Fibronectin 1 mediates pressure-induced aggressive phenotypes in colorectal cancer cells and cancer stem cells

Abstract

Background

Multiple interacting factors within the tumor microenvironment, including mechanical pressure, extracellular matrix components, hypoxia, and vascular architecture, are known to promote cancer progression. Although fibronectin 1 functions as a critical extracellular matrix glycoprotein in colorectal cancer, its specific interaction with mechanical pressure is not well characterized.

Methods

This study utilized a meta-analysis (PROSPERO: CRD42024571414) alongside an in vitro weight-induced compression model to evaluate the prognostic role of fibronectin 1 and its interaction with mechanical pressure during colorectal cancer progression.

Results

The meta-analysis demonstrated significantly elevated fibronectin 1 expression in colorectal cancer patients compared to controls (SMD = 0.90, 95% CI: 0.56–1.25, P < 0.001) and a positive association with distant metastasis (OR = 3.63, 95% CI: 1.21–10.95, P = 0.0219). Analysis of colorectal cancer tissues versus adjacent normal tissues (n = 19) revealed markedly increased fibronectin 1 expression in both tumor cells and stromal components. Single-cell RNA sequencing analysis (GSE302903) identified FN1 expression in diverse cell populations, including cancer-associated fibroblasts, endothelial cells, macrophages, and tumor cells. Furthermore, high fibronectin 1 levels were established as a poor prognostic factor in colorectal cancer (HR = 2.47, 95% CI: 1.88–3.25, P < 0.001). In vitro experiments showed that pressure enhanced viability, proliferation, migration, and invasion of colorectal cancer cells and cancer stem cells through fibronectin 1. RNA sequencing indicated significant pressure-induced gene expression changes in colorectal cancer cells and identified the activation of multiple pathways. Specifically, the combined effect of pressure and fibronectin 1 upregulated of TGFBR1, SMAD2, SMAD4, and ROCK1 gene and p-SMAD2 expression.

Conclusions

The expression of fibronectin 1 predicts poor overall survival in colorectal cancer and, together with mechanical pressure, facilitates colorectal cancer progression.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12935-026-04381-5.

Background

With approximately 1.9 million new cases reported in 2022, colorectal cancer (CRC) ranks among the three most common malignancies worldwide and is the second leading cause of cancer-related mortality [1, 2]. While the prognosis for early-stage CRC remains favorable, with survival rates exceeding 80%, the development of metastasis poses a significant challenge to patient outcomes. In stage IV disease, survival drops to approximately 13% despite the availability of advanced therapeutic modalities [3, 4]. Consequently, metastatic colorectal cancer remains a substantial burden in CRC management, necessitating a deeper understanding of the molecular and mechanical drivers of tumor progression.

The tumor microenvironment (TME) is a dynamic and intricate milieu comprising non-cancerous components, including mechanical pressure, extracellular matrix constituents, hypoxia, and blood vessels, all of which actively facilitate tumor growth, invasion, and metastasis [5, 6]. Notably, elevated pressure is consistently present throughout colorectal cancer development, arising from multiple sources such as the TME itself [7–10], colonic peristalsis [11–14], bowel obstruction [15, 16], and even colonoscopy used for diagnosis and treatment [17–20]. This mechanical stress represents an important component of the TME that influences various cellular populations within colorectal tumors, including cancer stem cells (CSCs) [21]. Mechanical compression activates various signaling pathways in cancer cells, thereby altering cellular architecture and promoting phenotypic changes, including enhanced proliferation, migration, invasion, and metastasis [22–25]. Recent evidence indicates that mechanical cues are active drivers of CRC development rather than passive byproducts, with solid stress magnitudes potentially varying across tumor context [26].

Several pressure-related mechanisms associated with tumor progression have been identified, including mechanotransduction [27–30] and epithelial-mesenchymal transition (EMT) [31–34]. Fibronectin 1 is a key molecule involved in both processes and also functions as a mechanoregulator of the extracellular matrix due to its conformational flexibility [35–39]. Although fibronectin 1 has been shown to regulate apoptosis, cell viability, invasion, and metastasis [40–44], the interplay between fibronectin 1 and mechanical compression remains poorly understood, particularly regarding TGFBR1-independent migration pathways that may involve integrin signaling [45]. Furthermore, the prognostic significance of fibronectin 1 in CRC has not yet been comprehensively evaluated through meta-analysis, and existing studies show considerable heterogeneity that requires careful interpretation [46]. Accordingly, we conducted a combined meta-analysis and in vitro study to investigate the relationship between mechanical compression and fibronectin 1 expression, as well as to assess the prognostic value of fibronectin 1 in CRC, while acknowledging limitations regarding physiological pressure range validation and the need for future in vivo confirmation.

Materials and methods

Systematic review and meta-analysis

Literature search. We searched the PubMed, Web of Science, and Embase databases. The search strategy was listed in the Supporting Information. The study completed its search on July 24, 2025. The registered protocol number in PROSPERO was CRD42024571414.

Inclusion and exclusion criteria. Studies were considered eligible if they met the following inclusion criteria: (1) patients had a pathologically confirmed diagnosis of CRC; (2) Fibronectin 1 levels were measured using immunohistochemistry (IHC), enzyme-linked immunosorbent assay (ELISA), real-time reverse transcription polymerase chain reaction (RT-PCR), or proteomic analysis; (3) studies reported the association between fibronectin 1 levels and clinicopathological features such as TNM classification, tumor stage, size, grade, lymphovascular invasion (LVI), perineural invasion (PNI), or 5-year mortality rate, or compared fibronectin 1 levels between CRC and non-cancer samples; and (4) studies were published in English. The exclusion criteria were: (1) titles, letters, cell or animal studies, reviews, and conference abstracts; (2) duplicate publications; or (3) insufficient or unavailable data.

Data extraction and quality assessment. Two investigators independently extracted data from the included studies, including: (1) first author, (2) year of publication, (3) study location, (4) study design, (5) sample size, (6) patient stage, (7) method of fibronectin 1 assessment, and (8) sample type. The quality of the included studies was assessed using the Newcastle–Ottawa Scale (NOS) [47].

Weight-induced pressure model in vitro study

Previous studies have described two common types of mechanical pressure: hydrostatic pressure [25, 48] and compressive pressure [25, 49, 50] (Figs. 1 and 2). In the present study, compressive pressure was applied using a weight-based system to generate defined levels of pressure (Table S1), in accordance with the International System of Units (SI) [51].

Fig. 1.

PRISMA flowchart and summary of meta-analysis findings. (a) PRISMA flow diagram for study selection. (b) Standardized Mean Difference (SMD) for fibronectin 1 levels between CRC patients versus control. (c) Pooled odds ratio (OR) of high/positive fibronectin 1 levels in colorectal cancer: non-survivors vs. survivors at 5-year follow-up. (d) Pooled hazard ratio (HR) for 5-year overall survival of high fibronectin 1 versus low fibronectin 1 levels in CRC patients. (e) Pooled odds ratio (OR) for high/positive fibronectin 1 levels in colorectal cancer patients with distant metastasis versus no distant metastasis

Fig. 2.

Fibronectin 1 is overexpressed in colorectal cancer (CRC) tissues and cell populations. (a) Fibronectin 1 protein expression is higher in CRC tissues (n = 19) than in adjacent normal tissues (P < 0.001). (b) Normal esophagus tissue as a positive control. Representative IHC images of CRC (c) and adjacent normal tissues (d). (e) UMAP plot of GSE302903 in GEO. (f) FN1 expression across cell populations. Student’s 𝘵-test, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Patient and tissue specimen

Colorectal cancer tissues were collected at Taipei Medical University Hospital (IRB No. N202404078) with informed consent from all participants.

Cell culture

Eight colorectal cancer cell lines (SW48, SW480, SW620, DLD1, CX1, HCT116, HT29, and H3347) were obtained from the American Type Culture Collection (ATCC). SW48, SW480, SW620, and CX1 cells were cultured in Dulbecco’s Modified Eagle Medium, high glucose (DMEM/HG) (Cat. No. SH30243.02, Cytiva, MA, USA). DLD1, DLD1_TCSC and H3347 cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 (Cat. No. SH30027.02, Cytiva, MA, USA). HCT116 and HT29 cells were cultured in McCoy’s 5 A medium (Cat. No. PM150710, Pricella, Elabscience, Texas, USA). WiDr_TCSC cells were cultured in Minimum Essential Medium (MEM) (Cat. No. SH30265.02, Cytiva, MA, USA). All media were supplemented with 10% fetal bovine serum (FBS) for non-CSCs and 5% FBS for CSCs (Cat. 10437028, Gibco™ Thermo Fisher Scientific, Waltham, MA, USA), along with 1% penicillin–streptomycin–amphotericin B (PSA). The cells were cultured at 37 °C in a humidified atmosphere with 5% CO2.

Gene knockdown by lentivirus-mediated shRNA

The shRNA sequences used were as follows: negative control shRNA TRCN0000072249 (NC_ShRNA), 5′-GCGGTTGCCAAGAGGTTCCAT-3′; FN1-targeting shRNA TRCN0000286357 (FN1_ShRNA_1), 5′-CGTGGTTGTATCAGGACTTAT-3′; and TRCN0000293839 (FN1_ShRNA_2), 5′-TGCAGCACAACTTCGAATTAT-3′; TGFBR1-targeting shRNA TRCN0000194693 (TGFBR1_ShRNA_1), 5′- CTCATGTTGATGGTCTATATC-3′; and TRCN0000196293 (TGFBR1_ShRNA_2), 5′- GAAGTTGCTGTTAAGATATTC-3′. Lentiviral shRNA-mediated FN1 and TGFBR1 knockdown were performed in SW48, SW620, DLD1_TCSC, and WiDr_TCSC cells. Plasmids were extracted from shRNA-containing bacterial stocks using the QIAGEN Plasmid Mini Kit (Cat. 12125, QIAGEN, Taipei, Taiwan). Lentiviral particles were generated by co-transfecting HEK293T cells with pCMV-ΔR8.91, pMD.G, and pLKO.1-shRNA using the PolyJet™ In Vitro DNA Transfection Reagent (Catalog SL100688, SignaGen Laboratories, Rockville, USA). After viral supernatant collection, target cells were infected with lentivirus in the presence of 8 µg/mL polybrene and selected with 2 µg/mL puromycin for 4–5 days. FN1 and TGFBR1 knockdown efficiency were confirmed by qRT-PCR and western blotting.

Real-time quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted using RNAzol from the TriRNA Pure Kit (Cat. No. GZX050/100/200, Geneaid). cDNA was synthesized from 1 µg of RNA using the ExcelRT™ Reverse Transcription Kit (RP1300, SMOBIO). Quantitative PCR (qPCR) was performed using Fast SYBR™ Green Master Mix (Cat. 4385612, Thermo Fisher). EEF1A1 was used as the internal control. PCR conditions were set according to the Fast SYBR™ Green protocol. Primer sequences are listed in Table S2.

Western blot

Cells were lysed using Cell Lysis Buffer (Cat. 9803 S, Cell Signaling Technology) and incubated on ice for 2 h. Lysates were centrifuged at 20,000 × g for 10–20 min, and the supernatants were collected. Protein concentration was determined using a BCA assay. Samples were mixed with 2X loading dye, denatured at 95 °C for 10 min, and 30–50 µg of protein was loaded onto SDS-PAGE gels for electrophoresis (90–150 V). Proteins were transferred onto PVDF membranes using either semi-dry or wet transfer. Membranes were blocked with 5% BSA in 0.01% TBST for 30–60 min, then incubated overnight at 4 °C with primary antibodies. After washing, HRP-conjugated secondary antibodies (1:10,000) were applied for 1–2 h. Bands were visualized using chemiluminescence.

Cell viability

Cells were seeded at a density of 3 × 10⁵ cells/mL within a circular area (Inner diameter: 27.4 mm) on 100-mm cell culture dishes. After 24 h of incubation, 10mL of 0.4% agar solution was added, followed by layering 10 mL of 1.2% agar on top. Weights were placed on top of the agar to apply compressive pressure, and cells were incubated for 12 h. After incubation, the weights and agar were removed to harvest the cells. A total of 1 × 10⁴ cells per well were seeded in 96-well plates. Cell viability was assessed using the PrestoBlueTM Cell Viability Reagent (Cat. No. A13261, Thermo Scientific, MA, USA), and fluorescence was measured using the SpectraMax M2e Microplate Reader (SpectraMax^® M2e Microplate Reader, Molecular Devices LLC., San Jose, CA, USA).

Cell proliferation

Cells (3 × 10⁵ cells/mL) were seeded within a circular area (Inner diameter: 27.4 mm) on 100-mm culture dishes. After 24 h, 10 mL each of 0.4% and 1.2% agar solutions were sequentially layered, and weights were placed on top to apply compressive pressure. Following 12 h of incubation, the agar and weights were removed, and the cells were harvested. Live cells were then seeded into 96-well plates (5 × 10³ cells/well), and fluorescence was measured using PrestoBlue™ (Cat. No. A13261, Thermo Scientific) with a SpectraMax M2e plate reader at 0, 24, 48, and 72 h.

Colony assay

After applying different amounts of pressure for 12 h, the cells (500 cells/well) were seeded into 6-well plates. After 14–18 days of incubation, colonies were formed. Colonies were fixed with methanol and stained with a 0.5% crystal violet solution. Colonies containing more than 50 cells were counted.

Cell migration

Inserts (Cat. 80209, ibidi) were placed on 100-mm culture dishes, and 100 µL of cell suspension (2–5 × 10⁴ cells/well) was added to each chamber. After cell attachment, cells were treated with mitomycin C (5 µg/mL) for 2 h. The insert was removed to create a gap. To apply pressure, 10 mL of 0.4% agar was added, followed by 10 mL of 1.2% agar and placement of weights on top. Cell migration was monitored by imaging the gap at 12-hour intervals.

Transwell invasion

Mitomycin C–treated cells in 2 h (1 × 10⁴ cells/ well) were seeded into the upper chamber of a transwell (Cat. 9328012, CellQART). A 400 µL layer of 0.6% agar was added above the cells, followed by weights. The lower chamber contained 600 µL of serum-free medium. After 24 h, invaded cells were stained with crystal violet, and images were captured from five randomly selected fields. For the ‘no agar’ group, a conventional transwell invasion assay was performed without agar or applied pressure. Cells were seeded in serum-free medium in the upper chamber, while complete medium was added to the lower chamber as a chemoattractant.

3D tumor spheroid invasion

A 200 µL suspension of cells (500 cells/well) was seeded into 96-well Clear Round Bottom Ultra-Low Attachment (Cat. No. 7007, Corning, Berlin, Germany). After 5 days of incubation, when spheroids had formed, 200 µL of 0.4% agar was added, followed by placement of a weight to apply compressive pressure. After 96 h of incubation, the weights were removed, and images were captured to calculate the relative area change (96 h/0 h).

Tumor sphere formation assay

Cells (5 × 10⁴ cells/well) were seeded into 6-well plates pre-coated with 1 mL of 1.2% agarose and cultured in medium containing 5% FBS. After 72 h, sphere formation was assessed, and spheres with a diameter ≥ 100 μm were counted.

Soft agar assay

Cells (5 × 10⁴ cells/well) were mixed with 0.4% low-melt agarose and plated onto 6-well plates pre-coated with 1.2% agarose. Subsequently, 2 mL of medium containing 5% FBS was added. After 14–21 days, colonies containing more than 20 cells were counted.

Immunohistochemistry (IHC)

Immunohistochemical staining was performed using an automated system (VENTANA, Roche; REF 950 − 223, 05424542001) according to the manufacturer’s instructions. Antigen retrieval was conducted using CC1 buffer at 100 °C for 72 min, followed by incubation with a fibronectin 1 antibody (Cat. No. GTX112794) for 1 h. Human esophageal tissue was used as a positive control. Fibronectin 1 expression was evaluated using the H-score, calculated as Σ (i × Pi), where staining intensity (i: 0–3) and the percentage of positive cells (Pi: 0–100%) were assessed. The total score ranged from 0 to 300.

Single-cell RNA-seq data analysis

Public single-cell RNA sequencing (scRNA-seq) data (GSE302903) were obtained from the Gene Expression Omnibus (GEO) and analyzed using the Seurat package in R. After quality control and normalization, highly variable genes were identified, followed by dimensionality reduction, clustering, and UMAP visualization. Cell types were annotated using canonical markers, and FN1 expression was evaluated across different cell populations. Differential expression analysis was performed using the Wilcoxon rank-sum test, with an adjusted P < 0.05 considered statistically significant.

RNA sequencing

RNA samples were collected from CX1 cells under control and pressure-treated conditions. RNA purity and integrity were assessed using a SimpliNano™ spectrophotometer (Biochrom, MA, USA) and Qsep 100 DNA/RNA Analyzer (BiOptic Inc., New Taipei City, Taiwan). For library preparation, 1 µg of total RNA per sample was used. Poly(A)+ mRNA was isolated, fragmented, and reverse-transcribed into cDNA. Strand-specific libraries were prepared by adapter ligation, size selection (300–400 bp), and PCR amplification. Library quality was evaluated using the Qsep 100, Qubit^® 2.0 Fluorometer (Thermo Scientific), and Agilent 2100 Bioanalyzer. Sequencing was performed on the Illumina NovaSeq 6000 platform to generate 150-bp paired-end reads.

Statistical analysis

Data for the meta-analysis were analyzed using RStudio version 4.4.1 [52]. Differences in fibronectin 1 expression between CRC patients and controls were expressed as standardized mean differences (SMDs) with 95% confidence intervals (CIs) [53]. Associations with clinicopathological features were assessed using odds ratios (ORs), and time-to-event outcomes were evaluated using hazard ratios (HRs) extracted from Kaplan–Meier curves with WebPlotDigitizer version 5. Heterogeneity was assessed using the Cochran Q test and I² index, with p ≤ 0.05 or I² > 50% indicating significant heterogeneity [53]. The Mantel-Haenszel random-effects model was applied to estimate pooled effects. Sensitivity analyses were performed by sequentially removing individual studies to assess the robustness of the results.

Data from the in vitro study are presented as mean ± standard deviation (SD) from at least three independent experiments. Student’s t-test was used for two-group comparisons, and one- or two-way ANOVA with post-hoc tests was used for three or more groups. The value of P < 0.05 was considered significant, and significance is expressed in figures as asterisks with *P < 0.05, **P < 0.01, and ***P < 0.001. Statistical analyses and graphs were generated using GraphPad Prism version 9.

Results

Study selection, characteristics, and quality assessment in systematic reviews and meta-analyses

During study selection, we screened 4,494 studies across three databases. Following selection criteria articles were eligible for full-text evaluation (Fig. 1a). Ultimately, 20 studies were included in the systematic review and meta-analysis (Table 1). Article characteristics are detailed in Table 1. Studies were conducted between 1982 and 2025, encompassing 2,605 patients from multiple countries (Japan, China, Korea, Brazil, Poland, USA). All included studies were nonrandomized controlled trials, comprising one case-control, nine cross-sectional, and ten cohort studies. Newcastle–Ottawa Scale (NOS) scores for all included studies ranged from 5 to 8 (Table 1).

Table 1.

Characteristics of included articles in systematic review and meta-analysis (n = 20)

Year	First Author	Location	Study design	Sample size	Stage	Technique	Sample	NOS score/total
1982	P Niemczuk [69]	England	Cohort	41	Dukes B, C	IHC	Tumor	7/9
1995	H Inufusa [70]	Japan	Case-control	99	Dukes A-D	IHC	Tumor, liver metastatic lesion	6/9
1997	N Hanamura [71]	Japan	Cross-section	29	I-III	IHC	Tumor	6/7
2008	Noboru Saito [72]	Japan	Cross-section	113	I-IV	ELISA	Serum, urine	6/7
2009	Daniela Cabibi [73]	Italy	Cross-section	102	Dukes B, C, D	IHC	Tumor	6/7
2013	Luciano de Souza Viana ADDIN EN.CITE [74]	Brazil	Cross-section	114	I-IV	IHC	Tumor	7/7
2014	Daniel J [75]	USA	Cohort	494	I-IV	IHC	Tumor	7/9
2014	Hiroyuki Kida ADDIN EN.CITE [76]	Japan	Cohort	146	I-IV	PCR	Tumor	8/9
2014	Suzana Angelica Silva Lustosa ADDIN EN.CITE [77]	Brazil	Cohort	114	I-IV	IHC	Tumor	7/9
2014	Jung-A Yun ADDIN EN.CITE [78]	Korea	Cohort	409	III	IHC	Tumor	8/9
2015	Manveen K Sethi ADDIN EN.CITE [79]	Korea	Cross-section	8	I-IV	Proteomic	Tumor, non-tumor tissue	5/7
2016	Wenzhong Yi ADDIN EN.CITE [64]	China	Cohort	107	I-IV	IHC, PCR	Tumor, normal tissues from non-CRC patients	8/9
2017	Zohreh Niknami [80]	Iran	Cross-section	45	I-IV	PCR	Tumor	6/7
2017	Hidehiko Takigawa ADDIN EN.CITE [81]	Japan	Cohort	44	I-IV	IHC	Tumor	8/9
2018	Xun Cai [43]	China	Cohort	108	I-III	PCR, IHC	Tumor, Adjacent normal tissue	8/9
2019	Margareta Žlajpah [82]	Slovenia	Cross-section	20	N/A	PCR	Tumor	5/7
2021	Aleksandar Bogdanovic ADDIN EN.CITE [83]	Serbia	Cohort	24	IV	PCR	Tumor, Non-tumor liver tissue	6/9
2021	Yuan-Chang Dai ADDIN EN.CITE [84]	Taiwan	Cross-section	150	I-IV	IHC	Tumor	7/7
2021	Łukasz Zadka ADDIN EN.CITE [85]	Poland	Cross-section	78	I-IV	IHC	Tumor	5/7
2025	Seçil Durel Avcıoğlu [86]	Türkiye	Cohort	360	I-IV	IHC, PCR, ELISA	Tumor, non-tumor tissue	8/9

Fibronectin 1 is overexpressed and predicts poor prognosis in colorectal cancer

Four of the eight studies compared fibronectin 1 levels between CRC and normal samples using mean ± standard deviation values (Table S3). Overall, fibronectin 1 expression was significantly higher in CRC patients than in controls (SMD = 0.90, 95% CI: 0.56–1.25, P < 0.001; Fig. 1b), although considerable heterogeneity was observed (Q = 16.13, I² = 75%, P < 0.01). Consistently, our IHC analysis of CRC tissues confirmed that fibronectin 1 protein expression was significantly higher in tumor tissues (n = 19) than in adjacent normal tissues (P < 0.001) (Fig. 2a). Fibronectin 1 is expressed across multiple cell populations in colorectal cancer (CRC). Immunohistochemical analysis showed higher fibronectin 1 expression in both tumor cells and stromal components compared with adjacent normal tissues (Fig. 2c and d). To further identify the cellular sources of FN1 in CRC, single-cell RNA-sequencing data from 12 CRC samples (GSE302903) were analyzed. FN1 expression was highest in cancer-associated fibroblasts (CAFs), followed by endothelial cells, macrophages, and tumor cells (Fig. 2e and f).

In addition, seven studies evaluated the prognostic significance of fibronectin 1 in relation to overall survival. Regarding 5-year overall survival, high or positive fibronectin 1 expression was significantly associated with increased mortality in CRC patients (OR = 4.12, 95% CI: 2.54–6.7, P < 0.001; Fig. 1c), with no evidence of heterogeneity (Q = 1.09, I² = 0%, P = 0.9). Similarly, fibronectin 1 overexpression was a significant predictor of poor 5-year overall survival (HR = 2.47, 95% CI: 1.88–3.25, P < 0.001; Fig. 1d), with no notable heterogeneity observed (Q = 1.33, I² = 0%, P = 0.72).

Fibronectin 1 is associated with distant metastasis in colorectal cancer

Twelve studies were included to evaluate the association between fibronectin 1 expression and clinicopathological features, including TNM classification, tumor stage, size, grade, LVI, and PNI. High or positive fibronectin 1 expression was not significantly related to mucinous histology (OR 0.72, 95% CI 0.19–2.64, Fig. S1a), tumor grade (OR 0.73, 95% CI 0.36–1.5, Fig. S1b), venous invasion (OR 0.76, 95% CI 0.13–4.43, Fig. S1c), perineural invasion (OR 0.71, 95% CI 0.12–4.2, Fig. S1d), lymphovascular invasion (OR 1.53, 95% CI 0.89–2.62, Fig. S1e), T stage (OR 1.34, 95% CI 0.78–2.32, Fig. S1f), and N stage (OR 2, 95% CI 0.88–4.56, Fig. S1g). However, fibronectin 1 expression was significantly correlated with distant metastasis (OR 3.63, 95% CI 1.21–10.95, P = 0.0219, Fig. 1e), though this association showed substantial heterogeneity (Q = 10.26, I² = 61%, P = 0.04).

Mechanical compression enhances cell viability and proliferation in colorectal cancer via fibronectin 1

Fibronectin 1 is a critical extracellular matrix glycoprotein that interacts with different components of the tumor microenvironment during cancer development [54, 55]. Single-cell RNA sequencing revealed that FN1 is predominantly expressed in CAFs within colorectal cancer tissues. However, immunohistochemical analysis showed that fibronectin 1 expression is also elevated in colorectal cancer cells compared to normal epithelial cells (Fig. 2c and d). Based on these observations, we investigated whether cancer cell-derived fibronectin 1 plays a functional role under mechanical pressure conditions. To further investigate the relationship between fibronectin 1 and mechanical pressure, an in vitro pressure-induced compression model using applied weights was established to generate varying levels of pressure in colorectal cancer cell lines with or without FN1 knockdown (Table S1, Fig. 3d). Among eight CRC cell lines, SW48 and SW620 were selected for FN1 knockdown using lentivirus-mediated shRNA due to their high FN1 mRNA and fibronectin 1 protein levels were observed (Fig. 3a-c).

Fig. 3.

FN1 knockdown in SW48 and SW620 cell lines and pressure-induced model using weights. (a) Fibronectin 1 expression at mRNA and protein levels across 8 colorectal cancer (CRC) cell lines. FN1 knockdown using lentivirus-mediated shRNA in SW48 (b) and in SW620 (c) cells. (d) Schematic of the pressure-induced model in CRC cell lines using applied weights. ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

After 12 h of pressure induction, pressures of + 384.2 and + 647.9 Pa significantly increased cell viability in SW48 and SW620 cells compared with the non-pressurized condition (+ 0 Pa). However, this effect was not observed in FN1 knockdown cells (Fig. 4a and f). Similarly, in the proliferation assay, pressure conditions (+ 384.2 Pa and + 647.9 Pa) significantly promoted proliferation in control cells but did not affect the FN1 knockdown groups (Fig. 4b-e and g-j). Consistently, colony formation was significantly increased in pressure-treated control cells compared to the non-pressurized condition, whereas no significant difference was observed in the FN1 knockdown groups (Fig. 5a-d).

Fig. 4.

Mechanical pressure enhances cell viability and proliferation via fibronectin 1 in CRC. (a, f) Pressure (+ 384.2 Pa and + 647.9 Pa) significantly increased cell viability in control groups compared to + 0 Pa, but not in FN1-knockdown groups. Pressure notably promoted proliferation in control groups of SW48 (b, c) and SW620 (g, h) cells, but not in their respective FN1-knockdown groups (d, e, i, j). ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Fig. 5.

Pressure promotes colony formation in CRC cells via fibronectin 1. (a, b) Mechanical pressure significantly increased the number of colonies in control SW48 and SW620 cells compared to + 0 Pa. In contrast, pressure notably reduced colony formation in FN1-knockdown SW48 and SW620 cells. (c, d) Images of colony formation under different pressure conditions in SW48 (c) and SW620 (d) cells, with and without FN1 knockdown. ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Mechanical compression promotes migration and invasion of colorectal cancer cells via fibronectin 1

Three pressure-induced models were developed in this study to investigate cancer cell migration, invasion, and 3D spheroid invasion (Fig. 6a-c). In the migration assay, exposure to pressures of + 384.2 Pa and + 647.9 Pa significantly enhanced cancer cell migration in the control groups compared to + 0 Pa. In contrast, no significant change in migration was observed in the FN1-knockdown cell lines under the same pressure conditions (Figs. 6d-e and 7a-b, S3a-b). At 24 h after pressure exposure in the migration assay, disruption of the cell monolayer structure was observed under compression conditions (Fig. S11), resulting in images of insufficient quality for reliable quantitative analysis. Similarly, in the invasion assay using comparable pressure levels (+ 388.7 Pa and + 643.7 Pa), invasion was significantly increased in the control groups, whereas the FN1-knockdown groups showed no significant response to pressure (Figs. 6f-g and 7c-d, S3c-d).

Fig. 6.

Pressure stimulates migration, invasion, and 3D spheroid invasion in CRC cells via fibronectin 1. Mechanical pressure was applied using weight-based models to assess migration (a), chamber invasion (b), and 3D spheroid invasion (c). (d, e) Pressure (+ 384.2 Pa, + 647.9 Pa) significantly enhanced cell migration in control SW48 and SW620 cells compared to + 0 Pa, whereas no effect was observed in FN1-knockdown cells. (f, g) Similarly, pressure (+ 388.7 Pa, + 643.7 Pa) considerably increased the number of invaded cells in control groups, with no significant change in FN1-knockdown cells. (h, i) Pressure (+ 346.7 Pa) substantially induced 3D spheroid invasion in control SW48 and SW620 cells, but had no effect in FN1-knockdown cells. ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Fig. 7.

Pressure stimulates migration, invasion, and 3D spheroid invasion in SW48 cells via fibronectin 1. Migration of SW48 (a) and SW48 FN1_ShRNA_1 (b) cells under pressure (+ 647.9 Pa) and no pressure (+ 0 Pa) for 12 h. Chamber invasion of SW48 (c) and SW48 FN1_ShRNA_1 (d) cells under pressure (+ 643.7 Pa) and no pressure (+ 0 Pa). 3D spheroid invasion of SW48 (e) and SW48 FN1_ShRNA_1 (f) cells under pressure (+ 346.7 Pa) and no pressure (+ 0 Pa)

Beyond the 2D models, a 3D spheroid invasion assay was performed to assess the effect of pressure on tumor spheroid behavior. Pressure at + 346.7 Pa significantly promoted cancer cell invasion into the surrounding matrix in the control groups. In contrast, the FN1-knockdown groups did not exhibit a significant increase in invasion under the same conditions (Figs. 6h-j and 7e-f, S3e-f).

Pressure enhances fibronectin 1-mediated viability, proliferation, migration, and invasion in colorectal cancer stem cells

Because pressure acts as an external factor, it can influence different cancer cell populations, including CSCs [21]. Two colorectal CSC lines, DLD1_TCSC and WiDr_TCSC, which exhibited greater stemness characteristics than their parental cell lines (DLD1_PT and WiDr_PT) (Fig. S4), were used in pressure-induced models to investigate the effects of pressure on CSC viability, proliferation, migration, and invasion (Fig. 8a-b).

Fig. 8.

Pressure induces cell viability and proliferation in DLD1_TCSC and WiDr_TCSC via fibronectin 1. (a, b) FN1 knockdown in DLD1_TCSC (a) and WiDr_TCSC (b) was confirmed at both the mRNA and protein levels. Pressure (+ 647.9 Pa and + 1411.1 Pa) significantly increased cell viability in DLD1_TCSC (c) and WiDr_TCSC (d) compared with + 0 Pa, but this effect was not observed in FN1_shRNA CSCs. (e-h) Compared with + 0 Pa, pressure meaningfully promoted proliferation in DLD1_TCSC (e) and WiDr_TCSC (f), whereas no increase was detected in FN1_shRNA CSCs (g, h). ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Exposure to pressure (+ 384.2, + 647.9, and + 1411.1 Pa) significantly increased cell viability (Fig. 8c-d), proliferation (Fig. 8e-h), and colony formation (Fig. 9a-d) in DLD1_TCSC and WiDr_TCSC compared with + 0 Pa. These effects were abolished in FN1-knockdown CSCs. Regarding migration and invasion, pressure (+ 384.2 Pa and + 647.9 Pa) markedly enhanced migration, invasion, and 3D spheroid invasion in both DLD1_TCSC (Fig. 10a-f) and WiDr_TCSC groups (Fig. S5a-f) compared with + 0 Pa, whereas FN1-knockdown DLD1_TCSC and WiDr_TCSC groups showed no response to pressure in these assays.

Fig. 9.

Pressure drives colony formation in colorectal cancer stem cells through fibronectin 1. Pressure (+ 384.2, + 647.9, + 1411.4 Pa) significantly elevated colony formation in DLD1_TCSC (a) and WiDr_TCSC (b), but not in the FN1-knockdown CSCs group compared with + 0 Pa. (c, d) Images of colony formation under different pressure conditions in DLD1_TCSC (c) and WiDr_TCSC (d) cells, with and without FN1 knockdown. ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Fig. 10.

Pressure stimulates migration, invasion, and 3D spheroid invasion in DLD1_TCSC via fibronectin 1. (a, b) Pressure (+ 384.2 Pa, + 647.9 Pa) significantly increased migration in DLD1_TCSC cells compared with + 0 Pa, whereas the same pressures had no effect in DLD1_TCSC FN1_shRNA cells. (c, d) Pressure (+ 388.7 Pa, + 643.7 Pa) notably enhanced the number of invaded cells in DLD1_TCSC compared with + 0 Pa, but no difference was observed in DLD1_TCSC FN1_shRNA cells. (e, f) In the 3D spheroid model, pressure (+ 346.7 Pa) considerably promoted invasion in DLD1_TCSC compared with + 0 Pa, but not in DLD1_TCSC FN1_shRNA cells. ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

TGFBR1, SMAD2, SMAD4, and ROCK1 are upregulated by combined effect of mechanical compression and fibronectin 1 in colorectal cancer cells and cancer stem cells

Based on a literature review of the mechanisms linking fibronectin 1 and pressure, several signaling pathways have been implicated, including integrins [56], transforming growth factor-beta (TGF-beta) [57], EMT [58], and mechanotransduction [59], which may mediate the downstream effects of fibronectin 1 under mechanical stress. Furthermore, to explore pressure-related pathways in colorectal cancer cells, we performed RNA sequencing on the CX1 cell line, which also showed increased proliferation (Fig. 11b), colony formation (Fig. 11c), migration (Fig. 11e), and invasion (Fig. 11d and f) under mechanical compression, similar to SW48 and SW620 cells. Compared with the + 0 Pa group, exposure to + 647.9 Pa markedly altered gene expression in CX1 cells (Fig. 11g–i). Differentially expressed genes were subsequently analyzed using Ingenuity Pathway Analysis (IPA), which identified pressure-associated pathways, including RHO GTPase and TGF-beta signaling (Fig. 11j; and Supplementary File, full IPA analysis of CX1 cells under pressure). Given that compression at + 647.9 Pa significantly altered CRC cells and CSCs behavior compared with the + 0 Pa condition, qPCR and western blot analyses were performed to examine and validate the combined effects of mechanical compression and fibronectin 1.

Fig. 11.

Effects of mechanical pressure on the CX1 cell line and RNA-seq analysis under compression. Mechanical pressure notably decreased CX1 cell viability (a), while significantly enhancing proliferation (b), colony formation (c), invasion (d), migration (e), and 3D spheroid invasion (f). RNA-seq analysis comparing pressure-treated and control CX1 cells (n = 3/group) is presented as a volcano plot (g), p-value distribution (h), heatmap (i), and Ingenuity Pathway Analysis analysis (j). Student’s 𝘵-test, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Under + 647.9 Pa, the FN1 mRNA expression was markedly upregulated compared with + 0 Pa in SW48, SW620 cells (Fig. 12a and f), DLD1_TCSC, and WiDr_TCSC (Fig. S6a, S6f) by qPCR, but this change was not observed at the fibronectin 1 protein level by western blotting (Fig. 12k-l, S6k-l, S9e-h). In these cell lines, TGFBR1 mRNA and TGFBR1 protein expression were noticeably increased in response to both fibronectin 1 and pressure (Fig. 12b, g and k-l, S6b, S6g, S6k-l). Subsequently, SMAD2 and SMAD4 mRNA expression, together with SMAD2, p-SMAD2, and SMAD4 protein expression were considerably elevated only under pressure (+ 647.9 Pa) in the control groups (Fig. 12c-d, h-i and k-l, S6c-d, S6h-I, S6k-l). Furthermore, ROCK1 mRNA and ROCK1 protein expression, a downstream effector of RHO GTPase signaling, were significantly enhanced following exposure to + 647.9 Pa in the control groups, but remained unchanged in the FN1 knockdown groups (Fig. 12e, j and k-l, S6e, S6j, S6k-l).

Fig. 12.

Pressure-induced fibronectin 1 signaling upregulates ROCK1, SMAD4, SMAD2, p-SMAD2, and TGFBR1 in colorectal cancer. Pressure increased the FN1 mRNA expression in SW48 (a, k) and SW620 cells (f, l) but not in fibronectin 1 protein level. The cooperative effect of fibronectin 1 and pressure upregulated the mRNA of TGFBR1, SMAD2, SMAD4, and ROCK1 and their protein expression as well as SMAD2 phosphorylation, in SW48 (b–f, k) and SW620 cells (g–j, l). ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

TGFBR1 is involved in pressure-induced proliferation and colony formation, but not cell viability, migration, or invasion in colorectal cancer cells and cancer stem cells

To further investigate the downstream effects of the combined action of fibronectin 1 and mechanical compression in colorectal cancer, TGFBR1 was knocked down using lentiviral transduction in SW48 (Fig. 13a, S10a), SW620 (Fig. 13b, S10b), DLD1_TCSC (Fig. S7a, S10c), and WiDr_TCSC (Fig. S7b, S10d). Mechanical compression (+ 647.9 Pa) significantly enhanced proliferation and colony formation in control cells, but these effects were abolished in the TGFBR1 knockdown groups (Fig. 13e-h, S7e-h). In contrast, no significant differences were observed in cell viability (Fig. 13c-d, S7c-d), migration (Fig. 14a-b, S8a-b), or invasion (Fig. 14c-f, S8c-f) between the control and TGFBR1-knockdown groups under mechanical compression, indicating that TGFBR1 signaling specifically mediates proliferative responses rather than migratory or invasive phenotypes in this model.

Fig. 13.

Pressure promotes colorectal cancer cell proliferation and colony formation through TGFBR1 signaling, but does not affect cell viability. Knockdown of TGFBR1 expression was confirmed in SW48 (a) and SW620 (b) cells. Mechanical compression significantly increased cell viability in both groups (c, d). However, TGFBR1 knockdown significantly abolished the pressure-induced increase in cell proliferation (e, f) and colony formation (g, h) compared with the control group. Student’s 𝘵-test and ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Fig. 14.

TGFBR1 is not involved in pressure-induced migration, invasion, or spheroid invasion in colorectal cancer cells. No significant differences were observed in migration (a, b), invasion (c, d), or spheroid invasion (e, f) between the TGFBR1 knockdown and control groups under mechanical compression in SW48 and SW620 cells. ANOVA, significant at P < 0.05 with *P < 0.05, **P < 0.01, and ***P < 0.001, ns = non-significant

Discussion

Metastasis remains a major challenge in the management of CRC due to its high incidence and associated mortality [3, 4, 60]. One of the key elements involved in the metastatic process is the TME, which comprises factors such as mechanical pressure and fibronectin 1. In the present study, meta-analysis, IHC, and single-cell RNA sequencing consistently demonstrated that fibronectin 1 was overexpressed in colorectal cancer cells and stromal components compared with normal tissues, and that elevated fibronectin 1 levels were positively associated with distant metastasis. In addition, high fibronectin 1 levels were identified as a poor prognostic factor in CRC. Sensitivity analysis was conducted to address substantial heterogeneity (I² > 50%) as presented in the Supporting Information (Fig. S2), and no significant changes were found in the primary results. Publication bias was not assessed due to the inclusion of fewer than 10 studies were included in each analysis. The interaction between fibronectin 1 and mechanical pressure, a physical component of the TME, was evaluated using a weight-induced in vitro model. Pressure promoted colorectal cancer cells and CSCs viability, proliferation, colony formation, migration, and invasion through fibronectin 1. Furthermore, mRNA of TGFBR1, SMAD2, SMAD4, and ROCK1 and their protein expression, together with increased SMAD2 phosphorylation, were upregulated through fibronectin 1 under pressure stimulation.

On the other hand, fibronectin 1 is a multidomain glycoprotein composed of three repeating units (Type I, Type II, and Type III) that mediate interactions with extracellular molecules [61–63]. Previous studies have demonstrated that fibronectin 1 promoted proliferation, migration, invasion, and chemoresistance in CRC [40, 64]. However, these studies did not account for mechanical pressure as an external factor, nor did they investigate the role of fibronectin 1 in colorectal cancer stem cells. The association between fibronectin 1 and pressure in our study suggests that fibronectin 1 plays a more complex role in CRC progression.

Mechanical pressure has been reported to induce drug resistance, migration, invasion, and metastasis in several cancer types, including breast cancer [24], brain tumors [50], and ovarian cancer [25]. Cancer cells respond to compressive stress through diverse signaling pathways and structural adaptations. For example, pressure enhanced invasion via PIEZO1 in breast cancer [24], and promoted migration through the MEK1/ERK1 pathway in brain tumors [50], and upregulated CDC42, thereby enhancing proliferation, survival, and chemoresistance in ovarian cancer [25]. However, limited studies have investigated the biological effects and underlying mechanisms of pressure in CRC, despite the fact that pressure is a naturally occurring factor in CRC progression, arising from various sources [7–20]. While the applied pressure range (300–1400 Pa) is consistent with in vitro mechanobiology protocols, its physiological relevance to human CRC solid stress warrants further correlation with in vivo measurements [26, 65]. RNA sequencing of colorectal cancer cells under mechanical compression versus control conditions provided a global view of pressure-induced transcriptional changes and identified multiple pathways potentially associated with aggressive phenotypes (Fig. 11g–j). Notably, while TGFBR1 signaling contributed to proliferation, the persistence of migration and invasion following TGFBR1 knockdown implies fibronectin 1 may also utilize TGFBR1-independent mechanotransduction effectors, such as integrin signaling, to drive invasion [45]. To our knowledge, comprehensive transcriptomic profiling of pressure responses in colorectal cancer has been limited in previous studies, highlighting the novelty and relevance of these findings. Additionally, the context-dependent role of TGF-beta in CRC progression, acting as either tumor-suppressive or promoting depending on stage, should be considered when interpreting these signaling results [66].

From a clinical perspective, two potential strategies may be considered. First, fibronectin 1 may serve as a prognostic biomarker in colorectal cancer, though heterogeneity suggests validity may vary across patient subgroups. Second, minimizing avoidable artificial sources of mechanical pressure during clinical procedures, such as endoscopy or surgery, may warrant further investigation, particularly given differences in intraabdominal pressure settings during laparoscopic procedures [67]. However, the present study was designed to evaluate the general biological effects of compressive stress rather than to replicate a specific clinical scenario. Therefore, our findings provide preliminary insight into the potential role of compressive forces in colorectal cancer progression and highlight the need for future clinical studies to evaluate the effects of pressure generated during endoscopy, surgery, or bowel obstruction on disease progression.

Furthermore, the study had several limitations. Firstly, some data in the meta-analysis were indirectly extracted from figures using WebPlotDigitizer and therefore may not exactly match the original data, although extraction reliability was prioritized to ensure accuracy. Secondly, the cut-off value of fibronectin 1 was not reported because of differences in measurement methods, and some studies did not provide criteria for defining the cut-off. Finally, the present study was primarily based on in vitro models, and in vivo validation was not performed because accurately reproducing controlled mechanical compression in animal models remains technically challenging. Future work should employ appropriate in vivo models, such as orthotopic models with induced obstruction, to confirm the pressure- fibronectin 1 axis within a complex TME [68].

Conclusion

Our meta-analysis identifies fibronectin 1 is a poor prognostic marker in CRC and is associated with distant metastasis. Mechanistically, fibronectin 1 and mechanical pressure cooperate to drive cancer progression by enhancing proliferation, migration, and invasion in both non-CSCs and CSCs. While TGFBR1 signaling specifically mediates pressure-induced proliferation and colony formation, the abolition of migration and invasion by FN1 knockdown despite preserved TGFBR1 function suggests the involvement of TGFBR1-independent mechanotransduction effectors. These findings highlight the potential clinical benefit of minimizing mechanical pressure during diagnostic and therapeutic procedures and support the use of fibronectin 1 expression as a biomarker for risk stratification and management in CRC.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CRC

Colorectal cancer

TME

Tumor Microenvironment

TNM

Tumour-node-metastasis

CSCs

Cancer stem cells

TCSC

Tissue cancer stem cell

EMT

Epithelial–mesenchymal transition

LVI

Lymphovascular invasion

PNI

Perineural invasion

IHC

Immunohistochemistry

ELISA

Enzyme-linked immunosorbent assay

RT-PCR

Real-time reverse transcription polymerase chain reaction

NOS

Newcastle–Ottawa Scale

SMDs

Standardized mean differences

HRs

Hazard ratios

ORs

Odds ratios

CI

Confidence intervals

Pa

Pascal

Author contributions

Viet Cuong Nguyen conducted the literature search, collected data from the databases, performed the meta-analysis, carried out experiments, and wrote the manuscript. Kuang-Chao Cheng analyzed the data and carried out experiments. Yu-Hsin Lin, Thi-Luu Ho, Thuy-Tien Thi Phan wrote and revised the manuscript. Yao-An Shen collected data from the databases and revised the manuscript. Li-Jen Kuo and Chi-Long Chen provided research ideas, performed the meta-analysis, and revised the manuscript.

Funding

This work was supported by grants from the National Science and Technology Council, Taiwan (NSTC 114-2320-B-038-001 to Chi-Long Chen; NSTC 113-2823-8-038-001, NSTC 113-2628-B-038-012, and NSTC 114-2628-B-038-009 to Yao-An Shen) and grants from the National Health Research Institutes (NHRI-EX114-11404E and NHRI-EX115-11404ΕΙ to Yao-An Shen).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Colorectal cancer patient-derived tumor tissues were collected at Taipei Medical University Hospital (IRB Number: N202404078). All participants provided written informed consent prior to inclusion in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.