Quantitation of the physicochemical properties of myelin using Nile Red fluorescence spectroscopy

Abstract

Myelin is a vital structure that is key to rapid saltatory conduction in the central and peripheral nervous systems. Much work has been done over the decades examining the biochemical composition and morphology of myelin at the light and electron microscopic levels. Here we report a method to study myelin based on the fluorescent probe Nile Red. This lipophilic dye readily partitions into live and chemicallyfixed myelin producing bright, well‐resolved images of the sheath. Using spectral confocal microscopy, a complete emission spectrum of Nile Red fluorescence can be acquired for each pixel in an image. The solvatochromic properties of Nile Red cause its emission spectrum to change depending on the polarity of its local environment. Therefore, measuring spectral shifts can report subtle changes in the physicochemical properties of myelin. We show differences in myelin polarity in central versus peripheral nervous system and in different regions of central nervous system white matter of the mouse brain, together with developmental and sex variations. This technique is also well suited for measuring subtle changes in myelin properties in live ex vivo white matter specimens. We also demonstrate how light deprivation induces a myelin polarity change in adult mouse optic nerve underscoring a continuing myelin plasticity in response to axonal activity well into adulthood. The Nile Red spectroscopic method allows measurement of subtle physicochemical changes in myelin that can importantly influence its electrical properties and by extension, conduction velocities in axons.

Nile Red is a fluorescent dye that brightly labels fatty substances like the insulating wrap around nerve fibers in the nervous system, called “myelin.” By carefully measuring the changes in the spectrum of light emitted by this probe when labeling myelin, we can detect subtle biochemical differences that reflect physiological changes or changes indicative of early disease.

Abbreviations

aCSF

artificial cerebrospinal fluid

ANOVA

analysis of variance

CNS

central nervous system

COVID

coronavirus disease

DMSO

dimethyl sulfoxide

NA

numerical aperture

NR

Nile Red (9‐(diethylamino)‐5H‐benzo[a]phenoxazin‐5‐one)

NRER

Nile Red emission ratio

PBS

phosphate‐buffered saline

PFA

paraformaldehyde

PNS

peripheral nervous system

Tukey's HSD

honestly significant difference

1. INTRODUCTION

The intricate architecture of the nervous system, with its vast network of neurons and supporting glial cells, is foundational to the complex functionality and adaptability observed in mammalian species. Central to this architecture is the process of myelination, wherein myelin sheaths, composed predominantly of lipids, envelop axons to ensure rapid and efficient saltatory conduction (Stadelmann et al., 2019). This lipid‐rich insulation is pivotal not only for enhancing nerve conduction velocity but also for maintaining the integrity and functionality of neural circuits (Bonetto et al., 2021). Previously, we reported a highly sensitive method for detecting physicochemical changes in myelin using the fluorescent solvatochromic dye Nile Red (NR) coupled with spectral confocal microscopy (Teo et al., 2021), particularly within the context of myelin pathology. Nile Red has emerged as a pivotal tool in this exploration. Its molecular structure, containing a diethylamino group as an electron donor and a carbonyl group as an electron acceptor, forms a push‐pull chromophore that responds sensitively to changes in environmental polarity (Ma et al., 2024). This responsiveness allows NR to act as a reliable reporter of polarity in a variety of environments, including cell membranes (Mukherjee et al., 2007) and myelin (Arnaud et al., 2009), providing insights into the subtle biochemical shifts that occur within these lipid‐rich environments. Moreover, the capacity of this method to estimate the dielectric constant of myelin introduces a novel approach that sheds light on subtle functional properties of this vital structure (Teo et al., 2021).

Traditional studies of myelin, typically focused on pathology (including ultrastructure) and bulk biochemistry, might not fully elucidate the nuanced dynamics of its properties in the intact state, particularly in terms of regional differences, developmental stages, responses to physiological stimuli, that is, adaptive myelination, and subtle effects of pathological states (Bartzokis, 2004; Caprariello et al., 2018; Cullen et al., 2019; Forbes & Gallo, 2017; Norton, 1984). The current study focuses on the ability of NR to report quantitative changes in myelin polarity under various physiological conditions within the mouse peripheral (PNS) and central nervous systems (CNS). Additionally, our study introduces the capability of high‐resolution spectroscopic imaging of live myelin, an approach that offers a new dimension to understanding myelin's physiology in its native state. We also explore the influence of myelin water on measured polarity. Understanding these aspects is instructive for elucidating the functionality of myelin given its strong dependence on passive electrical characteristics, including capacitance, which in turn is governed by polarity/dielectric constant.

2. MATERIALS AND METHODS

2.1. Animals and animal care

All animal experiments were approved by the Animal Care Committee at the University of Calgary (ethics approval #AC22‐0149) using standards set out by the Canadian Council on Animal Care. C57Bl6 (Charles River Laboratories, RRID:IMSR_CRL:027) mice were used for all experiments. Mice were housed in Tecniplast Greenline GM500 cages with ad libitum access to water and standard chow (0006955, Pico‐Vac® Mouse Diet 20, LabDiet). Each cage contained up to five same‐sex littermates.

2.2. Mouse tissue harvesting

Central nervous system tissues, including optic nerves (12 male mice, 15 weeks old), brains (31 mice, 3 female and the rest male mice, 6 were 31 weeks old, the rest were 13–15 weeks old), and dorsal columns (8 male mice, 15 weeks old), were harvested with the protocol varying based on the developmental stage of the mice (neonatal vs. adult). Adult animals were deeply anesthetized by 600 mg/kg of sodium pentobarbital (DIN. no. 02333708 Rafter 8). Intracardiac perfusion was performed with 12 mL of room‐temperature phosphate‐buffered saline (PBS), followed by 12 mL of ice‐cold 4% paraformaldehyde (PFA). Tissues were then postfixed in 4% PFA at 4°C overnight. Cryoprotection was achieved through a sequential sucrose treatment, initially in 20% sucrose, (Cat. No. S5‐500, Fisher Scientific) until tissue descent, followed by immersion in 30% sucrose. Brain tissues were then encapsulated in an optimal cutting temperature compound and frozen in isopentane cooled by dry ice. Coronal sections ranging from 20 to 100 μm thickness were cut using a cryostat and collected on VWR Superfrost Plus Micro Slides (Cat. No.48311‐703), ensuring three region‐matched sections per slide. For optic nerves and dorsal columns of adult mice, a similar perfusion and fixation protocol was employed. A 1.2 cm segment of the cervical spine was excised and either fixed as above or transferred for live imaging. Dorsal roots and sciatic nerves were harvested in a similar manner. Five neonatal male mice were euthanized by exposure to 10–15 min of profound hypothermia/hypercarbia using an ice block placed in a CO₂ chamber, after which their movement gradually ceased and rigor set in. At this time, a tail pinch test was conducted to confirm unresponsiveness to deep pain, then animals were sacrificed by decapitation, and dorsal roots and sciatic nerves were carefully harvested with minimal delay.

2.3. Ex vivo live imaging of dorsal columns

Tissues were harvested from three adult male mice under deep anesthesia as above, which included intracardiac perfusion with chilled, oxygenated artificial cerebrospinal fluid (aCSF). After cranial access through a craniofacial incision, a cervical laminectomy from the atlas to the T3 vertebra exposed the dorsal column. The brain–spinal cord complex, with the spinal cord still seated on intact vertebrae anteriorly, was then immersed in chilled, oxygenated aCSF. The cervical spinal segment was trimmed to 1–1.2 cm, focusing on the cervical area while preserving the ventral vertebral integrity for stable mounting in the imaging chamber. This arrangement prevented tissue tilt and facilitated motion‐free imaging. After the staining process, the spinal cord explants were secured within a chamber and were perfused with oxygenated aCSF at 35°C for 1 h before imaging. After the imaging was complete, the tissues were quickly fixed at room temperature at 4% PFA for 5 min, then transferred to a 4°C refrigerator for overnight storage. The following day, the fixative was replaced with 1x PBS. Additionally, two adult male explants remained in vials with temperature‐controlled oxygenated aCSF for up to 4 h, serving one‐time point collection experiments. After this period, tissues were fixed with 4% PFA.

2.4. Nile Red staining

A stock solution of NR (Cat. No. 19123, Millipore Sigma, USA) was prepared at a concentration of 6 mM in dimethyl sulfoxide (DMSO, Cat. No. 317275, EMD Millipore) and stored at −20°C for future use. Depending on the type of specimen (e.g., brain sections vs. intact optic nerves), the working concentration of NR varied from 10 to 40 μM in PBS. Fixed frozen tissue sections were stained with NR for a duration of 10 min, followed by a 5 min wash in PBS to remove excess dye. Subsequently, the stained tissue sections were placed in a PBS bath on a glass microscope slide for imaging. A water‐immersion objective was employed without mounting media or coverslips.

2.5. Spectral image analysis

Spectral fluorescence images were acquired using an upright Nikon A1RMP spectral confocal microscope using 488 nm laser excitation (linearly polarized) with a 25× 1.1NA water immersion objective. Fluorescence emission was collected over a spectral range of 510–740 nm, at a resolution of 10 nm, with each resulting spectrum being the average of four scans to enhance the signal‐to‐noise ratio. Subsequently, each spectrum underwent normalization to facilitate comparison and spectral decomposition, employing ImageTrak image analysis software written by PKS (available at http://stysneurolab.org/imagetrak/).

To visualize multichannel spectral images, we converted them into “truecolor” images, simulating the perception of the human eye. To further emphasize subtle spectral shifts, we applied spectral decomposition by selecting two reference spectra, representing the highest and lowest polarity in the images, then translated these data into pseudocolor images, assigning a pixel color from violet to red depending on its similarity to the two selected bracketing reference spectra. Thus, violet denotes areas of low polarity, while red marks high polarity zones. To quantify spectral shifts, regions of interest were first manually drawn to enclose structures of interest (e.g., dashed regions in Figure 2 outlining the corpus callosum), then the Nile Red emission ratio (NRER) was computed by splitting the emission spectrum at 600 nm into two bands and computing the ratio of areas above and below this split point for each 2 × 2‐pixel kernel in the image. All such kernel ratios were averaged into a single mean NRER.

FIGURE 2.

Nile Red spectroscopy reveals regional (a, b: 3 months old corpus callosum), age‐related (c, d: medial corpus callosum), and sex (e, f: medial corpus callosum) differences in myelin polarity within the adult mouse corpus callosum. Analyzed areas indicated by dashed rectangles. Myelin from older mice tended to have a lower polarity, while female mice exhibited higher polarity in the medial corpus callosum compared to males of the same age (3 months old). NRER, Nile Red emission ratio. Scale bar: 40 μm.

2.6. Optic nerve and dorsal root dehydration

Optic nerves and dorsal roots were dehydrated by initially rinsing with distilled water to eliminate any residual PBS, followed by placement on a warm plate set at 27°C for 1 h. This temperature was selected to encourage gradual dehydration without compromising the tissue's integrity. The final step entailed an overnight placement in a vacuum desiccator at 20°C to ensure complete drying. For imaging dehydrated tissues, an extra‐long working distance 20× air objective lens with an NA of 0.45 was used for spectral imaging.

2.7. Statistical analysis

Statistical significance was calculated using Student's t‐tests for the comparison of two groups unless indicated otherwise, or one‐way ANOVA for multiple groups, followed by Tukey's HSD for multiple comparisons. The Shapiro–Wilk test was used to test for normality and Bartlett's test for homogeneity of variance. All statistical calculations were performed using Python v3.11.5 and SciPy v1.11.3. Each data point in the graphs and quoted Ns represents a unique animal. No formal power calculations were performed. The number of animals used was based on our previous work with similar experiments (Jhelum et al., 2020). No test for outliers was performed and no animals or data points were excluded.

3. RESULTS

Paraformaldehyde‐fixed myelinated axons from the mouse CNS (optic nerve, medial corpus callosum, dorsal column) and PNS (dorsal roots, sciatic nerve) were stained with NR, and spectral images were acquired. Representative pseudocolored spectral micrographs are shown in Figure 1a. Myelin from the larger PNS axons stained brightly and is well visualized, together with nodes of Ranvier and Schmidt‐Lanterman incisures characteristic of larger myelinated fibers (Reynolds et al., 1994). The smaller CNS axons were less distinct, but myelin figures were still easily identifiable. The pseudocoloring visually shows subtle differences in polarity, with optic nerve myelin being more polar (redder hues) compared to the other CNS regions, and to PNS axons where the myelin was noticeably bluer. Representative spectra are plotted in Figure 1b showing a slightly more red‐shifted emission in optic nerve myelin, indicative of a more polar environment (Tuck et al., 2009). A convenient way to quantitate such a spectral shift is to calculate the ratio of the areas under the normalized spectrum above and below the emission peak, in this case, 600 nm. The Nile Red emission ratio (NRER), aggregated from all CNS versus PNS regions, revealed that on average, CNS myelin was more polar than in the PNS (Figure 1c). Analyzing CNS and PNS regions separately shows that peripheral myelin was similar in dorsal roots and sciatic nerve, whereas in the CNS, there were significant regional differences underscoring heterogeneity despite similar morphological appearance (Figure 1d) (Peters et al., 1991; Tennekoon et al., 1977; Zhang et al., 1998).

FIGURE 1.

Myelin polarity reported by Nile Red spectroscopy in the CNS and PNS. (a) Pseudocolor micrographs visually illustrating myelin polarity of paraformaldehyde‐fixed central versus peripheral myelinated axons from adult mice. In general, CNS axons had more polar myelin (redder hues) than in the PNS, shown more directly by the spectral overlay (b), and quantitatively in the bar graph (c). In addition to spectral information, distinct features of PNS myelin such as Schmidt‐Lanterman incisures (red arrowhead) and nodes of Ranvier (yellow arrowhead) can be readily visualized. (d) Quantitative summary of polarity differences in various regions of the CNS and PNS. p values by Student's t‐test (c) or one‐way ANOVA with Tukey's HSD multiple comparisons (d). ANOVA, analysis of variance; CNS, central nervous system; NRER, Nile Red emission ratio; PNS, peripheral nervous system. Scale bar: 10 μm.

Regional differences within the CNS were further dissected by analyzing the medial versus lateral and anterior versus posterior corpus callosum in 3‐month‐old age‐matched male mice. Myelin polarity varied significantly in different regions of the same structure (Figure 2a,b). White matter in general, and myelin in particular, are known to change with aging (Bartzokis, 2004; Wozniak & Lim, 2006). Figure 2c,d shows age‐dependent changes in polarity of the medial corpus callosum in male mice at 12 versus 31 weeks of age, with the myelin of the corpus callosum exhibiting a significantly lower polarity in the older mice. There were also sex differences in myelin polarity in the medial corpus callosum, with a higher polarity measured in 3‐month‐old female compared to male mice (Figure 2e,f).

We also examined changes in myelin polarity during early myelination in the PNS, where individual sheaths could be more easily resolved. Thin sheaths beginning to envelope dorsal root axons at postnatal day 2 (Lizarraga et al., 2007) had a significantly higher polarity than thicker, more mature myelin at 1 month of age (Figure 3a,b). Interestingly, in contrast to immature thin myelin, thicker adult PNS myelin was not homogeneous, showing a polarity gradient from the abaxonal (less polar) to the adaxonal (more polar) surface. While spinal roots and peripheral nerves are both considered part of the PNS, as Schwann cells provide myelination in both regions, on average sciatic nerve myelin was more polar than in dorsal roots (Figure 3c). The larger PNS axons also allowed us to study myelin polarity differences as a function of axon diameter, which was not possible in the much finer central fibers. We arbitrarily divided dorsal root axons into small (<3 μm in diameter) versus large (>3 μm) with the former exhibiting a significantly lower polarity than larger fibers (Figure 3d,e).

FIGURE 3.

Developmental progression and heterogeneity of myelin polarity in mouse dorsal root axons. (a) Very young (postnatal day 2) and adult (30 days old) roots are shown in truecolor in the top panels, and in pseudocolor below to emphasize spectral differences. As the animals matured, dorsal root myelin became thicker and assumed a much less polar character reflected by the bluer hues. Schmidt‐Lanterman incisures are evident in the mature axons. (b) Bar graph shows quantitative differences. (c) In the adult, myelin polarity differences were also observed between dorsal roots and sciatic nerve axons. (d, e) Myelin also differed in smaller (yellow arrows) versus larger dorsal root axons (green asterisk) in the adult, with the former exhibiting a lower polarity than the latter. NRER, Nile Red emission ratio. Scale bar: 10 μm.

Being a non‐toxic vital dye, NR is well suited for labeling live cells and tissues (O'Rourke et al., 2009). Figure 4 shows an example of live ex vivo mouse dorsal column stained with NR and maintained for several hours in a heated, oxygenated perfusion chamber. Brightly stained myelin sheaths are nicely resolved and show little morphological change after 4 h. Despite this apparent structural stability, quantitative polarity measurements indicated a gradual drift to lower values beginning at about 2 h, reflecting a very subtle change in myelin nanostructure and/or biochemical composition. In contrast, while fixation generally preserves morphology well, this processing step significantly reduced myelin polarity (Figure 4d).

FIGURE 4.

Nile Red (NR) spectroscopy in live ex vivo CNS axons. (a) Mouse dorsal columns maintained in a warmed oxygenated perfusion chamber showed excellent morphological integrity. (b) Morphology was preserved for several hours ex vivo and was similar to that seen with paraformaldehyde fixation. (c) However, despite the apparent stability of microscopic appearance of dorsal column myelin, NR spectroscopy revealed a gradual shift of polarity to lower values (T is the number of hours maintained ex vivo, at the end of which the dorsal columns were fixed). (d) Fixation, while preserving morphology, induced the largest spectroscopic shift (T0 implies the unfixed samples were measured immediately without further incubation). NRER, Nile Red emission ratio. Scale bar: a, 50 μm; b, 20 μm.

Myelin is a highly organized anisotropic liquid crystalline substance (Beaulieu & Allen, 1994; de Campos Vidal et al., 1980; García‐García et al., 2024), therefore, it is likely that lipophilic probes such as NR would partition into myelin lamellae in a similarly ordered manner. Since most fluorophores possess a dipole moment which imparts preferential excitation depending on the angle of a linearly polarized exciting field (Tajalli et al., 2008), this can be probed by exciting a myelinated axon with the laser delivered at various polarization angles. This effect is shown in Figure 5. In panel A, dorsal column axons stained with NR were imaged using 2‐photon excitation at 900 nm with fluorescence collected in a single channel in non‐spectral mode. Signal intensity was highest with the linearly polarized laser oriented parallel to the axons (0°) and diminished monotonically as the polarization was rotated to 45° and finally at right angles (90°) to the direction of the myelin sheaths (Figure 5b), indicating that the NR molecules themselves were strongly aligned after partitioning into the myelin lamellae. A similar effect was observed in larger PNS dorsal root axons which were imaged using conventional 1‐photon excitation and rotating the sample rather than the laser field which produces the same effect (Figure 5c,d). As with CNS axons, myelin of peripheral axons also exhibited a significant reduction in intensity when excited with polarized light oriented at right angles to the direction of the myelin sheaths. The 90°‐excited axons also exhibited myelin with a noticeably red‐shifted hue indicating a concomitant spectral shift to longer wavelengths (see below).

FIGURE 5.

(a, b) Anisotropy of myelin demonstrated in paraformaldehyde‐fixed Nile Red‐stained dorsal columns using 2‐photon excitation with linearly polarized pulsed laser light. The highest signal was obtained when the polarization of the laser field was aligned with the direction of the myelin structures (0°) and monotonically diminished as the polarization angle was rotated to 90°. This reflects the highly anisotropic architecture of myelin lamellae and the orderly insertion of dye molecules into the lipid bilayers. (c, d) Peripheral myelin exhibited similar anisotropy with 1‐photon excitation and confocal microscopy, with rotation of the sample rather than the laser polarization which induces the same effect. (e, f) Dehydration of the sample induced marked blue shift of the NR emission spectrum indicating a substantial reduction of tissue polarity. In the absence of water, myelin anisotropy was greatly reduced as shown by the diminished effect of polarization angle. (g, h) Unexpectedly, changes in polarization angle also induced a shift in the NR emission spectrum, but only in the hydrated state (see main text). NRER, Nile Red emission ratio. Scale bar: a, 10 μm; c, 50 μm; e, 100 μm.

Nile Red emission spectrum is strongly influenced by the polarity of its immediate environment (Ghoneim, 2000). Within myelin lamellae, the net environmental polarity will be determined by the chemical composition of myelin membrane lipids, together with the influence of highly polar water molecules in close proximity within the major dense line (the cytoplasmic spiral of myelin) and the intraperiod line (the extracellular myelin compartment) (Kister & Kister, 2022). In order to distinguish between the effects of lipids versus water, we imaged NR‐stained dorsal roots in the PFA‐fixed hydrated state, and after strong dehydration. Figure 5e shows the dramatic shift to much lower polarity after removal of water, indicated by the violet hue of the pseudocolored images. As mentioned in the previous section, 90°‐excited dorsal root myelin was redder indicating that NR molecules excited by this polarization angle were likely partitioned into a different nanoenvironment which had a more polar character. This is more directly shown by the spectral overlay in Figure 5g where the 90° spectrum (green trace) was red‐shifted compared to 0° (red trace). This is quantitated using emission ratios in Figure 5h (“hydrated” bars). Notably, both the intensity changes (Figure 5f) and spectral shifts (Figure 5g, black and gray traces) disappeared after dehydration, underscoring the critical importance of water for the maintenance of anisotropic properties of myelin. Finally, the influence of water on the polarity of the tissue is also illustrated in optic nerve (Figure 6a,b).

FIGURE 6.

As with peripheral nervous system axons, dehydration of optic nerve (a central white matter tract) induced a significant shift toward lower polarity as reported by Nile Red fluorescence and the notable shift toward bluer hues of the pseudocolor image (a). Nile Red emission ratios are shown quantitatively in the bar graph (b). NRER, Nile Red emission ratio. Scale bar: 100 μm.

In Figure 7, we show a proof‐of‐principle experiment illustrating how NR spectroscopy can report physiological changes in myelin polarity. It has been shown that visual stimulation alters myelination in optic nerves (Etxeberria et al., 2016). After being reared in a customary 12‐h alternating light–dark cycle, we placed a cohort of six 13‐week‐old adult male mice into complete darkness for a 2‐week period. Six age‐matched mice were maintained in cages with regular lighting. Optic nerves were harvested at 15 weeks of age and imaged with NR spectroscopy after PFA fixation. In the hydrated state, no significant differences were detected in optic nerve polarity in the control versus dark‐exposed animals. However, after dehydration, dark‐exposed optic nerves exhibited a significantly increased polarity.

FIGURE 7.

Sensory deprivation induced a change in myelin polarity in adult mice. (a) Two weeks of darkness induced a significant shift in optic nerve polarity toward higher values in adult (13‐week‐old) mice compared to an age‐matched cohort maintained in a standard 12‐h light cycle environment. However, this polarity change was only evident in dehydrated samples (b vs. c) suggesting that a biochemical change in lipid polarity, rather than water content, was mainly responsible. These results indicate that subtle modulation of the biochemistry of myelinated axons can occur even in the postdevelopmental stage in response to action potential traffic, changes that can be measured by Nile Red spectroscopy. NRER, Nile Red emission ratio. Scale bar: 100 μm.

4. DISCUSSION

The ability of NR to spectroscopically report changes in the polarity of myelin is based on two unique properties: (1) its lipophilic character, which promotes partitioning of dye molecules from the aqueous environment preferentially into the lipid‐rich multilamellar domains of the myelin sheath and (2) its solvatochromic properties that underpin shifts in its emission spectrum dependent on the polarity of the local environment. Together, these result in a dye that brightly stains PNS and CNS myelin (Figure 1) and also has the ability to sensitively report its physicochemical characteristics. There exist other dyes with similar properties including Prodan, Laurdan, Nile Blue, coumarin 102, and other 7‐amino coumarins, to name a few (Gilani et al., 2012; Hessz et al., 2014; Klymchenko, 2023; Parasassi et al., 1998; Wagner, 2009). These probes are part of a larger family of environmentally sensitive fluorophores whose emission spectra and lifetimes can depend on a variety of factors such as polarity, viscosity, temperature, pH, membrane tension, and other properties of their local environment (Lakowicz, 2006a; Ma et al., 2024). The mechanisms by which the photophysical behavior of these dyes is modulated by their surroundings are complex (Lakowicz, 2006b). In the case of NR, intramolecular charge transfer is thought to occur in the excited state, resulting in charge separation across the length of the NR molecule between the electron‐donating diethylamino group and the electron‐accepting quinoid moiety (Ghoneim, 2000; Guido et al., 2010; Ya. Freidzon et al., 2012). The resulting large excited‐state dipole moment will then preferentially interact with more polar solvent molecules, inducing a reorientation of solvent dipoles around the excited fluorophore, thus producing a shift in the emission spectrum (Kawski et al., 2008; Krishna, 1999). In general, the more polar the solvent, the more red‐shifted and weaker the emitted fluorescence (Greenspan & Fowler, 1985).

We reasoned that because of the hydrophobic nature of myelin, and the highly lipophilic properties of NR, the former becomes the “solvent” for the dye molecule, which can then report physicochemical properties of the myelin membranes. With respect to biological membranes, lipophilic dyes such as NR will tend to partition into the lipid bilayers in an ordered fashion, aligning with the acyl chains of membrane phospholipids (Klymchenko, 2023). This is confirmed by the dramatic effects of altering the polarization angle of the exciting laser field with respect to the orientation of the myelin sheaths (Figure 5). Fluorophores are maximally excited when their dipole moments are aligned with the polarization angle of the exciting light (Kawski et al., 2008; Tajalli et al., 2008). When the myelin sheath is oriented at right angles to the polarization of the laser, NR fluorescence was greatly reduced, indicating that the NR molecules were largely oriented anisotropically within the lipid bilayers. It stands to reason that if the bilayers become disordered, so will the NR dye molecules, so ratiometric polarized images acquired at 0° and 90° could provide additional information about membrane integrity. By virtue of their polarity‐sensing properties with induced spectral shifts, dyes such as NR can report liquid ordered versus disordered phases of lipid membranes, and therefore, the nanoscopic integrity of cell membranes such as myelin lamellae, with exquisite sensitivity (Klymchenko, 2023). One unexpected observation was the effect of polarization angle on the emission spectrum of NR (Figure 5g,h). In the hydrated state that more closely approximates the physiological condition, excitation with suboptimal polarization (90°) not only produced much less intense fluorescence but also a more red‐shifted spectrum suggesting that NR molecules excited by this polarization angle emitted from a more polar environment. These differences disappeared after dehydration. Taken together, these data suggest two distinct compartments in the myelin membranes, promoting different orientations of the NR dye molecules, and exhibiting different local polarities. The nature of these two populations is unknown but could offer an even more detailed reporting of myelin physicochemical makeup by analysis of both relative intensities, and spectra, of image pairs acquired at orthogonal polarization angles.

The local polarity experienced by an NR molecule partitioned into a membrane, which includes multiple wraps of myelin lamellae, is subject to two influences. First, the composition of lipids in the immediate vicinity of the probe, and second, the amount of water just outside the lipid bilayer. Lipids exhibit different polarities, with cholesterol being less polar, while phospholipids, for instance, display higher polarity by virtue of their polar heads (Diaz et al., 2008). This in turn exerts a predictable influence on the spectral shape of NR fluorescence, which emits at shorter wavelengths in the presence of cholesterol and shows a redder emission in phospholipid‐rich environments (Teo et al., 2021). Therefore, NR also has the potential to report myelin lipid biochemical changes. However, because of its very high dielectric constant (≈78; Fernández et al., 1995), water contained in the nanometer‐wide spaces of intra‐ and extracellular myelin spirals exerts a strong additional effect on the NR spectrum. This is shown in Figures 5 and 6, where strong dehydration of CNS and PNS axons greatly reduced the polarity of the tissue. Nonetheless, even in the dehydrated state, we observed a significant difference in myelin polarity between central optic nerve axons and peripheral dorsal roots, likely indicative of differences in lipid composition between CNS and PNS myelin consistent with what has been reported: PNS myelin contains more relatively non‐polar sphingomyelin to CNS, whereas CNS myelin has comparatively more polar sulfatide‐containing lipids (Greenfield et al., 1973; Horrocks, 1967; Norton & Poduslo, 1973; Voet et al., 2016). One limitation of the NR spectroscopic technique is the inability to distinguish between polarity effects of lipids versus water contained within the myelin. However, comparing hydrated versus dehydrated samples can give some indication of the contribution of each to overall polarity of the specimen, with the ability to interrogate myelin lipids (and proteins) in the dehydrated state.

The above limitation notwithstanding, our method revealed subtle and unexpected differences between CNS and PNS (Figure 1) and also emphasized regional, age‐related, developmental, and sex differences (Figures 1, 2, 3). Being a relatively non‐toxic dye, NR was also well suited for studying live preparations. Live ex vivo central and peripheral axons are easily and rapidly stained by exposure to NR‐containing perfusate, yielding bright high‐resolution images of myelin that can be followed over time (Figure 4). While morphological integrity appeared well preserved over 4 h of ex vivo perfusion and imaging at least at the light level, NR spectroscopy indicated a gradual reduction of myelin polarity with time. Fixation with PFA induced the largest reduction; since changes in lipid composition would not be expected to occur with chemical fixation, this is most likely because of shrinkage of tissue induced by cross‐linking fixatives, which are known to narrow intra‐myelinic spaces and would therefore reduce the water content (Kirschner & Hollingshead, 1980; Korogod et al., 2015; Thavarajah et al., 2012), in turn inducing a drop in measured polarity.

Myelination of axons was a remarkable evolutionary advance, allowing fibers to occupy far less space, while resulting in electrical characteristics that allow efficient saltatory action potential propagation (Kister & Kister, 2022). In addition, oligodendrocytes in the CNS provide substrate for axonal mitochondria for energy production, and this process requires an intact myelin sheath (Fünfschilling et al., 2012; Lee et al., 2012). A similar requirement for a preserved relationship between axon, myelin, and Schwann cell likely also exists in the PNS to ensure normal functioning of peripheral axons (Fledrich et al., 2019). Originally thought to be a rather static element after the developmental period, recent studies have shown that myelin continues to be dynamically modulated well into adulthood through a process termed activity‐dependent or adaptive myelination (Bonetto et al., 2021; Gibson et al., 2014; Mount & Monje, 2017). Thus, a broad array of activities such as motor learning, sensory stimulation, formation of memories and even social activities, or direct electrical stimulation of myelinated tracts, can have a significant influence on myelination (reviewed in Knowles et al., 2022). These fundamental insights that brought to light the novel concept of “myelin plasticity” (Bonetto et al., 2020), rely largely on detailed morphological assessment of fine myelin structure. However, it is likely that even more subtle modulation of the physicochemical properties of this element, for instance, by changing its lipid composition and/or water content, could in turn alter its electrical characteristics and fine‐tune conduction velocities. It has been proposed that even small changes in conduction can alter synchronization across brain regions with profound effects on cognition and behavior (Pajevic et al., 2014). The proof‐of‐principle experiment in Figure 7 shows how the NR spectroscopic method can report very subtle changes in optic nerve polarity which increases after a 2 week period of exposure to darkness. This would result in an increase in myelin capacitance, in turn causing a slowing of conduction velocity (Teo et al., 2021), consistent with previous reports that also showed slowed conduction in optic nerves from eyes deprived of light input (Etxeberria et al., 2016). The published experiments were performed in young mice where developmental myelination was still in progress and would likely be more subject to environmental influence. In our case, we showed that even well into adulthood (13 weeks of age), when myelination is complete, changes in sensory input, and therefore action potential traffic, can cause significant polarity changes. This raises the important concept of myelin plasticity that can be induced not only by altering its structure, for example, thickness, internodal length, changes in nodal gap length (Bertram & Schröder, 1993; Ford et al., 2015; Johnson et al., 2015; Nabel et al., 2024; Yang et al., 2020) but also by more subtle changes in myelin polarity—and therefore capacitance in turn altering conduction velocity—that can be modulated by water content and the proportions of polar versus less polar lipids such as cholesterol.

The comparatively long dimensions of the myelin sheath, the long distances of the axon cylinder from its neuronal soma, together with the axo‐myelinic architecture required to support survival and physiological operation, make the myelinated axon vulnerable to numerous inherited and acquired disorders that often disrupt the normal structure and function of the sheath. Prominent examples include ischemia, trauma to the brain or spinal cord, multiple sclerosis, numerous peripheral neuropathies, and a number of genetically determined dysmyelinating disorders affecting both central and peripheral axons (Alizadeh et al., 2015; Stavrou et al., 2021). Abnormalities of myelin may even extend to neuropsychiatric conditions, such as schizophrenia, autism, and major depression, where subtle white matter pathology has been reported (Fields, 2008; Valdés‐Tovar et al., 2022). In vivo imaging of mouse brain has shown that CNS myelin is remarkably stable morphologically (Hughes et al., 2018). However, such studies cannot report more subtle biochemical or nanostructural changes in the sheath, ones that could exert a significant influence on a myelinated axon's conduction properties, and in turn on network dynamics and brain function. For instance, using the NR spectroscopic method, we showed subtle changes in the dielectric constant of myelin in still‐myelinated areas of the multiple sclerosis brain which would result in a widespread slowing of conduction velocity (Teo et al., 2021). We proposed that such alterations could underpin the diffuse non‐focal clinical symptoms such as fatigue, cognitive impairment, and mood disorders very frequently seen in multiple sclerosis patients (Linnhoff et al., 2019). A related diffuse process may be operating to explain “brain fog” after cancer chemotherapy, or delayed cognitive impairment in long COVID syndrome, both of which have been associated with white matter changes (Fernández‐Castañeda et al., 2022; Menning et al., 2018). Indeed, abnormalities of myelin are now recognized in a wide array of neurological disorders including neurodegenerative diseases, such as Alzheimer's, epilepsy, and a variety of mood disorders (Knowles et al., 2022). Interestingly, influences originating outside the brain such as via the gut–brain axis have also been reported to affect oligodendrocytes and myelin via specific metabolites synthesized by gut microbiota (Needham et al., 2022). The method described in this report represents a powerful tool for the sensitive study of subtle myelin changes in a variety of physiological and pathological conditions, even in the absence of overt morphological myelin alterations. Nile Red spectroscopy of myelin adds unique and complementary information to other myelin imaging methods such as ScoRe (Gonsalvez et al., 2019), coherent anti‐Stokes Raman scattering (Poon et al., 2018), and birefringence microscopy (Blanke et al., 2024), which are less labor‐intensive than conventional electron microscopy.

AUTHOR CONTRIBUTIONS

W. Teo: Conceptualization; methodology; formal analysis; investigation; writing – review and editing; validation. M. L. Morgan: Investigation; methodology; writing – review and editing; validation; formal analysis. P. K. Stys: Conceptualization; funding acquisition; writing – original draft; methodology; validation; writing – review and editing; software; project administration; data curation; supervision; resources.

FUNDING INFORMATION

This work was funded by the Multiple Sclerosis Society of Canada (Grant/Award No. MSSC‐3563), the Canada Foundation for Innovation, and the Canadian Institutes of Health Research (Grant/Award No. 148833).

CONFLICT OF INTEREST STATEMENT

The authors have no conflict of interest to declare.

Supporting information

Data S1.

Teo, W. , Morgan, M. L. , & Stys, P. K. (2025). Quantitation of the physicochemical properties of myelin using Nile Red fluorescence spectroscopy. Journal of Neurochemistry, 169, e16203. 10.1111/jnc.16203

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.