Depletion of tropomyosin 1.6 alleviates TGF-β1-induced fibroblast activation by promoting the α-SMA-MMP9 matrix-degrading structure

Summary

Fibroblasts can be transformed into myofibroblasts under pro-fibrotic conditions, which are characterized by increased contractility and reduced matrix degradation. The relationship between contractile activity and matrix degradation is not fully understood. To mimic physiological conditions, fibroblasts were cultured on a collagen gel with low rigidity. We reveal that a tropomyosin isoform, tropomyosin 1.6 (Tpm1.6), plays a pivotal role in the phenotypic switch upon TGF-β1. Tpm1.6 was specifically upregulated by TGF-β1 in renal fibroblasts. Tpm1.6-silencing decreased TGF-β1-induced myofibroblast markers and contractility, and promoted collagen degradation. Remarkably, in Tpm1.6-silenced fibroblasts, TGF-β1 triggered the formation of distinct α-SMA dots enriched with MMP9, promoting collagen degradation. The targeted silencing of Tpm1.6 in activated myofibroblasts by the induction of promoters of Foxd1 and Acta2 mitigated unilateral ureteral obstruction-induced renal fibrosis and preserved proximal tubule differentiation. Our study highlights the crucial role of Tpm1.6 in TGF-β1-induced myofibroblast activation and collagen degradation, suggesting a potential therapeutic approach for chronic kidney disease.

Graphical abstract

Highlights

• •Tpm1.6 is upregulated by TGF-β1 in renal fibroblasts
• •Silencing Tpm1.6 reduces contractility and promotes robust collagen degradation
• •Tpm1.6 depletion induces α-SMA–MMP9 dot structures, facilitating matrix breakdown
• •Tpm1.6 silencing alleviates renal fibrosis and preserves tubular differentiation

Biochemistry; Molecular biology; Cell biology

Introduction

End-stage renal disease is marked by chronic renal inflammation and progressive tubulointerstitial fibrosis, driven by myofibroblast activation, which is characterized by excessive extracellular matrix (ECM) production and a contractile phenotype with increased α-SMA, N-cadherin, and β1 integrin. Myofibroblasts could originate from various cell sources, including tissue-resident fibroblasts,¹ perivascular mesenchymal cells,^1,2 and macrophages³ during fibrosis. Previous findings and our studies also showed that the mechanical properties of substrates and mechanosensitive components play crucial roles in myofibroblast activation.^4,5,6 However, most in vitro studies on myofibroblast activation were conducted on a stiff culture dish. Thus, the involvement of the mechanical properties of the microenvironment in myofibroblast activation from different precursors still needs to be clarified.

Actomyosin-associated contractile units (CUs) are cellular components that participate in generating contractile force, including actin, myosin, tropomyosin, troponin, and so forth. For fibroblasts, the function of CUs is to exert mechanical forces on surrounding ECM and neighboring cells, and thus plays a crucial role in the developmental process, wound healing, and tissue remodeling.^7,8,9 TGF-β1 induces a contractile phenotype of fibroblasts through Rho-ROCK and myosin light-chain kinase (MLCK).^10,11 The inhibition of actomyosin activity has been shown to suppress fibrosis in the lung, liver, and kidney.^12,13 However, the specific CU involved in myofibroblast activation during fibrosis remains unknown. Tropomyosin (Tpm) serves to stabilize actin filaments and myosin activity for the regulation of cell motility, adhesion, and vesicle transport via the modulation of actomyosin contractility.^14,15,16 Tpm1.6/1.7 isoforms are associated with stress fibers and specifically expressed in fibroblasts, suggesting a crucial role in the contractile phenotype of fibroblasts.^17,18,19,20

Actomyosin contractility and matrix digestion are two of the functional phenotypes in fibroblasts. The activation of Src in epithelial cells suppressed stress fiber and promoted podosome formation.²¹ Burgstaller and Gimona demonstrated that the cortactin-Arp2/3 microdomain in the podosome leads to the dispersal of myosin/Tpm and reduction of cell contractility via p190RhoGAP.²² Those studies supported the notion of an opposite and antagonistic relationship between actomyosin contractility and matrix degradation. Under TGF-β1 stimulation, fibroblasts display a contractile phenotype even when cultured on a collagen gel. In this study, we wished to identify a pivotal CU that contributes to TGF-β1-induced contractility in fibroblasts, and try to reverse the phenotype by CU-depletion. We cultured renal fibroblast on collagen gel to mimic the initial stage of myofibroblast activation during the progression of renal fibrosis, and tried to find a CU component in determining the switch between the contractile phenotype and collagen digestion phenotype induced by TGF-β1 in myofibroblasts. Our study revealed that Tpm1.6 is required for TGF-β1-induced myofibroblast activation, and the depletion of Tpm1.6 promotes phenotypic switch from contractility into matrix-degradation upon TGF-β1. Finally, this strong collagen degradation is based on matrix metalloprotease (MMP)-contained α-SMA dots.

Results

TGF-β1-induced myofibroblast activation on collagen gel is observed only in renal fibroblasts

Myofibroblast activation plays a key role in fibrosis progression; however, it is unclear about the origin of cell precursors as well as whether the matrix stiffness affects myofibroblast activation. In renal fibrosis, fibroblasts, macrophages, epithelial cells, and pericytes have been reported to be the precursor of myofibroblasts.^2,4,23 However, whether these cells can be activated on soft collagen gel (CG), mimicking the physiological interstitial microenvironment, has not been explored before. We examined TGF-β1-induced myofibroblast activation by the elevation of α-SMA when different types of cells (fibroblasts, pericytes, epithelial cells) were cultured on CG composed of a fibrillar collagen scaffold (<100 Pa). Collagen-gel coated dishes (Co) served as the stiff control, which also contained fibrillar collagen fibrils on the surface. The renal fibroblasts (NRK-49F) showed typical spindle shape morphology on both Co and CG substrates upon TGF-β1 induction, whereas primary renal fibroblasts derived from B6 mouse kidneys cultured on CG displayed less extension, but showed spindle morphology upon TGF-β1 treatment (Figures 1A and 1B). Both renal fibroblasts cultured on CG showed the phenotype of TGF-β1-stimulated myofibroblast activation, as manifested by elevation of α-SMA, N-cadherin, and β1 integrin (Figures 1C, 1E, and 1K). However, TGF-β1 treatment could not induce an increase in α-SMA level in pericytes (CCL226) cultured on CG (Figures S1A–S1C). Additionally, renal epithelial cells (NRK-52E and LLC-PK1) exhibited limited cell growth and elevated apoptosis on soft CG. TGF-β1 treatment particularly enhanced apoptosis in LLC-PK1 cells cultured on CG (Figures S1D and 1G). We found that TGF-β1-induced epithelium-to-mesenchymal transition, as reflected by augmentation of α-SMA levels, was significantly ameliorated by CG (Figures S1E–S1F, 1H–1I). If fibroblasts were cultured on monomeric collagen-coated polyacrylamide gels with a rigidity of 0.2 kPa, TGF-β1-induced myofibroblast activation was blocked (Figures 1O and 1P). A matrix that can be remodeled is essential for myofibroblast activation. Among the tested cell types, only fibroblasts retained the ability of TGF-β1-induced myofibroblast activation on CG, suggesting that fibroblasts are the potential precursor of myofibroblasts during the onset of fibrosis when matrix stiffness is low.

Figure 1.

Actomyosin contractility is required for TGF-β1-induced myofibroblast activation under soft collagen gel

(A–J) NRK-49F cells (A) and primary renal fibroblast cells (B) were cultured on collagen gel-coated dish (Co) and collagen gel (CG) with and without TGF-β1 (10 ng/mL) for 24 h. The protein levels of myofibroblast marker (α-SMA) and actomyosin-associated contractile units (CUs; myosin IIA, IIB, and Tpm1) in NRK-49F cells (C, G) and primary renal fibroblasts (E, H) were analyzed by western blotting and quantification results of western blotting are shown in (D), (F), (I), and (J). The arrow in (G) and (H) shows the different isoforms of Tpm1 in NRK-49F cells and primary fibroblasts. Data are represented as mean ± SEM; n = 5 (D), 3 (F), 8 (I), 3 (J) independent experiments were performed; two-way ANOVA; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns (not significant). Scale bars, 20 μm.

(K–N) Protein levels of α-SMA, N-cadherin, β1-integrin, and Tpm1 in NRK-49F cultured on CG with or without the co-treatment of actomyosin blockers (BB-blebbistatin, ML7, Y27632) for 24 h, and the respective quantification results are shown in (M) and (N). Data are represented as mean ± SEM; n = 3 independent experiments were performed; two-way ANOVA; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

(O and P) NRK-49F cells were cultured on culture dish and polyacrylamide gels with a rigidity of 0.2 kPa with or without TGF-β1 induction for 24 h. Western blotting results showed the α-SMA levels, and the quantification results of α-SMA levels are shown in (P). Data are represented as mean ± SEM; n = 3 independent experiments were performed; two-way ANOVA; ∗∗p < 0.01.

Tropomyosin-1 (Tpm1) expression is elevated in TGF-β1-induced fibroblast-to-myofibroblast activation

Since actomyosin contractibility may be accompanied with myofibroblast activation, we wished to investigate the role of actomyosin contractility and CUs in myofibroblast activation when cells were cultured on CG. First, we examined the levels of myosin IIA/IIB and Tpm1 in NRK-49F cells and primary renal fibroblasts cultured under Co or CG conditions with or without TGF-β1 treatment. Myosin IIA and IIB, but not Tpm1 as recognized by exon 1a-specific antibody (TM311), were downregulated when cells were cultured on CG. At our hand, TM311 recognized Tpm2.1 and different isoforms of Tpm1, as indicated by upper and lower bands, respectively (Figures 1G and 1H). Tpm1 was enhanced by TGF-β1 in NRK-49F cells and primary renal fibroblasts cultured on both Co and CG (Figures 1G–1J). Immunofluorescence staining showed that partial Tpm1 was recruited into stress fiber after TGF-β1 stimulation (Figure S2A). TGF-β1-induced increase in Tpm1 as well as α-SMA level was ameliorated by inhibitors of TGFβRI (SB431542) and smad3 (SIS3) in NRK-49F cells (Figures S2B–S2D). To assess the importance of actomyosin-mediated contractile force in TGF-β1-induced myofibroblast activation, actomyosin activity blockers (blebbistatin, ML7 and Y27632) were applied on NRK-49F cells cultured on Co or CG upon TGF-β1. The results showed that actomyosin inhibition suppressed TGF-β1-induced myofibroblast activation (Figures 1K–1N and S2E–S2F). In sum, actomyosin-associated contractility is required for TGF-β1-induced myofibroblast activation.

TGF-β1 specifically increases the expression of Tpm1 isoform 1.6 in renal fibroblasts

There are various isoforms of Tpm1 due to alternative promoters and splicing sites (Figures 2A). We first analyzed specific exons of Tpm1 by quantitative PCR. Exon 1a, 2a, 2b, 9a and 9d of Tpm1 were elevated by TGF-β1 in NRK-49F cells under CG condition, whereas exon 9d of Tpm3 was not altered (Figure S3A). The protein level of Tpm3 was also not altered by TGF-β1 stimulation (Figure S3B). To characterize the specific isoform of Tpm1 involved in TGF-β1-induced myofibroblast activation, we analyzed different Tpm1 isoforms by the combination of different exon primer sets (e.g., forward primer of exon 1a in combination with reverse primer of exon 9d). The amplicon of paired primers for 1a to 9d and 2b to 9d, but not 1a to 9a, 1b to 9d, or 2a to 9d, was significantly upregulated by TGF-β1 in both NRK-49F cells and primary renal fibroblasts cultured on CG. This indicates that the TGF-β1-induced elevated Tpm isoforms are Tpm1.6/1.7, not Tpm1.1/1.2/1.3/1.4/1.5/1.8/1.9 (Figures 2B, 2C, S3C, and S3D). We also checked the transcripts of Tpm1 in CCL-226 (pericyte cell line) and found the dominance of transcripts of 1a to 9d and 2b to 9d (Figure S3E). Moreover, we employed two Taqman probes conjugated with different fluorescence reporters to distinguish Tpm1.6 (6FAM-exon 6b targeted sequence-BHQ1) and Tpm1.7 (TAMRA-exon 6a targeted sequence-BHQ2) in total RNA extracts (Figure 2D). The results showed that the PCR amplification of 1a to 9d and 2b to 9d released solely the 6FAM into reaction mixtures, which was increased by TGF-β1 treatment (Figures 2E and 2F). The results of exon 1a-initiated Sanger sequencing also demonstrated the presence of only mRNA sequence of Tpm1.6 in NRK-49F cells with or without TGF-β1 treatment (Figure S3F). Taken together, Tpm1.6 is the dominant Tpm isoform and is specifically upregulated by TGF-β1 in renal fibroblasts.

Figure 2.

Silencing of the renal fibroblast-dominant isoform, Tpm1.6, suppresses TGF-β1-induced myofibroblast activation

NRK-49F cells and primary renal fibroblasts were cultured on CG with or without TGF-β1 for 24 h, and total RNA was extracted and analyzed by conventional PCR (B, C) or quantitative PCR (E, F). NRK-49F cells received liposome carried biogenic RNA interference against exon 2b and 9d of Tpm1, and then were cultured on Co or CG with or without TGF-β1 for 24 h. Total RNA was extracted and analyzed by quantitative PCR (G), whereas cell lysates were analyzed by western blotting (H, I).

(A) Schematic graph displays Tpm1 isoforms derived from alternative promoters and splicing exons.

(B and C) PCR results represent the levels of different transcripts of Tpm1 in NRK-49F cells (B) and primary renal fibroblasts (C). Isoform1.6/1.7 of Tpm1 (Tpm1.6/1.7) was specifically upregulated by TGF-β1 in renal fibroblasts.

(D) Forward/reverse primers and Taqman probes conjugated with fluorescent reporter were used for quantitative PCR studies in order to detect transcripts of specific Tpm1.6 or Tpm1.7. To identify Tpm1.6 and 1.7, we utilized 6FAM-BHQ1-and TAMRA-BHQ2-conjugated Taqman probes targeting exon 6b and 6a, respectively. The polymerase in Tpm1.6/1.7 amplification (1a to 9d or 2b to 9d) passes the exon 6 and releases the reporter fluorescence from Taqman probes by nucleotides cleavage.

(E and F) Quantitative PCR showed the two distinct exon 6 signals in NRK-49F cells (E, n = 4) or primary renal fibroblasts (F, n = 3). The level of Tpm1.6, the dominant isoform in renal fibroblasts, is specifically upregulated by TGF-β1. Data are represented as mean ± SEM; n = 4 (E), 3 (F) independent experiments were performed; Welch’s t test; ∗p < 0.05, ∗∗∗p < 0.001, u.d. (undetermined); data were compared to control group.

(G) Quantitative PCR showed the RNA levels of individual Tpm1 exon in NRK-49F cells harboring biogenic RNA interference against exon 2b and 9d of Tpm1. miRNC indicates non-silencing control and miR2b/miR9d indicate specific exon silencing groups. The results indicate that exon 2b-targeting silencing shows more specificity for Tpm1.6. Data are represented as mean ± SEM; n = 3 independent experiments were performed; one-way ANOVA; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns (not significant); data were compared to miRNC.

(H and I) Western blot represents the protein levels of myofibroblast proteins (α-SMA, N-cadherin, β1-integrin) and contractile units (myosin IIA, IIB, and Tpm1) in NRK-49F cells harboring biogenic RNA interference against exon 2b and 9d of Tpm1. Quantification results of western blotting are shown in (I). Data are represented as mean ± SEM; n = 3 independent experiments were performed; two-way ANOVA; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ns (not significant).

The silencing of Tpm1.6 suppresses TGF-β1-induced myofibroblast activation and promotes collagen gel degradation

To evaluate the functional role of Tpm1.6 in TGF-β1-induced myofibroblast activation, Tpm1.6 was silenced by liposomes-carried biogenic RNA interference against exon 2b or 9d of Tpm1 in NRK-49F cells (miR2b cells or miR9d cells). The silencing of Tpm1.6 did not change the spindle shape of renal fibroblasts and did not alter the cell proliferation and viability (Figures S4A–S4C). Quantitative PCR showed specific inhibition of exon 2b or 9d in miR2b or miR9d cells, respectively. Results indicate that exon 2b-silencing is more specific for Tpm1.6, whereas exon 9d-silencing may reduce not only Tpm1.6 but also the unexpected isoforms, Tpm 1.3/1.4, which contains exon 2a (Figure 2G).

The silencing of exon 2b or 9d did not alter the TGF-β1-induced increase in protein levels of myofibroblast markers (α-SMA, N-cadherin, β1-integrin) in renal fibroblasts cultured on Co, whereas on CG, the depletion significantly reduced the TGF-β1-induced increase in myofibroblast proteins (Figures 2H and 2I). However, exon 2b-silencing exhibited a better effect than exon 9d. To evaluate the TGF-β1-induced downstream signaling, we checked the phosphorylation of smad2/3 after Tpm1.6-silencing. The TGF-β1-induced phosphorylation of smad2/3 was inhibited by silencing Tpm1.6 (Figure S4D). Taken together, silencing of Tpm1.6 via exon 2b-targeting inhibits TGF-β1-induced myofibroblast activation under CG conditions. Hereafter, “Tpm1.6-kd” refers to Tpm1.6 knockdown achieved by RNA interference specifically targeting exon 2b (miR2b), which selectively silences Tpm1.6 expression in NRK-49F cells.

To test if the silencing of Tpm1.6 affects cell contractility, a collagen contraction assay was employed. Tpm1.6-silenced NRK-49F cells were cultured in CG with or without TGF-β1, and the CGs were released for the indicated time to visualize the change of gel area. We found that cells harboring exon 2b-silencing showed the most decreased gel area upon TGF-β1 at 4 h (Figures 3A–3B). Because the gel contraction was not accompanied by increased gel opaqueness and there was also strong degradation of gel observed at 48 h, such results suggest that Tpm1.6-silenced cells display an active collagen degradation capability upon TGF-β1 induction (Figure 3A). The conventional gel contraction assay may not be suitable for cell contractility measurement particularly when cultured cells display high matrix-degradation capability.

Figure 3.

Silencing Tpm1.6 through exon 2b-targeted RNA interference prevents TGF-β1-induced cell contractility but promote collagen degradation

(A and B) Silenced-control and Tpm1.6-silenced NRK-49F cells were cultured in CG with or without TGF-β1 for 24 h and the collagen gels were released for indicated time to visualize the gel contraction/degradation, and the quantification results are shown in (B). Cells harboring exon 2b-silencing showed most decreased gel area without increase opaqueness at 4 h and strong degradation of gel at 48 h, indicating an active process of gel digestion is present. Data are represented as mean ± SEM; n = 3 independent experiments were performed; RM one-way ANOVA; ∗p < 0.05.

(C–E) Cell contractility is measured by a micropost array detector. NRK-49F cells harboring biogenic RNA interference against Tpm1 were cultured on PDMS-based microposts with or without TGF-β1 for 24 h. Each post bends in response to the applied traction force as calculated by finite element model analysis. The force map represents the contractility distribution of silenced-control and Tpm1.6-silenced cells, and the color bar shows the force level (nN). The total cell force and force per post (nN) were quantified and shown in (D) and (E). Data are represented as mean ± SEM; n = 9–15 cells; two-way ANOVA; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns (not significant).

Tpm1.6-depletion reduces TGF-β1-induced cell contractility

To assess cell contractility, we employed mPAD. Results showed that silenced-control cells exhibited stronger contractile force around cell protrusions in presence of TGF-β1 (71.55 nN–192.8 nN). Tpm1.6-kd did not change cell contractility (80.14 nN) but markedly attenuated TGF-β1-induced increase in cell contractility (96.54 nN). In contrast, exon 9d-silencing did not prevent TGF-β1-induced increase in cell contractility (Figures 3C–3E). The suppression of TGF-β1-induced cell contractility in Tpm1.6-kd fibroblasts was confirmed by another targeting sequence of exon 2b (Tpm1.6-kd#2).

In order to assess the stiffness of single collagen fibril exerted by cell contractility, we employed an advanced technology, confocal microscope co-axis with atomic force microscope (AFM). LifeAct-RFP-transfected cells (red) were cultured on FITC-conjugated CGs (green) for confocal microscopy-based visualization. The structure and rigidity of cells and peri-cellular fibrils were analyzed by AFM. TGF-β1 treatment promoted more fibrils toward the cells, but did not change the diameter of collagen fibrils (Figures 4A and S4E). The stiffness of silenced-control cells and fibrils was 8.7 ± 2.3 KPa and 8.3 ± 1.3 KPa, respectively. TGF-β1 markedly enhanced cell stiffness to 32.4 ± 6.4 KPa, as well as the stiffness of fibril exerted by cells (40.7 ± 3.3 KPa; Figures 4B and 4C). Silenced cells did not alter stiffness of cell and collagen fibril; however, Tpm1.6-kd significantly reduced TGF-β1-induced increase in stiffness of cell and collagen fibril. Exon 9d-silencing did not affect TGF-β1-indcued increase in cell rigidity, but slightly attenuated the stiffness of peri-cellular fibrils. The obscure appearance of FITC-conjugated collagen after TGF-β1 treatment in Figure 4A is likely attributable to collagen degradation mediated by MMP9 activity in Tpm1.6-kd cells, as well as the potential remodeling of collagen fibrils due to altered cell contractility under TGF-β1 induction. Taken the results of mPAD and AFM, Tpm1.6-silencing markedly reduced TGF-β1-induced cell contractility.

Figure 4.

Tpm1.6-silencing reduces TGF-β1-induced augmentation of cell contractility and promote collagen deformation, which is mediated by MMPs

(A–C) Cell contractility is measured by a confocal plus atomic force microscope (AFM) system. LifeAct-expressing NRK-49F cells harboring biogenic RNA interference against Tpm1 were cultured on FITC-conjugated CG for 4 days with or without TGF-β1. Cells (LifeAct, red) and collagen fibrils (FITC, green) were visualized under a confocal microscope and then subjected to co-axis atomic force microscopic analysis to assess the structure and rigidity of cells and peri-cellular collagen fibrils. AFM-Height images show the relative height information (μm) in topography. AFM-3D images display the surface morphology in three-dimensional presentation. AFM+Confocal indicates the merge of images between the confocal microscope and AFM-Height. Quantification of the rigidity of cells and peri-cellular collagen fibrils was shown in (B) and (C), respectively. Data are represented as mean ± SEM; n = 9–72 cells or fibers; two-way ANOVA; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Scale bars, 20 μm.

(D–G) Silenced-control and Tpm1.6-silenced NRK-49F cells were cultured on FITC-conjugated CG with or without TGF-β1 for 24 h and co-treated with actomyosin blocker (BB-blebbistatin) or pan-MMP inhibitor (GM6001). The scaffold was fixed, and CGs were incubated with Cy3-collagen hybridizing probes (CHP) for 2 h to detect the deformed collagen fibrils caused by cell traction force or enzymatic digestion. Fluorescent images showed the deformed collagen fibrils (Cy3-CHP, red) and total collagen fibrils (FITC, green) in culture of silenced-control (D) or Tpm1.6-silenced (F) NRK-49F cells, and the quantification of intensity of CHP-positive fibrils vs. FITC-collagen fibrils is shown in (E) and (G). TGF-β1-induced deformation of collagen fibrils by Tpm1.6-silenced NRK-49F cells is completely inhibited by GM6001, indicating that MMP-triggered collagen degradation is involved. Data are represented as mean ± SEM; n = 8 independent experiments were performed; two-way ANOVA; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns (not significant). Scale bars, 20 μm.

Tpm1.6-depletion triggers TGF-β1-induced collagen deformation via MMP

To assess collagen degradation, we employed Cy3-conjugated collagen-hybridizing probe (CHP) which recognizes deformed collagen fibrils caused by traction force or enzymatic digestion.^24,25 Silenced-control and Tpm1.6-kd cells were cultured on FITC-conjugated CGs for 24 h and followed by fixation and CHP incubation. Control cells exhibited minimal degree of collagen aggregation and little CHP staining. TGF-β1 enhanced the collagen aggregation and triggered the augmentation of CHP intensity, which overlapped with the collagen fibril. Inhibition of actomyosin activity by BB completely blocked TGF-β1-induced collagen aggregation and CHP intensity, whereas pan-MMP inhibitor (GM6001) did not alter TGF-β1-induced collagen aggregation and CHP intensity (Figures 4D and 4E), indicating that the TGF-β1-indcued augmentation of CHP intensity is predominantly contributed by cell contractility. Interestingly, Tpm1.6-kd did not alter TGF-β1-increased collagen aggregation and CHP intensity. BB markedly blocked TGF-β1-induced collagen aggregation and CHP intensity, whereas GM6001 treatment completely alleviated TGF-β1-induced CHP intensity, suggesting that TGF-β1-induced collagen deformation in Tpm1.6-kd cells is mediated by the activation of MMPs (Figures 5F and 5G).

Figure 5.

Tpm1.6-silencing triggers the formation of canonical podosomes and α-SMA dots upon TGF-β1 induction

Silenced-control and Tpm1.6-silenced NRK-49F cells were cultured on CG with or without TGF-β1 for 24 h, and the cultured cells underwent immunofluorescence studies (A-F), transmission electron microscope (TEM) inspection (H-J), and the lysates were harvested for immunoprecipitation (G).

(A) Canonical podosomes were represented by cortactin (red), phalloidin (green) staining; non-canonical podosomes were shown by α-SMA dots (cyan). Tpm1.6-silenced NRK-49F cells displayed more canonical and non-canonical podosomes with TGF-β1 induction. Scale bar, 20 μm.

(B and C) Quantitative results show the size (${\mu \mathrm{m}}^2$) of canonical podosomes and α-SMA dots in silenced-control and Tpm1.6-silenced cells. Data are represented as mean ± SEM; n = 6–11 cells; two-way ANOVA; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

(D) Fluorescent images showed the localization of α-SMA (green) and MMP9 (red), representing MMP9 enrichment in non-canonical podosomes in Tpm1.6-silenced cells. Scale bars, 20 μm.

(E) Quantitative results show colocalization of MMP9 and α-SMA in silenced-control and Tpm1.6-silenced NRK-49F cells with or without TGF-β1 based on the Pearson’s correlation coefficient. Data are represented as mean ± SEM; n = 5 independent experiments were performed; two-way ANOVA; ∗∗∗∗p < 0.0001.

(F) Three-dimensional image is observed by xz optical sectioning using confocal microscopy. The α-SMA dots in TGF-β1-treated Tpm1.6-kd cells showed the presence at the basal region of cells, where there is the interaction of cells and collagen matrix. Scale bars, 20 μm.

(G) The association between α-SMA and MMP9 is evaluated by immunoprecipitation. The α-SMA antibody was used for capture and precipitation, and further, the MMP9 antibody was used for immunoblot. TGF-β1-treated Tpm1.6-kd group showed an obvious association between α-SMA and MMP9.

(H) The cell-collagen scaffolds were embedded in Spurr resin after fixation and dehydration. The embedded scaffolds were sectioned with a thickness of 70 nm and then incubated with α-SMA antibody and immunogold-conjugated secondary antibodies. The images of TEM showed the detailed structure of the cell-matrix adhesion site and α-SMA-positive dots. G, CG; N, nucleus; m, cell membrane; M, mitochondria. Scale bars, 0.5 μm (upper) and 0.2 μm (lower).

(I and J) Quantitative results show the size of α-SMA-enriched dots and the number of captured golds per punctate. Data are represented as mean ± SEM; n = 33–73 aggregates; two-way ANOVA; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Tpm1.6-depletion triggers TGF-β1-induced collagen degradation via MMP9-enriched α-SMA dots

To evaluate whether podosome formation is involved in active collagen degradation capability in Tpm1.6-kd fibroblasts. Silenced-control and Tpm1.6-kd fibroblasts were cultured on CGs for 24 h with or without TGF-β1, and the canonical podosome structure was subsequently analyzed by cortactin- and pholloidin-staining. Control cells displayed little amount of podosome structure. TGF-β1 or Tpm1.6-depletion slightly increased the podosome formation, but markedly enhanced podosome dots and podosome belt in TGF-β1-treated Tpm1.6-kd cells. (Figure 5A). When the cultured cells were co-stained with α-SMA and cortactin, we have very surprising observations. In control cells, TGF-β1 increased α-SMA level, but the expression of α-SMA was exclusive of cortactin staining. In Tpm1.6-kd cells, we observed the presence of α-SMA dots. Interestingly, TGF-β1 triggered a marked increase of α-SMA dots with a larger size near the cell membrane (Figures 5A–5C). The presence of α-SMA dots were verified by another silenced sequence of exon 2b of Tpm1.6 (Figure S5A). These α-SMA dots showed few co-staining with cortactin, Tks5, and WASP, suggesting that they are not canonical podosomes (Figures 5A and S5B). Three-dimensional observation by confocal microscopy showed that the α-SMA dots were expressed at the basal region where there is interaction between cells and the collagen substrate (Figure 5F).

To explore the possible role of α-SMA dots, we co-stained MMP9. MMP9 was expressed at a low level in control cells. TGF-β1 increased the levels of MMP9, which was not co-stained with α-SMA. Tpm1.6-kd did not alter the expression of MMP9, whereas TGF-β1 not only increased the MMP9 levels but also markedly induced the colocalization of MMP9 and α-SMA dots (Figures 5D and 5E). These findings indicate that Tpm1.6-depletion facilitates the formation of MMP-containing α-SMA dots upon TGF-β1. The association between α-SMA and MMP9 in TGF-β1-treated Tpm1.6-silenced cells was observed by immunoprecipitation (Figure 5G).

To visualize the detailed structure of α-SMA dots, we employed transmission electron microscope (TEM) and α-SMA-targeted immunogold. Control and Tpm1.6-kd cells were cultured on CGs for 24 h and followed by fixation, embedding, sectioning, and subjected for TEM examination. The diameter of α-SMA dots at the cell-matrix adhesion site in control cells was around 0.1 μm, and TGF-β1 did not alter the size. In Tpm1.6-kd cells, the size of α-SMA dots was not altered, but TGF-β1 induced larger α-SMA dots (Figures 5H–5J), consistent with immunofluorescence studies. Because such α-SMA dots are enriched in MMP9, but devoid of cortactin, we further test whether they encompass the capability of collagen degradation.

To assess the collagen-degraded capability in cells, we analyzed the levels of degraded collagen protein released into culture media. Control and Tpm1.6-kd fibroblasts were cultured on CGs for 24 h, and then CGs were either released or treated with GM6001 for 48 h. The protein components in culture media were precipitated and analyzed by western blotting. Only collagen monomer and polymer were present in culture media derived from control cells. TGF-β1 induced an increase in levels of collagen polymer (above 100 kDa). Tpm1.6-kd did not affect collagen release in culture media; however, TGF-β1 markedly increased the levels of collagen degradation products released into the media (Figures 6A and 6F). To deplete physical force of CG substrates, we released the attachment between gel and culture dish. The release of CG for two days markedly triggered TGF-β1-induced collagen degradation as reflected by reduced level of monomer and augmentation of the smallest collagen fragment in culture media (Figures 6A and 6F). The western blotting results were verified by another type I collagen antibody (Figure S4F). GM6001 did not alter the pattern of collagen degradation in the culture media in control and Tpm1.6-kd cells, but completely blocked TGF-β1-induced collagen degradation in Tpm1.6-kd cells (Figures 6B and 6G). To examine the function of MMP9-containing α-SMA dots in collagen degradation, we employed a pharmacological MMP9 inhibitor and biogenic silencing of α-SMA (shSMA). Both the inhibition of MMP9 and silencing of α-SMA completely blocked TGF-β1-induced collagen degradation only in Tpm1.6-kd cells (Figures 6C, 6D, 6H, and 6I). These results indicated that α-SMA dots serve collagen degradation in Tpm1.6-kd fibroblasts upon TGF-β1 induction. Additionally, silencing of typical podosomal protein (siCortactin) also attenuated TGF-β1-induced collagen degradation in Tpm1.6-kd cells (Figures 6E and 6J). Moreover, we cultured control and Tpm1.6-kd fibroblasts on Cy3-conjugated gelatin substrate, and further observed the matrix-degraded behavior by the fluorescence loss. Tpm1.6-silencing promotes obvious degradation of gelatin upon TGF-β1, and MMP9 inhibition blocked TGF-β1-induced degradation (Figure 6K). The degraded area was also observed the colocalization of α-SMA dots (Figure 6L). Taken together, TGF-β1 predominately triggers matrix degradation in Tpm1.6-kd cells via MMP9-containing α-SMA dots. These findings reveal that this structure contributes to the collagen degradation.

Figure 6.

Tpm1.6-silencing promotes collagen digestion upon TGF-β1 stimulation via induction of α-SMA puncta

Silenced-control and Tpm1.6-silenced NRK-49F cells were cultured on CG for 24 h, and then the CG was either released (A) or was treated with MMP inhibitors (GM600 or MMP9i, B-C) for 48 h with or without TGF-β1. Silenced-control and Tpm1.6-silenced cells were transfected with shSMA (D) or siCortactin (E) and cultured on unreleased CG with or without TGF-β1 for 48 h. The culture media was harvested for Western blotting analysis to detect collagen degradation.

(A and F) Culture media collected from cells cultured on unreleased or released CG. Western blotting results showed Tpm1.6-silencing markedly increased collagen fragmentation (smaller than the size of monomer) upon TGF-β1. The release of CG to reduce traction force triggered the collagen degradation, and the quantification is shown in (F). Data are represented as mean ± SEM; n = 3 independent experiments were performed; two-way ANOVA; ∗p < 0.05, ∗∗∗∗p < 0.0001.

(B, C, G, H) Western blotting results showed the effects of GM6001 (B) or MMP9 inhibitor (C) on TGF-β1-induced collagen degradation in Tpm1.6-silenced cells cultured on unreleased CG. Both GM6001 and MMP9 inhibitor completely suppressed TGF-β1-induced CG degradation in Tpm1.6-silenced cells, and the quantification is shown in (G) and (H), respectively. Data are represented as mean ± SEM; n = 3 independent experiments were performed; two-way ANOVA, ∗∗∗∗p < 0.0001.

(D, E, I, J) Western blotting results showed the inhibition of α-SMA (non-canonical podosome) or cortactin (canonical podosome marker) markedly reduced TGF-β1-induced collagen degradation exerted by Tpm1.6-silenced cells, and the quantification is shown in (I) and (J), respectively. Data are represented as mean ± SEM; n = 3 independent experiments were performed; one-way ANOVA (I), two-way ANOVA, ∗∗p < 0.01, ∗∗∗∗p < 0.0001.

(K) Cy3-conjugated gelatin was coated on a glass coverslip, and further silenced-control and Tpm1.6-silenced NRK-49F cells were cultured on it with or without TGF-β1 and MMP9 inhibitor for 48 h. The fluorescence loss indicates the gelatin degradation. TGF-β1-treated Tpm1.6-kd fibroblasts showed a matrix-degraded manner, and the treatment of MMP9i blocked the matrix degradation. Scale bars, 50 μm.

(L) The observation of α-SMA dots (cyan) and the degraded region in the TGF-β1-treated Tpm1.6-kd group. Scale bars, 20 μm.

(M–P) Silenced-control and Tpm1.6-silenced NRK-49F cells were transfected with liposome-embedded shRNA and siRNA against α-SMA (shSMA; M) and cortactin (siCortactin; N), and the cell lysates were analyzed by western blotting. Western blotting showed the protein levels of α-SMA or cortactin in cells silenced by shSMA or siCortactin, and quantification is shown in (O) and (P), respectively. Data are represented as mean ± SEM; n = 3 independent experiments were performed; Welch’s t test, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; data were compared to siNC.

Tpm1.6-depletion in Foxd1-lineage cells attenuates unilateral ureteral obstruction-induced myofibroblast activation and renal fibrosis

Foxd1-lineage cells had been identified as the critical mesenchyme for myofibroblast activation during renal fibrosis, including fibroblasts.^26,27 To more extensively target myofibroblast precursor, we generated a mouse model, FoxD1GC^−/+;Acta2^TpmKD, which expresses Tpm1.6-targeted miRNA mimetic (the target sequences is same as Tpm1.6-kd in vitro) in myofibroblasts after renal fibrosis induced by unilateral ureteral obstruction (UUO). In FoxD1GC^−/+;Acta2^TpmKD mice, Foxd1-lineage cells drove the Cre recombinase-mediated excision of a loxP-flanked sequence. This was followed by Acta2 promoter-driven production of miRNA mimetics and functional reporters, displaying fluorescence shift and secretory luciferase (SecNLuc), as illustrated in Figure 7A.

Figure 7.

Mesenchymal-targeted silencing of Tpm1.6 by induction of promoters of Foxd1 and Acta2 attenuates UUO-induced renal fibrosis

To conditionally silence Tpm1.6 in the mesenchymal population (Foxd1-lineage), we generated a transgenic mice model by mating between mesenchymal population-driven cre mice and the floxed mice containing Acta2-driven Tpm1.6-targeted miRNA (same sequence as the exon 2b-targeted RNA interference in vitro).

(A) Schematic shows the generation of FoxD1GC^−/+; Acta2^TpmKD mice. Acta2 promoter-driven FusionRed, SecNLuc, and Tpm1.6-targeted miRNA are achieved in Foxd1-lineage cells by the cre recombinase-mediated excision of a loxP-flanked sequence. EmGFP is expressed in Acta2-expressing cells of FoxD1GC^−/−; Acta2^TpmKD mice or in Acta2-positive non-Foxd1-lineage cells of FoxD1GC^−/+; Acta2^TpmKD mice.

(B) Detection of SecNLuc (RLU), which was secreted into plasma derived from FoxD1GC^−/−; Acta2^TpmKD and FoxD1GC^−/+; Acta2^TpmKD mice with or without UUO surgery for 3 or 7 days. Data are represented as mean ± SEM; n = 3 mice.

(C) Fluorescent images showed the EmGFP and FusionRed-expressed population in kidney tissues. EmGFP is expressed at small arteries in sham kidneys and is upregulated in the myofibroblast population of UUO-challenged kidneys, and the FusionRed-positive population is increased in FoxD1GC^−/+; Acta2^TpmKD mice kidneys after UUO for 14 days. Scale bars, 100 μm.

(D) Primary kidney fibroblasts derived from the kidney cortex of FoxD1GC^−/−; Acta2^TpmKD and FoxD1GC^−/+; Acta2^TpmKD mice under CG culture with or without TGF-β1. The signals of EmGFP and FusionRed in fibroblasts were observed. Scale bar, 100 μm.

(E) The parenchymal thickness of the kidney in FoxD1GC^−/−; Acta2^TpmKD and FoxD1GC^−/+; Acta2^TpmKD mice after UUO surgery for 14 days. Data are represented as mean ± SEM; n = 3–4 mice; Welch’s t test, ∗∗p < 0.01; data were compared to FoxD1GC^−/−; Acta2^TpmKD. Scale bars, 1 cm.

(F–H) Representative images of Sirius red and Masson’s trichrome-stained kidney tissues in FoxD1GC^−/−; Acta2^TpmKD and FoxD1GC^−/+; Acta2^TpmKD mice with or without UUO surgery for 14 days, and the quantification of fibrotic region based on Sirius red-stained collagen fibrils is shown in (H). Data are represented as mean ± SEM; n = 4–9 mice; two-way ANOVA, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Scale bars, 100 μm.

(I) Fluorescent images represented the myofibroblast population (α-SMA-positive cells) and the localization of SGLT2, AQP1, and LRP1 in FoxD1GC^−/−; Acta2^TpmKD and FoxD1GC^−/+; Acta2^TpmKD mice kidneys with or without UUO surgery. Scale bars, 100 μm.

To assess the miRNA induction, we detected the luciferase activity and fluorescence shift in FoxD1GC^−/+;Acta2^TpmKD mice after UUO for 3 to 14 days. We showed that FoxD1GC^−/−;Acta2^TpmKD mice exhibited minimal levels of SecNLuc in plasma, whereas FoxD1GC^−/+;Acta2^TpmKD mice displayed markedly enhanced levels of SecNLuc in plasma after UUO for 3 and 7 days (Figure 7B). In the kidneys of FoxD1GC^−/−;Acta2^TpmKD mice, EmGFP was only expressed in small arteries, and there was no signal of FusionRed. UUO induced a marked increase in EmGFP-positive cells in the renal interstitium. In kidneys of FoxD1GC^−/+;Acta2^TpmKD mice, the distribution and levels of EmGFP and FusionRed were not altered; however, UUO promoted an increase in FusionRed signal (Figure 7C). Based on the results of plasma luciferase and kidney fluorescence shift between FoxD1GC^−/−;Acta2^TpmKD and FoxD1GC^−/+;Acta2^TpmKD mice, we confirmed that Tpm1.6-targeted miRNA mimetic may be produced in UUO-challenged kidneys of FoxD1GC^−/+;Acta2^TpmKD mice. Moreover, we cultured renal fibroblasts, which were derived from kidney cortex of FoxD1GC^−/−;Acta2^TpmKD and FoxD1GC^−/+;Acta2^TpmKD mice, on CG. TGF-β1 triggered a strong EmGFP signal in fibroblasts derived from control mice, whereas TGF-β1 promotes FusionRed signal in fibroblasts derived from FoxD1GC^−/+;Acta2^TpmKD mice (Figure 7D). To examine the silencing effect of Tpm1.6 in these primary fibroblasts, we analyzed exon 2 expression in total RNA extracts. The expression of exon 2b, not exon 2a, was silenced in fibroblasts derived from FoxD1GC^−/+;Acta2^TpmKD mice (Figure S6A). Taken together, we assessed not only the presence of functional reporters but also the efficacy of Tpm1.6-silencing in this murine model.

To evaluate the effect of the conditional depletion of Tpm1.6 on UUO-induced morphological differences in the kidney, we compared the parenchymal thickness of kidneys after 14 days of obstruction. Kidneys from FoxD1GC^−/+;Acta2^TpmKD mice showed significant recovery from UUO-induced parenchymal thinning of renal tissue (Figure 7E). To study the progression of fibrosis, we examined collagen deposition, myofibroblast activation, and tubular epithelial markers in mice kidneys. There was marked deposition of collagen fibrils in the kidneys of FoxD1GC^−/−;Acta2^TpmKD mice, which can be significantly attenuated by the targeted depletion of Tpm1.6 (Figures 7F–7H). On the other hand, the targeted depletion of Tpm1.6 significantly suppressed the UUO-induced increase of myofibroblasts (Figure 7I). To examine whether Tpm1.6-depletion may affect phenotype of renal proximal tubular cells, immunofluorescence for AQP1, SGLT2, and LRP2 was employed. UUO induced a marked loss of AQP1, SGLT2, and LRP2 after 7 days in kidneys. Mesenchymal depletion of Tpm1.6 alleviated UUO-induced decrease of AQP1, SGLT2, and LRP2 and retained the luminal distribution of SGLT2 and LRP2 in renal proximal tubules within 7 days (Figures 7I, S6B, and S6C). In summary, the mesenchymal depletion of Tpm1.6 decreased UUO-induced collagen deposition and myofibroblast activation and maintained the phenotype of proximal tubules.

Discussion

In our study, we investigated myofibroblast activation under an in vivo-like environment to mimic the initial stage of fibrosis progression. Fibroblasts were activated into myofibroblasts, which facilitated by actomyosin-associated CUs. We found that actomyosin contractility plays a crucial role in myofibroblast activation, suggesting a reciprocal relationship between the remodeled matrix and fibroblast activation. Kramann’s study also supported the notion that fibroblasts are the primary precursors of myofibroblasts in renal fibrosis.¹ Thus, fibroblasts likely serve as the initial precursor population for myofibroblast activation through actomyosin-mediated matrix remodeling during the onset of fibrosis progression.

We also confirm the dominant Tpm1 transcripts in pericyte cell line, CCL-226, and found that Tpm1.6/1.7 are major isoforms (Figure S3E). Although we did not observe the pericyte activation under CG condition (Figures S1A–S1C), recent single-cell sequencing has revealed that both fibroblasts and perivascular fibroblast-like cells (pericytes) are origins of scar-forming cells during human and murine renal fibrosis.¹ Foxd1-lineage cells, including fibroblasts and pericytes, had been identified as the critical population for myofibroblast activation during renal fibrosis.^26,27,28 From therapeutic perspective, we aimed to target critical mesenchymal precursors of myofibroblast activation and thus chose Foxd1-cre as our model.

In western blotting results, we sometimes observed α-SMA appearing as a doublet and TGF-β1 consistently increased the intensity of the upper band, particularly on CG (Figure 2H). Post-translational modifications may participate in such regulations such as phosphorylation, methylation, and glycosylation, which have been shown to be present in α-SMA. However, the role of these modifications in the nature of α-SMA remains poorly understood. There is a hypothesis that these modifications may contribute to the protein stability of α-SMA when fibroblasts are activated.

Targeting exon 2b and 9d for Tpm1 displayed different outcomes in myofibroblast activation and collagen degradation. Tpm1.6 is the isoform associated with α-SMA-associated stress fibers is implicated in epithelium-to-mesenchymal transition.^19,29 The presence of the critical Tpm may affect the fibroblast phenotype. Transcripts amplification, Taqman probe targeting in qPCR, and even Sanger sequence clearly showed that Tpm1.6 is the dominant isoform in renal fibroblasts. Figure 2G revealed that both exon 2a and exon 2b were silenced in miR9d cells, indicating alterations in other isoforms (Tpm1.3/1.4). However, the level of exon 9a of Tpm1 did not change in miR9d cells, suggesting that there is partial compensatory effect of other exon 9-contained isoforms (exon 9b, 9c) or other Tpm family (Tpm2, Tpm3, Tpm4). The cross-effect of other isoforms and off-target effects in miR9d may be the reason for affecting the minor change of phenotype. The presence of shared exons within isoforms poses a limitation in recognizing Tpm isoforms and targeted therapy. Nevertheless, we still figure out that exon 2b is the more specific target for Tpm1.6 silencing in fibroblasts.

Tpm1.6-depletion blocks TGF-β1-induced cell contractility, while simultaneously promoting collagen degradation through α-SMA dots in response to TGF-β1. Additionally, the reduction of external force by gel release enhances TGF-β1-induced collagen degradation in Tpm1.6-silenced cells (Figure 3A). Other findings from our lab show that increased matrix stiffness reduces podosome formation by upregulating stress fibers and cell contractility in renal fibroblasts, clearly demonstrating a mutual exclusivity between stress fiber and podosome formation (under submission). Such findings highlight that the phenotypic change is controlled by Tpm1.6 and the mechanical properties of the substratum.

The function of Tpm is to stabilize the actin filaments. It is likely that a specific isoform of Tpm may interact with diverse actin isoforms. Tpm1.6-depletion significantly changes TGF-β1-induced α-SMA expression pattern via the reduction of stress fiber-associated α-SMA and induction of membrane-associated α-SMA dots. Tpm1.6 seems to display more affinity with contractile α-SMA. Actin depolymerizing factor (ADF) and filamin have been reported to compete with Tpm for the association of actin filaments with isoform-specificity.^30,31,32 Tpm regulates the activity of myosin motors and the intracellular sorting of myosin II depending on isoforms.^33,34,35,36 Tropomodulin maintains actin homeostasis through distinct affinities for Tpm isoforms.^37,38,39 Taken together, distinct Tpm/actin affinity and other regulatory proteins may be involved in phenotypic change from contractile to matrix-degrading machinery exerted by Tpm1.6-depletion.

The nature of MMP9 and related metalloproteinases exhibit a dual-edged effect in fibrogenesis, with significant complexity in delineating their precise mechanistic contributions.⁴⁰ In this study, we identify a structural entity, α-SMA/MMP9 aggregates, that provides a mechanistic framework for investigating matrix remodeling dynamics in fibrotic progression. In TGF-β1-treated Tpm1.6-silenced fibroblasts, α-SMA silencing also reduces TGF-β1-induced collagen degradation (Figure 6D), suggesting that the role of α-SMA in α-SMA dots may involve serving as a carrier for MMP9 into the region where matrix degradation will occur. Moreover, Tpm1.6-silencing suppresses the canonical signaling of TGF-β1, p-smad2/3, indicating that the downregulation of α-SMA in Tpm1.6-silencing cells may be caused by the blockage of upstream signaling, rather than MMP degradation (Figure S4D).

TGF-β superfamily encompasses a diverse group of signaling molecules, including TGF-β1 and BMP, which activate distinct smad signaling pathways. TGF-β1 predominantly triggers p-smad2/3, while BMP promotes p-smad1/5/8 pathway. Recent studies revealed the activation of non-canonical smad1/5/8 axis upon TGF-β1, suggesting an alternative signaling mechanism.^41,42 Additionally, the TGF-β superfamily was implicated in osteoclast maturation, which is known for high collagen-degrading capability. This degraded phenotype bears resemblance to the findings observed in Tpm1.6-depleted fibroblasts. The involvement of smad1/5/8 signaling in osteoclast maturation may contribute to collagen degradation; therefore, the TGF-β1-smad1/5/8 axis may be responsible for the degraded phenotype in Tpm1.6-depleted fibroblasts. To fully comprehend the relationship between different smad regulations and collagen degradation, further investigation is warranted.

Our findings demonstrated that Tpm1.6 is a specific isoform in renal fibroblasts, which play pivotal roles in determination of contractile phenotype or collagen degradation upon TGF-β1. Tpm1.6 contributed to cell contractility, and the depletion switched the contractile phenotype to a matrix-degrading phenotype upon TGF-β1. Because the α-SMA dots are highly colocalized with MMP9, but not cortactin, such collagen digesting structure may be a kind of non-canonical podosome. The matrix-degrading phenotype is mediated by non-canonical podosomes. Such findings point out the crucial role of Tpm1.6 in the initial stage of renal fibrosis and demonstrate a way for the prevention of fibrosis progression and potential fibrosis therapy.

Limitations of the study

This study offers valuable insights into the role of Tpm1.6 in myofibroblast activation and collagen degradation, yet certain limitations exist. First, while TPM3 expression showed no compensatory upregulation in exon 9d-targeted cells, we did not assess TPM2 or TPM4, limiting our understanding of broader tropomyosin family interactions. Second, we did not test additional exon 9d-targeting sequences to clarify phenotypic differences between exon 2b and 9d targeting, despite validating exon 2b specificity with a second sequence. Third, our electron microscopy data could not fully characterize α-SMA dots, lacking the resolution to distinguish organized filaments from aggregates; higher-resolution imaging such as super-resolution microscopy, is needed. Finally, while we propose that α-SMA dots may facilitate MMP9 activation by localizing it to matrix interaction sites, the precise molecular mechanisms underlying this process remain speculative and require further investigation through advanced imaging or proteomic approaches. These gaps indicate directions for future research on Tpm1.6-specific effects and matrix degradation mechanisms in myofibroblasts.

Resource availability

Lead contact

Additional details and inquiries regarding materials and data should be addressed to, and will be handled by, the lead contact, Ming Jer Tang (mjtang1@mail.ncku.edu.tw).

Materials availability

The transgenic mice [B6; 129S4-Foxd1^{tm1(GFP/cre)Amc};Tg(Acta2-GFP, -FuRed/SecLuc/Tpm-miR)^Tg/0/J mice and B6-Tg(Acta2-GFP, -FuRed/SecLuc/Tpm-miR)^Tg/0/J mice] generated in this study is available from the lead contact upon reasonable request.

Data and code availability

• •All data associated with this study are publicly available in the Mendeley Data repository under the accession number (https://doi.org/10.17632/3cgf8d8cxm.1). The dataset includes processed figures, raw western blot images, and other materials used in the preparation of this article. Details and resource identifiers are provided in the key resources table. Additional information required to access or reanalyze the data may be obtained from the lead contact upon reasonable request.
• •This study did not generate original software, code, or computational models. All data analyses were performed using established, commercially available software, which is fully listed and referenced (including version numbers and RRIDs) in the key resources table.
• •All unique reagents, cell lines, mouse strains, and analytical tools used are detailed with catalog numbers and/or RRIDs in the key resources table. Additional materials and information are available from the lead contact upon reasonable request.

Acknowledgments

The authors appreciate for members in National Cheng Kung University Medical College Core Research Laboratory for technical experimental help. Funding was provided by the Ministry of Education (Higher Education SPROUT Project: International Center for Wound Repair and Regeneration), Ministry of Science and Technology (MOST 109-2634-F-006-021, MOST 110-2634-F-006-018, MOST 111-2634-F-006-009 to M.J.T.; MOST 109-2320-B-006-019 to C.H.K.).

Author contributions

C.L.W. and M.J.T. designed research; C.L.W., G.H.L., B.Y.C. and T.T.C. performed research; Y.K.W. and S.-L.L. contributed analytic tools; C.L.W. analyzed data; C.L.W. and M.J.T. wrote the article in consultation with C.H.K. and T.Y.W.

Declaration of interests

Dr. Tang and Dr. Lee hold a patent on “Method for measuring cellular mechanics in an in vitro fibrosis model” (patent no. I860001), which is utilized in the method details section of this article. No immediate family members of the authors hold relevant patents related to this work.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-α-SMA (clone 1A4)	Sigma-Aldrich	Cat# A5228; RRID: AB_262054
Rabbit polyclonal anti-Collagen-1a1	Novus Boster	Cat# NBP1-30054; RRID: AB_1968486 Cat# PA2140-2 RRID: AB_3082895
Mouse monoclonal anti-β1 integrin (clone 18)	BD Biosciences	Cat# 610468; RRID: AB_397840
Mouse monoclonal anti-N-cadherin (clone 13A9)	Merck Millipore	Cat# 05-915; RRID: AB_441927
Rabbit polyclonal anti-Myosin IIA	Sigma-Aldrich	Cat# M8064; RRID: AB_260673
Rabbit polyclonal anti-Myosin IIB	Sigma-Aldrich	Cat# M7939; RRID: AB_260669
Mouse monoclonal anti-Tropomyosin-1 (clone TM311)	Sigma-Aldrich	Cat# T2780; RRID: AB_261632
Sheep polyclonal anti-Tropomyosin-3	Sigma-Aldrich	Cat# AB5447; RRID: AB_177476
Rabbit monoclonal anti-Phospho-Smad2 (Ser465/467)/Smad3 (Ser423/425) (D27F4)	Cell signaling	Cat# 8828; RRID: AB_2631089
Rabbit polyclonal anti-Smad2/3	Cell signaling	Cat# 3102; RRID: AB_10698742
Rabbit polyclonal anti-Cortactin	Cell signaling	Cat# 3503; RRID: AB_2115160
Rabbit polyclonal anti-MMP9	Abcam	Cat# ab38898; RRID: AB_776512
Rabbit polyclonal anti-SGLT2	Proteintech	Cat# 24654-1-AP; RRID: AB_2750601
Rabbit monoclonal anti-AQP1	Novus	Cat# NBP2-67162; RRID: AB_3353461
Rabbit polyclonal anti-LRP2	Abcam	Cat# ab76969; RRID: AB_10673466
Goat anti-Mouse IgG (H + L)-HRP	CROYEZ	Cat# C04001;RRID: AB_10015289
Goat anti-Rabbit IgG (H + L)-HRP	CROYEZ	Cat# C04003;RRID: AB_2722564
Donkey anti-Sheep IgG (H + L) Secondary Antibody, HRP	Invitrogen	Cat# A16041;RRID: AB_2534715
Invitrogen Goat anti-Mouse IgG (H + L) Alexa Fluor 488	Molecular Probes	Cat# A11001;RRID: AB_2534069
Invitrogen Goat anti-Mouse IgG (H + L) Alexa Fluor 594	Molecular Probes	Cat# A11005;RRID: AB_141372
Invitrogen Goat anti-Mouse IgG (H + L) Alexa Fluor 405	Molecular Probes	Cat# A31553;RRID: AB_221604
Invitrogen Goat anti-Rabbit IgG (H + L) Alexa Fluor 488	Molecular Probes	Cat# A11008;RRID: AB_143165
Invitrogen Goat anti-Rabbit IgG (H + L) Alexa Fluor 594	Molecular Probes	Cat# A11012;RRID: AB_2534079
Invitrogen Goat anti-Rabbit IgG (H + L) Alexa Fluor 405	Molecular Probes	Cat# A31556;RRID: AB_221605
Bacterial and virus strains
One Shot® TOP10 Chemically Competent E. coli	Thermo Fisher	Cat# C404010
Biological samples
Primary renal fibroblasts	C57BL/6 mouse kidney tissue	Isolated as described in STAR Methods section
Chemicals, peptides, and recombinant proteins
TGF-β1	PeproTech	Cat# 100-21C
Rat tail collagen	Corning	Cat# 354236
(±)-Blebbistatin	Tocris Bioscience	Cat# 1760 CAS: 674289-55-5
ML 7 hydrochloride	Tocris Bioscience	Cat# 4310 CAS: 110448-33-4
Y-27632 dihydrochloride	Tocris Bioscience	Cat# 1254 CAS: 129830-38-2
GM6001 (pan-MMP inhibitor, 20 μM)	Sigma-Aldrich	Cat# 364205 CAS: 142880-36-2
SIS3	Tocris Bioscience	Cat# 5291 CAS: 521984-48-5
SB431542	Sigma-Aldrich	Cat# 616461 CAS: 301836-41-9
Critical commercial assays
QCM™ Gelatin Invadopodia Assay (Red)	Millipore	Cat# ECM671
Cy3-CHP	3Helix	Cat# RED300
Deposited data
Mendeley Data	https://doi.org/10.17632/3cgf8d8cxm.1	N/A
Experimental models: Cell lines
Rat: NRK-49F	Bioresource Collection and Research Center (BCRC)	BCRC Number: 60084
Mouse: CCL-226 (Clone 8)	the Global Bioresource Center (ATCC)	C3H/10T1/2
Rat: NRK-52E	Bioresource Collection and Research Center (BCRC)	BCRC Number: 60086
Porcine: LLC-PK1	Bioresource Collection and Research Center (BCRC)	BCRC Number: 60080
Experimental models: Organisms/strains
C57BL/6 mice	Laboratory Animal Center, College of Medicine, National Cheng Kung University	N/A
B6; 129S4-Foxd1tm1(GFP/cre)Amc/J cre mice (FoxD1GC mice)	Shuei-Liong Lin’s lab from National Taiwan University	RRID:IMSR_JAX:012463
B6-Tg(Acta2-GFP, -FuRed/SecLuc/Tpm-miR)Tg/0/J mice (Acta2TpmKD mice)	National Laboratory Animal Center in Taiwan	N/A
B6; 129S4-Foxd1tm1(GFP/cre)Amc;Tg(Acta2-GFP, -FuRed/SecLuc/Tpm-miR)Tg/0/J mice (FoxD1GC−/+;Acta2TpmKD)	laboratory-bred in National Laboratory Animal Center in Taiwan	N/A
Oligonucleotides
siNC: UUCUCCGAACGUGUCACGUTT	GenePharma	N/A
siCortactin (four siRNAs mixture): GCCAUGAGUACCAGUCAAATT, GUGGCAAAUACGGAAUUGATT, CCCAGAAAGACUAUGUAAATT, CUGCUGGCGAUGAUGAAAUTT;	GenePharma	N/A
qPCR primers (listed in Table S1)	Protech Technology Enterprise Co., Ltd.	N/A
Recombinant DNA
pcDNA™6.2-GW/EmGFP-miR vector in BLOCK-iT™ Pol II miR RNAi Expression Vector Kit	Thermo Fisher	Cat# K493600
shSMA#1: pLKO.1-Puro vector with GCAAGTGATCACCATCGGAAA	RNAiCore in Academia Sinica	TRCN0000116933
shSMA#2: pLKO.1-Puro vector with GCATCCACGAAACCACCTATA	RNAiCore in Academia Sinica	TRCN0000091664
Software and algorithms
ImageJ	Schneider et al.43	https://imagej.nih.gov/ij/
MATLAB	Mathworks	https://www.mathworks.com/products/MATLAB.html
Prism 9	GraphPad Software	https://www.graphpad.com/
NanoScope Analysis (V1.9)	Bruker Corporation	https://www.bruker.com/en.html
Other
TOOLS Mycoplasma Detection Kit	BioTools	Cat# TTB-GBC8
DMEM, powder, high glucose	Gibco	Cat# 12100046
DMEM, powder, low glucose	Gibco	Cat# 31600034
Penicillin G sodium salt	Sigma-Aldrich	P3032-25MU; CAS: 69-57-8
Streptomycin sulfate salt	Sigma-Aldrich	Cat# 3810-74-0; CAS: 3810-74-0
dimethyl sulfoxide (DMSO)	Aventor	Cat# 0231-500 ML; CAS: 67-68-5
blasticidin	Sigma-Aldrich	Cat# 15205
Puromycin dihydrochloride	Sigma-Aldrich	Cat# P8833; CAS: 58-58-2
Pierce ECL Western blot substrate	Thermo Fisher	Cat# 32106
Ponceau S	Sigma-Aldrich	Cat# P7170 CAS: 6226-79-5
SuperBlock buffer	Thermo Fisher	Cat# 37516
Phalloidin-TRITC	Thermo Fisher	Cat# P1951
Phalloidin-/Alexa 647 nm	Thermo Fisher	Cat# A22287
Hoechst 33258 (Sigma-Aldrich)	Thermo Fisher	Cat# H3569
TRIzol	Invitrogen	Cat# 15596018
PrimeScript RT Reagent Kit reverse transcriptase	Takara Bio	Cat# RR037B
Taqman probes (6-FAM and TAMRA)	Protech Technology Enterprise	N/A
SYBR Green	Applied Biosystems	Cat# 4309155
Liquid PDMS mixture	BingBond	Cat# PM5040C
PF127	Invitrogen	Cat# P6866
FITC	Sigma-Aldrich	Cat# F3651
DAB substrate kit	Abcam	Cat# ab64238
CCK-8 Plus	IMT Formosa New Materials	Cat# 291
Dynabeads™ protein G	Thermo Fisher	Cat# 10004D
Hilymax liposome	Dojindo Molecular Technologies	Cat# H357
StepOnePlus real-time PCR machine	Applied Biosystems	N/A
FACS Aria	BDbiosciences Core research laboratory in National Cheng Kung University Hospital	N/A
Confocal microscope (FV3000)	Olympus Corporation	FV3000
The Silicon Chips for mPAD	Professor Yang-Kao Wang from Department of Cell Biology and Anatomy, National Cheng Kung University	N/A
Atomic force microscopy (Cantilever: PFQNM-LM-A-CAL)	BioScope Resolve, Bruker Corporation	N/A
JEM1400 electron transmission microscope	JEOL Co.,Ltd	JEM1400

Experimental model and study participant details

Cell lines and pharmacological treatment

NRK-49F (Cat# 60084), CCL-226 (C3H/10T1/2, Clone 8), NRK-52E (Cat# 60086) and LLC-PK1 (Cat# 60080) cells were purchased from Bioresource Collection and Research Center (BCRC) and the Global Bioresource Center (ATCC). NRK-49F cells (renal fibroblast), primary renal fibroblasts at passage number 6 or below, CCL-226 cells (pericyte), NRK-52E and LLC-PK1 cells (renal epithelium) used for the studies. All cell lines used in this study, including NRK-49F (rat renal fibroblast), CCL-226 (mouse pericyte), NRK-52E (rat renal epithelial cell), and LLC-PK1 (porcine renal epithelial cell), were obtained either from the BCRC or from the ATCC/Global Bioresource Center, both of which are internationally recognized repositories where cell lines undergo standard authentication procedures before distribution. While specific short tandem repeat (STR) profiling was not performed for these established rodent and porcine cell lines, the sourcing from such reputable biobanks ensures a high level of quality control and species verification, which is the accepted practice for primary and continuous non-human lines in the field. Cells were tested negative for Mycoplasma assay using TOOLS Mycoplasma Detection Kit (BioTools). Cells were cultured on culture dish, polyacrylamide gel, collagen gel (CG) or gel-coated dish (Co). TGF-β1 (10 ng/mL; PeproTech) is a typical treatment for myofibroblast activation after cell attachment. Primary renal fibroblasts were expanded from the kidney of C57BL/6 mice as described by previous study.⁴⁴ Briefly, sections of renal cortex were adhered on gelatin-coated dish and cultured in 1xDMEM/20% FBS condition for 3 days. Fibroblasts proliferated from the explants were cultured and underwent the following experiments under the maximal of six passages. Primary fibroblasts were confirmed by α-SMA and collagen-1a1 staining on culture dish.

To validate the important role of CU-mediated force in TGF-β1-induced myofibroblast activation, blebbistatin (actomyosin blocker, 20 μM; Cat# 1760), ML-7 (MCLK inhibitor, 10 μM; Cat# 4310/10), and Y-27632 (ROCK inhibitor, 20 μM; Cat# 1254/1) were used for treatment in in vitro. To examine the signal transduction of TGF-β1, we used SIS3 (smad3 inhibitor, 10 μM; Tocris, Cat# 5291/10) and SB431542 (TGFβRI inhibitor, 20 μM; Sigma-Aldrich, Cat# 616461) for treatment. To examine the role of potential CU, Tpm1.6, in myofibroblast activation, we silenced Tpm1.6 in NRK-49F cells using the pcDNA™6.2-GW/EmGFP-miR vector from the BLOCK-iT™ Pol II miR RNAi Expression Vector Kit (Thermo Fisher; Cat# K493600). Hilymax liposome (Dojindo Molecular Technologies; Cat# H357)-embedded pcDNA™6.2-GW/EmGFP-miR expressing vector with non-silencing control (GAAATGTACTGCGCGTGGAGACGTTTTGGCCACTGACTGACGTCTCCACGCAGTACATTT) or Tpm1.6-targeted sequences (exon 2b: CAAATACTCCGAGGCTCTCAA; exon 2b#2: TGGTGTCACTGCAAAAGAAAC; exon 9d: CTTTACTGGAGCTAAACAACA) was used for transfection. After selection of blasticidin for 1 week, cells were sorted by the signals of EmGFP using FACS Aria (Core research laboratory in National Cheng Kung University Hospital) for the enhancement of purity and efficiency of silenced cells. Due to the higher specificity of exon 2b-targeting in Tpm1.6-silencing (Figure 2G), the exon 2b-targeting group was labeled as the Tpm1.6-kd group. For the silencing of α-SMA and cortactin, we transfected silenced-control and Tpm1.6-silenced cells with Hilymax liposome-embedded sh-α-SMA constructs (shNC: pLKO.1-Puro vector, shSMA#1: pLKO.1-Puro vector with GCAAGTGATCACCATCGGAAA, shSMA#2: pLKO.1-Puro vector with GCATCCACGAAACCACCTATA; RNAiCore in Academia Sinica), and cortactin siRNA (siNC: UUCUCCGAACGUGUCACGUTT; four siRNAs mixture: GCCAUGAGUACCAGUCAAATT, GUGGCAAAUACGGAAUUGATT, CCCAGAAAGACUAUGUAAATT, CUGCUGGCGAUGAUGAAAUTT; GenePharma). The silenced efficiency was confirmed by Western blotting (Figures 6M and 6N).

Animals

All mouse experiments were approved by the Animal Care and Use Committee at National Cheng Kung University with the animal experimental guidelines (IACUC number: 109326). Mice were housed under a 12-h light/dark cycle; surgery was conducted at surgery room of Laboratory Animal Center. Because Foxd1-lineage cells had been identified as the critical population for myofibroblast activation during renal fibrosis,^26,27 we generated B6; 129S4-Foxd1^{tm1(GFP/cre)Amc};Tg(Acta2-GFP, -FuRed/SecLuc/Tpm-miR)^Tg/0/J mice (FoxD1GC^−/+;Acta2^TpmKD) by B6; 129S4-Foxd1^{tm1(GFP/cre)Amc}/J cre mice (FoxD1GC mice; RRID:IMSR_JAX:012463) and B6-Tg(Acta2-GFP, -FuRed/SecLuc/Tpm-miR)^Tg/0/J mice (Acta2^TpmKD mice) to silence Tpm1.6 in Foxd1-lineage cells (Figure 7).

FoxD1GC mice were from by Shuei-Liong Lin’s lab from National Taiwan University; Acta2^TpmKD mice (generated by National Laboratory Animal Center in Taiwan) obtained a bacterial artificial chromosome (BAC) vector containing Acta2 promoter-driven emerald green fluorescent protein (EmGFP), red fluorescent protein (FusionRed), secreted nano-luciferase (SecNLuc), and Tpm1.6-targeted miRNA mimetic which sharing the same target sequences in Tpm1.6-kd cells (miR2b) in vitro (Figure 7). FoxD1GC^−/+;Acta2^TpmKD mice were obtained by the mating between FoxD1GC mice and Acta2^TpmKD mice, and offsprings were genotyped by sequence of cre recombinase and FusionRed. In FoxD1GC^−/+;Acta2^TpmKD mice, Foxd1-lineage cells drove Cre recombinase-mediated excision of a loxP-flanked sequence. This was followed by Acta2 promoter-driven the production of miRNA mimetic and functional reporters, displaying fluorescence shift and SecNLuc. Acta2^TpmKD mice are employed as control group. The studies were performed on a mixed background consisting of 129/B6. For fibrosis induction, randomly decided 8-week-old male mice (n = 3–5 for sham group; n = 4–9 for surgery group), weighing 20–25g, were received unilateral ureteral obstruction (UUO) surgery as previously described.^4,45 Briefly, the left ureter was ligated by 5-0 surgical nylon sutures after flank incision for 3, 7, or 14 days. The kidneys were confirmed to exhibit a swelling phenotype and were harvested at individual time points. Sirius red and Masson’s trichrome staining were utilized for the determination of fibrotic region; α-SMA-, SGLT2-, AQP1, LRP2-staining was used for the observation of myofibroblast population and the phenotype of proximal tubules.

Method details

Western blotting

Protein samples derived from cells were separated by SDS-PAGE and transferred onto PVDF membranes, followed by HRP-conjugated secondary antibodies (CROYEZ) were further used for the recognition of primary antibodies, and visualized by the Pierce ECL Western blot substrate (Thermo Fisher). The intensity of target proteins was analyzed by ImageJ and further normalized with internal control (GAPDH or β-actin for cell lysates; Ponceau S for culture media).

Proteins on PVDF membranes were captured by antibodies against β1 integrin (BD Biosciences; Cat# 610468), N-cadherin (Merck Millipore; Cat# 05–915), myosin IIA (Sigma-Aldrich; Cat# M8064)/IIB (Sigma-Aldrich; Cat# M7939), Tpm-1 (TM311, Sigma-Aldrich; Cat# T2780), α-SMA (Sigma-Aldrich; Cat# A5228), Tpm-3 (Sigma-Aldrich; Cat# AB5447), p-smad2/3 (Cell signaling; Cat# 8828), smad2/3 (Cell signaling; Cat# 3102), and collagen 1a1 (Novus, Cat# NBP1-30054; Boster, Cat# PA2140-2).

Immunofluorescence staining

OCT-embedded tissue sections and cultured cells were fixed by 4% paraformaldehyde, and washed by PBS thrice. Tissue sections and fixed cells were permeabilized by 0.1% Triton X-100, for 15 min. Subsequently, washed and blocked by SuperBlock buffer (Thermo Fisher). After staining of primary and secondary anibodies, fluorescence was captured by confocal microscope (FV3000, Olympus Corporation). Primary antibodies against Tpm-1 (Sigma-Aldrich; Cat# T2780), α-SMA (Sigma-Aldrich; Cat# A5228), cortactin (Cell signaling; Cat# 3503), MMP9 (abcam; Cat# ab38898), SGLT2 (Proteintech; Cat# 24654-1-AP), and AQP1 (Novus; Cat# NBP2-67162) were used for the staining, and further captured by Alexa 405nm/488nm/594 nm conjugated secondary antibodies (Thermo Fisher). Fluorescent studies including F-actin and nucleus observation co-stained with phalloidin-TRITC/Alexa 647 nm (Thermo Fisher) and Hoechst 33258 (Sigma-Aldrich) under the incubation of secondary antibodies. The podosome formation was assessed by double-positive staining of cortactin and F-actin.

PCR and quantitative PCR

Total cell RNAs were extracted by TRIzol (Invitrogen), and complementary DNA was synthesized from 1000 ng of total RNA with oligo-dT/random primers and the reverse transcriptase (Takara). PCR studies were performed by the amplification of exons or transcripts of Tpm. The cyclic number in conventional PCR is twenty-five. Taqman probes, SYBR Green and StepOnePlus real-time PCR machine (Applied Biosystems) were used in quantitative PCR studies. Ct values were normalized to β-actin, and relative expression was calculated from 2ΔΔCt calculation. Tpm1.6 and 1.7 were determined independently by using Taqman probe conjugated with 6-FAM-BHQ-1 or TAMRA-BHQ-2. The positive signals of 6-FAM and TAMRA were verified by over-amplification using amplified PCR product. PCR studies were performed by the amplification of the exons or transcripts of Tpm1 (1a, 2a, 2b, 9a, 9d), exon 9d of Tpm3, and β-actin primers at 95°C for 30 s, 60°C for 30 s and 72°C for 30 s (25 cycles for conventional PCR in Figures 2B and 2C; 40 cycles for quantitative PCR studies). The sequences primers are listed in Table S1.

Preparation of collagen gel and collagen gel contraction assay

The collagen scaffolds (final collagen concentration: 1 mg/mL) were generated by the mixing of Rat tail type I collagen (Corning), 5.7x DMEM, 2.5% NaHCO₃, 0.1M HEPES, 0.17M CaCl₂, and 1N NaOH, and the mixture was poured on culture dish and waited for polymerization at least for 30 min. Collagen gel solution was used for the rinse of the surface of culture dish in the collagen gel-coated dish which is the control condition of collagen gel.

For contraction assay, before gelation, gel solution was mixed with cells. The gel-cell mixture was plated on 6-well dish and underwent following culture and treatment. After 24 h incubation and treatment, cell-containing collagen gel was released by tips from plates to remove the interaction between gel and culture dish, and then the area of collagen gels was observed.

Culture media precipitation

Harvested culture media derived from cells cultured on soft collagen gel (NRK-49F cells or NRK-49F cells harboring different RNA interference) for 2 days. Removed cell debris by centrifugation at 250g for 5 min. Total protein in culture media were precipitated by cold acetone with 2-fold volume and further waited for 1 h at −20°. Collected the protein lysates by centrifugation at 13,200g for 10 min and suspend the proteins by sample buffer for SDS-PAGE. Ponceaus S staining was used as the loading control of culture media.

Micropost array detector (mPAD)

Polydimethylsiloxane (PDMS) micropost arrays were fabricated on cover glass (22 mm × 22 mm) as previously mentioned.⁴⁶ The silicon chips were kindly provided from Professor Yang-Kao Wang (Department of Cell Biology and Anatomy, National Cheng Kung University). Liquid PDMS mixture (PM5040C A, BINGBONG) was poured onto a silicon chip, and then cured at 120°C overnight. The cured PDMS were peeled off and generated an array of holes, thereby creating a negative mold. The negative mold was ozonized and silanized to avoid bonding with PDMS. Next, liquid PDMS mixture was poured onto the negative mold, cured at 120°C overnight, and PDMS was peeled off from the negative mold, then the array of microposts was produced. Additionally-produced PDMS stamps were sterilized by ozone, followed by incubation in aqueous solution of fibronectin for 1h. The PDMS stamp coated with fibronectin were placed on the sterilized microposts for adhesion coating and then labeled with DiI for 1h. Finally, the coated micropost arrays were blocked with PF127 (Invitrogen), and ready for following in vitro studies. Quantitative analysis of cellular contractility was based on the top and bottom images of DiI-labeled posts by confocal microscope (FV3000, Olympus Corporation). The bending degree of each post in response to applied traction force as calculated by finite element model analysis. Olympus FV-3000 software and custom-developed MATLAB program were used for analysis of cell traction force.⁴⁷

Co-axis system of atomic force microscope and confocal microscope

NRK-49F cells or NRK-49F cells harboring RNA interference were transfected with LifeAct RFP-expressed vector which allowing RFP incorporate with actin filament for the cell position at the live-cell imaging system. LifeAct RFP-expressing cells were cultured on FITC-conjugated collagen gel which can be observed without following staining. The preparation of FITC-conjugated collagen gel is same as collagen gel preparation above, but the collagen stock was conjugated by FITC and further dialysis. After cell culture and treatment under FITC-conjugated collagen gel condition for 4 days, the cell and collagen fibrils were observed and detected by co-axis system of atomic force microscopy and confocal microscopy. After probe calibration and position adjustment, this living in vitro system was scanned using optical system that combined laser-scanning confocal microscopy (FV3000, Olympus Corporation) and atomic force microscopy (BioScope Resolve, Bruker Corporation; cantilever: PFQNM-LM-A-CAL), performed as previously mentioned with minor modification.⁴⁸ Briefly, AFM studies were operated in the mode of FASTForce Volume at a scanning rate of 20 Hz. Indentation force was adjusted to maintain 2 nm deformation. Force-distance curves in the indicated area were selected and analyzed with NanoScope Analysis (V1.9, Bruker Corporation) based on Sneddon model. The diameter of collagen fibrils was measured by AFM-Height detection. The individual values of cell stiffness and fibril stiffness was displayed on the figures as dots.

Transmission electron microscope (TEM) inspection

NRK-49F cells or NRK-49F cells harboring RNA interference were cultured on collagen gel with or without TGF-β1 induction for 24 h. The cell-matrix scaffold was fixed by a solution of 2.5% glutaraldehyde for 1 h, which served to crosslink and stabilize the structure. Further, a solution of 0.1M cacodylate buffer was applied to enhance the membrane permeability, followed by a solution of 1% osmium to further fix the cells and proteins in place. After the fixation process was complete, the scaffolds were washed by PBS and gradually dehydrated using a series of ethanol solutions ranging from 70% to 100%. The scaffolds were embedded in Spurr resin to provide support for sectioning. The embedded samples were sectioned onto a grid with a thickness of 70 nm using a Leica EM UC7 microtome. The grid with samples was blocked for 30 min, then incubated with an antibody against α-SMA (Sigma-Aldrich; Cat# A5228) for 2 h, followed by incubation with a nanogold-conjugated secondary antibody for 1 h. To enhance the contrast of the sections, grids were stained with uranyl acetate and lead citrate and the resulting samples were then investigated by a JEM1400 electron transmission microscope to observe the structure features of the scaffold and α-SMA-positive aggregates at high resolution.

Immunohistochemistry

Paraffin-embedded tissue sections were deparaffinized by xylene three times and further soaked in serial gradient of ethanol solution with different concentration (100%–50%). Antigen retrieval was conducted in citrate buffer (pH = 6.0) with heat by boiling water. The activity of endogenous peroxidase was removed by 6% H₂O₂ for 10 min. After blocking by SuperBlock buffer (Thermo Fisher), primary antibody against LRP2 was applied on sections overnight. The sections were washed by PBST three times and applied by HRP-secondary antibody for 1 h. The signal of target proteins was enhanced by DAB substrate kit (abcam; Cat# ab64238) for 10 min, and further the nucleus was stained by hematoxylin for 5 min.

Preparation of polyacrylamide gel

Preparation of polyacrylamide gel was according to Wei et al.⁴⁵ The stiffness (0.2, 2, 20 kPa) of polyacrylamide gels was modified by different the concentration of bisacrylamide and acrylamide with TEMED and ammonium persulfate (Sigma-Aldrich). Acrylic acid (0.3%) was added into PA gel for substrate crosslinking. After polymerization, the polyacrylamide gels were washed by 0.1M MES [2-(N-morpholine) ethansulfonic acid] (pH = 6). The freshly-prepared conjugation reagent, EDC [1-ethyl-3-(dimethalaminopropyl)] was applied onto the surface of the polyacrylamide gel for 15 min to activate the 0.3% acrylic acid. After following wash by 0.1 M MES, 100 μg/mL monomeric type I collagen (Corning) in 0.1M MES was applied for the coating of polyacrylamide gels for overnight at 4°C. The coated polyacrylamide gels were freshly-prepared for the following in vitro studies.

Collagen hybridizing peptide (CHP) recognition

NRK-49F cells or NRK-49F cells harboring RNA interference were cultured on FITC-conjugated collagen gel with or without TGF-β1 induction for 24 h and co-treated with DMSO, blebbistatin (actomyosin blocker, 20 μM; Tocris), and GM6001 (pan-MMP inhibitor, 20 μM; Sigma-Aldrich). The cell-matrix scaffolds were fixed by 4% paraformaldehyde for 1 h and washed by PBS thrice. The 5 μM Cy3-CHP (3Helix) was prepared by heating to 80°C for 5 min. Heated and cooling Cy3-CHPs were used to apply on fixed cell-matrix scaffolds overnight at 4°C. After staining, gels were washed with four rounds for 30 min in PBS, followed by the staining of nucleus (Hoechst 33258) and F-actin (Alexa Fluor™ 647 Phalloidin) for cell position. The cell-matrix was observed by confocal microscope (FV3000, Olympus Corporation). The FITC-conjugated collagen fibrils as total fibrils were used for the normalization of collagen deformation. The intensity of CHP-positive fibrils vs. the intensity of FITC-collagen fibrils was analyzed by ImageJ.

Cell counting Kit-8 (CCK-8) assay

Silenced-control or Tpm1.6-kd cells were cultured on CG and incubated for 4 h, followed by TGF-β1 treatment. After 24 h of TGF-β1 induction, CCK-8 (10%; IMT Formosa New Materials)-contained growth medium was replaced for dehydrogenase activity. After incubation with CCK-8 solution for 2 h, the absorbance was measured at 450 nm.

Cy3-gelatin degradation assay

The QCM™ gelatin invadopodia assay kit was used (Millipore). Poly-L-lysine was coated on 12 mm round coverslip for 20 min. Diluted glutaraldehyde solution was incubated with Poly-L-lysine-coated glass for surface activation. Heated Cy3-conjugated gelatin mixture (60°C) was applied on coverslip for 10 min. Further wash and disinfect by PBS and 75% ethanol. After quenching residual free aldehydes by growth media, silenced-control or Tpm1.6-kd cells were cultured on Cy3-gelatin-coated coverslip for 48 h. Fix the cells by 4% paraformaldehyde and mount by mounting medium with DAPI (abcam). The fluorescence loss was captured by confocal microscopy (FV3000, Olympus Corporation).

Immunoprecipitation

Silenced-control or Tpm1.6-kd cells were cultured on 100 mm culture dishes with CG substrates. After TGF-β1 induction for 24 h, the cell-CG mixture was incubated with 1 mg/mL type I collagenase at 37°C for 20 min. The seperated cells was centrifuged at 300 g for 5 min, and furether lysed cells by non-redusing lysis buffer. 500 μg lyastes were used for each immunoprecipitation by mixing with Dynabeads™ protein G (Thermo Fisher; Cat# 10004D), α-SMA antibody (Sigma-Aldrich; Cat# A5228) overnight. The captured proteins were separated on magnetic rack, and further washed by non-reducing lysis buffer. Finally, the beads and separated protein lysates were mixed in sample buffer and heated at 95°C, and further removed beads on magnetic rack. The separated protein lysates would analyze by SDS-PAGE.

Quantification and statistical analysis

Statistics

GraphPad Prism (version 9.0) software was used for all statistical analyses. Unpaired two-tailed Student’s t-tests with Welch’s correction were applied to compare data between two groups; ordinary one-way ANOVA, followed by Tukey’s multiple comparisons, was used to compare multiple groups; and ordinary two-way ANOVA with Tukey’s multiple comparison test was applied for analysis involving multiple variables. Quantitative results are presented as mean values with standard error of the mean (SEM), along with the experimental unit (n), as specified in each figure legend. Statistical significance was defined as follows: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

All statistical details—including the specific tests used, exact value of n and what n represents (e.g., number of animals, number of cells, or number of independent experiments), as well as definitions of center (mean), and dispersion or precision measures (SEM)—are provided in the figure legends and figures for each experiment.

Published: August 6, 2025