Animal experiments were approved by the Institutional Animal Care and Use Committee of Gachon University (IACUC number: GIACUC-R2019030) and the Institutional Animal Care and Use Committee of Yonsei University (IACUC number: 2018-0145), and the experiments were compliant with the Guide for the Care and Use of Laboratory Animals.
Reaching puberty, mammalian females are considered sexually mature after menarche, which occurs as early as 4 weeks after birth in mice [23, 24]. Reproductively senescent perimenopause females (between 9 and 12 months in mice) undergo a drastic decrease in sex hormone levels, which significantly contributes to the initiation of programmed aging [25,26,27]; menopause marks the final and irreversible termination of the female reproductive life span . Survivorship drops off markedly for mice over 24 months of age and can be considered very old . Based on these dispositions, female mice at 5 weeks were defined as young and two aged ones at 12 and 25 months as old and very old, respectively, in this study.
In late July 2019, C57BL6 female mice at 3 weeks, 12 months, and 25 months were obtained from the Laboratory Animal Resource Center, Daejeon, Korea, maintained in a specific pathogen-free (SPF) facility under a 12-h light/dark cycle. Mice were acclimated to the laboratory animal facility at Gachon University for 2 weeks before starting the experiment. Mice were fed with a normal Chow diet. Each individual mouse was tail-tagged for identification and randomly assigned to the treatment groups, which allows for matching different types of data including microbial structure, RNA sequencing, and aging-associated phenotypes.
Dbn1Het C57BL/6 mice (16-week-old) used for validation of Dbn1 for its involvement in skin properties were available from Korea Mouse Phenotyping Center (KMPC). Dbn1WT and Dbn1Het mice were maintained by crossbreeding Dbn1Het; complete loss of the Dbn1 gene (Dbn1-/-) resulted in embryonic lethality. Four weeks after birth, offspring were separated from the cage housing the mother and genotyped using the following primers: 5′–GTCCTTCCTCTTGGTCATTCCC and 5′–TGGAGAAACCAGGGAGATGTTG (for wild type), 5′–GCTACCATTACCAGTTGGTCTGGTGTC and 5′–CAGAGCCCAAGACTAATACACCC (for knockout).
Fecal microbiota transplantation
To prepare donor samples for fecal microbiota transplantation (FMT), we collected fresh feces after 2 weeks of acclimation by gently rubbing the belly from young (Y; 5 weeks: adolescence), old (O; 12 months: perimenopause), and very old (VO; 25 months: senescence) C57BL6 female mice, respectively, and maintained them under the strict anaerobic condition throughout the process. The feces were pooled by groups and mixed with PBS supplemented with 10% glycerol (30 mg feces/mL PBS; ca. 5 × 108 colony-forming units/mL) and thoroughly homogenized via vortexing. After sedimentation for 10 min, the supernatant was filtered through a 40-μm cell strainer and the homogenous aliquots were stored in a refrigerator at −80 °C.
Approximately 100 μL of the microbial suspension was transferred into each C57BL6 female mouse by oral gavage twice a week for a total of 8 weeks using autoclaved feeding needles. Superscript was used to describe the FMT donor. Thus, if feces from the young mice (Y) are given to 12-month-old or 25-month-old groups, they are designated as OldY or vOldY, respectively. As a control (C), 100 μL of PBS with 10% glycerol was administrated to old (OldC) or very old (vOldC) mice. In the same manner, feces from old or very old mice (O or VO) were respectively given to 12- or 25-month-old groups, resulting in OldO or vOldVO mice. All the recipient mice were housed under the SPF condition and weighed twice a week during the FMT timeline. Examination of passive avoidance, grip strength, skin properties, and visual abilities was performed 3 days after the last FMT. Then, all mice were euthanized and sacrificed for the measurement of physiological changes.
To generate microbiota-depleted old (xOld) mice by clearing out pre-existing gut microbiota, drinking water containing antibiotics (1 g/L ampicillin, 0.5 g/L vancomycin, 1 g/L metronidazole, 1 g/L neomycin, and 2.5 mg/L amphotericin B) was given to mice ad libitum during 7 days and withdrawn 24 h before FMT. When we tried the bacterial clearance in very old mice to make xvOld, they failed to tolerate the antibiotic treatment and died.
Fecal microbiota analysis
DNA was extracted using a Fast DNA SPIN kit for fecal samples (MP Biomedicals, Irvine, CA, USA), according to the manufacturer’s instructions. The V3-V4 region of the 16S ribosomal RNA (rRNA) gene was sequenced on an Illumina MiSeq platform by Chun Laboratory (Seoul, Korea) using sequence-specific primers (337F: CCTACGGGA(N)GGCWGCAG, 806R: GACTACHVGGGTM(A)TCTAAT). Raw sequencing data were analyzed using the QIIME2 pipeline (version 2017.10). The q2-demux plugin with DADA was used to de-multiplex and trim the adaptor sequences, and sequence variants were clustered into amplicon sequence variants (ASVs). The phylogenetic tree was constructed using q2-alignment and q2-phylogeny plugins. Sequence variants were taxonomically assigned using the SILVA reference database . The feature table and phylogenetic distances were imported into R for downstream analysis.
Alpha diversity indices including the number of observed ASVs, Chao1, and Shannon’s diversity were calculated using phyloseq and vegan R packages. Linear discriminant analysis effect size (LEfSe) was used to identify microbial biomarkers using the factorial Kruskal-Wallis test (P < 0.05) ; the threshold of logarithmic LDA score was 2.0. Spearman’s correlation coefficient was calculated for bacterial species that have more than 0.5% relative abundance in 12-month-old recipient mice using the psych R package.
Passive avoidance test
To examine the cognitive response of mice, a passive avoidance test was performed. A mouse was placed in a light chamber for 30 s and allowed to move to a dark chamber upon gate opening. The gate was immediately closed after the mouse moved to the dark chamber, and an electric foot shock (0.5 mA) was administered for 2 s. The time between the gate opening and entering of the mouse into the dark chamber was recorded as a training time. Then, the mouse was returned to the home cage. After 24 h, the mouse was returned to the light chamber for 30 s, and then, the gate of the dark chamber was opened. The time for the mouse to enter the dark chamber after the gate opening was recorded as a retention time.
Measurement of grip strength
Forelimb grip strength test was performed using GSM Grip-Strength Meter (Ugo Basile, 47200). The mouse was placed on an acrylic panel and the peak grip force of the mouse on the metal wire of the apparatus was measured in grams. The peak force was averaged from three independent trials.
Measurement of skin properties
The day before measuring skin hydration phenotypes, mouse dorsal hair was carefully shaved. Skin hydration was measured using MoistureMeterSC (Delfin) which registers a probe to the skin surface until skin hydration is stabilized. The capacitance is expressed in arbitrary units (AU). Transepidermal water loss (TEWL) was measured using VapoMeter (Delfin) that registers a probe to the skin surface until TEWL is stabilized (less than 1 min). The result is expressed in g/hm2.
Spatial frequency thresholds (i.e., visual acuity) were assessed by optokinetic nystagmus (OKN) using a virtual optokinetic system (OptoMotry, Cerebral Mechanics, Medicine Hat, Alberta, Canada). A video camera at the ceiling of the device records and transfers images to the connected computer. The clockwise movement of the drift grid tracked the movement of the left eye, and the counterclockwise movement tracked the reaction of the right eye of the mouse. The experimenter judged whether the head and body of the mouse were tracking the direction of the drift grid rotation. If tracking was unclear or absent, the process was repeated. The maximum spatial frequency capable of driving head tracking was determined.
For the intraocular pressure (IOP) test, mice were anesthetized using an intraperitoneal injection of xylazine (10 mg/kg; Rompun®, Bayer Animal Health) and zolazepam and tiletamine (30 mg/kg; Zoletil 50®, Virbac, Carros, France). IOP was measured using a rebound tonometer (Icare® TONOLAB tonometer, Colonial Medical Supply, Franconia, NH, USA), according to the manufacturer’s instructions. One trial result was obtained after six consecutive measurements, and the mean of consecutive trials was used for analyses.
Electroretinogram (ERG) analysis was performed using Micron Ganzfeld ERG (Phoenix Research Labs, Pleasanton, USA). The mouse was dark-adapted at least 12 h before the experiment for scotopic testing (rod cell response). After anesthesia, the pupil dilated as previously described. Once the pupil was fully dilated, we applied hypromellose 2.5% (Goniovisc®) and inserted the electrodes. ERG was recorded using Micron Ganzfeld ERG, according to the manufacturer’s instructions. Scotopic ERG was obtained with increasing flash intensity at a range of −1.7 log cd/s/m2 to 1.9 log cd/s/m2. Mice were light-adapted for 15 min prior to the cone cell response experiment. Photopic ERG was performed with increasing flash intensity at a range of −0.5 log cd/s/m2 to 4.1 log cd/s/m2. Values were based on the average of 10 responses to light stimuli. The implicit times of rod and cone cell response were determined.
For blood sample collection, the three groups of FMT recipients were euthanized using CO2 gas, and blood was directly obtained from the heart using 26G 1 mL syringe. To isolate serum from the total blood, blood samples were placed in the MiniCollect® tube (Greiner Bio-One, Kremsmünster, Austria) and centrifuged at 1 × 104 rpm for 5 min at 25 °C. The levels of aspartate aminotransferase (GOT), alanine aminotransferase (GPT), creatinine phosphokinase (CPK), alkaline phosphatase (ALP), blood urea nitrogen (BUN), amylase (AMYL), creatinine (CRE), albumin (ALB), triglyceride (TG), total bilirubin (TBIL), total cholesterol (TCHO), r-glutamyltransferase (GGT), lactate dehydrogenase (LDH), high-density lipoprotein-cholesterol (HDLC), and creatine kinase-mb (CKMB) in the serum samples (10 μL each) were measured using DRI-chem 4000i (Fuji, Minato, Japan) automated clinical chemistry analyzer.
The mouse tissues were fixed with ice-cold 4% paraformaldehyde and embedded in paraffin. The tissues were cut into 5-μm thick sections; deparaffinized with xylene three times for 20 min each, 100% EtOH three times for 10 min each, 90% EtOH twice for 10 min each, and 75% EtOH for 10 min; and stained with hematoxylin and eosin (H&E). Slides containing the stained tissue were dehydrated and mounted on Shandon Synthetic Mount (Thermo, Inc., Waltham, MA, USA). The diameter of the myofiber in the hindlimb skeletal muscle was determined using digital images of the sections using the ToupView program (http://www.touptek.com).
Immunohistochemical and immunofluorescence staining
For immunohistochemistry, the tissues were fixed with ice-cold 4% paraformaldehyde in PBS and mounted on paraffin blocks. Samples were cut into 3 μm sections, deparaffinized, and rehydrated in PBS. Antigens were then retrieved for 15 min at high pressure in Target Retrieval Solution (Dako, Santa Clara, CA, USA). Then, the specimens were chilled on ice for 1 h, washed with PBS three times for 5 min each and blocked with 3% H2O2 in PBS for 30 min to eliminate the endogenous peroxidase. The slides were washed with PBS, blocked for 2 h at 25 °C with Serum-Free Protein Block (Dako), probed at 4 °C overnight with the primary antibodies (1/1000 dilution), stained for 30 min with horseradish peroxidase-conjugated anti-mouse (Dako) or anti-rabbit IgG (Dako) secondary antibody, and developed with Liquid DAB+ Substrate Chromogen System (Dako). Finally, the specimens were counterstained with Mayer’s hematoxylin (Dako) and mounted on Shandon Synthetic Mount (Thermo).
For immunofluorescence, tissue sample preparation and blocking were performed in the same manner as described in immunohistochemistry. Then, the slides were probed at 4 °C overnight with the primary antibodies (1/1000 dilution), labeled for 2 h with Cy3-conjugated goat anti-rabbit IgG (Thermo) or A488 donkey anti-mouse IgG (Thermo), and stained with DAPI (Sigma, Burlington, MA, USA) for 15 min. Finally, the slides were mounted on ProLong Gold antifade reagent (Thermo), imaged on an Axio Imager M2 fluorescence microscope (Zeiss, Oberkochen, Land Baden-Württemberg, Germany), and analyzed using Zeiss Zen Blue Edition. The following primary antibodies were used for staining: anti-Ki67 (Abcam, Cambridge, UK), anti-DBN1 (Novus, Denver, CO, USA), anti-ITGB4 (Abcam), and anti-KRT10 (Abcam).
Multiplex cytokine and chemokine bead array assays
Multiplex bead array assays for serum cytokine quantification were performed using (#740446; BioLegend, San Diego, CA, USA) according to the manufacturer’s instructions. Serum samples were analyzed via a CytoFLEX platform (Beckman Coulter, Brea, CA, USA).
Immune cell preparation
Single-cell suspensions of the spleen, lymph nodes, and liver were prepared by mechanical homogenization of the organs and passing through 40-μm or 70-μm cell strainer (BD Biosciences, San Jose, CA, USA). For the spleen, erythrocytes were removed using ACK lysing buffer (Thermo). Liver cells were purified by density gradient centrifugation on a 67% and 44% Percoll gradient (GE Healthcare, Chicago, IL, USA).
Cells were seeded in a 96-well plate and stained with fluorochrome-conjugated antibodies for 20 min at 4 °C in the dark. Cells were subsequently washed with PBS containing 2% fetal bovine serum. The following antibodies were used: CD8 (53-6.7; BD Biosciences), CD62L (MEL-14; BD Biosciences), CCR7 (4B12; eBioscience, San Diego, CA, USA), CD103 (2E7; eBioscience), CD25 (PC61; BioLegend), CD45.2 (104; BioLegend), CD44 (IM7; BioLegend), CD69 (H1.2F3; BioLegend), and CD4 (RM4-5; BioLegend). Cells were stained with LIVE/DEAD® Fixable Blue Dead Cell Stain (Invitrogen, Carlsbad, CA, USA) to discriminate dead cells. All stained samples were acquired using CytoFLEX LX instrument (Beckman Coulter), and the data were analyzed using the FlowJo software (BD Biosciences).
Total RNA was isolated from the distal colon, quadriceps femoris muscle (quad muscle), and back skin, respectively, using Trizol reagent (Invitrogen). RNA quality was assessed by Agilent 2100 bioanalyzer using the RNA 6000 Nano Chip (Agilent Technologies, Amstelveen, The Netherlands), and RNA quantification was performed using ND-2000 Spectrophotometer (Thermo). For messenger RNA sequencing, a complementary DNA library was constructed using QuantSeq 3’ mRNA-Seq Library Prep Kit (Lexogen, Inc., Austria), according to the manufacturer’s instructions. Briefly, 500 ng of the total RNA was prepared and an oligo-dT primer containing an Illumina-compatible sequence at its 5′ end was hybridized to the RNA and reverse transcribed. After degradation of the RNA template, second strand synthesis was initiated using a random primer containing an Illumina-compatible linker sequence at its 5′ end. The double-stranded DNA library was purified using magnetic beads to remove all reaction components. The library was amplified to obtain the complete adapter sequences required for cluster generation. The finished library was purified to remove PCR components. High-throughput sequencing was performed using single-end 75-bp reads using NextSeq 500 (Illumina, Inc., USA).
QuantSeq 3′ mRNA-Seq reads were aligned using Bowtie2. Bowtie2 indices were either generated from the genome assembly sequence or representative transcript sequences for aligning to the genome and transcriptome. The alignment file was used for assembling transcripts, estimating their abundances, and detecting differential expression of genes. Raw read counts were normalized and differentially expressed genes (DEGs) were identified using the edgeR package (P < 0.01). The DEGs were determined based on read counts from unique and multiple alignments using coverage in Bedtools. Functional annotation of DEGs was performed using DAVID (http://david.abcc.ncifcrf.gov/) and MEDLINE (http://www.ncbi.nlm.nih.gov/) databases.
Complete hierarchical clustering based on Euclidean distance was performed to classify co-occurred species into clusters using the hclust function in the R package, and four clusters were obtained by cutting off the dendrogram tree at height 6. Spearman’s correlation coefficient was calculated to examine the relationship between DEGs and microbial composition. P values were corrected for multiple comparison problems using Benjamini-Hochberg method with a 0.05 cutoff.
Statistical analyses of host phenotype data were performed using Prism software (GraphPad). Raw data were subjected to one-way ANOVA to evaluate statistical significance between at least three groups, and pairwise comparison was conducted using Student’s t test. Means were considered significantly different at P < 0.05. Statistical details, including sample sizes for each experiment, are provided in the relevant figure legends.