Exploring the Dynamics of Brain Extracellular Space

Main Text

Extracellular space (ECS) morphology in the brain parenchyma is not static and can undergo relatively fast structural changes that occur on the timescale of seconds or faster, for example, during seizures, cortical spreading depression (SD), or UP and DOWN state dynamics in the neocortex. The article by Hrabe and Hrabětová (1) in this issue of Biophysical Journal significantly improves temporal resolution of the integrative optical imaging (IOI) (2, 3, 4) method and enables characterization of such fast dynamical changes in the ECS. The so-called time-resolved IOI (TR-IOI) (1) method utilizes a relatively small fluorescent probe and extends the IOI theory to include time-dependent effective diffusivity. It is then applied to the problem of SD, as will be described below.

Exploration of ECS has greatly increased in recent decades and is now considered an important frontier in our attempts to understand the brain (5). Besides the increasing appreciation of its importance for understanding brain function, this growth in ECS research is also fueled by the emergence of new techniques for studying it. Particularly efficient are techniques that are based on diffusion measurements, for example, real-time iontophoresis (RTI) (6), IOI, or diffusion-weighted MRI (DW-MRI) (7) because the complex geometry of the ECS can be characterized by analyzing the way molecules diffuse within it. Before TR-IOI, other tracer methods (RTI and IOI) could provide some temporal information, but the time resolution in those methods is on the order of tens of seconds, if not minutes, which makes them unsuitable to address many important dynamical phenomena in the brain.

The key factors that allow TR-IOI to achieve such a substantial increase in time resolution of IOI are the use of a small and bright fluorescent probe (<3 nm; 3 kDa dextran tagged with Texas Red) and a new mathematical framework that utilizes a time-dependent effective diffusion constant. In fact, the authors generalize their procedure to use diffusion tensor but with an important caveat that one of the principal axes of the tensor is orthogonal to the imaging plane. This limits the applicability of their anisotropic diffusion framework because in many situations, the principal axes of the diffusion are not known or may vary in time and across the imaging volume. In (1), the authors aim for simplicity and are thus avoiding the estimation of tensors with arbitrary orientation, which would demand acquiring fluorescent images in multiple, tilted, and mutually nonparallel planes. An important part of their simplification is the mathematical formalism for normalization of the three-dimensional cloud of the diffusing probes, which enables the estimation of the time-dependent effective diffusion tensor, $D_{ij}^{\ast }\left(t\right)$, directly from the time-dependent moments of the two-dimensional fluorescent images, ${\mu }_{ij}\left(t\right)$, using $D_{ij}^{\ast }\left(t\right)=\left(1/2\right)d{\mu }_{ij}\left(t\right)/dt$.

To test TR-IOI, the authors use it to study the effects of SD on ECS. It has been known for a long time that SD significantly reduces the volume fraction of the ECS (8), and many subsequent studies were made to address the temporal and other aspects of such change. For example, DW-MRI was used to elucidate the time course and the spatial extent of the induced SD (9), which, importantly, can also be used to study SD in vivo; however, DW-MRI alone cannot resolve if the decrease in diffusivity is due to intracellular or extracellular changes. With RTI, it is also possible to measure the absolute volume fraction of the ECS, not just the relative changes, but it requires steady conditions for the duration of the diffusion measurement, which is often longer than the time course of SD. By shortening the time of the diffusion measurements to 20 s, RTI was successfully applied to measuring the effects of SD on ECS (10). With the TR-IOI technique, Hrabe and Hrabětová (1) are aiming to achieve temporal resolution of ∼1 s during the entire SD episode. Fig. 1 A in (1) depicts the basic setup in which SD is induced at Site 1 by injection of KCl to Schaffer collaterals in the in vitro preparation of a rat hippocampal slice. Upon arriving at Site 2, the SD impacts the diffusion of the probes, which is recorded as a time series of fluorescent images from which ${\mu }_{ij}\left(t\right)$ and $D_{ij}^{\ast }\left(t\right)$ are calculated. The results in Fig. 3 B show a sudden drop in extracellular diffusivity in a matter of few seconds, and this time resolution might have been even further blurred by the fact that the propagating front of SD takes a finite time to affect the whole cloud of diffusing probe particles. A more fundamental limitation to the time resolution is the fact that the effective diffusivity is itself time-dependent even when the restricting environment is static, for example, dependent on the time to reach the tortuosity limit. This makes the interpretation of the obtained instantaneous effective diffusion tensor, $D_{ij}^{\ast }\left(t\right)$, difficult. As the authors point out, their formalism contains other assumptions inherited from IOI, which can affect the accuracy of their results, as can be gleaned from Fig. 3 B in which the diffusivity of the trapped dextran molecules becomes slightly negative. Nevertheless, the qualitative timeline of SD effects and the subsequent recovery are in a good agreement with the expected results.

The exploration and our understanding of the role of ECS in brain function are in their infancy, and many challenges in measuring ECS properties still remain. Diffusion-based techniques are an important tool for exploring ECS, but no single probe or technique will suffice to reveal all the aspects of its morphology or function. Besides having many different factors that can cause the hindrance of the diffusing molecules or even anomalous diffusion, the ECS morphology itself is highly heterogeneous. It is tens of nanometers in width (distance between the neighboring outer cell membranes) on average, but in some regions, it can contain larger pools of interstitial fluid, similar to that of cerebrospinal fluid, whereas in others it can be as tight as 2 nm, for example, where gap junctions form large plaques of connexin channels. The biophysics of such tight ECS is largely unexplored, and research in gap junctions is almost exclusively focused on their role in intracellular communication. By allowing the intracellular diffusion of very small fluorescent probes (<1 kDa), gap junctions can also violate the assumption of IOI that the long-range diffusion occurs strictly through the ECS. This is usually not a problem, but the presence of significant hydrodynamic bulk flow in larger ECS regions and elsewhere can potentially confound estimates of the effective diffusion constant, volume fraction, tortuosity, probe uptake, and other parameters inherent to RTI, IOI, and TR-IOI models. Although controversies exist about the relative importance of the convective versus diffusive transport in the brain interstitial space as well as about its relationship with the glymphatic system, a recent study suggests that the bulk flow might be an important component of the transport of larger molecules (see (11) and references within).

The ECS microenvironment is generally complex, containing extracellular matrix, ions, proteins, hormones, and other signaling molecules. Its morphology significantly impacts the neuronal as well as neuron-glia interactions. Being defined by the boundaries of individual neural and glial cells, the ECS has roughly a foam-like appearance in the regions of the tissue where the cell bodies are densely packed. In other regions, because of very complex shapes of individual neurons and glia, the morphology of ECS is also expected to be extremely complex, particularly in the neuropil, which has a relatively small number of cell bodies. Nevertheless, a detailed study of the high-resolution electron microscopy (EM) images (12) reveals a high degree of organization of ECS even there when analyzed in terms of planar and cylindrical pores (3) (called “sheets” and “tunnels” in (12)). Further research might reveal how such and other features of the ECS morphology lead to more efficient signaling through volume transmission as well as uncover a potential hierarchical network of pores that work in concert with the glymphatic system. What EM images cannot reveal is the relevant temporal dynamics of the ECS, and to fully understand its functioning, a high-resolution temporal imaging might be needed. The work by Hrabe and Hrabětová (1) is an important step in that direction.

Editor: Alexander Berezhkovskii.