Oligomeric amyloid beta prevents myelination in a clusterin-dependent manner

Background: White matter loss is a well-documented phenomenon in Alzheimer’s disease (AD) patients that has been recognized for decades. However, the underlying reasons for the failure of oligodendrocyte progenitor cells (OPCs) to repair myelin deficits in these patients remain elusive. A single nucleotide polymorphism (SNP) in Clusterin has been identified as a risk factor for late-onset Alzheimer’s disease and linked to a decrease in white matter integrity in healthy adults, but its specific role in oligodendrocyte function and myelin maintenance in Alzheimer’s disease pathology remains unclear. Methods: To investigate the impact of Clusterin on OPCs in the context of Alzheimer’s disease, we employed a combination of immunofluorescence and transmission electron microscopy techniques, primary culture of OPCs, and an animal model of Alzheimer’s disease. Results: Our findings demonstrate that Clusterin, a risk factor for late-onset AD, is produced by OPCs and inhibits their differentiation into oligodendrocytes. Specifically, we observed upregulation of Clusterin in OPCs in the 5xFAD mouse model of AD. We also found that the phagocytosis of debris, including amyloid beta (Aβ), myelin, and apoptotic cells leads to the upregulation of Clusterin in OPCs. In vivo experiments confirmed that Aβ oligomers stimulate Clusterin upregulation and that OPCs are capable of phagocytosing Aβ. Furthermore, we discovered that Clusterin significantly inhibits OPC differentiation and hinders the production of myelin proteins. Finally, we demonstrate that Clusterin inhibits OPC differentiation by reducing the production of IL-9 by OPCs. Conclusion: Our data suggest that Clusterin may play a key role in the impaired myelin repair observed in AD and could serve as a promising therapeutic target for addressing AD-associated cognitive decline.


Background
Alzheimer's Disease (AD) is a neurodegenerative disease with limited therapeutic options to effectively prevent disease progression, and patients inevitably succumb to debilitating dementia [1].White matter loss has been documented in AD patients, but how this myelin loss contributes to disease progression remains unclear [2].Recent studies indicate that generation of new myelin-forming oligodendrocytes is critical for memory consolidation and recall [3,4].Additionally, remyelination therapeutics have been Institutional Review Board at the University of Florida.Clinical details of all human specimens analyzed are details in Supplemental Table 1.

OPC Proliferation
OPCs were plated in proliferation media (described above) supplemented with 8µg/mL clusterin (Sino Biological 50485-M08H) or an equivalent volume of vehicle (H 2 0 or PBS) for the control samples.OPCs were allowed to proliferate for 40-72 hours.Cell number was assessed using the Cell Counting Kit-8 (Dojindo, CK04) according to manufacturer's instructions.Optical density (OD) was measured at 450nm.For data analysis, the OD of a media only control was subtracted from the OD of all experimental samples.The OD of wells treated with clusterin were then normalized to the biologically identical control well.

OPC Differentiation
OPCs were plated in proliferation media (described above) or differentiation media consisting of DMEM/F12 (Gibco, 11320082), B27 (Gibco, 17504044), Pen-Strep (Gibco, 15140122), 10ng/mL FGF (Peprotech, 450 − 33), 10ng/mL CNTF (Peprotech, 450 − 13), and 40ng/mL T3 (Sigma, T6397).For clusterin conditions, 8µg/mL clusterin (Sino Biological 50485-M08H) was added to the differentiation media and an equivalent volume of vehicle (H 2 0 or PBS) was added to the differentiation and proliferation control samples.For IL-9 rescue experiments, 100ng/ml IL-9 (Peprotech 219 − 19) was added to the differentiation media and an equivalent volume of vehicle (0.1% BSA) was added to the differentiation and proliferation control samples.For VEGF inhibition experiments, 10µg/ml of a VEGF function-blocking antibody (R&D Systems MAB9947-SP) was added to differentiation media and 10µg/ml of control IgG (Biolegend 403501) was added to the differentiation and proliferation control samples.OPCs were allowed to differentiate for 48-72 hours and were then subsequently processed for RNA extraction and qPCR or for immuno uorescence.

Aβ Preparation and in vitro treatment
Aβ oligomers were prepared as previously described [18].Brie y, human Amyloid Beta 1 − 42 (Echelon Biosciences, 641 − 15) was dissolved in HFIP (Sigma, 52517) to make a 1mM solution and was allowed to desiccate overnight.The resulting peptide lm was diluted to a 5mM solution in DMSO and subsequently diluted to a 100µM solution in phenol-free F-12 cell culture media (Gibco, 11039-021) and allowed to incubate overnight at 4°C.For the analysis of clusterin expression following Aβ-treatment, OPCs were treated with 3µM Aβ or a vehicle control for 4 hours.For experiments that included samples treated with CytoD, cells were pretreated for 30 minutes with 1µM CytoD (Millipore-Sigma, C8273) or DMSO and subsequently treated with 3µM Aβ or vehicle control (also containing CytoD or DMSO) for 4 hours.For CypHer-labeled Aβ experiments, 400µM Aβ oligomers in DMSO and phenol-free F-12 cell culture media was incubated with an equivalent volume of 0.1M sodium bicarbonate (Fisher Scienti c, S233-500) and 400µM CypHer5e (Cytiva, PA15401) for 30 minutes at room temperature.Following incubation, the solution was spun through a buffer exchange column (Thermo Scienti c, 89882) to remove any excess dye.

Myelin Preparation and in vitro treatment
Myelin was prepared from mouse brains as previously described [19].Puri ed myelin was passed through an insulin syringe prior to use to ensure cells were treated with a homogenous solution.Cells were treated with 100µg/ml myelin or vehicle control for 4 hours.For two replicates that included samples treated with CytoD, cells were pretreated for 30 minutes with 1µM CytoD (Millipore-Sigma, C8273) or DMSO and subsequently treated with 100µg/ml myelin or vehicle control (also containing CytoD or DMSO) for 4 hours.For the remaining two replicates that included samples treated with CytoD, no pretreatment was performed.Pretreatment with CytoD did not alter clusterin expression when compared to no pretreatment with CytoD, so all experiments were combined and plotted in Fig. 2F.Cells treated with myelin were washed once with PBS prior to RNA preparation.
Cytokine and H 2 0 2 in vitro treatment OPCs were treated with 10ng/ml TNFα, 10ng/ml IFNγ, 10µM H 2 0 2 or the relevant control for 3 hours and then processed for qPCR.

ELISA
OPCs were grown as described above.A 10cm dish of approximately 2 million OPCs was treated with 3µM Aβ or an equivalent volume of vehicle for 72 hours.Cells were lysed for 15 minutes in RIPA buffer (PBS, 1% Triton X-100, 0.5% deoxycholic acid, 1% sodium dodecyl sulfate) supplemented with 1x protease inhibitor (MedChem Express HY-K0010).The insoluble material was removed by spinning the lysate at 15,000g for 15 minutes.The resulting lysate was used to quantify the amount of clusterin present in each sample using the Mouse Clusterin ELISA Kit (Thermo Fisher EM18RB) according to manufacturer's instructions.

Luminex Assay
OPCs were grown as described above.Cells were treated with 8µg/ml clusterin for 24 hours (1 replicate) or 72 hours (1 replicate).Media was then collected and spun to remove all cellular debris.Media was concentrated using a 3KD cutoff concentrator column (Thermo Fisher 88526).The resulting concentrate was analyzed using the Milliplex 32-plex Mouse Cytokine/Chemokine Panel (Millipore Sigma MCYTMAG-70K-Px32) according to manufacturer's instructions.

Apoptotic cell preparation and in vitro treatment
To create apoptotic cells, Jurkats were treated with 150mJ of UV energy and incubated for 2-4 hours at 37°C in complete media consisting of RPMI 1640 Media (Gibco, 11875101), 10% FBS (R&D Systems, S12450H), and Pen-Strep (Gibco, 15140122).Apoptotic cells were washed and added to cultured OPCs at approximately a 1:1 ratio for 6 hours.OPCs were washed once prior to RNA isolation.

Aβ Injections
C57BL/6J mice (8-14 weeks) were injected with CypHer-Aβ (ipsilateral) and NHS-Fluorescein (contralateral, Thermo Fisher 46410) as previously described [20].Brie y, mice were anesthetized with a mixture of ketamine and xylazine and a small burr hole was drilled in the skull.1µL of 100µM Aβ was injected at a speed of 200nL/minute into the right hemisphere.1µL of 200µM NHS-Fluorescein diluted in PBS was injected into the left hemisphere at the same speed.Injections were targeted for 2mM lateral, 0mM anterior, and − 1.5mM deep relative to bregma.Mice were given ketoprofen following surgery and were euthanized 12 hours post-injection.
Immuno uorescence and electron microscopy.
Mice were deeply anesthetized with pentobarbitol and subsequently perfused with 5 units/mL heparin in saline followed by 10% buffered formalin, each for approximately one minute.For brain tissue, brains were rapidly dissected and post-xed in 10% buffered formalin overnight at 4°C.Tissue was then transferred into 30% sucrose in PBS and allowed to sink for at least 24 hours.Brains were frozen in OCT, sectioned, and stored in PBS plus 0.02% NaAz until further staining.Tissue or cultured cells were blocked with PBS, 1% BSA, 0.5% Triton-X 100, 2% normal donkey serum, and 1:200 CD16/CD32 (14-0161-82, 1:200, eBioscience) for at least one hour at room temperature.For stains utilizing a mouse primary antibody, tissue was blocked in Mouse on Mouse Blocking Reagent (MKB-2213, Vector Laboratories) according to manufacturer's instructions for at least 1 hour at room temperature.Samples were incubated in primary antibodies overnight at 4°C with gentle agitation.Samples were then washed three times in TBS containing 0.3% Triton-X 100 and incubated in secondary antibodies overnight at 4°C with gentle agitation.Following secondary incubation, samples were stained with Hoechst (1:700, ThermoFisher Scienti c, H3570) for 10 minutes at room temperature, washed three times in TBS containing 0.3% Triton-X 100, and mounted on slides using Aqua Mount Slide Mounting Media (Lerner Laboratories).Images were collected on a Leica TCS SP8 confocal microscope and processed using Fiji.
For quantitative ultrastructural analysis, mice were deeply anesthetized and transcardially perfused with 0.9% NaCl followed by a Millonig's buffer solution (pH 7.3) containing 5% glutaraldehyde and 4% paraformaldehyde.Following the post-xation, brains were harvested, vibratome sectioned and processed for standard electron microscopic analysis as previously described [21].Brie y sagittal brain sections, containing the corpus callosum at the level of the fornix, were xed in 1% cacodylate buffered osmium tetroxide, dehydrated in graded ice-cold ethanol and embedded in PolyBed 812 embedding resin (Polysciences, Inc., Fort Washington, PA).Ultrathin (70nm) sections were stained with uranyl acetate and lead citrate and imaged using a JEOL JEM 1400Plus transmission electron microscope (JEOL, Peabody, MA) equipped with a Gatan OneView CMOS camera (Gatan Inc., Pleasanton, CA).Electron micrographs were collected at 10,000X; a minimum of 100 myelinated axons per mouse were used to calculate gratios.

Immunohistochemistry (Human)
5 µm thick tissue sections of formalin xed, para n embedded (FFPE) brain tissue specimens were rehydrated in Xylene and descending alcohol series and heat-induced epitope retrieval (HIER) was performed in a pressure cooker (Tintoretriever, Bio SB) for 15 min at high pressure in a 0.05% Tween-20 solution.Endogenous peroxidase was quenched by incubation of sections in 1.5% hydrogen peroxide/0.005%Triton-X-100 diluted in pH 7.4 sterile phosphate buffered saline (PBS) (Invitrogen) for 20 min, following multiple washes in tap water and subsequently, 0.1 M Tris, pH 7.6.Non-speci c antibody binding was minimized with sections incubated in 2% FBS/0.1 M Tris, pH 7.6.Clusterin (Proteintech) primary antibody was diluted in 2% FBS/0.1 M Tris, pH 7.6 at a dilution of 1:500.Sections were incubated with primary antibody over night at 4°C, washed one time in 0.1 M Tris, pH 7.6, followed by 2% FBS/0.1 M Tris, pH 7.6 for 5 min, incubated in goat ant-rabbit IgG HRP Conjugated secondary antibody (Millipore Sigma) for 1 hour, additionally washed one time in 0.1 M Tris, pH 7.6, followed by 2% FBS/0.1 M Tris, pH 7.6 for 5 min, and incubated in VectaStain ABC Peroxidase HRP Kit (diluted in 2% FBS/0.1 M Tris, pH 7.6 at 1:1000) for 1 hour.After a nal wash in 0.1 M Tris, pH 7.6 for 5 min, immunocomplexes were visualized using the Vector Laboratories ImmPACT DAB Peroxidase (HRP) 3,3′diaminobenzidine. Tissue sections were counterstained with hematoxylin (Sigma Aldrich, St. Louis, MO) for 2 minutes, dehydrated in ascending alcohol series and Xylene and cover slipped using Cytoseal 60 mounting medium (Thermo Scienti c).For analysis of stains, slides of frontal cortex specimens stained with Clusterin antibodies were scanned on an Aperio AT2 scanner (Leica biosystems) at 20x magni cation and digital slides analyzed using the QuPath platform (version 0.3.1,https://QuPath.github.io/,PMID: 29203879) on a Dell PC (Intel® Xeon® W-1270 CPU @ 3.40GHz/ 64 GB RAM/ 1 TB SSD Hard Drive) running Windows 10.Cortex and white matter were annotated for regional analysis.After exclusion of tissue and staining artifacts we used the 'Positive Pixel Detection' tool (Downsample factor 2, Gaussian sigma 1 µm, Hematoxylin threshold ('Negative') 1.5 OD units, DAB threshold ('Positive') 0.5 OD units) to determine the percentage of area covered by Clusterin staining.

Combined in situ Hybridization and Immunohistochemistry (Human)
For in-situ hybridization, 5µm thick para n-embedded tissue sections on slides were rehydrated in xylene and series of ethanol solutions (100%, 90%, and 70%).Following air drying for 5 min at RT, slides were incubated with RNAscope® Hydrogen peroxide for 10 min at RT, followed by 3 washing steps in distilled water.Antigen retrieval was performed in a steam bath for 15 min using RNAscope® 1x target retrieval reagent.After a rinse in distilled water and incubation in 10% ethanol for 3 min slides were air dried at 60°C.Subsequently, slides were incubated with RNAscope® Protease plus reagent for 30 min at 40°C in a HybEZ™ oven, followed by 3 washes in distilled water.Slides were then incubated with the following RNAscope® probe for 2 hours at 40°C in a HybEZ™ oven: Hs-Clusterin (cat.number 606241).Following washes with 1X Wash buffer, slides were incubated with RNAscope®AMP1 solution for 30 min at 40°C.Subsequent incubations with other RNAscope® AMP solutions, followed each by two washes with 1x RNAscope® Wash buffer were completed as follows: AMP2-15 min at 40°C; AMP3-30 min at 40°C; AMP4-15 min at 40°C; AMP5-30 min at RT and AMP6-15 min at RT.After two washes in 1x RNAscope® wash buffer slides were incubated in RNAscope® Fast RED-B and RED-A mixture (1:60 ratio) for 10 min at RT, followed by two washes in tap water.For in situ hybridization/immunohistochemistry double labeling, sections were incubated in 2% FBS/0.1 M Tris, pH 7.6 for 5 min following RNAscope® Fast RED incubation and two washes in tap water.Primary antibodies were diluted in 2% FBS/0.1 M Tris, pH 7.6 at the following dilutions: Ab5 (PMID: 16341263), 1:1000.Sections were incubated with primary antibody over night at 4°C, washed two times in 0.1 M Tris, pH 7.6 for 5 min each and incubated with biotinylated secondary antibody (Vector Laboratories; Burlingame, CA) diluted in 2% FBS/0.1 M Tris, pH 7.6 for 1 hour at room temperature.An avidin-biotin complex (ABC) system (Vectastain ABC Elite kit; Vector Laboratories, Burlingame, CA) was used to enhance detection of the immunocomplexes, which were visualized using the chromogen 3,3′diaminobenzidine (DAB kit; KPL, Gaithersburg, MD).Tissue sections were counterstained with hematoxylin (Sigma Aldrich, St. Louis, MO), air dried at 60°C for 30 min and cover slipped using EcoMount™ mounting medium (Biocare Medical).

Statistical Analysis
Statistical analysis of all data was done using Prism 9 (Graphpad software).Signi cance was set at p < 0.05.

Results
Clusterin is upregulated by OPCs in 5xFAD and AD.
Clusterin is upregulated in the brains of patients with Alzheimer's disease and is a signi cant risk factor for Late-Stage AD [22].Recent transcriptional analysis has shown that OPCs express clusterin in mice and humans but the function in this cell type is unknown [14,15].Here, we investigated the potential role of clusterin in OPCs.We con rmed that clusterin expression is increased in late-stage AD patients vs normal aging controls (Fig. 1A-B).We found that cells expressing clusterin RNA could be found directly surrounding Aβ plaques (Fig. 1C).This observation is conserved in pre-clinical models of AD, as clusterin is also found to be upregulated in multiple brain regions of the 5xFAD mouse model of AD (Fig. 1D-E).Because of the emerging role of OPCs and myelin in AD pathology, we investigated whether OPCs expressed clusterin in AD [5,23].We found that OPCs express clusterin both in the AD brain, as well as in normal aging (Fig. 1F).To further pro le the OPC expression of clusterin in AD pathology, we quanti ed the expression of clusterin in OPCs in 5xFAD mice.We found that OPCs upregulate clusterin protein in 5xFAD mice compared to wildtype mice (Fig. 1G-H), mirroring human sequencing data demonstrating that OPCs upregulate clusterin RNA production in AD patients [15].This data demonstrates that clusterin, a risk factor for late-onset AD, is upregulated in the parenchyma of AD patients and can be found in cells surrounding Aβ plaques.Furthermore, this data shows that human OPCs express clusterin in normal aging as well as AD, and that OPCs upregulate clusterin production in 5xFAD.

Phagocytosis of oligomeric Aβ and cellular debris results in an upregulation of clusterin expression by
OPCs.
Given our data showing that OPCs express the AD-risk factor clusterin, we wanted to determine if OPCs could contribute to AD pathology and interact with Aβ plaques.Using 5xFAD mice, we found that OPCs were surrounding and extending their processes into Aβ plaques (Fig. 2A).Because clusterin is known to facilitate debris clearance, we next wondered if OPCs were involved in the engulfment of Aβ in the brain [12].To test this, we injected wildtype (WT) mice with Aβ oligomers labeled with CypHer5e, a dye that only uoresces when it has entered the acidic environment of the lysosome.We found that OPCs around the injection site phagocytosed Aβ within 12 hours of injection (Fig. 2B, Supplementary Fig. 1).We also found that injecting Aβ was su cient to drive an upregulation of clusterin expression in the brain 12 h after injection (Fig. 2C-D).
We next investigated whether OPCs could contribute to this Aβ-induced upregulation of clusterin.We treated primary OPCs in vitro with Aβ oligomers for 4 hours and observed a rapid increase in clusterin transcript expression (Fig. 2E).We subsequently used an ELISA assay to con rm that OPCs also upregulate the production of clusterin protein following treatment with Aβ (Fig. 2F).We found that phagocytosis of Aβ was necessary to drive this upregulation of clusterin, as treatment with cytochalasin D (CytoD), an actin polymerization inhibitor that blocks phagocytosis, prevented Aβ from increasing clusterin production in OPCs (Fig. 2E) [24][25][26].Interestingly, we found that these changes in clusterin were not speci c to the phagocytic clearance of Aβ oligomers, but were rather generalizable to instances in which OPCs engulfed large debris.We found that treatment of OPCs with myelin debris and apoptotic cells also produced a similar upregulation of clusterin that was abrogated by exposure to CytoD (Fig. 2G-H).In order to eliminate the possibility that this clusterin upregulation was simply a response to any cellular stressor present in AD, we treated OPCs with other factors known to be upregulated in Alzheimer's disease, including TNFα, IFNγ, and reactive oxygen species (ROS), which all failed to produce any change in clusterin expression (Fig. 2I-J) [27][28][29].
Clusterin does not promote the phagocytosis of oligomeric Aβ.
In previous studies, clusterin has been shown to increase the phagocytic capacity of both professional and non-professional phagocytes alike [24,25].This led us to investigate whether clusterin altered the kinetics of phagocytosis in OPCs.The addition of exogenous clusterin did not change the ability of OPCs to engulf Aβ in vitro (Fig. 3A-B), indicating that phagocytosis regulates clusterin expression in OPCs, but clusterin does not subsequently regulate the phagocytosis of Aβ.In sum, these results indicate that OPCs can clear extracellular debris, including Aβ, and that phagocytosis of protein or cellular debris is responsible for driving clusterin expression in OPCs.

Exogenous clusterin inhibits OPC differentiation.
The formation of new myelin has been increasingly recognized as a critical component of memory function [3,4].In fact, a recent study demonstrated that drugs promoting nascent myelin formation results in improved memory performance in a model of AD, indicating the importance of understanding what factors are preventing the differentiation of OPCs into oligodendrocytes [5].A multitude of studies have demonstrated that cellular debris and protein aggregation prevent OPC differentiation, although the mechanism by which this occurs remains unclear [30,31].This data, along with our observation that debris clearance drives clusterin expression in OPCs (Fig. 2E-H), made us question whether clusterin might be inhibiting OPC differentiation.Consequently, we treated differentiating OPCs with exogenous clusterin and observed a striking decrease not only in genes encoding myelin proteins (Mbp, Plp1, Cnp), but also in Myrf, the master transcriptional regulator of the OPC differentiation program 72 hours after treatment (Fig. 4A-D).We subsequently observed fewer MBP-positive oligodendrocytes when OPCs were differentiated for 72 hours in the presence of clusterin compared to clusterin-free media (Fig. 4E-F).Importantly, this decrease in differentiation is not due to OPC death, as there was no difference in OPC number following clusterin treatment (Fig. 4G).Overall, this data shows that clusterin is regulated by the phagocytosis of debris and is a potent inhibitor of OPC differentiation.
We next investigated what factors might be mediating clusterin's inhibition of OPC differentiation.OPCs have been shown to produce a variety of growth factors and cytokines that can signi cantly alter their local environment [32][33][34][35][36]. Additionally, clusterin has been shown to regulate to production of cytokines [37][38][39].Based on this data, we wondered if clusterin might be inhibiting OPC differentiation by affecting growth factor and cytokine production.We performed a Luminex Assay on the supernatant of OPCs treated with vehicle or clusterin and found that clusterin altered the secretion of multiple proteins, mostly notably increasing the production of VEGF and signi cantly decreasing the production of IL-9 (Fig. 5A).Since growth factors from the VEGF family have been shown to induce OPC proliferation [40], we tested whether the increase in VEGF following clusterin treatment was responsible for keeping OPCs in an undifferentiated state.However, treatment of OPCs with a combination of clusterin and an anti-VEGF antibody failed to rescue the differentiation defect observed with clusterin treatment (Supplementary Fig. 2A-C).
The next candidate we assessed as the potential mechanism for the effects of clusterin on OPCs was IL-9, since its production was reduced by more than 90% following clusterin treatment (Fig. 5A).IL-9 is a relatively understudied cytokine known to be produced by T-cells [41].Surprisingly, we found that the addition of exogenous IL-9 was su cient to rescue the differentiation de cits induced by clusterin treatment (Fig. 5B-D).Overall, this data demonstrates that clusterin blocks the differentiation of OPCs by inhibiting the production of IL-9, and that IL-9 is an important factor in proper differentiation of OPCs.

Clusterin deletion improves myelination in vivo.
Given our data showing that the expression of myelin genes was inhibited by clusterin, we next wanted to test if clusterin deletion would affect myelination in vivo.We performed transmission electron microscopy on the corpus callosum in 9-month-old clusterin knockout (Clu-/-) and WT mice.We measured the g-ratio of myelinated axons and found that Clu-/-mice had a lower g-ratio than control animals, indicating thicker myelination (Fig. 6A-B).This result shows that, at baseline, clusterin expression plays an inhibitory role in myelination.Since clusterin expression increases in the 5xFAD model, we next explored if deletion of clusterin could improve the myelination de cits observed in these mice [42][43][44].To do this, we bred Clu-/-mice to 5xFAD mice to generate 5xFAD; Clu-/-mice.We again performed transmission electron microscopy and measured the g-ratio of myelinated axons in the corpus callosum at 9 months old.Just as with the WT mice, we observed a decrease in the g-ratio of myelinated axons in the 5xFAD; Clu-/-mice compared to the 5xFAD controls (Fig. 6C-D).Overall, our results indicate that clusterin serves as an inhibitor of myelination, both under homeostatic conditions and in the context of 5xFAD pathology.

Discussion
Since the discovery of Aβ plaques in the brains of AD patients, therapeutic development has been focused on reducing plaque load.However, no therapeutic candidate, even when successful at reducing plaques, has succeeded in slowing disease progression [45].Over recent years, a growing body of literature has speculated that other altered biological processes may be the driving force behind the clinical decline observed in AD patients [46].Increasingly, OPCs and myelin have been implicated in the etiology and progression of AD symptoms.A recent study demonstrated that the ablation of senescent OPCs was su cient to improve the memory impairment observed in the APP/PS1 model of AD [23].Additionally, multiple studies have shown that inhibiting OPC differentiation prevents memory consolidation and recall, and that therapeutically increasing myelination can improve memory performance in AD models [3][4][5]47].Here, we offer a potential mechanism for the pervasive myelin de cits observed in Alzheimer's disease and the memory decline associated with these de cits [48,49].We demonstrate that clusterin is upregulated in the brains of AD patients as well as in a mouse model of AD.We found that human OPCs express clusterin in both normal aging as well as AD patients, and that OPCs in 5xFAD mice upregulate the production of the clusterin protein.Phagocytosis of debris, including Aβ oligomers, myelin, and apoptotic cells, e ciently drives the upregulation of clusterin in OPCs.We show that clusterin, through its effects on the production of IL-9, is a potent inhibitor of OPC differentiation and the production of myelin proteins.Finally, we demonstrate that in vivo knockout of clusterin produces thicker myelin sheaths under normal conditions and abrogates myelin loss observed in 5xFAD mice.
A common single-nucleotide polymorphism (SNP) in clusterin has been recognized as a signi cant risk factor for late onset AD for over a decade [9].While there are con icting reports in the literature regarding how this SNP effects the function and accumulation of clusterin, there are multiple studies indicating that that increased levels of clusterin in the plasma of AD patients, regardless of the presence of a SNP at the CLU locus, correlates with a more rapid cognitive decline and an increase in brain atrophy [11,[50][51][52].These correlative studies demonstrating that excess levels of clusterin are associated with more severe AD symptoms are supported by studies indicating that removing clusterin in mouse models of AD results in a reduced plaque load as well as improved performance on memory tasks [53][54][55].
There are, however, reports in the literature indicting that clusterin might be involved in the clearance of Aβ plaques and the protection of neurons [56,57].These discrepancies likely indicate that clusterin is a multi-functional protein that may be bene cial when expressed at homeostatic levels, but could become detrimental when signi cantly upregulated in the context of disease pathology.Additionally, since there is an increasing amount of data indicating that the reduction of Aβ plaques in AD patients does not improve clinical symptoms, it is possible that clusterin's inhibition of OPC differentiation, as described here, may be one of the mechanism though which increased levels of clusterin might contribute to AD progression [58].
It remains to be determined exactly which cell types might be contributing to the increase in clusterin observed in Alzheimer's disease.While we have demonstrated that OPCs upregulate clusterin production in response to the phagocytosis of Aβ as well as other cellular debris, and human single cell sequencing data indicates that OPCs are the only cell type in the brain that show an increase in clusterin expression in AD patients compared to healthy controls, it is known that astrocytes produce a signi cant amount of clusterin [15,59].However, while the exact cellular sources of elevated levels of clusterin in AD is still an open question, the fact that the most abundant isoform of clusterin encodes for a secreted chaperone protein indicates that clusterin from any cellular source, whether OPCs or any other cell type, can impact the function and differentiation capacity of OPCs [60].
We were surprised to nd that clusterin decreased the differentiation of OPCs by blocking the production of IL-9.While IL9 receptor expression has been documented in OPCs, this is the rst time that OPCs have been shown to produce IL-9 and that IL-9 is important for the proper differentiation of OPCs [61].While Tcells are known to be the main producers of IL-9, there is increasing evidence in the literature that OPCs can also contribute to cytokine production [32,34,62,63].The ability of OPCs to function as immunomodulatory cells is garnering increasing support in the literature, and our data indicating the importance of IL-9 in the function of OPCs offers an intriguing avenue for additional investigation [64,65].
It has been known that clusterin is increased in the brains of AD patients for over three decades [66].However, the mechanism of clusterin's effects on the progression of AD remains unclear.With an antisense-oligonucleotide targeting clusterin having progressed to Phase 3 clinical trials for prostate cancer, our data demonstrating that clusterin negatively affects OPCs and myelin production offers a new therapeutic avenue to potentially improve myelin integrity and memory de cits plaguing AD patients [67,68].What's more, our data linking clusterin and myelin dysregulation is supported by MRI data demonstrating that the clusterin risk allele is associated with a decrease in white matter integrity in healthy young adults, prior to onset of any cognitive decline [69].

Conclusions
Overall, our data offers a novel mechanism for the association between clusterin as a genetic risk factor and Alzheimer's disease myelin pathology.Further, our study provides a foundation for development of therapeutic targeting Clusterin for the treatment of AD.

Declarations Figures
OPCs expresses the AD-risk factor clusterin.
)=3.767.C, Detection of clusterin RNA (in situ hybridization, red) and Aβ protein (immunohistochemistry, brown) in late-stage AD brain (late-stage AD n=2; from two independent experiment).Arrowheads indicate clusterin-expressing cells around plaques.D, Representative images of clusterin expression (white) in the cortex and hippocampus of WT and 5xFAD mice.Scale bar=30µm.E, Quanti cation of clusterin coverage in the cortex and hippocampus of WT and 5xFAD mice (depicted in C; WT n=6, 5xFAD n=6; from two independent experiments).Statistics calculated using an unpaired Student's t-test.Cortex: t(10)=5.049;Hippocampus: t(10)=3.119.F, In situ hybridization for OPCs (PDGFRAin green, OLIG2 in white) expressing Clusterin (CLU; red) in normal aging and late-stage AD brains (normal aging n=1, late-stage AD n=1; from one independent experiment).Scale bar=10µm.G, Representative images of clusterin expression (red) in OPCs (Pdgfrα; green and Olig2; white) in the cortex and hippocampus of WT and 5xFAD mice.Colocalization of Pdgfrα and clusterin depicted in purple.Scale bar=20µm.H, quanti cation of the percentage of Pdgfrα that colocalizes with clusterin in the cortex and hippocampus of WT and 5xFAD mice (depicted in G; WT n=9, 5xFAD n=9; from two independent experiments).Statistics calculated using an unpaired Student's t-test.Cortex: t(16)=2.761;Hippocampus: t(16)=3.765.