Bone maintenance and repair are achieved through the maintenance and differentiation of rare skeletal stem cells (SSCs). Despite the development of numerous lineage tracing models and/or cell surface markers to label and track the in vivo SSCs, the in vivo identity and function of skeletal stem/progenitor cells remains largely elusive because these labeled populations are highly heterogeneous. In fact, endogenous SSCs are heterogeneous, and distinct SSC populations maintain different bone compartments.^[1] Therefore, which and how different SSCs functionally orchestrate and form fracture‐repairing osteoblasts are largely open questions.

In the hypothesis paper in this issue, Matsushita et al.^[2] proposed the theory that mature skeletal cells (i.e., osteoblasts, adipocytes, and chondrocytes) can transform their identities into skeletal stem‐cell‐like cells in response to specific stimuli. Recently, Matsushita et al. developed an inducible Cxcl12‐CreER model and found that Cxcl12‐creER preferentially marks a quiescent CXCL12⁺LepR⁺ bone marrow stromal cells (BMSCs) known to encompass a subset of bone marrow SSCs. Interestingly, the majority of Cxcl12‐CreER⁺ BMSCs express adipocyte specific genes—specifically adiponectin (AdipoQ)—but they do not readily contribute to osteogenic differentiation. Thus, the authors termed these cells Cxcl12‐CreER⁺ adipogenic precursors. Upon injury however, Cxcl12‐CreER⁺ cells localize to injury sights, proliferate, and undergo osteogenic differentiation, suggesting that adipogenic Cxcl12‐CreER⁺ cells de‐differentiate into SSC‐like cells and re‐differentiate (cell plasticity) into osteoblasts. This cellular plasticity in bone regeneration appears to be unique to Cxcl12‐CreER⁺ adipocyte precursors as opposed to Osterix‐CreER⁺ osteogenic cells in response to injury.^[3] From these data, the authors hypothesize that Cxcl12‐CreER⁺ adipogenic precursors transform their identities into skeletal stem‐cell‐like cells and then become a source of new osteoblasts in response to injury. Although cell plasticity may occur in response to injury, multiple questions remain surrounding the identity and plasticity of Cxcl12‐CreER⁺ adipocyte precursors.

First, it is known that CXCL12⁺LepR⁺ BMSCs can be heterogeneous and contain multiple lineage progenitors. It is possible that Cxcl12‐CreER⁺ cells include quiescent osteogenic progenitors that are inactive under physiological conditions but that become active after injury. It is also possible that a subset of Cxcl12‐CreER⁺ cells can be bipotential—adipogenic and osteogenic—progenitors, and preferentially respond to injury and differentiate into osteoblasts. Since Cxcl12‐CreER can label both adipogenic and osteogenic progenitor cells, endogenous identification and tracking of these separate Cxcl12‐CreER⁺ populations is not possible using only the Cxcl12‐CreER mouse model.

Second, it remains unclear whether Cxcl12‐CreER⁺ adipocyte precursors are the major source of new osteoblast/osteocytes during injury repair. Previous studies have revealed that the ablation of skeletal stem/progenitor cells with osteogenic lineage potential results in a significant hinderance in bone formation and repair.^[4] On the other hand, ablation of bone marrow adipogenic lineage‐restricted stromal cells using adiponectin (AdipoQ)‐Cre with diphtheria toxin (DTR or DTA model systems) substantially increases bone formation and volume under homeostatic conditions.^{[5, 6]} Although under homeostatic conditions AdipoQ‐Cre and AdipoQ‐CreER lineage cells do not significantly contribute to osteogenic cells, it is currently unknown whether this remains true in response to injury. To truly determine whether adipogenic precursors can transdifferentiate and contribute to osteogenic cells during fracture healing, one would want to test this proposal using an inducible trigenic lineage reporter mouse model to label adipogenic progenitors (i.e., AdipoQ‐CreER) in combination with an osteoblast‐specific reporter such as osteocalcin (Ocn‐GFP) or Col1a1 2.3 kb (Col2.3‐GFP).

The hypothesis presented by Matsushita et al. explores a compelling new concept that mature skeletal cells can transform their identities into skeletal stem‐cell‐like cells in response to injury. However, more studies are required to test direct conversion of “mature” skeletal cells into stem‐cell‐like cells at single cell level. Additional lineage tracing models specific to adipogenic cells will need to be used and/or developed to answer these fundamental cell plasticity questions.

骨骼的维护和修复是通过稀有骨骼干细胞（SSC）的维护和分化来实现的。尽管开发了许多谱系追踪模型和/或细胞表面标记来标记和跟踪体内SSC，但由于这些标记的群体高度异质，因此骨骼干/祖细胞的体内特性和功能仍然难以捉摸。实际上，内源性SSC是异质的，并且不同的SSC群体维持不同的骨腔。^{[ 1 ]}因此，哪些以及哪种不同的SSC在功能上编排并形成修复骨折的成骨细胞是一个悬而未决的问题。

在本期的假设文件中，Matsushita等人。^{[ 2 ]} 提出了这样的理论，即成熟的骨骼细胞（即成骨细胞，脂肪细胞和软骨细胞）可以对特定刺激做出反应，将其身份转变为骨骼干细胞样细胞。最近，松下等人。建立了可诱导的Cxcl12-CreER模型，发现Cxcl12-creER优先标记了已知包含骨髓SSC子集的静态CXCL12 ⁺ LepR ⁺骨髓基质细胞（BMSC）。有趣的是，大多数Cxcl12-CreER ⁺ BMSC表达脂肪细胞特异性基因-特别是脂联素（AdipoQ），但它们并不容易促进成骨细胞分化。因此，作者称这些细胞为Cxcl12-CreER^+成脂前体。但是，在受伤时，Cxcl12-CreER ⁺细胞会定位到受伤的视点，增殖并经历成骨分化，这表明成脂的Cxcl12-CreER ⁺细胞会去分化为SSC样细胞，然后再分化为（细胞可塑性）成骨细胞。这种骨再生中的细胞可塑性似乎是Cxcl12-CreER ⁺脂肪细胞前体所特有的，而不是Osterix-CreER ⁺成骨细胞对损伤的反应。^{[ 3 ]}根据这些数据，作者假设Cxcl12-CreER ⁺成脂前体将其身份转变为骨骼干细胞样细胞，然后响应损伤而成为新成骨细胞的来源。尽管细胞可塑性可能因损伤而发生，但围绕Cxcl12-CreER ⁺脂肪细胞前体的身份和可塑性仍存在多个问题。

首先，已知CXCL12 ⁺ LepR ⁺ BMSC可以是异质的，并且包含多个谱系祖细胞。Cxcl12-CreER ⁺细胞可能包括在生理条件下没有活性但在损伤后变得有活性的静态成骨祖细胞。Cxcl12-CreER ⁺细胞的子集也可能是双潜能（成脂和成骨）祖细胞，并优先对损伤做出反应并分化为成骨细胞。由于Cxcl12‐CreER可以标记成脂祖细胞和成骨祖细胞，因此仅使用Cxcl12‐CreER小鼠模型无法对这些单独的Cxcl12‐CreER ⁺群体进行内源性鉴定和追踪。

其次，尚不清楚Cxcl12-CreER ⁺脂肪细胞前体是否是损伤修复过程中新成骨细胞/成骨细胞的主要来源。先前的研究表明，骨骼干/祖细胞具有成骨血统的潜在能力会导致骨骼形成和修复的显着障碍。^{[ 4 ]}另一方面，使用脂联素（AdipoQ）-Cre和白喉毒素（DTR或DTA模型系统）消融骨髓成脂谱系限制的基质细胞，可在稳态条件下显着增加骨形成和体积。^{[ 5，6 ]}尽管在稳态条件下，AdipoQ-Cre和AdipoQ-CreER谱系细胞对成骨细胞没有显着贡献，但目前尚不清楚在对损伤的反应中是否仍然如此。为了真正确定成脂前体在骨折愈合过程中是否能够分化并促进成骨细胞，人们希望使用可诱导的三基因谱系报告基因小鼠模型标记成骨祖细胞（即AdipoQ-CreER）与成骨细胞特异性组合来测试该提议。骨钙素（Ocn‐GFP）或Col1a1 2.3 kb（Col2.3‐GFP）等报告基因。

松下等人提出的假设。探索了一个令人信服的新概念，即成熟的骨骼细胞可以响应损伤而将其身份转变为骨骼干细胞样细胞。但是，需要更多的研究来测试“成熟”骨骼细胞在单细胞水平上直接转化为干细胞样细胞的能力。需要使用和/或开发其他特定于成脂细胞的谱系追踪模型，以回答这些基本的细胞可塑性问题。