图 1  脑组织中水分子跨膜过程概述　(a) 水分子跨血脑屏障交换; (b) 水分子跨细胞膜交换; (c) 水分子跨血液-脑脊液屏障交换, 在脑室中, 水通道蛋白AQP存在于脑脊液界面的室管膜细胞的基底外膜; (d) 水分子跨脑脊液-脑实质屏障的交换, 蛛网膜下腔的脑脊液通过动脉周围间隙流入大脑, 然后通过位于星形胶质细胞终足的水通道蛋白与脑间质液交换

Fig. 1.  The transmembrane process of water molecules in brain tissue can be summarized as: (a) Exchange of water molecules across the BBB (blood-brain barrier); (b) exchange of water molecules across cell membranes; (c) exchange of water molecules across the blood-CSF barrier. In the ventricle, AQP is present in the basal outer membrane of ependymal cells at the interface of cerebrospinal fluid; (d) exchange of water molecules across the CSF-brain parenchymal barrier. The cerebrospinal fluid in the subarachnoid space flows into the brain through the periarterial space and is then exchanged with the interstitial fluid via aquaporins located in the astrocyte endfeet.

图 2  水分子跨细胞膜被动(p)交换途径和主动(a)交换途径的示意图, 其中$ {k}_{\text{io}} $是细胞内水交换到细胞外空间的速率常数, $ {k}_{\text{oi}} $是细胞外水交换到细胞内空间的速率常数; II和III分别表示K⁺外流通道和Na⁺内流通道, I和IV表示水分子共运通道

Fig. 2.  Illustration of the passive (p) and potentially active (a) transcytolemmal water exchange pathways, where $ {k}_{\text{io}} $ is the exchange rate of water from the cell into the extracellular space, $ {k}_{\text{oi}} $ is the exchange rate of water from the outside into the intracellular space, II and III represent transporters K⁺ uses to re-exit and Na⁺ uses to re-enter the cell, respectively, and I and IV represent water co-transporters that H₂O uses to exit and enter the cell.

图 3  水分子稳态跨膜交换过程的模型, 左边方框表示细胞外空间(outside), 右边方框表示细胞内空间(inside)

Fig. 3.  Model of steady-state transmembrane exchange of water molecules, the left box represents the extracellular space, and the right box represents the intracellular space.

图 4  快门速度(shutter-speed, SS)概念示意图, SS定义为细胞内外纵向弛豫速率的差值, 其中, $ {k}_{{\mathrm{b}}{\mathrm{o}}} $表示血管内水分子到细胞间隙的流出速率, $ {k}_{{\mathrm{o}}{\mathrm{b}}} $表示细胞间隙的水分子到血管内的流入速率, $ {k}_{{\mathrm{p}}{\mathrm{e}}} $表示对比剂分子从血管内到细胞间隙的渗透速率, $ {k}_{{\mathrm{e}}{\mathrm{p}}} $表示对比剂分子从细胞间隙渗透到血管内的速率; Gd代表对比剂分子, 不能进入细胞内

Fig. 4.  Conceptual diagram of shutter speed (SS), SS is defined as the difference in longitudinal relaxation rates between intracellular and extracellular spaces. Here, $ {k}_{{\mathrm{b}}{\mathrm{o}}} $ denotes the efflux rate of water molecules from the blood vessels to the interstitial space, while $ {k}_{{\mathrm{o}}{\mathrm{b}}} $ denotes the influx rate of water molecules from the interstitial space to the blood vessels, $ {k}_{{\mathrm{p}}{\mathrm{e}}} $ represents the rate of contrast agent molecules diffusing from the blood vessels to the interstitial space, and $ {k}_{{\mathrm{e}}{\mathrm{p}}} $ represents the rate of contrast agent molecules diffusing from the interstitial space to the blood vessels. Gd represents the contrast agent molecules, which cannot enter the cells.

图 5  (a) 从磁共振信号获取的表观纵向弛豫速率$ {R}_{1{\mathrm{L}}} $, $ {R}_{1{\mathrm{S}}} $随细胞外对比剂浓度[CR₀]的变化曲线^[53], 不存在水交换时, $ {R}_{1{\mathrm{S}}} $的值和对比剂的浓度符合$ {r}_{1{\mathrm{o}}}\left[{\mathrm{C}}{{\mathrm{R}}}_{0}\right]+{R}_{1{\mathrm{o}}0} $, 在图中的偏左侧区域表示(k_io+k_oi)远大于快门速度$ \left|R{1}_{{\mathrm{i}}}-R{1}_{{\mathrm{o}}}\right| $, 此时处于快交换边界; (b) (a)图中左侧区域的放大图; (c) 表观弛豫速率快的组分的信号占比随对比剂浓度的变化曲线, 当其接近细胞外水含量占比时, 认为系统进入到慢交换边界, 此时, 对细胞内停留时间的计算可参考(1)式, 即$ {R}_{1{\mathrm{L}}}={R}_{1{\mathrm{i}}}+{\tau }_{{\mathrm{i}}}^{-1} $. 上述曲线根据(8)式—(10)式绘制, 其中, $ {p}_{{\mathrm{i}}}=0.85 $, $ {\tau }_{{\mathrm{i}}}=1.0{\mathrm{s}} $, $ {R}_{1{\mathrm{i}}}=0.67\;{{\mathrm{s}}}^{-1} $, $ {R}_{1{\mathrm{o}}0}=0.5\;{{\mathrm{s}}}^{-1} $, $ {r}_{1{\mathrm{o}}}=3.75\;{\mathrm{m}}{{\mathrm{m}}{\mathrm{o}}{\mathrm{l}}}^{-1}{\cdot}{\mathrm{L}}{\cdot}{{\mathrm{s}}}^{-1} $(注: 图中仿真曲线根据(8)式—(10)式绘制)

Fig. 5.  (a) The curves of apparent longitudinal relaxation rates $ {R}_{1{\mathrm{L}}} $ and $ {R}_{1{\mathrm{S}}} $ as a function of extracellular contrast agent concentration $ \left[{\mathrm{C}}{{\mathrm{R}}}_{0}\right] $ obtained from magnetic resonance signals^[53], in the absence of water exchange, the value of $ {R}_{1{\mathrm{S}}} $ is proportional to the contrast agent concentration, following the relation $ {r}_{1{\mathrm{o}}}\left[{\mathrm{C}}{{\mathrm{R}}}_{0}\right]+{R}_{1{\mathrm{o}}0} $, the region on the left side of the figure indicates that (k_io+k_oi) is much greater than the shutter speed $ \left|R{1}_{{\mathrm{i}}}-R{1}_{{\mathrm{o}}}\right| $, which corresponds to the fast exchange boundary; (b) an enlarged view of the left-side region in (a); (c) the curve showing the proportion of signal from the fast component of the apparent relaxation rate as a function of contrast agent concentration, when it approaches the proportion of extracellular water content, the system is considered to have reached the slow-exchange limit. At this point, the calculation of intracellular residence time can refer to Eq. (1), where $ {R}_{1{\mathrm{L}}}={R}_{1{\mathrm{i}}}+{\tau }_{{\mathrm{i}}}^{-1} $. The above curves are plotted according to Eqs. (2) to (4), where $ {p}_{{\mathrm{i}}}=0.85 $, $ {\tau }_{{\mathrm{i}}}=1.0{\mathrm{s}} $, $ {R}_{1{\mathrm{i}}}=0.67\;{{\mathrm{s}}}^{-1} $, $ {R}_{1{\mathrm{o}}0}=0.5\;{{\mathrm{s}}}^{-1} $, $ {r}_{1{\mathrm{o}}}=3.75\;{\mathrm{m}}{{\mathrm{m}}{\mathrm{o}}{\mathrm{l}}}^{-1}{\cdot}{\mathrm{L}}{\cdot}{{\mathrm{s}}}^{-1}, $ these are simulation curves made according to Eqs. (8) to (10).

图 6  DCE-MRI不同模型的示意图, 其中, S¹M和S²M分别为第1代和第2代的快门速度模型(SSM), 此外, 本文选择eTofts模型作为处理DCE-MRI的传统药代动力学模型代表, 在S¹M中, 忽略了血管水占比, 只考虑细胞膜内外水交换$ {k}_{{\mathrm{i}}{\mathrm{o}}} $和$ {k}_{{\mathrm{o}}{\mathrm{i}}} $; 在S²M中, 同时考虑$ {k}_{{\mathrm{i}}{\mathrm{o}}} $和$ {k}_{{\mathrm{o}}{\mathrm{i}}} $、$ {k}_{{\mathrm{b}}{\mathrm{o}}} $和$ {k}_{{\mathrm{o}}{\mathrm{b}}}{\mathrm{以}}{\mathrm{及}}{k}^{{\mathrm{t}}{\mathrm{r}}{\mathrm{a}}{\mathrm{n}}{\mathrm{s}}} $血管内$ {k}_{{\mathrm{p}}{\mathrm{e}}} $和$ {k}_{{\mathrm{e}}{\mathrm{p}}} $; 在eTofts模型中, 血管内外和细胞内外的水交换被假设为无穷快

Fig. 6.  The schematic diagrams of different models used in DCE-MRI. Among them, S¹M and S²M represent the first and second generation of the Shutter speed model (SSM). In addition, the eTofts model is selected here as a representative traditional pharmacokinetic model for processing DCE-MRI. In S¹M, the vascular water fraction is ignored, and only the water exchange between intracellular and extracellular spaces is considered (k_io and k_oi). In S²M, both vascular and cellular water exchanges are considered, with the permeability of the contrast agent from the vascular space to the interstitial space being accounted for(k_bo and k_ob). $ {k}^{{\mathrm{t}}{\mathrm{r}}{\mathrm{a}}{\mathrm{n}}{\mathrm{s}}} $, the product of the permeability of the contrast agent in the vasculature ($ {k}_{{\mathrm{p}}{\mathrm{e}}} $ and $ {k}_{{\mathrm{e}}{\mathrm{p}}}) $. In the eTofts model, water exchange between both the vasculature and the cells is assumed to be instantaneous.

图 7  细胞内水分子流出速率$ {k}_{{\mathrm{i}}{\mathrm{o}}} $是胶质瘤中AQP4表达高敏感、高特异性的潜在影像标志物　(a) 左上图为一位胶质瘤患者$ {k}_{{\mathrm{i}}{\mathrm{o}}} $参数图(彩色)的示意图, 其中白色箭头指向处即活检点位置, 对应活检点的免疫组化结果见左下图, 右图为来自19位胶质瘤患者共45个活检点的$ {k}_{{\mathrm{i}}{\mathrm{o}}} $参数和AQP4表达量间的相关性结果, 其中AQP4表达量通过切片中AQP4阳性的细胞所占的比例进行定量化; (b) 左图为在大鼠胶质瘤模型(C6细胞)中第1天注射生理盐水和第2天利用TGN020特异性抑制AQP4后, 肿瘤区域的$ {k}_{{\mathrm{i}}{\mathrm{o}}} $参数图, 右图为注射生理盐水和TGN020后肿瘤区域平均$ {k}_{{\mathrm{i}}{\mathrm{o}}} $的统计结果, TGN020特异性抑制AQP4后, $ {k}_{{\mathrm{i}}{\mathrm{o}}} $显著下降43%(n = 9); *表示统计分析p值小于0.05. 图片改自参考文献[41]

Fig. 7.  The intracellular water efflux rate ($ {k}_{{\mathrm{i}}{\mathrm{o}}} $) is a potentially high-sensitivity and high-specificity imaging biomarker for AQP4 expression in gliomas. (a) The upper left diagram shows a schematic of the $ {k}_{{\mathrm{i}}{\mathrm{o}}} $ map for a glioma patient, with the white arrow indicating the location of the biopsy site. The corresponding immunohistochemical results for the biopsy site are shown in the lower left diagram, the right diagram presents the correlation between $ {k}_{{\mathrm{i}}{\mathrm{o}}} $ parameters and AQP4 expression levels from 45 biopsy sites in 19 glioma patients, with AQP4 expression quantified as the proportion of AQP4-positive cells in the sections. (b) The upper left diagram shows the $ {k}_{io} $ parameter maps of the tumor region in a rat glioma model (C6 cells) on the first day after injection of saline and on the second day after specific inhibition of AQP4 with TGN020. The right diagram presents the statistical results of the average $ {k}_{{\mathrm{i}}{\mathrm{o}}} $ in the tumor region after injection of saline and TGN020, with a significant 43% decrease in $ {k}_{{\mathrm{i}}{\mathrm{o}}} $ after specific inhibition of AQP4 by TGN020 (n = 9). This figure was adapted from the Ref. [41].

图 8  胶质瘤细胞中AQP4表达水平与治疗抵抗存在相关　(a) 低AQP4的细胞亚型以胶质瘤干细胞特征的慢增殖细胞为主, 在替莫唑胺(TMZ)治疗下存活, 细胞核完整, 并且表达更多的治疗抵抗标志蛋白ZEB1; (b) 高AQP4的细胞亚型则以胶质瘤干细胞特征的快增殖细胞为主, 在TMZ治疗3天后细胞核受损, 表现出治疗敏感性, ZEB1为锌指增强子结合蛋白1; DAPI为4', 6-二脒基–2-苯基吲哚; CTV为指细胞增殖荧光示踪剂; 图片改自参考文献[41]

Fig. 8.  The expression level of AQP4 in glioma cells is correlated with treatment resistance: (a) The low-AQP4 cell subtype is mainly composed of slow-proliferating cells with glioma stem cell characteristics, which survive under temozolomide (TMZ) treatment with intact nucleus and express higher levels of the treatment resistance marker protein ZEB1; (b) in contrast, the high-AQP4 cell subtype is mainly composed of fast-proliferating cells with glioma stem cell characteristics. On the third day of TMZ treatment, some cells show nucleus structure damage, which are chemoradiation-sensitive cells. ZEB1 refers to zinc finger E-box-binding homeobox 1; DAPI refers to 4', 6-diamidino-2-phenylindole; and CTV refers to a cell proliferation fluorescent tracer. This figure was adapted from the Ref. [41].

图 10  扩散峰度随扩散时间的变化曲线示意图, 其中, 在较短的扩散时间下, 扩散峰度随扩散时间延长而上升, 其反映了细胞内水分子扩散受微结构的影响, 在较长的扩散时间下, 系统达到“粗粒化”, 扩散峰度随扩散时间下降, 其反映了水分子跨膜交换的信息. 此时, 交换时间$ {\tau }_{{\mathrm{e}}{\mathrm{x}}ex} $可利用图中公式对下降部分拟合得到

Fig. 10.  The schematic diagram of the change in diffusion kurtosis with diffusion time. At shorter diffusion times, the diffusion kurtosis increases with the diffusion time, reflecting the intra-compartmental microstructural effects. At longer diffusion times, the system reaches 'coarse-graining, ' and the diffusion kurtosis decreases with diffusion time, reflecting information about transmembrane water exchange. At this point, the exchange time ($ {\tau }_{{\mathrm{e}}{\mathrm{x}}} $) can be fitted from the decreasing part of the K(t) curve using the formula shown in the figure.

图 9  单扩散梯度和双扩散梯度编码的序列示意图　(a) SDE序列示意图, 在SDE中, 可以使用自旋回波或者受激回波进行水分子交换过程测量, 与自旋回波相比, 受激回波技术允许使用更长的混合时间, 而不会增加TE, 两个扩散梯度开始的时间间隔为Δ, δ为扩散梯度的持续时间, G为扩散梯度的强度; (b) DDE序列示意图. 两对扩散梯度的参数用下标1和2区分, 并由混合时间(mixing time, t_m)隔开

Fig. 9.  The schematic diagram of SDE and DDE sequence: (a) The schematic diagram of SDE sequence, in SDE, either spin echo or stimulated echo can be employed to measure the water molecule exchange process, compared to spin echo, the stimulated echo allows for longer mixing time (t_m) without increasing the echo time (TE), Δ is the time interval between the two diffusion gradient, δ is the duration of the diffusion gradient, and G is the intensity of the diffusion gradient; (b) the schematic diagram of DDE sequence, the parameters of the two pairs of diffusion gradients are distinguished by subscripts 1 and 2, and separated by the mixing time (t_m).

图 11  FEXI序列示意图及其测量双组分系统中水分子跨膜交换的原理简图, 在FEXI序列的过滤模块中, 通过扩散梯度的施加可“过滤”系统扩散率快的水分子信号, 在交换模块, 两个组分的水分子发生跨膜交换引起系统表观扩散率ADC的恢复并在检测模块进行测量

Fig. 11.  The schematic diagram of the FEXI sequence and its principle for measuring transmembrane water exchange in a two-component system. In the filter block, the application of diffusion gradients filters out the signals of water molecules with fast diffusivity in the system. In the mixing block, water exchange between the two components leads to the recovery of the system's apparent diffusion coefficient (ADC), which is measured in the detection block.

图 12  基于不同磁共振测量方法获取的不同生物样本的水分子跨膜交换速率或者交换时间, 图片改自参考文献[125]

Fig. 12.  Transcytolemmal water-exchange rate or exchange time of different biological samples obtained using various MRI or NMR methods. This figure was adapted from the Ref. [125].