Diffusion of water in biological tissues is sensitive to the underlying microscopic structure because macromolecules act as obstacles or sources of electrostatic fields that hinder or restrict the diffusion paths. The diffusion behavior in tissues is known to be non-Gaussian. Moreover, two different water regimes (with distinct activation energies) have been found in brain tissue and assigned to a relatively 'free', bulk-like compartment and water in contact with macromolecules [1,2].

Due to its macromolecular composition, we may thus assume that the extracellular matrix (ECM) component of tissues impacts water diffusion. For cartilage, the ECM is already known to play a role in the generation of MRI contrast [3] and to affect diffusion [4]. Despite its very different structure [5-7], it has been recently suggested that the ECM may also influence the MR signal in brain [8], and iontophoretic studies indicate an influence on ion diffusion in rat brain [9].

To more directly assess an ECM influence on water diffusion, we performed a series of experiments in post-mortem brain tissue samples before and after enzymatic removal of the ECM. Additional experiments were performed in cartilage, whose hyaluronan content is remarkably higher compared to the brain.

Two paraformaldehyde-fixed specimens from goat thalamus and another two from dog ear were collected in $$$7\mathrm{mm}$$$-diameter glass tubes and stored in $$$0.1\;\mathrm{M}$$$, pH $$$7.4$$$ Phosphate-Buffered Saline (PBS) solution at $$$4\;\mathrm{°C}$$$. Each one of the brain and cartilage samples, were treated with hyaluronidase from bovine testis (Sigma H3884), with up to $$$\approx{}4500\;\mathrm{units/ml}$$$, at $$$37\;\mathrm{°C}$$$ for $$$14\;\mathrm{days}$$$, to completely remove the hyaluronan-based ECM. The hyaluronidase-digested cartilage specimen was subsequently (after diffusion MR) also treated with $$$2\%$$$ collagenase (Fluka) at $$$37\;\mathrm{°C}$$$ for $$$14\;\mathrm{days}$$$ to separately digest collagen ECM components. The remaining two untreated samples were used as reference to asses potential changes during storage.

Non-localized PFG diffusion measurements were performed on a custom-built FEGRIS NT $$$125\;\mathrm{MHz}$$$ spectrometer [10] with a gradient amplitude up to $$$35000\;\mathrm{mT/m}$$$ along the main magnetic field. It permits to achieve very short diffusion times ($$$\Delta=2\;\mathrm{ms}; \delta=0.6\;\mathrm{ms}; T_E=4\;\mathrm{ms}; T_R=3\;\mathrm{s}$$$) to minimize restriction and exchange effects. The b-values were varied in 20 steps (up to $$$18000\;\mathrm{s/mm}^2$$$), and the temperature was varied between $$$20\;\mathrm{°C}$$$ and $$$0\;\mathrm{°C}$$$ in steps of $$$2\;\mathrm{°C}$$$. Additionally, $$$T_1$$$ was measured using an inversion-recovery sequence. Diffusion data were fitted to a bi-exponential model, while a mono-exponential fit was used for obtaining $$$T_1$$$. Activation energies and the pool size fractions were extracted from Arrhenius plots.

After the MR acquisitions, the brain samples were sliced and stained for hyaluronan-sensitive markers.

Arrhenius plots for both the brain and the cartilage samples are shown in Fig. 1. Activation energies and pool size fractions, as well as $$$T_1$$$ results are summarized in Fig. 2. It was not possible to determine the activation energy for a 'slow' water compartment in cartilage due to the relatively low signal contribution. $$$T_1$$$ was decreased by $$$\approx{}10\%$$$ to $$$\approx{}20\%$$$ after hyaluronan digestion both in thalamus and cartilage.

Diffusion coefficients and activation energies of both water fractions did not show a measurable change upon treatment in the thalamus sample. Staining results (Fig. 3) confirm that the hyaluronan digestion was effective in this sample. In cartilage, the diffusion coefficient decreased by $$$\approx{}7\%$$$ after hyaluronan digestion but increased by $$$\approx{}15\%$$$ following collagenase treatment, while the activation energy was only marginally affected.

The observed $$$T_1$$$ decrease upon treatment in the thalamus sample is consistent with previous imaging results [9]. However, the underlying microstructural variation is not reflected in a measurable effect on water diffusion. The cartilage sample did indicate an influence on water diffusion from the ECM, which might be explained by the substantially higher hyaluronan content in this tissue type.

Interestingly, the diffusivity decreased after hyaluronidase treatment, which might indicate that macromolecular fragments remain after digestion as obstacles to slow down water mobility. Such fragments might be removed upon subsequent collagenase treatment as indicated by an increase in water diffusivity. Due to the lower ECM content in brain tissues, the effects on diffusion might be too small to be detectable unless high spatial resolution is achieved to better study the quite heterogeneous distribution between different brain regions.