Involvement of metformin and aging in salivary expression of ACE2 and TMPRSS2

Abstract

SARS-CoV-2-related proteins, ACE2 and TMPRSS2, are determinants of SARS-CoV-2 infection. Although these proteins are expressed in oral-related tissues, their expression patterns and modulatory mechanisms in the salivary glands remain unknown. We herein showed that full-length ACE2, which has both a fully functional enzyme catalytic site and high-affinity SARS-CoV-2 spike S1-binding sites, was more highly expressed in salivary glands than in oral mucosal epithelial cells and the lungs. Regarding TMPRSS2, zymogen and the cleaved form were both expressed in the salivary glands, whereas only zymogen was expressed in murine lacrimal glands and the lungs. Metformin, an AMPK activator, increased stimulated saliva secretion and full-length ACE2 expression and decreased cleaved TMPRSS2 expression in the salivary glands, and exerted the same effects on soluble ACE2 (sACE2) and sTMPRSS2 in saliva. Moreover, metformin decreased the expression of beta-galactosidase, a senescence marker, and ADAM17, a sheddase of ACE2 to sACE2, in the salivary glands. In aged mice, the expression of ACE2 was decreased in the salivary glands, whereas that of sACE2 was increased in saliva, presumably by the up-regulated expression of ADAM17. The expression of TMPRSS2 in the salivary glands and sTMPRSS2 in saliva were both increased. Collectively, these results suggest that the protein expression patterns of ACE2 and TMPRSS2 in the salivary glands differ from those in other oral-related cells and tissues, and also that metformin and aging affect the salivary expression of ACE2 and TMPRSS2, which have the potential as targets for preventing the transmission of SARS-CoV-2.

Keywords: ADAM17; SARS‐CoV‐2; epithelial cells; saliva; salivary glands.

FIGURE 1

Full‐length ACE2 expression in salivary glands and effects of metformin on its expression. (A) Whole‐cell lysates prepared from human salivary glands (hSG) and human salivary and oral‐related cells were immunoblotted with anti‐ACE2 and anti‐GAPDH antibodies. Representative blots from three independent experiments are shown. Thick and thin arrows indicate full‐length ACE2 (fACE2) and the truncated ACE2 protein, respectively. (B) Whole‐cell lysates prepared from the salivary glands (SG: submandibular and sublingual glands), lacrimal glands (LG), and lungs (Lu) of young mice (N = 3) were immunoblotted with anti‐ACE2 and anti‐α‐tubulin antibodies. The bar graph shows integrated signal intensities in fACE2 normalized to that of α‐tubulin (N = 5). (C) Whole‐cell lysates prepared from the submandibular glands (SM) and sublingual glands (SL) of control (Cont) mice and mice treated with metformin (Met) (N = 3 each) for 8 weeks were immunoblotted with anti‐ACE2 and anti‐α‐tubulin antibodies. Representative blots from two independent experiments are shown. (D) Whole‐cell lysates prepared from the lungs of Cont mice and mice treated with Met (N = 3 each) for 8 weeks were immunoblotted with anti‐ACE2 and anti‐α‐tubulin antibodies. The bar graph shows integrated signal intensities for fACE2 normalized to that of α‐tubulin (N = 5). (E) ACE2 mRNA expression levels in epithelial cells of the salivary glands of Cont mice and mice treated with Met (N = 5 each) for 10 weeks. (F) Detection of sACE2 by ELISA in serum collected from Cont mice and mice treated with Met (N = 5 each) for the indicated periods. Values are presented as means ± standard deviations (SD). NS, not significant. **p < 0.01 (the unpaired Student's t‐test). A253, a human submandibular gland carcinoma cell line; HSC‐2, a human oral squamous carcinoma cell line; HGK, human primary gingival keratinocytes; HOK, human primary oral keratinocytes.

FIGURE 2

TMPRSS2 expression in salivary glands and effects of metformin on its expression. (A) Whole‐cell lysates prepared from human salivary glands (hSG) and human salivary and oral‐related cells were immunoblotted with anti‐TMPRSS2 and anti‐GAPDH antibodies. Representative blots from three independent experiments are shown. (B) Whole‐cell lysates prepared from the salivary glands (SG, submandibular and sublingual glands), lacrimal glands (LG), and lungs (Lu) of young mice (N = 3) were immunoblotted with anti‐TMPRSS2 and anti‐α‐tubulin antibodies. The bar graph shows integrated signal intensities for TMPRSS2 zymogen normalized to that of α‐tubulin (N = 5). (C) Whole‐cell lysates prepared from the submandibular glands (SM) and sublingual glands (SL) of control (Cont) mice and mice treated with metformin (Met) (N = 3 each) for 8 weeks were immunoblotted with anti‐TMPRSS2 and anti‐α‐tubulin antibodies. Representative blots from two independent experiments are shown. (D) Whole‐cell lysates prepared from the lungs of Cont mice and mice treated with Met (N = 3 each) for 8 weeks were immunoblotted with anti‐TMPRSS2 and anti‐α‐tubulin antibodies. The bar graph shows integrated signal intensities for TMPRSS2 normalized to that of α‐tubulin (N = 5). (E) TMPRSS2 mRNA expression levels in the epithelial cells of the salivary glands of Cont mice and mice treated with Met (N = 5 each) for 10 weeks. Values are presented as means ± SD. NS, not significant. **p < 0.01 (the unpaired Student's t‐test). A253, a human submandibular gland carcinoma cell line; HSC‐2, a human oral squamous carcinoma cell line; HGK, human primary gingival keratinocytes; HOK, human primary oral keratinocytes. Images detecting GAPDH and α‐tubulin proteins are identical, as shown in Figure 1.

FIGURE 3

Decreases in β‐Gal and ADAM17 expression by metformin in murine salivary glands. (A, B) Detection of β‐Gal expression by IHC in the submandibular glands (A) and sublingual glands (B) of control (Cont) mice and mice treated with metformin (Met) for 8 weeks. Sections were stained with an isotype and anti‐β‐Gal antibody. Representative images of five samples each are shown. The bar graph shows positive areas in sections (N = 5) adjusted as described below. (C, D) Detection of ADAM17 expression by IHC in the submandibular glands (C) and sublingual glands (D) of Cont mice and mice treated with Met for 8 weeks. Sections were stained with an isotype and anti‐ADAM17 antibody. Representative images of five samples each are shown. The bar graph shows positive areas in sections (N = 5) adjusted as described below. (A, C) Images of submandibular glands were quantified as the positive area (μm²) × 10⁻²/field of view (μm²). (B, D) Images of sublingual glands were quantified as the positive area (μm²)/number of ducts. Bars = 50 μm. Values are presented as means ± SD. **p < 0.01 (the unpaired Student's t‐test). (E) Detection of AQP5 and CK19 protein expression in A253 and MSEC. Whole‐cell lysates prepared from these cells were immunoblotted with anti‐AQP5, anti‐CK19, and anti‐GAPDH antibodies. A representative blot is shown. (F) MSEC were treated with the indicated concentrations of Met for 48 h. Whole‐cell lysates prepared from these cells were immunoblotted with anti‐ADAM17 and anti‐GAPDH antibodies. A representative blot is shown. The bar graph shows the integrated signal intensities of the ADAM17/GAPDH ratio. Data represent the mean ± SD of triplicate assays. **p < 0.01 versus untreated cells (0) (Dunnett's multiple comparison test).

FIGURE 4

Metformin affects the volume of saliva secreted and sACE2 and sTMPRSS2 levels in saliva. (A) Evaluation of the drinking volume of Cont (water; black line) and Met (metformin, red line) for 10 weeks. Values are presented as means per day per mouse (N = 5). (B, C) Body weight (B) and volume of saliva secreted adjusted by the weight of the salivary glands (SG) (C) in control (Cont)‐ and metformin (Met)‐treated mice for the indicated period (N = 5 each). (D) Detection of sACE2, sTMPRSS2, and α‐amylase in saliva collected from Cont mice treated and mice treated with Met for 6 weeks by Western blotting and immunoblotted with anti‐ACE2, anti‐TMPRSS2, and α‐amylase antibodies. A representative blot of two independent experiments is shown. The bar graph shows absolute integrated signal intensities (N = 4) adjusted to the loading protein amount. (B–D) Values are presented as means ± SD. NS, not significant. **p < 0.01 (the unpaired Student's t‐test).

FIGURE 5

Effects of aging and cellular senescence on the salivary expression of ACE2, TMPRSS2, and ADAM17 in mice. (A) Whole‐cell lysates prepared from the salivary glands (submandibular and sublingual glands) of young and aged mice (N = 3 each) were immunoblotted with anti‐ACE2 and anti‐α‐tubulin antibodies. The bar graph shows integrated signal intensities for fACE2 normalized to that of α‐tubulin (N = 5). (B) Sublingual gland sections collected from young and aged mice were stained with an isotype, anti‐TMPRSS2, and anti‐ADAM17 antibodies. Representative images of five samples each are shown. The bar graph shows values (N = 5) as the positive area (μm²)/number of ducts. Bars = 50 μm. (C) MSEC were passaged as indicated (N = 3), and Cdkn2a and Adam17 mRNA expression levels were quantified using real‐time PCR. (D) MSEC were passaged as indicated (N = 3), and whole‐cell lysates prepared from these cells were immunoblotted with anti‐TMPRSS2, anti‐p21^Waf1/Cip1, and anti‐GAPDH antibodies. A representative blot is shown. The bar graph shows the integrated signal intensities of the TMPRSS2/GAPDH and p21^Waf1/Cip1/GAPDH ratio. (E) Detection of sACE2 in saliva collected from young and aged mice (N = 3 each) by Western blotting. Immunoblotting with an anti‐ACE2 antibody and a representative blot of two independent experiments is shown. The bar graph shows absolute integrated signal intensities (N = 5) adjusted to the loading protein amount. (F) Detection of sACE2 by ELISA in serum collected from young and aged mice (N = 5). (G, H) Detection of sTMPRSS2 (G) and α‐amylase (H) in saliva collected from young and aged mice (N = 3 each) by Western blotting. Immunoblotting with anti‐TMPRSS2 and α‐amylase antibodies. A representative blot of two independent experiments is shown. The bar graph shows absolute integrated signal intensities (N = 5) adjusted to the loading protein amount. Values are presented as means ± SD. NS, not significant. (A, B, E–H) **p < 0.01 (the unpaired Student's t‐test). (C, D) **p < 0.01 versus passage 1 (Dunnett's multiple comparison test).