Superior Antiproliferative and Enhanced Synergistic Effects of a ROCK Inhibitor in Multiple Models for Keloid Disease

Abstract

Keloid disease is a common fibroproliferative skin disorder characterized by excessive scar tissue formation and frequent recurrence. Limited therapies and study models hinder progress in addressing this unmet clinical need. AMA0825, a ROCK (Rho-associated protein kinase) inhibitor, has shown promising antifibrotic and antiproliferative effects in other fibrotic conditions. This study investigated the therapeutic potential of AMA0825 using in vitro, ex vivo, and a 3-dimensional spheroid model of keloid disease, which partially reflects features of the keloid microenvironment. AMA0825 demonstrated potent antiproliferative activity against keloid fibroblasts, with a half-maximal growth inhibitory concentration of 28.19 ± 1.6 nM, significantly outperforming dexamethasone (half-maximal growth inhibitory concentration = 35.35 ± 2.6 μM) and triamcinolone (half-maximal growth inhibitory concentration = 37.84 ± 3 μM). This effect was confirmed by decreased Ki-67 expression and cell cycle arrest at the G1 phase. In the 3-dimensional spheroid model, AMA0825 effectively inhibited cell proliferation at nanomolar concentrations, exceeding the efficacy of dexamethasone. Although AMA0825 did not demonstrate significant antifibrotic activity at lower concentrations, it exhibited antifibrotic effects at higher concentrations. In addition, synergistic effects were observed when combined with dexamethasone. This study highlights the potential of ROCK inhibitors, particularly AMA0825, as an antiproliferative agent for keloid disease and underscores the value of 3-dimensional spheroid models for evaluating alternative therapeutic strategies.

Introduction

Keloid disease (KD) is a prevalent fibroproliferative disorder of the reticular dermis characterized by excessive collagen deposition and extracellular matrix (ECM) accumulation, leading to raised, disfiguring scars (Tan et al, 2019).This clinically aggressive disorder often presents therapeutic challenges owing to its poorly understood etiopathogenesis and high recurrence rates after treatment (Basson and Bayat, 2022; Ud-Din and Bayat, 2013). KD significantly impacts patients' lives, causing physical morbidity, psychosocial distress, and esthetic concerns. Conventional treatments, such as surgical excision and corticosteroids, frequently prove inadequate, resulting in persistent or recurrent keloid scars.

Given that KD is driven by both fibrosis and hyperproliferation of abnormal keloid fibroblasts (KFs), effective therapeutic strategies must target these dual mechanisms. Antiproliferative drugs, such as methotrexate, cyclophosphamide, and tyrosine kinase inhibitors, inhibit rapid cell division by interfering with DNA replication and signal transduction pathways (Olsen and Murray, 1989; Shariatifar et al, 2022; Taran et al, 2023). Conversely, antifibrotic drugs focus on reducing excessive collagen deposition by inhibiting fibroblast activation through various pathways, including TGF-β, Wnt/β-catenin, PDGF, Jak–signal transducer and activator of transcription signaling, and matrix metalloproteinase activity (Ahangari et al, 2022; Bai et al, 2021; Bellamri et al, 2023). This multipronged approach aims to disrupt the complex network of interactions that drive fibrosis (Norouzi-Barough and Bayat, 2021). The central role of both fibrosis and hyperproliferation in KD underscores the potential of combined antiproliferative and antifibrotic therapy for improved disease management and prevention.

ROCK (Rho-associated protein kinase) is a critical regulator of fibroproliferative processes, mediating both canonical and noncanonical TGF-β signaling (Kamaraju and Roberts, 2005; Xie et al, 2022).

ROCK inhibitors have shown promise in managing fibrosis in various organs (Ho et al, 2012; Holvoet et al, 2017; Kentrup et al, 2011; Shimizu et al, 2001; Tada et al, 2001) and have demonstrated antiproliferative effects in several studies (Al-Hilal et al, 2021; Chen et al, 2009; Dahal et al, 2010). Aberrant RhoA/ROCK pathway activation is implicated in skin fibrosis, suggesting that targeting this pathway with ROCK inhibitors could offer therapeutic benefits for KD (Dohi et al, 2019; Maeda et al, 2019). However, systemic ROCK inhibitors carry the risk of cardiovascular side effects (Enomoto et al, 2010; Ishihara et al, 2012).

This study focuses on AMA0825, a locally acting ROCK inhibitor known for its extended gut retention and rapid systemic metabolism, which minimizes systemic exposure and reduces the risk of cardiovascular complications (Boland et al, 2013). This localized action makes AMA0825 a particularly attractive candidate for the treatment of keloid scars.

Furthermore, to overcome the limitations of traditional 2-dimensional (2D) in vitro systems and the lack of an appropriate animal model for KD, we developed a 3-dimensional (3D) KF-derived spheroid (KFS) model. This model partially replicates aspects of the keloid microenvironment, providing a valuable platform for investigating pathobiology and evaluating potential therapies for KD.

Therefore, the aim of this study was to utilize in vitro, spheroid, and ex vivo keloid models to comprehensively evaluate the antiproliferative and antifibrotic efficacy of AMA0825 compared with those of conventional treatments such as dexamethasone (DEX), triamcinolone, imatinib (IMA), and SB525334 (a selective TGF-β1 receptor inhibitor). The experimental design schematic is shown in Figure 1.

Figure 1.

Schematic of experimental design.

Results

AMA0825 exhibits antiproliferative activity superior to DEX, triamcinolone, IMA, and SB525334

The in vitro CyQuant assay assessed the half-maximal growth inhibitory concentration (IC₅₀) and the potential antiproliferative effects of AMA0825, triamcinolone, DEX, IMA, and the ALK5 inhibitor SB525334 using primary human KFs. The IC₅₀ represents the concentration of a drug required to inhibit 50% of cell proliferation, providing a measure of potency.

The results, summarized in Table 1, show that AMA0825 significantly inhibited proliferation of primary KFs (K1–K3) with a mean IC₅₀ concentration of 28.19 ± 1.6 nM. In contrast, IMA, triamcinolone, and DEX exhibited higher IC₅₀ values of 38.96 ± 8.6 μM, 37.84 ± 3.0 μM, and 35.35 ± 2.6 μM, respectively, indicating substantially lower potency than AMA0825. Similarly, SB525334 showed an IC₅₀ of 53.11 ± 3.2 μM (Table 1).

Table 1.

Comparative Proliferation Assessment of AMA0825, Imatinib, TA, DEX, and SB525334 on Keloids Fibroblasts

IC50
Primary Cells	AMA0825 (nM)	Imatinib (μM)	TA (μM)	DEX (μM)	SB525334 (μM)
K1 (keloid from earlobe)	28.4 ± 0.9	45.34 ± 21.4	32 ± 6.4	34.15 ± 3.8	57 ± 7.1
K2 (keloid from earlobe)	28.28 ± 2.2	34.1 ± 3.3	36.2 ± 0.9	40.6 ± 1.9	51.7± 0.8
K3 (keloid from earlobe)	27.9 ± 1.8	37.45 ± 1.2	42.33 ± 1.7	31.31 ± 2.1	50.64 ± 1.9
Mean KD	28.19 ± 1.6	38.96 ± 8.6	37.84 ± 3	35.35 ± 2.6	53.11 ± 3.2

Abbreviations: DEX, dexamethasone; IC₅₀, half-maximal growth inhibitory concentration; KD, keloid disease; TA, triamcinolone.

IC₅₀ values were obtained after 24-hour exposure of 3 different primary keloid donors K1, K2, and K3. The IC₅₀ was calculated by nonlinear logarithmic regression using GraphPad Prism 5.0 software. Values are expressed as the mean ± SD from 3 independent experiments with sample size n = 3, each performed with 4 replicates per sample.

AMA0825 demonstrated a stronger antiproliferative potency against KFs, achieving comparable effects at nanomolar concentrations, whereas the other tested compounds required micromolar concentrations. DMSO, used at a final concentration of 0.2%, had no observable effect in this assay and served as a negative control.

Selective effect of AMA0825 and DEX on KF growth inhibition

Next, we performed a real-time cell analysis (RTCA) using Real-Time Cell Analysis (Agilent) to evaluate the effect of AMA0825 and DEX on primary KFs compared with that on normal fibroblasts (NFs) at a dose near their respective IC₅₀ concentrations (30 nM for AMA0825 and 40 μM for DEX). Figure 2a–c shows plots of the cell indices 24 hours after treatment for KFs (Figure 2a) and NFs (Figure 2b). After the 24-hour treatment period, both AMA0825 and DEX exhibited a significant decrease in cell indices against 3 KFs (K1, K2, and K3) and NFs (N1, N2, and N3) compared with the control. Figure 2a and b representatively illustrates RTCA plots of KFs and NFs 24 hours prior to treatment and 24 hours after treatment, respectively. Treatment with AMA0825 is indicated in pink, that with DEX is indicated in green, cells only (untreated) are indicated in blue, and culture media only without cells are indicated in black. GraphPad Prism (version 5.0) (GraphPad Software) was utilized to generate graphs of the cell index for RTCA 24 hours after treatment with AMA0825 and DEX, including untreated control cells in KFs and NFs. After treatment in KFs, we observed significantly lower cell index 24 hours after AMA0825 and DEX treatments than in untreated cells (P < .001) (Figure 2c). In NFs, both AMA0825 and DEX exhibited a similar trend as in KFs, reducing cell index after treatment. However, they displayed greater selectivity towards KFs with lower cell index than in NFs. A significant difference in cell index was noted between NFs and KFs after AMA0825 and DEX treatment. AMA0825-treated KFs showed lower cell index than NFs (P < .05) as well as after treatment with DEX (P < .05) (Figure 2c). KFs exhibited lower cell index at 24 hours across all treatment groups than NF control fibroblasts.

Figure 2.

Inhibition of proliferation by AMA0825/DEX in 2D-cultured keloid fibroblasts. Representative RTCA flow plot illustrates the impact of AMA0825 (30 nM) and DEX (40 μM) against (a) KFs and (b) NFs. (c) Average quantitative analysis of RTCA CI for KFs and NFs 24 hours after treatment. Statistical analyses were conducted on a sample size of n = 3 for both KFs and NFs, with 4 replicates for each sample, using GraphPad Prism (version 5.0). Two-way ANOVA followed by Bonferroni posthoc test was used to assess statistical significance. Data are shown as mean (cross in the bars) and whiskers (minimum to maximum values). Significance levels are indicated as follows: ∗P < .05 and ∗∗∗P < .001. 2D, 2-dimensional; CI, cell index; DEX, dexamethasone; KF, keloid fibroblast; NF, normal fibroblast; RTCA, real-time cell analysis.

AMA0825 and DEX suppress proliferation and fibrotic markers in KFs

The decrease in cell proliferation prompted us to confirm our 2D results using the cell proliferation marker Ki-67. The results in the representative Figure 3a showed KFs untreated and treated with 30 nM AMA0825 and 40 μM DEX, with DAPI and Ki-67 staining. Compared with the control, a noticeable decrease in Ki-67 expression was observed after treatment in both (Figure 3a). Treatment with AMA0825 led to a 43.35% reduction in the expression of the cell proliferation marker Ki-67 in KFs, compared with the control, which exhibited 78.68% Ki-67–positive cells. Similarly, DEX treatment resulted in a 45.46% decrease in Ki-67–positive cells, demonstrating a trend similar to that of AMA0825 treatment when compared with the control (Figure 3b). These results were consistent with our previous findings, demonstrating a decrease in cell proliferation after treatment.

Figure 3.

Effects of AMA0825/DEX treatment on Ki-67 expression using immunofluorescence staining with anti–Ki-67 antibody in 2D cultured keloid fibroblasts. (a) Ki-67 expression (green) in the nucleus of KF monolayers (K1, K2, and K3) using immunofluorescence staining with anti–Ki-67 antibody (1:250). Costaining was performed using DAPI (blue). (b) Bar graphs presenting the percentage of Ki-67+ in KF 24 hours after treatment by AMA0825 and DEX. Statistical analyses were conducted on a sample size of n = 3, with 3 replicates for each sample, using GraphPad Prism (version 5.0). One-way ANOVA followed by Dunnett's posthoc test was used to assess statistical significance. Significance levels are indicated as follows: ∗∗∗P < .001. Bars = 100 μm. Error bars represent the mean ± SD of 3 independent experiments. 2D, 2-dimensional; DEX, dexamethasone; KF, keloid fibroblast.

Furthermore, AMA0825 decreased collagen I expression by 51.43%, compared with the control (81.29%), whereas DEX led to a 49.98% reduction (Figure 4a and b). In a separate experiment, AMA0825 reduced αSMA expression by 49.02%, relative to the control (88.50%), whereas DEX resulted in a 51.64% decrease (Figure 4c and d).

Figure 4.

Effects of AMA0825/DEX treatment on collagen I and αSMA expression using immunofluorescence staining with anti–collagen I and αSMA antibodies in 2D cultured keloid fibroblasts. Keloid fibroblasts K1, K2, and K4 were stained for (a) collagen I (green) using immunofluorescence staining with anti–collagen I antibody (1:250), (b) with bar graphs representing the percentage collagen I⁺ cells, and for (c) αSMA (green) using immunofluorescence staining with anti-αSMA antibody (1:50), with (d) bar graphs representing the percentage αSMA⁺ cells 24 hours after treatment with AMA0825 and DEX. Costaining was performed using DAPI (blue) to visualize nuclei. Statistical analyses were conducted on a sample size of n = 3, with 3 replicates for each sample, using GraphPad Prism (version 5.0). One-way ANOVA followed by Dunnett's posthoc test was used to assess statistical significance. Significance levels are indicated as follows: ∗∗∗P < .001. Bars = 100 μm. Error bars represent the mean ± SD of 3 independent experiments. αSMA, α-smooth muscle actin; 2D, 2-dimensional; DEX, dexamethasone.

AMA0825 induces G1-phase cell-cycle arrest

To elucidate the effects of AMA0825 (30 nM) on cell cycle, the distribution of KFs in the various cell-cycle phases (G1, S, and G2) was determined by flow cytometry. The representative data show the accumulation of KFs in G1 phase compared with the control (Figure 5a and b). The accumulation of KFs in G1 phase occurred with concomitant decrease in S phase (26.82% vs 43.28% for control) and G2 phase (5.84% vs 10.92% for control) (Figure 5c).

Figure 5.

Modulation of cell cycle in 2D cultured keloid fibroblasts by AMA0825. After treatment with AMA0825, the cells (K1, K2, and K3) were incubated with propidium iodide (10 μg/ml) containing 200 μg/ml of RNase A. The amount of DNA was measured by flow cytometry. (a) The distribution of the cells through phase/point (G1, S, and G2) from left to right in KF untreated cells, and (b) cells treated with AMA0825, at 30 nM concentration. (c) Quantification analysis of G1, S, and G2 cell cycle from flow cytometry where the cell-cycle percentages were derived from flow cytometric analysis. (d) The mean percentage of cells in G1. Statistics were performed on sample size n = 3 with 3 replicates for each sample, using unpaired 2-tailed Student’s t-test and GraphPad Prism (version 5.0). Statistically different in relation to the untreated control cells is indicated as ∗∗∗P < .001. Error bars represent the mean ± SD of 3 independent experiments. 2D, 2-dimensional; KF, keloid fibroblast.

The percentages of G1 cell populations increased by a mean of 90.62% compared with 40.00% of untreated cells (P = .0015) (Figure 5d). These results suggest that growth inhibition effect of AMA0825 could be attributed to reduction in the progression rate of the cell cycle, with a concomitant decrease in proliferation.

Development and characterization of KFSs and NF-derived spheroids

To evaluate the biological activity of AMA0825, we employed a more reliable study model using KFs and NFs. Ultralow attachment plates coated with agar or poly-2-hydroxyethyl methacrylate were used to prevent cell attachment, inducing spheroid formation. Both KFSs and NF-derived spheroids (NFSs) formed within a day, taking a circular shape by day 2 and achieving maximum compactness by day 3 (Figure 6a–f). The diameter of NFSs increased with cell density, reaching 200.51 ± 8.96 μm³, 24.14 ± 7.74 μm³, and 29.66 ± 5.31 μm³ for 5k, 10k, and 20k NFs, respectively (Figure 6g–i and m).

Figure 6.

Morphological and volumetric analysis of spheroids generated from primary NFS and KFS cells at different seeding densities. Images of (a–c) N2 and (d–f) K1 primary cell spheroid synthesized in varying cell densities (5000; 10,000; and 20,000 cells) with continuous observation for 7 days. Bars = 100 μm. (g–i) Confocal images of N1-derived spheroids at 5000; 10,000; and 20,000 cell densities. (j–l) Confocal images of K1-derived spheroids at 5000; 10,000; and 20,000 cell densities. (m) Volume of NFS (N1, N2, and N3) and KFS (K1, K2, and K3) in μm³. Bars = 50 μm. Statistics were performed on a sample size of n = 3, with 3 replicates for each sample, using GraphPad Prism (version 5.0). Error bars represent the mean ± SD of 3 independent experiments for panel M. KF, keloid fibroblast; KFS, keloid fibroblast–derived spheroid; NFS, normal fibroblast–derived spheroid.

Nevertheless, KFSs were larger in size than NFSs but followed a similar pattern, with mean diameters of 218.39 ± 8.11 μm³, 269.84 ± 9.43 μm³, and 340.48 ± 5.96 μm³ in 5k, 10k, and 20k KFs, respectively (Figure 6j–m). The volume of both NFSs and KFSs are shown in Figure 6m. Furthermore, calcein/propidium iodide analysis after 7 days confirmed functionality, with high numbers of live cells and relatively low numbers of dead cells in both NFSs and KFSs (Figure 7a). In NFSs, the percentage of dead cells was highest at 20k NFS (52.17%) than at 10k (22.37%) and 5k (16.64%) NFSs. Similarly, the percentage of dead cells increased with the rise in cell density in KFSs from 22.06% at 5k to 29.46% at 10k and 63.81% at 20k cell density (Figure 7a). The increase in cell density led to higher cell death due to limited nutrients and the production of more necrotic cells.

Figure 7.

Assessment of viability and diameter of NFS and KFS spheroids across varying cell densities. (a) Live/dead cell staining (calcein/PI) of NFS/KFS spheroids with different cell densities (5000; 10,000; and 20,000 cells/well). Quantification of spheroids diameter over 7 days at optimized seeding densities of (b) 5000 cells/well for NFS and KFS and (c)10,000 for NFS and KFS. Statistics were performed on a sample size of n = 3, with 3 replicates for each sample, using GraphPad Prism (version 5.0). Error bars represent the mean ± SD of 3 independent experiments. KF, keloid fibroblast; KFS, keloid fibroblast–derived spheroid; NFS, normal fibroblast–derived spheroid; PI, propidium iodide.

Figure 7b and c illustrates the comparison between 5k and 10k NFS/KFS in terms of difference in size at different time points. The graphs reveal that the spheroids reached their maximum compactness on day 3 with the smallest diameter (174.99 ± 6.10 μm for 5k and 213.42 ± 8.64 μm for 10k NFSs) (Figure 7b and c). Similarly, KFSs displayed the same trend, achieving the highest compactness on day 3 with diameter of 206.98 ± 10.51 μm and 242.37 ± 9.31 μm for 5k and 10k, respectively (Figure 7b and c). The successful formation and characterization of NFSs and KFSs provide a reliable model for studying the biological activity of AMA0825. Sustained functionality and size dynamics of 5k and 10k spheroids ensure a robust and representative platform for treatment initiation. The lack of significant differences between 10k and 5k spheroids supports the use of 5k spheroids for further investigations.

Optimal treatment time point of 96 hours and dose-dependent effects of AMA0825 on 3D KFSs

We investigated the optimal treatment time point for AMA0825 on 5k KFS from donors K1 and K4, utilizing varying concentrations (0.5; 5; 50; 500; 5000; and 50,000 nM) of AMA0825. KFSs were treated on day 3 and monitored for up to 7 days using a fluorescence cell imager (ZOE, Bio-Rad), with images captured every 24 hours (Figure 8a). Twenty-four and 48 hours after treatment, no significant change in cell proliferation was observed (Figure 8a–c). However, at 72 hours, a notable decrease in cell proliferation was noted at higher concentrations (500; 5000; and 50,000 nM), with reductions of 51.25%, 62.50%, and 72.50%, respectively, compared with the untreated control (Figure 8a and d). The most substantial decrease in cell proliferation occurred at the 50 nM concentration after 96 hours, resulting in a 54.32% reduction compared with the untreated control. This reduction persisted at higher concentrations, with decreases of 66.66%, 74.07%, and 81.48% at 500; 5000; and 50,000 nM, respectively, compared with the control (Figure 8a and e). Thus, we conclude that the optimal treatment time is 96 hours (4 days) for subsequent experiments, as evidenced by these results. Furthermore, 5k KFSs from 4 donors were subjected to varying concentrations (5; 50; 500; 5000; and 50,000 nM) of AMA0825 after 4 days. The representative data of the spheroids treated at different concentration (5, 50, 500, and 5000 nM) are shown in Figure 9a, whereas additional concentrations were tested to enhance confidence. The results demonstrated a dose-dependent increase in percentage in cell proliferation inhibition at the 4-day treatment time points of all donors, with values of 61.22%, 77.55%, 91.83%, and 93.87% observed at concentrations of 50; 500; 5000; and 50,000 nM, respectively (Figure 9b). Notably, the positive control, DEX, exhibited a 59.18% of inhibition in KFS proliferation at a concentration of 80 μM, nearly equivalent to the 50 nM concentration of AMA0825. At lower concentrations (0.5 and 5 nM), minimal changes in the cell proliferation were observed. However, a decrease in cell proliferation was evident starting at a 50 nM concentration of AMA0825 in all donors compared with the control (Figure 9a and b). The decline in cell proliferation was consistently observed across all 4 KFSs at 500; 5000; and 50,000 nM concentrations in a dose-dependent manner (Figure 9b). Similarly, DEX at 80 μM concentration showed a significant effect on all KFSs compared with the control (Figure 9b). The effect of AMA0825 on spheroids at the nanomolar range was shown to be higher than that of DEX (Figure 9b).

Figure 8.

Optimizing AMA0825 treatment timing in spheroid model using CyQuant to detect DNA content as a surrogate for cell proliferation, along with cell imaging. (a) Representative figure demonstrating the impact of AMA0825 on K1 at concentrations ranging from 0.5 to 50,000 nM across various time points from 24 to 96 h. (b–e) Graphs represent proliferation data of K1 and K4 treated with different concentrations of AMA0825 at 24, 48, 72, and 96 h. Statistical analysis was performed on a sample size of n = 2, with 8 replicates per sample, using GraphPad Prism (version 5.0). One-way ANOVA followed by Dunnett's posthoc test was used to assess statistical significance. Significance levels are indicated as follows: ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001. Bars = 100 μm. Error bars represent the mean ± SD of 2 independent experiments for panels d and e. Ctrl, control; h, hour.

Figure 9.

Evaluation of dose-dependent responses to AMA0825/DEX in spheroid model after 96 hours using CyQuant DNA quantification as a surrogate for cell proliferation and cell imaging. (a) Representative figure illustrates the effect of AMA0825 and DEX on K1, K2, K3, and K4 at concentrations ranging from 0.5 to 50,000 nM after 96 h. (b) Representative data related to K1, K2, K3, and K4 treated with different concentrations of AMA0825 and 80 μM DEX at 96 h. Statistical analysis was performed on sample size n = 4 with 4 replicates for each sample using GraphPad Prism (version 5.0). One-way ANOVA followed by Dunnett's posthoc test was used to assess statistical significance. Significance levels are indicated as follows: ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001. Bars = 100 μm. Error bars represent the mean ± SD of 4 independent experiments for panel b. DEX, dexamethasone; h, hour; KFS, keloid fibroblast–derived spheroid.

Differential dose-dependent effects of AMA0825 and DEX on ECM remodeling and collagen reduction in spheroid model and biopsies

In KFSs, AMA0825 and DEX exhibited contrasting effects on ECM and collagen content. Visual assessment showed increased light green staining in the ECM of treated spheroids (Figure 10a), indicating collagen distribution. AMA0825 treatment over 4 days did not significantly alter ECM content compared with untreated controls in K1 and K2 spheroids. In contrast, DEX treatment led to a significant reduction in collagen levels (P = .0014), as shown in Figure 10b. This difference suggests that AMA0825 at this concentration, which showed antiproliferative effect, lacks the collagen-reducing effect observed with DEX in KFSs. To further assess the antifibrotic effects of AMA0825 at higher concentrations, we tested varying doses in the 3D spheroid model of 3 different donors. Interestingly, our results revealed that at 1000 nM, AMA0825 significantly reduced both collagen I and αSMA expressions, with P = .0124 and .0112, respectively (Figure 11a–d).

Figure 10.

Effect of AMA0825/DEX on collagen I in spheroid model assessed by Masson’s trichrome staining and quantitative analysis. (a) Representative figure of Masson’s trichrome staining of spheroids from K1 donor. Collagen appears light green, cytoplasm in red, and nuclei in dark blue, illustrating collagen distribution throughout the spheroids. Because collagen staining is not readily visible, image analysis was conducted using color deconvolution in FIJI-ImageJ to isolate and quantify the light green signal for collagen as described in Materials and Methods. Bar = 200 μm. (b) Quantification of collagen content in K1 and K2 donors was performed on a sample size of n = 2, with 6 replicates for each sample using GraphPad Prism (version 5.0). One-way ANOVA followed by Dunnett's posthoc test was used to assess statistical significance. Significance levels are indicated as follows: ∗P < .05. DEX, dexamethasone.

Figure 11.

Effect of AMA0825 on collagen I and αSMA expression in spheroid model assessed by immunofluorescence staining with anti–collagen I and αSMA antibodies and quantification. (a) Representative image demonstrating the effect of AMA0825 on collagen I (green) on K1, K2, and K4 spheroids using immunofluorescence staining with anti–collagen I antibody (1:250). (b) Corresponding graph representing collagen I⁺ cells. (c) Representative image showing the effect of AMA0825 on αSMA (green) using immunofluorescence staining with anti-αSMA antibody (1:50). (d) Corresponding graph quantifying αSMA⁺ cells. Statistical analyses were conducted on a sample size of n = 3 with 3 replicates for each sample. One-way ANOVA followed by Dunnett's posthoc test was used to assess statistical significance. Significance levels are indicated as follows: ∗P < .05. Error bars represent the mean ± SD of 2 independent experiments for panel b and 3 independent experiments for panels b and d. αSMA, α-smooth muscle actin; Ctrl, control.

Further quantitative analysis of keloid biopsies of 3 independent donors confirmed DEX's effect on tissue mass over an extended period. Keloid biopsies at 4 mm treated with AMA0825 (1 μM) or DEX (150 μM) every 3 days were evaluated by weight at day 0 and after 28 days. Images are shown to demonstrate the morphology of biopsies (Figure 12a). DEX treatment resulted in a significant tissue weight reduction by day 28 (P = .0319), whereas AMA0825 showed no comparable effect (Figure 12b). This aligns with prior studies (Syed et al, 2013a), supporting DEX's antifibrotic pathway distinct from that of AMA0825.

Figure 12.

Measurement of weight changes in keloid biopsies after treatment with AMA0825 or DEX. (a) Representative macroscopic images of the keloid biopsies K5, K6, and K7 at days 0, 14, and 28 days with and without AMA0825/DEX treatment. The images are shown to demonstrate the morphology of the biopsies. (b) Average weights of the keloid biopsies on days 0 and 48 with and without AMA0825/DEX treatment as indicated in the bar graph. Statistical analysis was performed on a sample size of n = 3, with 3 replicates for each sample using 1-way ANOVA followed by the Bonferroni posthoc test and GraphPad Prism (version 5.0). ∗P = .0319 indicates significant difference compared with untreated control group at day 0 (biopsy size = 4 mm). Error bars represent the mean ± SD of 3 independent experiments. DEX, dexamethasone.

Checkerboard assay further evidences the combinatorial effect of AMA0825 and DEX

Considering the potential differing action modes of AMA0825 and DEX, we sought to investigate their potential synergistic activity. Combining AMA0825 with DEX demonstrated an enhancement in cytotoxicity compared with individual drug treatments. Given the superior efficacy of AMA0825 over DEX in reducing the survival of KFs, from 3 independent donors, we employed checkerboard analysis to assess the combined effects of these drugs. Combinations of varying concentrations of AMA0825 (at half the IC₅₀, IC₅₀, and twice the IC₅₀ concentration) with DEX (40 μM) showed no significant impact on cytotoxicity. However, combining DEX with AMA0825 at higher concentrations (4 μM) notably increased cell susceptibility toward death. Calculation of the fractional inhibitory concentration (FIC) index values from survival curves indicated synergism between AMA0825 and DEX in KFs (FIC = 0.88 ± 0.1) (Table 2). This showed the synergism effect of DEX at 40 μM concentration with AMA0825 at 4 μM concentration (Table 3).

Table 2.

FIC Index for Treatment of Keloid Fibroblasts with a Combination of AMA0825 and DEX

FIC Index
Primary Cells	Keloid Fibroblasts (K1, K2, and K3)
AMA0825 + DEX	0.88 ± 0.1

Abbreviations: DEX, dexamethasone; FIC, fractional inhibitory concentration.

Combinatorial assay using K1, K2, and K3 in combination with AMA0825 and DEX was performed. Value is expressed as the mean ± SD from 3 independent experiments with sample size n = 3, each performed with 4 replicates per sample.

Table 3.

FIC Index for Treatment of Keloid Fibroblasts with a Combination of AMA0825 and DEX

Primary Cells	A IC50 (μM)	B IC50 (μM)	AB IC50 (μM)	BA IC50 (μM)	FIC Index
K1	4	40	3.41	1.75	0.8962
K2	4	40	3.12	1.02	0.7177
K3	4	40	3.80	1.34	0.984

Abbreviations: DEX, dexamethasone; FIC, fractional inhibitory concentration.

Combinatorial assay using K1, K2, and K3 in combination with A = AMA0825, B = DEX, AB = AMA0825 (0.001, 0.1, 0.5, 1, 50, 100, and 160 μM) + DEX (40 μM), and BA = DEX (0.001, 0.1, 0.5, 1, 50, 100, and 160 μM) + AMA0825 (4 μM) was performed. Value is expressed as the mean ± SD from 3 independent experiments with sample size n = 3, each performed with 4 biological replicates per sample.

Discussion

Keloid scars, a common connective tissue disorder disproportionately affecting individuals with skin of color, present a significant clinical challenge owing to their poorly understood etiology and limited treatment options (Ud-Din and Bayat, 2013). Emerging research suggests that ROCK inhibitors may play a crucial role in treating profibrotic disorders. In particular, the compound AMA0825 has shown considerable promise in preclinical models of intestinal fibrosis (Holvoet et al, 2017; Boland et al, 2013), prompting us to investigate its potential in treating keloids. Although ROCK inhibitors have been proposed for the treatment of dermal fibrosis, the current body of experimental evidence remains inconclusive. Therefore, the aim in this study was to investigate the effect of AMA0825 on KD using existing keloid study models, in vitro as well as an ex vivo model, plus using a recently in-house developed spheroid model of KFs.

We conducted our study to investigate the antifibrotic and antiproliferative effects of AMA0825 on KD. Our findings revealed that AMA0825 exhibits significant antiproliferative effects at low nanomolar concentrations, a property that could be crucial in preventing keloid recurrence. In particular, AMA0825 exhibits a significant growth inhibition on KFs, with an IC₅₀ of 28.19 ± 1.6 nM, outperforming conventional keloid steroid treatments such as triamcinolone acetonide and DEX, which are commonly used for their anti-inflammatory and antifibrotic effects (Basson and Bayat, 2022; Cain and Cidlowski, 2017).

Standard steroid dosages (25–100 μM) are associated with adverse local and systemic side effects (Lee and Kim, 2020). Although SB525334, a strong TGFβRI inhibitor, effectively blocks TGFβ-induced SMAD activation and reduces collagen production (Cheng et al, 2023; Heo et al, 2021), it carries the risk of severe side effects such as cardiac issues, cancer development, or immune system disruption (Itoh et al, 2009, Prud’homme, 2007). Similarly, although IMA is a potent antiproliferative agent that suppresses fibroblast-driven collagen production in fibrotic diseases (Distler and Distler, 2007), its efficacy is limited to high micromolar concentrations (IC₅₀ = 38.96 ± 8.6 μM). In contrast, AMA0825 is effective at nanomolar concentrations, indicating a potentially lower toxicity profile while maintaining its strong antiproliferative properties.

Interestingly, RTCA not only confirmed the superior growth inhibition of AMA0825 against KFs but also revealed its higher selectivity for KFs than for NFs. These findings were further validated using Ki-67 as a marker for proliferation, where a significant decrease in Ki-67 expression was observed, confirming the marked reduction in KF proliferation due to the effects of AMA0825. This is consistent with previous studies on ROCK inhibitors, such as Y-27632 and Y-33075, which demonstrated a decrease in cell proliferation in both murine and human hepatic stellate and hemangioma cells (Bachtler et al, 2023; Qiu et al, 2017).

The antifibrotic effects of both AMA0825 and DEX were further evident in KFs at their IC₅₀ concentrations. Similarly, a recent study by Ross et al (2024) demonstrated that at higher concentrations (10–20 μM), the ROCK inhibitor Y27632 inhibited collagen deposition in keloid-derived fibroblasts cultured in 2D.

Antiproliferative characteristics of AMA0825 were further explored by examining cell-cycle progression. ROCKs are critical for cell-cycle progression, particularly in cancer types where they regulate mitosis, chromosome alignment, and cell division (Fukata et al, 2001; Shi and Wei, 2007; Takeba et al, 2012; Wu et al, 2021). ROCKs also regulate cell proliferation, and inhibiting them leads to cell-cycle arrest and cellular senescence (David et al, 2012). ROCK inhibitors such as Fasudil have been shown to inhibit tumor cell migration, metastasis, and cell proliferation (Bhullar et al, 2018; Ridley, 2001; Takeba et al, 2012; Wu et al, 2021). Consistent with previous findings, AMA0825 (30 nM) induced G1-phase cell-cycle arrest in KFs, exerting its antiproliferative effects by inhibiting cell-cycle progression.

To explore and validate the in vitro studies, we developed and characterized 3D spheroids from KFs and NFs. Our multicellular spheroid model using ultralow attachment plates demonstrated low cell death rates, with spheroids maintaining viability for up to 7 days and sizes remaining under 500 μm³, even at the highest cell density. Although diffusion limitations have been a concern in previous keloid 3D models (Agudelo-Garcia et al, 2011; Jeon et al, 2019; Kunz-Schughart et al, 2016), our results showed that both 5k and 10k NFSs/KFSs exhibited cell death rates below 50% (16.64% and 22.37% in 5k and 10k NFSs, respectively and 22.06% and 29.46% in 5k and 10k KFSs, respectively), with spheroid sizes consistently staying below 500 μm³, even at the highest cell densities. These findings are consistent with those reported (Dirand et al, 2023).

Subsequent treatment of KFS with AMA0825 at different concentrations and different time points revealed the best treatment time of 4 days, which did not extend beyond day 7. AMA0825 inhibited KFS proliferation by 61.22% at 50 nM concentration after 4 days of treatment (Figure 6g). The inhibition of spheroid proliferation increased to 77.55, 91.83, and 93.87% with concentrations of 500; 5000; and 50,000 nM, respectively. In comparison, DEX (80 μM) reduced KFS proliferation by 59.18% after 4 days. These results highlight the longer treatment time required for AMA0825 in 3D cultures than in 2D, possibly due to differences in drug penetration, gene expression, and survival signaling (Breslin and O'Driscoll, 2016; LaBonia et al, 2016; Riedl et al, 2017).

Although AMA0825 demonstrated potent antiproliferative activity, it did not significantly reduce collagen deposition at low concentrations in 3D KFS models or ex vivo keloid explants studies. This suggests that higher concentrations may be necessary to achieve an antifibrotic effect or that ROCK inhibition primarily targets the proliferative aspect of KD at low concentrations. However, although AMA0825 did not produce a measurable reduction in collagen content in KFSs at low concentrations, its antifibrotic effect became evident at higher concentrations (1 μM) in 3D model. Notably, in 3D cultures, Y27632 at 10 μM significantly reduced collagen content in human orbital fibroblast spheroids (Hikage et al, 2021), which aligns with our findings. As the concentration of AMA0825 increased, we observed a reduction in both collagen content and αSMA expression, further supporting its antifibrotic potential.

Whereas AMA0825 had no antifibrotic effect at lower concentrations in both KFSs and keloid explants, DEX did reduce fibrosis within spheroids as well as keloid explants, confirming its antifibrotic activity (Kidzeru et al, 2025; Kim et al, 2016; Lu et al, 2022; Saarela et al, 2003; Syed et al, 2013a). These results suggest that whereas AMA0825 primarily contributes to the antiproliferative response at low concentration, DEX is responsible for reducing collagen content in KFS and keloid biopsies. The impact of steroidal anti-inflammatory drugs on KD treatment, such as DEX (Kidzeru et al, 2025; Syed et al, 2013a) and other corticosteroids (Ojeh et al, 2020), is well-documented. In addition, the enhanced effectiveness of combination therapy compared with single-agent steroid treatment in managing keloids is also well-recognized (Syed et al, 2013b). Response rates to direct steroid injections in patients with keloid vary from 50 to 100%, with recurrence rates from 9 to 50% (Berman and Bieley, 1995; Kiil, 1977), and approximately 50% of keloids exhibit corticosteroid resistance (Hietanen et al, 2019; Rutkowski et al, 2015).

Despite the significant implications of this study, several limitations are associated with the 3D culture approach. One major challenge is the time-consuming and labor-intensive nature of manually working with spheroids. The synthesis and characterization of all spheroids as well as the capture of images were conducted manually. This bottleneck could be effectively addressed by leveraging recent advancements in automated instruments capable of providing live-cell imaging and noninvasive analysis of cell death, morphology, and function. The incorporation of such technology would enhance the efficiency of the process, leading to more accurate and reproducible data.

Histological analysis of spheroids poses another difficulty, particularly owing to their small size. Although we implemented a protocol modification to monitor the spheroids, the process remains intricate. Our approach has been designed to function in the absence of cryostats, addressing equipment limitations. Another noteworthy challenge in our 3D study is the utilization of primary cells. Inconsistencies in spheroid formations have been observed across different skin biopsy sites. For instance, spheroids derived from earlobe biopsies exhibited behavior distinct from those derived from the neck or other sites. This variability could potentially be mitigated by ensuring the use of nearly identical keloid biopsy sites within a cohort or by employing cell lines instead of primary cells.

The sample size for each experimental condition was informed by previous studies involving primary fibroblast cultures and 3D spheroid models, where 3–4 independent biological replicates (derived from different donors) were typically sufficient to detect meaningful treatment effects with acceptable levels of variability (Kidzeru et al, 2025; Syed et al, 2013a). Despite the inherent logistical challenges posed by the rarity and heterogeneity of patient-derived keloid samples, our findings were consistently replicated across multiple independent donors and supported by technical replicates, reinforcing the robustness of the data.

One limitation of our study concerns the statistical methodology employed. In most experiments, technical replicates were averaged to produce a single value per biological replicate, thereby preserving statistical independence. However, we recognize that alternative approaches, such as mixed-effects models, could have offered a more nuanced analysis by accounting for variability across both technical and biological replicates.

Further investigations are needed to elucidate the distinct mechanisms of action of AMA0825 and DEX. Addressing the challenges mentioned earlier will be essential for refining the experimental methodology and deepening our understanding of these therapeutic agents in the context of fibrotic disease.

This study provides compelling evidence that AMA0825 is a potent antiproliferative and antifibrotic agent against KD. Although AMA0825 did not exhibit antifibrotic activity at the low concentrations, its combination with DEX offers a promising therapeutic strategy by targeting both the hyperproliferation and fibrosis that characterize this challenging condition. AMA0825 in our study demonstrated antiproliferative activity on KD, which may help reduce recurrence rates by targeting the hyperproliferative aspect of KD. On the other hand, DEX, with its antifibrotic effect, could address the fibrotic component of KD. Therefore, their combined use could offer a broad-spectrum antifibroproliferative approach, potentially representing a more effective treatment for KD. Notably, a checkerboard assay revealed potential synergism between AMA0825 and DEX, suggesting a combination therapy targeting both hyperproliferation and fibrosis in KD. We suggest that combining a conventional steroid regimen with the antiproliferative agent AMA0825 could be a beneficial therapeutic strategy.

Further studies are needed to explore the mechanisms underlying the interaction between AMA0825 and DEX, particularly at higher concentrations. Understanding these mechanisms could optimize dosing strategies and reduce potential cytotoxic effects.

In conclusion, we demonstrate that AMA0825 acts as a potent antiproliferative agent against KD. AMA0825 in combination with DEX holds promise as a synergistic treatment for KD, primarily through antiproliferative and antifibrotic effects. These findings contribute to the ongoing development of targeted therapies for this challenging condition.

Materials and Methods

Study samples and donors

Primary normal human skin fibroblasts (NFs) and KFs derived from skin were harvested from donations of excess skin (from keloid and normal skin control from breast reductions or normal abdominal skin tissues) by elective surgery at the Groote Schuur (Cape Town, South Africa), after obtaining due patient consent and with approval from the University of Cape Town Health Research Ethics Committee (HREC reference 374/2023). Four keloid tissue samples and 3 normal skin tissues were acquired during surgical removal treatment for fibroblast culture. The sociodemographic and clinical characteristics of participants are detailed in Table 4.

Table 4.

Summary of Sociodemographic and Clinical Data for Keloid and Normal Skin Primary Cell Donors

Conditions	Sample Code	Study Code	Age, y	Sex	Ethnicity	Skin Biopsy Site	Keloid Age, y	Cause
Keloid	KD44	K1	18	Female	Mixed race	Right earlobe keloid	3	Trauma
	KD45	K2	31	Female	African	Right earlobe keloid	4	Piercing
	KD52	K3	30	Male	African	Left earlobe keloid	9	Piercing
	KD53	K4	23	Female	Mixed race	Right and left earlobe keloid	3	Piercing
	KD21	K5	31	Female	Mixed race	Right and left earlobe keloid	N/A	N/A
	KD22	K6	21	Male	African	Left earlobe keloid	2	N/A
	KD23	K7	25	Male	African	Left earlobe keloid	N/A	N/A
Control	NSF10	N1	17	Female	African	Normal skin from breast reduction	N/A	N/A
	NSF-12	N2	47	Female	Mixed race	Normal skin from the abdomen	N/A	N/A
	NSF-13	N3	16	Female	African	Normal skin from breast reduction	N/A	N/A

Abbreviation: N/A, not available.

Race and ethnicity data were collected to ensure appropriate representation of the local population affected by the condition under investigation. Given that the study was conducted in South Africa, where the condition is disproportionately prevalent in Black African communities (5–10%) (Ogawa, 2022), collecting racial data was necessary to contextualize the findings and assess population-specific relevance. The collection was not mandated by the funder but was included to reflect the demographic characteristics of the study population. Race and ethnicity were self-identified by participants using standardized categories aligned with South African demographic classification norms. Participants selected their racial/ethnic identity during the consent and enrollment process through consent form.

The sample size for each experimental condition was informed by prior studies using primary fibroblast cultures and 3D spheroid models, where 3–4 independent biological replicates (donor-derived cell lines) were typically sufficient to detect meaningful treatment effects with acceptable variability (Kidzeru et al, 2025; Syed et al, 2013a). A formal power analysis was not performed owing to the exploratory nature of this preclinical work and the well-known challenges in obtaining sufficient quantities of primary human keloid tissue. The rarity and variability of patient-derived keloid samples pose significant logistical constraints, which we openly acknowledge as a limitation. Despite this, the consistency of treatment responses across available independent donors and multiple technical replicates supports the robustness of our findings.

Regarding missing data, no imputation was performed. All data included in the final analysis reflect successful and valid experimental replicates. Samples that failed viability, technical, or quality control criteria (eg, spheroid disintegration or imaging artifacts) were excluded and are documented in the raw dataset. Statistical analyses were performed only on complete datasets, and the number of biological and technical replicates is clearly reported in the figure legends and results section.

Primary cell culture

Keloid/normal skin tissues (KFs and NFs), obtained immediately after surgery, underwent aseptic removal of excess adipose tissue, and primary human fibroblasts were isolated using the collagenase digestion method, as described by Amini-Nik et al (2018). The exact procedure followed is thoroughly explained in our previous study (Kidzeru et al, 2025).

Spheroid culture

Primary keloid/normal cells from passages 3–6 were detached from the confluent culture dish. Different cell densities including 5000; 10,000; and 20,000 cells per well were seeded in ultralow attachment 96-well rounded-bottom plates (Corning spheroid microplates, Promega) in complete DMEM supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin. The medium was changed every 2 days and enriched with 10% fetal bovine serum and were cultured submerged at 37 °C in a humidified atmosphere at 5% carbon dioxide.

Synthesis and characterization of spheroid model

Spheroids (KFSs and NFSs) were synthesized using ultralow attachment plates (Corning spheroid microplates, Promega) monitored every day and maintained to day 7. Compactness and volume (3 independent diameters) were measured by capturing the image of them by ZOE (Bio-Rad), and images were analyzed further by FIJI-ImageJ (freeware). Functionality was assessed at days 3 and 7 using calcein/propidium iodide staining.

Fluorescence imaging

Fluorescent images were obtained using the confocal microscopy (Multi-functional Carl Zeiss LSM880 AiryScan) and cell imager ZOE (Bio-Rad) imaging reader with corresponding FIJI-ImageJ (freeware). Fluorescent signal images were obtained for live/dead stained cells using the calcein/propidium iodide in confocal images with ×10 objectives. A Z-stack of each spheroid was taken, and a Z-projection was obtained using focus stacking. The fluorescent signal images were used in cell imager, and CyQuant and calcein/propidium iodide and the volume (μm³) were measured by FIJI-ImageJ (freeware).

The 2D assessment of therapeutic effects by CyQuant assay

The CyQuant assay (Thermo Fisher Scientific) evaluated the impact of AMA0825 (Redx Pharma), DEX (265005, Sigma-Aldrich), triamcinolone (BP339, Sigma-Aldrich), IMA (SML1027, Sigma-Aldrich), and SB525334 (Redx Pharma) on KF and NF proliferation, which evaluated the DNA content, directly proportional to the number of the cells (Jones et al, 2001). Cells were seeded in 96-well plates (8 × 10³ cells/well) and grown to confluence. Then, the medium was removed, fresh medium was added, and the cells were treated with AMA0825 and drugs for 24 hours. After this incubation, the medium was removed, and the instructions of the manufacturer were strictly followed. To evaluate the IC₅₀ of AMA0825, IMA, triamcinolone, DEX, and SB525334 on 3 different KFs (K1, K2, and K3), cells were treated with different concentrations of AMA0825 (10, 20, 25, 30, 35, 40, 50, and 60 nM), IMA, triamcinolone, DEX, and SB525334 with range concentrations of 10, 20, 30, 40, 50, 60, 80, and 100 μM for 24 hours. The fluorescence has been read by Varioskan LUX multimode microplate reader (Thermo Fisher Scientific). The results are expressed as an average percentage growth inhibition for each compound tested in quadruplicate.

RTCA

We employed the label-free xCELLigence RTCA system (ACEA Biosciences) to assess cellular impedance measured as cell index (a measure of cell viability) after KF/NF treatment with AMA0825 and DEX (Syed and Bayat, 2012). Cells (8 × 10³ cells/well) from 3 keloid donors (K1, K2, and K3) and 3 normal donors (N1, N2, and N3) were seeded into 96-well E-plates (ACEA Biosciences) in 100 μl DMEM per well to give a final volume of 180 μl and cultured for 24 hours. Before the experiment, background signals resulting from culture media impedance (80 μl per well) were subtracted. Compounds were added in a 1:10 dilution ratio (ie, 20 μl/well) in 30 and 40 μM for AMA0825 and DEX, respectively, making the total reaction volume 200 μl/well. Control wells (ie, cells alone without treatment) were set up along with the experiments. The cell index was measured at intervals of 1 hour to observe and evaluate alterations in cell growth of the compound and drugs for a duration of 24 hours after treatment.

Immunofluorescence staining and fluorescence microscopy

Cells (K1, K2, and K3) treated with AMA0825 and DEX (30 and 40 μM) for 24 hours were fixed in 3–4% formaldehyde and subsequently in methanol at −20 °C for 10 minutes. They were then washed with PBS and permeabilized with 0.1% Triton X-100 for 5 minutes, followed by another wash with PBS, and subsequently blocked with 1% BSA for 1 hour. Cells were stained with primary antibodies directed against Ki-67 (1:250) (Abcam, SP6 16667), collagen I (1:250) (Abcam, ab34710), and αSMA (1:50) (Abcam, 5694) incubated overnight at 4 °C. A secondary conjugated antibody (Thermo Fisher Scientific, AlexaFluor 488) (1:1000) was added and incubated for 1 hour at room temperature. DAPI (Thermo Fisher Scientific, Life technology, Johannesburg, South Africa, 62247) was added for 10 minutes. Samples were observed, fluorescence images were obtained using the Cell Imager ZOE (Bio-Rad), and the results were read on the spectrometer (Varioskan LUX multimode microplate reader, Thermo Fisher Scientific). The same protocol and dilution ratios were used for collagen I (1:250) and αSMA (1:50) (Abcam, 5694) in the 3D cultures of K1, K2, and K4.

Cell-cycle analysis

To determine the changes on the cell cycle of KFs induced by AMA0825, cells were treated or not with AMA0825 (30 nM) for 24 hours and incubated in the presence of propidium iodide (BD 51-65874X) prior to the analysis by flowcytometry. K1, K2, and K3 cells were transferred to 12-well plates at a concentration of 1.5 × 10⁵ cells/well in a final volume of 1000 μl per well. After 24 hours, after reaching confluence, the cells were treated with AMA0825 at concentrations of 30 nM. After 24 hours, the cells were trypsinized and washed twice with PBS (pH 7.4). The cells were resuspended in 0.5 ml of propidium iodide (100 mg/ml in PBS) and were incubated for 20 minutes in the dark at room temperature. The fluorescence intensity of propidium iodide was analyzed with a FACSymphony (BD FACSymphony A5) flow cytometer where 10,000 events were counted per sample. Histograms and percentages of cells in G1, S, and G2 phases were obtained through FlowJo (version 10.1.0). The graphs were plotted using GraphPad Prism (version 5.0) (GraphPad Software) software.

The 3D treatment time point assessment for spheroids treated with AMA0825

The cultured KFSs from K1 and K4 donors (5 × 10³ cells/well) were treated at day 3 with different concentrations of AMA0825 (0.5; 5; 50; 500; 5000; 50,000 nM). CyQuant kits (the protocol was used as mentioned earlier) were used to analyze cell proliferation. The spheroids were monitored, and images were collected by cell imager (ZOE, Bio-Rad) to find the optimum treatment day of the compounds on spheroid model till day 7. The graphs were plotted using GraphPad Prism (version 5.0) (GraphPad Software) software.

The 3D assessment of therapeutic effects through CyQuant and cell imager and immunofluorescence

The KFSs from K1, K2, K3, and K4 donors (5 × 10³ cells/well) were treated at day 3 with different concentrations of AMA0825 (0.5; 5; 50; 500; 5000; 50,000 nM) and DEX at 80 μM (double amount of IC₅₀, as we are treating the spheroids). Ninety-six hours after treatment (day 7), the KFSs were harvested and subjected to CyQuant staining (CyQuant protocol). The images have been taken up through cell imager (ZOE, Bio-Rad Laboratories), and the results were read on spectrometer (Varioskan LUX multimode microplate reader, Thermo Fisher Scientific). The graphs were plotted using GraphPad Prism (version 5.0) (GraphPad Software) software. The same immunofluorescence protocol used for 2D was applied to 3D spheroid cultures of K1, K2, and K4. Briefly, cells treated with AMA0825 (30 nM) and DEX (40 μM) for 24 hours were fixed in 3–4% formaldehyde and methanol at −20 °C, permeabilized with 0.1% Triton X-100, and blocked with 1% BSA. Primary antibodies against collagen I (1:250, Abcam, ab34710) and αSMA (1:50, Abcam, 5694) were incubated overnight at 4 °C. AlexaFluor 488–conjugated secondary antibodies (1:1000, Thermo Fisher Scientific) and DAPI were applied before imaging with ZOE Cell Imager (Bio-Rad) and analysis through Varioskan LUX reader (Thermo Fisher Scientific).

Masson’s trichrome staining

Spheroid models (KFS) from K1 and K2 donors were fixed, processed, and stained using Masson’s trichrome stain. Untreated and treated spheroids (AMA0825 at 50 nM, DEX at 80 μM) were prestained with 5 μl of hematoxylin before processing to enhance visualization. Spheroids were processed, embedded in paraffin, and sectioned into 5-μm slices for trichrome staining, which was performed according to the manufacturer’s intructions (Newcomer Supply Trichrome Stain, Masson, Light Green Kit, Abcam) to visualize the ECM. For collagen staining, paraffin-embedded spheroid sections were deparaffinized with 3 changes of xylene, 1 minute each, and then hydrated through 2 changes each of 100% and 95% ethyl alcohol, with 10 dips each, followed by a thorough wash with distilled water. The sections were placed in preheated Bouin’s fixative solution for 1 hour and stained with Weigert’s iron hematoxylin solution and Biebrich scarlet-acid fuchsin. The slides were then immersed in phosphomolybdic acid, followed by light green SF Yellowish stain; quickly dipped in acetic acid; and dehydrated through 2 changes each of 95% and 100% ethyl alcohol, with 3 changes of xylene before being cover slipped with mounting medium. The slides were scanned, and images were captured using the Olympus VS120 Digital Slide Scanner (Olympus) at ×40 magnification, resulting in images with a pixel resolution of 0.16 μm/pixel. The sections were analyzed using FIJI-ImageJ software (freeware), following the protocol by Crowe and Yue (2019). Light green regions of interest were selected after color deconvolution and quantified. Statistical analyses were conducted using GraphPad Prism (version 5.0) (GraphPad Software), and the intensity of the green signal between groups was compared using an unpaired 2-tailed Student's t-test.

Ex vivo study

An air-exposed KD organ culture system was adopted with slight modification of a previously published protocol (Syed et al, 2013a). In brief, 4-mm punch biopsy fragments were prepared from keloid tissue (K5, K6, and K7). These were embedded in rat tail collagen gel matrix and cultured at the air–liquid interface in 96-well plate. Serum-free William's E medium, supplemented with 10 μg/ml insulin, 10 ng/ml hydrocortisone, 2 mmol/l l-glutamine, 100 IU/ml penicillin, and 10 μg/ml streptomycin, was used to maintain keloid biopsies for up to 4 weeks. The biopsies were treated by AMA0825 at 1 μM and a concentration of DEX (150 μM) shown to be effective in this model by Syed et al (2013a). Each biopsy was weighed on an analytical balance (Mettler Toledo ME204) on days 0 and 28. Images were taken by cellphone (iPhone 12 mini, zoom 0.5×) under the laminar flow hood. The datasets were analyzed using unpaired t-test to determine the statistically significant differences between the treated and untreated groups (the same group at 2 different days 0 and 28). The differences were considered statistically significant at P < .05. Data are presented as the mean ± SD of 3 donors carried out for 3 technical replicates.

Checkboard assay for activity against KFs using AMA0825 and DEX

KFs from 3 donors (K1, K2, and K3) were seeded in 96-well plates (8 × 10³ cells/well) and grown to confluence and followed by 24-hour treatment with different concentrations of AMA0825 and DEX. The synergism between AMA0825 and DEX was tested using the checkboard assay, combining DEX (0.001, 0.1, 0.5, 1, 50, 100, and 160 μM) with AMA0825 (30 nm, 60 nm, or 4 μM). In parallel, concentration variations of AMA0825 (0.001, 0.1, 0.5, 1, 50, 100, and 160 μM) were made in combination with DEX (40 μM). The interaction between the compounds used was analyzed by the calculation of FIC (Berenbaum, 1978; Horrevorts et al, 1987; Mariani et al, 2021). The IC₅₀ values obtained by the MTT viability curve were used in the calculations. The FIC equation is as follows: FIC = (IC₅₀AB/IC₅₀A) + (IC₅₀BA/IC₅₀B), where IC₅₀AB stands for the treatment with the IC₅₀ concentration of DEX in combination with the IC₅₀ concentration of AMA0825, IC₅₀A stands for the treatment with the IC₅₀ concentration of cisplatin, IC₅₀BA stands for the treatment with AMA0825 plus DEX, and IC₅₀B stands for the treatment with the IC₅₀ concentration of AMA0825. FIC was interpreted as follows: FIC ≤ 1.0 indicates synergism, 1.0 < FIC ≤ 2.0 indicates additive effect, and FIC > 2.0 indicates antagonism.

Statistical analysis

All quantitative data are presented as mean ± SD. For each experimental condition, data were derived from at least 3 independent experiments, each performed using 3–4 biological replicates from different donors, unless stated otherwise. Within each independent experiment, technical replicates were averaged to yield 1 value per biological replicate. Final group comparisons were performed using the mean of these biological replicates to ensure appropriate statistical representation.

The choice of statistical test was based on the experimental design and distribution characteristics: (i) for comparisons between 2 groups, an unpaired Student’s t-test was used, (ii) for experiments involving multiple treatment conditions compared with a single control, 1-way ANOVA followed by Dunnett’s posthoc test was applied, and (iii) for experiments comparing multiple variables across different treatment groups, such as time-course or dose–response studies, a 2-way ANOVA followed by Bonferroni posthoc test was employed to assess interactions and pairwise differences.

All statistical analyses were performed using GraphPad Prism (version 5) (GraphPad Software). A P < .05 was considered statistically significant. To account for experimental variability, repeated measures and donor-specific variability were considered during data interpretation. Although mixed-effects models were not applied, the design ensures biological relevance through the inclusion of multiple donors and independent experiments.

Ethics Statement

All participant recruitment and tissue sampling procedures adhered to the ethical guidelines approved by the University of Cape Town Human Research Ethics Committee (HREC reference 374/2023). Written informed consent was obtained from each participant.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author, AB, upon reasonable request. No large datasets were generated or analyzed for this study.

ORCIDs

Zeinab Ghasemishahrestani: http://orcid.org/0000-0002-2482-4296

Traci A. Wilgus: http://orcid.org/0000-0001-6715-7633

Nonhlanhla P. Khumalo: http://orcid.org/0000-0002-1686-1900

Ardeshir Bayat: http://orcid.org/0000-0002-4116-6491

Conflict of Interest

The authors state no conflict of interest. The drug was a gift from Redx Pharma, and the project was partially funded by Redx Pharma. The authors did not receive any financial compensation or other benefits from Redx Pharma in connection with this study or its publication.

accepted manuscript published online XXX; corrected proof published online XXX