Erlotinib may affect short-term memory, protein buildup, and brain inflammation in mouse models of Alzheimer’s disease

All experiments were approved by the institutional biosafety committee (IBC) and performed in accordance with approved animal protocols of the Korea Brain Research Institute (KBRI, approval no. IACUC-19-00049, IACUC-22-00044).

Erlotinib (Cayman Chemical, Cat no: 10483, Ann Arbor, MI, USA) was dissolved in dimethyl sulfoxide (DMSO) or vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) for in vitro or in vivo experiments, respectively. For in vivo experiments, erlotinib was intraperitoneally (i.p.) administered at a dose of 20 mg/kg, which is lower than the tumor-suppressing dose (25 mg/kg to 100 mg/kg (10)) and the anti-apoptosis dose (80 mg/kg and 100 mg/kg (11, 12)). The use of a dose of 20 mg/kg minimized the toxicity of this anti-cancer drug.

3 to 3.5-month-old 5xFAD mice (Stock No. 34848-JAX; B6.Cg-Tg (APPSwFlLon,PSEN1*M146L*L286V)6799Vas/Mmjax; Jackson Laboratory, Bar Harbor, ME) and 3- or 6-month-old PS19 mice (Stock No. 008169; B6;C3-Tg (Prnp-MAPT*P301S)PS19Vle/J; Jackson Laboratory) mice were selected as mouse models of AD (13, 14). In 5xFAD mice, five familial AD mutations under the control of the Thy1 promoter (APP^Swe,Lon,Flo and PS1^M146L,L286V) lead to Aβ overexpression. PS19 mice, a model of tauopathy, express the human P301S mutation (one N-terminal insert and four microtubule binding repeats (1N4R)) under the control of the mouse prion protein promoter. Tail genomic DNA was used for genotyping. Male C57BL6/N mice (3.5 months old, 25–30 g) were purchased from Orient-Bio Company (Gyeonggi-do, South Korea). To minimize the effects of hormones, only male mice were included in the behavioral experiments. The mice were housed at 22 ± 2°C, 50 ± 5% humidity, and a 12-h light/dark cycle in a pathogen-free facility with chow and water ad libitum and were randomly allocated to the erlotinib or vehicle (control) treatment group.

We prepared RIPA buffer-based soluble and insoluble tau protein as previously described with minor modifications (15). Briefly, 6-month-old PS19 mice were perfused with PBS, and the brains were extracted (n = 6-8 per group). The entorhinal cortex and hippocampus were dissected and homogenized in RIPA lysis buffer (Merck Millipore, Billerica, MA, USA) supplemented with a protease and phosphatase inhibitor cocktail (Thermo Scientific, Waltham, MA, USA). The homogenates were incubated at 4°C for 2 h and centrifuged at 20,000 × g at 4°C for 20 min. The supernatant was collected as the RIPA-soluble fraction and stored at −80°C until analysis. The pellet was washed once with 1 M sucrose in RIPA lysis buffer, resuspended in 2% SDS solution (1 mL per gram of tissue), and incubated at room temperature (RT) for 1 h. The suspension was centrifuged at 20,000 × g for 1 min at RT, and the supernatant was collected as the RIPA-insoluble fraction and stored at −80°C until analysis.

To measure the effects of erlotinib on tau pathology and tau kinases, we conducted western blotting of entorhinal cortex/hippocampal tissue from mice treated with erlotinib (20 mg/kg, i.p.) or vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.). The tissues were lysed in RIPA buffer (Merck Millipore, Billerica, MA, USA), and the protein concentration in the lysate was quantified relative to a standard BSA solution. Next, 20 μg of protein was electrophoretically separated on an 8% SDS gel and transferred to a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked with 5% skim milk or 5% BSA at RT for 1 h before incubation overnight with anti-p-EGFR (1:1000, Cell Signaling, Cat No. 2236), anti-AT8 (1:1000, Invitrogen, Cat No. mn1020), anti-p-PHF^Ser396 (1:1000, Invitrogen, Cat No. 44-752G), anti-p-PHF^Ser404 (1:1000, Invitrogen, Cat No. 44-758G), anti-Tau5 (1:1000, Invitrogen, AHB0042), anti-DYRK1A (l:1000, Abcam, Cat No. ab180910), anti-p-GSK3α/β (1:1000, Abcam, Cat No. ab75745), anti-p-CDK5 (1:1000, MybioSource, Cat No. MBS2534755), anti-CDK5 (1:1000, Abcam, ab40773) or anti-GAPDH (l:1000, Proteintech, Cat No. 104941-AP) antibodies at 4°C. Next, the membrane was incubated with HRP-conjugated goat anti-mouse or anti-rabbit IgG (both 1:1000, Enzo Life Sciences, Farmingdale, NY, USA) for 1 h. Detection was realized using ECL Western Blotting Detection Reagent (GE Healthcare, Chicago, IL, USA), and images were acquired and analyzed using Fusion Capt Advance software (Vilber Lourmat).

To investigate the effect of erlotinib on AD pathology, 3-month-old and/or 6-month-old PS19 mice, 3 to 3.5-month-old 5xFAD mice or 3.5-month-old WT mice were intraperitoneally (i.p.) injected with 20 mg/kg erlotinib or vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) daily for 14 or 21 days. At the end of the treatment period, the mice were perfused with PBS and fixed with 4% paraformaldehyde. After storage in 4% paraformaldehyde for 24 h at 4°C, the brain tissue was immersed in 30% sucrose in PBS for 72 h; finally, a cryostat microtome (Leica CM1850, Wetzlar, Germany) was used to obtain 30-μm-thick slices. After blocking in 10% normal goat serum (Vector Laboratories, Burlingame, CA, USA) for 2 h at RT, the slices were immunostained with the following primary antibodies at 4°C overnight: anti-p-EGFR (1:100, Invitrogen, Cat No. 44-788G), anti-AT8 (1:200, Invitrogen, Cat No. mn1020), anti-AT100 (1:200, Invitrogen, Cat No. mn1060), anti-AT180 (1:200, Invitrogen, Cat No. mn1040), anti-NFT (1:200, Abcam, Cat No. ab136407), anti-DYRK1A (l:200, Abcam, Cat No. ab180910), anti-p-GSK3α/β (1:200, Abcam, Cat No. ab75745), anti-p-CDK5 (1:200, LS bio, Cat No. LS-C354604-100), anti-6E10 (1:200, Biolegend, Cat No. 803002), anti-4G8 (1:500, Biolegend, Cat No. 800711), anti-GFAP (1:500, Neuromics, Cat No. RA22101; or Invitrogen, Cat No. 13-0300) or anti-Iba-1 (1:500, Wako, Cat No. 019-19741). After washing with PBST buffer, the slices were incubated for 2 h at RT with Alexa 555- or Alexa 488-conjugated secondary antibodies. Finally, the slices were mounted in Vectashield mounting solution containing DAPI (Vector Labs, Burlingame, CA) on glass slides, and images of the immunostained tissue were obtained by fluorescence microscopy (DMi8, Leica Microsystems, Wetzlar, Germany) and analyzed (ImageJ, US National Institutes of Health, Bethesda, MD, USA).

The Y-maze test was performed to assess the effects of erlotinib on short-term and spatial memory in 3-month-old 5xFAD and 6-month-old PS19 mice as previously described (16). The three arms of the Y maze had dimensions of 35 cm x 7 cm x 15 cm and met at 120° angles. A test consisted of a single mouse freely exploring the Y maze for 5 min. Spontaneous alternations during the period of exploration were recorded by a SMART video camera and manually counted. The alternation percentage was calculated by dividing the number of alternations by the number of alternation triads and multiplying by 100.

The NORT was performed as previously described to evaluate long-term and recognition memory (16). The NORT apparatus consisted of an open-field box with dimensions of 40 cm x 40 cm x 25 cm. The training and testing phases of the NORT were performed at an interval of 24 h. During training, a single mouse explored the box, which contained two identical objects, for 5 min. During testing, the mouse again explored the box for 5 min, but the box contained one familiar object and one novel object. The object locations in the testing phase were counterbalanced among the trials. Odor cues were eliminated between trials by using 70% ethanol to thoroughly clean the box and objects. Each trial was video recorded, and the video was reviewed to manually count the time of exploratory behavior (defined as the mouse pointing its nose toward an object). The novel object preference (%) was calculated from T_Novel (novel object exploration time) and T_Familiar (familiar object exploration time) as follows: [Object preference (%) = T_Novel/(T_Familiar + T_Novel) × 100].

The PAT was performed to analyze aversive memory as described previously (17) with the following modifications. A single mouse was acclimated for 5 min in the passive avoidance chamber (GEMINI Active and Passive Avoidance System, San Diego Instruments, San Diego, CA, USA), which comprised a lit compartment and a dark compartment with a gate between them, and then was returned to its home cage. The next day (day 2, training phase), the mouse was placed in the lit compartment, and after 10 s, the gate opened. Once the mouse entered the dark compartment, the gate closed, and the mouse received an electrical foot shock (3 mA for 3 s). After the foot shock, the mouse was left in the dark compartment for 10 s to enforce an association between the environment and the aversive stimulus and then returned to its home cage. Finally, the mouse was returned to the lit compartment on the following day (day 3, test phase), and after 10 s, the gate was opened. The latency in the lit compartment before the mouse entered the dark compartment was measured with a maximum (cut-off) time of 180 s, and the mouse was returned to its home cage. The mice that did not enter the dark compartment during the training phase (day 2) were excluded from the analysis.

After the behavioral tests, Golgi staining of brain tissues from 3-month-old 5xFAD and 6-month-old PS19 mice was performed using the FD Rapid Golgi Stain kit (FD NeuroTechnologies) as previously described (18, 19). Fresh brains were incubated in solutions A and B for 2 weeks at RT and solution C for 24 h at 4°C before slicing using a vibratome (VT1000S; Leica) at a thickness of 150 μm. The number of dendritic spines with a length greater than 20 μm was counted in bright-field microscopy (63× magnification; Axioplan 2; Zeiss) images of pyramidal neurons in cortical layers V and CA1 using ImageJ (NIH).

To investigate the effects of erlotinib on reactive astrogliosis in an in vitro model of AD, mixed glial cultures (MGCs) were prepared from postnatal day 1–2 PS19 and 5xFAD mice as previously described with modifications (20, 21). Briefly, tail genomic DNA was used to confirm the genotypes of the pups. Then, whole brains from the PS19 or 5xFAD pups were removed, minced and filtered through 70-μm nylon mesh, and cultured as MGCs in low-glucose DMEM with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin in a 5% CO₂ incubator at 37°C for 3 weeks. At day in vitro (DIV) 21, primary astrocyte cells (PACs) were isolated from the MGCs by shaking the MGCs in 75 T flasks at 250 rpm on a rotary shaker for 6 h at RT. After discarding the conditioned culture medium, the cells were dissociated by adding trypsin-EDTA (0.25%) and centrifuged at 2000 rpm for 10 min. The pellet containing PACs was collected and seeded at a density of 7×10⁵ cells/well for experiments. At DIV 26, the PACs were treated with vehicle (1% DMSO) or erlotinib (5 μM) for 24 h and then harvested. Total RNA was extracted from the PACs for further analysis.

To determine whether erlotinib regulates reactive gliosis in an in vitro model of AD, total RNA from erlotinib-treated PACs was reverse transcribed to cDNA using the Superscript cDNA Premix Kit II with oligo (dT) primers (GeNetBio, Chungman, Korea). The synthesized cDNA was then used as the template in real-time qPCR with Fast SYBR Green Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) in a QuantStudio 5 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA). The primer sequences for real-time qPCR are given in Table 1. The cycle threshold (Ct) values were normalized to the value for gapdh, and the fold change was calculated relative to the control.

To determine whether erlotinib regulates protein levels of proinflammatory cytokines in PACs from 5xFAD mice, IL-1β, COX-2, IL-6, and TNF-α protein levels were measured by ELISA. For this experiment, PACs from 5xFAD mice were treated with vehicle (1% DMSO) or erlotinib (5 μM) for 24 h. The supernatant of the cell lysate was collected and stored at -80°C until analysis. The protein levels of proinflammatory cytokines in the supernatant were measured using IL-1β, COX-2, IL-6, and TNF-α ELISA kits (IL-1β ELISA kit: 88-7013-88, Invitrogen, Waltham, MA, USA; COX-2 ELISA kit: DYC4198, R&D Systems, Minneapolis, MN, USA; IL-6 ELISA kit: 88-7064-88, Invitrogen, Waltham, MA, USA; TNF-α ELISA kit: 88-7324-88, Invitrogen, Waltham, MA, USA) according to the manufacturer’s instructions as previously described (22).

GraphPad Prism 7 software (GraphPad Software, San Diego, CA, USA) was used to analyze all data. For comparisons between two groups, Student’s t test or the paired t-test was used. One-way ANOVA was used for multiple comparisons. Repeated-measures two-way ANOVA by Bonferroni’s post hoc analysis was used for PATs, and p < 0.05 indicated a significant difference. Means ± SEM are presented (*p < 0.05, **p < 0.01, ***p < 0.001).

A previous study demonstrated that inhibition of EGFR improves spatial memory in APP/PS1 mice, a mouse model of AD (9). However, the effects of erlotinib on cognitive function have not been fully elucidated in a tau-overexpressing mouse model of AD. Prior to behavioral tests, the effect of erlotinib on EGFR phosphorylation, an on-target of erlotinib, was assessed. For these experiments, 3-month-old PS19 mice were treated with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining was conducted with an anti-p-EGFR antibody. Erlotinib treatment significantly suppressed EGFR phosphorylation in the entorhinal cortex and hippocampus in PS19 mice (Figure 1A).

We then examined whether erlotinib alters EGFR phosphorylation in an aged mouse model of AD. Six-month-old PS19 mice were treated with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 days, and immunofluorescence staining was conducted with an anti-p-EGFR antibody. Erlotinib treatment significantly downregulated EGFR phosphorylation in the entorhinal cortex and hippocampal DG region in 6-month-old PS19 mice (Figure 1B).

Next, we investigated whether erlotinib regulates learning and memory in 6-month-old PS19 mice. For these experiments, 6-month-old PS19 mice were treated with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 days, and behavior experiments were conducted on days 14–19. Erlotinib treatment significantly increased the alternation % in the Y-maze test compared with vehicle treatment (Figure 1C). However, erlotinib administration did not alter recognition memory in the NORT nor associative aversive memory in the PAT compared with vehicle treatment in 6-month-old PS19 mice (Figures 1D, E).

Next, to determine whether erlotinib improves short-term spatial memory by regulating dendritic spine formation, 6-month-old PS19 mice were treated with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 consecutive days, and Golgi staining of brain slices was conducted. Importantly, dendritic spine number in the hippocampal apical oblique (AO) and basal shaft (BS) regions was significantly increased in erlotinib-treated PS19 mice compared with vehicle-treated PS19 mice (Figure 1F), indicating that erlotinib improves hippocampal-dependent spatial memory by promoting hippocampal dendritic spine formation in 6-month-old PS19 mice. Collectively, these results suggest that administration of the EGFR inhibitor erlotinib improves tau-mediated cognitive impairments in PS19 mice by increasing hippocampal spinogenesis.

Since erlotinib improved cognitive function in PS19 mice, we investigated whether erlotinib modulates tau pathology in younger and aged transgenic PS 19 mice. First, 3-month-old PS19 mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining of brain slices was conducted with anti-AT8, anti-AT100, and anti-AT180 antibodies. Erlotinib treatment significantly decreased tau phosphorylation at residues Ser202/Thr205 (AT8) in the entorhinal cortex and hippocampal CA1 regions (Figure 2A). In addition, erlotinib administration significantly decreased tau phosphorylation at residues Thr212/Ser214 (AT100) in the entorhinal cortex but not in the hippocampus (Figure 2B). Furthermore, erlotinib treatment significantly decreased tau phosphorylation at residue Thr231 (AT180) in the entorhinal cortex and hippocampus in 3-month-old PS19 mice (Figure 2C).

To examine the effects of erlotinib on tau hyperphosphorylation in aged PS19 mice, 6-month-old PS19 mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 days, and immunofluorescence staining of brain slices was conducted with anti-AT8, anti-AT100, and anti-AT180 antibodies. Importantly, erlotinib treatment significantly decreased tau phosphorylation at Ser202/Thr205 (AT8) in the hippocampus but not in the entorhinal cortex in 6-month-old PS19 mice (Figure 3A). In addition, administration of erlotinib did not alter tau phosphorylation at Thr212/Ser214 (AT100) and Thr231 (AT180) in the entorhinal cortex and hippocampus in 6-month-old PS19 mice (Figures 3B, C).

To confirm the immunofluorescence staining results, we conducted western blotting of the soluble and insoluble fractions of brain tissues with an anti-AT8 antibody. This analysis showed that erlotinib treatment significantly reduced tau phosphorylation at Ser202/Thr205 (AT8) in the RIPA-soluble fraction but not in the RIPA-insoluble fraction of the entorhinal cortex of 6-month-old PS19 mice (Figure 3D). Additionally, erlotinib administration significantly suppressed tau phosphorylation at Ser202/Thr205 (AT8) in the RIPA-insoluble fraction but not in the RIPA-soluble fraction of the hippocampus in 6-month-old PS19 mice (Figure 3E). These data indicate that erlotinib selectively regulates tau hyperphosphorylation in PS19 mice.

In PS19 mice, tau filament development begins at 6 months of age (23). We therefore examined the effects of erlotinib treatment on PHF and NFT formation in 6-month-old PS19 mice. Six-month-old PS19 mice were treated with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 days, and RIPA-soluble and -insoluble fractions of brain tissues were obtained. The fractions were then analyzed by immunoblotting with anti-PHF-1 antibodies, including antibodies against p-PHF^Ser396 and p-PHF^Ser404. Erlotinib treatment significantly reduced RIPA-soluble p-PHF^Ser396 levels (45– 70 kDa) but did not alter the levels of RIPA-soluble tau fragments (20–35 kDa) and RIPA-insoluble p-PHF^Ser396 in the entorhinal cortex in 6-month-old PS19 mice (Figures 4A, B). In the hippocampus, erlotinib treatment significantly diminished RIPA-insoluble p-PHF^Ser396 levels but did not alter the levels of RIPA-soluble p-PHF^Ser396 and tau fragments (Figures 4C, D). Similar to the results for p-PHF^Ser396, erlotinib treatment significantly decreased RIPA-soluble p-PHF^Ser404 levels in the entorhinal cortex in 6-month-old PS19 mice but not RIPA-soluble tau fragments and RIPA-insoluble p-PHF^Ser404 levels (Figures 4E, F). In the hippocampus, erlotinib treatment did not alter the levels of RIPA-soluble and RIPA-insoluble p-PHF^Ser404 levels in 6-month-old PS19 mice (Figures 4G, H). Furthermore, we found that erlotinib did not alter total Tau levels (detected by Tau5) in the RIPA-insoluble fractions of the entorhinal cortex and hippocampus in 6-month-old PS19 mice (Figure 4I).

We further confirmed these results by performing immunofluorescence staining of brain slices with an anti-NFT antibody that recognizes a phosphorylation-independent epitope in residues 404–441 (human). Erlotinib treatment significantly reduced NFT levels in the entorhinal cortex and hippocampal CA1 and DG regions (Figure 4J). Collectively, these data suggest that the EGFR inhibitor erlotinib alleviates tau aggregation in 6-month-old PS19 mice.

Since erlotinib treatment suppressed tau pathology in 3-month-old and 6-month-old PS19 mice, we analyzed tau kinase levels to examine the mechanisms by which erlotinib regulates tau hyperphosphorylation. Three-month-old PS19 mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining of brain slices was conducted with an anti-DYRK1A antibody. Erlotinib administration significantly reduced DYRK1A fluorescence intensity in the entorhinal cortex and hippocampal CA1 region in 3-month-old PS19 mice but had no effect on DYRK1A fluorescence intensity in the hippocampal DG region (Figure 5A).

We then investigated the effects of erlotinib on tau kinase levels and activity in aged PS19 mice. Six-month-old PS19 mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 days, and western blotting of brain lysates was conducted with an anti-DYRK1A antibody. Erlotinib significantly reduced DYRK1A levels in the entorhinal cortex but not in the hippocampus in 6-month-old PS19 mice (Figures 5B, C).

To confirm these findings, we performed immunofluorescence staining of brain slices, which showed that erlotinib treatment significantly downregulated DYRK1A immunofluorescence intensity in the entorhinal cortex and hippocampal CA1 region compared with vehicle treatment (Figure 5D). In addition, erlotinib administration significantly reduced the level of the tau kinase p-GSK3α/β in the entorhinal cortex but not in the hippocampus in 6-month-old PS19 mice, as assessed by immunoblot analysis and immunofluorescence staining (Figures 5E–G). Moreover, erlotinib treatment significantly suppressed the level of the tau kinase p-CDK5 in the entorhinal cortex in 6-month-old PS19 mice as assessed by immunoblot analysis (Figures 5H, I). However, erlotinib treatment did not alter total CDK5 levels in 6-month-old PS19 mice (Figures 5J, K).

Finally, to further verify our findings, we conducted immunofluorescence staining and found that erlotinib treatment significantly downregulated tau kinase p-CDK5 immunostaining intensity in the entorhinal cortex and hippocampal DG regions in 6-month-old PS19 mice (Figure 5L). These data suggest that erlotinib suppresses tau phosphorylation by downregulating tau kinases in 3-month-old and 6-month-old PS19 mice.

Given the effects of erlotinib on tau-induced cognitive impairment and tau aggregation, we next investigated whether erlotinib alters tau-mediated neuroinflammatory responses. Six-month-old PS19 mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 days, and immunofluorescence staining of brain slices was performed with anti-GFAP (an astrogliosis marker) and anti-Iba-1 (a microgliosis marker) antibodies. Erlotinib treatment did not alter Iba-1 fluorescence intensity, Iba-1-positive cell number, or Iba-1-positive area in the brain in 6-month-old PS19 mice (Figure 6A). However, erlotinib treatment significantly reduced GFAP fluorescence intensity, GFAP-positive cell number, and GFAP-positive area in the entorhinal cortex but not the hippocampus in 6-month-old PS19 mice (Figure 6B). These data indicate that erlotinib downregulates astrogliosis but has no effects on microgliosis in 6-momth-old PS19 mice.

To further analyze the effect of erlotinib on astrocytic activation in vitro, PACs from PS19 mice were treated with vehicle (1% DMSO) or erlotinib (5 μM) for 24 h, and real-time PCR was conducted. Interestingly, we found that erlotinib treatment selectively downregulated il-1β and cox-2 mRNA levels in PACs from PS19 mice compared to vehicle treatment, whereas il-6 and tnf-α mRNA levels were not altered by erlotinib treatment (Figure 6C). Collectively, these data suggest that erlotinib suppresses astrogliosis and proinflammatory responses in 6-month-old PS19 mice and/or PACs from PS19 mice.

After confirming that erlotinib affects cognitive function, tau pathology, and astrogliosis in 3-month-old and 6-month-old PS19 mice, we examined the impact of erlotinib on cognitive function in another model of AD, 5xFAD mice, which overexpress Aβ. Prior to performing behavioral tests, the effect of erlotinib on EGFR phosphorylation, an on-target of erlotinib, was assessed. For this experiment, 3.5-month-old 5xFAD mice were treated with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining of brain slices was conducted with an anti-p-EGFR antibody. Erlotinib treatment significantly suppressed EGFR phosphorylation in the cortex and hippocampus in 3.5-month-old 5xFAD mice (Figure 7A).

We then examined the effect of erlotinib on cognitive function in 3-month-old 5xFAD mice. Three-month-old 5xFAD mice were treated with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 21 days, and the Y-maze test, NORT, and PAT were conducted. Similar to the effects of erlotinib treatment in PS19 mice, treating 3-month-old 5xFAD mice with erlotinib significantly enhanced alternation triads in the Y-maze test but did not alter recognition memory in the NORT nor aversive memory in the PAT compared with vehicle treatment (Figures 7B–D).

To explore the possible underlying mechanisms by which erlotinib improves short-term spatial memory in Aβ-overexpressing AD mice, we investigated the effects of erlotinib treatment on dendritic spine formation in 5xFAD mice. Strikingly, erlotinib treatment significantly increased hippocampal AO and BS dendritic spine numbers compared with vehicle treatment in 3-month-old 5xFAD mice (Figure 7E). These data suggest that erlotinib improves hippocampal-dependent short-term spatial memory by promoting hippocampal dendritic spinogenesis in 3-month-old 5xFAD mice.

We then investigated whether erlotinib regulates Aβ plaque accumulation in 3.5-month-old 5xFAD mice. The mice were injected with vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib daily for 14 days, and immunofluorescence staining of brain slices was conducted with an anti-6E10 antibody that detects residues 1–16 of Aβ. Importantly, erlotinib treatment significantly suppressed 6E10-positive Aβ plaque deposition in the cortex and hippocampus compared with vehicle treatment in 3.5-month-old 5xFAD mice (Figure 7F). To further confirm the effect of erlotinib on Aβ plaque deposition, immunofluorescence staining was conducted with an anti-4G8 antibody detecting residues 18–23 of Aβ. The results showed that erlotinib treatment significantly suppressed 4G8-positive Aβ plaque deposition in the cortex and hippocampus compared with vehicle treatment in 3.5-month-old 5xFAD mice (Supplementary Figure 1). Taken together, these data suggest that erlotinib ameliorates Aβ-mediated cognitive impairments and Aβ pathology in 3- to 3.5-month-old 5xFAD mice.

Since erlotinib ameliorated tauopathy in PS19 mice, we examined the effects of erlotinib on tau hyperphosphorylation in 3.5-month-old 5xFAD mice. The mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining of brain slices was conducted with anti-AT8 and anti-AT100 antibodies. Erlotinib treatment significantly decreased tau hyperphosphorylation at Ser202/Thr205 (AT8) in the cortex and hippocampus in 3.5-month-old 5xFAD mice (Figure 8A). In addition, erlotinib treatment significantly reduced tau hyperphosphorylation at Thr212/Ser214 (AT100) in the hippocampus but not the cortex (Figure 8B).

We then investigated the mechanisms by which erlotinib reduces tau hyperphosphorylation by examining tau kinase levels/activity. To address this, 3.5-month-old 5xFAD mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining of brain slices was conducted with anti-DYRK1A and p-GSK3α/β antibodies. Erlotinib treatment significantly reduced DYRK1A expression in the hippocampal DG region and GSK3α/β phosphorylation in the cortex and hippocampal CA1 region in 3.5-month-old 5xFAD mice (Figures 8C, D). Collectively, these results indicate that erlotinib treatment reduces tau hyperphosphorylation and tau kinase levels/activity in this mouse model of the early phase of AD.

To investigate whether erlotinib affects Aβ-induced microgliosis in 5xFAD mice, 3.5-month-old 5xFAD mice and wild-type (WT) mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining of brain slices was conducted with anti-Iba-1 and anti-GFAP antibodies. Iba-1 fluorescence intensity, Iba-1-positive cell number, and Iba-1-positive area were significantly greater in vehicle-treated 3.5-month-old 5xFAD mice than in vehicle-treated 3.5-month-old WT mice (). However, compared with the corresponding vehicle-treated mice, erlotinib treatment did not alter Iba-1 fluorescence intensity, Iba-1-positive cell number, or Iba-1-positive area in 3.5-month-old WT mice and 5xFAD mice (). 1 1

We then assessed the effects of erlotinib on Aβ-mediated astrogliosis in 5xFAD mice. GFAP fluorescence intensity, GFAP-positive cell number, and GFAP-positive area were significantly greater in vehicle-treated 3.5-month-old 5xFAD mice than in vehicle-treated WT mice (Figures 9A–C). Interestingly, erlotinib treatment significantly decreased GFAP fluorescence intensity in the cortex but not in the hippocampus in 3.5-month-old 5xFAD mice (Figures 9A–C).

To examine whether erlotinib modulates the AD-associated reactive astrocyte population, 3.5-month-old 5xFAD mice and WT mice were administered vehicle (5% DMSO + 10% PEG + 20% Tween80 + 65% D.W.) or erlotinib (20 mg/kg, i.p.) daily for 14 days, and immunofluorescence staining of brain slices was conducted with an anti-CXCL10 antibody to detect reactive astrocytes. Importantly, erlotinib treatment significantly decreased CXCL10 fluorescence intensity in the cortex and hippocampus in 3.5-month-old 5xFAD mice (Figures 9D–F). These data indicate that erlotinib treatment mitigates Aβ-induced astrogliosis but not Aβ-mediated microgliosis in 3.5-month-old 5xFAD mice.

We then analyzed the effect of erlotinib on the reactive state of astrocytes in vitro. PACs from 5xFAD mice were treated with vehicle (1% DMSO) or erlotinib (5 μM) for 24 h. Real-time PCR analysis showed that erlotinib treatment significantly decreased cxcl10 (a marker of reactive astrocytes) and gbp2 (a marker of proinflammatory A1 astrocytes) mRNA levels without altering s100a10 (a marker of anti-inflammatory A2 astrocytes) mRNA levels (Figure 9G). In parallel with these results, erlotinib treatment significantly downregulated proinflammatory cytokine il-1β, cox-2, il-6, and tnf-α mRNA levels in PACs from 5xFAD mice (Figure 9H).

Finally, to examine whether erlotinib alters protein levels of proinflammatory cytokines in PACs from 5xFAD mice, PACs were treated with vehicle (1% DMSO) or erlotinib (5 μM) for 24 h, and ELISA was conducted. Erlotinib treatment significantly decreased proinflammatory cytokine IL-1β, COX-2, IL-6, and TNF-α protein levels in PACs from 5xFAD mice (Figure 9I). Taken together, these data suggest that erlotinib suppresses reactive astrogliosis and proinflammatory cytokine levels in this mouse model of the early phase of AD and/or in PACs from these mice.