Amyloidosis is associated with thicker myelin and increased oligodendrogenesis in the adult mouse brain

Abstract In Alzheimer's disease, amyloid plaque formation is associated with the focal death of oligodendrocytes and soluble amyloid β impairs the survival of oligodendrocytes in vitro. However, the response of oligodendrocyte progenitor cells (OPCs) to early amyloid pathology remains unclear. To explore this, we performed a histological, electrophysiological, and behavioral characterization of transgenic mice expressing a pathological form of human amyloid precursor protein (APP), containing three single point mutations associated with the development of familial Alzheimer's disease (PDGFB‐APPSw.Ind, also known as J20 mice). PDGFB‐APPSw.Ind transgenic mice had impaired survival from weaning, were hyperactive by 2 months of age, and developed amyloid plaques by 6 months of age, however, their spatial memory remained intact over this time course. Hippocampal OPC density was normal in P60‐P180 PDGFB‐APPSw.Ind transgenic mice and, by performing whole‐cell patch‐clamp electrophysiology, we found that their membrane properties, including their response to kainate (100 µM), were largely normal. However, by P100, the response of hippocampal OPCs to GABA was elevated in PDGFB‐APPSw.Ind transgenic mice. We also found that the nodes of Ranvier were shorter, the paranodes longer, and the myelin thicker for hippocampal axons in young adult PDGFB‐APPSw.Ind transgenic mice compared with wildtype littermates. Additionally, oligodendrogenesis was normal in young adulthood, but increased in the hippocampus, entorhinal cortex, and fimbria of PDGFB‐APPSw.Ind transgenic mice as pathology developed. As the new oligodendrocytes were not associated with a change in total oligodendrocyte number, these cells are likely required for cell replacement.

White matter degeneration may occur early in human Alzheimer's disease pathology. Diffusion tensor imaging (DTI) studies, examining individuals in the preclinical stages of disease, determined that measures of fractional anisotropy increased and measures of mean diffusivity decreased in white matter regions such as the fornix, cingulum, and corpus callosum, and changes in these DTI parameters correlated with amyloid β1-42 load (Gold et al., 2014;Racine et al., 2014;Shi, Zhao, Wong, Wang, & Mok, 2015). Furthermore, in preclinical individuals carrying genetic mutations that increase their risk of developing Alzheimer's disease, a lower cerebral spinal fluid concentration of amyloid β1-42, indicative of increased amyloid plaque deposition (Grimmer et al., 2009), correlated with more severe white matter hyperintensities (Lee et al., 2016;Scott et al., 2015).
In vitro, rodent oligodendrocyte and oligodendrocyte progenitor cell (OPC) function is impaired by amyloid proteins. The exposure of cultured rat oligodendrocytes to amyloid β1-42 or amyloid β25-35 resulted in oxidative stress and cell death (Lee et al., 2004;Xu et al., 2001), and exposure to amyloid β1-42 impaired myelin sheath formation (Horiuchi et al., 2012). Additionally, exposure of a mouse OPC line (mOP) to amyloid β1-42 induced cell death of both differentiated and undifferentiated cells (Desai, Guercio, Narrow, & Bowers, 2011). The influence of amyloid pathology on OPC and oligodendrocyte health has also been examined in vivo, in mice carrying human pathological variants of amyloid precursor protein (APP) and PSEN1 (Desai et al., 2011). In APP/PSEN1 transgenic mice, amyloid plaques formed by 2 months of age (Radde et al., 2006), and expression of the OPC proteoglycan NG2 and the oligodendrocyte proteins CNPase and MBP increased in the hippocampus (Wu et al., 2017).
At 6 months of age, despite an increase in OPC density and proliferation and increased newborn oligodendrocyte density, focal demyelination was detected in association with amyloid plaques, and myelin aberrations were apparent including double ensheathment, excess cytoplasm in the inner loop, myelin out-folding, degenerating sheaths, and myelin ballooning (Behrendt et al., 2013). However, in these mice it is difficult to attribute specific changes in cells of the oligodendrocyte lineage with APP overexpression or amyloidosis, as PSEN1 modulates Notch signaling (Newman et al., 2014), and can directly impact oligodendrocyte maturation and myelination (Zhang, Tadesse et al., 2009). In triple transgenic (3xTg) mice that carry human pathological variants in APP Sw , PSEN1 M146V , and MAPT P301L ,

Significance
Alzheimer's disease is the leading cause of dementia and presents growing social and economic challenges. Once initiated, Alzheimer's disease can affect all cell types in the brain and treating this condition will likely require early detection and a combination therapy to halt disease progression. This study has determined that early amyloid pathology affects immature brain cells called oligodendrocyte progenitor cells (OPCs) and mature brain cells called oligodendrocytes, but does not prevent OPCs from producing new cells. Our findings suggest that OPCs are a possible target to protect against early neurodegeneration in Alzheimer's disease. myelin protein expression was reduced in the hippocampal CA1 at 2 months of age and Schaffer collateral axons were dystrophic and had granulated myelin (Desai et al., 2009), such that fewer myelinated CA1 axons were present by 6 months of age (Desai et al., 2010).
Although the expression of amyloid β1-42 was not increased until 6 months of age in the 3xTg mice (Desai et al., 2010), the viral delivery of intracellular targeted anti-amyloid β antibodies at 2 months of age, to prevent amyloid β aggregation, resulted in normal myelination at 6 months (Desai et al., 2010), suggesting that amyloid pathology was the primary driver of oligodendrocyte damage in these mice.
We aimed to determine whether pathological APP signaling, in the absence of pathological PSEN1 or MAPT, was sufficient to influence oligodendrocyte lineage cell function in the hippocampus, fimbria, or entorhinal cortex in early stages of disease. The hippocampus and entorhinal cortex are brain regions affected early in Alzheimer's disease (Du et al., 2001;Pennanen et al., 2004), and the fimbria is the major white matter tract that connects the hippocampal hemispheres to subcortical and cortical regions such as the thalamus and prefrontal cortex (Jin & Maren, 2015;Wyss, Swanson, & Cowan, 1980). Herein, we show that mice carrying the Swedish and Indiana mutations in APP (PDGFB-APP Sw.Ind transgenic mice) maintain a normal density of OPCs and oligodendrocytes in the hippocampus from P60 to P180, however, OPC behavior is altered by amyloid pathology. In young adulthood (P100), OPCs in the hippocampus of PDGFB-APP Sw.Ind transgenic mice have an increased response to GABA, displaying larger currents upon bath application of the neurotransmitter. Oligodendrocyte maturation also appears to be affected in the hippocampus of these mice, as the nodes of Ranvier are shorter and the paranodes longer, and this phenotype is associated with increased myelin thickness by P100. The number of new oligodendrocytes produced by adult OPCs was normal in early adulthood, but increased in the hippocampus, entorhinal cortex, and fimbria of PDGFB-APP Sw.Ind transgenic mice as amyloid pathology developed. As total oligodendrocyte density was unchanged by P180, it is likely that the newborn oligodendrocytes replace oligodendrocytes lost to pathology.

| Ethical approval
All animal experiments were approved by the Animal Ethics Committee of the University of Tasmania (13741 and 16151) and carried out in accordance with the Australian code of practice for the care and use of animals in science. Details of animal experiments are reported in accordance with the ARRIVE guidelines.

| Experimental design and procedures
The source for all materials used in this study is clearly identified in text or in Table S1
Each membrane was washed thrice in TBS-T before being incubated with goat anti-mouse HRP (1:10,000, Dako; RRID:AB_2617137) diluted in TBS-T containing 1% (w/v) skim milk powder. After washing in TBS-T the membrane was exposed to a 1:1 mix of Immobilon Western™ HRP Peroxidase Solution (Millipore) and Luminol Reagent (Millipore) to visualize the protein bands on an Amersham Imager 600 (GE Healthcare Life Sciences, UK). To control for protein loading, membranes were stripped by washing with PBS, TBS-T, and blot stripping buffer (ThermoScientific), before incubating with mouse anti-β-actin (1:1,000, Sigma; RRID:AB_476692) diluted in TBS-T containing 5% (w/v) skim milk. After washing, secondary goat antimouse HRP (1:10,000, Dako; RRID:AB_2617137) was applied and the protein bands and visualized as previously described. Human APP or amyloid β oligomer expression was calculated by measuring the integrated density of the APP protein band at ~100 kDa or the amyloid β oligomer band at ~12 kDa (Collins, King, Woodhouse, Kirkcaldie, & Vickers, 2015) and normalizing the signal to β-actin protein (~42 kDa) expression levels for each sample.

| Locomotor and cognitive testing
Behavioral testing was carried out on wildtype (WT) and APP transgenic littermates at P60, P90, or P180. Individual mice were tested once, and separate cohorts analyzed at each age. All behavioral testing was carried out during the dark phase of the light-dark cycle.
Mice were moved to the testing room 2 hr prior to the light cycle change and habituated to the room for 3 hr. All testing was carried out within the same 5-hr window of the dark phase. Sodium lights were used in the room and bright lights were used above the maze as needed. All trials were video recorded and animal movement tracked using automated tracking software (EthoVision XT 11, Noldus, Netherlands; RRID:SCR_000441). Male mice were tested prior to female mice, but the order of testing was otherwise randomized for each test. All equipment was cleaned with 70% ethanol between trials.

T-Maze
The T-Maze was performed using a protocol adapted from Deacon and Rawlins (2006). A mouse was placed in the start arm and once they chose to explore the left or right arm of the maze, retreat from that arm was blocked for 1 min. The mouse was then returned to the start arm and allowed to make another choice. This was repeated 10 times. Mice naturally exhibit exploratory behavior and tend to choose the arm not visited in the previous trial, therefore, returning to the same arm in successive trials was recorded as an error. The number of trials in which the mouse failed to alternate was recorded and is expressed as a proportion (%) of the number of trials [(errors/ total trial number) × 100]. If a mouse failed to complete the trial (e.g., did not leave the start arm), data from that mouse were excluded from analysis.

Open field
The open field assessment was performed using a protocol adapted from Wang et al. (2013), to assess locomotor and anxiety-like behavior. Each mouse was placed in an open square arena (30 cm 2 , with walls that were 20 cm in height) lit (200 lux) to create a bright center and dark perimeter, and the speed of movement and total distance moved was measured over a 10 min period.

Barnes maze
Mice underwent a shortened version of the Barnes maze protocol, adapted from Attar et al. (2013). On day 1, mice were placed in the brightly lit center (120 lux) of an elevated (30 cm above the ground), circular maze (100 cm diameter) that contained 20 holes evenly spaced around the circumference. After 1 min, the mice were gently directed to the escape box located underneath one of the holes in the circumference and left to habituate to the box for 5 min. On days 2 and 3, the maze was raised to 70 cm, and light intensity in the center of the maze increased to 160 lux. Distinct patterns were placed on each wall surrounding the maze, acting as spatial reference points that remained consistent throughout all trials. At the start of each trial, the mouse was placed at the center of the maze under a covered start box for 15-30 s before the box was removed, and the mouse left to explore until it found the escape box or 5 min elapsed.
If a mouse did not find the escape box prior to the end of the trial, it was given direction to the box and allowed to enter it. After entering the escape box, each mouse was left for 1 min before being returned to the home cage to await the next trial. Mice were trained to learn the location of the escape box across three trials per day with an inter-trial interval of 30-45 min. During training, approaching any hole that did not lead to the escape box was considered a primary error, and the number of primary errors made during a trial was measured as an indicator of learning (reviewed by Gawel, Gibula, Marszalek-Grabska, Filarowska, & Kotlinska, 2019).
Short-term and long-term memory were assessed 1 day and 2 weeks after the initial training, respectively. For each memory probe trial, mice were returned to the maze with the escape box removed and were left to explore the maze for 5 min. The maze was divided into four quadrants within the tracking software (EthoVision XT 11) and the quadrant containing the hole that previously led to the escape box was designated the target zone. The proportion of time spent within the target zone during the probe trial was measured as an indicator of intact memory for the location of the escape box.
Upon breakthrough, resting membrane potential (RMP), capacitance, membrane resistance, and the magnitude of the voltage-gated inward (sodium) current, elicited by a voltage step from −60 mV to 20 mV, were recorded as previously described (Clarke et al., 2012).
Cells with a voltage-gated sodium channel current <60 pA were classified as newly differentiated oligodendrocytes (Clarke et al., 2012) and were consequently removed from analysis. Access resistance was measured before and after each recording and was between 12 Recordings continued for a wash-out period (2-10 min) after drug application to ensure that the health of the cell was not affected during the recordings of KA-or GABA-evoked currents. The average steady state current magnitude, in the last 50 ms of the voltage step, was measured using clampfit 10.5 (molecular devices) and the evoked current (current in the presence of drug minus baseline current) reported. After recording the KA-or GABA-evoked current, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; AMPA/KA receptor antagonist, 10 µM, Sigma) or picrotoxin (PTX, GABA A receptor antagonist, 100 µM, Sigma) was bath applied for 2 min before reapplication of KA or GABA to the same cell.

| Confocal microscopy
For cell density quantification, confocal images were collected from the entorhinal cortex, hippocampus, and fimbria of n = 3 brain sections per mouse for each staining condition, using an Andor spinning disk confocal microscope with Nikon Software (Andor Technology Ltd., Belfast, Northern Ireland) or UltraView Nikon Ti spinning disk confocal microscope with Volocity software (Perkin Elmer, Waltham, USA). Images were collected using a 20x air objective (3 µm z-spacing) with standard excitation and emission filters for DAPI, FITC To measure node and paranode lengths, Hoescht 33342 nuclear labeling was initially used to identify the CA1 region of the hippocampus or the fimbria, and to ensure the unbiased selection of non-overlapping fields of view within each structure. Images were collected using an UltraView Nikon Ti spinning disk confocal microscope (100× oil objective; 0.5 µm z-spacing). When nodes and their flanking paranodes were intact within a single z-plane, they were measured manually using Fiji software (NIH, Washington DC, USA).
For the CA1 region of the hippocampus, 30-90 nodes and paranodes were measured, sampled from four fields of view per mouse. For the fimbria, 36-94 nodes and paranodes were measured, sampled from two fields of view per mouse. All measurements were made by a researcher blind to genotype.
Tissue was transferred to fresh 100% epon for 4 hr before being polymerized at 60°C for 72 hr.

| Statistical analyses
Statistical analyses were performed using GraphPad Prism 8.0 (La Jolla CA, USA; RRID:SCR_002798). Power analyses to determine sample size were not carried out a priori for the experiments included in this paper. Based on our previous experience with the techniques being applied, we aimed for sample sizes large enough to reliably detect a moderate effect size. For example, we aimed for a minimum of n = 12 mice per genotype per timepoint for behavioral analyses and n = 4 mice per genotype per timepoint for histological analyses. The large number of APP mice that died unexpectedly ( Figure 1) limited the number of mice available for our experimental analyses and for some experiments or timepoints we were only able to obtain n = 3 mice in a group. These small sample sizes are a limitation to our study, as we do not have sufficient power to detect differences with an effect size smaller than 0.5 (Sullivan & Feinn, 2012). However, we consistently show large differences between WT and APP group means (effect sizes >1.0) and post hoc power analyses revealed that most comparisons had a statistical power >80% (G*Power 3.1; RRID:SCR_013726), with the exception being our t-test comparison of average hippocampal node length per mouse (n = 3 mice per group), which only achieved a power of 58%.
The distribution of each data set was evaluated to determine whether the data were normally distributed using the d'Agostino and Pearson normality test or Shapiro-Wilk normality test where n ≥ 5.
Data that were normally distributed were analyzed by a parametric test [one-way analysis of variance (ANOVA) or two-way ANOVA for group comparisons with a Bonferroni post hoc test], and data that were not normally distributed were analyzed using a Mann-Whitney U test or Kolmogorov-Smirnoff test. For data sets with n = 3 in any group, we were unable to test for normality, however, we applied parametric tests to analyze these data as the non-parametric equivalents rely on ranking and are unreliable for small sample sizes (GraphPad Prism 8.0). A survival curve comparison was performed using a Log-Rank (Mantel-Cox) test. Data are presented as mean ± standard deviation (SD). Statistical significance was established as p < 0.05. Statistical details are reported in each figure legend and individual data points are presented on each graph. Data supporting these findings will be made available by the corresponding author upon reasonable request.

| APP mice develop histopathological features of Alzheimer's disease by P180
To confirm the expression of human APP in the brain of APP transgenic mice, we generated hippocampal protein lysates from WT, Prnp-MAPT P301S (MAPT) and APP mice at P30, P60, P90, and P180.
By performing a series of Western blots, to detect immature and mature human APP and oligomerized amyloid β (6E10 antibody), we determined that human APP and oligomerized amyloid β were already expressed in the APP mouse hippocampus at P30, and that expression was relatively stable over time, but increased by P180 (Figure 1a-c). Human APP (~100 kDa) and amyloid β (~12 kDa) were not detected in hippocampal lysates from WT or MAPT transgenic mice (Figure 1b,c), which acted as an additional negative control for this experiment. A non-specific protein band (~55 kDa) was detected in lysates from MAPT transgenic mice, however, this band does not correspond to human APP (Grant et al., 2019) but likely corresponds to human MAPT (Kalani et al., 2017;Pu et al., 2018), which shares a common epitope with amyloid β (Griner et al., 2019).
To identify the cells that overexpress human APP within the hippocampus, coronal brain cryosections from P180 WT and APP mice were immunolabeled to detect human APP (6E10 antibody) and the neuronal marker microtubule associated protein 2 (MAP2), the mature oligodendrocyte marker aspartoacylase (ASPA), or the OPC marker, platelet-derived growth factor α (PDGFRα) ( Figure S1). Human APP was not expressed by WT mice, but was F I G U R E 2 APP transgenic mice are hyperactive but show no overt learning and memory deficit by 6 months of age. Data are presented as mean ± SD for distinct WT and APP cohorts at each age (n = 8 P60 WT, 9 P60 APP, 10 P90 WT, 10 P90 APP, 11 P180 WT, and 11 P180 APP mice). *p < 0.05, **p < 0.01, ***p < 0.001 denote significant differences identified by Bonferroni's post hoc analyses. TA: target quadrant; OP: opposite quadrant; AL: adjacent left quadrant; AR: adjacent right quadrant. Scale bars represent 5 cm (a-b) and 20 cm (g-h) [Color figure can be viewed at wileyonlinelibrary.com] strongly expressed throughout the hippocampus and fimbria of APP mice ( Figure S1). More specifically, human APP was strongly expressed by pyramidal, dentate granule, and mossy neurons in the hippocampus and by all ASPA + mature oligodendrocytes, but not OPCs (PDGFRα + ), in the hippocampus and fimbria of P180 APP mice ( Figure S1).
To establish the time frame over which human APP expression caused the formation of amyloid plaques in the brains of APP transgenic mice, coronal brain cryosections from P60, P90, and P180 WT and APP mice were stained with thioflavin S (Figure 1dj, green), which binds to β-sheet structures and identifies amyloid β plaques (Bussière et al., 2004;Sun, Nguyen, & Bing, 2002).
Cryosections were co-labeled to detect the microglial marker Iba1 By quantifying the survival of WT and APP transgenic mice from birth until P180, we determined that ~60% of APP transgenic mice died prior to P180, compared with only ~3% of WT mice [Logrank (Mantel-Cox) test, p < 0.001, Figure 1k]. As we next aimed to characterize the behavioral consequences of APP Sw.Ind overexpression, it should be noted that the impaired survival of APP transgenic mice introduces an unavoidable bias into our analyses, skewing our characterization toward the less affected mice that survive to the older ages.

| APP mice exhibit hyperactive behavior by P60 but do not develop spatial learning deficits by P180
To compare the cognitive performance of WT and APP transgenic mice prior to and during plaque formation, WT and APP mice were Working memory performance was subsequently evaluated by assessing spontaneous alternation in the T-maze. We found that WT and APP mice performed similarly at P60, but that by P90 APP mice persistently made more repeated arm entries (errors) than their WT littermates (Figure 2f), suggesting that these mice have impaired working memory or attentional processing that is likely associated with their hyperactivity (Kim, Woo, Lee, & Yoon, 2017;Montarolo et al., 2019). When evaluating short-and long-term memory retention by WT (Figure 2g) and APP transgenic mice (Figure 2h), using the Barnes maze spatial navigation task, we found that regardless of age, APP mice and their WT littermates performed equally well in the short-term memory probe trial, 1 day after learning the location of the escape box [P60: WT 35.11 ± 6.9, APP 34.02 ± 9.2; P90: WT 40.35 ± 19.2, APP 44.90 ± 13.5; P180: WT 45.10 ± 12.8, APP 40.02 ± 11.8; mean ± SD, time in target zone (%)], and again 2 weeks later during the long-term memory probe trial (Figure 2i). This is highlighted by data showing that even at

| OPC density and membrane properties are unchanged but the response to GABA is increased at P100
To determine how OPC behavior might be affected by amyloid pathology, we first quantified the density of PDGFRα + OPCs in the hippocampus, entorhinal cortex, and fimbria of WT and APP mice F I G U R E 3 OPC density and membrane properties are normal in APP transgenic mice. (a-f) Coronal brain sections (30 µm) from P120 WT and APP mice were stained to detect PDGFRα + OPCs in the hippocampus, entorhinal cortex, and fimbria. (g) Quantification of OPC density in the hippocampus of P60, P120, 150 and P180 WT and APP transgenic mice [Two-way ANOVA, genotype: F (1, 30) = 13.66, p < 0.001; age: F (3, 30) = 0.14, p = 0.93; interaction: F (3, 30) = 0.91, p = 0.44; n = 6 P60 WT, 5 P60 APP, 4 P120 WT, 5 P120 APP, 6 P150 WT, 3 P150 APP, 6 P180 WT, and 3 P180  (Figure 3a-i). We found that OPC density was slightly reduced in the hippocampus of APP mice compared with control mice at P67, however, this difference was not maintained at later ages (Figure 3g).
In the entorhinal cortex ( Figure 3h) and fimbria (Figure 3i), OPC density was not affected by genotype and remained stable over time.
To determine whether amyloid pathology affected the membrane properties of OPCs, we performed whole-cell patch-clamp analysis of GFP + OPCs in the hippocampus of brain slices collected from WT, MAPT, or APP transgenic mice carrying the Pdgfrα-H2BGFP transgene. We report that neither the expression of MAPT P301S nor APP Sw.Ind altered the OPC membrane capacitance (an approximate measure of cell size; Figure 3j), membrane resistance (Figure 3k), or resting membrane potential (Figure 3l), which were equivalent for WT, MAPT, and APP transgenic mice at P30 and P100. Furthermore, the magnitude of the inwards voltage-gated sodium channel current recorded from P30 and P100 OPCs was equivalent for WT, MAPT, and APP transgenic mice (Figure 3m). APP mice are hyperactive by P60-a phenotype that may reflect altered neurotransmitter signaling or an inhibitory-excitatory imbalance in the brain Sanchez et al., 2012;Snowden et al., 2019;Verret et al., 2012), therefore, we next used whole-cell patch-clamp electrophysiology to examine the ability of OPCs in the hippocampus of WT, MAPT, and APP mice to respond to excitatory and inhibitory neurotransmitters. GFP + OPCs from WT mice were first held at −60 mV and KA (100 μM) was bath applied to activate the ionotropic AMPA/KA subtype of glutamate receptors. KA application evoked an inward current that was sensitive to the AMPA/KA receptor antagonist CNQX (Figure 4a-c). The KA-evoked current in OPCs at P30 and P100 had a non-ohmic voltage dependence and did not reverse (Figure 4a), which is consistent with previous reports showing that in OPCs sodium entry through AMPA/KA receptors inhibits the delayed rectifier voltage-gated potassium channel current at depolarized potentials (Borges & Kettenmann, 1995). We found that the amplitude of the KA-evoked current did not differ between P30 and P100, and was equivalent in WT, MAPT, and APP transgenic mice (Figure 4d,e).
To assess the response of OPCs to the inhibitory neurotransmitter GABA (100 μM), OPCs were held at 0 mV while GABA was bath applied. At P30 and P100, GABA evoked an outwardly rectifying current that was completely abolished in the presence of PTX (100 μM;

| Node of Ranvier length is decreased and paranode length increased in the hippocampus of P100 APP mice
To determine whether myelin integrity was affected in young adult APP transgenic mice, we first examined the morphology of the F I G U R E 4 OPCs from APP transgenic mice have a heightened response to GABA. (a) I-V relationship for the current evoked in hippocampal OPCs by the bath application of 100 µM KA (mean steady state baseline current was subtracted from the mean steady state current in the presence of KA) in P30 or P100 WT mice. Each cell was submitted to multiple voltage steps [Repeated measures two-way ANOVA, age: immunolabeled to detect the nodal protein NaV1.6 (red) and the paranodal protein Caspr (green). We measured the length of each of these structures from confocal micrographs, and found that node of Ranvier length was shorter in the hippocampus of APP transgenic  (Figure 5n-p) were quantified in the fimbria, each was found to be equivalent in WT and APP transgenic mice, suggesting that this phenotype is region specific.

| Myelin thickness is increased in the hippocampus of P100 APP transgenic mice
As paranode lengthening can result from myelin decompaction (Howell et al., 2006;Stojic, Bojcevski, Williams, Diem, & Fairless, 2018) or an increased number of myelin wraps (Jeffries et al., 2016;Snaidero et al., 2014), we next examined the ultrastructure of hippocampal myelin in P100 WT ( Figure 6a) and APP mice ( Figure 6b) by transmission electron microscopy. We found that axon density (Figure 6c), myelinated axon density (Figure 6d), and the proportion of axons that are myelinated (Figure 6e) was equivalent between WT and APP mice, suggesting that axon number and the proportion of axons that get myelinated during development is not affected by genotype. However, the g-ratio [axon diameter/ (axon + myelin diameter)] of myelinated axons in the hippocampus of APP mice was reduced relative to WT mice (Figure 6f-g), suggesting that APP mice have thicker hippocampal myelin. This was confirmed when we measured the diameter of the myelinated axons (Figure 6h-l) and found that this was equivalent in WT and APP mice, confirming that the physical size of the axons was not driving the change in myelin thickness. Furthermore, we found that the number of myelin lamellae (wraps) surrounding each axon was increased in APP mice ( Figure 6m). These data indicate that overexpressing human APP in neurons and oligodendrocytes, or early amyloid pathology influences the myelinating behavior of oligodendrocytes in the hippocampus.
OPCs differentiated to produce new PDGFRα-neg YFP + oli-  (Figure 7t) of APP F I G U R E 5 Nodes of Ranvier are shorter and paranodes longer in the hippocampus of APP transgenic mice. (a-d) 100× magnification confocal images of coronal brain sections (30 µm) from P107 WT and APP mice that were stained to detect Caspr (green; paranodes) and Nav1.6 (red; nodes of Ranvier) in the hippocampus and fimbria. Magnified panels depict example nodes of Ranvier that lay within a single zplane and were used for length measurements. Results are presented as mean ± SD. *p < 0.05, **p < 0.01 and ***p < 0.001 denote significant differences identified by Mann-Whitney test, Kolmogorov-Smirnov test, or unpaired t test. Scale bars represent 2.8 µm (a-d) or 1.4 µm (magnified nodes below a-d) [Color figure can be viewed at wileyonlinelibrary.com] mice relative to controls. To determine whether this corresponded to an increase in the number of newborn mature oligodendrocytes, we performed immunohistochemistry to detect YFP, PDGFRα and the mature oligodendrocyte marker, ASPA in the hippocampus and fimbria of P60 + 120 control and APP mice ( Figure S3). We found that the density of YFP + PDGFRα-neg ASPA-neg cells (presumably premyelinating oligodendrocytes) was equivalent in the hippocampus and fimbria of WT and APP mice ( Figure S3). By contrast, the density of YFP + PDGFRα-neg ASPA + mature oligodendrocytes was significantly increased in the hippocampus and fimbria of APP relative to control mice ( Figure S3). In P60 + 120 APP mice, a similar increase in newborn mature oligodendrocyte density was also detected in the retrosplenial cortex ( Figure S4), a region involved in learning and navigation (Vann, Aggleton, & Maguire, 2009), suggesting that oligodendrogenesis may be increased in a number of regions of the APP mouse brain.
Surprisingly, the addition of new oligodendrocytes did not alter the total density of oligodendrocytes in the hippocampus, entorhinal cortex or fimbria of control or APP mice (Figure 8). By performing immunohistochemistry on coronal brain cryosections from P120 or P180 WT and APP transgenic mice to detect the mature oligodendrocyte marker ASPA (Figure 8a-n), we determined that the density of ASPA + oligodendrocytes was higher in the fimbria (Figure 8q) and hippocampus (Figure 8o) than the entorhinal cortex (Figure 8p), but was consistent across age and between WT and APP transgenic mice. As microglia were previously shown to express ASPA in rat brain tissue (Madhavarao et al., 2004;Moffett et al., 2011), we also ensured that ASPA expression did not colocalize with the microglial marker Iba1 in WT or APP brain cryosections ( Figure S5).
As there are 213 ± 65 ASPA + mature oligodendrocytes per mm 2 in the hippocampus in P180 APP mice (Figure 8o; mean ± SD) and of these, 56.09 ± 4.5 per mm 2 are adult-born ASPA + oligodendrocytes ( Figure 7r; mean ± SD), ~26% of the mature ASPA + oligodendrocytes were born after P60. By contrast, only ~10.6% of the mature ASPA + oligodendrocytes present in the hippocampus of P180 WT mice were born after P60. Similarly, of the 2,018.2 ± 612.1 ASPA + oligodendrocytes per mm 2 of fimbria in P180 APP mice, 260.45 ± 28.69 per mm 2 are adult born (mean ± SD), indicating that ~12.9% were born after P60. In WT mice ~5.0% of ASPA + oligodendrocytes in the fimbria were born after P60. These data indicate that OPCs are active in APP mice and make a significant contribution to the total population of mature oligodendrocytes. These data may also suggest that amyloid pathology triggers oligodendrocyte death and replacement by P180, however, the total density of oligodendrocytes in the hippocampus of control and APP mice was variable. Considering this, it may not be possible to detect a difference in total mature oligodendrocyte density of <8% in the fimbria, between control and APP mice, or a difference of ~15% in total mature oligodendrocyte density in the hippocampus between WT and APP mice at P180, in the event that these cells were added to the population with no concurrent loss.

| D ISCUSS I ON
Herein we show that APP transgenic mice express an increased level of human APP in the hippocampus at P180, when they develop amyloid plaques (Figure 1). Behaviorally these mice were hyperactive by P60, which impacted their performance in other behavioral assessments, however, they did not develop spatial memory deficit by P180 ( Figure 2). OPC density is normal in the hippocampus of APP transgenic mice, but by P100, they respond more robustly to GABA (Figures 3 and 4). Additionally, developmental myelination was affected in APP transgenic mice, as the nodes of Ranvier along hippocampal axons were shorter and the paranodes longer in young adulthood, and this phenotype was associated with increased myelin thickness (Figures 5 and 6). OPCs in the hippocampus, entorhinal cortex, and fimbria of young adult APP transgenic mice also produced a normal number of new oligodendrocytes, however, as pathology developed oligodendrogenesis increased ( Figure 7). As this was not accompanied by a change in total oligodendrocyte number ( Figure 8), we propose that APP transgenic mice have a higher level of oligodendrocyte turnover than their WT littermates by P180.

| APP mice are hyperactive prior to amyloid plaque deposition
APP or J20 mice expressed a human variant of APP containing the Swedish (K670N/M671L) and Indiana (V717F) familial Alzheimer's disease linked mutations, in neurons and oligodendrocytes ( Figure S1), driven by the platelet-derived growth factor-beta chain (PDGF-β) promotor (Mucke et al., 2000). Both human APP (~100 kDa) and oligomerized amyloid β (~12 kDa) were present in the hippocampus of APP mice by P30, however, amyloid plaques did not form until P180 (Figure 1), consistent with previous reports (Meilandt et al., 2009;Mucke et al., 2000;Wright et al., 2013). The overexpression of human APP SwInd was also associated with premature mortality (Figure 1; Cheng et al., 2007;Cissé et al., 2011;Dubal et al., 2015;Murakami et al., 2011;Verret et al., 2012), which has been previously attributed to their increased susceptibility to spontaneous seizures . As ~60% of APP mice died by P180, and it is reasonable to assume that the mice that died had developed more severe pathology, the reduced survival of APP F I G U R E 6 Myelin thickness is increased in APP transgenic mice. (a-b) Representative electron micrographs from the CA1 region of the hippocampus of WT (a) and APP (b) mice at P90. (c) Quantification of total axon density (axons/µm 2 ) in WT (open circles) and APP (gray diamonds) mice [Two-tailed, unpaired t test, t (4) = 2.46]. (d) Quantification of myelinated axon density (axons/µm 2 ) in WT and APP mice [Two-tailed, unpaired t test, t (4) = 1.69]. (e) Quantification of the proportion of myelinated axons in the CA1 of WT and APP mice at P90 [Two-tailed, unpaired t test, t (4) = 0.46]. (f) Relative frequency distribution plot showing g-ratio measurements from myelinated axons in the CA1 of P90 WT and APP mice [Kolmogorov-Smirnoff test, K-S D = 0.2926; n = 310 WT and 472 APP axons analyzed from n = 3 mice per group]. (g) Quantification of average g-ratio for WT and APP mice [Two-tailed, unpaired t test: t (4) = 4.18]. (h-k) Representative high magnification electron micrographs of transected myelinated axons within CA1 of WT (h-i) and APP mice (j-k) at P90. (l) Average myelinated axon diameter within CA1 of WT and APP mice [Two-tailed, unpaired t test, t (4) = 2.27]. (m) Quantification of average number of myelin wraps for axons within CA1 of WT and APP mice at P90 [Two-tailed, unpaired t test, t (4) = 4.12]. Results are presented as mean ± SD, n = 3 WT mice and 3 APP mice. *p < 0.05 and ****p < 0.001 denote significant differences identified by Kolmogorov-Smirnov or unpaired t tests. Scale bars represent: 1 µm (a-b), 300 nm (h, j) or 100 nm (i, k) mice unavoidably biased our analyses, skewing our characterization toward the less affected mice that survived to the older ages.
By subjecting WT and APP and mice to a battery of behavioral tasks at P60, P90, and P180, we determined that APP mice were hyperactive by P60, developed episodic working memory deficits by P90, and exhibited anxiety-like behavioral traits but no spatial learning impairment by P180 (Figure 2). Most studies report that APP mice develop spatial learning deficits between 5 and 8 months of age (Cheng et al., 2007;Cissé et al., 2011;Flores et al., 2018;Harris et al., 2010;Mably et al., 2015;Mesquita et al., 2015;Sanchez et al., 2012: Wright et al., 2013, while others report no deficit in Barnes Maze performance before 12 months (Fujikawa et al., 2017;Nunes et al., 2015). By contrast, hyperactivity is a consistent behavioral feature of APP mice (Cheng et al., 2007;Cissé et al., 2011;Dubal et al., 2015;Flores et al., 2018;Fujikawa et al., 2017;Harris et al., 2010;Murakami et al., 2011;Sanchez et al., 2012;Verret et al., 2012;Wright et al., 2013), however, to our knowledge, we are the first to show a tendency toward increased anxiety-like behavior in this strain. Most studies report a decrease in anxiety-like behavior, or a disinhibition of caution in APP mice as early as 2 months of age (Cheng et al., 2007;Cissé et al., 2011;Dubal et al., 2015;Harris et al., 2010;Murakami et al., 2011;Sanchez et al., 2012) or no anxiety-related phenotype (Dekens et al., 2018;Fujikawa et al., 2017;Wright et al., 2013).  (Pervolaraki et al., 2019), which the authors suggest reflects altered decision-making processes, rather than a core deficit in emotional motivation. The hyperactivity exhibited by our APP mice may underpin similar vagaries in decision-making, and contribute to the apparent working memory deficit and anxiety-like behavior in these mice. Altered OPC and/or oligodendrocyte function could also contribute to behavioral abnormalities in APP mice, as both cell types influence neural circuit function, memory, and affective behavior (reviewed Pepper, Pitman, Cullen, & Young, 2018). The ablation of OPCs has been shown to produce anxiety-like behavior within 7 days (Birey et al., 2015), and while the electrophysiological response of OPC is altered by P100 in APP mice, any reduction detected in OPC density is minor, and APP mice do not display an anxiety-like phenotype until P180. Similarly, increased hippocampal myelin thickness and the onset of working memory deficits are detected early in APP mice, however, in other contexts, thick myelin enhances hippocampal-dependent memory formation (Jeffries et al., 2016), making it unlikely that this cellular change impairs working memory performance, but this cannot be entirely ruled out within this pathological context.

| OPCs from P100 APP transgenic mice have a heightened response to GABA
OPCs in the hippocampus of P100 APP mice responded more robustly to the bath application of GABA than those in control mice; this could reflect a change in GABA A receptor expression, or a change in the conductance of GABA A receptors. APP is a synaptic protein that can regulate GABAergic signaling, predominantly by modulating presynaptic metabotropic GABA B receptors or the reversal potential of chloride (reviewed by Tang, 2019). Experimentally manipulating the expression of APP decreases expression of the chloride transporter, KCC2 (K + -Clcotransporter 2; SLC12A5), which alters the effect of GABA A receptor activation on neuronal membrane potential Doshina et al., 2017) and modulates GABA A receptor subunit expression . In hippocampal neuron cultures from APP knockout mice, the amplitude of evoked unitary inhibitory post-synaptic currents is reduced, as is their response to the GABA A receptor agonist, isoguvacine . This phenotype was associated with a ~50% reduction in expression of F I G U R E 7 Adult oligodendrogenesis is elevated in the APP transgenic mouse brain. P60 control and APP mice received tamoxifen at P60, inducing recombination of the Cre-sensitive YFP reporter, effectively labeling OPCs and their progeny. Immunohistochemistry was performed on brain cryosections from mice at P60 + 7 (P67), P60 + 60 (P120), P60 + 90 (P150), and P60 + 120 (P180). (a-b) Low magnification confocal images indicating the hippocampus and fimbria in P60 + 120 control and APP mice [YFP (green), PDGFRα (red) and Hoechst 33342 ( Data are presented as mean ± SD. Control hippocampus n = 6 P60 + 7, 4 P60 + 60, 6 P60 + 90, and 6 P60 + 120; APP hippocampus n = 5 P60 + 7, 5 P60 + 60, 3 P60 + 90, and 3 P60 + 120; control entorhinal cortex n = 5 P60 + 7, 4 P60 + 60, and 4 P60 + 120; APP entorhinal cortex n = 5 P60 + 7, 3 P60 + 60 and 3 P60 + 120; control fimbria n = 6 P60 + 7, 4 P60 + 60, 6 P60 + 90 and 6 P60 + 120; APP fimbria n = 4 P60 + 7, 4 P60 + 60, 3 P60 + 90, and 3 P60 + 120. *p < 0.05, **p < 0.01 and ***p < 0.001 denote significant differences identified by Bonferroni's post hoc analyses. White arrows indicate YFP + PDGFRα-negative newborn oligodendrocytes. Scale bars represent 25 µm (c-j) or 33 µm (k-n) [Color figure can be viewed at wileyonlinelibrary.com] the α1 GABA A receptor subunit in the hippocampus of APP knockout mice, that could be rescued by potentiating KCC2 function . OPCs express few α1-containing GABA A receptors (Lin & Bergles, 2004), but, it is possible that the increased response to GABA detected in OPCs in the hippocampus of P100 APP mice is the result of an APP/KCC2 mediated increase in the expression of α1 or other GABA A receptor subunits. However, human APP was not expressed by OPCs in our mice, and it is unclear whether exogenous APP can alter KCC2 and/or GABA A receptor subunit expression within these cells.
APP also interacts with the presynaptic GABA B1a subunit of GABA B receptors to alter receptor trafficking, and APP deletion reduces the presynaptic GABA B -mediated inhibition of neurotransmitter release (Dinamarca et al., 2019). Secreted APP can also modulate neurotransmission by binding to presynaptic GABA B receptors (Rice et al., 2019). It is possible that APP-mediated changes to presynaptic GABA B signaling could affect neuron-OPC communication and OPC receptor expression, as OPCs can modify their expression of AMPA receptors in response to altered neuronal activity (Ge et al., 2006). OPCs have also been shown to express GABA B receptors (Luyt et al., 2007;Luyt, Varadi, & Molnar, 2003;Serrano-Regal et al., 2019), however, the currents evoked in OPCs by GABA application were completely antagonized by a selective GABA A receptor antagonist, excluding the direct involvement of OPC GABA B receptors in the heightened response to GABA.
A change in the response of OPCs to GABA may also be the consequence of the dysfunctional glutamatergic signaling associated with Alzheimer's disease (Findley, Bartke, Hascup, & Hascup, 2019). GABA A activation on OPCs is considered to be excitatory (Lin & Bergles, 2004), however, the activation of GABA A receptors can still negatively modulate glutamatergic signaling by increasing membrane conductance (shunting) and/or altering intracellular chloride concentration. The bath application of GABA can significantly reduce the response of hippocampal OPCs to KA (Lin & Bergles, 2004), indicating that a heightened OPC GABA A response could serve to dampen pathological glutamatergic signaling onto OPCs. How this might affect OPC behavior is unclear. The activation of GABA A receptors in rat OPCs cultures has no effect on OPC differentiation or myelin protein expression (MBP or MAG;Serrano-Regal et al., 2019).
By contrast, applying the GABA A receptor antagonist, GABAzine, to developmental mouse organotypic cortical slice cultures increased the number of OPCs and oligodendrocytes present at 6 days in vitro (Hamilton et al., 2017). This effect of GABAzine on oligodendrocyte lineage cell number was prevented by blocking neuronal action potentials with tetrodotoxin (TTX), suggesting that the effect of GABAzine was dependent on the release of an activity-dependent factor from neurons (Hamilton et al., 2017). As neither GABAzine nor the GABA A agonist muscimol altered node of Ranvier length in mouse organotypic cortical slice cultures (Hamilton et al., 2017;Zonouzi et al., 2015) it is likely that the changes detected in node length in our study were independent of the increased responsiveness of OPCs to GABA, additionally, the impact that GABAergic signaling has on myelin thickness has not been explored.

| Amyloid accumulation changes myelin ultrastructure
We have shown that paranodes are longer and nodes of Ranvier shorter for axons in the hippocampus of 3-month-old APP transgenic mice, compared with WT controls. Paranodes form at the end of each myelin internode, flanking the node of Ranvier, facilitated by contact-mediated signaling between proteins in the myelin sheath and the axon to maintain the clustered voltage-gated sodium channels at the nodes of Ranvier and tether the myelin internode to the axon (reviewed by Pepper et al., 2018). The effect of amyloid pathology on node of Ranvier length does not appear to be the result of altered axon diameter, as axon diameter was similar in WT and APP transgenic mice. Furthermore, we found no correlation between node of Ranvier length and node diameter (a proxy for axon diameter) in WT or APP transgenic mice [linear regression node length v node width deviation of slope from 0: WT slope = 0.04, F (1, 227) = 2.2, p = 0.14; APP slope = 0.04, F (1, 140), p = 0.22], a finding that is consistent with a previous data showing that node of Ranvier length and diameter did not correlate in the optic nerve or frontoparietal motor cortex of Sprague-Dawley rats (Arancibia-Cárcamo et al., 2017). However, APP can be found at the nodes of Ranvier (Xu et al., 2014) and can increase NaV1.6-mediated sodium currents (Li et al., 2016), making it possible that pathological human APP could directly induce narrowing of nodes of Ranvier in the hippocampus.
In the 3xTg transgenic mouse line, a qualitative decrease was noted in NaV1.6 expression in the CA1 region of the hippocampus and the entorhinal cortex of 6-month-old mice (Desai et al., 2009), which could be the result of myelin degradation and an overall reduction in node of Ranvier number, but could also reflect a change in nodal structure.
F I G U R E 8 Oligodendrocyte number is normal in APP transgenic mice. (a-b) Low magnification confocal images of the hippocampus and fimbria in confocal brain sections from P180 WT and APP mice immunolabeled to detect ASPA. (c-f) Confocal images from the hippocampus in coronal brain cryosections (30 µm) from P120 and P180 WT and APP transgenic mice showing cells labeled with the oligodendrocyte marker ASPA (red) and Hoechst 33342 (blue). (g-j) Confocal images from the entorhinal cortex in coronal brain cryosections from P120 and P180 WT and APP transgenic mice showing cells labeled with the oligodendrocyte marker ASPA (red) and Hoechst 33342 (blue). (k-n) Confocal images from the fimbria in coronal brain cryosections from P120 and P180 WT and APP transgenic mice showing cells labeled with the oligodendrocyte marker ASPA (red) and Hoechst 33342 (blue). (o) Quantification of the density of ASPA + oligodendrocytes in the hippocampus of WT (black bars, open circles) and APP (gray bars, gray diamonds) transgenic mice (oligodendrocytes per mm 2 as adjusted for x-y area only) [Two-way ANOVA, genotype: F (1, 16) = 1.78, p = 0.20; age: F (1, 16) = 1.07, p = 0.31; interaction: F (1, 16) = 2.01, p = 0.17]. (p) Quantification of the density of ASPA + oligodendrocytes in the entorhinal cortex of WT and APP transgenic mice (oligodendrocytes per mm 2 as adjusted for x-y area only) [Two-way ANOVA, genotype: F (1, 18) = 2.08, p = 0.16; age: F (1, 18) = 0.95, p = 0.34; interaction: F (2, 53) = 0.02, p = 0.87]. (q) Quantification of the density of ASPA + oligodendrocytes in the fimbria of WT and APP transgenic mice (oligodendrocytes per mm 2 as adjusted for x-y area only) [Two-way ANOVA, genotype: F (1, 16) = 0.40, p = 0.53; age: F (1, 16) = 0.08, p = 0.77; interaction: F (1, 16) = 0.64, p = 0.43]. Results are presented as mean ± SD. Hippocampus n = 5 P120 WT, 3 P120 APP, 7 P180 WT, and 5 P180 APP; entorhinal cortex n = 6 P120 WT, 3 P120 APP, 7 P180 WT, and 6 P180 APP; fimbria n = 6 P120 WT, 3 P120 APP, 6 P180 WT, and 5 P180 APP. Scale bars represent 30 µm [Color figure can be viewed at wileyonlinelibrary.com] It is likely that the increased paranode length measured for hippocampal axons in the APP transgenic mice is the direct result of APP transgenic mice having thicker myelin. On average, oligodendrocytes that myelinated axons within the hippocampus of APP mice produced an extra myelin lamella, relative to those in WT mice, which is consistent with previous reports of increased myelin thickness in the hippocampus of 2-month-old APP/PSEN1 transgenic mice (Wu et al., 2017). The extra layer of myelin must be anchored at the paranode, and could readily explain the increased paranode length, and perhaps, if the myelin encroaches on the node of Ranvier, the decrease in node of Ranvier length.

| Amyloid accumulation increases oligodendrocyte turnover
OPCs generate new oligodendrocytes in the gray and white matter of the adult mouse brain (Dimou, Simon, Kirchhoff, Takebayashi, & Gotz, 2008;Fukushima et al., 2015;Hill, Patel, Medved, Reiss, & Nishiyama, 2013;Kang, Fukaya, Yang, Rothstein, & Bergles, 2010;Rivers et al., 2008;Young et al., 2013), and while the rate of OPC proliferation and oligodendrocyte addition slows with aging (reviewed by Wang & Young, 2014), OPCs can rapidly proliferate and differentiate in response to a demyelinating event, to facilitate the replacement of lost oligodendrocytes and enable remyelination (Assinck et al., 2017;Baxi et al., 2017;Tripathi, Rivers, Young, Jamen, & Richardson, 2010;Zawadzka et al., 2010). By performing cre-lox lineage tracing using Pdgfrα-CreER T2 :: Rosa26-YFP transgenic mice, we have shown that the number of new mature oligodendrocytes that accumulate in the hippocampus, entorhinal cortex, and fimbria of APP transgenic mice between P60 and P180 is significantly higher than the number added to these regions in the WT mouse brain.
In the hippocampus, oligodendrocyte addition started to deviate between WT and APP transgenic mice between 4 and 5 months of age. However, the effect of amyloid pathology on oligodendrogenesis was delayed in the entorhinal cortex and fimbria, being seen between 5 and 6 months of age. Despite the increased addition of new mature oligodendrocytes to the brain of APP transgenic mice, total oligodendrocyte density in the hippocampus, entorhinal cortex and fimbria was equivalent in WT and APP mice, suggesting that the large number of new cells may serve to replace oligodendrocytes that are lost, rather than increasing total oligodendrocyte number.
Myelin abnormalities have been reported for a number of animal models of amyloid pathology (Chu et al., 2017;Tse et al., 2018), including focal myelin loss associated with amyloid plaques (Mitew et al., 2010;Schmued, Raymick, Paule, Dumas, & Sarkar, 2013). We found that the proportion of hippocampal axons that were myelinated in 3-month-old WT and APP transgenic mice was equivalent, however, the marked increase in oligodendrogenesis observed between 4 and 6 months of age in the APP transgenic mice may suggest that oligodendrocyte loss and myelination are a later feature of the pathogenesis. This time line differs significantly from that of APP Sw mice in which myelin loss was not reported until 18 months of age (Tse et al., 2018), but would be consistent with the marked reduction in myelinated axon number in the hippocampus of 6-month-old 3xTg mice, which occurred prior to plaque deposition (Desai et al., 2009), and the myelin aberrations detected in APP/PS1 mice at 6 months, coincident with plaque detection (Behrendt et al., 2013).
Understanding the cellular changes that occur in the early stages of Alzheimer's disease will enable improved early detection and the development of therapies to halt disease progression. We have determined that early amyloid pathology affects OPCs and myelin morphology, and that as pathology develops there is an increase in oligodendrogenesis without an accompanying change in total oligodendrocyte number, suggesting that new oligodendrocytes are produced to replace lost oligodendrocytes. Our findings contribute to the growing body of literature suggesting that OPCs represent a viable target to protect against early degeneration of the white matter tracts in Alzheimer's disease.

DECL AR ATION OF TR ANS PAREN C Y
The authors, reviewers and editors affirm that in accordance to the policies set by the Journal of Neuroscience Research, this manuscript presents an accurate and transparent account of the study being reported and that all critical details describing the methods and results are present.