Male 3xTg-AD mice [B6129-Psen1tm1MpmTg (APPSwe, tauP30L) 1Lfa/J], which overexpress mutant APP (APPSwe), Psen1 (PS1M146V), and hyperphosphorylated tau (tauP301L), and wild-type mice (B6129SF2/J) were purchased from Jackson Laboratories. The 3xTg-AD mice (Oddo et al. 2003) were generated by co-injecting two distinct transgene constructs encoding human APPSwe and tauP301L (4R/0 N) (controlled by murine Thy1.2 regulatory elements) into single-cell embryos harvested from mutant homozygous PS1M146V knock-in mice. Age matched wild-type mice of mixed genetic background 129/C57BL6 were used as controls. The animals were maintained on a 12-h light/dark cycle in temperature- and humidity-controlled rooms, and food and water were available ad libitum. Animal experiments were conducted in strict accordance with the ARRIVE guidelines and recommendations for the care and use of laboratory animals. All experiments were carried out according to the Directive 2010/63/EU and the Italian law (D.Lgs. 26/2014). The protocols were approved by the Italian Ministry of Health (authorization n. 552/2020-PR).

The TNFSF10-neutralizing monoclonal antibody (purified rat anti-mouse CD253) and the vehicle (purified rat IgG2ακ isotype control) were obtained from BD Biosciences, Franklin Lakes, New Jersey, USA. All other compounds were of the highest commercial grade available.

Thirty 3xTg-AD and fifteen wild-type mice were enrolled at 12 months of age and three study groups were used: (i) wild-type plus vehicle; (ii) 3xTg-AD plus vehicle (purified rat IgG2ακ isotype control; BD Biosciences); and (iii) 3xTg-AD plus TNFSF10-neutralizing antibody (purified rat anti-mouse CD253; BD Biosciences). Animals were simultaneously randomised to the treatments without taking any other variable into account to avoid bias. Animals (fifteen per experimental group) were administered with a TNFSF10-neutralizing antibody (concentration: 0.05 mg/ml; 200 μl/ mouse; intraperitoneally) or vehicle (concentration: 0.05 mg/ml; 200 μl/mouse; intraperitoneally) (mouse weight = 25 ± 5 g), twice a month (Monday at 12 a.m.) and sacrificed via CO2 inhalation or anesthetized and transcardially perfused with ice-cold 0.1 M PBS (pH 7.4) or ice-cold 4% paraformaldehyde (PFA) at 15 months of age, two weeks after the last injection. The treatment with the anti-TNFSF10 monoclonal antibody effectively neutralized TNFSF10 in 3xTg-AD aging mice (see Supplementary Fig. 1).

Brain and spleen samples of 3xTg-AD mice and age-matched wild-type mice were dissected in ice-cold Hank’s balanced salt solution (HBSS: 137 mM NaCl, 5.4 mM KCl, 0.45 mM KH2PO4, 0.34 mM Na2HPO4, 4 mM, NaHCO3, 5 mM glucose; pH 7.4) and then stored at − 80 °C until use. For protein extraction, brain and spleen tissues were homogenized and sonicated in a lysis buffer containing 150 mM NaCl, 50 mM Tris–HCl (pH 7.5), 5 mM EDTA, 1 mM Na3VO4, 30 mM sodium pyrophosphate, 50 mM NaF, 1 mM acid phenyl-methyl-sulphonyl-fluoride, 5 μg/ml aprotinin, 2 μg/ml leupeptin, 1 μg/ml pepstatin, 10% glycerol, and 0.2% TritonTM X-100. After sonication, the homogenates were centrifuged at 14,000 rpm for 10 min at 4 °C and the supernatant was collected. Protein content of the supernatant was quantified according to the Bradford Assay method (Bradford 1976).

Equal amounts of protein (30 μg) of hippocampus and spleen extracts were resolved onto SDS-PAGE gels and transferred onto Hybond-ECL nitrocellulose membranes (10600003, Amersham Life Science, Buckinghamshire, UK). Membranes were blocked at room temperature for 1 h with a blocking solution composed of 5% non-fat dry milk (Bio-Rad Laboratories, Segrate, Italy) in phosphate-buffered saline plus 0.05% Tween-20 (PBS-T) and were then probed at 4 °C overnight with the following appropriate primary antibodies: rabbit anti-TNFSF10 (1:200, ab231265, Abcam, Cambridge, UK) or a mouse anti-CD86 (1:500, sc-28347, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) or a mouse anti-CD206 (1:500, sc-58986, Santa Cruz Biotechnology Inc.). After that, the membranes were washed with PBS-T and incubated with the appropriate horseradish peroxidase-conjugated secondary antibodies (GE Healthcare) for 1 h at room temperature in 5% non-fat dry milk. For immunodetection, the membranes were exposed to film after enhanced chemiluminescence (ECL) (GE Healthcare). β-actin (1:1000, 4967, Cell Signaling Technology, Danvers, MA, USA.) was used as an internal control to validate the right amount of protein loaded on the gels. Densitometric analysis of band intensity was performed with the aid of ImageJ software version 1.53v (developed by NIH, freeware, available online: https://imagej.nih.gov/ij).

Brain tissues were cut with scissors into small pieces on ice and incubated with 0.4 mg ml–1 of collagenase type IV (Gibco, Cat. No. 17104019) for 30 min at 37 °C. After incubation, brain pieces were passed repeatedly through a 19-gauge needle with a syringe to obtain a homogeneous cell suspension, filtered through a 70 μm cell strainer, washed with ice-cold 1 × PBS, and centrifuged at 600 × g for 6 min. The supernatant was aspirated, and 37% Percoll (GE Heathcare, Uppsala, Sweden) was added. The samples were centrifuged at 800 × g for 30 min and the supernatant was discarded, and cell pellet was resuspended into cell staining buffer. Spleen tissues were minced with scissors into a petri dish with HBSS buffer. Excised spleen pieces were transferred and mashed through a 70 μm cell strainer and then washed with PBS to obtain a spleen single-cell suspension (SCS). SCS was centrifuged at 400 × g for 7 min at 4 °C and the supernatant was discarded. The pellet was resuspended in 1 × RBC lysis buffer for 5 min on ice. After washing with ice-cold PBS, the cell suspension was centrifuged at 400 × g for 7 min at 4 °C and the cell pellet was resuspended in PBS. Dead cells were excluded using VioBlue viability dye (Miltenyi Biotec, Bergisch Gladbach, Germany). Frequency of brain and spleen immune cells was analyzed by multicolor FACS analysis using the following fluorochrome-conjugated antibodies: anti-CD25-VioBright FITC (clone REA568, 1:50, 130–120–172, Miltenyi Biotec), or an anti-CD11b-PerCP-Vio700 (clone REA592, 1:50, 130–113–809, Miltenyi Biotec), or an anti-Ly-6C-PE-Vio770 (clone REA796, 1:50, 130–111–918, Miltenyi Biotec), or an anti-CD4-APC-Vio770 (clone REA604, 1:50, 130–119–132, Miltenyi Biotec) or an anti-CD8a-APC-Vio770 (clone REA601, 1:50, 130–120–806, Miltenyi Biotec) or an anti-CD45-VioGreen (clone REA737, 1:50, 130–110–803, Miltenyi Biotec), or an anti-CD279 (PD1)-PE (clone REA802, 1:50, 130–111–953, Miltenyi Biotec), or an anti-P2RY12 (clone S16007D, 1:200, 848003, BD Bioscences). Briefly, 1 × 106 cells were incubated with the antibodies mix for 30 min at RT in the dark and acquired by flow cytometer (Cytomics FC 500, Beckman Coulter). Analysis was performed using CXP Analysis software. Using a sequential gating strategy, the following cell populations were detected within the CD45+ immune cells: proinflammatory monocytes were characterized as CD11b+LY6ChighP2RY12− in the brain and as CD11b+LY6Chigh in the spleen, while exhausted T cells were identified as CD4+PD1+ or CD8+PD1+.

In brain and spleen samples were also evaluated the frequency of Treg cells (CD4+CD25+FoxP3+). For Tregs evaluation, 1 × 106 cells were fixed and permeabilized with FoxP3/transcription factor staining buffer set (cat. No. 00–5523-00, Thermo Fisher Scientific), and intracellular staining with anti-FoxP3-APC (clone REA788, 1:50, 130–111–679, Miltenyi Biotec) was performed according to the manufacturer's instructions. The results were expressed as percentage. To unequivocally identify the several subpopulations of interest, all gates were controlled using a stain that lacks just one of the fluorescent markers (FMO) allowing accurate gating.

Mice were deeply anesthetized and intracardially perfused with ice-cold 4% PFA. Brain and spleen samples were collected and fixed overnight in 10% neutral-buffered formalin (Bio-Optica). After overnight washing, samples were dehydrated in graded ethanol and paraffin-embedded, taking care to preserve their anatomical orientation. Five-micrometer-thick sections were serially cutted, mounted on silanized glass slides and air dried. To remove paraffin, slides were immersed in xylene twice, for 10 min, then rehydrated with graded ethanol, 100%, 95%, 70%, and 50%, twice per 10 min each, and transferred to distilled water. Antigens were retrieved in sodium citrate buffer (10 mM sodium citrate, 0.05% Tween-20, pH 6.0) by microwave for 10 min, followed by rinsing with distilled water. For detection of Aβ plaques (6E10), slides were incubated with 70% formic acid for 20 min. The slides were then washed in PBS containing 0.025% Tween-20 (PBST) twice for 5 min each, blocked in 5% BSA for 1 h at room temperature, in a humid chamber. The following primary antibodies were incubated overnight at 4 °C with BSA 1%: goat anti-Iba1 (1:100, NB100-1028, Novus Biologicals), or a mouse anti-CD86 (1:250, sc-28347, Santa Cruz Biotechnology Inc.), or a mouse anti-CD206 (1:500, sc-58986, Santa Cruz Biotechnology Inc.), or a rabbit anti-CD3 (1:100, ab16669, Abcam) or a mouse anti-FoxP3 (1:100, sc-166212, Santa Cruz Biotechnology Inc.) or a mouse anti-β-amyloid (clone 6E10, 1:200, SIG-39320, Covance, Princeton, NJ, USA) or a mouse anti-p-tau (1:100, sc-32275, Santa Cruz Biotechnology Inc). For immunoreactivity and fluorescence detection, after washing in PBS-T three times for 5 min each, sections were incubated with the appropriate fluorescent-labeled secondary antibodies for 1 h at room temperature in the dark: Alexa Fluor 546 donkey anti-goat (A11056, Thermo Fisher Scientific, Inc.), or Alexa Fluor 488 donkey anti-mouse (A21202, Thermo Fisher Scientific), or Alexa Fluor 488 donkey anti-rabbit (A21206, Thermo Fisher Scientific) or Alexa Fluor 546 donkey anti-mouse (A10036, Thermo Fisher Scientific). Finally, for staining of nuclei and stabilization of fluorescent signals, slides were covered in mounting medium with DAPI (F6057, Fluoroshield; Sigma-Aldrich, Milan, Italy) and secured with a coverslip. Images were acquired using a Zeiss Observer.Z1 microscope equipped with the Apotome.2 acquisition system (Zeiss LSM 700, Jena, Germany) with 5x, 20x, and 40 x objectives and were processed using ImageJ software. To quantify immunohistochemical staining, five sections/sample were analyzed. Mean Fluorescence Intensity (MFI) was used to evaluate the intensity of the immunofluorescence signal. Briefly, after converting the images to 8-bit, inverting to grayscale, and using the appropriate threshold function, the positive stained area in the images was measured (Shihan et al. 2021). Colocalization was quantified with ImageJ software using the Colocalization Colormap plugin (Jaskolski et al. 2005). Fluorescent images were converted to 8-bit, and after setting the auto threshold function, the Colocalization Colormap plugin was used to analyze the percentage of immunoreactivity in each image found at the same pixel. Results are presented as correlation index (Icorr).

Plaque size was measured using the Ellipse tool to quantify the area of individual plaques (μm2).

To assess the extent of positive immunolabelling, indicated as positive area, images were converted to 8-bit grayscale, and the appropriate threshold function was applied. The percentage of the immunolabelled area was then calculated (Christensen and Pike 2020).

Hippocampal lysates were assayed using the ProcartaPlex™ Mouse Th1/Th2/Th9/Th17/Th22/Treg Cytokine Panel, 17plex (EPX170-26087–901, Thermo Fisher Scientific, Vienna, Austria), according to the manufacturer’s instructions. The concentrations of analytes were detected with the Luminex MAGPIX instrument (Luminex Corporation, Austin, TX). Data were analyzed with xPONENT® software (Luminex Corporation, Austin, TX). Any analyte with a concentration outside the linear range was excluded from the analysis.

Statistical analysis was performed with Prism GraphPad version 9.0. After testing normality distribution of the data, an appropriate one-way analysis of variance (ANOVA) test was applied, and post-hoc Tukey’s multiple comparisons test was used to determine statistical significance. Aβ plaques data were analyzed using a two-tailed Student's t-test. Data were represented as means ± S.E.M. Significance was set at a p value < 0.05 or p < 0.01; shown as *, and ** respectively.