Adiponectin is associated with inflammaging and age-related salivary gland lipid accumulation

The proportion of lipocytes in SGs increased with age

We compared SG morphologies in the human parotid gland (PG) and submandibular gland (SMG) tissues of 3 adult (28.6 ± 6.6 y.o. in PG and 43.3 ± 1.5 y.o. in SMG) and aged (82.0 ± 4.3 y.o. in PG and 88.0 ± 4.3 y.o. in SMG) subjects. Hematoxylin and eosin (H&E) stained images of the human PG and SMG are shown in Figure 1A and the positive area of adipocyte, leukocyte infiltration, and acinar cells are quantified in Figure 1B–1D, respectively. Age-related SG change was assessed by determining the proportion of lipid droplets. The histological findings in old SGs demonstrated the presence of more lipocytes, chronic inflammation, fibrosis, and foci of leukocyte infiltration. Adult SGs exhibited a dense packed lobular acinar structure and well-organized ductal alignment, whereas aged SGs exhibited atrophied acinar cells and lipocyte and leukocyte infiltration.

Figure 1. Hematoxylin and eosin (H&E) stained tissue sections showing age-related morphological changes in human parotid and submandibular glands. (A) The morphology of adult and aged salivary glands, bar size: upper (200 um), lower (50 um). (B–D) The quantifications of lipocyte, leukocyte infiltration and acinar area. Results are presented as the mean ± SD and T-test was performed for statistical analysis ^**p < 0.01, ^*p < 0.05.

Age-related SG morphologic changes were associated with apoptosis and oxidative stress

During aging, substantial acinar cell loss occurs in SGs, and lipocytes accumulate in areas of loss. We found significant acinar cell loss and increases in 8-hydroxy-2′-deoxyguanosine (8-OHdG, a critical biomarker of oxidative stress) expression and apoptosis in aged SG tissues (Figure 2A), and more cellular senescence β-galactosidase (SA-β-gal) expression in aged PG and SMG tissues. Quantitative analysis also showed that the proportion of SA-β-gal positive senescent cells increased with age (P < 0.01, Figure 2B). TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) assays demonstrated old SG tissues contained more apoptotic cells than young tissues (P < 0.05, Figure 2C). Furthermore, immunohistochemical (IHC) staining for 8-OHdG was significantly greater in old SG tissues (P < 0.005, Figure 2D).

Figure 2. Cellular senescence in human salivary gland tissue. (A) β-Galactosidase assay, TUNEL assay, and the expression of 8-OHdG. (B) Quantitation of β-galactosidase positive areas. (C) Quantitative analysis of TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) positive areas. (D) Quantitative analysis of 8-OHdG (8-hydroxy-2' -deoxyguanosine) positive areas. Results are presented as the mean ± SD and T-test was performed for statistical analysis ^****p < 0.0001, ^***p < 0.001, ^**p < 0.01.

Differentially expressed genes (DEGs) in young and old SG tissues

To determine whether aging alters gene expressions, we subjected young and old SG tissues to microarray analysis. Fifteen DEGs in PG and 31 in SMG exhibited absolute fold changes of > 2 (P < 0.05, respectively). In aged PG tissues, 7 of the 15 DEGs were upregulated and 8 were downregulated (P < 0.05), and in aged SMG tissues, 21 of the 31 were upregulated and 10 were downregulated. The top 10 DEGs are listed in Table 1. The functions of these DEGs were determined by Gene Ontology analysis and a comprehensive literature review [7–22]. All DEGs (fold changes of > 1.5, p < 0.05) in PG and SMG were listed in Supplementary Figures 1 and 2.

KEGG pathway analysis of DEGs

To explore the molecular mechanism of salivary gland aging, we performed enrichment analysis using Kyoto Encyclopedia of Genes and Genomes (KEGG) database. KEGG is a comprehensive database that integrates genomic, chemical, and systemic functional information better to understand the molecular response networking of coding genes. The data showed that these DEGs of aged PG were enriched in metabolic pathway, MAPK signaling pathway, Oocyte meiosis, Toll-like receptor signaling pathway, And the DEGs of aged SMG were enriched in metabolic pathway, RAS signaling pathway, phagosome, hematopoietic cell lineage, rheumatoid arthritis (Figure 3).

Figure 3. KEGG pathway analysis of DEGs derived from comparison of adult and aged salivary glands.

ADIPOQ was associated with SG aging

ADIPOQ is commonly expressed in submandibular glands, and IHC staining showed it was upregulated in aged PG and SMG (Figure 4A). Quantitative analysis also showed ADIPOQ was significantly upregulated in aged SG tissues (P < 0.05, Figure 4B).

Figure 4. Immunohistochemical (IHC) staining of ADIPOQ in adult and aged human salivary gland tissues. (A) IHC staining of ADIPOQ in adult and aged parotid and submandibular gland tissues (B) Quantitative analysis of ADIPOQ expression in adult and aged parotid and submandibular gland tissues Results are presented as the mean ± SD and T-test was performed for statistical analysis ^*p < 0.05.

ADIPOQ regulated cellular senescence

To examine SG cell senescence, we compared the expressions of SG functional proteins in primary cultured cells isolated from adult and aged healthy SG tissues. Significantly more SA-β-gal positive cells were detected in primary cells cultured from aged SG tissues (P < 0.05, Figure 5A). And the proliferation rate of the aged cells was significantly lower than that of the adult cells as shown in Figure 5B. Also, Western blot and RT-PCR showed aged SG cells expressed significantly more ADIPOQ but less aquaporin 5 than adult SG cells (Figure 5C, 5E, 5G, 5I and 5F). The mRNA level of amylase was decreased in aged SG cells, but the protein level of amylase was similar in adult and aged SG cells (Figure 5D, 5H and 5F). Uncropped western blot images were provided as Supplementary Figure 3.

Figure 5. Cellular senescence and the expressions of aging-related functional proteins in primary cultured salivary gland (SG) cells. (A) β-Galactosidase stained adult and aged primary cultured SG cells. (B) The proliferation rate of adult and aged human SG cells. (C–E) The mRNA expressions of ADIPOQ, amylase, and AQP5 in adult and aged human parotid and submandibular gland tissues. (F–I) The protein expressions of ADIPOQ, amylase, and AQP5 in adult and aged human parotid and submandibular gland tissues. Results are presented as the mean ± SD and T-test was performed for statistical analysis, ^***p < 0.001, ^**p < 0.01, ^*p < 0.05.

Silencing of ADIPOQ increased cell proliferation and suppressed SG cell dysfunction and proinflammatory reactions

To examine the functional role of ADIPOQ, we silenced its expression in human SG epithelial cells using siRNA (Figure 6C and 6G, P < 0.05, Figure 6F). ADIPOQ siRNA-treated aged SG cells exhibited significantly fewer SA-β-gal positive cells than adult cells treated with scrambled siRNA (p < 0.05, Figure 6A). Furthermore, ADIPOQ siRNA-treated SG cells proliferated significantly more than scrambled siRNA-treated SG cells (P < 0.0001 and P < 0.001, Figure 6B). We also studied the expressions of the SG functional molecules amylase and aquaporin 5 by Western blot and qRT-PCR and found the expressions of both were higher in ADIPOQ siRNA treated SG cells than in scrambled siRNA treated SG cells (Figure 6D, 6E, 6H and 6I, P < 0.001).

Figure 6. Silencing of ADIPOQ reduced cellular senescence and enhanced the proliferation and expression of salivary functional proteins. (A) β-Galactosidase assays of scrambled-siRNA and ADIPOQ-siRNA transfected primary cultured SG cells. (B) Proliferations of ADIPOQ-siRNA and scrambled siRNA treated SG cells. (C–E) The mRNA expressions of ADIPOQ, amylase, and AQP5 in ADIPOQ-siRNA and scramble-siRNA treated SG cells. (F–I) The protein levels of ADIPOQ, amylase and AQP5 in ADIPOQ-siRNA and scramble-siRNA treated SG cells. Results are presented as the mean ± SD and T-test was performed for statistical analysis, ^****p < 0.0001, ^***p < 0.001, ^*p < 0.05.

RT-PCR showed that IL-8 and p16 mRNA levels in SG primary cultured cells were significantly lower in ADIPOQ siRNA transfected cells than in scrambled siRNA transfected cells (P < 0.05, Figure 7A and 7B). The IL-8 protein expression level was significantly lower in ADIPOQ siRNA transfected cells than in scrambled siRNA transfected cells (Figure 7D, P < 0.05 and Figure 7C). Furthermore, ADIPOQ knockdown tended to reduce the expression of p21 in human SMG primary cultured cells (Figure 7E and P = 0.09, Figure 7C). Uncropped western blot images were provided as Supplementary Figures 4 and 5.

Figure 7. Effects of ADIPOQ gene knockdown on inflammatory factors. (A, B) Expressions of IL8 and p16 (proinflammatory genes) in ADIPOQ silenced cells. (C–E) Expressions of IL8 and p21 (a proinflammatory protein) in ADIPOQ silenced cells. Results are presented as the mean ± SD and T-test was performed for statistical analysis ^*p < 0.05.

Preparation of human parotid and submandibular gland tissue specimens for cDNA microarray analysis

Human parotid gland and human submandibular gland tissues were obtained from patients that underwent parotidectomy for a benign parotid tumor. Tissue samples were examined by a pathologist, who excised normal tissues for cDNA microarray analysis. Human parotid gland and submandibular gland tissue samples from adult (28.6 ± 6.6 y.o. and 43.3 ± 1.5 y.o., respectively) and aged (82.0 ± 4.3 y.o. and 88.0 ± 4.3 y.o., respectively) subjects were examined and processed (n = 3, each). For the histopathological study, excised human parotid gland and human submandibular gland tissues were fixed in a 4% paraformaldehyde phosphate buffer solution, embedded in paraffin, and stained with H&E. In addition, clinical data (age, sex, and BMI) were retrospectively reviewed and summarized in Supplementary Table 1. All specimens were collected after obtaining informed consent and institutional review board approval (INHA 18 0503-560).

cDNA microarray analysis

Total RNAs from adult and aged SG tissues were extracted using the RNeasy Mini kit (Qiagen, Germany). RNA purity (The values of 1.7–2.2 of 260/230 and 260/280 ratio) and integrity (RIN >7) were confirmed using an ND-1000 Spectrophotometer (NanoDrop Technologies, Inc., Wilmington, DE, USA) and an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). The Affymetrix Whole Transcript Expression array process was performed according to the manufacturer’s instructions (GeneChip Whole Transcript PLUS reagent Kit). cDNA was synthesized using the GeneChip WT (Whole Transcript) Amplification kit. Sense cDNA was fragmented and biotin-labeled with TdT (terminal deoxynucleotidyl transferase) using the GeneChip WT Terminal labeling kit. Approximately 5.5 μg of labeled target DNA was hybridized to the Affymetrix GeneChip Human 2.0 Array for 16 hours at 45°C. Hybridized arrays were then washed and stained on a GeneChip Fluidics Station 450 and scanned using a GCS3000 Scanner (Affymetrix). Signal values were computed using Affymetrix^® GeneChip^™ Command Console software. Data were summarized and normalized using the robust multi-average (RMA) method implemented in Affymetrix^® Power Tools (APT). DEG analysis was performed using gene-level RNA analysis results. Statistical significances of expressional differences were determined using the independent t-test. The false discovery rate (FDR) was controlled by adjusting p-values using the Benjamini-Hochberg algorithm. To produce a DEG set, Hierarchical cluster analysis was performed using complete linkage and Euclidean distance as a measure of similarity. Gene-Enrichment and Functional Annotation analyses for the DEGs list using Gene Ontology (http://geneontology.org) and KEGG (http://kegg.jp). Data analysis and DEG visualization were conducted using R 3.3.3 (http://www.r-project.org).

Histological analysis and immunohistochemistry

Human PG and SMG tissue sections were dewaxed, hydrated, stained with H&E, and examined under a digital microscope (Olympus, Japan). The IHC study was performed as previously described [41] using antibodies for ADIPOQ, b-galactosidase, and 8-OHdG (1:1000; Santa Cruz, CA, USA). A blinded examiner evaluated three random fields per section, and stained areas were measured in pixels using Image J software (MD Anderson Cancer Center, TX, USA).

Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay

Apoptosis in human SG tissues was identified by TUNEL staining using an in-situ Cell Death Detection Kit (Roche, Germany). Numbers of apoptotic cells were quantified by digital microscopy (Olympus, Japan), and apoptotic indices were calculated using three random fields per sample.

Cell culture

For human SG epithelial cell culture, a small portion of non-tumor bearing gland tissue was washed with HBSS containing 1% antibiotics, chopped with a pair of fine scissors for 7 min, enzymatically digested with 0.25% collagenase type B (2.5 mg/mL) and DNase I (1 mg/mL) with gentle shaking at 37°C for 30 minutes, filtered through a 70 μm cell strainer, and centrifuged at 1500 rpm for 3 min. Cells were plated on a culture dish in Keratinocyte serum-free medium (Gibco, USA) containing L-glutamine, 2.5 μg of EGF (epidermal growth factor), 0.09 mM CaCl₂, and 1% antibiotics and incubated in a 5% CO₂ atmosphere at 37°C.

RNA interference

Small interfering RNAs (siRNAs) targeting ADIPOQ (Forward: 5′-CCA UGA CAC CAA CUG AUC AUU-3′, Reverse: 5′-UGA UCA GUU GGU GUC AUG GUU-3′) and scrambled siRNA (Forward 5′-CCU CGU GCC GUU CCA UCA GGU AGU U-3′, Reverse: 5′-CUA CCU GAU GGA ACG GCA CGA GGU U-3′) were purchased from Genolution (Seoul, Korea). SG cells were transfected with 40 nM of siRNA using G-fectin transfection reagent (Genolution) for 72 hr, and protein levels were assayed 72 hr after transfection.

Proliferation assay

After treating SG cells with siRNA for 72 hr, cells were detached, and plated in a 96-well plate in 100 μL complete medium at a density of 4 × 10³ cells/well and cellular proliferations were measured at four time points (days 0, 1, 3, and 5) by CELLOMAX^™ solution (PreCareGene, Korea). On each day, cells were treated with CELLOMAX^™ solution and incubated for one hour at 37°C in 5% CO₂. Absorbance was read at 450 nm with a microplate reader (Molecular Devices, USA).

Senescence-associated β-galactosidase staining

SA-β-gal-positive cells were identified using a senescence β-galactosidase staining kit (Sigma Aldrich, USA). In brief, fixed SG cells were washed 3 times with PBS and incubated with a staining mixture in the absence of CO₂ for 24 hr at 37°C. Blue-stained cells were visualized under a digital microscope, and senescent cells were counted in three random fields per slide using Image J (MD Anderson Center, USA).

RNA isolation and real time PCR

Total RNA was isolated from human salivary gland epithelial cells using the RNeasy Mini kit (Qiagen, Germany), and cDNA was synthesized from total RNA using the Tetro cDNA synthesis kit (Bioline, USA). In brief, the cDNA reaction mixture was incubated for 30 min at 45°C, heated for 5 min at 85°C, and cooled to 4°C. Using cDNA as a template, real-time PCR was performed in 96-well plates (Applied Biosystems, USA) using SYBR green II Master Mix (Takara Bio Inc., Japan) in a StepOne unit (Applied Biosystems, USA) using the following program: 95°C for 20 s followed by 40 amplification cycles of 95°C for 5 s and 60°C for 20 s. The primers used were: IL-8 forward, 5′-TTT TGC CAA GGA GTG CTA AAG A-3′ and reverse, 5′-AAC CCT CTG CAC CCA GTT TTC-3′; p16 forward, 5′-CAA CGC ACC GAA TAG TTA CG-3′, and reverse, 5′-CAG CTC CTC AGC CAG GTC-3′; Adipoq forward, 5′-CTA TGA TGG CTC CAC TGG TA-3′ and reverse, 5′-GAG CAT AGC CTT GTC CTT CT-3′; AQP5 forward, 5′-ACT GGG TTT TCT GGG TAG GG -3′ and reverse, 5′-GTG GTC AGC TCC ATG GTC TT-3′; Amylase forward, 5′-ACA TGG GGC TGG AGG AGC CT-3′, and reverse, 5′-TGG TGG CCC AAC CCA ATC AT-3′ and β-actin (the endogenous control) forward, 5′-AGC TGT GCT ATG TTG CCC TG-3′, and reverse, 5′-AGG AAG CAA GGC TGG AAG AG-3′.

Western blotting

Cells were homogenized in PRO-PREP^™ protein extraction solution (iNtRON Biotechnology, Korea), incubated on ice for 30 min, centrifuged at 13,000 rpm for 10 min at 4°C, and supernatants were collected. Primary antibodies for ADIPOQ, amylase, AQP5, p16, p21, and GAPDH (all from Santa Cruz Biotechnology, 1:1000) were used. Goat anti-mouse IgG-HRP was used as the secondary antibody (Santa Cruz Biotechnology, 1:5000). Proteins were visualized using SuperSignal^™ West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific, USA) using a ImageQunat^™ LAS 4000 unit (GE Healthcare, USA).