Downregulation of CFIm25 amplifies dermal fibrosis through alternative polyadenylation

Abstract

Systemic sclerosis (SSc; scleroderma) is a multisystem fibrotic disease. The mammalian cleavage factor I 25-kD subunit (CFIm25; encoded by NUDT21) is a key regulator of alternative polyadenylation, and its depletion causes predominantly 3'UTR shortening through loss of stimulation of distal polyadenylation sites. A shortened 3'UTR will often lack microRNA target sites, resulting in increased mRNA translation due to evasion of microRNA-mediated repression. Herein, we report that CFlm25 is downregulated in SSc skin, primary dermal fibroblasts, and two murine models of dermal fibrosis. Knockdown of CFIm25 in normal skin fibroblasts is sufficient to promote the 3'UTR shortening of key TGFβ-regulated fibrotic genes and enhance their protein expression. Moreover, several of these fibrotic transcripts show 3'UTR shortening in SSc skin. Finally, mice with CFIm25 deletion in fibroblasts show exaggerated skin fibrosis upon bleomycin treatment, and CFIm25 restoration attenuates bleomycin-induced skin fibrosis. Overall, our data link this novel RNA-processing mechanism to dermal fibrosis and SSc pathogenesis.

Graphical abstract

Figure 1.

CFIm25 is depleted in myofibroblasts of SSc skin. (A) Comparison of CFIm25 transcript levels in 61 SSc and 36 matched controls (shown as log-transformed signal intensity) indicates significant CFIm25 downregulation in SSc. (B) CFIm25 transcript levels in the skin of early and late Ssc patients (n = 54). The y axis in A and B indicates the log2-transformed signal intensity of CFIm25. Data were analyzed using a Student’s t test. (C) Dual immunostainings show CFIm25 (red) and α-SMA (green) expression in normal and SSc skin (n = 10). Arrowheads show CFIm25-positive cells, and arrows point to myofibroblasts; nuclei are blue. Scale bars = 50 µm. (D) The percentage of CFIm25⁺ fibroblasts were counted and calculated. (E) The number of CFIm25⁻ and α-SMA⁺ cells were counted. P value was calculated using two-tailed Student’s t test to determine the difference between normal and SSc cells. (F) CFIm25 protein expression was examined in primary fibroblasts derived from five unaffected controls and five SSc patients. For A and B, data are presented as box and whisker plot showing quartiles and range. For D and E, data are mean ± SEM and representative of two experiments with n = 10 per group. Data were analyzed using a Student’s t test; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure S1.

Dual immunostaining was performed for CFIm25 and α-SMA expression in normal and SSc skin (n = 10). (A) The number of CFIm25 and α-SMA double-positive cells was calculated for each sample. (B) The percentages of CFIm25⁻ and α-SMA⁺ cells in all α-SMA⁺ cells were counted and calculated. (C) The percentages of CFIm25⁻ and α-SMA⁺ cells in all CFIm25⁻ fibroblasts were counted and calculated. (D) The densitometry of CFIm25 and COL(I) were measured. P values were determined by a two-tailed Student’s t test to determine the difference between normal and SSc cells. Data are presented as mean ± SEM and representative of two experiments with n = 10 per group (A–C) and two experiments with n = 5 per group (D). Data were analyzed using Student’s t test. *, P < 0.05; ***, P < 0.001.

Figure 2.

RNA-seq of CFIm25 KD versus control in normal HDFs (n = 5). (A) Representative scatterplot of PDUIs in control and CFIm25 KD cells where mRNAs are significantly shortened (n = 971) or lengthened (n = 93) after CFIm25 KD in five HDFs. (B) Pie graph showing 971 (8.1%) genes had shortened 3′UTR, while only 93 (0.7%) genes showed 3′UTR lengthening. (C) TGF-β pathway was predicted to be the top upstream regulator of the CFIm25 KD APA profile, followed by angiotensinogen, Angiopoietin 2, and fibroblast growth factor pathways.

Figure 3.

KD of CFlm25 induces APA and upregulation of fibrotic factors in normal HDFs. Five primary normal HDFs were transfected with CFlm25 specific siRNA (si_CFIm25) or a si-RNA control. (A) Western blot confirmed effective KD of CFlm25 and showed increased protein levels of COL(I), TGFBR1, COL11A1, and SPARC. (B) The upper panel demonstrates the primer locations for monitoring APA, and the lower panel demonstrates normalized dPAS usage by RT qPCR showing 3′UTR shortening upon CFIm25 KD. (C–F) RT-qPCR demonstrates increased COL1A1, TGFBR1, COL11A1, and SPARC transcript levels in CFIm25 KD HDFs. Data are mean ± SEM of n = 3 independent experiments (five primary cell lines and two biological replications). P values were determined using a one-sample t test comparing to 0 (B) or a paired two-tailed Student’s t test comparing si_CFIm25 to si_Con (C–F). *, P < 0.05; **, P < 0.01.

Figure 4.

APA of CFIm25 targets was present in SSc skin. (A) Normalized dPAS usage performed by RT-qPCR showing COL1A1, COL11A1, TGFBR1, and SPARC had 3′UTR shortening in affected skin collected from SSc patients. n = 10. **, P < 0.01 versus 0. (B) The transcript levels of above four genes were analyzed using RT-qPCR. Data are presented as mean ± SEM and representative of n = 10 patients (two biological replications). P values were calculated using a one-sample t test compared with 0 (A) or a paired two-tailed Student’s t test (B). *, P < 0.05; **, P < 0.01. ns, not significant.

Figure 5.

CFIm25 is decreased in murine dermal fibrosis. (A) Immunohistochemistry showing CFIm25 (brown) and α-SMA (red) staining in the skin of mice subcutaneously treated with PBS or bleomycin (n = 10). Arrowheads show CFIm25⁺ cells, and arrows point to α-SMA⁺ myofibroblasts. Scale bars = 25 µm. (B and C) CFIm25 protein expression was determined in the skin lysate from (B) mice injected with PBS or bleomycin; and (C) 8- to 10-wk-old WT and TSK1 mice. COL(I) was used as a marker for fibrosis.

Figure 6.

Knocking out CFIm25 in fibroblasts promotes dermal fibrosis and APA. 6- to 8-wk-old Col1a1-creER^T2 and Col1a1-creER^T2-CFIm25^f/f mice were treated with tamoxifen for 5 d to induce Cre activation. (A) Western blot shows Cre expression and a downregulation of CFIm25 in primary fibroblasts isolated from Col1a1-creER^T2-CFIm25^f/f mice (n = 4). Tamoxifen-treated mice were then administrated with s.c. PBS/bleomycin 6 d a week for 4 wk (n = 10). (B) Western blot was performed to confirm the downregulation of CFIm25 and the upregulation of fibrotic makers in skin of PBS/bleomycin-treated Col1a1-creER^T2 and Col1a1-creER^T2-CFIm25^f/f mice. (C) Masson’s trichrome showing dermal fibrosis in different treatment groups; the data are quantitated in D. Scale bars = 125 µm. (E) The dPAS usage for Col1a1, Tgfbr1, and Col11a1 was determined using RT-qPCR. Data are mean ± SEM and were analyzed using one-way ANOVA followed by a Sidak's multiple comparisons test (D) or unpaired two-tailed Student’s t test (E). n = 2 independent experiments (six to eight mice per group, two biological replications). **, P < 0.01; ***, P < 0.001.

Figure 7.

CFIm25 OE lentivirus promotes 3′UTR lengthening and inhibits protein translation of profibrotic factor and ECMs in skin fibroblasts. Two HDFs were infected with control or CFIm25-overexpressing lentivirus at 25 MOI. (A) Western blots were performed to determine CFIm25 OE 2 d after lentivirus infection, and the expression of COL(I), TGFBR1, COL11A1 and SPARC 3 d after lentivirus infection. (B) The dPAS usage of COL1A1, TGFBR1, COL11A1, and SPARC were determined in HDFs infected with lentivirus for 3 d. Data are presented as mean log₂ (percentage of long transcript in CFIm25-overexpressing cells/percentage of long transcript in control cells) ± SEM; n = 3 independent experiments (two cell lines, two biological replications). P value was determined using a one-sample t test versus 0. *, P < 0.05; **, P < 0.01.

Figure 8.

CFIm25 OE prevents bleomycin-induced skin fibrosis. 6-wk-old mice were injected with s.c. GFP or CFIm25-IRES-GFP–overexpressing lentivirus (10 × 10⁷ PFU/ml, 50 µl per spot) 1 wk before, 1 wk after, and 3 wk after the first bleomycin injection, and mice were injected with repeated s.c. bleomycin six times a week for 4 wk. Skin was collected 28 d after the first bleomycin injection for analysis. (A) Western blot was performed to determine the expression of CFIm25, COL(I), fibronectin (FN), TGFBR1, COL11A1, and SPARC. (B) The dPAS usage of Col1a1, Tgfbr1, and Col11a1 was determined. Data are presented as log₂ (percentage of long transcript in CFIm25-overexpressing skin/percentage of long transcript in control skin). (C) Sircol assay was performed to determine the pepsin-soluble collage levels in skin. (D) Masson’s trichrome, α-SMA, and picrosirius red staining showing dermal fibrosis in different treatment groups. Scale bars =125 µm. (E) Histological dermal thickness. (F) The numbers of α-SMA⁺ fibroblasts per field (left panel) as well as the percentage of α-SMA⁺ fibroblasts over the total fibroblasts (right panel). (G) The percentage of picrosirius red–stained area was measured, and data are presented as fold change to GFP control. Data are mean ± SEM of n = 2 independent experiments (five mice per group; C, D, F, and G) with two biological replications (B). P value was determined using one sample t test (B) or unpaired two-tailed Student’s t test (C, D, F, and G). *, P < 0.05; **, P < 0.01.

Figure S2.

CFlm25 depletion induces upregulation of fibrotic factors in normal HDFs. Five primary normal HDFs were transfected with CFlm25-specific siRNA (si_CFIm25) or a siRNA control (si_Con). Western blot were performed to confirmed effective CFlm25 KD and increased protein levels of COL(I), TGFBR1, COL11A1, and SPARC. The densitometry of the Western blot was measured. Data are presented as mean (ratio to β-Actin) ± SEM and representative of three experiments with n = 5 cell lines per group. P values were determined using a paired two-tailed Student’s t test comparing si_CFIm25 with si_Con. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure S3.

CFIm25 protein levels are decreased in murine models of dermal fibrosis. The densitometry of CFIm25 and COL(I) protein expression was determined in the skin lysate of (A) mice injected with PBS or bleomycin (n = 3 experiments with three or four mice per group) and (B) 8- to 10-wk-old WT and TSK1 mice (n = 2 experiments with six mice per group). COL(I) was used as a marker for fibrosis. Data are presented as mean (ratio to GAPDH) ± SEM. P values were determined using paired two-tailed Student's t test. *, P < 0.05; ***, P < 0.001.

Figure S4.

Knocking out CFIm25 in fibroblasts promotes dermal fibrosis. 6- to 8-wk-old Col1a1-creER^T2 and Col1a1-creER^T2-CFIm25^f/f mice were treated with tamoxifen for 5 d to induce Cre activation. (A) Western blot was performed to determine Cre and CFIm25 expression in primary fibroblasts isolated from Col1a1-creER^T2-CFIm25^f/f mice (n = 2 experiments with four mice per group). Densitometry of the Western blot was measured and data were normalized to GAPDH. (B) Western blot was performed to confirm the downregulation of CFIm25 and the upregulation of fibrotic makers in skin of PBS/bleomycin-treated Col1a1-creER^T2 and Col1a1-creER^T2-CFIm25^f/f mice (n = 2 experiments with approximately five to eight mice per group). Densitometry of the Western blot was determined. Data are presented as mean (ratio to GAPDH) ± SEM. P values were determined using an unpaired two-tailed Student’s t test. *, P < 0.05; **, P < 0.01.

Figure S5.

CFIm25 OE attenuates fibrotic phenotype. (A) Two HDFs were infected with control or CFIm25-overexpressing lentivirus at 25 MOI. Western blot was performed to confirm CFIm25 OE 2 d after the lentivirus infection and the expression of COL(I), TGFBR1, COL11A1, and SPARC 3 d after lentivirus infection. The densitometry of the Western blot image was quantified for each of the genes. (B and C) 6-wk-old mice were injected with s.c. GFP or CFIm25-IRES-GFP-overexpressing lentivirus (10 × 10⁷ PFU/ml, 50 µl per spot) 1 wk before, 1 wk after, and 3 wk after the initial bleomycin injection. Then mice were injected with repeated s.c. bleomycin six times a week for 4 wk. Skin was collected 28 d after the initial bleomycin injection. (B) Subsequently, the CFIm25 transcript levels were determined. (C) The densitometry of the Western blot image was determined to quantify the expression of CFIm25, COL(I), fibronectin (FN), TGFBR1, COL11A1, and SPARC. Data are presented as mean ± SEM for n = 3 experiments and two cell lines per group (A) or n = 2 experiments and five mice per group (B and C). P values were determined using an unpaired two-tailed Student’s t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001.