Administration of recombinant FOXN1 protein attenuates Alzheimer’s pathology in mice

Abstract

Background

Alzheimer’s disease (AD) is the most common cause of dementia in older adults and characterized by progressive loss of memory and cognitive functions that are associated with amyloid-beta (Aβ) plaques and neurofibrillary tangles. Immune cells play an important role in the clearance of Aβ deposits and neurofibrillary tangles. T cells are the major component of the immune system. The thymus is the primary organ for T cell generation. T cell development in the thymus depends on thymic epithelial cells (TECs). However, TECs undergo both qualitative and quantitative loss over time. We have previously reported that a recombinant (r) protein containing FOXN1 and a protein transduction domain can increase the number of TECs and subsequently increases the number of T cells in mice. In this study we determined the ability of rFOXN1 to affect cognitive performance and AD pathology in mice.

Methods

Aged 3xTg-AD and APP/PS1 AD mice were injected with rFOXN1 or control protein. Cognitive performance, AD pathology, the thymic microenvironment and immune cells were then analyzed.

Results

Administration of rFOXN1 into AD mice improves cognitive performance and reduces Aβ plaque load and phosphorylated tau in the brain. This is related to rejuvenating the aged thymic microenvironment, which results in enhanced T cell generation in the thymus, leading to increased number of T cells, especially IFNγ-producing T cells, in the spleen and the choroid plexus (CP), enhanced expression of immune cell trafficking molecules in the CP, and increased migration of monocyte-derived macrophages into the brain. Furthermore, the production of anti-Aβ antibodies in the serum and the brain, and the macrophage phagocytosis of Aβ are enhanced in rFOXN1-treated AD mice.

Conclusions

Our results suggest that rFOXN1 protein has the potential to provide a novel approach to treat AD patients.

1. Introduction

Alzheimer’s disease (AD) is the most common form of dementia of older adults (2022). AD is characterized by progressive loss of memory and cognitive functions, which is associated with hallmark protein aggregates, amyloid-beta (Aβ) plaques and neurofibrillary tangles. Aβ plays a key role in the pathogenesis of AD because Aβ can accelerate neuronal cell death and neuronal tangle formation, affect synaptic function adversely and eventually cause neuron loss.

The brain has been traditionally considered immune privileged (Louveau et al., 2015). However, compelling data have suggested that the CNS is actually immunocompetent and not completely immune-privileged, and that neuroimmune communication plays an important role in CNS homeostasis and function, as well as brain repair, especially in pathological conditions including AD (Jorfi et al., 2023; Matejuk et al., 2021; Schwartz, 2017). Systemic immune deficiency contributes to cognitive dysfunction and accelerated AD pathology (Marsh et al., 2016; Rosenzweig et al., 2019).

T cells are the major component of the immune system. Although the role of T cells in AD pathogenesis is also controversial (Gate et al., 2020; Jorfi et al., 2023), multiple lines of evidence suggest that T cells play an important role in the CNS maintenance and repair. T cell deficiency is associated with increased neuronal loss in animal models of CNS injury or AD (Marsh et al., 2016; Wolf et al., 2009). In contrast, transplantation of T cells reduces AD pathology (Bryson and Lynch, 2016; Cao et al., 2009; Fisher et al., 2014; Marsh et al., 2016). Systemic T cells not only participate in CNS repair but are also needed for life-long brain plasticity (Miller, 2010; Wolf et al., 2009).

The thymus is the primary organ for T cell generation. Although the thymus continues to export T cells throughout life, it undergoes a profound atrophy with age, a process termed thymic involution/atrophy, resulting in decreased numbers and functional capacity of T cells in older adults, which has direct etiological linkages with many diseases (Lynch et al., 2009; Swain and Nikolich-Zugich, 2009). T cell development in the thymus depends on the thymic microenvironment, in which thymic epithelial cells (TECs) are the major component (Anderson and Takahama, 2012; Chidgey et al., 2007; Ciofani and Zuniga-Pflucker, 2007; Zediak and Bhandoola, 2005). However, TECs undergo both qualitative and quantitative loss over time, which is believed to be the major factor responsible for age-dependent thymic involution/atrophy (Lynch et al., 2009; Swain and Nikolich-Zugich, 2009). It has been reported that AD mice have premature immunosenescence (Giménez-Llort et al., 2012; Mate et al., 2014) that contributes to the AD pathology (Richartz et al., 2005). We have previously reported that rejuvenating the aged thymic microenvironment by mouse embryonic stem cell (mESC)-derived thymic epithelial progenitors (TEPs) results in an increased number of TECs and T cells, leading to attenuated AD pathology (Zhao et al., 2020).

FOXN1 is a pivotal regulator for TEC development (Manley and Condie, 2010; Nehls et al., 1994; Zhang et al., 2012); it is required not only for TEC development in fetal thymus, but also for maintenance of the postnatal thymus (Chen et al., 2009; Cheng et al., 2010; Corbeaux et al., 2010; Gallo et al., 2017; Nowell et al., 2011). We have produced a recombinant (r) protein containing FOXN1 and a protein transduction domain and shown that administration of the rFOXN1 protein increases the number of TECs and T cells in mice that have undergone hematopoietic stem cell transplantation (HSCT) (Song et al., 2016). Because rejuvenating aged thymic microenvironment can ameliorate AD pathology and because rFOXN1 can rejuvenate the aged thymic microenvironment, we determined that the ability of the rFOXN1 to affect AD pathology in AD mice in this study.

2. Material and Methods

2.1. Mice

Homozygous 3xTg-AD, B6129SF2 wild-type (WT), heterozygous APP/PS1, and C57BL/6-Tg (CAG-EGFP)131Osb/LeySopJ (GFP⁺) mice were purchased from Jackson Laboratory. Mice were housed in groups of 2–5 mice/cage and maintained under standard laboratory conditions: light from 7 A.M.-7 P.M., room temperature ~22°C, humidity ~62%, low fat rodent chow and tap water available ad libitum, and Beta Chip bedding changed twice a week. Bone marrow (BM) chimeric mice were generated as described (Rosenzweig et al., 2019; Shechter et al., 2009). In brief, recipient mice were irradiated (950 rad) with head shielding and then reconstituted with 3 × 10⁶ BM cells from GFP⁺ mice. GFP⁺ donor-derived leukocytes in the blood of the recipient mice were analyzed by flow cytometry 2-3 months later. The mice were used in accordance with a protocol approved by the Institutional Animal Care and Use Committee of the University of Connecticut.

2.2. RT-PCR and real-time qualitative RT-PCR (qRT- PCR)

Total RNA was extracted from tissues or cells using a Nucleo Spin RNA II kit (Macherey-Nagel, Düren, Gemany). The RNA was converted into complementary DNA using High-Capacity cDNA Reverse Transcription Kit (Invitrogen, USA). RT-PCR was performed with GoTaq^® Green Master Mix (Promega, USA). qRT- PCR was performed with the Power SYBR green master mix (Applied Biosystems, UK) using the 7500 real-time PCR system (Applied Biosystems, UK). The primers are shown in Supplemental Table 1.

2.3. Immunofluorescence

The brain tissues were incubated in a fixative solution (4% paraformaldehyde in phosphate-buffered saline) for 24 hours, followed by incubation in 30% sucrose solution overnight. The tissues were embedded in OCT medium and snap frozen. From each brain, 6-μm coronal slices were collected from five different pre-determined depths (with at least 3 slices at each depth), all together covering 600 μm throughout the region of interest of the hippocampus as described (Baruch et al., 2016). The sections were incubated with primary antibodies (Abs). The following primary Abs were used: mouse anti-Aβ (clone 6E10), mouse or rabbit anti-GFAP (Biolegend, USA), anti-Iba1 (Invitrogen), and anti-IFNγ (Bioss, USA), The dilution or concentration of each antibody is shown in Supplemental Table 2. After washing, the sections were incubated with fluorochrome-conjugated secondary Ab, counterstained with 4’, 6’-diamidino-2-phenylindole (DAPI) (Invitrogen) and observed under a Nikon A1R Spectral Confocal microscope (Nikon, Kanagawa, Japan) or Keyence microscope (KEYENCE, USA). The information for microscope settings is provided in Supplemental Tables 3 and 4. The percentage of positive areas was measured using ImageJ software (NIH, USA). Plaque numbers were also quantified from 10 slides in each brain, and they are presented as the average number of plaques per brain region. Prior to quantification, samples were coded to mask the identity of the mice, and plaque burden was quantified by an observer blinded to the identity of the treatment groups.

The thymus sections were stained with rabbit anti-mouse K5 polyclonal Ab (Covance Research Products, Denver, PA), and rat anti-mouse K8 monoclonal Ab (Throma I mAb, raised by P. Brulet and R. Kemler and obtained from the Developmental Studies Hybridoma Bank, University of Iowa, IA), followed by AlexaFluor-488- or 546-conjugated goat anti-rabbit IgG, or goat anti-rat IgG Ab (Invitrogen), and observed as described (Lai and Jin, 2009).

2.4. Immunohistochemical staining

The brain tissues were fixed and embedded as above. From each brain, 40-μm coronal slices were collected, incubated with 0.3% hydrogen peroxide for 10 minutes to remove endogenous peroxidase activity. After washing with PBS twice for 5 minutes, sections were blocked with normal goat serum and then incubated with the primary antibody [biotin-conjugated phospho-Tau (Ser202, Thr205) (AT8, Invitrogen)] overnight at 4 °C. The sections were then incubated with streptavidin conjugated horseradish peroxidase (Invitrogen) at room temperature for 1 hour, followed by development in fresh DAB substrate solution (Vector, USA). The sections were mounted on microscopic slides and observed under a Keyence microscope.

2.5. Flow cytometry analysis

Brains were dissected, and different brain regions were removed under a dissecting microscope. CP tissues were isolated from the lateral, third and fourth ventricles of the brain; and single-cell suspension was prepared as described (Baruch et al., 2016; Baruch et al., 2015). Spleens were mashed with the plunger of a syringe and treated with ammonium chloride potassium-lysing buffer to remove erythrocytes. A single-cell suspension of tissues was stained with fluorochrome-conjugated Abs directly or indirectly as described (Lai et al., 2012). For intracellular staining, the cells were first permeabilized with a BD Cytofix/Cytoperm solution for 20 minutes at 4°C. The following Abs were used: CD4, CD8, EpCAM1, CD45, Ly51, CD11b, F4/80, IFNγ, Ly6c, SRA1, IL-17, and IL-10 (BioLegend, San Diego, CA, or ThermoFisher Scientific). The samples were analyzed on an LSRFortessa X-20 Cell Analyzer (BD Biosciences). Compensation controls were set up with single staining for each of the antibodies. Debris were excluded, and lymphocytes included, using a forward scatter area (FSC-A) versus side scatter area (SSC-A) gate. Single cells were selected on an FSC-A versus FSC-W plot; 100,000 and 10,000 events were acquired for each sample of brain and spleen, respectively. Data analysis was performed using FlowJo software (Ashland, OR).

2.6. ELISA assay for Anti-Aβ₄₀ and anti-Aβ₄₂ Abs, cytokines, and soluble and insoluble Aβ proteins

At the end of studies, mice were euthanized by CO₂ inhalation, and blood was collected from the heart. Following centrifugation, serum was collected. Aβ₄₀ or Aβ₄₂ (Anaspec, USA) was coated on 96-well microplates overnight at 4 °C, then blocked with blocking buffer (2%BSA+5% goat serum in PBS) for 2 hours at room temperature. The serum samples were diluted into 1:1000 and added to the plates and incubated for 2 hours at room temperature. After washing, HRP-conjugated goat anti-mouse IgG Ab (Biolegend) was added to the plates and incubated for 1 hour. The reaction was developed by TMB substrate (Thermo Scientific, USA) and stopped with 0.1 N HCl. The microplate was read at 450 nm under a microplate reader (BMG LABTECH, CLARIOstar Plus, USA). The Ab concentrations were calculated using a standard curve generated with known concentrations of anti-Aβ Ab.

For the measures of cytokines and soluble and insoluble Aβ proteins, brain parenchyma was dissected, snap-frozen and kept at −70 °C until homogenization. The samples were homogenized, and then centrifuged at 5,000g for 8 min at 4 °C. The amount of IL-1β, IL-6, TNFα, Aβ_1-40 and Aβ_1-42 in the supernatant was then determined. The soluble Aβ fraction probably contained monomeric and oligomeric Aβ (Iaccarino et al., 2016). After the soluble fraction of Aβ_1-40 and Aβ_1-42 in the supernatant was collected, the pellets were further resuspended with guanidine hydrochloride (5 M) to extract the insoluble Aβ fraction as described (Baruch et al., 2015; Iaccarino et al., 2016). The final levels of IL-1β, IL-6, TNFα, Aβ_1-40 and Aβ_1-42 were determined and normalized to total protein content.

2.7. Western blot

Brain (cortex and hippocampus regions) were homogenized in RIPA lysis buffer containing protease inhibitors as described (Liu et al., 2021). Equal amounts of sample protein were separated on Tris-HCl polyacrylamide SDS gels and transferred to polyvinylidene fluoride membranes. Blots were placed in blocking solution with 10% non-fat milk in PBS with 0.05% Tween-20 (PBS-T) for 1 h, followed by incubation with various primary Abs with 5% nonfat milk in PBS-T for 3 h at room temperature or overnight at 4°C. Primary Abs include phospho- Tau (Ser202, Thr205) (AT8, Invitrogen), Tau (HT7, Invitrogen), β-actin, BACE1 (Santa Cruz), ADAM10 (Santa Cruz), ADAM17 (Millipore Sigma), Presenilin-1 (PS-1) (Novus Biologicals) and APP [to detect both full-length APP and C-terminal fragments (CTFs); CT695, ThermoFisher]. Blots were washed with PBS-T, incubated with horseradish peroxidase-conjugated secondary Abs, and then developed with Super Signal^® West Pico chemiluminescent Substrate (Thermo Scientific).

2.8. B cell ELISpot assay

MultiScreen-IP plates (Millipore, Billerica, MA) were washed with 70% ethanol, rinsed three times with PBS, coated with Aβ₄₀ (4 μg/ml) or Aβ₄₂ (4 μg/ml) at 4°C overnight. The plates were blocked with blocking buffer (2% BSA in RPMI medium). 1 × 10⁵ splenocytes were added into the plates and incubated for 48 hours. The plates were washed 6 times with 0.25% Tween 20 (Sigma, USA) in PBS, incubated with HRP-conjugated goat anti-mouse IgG (H+L) Ab (Biolegend) for 1 hour, developed with a AEC Peroxidase Substrate Kit (Vector), and counted for ELISpots (Zhao et al., 2020).

2.9. In vitro and in vivo amyloid phagocytosis assays

For in vitro amyloid phagocytosis assay, HiLyte Fluor 647 Beta-Amyloid (1-42) (Anaspec) was resuspended in Tris/EDTA (pH 8.2) at 20 mM and then incubated in the dark for 3 days at 37°C to promote aggregation. Macrophages in suspension were pretreated in low serum medium as described (Liu et al., 2021). The HiLyte Fluor 647 Beta-Amyloid was added and incubated for 5 hours. Cells were stained with macrophage markers; amyloid phagocytosis by the macrophages was determined by flow cytometry (Liu et al., 2021). To determine whether IFNγ from T cells plays a role in enhanced amyloid phagocytosis, macrophages were cultured with CD4⁺ T cells isolated from rFOXN1 protein- or control protein-treated 3XTg-AD mice in the presence of neutralizing anti- IFNγ or isotype antibody (rabbit IgG) (Invitrogen) for 4 hours. The HiLyte Fluor 647 Beta-Amyloid was then added to the cultures and incubated for an additional 5 hours. Amyloid phagocytosis by the macrophages was determined by flow cytometry as above.

For in vivo amyloid phagocytosis assay, mice were injected intraperitoneally with 10 mg kg⁻¹ methoxy-X04 (R&D Systems) in 50% DMSO/50% NaCl (0.9%), pH 12 (Heneka et al., 2013). Three hours later, the brains were harvested; single-cell suspension was stained with Abs against CD11b and CD45 and analyzed for methoxy-X04 fluorescence in CD45^hiCD11b⁺ monocyte-derived macrophages by flow cytometry.

2.10. Barnes maze

Barnes Maze was conducted as previously described (Clinton et al., 2007; Martinez-Coria et al., 2015). Briefly, each mouse was placed in the center of the maze and subjected to aversive stimuli. Mice were trained 4 training trials per day for 5 days, and a probe test was performed 24 h after the last training trial. The latency and number of errors were recorded for the training tail and probe test.

2.11. Novel object recognition (NOR) test

A NOR test was conducted as previously described (Liu et al., 2021; Zhao et al., 2020). Briefly, mice were trained by allowing them to explore two identical objects placed at opposite ends of the arena for 10 min. 24 h later, mice were tested with one copy of the familiar object and one novel object of similar dimensions for 3 min. The time spent on exploring and sniffing of each object was recorded manually in real-time. Both the Barnes Maze and NOR test were performed by an observer blinded to the identity of the treatment groups. The NOR index represents the percentage of time mice spent exploring the novel object.

2.12. Locomotor activity

Open fields are composed of two or four 12″×12″×12″ transparent acrylic chambers aligned horizontally or vertically, with each chamber housing a single mouse to be analyzed. A 10 mins video for each test was recorded in the MOV format with a resolution of 1920×1080 pixels at 30 frames per second, and the locomotor activity was analyzed by a MATLAB script (version R2018b) as previously described (Zhang et al., 2020), running on a PC with Microsoft Windows 10.

2.13. Statistical analysis

For comparing means of two groups, two-tailed Student’s t-test was used. For comparing means of multiple groups, significance was determined using one-way ANOVA with Dunnett test, two-way ANOVA with Tukey test, or three-way ANOVA with Tukey test by using SPSS29 software (IBM Corp., Armonk, NY). Differences with P < 0.05 were considered statistically significant. Statistical analysis data including N numbers, T/F values, degrees of freedom and exact p values are shown in figure legends or Supplemental Tables 5 and 6.

3. Results

3.1. rFOXN1-treated AD mice have an improved cognitive performance

We have previously reported that rejuvenating the aged thymic microenvironment by mESC-TEPs can attenuate AD pathology and improve cognitive performance (Zhao et al., 2020). We have also demonstrated that intrathymic (i.t.) injection of rFOXN1 protein into mice that have undergone HSCT results in an increased number of TECs and thymocytes; and 40 μg is the optimal dose (Song et al., 2016). We therefore used this condition to determine whether rFOXN1 could affect cognitive performance and AD pathology. Animal models of AD, such as 3XTg and APP/PS1 mice, have been invaluable for studying pathogenesis and evaluating potential therapeutic interventions. 3XTg-AD mice aged 12 months, an age of advanced cerebral pathology (Oddo et al., 2003), were injected i.t. with 40 μg rFOXN1 protein at days 0, 15, and 30 (Figure 1A). Equal amount of rMyoD protein was used as a negative control (Song et al., 2016). Age-matched non-transgenic WT mice that were injected i.t. with 40 μg rFOXN1 or control rMyoD protein as in Figure 1A were also used as controls. Two months later, the mice were evaluated for spatial learning and memory. Barnes maze, a hippocampal-dependent spatial task (Cohen and Stackman, 2015; Maras et al., 2014), is one of the most sensitive tests for detecting cognitive deficits in 3XTg-AD mice (Stover et al., 2015). We found that rFOXN1 protein-treated 3XTg-AD mice had greater Barnes maze learning curves than control protein-treated mice (Figure 1B). rFOXN1 protein-treated 3XTg-AD mice also had decreased latency to find the target zone during the probe trial conducted 24 hours after the final training session (Figure 1C), indicating an improved memory performance. In addition, the number of errors committed in rFOXN1 protein-treated 3XTg-AD mice was significantly reduced (Figure 1D). In contrast, rFOXN1 and control protein-treated WT mice were not significantly different in the spatial learning and memory (Figure 1B-D).

Figure 1. rFOXN1 protein treatment improves cognitive performance in 3XTg-AD mice.

3XTg-AD mice (12-month-old) were injected i.t. with 40 μg rFOXN1 or control rMyoD protein at days 0, 15, and 30. Age-matched non-transgenic WT mice that were injected i.t. with 40 μg rFOXN1 or control rMyoD protein at days 0, 15, and 30 were also used as controls. Two months later, the mice were evaluated for cognitive performance by Barnes Maze and Object Recognition tests. (A) Timeline for the injection and behavioral tests. (B) The escape latency during the training period, and (C and D) the escape latency and the number of errors committed during the probe trial are shown. (E) NOR index was determined as the time spent interacting with the novel object divided by the total time of exploration during the testing phase. (F) Travel distance (m) was measured for locomotory activity analysis. (B) Overall groups comparisons (n=24/group for the escape latency during the training period) were carried out by using three-way ANOVA (treatment × genotype× training days) followed by post-hoc Tukey’s HSD test. (C-F) Overall groups comparisons (n=24/group for probe trail of Barnes Maze, NOR test, and locomotory activity) were carried out by using two-way ANOVA (treatment × genotype) followed by post-hoc Tukey’s HSD test. The data were from three independent experiments and expressed as mean ± SD or mean ± SEM for escape latency during the training period. *p < 0.05 versus control MyoD protein group.

The NOR test can be used to examine learning and memory in rodents based on their spontaneous tendency to have more interactions with a novel than with a familiar object (Arsenault et al., 2011); it is a more cortically-dependent novel object recognition preference task (Cohen and Stackman, 2015; Maras et al., 2014). Consistence with the results with the Barnes maze task, rFOXN1 protein-treated 3XTg-AD mice (but not WT mice) performed significantly better than control protein-treated mice (Figure 1E).

It has been reported that 3xTg-AD mice at 12-month-old have decreased locomotor activity compared with their age- and gender-matched non-transgenic control mice (Filali et al., 2012; Pairojana et al., 2021; Rothman et al., 2012) although increased locomotor activity in 3xTg-AD mice was also observed (Pietropaolo et al., 2008). We found that 3xTg-AD have decreased locomotor activity compared with WT mice, consistent with the reports (Filali et al., 2012; Pairojana et al., 2021; Rothman et al., 2012). rFOXN1 treatment increased locomotor activity in AD mice (Figure 1F).

We also examined the effect of rFOXN1 protein treatment in another AD model, APP/PS1 mice, which develop Aβ-plaque pathology at a more advanced age than do 3XTg-AD mice. Similarly, rFOXN1-treated APP/PS1 mice had greater Barnes maze learning curves, decreased latency to find the target zone during the probe trial, decreased number of errors committed, performed better in the NOR test, and increased locomotor activity, as compared to control protein-treated APP/PS1 mice (Supplemental Figure 1A-F). Taken together, our results suggest that rFOXN1 protein treatment leads to improved spatial learning and memory in 3xTg- and APP/PS1 AD mice but not in WT mice.

3.2. rFOXN1-treated AD mice have reduced AD pathology

We then determined whether rFOXN1 protein treatment results in improved AD pathology. After the Barnes maze and NOR tests, the brains were harvested and subjected to immunofluorescence analysis. rFOXN1 protein treated 3XTg-AD mice had a reduced cerebral Aβ plaque load in the hippocampus (Hp) including dentate gyrus (DG) and cerebral cortex (Figure 2A-C and Supplemental Figure 2A), areas showing robust Aβ-plaque pathology in AD mice. Astrogliosis, as assessed by glial fibrillary acid protein (GFAP) immunoreactivity, was also reduced in rFOXN1 protein-treated mice, as compared to control protein-treated mice (Figure 2A, D and Supplemental Figure 2A). In addition, we analyzed the mRNA and protein expression levels of inflammation cytokines including IL-1bβ, IL-6 and TNFα in the brain by qRT-PCR and ELISA. The levels of these cytokines between rFOXN1 and control protein-treated 3xTg-AD mice were not significantly different (Supplemental Figure 3).

Figure 2. rFOXN1 protein treatment attenuates AD pathology.

(A-C) 3XTg-AD mice (12-month-old) were injected i.t. with rFOXN1, or control protein as in Figure 1. Two and a half months later, the mice were evaluated for (A-D) brain pathology, and (E) soluble and insoluble Aβ levels. (A-D) The brains were immunostained for Aβ (in red), GFAP (in green) and DAPI nuclear staining. Mean Aβ area and plaque numbers, as well as GFAP area in the hippocampal DG and cortex were measured. (A) Representative immunofluorescent images, and (B-D) quantitative analysis of Aβ and GFAP. (E) Levels of soluble and insoluble Aβ_1-40 and Aβ_1-42 in the cerebral brain parenchyma of the mice were quantified by ELISA. (F) The levels of p-tau and total tau in the total protein lysates were analyzed. (left) Representative Western blot images, (right) quantification of Western blot data that was normalized to β-actin and control-group. (G-I) 14-month-old APP/PS1 mice were injected i.t. with rFOXN1, or control protein. Two months later, the mice were evaluated for brain pathology. (G) Representative immunofluorescent images, and (H) quantitative analysis of Aβ from the immunofluorescence. (I) Levels of soluble and insoluble Aβ_1-40 and Aβ_1-42 in the cerebral brain parenchyma of the mice were quantified by ELISA. Group comparison (n=18) was carried out by using two-tailed Student’s t-test. The data were from three independent experiments and expressed as mean ± SD. *p < 0.05 versus control MyoD protein group.

Because impaired synaptic plasticity and memory deficits in AD are related to elevated cerebral Aβ_1-40/Aβ_1-42 levels (Shankar et al., 2008), we measured both soluble and insoluble Aβ levels in the AD mice by ELISA. Consistent with the immunofluorescence results, rFOXN1 protein-treated mice had reduced both soluble and insoluble Aβ levels, as compared to control protein-treated mice (Figure 2E).

Studies have shown that cognitive impairments in AD are also closely related to tau pathology in which hyperphosphorylated tau protein accumulated in the brain (Clinton et al., 2010). We then detected phosphorylated (p) tau in the brain lysates using Western blot with antibody AT8 to p-tau at Ser202/Thr205 residues (Joly-Amado et al., 2020; Stojakovic et al., 2021). rFOXN1-treated mice had reduced p-tau (but not total tau), as compared to control-treated mice (Figure 2F). The reduced p-tau level in rFOXN1-treated mice was confirmed by immunohistochemical staining (Supplemental Figure 2B). Similar results were also obtained when antibody against other p-tau (Thr231) was used (data not shown). The results suggest that rFOXN1 also reduces tau pathology.

rFOXN1-treated APP/PS1 mice also had a reduced hippocampal Aβ plaque load (Figure 2G, H), and reduced both soluble and insoluble Aβ levels (Figure 2I), as compared to control protein-treated mice. Taken together, our results suggest that rFOXN1 protein treatment results in clearance of Aβ plaques and p-tau, and reversal of cognitive decline in 3XTg- and APP/PS1 AD mice.

3.3. Reduced Aβ load in rFOXN1-treated mice is not a result of decreased amyloid precursor protein (APP) expression or Aβ Production

We then determined whether the reduced Aβ plaque load in rFOXN1-treated AD mice were due to reduced Aβ production and processing. We first examined the protein levels of APP and Presenilin-1 (PS-1) by Western blot (Marsh et al., 2016). No differences in expression levels of APP and PS-1 as well as the APP cleavage products β-CTF and α-CTF were observed between rFOXN1- and control protein-treated 3XTg-AD mice (Supplemental Figure 4A, B). Furthermore, the expression levels of the key APP processing genes and proteins Adam10, Adam17, Bace1, and Bace2 were not significantly different between rFOXN1- and control protein-treated 3XTg-AD mice (Supplemental Figure 4C-E). The results suggest that the observed reduction in Aβ plaque load in rFOXN1-treated AD mice is not due to decreased APP production and/or processing, but rather likely mediated via increased Aβ clearance.

3.4. Administration of rFOXN1 protein rejuvenates the aged thymic microenvironment, and increases the number of TECs and thymocytes in AD mice

The young thymus has a distinct corticomedullary junction; in contrast, the aged thymus has abnormalities of thymic architecture characterized by the loss of corticomedullary distinction and areas of absent TECs (Min et al., 2007; Takeoka et al., 1996). We investigated whether rFOXN1 protein treatment could rejuvenate the aged thymic architecture. Frozen thymic sections were prepared and subjected to H&E staining and immunofluorescent staining with anti-K8 and K5 Abs. As shown in Figure 3A and B, control protein treated thymus in 12-month-old WT and 3XTg-AD had a typical aged thymic architecture. In contrast, rFOXN1 treated aged thymus had a young thymic architecture with a distinct corticomedullary junction and increased areas of K8⁺K5⁻ cortical TECs (cTECs) and K8⁻K5⁺ medullary TECs (mTECs).

Figure 3. rFOXN1 rejuvenates the aged thymic architecture, increases the number of TECs and thymocytes in 3XTg-AD mice.

WT and 3XTg-AD mice (12-month-old) were injected i.t. with 40 μg rFOXN1 or control rMyoD protein as in Figure 1. Two and a half months later, the thymi were harvested. (A, B) Frozen thymic sections were prepared and subjected to (A) H&E staining (for both WT and AD mice) and (B) immunofluorescent staining with anti-K8 and K5 antibodies (for AD mice). (C-F) The thymi were analyzed by flow cytometry. (C, D) The percentages and numbers of total TECs (CD45⁻EpCAM⁺ MHC II⁺), cTECs (CD45⁻EpCAM⁺MHC II⁺Ly51⁺), and mTECs (CD45⁻EpCAM⁺MHC II⁺Ly51⁻) are shown. (C) Representative flow cytometric profiles showing the percentage of EpCAM⁺MHC II⁺ TECs in CD45⁻ thymic stromal cells and Ly51⁺ cTEC and Ly51⁻ mTECs in CD45⁻EpCAM⁺MHC II⁺ TECs from the AD mice. (D) The numbers of total TECs, cTECs and mTECs. (E, F) The percentages and numbers of CD45⁺ total thymocytes and thymocyte subsets. (E) Representative flow cytometric profiles showing the percentage of CD4 and CD8 DN, DP, and SP thymocytes from the AD mice. (F) The numbers of CD4 and CD8 DP, CD4 SP and CD8 SP thymocytes. (D and F) Overall groups comparisons (n=18) were carried out by using two-way ANOVA (treatment × genotype) followed by post-hoc Tukey’s HSD test. The data were from three independent experiments and expressed as mean ± SD. *p < 0.05 versus control MyoD protein group. ^#p < 0.05, control MyoD protein-treated WT control group versus control MyoD protein-treated 3XTg-AD group.

We then used flow cytometry to analyze the number of TECs. TECs can also be divided into cTECs and mTECs based on the expression of Ly51. We found that the numbers of total TECs (CD45⁻EpCAM⁺), cTECs (CD45⁻EpCAM⁺MHC II⁺Ly51⁺) and mTECs (CD45⁻EpCAM⁺MHC II⁺Ly51⁻) in control-treated 3XTg-AD mice were lower than those in age- and sex-matched WT mice (Figure 3C, D), suggesting that AD pathology reduces the number of TECs. Consistent with the immunofluorescent results, rFOXN1 increased the numbers of total TECs, cTECs and mTECs in 3XTg-AD mice, to the levels comparable to those in rFOXN1-treated WT mice (Figure 3C, D). mTECs can be further divided into mTECs^lo and mTECs^hi subsets based on the expression level of MHC II. rFOXN1 treatment increased the number of both mTECs^lo and mTECs^hi (Figure 3D).

We also analyzed thymocytes and thymocyte subsets. Thymocytes can be divided into four major subsets: CD4 and CD8 double negative (DN), double positive (DP), CD4 single positive (SP), and CD8 SP thymocytes. The numbers of thymocytes in control-treated 3XTg-AD mice were also lower than those in age- and sex-matched WT mice (Figure 3E, F). It has been reported that aged thymus has increased percentage of DN thymocytes. rFOXN1 protein decreased the percentages of DN thymocytes but increased the percentages of CD4 and CD8 DP and SP thymocytes in aged 3XTg-AD and WT mice (Figure 3E and data not shown). Because the rFOXN1 protein increased the number of total thymocytes, it increased the numbers of each thymocytes subsets including DN thymocytes (Figure 3F).

Like the data in 3XTg-AD mice, rFOXN1-treated aged APP/PS1 thymus had a young thymic architecture with a distinct corticomedullary junction, increased numbers of total TECs, cTECs and mTECs as well as each thymocytes subsets (Supplemental Figure 5). Taken together, our results suggest that AD mice have reduced numbers of TECs and thymocytes and that rFOXN1 treatment rejuvenates the aged thymic architecture and increases the number of TECs and thymocytes in AD mice.

3.5. rFOXN1-treated AD mice have increased numbers of peripheral T cells and enhanced choroid plexus (CP) activity

We then analyzed T cells in the periphery in the mice and found that the numbers of splenic CD4⁺ and CD8⁺ T cells in 3XTg-AD mice were also lower than those in WT mice (Figure 4A-D), in agreement with other reports (Kapadia et al., 2018; Marchese et al., 2014). We have previously demonstrated that the enhanced thymopoiesis in rFOXN1-treated mice results in increased number of T cells in the spleen of the mice that have undergone HSCT (Song et al., 2016). Consistent with the previous results, rFOXN1 increased the numbers of CD4⁺CD8⁻ and CD4⁻CD8⁺ splenic T cells in 3XTg-AD mice to the levels comparable to those in rFOXN1-treated WT mice (Figure 4A-D) but decreased the number of CD4⁻CD8⁻ splenocytes (data not shown). rFOXN1 treatment also led to increased numbers of CD4⁺CD8⁻ and CD4⁻CD8⁺ splenic T cells in APP/PS1 mice (Supplemental Figure 6).

Figure 4. rFOXN1 protein increases peripheral T cells and CP activity in 3XTg-AD mice.

3XTg-AD and WT mice were injected i.t. with rFOXN1, or control protein as in Figure 1. Two and half months later, the spleen and CP were harvested. (A-F) The splenocytes were analyzed for the percentages of (A-D) CD4 and CD8 T cells, and (E, F) IFN-γ-producing CD4 T cells by flow cytometry. (G-I) The CP was analyzed for (G, H) the percentage of IFN-γ-producing CD4 T cells by flow cytometry, (I) the mRNA expression levels of IFNγ, CCL2, CCL25, CXCL12, and ICAM1 by qRT-PCR., and (J) IFNγ protein expression level by immunofluorescence. The expression levels of the genes in control protein-treated mice were defined as 1. Overall groups comparisons (B, D, F, H and I, n=18) were carried out by using two-way ANOVA (treatment × genotype) followed by post-hoc Tukey’s HSD test. The data were from three independent experiments and expressed as mean ± SD. *p < 0.05 versus control MyoD protein group. ^#p < 0.05, control MyoD protein-treated WT control group versus control MyoD protein-treated 3XTg-AD group.

We and other have reported that reduced AD pathology involves an IFNγ-dependent immunological response (Baruch et al., 2016; Liu et al., 2021; Zhao et al., 2020). We therefore analyzed IFNγ-producing T cells in the spleen. The percentage of IFNγ-producing CD4 T cells in rFOXN1-treated 3XTg- and APP/PS1 AD mice was significantly higher than that in control protein-treated mice (Figure 4E, F and Supplemental Figure 6C). However, the percentage of IFNγ-producing CD4 T cells in rFOXN1- and control protein-treated WT mice was not significantly different (Figure 4E, F). We also analyzed IL-17a and IL-10-expressing CD4 cells in the spleen; there were not significant differences in the percentage of these cells between rFOXN1- and control treated AD and WT mice (Supplemental Figure 7 and data not shown).

The CP, the epithelial layer that forms the blood-CSF barrier, is a selective gateway for leukocyte entry to the brain (Baruch et al., 2015). It has been shown that AD mice have a defect in the CP gateway, as indicated by significantly lower levels of immune cells trafficking molecule expression in the CP (Baruch et al., 2015). In contrast, IFNγ signaling enhances the expression of the trafficking molecules (Kunis et al., 2013). We examined IFNγ availability at the CP and found that rFOXN1-treated 3XTg- and APP/PS1AD mice (but not WT mice) had a significantly higher percentage of IFNγ-producing CD4⁺ immune cells in the CP than control protein-treated mice as determined by flow cytometry (Figure 4G, H and Supplemental Figure 6D). qRT-PCR and immunofluorescence analyses confirmed higher IFNγ mRNA and protein expression levels in this compartment of rFOXN1-treated AD mice (Figure 4I, J and Supplemental Figure 6E).

Because increased IFNγ availability can enhance CP activity (Baruch et al., 2016; Baruch et al., 2015), we examined the mRNA expression levels of a panel of immune cell trafficking molecules by qRT-PCR. We found that the expression levels of the genes for chemokine C-C motif ligand 2 (CCL2), CCL25, and C-X-C motif chemokine 12 (CXCL12), and intercellular adhesion molecule 1 (ICAM1) in the CP of rFOXN1-treated AD mice (but not WT mice) were significantly higher than those in control protein-treated mice (Figure 4I). Collectively, our results suggest that rFOXN1 treatment increases the percentage of T cells, especially IFNγ-producing T cells in the spleen and the CP, and enhanced CP activity.

3.6. rFOXN1-treated AD mice have an increased number of anti-Aβ Ab-producing B cells in the spleen and increased level of anti-Aβ Ab in the serum and the brain

It is well known that T cells can help B cell immune responses. We then determined whether the increased number of T cells in rFOXN1-treated AD mice led to increased production of anti-Aβ Ab-producing B cells. We used Aβ₄₀ and Aβ₄₂ as antigens for an ELISpot assay to measure anti-Aβ Ab-producing B cells in the spleen. The number of anti-Aβ Ab-producing B cells in rFOXN1-treated 3XTg-AD mice was higher than that in control protein-treated mice (Figure 5A, B).

Figure 5. rFOXN1 protein treatment results in an increased number of anti-Aβ Ab-producing B cells in the spleen and increased levels of anti-Aβ Abs in the serum and the brain of 3XTg-AD mice.

3XTg-AD mice were injected i.t. with rFOXN1, or control protein as in Figure 1. Two and half months later, the spleen, serum, and brain were harvested. (A, B) The splenocytes were placed on culture plates coated with Aβ₄₀ (4 μg/ml) or Aβ₄₂ (4 μg/ml), Anti-Aβ Ab-producing B cells were measured by an ELISpot assay. (C, D) The levels of anti-Aβ₄₀ and anti-Aβ₄₂ Abs in (C) the serum and (D) the brain were measured by ELISA. Group comparison (n=18) was carried out by using two-tailed Student’s t-test. Data are shown as mean ± SD. *p < 0.05 versus control protein MyoD group.

We also analyzed the levels of anti-Aβ Abs in the serum. Consistent with the ELISpot results, the levels of both anti-Aβ₄₀ and anti-Aβ₄₂ Abs in the serum of rFOXN1-treated 3XTg- and APP/PS1 AD mice were higher than those in control protein-treated mice (Figure 5C and Supplemental Figure 8A). Furthermore, the levels of both anti-Aβ₄₀ and anti-Aβ₄₂ Abs in the brain of rFOXN1-treated 3XTg- and APP/PS1 AD mice were also higher than those in control-treated mice (Figure 5D and Supplemental Figure 8B). The results suggest that increased T cell number in rFOXN1-treated mice help to generate anti-Aβ Ab-producing B cells that secret anti-Aβ Abs into the serum and the brain.

3.7. rFOXN1-treated AD mice have an increased number of monocyte-derived macrophages in the brain and enhanced phagocytic capacity

We and others have reported that increased CP activity can result in recruitment of monocyte-derived macrophages to the brain to attenuate AD pathology (Baruch et al., 2016; Baruch et al., 2015; Liu et al., 2021; Zhao et al., 2020). Since rFOXN1 protein treatment increases CP activity, we analyzed whether there was an increased percentage of monocyte-derived macrophages in the brain. CD45^hiCD11b⁺ cells represent a myeloid population enriched with CNS-infiltrating monocyte-derived macrophages in the brain (Baruch et al., 2016; Baruch et al., 2015). rFOXN1 protein-treated 3XTg-AD mice (but not WT mice) had an elevated percentage of CD45^hiCD11b⁺ cells in the brain, as compared to control protein-treated mice (Figure 6A, B). The CD45^hiCD11b⁺ cells in the rFOXN1 protein-treated AD mice (but not WT mice) also had a higher percentage of lymphocyte antigen 6c (Ly6C) positive cells in CD45^hiCD11b⁺ cells than control-treated mice (Figure 6A, C and data not shown). In addition, the CD45^hiCD11b⁺ cells in the rFOXN1 protein-treated AD mice expressed a higher level of scavenger receptor A (SRA1) (Figure 6D). We also analyzed CD45^loCD11b⁺ microglia in the brain by flow cytometry and Iba1⁺ microglia by immunofluorescence; the percentage of microglia between rFOXN1 and control protein-treated 3XTg-AD mice were not significantly different (Supplemental Figure 9A-C). However, the number of astrocytes in rFOXN1-treated AD mice decreased (Supplemental Figure 9B, C), consistent with the results in Figure 2A, D.

Figure 6. rFOXN1 protein treatment increases the proportion and the function of macrophages in 3XTg-AD mice.

3XTg-AD and WT mice were injected with rFOXN1, or control protein as in Figure 1. Two and a half months later, the brain and the spleen were harvested. (A-D) The brain was analyzed for the (A, B) percentage of CD45^hiCD11b⁺ cells, (A, C) the percentage of Ly6C⁺ in CD45^hiCD11b⁺ cells, and (D) the expression of SRA1 by CD45^hiCD11b⁺ cells. (E) Quantification of Aβ phagocytosis of CD45^hiCD11b⁺ cells by flow cytometry 3 hours after intraperitoneal injection of methoxy-XO4. (F-I) The splenocytes were analyzed for (F, G) the percentage of F4/80⁺ macrophages, and (H, I) the macrophages were analyzed for the ab ility to phagocytose HiLyte Fluor 647-Aβ₄₂. (J) F4/80⁺ Macrophages were isolated from untreated 3XTg-AD mice and cultured with CD4⁺ T cells from rFOXN1 protein- or control protein-treated 3XTg-AD mice in the absence or presence of neutralizing anti-IFNγ or isotype antibody. The macrophages were then analyzed for the ability to phagocytose HiLyte Fluor 647-Aβ₄₂. Overall groups comparisons (B, G, I, n=18) were carried out by using two-way ANOVA (treatment × genotype) followed by post-hoc Tukey’s HSD test, J (n=18) was carried out by using one-way ANOVA, and group comparison (C, D, E, n=18) was carried out by using two-tailed Student’s t-test. The data were from three independent experiments and expressed as mean ± SD. *p < 0.05 versus control MyoD protein group. ^#p < 0.05, rFOXN1 protein treated 3XTg-AD mice in the absence of neutralizing anti-IFNγ versus isotype antibody.

To ascertain rFOXN1 treatment results in enhanced migration of monocyte-derived macrophages from periphery into the brain, we used BM chimeric mice. 3xTg-AD mice aged 11 months were irradiated with head shielding as described (Rosenzweig et al., 2019; Shechter et al., 2009) and injected with BM from GFP⁺ mice. One months later, the mice were injected i.t. with 40 μg rFOXN1 or control protein as in Figure 1. Two and a half months later, the peripheral blood was analyzed for GFP⁺ donor-derived cells. Similar levels of GFP⁺ leukocytes were observed in both groups (Supplemental Figure 10A), suggesting that adequate blood chimerism (>85%) had been obtained and rFOXN1 did not affect the chimerism. We then analyzed GFP⁺CD45^hiCD11b⁺ cells in the brain. rFOXN1-treated chimeric mice had higher percentages of donor-derived CD45^hiCD11b⁺ cells in the brain (Supplemental Figure 10B), confirming that the increase was due to enhanced migration of monocyte-derived macrophages from periphery into the brain.

We then analyzed the phagocytic capacity of the monocyte-derived macrophages in the brain. A fluorescent derivative of Congo red known as methoxy-X04, which crosses the blood brain barrier and has nanomolar binding affinity for Aβ, has been used in in vivo phagocytic assay (Heneka et al., 2013). We injected methoxy-X04 into the rFOXN1 or control protein-treated 3XTg-AD mice. Three hours later, mice were analyzed for methoxyX04 fluorescence in CD45^hiCD11b⁺ monocyte-derived macrophages by flow cytometry. A three-fold increase in Aβ phagocytosis was found in rFOXN1-treated 3XTg-AD mice, as compared to control protein-treated mice (Figure 6E). Total methoxy-X04 fluorescence in the brain homogenates between the rFOXN1 and control-treated mice was not significantly different (Supplemental Figure 11), which excluded the possibility that the difference in phagocytosis was due to varying cerebral levels of methoxy-X04. Together, our results suggest that rFOXN1 treatment not only increases the migration of monocyte-derived macrophages into the brain, but also enhances their Aβ phagocytosis.

We also analyzed macrophages in the spleen and found the percentage of F4/80⁺ macrophages in rFOXN1 protein-treated AD mice were higher than that in control protein-treated mice (Figure 6F, G). Furthermore, F4/80⁺ macrophages in rFOXN1 protein-treated AD mice were more able to phagocytose Aβ₄₂ than those in control protein-treated mice (Figure 6H, I). Although rFOXN1 protein-treated WT mice also had a higher percentage of F4/80⁺ macrophages than control protein-treated WT mice, the macrophages in both WT groups had a similar ability to phagocytose Aβ₄₂ (Figure 6G, I).

Since we have shown that the percentages of IFNγ-producing CD4 T cells in the spleen and the CP of rFOXN1-treated AD mice (but not in WT mice) were significantly higher than those in control protein-treated AD mice (Figure 4E-H), we determined whether IFNγ from the T cells plays a role in the enhanced amyloid phagocytosis of macrophages. F4/80⁺ Macrophages from untreated 3XTg-AD mice were cultured with CD4⁺ T cells isolated from rFOXN1 protein- or control protein-treated 3XTg-AD mice in the absence or presence of a neutralizing anti-IFNγ or isotype antibody. Aβ₄₂ phagocytosis by the macrophages was then analyzed. As shown in Figure 6J, CD4⁺ T cells from rFOXN1 protein-treated AD mice significantly enhanced the ability of the macrophages to phagocytose Aβ₄₂ and the anti-IFNγ antibody, at least partly, neutralized the enhanced phagocytosis.

The increased number of CD45^hiCD11b⁺ monocyte-derived macrophages in the brain and the enhanced methoxyX04 fluorescence in the macrophages was also observed in rFOXN1 protein-treated APP/PS1 mice (Supplemental Figure 12A, B). Furthermore, these mice also had higher percentages of macrophages in the spleen and the macrophages were more able to phagocytose Aβ₄₂, as compared those in control protein-treated APP/PS1 mice (Supplemental Figure 12C, D).

4. Discussion

In this study, we have demonstrated that administration of rFOXN1 into 3XTg- and APP/PS1 AD mice rejuvenates the aged thymic architecture and increases the number of TECs, resulting in increased number of T cells, especially IFNγ-producing T cells, in the spleen and the CP. Consequently, the rFOXN1-treated AD mice have an increased production of anti-Aβ Abs in the serum and the brain, and increased migration of monocyte-derived macrophages from periphery into the brain, as well as enhanced ability of the macrophages to phagocytose Aβ, which likely contribute to the improved cognitive performance and reduced AD pathology in the rFOXN1-treated AD mice.

It has been reported that sporadic AD patients primarily accumulate Aβ because of impaired clearance (Tarasoff-Conway et al., 2015). Our rFOXN1-treated AD mice do not have significant changes in the expression levels of the APP, PS-1, and the key APP processing genes as compared to control protein-treated AD mice. The results suggest that the reduced Aβ plaque load in rFOXN1-treated AD mice is not due to decreased APP production and/or processing, but rather likely mediated via increased Aβ clearance.

Aging has a deleterious effect on the immune system (Lynch et al., 2009; Swain and Nikolich-Zugich, 2009; Taub and Longo, 2005). Of all age-related changes in the immune system, atrophy of the thymus is among the most dramatic, ubiquitous, and recognizable (Lynch et al., 2009; Swain and Nikolich-Zugich, 2009; Taub and Longo, 2005). The thymus is a specified immune organ that provides an inductive environment for the generation of T cells. However, the thymus undergoes a profound atrophy with age, resulting in fewer T cells in older adults. This reduction has direct etiological linkages with many diseases (Lynch et al., 2009; Swain and Nikolich-Zugich, 2009; Taub and Longo, 2005), including acceleration of the development and progression of AD (Marsh et al., 2016). It has been reported that AD mice have reduced CD4 and CD8 T cells (Kapadia et al., 2018; Marchese et al., 2014). AD mice also have premature immunosenescence (Giménez-Llort et al., 2012; Mate et al., 2014) that contributes to the AD pathology (Richartz et al., 2005), which may be caused by premature thymic involution. We show here that 3XTg- and APP/PS1 AD mice have reduced numbers of peripheral T cells, which are consistent with these reports (Giménez-Llort et al., 2012; Kapadia et al., 2018; Marchese et al., 2014; Mate et al., 2014). Our results also suggest that the reduced TECs play a critical role in reduced T cells in AD mice because rFOXN1 treatment increases the number of TECs, leading to increased numbers of thymocytes and splenic T cells. We have previously shown that transplantation of embryonic stem cell-derived thymic epithelial progenitors rejuvenates the aged thymic microenvironment, leading to enhanced generation of T cells and attenuated AD pathology (Zhao et al., 2020). Our current results for the effect of rFOXN1 protein on the aged thymic microenvironment and AD pathology further support our notion that rejuvenating the aged thymic microenvironment can ameliorate AD pathology.

We have shown that rFOXN1-treated AD mice have an increased number of anti-Aβ Ab-producing B cells in the spleen and increased level of anti-Aβ Ab in the serum and the brain. It has been reported that CD4 T cells help B cells to generate anti-Aβ Abs to neutralize the toxin of Aβ (Parker, 1993). It is possible that increased generation of CD4 T cells leads to increased generation of anti-Aβ-producing B cells.

T cells also play an important role in adaptive–innate immunity crosstalk and help in CNS repair (Marsh et al., 2016; Rosenzweig et al., 2019). Monocytes and macrophages are the central cells of the innate immune system. Bone marrow-derived monocytes can differentiate into macrophages (Theriault et al., 2015). It has been reported that the recruitment of monocyte-derived macrophages to the brain (Schwartz and Baruch, 2014) can lead to reduced AD pathology (Baruch et al., 2016; Baruch et al., 2015; Simard et al., 2006; Zhao et al., 2020). We have shown that administration of rFOXN1 increases the proportion of monocyte-derived macrophages in the brain, which is related to increased expression of IFNγ, CCL2, CCL25, CXCL12, and ICAM1 in the CP. It will be important to determine whether the increased expression of IFNγ and the immune cell trafficking molecules are responsible for the enhanced migration of the monocyte-derived macrophages in the brain. Blood-brain barrier leakage have been observed in AD mice and patients (Nehra et al., 2022). It is also possible that rFOXN1 enhances the migration of monocyte-derived macrophages in the AD mice by affecting blood-brain barrier leakage, which remains to be investigated.

In addition to the increased migration of macrophages into the brain, their phagocytic capacity is also enhanced in rFOXN1-treated AD mice as shown by the in vitro and in vivo amyloid phagocytosis assays (Figure 6). This is related to increased expression of SRA1 and Ly6C on the macrophages. SRA1 is one of the principal receptors involved in Aβ clearance by immune cells (El Khoury et al., 1996; Zuroff et al., 2017). SRA1 can bind both soluble and fibrillar Aβ₄₂ in vitro (Frenkel et al., 2013; Yang et al., 2011) and facilitate its subsequent uptake. In contrast, lack of functional SRA1 in macrophages reduces Aβ₄₂ clearance (Frenkel et al., 2013). It has also been reported that Ly6C is related to myeloid cell neuroprotection in AD (Naert and Rivest, 2013). It is possible that rFOXN1-treated AD mice have increased number of functional T cells, leading to enhanced macrophage functions including increased expression of SRA1 and Ly6C and enhanced phagocytic ability.

Interestingly, the effects that rFOXN1 treatment leads to increased percentage of IFNγ-producing CD4 T cells in the spleen and the CP, enhanced CP activity, increased migration of monocyte-derived macrophages into the brain and enhanced phagocytic capacity of the macrophages only occur in AD mice but not in WT mice, suggesting that they are AD-specific. It is possible that AD mice have defects in T cell functions and rFOXN1 treatment corrects the defects by increasing the number of functional T cells. Indeed, we have shown that CD4⁺ T cells from rFOXN1 protein-treated AD mice significantly enhanced the ability of macrophages to phagocytose beta-amyloid, as compared to those from control protein-treated AD mice. IFNγ produced by the CD4 T cells plays an important role in the enhanced phagocytosis (Figure 6J). Our results are consistent with the reports that T cells can activate macrophages via releasing IFNγ (Ma et al., 2003).

Since it has been reported that female 3xTg-AD and APP/PS1 mice have more aggressive_Aβ pathology than male mice (Hirata-Fukae et al., 2008; Wang et al., 2003), we used female AD mice in the studies for this paper. However, we have observed a similar trend that rFOXN1 reduces Aβ pathology and regulates immune cells in male 3xTg-AD and APP/PS1 mice (data not shown).

Collectively, we have demonstrated that administration of rFOXN1 protein attenuates AD pathology and improves cognitive performance. rFOXN1 protein not only provides a novel approach to study the role of immune cells and molecules in AD development but also has the potential to provide a novel approach to treat AD patients.

• rFOXN1 protein improves cognitive performance and reduces Aβ plaque load in AD mice

• rFOXN1 increases T cells that regulate other immune cells to clear Aβ in the brain

• rFOXN1 protein has the potential to provide a novel approach to treat AD patients

List of abbreviations:

AD

Alzheimer’s disease

Aβ

amyloid-beta

CNS

central nervous system

TECs

thymic epithelial cells

CP

choroid plexus

APP

amyloid precursor protein

Ab

antibody

qRT-PCR

real-time qualitative RT-PCR

K5

keratin 5

K8

keratin 8

NOR

Novel object recognition

sAβ

soluble Aβ protein

i. t.

intrathymically

WT

wild-type

Ctrl

control protein

DG

dentate gyrus

GFAP

glial fibrillary acid protein

CCL

C-C motif ligand

CXCL

C-X-C motif chemokine

ICAM1

intercellular adhesion molecule 1

SRA1

scavenger receptor A