The study protocol received ethical approval from the Ethics Committee of Traditional Chinese Medicine Hospital of Xinjiang Uygur Autonomous Region (Approval No: 2025XE-GS231) and was conducted in accordance with the principles outlined in the Declaration of Helsinki. All participants provided written informed consent before their enrollment in the study, after which peripheral blood samples were drawn. Participants were then categorized into three groups: healthy individuals (Control group, n=6), patients diagnosed with IPF presenting with Qi deficiency and blood stasis syndrome (Patient group, n=6), and IPF patients with the same TCM syndrome who had completed a course of treatment with YZTLF (YZTLF group, n=6). YZTLF composition was shown in Table 1. Diagnosis of IPF followed the official ATS/ERS/JRS/ALAT clinical practice guidelines. TCM syndrome differentiation (Qi deficiency and blood stasis) was confirmed by two senior TCM clinicians in accordance with the Guiding Principles for Clinical Research of New Chinese Medicines, based on characteristic symptoms including fatigue, shortness of breath, spontaneous sweating, sallow complexion, and a cyanotic tongue with ecchymosis. Patients were enrolled based on the following inclusion criteria: (1) age ≥ 40 years; (2) a confirmed or highly suspected diagnosis of IPF according to ATS/ERS/JRS/ALAT criteria within the preceding 5 years, validated by high-resolution computed tomography (HRCT) or lung biopsy within 12 months prior to enrollment; (3) pulmonary function characterized by a forced vital capacity (FVC) ≥ 50% of the predicted value and a diffusing capacity for carbon monoxide (DLCO) between 30% and 79% of the predicted value; and (4) concordance with the TCM diagnostic criteria for Qi deficiency and blood stasis syndrome. Patients were excluded if they presented with severe hepatic, renal, or cardiac dysfunction (NYHA class III–IV), malignancies, or coagulation disorders. Additionally, individuals currently using immunosuppressants or glucocorticoids, or those participating in other clinical trials, were excluded from the study.

Samples of whole blood were drawn into vacuum EDTA tubes and subsequently processed within a two-hour window. For serum isolation, the samples were incubated at ambient temperature for 30 minutes to permit coagulation, after which they were centrifuged for 15 minutes at 3, 000×g and 4 °C. The resulting serum supernatant was then carefully pipetted into sterile cryovials and promptly frozen at -80 °C to await subsequent analysis.

The Chinese herbal medicines were placed in a clay pot and soaked in tap water for 20 minutes. The mixture was brought to a boil and simmered for 30 minutes, after which the liquid was filtered. For the second decoction, water was added to the residue, boiled, and simmered for 25 minutes before filtering. The two filtrates were combined to obtain the final decoction for clinical trials. For animal experiments and in vitro, the decoction was concentrated and adjusted to a standard physiological pH (7.2–7.4), and filtered through a 0.22 μm sterile membrane to obtain the stock solution. Subsequent CCK-8 assays confirmed that the working concentration maintained cellular viability, ruling out non-specific cytotoxicity caused by physicochemical alterations.

Human lung fibroblast MRC-5 cells were grown in Dulbecco’s Modified Eagle Medium (DMEM) that contained 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. These cultures were kept in or maintained at 37 °C in a 5% CO₂ atmosphere. When the cell density reached 80-90%, they were passaged with a 0.25% trypsin-EDTA solution.

To establish an in vitro model of IPF, fibrotic activation was induced by treating the MRC-5 cells with 5 ng/mL of recombinant human TGF-β1 (R&D Systems) for a duration of 48 hours (23) experimental groups were designated as follows: Normal control (NC, untreated MRC-5 cells), Model (TGF-β1-induced cells), si-NC (siRNA negative control), siHIF-1α (Model + 50 nM HIF-1α-targeting siRNA), YZTLF (Model + 200 mg/L YZTLF), and Combination (Model + YZTLF + si-HIF-1α). YZTLF decoction, sourced from the Traditional Chinese Medicine Hospital of Xinjiang Uygur Autonomous Region, was first centrifuged for ten minutes at 12, 000 ×g. For sterilization, the resulting liquid was subsequently passed through a 0.22-μm filter. This sterile supernatant was then brought to its final working concentration of 1, 000 mg/L (total extract) by dilution with DMEM before being administered to the cells for 24 hours. The transfection of siRNA into the cells was accomplished with the Lipofectamine 3000 reagent (Thermo Fisher Scientific).

For an investigation into LSH-mediated regulation of SCD1 by HIF-1α, several experimental cohorts were prepared. These included a Model cohort, si-NC (Model + si-NC), si-HIF-1α (Model + HIF-1α siRNA), oe-NC (Model + empty vector), oe-LSH (Model + LSH overexpression plasmid), and a combination of si-HIF-1α + oe-LSH (Model + HIF-1α siRNA + LSH plasmid). Each well received 2 µg of plasmid DNA for transfection. All experimental conditions were conducted in triplicate, and for subsequent analysis, cellular materials were collected 48 hours after the treatments.

All experimental procedures involving animals were conducted in compliance with the guidelines of Regulations on the Administration of Laboratory Animals (State Council of China), and received approval from Animal Ethics Committee of Jiangxi Hengqing Testing Technology Co., Ltd. (Approval No: JXHQ2025005). Male Sprague-Dawley rats, weighing between 200–250 g, were used for this study. The animals were housed in a controlled environment under a 12-hour light/dark cycle with a stable room temperature of 22 ± 1 °C and a relative humidity of 50 ± 10%. Standard rodent chow and water were available ad libitum for the duration of the study.

To induce a model of pulmonary fibrosis, isoflurane was used to anesthetize the rats, which were then placed in a supine position. The neck area was shaved and sterilized with ethanol before a midline cervical incision was performed to reveal the trachea. A 1-mL syringe was utilized to slowly administer a 5 mg/kg bleomycin solution directly into the trachea, with the needle inserted between the cartilage rings at a near-parallel angle. Post-injection, the wound was sutured, and rats were held upright for 1–3 minutes while gently rotating and massaging the thorax to ensure homogeneous distribution of bleomycin. Animals were monitored postoperatively for recovery and maintained for 28 days before tissue and blood collection.

The rats were randomly assigned to the following experimental groups (n=6 per group), including Control (healthy rats, no treatment), Model (bleomycin-induced IPF), YZTLF-L (Model + low-dose YZTLF, 5 g/kg/day), YZTLF-M (Model + medium-dose YZTLF, 10 g/kg/day), YZTLF-H (Model + high-dose YZTLF, 20 g/kg/day), and Model + nintedanib (0.05 g/kg/day). YZTLF, prepared as a decoction using traditional methods, was filtered (0.22-μm pore) and diluted in distilled water for daily oral gavage (5 mL/kg). Nintedanib (MedChemExpress) was suspended in 0.5% carboxymethyl cellulose and administered via the same route. Treatments commenced 24 hours post-bleomycin instillation and continued for 28 days. Post-experiment, lung tissues and serum were collected for histopathological and biochemical analyzes.

The dosage was calculated based on the body surface area normalization method, using the standard conversion factor of 6.3 between humans and rats (Rat dosage = Human dosage × 6.3), to ensure clinical relevance.

ChIP was performed to assess the binding of HIF-1α and LSH to specific DNA regions. Initially, MRC-5 cells were cross-linked through a 10-minute incubation with 1% formaldehyde at ambient temperature; this process was quenched by adding 0.125 M glycine. After a PBS wash, the cells underwent lysis using a designated ChIP lysis buffer. The resulting chromatin was then fragmented into 200–1000 base pair segments via sonication. The immunoprecipitation step was performed by incubating the fragmented chromatin overnight at 4 °C. One of the following specific primary antibodies was used for each reaction: Anti-HIF-1α (Proteintech, 20960-1-AP), Anti-LSH (Santa Cruz Biotechnology, c-365815X), or a Normal rabbit IgG (CST, 2729) negative control. The immune complexes were subsequently isolated from the solution using Protein A/G magnetic beads. Following washes and cross-link reversal, DNA was purified with a commercial kit (QIAGEN, 28104). Analysis of the purified DNA was conducted via electrophoresis on a 1.5% agarose gel, with the bands being documented using a gel imaging system (Bio-Rad ChemiDoc MP). Binding was confirmed by the enrichment of specific DNA bands in the experimental samples relative to the IgG control.

To assess the regulatory effect of LSH on the activity of the SCD1 promoter, specific reporter constructs were developed. Fragments of the SCD1 promoter (–1500 to +100 bp) for both wild-type (WT) and mutant (MT) versions were inserted into the pGL3-Basic vector (Promega). The mutant construct was created using site-directed mutagenesis, which was designed to interrupt the putative binding sites for HIF-1α/LSH.

MRC-5 cells were co-transfected with a combination of plasmids: 500 ng of either the WT- or MT-pGL3 reporter, 50 ng of the pRL-TK control plasmid, and 1 µg of either an LSH overexpression plasmid or an empty vector. The transfection procedure was carried out with Lipofectamine 3000 (Thermo Fisher). After 48 hours, the resulting reporter activity was measured using the Dual-Luciferase Reporter Assay System (Promega). To control for variations in transfection efficiency across samples, the signal from firefly luciferase was adjusted by normalizing it to the corresponding Renilla luciferase signal.

The viability of MRC-5 cells was assessed using a CCK-8 assay. Initially, cells were plated in 96-well plates at a concentration of 5 × 10⁴ per well and grown to their logarithmic phase. YZTLF drug was sterilized by filtration through a 0.22-µm membrane to remove any particulates. A dilution series of the drug was then created, yielding concentrations of 0, 10, 50, 100, 200, and 400 mg/L. Cultures were treated with this range of concentrations for 24 hours. After the treatment period, 10 µL of CCK-8 reagent was introduced to each well, and the plates were subjected to a final two-hour incubation at 37 °C. Viability was subsequently quantified by reading the absorbance at 450 nm on a microplate reader.

The Iron Assay Kit (Colorimetric, Abcam, ab83366) was used to measure Fe²⁺ levels. For sample preparation, particulate matter was first pelleted from serum and cell supernatants via centrifugation (3000 × g for 15 minutes at 4 °C). To reduce matrix effects, samples were subsequently diluted 1:10 with the kit’s buffer. The final iron concentration was calculated based on the absorbance reading at 593 nm, measured on a BioTek Synergy H1 microplate reader, and compared against a standard curve of iron standards (0–100 µM).

The concentration of GSH was measured with a Glutathione Assay Kit (Sigma-Aldrich, MAK259) based on a kinetic enzymatic recycling method. To begin, samples were deproteinized using a 1:1 volume of 5% sulfosalicylic acid and then centrifuged (10, 000 × g for 10 minutes) to clarify. The resulting supernatant was combined with 5, 5′-dithiobis (2-nitrobenzoic acid) (DTNB) and glutathione reductase. The reaction progress was tracked by monitoring the change in absorbance at 412 nm. Final GSH values were determined by comparison to a standard curve (0–20 µM) and then adjusted based on the total protein content of each sample, which was measured independently with a BCA assay (Thermo Fisher, 23225).

To assess lipid peroxidation, malondialdehyde (MDA) levels were quantified with a TBARS Assay Kit (Cayman Chemical, 10009055). In the procedure, samples were combined with thiobarbituric acid (TBA) and incubated at 95 °C for 60 minutes to facilitate the formation of the MDA-TBA adduct. After the mixture cooled, it was centrifuged for 10 minutes at 1600 × g. The absorbance of the resulting supernatant was then measured at 532 nm. Quantification was achieved by comparing the results to a standard curve generated from 1, 1, 3, 3-tetramethoxypropane (0–50 µM).

A fluorescence-based flow cytometry assay was employed to measure intracellular ROS. For this purpose, cells were labeled with 10 µM of the ROS-sensitive fluorescent probe DCFH-DA (Beyotime, S0033). This labeling step was carried out in the dark for 30 minutes at 37 °C using serum-free medium. After the incubation, cells were rinsed two times with phosphate-buffered saline (PBS; HyClone, SH30256.01) and resuspended in warm PBS. Analysis was promptly performed on a BD FACSCanto II flow cytometer (BD Biosciences). The fluorescent signal was detected with an excitation of 488 nm and emission of 525 nm. FlowJo software (v10.8.1) was used for data processing, with gating based on an unstained negative control.

The extent of cellular apoptosis was determined with an Annexin V-FITC/PI Apoptosis Detection Kit (BD Biosciences, 556547). To prepare samples, 1 × 10⁶ cells were collected via mild trypsinization (Thermo Fisher, 25200072), rinsed twice in cold PBS, and resuspended in 100 µL of binding buffer. Cells were then co-stained with 5 µL of Annexin V-FITC and 5 µL of propidium iodide (PI). This labeling was allowed to incubate for 15 minutes at ambient temperature while shielded from light. Before cytometric analysis, each sample was diluted with an additional 400 µL of binding buffer. Data were collected within an hour of staining using the FL1 (530/30 nm) and FL3 (>670 nm) detection channels.

First, cells were lysed on ice using RIPA buffer (Beyotime, ST506) supplemented with 1 mM PMSF to inhibit protease activity. The cell lysates were then clarified by centrifugation at 12, 000 × g for 15 minutes at 4 °C. The total protein content of the resulting clarified supernatant was then precisely determined using a BCA Protein Assay Kit (Thermo Fisher Scientific, 23225). For each sample, 20–40 µg of protein was separated on 8–12% SDS-PAGE gels and then transferred onto PVDF membranes (Millipore, IPFL00010). The membranes were then treated with a blocking solution (5% non-fat milk in TBST) for one hour at ambient temperature to minimize non-specific signals. After blocking, the membranes were probed overnight at 4 °C with one of these primary antibodies: HIF-1α (1:600; Proteintech, 20960-1-AP), LSH (1:800; CST, sc-365815X), SCD1 (1:10000; Proteintech, 10627-1-AP), ACSL4 (1:40000; Proteintech, 22286-1-AP), GPX4 (1:3000; CST, 52455), α-SMA (1:1000; CST, 19245), Collagen I (1:800; CST, 84336), or β-actin (1:50000; Proteintech, 66009-1-Ig). Following a series of washes, the membranes were treated for one hour at ambient temperature with a suitable HRP-conjugated secondary antibody (1:10000; Proteintech, SA00002-2). Protein bands were visualized using an enhanced chemiluminescence (ECL) substrate (Thermo Fisher Scientific, 34580) and imaged using the Tanon-5200 Multi-Automatic Chemiluminescence Imaging Analysis System (Shanghai Tanon Technology Co., Ltd.). Densitometric analysis was performed using ImageJ software (NIH, USA), and the band intensity of target proteins was normalized to that of β-actin.

For qRT-PCR analysis, TRIzol Reagent (Invitrogen, 15596026) was utilized for the initial extraction of total RNA from cell or tissue samples. A NanoDrop spectrophotometer (Thermo Fisher) was then employed to assess the purity and concentration of the isolated RNA. Following this, first-strand complementary DNA (cDNA) was synthesized from 1 µg of the total RNA template using the PrimeScript RT Reagent Kit (Takara, RR037A). Quantitative PCR was conducted in triplicate on a QuantStudio 5 Real-Time PCR System (Applied Biosystems) with SYBR Green Master Mix (Thermo Fisher, A25742). Thermocycling was performed with the following profile: an initial denaturation step at 95 °C for 10 minutes, followed by 40 cycles alternating between 95 °C for 15 seconds and 60 °C for 1 minute. Primers were designed using the Primer-BLAST tool (NCBI) and were experimentally validated for specificity (Table 2); this was additionally confirmed via melt curve analysis after every run. The relative mRNA expression levels were calculated using the 2^-ΔΔCt method, where ΔCt = Ct_target - C_t, _GAPDH and ΔΔCt = ΔCt_sample - ΔCt_control. GAPDH was used as the stable internal reference gene for normalization. Full details of the specific primer sequences used in this study are listed in Table 2.

For the H&E staining protocol, paraffin-embedded tissue blocks were initially sectioned to a thickness of 4–5 µm. The resulting tissue sections were then prepared for staining by first undergoing deparaffinization with xylene. Following this step, the sections were rehydrated by passing them through a descending ethanol gradient, which concluded with a final rinse in distilled water. The staining protocol began by submerging the slides in Mayer’s hematoxylin for an eight-minute period, followed by a rinse with tap water. A one-minute counterstain with 1% eosin Y was then applied. After staining, the sections were processed through a dehydration series, cleared, and mounted. A light microscope was used to examine the prepared slides to evaluate general histopathological changes. Histological evaluation and scoring were performed by two independent investigators who were blinded to the experimental group allocations.

To evaluate tissue fibrosis, tissue was first sectioned at 4–5 µm. The resulting sections were then mounted on glass slides, cleared of paraffin, and rehydrated for staining. The procedure consisted of three main steps: a 10-min immersion in Weigert’s iron hematoxylin, followed by a 10-min immersion in Biebrich scarlet-acid fuchsin, and concluding with a 10-minute differentiation step using phosphomolybdic-phosphotungstic acid. Following this, a five-minute counterstain with aniline blue was applied. After a brief wash in 1% acetic acid, the samples were processed through a dehydration series, cleared in xylene, and finally mounted with a coverslip for analysis. This staining technique allows for the histological assessment of fibrosis by coloring collagen blue, muscle/cytoplasm red, and nuclei black.

For statistical analysis, all in vitro experiments were performed with at least three independent biological replicates (n = 3). For clinical and in vivo experiments, the total sample size was six (n = 6) per group. Specifically, Western blot analyses were performed using three randomly selected biological replicates (n = 3). Data are presented as the mean ± standard deviation (SD). Prior to parametric analysis, data distribution normality was verified using the Shapiro-Wilk test, and homogeneity of variance was assessed using Levene’s test. To determine the statistical significance of differences between experimental groups, a one-way analysis of variance (ANOVA) was used, with Tukey’s post-hoc test applied for multiple comparisons. A result was deemed statistically significant if the p-value was less than 0.05. All graphs and charts were created with GraphPad Prism software (Version 5.0, USA).

To investigate the regulatory mechanism of the HIF-1α/LSH/SCD1 axis, we first measured the protein and mRNA expression of HIF-1α, LSH, and SCD1 in blood samples from patients and healthy controls. Compared to the healthy control group, the expression levels of both HIF-1α and LSH protein and mRNA were significantly elevated in patients. Following treatment with YZTLF, their expression was markedly reduced. In contrast, the expression of SCD1 protein and mRNA exhibited the opposite trend (Figures 1A–G). Furthermore, the ferroptosis-related markers ACSL4 and Fe²⁺ were significantly elevated in patients. YZTLF treatment reduced the levels of both ACSL4 and Fe²⁺, indicating a decrease in the overall level of ferroptosis (Figures 1H, I).

To elucidate the regulatory relationships among HIF-1α, LSH, and SCD1, ChIP assays were performed to assess HIF-1α-mediated DNA methylation at the LSH locus, followed by dual-luciferase reporter assays to evaluate LSH-dependent SCD1 regulation. ChIP results revealed enrichment of specific DNA fragments encompassing the target region (LSH) in samples immunoprecipitated with anti-HIF-1α and anti-LSH antibodies, while negative (NC) and positive (PC) controls displayed expected patterns, confirming direct binding of HIF-1α to the LSH promoter region (Figure 2A). Dual-luciferase assays demonstrated significant Fluc/Rluc ratio elevation upon LSH overexpression in wild-type constructs, with no effect observed in mutant variants, indicating LSH-specific regulation of SCD1 transcription (Figure 2B). These findings establish the functional interplay among HIF-1α, LSH, and SCD1, with detailed mechanistic validation provided in subsequent experiments.

To investigate whether YZTLF suppresses ferroptosis via the regulation of the HIF-1α-mediated LSH/SCD1 axis, we initially determined the optimal dosage using CCK-8 assay. We identified 200 mg/L as the concentration yielding maximal efficacy without cytotoxicity (Figure 3A). Subsequently, TGF-β-induced IPF model was established in MRC-5 cells, accompanied by HIF-1α silence. The data revealed that TGF-β induction significantly upregulated the protein and mRNA levels of HIF-1α and LSH, while concurrently downregulating SCD1 expression. Notably, both HIF-1α silencing and YZTLF treatment significantly reversed these alterations, and the combined intervention demonstrated a synergistic effect (Figures 3B–H). To further elucidate the impact of HIF-1α and YZTLF on ferroptosis, we assessed intracellular ferrous iron levels alongside the expression of ferroptosis-related proteins ACSL4 and GPX4. Compared to the model group, both HIF-1α knockdown and YZTLF treatment attenuated ferrous iron accumulation and reduced ACSL4 levels, with the combined treatment yielding better efficacy. Conversely, GPX4, an antioxidant enzyme against ferroptosis, exhibited an inverse trend (Figures 3I–K).

Given that oxidative stress is a critical driver of ferroptosis, we further evaluated ROS levels and the oxidative stress markers GSH and MDA. Our findings indicate that HIF-1α silencing and YZTLF treatment significantly decreased ROS and MDA levels in the IPF model while increasing GSH levels. Furthermore, the combined therapy exerted a synergistic antioxidant effect (Figures 3L–O). Furthermore, fibroblast proliferation, apoptosis, and fibrosis levels were examined. The results demonstrated that both HIF-1α silencing and YZTLF treatment exhibited pro-proliferative and anti-apoptotic effects, with the combined treatment showing enhanced efficacy (Figures 3P–R). Regarding fibrosis, both interventions displayed anti-fibrotic effects; however, YZTLF treatment demonstrated superior efficacy. This was evidenced by the expression levels of α-SMA and Collagen I in fibroblasts (Figures 3S, T).

To further elucidate the interplay among LSH, HIF-1α, and SCD1 in IPF pathogenesis, we conducted functional studies using HIF-1α knockdown and LSH overexpression. Western blot and qPCR analyzes revealed markedly elevated HIF-1α levels in the model, model+si-NC, and model+oe-NC groups. HIF-1α knockdown significantly reduced its expression, while LSH overexpression also increased HIF-1α levels, likely due to ROS accumulation from exacerbated ferroptosis activating HIF-1α stabilization (24, 25). Combined HIF-1α silencing and LSH overexpression partially reversed this effect and the HIF-1α level was higher than alone HIF-1α silencing, confirming a regulatory feedback loop (Figures 4A, B).

LSH protein and mRNA levels were substantially upregulated in the model group, with further elevation observed in the model+oe-LSH group. HIF-1α knockdown significantly suppressed LSH expression, demonstrating HIF-1α-dependent positive regulation of LSH (Figures 4C, F). Conversely, SCD1 levels were reduced in the model group but increased upon HIF-1α silencing. LSH overexpression suppressed SCD1 expression, while dual intervention (siHIF-1α + oe-LSH) restored SCD1 levels, confirming LSH-mediated negative regulation of SCD1 and HIF-1α regulates SCD1 through LSH (Figures 4D, G).

Iron ion levels, a direct indicator of ferroptosis, were elevated in the model group. HIF-1α knockdown reduced iron accumulation, whereas LSH overexpression exacerbated it, even counteracting the iron-lowering effect of siHIF-1α (Figure 5A). Consistently, ACSL4 expression mirrored iron ion trends, while GPX4 showed an inverse pattern, validating HIF-1α/LSH/SCD1 axis-mediated ferroptosis regulation (Figures 5B-D). Apoptosis assays revealed that HIF-1α silencing attenuated TGF-β-induced apoptosis in MRC-5 cells, whereas LSH overexpression amplified apoptotic rates. Notably, LSH overexpression reversed the anti-apoptotic effect of HIF-1α knockdown (Figures 5E, F). Cell viability assays corroborated these findings, with HIF-1α inhibition enhancing survival and LSH overexpression diminishing cellular activity (Figure 5G).

Ferroptosis induces ROS-mediated lipid peroxidation, impairing autophagosome formation and recruitment of downstream proteins. HIF-1α silencing markedly decreased ROS levels, suggesting suppression of the HIF-1α/LSH/SCD1 pathway mitigates ROS accumulation during ferroptosis. Compared to the NC group, LSH overexpression elevated ROS levels, while HIF-1α knockdown attenuated this increase (Figures 6A, D). Oxidative stress markers also mirrored these trends, with MDA levels paralleling ROS elevation and GSH depletion, confirming their consistent regulation by the HIF-1α/LSH/SCD1 pathway (Figures 6B, C).

An intratracheal bleomycin-induced IPF rat model was established to evaluate YZTLF and nintedanib (50 mg/kg). Through 28 days, the body weight of rats was monitored as a general indicator of systemic health. The results showed that bleomycin induction caused significant weight loss compared to the control group. However, treatment with YZTLF and nintedanib effectively mitigated this weight reduction in a dose-dependent manner, suggesting a protective effect against systemic toxicity (Supplementary Figure S1). Nintedanib, a first-line tyrosine kinase inhibitor for IPF associated with diarrhea (discontinued in 5% of patients), served as the positive control (26). Bleomycin significantly increased HIF-1α and LSH protein/mRNA levels while reducing SCD1. YZTLF and nintedanib reversed these alterations, with high and medium dose groups showing superior therapeutic effects compared to the low-dose group. However, the high-dose group did not show a significantly greater effect than the medium-dose group, likely due to optimal concentration saturation (Figures 7A–G).

Bleomycin increased iron ion levels and ACSL4 expression while suppressing GPX4. Treatment with YZTLF and nintedanib attenuated these changes, reducing ferroptosis. Medium and high dose groups exhibited comparable efficacy, both outperforming the low-dose group, though nintedanib demonstrated stronger effects (Figures 7H–J). The levels of ROS and MDA in the lung tissue of rats treated with bleomycin increased significantly, while the level of GSH decreased significantly. The formula also alleviated pulmonary oxidative stress, normalizing elevated ROS and MDA levels and restoring GSH in bleomycin-treated rats (Figures 7K–M).

HE staining revealed collapsed alveolar structures, thickened septa, and inflammatory infiltration in the bleomycin group, indicative of lipid peroxidation and ferroptosis-driven membrane damage. Yizong Tongluo Formula improved pathology dose-dependently: low-dose partially reduced inflammation, medium-dose preserved alveolar integrity, and high-dose nearly reversed the damage caused by bleomycin. Nintedanib achieved optimal alveolar repair with minimal inflammation (Figure 7N). Masson staining showed extensive collagen deposition in the model group, reflecting fibroblast activation and ECM accumulation. Yizong Tongluo Formula dose-dependently reduced fibrosis, likely via SCD1-mediated membrane stabilization (Figure 7O).