New paradigm to assess brain cell morphology by diffusion-weighted MR spectroscopy in vivo

Abstract

The brain is one of the most complex organs, and tools are lacking to assess its cellular morphology in vivo. Here we combine original diffusion-weighted magnetic resonance (MR) spectroscopy acquisition and novel modeling strategies to explore the possibility of quantifying brain cell morphology noninvasively. First, the diffusion of cellspecific metabolites is measured at ultra-long diffusion times in the rodent and primate brain in vivo to observe how cell long-range morphology constrains metabolite diffusion. Massive simulations of particles diffusing in synthetic cells parameterized by morphometric statistics are then iterated to fit experimental data. This method yields synthetic cells (tentatively neurons and astrocytes) that exhibit striking qualitative and quantitative similarities with histology (e.g., using Sholl analysis). With our approach, we measure major interspecies difference regarding astrocytes, whereas dendritic organization appears better conserved throughout species. This work suggests that the time dependence of metabolite diffusion coefficient allows distinguishing and quantitatively characterizing brain cell morphologies noninvasively.cell morphology | noninvasive histology | diffusion-weighted NMR spectroscopy | numerical simulations | metabolites T he brain is one of the most complex organs, and it has defined an inexhaustible field of research over the last centuries. Unfortunately, brain's complexity is paralleled by the difficulty in examining it noninvasively. Some fundamental questions regarding morphological modifications of neurons and astrocytes along brain development, aging, or disease, as well as interspecies differences, can only be investigated postmortem using histology, the current gold standard to study cellular morphology. The development of a noninvasive neuroimaging tool to evaluate and monitor brain cell morphology under normal and pathological conditions in vivo would thus represent a major breakthrough.MRI and magnetic resonance spectroscopy (MRS) techniques have opened new doors for examining brain tissues in vivo at both meso-and macroscales. Diffusion-weighted (DW)-MRI and -MRS, which allow the investigation of the diffusion process of endogenous molecules in biological tissues at these scales (1), have made it clear that cell architecture has a critical influence on molecular displacement (2-5). To quantitatively evaluate the impact of cell structure on measured molecular diffusion, mainly two modeling strategies have been developed. The first approach consists in performing numerical simulations of many particles diffusing in arbitrary geometries (e.g., defined by 3D meshes) mimicking "realistic" cell architectures (6-9). Because these realistic geometries are generally built directly from microscopy data rather than being described and generated by a (small) set of parameters, and because simulations are extremely computationally demanding, this approach does not seem adapted to fit experimental data. The second approach consists in simplifying cell architec...

Modeling diffusion of intracellular metabolites in the mouse brain up to very high diffusion‐weighting: Diffusion in long fibers (almost) accounts for non‐monoexponential attenuation

“…Diffusivity values reported in Table 1 are in very good agreement with those recently extracted from ADC measurements performed at very long t d (and low b) in mouse brain (Table 1 in (17)). However, the only exception is again NAA, for which the estimated D intra here is significantly higher than that reported in (17).…”

“…However, the only exception is again NAA, for which the estimated D intra here is significantly higher than that reported in (17). This difference can be well explained by considering the effect of a slowly diffusing, highly restricted NAA pool on ADC time dependence at long t d .…”

“…5 shows the ADC time dependence simulated from a synthetic tissue of 2500 cell-graphs generated according to the new modeling paradigm and the morphometric statistics reported in (17) for NAA, adding a variable small fraction v of not-diffusing NAA. When fitting these long t d ADCs using the modeling approach described in (17), the effect of a small fraction of slowly diffusing NAA pool is to decrease the value of the estimated D intra , while not affecting the evaluation of the morphometric statistics describing the NAA compartment morphology, demonstrating the robustness of the approach to extract morphometric statistics despite the potential presence of a small immobile metabolite fraction. We found that the true D intra value is underestimated according to the linear relation: D intra * = D intra - 0.64 v , where D intra * is the estimated D intra from ADC time dependence at long t d .…”

Axon Diameters and Myelin Content Modulate Microscopic Fractional Anisotropy at Short Diffusion Times in Fixed Rat Spinal Cord

“…In fact, the values reported in [44,45]. Further studies are needed to establish which underlying biological components give rise to such high µFA in gray matter, but dendrites, astrocyte branches, and nonmyelinated or myelinated axons traversing gray matter could be suspected [93,99]. Time-dependent or spectroscopic experiments on metabolites could provide insight into such questions in the future.…”

Influence of the Size and Curvedness of Neural Projections on the Orientationally Averaged Diffusion MR Signal

“…At the extreme end of the spectrum, for structures so short that $\mathrm{scriptl}\gg \sqrt{D\delta }$, entering into regime C, no curvature is tolerated if the tail q −1 is to appear. These considerations will have to be revisited if one measures the diffusion of molecules other than water, due to differences in their diffusion characteristics [48, 49, 50]. …”