Transgelin Expression in Activated Cancer-Associated Fibroblasts Regulates Stromal Contractility and Promotes Colon Cancer Progression.

1
Introduction
Cancer is a challenge and the big health problem to the whole world. A number of conventional therapies are available for the treatment of cancer including radiation therapy, chemotherapy, gene silencing, nutrient supplementation, surgery, etc. [1, 2, 3, 4, 5]. These approaches to treatment can produce significant adverse effects, hence limiting their therapeutic efficacy and compliance among patients. On the other hand, it has been observed that transgelin (TAGLN) protein promotes cancer progression via activation of cancer‐associated fibroblasts. Transgelin has been detected as a prognostic marker and a known actin‐binding protein that deforms the cytoskeleton presenting. Therefore it can be utilized for the diagnostic and therapeutic implication in cancer treatment [6, 7]. In particular, its role in cancer‐associated fibroblasts may be critical for modulating the tumor microenvironment.
Fibroblasts are primarily activated in cancer tissues, promoting tumor growth, metastasis, and therapy resistance through their biological and biophysical functions. Cancer‐associated fibroblasts (CAFs) provide a scaffold substrate on which tumor cells proliferate and induce paracrine factors that enhance tumor development [8, 9, 10]. CAFs also determine the biophysical features of tumors through active cell contractility and robust substrate secretion ability, which promote tumor cell migration and induce clinical symptoms such as intestinal obstruction [11, 12, 13]. Therefore, the identification and functional investigation of molecules involved in the pathophysiological features of CAFs are crucial for establishing a targeted therapies based on CAFs. In addition to local tissue‐resident fibroblasts, mesenchymal stem cells from distant sites may also be progenitors of CAFs [14, 15]. We previously reported that colonic subperitoneal fibroblasts (cSPFs) promote tumor growth and metastasis, and that conditioned medium (CM) stimulation of cSPFs upregulates the expression of various clinical biomarkers [16]. This tumor‐promoting ability is significantly stronger than that of submucosal fibroblasts in colonic tissue, suggesting that the heterogeneity of resident fibroblasts leads to variable tumor microenvironments, which may explain some clinical phenomena, e.g., the tumor spread–dependent clinical outcomes in gastrointestinal cancers. Additionally, among the upregulated genes in cSPFs after CM stimulation, transgelin, a member of the calponin family, has been identified as a prognostic marker and a known actin‐binding protein that deforms the cytoskeleton [17, 18, 19].
Therefore, tumor‐promoting CAF progenitors residing just beneath the peritoneal surface and transgelin, upregulated by cancer stimuli, may contribute to cancer development and metastasis. Transgelin is primarily expressed in smooth muscle cells (SMCs) of normal tissues and serves as a differentiation marker for SMCs [20]. Increased transgelin expression has been reported in some solid tumor tissues and is associated with lymph node metastasis in cancer tissues [21, 22]. However, its function in colon cancer has been investigated exclusively in cancer cell lines (HCT116 and SW480), and its anti‐oncogenic role in cancer cells has also been explored [23]. The discrepancy in the relationship between transgelin expression, biological function, and clinical outcomes can be attributed to the distribution of transgelin expression within cancer tissues, which comprise a variety of cell types.
We have previously shown that tumor invasion beyond the peritoneal elastic lamina of colorectal cancer induces peritoneal indentation and bowel obstruction, as well as peritoneal influences prognosis [24]. Furthermore, from the upregulated cSPF gene set after cancer CM stimulation, we identified several clinically available biomarker genes, including transgelin [13, 16, 25, 26]. Although transgelin expression distribution is not pathologically clear, our data suggest that transgelin upregulation in activated fibroblasts in cancer tissues plays an essential biological role associated with poor prognosis and bowel obstruction. Thus, transgelin can be a good candidate for CAF‐targeted therapies.
In this study, we aimed to investigate the pathophysiological role of transgelin in colon cancer to identify new targets for cancer therapy.

2
Material and Methods
2.1
Establishment of Primary Cell Lines
cSPFs were established from surgically resected material based on the clinical diagnosis of colorectal cancer. Only patients who did not undergo preoperative chemotherapy or radiotherapy were included to establish primary cell lines. Tissue fragments were collected, and primary cSPFs were established from the serosa of noncancerous areas of surgically resected sigmoid and rectal cancers. Further details can be found in previous studies [26]. The collected tissues were washed with phosphate‐buffered saline (PBS; KOHJIN BIO, Saitama, Japan) and shaken three times for 20 min at 37°C in PBS + 0.25% trypsin/EDTA (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). After shaking, the tissue pieces were removed and centrifuged at 1500 rpm for 2 min, and the supernatant was discarded. The remaining precipitate was incubated in MF medium (Toyobo, Osaka, Japan) supplemented with 1% penicillin/streptomycin (Sigma‐Aldrich, St. Louis, MO, USA) and then seeded onto a 10‐cm dish (Corning Life Sciences, Corning, NY, USA). The medium was changed daily, and after confirming that the cells had grown to confluence in a 10‐cm dish, the cells were collected and reseeded in a 10 cm dish at a fivefold dilution. The cells were then counted at passage 1. Passages 2–8 were used in subsequent experiments.

2.2
Cell Lines
The colorectal cancer cell line used was DLD‐1 (human colon adenocarcinoma cell line; ATCC, Manassas, VA, USA, RRID: CVCL_0248). DLD‐1 cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma‐Aldrich, St. Louis, MO, USA) containing 100 U/mL penicillin, 100 μg/mL streptomycin (Sigma‐Aldrich, St. Louis, MO, USA), and 10% fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) at 37°C and 5% CO2. Primary cultures of the harvested fibroblasts were cultured in MF medium at 37°C and 5% CO2. All experiments were performed with mycoplasma–free cells.

2.3
Fibroblast Stimulation by Cancer‐Cell‐Conditioned Medium
After seeding 1.0 × 106 DLD‐1 cells in a 10 cm dish and culturing in DMEM for 48 h, the medium was replaced with serum‐free DMEM, and the culture supernatant was collected after 24 h. The supernatant was processed using a 0.22 μm filter (Millipore, Merck KGaA, Darmstadt, Germany) and frozen at −80°C. We seeded 5.0 × 105 cSPFs onto 10 cm dishes, cultured in MF medium for 48 h, replaced the medium with serum‐free DMEM, and added cancer CM after another 24 h. The control group was treated with serum‐free DMEM. cSPFs were collected, and RNA and proteins were extracted.

2.4
Gene Knockdown Using Lentiviral Vectors
Gene knockdown cells were prepared using lentiviral vectors as previously described [27]. Briefly, 1.0 × 106 producer cells were seeded in a 6 cm dish (FALCON, Corning Incorporated, Corning, NY, USA) and cultured in RPMI 1640 for 24 h. The cells were then incubated in RPMI 1640 for 1 h. Next, 5 μg of lentiviral vector [either sh‐Luc (RIKEN), sh‐transgelin#1, or sh‐transgelin#2] was added to 500 μg of Opti‐MEM I Reduced Serum Medium (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) along with 3 μg pCMV‐VSV‐G‐RSV‐Rev and 3 μg pCAG‐HIV (Riken).
This solution was designated as Solution A. In addition, 20 μL Lipofectamine 2000 reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) was added to 500 μL Opti‐MEMI Reduced Serum Medium. This solution was designated as solution B. Solutions A and B were incubated at room temperature for 5 min and then mixed. The mixture was incubated at room temperature for 20 min, then added to producer cells, which were incubated for 24 h at 37°C and 5% CO2. After 24 h, the medium was replaced with RPMI 1640, and the cells were cultured for another 24 h. The culture supernatant was passed through a 0.45 μm filter (Millipore, Merck KGaA, Darmstadt, Germany), and 32 μL containing 1 μg/μL polybrene (Santa Cruz Biotechnology, Dallas, TX, USA) was added to the cSPFs seeded the previous day. The sh‐Luc, sh‐TAGLN #1, and sh‐TAGLN #2 sequences are listed in Table S1.

2.5
Quantitative Real‐Time Polymerase Chain Reaction (RT‐qPCR)
After 24 h following the addition of cancer cell–culture supernatant, cSPFs were washed with 10 mL PBS, cells were detached from a 10 cm dish with 1 mL 0.05% trypsin/EDTA (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), and pellets were prepared. RNA was extracted from the pellet using NucleoSpin RNA Plus (Takara Bio, Kusatsu, Shiga, Japan) according to the manufacturer's protocol. cDNA was synthesized from the extracted RNA using a PrimeScript RT Reagent Kit (Takara Bio, Kusatsu, Shiga, Japan). Briefly, total RNA was adjusted to a concentration of 100 ng/μL, and 500–1000 ng of RNA was used for each RT reaction. RT‐qPCR was performed using a Smart Cycler System (Takara Bio, Kusatsu, Shiga, Japan) with SYBR Premix Ex Taq (Takara Bio, Kusatsu, Shiga, Japan). mRNA expression levels were standardized using GAPDH.
Changes in gene expression levels between comparison groups were calculated using the calibration curve method. Primer sequences are listed in Table S2.

2.6
Western Blotting
After 24 h of CM addition, cSPFs were washed with 10 mL PBS, and the cells were detached from a 10 cm dish with 1 mL 0.05% trypsin/EDTA (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). Pellets were prepared, and RIPA buffer (Cell Signaling Technology, Danvers, MA, USA) with Complete Protease Inhibitor Cocktail (Roche, Basel, Switzerland) and PhosSTOP (Sigma–Aldrich, Merck KGaA, Darmstadt, Germany) were added. Western blotting was performed using a simple protein protocol. Protein detection was performed using an automated Jess capillary electrophoresis system (Simple Western system and Compass software; ProteinSimple, San Jose, CA, USA), according to the manufacturer's instructions. The primary antibodies used in this study were rabbit anti–human GAPDH (Cell Signaling Technology, Danvers, MA, USA) and mouse anti–human transgelin antibody (OriGene Technologies, Rockville, MD, USA).

2.7
Cell Contraction Assay
The collagen gel contractility of cSPFs was examined using a Cell Contraction Assay (Cell Biolabs Inc., San Diego, CA, USA). cSPFs were suspended in a mixture of 500 μL collagen solution, 5 × DMEM, and neutralization solution and seeded in 24‐well plates. After incubation at 37°C for 1 h, DMEM was added to the polymerized collagen gel and incubated for an additional 24 h incubation. The DMEM was removed, and CM or serum‐free DMEM (control) was added. Photographs of the gels were taken 48 h after CM stimulation, and the gel area was calculated using the WinROOF 2021 software. The number of cells encapsulated in the collagen gel was examined under five different cell count conditions (5 × 104, 1 × 105, 2 × 105, 4 × 105, and 8 × 105), and shrinkage was easily observed when the cells were stimulated with the cancer cell‐culture supernatant.

2.8
Xenograft Transplantation and Tumor Formation Assay
For the co‐transplantation of DLD‐1 cells (passage 3) and cSPFs into immunocompromised mice, 8‐ to 10‐week‐old SCID mice (Nippon Clare Co. Ltd., Tokyo, Japan) were used, with three groups per experiment: the DLD‐1 cell–alone group, the DLD‐1 cell and cSPF (control) co‐transplant group, and DLD‐1 cell and cSPF (transgelin knockdown) group. Luciferase was used as a control vector in the same manner. Each of the three groups was administered two immunocompromised mice at two dorsal sites. In the single‐transplant group, 1 × 106 DLD‐1 cells were prepared, and in the co‐transplant group, 1 × 106 DLD‐1 cells and cSPFs were prepared, suspended in 100 μL of PBS, and administered. Short and long tumor diameters were measured weekly for 4 weeks following cell administration. The tumor diameter was calculated using the formula [short diameter2 × long diameter/2], and the experimental results were evaluated by plotting changes in tumor diameter over 4 weeks.
Two different vectors (sh‐transgelin #1 and sh‐transgelin #2) were used for the transgelin knockdown of transgelin expressed in cSPFs in this experiment. However, because it was difficult to produce cells that could be treated simultaneously, the experiments were divided into two separate sets, depending on the vector used for knocking down transgelin.

2.9
Gene Expression Analysis
Gene expression analysis of cSPFs was performed using RNA sequencing (RNA‐seq). Three cSPFs, each established from a separate cancer tissue sample, were transfected with sh‐Luc or sh‐transgelin #1. After introducing the sh vector, RNA was extracted from stimulated and unstimulated cells by adding CM. The final analysis was performed using a 3 (cSPF type) × 2 (two vectors) × 2 (with or without CM) design for 12 samples. Total RNA samples were sent to Rhelixa Corporation for Bioanalyzer (Agilent) quality control and next‐generation sequencing analysis. Libraries were prepared using the NEBNext Poly(A) mRNA magnetic isolation apparatus (New England Biolabs [NEB], Hitchin, UK) and the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB). Rhelixa (Tokyo, Japan) was sequenced using a NovaSeq 6000 (Illumina, San Diego, CA, USA). The quality of raw paired‐end sequence reads was assessed using FastQC (version 0.11.7). DNA reads were mapped to the human reference genome using HISAT2 (version 2.1.0) with the default settings. For the downstream expression analysis, gene counts were generated from the mapped HISAT2 output and genes using featureCounts (version 1.6.3). The number of reads for each gene was normalized to the number of transcripts per million. Fragments per kilobase of transcript per million mapped reads (FPKM), FPKM‐upper quartile (FPKM‐UQ), and transcripts per million (TPM) values were calculated. For data visualization, the data were imported into several bioinformatics analysis programs such as iDEP.96 [28]. Differentially expressed genes (DEGs) were determined using DESeq2 with either a false discovery rate (FDR) < 0.2 and absolute log2‐fold‐changes > 1. Functional pathway and gene set enrichment analyses were performed using online tools, such as iDEP.96 (DEG data are shown in Table S3). The sequencing coverage and quality statistics for each sample are summarized in Table S3.

2.10
Patient Characteristics
This study included 388 consecutive patients who underwent curative surgery for colorectal cancer at the National Cancer Center Hospital East between April 2014 and March 2020. Surgical specimens were obtained from the institutional biobank, and only patients who had provided consent for biobank use, including those who consented through an opt‐out procedure approved by the institutional review board, were included. Clinical data and other patient information were collected from the hospital's electronic medical records. To monitor postoperative progress, all patients were followed up every 6 months for the next 5 years. Stage classification of colorectal cancer was performed using the eighth edition of the Japanese Classification of Colorectal, Appendiceal, and Anal Carcinoma.

2.11
Construction of Area‐Specific Tissue Microarrays
One tissue core, measuring 2 mm in diameter, was taken from a random region of each of the 388 colorectal cancer tissues and implanted into the recipient block using a tissue microarray (TMA; Azumaya, Tokyo, Japan). As the tissue was collected from a tumor area without any influence of the pathological diagnosis, areas of normal tissue or adipose tissue in the peripheral tumor area were included in some cases. Therefore, the following exclusion criteria were established using microscopy:
Cancerous tissue comprising less than 30% of the area.

Smooth muscle tissue comprising more than 30% of the area.

More than 30% of the tissue is missing.

Photographs of excluded cases are shown in Figure S1. Transgelin expression was strong in normal smooth muscle tissue and was included as an exclusion criterion (Figure S2 and the Transgelin expression is predominant in cancer stromal fibroblasts in human colorectal and other cancer tissues section). Twenty‐nine cases were excluded based on these exclusion criteria, and 359 cases were histologically and immunohistochemically analyzed.

2.12
Antibodies, Reagents, Immunohistochemistry, and Fluorescence Multiplex Immunohistochemistry
Immunohistochemical staining was performed using an autostainer (Ventana Benchmark, Tucson, AZ, USA). Paraffin‐embedded 5 μm tissue or microarray sections were incubated with the primary antibody for 1 h. The slides were counterstained with hematoxylin and diaminobenzidine as chromogenic agents before mounting. Fluorescence multiplex immunohistochemistry was performed using a tyramide signal amplification (TSA; PerkinElmer, Waltham, MA, USA) system with an Opal IHC kit (PerkinElmer, Waltham, MA, USA) according to the manufacturer's instructions, as previously described [29]. Quadruple fluorescence staining was conducted to examine transgelin and alpha‐smooth muscle actin (α‐SMA) expression for similarities. The primary antibodies used were mouse monoclonal anti‐human transgelin antibody (OriGene Technologies, Rockville, MD, USA; used at 1:200 dilution), rabbit monoclonal anti–human α‐SMA (SPRING BIOSCIENCE, Pleasanton, CA, USA; used at 1:200 dilution), Desmin (Nichirei, Tokyo, Japan; used without dilution), 4′,6‐diamidino‐2‐phenylindole (DAPI; Akoya Biosciences, Menlo Park, CA, USA; used without dilution), and AE1/3 antibody (Dako, Agilent Technologies, Santa Clara, CA, USA; used without dilution). Protein localization was examined using an all‐in‐one fluorescence microscope (Keyence, Osaka, Japan). A single staining with each antibody was performed before fluorescence immunostaining to ensure no leakage (Figure S3).

2.13
Evaluation of Area‐Specific Tissue Microarray Sections
High‐resolution slide images with hematoxylin and eosin and immunohistochemical staining were acquired from all tissues using a NanoZoomer 2.0‐HT slide scanner (Hamamatsu Photonics, Hamamatsu, Japan). All cores were analyzed using the Viewer Software (Hamamatsu Photonics, Hamamatsu, Japan). Images were taken at 4 × magnification and saved as JPEG files. The percentage of transgelin, α‐SMA, and desmin‐positive area ratios in the whole tissue was calculated using morphometric software, as described in a previous report [30] (WinROOF; Mitani Corporation, Tokyo, Japan). Figure S4 illustrates some of the data from the image analysis.

2.14
Statistical Analysis
For RT‐qPCR and cell contraction assays, Student's t‐test was used. The xenograft transplantation study used a one‐way factorial analysis of variance with a Bonferroni–Dunn post hoc test. Progression‐free survival (PFS) curves were plotted using the Kaplan–Meier method, and differences between the curves were analyzed using the log‐rank test. Transgelin expression and clinical characteristics were analyzed using the chi‐square test. p < 0.05 was considered statistically significant.

3
Results
3.1
Transgelin Expression Is Predominant in Cancer Stromal Fibroblasts in Human Colorectal and Other Cancer Tissues
The results of transgelin immunostaining in normal colon tissue are shown in Figure 1A. In normal tissues, transgelin expression is predominantly found in smooth muscle tissues, including the muscularis mucosa, the proper muscle layer, and vascular smooth muscle. Additionally, pericryptal fibroblasts were observed in the mucosal area. Only a faint expression was observed in epithelial cell components.
Immunostaining results for transgelin in colon cancer tissues are shown in Figure 1B,C. In colon cancer tissues, transgelin was expressed more predominantly in stromal fibroblasts than in cancer cells. The solid tumor areas showed faint transgelin expression (Figure 1B). In contrast, colon cancer tissues with prominent stroma (Figure 1C) showed high transgelin expression in spindle stromal fibroblasts.
Quadruple fluorescent antibody staining was performed to determine the cell types that expressed transgelin. Co‐expression of transgelin and α‐SMA was observed in spindle cells in cancer stroma, suggesting predominant transgelin expression in CAFs (Figure 1E). To confirm this, we compared the expression regions of transgelin and α‐SMA with respect to three regions in each of the eight different tissues (a total of 24 regions, Figure S5). Image analysis revealed that 84.9% of the expression regions of transgelin and 83.7% of those of α‐SMA showed co‐expression. Higher magnification showed that α‐SMA and transgelin were both expressed in fibroblasts in the stroma (Figure 1E). Double immunofluorescence staining of transgelin and AE1/3, in which colon epithelial and cancer cells were stained, revealed different expression regions (Figure 1F). Importantly, transgelin expression in CAFs was also found in liver metastases and coincided with α‐SMA expression (Figure S6). Furthermore, transgelin expression in CAFs was observed in CAFs of other cancers (Figure S7). Based on these results, we focused on transgelin in fibroblasts and its biological role in cancer tissue activation.

3.2
Transgelin Knockdown Does Not Affect the Expression of Fibroblast Activation Markers
We previously reported that genes upregulated in cSPFs by CM stimulation in vitro were also found in the cancer stroma of human colon cancer tissue [26]. Here, we confirmed transgelin and α‐SMA expression upregulation in cSPF cells after CM stimulation (Figure 2A). Next, we conducted a transgelin‐knockdown experiment to investigate the biological effect of actin‐binding transgelin on α‐SMA and other activation markers of fibroblasts. Transgelin‐knockdown cSPFs were generated by introducing sh‐vectors into the cSPFs. Two types of vectors, sh‐transgelin #1 and sh‐transgelin #2, were used, with sh‐Luc (luciferase) employed as a control. The mRNA and protein expression of transgelin was significantly suppressed in cSPFs transfected with sh‐transgelin #1 and #2 compared with that in cSPFs transfected with sh‐Luc. Transgelin expression in cSPFs was significantly decreased not only in the steady state but also in the activated state following CM stimulation (Figure 2B,C, Figure S8). Morphological changes after transgelin knockdown showed a tendency for increased circularity without significant differences (Figure S9). After examining the effects of transgelin knockdown on common fibroblast activation markers, we found that the expression of α‐SMA, COL1A1, and TNC was unaffected, and CM stimulation‐induced upregulation persisted after transgelin knockdown (Figure 2D).

3.3
Transgelin in cSPFs Has a Pathophysiological Function
We investigated the pathophysiological functions of cSPFs using a contraction assay. To determine the appropriate number of cells for the contraction assay, we employed assays consisting of various numbers of cells ranging from 5 × 104 to 8 × 105, and the degree of contraction following CM stimulation was measured. The maximum contraction with the smallest gel area, having a 0.75‐fold difference from the steady‐state condition, was obtained at 4 × 105, which was selected for subsequent experiments (Figure 3A). Unstimulated cSPFs with sh‐transgelin #1 and sh‐transgelin #2 significantly increased the gel area by 1.29‐ and 1.37‐fold, respectively, compared to controls. Moreover, the gel area was significantly increased in the cSPFs stimulated with sh‐transgelin #1 and sh‐transgelin #2. This indicates that transgelin in cSPFs possesses physiological functions in fibroblasts (Figure 3B,C).

3.4
Effects of Fibroblast‐Based Transgelin on Tumor Growth in Vivo
After comparing tumor growth by measuring the short and long diameters of the tumors over time and calculating the tumor diameter (Figure 3D), we found that tumor growth was significantly enhanced by the co‐injection of cSPFs (luc) compared with the injection of DLD‐1 cells alone. This increase in tumor growth was significantly attenuated by co‐injection of DLD‐1 cells with cSPFs (transgelin knockdown). This tumor growth pattern was consistent between sh‐transgelin #1 and #2 (Figure 3E). Images of the mice at four weeks post‐transplantation are shown in Figure 3F. The photographs show similar findings.

3.5
cSPF Gene Expression Analysis With Transgelin Knockdown
cSPFs from three human colonic tissues were established and transfected with the control and transgelin‐knockdown vectors. Tissues were stimulated with CM (cSPF(+)) or not (cSPF(−)). Next, 12 RNA samples were analyzed (Figure 4A). Principal component analysis revealed that the gene expression profiles of the samples were clustered by their origin, regardless of whether they underwent transgelin knockdown or CM stimulation, indicating that the gene expression patterns of individuals were unaffected by transgelin knockdown (Figure 4B). Similarly, heat map analysis showed that each sample exhibited different gene expression patterns among the individuals (Figure 4C). The gene list is listed in Table S4.

3.6
Transgelin Expression in Cancer Stroma Is Associated With Poor Prognosis
Of the 388 patients, one was excluded due to residual tumor after surgery, and 29 were excluded due to a small tumor component (30% >) or large defect in TMA sections (30% <) (Figure 5A). Therefore, 359 cases were included in the analysis. Pathological characteristics and recurrence rates were examined to confirm the validity of patient selection (Table S5). The pathological findings and recurrence rates were consistent with previously reported results. Patient background characteristics are listed in Table S6.
Consistent with the co‐expression of transgelin and α‐SMA on immunofluorescence observation in Figure 1D,E, the transgelin‐positive area ratio in the tissue core strongly correlated with the α‐SMA‐positive area ratio (Figure 5B, Pearson's correlation coefficient = 0.77, p < 0.01). The high transgelin‐positive area ratio group was significantly associated with a shorter PFS (Figure 5C). Similarly, α‐SMA‐positive area ratios were also examined; a high α‐SMA‐positive area ratio was significantly associated with a shorter PFS (Figure 5D). The associations between low and high transgelin expression and the pathological features are listed in Table 1. The results showed that transgelin expression was associated with T Stage, M Stage, vascular invasion, and p Stage. The incidence rates of transgelin iatrogenic metastatic recurrence were compared (Table S7); patients with a high level of transgelin expression were approximately three times more likely to have lung metastatic recurrence and five times more likely to have liver metastatic recurrence than those with low transgelin expression.

4
Discussion
Transgelin has been reported to be expressed in the glands and muscles of normal organs. However, contradictory reports have been published regarding the association between transgelin expression and clinicopathological findings [23, 31, 32]. In the assessment of epithelial cells, transgelin expression in cancer cells was lower than that in the normal epithelium, and this lower expression in cancer cells is a potential prognostic factor [32]. In contrast, transcriptome analyses have revealed that transgelin overexpression in several cancers is an indicator of poor prognosis [7, 33, 34, 35, 36, 37]. We have also reported that transgelin mRNA overexpression in colon cancer tissues is a prognostic factor [16]. While Zhao et al. focused only on transgelin protein expression in cancer and epithelial cells, and others focused on transgelin mRNA expression in bulk tumor tissue, we speculated that transgelin expression in the cancer stroma might influence these contradictory results in cancer tissue [32]. Through detailed immunohistochemical and immunofluorescent examinations, we concluded that transgelin was predominantly expressed in fibroblasts in cancer tissue and was nearly 80% identical to that of α‐SMA, which is a marker of CAFs. Moreover, this trend has been observed in the stroma of other cancers and metastatic sites. Therefore, transgelin can be used as a novel marker for CAFs. However, further pathological investigations are required to establish its role as a CAF marker.
Transgelin serves as an actin‐binding protein, altering the cytoskeletal structure and morphology [38]. However, its biophysical role in fibroblasts remains to be elucidated. In this study, transgelin expression in fibroblasts was found to be significantly associated with cell contractility. We have previously reported that tumor invasion around the serosal surface of the colonic wall, where cSPFs reside, induces indentation of the serosal surface, which results in bowel obstruction through intimate interactions between cancer cells and cSPFs [13, 24]. Our contraction assay results using transgelin KD revealed that this phenomenon can be induced by transgelin upregulation in cSPFs. Furthermore, using transgelin‐knockdown cSPFs, we first elucidated its tumor‐promoting function by co‐injection with cancer cells into immunodeficient mice. As transgelin knockdown did not significantly influence the expression of other genes in the RNA sequence data, the physical properties induced by transgelin knockdown in cSPFs may also have suppressed tumor growth in the co‐injection model. Our results using cSPFs were consistent with the clinicopathological data of transgelin. Its expression in cancer stroma is strongly correlated not only with prognosis but also with distant metastasis.
We also identified transgelin in fibroblasts as a functional biomarker of colon cancer. Importantly, transgelin overexpression was found in other cancers and metastatic sites. Additionally, our findings suggest that transgelin may play a crucial role in various human cancers through its biological and biophysical functions in CAFs. Furthermore, transgelin in CAFs can be an attractive target in various organ cancers. Alleviation of this physical abnormality in the cancer microenvironment represents a novel strategy for cancer therapy, involving the activation of the tumor‐immune microenvironment and the promotion of drug delivery [39]. Importantly, transgelin knockdown in cSPFs showed minimal alterations in the transcriptome profile. Therefore, transgelin modulation in CAFs may have a minimal influence on other molecules associated with adverse effects [7, 40, 41, 42]. Recently, we contributed to the development of CAF‐targeted micelles that may carry loads on the transgelin‐knockdown construct. Future studies should explore CAF‐targeted therapeutic approaches based on our findings. CAF‐targeted micelles delivering transgelin‐knockdown constructs represent a promising strategy that warrants further preclinical validation for potential clinical application.
This study has some limitations. We only assessed the phenotype of transgelin knockdown; therefore, overexpression or knockout of the gene could not be confirmed. Further studies are warranted to determine whether transgelin expression promotes cancer progression, whether cancer progression increases transgelin expression, and whether it could serve as a therapeutic target for human colonic cancer.
In conclusion, our findings demonstrate that transgelin expression is predominantly localized to cancer‐associated fibroblasts in colon cancer tissues and is significantly associated with clinicopathological features of disease progression. These results suggest that transgelin expression in CAFs may contribute to stromal contractility and tumor progression in colon cancer. Collectively, our study highlights the potential biological relevance of transgelin in CAFs and provides a rationale for further investigations into its functional role and therapeutic relevance.

Author Contributions

Ryuji Okamoto: conceptualization, investigation, writing – original draft, writing – review and editing, formal analysis. Masaaki Ito: writing – review and editing, supervision. Genichiro Ishii: writing – review and editing, supervision.

Funding
This work was supported by JSPS KAKENHI (Grant Number JP23725350, 18007279).

Ethics Statement
The study was approved by the Ethics Review Committee of the National Cancer Hospital (no. 2023–267) and conducted in accordance with the relevant guidelines and regulations. Surgical specimens were obtained from the institutional biobank, and all patients had provided consent for biobank use, either through written informed consent or an opt‐out procedure approved by the ethics committee. The animal experiments were approved by the Animal Ethics Committee of the National Cancer Centre Hospital East (K21‐014).

Conflicts of Interest
The authors declare no conflicts of interest.

Supporting information

Supplementary Figure S1: Examples of excluded tissues. (A) A piece of tissue taken from a piece of normal tissue. (B) Tissues with high smooth muscle content. (C) A largely deficient tissue.

Supplementary Figure S2: Immunostaining images of transgelin and α‐SMA in the smooth muscle in TMA. In TMA, transgelin and α‐SMA showed strong expression in smooth muscle tissue. Owing to the strong expression of transgelin in smooth muscle cells, areas of smooth muscle greater than 30% were excluded from the evaluation. α‐SMA: alpha‐smooth muscle actin; TMA: tissue microarray.

Supplementary Figure S3: Verification of fluorescence leakage using single staining of fluorescent immunostaining. Fluorescence leakage to other wavelengths was checked during fluorescence staining evaluation. Desmin or AE 1/3 was stained with Opal 520, alpha‐smooth muscle actin (α‐SMA) with Opal 570, and transgelin with Opal 690. A small amount of nonspecific fluorescence was observed on the filter of Cy5 during desmin staining, which was not sufficient to affect the observation.

Supplementary Figure S4: Method for calculating immunostaining‐positive areas. From left to right: H.E., transgenin immunostaining, area positive for transgelin (Transgelin‐PN), and area positive in whole tissue. The positive percentage for transgelin was calculated as Transgelin‐PN/Whole Tissue. The analysis was performed using WinRoof 2021.

Supplementary Figure S5: Comparison of transgelin‐ and α‐SMA‐expressing regions in colorectal cancer tissue. Merged photographs of transgelin and α‐SMA expression shown using fluorescence staining in colorectal cancer tissue. Tissues of seven specimens were evaluated, and three area images per specimen were analyzed. α‐SMA: alpha‐smooth muscle actin.

Supplementary Figure S6: Immunostaining for transgelin in colorectal cancer liver metastases. #1, #2, and #3 show tissues of different colorectal cancer metastases to the liver. Liver metastases from colorectal cancer showed elevated transgelin expression in the carcinoma stroma. Alpha‐smooth muscle actin (α‐SMA) expression was observed in the same region.

Supplementary Figure S7: Transgelin expression in gastric, hepatocellular, pancreatic, and lung cancer. Expression of transgelin in gastric cancer (A), hepatocellular carcinoma (B), pancreatic cancer (C), and lung cancer (D). The lung cancer data are shown for squamous cell carcinoma (SCC) and adenocarcinoma. Hepatocellular carcinoma showed transgelin expression in the stroma, including the capsule surrounding the carcinoma.

Supplementary Figure S8: Original figure of total protein western blot Figure 2C. GAPDH (A), Transgelin (B). Protein detection was performed using an automated Jess capillary electrophoresis system.

Supplementary Figure S9: Effects of transgelin on cell morphology. Cells were assessed for their long diameter (A), short diameter (B), circularity (C), and roundness (D). sh Luc, sh transgelin #1 (#1), or sh transgelin #2 (#2) were transfected, and cell morphometry was measured using WinRoof 2021. There was an increasing trend in circularity; however, no significant differences were observed.

Supplementary Table S1: Short hairpin RNA sequences used for knockdown experiments.

Supplementary Table S2: Primers used for quantitative real‐time polymerase chain reaction.

Supplementary Table S3: Quality control assessment of the total RNA sequencing data

Supplementary Table S4: Genes with altered expression in the RNA sequence analysis via transgelin knockdown.

Supplementary Table S5: Correlation between pathological parameters and recurrence rate.

Supplementary Table S6: Characteristics of the participants.

Supplementary Table S7: Transgelin expression and recurrence of distant metastasis.