Exposure of endothelial cells to doxorubicin inhibits extracellular matrix production by dermal fibroblasts in a paracrine manner
Zhu Jiang, Giulia Sorrentino, Madalena Lopes Natário Pinto Gomes, Amber Swan-Taylor, Suat Simsek, Joris J T H Roelofs, Hans W M Niessen, Paul A J Krijnen
TL;DR
Exposure of endothelial cells to doxorubicin harms skin tissue by disrupting fibroblast function and extracellular matrix production.
Contribution
The study reveals a paracrine mechanism linking endothelial dysfunction to fibroblast activation and ECM disruption after doxorubicin exposure.
Findings
Doxorubicin induces endothelial-to-mesenchymal transition in endothelial cells.
Conditioned medium from doxorubicin-treated endothelial cells causes fibroblast senescence and reduced collagen production.
Altered cytokine profiles suggest impaired communication between endothelial cells and fibroblasts.
Abstract
Doxorubicin (Dox) is a potent chemotherapeutic with known vascular toxicity and connective-tissue damage. Endothelial cells (EC) and fibroblasts crosstalk is essential for vascular homeostasis and extracellular matrix (ECM) remodeling. This study aimed to explore whether Dox induces endothelial-to-mesenchymal transition (EndMT) and the paracrine effects of Dox-exposed EC on fibroblasts activation, senescence, and ECM synthesis. Human umbilical vein endothelial cells (HUVECs) were treated with Dox, and conditioned medium (CM) from EC was applied to human dermal fibroblasts for short- and long-term culture. Dox induced EndMT in ECs. Fibroblasts exposed to CM from Dox-treated EC exhibited early activation with increased fibroblast activation protein (FAP) and α-smooth muscle actin (α-SMA) at day 3, followed by a progressive senescent phenotype marked by elevated p21 and reduced Lamin B1 at…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8| Gene | Protein | Primers | Sequence (5′-3′) |
|---|---|---|---|
|
| MMP1 |
Forward Reverse |
TCGGGGAGAAGTGATGTTCT GTCGGCAAATTCGTAAGCAG |
|
| MMP9 |
Forward Reverse |
CGGACCAAGGATACAGTTTGTT TCAGGGCGAGGACCATAGA |
|
| YWHAZ |
Forward Reverse |
CATCTTGGAGGGTCGTCTCA ACTTTGCTCTCTGCTTGTGAA |
|
| GPI |
Forward Reverse |
GACGGCGAAGGAGTGGTTTC AGGGCAATGGAGAGTCCGAT |
- —China Scholarship Council10.13039/501100004543
- —European Union’s Horizon 2020 Marie Skłodowska-Curie Actions—Innovative Training Networks
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsChemotherapy-induced cardiotoxicity and mitigation · Chemotherapy-related skin toxicity · Breast Cancer Treatment Studies
Doxorubicin (Dox) is a widely used anthracycline chemotherapeutic agent, effective against various malignancies (Bloom et al. 2016). However, its clinical application is significantly limited by well-documented dose-dependent cardiotoxicity (Bhatia 2020). Increasing attention has also been directed toward its systemic toxicity, affecting multiple organs and the vasculature (Smith et al. 2010; Soultati et al. 2012). In the skin, Dox has been associated with impaired wound healing (Lawrence et al. 1986), skin necrosis (Yilmaz et al. 2002), and reduced regenerative capacity (Korać and Buzadzić 2001).
The skin is a highly vascularized organ, where dynamic interactions between vascular cells are essential for maintaining tissue integrity and regulating injury response. Endothelial cells (ECs), which form the inner layer of blood vessels, play important roles in vascular barrier function, inflammatory regulation, and paracrine signaling (Vita 2011). During intravenous chemotherapy, ECs are directly exposed to Dox and are particularly susceptible to damage. Fibroblasts, key mediators of extracellular matrix (ECM) synthesis, wound healing, angiogenesis, and fibrosis, are also critical for tissue repair and homeostasis (Kirk et al. 2021). Dysregulated fibroblast activity during chemotherapy can exacerbate disease progression. In Dox-induced cardiotoxicity models, activated fibroblasts are primary contributors to ECM accumulation and myocardial fibrosis (Huyan et al. 2024). In vitro studies have demonstrated that Dox can trigger pro-inflammatory responses and induce cellular senescence in fibroblasts (Espitia-Corredor et al. 2022).
One key mechanism of endothelium dysfunction is the endothelial-to-mesenchymal transition (EndMT), in which ECs lose endothelial markers and acquire mesenchymal properties. EndMT has been linked to pro-fibrotic activity and development of fibrosis in various pathological progression, including cardiac fibrosis (Zeisberg et al. 2007) and pulmonary fibrosis in mice (Hashimoto et al. 2010), as well as post myocardial infarction remodeling in human (Jiang et al. 2024). Additionally, under pathological conditions, ECs released a complex secretome of proteins and biomolecules which is crucial for endothelial behavior regulation, intercellular communication and shaping tissue microenvironment (Mathivanan et al. 2010). For instance, ECs undergoing EndMT in obese adipose tissue released paracrine or endocrine signals that modulated neighboring vascular and stromal cells (Haynes et al. 2019).
Moreover, dysfunctional endothelial cells can adopt a senescence-associated secretory phenotype, characterized by the release of cytokines and growth factors that promote inflammation, thrombosis, and vascular dysfunction (Fyhrquist et al. 2013; Han and Kim 2023). In chronic kidney disease, an altered endothelial secretome has been implicated in tubulointerstitial fibrosis (Lipphardt et al. 2017).
Although vascular injury is increasingly recognized as a driver of Dox-induced fibrosis, its role in modulating fibroblast behavior in the skin remains poorly understood. In particular, the impact of Dox on skin vasculature and dermal ECM homeostasis has yet to be fully elucidated.
In this study, we investigated the effects of Dox on endothelial-fibroblast crosstalk using a human cell model. EndMT was assessed in HUVECs following Dox treatment, and the endothelial secretome was collected. The Dox-induced EC secretome was applied to human dermal fibroblast to evaluate alterations in activation markers, senescence indicators and ECM remodeling over short-term (3 days) and long-term (21 days) periods.
Materials and methods
Materials and cells
Chemicals and reagents used in this study are listed in Table S1. Human Umbilical Vein Endothelial Cells (HUVECs) were purchased from ScienCell (#8000, Carlsbad, California) and cultured in Endothelial Cell Medium with endothelial cell growth supplement, antibiotic solution, and fetal bovine serum (FBS) (ECM; ScienCell). Healthy skin fibroblasts were obtained from the Burn Research Lab, Alliance of Dutch Burn Care, Beverwijk, The Netherlands. Fibroblasts were isolated from 7 donors (5 female, 1 male, 1 unknown sex; mean age 50 ± 6 yr). The use of the coded, post-operative residual tissue materials was approved by patients through the informed opt-out-plus protocol of the Red Cross Hospital (https://www.coreon.org/wp-content/uploads/2023/06/Code-of-Conduct-for-Health-Research-2022.pdf). Fibroblast isolation was performed as previously described (van den Bogaerdt et al. 2002). The fibroblasts were cultured in Dulbecco’s Modified Eagle Medium (DMEM, 41966029, ThermoFisher, USA), supplemented with 10% FBS (F7524, Sigma-Aldrich, St. Louis, Missouri), 1% penicillin-streptomycin (15140-122, Thermo Scientific, USA) and 1% GlutaMAX (35050-038, Thermo Scientific, USA), referred to as complete DMEM.
Doxorubicin hydrochloride (#DM1515, Sigma-Aldrich, St. Louis, Missouri) was dissolved in Dimethylsulfoxide (DMSO) to prepare a 15-mM stock solution and stored at −80 °C.
Cell culture
The experimental set up was shown in Fig. 1. The HUVECs and fibroblasts were cultured at 37 °C in a humidified 5% CO_2_ environment. Cells used in all experiments were passage 2 to 4.
Experimental design. Shown is a graphical outline of the experimental design of this study. For evaluating Dox-induced EndMT, HUVECs were exposed to Dox, (10 µM for 1 h; or 1 µM for 24 h), followed by a 7-day recovery in Dox-free medium to assess EndMT using immunofluorescent microscopy (IF). To assess the paracrine effects of Dox-exposed EC on dermal fibroblasts, conditioned medium (CM) was collected from HUVEC cultures, 24 h after exposure to Dox or DMSO. This CM was then applied to primary human dermal fibroblasts up to 21 days of culture, with or without supplemented magnesium ascorbyl phosphate (VitC). The effects of CM on fibroblast activation and senescence were assessed at 3 and 21 days (WB or/and IF); on collagen synthesis (WB, collagen assay, eosin staining) was evaluated at 21 days; and on MMP1 and MMP9 gene expression (qPCR) at 2 days.
HUVECs were seeded in 6-well or 24-well plates. For immunofluorescent analysis, cells were cultured on glass coverslips in 24-well plates. When the cells reached approximately 90% confluence, they were exposed to endothelial cell medium containing Dox at a final concentration of 10 µM for 1 h or 1 µM for 24 h. These Dox concentrations were chosen to model acute and prolonged exposure, respectively, based on the pharmacokinetic studies indicating that peak plasma Dox concentration in patients can exceed 10 µM before gradually declining (Mross et al. 1988). As Dox was dissolved in DMSO, and equivalent DMSO concentrations were used as controls. After exposure, cells were cultured in fresh Dox-free endothelial cell medium for 7 days with medium changes every 2 days.
To obtain the secretome of endothelial cells, after Dox or DMSO exposure HUVECs were cultured in fresh Dox-free complete DMEM for 24 h, after which the conditioned medium was collected. Fibroblasts were seeded in 24-well plates (on glass coverslips for immunofluorescence) and, at ∼90% confluence, cultured in conditioned medium for 2 to 21 days with medium refreshed every 2 days. Given the essential role of vitamin C (ascorbic acid) as a cofactor in ECM synthesis, magnesium ascorbyl phosphate (VitC), a stable vitamin C derivative that is enzymatically converted to ascorbic acid in cells, was added to the conditioned medium for the 21-day cultures at a final concentration of 10 µg/ml.
To evaluate the effects of IL-6 on fibroblast ECM production, fibroblasts were treated with recombinant human IL-6 at 330 pg/ml, corresponding to those measured in Dox-induced EC secretome, or 3300 pg/ml, representing a supraphysiological 10-fold higher concentration, with parallel controls maintained in complete DMEM without IL-6. Cells were cultured for up to 21 days with medium changes every 2 days, with or without vitamin C supplementation as indicated.
To assess the direct effects of doxorubicin on fibroblasts, dermal fibroblasts were exposed at ∼90% confluence to 10 µM Dox for 1 h or 1 µM Dox for 24 h, with DMSO-treated cells serving as controls. Following exposure, cells were cultured in Dox-free complete DMEM for up to 21 days with medium changes every 2 days, with or without Vit C supplementation as indicated.
Western blot
The cells were scraped from wells and lysed with sample buffer containing 4% (v/v) SDS, 18% (v/v) Glycerol, and 0.2 M Tris-HCl (pH 6.8). Protein concentration of prepared samples was determined using Pierce BCA Protein Assay kit (23225, Thermo Scientific, USA) and absorbance at 562 nm was measured with CLARIO star Plus microplate reader. Equal amounts of protein were calculated and prepared for loading by mixing samples with Laemmli buffer [4% v/v SDS, 18% v/v Glycerol, 2% v/v β-mercaptoethanol, 0.002% w/v Bromophenol blue dissolved in 0.2 M (pH 6.8) Tris-HCl], followed by heating for 5 min at 95 °C. For the supernatant samples’ preparation, equal volumes of the supernatant were mixed with Laemmli buffer in a 1:1 ratio and heated at 95 °C for 5 min.
Samples were separated using NuPAGE Bis-Tris Mini Protein Gels (4% to 12%, 1.0 to 1.5 mm, NP0322BOX, Thermo Scientific, USA) at 80 V for 20 min, followed by 120 V for 45 min. Proteins were transferred from gel onto a PVDF membrane at 120 V for 1 h. After extensive washing with TBST, membranes were blocked with 5% (w/v) skim milk in TBST for 1 h at RT.
Membranes were incubated overnight at 4 °C with primary antibodies: mouse-anti-α-SMA (Dako; M0851, 1:1000), rabbit-anti-FAP (NOVUS; NBP2-66844, 1:1000), rabbit-anti-Collagen I (Abnova; PAB10190, 1:1000 dilution), or rabbit-anti-S100A4 (abcam; ab197896, 1:1000). Following 3 washes with TBST [TBS with 0.1%(v/v) tween 20], the membranes were incubated with secondary antibodies: rabbit-anti-mouse IgG-HRP (Dako; P0260, 1:2500 dilution) for α-SMA or goat-anti-rabbit IgG-HRP (Dako; P0448, 1:2500 dilution) for FAP, Collagen I and S100A4, for 1 h at RT. Detection was performed using Pierce ECL Western Blotting Substrate (32106, Thermo Scientific, USA).
For proteins with similar molecular weights, membranes were stripped using stripping buffer [2% (v/v) SDS, 0.7% (v/v) β-mercaptoethanol, and 60 mM Tris-HCl, pH 6.7]. After extensive washing with TBST, the membranes were re-blocked with a solution of 5% (w/v) skim milk in TBST for 1 h at RT and incubated overnight at 4 °C with primary antibodies: rabbit-anti-p21 (Cell Signaling, CS2947, 1:1000 dilution), rabbit-anti-Lamin B1 (abcam; ab16048, 1:1000 dilution), or mouse-anti-β-actin (Sigma; A1578, 1:50000 dilution). β-actin served as the internal reference. After washing 3 times with TBST, membranes were incubated with secondary antibodies: rabbit-anti-mouse IgG-HRP (Dako; P0260, 1:2500 dilution) for β-actin or goat-anti-rabbit IgG-HRP (Dako; P0448, 1:2500 dilution) for p21 and LaminB1, for 1 h at RT. Visualization steps were performed as described above. Western blot bands were quantified using Image J. Target protein intensities (α-SMA: 44 kDa; FAP: both 80 and 88 kDa; S100A4: 12 kDa; Lamin B1: 68 kDa; p21: 21 kDa; Collagen I: 139 kDa) were normalized to housekeeping protein β-actin (44 kDa) within the same lane. Fold changes were calculated relative to corresponding control samples.
Immunofluorescent staining
For immunofluorescent staining, cells were fixed after treatment with 4% paraformaldehyde for 10 min at RT and subsequently permeabilized using 0.1% (v/v) Triton X-100 in PBS for 10 min. After extensive PBS washes, cells were blocked with a solution of 1% BSA and 0.1% NaN_3_ in PBS and then incubated with primary antibodies: mouse-anti-α-SMA (Dako; M0851, 1:1000) or rabbit-anti-FAP (NOVUS; NBP2-66844, 1:1000) for 2 h at RT. After washing 3 times with PBS, cells were incubated with secondary antibodies: goat-anti-rabbit IgG(H + L) A633 (Thermo Fisher Scientific; A21070, 1:500 dilution) for FAP or goat-anti-mouse IgG(H + L) A488 (Thermo Fisher Scientific; A11029, 1:500 dilution) for α-SMA, for 1 h at RT. Hoechst (1:1000) was used for nuclear counterstaining. After staining, cells were mounted with Prolong Gold antifade reagent (P36934; Thermo Fisher Scientific) onto glass slides. Cells that were stained without a primary antibody and without any antibody served as negative controls. A Leica DMi8 inverted microscope equipped with a Leica TCS SP8 X DLS camera was used to image the slides at 600× magnification.
Collagen detection and morphology quantification
Eosin staining was used to visualize and quantify ECM produced by fibroblasts in culture. After 21 days of culture, fibroblasts were washed 3 times with PBS, stained with 200 µl of 1% eosin solution for 2 min, and rinsed 5 times with PBS. Images of each well were acquired using a digital camera, and staining intensity was quantified using QuPath v0.6.0. Insoluble and soluble collagen levels were measured using Sircol Collagen Kits (S1000, S2000, Biocolor Ltd, UK) following the manufacturer’s instructions.
For morphological assessment, brightfield images of fibroblasts in culture were captured at 200× magnification using an inverted light microscope equipped with a digital camera. Cellular morphology was quantified by determining the elongation factor, calculated as the ratio of maximum cell length to maximum cell width. Per image the average of 10 to 20 cells was used.
RNA isolation, cDNA synthesis, and qRT-PCR
Fibroblasts were harvested after a 2-day culture of conditioned medium. RNA was extracted using TRI Reagent (T9424, Sigma-Aldrich) following the acid guanidinium thiocyanate-phenol-chloroform extraction method (Chomczynski and Sacchi 1987).cDNA synthesis was performed using M-MLV reverse transcriptase (28025013, Thermo Fisher Scientific) according to the manufacturer’s protocol. qRT-PCR was conducted with cDNA using primers listed in Table 1. The primers were synthesized using Eurogentec (Maastricht, The Netherlands). Reactions were performed with sensiFAST SYBR master mix (BIO-98005, Bioline reagents, London, UK) on the Roche LightCycler 480 platform. YWHAZ and GPI served as the internal reference genes. Reaction without template DNA served as a negative control. Cycling conditions as follows: 95 °C for 2 min, followed by 50 three-step cycles of 95 °C for 5 s, 65 °C for 10 s, and 72 °C for 10 s.
qRT-PCR results were analyzed using LC480 Conversion, LinRegPCR and LightCycler 480.
Cytokine measurement in Dox-induced EC secretome
The cytokine composition of the conditioned media (CM) from Dox- and DMSO-treated ECs was measured using LEGENDplex Human Essential Immune Response Panel (13plex, 740930, BioLegend), following the manufacturer’s instructions.
Statistics
GraphPad Prism (version 9, San Diego, California) was used for the design and generation of graphs. Statistical analyses were conducted using GraphPad Prism, SPSS (version 26.0, Armonk, New York), and Excel (2016, Microsoft, USA). Each experiment was replicated a minimum of 3 times. t-tests were applied to compare 2 groups. One-way analysis of variance (ANOVA) was applied to compare 3 or more groups. Data were presented using mean ± standard deviation (SD), P-value below 0.05 was considered statistically significant.
Results
Dox treatment induced EndMT
To investigate whether Dox treatment induces EndMT, HUVECs were treated with 10 µM Dox for 1 h or 1 µM Dox for 24 h followed by culture in Endothelial Cell Medium for 7 days. Immunofluorescence analysis showed a significant upregulation of fibroblast biomarker S100A4 in HUVECs treated with both Dox concentrations compared with cells treated with DMSO-containing medium or untreated cells (control) (Fig. 2A). S100A4 expression was respectively 6.5 ± 5.0-fold and 6.2 ± 3.1-fold higher in 10 µM and 1 µM Dox-treated cells compared with untreated cells (both *P *< 0.05) (Fig. 2B). CD31 expression (marker of endothelial cells) showed no significant changes compared with DMSO-treated or untreated cells (Fig. 2C). Exposure to the vehicle DMSO had no significant effects on S100A4 expression compared with untreated cells (data not shown).
*Dox treatment-induced effects on endothelial-to-mesenchymal transition (EndMT). A) Representative immunofluorescence images of S100A4 and CD31 staining in HUVECs show increased S100A4 expression 7 days after Dox treatment (10 µM, 1 h), indicating induced EndMT. White arrows indicate CD31-positive cell membranes (red), whereas double-head white arrows indicate cytoplasmic S100A4 positivity (green). Nuclei were counterstained with Hoechst in blue, as shown in the merged images. B, C) Quantification of S100A4 and CD31 expression in HUVECs treated with Dox (10 µM for 1 h or 1 µM for 24 h), assessed by quantifying the relative fluorescence intensity from immunofluorescence images. Fifteen images per condition were analyzed across 3 individual experiments (n = 3), using identical imaging settings. Data are represented as Mean ± SD. P<0.05, ns indicates not significant.
Paracrine effects of Dox-exposed EC on fibroblast activity and myofibroblast transition
To investigate the effect of the secretome from Dox-treated endothelial cells on fibroblast activation and myofibroblast transition, fibroblasts from 7 donors were cultured in conditioned medium (CM) from HUVECs treated with 10 µM Dox for 1 h (CM^Dox10^) or 1 µM Dox for 24 h (CM^Dox1^) or equivalent DMSO concentrations (CM^DMSO10^ or CM^DMSO1^), for 3 days (short-term) or 21 days (long-term) (Fig. 1). For the 21-day cultures, the CM was supplemented with 10 µg/ml VitC.
Fibroblast activation and myofibroblast transition were assessed via quantification of the expression of fibroblast activation marker FAP, myofibroblast marker α-SMA, and fibroblast marker S100A4. Immunofluorescence staining revealed cytoplasmic localization of FAP (red) and filamentous α-SMA (green) structures characteristic of fibroblast activation and myofibroblast differentiation (Fig. 3A).
*Paracrine effects of Dox-exposed EC on fibroblast activity and myofibroblast transition. A) Representative immunofluorescence images showing FAP (red) and α-SMA (green) in human dermal fibroblasts treated with conditioned medium (CM) harvested from HUVECs that were exposed to Dox (10 µM, 1 h) for 3 days. White arrows indicate cytoplasmic FAP; double-headed white arrows indicate filamentous α-SMA. Nuclei were counterstained with Hoechst in blue, as shown in the merged images. B) Representative Western blot (WB) images showing bands for FAP, α-SMA, S100A4, and β-actin in human dermal fibroblasts treated with CM harvested from HUVECs exposed to Dox (Dox-CM) or equivalent doses of DMSO (DMSO-CM) for 3 days or 21 days (with or without magnesium ascorbyl phosphate (VitC)). C, D) Quantification of WB results for FAP, α-SMA and S100A4 expression, relative to β-actin, in human dermal fibroblasts from 7 individual donors (n = 7), treated with CM harvested from HUVECs exposed to Dox (Dox-CM) or equivalent doses of DMSO (DMSO-CM), with/without VitC, for 3 or 21 days. The results for CM of HUVECs exposed to 10 µM Dox (or DMSO) for 1 h are shown in (C), and for CM of HUVECs exposed to 1 µM Dox (or DMSO) for 24 h are shown in (D). Data is shown normalized to fibroblasts treated with DMSO-CM. P < 0.05.
WB analysis confirmed the presence of FAP [distinct bands at ∼80 to 88 kDa, representing different forms of FAP (Lee et al. 2004)], α-SMA (42 kDa) and S100A4 (12 kDa), both after 3 and 21 days (Fig. 3B). After 3 days, CM^Dox10^-treated fibroblasts had significantly increased FAP (1.3 ± 0.3-fold, *P *< 0.05) and α-SMA (1.4 ± 0.3-fold, *P *< 0.05) protein levels, whereas S100A4 protein levels did not differ significantly compared with CM^DMSO10^-treated cells (Fig. 3C). Similarly, CM^Dox1^-treated fibroblasts showed significantly higher FAP (1.4 ± 0.4-fold, *P *< 0.05) and α-SMA (1.4 ± 0.5-fold, *P *< 0.05) levels, with similar S100A4 levels compared with CM^DMSO1^-treated cells (Fig. 3D).
After 21 days, sustained significant upregulation of FAP (1.5 ± 0.3-fold, *P *< 0.05) was observed in CM^Dox10^-treated fibroblasts, in the absence of VitC, but not when VitC was present (Fig. 3C). Although in CM^Dox1^-treated cells, no significant changes in FAP or α-SMA were observed compared with CM^DMSO1^-treated cells, either with or without VitC (Fig. 3D).
These results show that Dox-induced EC secretome promotes dermal fibroblast activation and myofibroblast transition, especially in response to higher Dox concentrations.
Paracrine effects of Dox-exposed EC on fibroblast senescence
To evaluate fibroblast senescence in response to Dox-induced EC secretome, fibroblasts from 7 donors were cultured for either 3 or 21 days in CM. Cells were then assessed for morphological changes and analyzed by WB for protein levels of the senescence marker p21 and the negative senescence marker Lamin B1 (Hernandez-Segura et al. 2018).
Both CM^Dox10^- and CM^Dox1^ treatments induced a significant elongation of fibroblasts compared with CM^DMSO10^- and CM^DMSO1^ treatments, both with and without VitC (Fig. 4A, left). The cellular elongation factors in response to CM^Dox10^ [17.7 ± 2.3 (−VitC) and 18.7 ± 2.2 (+VitC)] and CM^Dox1^ [17.8 ± 2.0 (−VitC) and 21.0 ± 3.4 (+VitC)] more than doubled compared with CM^DMSO10^ [8.5 ± 0.9 (−VitC) and 8.0 ± 0.7 (+VitC)] and CM^DMSO1^ [8.3 ± 1.5 (−VitC) and 8.5 ± 0.4 (+VitC)] (*P *< 0.0001 for all; Fig. 4A, right). VitC had no significant additional effects on cell morphology.
*Paracrine effects of Dox-exposed EC on fibroblast senescence. (A) Left: Representative bright field images of human dermal fibroblasts treated with conditioned medium (CM) harvested from HUVECs exposed to Dox (Dox-CM; either 10 µM for 1 h or 1 µM Dox for 24 h) or equivalent doses of DMSO (DMSO-CM), for 21 days (with or without magnesium ascorbyl phosphate (VitC)). The differences in cellular morphology are visible. Right: Quantification of the elongation factor, calculated as the ratio of the maximum to minimum cell diameter in the treatments described under (A) left. B) Representative Western blot (WB) images showing bands of Lamin B1, p21, and β-actin in human dermal fibroblasts treated with conditioned medium (CM) from HUVECs exposed to Dox (Dox-CM; either 10 µM for 1 h or 1 µM Dox for 24 h) or equivalent doses of DMSO (DMSO-CM), for 3 days or 21 days (with or without magnesium ascorbyl phosphate (VitC)). C, D) Quantification of WB results for p21 and Lamin B1 expression, relative to β-actin, in human dermal fibroblasts from 7 individual donors (n = 7), treated with CM harvested from HUVECs exposed to Dox (Dox-CM) or equivalent doses of DMSO (DMSO-CM), with/without VitC, for 3 or 21 days. The results for CM of HUVECs exposed to 10 µM Dox (or DMSO) for 1 h are shown in (C) and for CM of HUVECs exposed to 1 µM Dox (or DMSO) for 24 h are shown in (D). Data is shown normalized to fibroblasts treated with DMSO-CM. P < 0.05.
WB analysis showed p21 as a distinct band at 21 kDa and Lamin B1 at 68 kDa (Fig. 4B). In CM^Dox10^-treated fibroblasts p21 expression significantly increased compared with CM^DMSO10^-treated cells at both time points, with a 2.0 ± 1.1-fold at 3 days and a further increase to 2.5 ± 1.0-fold at day 21 (*P *< 0.05 for both) (Fig. 4C). The presence of VitC enhanced this CM^Dox10^-induced increase in p21 expression to 7.0 ± 3.9-fold compared CM^DMSO10^-treated cells (*P *< 0.05). In parallel, CM^Dox10^ significantly decreased Lamin B1 protein levels to 0.6 ± 0.2-fold at 3 days and further reduced to 0.4 ± 0.1-fold by day 21 (*P *< 0.05 for both) compared with CM^DMSO10^ treatment. This decrease appeared more pronounced in the presence of VitC, with Lamin B1 expression decreasing to 0.2 ± 0.1-fold (*P *< 0.05) (Fig. 4C).
A similar trend was observed in response to CM^Dox1^ treatment where p21 expression was significantly elevated at day 3 (2.4 ± 1.0-fold) and at day 21 (2.5 ± 1.5-fold) (*P *< 0.05 for both), with further enhancement by VitC supplement (5.1 ± 1.6-fold, *P *< 0.05). Lamin B1 levels decreased to 0.6 ± 0.3-fold at day 3 and 0.5 ± 0.1-fold at day 21 (*P *< 0.05 for both) and decreased further with VitC to 0.2 ± 0.1-fold (*P *< 0.05) (Fig. 4D).
These results indicate that Dox-induced EC secretome promotes a senescent phenotype in dermal fibroblasts.
Paracrine effects of Dox-exposed EC on collagen production and ECM formation
To evaluate the effects of Dox-induced EC secretome on fibroblast function, fibroblasts from 7 donors were cultured in CM for 21 days with or without VitC, and subsequently analyzed for ECM formation and collagen production.
Eosin staining intensity was used to quantify ECM formation by fibroblasts (Fig. 5A, left). In the absence of VitC, eosin staining intensity after CM^DMSO10^-treatment was 0.3 ± 0.1, which was significantly decreased by CM^Dox10^-treatment (0.1 ± 0.1; P < 0.05) (Fig. 5A, right). The presence of VitC significantly increased ECM deposition by CM^DMSO10^-treated cells to 0.6 ± 0.1 (P < 0.01), which again was significantly lowered by CM^Dox10^-treatment (0.2 ± 0.1; P < 0.01). Similarly, CM^Dox1^-treatment significantly decreased eosin staining intensity to 0.1 ± 0.1 (−VitC) and 0.2 ± 0.1 (+VitC) compared with CM^Dox10^-treatment [0.4 ± 0.1 (−VitC) and 0.5 ± 0.1 (+VitC); P < 0.05] (Fig. 5A, right).
*Paracrine effects of Dox-exposed EC on fibroblast collagen production and ECM formation. A) Left: Representative eosin staining images of human dermal fibroblasts treated with conditioned medium (CM) harvested from HUVECs exposed to Dox (Dox-CM; either 10 µM for 1 h or 1 µM Dox for 24 h) or equivalent doses of DMSO (DMSO-CM), for 21 days [with or without magnesium ascorbyl phosphate (VitC)]. Formed ECM stains red. Right: Quantification of the eosin staining intensities in the treatments described under (A) left. B, C) Quantification of insoluble collagen and soluble collagen levels in the harvested fibroblasts + ECM after the treatment described under (A) left. The results for CM of HUVECs exposed to 10 µM Dox (or DMSO) for 1 h are shown in (B) and for CM of HUVECs exposed to 1 µM Dox (or DMSO) for 24 h are shown in (C). D) Representative Western blot (WB) images showing Collagen I expression in the supernatants of human dermal fibroblasts treated as described in (A). Supernatant was collected daily and from each donor the samples from days 1 to 7, 8 to 14, and 15 to 21 were combined to form 3 groups over time. E, F) Quantification of supernatant collagen I levels from human dermal fibroblasts, from 7 individual donors (n = 7), treated as described in (A). The results for CM of HUVECs exposed to 10 µM Dox (or DMSO) for 1 h are shown in (E) and for CM of HUVECs exposed to 1 µM Dox (or DMSO) for 24 h are shown in (F). Data are shown normalized to the supernatant of fibroblasts treated with DMSO-CM. *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001 and ns indicates not significant.
In addition, soluble and insoluble collagen in the cells and ECM were quantified using Sircol collagen assay. Soluble collagen reflects newly synthesized, non-crosslinked collagen, whereas insoluble collagen represents mature, crosslinked fibers integrated into the ECM. No significant difference in insoluble collagen was measured between CM^Dox10^-treated and CM^Dox10^-treated cells, independent of VitC (Fig. 5B). However, soluble collagen levels after CM^Dox10^-treatment decreased to a significant 0.2 ± 0.3-fold of CM^DMSO10^-treated cells (P < 0.05), only in the presence of VitC. In CM^Dox1^-treated cells, in the presence of VitC, insoluble collagen levels were significantly reduced to 0.6 ± 0.3-fold of CM^DMSO1^-treated cells (P < 0.05), whereas soluble collagen levels were significantly decreased both in the absence (0.5 ± 0.3-fold; P < 0.05) and presence of VitC (0.1 ± 0.1-fold; P < 0.0001) compared with CM^DMSO1^ (Fig. 5C).
To assess whether the observed inhibitory effects on ECM formation are related to an inhibitory effect of collagen crosslinking rather than production, the levels of collagen type 1 (collagen-1) protein were quantified via WB in the supernatants of the cultures that were collected across 3 time windows (days 1 to 7, 8 to 14, and 15 to 21) (Fig. 5D). This time-resolved analysis accounts for the cumulative and dynamic secretion of ECM proteins, thereby enabling a more accurate assessment of temporal changes in collagen release and potential impairment in collagen crosslinking or stabilization.
Collagen-1 levels in the supernatant of CM^Dox10^- and CM^Dox1^-treated cells in all 3-time windows were significantly lower than those in the supernatant of CM^DSMO10^- and CM^DMSO1^-treated cells, regardless of the presence of VitC (Fig. 5E, P < 0.05 for all). The presence of VitC significantly enhanced the levels of collagen-1 (2.0 ± 1.3-fold or higher) in the supernatant of both CM^DMSO10^- and CM^DMSO1^-treated fibroblasts (Fig. 5E; P < 0.05 for all). Of note, collagen-1 levels in the supernatant after CM^DMSO^-treatment did not differ significantly relative to fibroblasts cultured in standard DMEM (data not shown). Although VitC also significantly enhanced collagen-1 levels in the supernatant of CM^Dox10^- and CM^Dox1^-treated cells (P < 0.05; apart from not significant at days 1 to 7 in CM^Dox10^), this effect was less pronounced.
These results suggest that the Dox-induced EC secretome significantly impairs collagen production and ECM formation.
Direct effects of Dox on collagen production and ECM formation
To evaluate the direct effects of Dox on fibroblast collagen production and ECM formation, fibroblasts from 7 donors were treated with 10 µM Dox for 1 h or 1 µM Dox for 24 h followed by culture in Dox-free medium for 21 days with or without VitC. ECM formation and collagen production were subsequently assessed.
In the absence of VitC, eosin staining intensity was significantly decreased in 10 µM Dox-treated cells compared with DMSO-treated cells (0.26 ± 0.02 vs 0.34 ± 0.04; P < 0.0001). VitC markedly increased ECM deposition in DMSO-treated cells to 0.53 ± 0.09 (P < 0.001), which was significantly lowered by Dox-treatment (0.35 ± 0.04; P < 0.001). Similarly, 1 µM Dox treatment significantly decreased eosin staining intensity to 0.24 ± 0.01 (−VitC) and 0.30 ± 0.02 (+VitC) compared with DMSO-treatment [0.35 ± 0.04 (−VitC) and 0.53 ± 0.08 (+VitC); P < 0.001] (Fig. 6A).
*Dox effects on fibroblast collagen production and ECM formation. A) Left: Representative eosin staining images of human dermal fibroblasts were exposed to Dox (Dox: either 10 µM for 1 h or 1 µM for 24 h) or equivalent doses of DMSO (DMSO), followed by 21-day culture in Dox-free medium with or without magnesium ascorbyl phosphate (VitC). Right: Quantification of the eosin staining intensities in the treatments described under (A) left. B) Left: Representative Western blot (WB) images showing collagen I expression in the supernatants of human dermal fibroblasts treated as described in (A). Supernatant was collected daily, for each donor, samples from days 1 to 21 were pooled to generate a single sample. Right: Quantification of supernatant collagen I levels from human dermal fibroblasts, from 7 individual donors (n = 7), treated as described in (A). Data are normalized to the supernatant of fibroblasts cultured in complete DMEM with DMSO. *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001 and ns indicates not significant.
Collagen-1 levels in the supernatant collected between days 1 and 21 were significantly lower in both 10 µM and 1 µM Dox-treated fibroblasts compared with DMSO-treated cells, regardless of the presence of VitC (Fig. 6B, P < 0.05 for all). VitC significantly enhanced the collagen-1 levels in the supernatant of DMSO-treated fibroblasts (2.9 ± 0.9-fold or higher, P < 0.05). Although a less pronounced but significant increase was also observed in Dox-treated fibroblasts (P < 0.05) (Fig. 6B).
Paracrine effects of Dox-exposed EC on the transcriptional levels of MMP1 and MMP9
Matrix metalloproteinases (MMPs) are ECM-degrading enzymes that can be produced by fibroblasts, and MMP1 and MMP9 have been shown to be involved in skin ageing (Quan et al. 2023). To evaluate the effects of Dox-induced EC secretome MMP1 and MMP9 expression, fibroblasts from 7 donors were cultured in CM for 2 days, with or without VitC, and subsequently harvested for qRT-PCR analysis.
CM^Dox10^-treatment resulted in a significant 3.2 ± 1.9-fold increase in MMP1 transcription (P < 0.05), and a borderline significant 2.4 ± 1.6-fold increase in MMP9 transcription (P = 0.0644) compared with CM^DMSO10^. These effects were enhanced in the presence of VitC, where CM^Dox10^-treatment increased MMP1- and MMP9 transcriptional levels 4.0 ± 2.2-fold and 5.2 ± 6.3-fold respectively (*P *< 0.05 for both) (Fig. 7A). Similarly, CM^Dox1^-treatment significantly increased MMP1- and MMP9 transcriptional levels 4.1 ± 3.8-fold and 7.6 ± 5.4-fold (P < 0.05 for both) compared with CM^DMSO1^, whereas VitC supplementation amplified this response especially for MMP1, whose transcriptional levels increased 16.9 ± 22.4-fold compared with CM^DMSO1^, whereas MMP9 remained at 7.6 ± 5.4-fold (*P *< 0.05 for both) (Fig. 7B).
*Paracrine effects of Dox-exposed EC on the transcriptional levels of MMP1 and MMP9. A—B) Quantification of MMP1 and MMP9 transcriptional expression relative to YWHAZ and GPI in human dermal fibroblasts, from 7 donors (n = 7), treated with conditioned medium (CM) harvested from HUVECs exposed to Dox [Dox-CM; either (A) 10 µM for 1 h or (B) 1 µM Dox for 24 h] or equivalent doses of DMSO (DMSO-CM), for 2 days [with or without magnesium ascorbyl phosphate (VitC)]. Data are shown normalized to fibroblasts treated with DMSO-CM. P < 0.05 and ns indicates not significant.
Inflammatory cytokine changes in Dox-induced EC secretome and IL-6 effects on ECM formation
The cytokine composition of the conditioned medium (CM) from Dox-treated ECs was analyzed using a multiplex bead-based immunoassay, which simultaneously quantified IL-4, IL-2, CXCL10 (IP-10), IL-1β, TNF-α, CCL2 (MCP-1), IL-17A, xl, IL-10, IL-6, IFN-γ, IL-12p70, CXCL8 (IL-8) and free active TGF-β1. Among the 13 cytokines measured, IL-4, IL-2, TNF-α, IL-17A, IL-10, IFN-γ, IL-12p70 and free active TGF-β1were below the detection limit (<2.0 pg/ml).
Compared with CM^DMSO10^ and CM^DMSO1^-treated fibroblasts, CXCL10 (IP-10) was significantly reduced only in CM^Dox1^ (4.1 ± 0.3 vs 2.4 ± 0.4 pg/ml; P < 0.05) but remained unchanged in CM^Dox10^. Both CM^Dox10^ and CM^Dox1^ treated fibroblasts showed significant reductions in IL-1β (5.8 ± 0.2 vs 1.6 ± 0.6 pg/ml and 5.2 ± 0.9 vs 0.6 ± 1.1 pg/ml, respectively; P < 0.001 for both) and CCL2 (MCP-1) (4585 ± 240 vs 1917 ± 122 pg/ml and 4603 ± 263 pg/ml vs 659 ± 27 pg/ml, respectively; P < 0.0001 for both), alongside significant increases in IL-6 (213 ± 13 vs 328 ± 21 pg/ml and 162 ± 12 vs 327 ± 11 pg/ml, respectively; P < 0.0001 for both). CXCL8 (IL-8) levels remained unchanged in both conditions (Fig. 8A and B).
*Cytokine measurements in conditioned media and IL-6 effects on ECM formation. A, B) Cytokine [CXCL10 (IP-10), IL-1β, CCL2 (MCP-1), IL-6, CXCL8 (IL-8)] measurements in conditioned medium (CM) from HUVECs treated with 10 µM Dox for 1 h (CMDox10) or 1 µM Dox for 24 h (CMDox1) or equivalent DMSO concentrations (CMDMSO10 or CMDMSO1), data from 3 individual experiments (n = 3). C) Left: Representative eosin staining images of human dermal fibroblasts cultured for 21 days in complete DMEM containing recombinant human IL-6 (330 or 3300 pg/ml) or vehicle control (complete DMEM without IL-6) with or without magnesium ascorbyl phosphate (VitC). Formed ECM stains red. Right: Quantification of the eosin staining intensities in the treatments described under (A) left. D) Left: Representative Western blot (WB) images showing collagen I expression in the supernatants of human dermal fibroblasts treated as described in (C). Supernatant was collected daily, for each donor, samples from days 1 to 21 were pooled to generate a single sample. Right: Quantification of supernatant collagen I levels from human dermal fibroblasts, from 7 individual donors (n = 7), treated as described in (C). Data are normalized to the supernatant of fibroblasts cultured in complete DMEM without IL-6. *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001 and ns indicates not significant.
To assess whether IL-6 mediates fibroblast responses similar to Dox-induced EC secretome, fibroblasts were treated with recombinant human IL-6 at 330 or 3300 pg/ml. ECM formation was quantified by eosin staining at 21 days of culture (Fig. 8C), and collagen levels in the culture supernatant collected between days 1 and 21 were assessed by WB (Fig. 8D). At 330 pg/ml, IL-6 did not significantly affect ECM formation compared with control, either in the absence of VitC (control vs IL-6: 0.34 ± 0.04 vs 0.34 ± 0.04) or in its presence (control vs IL-6: 0.53 ± 0.09 vs 0.53 ± 0.08) (Fig. 8C). Consistently, collagen levels in the culture supernatant were unchanged (Fig. 8D).
In contrast, IL-6 at 3300 pg/ml induced a mild but significant increase in ECM formation both without VitC (control vs IL-6: 0.34 ± 0.04 vs 0.36 ± 0.04, P < 0.01) and with VitC (control vs IL-6: 0.53 ± 0.09 vs 0.58 ± 0.08, P < 0.001) (Fig. 8C), whereas collagen levels in the supernatant remained unaffected (Fig. 8D). As expected, VitC significantly enhanced ECM deposition and collagen secretion across all experimental conditions.
Discussion
In this study, we investigated the effects of acute (1 h at10 µM) and prolonged (24 h at 1 µM) Dox exposure on ECs, as well as the subsequent impact of their secretome on dermal fibroblast function. Our findings revealed that Dox treatment induced EndMT in HUVECs and that the Dox-induced EC secretome modulated fibroblast activation and senescence induction, ultimately resulting in suppressed collagen production and ECM synthesis during prolonged incubation.
HUVECs exposed to Dox exposure and subsequently cultured for 7 days of recovery in Dox-free medium exhibited a significant upregulation of the fibroblast marker S100A4 while preserving CD31 expression, indicating a partial or early-stage EndMT phenotype that has been reported to be functionally relevant (Greenspan and Weinstein 2021). The vascular implications of EndMT have been well-documented, leading to disassembly of endothelial intercellular junctions, compromised endothelial barrier integrity, impaired angiogenic capacity, and fibrosis (Cho et al. 2018; Haynes et al. 2019; Sun et al. 2020). In vivo, studies have shown that Dox treatment reduced cardiac microvascular density and impaired cardiac function in mice (Yin et al. 2016), with endothelial fate tracing confirming that EndMT-derived cells contributed to Dox-induced cardiac fibrosis (Tsai et al. 2019). The EndMT following Dox treatment observed in this study reinforced these observations, suggesting that Dox-induced EndMT may be a potential mechanism in microvascular rarefaction and tissue hypoperfusion, both of which are critical contributors to tissue aging and impaired wound healing (Zeng and Chen 2019; Wang-Evers et al. 2021). Beyond direct vascular effects, through their secretome endothelial cells exert potent paracrine effects on adjacent and even distal stromal cell populations. For instance, during physiological skin repair, endothelial‐derived factors such as VEGF, EGF and PDGF promote fibroblast proliferation and ECM synthesis, facilitating re-epithelialization and collagen deposition in murine wounds (Lu et al. 2021). However, under pathological conditions, such as EndMT in hypertrophic scars or metabolic stress in obesity, the EC secretome alters toward TGF-β, bFGF, and endothelin-1, driving fibroblast activation/conversion and excessive matrix accumulation in dermis (Tan et al. 2024), adipose tissue, and pressure overloaded (Zeisberg et al. 2007) or diabetic (Widyantoro et al. 2010) hearts. In our study, the endothelial cells exposed to Dox showed markedly elevated IL-6 secretion, but reduced IL-1β, CCL2 (MCP-1), and CXCL10 levels, indicating a shift in cytokine profile in their secretome. Although IL-6 upregulation is consistent with findings in Dox-treated human coronary artery endothelial cells, the study also reported modestly elevated levels of IL-1β, IL-8, and CCL2 expression (Sinitskaya et al. 2024), suggesting that Dox-induced cytokine responses may vary depending on endothelial cell type, exposure duration, or concentration. IL-6 has previously been linked to EndMT-associated phenotypes in endothelial cells. For instance, inhibition of autophagy in human microvascular endothelial cells induced EndMT, which could be prevented with an IL-6-neutralizing antibody, indicating that the EndMT was IL-6-dependent (Takagaki et al. 2020). In fibroblasts, IL-6 has been associated with profibrotic activation, myofibroblast differentiation, and senescence. However, many of these studies employed cardiac or pulmonary fibroblasts, or used supraphysiological IL-6 levels (Li et al. 2022; Cao et al. 2025). In the skin, sustained IL-6 elevation is more closely associated with fibroblast senescence and delayed wound healing than with enhanced ECM deposition (Johnson et al. 2020). This indicates that the effects of IL-6 likely depend on fibroblast origin, inflammatory context, and cytokine concentration. In this study, recombinant IL-6 at concentrations corresponding to those detected in the Dox-induced EC secretome was not associated with changes in fibroblast ECM formation, collagen secretion, or fibroblast morphology following prolonged culture, whereas supraphysiological IL-6 exposure induced only a mild increase in ECM formation. These findings indicate that IL-6 alone, at physiologically relevant concentrations, is unlikely to account for the impaired ECM production observed following exposure to the Dox-induced EC secretome, whereas supporting a dose- and context-dependent modulatory role of IL-6 in dermal fibroblast function. Beyond IL-6, the Dox-induced EC secretome was characterized by reduced IL-1β and CCL2 levels, cytokines that mediate early monocyte/macrophage recruitment (Liu et al. 2013) to clear debris and guide fibroblast-mediated matrix remodeling (Deshmane et al. 2009) during normal tissue repair. The inhibitory effect of Dox on their production by EC observed in this study may impair critical early phases of wound healing. Alterations in this cytokine balance may therefore reflect a broader inflammatory dysregulation rather than the action of a single dominant mediator. In this context, the altered cytokine profile observed in the Dox-induced EC secretome may represent an endothelial inflammatory phenotype that contributes to the delayed wound healing observed in individuals undergoing chemotherapy (Słonimska et al. 2024).
CM from Dox-treated ECs profoundly affected human dermal fibroblasts’ behavior. Fibroblasts exposed to Dox-induced EC secretome for 3 days showed enhanced expression of FAP, indicating increased activation of fibroblasts. Activated fibroblasts (FAP+) contribute to pro-fibrotic activation and tissue remodeling and have been implicated in infarct tissue remodeling in myocardial infarction patients (Jiang et al. 2024) and fibrotic tissue formation during skin wound repair in mice (Rinkevich et al. 2015). Concurrently, the expression of α-SMA, a key marker of myofibroblasts, was significantly upregulated, indicating a phenotypic transition toward a contractile, fibrosis-driving state. This myofibroblast phenotype is essential for ECM remodeling and wound contraction (Darby et al. 2014). A similar pro-fibrotic activation was shown to underlie Dox-induced cardiac fibrosis via PI3K/Akt-and Smad-dependent upregulation of α-SMA, TGF-β, and collagen (Levick et al. 2019; Podyacheva et al. 2022; Patricelli et al. 2023). Notably, free active TGF-β1 was undetectable in our CM.
The early activation of fibroblasts likely represented an acute reactive fibrotic response to paracrine stress signals from dysfunctional ECs. However, prolonged exposure to Dox-induced EC secretome did not show an elevated ECM formation. Instead, fibroblasts displayed signs of senescence, including upregulation of p21 and downregulation of Lamin B1, both indicative of impaired proliferative capacity and altered secretory profiles that may further disrupt tissue repair.
Vitamin C, a well-known enhancer of collagen synthesis and ECM stabilization through the promotion of collagen molecule hydroxylation, crosslinking, and collagen expression, was supplemented in long-term culture to promote ECM formation (Pullar et al. 2017). Although VitC indeed supported collagen production in fibroblasts cultured with normal EC secretome (DMSO-induced EC secretome), fibroblasts exposed to the Dox-induced EC secretome showed significantly suppressed collagen deposition despite VitC supplementation. Quantification analysis confirmed a reduction in both intracellular and secreted collagen synthesis, accompanied by impaired ECM deposition visualized by eosin staining, which reflected a cumulative impact of Dox-induced endothelial dysfunctional secretome on connective tissue formation. Morphological elongation of Dox-induced EC secretome-exposed fibroblasts coincided with disrupted ECM production and increased senescence markers. Similar phenotypic changes have been observed in fibroblasts from aged skin and fibrotic tissues, where fibroblasts lose their matrix production capacity and adopt a senescence-associated secretory phenotype (Pitiyage et al. 2011; Gerasymchuk et al. 2022).
Furthermore, qPCR analysis revealed a transcriptional upregulation of matrix metalloproteinases MMP1 and MMP9 in fibroblasts treated with the Dox-induced EC secretome. Both enzymes are responsible for ECM degradation and are commonly associated with aged human skin (Fisher et al. 2023). Elevated MMP1, which primarily targeted interstitial collagens (type I, II, and III), has been linked to collagen breakdown and skin aging in UV-exposed human skin (Varani et al. 2006). Although MMP9 degrades denatured collagens and basement membrane proteins, particularly type IV and V collagen, involved in vascular morphogenesis and angiogenesis (Ardi et al. 2009; Wang and Khalil 2018). Although MMP1 and MMP9 are essential for wound healing (Caley et al. 2015), their overactivity can exacerbate collagen degradation, weaken tissue tensile strength, and contribute to the abnormal tissue remodeling, commonly observed in aging and fibrotic tissue pathology (Visse and Nagase 2003). The transcriptional upregulation of both MMP1 and MMP9 observed in this study may disrupt the dermal ECM and contribute to the vascular basement membrane disruption, potentially exacerbating vascular dysfunction in Dox-treated patients.
Clinically, the mode of Dox administration has been adjusted to mitigate cardiotoxicity, with continuous intravenous infusion commonly used instead of rapid bolus injection to reduce peak plasma concentration, which have been linked to cellular damage and morphological changes in the cardiac tissue (Legha et al. 1982). In our study, however, despite differences in treatment duration and concentration, both acute high-dose exposure (10 µM for 1 h) and prolonged low-dose exposure (1 µM for 24 h) elicited comparable endothelial responses, including EndMT induction and the secretome that promoted fibroblast activation and senescence, ultimately impairing ECM production. These findings suggest that lowering peak plasma concentrations alone may not be sufficient to prevent the adverse effects of Dox on endothelial- and fibroblast function we observed in our study.
In conclusion, these findings provide novel insights into Dox-induced endothelial dysfunction and its downstream paracrine effects on dermal fibroblast function. Dox induced EndMT impairs endothelial function and may compromise microvascular function, which is essential for nutrient delivery, oxygenation, and wound healing. Moreover, the conditioned-medium-based approach in this study enabled clear attribution of fibroblast responses to endothelial-derived soluble factors in a doxorubicin-free environment, independent of juxtracrine interactions or direct drug toxicity. The exposure to Dox alters the secretome of endothelial cells, which initially promotes fibroblast activation, followed by fibroblast senescence and reduced ECM production. These pathological effects of Dox may underlie the impaired connective tissue regeneration and ageing of the skin observed in Dox-treated murine models (Demaria et al. 2017) and delayed skin wound healing observed in patients following chemotherapy (Deptuła et al. 2019; Słonimska et al. 2024). These results underscore the need for therapeutic strategies aimed at preserving vascular health and regulating endothelial-fibroblast interactions to mitigate ECM disruption and promote effective wound healing in patients undergoing Dox-based chemotherapy.
Supplementary Material
kfag027_Supplementary_Data
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Ardi VC , Van den Steen PE, Opdenakker G, Schweighofer B, Deryugina EI, Quigley JP. 2009. Neutrophil mmp-9 proenzyme, unencumbered by timp-1, undergoes efficient activation in vivo and catalytically induces angiogenesis via a basic fibroblast growth factor (FGF-2)/FGFR-2 pathway. J Biol Chem. 284:25854–25866.19608737 10.1074/jbc.M 109.033472 PMC 2757987 · doi ↗ · pubmed ↗
- 2Bhatia S. 2020. Genetics of anthracycline cardiomyopathy in cancer survivors: JACC: cardiooncology state-of-the-art review. JACC Cardio Oncol. 2:539–552.33364618 10.1016/j.jaccao.2020.09.006PMC 7757557 · doi ↗ · pubmed ↗
- 3Bloom MW , Hamo CE, Cardinale D, Ky B, Nohria A, Baer L, Skopicki H, Lenihan DJ, Gheorghiade M, Lyon AR, et al 2016. Cancer therapy-related cardiac dysfunction and heart failure: part 1: definitions, pathophysiology, risk factors, and imaging. Circ Heart Fail. 9:e 002661.26747861 10.1161/CIRCHEARTFAILURE.115.002661 PMC 4709035 · doi ↗ · pubmed ↗
- 4Caley MP , Martins VL, O’Toole EA. 2015. Metalloproteinases and wound healing. Adv Wound Care (New Rochelle). 4:225–234.25945285 10.1089/wound.2014.0581 PMC 4397992 · doi ↗ · pubmed ↗
- 5Cao C , Xu W, Lei J, Zheng Y, Zhang A, Xu A, Lin F, Zhou M. 2025. The IL-6 autocrine loop promoting IFN-γ-induced fibroblast senescence is involved in psychological stress-mediated exacerbation of vitiligo. Inflamm Res. 74:72.40299037 10.1007/s 00011-025-02035-2 · doi ↗ · pubmed ↗
- 6Cho JG , Lee A, Chang W, Lee MS, Kim J. 2018. Endothelial to mesenchymal transition represents a key link in the interaction between inflammation and endothelial dysfunction. Front Immunol. 9:294.29515588 10.3389/fimmu.2018.00294 PMC 5826197 · doi ↗ · pubmed ↗
- 7Chomczynski P , Sacchi N. 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem. 162:156–159.2440339 10.1006/abio.1987.9999 · doi ↗ · pubmed ↗
- 8Darby IA , Laverdet B, BontéF, Desmoulière A. 2014. Fibroblasts and myofibroblasts in wound healing. Clin Cosmet Investig Dermatol. 7:301–311.10.2147/CCID.S 50046 PMC 422639125395868 · doi ↗ · pubmed ↗