DNMT1-mediated SPINT2 expression drives early senescence by suppressing c-Met signaling in human fibroblasts

Abstract

Cellular senescence is a critical process involved in aging and related disorders, yet the molecular triggers of early senescence remain elusive. Here, we identify DNA methyltransferase 1 (DNMT1) downregulation as a key trigger of early senescence and establish serine protease inhibitor Kunitz type 2 (SPINT2) as its critical downstream effector. Using replicative and oxidative stress-induced senescence models of primary human diploid fibroblast, we observed persistent upregulation of SPINT2 and inverse downregulation of DNMT1, preceding senescence-associated β-galactosidase activity, a conventional senescence marker. Pharmacological inhibition and siRNA-mediated knockdown of DNMT1 significantly increased SPINT2 expression and induced senescence, showing mitigated effects by SPINT2 knockdown. Furthermore, SPINT2 overexpression alone induced senescence. Methylation-specific sequencing identified four CpG sites in SPINT2 promoter, that became hypomethylated at early transition of senescence and upon DNMT1 suppression. Functional analyses revealed that DNMT1-mediated SPINT2 expression induced c-Met inhibition, triggering senescence. Transcriptomic profiling identified 17 commonly deregulated c-Met signaling genes in both senescence models, with COL27A1, STAM2, and CBL validated as key downstream targets of SPINT2/c-Met signaling. These findings establish DNMT1-mediated SPINT2 upregulation as a novel epigenetic mechanism driving senescence initiation via c-Met inhibition, providing insights into the early stage of senescence and potential therapeutic targets for aging-related diseases.

Keywords: DNA methyltransferase 1 (DNMT1); aging; cellular senescence; senescence; serine protease inhibitor Kunitz type 2 (SPINT2).

Figure 1

An inverse correlation between DNMT1 and SPINT2 emerges at the early transition (E-to-M) to senescence in RS of primary HDF. (A, B) Heatmap (A) and correlation plot (B) depicting DNMT1 and SPINT2 expression dynamics throughout RS process. Transcriptomic data were obtained from our previously established time-series RS model of primary HDF (GSE41714). The different stages of RS—early (E), middle (M), advanced (A), and very advanced (VA)—are indicated. (C–E) RS of primary HDFs was induced as described in the 'Experimental Procedures' section. (C) qRT-PCR analysis of DNMT1 and SPINT2 mRNA expression. ^**p < 0.01 vs. DT2 by Student’s t-test. (D) Western blot analysis of DNMT1, SPINT2, and p21 protein expression. (E) Cell size (upper) and granularity (lower) were assessed by side scatter (SSC) and forward scatter (FSC), respectively, using flow cytometry. ^**p < 0.01 vs. DT2 by Student’s t-test. Bar graphs represent mean ± standard deviation (SD) from three independent cultures. Representative SSC and FSC patterns of DT2 and DT11 cells are shown in the right panel. (F) SA-β-gal assay was performed at each stage. Representative images are shown and quantifications of SA-β-gal-positive cells are presented below each image. ^**p < 0.01 vs. DT2 by Student’s t-test.

Figure 2

The inverse relationship between DNMT1 and SPINT2 expression is also observed in the OSIS model of primary HDF. (A–C) OSIS was induced by exposing primary HDFs (DT2) to the indicated concentrations of H₂O₂ for 3 days. (A) SA-β-gal activity was quantified (upper panel), and representative staining images are shown (lower panel). ^**p < 0.01 vs. HDF DT2 control (C) by Student’s t-test. (B) qRT-PCR analysis of DNMT1 and SPINT2 mRNA levels. ^**p < 0.01 vs. C by Student’s t-test. (C) Western blot analysis of DNMT1, SPINT2, and p21 protein expression. (D, E) HDFs (DT2) were treated with 200 μM H₂O₂ for the indicated time points. (D) qRT-PCR analysis of DNMT1 and SPINT2 mRNA levels. ^**p < 0.01 vs. control (C) by Student’s t-test. (E) Western blot analysis of DNMT1, SPINT2, and p21 protein expression.

Figure 3

DNMT1 inhibition induces senescence via SPINT2 upregulation. (A–C) HDFs (DT2) were treated with indicated concentrations of 5-AzC for 3 days. (A) Western blot analysis of DNMT1, SPINT2, and p21 protein expression. (B) qRT-PCR analysis of SPINT2 mRNA expression. ^**p < 0.01 vs. DMSO control (V) by Student's t-test. (C) Quantification of SA-β-gal activity ^**p < 0.01 vs. V by Student’s t-test. (D–H) HDFs (DT2) were transfected with siRNAs targeting the indicated genes for 4 days. (D) qRT-PCR analysis of DNMT1 and SPINT2 mRNA levels. ^**p < 0.01 vs. negative control siRNA (NC) by Student’s t-test. (E) Western blot analysis of DNMT1, SPINT2, and p21 protein expression. (F) Quantification of SA-β-gal activity. ^**p < 0.01 vs. NC by Student’s t-test. (G) Western blot analysis of DNMT1, SPINT2, and p21 protein expression. (H) Quantification of SA-β-gal activity following co-transfection ^**p < 0.01 vs. NC by Student’s t-test. (I, J) HDFs (DT2) were infected with the recombinant lentivirus encoding SPINT2 cDNA for 4 days. GFP-encoded recombinant lentivirus was used as a control. (I) Western blots analysis. (J) Quantification of SA-β-gal activity. ^**p < 0.01 vs. GFP by Student’s t-test.

Figure 4

DNMT1 regulates SPINT2 expression via promoter methylation. (A) Schematic representation of SPINT2 promoter constructs used in the luciferase reporter assay, with indicated deletion sites (FL, full-length of −1700 to +300; ΔP1, deletion of −1700 to −839; ΔP1+ΔP2, deletion of −1700 to −268; ΔP2, deletion of −839 to −268). (B) HDFs (DT2) were transfected with pGL4-basic or pGL4-SPINT2 promoter reporter constructs for 4 days, followed by luciferase assay. (C) HDFs (DT2) were transfected with siRNAs targeting negative control (NC) or DNMT1 (siDNMT1) for 24 hours, then transfected with SPINT2 promoter reporter constructs for 2 days. Cell lysates were analyzed for luciferase activity. (D) Schematic representation of CpG islands in the SPINT2 promoter and selected regions used for DNA methylation-specific sequencing (MSS) and methylation-specific PCR (MSP) analyses. (E) MSS analysis of DNA clones from DT2 (early-stage) and DT5 (middle-stage) HDFs. Methylated CpG sites are shown in black, and unmethylated sites in white. Seventeen DNA clones were sequenced per group. (F) Methylation-specific PCR (MSP) analysis of the four CpG methylation hot spots (−445, −366, −363, and −358) identified by MSS. SM indicates DNA size marker. (G) HDFs (DT2) were transfected with siRNAs for negative control (NC) or DNMT1 (siDNMT1) for 3 days and then subjected to MSP analysis against the 4 CpG methylation hot spots.

Figure 5

SPINT2 induces senescence by inhibiting c-Met signaling. (A) Western blot analysis of phosphorylated c-Met (p-c-Met), total c-Met, SPINT2, and p21 protein expression in the RS model HDFs. (B) Western blot analysis of p-c-Met, c-Met, SPINT2, and p21 in the time-series OSIS model. (C) HDFs (DT2) were infected with recombinant lentivirus encoding SPINT2 cDNA for 4 days, followed by Western blot analysis. GFP-encoded recombinant lentivirus was used as a control. (D-E) HDFs (DT2) were transfected with siRNAs targeting negative control (NC) or c-Met (si-c-Met) for 4 days. (D) Western blot analysis of DNMT1, SPINT2, and p21 protein expression. (E) Quantification of SA-β-gal activity. ^**p < 0.01 vs. NC by Student’s t-test. (F) HDFs (DT2) were transfected with siRNAs targeting DNMT1 (siDNMT1) and/or SPINT2 (siSPINT2) for 4 days. Western blot analysis was performed. (G) M-stage HDFs (DT4) were transfected with siRNAs targeting SPINT2 (siSPINT2) for 4 days. E-stage HDFs were also used as control. Western blot analysis. (H) Heatmap presentation of differentially expressed c-Met signaling genes based on transcriptomic data from RS and OSIS models (GSE41714 and GSE80332, respectively). (I) Venn diagram depicting the overlap in deregulated c-Met signaling genes between the RS and OSIS models. Among these, 15 genes were commonly downregulated, and 19 genes were commonly upregulated. (J) HDFs (DT2) were transfected with c-Met siRNA (si-c-Met) for four days, followed by qRT-PCR analysis of c-Met target gene expression. ^**p < 0.01 vs. NC by Student’s t-test. (K) HDFs (DT2) were infected with recombinant lentivirus encoding SPINT2 cDNA for 4 days, followed by qRT-PCR analysis of c-Met target genes. GFP-encoded recombinant lentivirus was used as a control. ^**p < 0.01 vs. GFP by Student’s t-test. (L) Schematic diagram of early senescence modulation.