Morphology-based prediction of osteogenic differentiation potential of human mesenchymal stem cells.
Morphology-based prediction of osteogenic differentiation potential of human mesenchymal stem cells.
复制标题
DOI：
10.1371/journal.pone.0055082
复制
发表时间：
2013
期刊：
影响因子：
3.7
通讯作者：
Kato R
中科院分区：
文献类型：
--
作者：
Matsuoka F;Takeuchi I;Agata H;Kagami H;Shiono H;Kiyota Y;Honda H;Kato R
人骨髓间充质干细胞（hBMSCs）是广泛应用于临床骨再生的细胞来源。获得最大的治疗效果取决于移植干细胞的成骨分化潜能。然而，目前仍然没有实用的方法来表征这种潜在的非侵入性或先前的。监测细胞形态是评估成骨潜能的一种实用且无创的方法。不幸的是，这种基于图像的方法在历史上是定性的，需要有经验的解释。通过将显微镜的非侵入性与最新技术相结合，允许更高的通量和定量成像指标，我们研究了形态测量特征在定量预测细胞成骨潜能方面的适用性。我们应用计算机器学习，结合细胞形态特征及其相应的生化成骨测定结果，建立成骨分化预测模型。利用BioStation CT在hBMSCs成骨分化培养过程中自动获取的9990张图像数据集，提取666个形态特征作为参数。实验测量了两种常用的成骨标志物碱性磷酸酶（ALP）活性和钙沉积，并将其作为真实的生物分化状态来验证预测的准确性。利用分化培养过程中随时间变化的形态特征，预测结果与实验定义的分化标记值高度相关（两种标记预测的R均为0.89）。我们基于形态学的预测在两种情况下的临床适用性进行了进一步检验：一种仅使用历史细胞图像，另一种使用历史图像和患者自身的细胞图像来预测新患者的细胞潜能。通过将患者自身细胞特征纳入模型，大大提高了预测精度，为临床应用提供了切实可行的策略。因此，我们的研究结果为在再生医学中使用相位对比细胞形态学的定量时间序列进行非侵入性细胞质量预测的可行性提供了强有力的证据。
Human bone marrow mesenchymal stem cells (hBMSCs) are widely used cell source for clinical bone regeneration. Achieving the greatest therapeutic effect is dependent on the osteogenic differentiation potential of the stem cells to be implanted. However, there are still no practical methods to characterize such potential non-invasively or previously. Monitoring cellular morphology is a practical and non-invasive approach for evaluating osteogenic potential. Unfortunately, such image-based approaches had been historically qualitative and requiring experienced interpretation. By combining the non-invasive attributes of microscopy with the latest technology allowing higher throughput and quantitative imaging metrics, we studied the applicability of morphometric features to quantitatively predict cellular osteogenic potential. We applied computational machine learning, combining cell morphology features with their corresponding biochemical osteogenic assay results, to develop prediction model of osteogenic differentiation. Using a dataset of 9,990 images automatically acquired by BioStation CT during osteogenic differentiation culture of hBMSCs, 666 morphometric features were extracted as parameters. Two commonly used osteogenic markers, alkaline phosphatase (ALP) activity and calcium deposition were measured experimentally, and used as the true biological differentiation status to validate the prediction accuracy. Using time-course morphological features throughout differentiation culture, the prediction results highly correlated with the experimentally defined differentiation marker values (R>0.89 for both marker predictions). The clinical applicability of our morphology-based prediction was further examined with two scenarios: one using only historical cell images and the other using both historical images together with the patient's own cell images to predict a new patient's cellular potential. The prediction accuracy was found to be greatly enhanced by incorporation of patients' own cell features in the modeling, indicating the practical strategy for clinical usage. Consequently, our results provide strong evidence for the feasibility of using a quantitative time series of phase-contrast cellular morphology for non-invasive cell quality prediction in regenerative medicine.
登录
查看更多内容
影响因子：
12.3
作者：
Carpenter AE;Jones TR;Lamprecht MR;Clarke C;Kang IH;Friman O;Guertin DA;Chang JH;Lindquist RA;Moffat J;Golland P;Sabatini DM
通讯作者：
Sabatini DM
影响因子：
2.5
作者：
HOERL, AE;KENNARD, RW
通讯作者：
KENNARD, RW
影响因子：
4.2
作者：
Hayashi, Ousuke;Katsube, Yoshihiro;Ito, Hiromoto
通讯作者：
Ito, Hiromoto
影响因子：
3
作者：
Misselwitz B;Strittmatter G;Periaswamy B;Schlumberger MC;Rout S;Horvath P;Kozak K;Hardt WD
通讯作者：
Hardt WD
影响因子：
7.6
作者：
Agata, H.;Asahina, I.;Ueda, M.
通讯作者：
Ueda, M.