Estrogen-dependent gene regulation: Molecular basis of TIMP-1 as a sex-specific biomarker for acute lung injury

Abstract

Increased circulating tissue inhibitor of metalloproteinases-1 (TIMP-1) levels have been observed in patients with acute lung injury (ALI). However, the sex-specific regulation of TIMP-1 and the underlying molecular mechanisms have not been well elucidated. In this study, we found that plasma TIMP-1 levels were significantly higher in COVID-19 and H1N1 patients compared with those in healthy subjects (n = 25). TIMP-1 concentrations were significantly different between males and females in each disease group. Among female but not male patients, TIMP-1 levels significantly correlated with the PaO₂/FiO₂ ratio and hospital length of stay. Using the mouse model of ALI induced by the H1N1 virus, we found that TIMP-1 is strikingly induced in PDGFRα-positive cells in the murine lungs. Moreover, female mice showed a higher Timp-1 expression in the lungs on day 3 postinfection. Mechanistically, we observed that estrogen can upregulate TIMP-1 expression in lung fibroblasts, not epithelial cells. In addition, overexpression of estrogen receptor α (ERα) increased the TIMP-1 promoter activity. In summary, TIMP-1 is an estrogen-responsive gene, and its promoter activity is regulated by ERα. Circulating TIMP-1 may serve as a sex-specific marker, reflecting the severity and worst outcomes in female patients with SARS-CoV2- and IAV-related ALI.

Keywords: COVID‐19; PDGFRA; fibroblasts; gene regulation; influenza A virus; sex differences.

FIGURE 1

Plasma TIMP‐1 levels in patients with COVID‐19 and healthy subjects. (a) Plasma TIMP‐1 levels from healthy subjects (n = 25) and COVID‐19 patients (n = 60). (b) Plasma TIMP‐1 levels from male healthy subjects (n = 7) and male COVID‐19 patients (n = 30). (c) Plasma TIMP‐1 levels from female healthy subjects (n = 18) and female COVID‐19 patients (n = 30). (d) Comparison of plasma TIMP‐1 levels in female (n = 30) and male COVID‐19 patients (n = 30).

FIGURE 2

Plasma TIMP‐1 levels in patients with H1N1 and healthy subjects. (a) Plasma TIMP‐1 levels from healthy subjects (n = 25) and H1N1 patients (n = 82). (b) Plasma TIMP‐1 levels from male healthy subjects (n = 7) and male H1N1 patients (n = 44). (c) Plasma TIMP‐1 levels from female healthy subjects (n = 18) and female H1N1 patients (n = 38). (d) Comparison of plasma TIMP‐1 levels in males (n = 44) and female H1N1 (n = 38).

FIGURE 3

Correlation analyses between plasma TIMP‐1 and the clinical parameters in patients infected with either COVID‐19 or H1N1. Spearman correlation coefficient was used to test relationships between (a) TIMP‐1 and PaO₂/FIO₂ ratio; male patients (n = 74), (b) TIMP‐1 and PaO₂/FIO₂ ratio; female patients (n = 64), (c) TIMP‐1 and hospital length of stay; male patients (n = 61), (d) TIMP‐1 and hospital length of stay; female patients (n = 53).

FIGURE 4

Sex‐biased TIMP‐1 expression in murine lungs during H1N1 viral infection. (a) The human lung single cell RNA‐sequencing data generated using the Human Protein Atlas database shows the relative TIMP‐1 mRNA expression in healthy lung tissues. (accessed on 07/30/2024). (b) The mouse lung single cell RNA‐sequencing data generated using Lung Gene Expression iN Single‐cell (LungGENS) database shows the relative TIMP‐1 mRNA expression in murine lungs at postnatal 28 days (accessed on 07/30/2024). (c) WT mice infected with 10³ PFU A/California/04/2009 H1N1 for 3 days (5–7 mice/group). Lung sections from H1N1‐infected WT mice on day 3 p.i. were immunostained for TIMP‐1 (red) and PDGFRα (green). The nuclei were stained with DAPI. Immunostained lung sections were examined with a confocal microscope and merged images. Images shown are representative of sections from 5 to 8 mice per experimental group. Scale bar = 50 μm and 10 μm as indicated. (d) Timp‐1 mRNA levels in female and male lungs (3–4 mice/group). (e and f) Lung sections from H1N1‐infected female and male mice (4 mice/group) on day 3 p.i. were immunostained for TIMP‐1 (green). The nuclei were stained with DAPI. Scale bar = 100 μm and 50 μm as indicated (e). TIMP‐1 positive cells were counted (f). ns, p > 0.05.

FIGURE 5

Estrogen induces TIMP‐1 expression in lung fibroblasts. (a) A549, BEAS‐2B, and IMR‐90 cells were incubated with 100 nM E2 for 6 h. TIMP‐1 mRNA expression was measured using qPCR. (B and C) IMR‐90 (b) and MLg cells (c) were treated with 100 nM E1 or E2 for 6 h. TIMP‐1 mRNA expression was measured using qPCR. (d) MLg cells were treated with or without 100 nM E1 or E2 for 24 h as indicated. MLg cells were immunostained for TIMP‐1 (green). The nuclei were stained with DAPI. Scale bar = 100 μm. ns, p > 0.05.

FIGURE 6

ERα regulates TIMP‐1 promoter activity. (a) Predicted ERα binding sites in the human TIMP‐1 promoter region. Fragments (F1–F3) were subcloned into a luciferase reporter. (b) The luciferase activity containing F1, F2, or F3 was measured using HEK293T 48 h after they were co‐transfected with ERα or ERβ as indicated. The pGL‐3basic luciferase reporter (Promega) empty backbone plasmid was used as a negative control. Luciferase reporter containing three copies of vitellogenin estrogen response element (3 × ERE) was used as a positive control.