miR-338-3p promotes radiation recall like dermatitis by suppressing pleiotrophin via the PI3K/Akt/Bcl2 pathway
Yu Min, Jingjing Wang, Yahui Feng, Ping Yang, Xiaopeng Xu, Jun Dai, Shuyu Zhang, Xingchen Peng
TL;DR
This study shows that miR-338-3p contributes to radiation recall dermatitis by suppressing a key protein, offering a potential new treatment target.
Contribution
The study identifies miR-338-3p as a novel regulator of radiation recall dermatitis through the PTN/PI3K/Akt/Bcl2 pathway.
Findings
miR-338-3p is significantly upregulated in both rat models and human patients with radiation recall dermatitis.
Inhibiting miR-338-3p or activating PTN reduces the severity of radiation recall-like dermatitis in rats.
miR-338-3p suppresses PTN, leading to reduced PI3K/Akt/Bcl2 signaling and increased cell death.
Abstract
Radiation recall dermatitis (RRD) is a rare but severe inflammatory reaction induced by certain drugs in previously irradiated skin, which can markedly impair quality of life and disrupt cancer treatment. However, the molecular mechanisms underlying RRD remain poorly understood. In this study, a radiation recall–like dermatitis model (RRLD) was established in Sprague–Dawley rats by localized skin irradiation followed by subcutaneous administration of non-toxic concentrations of chemotherapeutic agents, including 5-fluorouracil (5-Fu, 10 µg/µL) or Epirubicin (EPI, 0.05 µg/µL), into the irradiated area. Transcriptomic and miRNA sequencing of rat skin tissues from RRLD rats identified a panel of dysregulated miRNAs, among which miR-338-3p was the most prominently upregulated and was further corroborated by elevated miR-338-3p levels in serum samples from patients with RRD. Dual-luciferase…
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.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7- —Noncommunicable Chronic Diseases-National Science and Technology Major Project
- —Regional Innovation and Development Joint Fund Key Project of the National Natural Science Foundation of China
- —National Natural Sciences Foundation of China
- —Sichuan Provincial Science and Technology Department Key Research and Development Program
- —Sichuan Science and Technology Program
- —Science and Technology Project of Sichuan Provincial Health Commission
- —International Science and Technology Cooperation Program of Chengdu Science and Technology Bureau
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-related skin toxicity · Effects of Radiation Exposure · Bioactive Natural Diterpenoids Research
Introduction
Radiotherapy (RT) is a crucial cancer treatment modality aimed at maximizing tumor cell destruction and synergistic amplification of therapeutic efficacies of other anti-tumor therapies [1–3]. However, RT inevitably causes acute and chronic damage to normal tissues [4]. As the largest organ and primary barrier, the skin is directly affected by radiation, leading to adverse reactions such as pain, redness, dryness, itching, and even necrosis [5, 6]. RT-related skin injury can compromise the therapeutic effects and diminish the quality of life (QoL) in cancer survivors [7].
Notably, a distinct phenomenon termed Radiation recall dermatitis (RRD), apart from classical radiation dermatitis, has been recognized. RRD occurs when a previously irradiated area of the skin becomes inflamed upon certain drug administration [8, 9]. The manifestations of RRD range from mild erythema to severe ulceration and necrosis, depending on the severity [8, 9]. This less common condition was first reported by D'Angio et al. in 1959 [10]. Emerging evidence suggests that RRD is frequently triggered by subsequent anticancer drugs, particularly fluoropyrimidines, anthracyclines, and gemcitabine, following RT [11]. The incidence of RRD varies with different drugs, with nearly 10%−20% of patients treated with Capecitabine and Epirubicin (EPI) after RT developing RRD, respectively [12]. Importantly, RRD appears to occur disproportionately in breast cancer patients compared with other malignancies [8, 13–16]. In a comprehensive literature review comprising 129 reported RRD cases, breast cancer was the most associated primary tumor type, accounting for more than half of all cases, and various systemic agents were implicated as triggers [8]. Although RRD is less prevalent than other acute radiation-induced skin toxicities in routine clinical practice, its unpredictable onset and often prolonged course can substantially compromise the QoL of cancer survivors, particularly in breast cancer patients who commonly receive multimodal therapy [17]. Several hypotheses have been proposed to explain the pathogenesis of RRD, including depletion or functional impairment of irradiated skin stem cells, vascular endothelial injury, localized hypersensitivity and immune-mediated reactions, as well as cumulative DNA damage and oxidative stress in keratinocytes induced by subsequent drug exposure [18–20]. However, most existing evidence is derived from descriptive clinical observations, and molecular-level mechanistic studies remain limited. Consequently, no consensus mechanism has yet been established, which substantially hinders the development of effective preventive or therapeutic strategies for RRD [8, 14, 21–23].
Notably, as members of non-coding RNAs, microRNAs (miRNAs) regulate gene expression by binding to the 3'-UTR of target mRNAs, inhibiting translation, and promoting mRNA degradation [24]. Compelling evidence supports that miRNAs play critical roles in DNA damage repair, cell proliferation, apoptosis, and radiation-related signal transduction, making them potential targets for radiation protection [25–27]. In addition to serving as potential therapeutic targets, miRNAs hold promise as novel diagnostic biomarkers. Their ability to enter the circulatory system compared to other RNAs makes them ideal candidates for clinical applications. miRNA levels can be measured in patient serum samples, highlighting their potential as predictive biomarkers for early disease diagnosis [28–30].
In the present study, we hypothesized that radiation recall–associated skin injury was mediated, at least in part, by dysregulation of specific miRNAs and their downstream target genes, which collectively altered skin-cell survival pathways following radiotherapy and subsequent chemotherapy administration. Accordingly, we established a radiation recall–like dermatitis (RRLD) model using localized skin irradiation in Sprague–Dawley (SD) rats followed by subcutaneous administration of chemotherapeutic agents (5-fluorouracil, 5-Fu, or EPI) after resolution of acute radiation injury. Concurrently, in vitro recall-like injury models were developed using human skin-derived cell lines and primary rat epidermal cells subjected to X-ray irradiation followed by drug challenge. The transcriptomic and miRNA sequencing of skin tissues from RRLD rats and matched controls identified sets of differentially expressed mRNAs and miRNAs. Notably, miR‑338‑3p emerged as the most significantly upregulated miRNA under RRLD conditions, which was validated in serum from RRD patients. Furthermore, the functional assays demonstrated that miR‑338‑3p directly targets the gene encoding Pleiotrophin (PTN), and that modulation of the miR‑338-3p/PTN axis significantly influences skin cell apoptosis and viability, in part via suppression of the PI3K/Akt/Bcl2 signaling pathway. Importantly, in vivo interventions, knockdown of miR‑338‑3p, or overexpression of PTN, markedly ameliorated RRLD‑associated skin damage. Together, these findings demonstrate a potential role for the miR-338-3p/PTN/PI3K/Akt/Bcl2 axis in RRLD and provide a preclinical basis for future mechanistic and translational studies of clinical RRD.
Results
Establishment of the RRLD model in vitro and vivo
To establish reproducible RRLD models in vivo and in vitro, previously irradiated skin or cells were challenged with chemotherapeutic agents following a defined recovery period. In vivo, rats receiving 30 Gy electron beam irradiation to the hindlimb skin exhibited complete or near-complete resolution of acute radiation injury by day 45, as evidenced by macroscopic recovery and histological normalization (Fig. S1). Subsequent administration of EPI or 5-fluorouracil (5-Fu) elicited inflammatory skin responses preferentially confined to previously irradiated sites (Fig. 1a). The irradiated skin in the EPI and 5-Fu groups developed erythema, dry desquamation, and occasional scabbing, with injury scores ranging from 2.5 to 4, whereas no lesions were observed in saline-treated controls (Fig. 1b-d, S2). Histopathological analysis demonstrated marked epidermal thickening, partial hyperkeratosis, neutrophil infiltration, and loss of skin appendages in RRLD lesions (Fig. 1b, S2, S3). Ultrastructural examination further revealed disrupted desmosomal junctions and disorganized intercellular architecture in RRLD lesions, contrasting with preserved epidermal integrity in controls (Fig. 1e). In addition, TUNEL staining revealed an increased number of apoptotic cells in RRLD tissues, reflecting enhanced cellular injury following delayed chemotherapeutic challenge of previously irradiated skin (Fig. S4). Thus, these findings indicate that delayed chemotherapeutic challenge after irradiation can induce recall-like injury reactivation in previously irradiated skin.Fig. 1. Construction and validation of the RRLD model in vivo. a Study protocol for the construction and validation of the RRLD model in vivo in SD rats. b Representative images of rat hindlimb skin and corresponding hematoxylin–eosin (H&E) stained sections following different interventions (the area within the white dashed lines indicates the skin injury site and epidermal thickness and the area within the yellow dashed lines indicates inflammatory cell infiltration), with corresponding (c) skin injury scores (n = 3) and d quantitative measurements of epidermal thickness. Scale bar = 50 µm. e Representative transmission electron microscopy (TEM) images showing ultrastructural changes in different treatment groups. Scale bar = 2 µm. Results are presented as mean ± SD. 0.05 µg/µL EPI (10 µg/µL 5-Fu) was used for the in vivo RRLD model. For comparisons between multiple groups, one-way ANOVA followed by post-hoc multiple-comparison tests was used. ***P < 0.001; ns, not significant
In vitro, RRLD models were established in HaCaT keratinocytes, WS1 dermal fibroblasts, and primary rat epidermal cells following X-ray irradiation and delayed drug exposure. After irradiation, cell viability stabilized at approximately 85–93% during the recovery phase, indicating resolution of acute radiation cytotoxicity (Fig. S5). Drug concentrations inducing minimal baseline toxicity (≈85% viability) were subsequently selected for model construction (Fig. S6). Upon EPI or 5-Fu challenge after irradiation, a pronounced reduction in cell viability was observed across all cell types (Fig. 2a, b). Specifically, viability decreased to 68.30% and 59.54% in HaCaT cells, 71.79% and 69.66% in WS1 cells, and 68.95% and 62.21% in primary rat epidermal cells in the RRLD (EPI) and RRLD (5-Fu) groups, respectively. This was accompanied by sustained elevation of intracellular reactive oxygen species (ROS) (Fig. 2c, S7, S8) and a significant increase in apoptosis rates (Fig. 2d-g, S9). Notably, irradiation alone or drug treatment alone did not significantly alter Bcl2 or Bax expression, whereas combined irradiation followed by chemotherapeutic challenge resulted in marked Bcl2 downregulation and Bax upregulation (Fig. 2h-j). These data indicate that delayed chemotherapeutic exposure after irradiation leads to pronounced oxidative stress and cellular damage, accompanied by increased cell death and disrupted pro-survival signaling, consistent with a radiation recall–like injury phenotype in vivo.Fig. 2. Construction and validation of the RRLD model in vitro. a Study protocol for the construction and validation of the RRLD model in vitro. b Cell viability of different HaCaT and WS1 cells, as well as primary rat epidermal cells treated with EPI or 5‑Fu after irradiation, as assessed by CCK‑8 assay (n = 6). c Representative fluorescence microscopy images of reactive oxygen species (ROS) in HaCaT cells treated with EPI or 5‑Fu after irradiation using DCFH‑DA (n = 3). Scale bar = 100 µm. d Flow cytometry analysis of apoptosis rates in HaCaT cells across different treatment groups and f corresponding quantitative analysis of apoptosis rates (n = 4). e Flow cytometry analysis of apoptosis rates in WS1 cells across different treatment groups and g corresponding quantitative analysis of apoptosis rates (n = 4). h Expression levels of apoptosis‑related proteins in different treatment groups (β-actin was the loading control, n = 3) and corresponding quantification of relative Bax/Bcl2 ratio in (i) HaCaT and j WS1 cells. Results are presented as mean ± SD. HaCaT cells were treated with 0.05 µM EPI (1 µM 5-Fu), and WS1 and primary rat epidermal cells were treated with 0.05 µM EPI (5 µM 5-Fu) for in vitro RRLD models. For comparisons between multiple groups, one-way ANOVA followed by post-hoc multiple-comparison tests was used. ***P < 0.001
Identification of RRLD-associated miRNAs and target genes
To elucidate the molecular mechanisms underlying RRLD, integrated mRNA and miRNA sequencing was performed using skin tissues from SD rats subjected to different treatments (Fig. 3a). Principal component analysis revealed clear transcriptomic segregation between RRLD and non-RRLD groups, with the control (CON) and irradiation-only (IR) samples clustering closely, whereas the RRLD (EPI) and RRLD (5-Fu) samples formed distinct and well-separated clusters (Fig. S10), indicating profound RRLD-specific transcriptional reprogramming. Differential expression analysis showed extensive alterations in both miRNA and mRNA profiles in RRLD skin compared with CON and IR groups (Fig. S11, S12). Venn analyses identified a shared RRLD signature comprising 29 upregulated and 15 downregulated miRNAs (Fig. 3b), together with 407 upregulated and 932 downregulated differentially expressed mRNA common to both chemotherapeutic triggers (Fig. 3c). Heatmap visualization further confirmed the consistent expression patterns of these RRLD-associated miRNAs and mRNA across samples (Fig. 3b, c). To identify key regulatory relationships, an integrated miRNA–mRNA interaction network was constructed based on inverse expression patterns (Figs. S13, S14). This network revealed 27 upregulated miRNAs targeting 279 downregulated genes and 12 downregulated miRNAs targeting 53 upregulated genes, highlighting coordinated post-transcriptional regulation during RRLD occurrence.Fig. 3. Identification of key regulatory miRNAs and target genes involved in RRLD by mRNA and miRNA sequencing. a Schematic overview of the study design, including transcriptome (mRNA) and miRNA sequencing followed by integrated bioinformatic analysis. b Venn diagram showing differentially expressed miRNAs among treatment groups (n = 3). c Venn diagram showing differentially expressed mRNAs among treatment groups (n = 3). d Relative expression levels of 13 candidate miRNAs in rat skin tissues: comparison between IR and RRLD (EPI or 5‑Fu) groups, as measured by qRT‑PCR (IR: n = 3; RRLD: n = 6). e Relative expression levels of the same candidate miRNAs in human peripheral blood serum samples (RT vs. RRD) measured by qRT‑PCR (IR: n = 17; RRD: n = 8). RT: Radiotherapy; RRD: radiation recall dermatitis. f Relative expression of miR‑338‑3p in HaCaT cells under different treatments (n = 3). g Relative expression of miR‑338‑3p in WS1 cells under different treatments (n = 3). h Venn diagram showing that both the human and rat 3′‑UTRs of the candidate genes PTN and TTK harbor predicted binding sites for miR‑338‑3p. i Relative expression levels of PTN and TTK mRNAs in HaCaT cells under different treatments (n = 3). j Relative expression levels of PTN and TTK mRNAs in WS1 cells under different treatments (n = 3). k Serum levels of PTN (pg/mL) in HNSCC RT and RRD groups determined by ELISA assay (n = 3). HNSCC: head and neck squamous cell carcinoma; NPC: Nasopharyngeal carcinoma. Results are presented as mean ± SD. 0.05 µg/µL EPI (10 µg/µL 5-Fu) was used for *the *in vivo RRLD model. HaCaT cells were treated with 0.05 µM EPI (1 µM 5-Fu), and WS1 was treated with 0.05 µM EPI (5 µM 5-Fu) for in vitro RRLD models. For comparisons between two groups, Student’s t-test was used. For comparisons between multiple groups, one-way ANOVA followed by post-hoc multiple-comparison tests was used. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant
To enhance translational relevance, miRNAs within the network were cross-referenced with their human homologs, identifying 13 miRNAs (miR-16-5p, miR-18a-5p, miR-183-5p, miR-200c-5p, miR-20b-5p, miR-22-3p, miR-22-5p, miR-29a-5p, miR-338-3p, miR-34a-5p, miR-34c-5p, miR-484, and let-7b-5p) with identical mature sequences in rats and humans. Serum qRT-PCR analysis in human subjects demonstrated that four miRNAs, miR-200c-5p, miR-20b-5p, miR-22-5p, and miR-338-3p, exhibited expression trends consistent with sequencing results, among which miR-338-3p showed the most pronounced upregulation (P < 0.001) (Fig. 3d, e). Consistently, miR-338-3p expression was unchanged in CON and IR groups but significantly elevated in RRLD cell models, further validating its RRLD-specific induction (Fig. 3f, g). To identify functional downstream targets of miR-338-3p, predicted human target genes were intersected with rat different expressed genes (DEGs) derived from sequencing analysis. Among 46 candidate genes, only PTN and TTK protein kinase possessed conserved miR-338-3p binding sites within their 3′-untranslated regions in both species (Fig. 3h). Subsequent results showed that the mRNA levels of PTN were consistently and significantly downregulated in RRLD models across HaCaT and WS1 cells, whereas TTK exhibited variable and less consistent suppression effects (Fig. 3i, j). Importantly, ELISA analysis demonstrated markedly reduced serum PTN levels in clinical RRD patients compared with radiotherapy-only controls (40.51 ± 9.56 vs. 182.47 ± 71.10 pg/mL, P < 0.001) (Fig. 3k), supporting the clinical relevance of PTN dysregulation. The inclusion and exclusion criteria were summarized in the Supplementary Materials and Methods and Table S5, 6.
The regulatory effect of miR-338-3p on the biological behavior of RRLD
To determine the functional role of miR-338-3p in RRLD pathogenesis, gain- and loss-of-function approaches were applied in keratinocytes and dermal fibroblasts. Efficient overexpression and knockdown of miR-338-3p were achieved in HaCaT and WS1 cells, respectively (Fig. 4a, b). Colony formation assays were used to evaluate the impact of miR-338-3p on cellular radiosensitivity. In HaCaT cells, miR-338-3p overexpression significantly reduced clonogenic survival at irradiation doses of 2, 4, and 6 Gy compared with negative controls (Fig. 4c). Quantitative modeling revealed a decrease in the mean lethal dose (D_0_) from 1.27 to 0.97, corresponding to a sensitizer enhancement ratio (SER) of 1.31. Conversely, the inhibition of miR-338-3p in cells significantly increased colony formation across the same dose range (Fig. 4d), with D_0_ increasing from 1.32 to 1.72 (SER = 0.77). Consistent with these findings, miR-338-3p overexpression was associated with increased LDH release and reduced cell viability following irradiation, whereas miR-338-3p inhibition partially preserved cell viability (Fig. S15). Apoptosis analyses further demonstrated that overexpression of miR-338-3p significantly increased apoptotic rates in radiation-recall-like injury cells (Fig. 4e–h, S16). In contrast, suppression of miR-338-3p markedly reduced apoptosis compared with inhibitor controls (Fig. S17). These results indicate that miR-338-3p promotes radiation-associated cytotoxicity in human skin cells.Fig. 4. Regulatory role of miR-338-3p in RRLD models. a Relative miR-338-3p expression levels after transfecting with miR-338-3p mimics (n = 3). b Relative miR-338-3p expression levels after transfecting with miR-338-3p inhibitor (n = 3). c Effect of miR-338-3p overexpression on radiosensitivity of HaCaT cells evaluated by colony-formation assay (plating efficiency normalized to negative control, n = 3). d Effect of miR-338-3p knockdown on radiosensitivity of HaCaT cells by colony-formation assay (n = 3). e Flow cytometry analysis of apoptosis in the RRLD model of HaCaT cells with miR-338-3p mimics and f corresponding quantitative analysis of apoptosis rates (n = 4). g Flow cytometry analysis of apoptosis in WS1 cells with miR-338-3p mimics and h corresponding quantitative analysis of apoptosis rates (n = 4). i Study protocol for evaluating effects of miR-338-3p knockdown in RRLD rats. j Representative photographs of rat hindlimb skin showing effects of miR-338-3p knockdown by antagomir in the RRLD model and k corresponding quantitative skin injury scores (n = 4). l Representative hematoxylin–eosin (H&E) stained images of rat hindlimb skin illustrating effects of miR-338-3p knockdown on histopathology (Epidermal thickness is indicated by area within the white dashed lines, Scar bar = 100 µm) and m corresponding quantitative measurements of epidermal thickness from H&E sections (n = 4). Results are presented as mean ± SD. 10 µg/µL 5-Fu was used for the in vivo RRLD model. HaCaT cells were treated with 1 µM 5-Fu, and WS1 was treated with 5 µM 5-Fu for in vitro RRLD models. For comparisons between two groups, Student’s t-test was used. **P < 0.01; ***P < 0.001. NC: negative control
To validate these findings in vivo, miR-338-3p was inhibited in RRLD rat models by using the antagomir. Effective knockdown of miR-338-3p in irradiated skin was confirmed by qRT-PCR (Fig. 4i, S18). Following the 5-Fu challenge, rats receiving the miR-338-3p antagomir exhibited markedly attenuated skin injury compared with antagomir controls. Notably, lesion scores in the knockdown group were significantly reduced relative to controls, with milder erythema and desquamation (Fig. 4j, k). Histological examination corroborated these observations, revealing reduced epidermal thickness as well as lower counts of neutrophil infiltration in miR-338-3p–suppressed tissues (Fig. 4l, m, S19). Collectively, both in vitro and in vivo data demonstrate that miR-338-3p acts as a critical regulator of radiation-associated skin injury by enhancing radiosensitivity and apoptosis, and that its inhibition effectively mitigates RRLD severity.
Biological effect of PTN on RRLD in vitro and in vivo
Thus, we further assessed the impact of PTN on the injury of HaCaT cells. Then, the overexpression and knockdown of PTN cell lines were established with the protein level validation (Fig. 5a, b, S20). Overexpression of PTN resulted in a significant increase in the number of cell colony clusters at irradiation doses of 2, 4, and 6 Gy. The D_0_ was 1.25 for the control group and 1.79 for the PTN overexpression group, with the SER of 0.70 (Fig. 5c). Conversely, PTN knockdown significantly reduced the number of cell colony clusters at the same doses, with a D0 of 1.35 for the control group, 0.93 for the PTN knockdown group, and an SER of 1.45 (Fig. 5d). Besides, Ad-PTN overexpression would help reduce the cell injury and cell viability under different irradiation doses (Fig. S21). These findings indicated that PTN overexpression reduced radiosensitivity in HaCaT cells, while PTN knockdown enhanced radiosensitivity and increased cell death. Moreover, overexpression of PTN inhibited apoptosis in RRLD (EPI) and RRLD (5-Fu) models in vitro (Fig. 5e-h, S22), whereas knockdown of PTN would promote apoptosis rates of cells (Fig. S23).Fig. 5. Regulatory role of PTN in RRLD models. a PTN protein expression levels in PTN overexpression or PTN knockdown groups via adenoviral transduction by western blotting (β-actin was the loading control, n = 3) and b corresponding quantification of relative PTN protein expression levels in HaCaT and WS1 cells. c Effect of PTN overexpression (OE) on radiosensitivity of HaCaT cells assessed by colony‑formation assay (n = 3). d Effect of PTN knockdown on radiosensitivity of HaCaT cells assessed by colony‑formation assay (n = 3). e Apoptosis rate of HaCaT cells with PTN OE and f corresponding quantitative analysis of apoptosis rates (n = 4). g Apoptosis rate of WS1 cells with PTN overexpression and h corresponding quantitative analysis of apoptosis rates (n = 4). i Study protocol for evaluating effects of PTN OE in RRLD rats. j Representative photographs of rat hindlimb skin showing the effects of PTN overexpression (Ad‑PTN OE) in the in vivo RRLD model and k corresponding quantitative skin‑injury scores (n = 4). l Representative hematoxylin–eosin (H&E) stained images of rat hindlimb skin following PTN overexpression (Epidermal thickness is indicated by area within the white dashed lines, Scar bar = 100 µm) and m corresponding quantitative measurements of epidermal thickness from H&E sections (n = 4). 10 µg/µL 5-Fu was used for the in vivo RRLD model. HaCaT cells were treated with 1 µM 5-Fu, and WS1 was treated with 5 µM 5-Fu for in vitro RRLD models. Results are presented as mean ± SD. For comparisons between two groups, Student’s t-test was used. **P < 0.01; ***P < 0.001. NC, negative control; Ad, adenovirus; ns, not significant
To further explore the effect of PTN overexpression on skin injury in RRLD rats, we subcutaneously and intradermally injected the PTN overexpression adenovirus (Ad-PTN OE) into the rats (Fig. 5i). The successful overexpression of PTN in the rat skin tissue was confirmed by immunohistochemistry (IHC) staining, which demonstrated an increase in PTN protein levels (Fig. S24). According to the RRLD rat modeling process, 4 days before the subcutaneous injection of 5-Fu, the rats were administered multiple subcutaneous and intradermal injections of Ad-PTN OE. Subsequently, the rats were treated with 5-Fu as per the modeling procedure. Experiments showed that after 7 days of continuous 5-Fu administration, the NC group exhibited significant skin lesions, whereas the Ad-PTN OE group showed only mild erythema and dryness (Fig. 5j). The lesion scores of the PTN OE group were significantly lower than those of the NC group (Fig. 5k). Histological examination of the skin tissues from the lesion sites showed reduced epidermal thickness in the PTN OE group compared to the NC group, which supported that PTN overexpression could mitigate radiation-recall-like skin injury in vivo (Fig. 5l, m, S25).
miR-338-3p negatively regulates the expression of PTN
To define the molecular mechanism through which miR-338-3p contributes to RRLD, bioinformatic analyses were first performed to identify its downstream targets. TargetScan and miRanda consistently predicted conserved miR-338-3p binding sites within the 3′-untranslated region (3′-UTR) of both human and rat pleiotrophin (PTN) mRNAs (Fig. 6a). Direct interaction between miR-338-3p and PTN was confirmed using a dual-luciferase reporter assay. Co-transfection of miR-338-3p mimics with the wild-type PTN 3′-UTR construct resulted in a significant reduction in luciferase activity compared with negative controls, whereas mutation of the miR-338-3p binding site completely abolished this effect (P = 0.912 for PTN-MUT) (Fig. 6b). These data demonstrate that PTN serves as a direct target of miR-338-3p. Consistent with this interaction, modulation of miR-338-3p expression inversely regulated PTN levels in vitro. Overexpression of miR-338-3p significantly downregulated PTN mRNA and protein expression in both HaCaT and WS1 cells, whereas inhibition of miR-338-3p led to a marked upregulation of PTN (Fig. 6c-h). In vivo, immunohistochemical analysis of rat skin tissues showed a significantly reduced PTN immunoreactivity score following miR-338-3p agomir administration, while miR-338-3p antagomir treatment significantly increased PTN expression compared with controls (Fig. 6i-l). These findings suggest that miR-338-3p negatively regulates PTN expression both in vitro and in vivo.Fig. 6miR‑338‑3p participates in the process of RRLD by targeting PTN. a Predicted miR‑338‑3p binding site sequences in the 3′‑UTR of human and rat PTN and Dual‑luciferase reporter assay in WS1 cells co‑transfected with miR‑338‑3p mimics or negative control (NC) and wild‑type (WT) or mutant (MUT) PTN‑3′‑UTR plasmids and b corresponding relative luciferase activity (n = 4). c The mRNA expression levels of PTN in HaCaT and WS1 cells after transfecting with miR‑338‑3p mimics or (d) inhibitors (n = 3). e The protein expression levels of PTN in HaCaT cells after transfecting with miR‑338‑3p mimics and g corresponding relative PTN protein expression (β-actin as the loading control; n = 3). f The protein expression levels of PTN in HaCaT cells after transfecting with miR‑338‑3p inhibitors and h corresponding relative PTN protein expression (β-actin as the loading control; n = 3). i Representative immunohistochemistry (IHC) images of PTN in non-RRLD rat skin tissue from the miRNA agomir NC group and miR-338-3p agomir group, with (k) corresponding immunoreactive scores (blue arrow indicates PTN immunoactivity, n = 4). j Representative IHC images of PTN in non-RRLD rat skin tissue from the miRNA antagomir NC group and miR-338-3p antagomir group, with (l) corresponding immunoreactive scores (blue arrow indicates PTN immunoactivity, n = 4). m Flow cytometry analysis of apoptosis in HaCaT and WS1 cells in the RRLD model under different treatments: miRNA mimics NC + Ad‑NC, miR‑338‑3p mimics + Ad‑NC, miRNA mimics NC + Ad‑PTN OE, and miR‑338‑3p mimics + Ad‑PTN OE and corresponding quantitative analysis of apoptosis rates in (n) HaCaT and (o) WS1 cells (n = 4). Results are presented as mean ± SD. 10 µg/µL 5-Fu was used for the in vivo RRLD model. HaCaT cells were treated with 1 µM 5-Fu, and WS1 was treated with 5 µM 5-Fu for in vitro RRLD models. For comparisons between two groups, Student’s t-test was used. For comparisons between multiple groups, one-way ANOVA followed by post-hoc multiple-comparison tests was used. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001. Ad, adenovirus; NC, negative control; OE, overexpression
To determine whether PTN mediates the anti-apoptotic effects of miR-338-3p in radiation-recall-like injury cells, rescue experiments were conducted. As shown in Fig. 6m-o, miR-338-3p overexpression significantly increased apoptosis compared with controls. However, restoration of PTN expression markedly attenuated miR-338-3p–induced apoptosis. This partial reversal indicates that PTN is a functional downstream effector of miR-338-3p in regulating radiation-recall-like cell death. Additionally, to explore whether miR-338-3p influences skin injury in RRLD rats by targeting PTN, we divided SD rats into four groups randomly: miRNA agomir NC + Ad-NC, miR-338-3p agomir + Ad-NC, miRNA agomir NC + Ad-PTN OE, and miR-338-3p agomir + Ad-PTN OE. As per the previously described animal experimental procedure, miR-338-3p agomir and/or Ad-PTN OE were injected subcutaneously and intradermally in rats with RRLD. The results showed that after continuous 5-Fu administration for 7 days, the miRNA agomir NC + Ad-NC group developed obvious skin lesions, mainly characterized by erythema and dry desquamation, with two rats showing scabbing. In the miR-338-3p agomir + Ad-NC group, the skin lesions worsened, with significant erythema and scabbing, and the skin injury scores ranged from 3.5 to 4.0, which was significantly higher than that of the miRNA agomir NC + Ad-NC group. In the miR-338-3p agomir+Ad-PTN OE group, the skin injury was less severe compared to the miR-338-3p agomir+Ad-NC group (Fig. 7a-c). The results showed that compared to the miRNA agomir NC + Ad-NC group, the epidermal thickness was significantly increased in the miR-338-3p agomir + Ad-NC group. The epidermal thickening was alleviated in the miR-338-3p agomir + Ad-PTN OE group compared to the miR-338-3p agomir + Ad-NC group (Fig. 7d). The Bcl2 IHC staining showed that skin tissue in the miR-338-3p agomir group presented the lowest expression level, whereas miR-338-3p agomir + Ad-PTN OE would help the injury recovery with a higher expression level of Bcl2 (Fig. S26). TUNEL staining revealed a significant increase in apoptotic cells in the irradiated/drug-treated skin compared with controls (Fig. S27). Therefore, these results indicate that PTN is a direct and functional target of miR-338-3p and is significantly associated with apoptosis and skin injury in RRLD models.Fig. 7miR-338-3p targets PTN to induce apoptosis via the PI3K/Akt/Bcl2 signaling pathway. a Representative images of RRLD rat hindlimb skin showing effects of miR‑338‑3p overexpression with or without PTN overexpression in the in vivo RRLD model (n = 4). b Representative hematoxylin–eosin (H&E) images of RRLD rat hindlimb skin illustrating histopathological changes under the same treatments (Epidermal thickness is indicated by area within the white dashed lines, Scale bar = 100 µm) and c corresponding quantitative skin‑injury scores and d epidermal thickness from H&E images (n = 4). e Western blot analysis of PTN protein and proteins in the PI3K/Akt/Bcl2 pathway in RRLD (HaCaT) and RRLD (WS1) transfected with miR‑338‑3p inhibitor or mimics, and f corresponding relative protein expression levels of markers in RRLD HaCaT and WS1 cells (β-actin as the loading control; n = 3). g Scheme of miR-338-3p targeting PTN to induce the apoptosis of RRLD in rats. Results are presented as mean ± SD. 10 µg/µL 5-Fu was used for the in vivo RRLD model. HaCaT cells were treated with 1 µM 5-Fu, and WS1 was treated with 5 µM 5-Fu for in vitro RRLD models. For comparisons between two groups, Student’s t-test was used. For comparisons between multiple groups, one-way ANOVA followed by post-hoc multiple-comparison tests was used. *P < 0.05; **P < 0.01; ***P < 0.001. Ad, adenovirus; NC, negative control; OE, overexpression
miR-338-3p targets PTN to regulate the RDD via PI3K/Akt/Bcl2 pathway
To elucidate the downstream signaling mechanisms underlying the miR-338-3p–PTN axis in RRLD, pathway enrichment analysis was performed. KEGG analysis revealed that RRLD-associated differentially expressed genes were predominantly enriched in multiple signaling pathways related to apoptosis, oxidative stress, metabolism, and cell survival, including necroptosis, oxidative phosphorylation, JAK–STAT, TNF, Wnt, FoxO, and PI3K/Akt pathways (Fig. S28). Among these, the PI3K/Akt pathway was of particular interest given its established interaction with PTN and the central role of Bcl2 as a downstream anti-apoptotic effector.
To investigate whether miR-338-3p regulates the PI3K/Akt/Bcl2 signaling pathway by targeting PTN in RRLD, we used western blotting to assess the protein expression levels of PTN, PI3K, AKT, and Bcl2 in radiation-recall-like skin injury cells with miR-338-3p overexpression or knockdown. The inhibition of miR-338-3p resulted in upregulation of PTN, phosphorylated PI3K (p-PI3K), phosphorylated Akt (p-Akt), and Bcl2 protein levels, with no significant impact on the total protein levels of PI3K and Akt (Fig. 7e, f). Conversely, overexpression of miR-338-3p decreased PTN, p-PI3K, p-Akt, and Bcl2 protein levels without altering the total amounts of PI3K and Akt, indicating that miR-338-3p preferentially affects the activation state of the PI3K/Akt pathway rather than overall protein abundance.
To functionally validate the involvement of the PI3K/Akt/Bcl2 pathway, rescue experiments were further conducted using the Akt activator (SC79) in RRLD HaCaT cells. miR-338-3p overexpression significantly reduced Akt phosphorylation and Bcl2 expression compared with controls (Fig. S29). SC79 treatment alone robustly increased the p-Akt/Akt ratio and Bcl2 levels, consistent with effective pathway activation and an anti-apoptotic response. Importantly, co-treatment with SC79 partially restored Akt phosphorylation and Bcl2 expression in miR-338-3p–overexpressing cells and significantly decreased apoptosis relative to miR-338-3p overexpression alone (Fig. S30). Collectively, these data demonstrate that miR-338-3p promotes apoptosis and RRLD progression by targeting PTN and subsequently suppressing the PI3K/Akt/Bcl2 signaling pathway, establishing a mechanistic link between miRNA dysregulation and radiation recall-like skin injury (Fig. 7g).
Discussion
Although RRD is a relatively rare complication following radiotherapy, its substantial impact on treatment efficacy and QoL has garnered growing attention from oncologists [17, 31]. The lack of mechanistic studies and nonspecific pathological features has made it challenging to understand, diagnose, and treat RRD effectively [11, 12, 32, 33]. Moreover, the nonspecific pathological features of RRD further complicate its diagnosis and treatment [17, 31]. In this study, we established reproducible RRLD models both in vivo and in vitro that reflect features of the clinical radiation-recall phenomenon. Through comprehensive transcriptomic and miRNA analyses, miR-338-3p emerged as one of the miRNAs most prominently associated with recall-like injury responses; it targets PTN and is linked with altered PI3K/Akt/Bcl2 pathway signaling and increased cellular injury in our models. To our knowledge, this is the first study to establish and reveal the potential molecular mechanisms underlying RRLD, offering potential therapeutic targets for the clinical management of RRD. Literature reviews have described an elevated risk of RRD in patients receiving sequential chemotherapy with agents such as 5-Fu or EPI following radiotherapy [11, 34]. Consistently, we administered 5-Fu or EPI subcutaneously to SD rats after the resolution of acute radiation-induced skin injury. Under these conditions, the temporal sequence and spatial confinement of recall-like responses in our model broadly align with defining features of clinical RRD and provide a reproducible platform for further mechanistic investigation.
Clinically, RRD presents as an acute inflammatory reaction confined to previously irradiated skin and is triggered by subsequent systemic agents, with a wide spectrum of manifestations ranging from erythema, pruritus, and desquamation to blistering or necrosis [8, 35]. Meanwhile, limited biopsies from RRD patients have demonstrated nonspecific epidermal changes, keratinocyte necrosis, mixed inflammatory infiltrates, dermal fibrosis, and vascular alterations, without a single pathognomonic signature, which reflects the complexity and variability of this process [8, 36, 37]. Moreover, a recent scoping review of RRD emphasized that current knowledge of RRD is largely derived from case reports and small series, and that systematic histopathological characterization in large patient cohorts remains lacking [9]. As a result, clinical diagnosis and grading of RRD still rely predominantly on macroscopic cutaneous manifestations, such as erythema, dry or moist desquamation, ulceration, bleeding, and necrosis, rather than on standardized microscopic criteria. In the current RRLD rat model, the reappearance of localized skin injury after chemotherapeutic challenge similarly exhibited epidermal disruption and inflammatory infiltration limited to irradiated regions, and the temporal course of pronounced injury followed by progressive recovery recapitulated the typical clinical timeline of RRD onset and resolution. However, intrinsic interspecies differences in skin architecture, thickness, and repair mechanisms between rodents and humans can influence lesion histology and severity, and thus warrant cautious interpretation when extrapolate preclinical findings to human radiation recall pathology. Notably, RRLD-affected rat skin exhibited a marked increase in the proportion of TUNEL-positive cells, reflecting enhanced cellular injury within the local cutaneous microenvironment following chemotherapeutic rechallenge of previously irradiated tissue. Similarly, in vitro experiments demonstrated that cells exposed to prior irradiation, after a recovery interval analogous to the clinically latent phase, displayed elevated intracellular ROS levels, reduced cell viability, and activation of apoptosis-related signaling upon treatment with recall-inducing agents. These observations are consistent with the prevailing hypothesis that irradiation induces persistent alterations in tissue homeostasis, thereby increasing susceptibility to secondary insults through dysregulated oxidative stress responses and impaired cellular recovery capacity [38]. Importantly, the implementation of an appropriate recovery period and the use of sublethal drug concentrations ensured that the observed effects represented radiation recall–like responses, rather than residual radiation injury or direct pharmacological cytotoxicity. Although the in vitro model does not recapitulate immune and vascular components present in vivo, the concordant patterns of cellular injury across both systems provide a complementary framework linking epidermal cell dysfunction with tissue-level pathology in RRLD.
To further explore the potential mechanisms of RRLD, we evaluated the differential expression of several candidate miRNAs that were validated in human samples and exhibited trends concordant with the sequencing data, including miR-200c-5p, miR-20b-5p, miR-22-5p, and miR-338-3p. Although miR-200c-5p and miR-20b-5p showed concordant expression trends, only miR-22-5p and miR-338-3p reached statistical significance in the serum of patients with RRD, with miR-338-3p displaying the most pronounced upregulation. Previous literature has shown that miR-200c-5p and miR-22-5p have been implicated in the regulation of epithelial integrity, inflammatory responses, and cellular stress pathways, suggesting that their dysregulation may contribute to the pathological processes associated with RRLD [39–42]. For instance, aberrant expression of the miR-200c family has been linked to epithelial–mesenchymal transition and epithelial barrier function, whereas miR-22-5p has been implicated in cellular responses to oxidative stress and cytotoxic injury, both of which are highly relevant to tissue susceptibility following radiation and subsequent chemotherapeutic challenge [41, 42]. Among these candidates, miR-338-3p emerged as a key regulator associated with both the initiation and severity of RRLD in vitro and in vivo, and was shown to potentiate the stimulatory effects of chemotherapeutic agents in previously irradiated HaCaT and WS1 cells. Integration of miRNA–mRNA interaction analyses identified PTN as a direct downstream target negatively regulated by miR-338-3p, thereby implicating this axis in the pathogenesis of RRLD. When RRLD cells were transfected with miR-338-3p mimics, both radiosensitivity and apoptosis-related responses were significantly enhanced, whereas transfection with miR-338-3p inhibitors resulted in a corresponding reduction in apoptosis-related signaling. These observations are consistent with previously reported biological functions of miR-338-3p in tumor models [43–46]. Moreover, pharmacological or genetic inhibition of miR-338-3p markedly alleviated skin injury in rats with RRLD, providing preliminary mechanistic evidence supporting miR-338-3p as a potential therapeutic target in RRLD. Of note, PTN, a prototypical secreted growth factor, has been reported to be highly expressed in damaged tissues and to participate in tissue repair and regeneration processes [47, 48]. In the context of radiation recall skin injury, elevated PTN expression exerted a protective effect by attenuating the occurrence and severity of RRLD through modulation of the PI3K/Akt/Bcl2 signaling pathway. Collectively, our results provide a preclinical rationale for exploring circulating miR-338-3p as a potential liquid biopsy–based marker of RRD susceptibility in patients receiving sequential chemotherapy after radiotherapy.
Nevertheless, several limitations of this study should be acknowledged. First, the number of clinical samples was relatively limited, reflecting the low incidence of RRLD, and larger, multi-center cohorts will be required to further assess the clinical relevance and translational value of our findings. Second, whether the longitudinal dynamics of circulating miR-338-3p and its target PTN provide additional diagnostic or predictive value for RRLD remains to be determined. At the mechanistic level, although increased intracellular ROS was observed following chemotherapeutic rechallenge of previously irradiated cells, it remains unclear whether oxidative stress represents a primary driver of recall-like injury or predominantly reflects a downstream consequence of cellular damage. From a modeling perspective, although our model reproduces the temporal pattern of clinical radiation recall by triggering localized skin injury after a clinically silent post-radiotherapy interval, it relies on localized subcutaneous chemotherapy rather than systemic administration and thus does not fully mirror the pharmacokinetic features of clinical chemotherapy. Moreover, whether the associated histopathological alterations in rat skin correspond to those observed in human RRD remains to be clarified through future studies incorporating well-characterized patient biopsy specimens. Finally, the integration of single-cell transcriptomic and proteomic profiling, multicolor flow cytometry, and other multi-omics approaches will be essential to resolve cellular heterogeneity, delineate apoptotic and non-apoptotic injury programs, and characterize the dynamic immune microenvironment underlying recall-like skin injury at single-cell resolution.
In summary, we established a reproducible RRLD rat model, which can be triggered by chemotherapy agents. Aberrant regulation of miR-338-3p and its target gene PTN was consistently observed in RRLD models and was found to be closely associated with the development of recall-like responses. Modulation of miR-338-3p expression, accompanied by restoration of PTN levels, attenuated chemotherapy-induced cellular injury in previously irradiated HaCaT and WS1 cells, at least in part through regulation of the PI3K/Akt/Bcl2 signaling pathway. Collectively, these findings provide novel mechanistic insight into the molecular basis of RRLD and suggest that the miR-338-3p–PTN axis may represent a promising avenue for the development of early risk stratification strategies and potential therapeutic interventions for clinical RRD.
Materials and methods
Cell line
The immortalized human keratinocyte cell line (HaCaT) was purchased from iCell Bioscience Inc. (Shanghai, China). The human skin fibroblast cell line WS1 was obtained from the American Type Culture Collection (ATCC). Primary rat epidermal cells were isolated from neonatal Sprague–Dawley (SD) rats (purchased from Chengdu Dashuo Laboratory Animal Co., Ltd.) and used as a mixed skin-derived cell population rather than a purified keratinocyte population, to better reflect the cellular complexity of skin tissue in vivo. The animals were euthanized by cervical dislocation, and the carcasses were briefly immersed in 75% ethanol for 10 min for disinfection. Under sterile conditions, skin was harvested from the limbs, and subcutaneous tissue was carefully removed. The skin was rinsed 3–5 times in PBS containing 1% penicillin‑streptomycin. The cleaned skin was minced into ~ 1 cm^2^ pieces and transferred into a digestion solution composed of 0.25% trypsin and type‑IV collagenase at a final collagenase concentration of 1 mg/mL. Tissue fragments were incubated at room temperature for 2 h to ensure full enzymatic dissociation. After digestion, the tissue suspension was plated in culture dishes containing complete medium supplemented with 1 mg/mL type‑IV collagenase, and incubated at 37 °C with 5% CO_2_ for 24 h. After the first 24 h, the culture was gently pipetted to detach cells and remaining tissue, and the cell suspension was passed through a 70 µm cell strainer to remove undigested debris. The filtrate was centrifuged at 800 × rpm for 3 min at room temperature, the supernatant discarded, and the pellet resuspended in fresh complete medium. HaCaT cells, WS1 cells, and primary rat epidermal cells were cultured in DMEM high-glucose medium supplemented with 10% fetal bovine serum (FBS) at 37 °C in a humidified incubator with 5% CO_2_. The HaCaT and WS1 cell lines used in this study were authenticated. Meanwhile, all cells used in this study were tested negative for mycoplasma contamination before use.
Animal
Male Sprague–Dawley (SD) rats (5 weeks old, weighing 150 g ± 20 g) were purchased from Chengdu Dossy Experimental Animal Co., Ltd. The rats were kept under conditions of 25 ± 2 °C with a humidity of 45–60%, with a 12-h light/dark cycle, and had free access to food and water. The research protocol was initially approved by the Animal Ethics Committee of West China School of Public Health and West China Fourth Hospital of Sichuan University (No. HXGW-EC-Y2025021).
Radiation-induced skin injury scoring criteria
Radiation‑induced skin injury was scored according to the radiation skin injury scale (Table S1). Normal skin: 1 point; mild erythema and dryness: 1.5 points; moderate erythema and dryness: 2 points; obvious erythema and dry desquamation: 2.5 points; dry desquamation and mild dry crusting: 3 points; dry desquamation, dry scab, mild epidermal scarring: 3.5 points; patchy moist desquamation with moderate patchy ulceration: 4 points; confluent moist desquamation, ulceration, and large, deep crusting: 4.5 points; open wound with full-thickness skin loss: 5 points; skin necrosis: 5.5 points. Scoring was performed by two independent observers (blinded to the experimental group) to reduce subjective bias. In case of discrepancy between the two observers, the final score was determined by consensus or by taking the average of the two values. All assessments were done under the same lighting and observation conditions.
RRLD construction in vivo
Initially, to determine the rational concentrations of chemo drugs, the right thigh and foot of the SD rats were irradiated with a 30 Gy electron beam using a Linac AccStar linear accelerator, with a 1 cm tissue compensation membrane and lead shielding for non-irradiated areas. Skin injury was evaluated every 3 days, and injury scores were recorded and photographed according to radiation skin injury criteria. To determine non-toxic concentrations of chemotherapeutic agents for the RRLD rat model, EPI and 5-Fu were each dissolved in 0.9% NaCl (physiological saline) immediately before use. SD rats received daily subcutaneous injections into the hind limb skin of either EPI (0.05, 0.1, 0.25, or 0.5 µg/µL; injection volume: 100 µL per site per day, corresponding to 5, 10, 25, or 50 µg EPI per injection) or 5-Fu (10, 15, 20, or 25 µg/µL; injection volume: 100 µL per site per day, corresponding to 1,000, 1,500, 2,000, or 2,500 µg 5-Fu per injection) for seven consecutive days (n = 3 per concentration). No skin lesions were observed within 14 days at EPI concentrations of 0.05 and 0.1 µg/µL, whereas 0.25 µg/µL caused erythema and skin dryness, and 0.5 µg/µL resulted in hair loss and crust formation. In the 5-Fu groups, doses of 10 and 15 µg/µL produced no visible skin damage, 20 µg/µL induced mild dryness and erythema in one rat, and 25 µg/µL caused varying degrees of dermatitis in all animals. Therefore, 0.05 µg/µL EPI and 10 µg/µL 5-Fu were selected for subsequent RRLD model induction to exclude chemotherapeutic agent-intrinsic skin toxicity (Table S2, S3). With non-toxic concentrations thus determined, post-irradiation skin injury was monitored until spontaneous recovery reached a score of ≤ 2. Rats were then injected with 0.05 µg/µL EPI or 10 µg/µL 5-Fu on both irradiated and non-irradiated thighs (100 µL per site per day for 7 consecutive days, n = 8) to induce the RRLD in vivo. Injection of non-irradiated skin served as an internal control to verify the recall-specific nature of the skin response. The occurrence and severity of RRLD were evaluated using skin injury scoring, H&E staining, TEM, and immunofluorescence analysis.
RRLD construction in vitro
The RRLD model was established in vitro by irradiating HaCaT, WS1, and primary rat epidermal cells with X-rays, followed by longitudinal assessment of cell viability and cytotoxicity using CCK-8 and LDH assays from 1 to 14 days post-irradiation. To determine optimal concentrations of EPI and 5-Fu for RRLD induction while minimizing direct cytotoxic effects, cells were treated with a range of drug concentrations, and cell viability was measured by CCK-8 assay at 48 h after drug exposure. to ensure sufficient cell survival while allowing recall-like sensitization to be detected, rather than reflecting acute drug toxicity. As shown in Fig. S6, treatment with 0.05 μM EPI resulted in nearly 85% survival across all three types of cells; 1 μM 5-Fu achieved nearly 85% viability in HaCaT cells, and 5 μM 5-Fu resulted in nearly 85% viability in WS1 and primary rat epidermal cells. Accordingly, 0.05 μM EPI combined with 1 μM 5-Fu was used to induce the RRLD (HaCaT) model, and 0.05 μM EPI (5 μM 5-Fu) was used to induce the RRLD (WS1) and RRLD (primary rat epidermal cells) models. On day 8 post-irradiation, cells were exposed to the selected concentrations of EPI and 5-Fu for 48 h to mimic the drug stimulation effects in vivo. Successful establishment of the RRLD model in vitro was measured by CCK-8 cell viability assays, ROS fluorescence staining, apoptosis rate in flow cytometry, and protein level.
Statistical analysis
Data derived from this research were analyzed by using GraphPad Prism (version 8.0.0, San Diego, USA). Independent sample t-tests were used to compare the differences between the two groups, and one‑way ANOVA followed by post‑hoc tests was performed for comparisons among three or more groups. Pearson correlation analysis was conducted to assess the RNA expression correlations. A two-tailed P-value < 0.05 indicated statistical significance. ns: no significant; : P* < 0.05; : P < 0.01; ***: P < 0.001.
Supplementary Information
Supplementary Material 1.
The reference list from the paper itself. Each links out to its DOI / PubMed record.