IFN-γ and TNF-α Impair Lung Development by Upregulating SMAD7 to Inhibit TGF-β Signaling Pathway and ECM Dysregulation

Abstract

Inflammation plays a pivotal role in neonatal lung injury and is closely associated with the pathogenesis of bronchopulmonary dysplasia (BPD) in preterm infants, although the underlying molecular mechanisms remain incompletely understood. Our study detected elevated serum levels of interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α) in preterm neonates as early as postnatal day 1 among those who later developed moderate-to-severe BPD. In pulmonary fibroblasts, co-treatment. with IFN-γ and TNF-α significantly downregulated α-smooth muscle actin (α-SMA) and disrupted extracellular matrix (ECM) homeostasis, evidenced by reduced collagen type I alpha 1 (COL1A1), collagen type III alpha 1 (COL3A1), and elastin expression, but elevated fibronectin 1 (FN1) and matrix metalloproteinase-1. Furthermore, dual-cytokine exposure attenuated SMAD2/3 phosphorylation and nuclear translocation, while upregulating SMAD7. Parallel experiments using E19.5 fetal mouse lung explants recapitulated these changes, showing decreased COL1A1, elevated SMAD7, and BPD-like histopathological alterations, including alveolar simplification and enlarged airspaces. Mechanistically, IFN-γ and TNF-α synergistically promoted SMAD7 overexpression, which competitively bound to SMAD2/3 and suppressed TGF-β signaling, ultimately leading to ECM dysregulation. These data delineate a novel inflammatory axis impairing lung development, highlighting SMAD7 and TGF-β pathways as promising intervention targets.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10753-025-02357-7.

Introduction

Bronchopulmonary dysplasia (BPD) has emerged as a predominant chronic lung disease affecting preterm neonates and serves as a significant indicator of poor outcomes related to premature delivery [1]. Each year in the U.S., about 50,000 infants with extremely low gestational age are born, among whom approximately 35% eventually develop this condition [2]. Research conducted across multiple centers in China’s Guangdong province demonstrated that 63.7% of extremely low birth weight infants needed supplemental oxygen at 28 days postnatal age [3], with a BPD diagnosis confirmed in 32.3% of extremely preterm cases [4]. The disease pathogenesis involves maladaptive repair processes following repetitive pulmonary injuries occurring during crucial developmental phases, either before or after birth.

Affected neonates exhibit heightened respiratory compromise, extended hospital stays [5], and persistent airway obstruction that frequently evolves into chronic respiratory dysfunction in later life [6]. Growing evidence suggests that inflammatory events occurring during gestation or the perinatal period play a pivotal role in predisposing to neonatal lung damage and show a strong correlation with BPD development. Multiple investigations have detected increased concentrations of inflammatory mediators such as TNF-α, IFN-γ, and IL-6 in both amniotic fluid and neonatal bloodstream, which may contribute to impaired alveolar formation and abnormal pulmonary vascular development [7, 8]. Nevertheless, the exact biological pathways through which these inflammatory molecules regulate the characteristic pulmonary developmental abnormalities in BPD remain incompletely elucidated.

The extracellular matrix (ECM) constitutes a sophisticated three-dimensional framework generated by stromal cells, encompassing structural proteins (collagens), adhesive glycoproteins, glycosaminoglycans, and other macromolecular components that orchestrate tissue patterning during embryogenesis, physiological maintenance, and regenerative processes. Within the pulmonary microenvironment, this dynamic ECM architecture serves dual roles: providing critical structural integrity for parenchymal cells while simultaneously mediating mechanochemical signaling crucial for pulmonary morphogenesis [9]. Lung morphogenesis progresses through five consecutive developmental phases: the embryonic period (weeks 4–7), pseudoglandular phase (weeks 5–17), canalicular stage (weeks 16–26), terminal saccular period (weeks 24–38), and alveolar maturation phase (week 36-postnatal) during which precisely regulated ECM remodeling governs cellular proliferation and differentiation events [10, 11]. Conversely, developmentally programmed ECM synthesis and degradation represent fundamental requirements for proper organogenesis. Disruptions in these tightly coordinated ECM-cell interactions can result in aberrant pulmonary development, as observed in bronchopulmonary dysplasia. Notably, pathological reactivation of developmental pathways may underlie various pulmonary disorders including idiopathic interstitial pneumonia, pulmonary vascular remodeling diseases, and thoracic malignancies, all of which exhibit characteristic ECM compositional changes [12, 13].

The transforming growth factor-β (TGF-β) signaling cascade serves as a master regulator of extracellular matrix homeostasis, critically governing both ECM biosynthesis and accumulation processes [14]. Mechanistically, TGF-β induces phosphorylation of SMAD2/3 proteins, which subsequently form heteromeric complexes with SMAD4 and translocate to the nucleus to activate transcription of key ECM components including COL1A1, COL3A1 and FN1 genes [15, 16]. This pathway simultaneously modulates ECM turnover by enhancing TIMP1/2 expression while suppressing MMP2/9 activity, thereby shifting the balance toward matrix accumulation [17, 18]. Signal initiation occurs when TGF-β ligands engage specific cell surface receptors (types I and II serine/threonine kinase receptors), inducing receptor phosphorylation and subsequent activation of downstream effectors. The SMAD protein family, consisting of eight structurally related members, mediates these signals through three functionally distinct subgroups: (1) receptor-activated SMADs (R-SMADs: SMAD1/2/3/5/8), (2) the common-mediator SMAD4, and (3) inhibitory SMADs (I-SMADs: SMAD6/7) that negatively regulate the pathway [19]. Notably, genetic ablation of Smad3 in murine models leads to profound pulmonary abnormalities characterized by defective alveolarization, airspace enlargement, and marked reduction in lung tropoelastin expression - pathological features that closely recapitulate the hallmarks of both bronchopulmonary dysplasia and emphysema [20].

This study revealed that preterm infants with moderate-to-severe BPD exhibited significantly increased serum levels of IFN-γ and TNF-α on postnatal day 1. The synergistic action of IFN-γ and TNF-α upregulates SMAD7, thereby inhibiting TGF-β-Smad2/3 signaling and suppressing fibroblast activation and ECM remodeling. Additionally, lung explants treated with IFN-γ and TNF-α showed reduced alveolar number and enlarged alveolar diameter, mimicking key pathological features of BPD.

Materials and Methods

Clinical Data Collection

This study included very preterm infants admitted to the Department of Neonatology at the Third Affiliated Hospital of Guangzhou Medical University between February and December 2021. The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Third Affiliated Hospital of Guangzhou Medical University (Approval No. [2020] 097). Written informed consent was obtained from all participants or their legal guardians. Inclusion criteria included infants with a gestational age of less than 32 weeks who were admitted to the neonatal intensive care unit within 24 h of birth. Exclusion criteria comprised transfer or death before 36 weeks’ postmenstrual age, incomplete serum sample collection, unsuccessful propensity score matching, or the presence of comorbidities known to affect lung development—such as congenital heart disease or severe congenital infections. Collected clinical data encompassed perinatal factors (gestational age, birth weight, delivery mode, amniotic fluid status, sex, Apgar scores, singleton/twin status), respiratory support details (FiO₂ on postnatal days 1, 7, 14, 21, and 28; cumulative oxygen exposure; durations of invasive/non-invasive ventilation), and initial white blood cell counts. Bronchopulmonary dysplasia (BPD) was diagnosed according to the 2018 NICHD revised criteria [21]. Based on clinical outcomes and prognosis, infants without BPD and those with Grade I BPD were categorized into a Non/Mild BPD group, whereas infants with Grade II or III BPD were classified into a Moderate/Severe BPD group.

Luminex Multiplex Assay

All preterm infants included in the study underwent arterial or venous blood collection for 0.5 mL per time on the postnatal day 1, 14, and 28 respectively. The samples were centrifuged at 3,200 rpm for 10 min at room temperature to separate serum, which was then aliquoted into EP tubes and stored at − 80 °C until analysis. After completing specimen collection, serum levels of TNF-α, IFN-γ, IL-10, CCL1, VEGF, and PDGF-BB were simultaneously quantified using Luminex technology (Kit: Shanghai Univ Biotechnology Co., Ltd., Cat. No. LXSAHM-22), strictly following the manufacturer’s protocols.

Cell Culture

The human alveolar epithelial cell line A549 and the lung fibroblast cell line MRC-5 were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). A549 cells were maintained in DMEM (Gibco, Grand Island, NY), and MRC-5 cells were cultured in MEM (Gibco); both media were supplemented with 10% fetal bovine serum (FBS; Gibco). Cells were passaged using 0.25% trypsin-EDTA and maintained under standard culture conditions (37 °C, 5% CO₂). Primary alveolar epithelial type I (AT1) cells were isolated from the lungs of 6-week-old C57BL/6J mice. The isolation procedure included enzymatic dissociation of alveolar epithelial cells via intratracheal dispase infusion, followed by agarose solidification, mechanical mincing, and DNase I treatment. The isolated cells were cultured in supplemented Dulbecco’s Modified Eagle Medium/Ham’s F-12 under standard conditions (37 °C, 5% CO₂), yielding approximately 1–3 × 10⁶ cells per mouse with characteristic cobblestone morphology [22]. These cells exhibited limited proliferative capacity in vitro, sustaining for up to two weeks. For experimental treatments, cells were exposed to commercially available cytokines—IFN-γ, TNF-α, and TGF-β (Selleck Chemicals)—at specified concentrations and durations as indicated in the results.

Fetal Lung Explant Culture

Mouse embryonic lung tissues at gestational day 19.5 were collected under sterile conditions and sectioned into 5 mm³ pieces. After thorough washing with antibiotic-supplemented PBS (1% penicillin-streptomycin) to eliminate residual blood components, tissue explants were placed on Transwell membranes in 12-well culture plates. The explant medium consisted of Minimum Essential Medium enriched with 10% fetal bovine serum, 1% penicillin-streptomycin, 2 mM L-glutamine, and 25 mM HEPES buffer, with three tissue pieces evenly arranged per well. Cultures were incubated under standard conditions (37 °C, 5% CO₂) with semi-daily medium replacement (50% volume). For cytokine treatments, 20 ng/mL concentrations of IFN-γ, TNF-α, or their combination were administered for 72 h before tissue processing for histological (hematoxylin-eosin staining) and immunohistochemical evaluation. All procedures involving animals were carried out in compliance with the ARRIVE guidelines and were approved by the Animal Experiment Ethics Committee of Guangzhou Medical University (Approval No. G2024-1274). Every effort was made to minimize animal pain, suffering, and distress.

CCK-8 Cell Viability Assay

The cell proliferation assay was performed as follows: initially, cells were plated in 96-well culture plates at a density of 5 × 10³ cells per well and incubated for 48 h with the designated treatments. Cellular viability was then evaluated using the CCK-8 assay kit (Vazyme Biotech Co., Nanjing), strictly adhering to the supplier’s instructions. Absorbance measurements were conducted at a wavelength of 450 nm employing a BioTek ELx800 absorbance microplate reader.

RNA Sequencing

RNA sequencing was conducted by Novogene Co., Ltd. (Suzhou, China). Initially, RNA concentration and integrity were evaluated using an RNA Nano 6000 Assay Kit (Agilent Technologies, Santa Clara, CA, USA) in conjunction with the Bioanalyzer 2100 system (Agilent Technologies). Subsequently, cDNA library preparation was performed following standard protocols, with library quality verification conducted using the AMPure XP system (Beckman Coulter, Brea, CA, USA). Following quality control, the prepared libraries were subjected to high-throughput sequencing on an Illumina Novaseq platform (Illumina, San Diego, CA, USA), generating 150 base-pair paired-end reads. For differential expression analysis, the edgeR package in R was employed to identify significantly differentially expressed genes (DEGs), applying stringent thresholds of |log2 fold change| > 1 and adjusted p-value (false discovery rate) < 0.05. Functional annotation of the identified DEGs was performed through Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis using the clusterProfiler package. Additionally, Gene Set Enrichment Analysis (GSEA) was implemented to evaluate whether predefined gene sets exhibited statistically significant, concordant differences between experimental groups, providing complementary pathway-level insights to the DEG analysis.

Western Blot

Western blot analysis was conducted according to standard methods [23], using whole-cell lysates prepared as described previously. Primary antibodies were used as follows: α-smooth muscle actin (α-SMA, Cat#19245, 1:1000), collagen type I alpha 1 chain (COL1A1, Cat#91144, 1:1000), SMAD2 (Cat#5339, 1:1000), phosphorylated SMAD2 (p-SMAD2, Ser465/467, Cat#18338, 1:1000), SMAD3 (Cat#9523, 1:1000), phosphorylated SMAD3 (p-SMAD3, Ser423/425, Cat#9520, 1:1000), and SMAD4 (Cat#38454, 1:1000) from Cell Signaling Technology (CST, USA); SMAD7 (Cat#ab216428, 1:1000) from Abcam (UK); β-actin (1:1000) as loading control from Beijing Boaosen Biotechnology (Beigjing, China).

RNA Extraction and Quantitative Real-time PCR Analysis

Total RNA was extracted from cells/tissues with TRIzol reagent (Thermo Fisher Scientific) following the supplier’s instructions. Subsequently, 1 µg of RNA was reverse-transcribed into cDNA using a commercial kit (Vazyme, Nanjing). qRT-PCR was carried out in a 20 µL reaction mixture containing 2 µL cDNA, 0.4 µM primers, and SYBR Green Premix (Vazyme) on a QuantStudio 6 Flex system. Thermal cycling conditions included an initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s. Amplification specificity was confirmed by melt curve analysis. Relative gene expression, normalized to GAPDH, was determined using the 2^(-ΔΔCt) method.1 Primer sequences are listed in Table 1.

Table 1.

Clinical characteristics of the eligible preterm infants

Characteristics	Non/Mild BPD group(n = 9)	Moderate/Severe BPD group (n = 9)	P
Gestational age (weeks), Maen ± SD	27.2 ± 1.21	27.08 ± 2.05	NS
Birth weight (grams), Maen ± SD	822 ± 182	814 ± 120	NS
Gender (female), No.(%)	4(44.44)	3(33.33)	NS
Cesarean section, No.(%)	7(77.78)	7(77.78)	NS
Twin birth, No.(%)	2(22.22)	3(33.33)	NS
Apgar score at 1 min, M(Q1-Q3)	7(6–8)	8(7–8)	NS
Apgar score at 5 min, M(Q1-Q3)	9(8–10)	9(9–10)	NS
Apgar score at 10 min, M(Q1-Q3)	10(9.5–10)	10(9–10)	NS
FiO₂(%) on day 1, M(Q1-Q3)	30(23–37.5.5)	28(25.5–45)	NS
FiO₂(%) on day 7, M(Q1-Q3)	21(21–26.5.5)	23(21.5–30)	NS
FiO₂(%) on day 14, M(Q1-Q3)	23(21–27)	25(22.5–30)	NS
FiO₂(%) on day 21, M(Q1-Q3)	25(21–28)	28(21.5–29)	NS
FiO₂(%) on day 28, M(Q1-Q3)	25(21–29)	28(23.5–32.5)	NS
Oxygen dependence (days), M(Q1-Q3)	68(54–80.5.5)	81(65.5–114.5.5.5)	NS
Non-invasive ventilation (days), M(Q1-Q3)	48(41–56)	61(57.5–70)	0.012
Invasive ventilation (days), M(Q1-Q3)	10(5.5–14.5)	11(2.5–38.5)	NS
WBC on day 1 (×10⁹/L), M(Q1-Q3)	6.94(3.18–13.4)	7.08(4.15–7.52)	NS

Immunofluorescence Protocol

Following fixation (4% PFA, 15 min), permeabilization (0.1% Triton X-100, 10 min), and blocking (5% BSA, 30 min), cells were incubated overnight at 4 °C with primary antibodies (α-SMA, 1:200, CST #19245; SMAD2/3, 1:200, CST #8685). After PBS washes, Alexa Fluor-conjugated secondary antibodies (1:200) were applied for 1 h at room temperature. DAPI (5 min) was used for nuclear visualization before mounting with ProLong Gold (Thermo Fisher Scientific, USA).

Co-immunoprecipitation

MRC5 cells were disrupted using RIPA lysis buffer (Thermo Fisher Scientific, USA) containing a cocktail of protease and phosphatase inhibitors. After pre-clearing, the lysates were subjected to centrifugation at 12,000 ×g for 15 min at 4 °C. The resulting supernatant was incubated overnight at 4 °C with gentle agitation in the presence of an anti-SMAD3 primary antibody (Cell Signaling Technology, #9523). Subsequently, Protein A/G magnetic beads were introduced and allowed to bind for 6 h at 4 °C to isolate the immune complexes. The beads were then washed three times with lysis buffer before proceeding to immunoblotting.

Histology and Immunohistochemistry

Lung tissue samples were carefully washed with phosphate-buffered saline (PBS) and then immersion-fixed in 4% paraformaldehyde. After embedding in paraffin, thin Sect. (5 μm thickness) were cut and subjected to H&E staining. For immunohistochemistry (IHC), tissue sections were first deparaffinized, followed by antigen unmasking and nonspecific binding blocking. Subsequently, they were exposed to primary antibodies targeting COL1A1 (Cell Signaling Technology, #72026) or SMAD7 (Thermo Fisher, #42–0400) and kept at 4 °C overnight. Following PBS washes, the sections were treated with horseradish peroxidase (HRP)-linked secondary antibodies for 1 h at room temperature. Finally, DAB substrate (Vector Laboratories, USA) was applied for signal development, and hematoxylin was used to counterstain nuclei.

Statistical Analysis

Data analysis was conducted utilizing SPSS 25 and GraphPad Prism 8. Continuous variables are presented as mean ± SD unless specified otherwise. For comparisons involving two groups, independent samples t-tests or Mann-Whitney U tests were employed based on data distribution characteristics. Multigroup analyses were performed using one-way ANOVA with appropriate post-hoc tests. Categorical variables were assessed through Pearson’s chi-square test or exact probability tests depending on sample size requirements. When parametric test assumptions (normality and equal variance) were not met, corresponding non-parametric alternatives were applied. P < 0.05 was considered statistically significant.

Result

Elevated Serum IFN-γ and TNF-α Levels at Birth in Very Preterm Infants with Moderate/Severe BPD

Previous studies have demonstrated a significant association between BPD and inflammatory cytokines [7, 24]. However, the inflammatory factors at different perinatal time points remain to be further elucidated. In this study, at first, we employed Luminex multiplex assay to systematically analyze the concentration of serum inflammatory cytokines in very preterm infants on day 1 (at birth), 14, and 28 after birth; and made comparison between Moderate/Severe BPD group and Non/Mild BPD group to identify the major elevated cytokines.

To minimize confounding effects of gestational age and birth weight, we performed propensity score matching (PSM) prior to serum analysis. The final cohort comprised 18 matched infants, with baseline characteristics presented in Table 2. The Moderate/Severe BPD group required significantly longer non-invasive ventilation compared to controls. Cytokine profiling demonstrated markedly higher serum TNF-α and IFN-γ levels on postnatal day 1 in the Moderate/Severe BPD group (Fig. 1a-b), while no significant differences were observed for these cytokines at later time points (day 14/28) or for IL-10, CCL1, PDGF-BB, and VEGF levels at any time point (Fig. 1c-f). These findings indicate that early elevation of IFN-γ and TNF-α may serve as predictive biomarkers for Moderate/Severe BPD development in very preterm neonates.

Table 2.

Primers of RT-PCR

Species	Gene	Forward primer (5'-3')	Reverse primer (5'-3')
Homo	GAPDH	CTGGGCTACACTGAGCACC	AAGTGGTCGTTGAGGGCAATG
Homo	ACTA2	GTGTTGCCCCTGAAGAGCAT	GCTGGGACATTGAAAGTCTCA
Homo	COL1A1	GTGCGATGACGTGATCTGTGA	CGGTGGTTTCTTGGTCGGT
Homo	COL1A2	GGCCCTCAAGGTTTCCAAGG	CACCCTGTGGTCCAACAACTC
Homo	COL3A1	TTGAAGGAGGATGTTCCCATCT	ACAGACACATATTTGGCATGGTT
Homo	FN1	AGGAAGCCGAGGTTTTAACTG	AGGACGCTCATAAGTGTCACC
Homo	ELN	GCAGGAGTTAAGCCCAAGG	TGTAGGGCAGTCCATAGCCA
Homo	MMP1	AAAATTACACGCCAGATTTGCC	GGTGTGACATTACTCCAGAGTTG
Homo	MMP2	TACAGGATCATTGGCTACACACC	GGTCACATCGCTCCAGACT
Homo	SMAD2	TCATAGCTTGGATTTACAGCCAG	TTCTACCGTGGCATTTCGGTT
Homo	SMAD3	CCATCTCCTACTACGAGCTGAA	CACTGCTGCATTCCTGTTGAC
Homo	SMAD7	GGACAGCTCAATTCGGACAAC	GTACACCCACACACCATCCAC

Fig. 1.

Elevated serum IFN-γ and TNF-α levels in Moderate/Severe BPD infants at birth. The serum TNF-α (a), IFN-γ (b), IL-10 (c), CCL1 (d), PDGF-BB (e), and VEGF (f) in very preterm infants were detected using Luminex multiplex assay on the first day, 14th day, and 28th day after birth. (n = 9; Mean ± SD; *P < 0.05, **P < 0.01, t-test)

IFN-γ and TNF-α Inhibit the Fibroblast Activation

To investigate the impact of IFN-γ and TNF-α on lung development, the in vitro experiments were adopted. The A549 cell line (lung epithelial), primary AT1 cells and MRC-5 cell line (lung fibroblast) were treated with different dose of IFN-γ, TNF-α, and their combination (IFN-γ + TNF-α). None of them induced cell death or affected cell proliferation (Fig. 2a-c). However, morphology of MRC-5 cells changed significantly from spindle or irregularly triangular shape to flat shape, especially treatment with IFN-γ + TNF-α for 48 h (Fig. 2d). Significant changes in cell morphology seem to imply fibroblasts have lost their mesenchymal cell properties to some extent. Furthermore, the expression of α-smooth muscle actin (α-SMA) were tested, which is a hallmark of fibroblast activation. The immunofluorescence staining showed IFN-γ + TNF-α treatment remarkably decreased the expression of α-SMA, while compared with the control, TGF-β, IFN-γ + TGF-β, or TNF-α + TGF-β treatment groups (Fig. 2e). The above results indicated IFN-γ combined with TNF-α significantly inhibit the fibroblast activation. In order to explore the underlying mechanism, the MRC-5 treated with IFN-γ, TNF-α or their combination was tested by RNA-seq. Notably, the transcriptomic changes were most prominently enriched in pathways involving extracellular matrix components and related functions (Fig. 2f).

Fig. 2.

IFN-γ and TNF-α suppress fibroblast activation. (a, b, c) A549, MRC-5 and primary alveolar epithelial type I (AT1) cells were exposed to IFN-γ, TNF-α (10, 20, 40 ng/mL), or their combination for 48 h, followed by CCK-8 viability assays. (d) Phase-contrast microscopy (20×; scale bar = 100 μm) of MRC-5 cells treated with IFN-γ (20 ng/mL), TNF-α (20 ng/mL), or both for 24 h and 48 h. (e) Immunofluorescence staining (40×; scale bar = 50 μm) of MRC-5 cells incubated with IFN-γ (20 ng/mL), TNF-α (20 ng/mL), and TGF-β (10 ng/mL) for 48 h. (f) Transcriptomic profiling (RNA-seq) of MRC-5 cells after 24 h treatment with IFN-γ (20 ng/mL), TNF-α (20 ng/mL), or their combination. Gene Ontology (GO) enrichment analysis comparing IFN-γ + TNF-α-treated cells to controls

IFN-γ and TNF-α Induce Fibroblast ECM Remodeling

To validate the gene enrichment analysis findings, we examined the effects of IFN-γ and TNF-α on ECM-related gene expression. RT-PCR analysis of MRC-5 cells treated with IFN-γ, TNF-α, or their combination for 24 h revealed a pronounced reduction in COL1A1, COL3A1, and ELN expression, whereas MMP1 and FN1 were markedly upregulated in the IFN-γ + TNF-α group compared to single treatments or controls (Fig. 3a). Consistent with these results, Western blot analysis showed dose-dependent decreases in α-SMA and COL1A1 protein levels following 24 h and 48 h of IFN-γ + TNF-α treatment (Fig. 3b–c). To further assess these effects in a developmental context, E19.5 fetal mouse lung explants were exposed to IFN-γ, TNF-α, or their combination for 72 h. Immunohistochemical analysis revealed significant downregulation of COL1A1 in the lung explants treated with IFN-γ + TNF-α group (Fig. 3d, g). Additionally, this group exhibited impaired alveolarization, characterized by reduced alveolar number and increased mean alveolar diameter (Fig. 3e–f). Collectively, these results indicate that the synergistic action of IFN-γ and TNF-α disrupts lung development by inhibiting fibroblast activation and altering ECM remodeling.

Fig. 3.

IFN-γ and TNF-α mediate extracellular matrix remodeling in fibroblasts. (a) RT-qPCR analysis of COL1A1, COL1A2, COL3A1, ACTA2, ELN, FN1, MMP1, and MMP2 mRNA expression in MRC-5 cells treated with IFN-γ(20 ng/ml), TNF-α(20 ng/ml), or their combination for 24 h. (b-c) Western blot analysis of COL1A1 and α-SMA protein expression after 24 h and 48 h treatment. (d) Representative images of HE staining and COL1A1 immunohistochemistry (40×; scale bar = 50 μm) in E19.5 fetal mouse lung explants treated for 72 h. (e-g) Quantitative analysis of alveolar number, mean alveolar diameter, and the COL1A1-positive area. Data are presented as mean ± SD. Significance was determined by one-way ANOVA (*P < 0.05, **P < 0.01, ***P < 0.001).

IFN-γ and TNF-α Decrease the ECM Through Inhibition of TGF-β/SMAD2/3 Signaling Pathway

GSEA of RNA-seq data identified significant alterations in TGF-β signaling pathway-related gene sets in the IFN-γ + TNF-α group compared to controls (Fig. 4a). Given the critical role of TGF-β in ECM deposition [14], we further examined its regulation under pro-inflammatory conditions. Exogenous TGF-β1 was applied to MRC-5 cells alongside IFN-γ, TNF-α, or their combination. RT-PCR analysis revealed that IFN-γ + TNF-α co-treatment significantly downregulated COL1A1 and α-SMA mRNA levels compared to other groups, whereas SMAD2 and SMAD3 expression remained unchanged (Fig. 4b). At the protein level, short-term TGF-β1 treatment (2 h) robustly induced SMAD2/3 phosphorylation regardless of cytokine exposure, accompanied by a modest increase in COL1A1 and α-SMA (Fig. 4c). However, prolonged treatment (24–48 h) showed that IFN-γ + TNF-α co-treatment markedly attenuated TGF-β1-induced SMAD2/3 phosphorylation and ECM protein expression, while IFN-γ or TNF-α alone had no significant effect (Fig. 4d–e). Immunofluorescence staining further confirmed these findings: TGF-β1 promoted SMAD2/3 nuclear translocation, but this effect was abolished by IFN-γ + TNF-α co-treatment (Fig. 4f–g). Together, these data demonstrate that IFN-γ and TNF-α synergistically inhibit TGF-β/SMAD signaling, thereby suppressing ECM production.

Fig. 4.

IFN-γ and TNF-α attenuate ECM production by inhibiting TGF-β/SMAD2/3 signaling pathway.(a) GSEA analysis between control and IFN-γ + TNF-α treated groups. (b) RT-qPCR analysis of COL1A1, ACTA2, SMAD2, and SMAD3 mRNA expression in MRC-5 cells treated with IFN-γ(20 ng/mL), TNF-α(20 ng/mL) and TGF-β (10 ng/mL) for 24 h. (c-e) Western blot analysis of COL1A1, α-SMA, total SMAD2/3, and phosphorylated SMAD2/3 (p-SMAD2/3) after 2 h, 24 h, and 48 h treatment. (f) Immunofluorescence staining of SMAD2/3 (40×; scale bar = 50 μm) in MRC-5 cells pre-treated with IFN-γ (20 ng/mL), TNF-α(20 ng/mL), or IFN-γ (20 ng/mL) + TNF-α(20 ng/mL) for 47 h followed by 1 h TGF-β stimulation. (g) Quantitative analysis of SMAD2/3 nuclear translocation. Data are presented as mean ± SD. Significance was determined by one-way ANOVA (*P < 0.05, **P < 0.01, ***P < 0.001).

IFN-γ and TNF-α Inhibit SMAD2/3 Phosphorylation Via Upregulating SMAD7

The regulation of the TGF-β signaling pathway involves multiple molecules and pathways, among which SMAD7 is a key negative regulator [25]. Our RNA-seq analysis revealed significant upregulation of SMAD7 expression in the IFN-γ + TNF-α group comparing with the control, IFN-γ or TNF-α(Fig. 5a). We further validated this result. MRC-5 cells were treated with IFN-γ, TNF-α and IFN-γ + TNF-α for 1 h, 12 h, 24 h and 48 h. RT-PCR analysis demonstrated that SMAD7 expression increased sharply in IFN-γ + TNF-α treatment group compared to other groups (Fig. 5b). Moreover, Co-treated with TGF-β1, IFN-γ + TNF-α treatment group markedly increased the SMAD7 at 24 h and 48 h(Fig. 5c). Previous studies proved SMAD7 competitively inhibited the binding of R-SMADs (including SMAD1, 2, 3, 5, and 8) to receptors and disrupted the interactions of R-SMADs with Co-Smad (SMAD4), thereby inhibiting the TGF signaling pathway [25]. As shown in Fig. 5d, co-immunoprecipitation revealed increased binding of SMAD3 to SMAD7, and decreased binding to SMAD4 in the IFN-γ + TNF-α co-treatment with TGF-β1. Conversely, TGF-β1 treatment enhanced the interactions between SMAD3 and SMAD4, and attenuated the interactions with SMAD 7. Consistently, IFN-γ + TNF-α treatment markedly increased SMAD7 expression in fetal mouse lung explants, as demonstrated by immunohistochemistry (Fig. 5e-f). These results indicated IFN-α combined with TNF-α inhibited TGF-β signaling pathway by upregulating SMAD7 expression.

Fig. 5.

IFN-γ and TNF-α suppress SMAD2/3 phosphorylation through SMAD7 upregulation. (a) Heatmap of key TGF-β signaling pathway components from RNA sequencing. (b) Time-course analysis (0, 1 h, 12 h, 24 h, 48 h) of SMAD7 mRNA expression by RT-qPCR in MRC-5 cells treated with IFN-γ(20 ng/mL) + TNF-α(20 ng/mL). (c) Western blot analysis of SMAD7 protein expression after IFN-γ(20 ng/mL), TNF-α(20 ng/mL) and TGF-β (10 ng/mL) treatment for 24–48 h. (d) Co-immunoprecipitation assay showing SMAD4 and SMAD7 binding to SMAD3 under indicated treatment. (e) Representative images of HE staining and SMAD7 immunohistochemistry (IHC) in E19.5 fetal mouse lung explants after 72 h treatment. (f) Quantitative analysis of SMAD7 IHC staining intensity. Data are presented as mean ± SD. Significance was determined by one-way ANOVA (*P < 0.05, **P < 0.01, ***P < 0.001).

Discussion

Extensive research has established a significant link between prenatal or perinatal inflammatory exposure and bronchopulmonary dysplasia (BPD) pathogenesis. A comprehensive meta-analysis incorporating 158 clinical studies reaffirmed chorioamnionitis as an independent risk factor for BPD in premature infants [26]. Subsequent investigations further demonstrated a positive correlation between the histological grading of chorioamnionitis and BPD occurrence [27, 28]. Consistently, numerous studies have indicated that elevated levels of TNF-α and IFN-γ in the early postnatal period are associated with the development of BPD [7, 29–31]. Our study extends these observations by revealing that very preterm infants who later developed moderate-to-severe BPD exhibited significantly elevated serum levels of both IFN-γ and TNF-α as early as postnatal day 1. This early cytokine surge may serve as a predictive biomarker for subsequent BPD progression. Notably, a recent multi-omics study on hyperoxia-induced BPD models highlighted the sustained upregulation of TNF-α signaling networks as a key driver of persistent lung injury [32].Importantly, our in vitro and ex vivo experiments demonstrated that the synergistic, rather than individual, action of IFN-γ and TNF-α potently inhibited pulmonary fibroblast activation and disrupted extracellular matrix (ECM) homeostasis.

The lung extracellular matrix (ECM) is a dynamic structure that actively directs lung morphogenesis [33]. Our data demonstrate that co-treatment with IFN-γ and TNF-α significantly altered the expression of key ECM components. This inflammation-induced ECM disorganization is a central feature of BPD pathology, as recently corroborated by proteomic analyses of human BPD lungs showing persistent dysregulation of matrisome proteins into early childhood [34].Our findings that IFN-γ and TNF-α synergistically inhibit fibroblast activation and ECM production appear to contrast with some previous reports describing pro-fibrotic effects of TNF-α in certain contexts [35, 36].This discrepancy may be explained by several critical factors: First, the cellular context is crucial. While some studies in epithelial cells or alternative fibroblast lines have reported TGF-β co-activation, our model uses lung fibroblasts (MRC-5) and focuses on the developmental impact during alveologenesis, a process distinct from adult fibrotic repair. Second, the cytokine milieu and concentration are decisive. Low-dose or sequential exposure may produce different outcomes than the high-dose, concurrent dual-cytokine challenge used here to mimic the robust inflammatory insult in preterm infants. Third, the timing of assessment is vital. We observed that the inhibitory effect on TGF-β signaling was most pronounced after 24–48 h of sustained exposure, suggesting that prior short-term studies might have captured only initial signaling events.

The TGF-β signaling pathway is a master regulator of ECM homeostasis, critically governing both ECM biosynthesis and degradation [14]. Either hyperactivation or deficiency in TGF-β signaling can disrupt normal lung development [20, 37–41]. The core mechanistic insight of our study is the elucidation of the “IFN-γ + TNF-α–SMAD7–TGF-β” axis. We demonstrate that this cytokine combination synergistically induces the transcription of SMAD7, a potent intracellular antagonist of TGF-β signaling. The novelty of our finding lies in identifying SMAD7 as the central node transducing synergistic inflammatory signals into TGF-β pathway suppression. Recent work has illuminated the context-specific regulation of SMAD7, for instance, in cardiac fibroblasts, inflammatory priming can elevate SMAD7 to limit fibrotic responses [42]. Our data establish a parallel mechanism in the developing lung. The induced SMAD7 protein competitively binds to activated SMAD2/3, disrupting the formation of the functional SMAD2/3-SMAD4 complex and preventing its nuclear translocation. Consequently, despite the presence of TGF-β ligand, the expression of critical target genes such as COL1A1 and α-SMA is profoundly suppressed. This mechanism is consistent with recent work by Callaway et al., which emphasized the critical role of TGF-β signaling in maintaining alveolar type 1 cell fate and orchestrating the alveolar matrisome [43]. Our findings thus position SMAD7 as a crucial node integrating inflammatory signals with TGF-β-mediated lung development. Beyond this core axis, it is plausible that other mechanisms contribute to the overall phenotype. For example, the observed upregulation of MMP1 could be mediated through parallel pathways such as NF-κB, which is robustly activated by TNF-α. Future studies should explore potential crosstalk between the SMAD7 axis and these parallel pathways.

Our study has several limitations. The clinical sample size was relatively small, and in vivo validation using conditional Smad7 knockout models will be essential to confirm its pathogenic role. Nevertheless, our findings provide compelling evidence for a novel inflammatory pathway in BPD pathogenesis. In summary, we have delineated a novel inflammatory pathway in which IFN-γ and TNF-α act synergistically to upregulate SMAD7, leading to a state of TGF-β resistance, impaired fibroblast activation, dysregulated ECM homeostasis, and ultimately disrupted alveolar development. Targeting this specific inflammatory axis, particularly the upstream signaling that drives SMAD7 overexpression, could represent a more precise therapeutic strategy for mitigating lung injury in preterm infants without globally inhibiting TGF-β, which plays complex and stage-dependent roles in lung development.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

XTL, WLH and JWW: data curation, methodology, validation and writing original draft. LZ and XJL: methodology and visualization. ZTM and CHJ: Formal analysis and software. ZWS and FW: conceptualization, funding acquisition, project administration and Writing review & editing.

Funding

This work was supported by Science and Technology Projects in Guangzhou (grant number 2024A03J0145 and 2023A03J0381) and Guangzhou Municipal Health Commission (grant number 20241A011087).

Data Availability

The data analyzed in this study are available within the article or from the corresponding authors upon reasonable request.

Declarations

Ethics Approval and Consent to Participate

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Third Affiliated Hospital of Guangzhou Medical University (Approval No. [2020] 097). Written informed consent was obtained from all patients involved in the study. All animal experiments involved in this study were conducted in accordance with the ARRIVE guidelines and approved by the Animal Experiment Ethics Committee of Guangzhou Medical University (Approval No. G2024-1274). Every effort was made to minimize pain, suffering, and distress in the animals.

Consent for Publication

Not applicable.

Clinical Trial Number

Not applicable.

Competing interests

The authors declare no competing interests.