Animals
Male DBA/1J mice were purchased from Sankyo Labo Service (Tokyo, Japan). They were housed in groups of 4 to 6 and maintained on a light/dark cycle of 12:12 hours with food and water available ad libitum.
Detecting the circumventricular organs using fluorescein isothiocyanate
To visualize the sCVOs of DBA/1J mice, the fluorescein isothiocyanate (FITC) method was used according to a previous report (32). Briefly, 12-week-old mice were anesthetized using isoflurane (3% in 100% O2) and transcardially perfused with the following: first, 0.1 M phosphate-buffered saline (PBS), 5 ml; second, FITC in PBS (0.1 mg/ml), 25 ml; third, PBS, 12.5 ml; and finally, 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB), 40 ml. Dissected brains were postfixed overnight and cryo-protected. Brain blocks were embedded in OCT Compound (Sakura Finetek, Tokyo, Japan) and stored at −80°C. Consecutive coronal sections (20 µm) were obtained throughout brain regions containing the third ventricle and medulla using a cryostat (CM1850; Leica Biosystems, Tokyo, Japan). Every second section was embedded in anti-fading Aqua Poly/Mount (18606; Polysciences, Warrington, PA, USA) onto coverslips.
Collagen-induced arthritis
After acclimation for 1 week, 7-week-old mice were used. According to a previous report, the immunization procedure comprised two intradermal injections at the base of the tail on post-immunization days (PID) 0 and 21 (33,34). In the CIA group (n = 99), a first immunization of bovine type II collagen (200 µg/mouse; Collagen Research Center, Tokyo, Japan) dissolved in 0.1 M acetic acid (4 mg/mL) emulsified in complete Freund’s adjuvant (CFA; Becton Dickinson and Company, Franklin Lakes, NJ, USA) was administered on PID 0, with a booster immunization of bovine type II collagen dissolved in 0.1 M acetic acid emulsified in incomplete Freund’s adjuvant (IFA; Becton Dickinson and Company). In the Freund’s adjuvant group (FA group; n = 12), 0.1 M acetic acid without type II collagen emulsified in CFA (PID 0) and IFA (PID 21) were administered in the same manner. In the saline group (n = 52), an equivalent volume of saline was administered. Immunization was performed in a blinded fashion in the CIA and FA groups, but not the saline group, because of differences in appearance of the saline and emulsions. According to a previous report, arthritis severity was determined using arthritis scores for all four limbs on the following scale: 0, normal; 1, swelling of digits alone or localized swelling of wrist and ankle joints; 2, swelling of both digits and wrist or ankle joints; and 3, swelling of a whole limb (34). “Total arthritis score” defined the sum of the scores for all four limbs. Brain analyses (described below) were performed on PID 21 and 35 and considered to represent the pre-onset and establishment phases, respectively, while those on PID 56 and 84 represented chronic phases. Several mice with ulceration around the anus caused by CFA (CIA, 6 of 99; FA, 1 of 12) were excluded from further analyses. Finally, 93 CIA mice, 11 FA mice, and 52 saline mice were used in this study.
Tissue preparation
For immunohistochemistry, mice under anesthesia were transcardially perfused with PBS followed by 4% PFA in 0.1 M PB. After post-fixation, the brain was cryo-protected. Brain blocks were embedded in OCT compound and stored at −80°C. Sections containing the SFO, OVLT, or AP were obtained using a cryostat at a thickness of 20 µm. For in situ hybridization, mice under anesthesia were transcardially perfused with PBS. The unfixed medulla was dissected and frozen in isopentane on dry ice, and 16 µm coronal sections were obtained using a cryostat.
Immunohistochemistry
Sections were washed in PBS and then incubated in blocking solution containing 1% bovine serum albumin and 0.3% Triton X-100 in PBS for 1 h at room temperature. Subsequently, sections were incubated for 21 h at 4°C with rabbit anti-mouse ionized calcium-binding adaptor molecule-1 (Iba-1) (1:4000; Wako Chemicals, Osaka, Japan), mouse anti-mouse glial fibrillary acidic protein (GFAP) (1:2000, G3893; Sigma–Aldrich, St. Louis, MO, USA), and/or American-hamster anti-mouse CD31 (1:100, 2HB; Developmental Studies of Hybridoma Bank, Iowa University, Iowa City, IA, USA). After rinsing in PBS, sections were incubated for 2 h at room temperature with the following secondary antibodies: Alexa Fluor 488-conjugated goat anti-mouse IgG (1:1000; Thermo Fisher Scientific, Rockford, IL, USA), Alexa Fluor 568-conjugated goat anti-rabbit IgG (1:1000; Thermo Fisher Scientific), and/or Alexa Fluor 647-conjugated goat anti-American-hamster IgG (1:400; Jackson ImmunoResearch, West Grove, PA, USA). Sections were then washed with PBS and incubated with 4′,6-diamidino-2- phenylindole (DAPI) (1 μg/ml; Dojindo, Kumamoto, Japan) for nuclear staining. Slices were embedded in anti-fading Aqua Poly/Mount on coverslips.
Multiplex fluorescent in situ hybridization
Multiplex fluorescent RNAscope (Advanced Cell Diagnosis [ACD], Hayward, CA, USA; Medical & Biological Laboratories, Nagoya, Japan) was performed using probes for Iba-1 (Mm-Aif1, #319141) and IL-1β (Mm-Il1b-C2, #316891C2), in accordance with the manufacturer’s instructions. Briefly, after fixation in 10% neural-buffered formalin at 4°C for 15 min, sections were washed with PBS, incubated in ethanol, and air-dried. Sections were incubated with protease III (diluted 1:1 with PBS) for 30 min. After additional PBS washing, probe hybridization and amplification steps were performed. Iba-1 (Alexa 488) and IL-1β (Atto 550) probe-stained sections were incubated with DAPI and mounted with Aqua Poly/Mount on coverslips.
Image acquisition
All fluorescence images were obtained using laser scanning confocal microscopy (FV1200; Olympus, Tokyo, Japan). Grayscale (16 bit) images were captured with a c-MOS camera (1024 × 1024 pixels, DP80; Olympus) and saved in TIFF format.
Quantitative analysis for immunohistochemistry
All image analyses were performed by a blinded examiner using ImageJ (National Institute of Mental Health, Bethesda, MD, USA). Images for Iba-1 immunosignal were captured with a 20× objective lens to identify microglia cells using the “triangle methods” with the same threshold value for all analyses. The number of microglia was counted using the “analyze particle” function in ImageJ by setting “size (pixel2)” to 75 - infinity and “circularity” to 0.0–1.00. The area for the sCVOs was determined using DAPI staining in each image by identifying areas with high DAPI-positive signal density. After measuring the sCVO area, the ratio of total Iba-1 immunopositive area to sCVO area (in %) and specific number of microglia (in/mm2) were calculated. Values were calculated in duplicate from two sections per animal.
For evaluation of microglial morphology, quantitation was performed on immunostained images using a 40× objective lens. Regions of interest were placed on the four main divisions of the AP based upon GFAP immunostaining, as described previously (Supplementary Figure 1) (10,35). Binary images in each region of interest were acquired using the same threshold algorithm (at least eight regions from two slices per mouse). To extract single cell images, binary images were segmented using the “analyze particle” function by setting “size (pixel2)” as 300-infinity and “circularity” as 0.0–1.00. The following twelve morphological parameters were measured in each cell: perimeter, area, ratio of perimeter to area, ferret length, minimum ferret length, maximum and minimum diameter of approximate ellipse, aspect ratio, (minimum diameter/maximum diameter), ratio of width to height, circularity, roundness, and solidity (Supplementary Figure 2).
Quantitative analysis for multiplex fluorescent in situ hybridization
A blinded examiner performed the following image analyses. First, using the “max entropy” threshold method in ImageJ, separate binary images were created for each IL-1β and Iba-1 mRNA signals obtained after RNAscope processing. The number of puncta for IL-1β and Iba-1 located within DAPI-positive nuclei within the AP were counted. According to the scoring guideline for RNAscope images provided by ACD, IL-1β-positive microglia cells (identified by nuclear Iba-1 and IL-1β mRNA signals [IL-1β+Iba-1+DAPI+]) were classified using the following IL-1β expression scales: “IL-1 βnegIba-1+DAPI+”, no-expression; “IL-1βlowIba-1+DAPI+”, co-existence of 1–3 nuclear puncta; and “IL-1βhighIba-1+DAPI+”, co-existence of 4 or more nuclear puncta. Values from two slices from a single mouse were averaged.
Quantitative real-time polymerase chain reaction
RNA extraction and real-time polymerase chain reaction (PCR) were performed as described previously (34). Briefly, total RNA was extracted from four amputated limbs using a RNeasy Lipid Tissue Mini Kit (Qiagen, Tokyo, Japan). Real-time PCR was performed using an Applied Biosystem StepOnePlus Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA) with Taqman probes and the following primers: IL-1β (Mm01336189_m1), IL-6 (Mm00446190_m1), and Actb (Mm00607939_s1). Expression levels normalized to Actb were analyzed using the ΔΔCT method. mRNA expression levels were represented as values relative to the average of the saline group.
Statistical analysis
Data were expressed as mean ± SEM. All statistical analyses were performed using R (version 3.6.1; the R foundation for Statistical Computing, Vienna, Austria) and EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan)(36). Sample sizes for the experiments on PID 21, 56, and 84 were calculated using expected effect size and variance based on data of Iba-1-immuno-stained area (%) in the AP on PID35. The Kolmogorov–Smirnov test was used as a test of normality. Unpaired t-test (two-sided) was used for comparison between two groups. When the normal distribution was not confirmed, the Mann–Whitney U test was used to compare the mean ranks of two groups. Three groups were compared by one-way analysis of variance (ANOVA) followed by a Bonferroni post-hoc test. Correlation analysis was performed using Spearman’s rank correlation. To classify cells according to morphological parameters, principal component analysis (PCA) and hierarchical clustering analysis (HCA) were used. Frequencies of categorical variables were compared using the chi-square test. Differences were considered significant when the p value was < 0.05.