipated tissue abnormalities in BMLs, such as an increased bone volume fraction 24 and overlying cartilage loss 25, 26.
我们研究中切除的BML和非BML区域的组织学特征与 BML 中预期的组织异常一致,例如骨体积分数增加24和覆盖软骨损失25 , 26 。
We first showed that MSCs were proportionally increased in more diseased OA bone; this was initially surprising but not entirely unexpected considering previously published reports of the increase in synovial fluid MSCs in relation to OA severity 34, 35. Consistent with these findings, Harris et al recently documented aberrant MSC accumulation in the joints of patients with advanced OA 36. Additionally, an increase in subchondral bone MSCs was recently documented in a mouse anterior cruciate ligament transection model of OA 37.
我们首先证明,在更多患病的 OA 骨中,间充质干细胞 (MSC) 成比例增加;考虑到先前发表的关于滑液 MSC 增加与 OA 严重程度相关的报告,这最初令人惊讶,但并非完全出乎意料34 、 35 。与这些发现一致的是,Harris 等人最近记录了晚期 OA 患者关节中异常的 MSC 积聚36 。此外,最近在 OA 小鼠前十字韧带横断模型中记录了软骨下骨 MSC 的增加37 。
In accordance with data from the studies by Harris et al and Zhen et al and with their proposed mechanism for increased MSCs in OA joints 36, 37, we observed that the OA MSC chemokine receptor transcript profile was consistent with the notion of their potential recruitment from deeper marrow cavities toward the joint surface.
根据Harris等人和Zhen等人的研究数据以及他们提出的OA关节中MSCs增加的机制36 , 37 ,我们观察到OA MSC趋化因子受体转录谱与它们从关节中潜在招募的概念是一致的。骨髓腔朝向关节面更深。
BML MSCs may indeed be recruited to more damaged areas of cartilage and superficial subchondral bone due to higher concentrations of SDF‐1 in these regions, which is the result of diffusion to subchondral bone from OA synovial fluid via thinned, damaged cartilage 38, 39. Our data suggest that once at the site of damage, MSC CXCR4 expression may be down‐regulated to prevent further migration.
由于这些区域中 SDF-1 浓度较高, BML MSC 确实可能被招募到软骨和浅层软骨下骨的更多受损区域,这是通过变薄、受损的软骨从 OA 滑液扩散到软骨下骨的结果38 , 39 。我们的数据表明,一旦到达损伤部位,MSC CXCR4表达可能会下调,以防止进一步迁移。
Furthermore, OA MSCs up‐regulated CXCR1 (receptor for IL‐8) and CCR6 (receptor for MIP‐3α), 2 chemokines that are known to be abundant in OA synovial fluid 36, 40 and have been shown to be potent inducers of bone marrow MSC migration 41. Therefore, both the gene expression data and the immunohistochemical staining pattern, where MSCs were abundant in regions underlying cartilage defects, support the notion of their migratory response 42 toward areas of cartilage loss where the influences of inflammatory synovial fluid chemokine gradients are the strongest. In progressive OA, however, this response appears to be inadequate, pointing toward the possibility of a defect in MSC recruitment following skeletal damage.
此外,OA MSC 上调CXCR1 (IL-8 受体)和CCR6 (MIP-3α 受体),这两种趋化因子已知在 OA 滑液中丰富36 、 40 ,并已被证明是骨形成的有效诱导剂骨髓间充质干细胞迁移41 .因此,基因表达数据和免疫组织化学染色模式(其中间充质干细胞在软骨缺损区域中丰富)都支持它们向软骨损失区域迁移反应的概念42 ,其中炎症滑液趋化因子梯度的影响最强。然而,在进行性骨关节炎中,这种反应似乎不够充分,表明骨骼损伤后 MSC 募集可能存在缺陷。
In the current study, a lower MSC calcium production capacity of BML MSCs compared with non‐BML MSCs was observed, which could explain histologic findings of reduced tissue mineral density in BML bone despite a higher cross‐sectional bone area.
在当前的研究中,观察到与非BML MSC 相比, BML MSC 的 MSC 钙生成能力较低,这可以解释尽管横截骨面积较高,但BML骨中组织矿物质密度降低的组织学结果。
Inappropriate mineralization of BML bone could also be attributable to the defect in the capacity of BML MSCs to regulate bone remodeling. Compared with non‐BML MSCs, BML MSCs expressed less RANKL surface protein in 4 of 5 matched MSC cultures tested.
BML骨的不适当矿化也可能归因于BML MSC 调节骨重塑的能力缺陷。与非BML MSC 相比, BML MSC 在测试的 5 个匹配 MSC 培养物中的 4 个中表达较少的 RANKL 表面蛋白。
Shifts in RANKL expression, at both the messenger RNA and protein levels, have been previously documented for OA subchondral bone osteoblasts and explained by their “different stages of attempts to repair” 43. This further supports the concept of “uncoupled” bone formation and resorption by subchondral bone in OA 37, conceivably altering the biomechanical and load‐distribution properties of OA bone, putting cartilage at higher risk of injury. Such alterations support the need for further development of novel therapies targeting subchondral bone homeostasis for the treatment of OA 2, 44, 45, 46. In this context, our findings indicate that the MSC population is affected by the OA process and may therefore be an important therapeutic target for modulation in early disease.
先前已记录了 OA 软骨下骨成骨细胞在信使 RNA 和蛋白质水平上的 RANKL 表达变化,并通过其“尝试修复的不同阶段”进行了解释43 。这进一步支持了OA中软骨下骨“非耦合”骨形成和吸收的概念,可以想象,改变了 OA 骨的生物力学和负载分布特性,使软骨面临更高的损伤风险。这些改变支持需要进一步开发针对软骨下骨稳态的新疗法来治疗 OA 2 , 44 , 45 , 46 。在这种情况下,我们的研究结果表明 MSC 群体受到 OA 过程的影响,因此可能是调节早期疾病的重要治疗靶点。
Our histologic data showed that in the femoral heads of patients with OA, CD271+ MSCs surrounded vessels that had penetrated up to the cement line. MSCs can indeed act as promoters of angiogenesis 10 and are closely associated with pericytes and catecholaminergic nerve fibers 13, 47. OA neurovascular changes at the osteochondral junction, including vessels and both sensory and sympathetic nerves breaching the tidemark, are now considered to be a possible source of OA joint pain 48, 49.
我们的组织学数据显示,在 OA 患者的股骨头中,CD271+ MSC 围绕已渗透至骨水泥线的血管。间充质干细胞确实可以充当血管生成的促进者10 ,并且与周细胞和儿茶酚胺能神经纤维13 、 47密切相关。骨软骨连接处的 OA 神经血管变化,包括血管以及突破潮线的感觉神经和交感神经,现在被认为是 OA 关节疼痛的可能根源48 , 49 。
Based on our immunohistochemistry data, it is not unreasonable to suggest that MSCs in patients with advanced OA could also take part in pathologic subchondral neurovascular ingrowth (via their angiogenic actions and vessel‐stabilizing functions) and hence contribute to the development of joint pain.
根据我们的免疫组织化学数据,有理由认为晚期 OA 患者中的 MSC 也可能参与病理性软骨下神经血管向内生长(通过其血管生成作用和血管稳定功能),从而导致关节疼痛的发生。
This study is limited by the number of OA patients recruited for the MRI study and the amount of material that could be distributed to all of the experimental arms. Although the study was sufficiently powered to detect numeric and functional differences in paired BML/non‐BML MSC populations, some statistical analyses, such as gene expression validation using flow cytometry and TaqMan qPCR, were not possible in all cultures.
这项研究受到 MRI 研究招募的 OA 患者数量以及可以分配给所有实验组的材料数量的限制。尽管该研究有足够的能力检测配对的BML /非BML MSC 群体的数字和功能差异,但一些统计分析(例如使用流式细胞术和 TaqMan qPCR 的基因表达验证)在所有培养物中都是不可能的。
Although the expression of GDF5, a growth factor and known OA susceptibility gene 22, 50, was found to be different in OA MSCs compared with healthy control MSCs and OP MSCs, further work is required to assess its role in influencing MSC activity at the site of damage. This is in contrast to our data for CXCR1 and CCR6, the expression of which was confirmed to be OA‐specific.
虽然GDF5 (一种生长因子和已知的OA易感基因22 , 50 )在OA MSC中的表达与健康对照MSC和OP MSC相比有所不同,但仍需要进一步的工作来评估其在影响该位点MSC活性中的作用。的损坏。这与我们的CXCR1和CCR6数据形成鲜明对比,CXCR1 和 CCR6 的表达被证实是 OA 特异性的。
Finally, although we were able to observe a difference in CD271 immunohistochemical staining distribution between BML and non‐BML samples, tissue architectural heterogeneity prevented us from making a statistical evaluation of these data and comparing it with our flow cytometry findings.
最后,虽然我们能够观察到BML和非BML样本之间 CD271 免疫组织化学染色分布的差异,但组织结构异质性阻止我们对这些数据进行统计评估并将其与我们的流式细胞术结果进行比较。
In summary, our data show that in subchondral bone from patients with late‐stage hip OA, MSCs are increased in number in the areas of damage but exhibit functional and gene expression perturbations that could lead to further damage escalation.
总之,我们的数据表明,在晚期髋关节 OA 患者的软骨下骨中,损伤区域的 MSC 数量增加,但表现出功能和基因表达扰动,可能导致损伤进一步升级。
In relation to the development of novel therapy for early OA, our work emphasizes the abundance of subchondral bone MSCs in humans and provides initial insight into potential candidate pathways that can be targeted in order to normalize or improve the MSC pool.
在开发早期 OA 的新疗法方面,我们的工作强调了人类软骨下骨 MSC 的丰富性,并提供了对潜在候选途径的初步见解,这些途径可以作为目标,从而使 MSC 库正常化或改善。
New therapies targeting the bone–cartilage interface 14 and aimed at reestablishment of a functional cartilage surface zone 10 could delay progression of the disease, particularly if they are combined with other interventions such as correction of joint biomechanics.
针对骨-软骨界面14并旨在重建功能性软骨表面区域10的新疗法可以延缓疾病的进展,特别是如果它们与其他干预措施(例如校正关节生物力学)相结合。
AUTHOR CONTRIBUTIONS 作者贡献
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published.
所有作者都参与了文章的起草或对重要知识内容的批判性修改,并且所有作者都批准了最终版本的出版。
Dr. Jones had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
琼斯博士可以完全访问研究中的所有数据,并对数据的完整性和数据分析的准确性负责。
Study conception and design
研究构思和设计
Campbell, Churchman, Ponchel, Jones.
坎贝尔、丘奇曼、庞切尔、琼斯。
Acquisition of data 数据采集
Campbell, Churchman, Gomez, Ponchel, Jones.
坎贝尔、丘奇曼、戈麦斯、庞切尔、琼斯。
Analysis and interpretation of data
数据分析和解释
Campbell, Churchman, McGonagle, Conaghan, Ponchel, Jones.
坎贝尔、丘奇曼、麦格纳格尔、科纳汉、庞切尔、琼斯。
Supporting information 支持信息
Supplementary Table 1. MRI Sequence Setting for Femoral Head Imaging
补充表 1.股骨头成像的 MRI 序列设置
Supplementary Table 2. Reagents used for immunohistochemistry
补充表2.用于免疫组织化学的试剂
Supplementary Table 3. Antibody Conjugates and Markers Used for Flow Cytometry
补充表 3.用于流式细胞术的抗体偶联物和标记物
Supplementary Table 4 Assays used in the Taqman low density array (TLDA)
补充表 4 Taqman 低密度阵列 (TLDA) 中使用的检测
Supplementary Figure 1 CD271 cell distribution in BML sections assessed by immunohistochemistry. Light microscopy photomicrographs. (A) BML section showing cartilage fissuring and thinning with prominent subchondral CD271+ staining. (B) High‐magnification image of rectangular area in (A) showing CD271+ staining within subarticular end‐plate immediately beneath chondral lesion. Magnification bars: 500 μm (A), 200 μm (B).
补充图 1通过免疫组织化学评估的 BML 切片中 CD271 细胞的分布。光学显微镜照片。 (A) BML 切片显示软骨开裂和变薄,软骨下 CD271 +染色明显。 (B) (A) 中矩形区域的高倍放大图像,显示软骨病变正下方关节下终板内的 CD271 +染色。放大倍率条:500 μm (A)、200 μm (B)。
Supplementary Figure 2 Surface expression of CXCR4, CXCR1 and RANKL proteins by flow cytometry. Cultures were grown from magnetically‐selected CD271+ cells. Live MSCs were gated as DAPI‐negative, CD45‐CD73+CD90+ cells. Histograms for representative cultures are shown for the different markers. (A) BML (empty histograms) and non‐BML (filled histograms) donor‐matched cultures and (B) MSCs cultures from HC, OP, MSC femoral heads.
补充图 2通过流式细胞术检测 CXCR4、CXCR1 和 RANKL 蛋白的表面表达。培养物由磁性选择的 CD271 +细胞生长。活 MSC 被门控为 DAPI 阴性、CD45 - CD73 + CD90 +细胞。显示了不同标记的代表性培养物的直方图。 (A) BML (空直方图)和非BML (实心直方图)供体匹配培养物和 (B) 来自 HC、OP、MSC 股骨头的 MSC 培养物。
BML = bone marrow lesion; HC = healthy control; MSC = mesenchymal stem cell; OA = osteoarthritic; OP = osteoporotic.
BML = 骨髓病变; HC = 健康对照; MSC=间充质干细胞; OA=骨关节炎; OP=骨质疏松症。
ACKNOWLEDGMENTS 致谢
We thank Mark Emerton, David MacDonald, and Dr. Rahul Singh for providing femoral head specimens as well as Prof. Peter Giannoudis and Dr. Argiris Papathanasopoulos for providing normal bone specimens. We gratefully thank Drs.
我们感谢 Mark Emerton、David MacDonald 和 Rahul Singh 博士提供股骨头标本,感谢 Peter Giannoudis 教授和 Argiris Papathanasopoulos 博士提供正常骨标本。我们衷心感谢Drs。
Richard Hodgson and Andrew Grainger for interpretation of MRIs, Dr. Sarah Kingsbury for guidance with project development, Mike Shires and Karen Henshaw for processing of histologic samples and for their advice, and Dr. Elizabeth Hensor for her help with the statistical analysis.
Richard Hodgson 和 Andrew Grainger 对 MRI 进行解释,Sarah Kingsbury 博士对项目开发提供指导,Mike Shires 和 Karen Henshaw 对组织学样本进行处理并提供建议,Elizabeth Hensor 博士对统计分析提供帮助。
We also thank Dr. Tom Baboolal, Richard Cuthbert, Dr. Sally Boxall, Dr. Yasser El‐Sherbiny, and Rekha Parmar for providing technical assistance.
我们还感谢 Tom Baboolal 博士、Richard Cuthbert 博士、Sally Boxall 博士、Yasser El-Sherbiny 博士和 Rekha Parmar 提供的技术援助。
Notes 笔记
Supported in part by the Innovative Medicines Initiative (BTCure grant 115142‐2) and the Wellcome Trust (project WELMEC; grant WT 088908/Z/09/). Dr. Campbell's work was supported by the University of Ottawa (Scholarship for New Faculty).
部分由创新药物计划(BTCure 拨款 115142‐2)和 Wellcome 信托基金(WELMEC 项目;拨款 WT 088908/Z/09/)支持。坎贝尔博士的工作得到了渥太华大学(新教师奖学金)的支持。
Dr. Churchman's work was supported by the NIHR Leeds Musculoskeletal and Biomedical Research Unit. Drs. McGonagle, Conaghan, Ponchel, and Jones’ work was supported in part by the NIHR Leeds Musculoskeletal and Biomedical Research Unit.
Churchman 博士的工作得到了 NIHR 利兹肌肉骨骼和生物医学研究单位的支持。博士。 McGonagle、Conaghan、Ponchel 和 Jones 的工作部分得到了 NIHR 利兹肌肉骨骼和生物医学研究中心的支持。
REFERENCES 参考
Nho SJ、Kymes SM、卡拉汉 JJ、费尔森 DT。美国髋骨关节炎的负担:流行病学和经济考虑。 J Am Acad Orthop Surg 2013; 21补编 1 :S1–6。 [考研] [谷歌学术]
曼塞尔 JP、柯林斯 C、贝利 AJ。骨关节炎的主要焦点应该是骨,而不是软骨。自然临床实践风湿病2007; 3 :306-7。 [考研] [谷歌学术]
戈德林 MB、戈德林 SR。关节软骨和软骨下骨在骨关节炎的发病机制中。安纽约学术科学2010; 1192 :230–7。 [考研] [谷歌学术]
Felson DT、Chaisson CE、Hill CL、Totterman SM、Gale ME、Skinner KM 等。骨髓病变与膝骨关节炎疼痛的关联。安实习医学2001; 134 :541–9。 [考研] [谷歌学术]
Xu L,Hayashi D,Roemer FW,Felson DT,Guermazi A。与骨关节炎相关的软骨下骨髓病变的磁共振成像。塞明关节炎大黄2012; 42 :105–18。 [ PMC 免费文章] [ PubMed ] [ Google Scholar ]
Driban JB、Lo GH、Lee JY、Ward RJ、Miller E、Pang J 等。骨关节炎膝关节的定量骨髓病变大小与软骨损伤相关,并可预测纵向软骨损失。 BMC 肌肉骨骼疾病2011; 12 :217。 [ PMC 免费文章] [ PubMed ] [ Google Scholar ]
Zanetti M,Bruder E,Romero J,Hodler J。骨关节炎膝盖的骨髓水肿模式:MR 成像与组织学结果之间的相关性。放射学2000; 215 :835–40。 [考研] [谷歌学术]