Co-Ultramicronized Palmitoylethanolamide/Luteolin Restores Oligodendrocyte Homeostasis via Peroxisome Proliferator-Activated Receptor-α in an In Vitro Model of Alzheimer’s Disease

Oligodendrocytes are cells fundamental for brain functions as they form the myelin sheath and feed axons. They perform these critical functions thanks to the cooperation with other glial cells, mainly astrocytes. The astrocyte/oligodendrocyte crosstalk needs numerous mediators and receptors, such as peroxisome proliferator-activated receptors (PPARs). PPAR agonists promote oligodendrocyte precursor cells (OPCs) maturation in myelinating oligodendrocytes. In the Alzheimer’s disease brain, deposition of beta-amyloid (Aβ) has been linked to several alterations, including astrogliosis and changes in OPCs maturation. However, very little is known about the molecular mechanisms. Here, we investigated for the first time the maturation of OPCs co-cultured with astrocytes in an in vitro model of Aβ1–42 toxicity. We also tested the potential beneficial effect of the anti-inflammatory and neuroprotective composite palmitoylethanolamide and luteolin (co-ultra PEALut), which is known to engage the isoform alfa of the PPARs. Our results show that Aβ1–42 triggers astrocyte reactivity and inflammation and reduces the levels of growth factors important for OPCs maturation. Oligodendrocytes indeed show low cell surface area and few arborizations. Co-ultra PEALut counteracts the Aβ1–42-induced inflammation and astrocyte reactivity preserving the morphology of co-cultured oligodendrocytes through a mechanism that in some cases involves PPAR-α. This is the first evidence of the negative effects exerted by Aβ1–42 on astrocyte/oligodendrocyte crosstalk and discloses a never-explored co-ultra PEALut ability in restoring oligodendrocyte homeostasis.


Introduction
Aberrant myelination of neuronal axons is a common hallmark of several neurodegenerative diseases [1]. Physiologically, myelination consists in wrapping axons in a sheath of lipid called myelin, which is able to increase the rate of action potentials along the axon [2]. Thus, any impairment in myelin formation may alter neurotransmission and lead to reduced cognitive functions and dementia [3].
Myelination is performed by oligodendrocytes, brain cells belonging to the macroglia family of glial cells [3]. They originate from oligodendrocyte precursor cells (OPCs), mainly localized in the ventricular zones of the central nervous system (CNS). From there, they Otherwise specified, all reagents and plasticware used in the above-described p col were purchased from Sigma Aldrich, Saint Louis, MO, USA.

Drugs and Schedule of Treatments
Human Aβ1-42 was purchased from AnaSpec (Fremont, CA, USA) and dissolve sterile NaOH 10 mM at a concentration of 221.5 μM following the manufactu Primary astrocytes were treated with Aβ 1-42 1 µM in the presence or absence of the following substances: co-ultra PEALut (3 µm) and GW6471 (3 µM). Concentrations and schedule of treatments were chosen based on the available literature [23,32,40,50,51] and according to the results of neutral red uptake assays. All reagents were lipopolysaccharide-free.
After 48 h of treatment, astrocytes and oligodendrocytes were collected separately and processed for analyses.

Neutral Red Uptake Assay
Respective astrocyte and oligodendrocyte viability was tested by neutral red uptake assay, as previously described [42]. Cells were bathed for 3 h in 50 µg/mL neutral red solution (Sigma-Aldrich) at +37 • C. Later, cells were rinsed first in Ca 2+ -and Mg 2+ -free PBS and then in a de-staining solution containing ethanol, deionized water, and glacial acetic acid (50:49:1 v/v). Cell absorbance (540 nm) was read by a microplate spectrophotometer (Epoch, BioTek, Winooski, VT, USA). The proportional number of viable cells was calculated and expressed as percentage to control sample (CTRL).

Protein Extraction and Western Blot Analysis
Western blot analysis was performed on protein lysates of primary astrocytes, following our published protocol [42]. Cells were lysed in ice-cold hypotonic buffer containing 150 mM NaCl, 1 mM EDTA, 50 mM Tris/HCl pH 7.5, 1% triton X-100 complete of inhibitors of protease (0.1 mM leupeptin, 10 µg/mL aprotinin, and 1 mM phenylmethylsulfonyl fluoride, all from Sigma-Aldrich). Homogenates were incubated for 40 min at +4 • C, then centrifuged at 18,440× g for 30 min. Supernatants containing proteins were stored at −80 • C until use. An equivalent amount of each sample (50 µg), calculated by Bradford assay, was resolved through 12% acrylamide SDS-PAGE Stain-Free precast gels (Bio-Rad Laboratories, Segrate, Italy). All samples were run concurrently. Proteins were transferred from gels to nitrocellulose papers using a trans-blot SD semidry transfer cell (Bio-Rad Laboratories). After 1 h incubation in a blocking solution containing 5% no-fat dry milk and 0.1% Tween 20 in TBS (TBS-T) at room temperature (Tecnochimica Moderna, Rome, Italy), membranes were left overnight at +4 • C in a blocking solution containing one of the following primary antibodies: rabbit anti-glial fibrillary acidic protein (GFAP) 1:25,000 (Abcam, Cambridge, UK), mouse-anti-glutamine synthetase (GS) 1:500 (Millipore, Burlington, MA, USA), and mouse anti-high mobility group box 1 (HMGB1) 1:1000 (BioLegend, San Diego, CA, USA). The following morning, membranes were rinsed using 0.05% TBS-T and incubated with a blocking solution containing the appropriate secondary horseradish peroxidase (HRP)conjugated antibody for 1 h at room temperature. Lastly, membranes were washed with T-TBS and briefly bathed in a specific solution of the enhanced chemiluminescence (ECL) kit (GE Healthcare Life Sciences, Milan, Italy). Immunocomplexes were detected by a Chemidoc XRS and analyzed by Image Lab software (Bio-Rad, Hercules, CA, USA). Values were normalized to the total protein content. Experimental conditions are summarized in Table 1.

Real Time-Quantitative Polymerase Chain Reaction (RT-qPCR)
RT-qPCR was carried out as previously described [52] and summed up in Table 2. The TRI-Reagent (Sigma-Aldrich) was added to primary astrocyte samples to extract total mRNA. An equal amount of mRNA (1 µg) from each experimental group was quantified by a D30 BioPhotometer spectrophotometer (Eppendorf AG, Hamburg, Germany) and reverse transcribed using a QuantiTect Reverse Transcription Kit (Qiagen, Valencia, CA, USA). RT-PCR reactions (20 µL) were set mixing specific primers, cDNA (20 ng) with the iTaq Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) and run in 96-well plates using the CFX96 Touch thermocycler (Bio-Rad, Hercules, CA, USA). All samples were run concurrently in triplicate. A two-step thermal protocol was used for 40 cycles (10 s at 95 • C and then 30 s at 60 • C) preceded by a polymerase activation step (3 min at 95 • C). Gene-specific amplification was controlled by including a melting curve analysis of the amplification products at the end of the reaction. The Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) was used as reference gene to normalize each target amplicon. Data were analyzed using the Pfaffl method, correcting the delta Ct values for each primer efficiency [53]. All primer sequences are listed in Table 2.

Immunofluorescence
Immunofluorescence was performed according to our previous studies [54]. Briefly, semi-porous membranes containing primary oligodendrocytes were carefully cut out and laid down on glass slides (Thermo Scientific, Rockford, IL, USA). Cells were fixed in 4% paraformaldehyde (Santa Cruz, Dallas, TX, USA) in PBS 1X for 20 min at room temperature. After being washed three times with PBS 1X, they underwent a 10 min bath in PBS 1X with 0.25% Triton X-100 (Sigma-Aldrich) to promote cell permeabilization. Then, cells were submerged for 2 h in a blocking solution containing 5% BSA and 0.25% Triton X-100 in PBS 1X and subsequently incubated overnight at 4 • C with the blocking solution added of the following primary antibodies: rabbit anti-GFAP (1:1500, Abcam), rabbit anti-Olig2 Nuclei were stained with Hoechst (1:500, Thermo Fisher Scientific). Lastly, cells were rinsed 3 times in PBS 1X before being covered with Fluoromount aqueous mounting medium (Sigma-Aldrich) and a glass coverslip. Experimental conditions are summarized in Table 3.
The specificity of the signal was confirmed in cells that underwent the same protocol procedure except for the incubation with the primary antibody. Data were collected using an Eclipse E600 microscope equipped with Nikon Plan 20× and 40× objectives (Nikon instruments, Rome, Italy) connected to a Qimaging camera (Surrey, BC, Canada), and controlled by the software NIS-Elements-BR 3.2 64 bit (Nikon instruments). Parameters of gain and exposure were kept constant to allow comparisons between samples within the same experiment.

Image Analysis
In order to calculate the percentage of MBP + /Olig2 + oligodendrocytes, cells were counted manually from six random fields per condition for each independent experiment using the Fiji software by two investigators blinded to treatments.
The morphology of MBP + oligodendrocytes was analyzed by measuring the surface, expressed in µm 2 , of each individual cell [55].
The complexity of MBP + oligodendrocytes was analyzed by using the Sholl analysis plugin [56,57]. Fluorescent pictures were converted into binary 8-bit images, background subtracted, and the free edges were detected. Starting radius was set at 10 µm and step size at 5 µm, thus creating 5 concentric radii up to 30 µm distance from the soma. The number of cell processes intersecting those circles was plotted as a function of the radial distance from the cell soma. Several descriptors of cell arborization were analyzed, which are the maximum intersections (calculated as the highest number of processes/branches in each cell arbor), the sum of intersections (calculated as the sum of intersections detected at each concentric circle), and the mean of intersections (calculated as the sum of intersections divided by intersecting radii) [23,58,59]. Images were analyzed by an investigator blinded to treatments.

Statistical Analysis
Statistical analysis was performed using GraphPad Prism software version 6.0 (Graph-Pad Software, San Diego, CA, USA). Data were analyzed by either the Student's t-test, in case just two experimental groups were compared, or one-way analysis of variance (ANOVA) in all other experiments. Upon detection of a main significant effect from ANOVA, a post hoc test was carried out as either the Dunnett's test, in case groups needed to be compared only to the CTRL, or the Student Newman Keuls (SNK) test in case of multiple comparisons among groups. Differences between mean values were considered statistically significant when p < 0.05.

Characterization of OPCs Maturation In Vitro and Effects of Treatments on the Viability of OPCs and Astrocytes
To monitor the maturation of OPCs in our experimental conditions, we performed a longitudinal immunofluorescence study testing the appearance of the MBP signal. We studied the MBP + /Olig2 + cell ratio, which is considered an index of the level of maturation of OPCs [60]. Indeed, MBP is the major protein constituent of myelin, and it is produced only by mature oligodendrocytes, whereas Olig2 is a specific transcriptional factor expressed by oligodendrocytes at all developmental stages, from precursors to mature cells [60][61][62]. According to literature data, our results show that primary oligodendrocytes cultured alone survive approximately 10 days and start to express MBP between the second and third day in vitro (DIV). MBP expression progressively increases to reach the highest value at about the seventh DIV (Figure 1b). Provided these data, we decided to study the effects of treatments on the initial phase of OPCs maturation, that is, the third DIV. Figure 1c-e shows the results of viability assays performed to choose the appropriate concentrations of treatments to be applied for 48 h to astrocytes. In Figure 1, we also report the effects of the selected concentrations and of the relative vehicles used to solubilize each substance on astrocyte (f) and oligodendrocyte (g) viability. No significant variation versus control was observed when co-ultra PEALut or GW6471 were provided alone.

Aβ 1-42 Triggers Astrocyte Reactivity
To study the effects of human fibrillary Aβ 1-42 1 µm on astrocyte function in our transwell culture system, we performed immunofluorescence and western blot experiments analyzing the expression levels of key proteins that are commonly related to a reactive and proinflammatory phenotype [63,64]. Our results show that 48 h treatment with Aβ 1-42 triggers this astrocytic phenotype as documented by the increased expression of GFAP, GS, and HMGB1 ( Figure 2).

Co-Ultra PEALut Prevents the Astrocyte Reactivity and the Reduction in Growth Factors Transcription Induced by Aβ1-42 Exposure
Experiments investigated whether co-ultra PEALut exerts beneficial effects in primary astrocytes exposed to Aβ1-42, focusing on astrocytic factors able to affect OPCs survival and development. The possible engagement of the PPAR-α in such co-ultra PEALut effects was also explored.

Co-Ultra PEALut Prevents the Astrocyte Reactivity and the Reduction in Growth Factors Transcription Induced by Aβ 1-42 Exposure
Experiments investigated whether co-ultra PEALut exerts beneficial effects in primary astrocytes exposed to Aβ 1-42 , focusing on astrocytic factors able to affect OPCs survival and development. The possible engagement of the PPAR-α in such co-ultra PEALut effects was also explored.

Co-Ultra PEALut Controls the Oligodendrocyte Morphological Changes Induced by Aβ 1-42 Exposure
Oligodendrocytes change their cell morphology in various CNS pathological conditions, including AD [65][66][67][68][69][70]. As we observed that Aβ 1-42 increased MBP + cells, we decided to clarify the Aβ 1-42 effects by performing a thorough morphological analysis of oligodendrocytes. We analyzed the cell surface area and the number of arborizations of MBP + oligodendrocytes. As shown in Figure 5, we observed a reduction in the cell surface area of oligodendrocytes when these cells were co-cultured with astrocytes challenged with Aβ 1-42 but not with the vehicle. Co-ultra PEALut significantly prevented such Aβ 1-42 -induced effect. Concurrent treatment with the PPAR-α antagonist GW6471 did not block the effect of co-ultra PEALut, suggesting the involvement of other mechanisms.
The Sholl analysis revealed a significant reduction both in the maximum number of intersections and the sum of intersections of MBP + cell branches when oligodendrocytes were indirectly exposed to Aβ 1-42 (Figure 6a-d). Similarly, the mean number of intersections was reduced (Figure 6e). Co-ultra PEALut prevented such Aβ 1-42 -induced modifications in oligodendrocyte morphology through PPAR-α involvement. Indeed, GW6471 blocked the observed co-ultra PEALut effect (Figure 6a,c-e). By counting the number of arborizations from the MBP + cell soma outwards every 5 µm (Figure 6f), we found that Aβ 1-42 -exposed oligodendrocytes showed a lower number of arborizations at three distance measures (10, 15, and 20 µm) compared to control cells, reaching the greatest and most significant reduction at 20 µm from the soma (Figure 6f). Co-ultra PEALut treatment was effective in preventing these Aβ 1-42 -induced morphological changes. The co-treatment with GW6471 significantly blunted such co-ultra PEALut effect, indicating an engagement of the PPAR-α in the co-ultra PEALut mechanism of action ( Figure 6f). not with the vehicle (Figure 4a,b). Olig2 + cell count was not affected by treatment, suggesting that no proliferation or death of oligodendrocytes occurred (Figure 4a,c). Coherently, the MBP + /Olig2 + ratio increased in oligodendrocytes indirectly exposed to Aβ1-42 ( Figure 4d). Co-ultra PEALut significantly prevented such an Aβ1-42-induced effect ( Figure  4a-d). Concurrent treatment with GW6471 did not reverse the co-ultra PEALut effect indicating that other mechanisms of action could be involved.  to clarify the Aβ1-42 effects by performing a thorough morphological analysis of oligodendrocytes. We analyzed the cell surface area and the number of arborizations of MBP + oligodendrocytes. As shown in Figure 5, we observed a reduction in the cell surface area of oligodendrocytes when these cells were co-cultured with astrocytes challenged with Aβ1- 42 but not with the vehicle. Co-ultra PEALut significantly prevented such Aβ1-42-induced effect. Concurrent treatment with the PPAR-α antagonist GW6471 did not block the effect of co-ultra PEALut, suggesting the involvement of other mechanisms. The Sholl analysis revealed a significant reduction both in the maximum number of intersections and the sum of intersections of MBP + cell branches when oligodendrocytes were indirectly exposed to Aβ1-42 (Figure 6a-d). Similarly, the mean number of intersections was reduced (Figure 6e). Co-ultra PEALut prevented such Aβ1-42-induced modifications in oligodendrocyte morphology through PPAR-α involvement. Indeed, GW6471 blocked the observed co-ultra PEALut effect (Figure 6a,c-e). By counting the number of arborizations from the MBP + cell soma outwards every 5 μm (Figure 6f), we found that Aβ1-42-exposed oligodendrocytes showed a lower number of arborizations at three distance measures (10, 15, and 20 μm) compared to control cells, reaching the greatest and most significant reduction at 20 μm from the soma (Figure 6f). Co-ultra PEALut treatment was effective in preventing these Aβ1-42 -induced morphological changes. The co-treatment with GW6471 significantly blunted such co-ultra PEALut effect, indicating an engagement of the PPAR-α in the co-ultra PEALut mechanism of action ( Figure 6f).

Discussion
Neuroglia is an extremely heterogeneous population of cells present in the nervous system. The heterogeneity of glial cells correlates with the multiplicity of functions that they perform. Regardless of their morphology and functions, the common fundamental role of all types of glial cells in the maintenance of CNS homeostasis [71]. Among glial

Discussion
Neuroglia is an extremely heterogeneous population of cells present in the nervous system. The heterogeneity of glial cells correlates with the multiplicity of functions that they perform. Regardless of their morphology and functions, the common fundamental role of all types of glial cells in the maintenance of CNS homeostasis [71]. Among glial cells, astrocytes have recently gained great attention for their homeostatic support that takes place at all levels of brain organization. Therefore, alterations affecting astrocytes are nowadays considered to be implicated in the etiology or progression of almost all disorders affecting the nervous system, including AD [72][73][74][75]. For this reason, neuropharmacologists look at these cells as possible targets for the development of new therapeutics [76,77].
One of the less explored functions of astrocytes is their supportive role in the maturation of oligodendrocytes. Astrocytes establish physical and functional connections with oligodendrocytes, thus allowing, among many other processes, the proper formation of myelin [10]. Changes in this cell-cell communication may alter myelination, impairing neurotransmission, ultimately leading to cognitive decline and dementia [3]. Indeed, aberrant myelination is considered one of the key features of AD [17], although it remains an overlooked field of study. To contribute to filling this gap, here we set up an in vitro model of Aβ 1-42 toxicity to study the interaction between astrocytes and oligodendrocytes, with particular emphasis on astrocyte reactivity and the release of trophic factors required for OPCs maturation, which is predictive of their ability to form myelin. By using semi-porous inserts to grow the two populations of glial cells separately, we treated astrocytes with human Aβ 1-42 and studied the resulting changes in the two cell populations, with special reference to OPCs maturation. We observed that 48 h exposure to human fibrillary Aβ 1-42 induced astrocytes to acquire a reactive and proinflammatory phenotype. We indeed observed an increase in the expression of key proteins related to astrocyte reactivity, such as GFAP, GS, and HMGB1. Similarly, elevated transcriptional levels of p50NFkB, IL-6, and IL-1β, well-known proinflammatory mediators, have been observed. Intriguingly, the elevation of such meditators has already been linked to a toxic environment that damages all cells, including oligodendrocytes [78][79][80]. Furthermore, our data reveal that Aβ 1-42 -exposure causes a significant reduction in the astrocytic transcriptional levels of FGF2 and TGF-β, two factors strictly implicated in OPCs maturation [11,12]. Based on these results and literature evidence indicating Aβ-induced altered myelination [81,82], we predicted to observe a defective maturation of co-cultured OPCs in our experimental condition. Our immunofluorescence experiments instead revealed a higher number of MBP + cells and MBP + /Olig2 + ratio in oligodendrocytes indirectly exposed to human Aβ 1-42 compared with vehicle. In agreement with these counterintuitive results, direct treatment of oligodendrocytes with oligomeric Aβ 1-42 promoted MBP upregulation in primary rat oligodendrocyte cultures [23]. Furthermore, a marked increase in oligodendrocyte differentiation was found in the corpus callosum and hippocampus of transgenic models of AD, namely 3xTg-AD and APP/PS1 mice, compared with the wild-type counterparts [83]. As suggested by some researchers, a possible explanation of the Aβ effect on oligodendrocyte MBP expression could be linked to the reaction of these cells to the toxic stimulus [5]. Oligodendrocytes, indeed, may react by increasing their rate of differentiation in the attempt to restore the lost homeostasis [84,85]. By performing an in-depth morphological analysis, here we unveil that this push to the maturation operated by the toxic insult with Aβ 1-42 actually causes severe morphological changes to oligodendrocytes. We observed that MBP + oligodendrocytes lose their proper dimension and complexity as a consequence of the toxic insult. A significant reduction in the cell surface area, the maximum number of intersections, and both the sum and mean of intersections of MBP + cell branches were detected in oligodendrocytes co-cultured with astrocytes challenged with Aβ 1-42 . All these structural changes highlight oligodendrocyte loss of planarity and their confused organization of branches. Data obtained agree with some recent in vivo findings in 3xTg-AD mice. In particular, oligodendrocytes of six-month-old transgenic mice appeared atrophic when compared to non-transgenic age-matched animals [55].
The new and extremely interesting evidence that emerges from the present results concerns the ability of co-ultra PEALut to restore oligodendrocyte homeostasis impaired by the Aβ 1-42 challenge. Our and other groups have extensively documented the neuroprotective and beneficial effects of PEA in several preclinical and clinical settings [41,[86][87][88]. We also demonstrated the ability of both PEA and co-ultra PEALut to modulate glial reactivity and dampen neuroinflammation, thus fostering neuron survival in different AD models [32,40,[42][43][44].
Here, we provide the very first evidence about co-ultra PEALut's ability to restore the pathological changes occurring in both astrocytes and oligodendrocytes after a toxic insult with the human Aβ 1-42 . We also show that co-ultra PEALut exerts some of these effects with a mechanism mediated by the PPAR-α. The activation of PPAR-α has been linked to NFκB inhibition, reduction of proinflammatory mediators, increased transcription of antioxidant enzymes, and promotion of cell survival through the inhibition of caspase 3 [89]. All these pathways could theoretically explain our results; however, further experiments are needed to clarify the effects of co-ultra PEALut.
By increasing FGF2 and TGF-β mRNA levels and blunting the expression of proinflammatory mediators, co-ultra PEALut normalized astrocyte functions and preserved the morphological features of oligodendrocytes indirectly exposed to Aβ 1-42 . It is worth noting that co-ultra PEALut provided concurrently with Aβ 1-42 , apart from preventing Aβ 1-42 -induced branching loss throughout the entire cell, even increases the number of intersections at 20 µm from the soma compared with the control group. Interestingly, the total area occupied by oligodendrocytes and the number of cellular arborizations are associated with the ability of these cells to create the myelin sheath around neuronal axons [90,91]. We are aware that further studies are required to better clarify the molecular mechanisms. However, our data are promising as they suggest the potential use of co-ultra PEALut in neurological disorders characterized by impaired myelination. In agreement, recent studies showed the co-ultra PEALut's ability to promote in vitro OPCs differentiation [37,38,46] and to reduce the development of clinical signs in an in vivo experimental model of MS [39].
This study has some limits: first, our in vitro system is not able to fully reproduce the physiological environment where myelination occurs, which also comprehends neurons. Secondly, the mechanism of action of co-ultra PEALut could be further explored since not all the observed effects here are mediated by the PPAR-α. Indeed, other receptors have been proposed as PEA targets, including the transient receptor potential vanilloid type 1 channel and the orphan G-protein coupled receptor 55; PEA could engage the endocannabinoid enzyme fatty acid amide hydrolase acting as a false substrate and, in turn, indirectly increasing anandamide concentration [77,92]. Finally, further studies are needed to characterize additional markers targeting myelin and to explore OPCs differentiation and maturation longitudinally.
In conclusion, our data show a novel and never-explored ability of co-ultra PEALut in counteracting the negative effects exerted by Aβ on oligodendrocyte homeostasis. Altogether, our findings open new opportunities for the adjuvant treatment of AD, and co-ultra PEALut is a promising composite capable of exerting beneficial effects in both experimental and clinical settings. Since it is already licensed for human use, the potential to proceed to clinical use rapidly is reasonable.

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