Towards comprehensive myelin assessment: combining multi-contrast quantitative MRI with spatially resolved lipodomics in the human brain

本项目的目的是建立基质辅助激光解吸/电离质谱仪成像(MALDI-MSI)，作为验证和开发基于定量磁共振成像(QMRI)的髓鞘标志物的工具。髓磷脂是一种富含脂质的物质，对健康的大脑功能至关重要。体内髓鞘的研究依赖于定量磁共振成像。不同的qMRI参数对髓鞘形成敏感，但特异性只能通过生物物理或数据驱动的建模方法来实现。这类模型的开发和验证都依赖于来自组织学的可靠的“地面真实”髓鞘图谱。目前，经典组织学或免疫组织化学的染色强度经常被用作组织髓鞘含量的近似值。在初步研究中，我们观察到不同的组织学方法在皮质中产生的髓鞘形成模式不同。这种差异可能与不同脑组织类型的髓鞘组成的差异性有关。在定量磁共振成像的背景下，考虑这种成分是至关重要的，因为已知不同的脂类在其磁共振性质上是不同的。到目前为止，将脂质组学与定量磁共振成像联系起来的研究仅限于使用模体进行的研究。在这里，我们想开发一种可以捕捉人脑组织切片中髓鞘生化成分的方法：MALDI-MSI。在以前收集的试点数据的基础上，我们将为MALDI-MSI数据的获取和分析制定一个优化的工作流程。因此，我们将解决一些与MALDI-MSI脂类定量相关的方法学挑战：识别最重要的峰下的脂类，通过应用内部标准校准脂类图谱，基于其脂类成分对脑组织进行聚类，以及脂类浓度图谱与qMRI图谱的联合配准。我们将解决以下研究问题：a)髓鞘定量的最佳MALDI-MSI衍生指标是什么？B)组织脂质成分对不同的QMR参数有何影响？D)是否可以从qMRI参数的组合中推断组织的脂质成分？脱髓鞘神经系统疾病的诊断、患者分层和治疗受到缺乏准确量化体内髓鞘形成的工具的限制。目前可用的衡量标准存在特异性低的问题，因此不能轻易地跨个人或随时间进行比较。这限制了它们在临床研究中的可解释性，并阻碍了它们对个性化医学的适用性。人脑组织切片中的qMRI数据和脂质组学的结合是朝着满足对可靠的针对组织髓鞘形成的MR标记的需求迈出的重要一步。该项目收集的数据将作为今后建立生物物理模型和数据驱动模型的独特参考材料。

The aim of this project is to establish matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) as a tool for the validation and development of quantitative MRI (qMRI) – based markers of myelin. Myelin is a lipid-rich substance that is crucial for healthy brain functioning. The study of myelin in vivo relies on qMRI. Various qMRI parameters are sensitive to myelination, but specificity can only be achieved by biophysical or data-driven modeling approaches. Both the development and validation of such models rely on reliable ‘ground-truth’ myelin maps from histology. Currently, staining intensity from classical histology or immunohistochemistry is frequently used as an approximation for tissue myelin content. In preliminary research, we have observed that different histological methods differ in the patterns of myelination they yield in the cortex. This difference is likely related to variability in myelin composition across different brain tissue types. Considering this composition is crucial in the context of qMRI, as different lipid species are known to differ in their MR properties. So far, relating lipidomics to qMRI has been limited to studies conducted with phantoms.Here, we want to exploit a method that can capture the biochemical composition of myelin in sections of human brain tissue: MALDI-MSI. Building upon previously collected pilot data, we will develop an optimized workflow for the acquisition and analysis of MALDI-MSI data. Hereby, we will tackle a number of methodological challenges related to lipid quantification with MALDI-MSI: identification of the lipids underlying most important peaks, calibration of the lipid maps through application of internal standards, clustering of the brain tissue based on its lipid composition, and the co-registration of lipid concentration maps with qMRI maps. We will address the following research questions: a) What is the best MALDI-MSI-derived metric for myelin quantification? b) How does the tissue lipid composition impact on different qMR parameters? d) Can tissue lipid composition be inferred from a combination of qMRI parameters? Diagnosis, patient stratification and treatment of demyelinating neurological diseases is constrained by a lack of tools to accurately quantify myelination in vivo. Currently available measures suffer from low specificity and can therefore not easily be compared across individuals or over time. This limits their interpretability in clinical studies and hinders their applicability to personalised medicine. The combination of qMRI data and lipidomics in human brain tissue sections is an important step towards meeting the need for reliable MR markers that are specific to tissue myelination. The data collected in this project will act as unique reference material for future efforts in establishing biophysical and data-driven models.