Tracking labeled stem cells in TBI model by cellular MRI

实验模型创伤性脑损伤的病理表现包括急性炎症反应、血脑屏障破坏、出血、脱髓鞘、轴突横断和慢性轴突神经元丢失和神经胶质瘤。干细胞（SC）治疗是一种潜在的治疗方法，可以作为替代疗法，也可以通过释放生长因子和抗炎细胞因子的旁分泌作用治疗TBI损伤。啮齿动物创伤性脑损伤模型的实验研究有限，通常在实验损伤后24至72小时内给予单剂量细胞。在损伤的急性炎症和水肿期，细胞递送的最佳时间和剂量以最大限度地恢复功能和移植存活尚不清楚。我们通过系列MRI评估了控制性皮质冲击（CCI）诱导大鼠脑TBI的自然历史。雌性Wistar大鼠左运动皮质（Bregma +1外侧2mm）行CCI，冲击时间50 msec，深度2mm，速度5 m /s。MRI在7特斯拉下进行。获得CCI后不同时间点分辨率为117x117x500m的T2w和T2*w图像（第2、9、30和58天）和行为（第1、14、28和56天）。我们观察到，尽管相同的CCI损伤环境，但在第2、9和30天CCI诱导病变的外观和体积变化很大。第2天CCI侧到对侧皮质体积的变化百分比与随后9日和30日的检查之间几乎没有相关性（CCI皮质/对侧皮质体积的平均变化百分比：第2天=2611%；第9天= -214%；第30天= -1014%）。在CCI后第2天至第9天，45%的大鼠发生了CCI病变内的出血转化。MRI和病理学显示，在使用相同的冲击参数和技术时，皮质损失、胶质瘤、轴突损失和囊肿形成有显著差异。虽然CCI病变的差异可能归因于技术的差异，但类似病变在第2天和第30天之间的差异表明了CCI大鼠模型固有的生物学变异性。在第2天至第9天之间出血演变的可能性也引起了对先前使用磁标记细胞跟踪其递送到CCI病变而不获得干细胞输注前MRI的解释的关注，因为MRI和普鲁士蓝染色显示出血相同。基于这项研究，MRI病变或皮质体积可作为实验性TBI新疗法的结果测量指标，然而，在随机平行组试验设计中，检测到治疗结果在α 0.05和功率=0.8时皮质体积差异的10 - 20%的能力需要在22-64名受试者之间才能看到显著差异。病变体积或皮质差异百分比与30天的行为测量没有相关性。我们开发了一种图像分析方法，用于融合MRI弥散张量成像（DTI）结果与免疫组织学染色。在模型中，分数各向异性（FA）的增加被认为与脑卒中或脑外伤后的轴突重组和再生长有关，但这种变化的组织学证据似乎很薄弱。CCI后2个月大鼠的组织学切片染色GFAP（星形胶质细胞）、SMI31和SMI32（神经丝）、MAP2（微管）和MBP（髓鞘）。基于组织学的傅里叶分析捕获了DTI测量的微观各向异性和取向。在CCI损伤的白质（WM）中，FA下降，与轴突和髓鞘损伤一致，尽管星形胶质细胞也具有高度的各向异性。在灰质中，FA增加，这与相干星形胶质细胞增生一致，并且在较小程度上与树突组织一致。这些结果表明，造成正常脑各向异性的微观细胞和亚细胞成分（轴突和髓鞘）不一定是造成损伤脑各向异性的唯一因素。星形胶质细胞的重塑和迁移也与dti测量的各向异性一致。这种分析方法应该有助于提供微观各向异性和dti衍生各向异性之间定量和具体的关系。

The pathology of traumatic brain injury in experimental models includes acute inflammatory reaction, blood brain barrier disruption, hemorrhage, demyelination, axonal transection and chronically with axonal neuronal loss and gliosis. Stem cell (SC) therapy is a potential treatment either as replacement therapy or via paracrine effect with release of growth factors and anti-inflammatory cytokines for TBI injury. Experimental studies in rodent models of TBI have been limited and usually a single dose of cells is administered within 24 to 72 hours after experimental injury. The optimal timing and dose of cell delivery to maximize functional recovery and transplantation survival during the acute inflammatory and edematous phase of damage is unknown. We evaluated the natural history of control cortical impact (CCI) to induce TBI in the rat brain by serial MRI. CCI was performed in left motor cortex (Bregma +1 lateral 2 mm) with impact time 50 msec at depth of 2mm and velocity 5 meters/seconds in female Wistar rats. MRI was performed at 7 Tesla. T2w and T2*w images with image resolution of 117x117x500m (days 2,9,30 and 58) and behavior (days 1,14, 28 and 56) were obtained at various time points after CCI. We observed that the appearance and volume of CCI-induced lesions at days 2, 9, and 30 was highly variable despite identical CCI injury settings. There was little correlation between the percent change of CCI side to contralateral side cortical volumes on days 2 to subsequent exams on 9, and 30 (mean percent change CCI cortical/contralateral cortex volume: Day 2=2611%; Day 9 = -214%; Day 30 = -1014%). Hemorrhagic conversion within the CCI lesion occurred in 45% of rats between days 2 and 9 post CCI. MRI and pathology demonstrate significant variation in cortical loss, gliosis, axonal loss and cyst formation using identical impact parameters and techniques. Although the variation in CCI lesions may be attributed to differences in technique, the divergence of similar lesions between days 2 and 30 demonstrates the inherent biological variability of the CCI rat model. The possibility of hemorrhagic evolution between days 2 and 9 also raised concerns about previous interpretation of using magnetically labeled cells to track their delivery to CCI lesions without obtaining a pre-stem cell infusion MRI, as hemorrhage appears identical on MRI and Prussian blue staining. Based on this study, MRI lesion or cortical volumes could be used as an outcome measure for novel therapies of experimental TBI, however the ability to detect 10 to 20% difference in cortical volume with alpha 0.05 and power =0.8 as a result of treatment, would require between 22-64 subjects would be needed to see significant differences in a randomized parallel group trial design. There was no correlation of lesion volume or percent cortical differences and behavioral measures at 30 days. We have developed an image analysis method for fusing MRI diffusion tensor imaging (DTI) findings to immunohistological stains. In models, increased Fractional Anisotropy (FA) has been argued to be associated with axonal reorganization and re-growth following stroke or TBI, but the histological evidence for such changes appeared tenuous. Histological sections from rats at 2 months post CCI were stained for GFAP (astrocytes), SMI31 & SMI32 (neurofilaments), MAP2 (microtubules), and MBP (myelin). The histological-based Fourier analysis captured the microscopic anisotropy and orientation as measured by DTI. In the CCI injured white matter (WM), FA decreased and was consistent with axonal and myelin injury, although astrocytes also had high anisotropy. In gray matter, FA increased and this was consistent with coherent astrogliosis and to a lesser extent, dendrite organization. These results demonstrate the microscopic cellular and subcellular elements that contribute to anisotropy in normal brain (axons and myelin) are not necessarily the only contributors to anisotropy in the injured brain. Astrocyte remodeling and migration also appears consistent with the DTI-measured anisotropy. This analytical method should help provide a quantitative and specific relationship between the microscopic anisotropy and DTI-derived anisotropy.