Variable steady inflammation and inflammatory responses in precision-cut lung slices from various IPF lung Regions Authors: Carlos Machahua, Thomas M Marti, Namrata Kewalramani, Patrick Dorn, Lavinia Neubert, Christina Petzold-Mügge, Manuela Funke-Chambour Published: 2025 ABSTRACT AbstractBackgroundThe role of inflammation in the pathogenesis of idiopathic pulmonary fibrosis (IPF) remains controversial. While immune activation is evident in IPF lungs, immunosuppressive therapies have historically been associated with adverse outcomes. Investigating the transcriptional response of IPF lung tissue to inflammatory stimuli may help elucidate the contribution of dysregulated inflammatory pathways to the underlying disease mechanism.MethodsTo test the inflammatory response of IPF lung tissue, human precision-cut lung slices (PCLS) were generated from 3 different regions of IPF lungs: non-fibrotic (NF), transition (T), and dense fibrotic area (F). For control, we obtained PCLS from normal lung tissue (N). PCLS were incubated with or without 100 ng/ml of lipopolysaccharide (LPS) for 48 h. Bulk RNA sequencing was performed and evaluated by differential gene expression (DGE), gene ontology (GO), gene set enrichment analysis (GSEA) and network analysis. Cytokines levels in supernatants, such as tumour necrosis factor (TNF), interleukin (IL)-6 and IL-8, were quantified by ELISA.ResultsOur findings confirm metabolic reprograming, including suppression of fatty acid/cholesterol metabolism and ribosomal genes, as characteristic hallmark of all IPF regions compared to N tissues, with a fibrosis severity-dependent gradient. Interestingly, untreated NF- and T-IPF PCLS showed upregulated inflammatory genes associated with toll-like receptors (TLR), TNF and IL-6 signature when compared to N and F-IPF. However, when challenged with LPS, most of the inflammatory genes displayed attenuated activation in IPF tissues, such as markers for T cells activation (INF-γ and α, TNF via NF-κB, TLR). TNF release was greater in N tissue compared to NF- and T-IPF but less for IL-6 and IL-8 if compared to IPF tissue.ConclusionOur findings show steady inflammation in normal appearing and transitional areas of IPF lung tissue. These inflammatory markers equally participate in the pro-fibrotic signalling through signal transducers and transcription factors, which might result in dual pro-fibrotic and inflammatory effects during disease progression.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12931-025-03389-1.Keywords: LPS, Inflammation, Lung fibrosis, IPF regions, Transcriptome Background Progressive pulmonary fibrosis (PPF) is a heterogenic group of interstitial lung diseases (ILD) that share a progressive disease course with overlapping pathomechanisms, genetic phenotypes and clinical manifestations [1]. Idiopathic pulmonary fibrosis (IPF) has been described as the archetype of PPF, with a fatal prognosis and known risk factors, while the origin remains unknown [2]. IPF is characterized by advanced lung parenchymal remodelling: the disruption of alveolar structures and the replacement of alveolar epithelial cells (AECs) by hyperplastic basaloid cells, continuous extracellular matrix (ECM) deposition and the accumulation of myofibroblasts [3]. IPF shows a characteristic heterogeneous distribution of fibrosis, with predominantly subpleural and basal fibrotic regions. While cell composition changes between histologically distinct regions [4], some authors have shown that gene expression changes occur in IPF lungs even before fibrosis is histologically evident [5]. The cytokine response upon various stimuli in between different regions is unknown. Although inflammation has been observed in patients with IPF [6–10], it is currently not considered a primary contributor to the disease’s pathogenesis. Moreover, immunosuppressive drugs increased mortality in IPF compared to placebo [11]. Monoclonal antibodies against interleukin (IL)−13 and IL-4, that are type II cytokines involved in fibroblast to myofibroblast differentiation [12], did not reduce lung functional decline in IPF [13, 14]. Interferon gamma (IFN-γ), aimed to enhance effects of anti-fibrotic drugs according to preclinical studies [15], did not improve IPF patient survival in clinical trials [16]. Despite numerous discouraging clinical trials results, multiple lines of evidence point to a role of dysregulated inflammatory mechanisms in IPF pathogenesis. This is highlighted by the recent success of dual anti-inflammatory and anti-fibrotic drugs in PPF and IPF by phosphodiesterase-4 (PDE4) inhibition [17, 18]. Various inflammatory cells are found in IPF lungs. In bronchoalveolar lavage (BAL) from IPF patients, M1 and M2 alveolar macrophage (AM) activity has been described with increased inflammatory chemokine (C-C motif) ligand 13 (CCL13), CCL18, chemokine (C-X-C motif) ligand 5 (CXCL5), CXCL7 and CXCL10 release compared to healthy controls [6]. Calprotectin, an inflammatory marker found in neutrophils, was increased in blood from IPF patients from two independent cohorts and had significant negative correlation with diffusing capacity for carbon monoxide (DLCO) [9]. S100 calcium-binding protein A12 (S100A12), a member of the same protein family as calprotectin, has been associated with greater disease severity in patients with IPF [7]. Studies using single-cell RNA sequencing (scRNA-seq) found different overrepresented subsets of inflammatory cells in IPF lungs compared to healthy controls, e.g. AM (SSP1+), T cells (FOXP3+) and B cells (naive and memory) [10]. RNA-Seq analyses in IPF lung explants from macroscopically normal-appearing and scarred areas compared to normal tissue revealed increased expression of inflammatory markers [8]. Inflammation might contribute to IPF disease acceleration, most evident in patients with acute exacerbation (AE). AE in IPF (AE-IPF) is the acute worsening of the disease with the appearance of new consolidations [19]. During AE pro- and anti-inflammatory cytokines are increased [20]. Among others, bacterial or viral infections can induce secondary AE-IPF [21]. Evaluating the prevalence of bacterial infection in IPF patients, Moghoofei et al. showed that Pseudomonas aeruginosa (P. aeruginosa), Klebsiella pneumoniae (K. pneumoniae) and Haemophilus influenzae (H. influenzae) have similar coinfection patterns, which enhance the risk of death by AE in IPF [22]. Meanwhile, bacterial colonisation has been associated with IPF patients’ outcome, as an increased bacterial load is associated with a poor prognosis [23]. These studies might suggest involvement of underlying inflammation in the pathogenesis of IPF disease and progression, which can be influenced by bacterial exposure. A better understanding of the role of inflammation in IPF lungs, whether triggered by infection or other related factors, may help refine treatment strategies e.g. by targeting specific inflammatory mechanisms in lung fibrosis. Precision-cut lung slices (PCLS) are an organotypic model that comprises several lung cell types displayed in their intact lung parenchymal structure with cell-cell and cell-matrix interactions [24]. Human normal PCLS show a consistent inflammatory response against lipopolysaccharides (LPS) [25–27] as well as early fibrotic changes after treatment with transforming growth factor beta 1 (TGF-β1) or other pro-fibrotic factors [28–31]. While PCLS is a suitable model to mimic early IPF disease mechanisms, the response to pro-inflammatory stimuli such as LPS has yet to be elucidated. Investigating the transcriptional changes that occur in IPF lungs upon exposure to bacterial endotoxins may offer valuable insights into the dysregulated inflammatory mechanisms of the underlying disease. In our current study, we have treated PCLS from three histologically different IPF tissue regions with LPS, an endotoxin obtained from K. pneumoniae, to assess the inflammatory response in tissue and to compare them to normal lung tissue obtained from patients undergoing tumour resection. For comparison, untreated PCLS from IPF and normal lung tissue have been examined to uncover key mechanistic disparities, and between various regions of the IPF lung. Methods Donor’s characteristics Human normal lung tissues (n = 8) were obtained between 2022 and 2023 from the distal area (> 5 cm) of lung resection. Lung tissue samples were obtained by the Department of General Thoracic Surgery at the Inselspital, Switzerland, for diagnostic purposes. The remaining tissue was used for research purposes. Ethical approval from the cantonal authority was obtained (approval number: KEK-BE_2024 − 01841) and written informed consent was acquired from all donors before the intervention. Relevant donor information is displayed in Table 1. Table 1Patient characteristics: Normal lung donorsDonorAge (yr.)SexSmoking statusDiagnostic143Munknownlung metastasis275Mformer, 80PYNSCLC346Funknownlung metastasis465MunknownNSCLC-SCLC562Mformer, 70PYNSCLC655Msmoker, 90PYNSCLC772Mneverlung metastasis857Mformerlung metastasisF Female, M Male, NSCLC Non-small cell lung cancer, PY Pack-year index Patient characteristics: Normal lung donors F Female, M Male, NSCLC Non-small cell lung cancer, PY Pack-year index IPF lung tissues were obtained between 2023 and 2024 from patients who underwent lung transplantation in the Hannover Medical School Hospital, Germany. Fibrotic regions were pre-selected by the AG Lungenforschung in the Institute of Pathology of the Hannover Medical School. Donors signed a written informed consent prior to the intervention, and sample collection was approved by the local ethical committee (8867_BO_K_2020). Table 2 shows relevant information about the IPF patient group. Table 2Patients characteristics: ILD donorsDonorAge (yr.)SexDiagnostic160MIPF265MIPF360MIPF461MIPF556MIPFIPF Idiopathic pulmonary fibrosis, M Male Patients characteristics: ILD donors IPF Idiopathic pulmonary fibrosis, M Male PCLS generation PCLS were generated following previously published protocols [31]. Normal (N) samples were processed within the first 2 h of gathering and maintained in resting medium (Dulbecco’s Modified Eagle Medium, DMEM + Antibiotic/Antimycotic 1x) supplemented with insulin-transferrin-sodium selenite (ITS) liquid media supplement 1x (Sigma-Aldrich, #I3146). IPF lung tissues were pre-filled with agarose at the originating site and sent in culturing medium at 4 °C to be processed within the first 12 h. The following three different regions were selected according to the fibrotic extent: (1) non-fibrotic area (NF-IPF), (2) transition area (T-IPF), and (3) mainly fibrotic area (F-IPF). Tissue blocks from NF- and T-IPF regions were obtained from all IPF donors (n = 5), while F-IPF was obtained only from four IPF donors for technical reasons (n = 4). Slicing of lung tissue blocks was performed at the University of Bern using a Compresstome® tissue slicer (Precisionary Instruments LLC); all PCLS were cut at 400 μm thickness and 4 mm diameter. Samples were incubated with 250 µl DMEM supplemented with 0.1% foetal bovine serum (FBS) in a 48-well plate at 37 °C with 5% CO2. Supplemental Table 1 summarises the analysis-specific sample use of all donor’s samples. Viability tests Prior to experiments, PCLS were tested for viability and cytotoxicity with different doses (0 to 10 µg/ml) of LPS from K. pneumoniae (Sigma-Aldrich, #L4268) in DMEM supplemented with 0.1% FBS, using CyQUANT™ LDH cytotoxicity assay kit (Thermo Fisher, #C20301) and XTT cell viability assay (Thermo Fisher, #X12223); assays were adjusted for a 48-well plate setting. After 24 h of incubation, 50 µl medium was mixed with 50 µl of reaction mixture to calculate the amount of LDH released, the reaction was stopped after 15 min with 50 µl stop solution and read in a Varioskan LUX multimode microplate reader (Thermo Fisher). Meanwhile, to assess viability by cellular redox potential, 70 µl of working solution was added in each test well, absorbance was measured 4 h after incubation. Some PCLS were incubated with 25 µl of lysis buffer as a positive control for cell death, and only medium without PCLS were used to subtract background signal. The percentage relative absorbance to the untreated control was used to calculate viability and cytotoxicity as per manufacturer’s recommendations. LPS treatment For the dose-response analysis, a group of PCLS were incubated with increasing doses of LPS in DMEM supplemented with 0.1% FBS, from 0 to 1000 ng/ml; each condition was replicated 3 times. Supernatants were collected in a 1.5 ml Eppendorf tubes 24 h later and stored at −80 °C until analysis. To assess the inflammatory response, PCLS were kept in culture with DMEM supplemented with 0.1% FBS for 3 days prior to treatment, with medium change on day 1. Then, PCLS were treated with LPS (100 ng/ml) or left untreated for the control group. Supernatants were collected 24 h after LPS addition in a 1.5 ml Eppendorf tube, while PCLS (4 units) were collected in cryotubes and snap-frozen in liquid N2 48 h after LPS treatment. All samples were stored at −80 °C until processing. Bulk RNA-sequencing (RNA seq) RNA extraction for bulk RNA-seq was done following Stegmayr et al. [32] procedure. Briefly, PCLS were transfer to a 2 ml Eppendorf tube with 400 µl TRIZol (Thermo Fisher, # 15596026) and stainless-steel beads (Qiagen, #69997), and homogenised in a TissueLyser II (Qiagen, USA). Lysated was transfer to RNeasy Mini spin column (Qiagen, # 74104), and resulting RNA was recovered in 20 µl RNAse free water. All steps post RNA extraction were performed at the Next Generation Sequencing Platform, University of Bern. N and IPF samples were sequenced in two separated batches, following the same library preparation and sequencing method for both groups, and adding two bridge samples from the N group to the IPF samples in order to take in account possible batch effect. Briefly, RNA quantity and quality control was assessed using a Thermo Fisher Scientific Qubit 4.0 fluorometer with the Qubit RNA BR Assay Kit (Thermo Fisher Scientific, Q10211) and an Advanced Analytical Fragment Analyzer System using a Fragment Analyzer RNA Kit (Agilent, DNF-471), respectively. For the library preparation, Poly(A) selection was made with 200 ng or 100 ng of input RNA from N and IPF samples, respectively. Poly(A) RNA was enriched by hybridization to oligo(dT) magnetic beads using a Poly(A) RNA Selection Kit V1.5 from Lexogen (Lexogen, SKU 157), following their user guide (Lexogen, 157UG069V0101); then cDNA libraries were generated using a CORALL Total RNA-Seq V2 library Prep. Kit with UDIs 12nt set A2 & A3 (Lexogen, SKU 176.96) and 14 PCR cycles according to the protocol in the Lexogen user guide 171UG394V0100. Quality and length assessments of the generated cDNA were performed using a Thermo Fisher Scientific Qubit 4.0 fluorometer with the Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific, Q32854) and an Agilent Fragment Analyzer (Agilent) with a HS NGS Fragment Kit (Agilent, DNF-474), respectively. Regarding the first batch, cDNA libraries were pooled and sequenced with a loading concentration of 700 pM, paired-end and dual indexed, using an illumina NextSeq 1000/2000 P2 Reagent kit (100 cycles) (illumina, 20046811) on an illumina NextSeq 1000 instrument. The read set-up was as follows: read 1: 63 cycles, i7 index: 12 cycles, i5: 12 cycles and read 2: 51 cycles. Meanwhile, for the second batch, the cDNA libraries were equimolar pooled and sequenced 50 bp paired end using an illumina NovaSeq 6000 S1 Reagent Kit v1.5 (100 cycles; illumina, 20028319). In both cases, the quality of the sequencing runs was assessed using illumina Sequencing Analysis Viewer (illumina version 2.4.7 and 2.5.12, for N and IPF samples, respectively) and all base call files were demultiplexed and converted into FASTQ files using illumina bcl2fastq conversion software v2.20. A minimum of 25 M reads/cell were generated for each library. ELISA for cytokine measurements Supernatants from the experiments were analysed for human tumour necrosis factor (TNF) (#DY210-05, R&D System, USA), human IL-6 (#DY206-05, R&D System), and human IL-8 levels (#DY208-05, R&D Systems) by ELISA. Samples were pre-diluted and measured in triplicates, and the ELISA assays were performed following the manufacturer’s recommendations. Statistical analysis RNA-seq data processing was performed by the Interfaculty Bioinformatics Unit at the University of Bern, following previously published approaches [33]. Briefly, the quality of the RNA-seq data was assessed using FastQC v.0.12.1 and RSeQC v.5.0.3. Mapping to the reference genome (Homo_sapiens.GRCh38.107) and counting the number of reads overlapping with each gene were done using HISAT2 v.2.2.1 and FeatureCounts v.2.0.1, respectively. Differential gene expression (DGE) analysis between groups was performed using DESeq2 v1.42.0. Principal component analysis (PCA) revealed minimal differences between replicate samples across both batches (data not shown); therefore, batch correction was not required. To account for potential variability between donor regions, we applied a multi-factor design in DESeq2, enabling estimation of the condition effect while controlling for donor region differences. To identify gene ontology (GO) and gene set enrichment analysis (GSEA) using gene sets from Kyoto encyclopedia of genes and genomes (KEGG) and the molecular signatures database (MSigDb), ClusterProfiler v4.10.0 was used. PCA, volcano plots and clustered hierarchical heatmaps were obtained through the interactive Shiny application v1.6.0. Significant differentially expressed genes (DEGs) were considered when Log2 fold change was greater than ± 0.5 and the p-value adjusted for multiple testing was lower than 0.05. Later, we performed a weighted gene co-expression network analysis (WGCNA) and analysed the correlations using the WGCNA R package (version 1.72) [34, 35]; we performed a block-wise network construction and module detection. This analysis generated “Modules” or groups of genes with a strong connection to each other. Mean Z-score of all genes integrating a module was calculated and plotted for each sample in each category (treatment, region, donor and condition). Pearson’s linear correlation was calculated between Module membership and gene significance for each category. GO and GSEA analysis were conducted with the same package as for DGE (ClusterProfiler v4.10.0). All analyses were run in R version 4.3.2 (2023-10-31) [36]. In addition, within the hallmark Inflammatory Response comparing untreated or LPS-treated N and all three IPF PCLS, clusters of the most upregulated genes (Log2FC 0.5, adj. p-value < 0.05) were re-selected in “Sets” and re-analysed using the STRING database (version 12.0, SIB, Switzerland) to obtain their functional annotation through Gene Ontology (GO) for biological process. Regarding viability/cytotoxicity tests and ELISA results, groups were compared by two-way ANOVA, with Tukey’s multiple comparison test as post hoc in GraphPad Prism v. 10 (GraphPad Software Inc., US). Graphs show the mean and ± standard deviation for all datasets. A p-value < 0.05 was considered significant. Results Untreated non fibrotic IPF lung regions (NF-IPF) show increased inflammation compared to normal (N) lung tissue To compare the inflammatory response of normal human lung tissue with different regions of IPF lungs, we treated PCLS according to our experimental plan (Fig. 1A). The potential cytotoxic effect of LPS on the PCLS and its impact on their viability was previously assessed, showing no significant differences between untreated samples and the tested doses in both assays and between N group and different IPF regions (Supplemental Fig. 1A-B). Comparing untreated PCLS, PCA of the 500 most significant genes differentiated the normal parenchyma from the most fibrotic regions (Supplemental Fig. 2 A, PC1 = 50.3%). The scatter plot demonstrated that samples from the normal (N) patient group clustered together and were distinct from all three IPF regions, while IPF regions showed considerable overlap, particularly between the NF and T-IPF areas compared to the F-IPF region. First, we compared the transcriptomic signature of all three untreated IPF PCLS vs. N group independently, to elucidate dysregulation in all three IPF regions. Volcano plot showed that IPF PCLS presented substantially more downregulated than upregulated genes compared to the N patient group (Fig. 1B). 1451 genes in the DGE analysis were dysregulated in NF, T and F-IPF areas when compared to N lung tissue (Fig. 1C). GSEA showed these DEGs belonged to 18 common annotations (Supplemental Fig. 2B), some of them related to the suppression in fatty acid metabolism (FASN, MVD, HMGCS1, PPARG, FABP7) and oxidative phosphorylation (UQCR10, ATP6V0E1, COX5A, PINK), or the activation of epithelial-mesenchymal transition (EMT) (SERPINE2, CCN2, COL5A3, PDGFRB, WIPF1) and hypoxia (IR52, CA12, CCN5, TMEM45A) (Supplemental Fig. 2 C). Increased inflammation in untreated IPF regions (NF-, T-, F-IPF) compared to normal (N) PCLS. A Experimental design. B Volcano plot comparing DGE of untreated NF-IPF and N tissue. C Venn diagram comparing all three IPF regions vs. N group in untreated condition. D Common annotation in GSEA for untreated NF- and T-IPF PCLS vs. N PCLS. E Clustered heatmap of the hallmark “Inflammatory Response” showing three distinct sets of upregulated genes: N PCLS (Set 1), NF- and T-IPF regions (Set 2), and F-IPF PCLS (Set 3). F GO analysis reveal that genes are associated with T cells activity and INF-γ activation (Set 1), leucocyte chemotaxis and response to LPS (Set 2), and with angiogenesis and neutrophils chemotaxis (Set 3). DGE parameters: Log2FC ± 0.5, adj. p-value < 0.05 The integrative network analysis allowed us to identify a set of 1031 genes in Module 3 that were essentially downregulated in IPF lung tissue compared to N lung tissue group, observed in both untreated and LPS-treated (e.g. SFPTC, SFTPA1, CTSH, LAMP3, FASN, PDP2, LDLR, LPIN1). These genes showed overall positive mean Z-scores in the N group compared to negative values in all three IPF PCLS (Supplemental Fig. 2D-E). Linear correlation (r) and p-value between Module membership in each Module and the gene significance for each parameter (condition, donor, region, and treatment) are represented in Supplemental Table 2. In Module 3, GSEA showed that downregulated genes in all three IPF regions were related to fatty acid, steroids, glycerophospholipid metabolism (CYP51A1, MVK, HMGCR, STARD4, NPC2), biosynthesis of amino acids (SHMT1, ARG2, CTH, PSPH), tRNA biosynthesis (WARS1, GARS1, IARS1), and ribosome (MRPL14, MRPS15, MRPL21) (Supplemental Fig. 2 F). Interestingly, an overall Inflammatory Response was activated only in NF and T-IPF tissue compared to N tissue (Fig. 1D). We observed that the genes belonging to this hallmark provided a differential pattern of inflammatory activation depending on the tissue origin (Fig. 1E). This sub-analysis reflected greater activation of T cell-mediated pathways (TNFSF9, IL18, CD55, SLC7A1) in N PCLS (Set 1); leukocyte migration (CCL2, CCL22, CCL8, IL10, GPR183, CCR7); response to molecules of bacterial origin (BDKRB1, PDE4B) with LPS and Toll-like receptors (TLRs) (TLR4 LY96, IRF5, IRF9, MAP3K8, AKT3) in NF and T-IPF PCLS (Set 2). Moreover, increased neutrophil chemotaxis (CX3CL1, CXCL6, CXCL8, CCL20) and angiogenesis (ITGB8, EREG, LIF) were found in F-IPF (Fig. 1F, Set 3). In addition, GSEA showed that other inflammatory mechanisms were equally activated in NF- and T-IPF regions compared to N group: hallmarks associated to IL-6, Janus kinase (JAK)/signal transducer, activator of transcription (STAT) 3 signalling (CXCL13, IL6, MAP3K8, CXCL1, STAT2) and IL-2/STAT5 signalling (SNX9, MYO1E, IL2RA, SELL) (Fig. 1D). Interestingly, TNF signalling via nuclear factor kappa B (NF-κB) (TNFAIP2, TNFAIP8, PLAU, GPR183, EGR3, CXCL1, IRS2) was the only inflammatory annotation activated in all three IPF regions compared to N PCLS (Supplemental Fig. 2 C). Moreover, INF-γ response was partially activated: while the majority of the genes facilitating type II INF signalling were upregulated in NF- and T-IPF (VCAM1, CD69, IRF5, IRF8, IRF9, JAK2, CASP3) compared to the N group, some genes were downregulated (PTPN1, TRAFD1, HLA-G, ICAM1) (Supplemental Fig. 2 C). In contrast, F-IPF showed consistent downregulation in INF-γ and α signalling compared to the N group. Comparing untreated IPF regions in between, NF- and T-IPF regions showed similar transcriptomic profiles but were different compared to the F-IPF region (Supplemental Fig. 2G). Thirty-two annotations were common in the GSEA analysis comparing “NF vs. F-IPF” and “T vs. F- IPF”. These included mechanisms related to metabolic pathways (oxidative phosphorylation, adipogenesis), inflammatory response, IL-6 signalling, TNF-related genes (SGK1, CCL4, CSF1), INF-γ (TMT1B, CASP1, SLAM7), unfolded protein response (UPR), proteasome (UBE2D4, EIF4G1, ATP6V0D1), TLR downstream target genes (CCL4, IL12, CCL3), genes involved in viral response (TLR7, TLR8, MYD88) (data not shown), or EMT that were activated in these two IPF regions compared to the mainly fibrotic area (Supplemental Fig. 2H-I). Pro-inflammatory and pro-fibrotic signals are activated in untreated NF- and T-IPF PCLS compared to F-IPF and N tissue Upregulation of the inflammatory response in the NF-IPF tissue compared to N and F-IPF PCLS was confirmed by the integrative network analysis if directly compared. 416 genes in Module 6 were upregulated in untreated NF- and T-IPF (e.g. VCAN, VCAM1, COL3A1, WNT2, TLR4, CD4, CXCL12, CCL18) compared F-IPF and N PCLS and remained still activated after LPS treatment (Fig. 2A-B). Pro-inflammatory and fibrotic signals are activated in untreated NF and T-IPF PCLS compared to F-IPF and N PCLS. A Mean Z-score representation for Module 6 composed of 416 genes, upregulated in NF- and T-IPF compared to N group and F-IPF. B Heatmap with the 60 most significant genes in this Module. C. GO pointed out the activation of leukocyte and IL-6 production. D GSEA (KEGG) identify activation of chemokine signalling pathways, phagosome and complement in NF- and T-IPF. E Mean Z-score plot for Module 8 (274 genes) representing genes upregulated in all IPF regions compared to N group in both conditions. F Heatmap for the 60 most significant genes in Module 8. G GO annotations for biological process in Module 8 indicate EMT and ECM organisation. H GSEA (MsigDB) for the hallmarks in Module 8 confirm these finding. Values in GO and GSEA are represented in gene ratio The GO analysis for biological process showed that these genes were part of the leukocyte activation (CCR1, CXCL9, IL17RA, CSF1R), extracellular signal-regulated kinases 1 (ERK1), 2 cascade (TLR4, CD4, GPR183), regulation of IL-6 (TLR6, AFI1, IL17RA, CLEC7A), and TNF production (FCGR3A, PTPN6, CD33, IL27RA) (Fig. 2C). Besides disclosing chemokine and NF-κB signalling pathways, GSEA revealed that these genes were involved in the activation of NET formation (CLEC7A, FPR1, FPR3, FCGR3A), complement and coagulation cascade (C1QB, ITGAM, C5AR1), and phagosome (TLR6, ATP6V1B2, HLA-DQA1) (Fig. 2D). Beyond the pre-activated inflammatory activity in untreated NF-IPF and T-IPF, network analysis showed presence of some pro-fibrotic signals, grouped in 274 genes in Module 8 (e.g. TGFB1, ADAM19, PDGFRB, FGFR1, HGF, NOTCH1, SNAI2, FZD1). This Module was highly upregulated in all three IPF regions compared to N group that remained upregulated after stimulation with LPS (Fig. 2E-F). E.g. genes were associated to the regulation of Bone Morphogenetic Protein (BMP) (GREM1, FZD1), Wnt signalling pathway (NOTCH1, SNAI2, SFRP2), and ECM organisation (GREM1, SFRP2) in the GO for biological process (Fig. 2G). GSEA showed that these genes promoted EMT (COL6A2, COL8A2, PDGFRB, TGFB1) in IPF PCLS, while some of these genes belonged to Notch signalling (FZD7, FZD1) amongst other hallmarks (Fig. 2H). After LPS stimulation, N PCLS showed increased inflammatory-related gene expression compared to IPF lung tissues We challenged PCLS with LPS to generate an inflammatory response and compare the transcriptomic profiles of lung tissues from “healthy” origin and histologically different IPF lung regions. The comparison of the DEGs between all LPS-treated PCLS vs. their matching untreated condition showed that, LPS caused more transcriptomic changes in N compared to any IPF region (Fig. 3A, Supplemental Fig. 3A). LPS treatment resulted in 1439 unique DEGs in N tissue compared to a few uniquely activated genes in IPF lung tissue, especially in F-IPF. These dysregulated genes in N tissue after LPS treatment were related to the activation of sphingolipid signalling pathway (SGPP2, S1PR3), NET formation at the receptor level (TLR2, FPR2, NCF1, CR1L), the suppression of folate transport, metabolism (FOLR2, FOLR3) and genes linked to mismatch repair of DNA (EXO1, RFC5, RFC4) according to GSEA (Fig. 3B). Increased inflammation in LPS-treated N tissues compared to different IPF regions. A Venn diagram of total DGE in the comparison LPS vs. untreated condition in N PCLS and all IPF regions, sharing 80 common genes. B Enrichment score of significant GSEA annotation in LPS vs. untreated in N group. C Common GSEA annotations in all lung tissues. D PCA of the 500 most significant genes in LPS-treated lung tissues. E Venn diagram for the GSEA (MsigDB) in the direct comparison LPS-treated IPF regions vs. N PCLS. F Among the common 23 annotations in LPS-treated NF-, T- and F-IPF regions compared to N group, hallmark “Inflammatory Response” is downregulated. G Clustered heatmap of the hallmark “Inflammatory Response” showing three sets of upregulated genes: N PCLS (Set 1), NF- and T-IPF (Set 2), and F-IPF (Set 3). H GO analysis for biological process indicates that genes in Set 1 are associated with response to LPS, T-cell activation and INF-γ signalling, while genes in Set 2 and 3 are related to angiogenesis and GPCR signalling pathway. DGE parameters: Log2FC ± 0.5, adj. p-value < 0.05 In addition, this approach showed 80 DEGs linked to LPS response, which were commonly dysregulated in all four comparisons and similarly activated (IRAK3, CXCL1, IL1B, PTGES, TNFAIP6) or suppressed (SCN7A, COL6A6, FOLR2). Among the common annotations in all lung tissues, the activation of essential genes related to TNF, JAK/STAT and TLR signalling (CSF3, IL1B, CCL4, CXCL3), the suppression of pyruvate metabolism (ADH1B, ACAT1, LDHB) and cell cycle (E2F2, CDK1, CDKN2C, CCNA2) (Fig. 3C) were common mechanisms induced by LPS. Moreover, we directly compared LPS-treated IPF regions versus LPS-treated N tissue. PCA for the 500 most significant genes showed that PC1 explained 63.6% of the variance separating N group from all IPF regions, while PC2 accounted for 9.61%, vaguely separating the IPF regions from each other (Fig. 3D). 611 DEGs were dysregulated only in NF-IPF compared to N group after LPS treatment, while the differences between LPS-treated F-IPF and N tissue involved up to 2354 DEGs uniquely dysregulated in this comparison (Supplemental Fig. 3B-C,). GSEA revealed that some genes related to TGF-β signalling (TGFBR1, GREM1, THBS1, INHBA) remained activated in NF-IPF after LPS stimulation, as well as the phosphoinositide 3-kinase (PI3K)/Akt signalling pathway, (HGF, IGF1, IGF1R, NGFR, FGF1, ITGA6, ITGB3, ITGA7) with genes that translate to growth factors or integrins (Supplemental Fig. 3D). In the GSEA particular annotations comparing LPS-treated F-IPF to N group revealed an incomplete activation of inflammatory mechanisms linked to TLR, JAK/STAT, and NOD-like receptor (NLR) signalling pathways, not only in genes encoding signalling proteins (NFKB1, NFKBIA, JAK1, MYD88, IRF7, MAP2K2, STAT1, CASP4, TNFAIP3) (Supplemental Fig. 3E), but also in major target cytokines and chemokines (IL12B, IL15RA, TNF, IL1B, CCL3, CCL5) of these inflammatory pathways (data not shown). Interestingly, while overall TLR signalling annotation was supressed in F-IPF compared to N PCLS, the gene TLR5 was upregulated in the most fibrotic area compared to N group (data not shown). With respect to the common hallmarks in the GSEA that all three comparisons in LPS-treated PCLS shared, we found that the INF-γ and α signalling (TRAFD1, PSME1, CSF1, IFI27, ICAM1, CD40) and genes belonging to UPR (PAIP1, ATF4, IMP3) were suppressed in all three IPF tissues compared to LPS-treated N PCLS (Fig. 3E-F). Interestingly, the overall Inflammatory Response was downregulated in all three LPS-treated IPF regions compared to N PCLS (Fig. 3G). We found that the subset of upregulated genes in N group compared to all three IPF regions were associated with the positive regulation of INF-γ production, T cell and mononuclear cell proliferation (IL12B, TNFSF9, IL1B, CD55, SLC7A1, CSF1). Meanwhile, the few upregulated genes in the NF, T and F-IPF areas were associated with smooth muscle cell migration (SERPINE1, HAS2, LPAR1, ITGB3), angiogenesis, response to wounding (ITGB8, ADORA2B, ITGB3, CX3CL1), and G protein-coupled receptor (GPCR) signalling pathways (PTGER4, CXCL9, RGS1, LPAR1) according to GO analysis for biological process (Fig. 3H). When comparing the inflammatory response among IPF regions, only 62 DEGs overlapped between F-IPF and NF-IPF regions in the transcriptomic profile (Supplemental Fig. 3F-G). GSEA indicated that LPS-treated NF-IPF activated few more inflammatory pathways than F-IPF regions: TNF, IL-6/JAK/STAT3, and INF-γ and α signalling (IFIH1, IFIT3, IL15RA, FAS, JAK2, STAT1) were some of the activated annotations in NF-IPF. Differential inflammatory gene patterns in IPF and N PCLS entail differential release of inflammatory cytokines The integrative network analysis revealed 74 genes in Module 30 (e.g. TNFRSF9, TNF, TNIP3, NFKB1, BIRC3, NFKBIA, TRAF1, TNIP1) specifically upregulated in N tissue after LPS stimulation compared to all three IPF regions, highlighting the set of genes that were not fully displayed in the inflammatory response of IPF lungs (Fig. 4A-B). GO for biological process showed that these DEGs belonged to the canonical NF-κB signal transduction (NFKB1, NFKBIA, TNF, TNIP3, TNIP1, TRAF1), leukocyte and T cell proliferation (IL23A, EBI3, TNFRSF9, CSF2, BID) (Fig. 4C). Moreover, besides TNF and NF-κB signalling pathway, GSEA indicated that these genes in Module 30 were also closely related to TLR signalling pathway; Th17 cell differentiation and NLR signalling pathway (CXCL2, BIRC3, DEFB4A) (Fig. 4D). Different inflammatory response are observed between IPF and N tissue after LPS treatment. A Mean Z-score of Module 30 (74 genes) highly upregulated in N tissue after LPS treatment compared to all three IPF lung PCLS. B Clustered heatmap representation of the 60 most significant genes in Module 30. C Gene ratio of GO for biological process in Module 30. D Gene ration of most significant GSEA (KEGG) annotations. E Mean Z-score of Module 11 (232 genes) activated in LPS-treated PCLS compared to untreated. F Clustered heatmap of 60 most relevant genes in Module 11 (G) GO annotation for biological process in Module 11. H GSEA (KEGG) for Module 11. I ELISA of cytokines TNF, IL-6 and IL-8 in normal tissue and F-IPT PCLS 24 h after LPS addition (100 ng/ml). (*) p < 0.05, (***) p < 0.001, (****) p < 0.0001 While the previous Modules highlighted that TNF expression was elevated in normal tissue following LPS stimulation compared to all three IPF PCLS samples, a distinct set of 232 cytokine-related genes in Module 11 (e.g. WNT5A, IL6, IL1B, IRAK3, TLR2, FGF2, CCL20) was strongly upregulated in all lung tissues—regardless of their origin—after LPS treatment, relative to their corresponding untreated controls (Fig. 4E-F). GO and GSEA situated genes in Module 11 within IL-6/JAK/STAT3 signalling axis, regulation of ERK1 and 2 cascade, and cellular response to IL-1; in addition to common genes in the TNF signalling pathways that are equally activated in all three lung tissues after the inflammatory stimulus (Fig. 4G-H). To assess whether the different PCLS samples elicited a similar release of the three key inflammatory cytokines involved in LPS-mediated TLR signalling—TNF, IL-6, and IL-8—we measured their levels by ELISA 24 h after incubation with increasing concentrations of LPS. Results for TNF showed that at concentrations > 100 ng/ml LPS, N tissue released greater amounts of this cytokine compared to T- (p < 0.05) and F-IPF PCLS (p < 0.01) (Supplemental Fig. 4 A). Levels of IL-6 in supernatant increased in T- (p < 0.05) and F-IPF (p < 0.01) compared to N PCLS, in a dose-dependent manner and demonstrating pronounced efficacy at lower doses in the latter (1 ng/ml). Similarly, IL-8 levels in supernatant were increased in T- and F-IPF PCLS compared to N tissue, with differences becoming apparent even at lower doses in the latter. Of note, despite the observed statistical differences at higher doses (1000 ng/ml LPS), the secretion profiles of TNF, IL-6 and IL-8 were not significantly different between N tissue and NF-IPF PCLS (p-value = 0.95, 0.17, 0.39, respectively). Under the optimized concentration conditions of 100 ng/ml LPS, human PCLS showed differences in the release of these cytokines depending on the region and/or origin. Consistent with our transcriptomic findings, F-IPF regions released lower levels of TNF 24 h after LPS stimulation compared to N tissue (p < 0.05, Fig. 4I), while similar levels of this cytokine were found in NF and T-IPF compared to N PCLS after LPS treatment (Supplemental Fig. 4B). IL-6 levels were substantially increased in all three LPS-treated IPF regions compared to N tissue (p < 0.05 vs. NF; p < 0.001 vs. T and F-IPF; Fig. 4I, Supplemental Fig. 4B). Likewise, protein levels of IL-8 demonstrated higher values in supernatant from F and T-IPF 24 h post-incubation with LPS compared to N PCLS (p < 0.01, Fig. 4I, Supplemental Fig. 4B). Finally, all transcriptomic findings and ELISA readouts from IPF PCLS are summarised in Fig. 5 highlighting the most pronounced changes. Summary of findings in the transcriptomic profile of IPF tissue before and after LPS treatment. IPF PCLS exhibited progressive metabolic dysregulation compared to N tissue, accompanied by steady inflammation predominantly in NF- and T-IPF areas. The challenge with an inflammatory agent (LPS) activates less inflammatory mechanism in all three IPF PCLS, focused on few signalling proteins and transcriptomic factors associated to JAK/STAT, which drive to a differential cytokine release, lower TNF but greater IL-6 in F-IPF compared to N PCLS. (Created in BioRender. Kewalramani, N. (2025) https://BioRender.com/xxtuvha) Discussion Inflammation is the initiating mechanism for tissue repair and self-renewal processes [37], and its resolution is essential for correct wound healing [38]. Inflammatory cytokines, such as TNF, orchestrate the transition from inflammation to fibrosis, eventually yielding control to TGF-β1 [39]. Debate surrounding the role of inflammation in IPF has been fuelled in part by the results of the PANTHER-IPF study, which showed that immunosuppression increased mortality in IPF patients compared to placebo [11]. The cause of this finding is unclear and early toxicity of higher steroid doses in frail elderly patients as well as including patients with short telomere length might have contributed to this outcome [40, 41]. Immune mechanisms might still regulate the development of fibrosis based on experimental and clinical findings [42]. Our study demonstrates that inflammatory genes associated with monocyte/leucocyte activation and chemotaxis, host defence, and TNF-related signalling show transcriptional activation in normal-appearing lung parenchyma from IPF lungs (NF-IPF) compared to normal lung tissue (N), reflecting spatial different stages of disease progression. When exposed to inflammatory triggers such as the bacterial endotoxin LPS, IPF lungs do not fully recapitulate the typical inflammatory response observed in healthy lung tissue. However, especially upon such stimulation, levels of IL-6 and IL-8 are elevated in IPF lungs, potentially contributing to disease pathogenesis through their combined pro-inflammatory and pro-fibrotic effects. Untreated PCLS from all three IPF regions exhibit dysregulated metabolic pathways compared to the normal (N) group, with particularly pronounced suppression of fatty acid and acyl-CoA biosynthesis, cholesterol metabolism, and the peroxisome proliferator-activated receptors (PPARs) pathway. The silencing of important enzymes of the de novo synthesis of fatty acids, including fatty acid synthase (FASN), has been reported in type II AECs form IPF lungs [43]. A prospective study in fibrotic ILD patients has shown a positive correlation between higher levels of omega-3 fatty acids and higher DLCO [44]. Additionally, the suppression of these metabolic pathways has been related to loss of surface tension in the alveoli in experimental chronic lung disease models [45], as fatty acids are essential for surfactant protein production [46]. Consistently, we found downregulated gene expression for various surfactant proteins in PCLS from all three IPF regions (SFTPD, SFTPC, SFTPA1, SFTPB) (data not shown) compared to N lung tissue. Additionally, genes encoding enzymes responsible for the aminoacylation of tRNA with their corresponding amino acids, as well as genes encoding ribosomal proteins, were downregulated in all IPF PCLS compared to the N group, consistent with previous scRNA-seq observations [47]. New reports in human lung fibroblasts demonstrate the overproduction of ECM proteins may not be affected by the suppression of ribosomal protein genes but may rather generate protein product instability [48], which might parallel dysregulated protein renewal pathways in untreated F-IPF PCLS. Altogether, our findings confirm an increased endoplasmic reticulum (ER) stress specifically in fibrotic areas, together with poor protein clearance. The hallmark Inflammatory Response and inflammatory-related mechanisms, such as IL-6/JAK/STAT3, IL-2/STAT5, or TNF signalling were overexpressed in untreated NF- and T-IPF compared to normal tissue. Our study underscores the potential contribution of steady inflammation to IPF development, further supporting longstanding evidence that implicates inflammation as a driver of lung fibrosis [49–52]. Despite previous unsuccessful attempts with immunosuppressive therapies, emerging clinical trials involving agents with both anti-fibrotic and immunomodulatory properties are demonstrating promising benefits in patients with IPF and progressive pulmonary fibrosis (PPF) [17, 18]. PDE4B is the predominant PDE subtype in monocytes and neutrophils, orchestrating LPS-induced inflammation in these cells through NF-κB phosphorylation via cyclic adenosine monophosphate (cAMP)/protein kinase A (PKA) [53]. Due to its cAMP degrading function, PDE4B has been linked to pro-fibrotic mechanism in lung fibroblasts, including proliferation and myofibroblasts differentiation, by promoting the canonical TGF-β1 signalling [54]. The anti-fibrotic potential of PDE4 inhibitors has been attributed to the restoration of endogenous anti-fibrotic mediators that signal through cAMP, thereby reducing chemokine release and the pro-fibrotic phenotype in preclinical models [55]. These agents have also shown superior therapeutic effects compared to glucocorticoids in a mouse model of lung fibrosis [56]. Recent phase 3 clinical trials demonstrated beneficial effects on lung function by selectively inhibiting PDE4B in patients with IPF and PPF [17, 18]. Consistent with these findings, our results confirm the upregulation of PDE4B in IPF lung tissue, particularly in the transitional area (data not shown) containing fibroblast foci which are the main site of ECM production [57]. Untreated NF-IPF and T-IPF demonstrated activation of the LPS-induced immune response and pattern recognition receptor (PRR) signalling. Our results showed that the gene for TLR4, the main receptor for LPS, was upregulated in NF- and T-IPF compared to N group. Although the role of TLR4 in IPF remains controversial, activation of this pathway reveals distinct TGF-β1–induced responses in normal lung fibroblasts compared to those from IPF patients [58]. Emerging evidence indicates that increased soluble TLR4 linked to activated monocytes may signal a higher risk of AE-IPF [59]. Moreover, TLR4 + hybrid M1/M2 monocytes have been reported in systemic sclerosis (SSc) patients with ILD [60]. Genes facilitating INF-γ signalling and TLR-mediated viral response were upregulated in untreated NF- and T-IPF PCLS in our study, but INF-γ and α were entirely repressed in F-IPF PCLS, in agreement with previous reports [61, 62]. Although circulating INF-γ is reduced in IPF patients [63], signalling proteins in type II INF signalling such as JAK2 also mediate myofibroblast activation upon TGF-β activation, and inhibition attenuates experimental lung fibrosis [64]. Despite of the important anti-fibrotic effects of INF observed in preclinical assays [15], INF-γ−1b supplementation was unsuccessful in clinical trials [16]. These findings suggest that the crosstalk between components of the JAK/STAT signalling pathway and pro-fibrotic downstream effectors could constitute critical points of intervention. A similar controversy, along with the expression pattern ‘NF/T-IPF >N >F-IPF’, characterizes the state of TNF signalling in our dataset. Previous reports in spatial transcriptomic (ST) RNA-seq of formalin-fixed paraffin embedded (FFPE) IPF lungs showed downregulation of TNF signalling via NF-κB compared to healthy controls [62], while the levels of TNF in serum of IPF patients and healthy adults showed no significant differences [65]. Meanwhile, our results indicate upregulation of genes associated to signalling proteins and chemokines resulting from TNF stimulation, previously linked with fibrosis pathogenesis [66] and associated with worse prognosis in IPF [67]. Interestingly, lung fibroblasts have been shown to retain short-term inflammatory memory after secondary TNF treatment, which may drive persistent transcription of inflammatory mRNAs [68]. These finding might explain our observations in upregulated NF-κB-related proteins and interferon-regulatory factors (IRFs) genes in NF- and T-IPF PCLS without specific TNF stimulation. If the steady inflammatory gene expression might lead to an increased inflammatory response in IPF lungs is unclear. While LPS stimulation in murine and human PCLS has been shown to reliably trigger cytokine-mediated inflammation [25–27, 68], variations in this response may emerge under disease-specific conditions. This possible incomplete response in IPF PCLS combines two resulting effects in our dataset: inflammatory mechanisms that remain downregulated in IPF with or without LPS, and mechanisms previously expressed that are not fully activated after LPS stimulation. Despite persistent expression of pro-inflammatory genes, LPS does not amplify the inflammatory response in IPF PCLS possible due to partial NF-κB/TNF pathway activation and stromal cell desensitization, resulting instead in selective cytokine release (e.g., IL-6, IL-8) that may potentially contribute to fibrosis progression. For instance, our findings demonstrate that F-IPF tissue not only has impaired type I INF signalling, but also cannot activate this pathway after LPS stimulation similar to N tissue. Only a few genes associated with type II INF signalling, such as VCAM1, CMKLR1, or CFH, whose proteins have been identified as biomarkers of lung fibrosis [69–71], are highly expressed with or without LPS in NF- and T-IPF PCLS compared to normal tissue. It has been described that INF-α has anti-inflammatory properties, as demonstrated in IFNAR1-/- mice challenged with LPS from P. aeruginosa, suggesting that the lack of this mechanism can lead to an exacerbated inflammatory response, manifested in myeloid cell and neutrophil recruitment [72]. Notably, transcriptomic profiles of peripheral blood mononuclear cells (PBMCs) from AE-IPF patients showed a significant downregulation of IFN-regulated signature genes compared to stable IPF patients [73]. Concurrent with reduced INF, which results in a compromised antiviral defence, genes related to T-cell mediated immunity and activation are also downregulated in IPF PCLS with or without LPS compared to N PCLS. In IPF, multiple studies on T cells suggest that their roles are highly context-dependent, varying with both the disease stage and the specific T cell subtype involved [74]. In contrasts, some inflammatory signals were activated in untreated conditions but failed to exhibit a full response in IPF PCLS after LPS treatment, including TLR signalling and related cytokines. Only TLR5 and its signalling genes were upregulated in F-IPF, while in TLR4 was significantly increased in NF- and T-IPF after LPS stimulation. Partial TNF activation via NF-κB shows a comparable pattern in IPF PCLS: Genes encoding chemokines, previously upregulated in untreated conditions, are not activated in LPS-treated T-IPF. Other group of genes associated with membrane proteins of the TNF superfamily were found downregulated in all IPF PCLS with or without LPS, possibly associated with the reduction in T cell activity [75]. Only few genes associated to transcriptomic factors such as EGR3, NR4A1 and SGK1, linked to TGF-β signalling [66, 76, 77] maintained their upregulated expression profile following LPS stimulation. This partial activation suggests two inflammatory phenotypes in lung tissue depending on their origin: a complete TNFR1-induced NF-κB signalling, with higher TNF releases observed only in LPS-treated N tissues; and another centred-on IL-6/IL-8 signalling, observed in all PCLS after LPS treatment. Given that TNF contributes to the transcriptional regulation of IL-6 [78], it is quite intriguing how the production of this cytokine is considerably higher in T- and F-IPF PCLS compared to N tissues. TLRs can drive IL-6 production via the activation of inducible inhibitor of NF-κB zeta (IκBζ) [79], an enhancer of NF-κB secondary responses to genes encoding inflammatory mediators such as IL6 [80]. In addition, some authors have shown that the constitutive expression of NFKBIZ, encoding IκBζ, in type II AECs mediates the secretion of IL-6 after the fibrotic stimulus of TGF-β1 [62]. Hence, considering these findings, we suggest that the contrast in IL6 gene expression and IL-6 release after LPS stimulation in N group and IPF PCLS in our results could be associated to the intervention of TNF-independent mechanisms related to the pre-existing fibrotic environment. In mesenchymal cells, IL-6 production is enhanced by constitutive activation of STAT3, a transcription factor that mediates both pro- and anti-inflammatory signals [81]; likely through its role in the retaining inflammatory memory following secondary inflammatory stimuli [68]. Growth factors can also induce IL-6 secretion in stromal cells via Mitogen-Activated Protein Kinase (MAPK) pathway independently of NF-κB [82]. Of note, MAPK signalling controls post-transcriptional regulation by influencing cytokine mRNA stability [83]. These results underline the potential of IL-6 and IL-8 as biomarker for IPF progression or as therapeutic target. Although IL-6 levels are elevated in AE-IPF patients and associated with worse prognosis [84, 85], the inhibition of anti-IL-6 receptor (IL-6R) showed inconclusive results [86]. Detailed studies are crucial to evaluate JAK/STAT and MAPK signalling as potential therapeutic pathways for modulating IL-6 and IL-8 release in IPF. Moreover, the localized activation of inflammatory genes in IPF PCLS suggests that anti-inflammatory interventions may benefit from region-specific targeting. Studies in chronic obstructive pulmonary disease (COPD) have shown that inhaled PDE4 inhibitor possess a favourable safety profile without side effects previously observed with systemic administration [87]. Aerosol delivery is also being explored as an alternative route for administering current standard-of-care therapies for IPF [88]. It is important to acknowledge certain limitations inherent to this study. First, our transcriptomic analysis does not consider the individual contribution of each cell type. Despite each isolated cell have particular ways to manifest inflammation [89], studies using co-culture systems have shown that intercellular interactions play a critical role in amplifying cytokine release upon inflammatory challenge [90]. We believe that our PCLS findings reflect the importance of this cross-talk, even though further analysis will be needed to determine precise signalling components suitable for immune or fibrotic pathway modulation. Second, a loss of diversity in IPF lung tissue and the emergence of new fibrotic niches have been reported [4]. Thus, we acknowledge that cell composition differs among IPF regions, and the proportion of different cell types might have influenced our final transcriptomic profile. Third, PCLS model limits our findings to the primary response against potential respiratory pathogens from resident cells. AMs play a central role in host defence mechanisms in the lung [91]. Additionally, type I AECs [90], lung fibroblasts [92] and endothelial cells [93] can actively contribute to the innate defence process as well, as they present PRRs, are able to perceive cell damage and initiate an immune response. Our PCLS model recapitulates key mechanisms involved in the first line of defence in normal lung tissue and reveals functional impairments across IPF regions. Finally, we acknowledge that our “normal” lung tissue may be influenced by the surrounding tumorigenic tissue, and due to limited availability of IPF tissue, sample size is low, although statistical differences could nevertheless be detected for the presented readouts. Conclusions To sum up, our findings demonstrate that IPF lung tissue exhibits a steady inflammatory profile compared to normal lung tissue, while showing spatial heterogeneity. However, upon LPS stimulation, the inflammatory response in IPF lungs shows only partial activation of inflammatory genes compared to normal lung, reflected in impaired antiviral defence mechanisms, and reduced T cell activation, leading to the expression of cytokines that exert both pro-inflammatory and pro-fibrotic effects. Despite the upregulation of inflammatory genes in IPF tissue, subsequent LPS exposure fails to elicit a heightened inflammatory response. However, signal transducers and transcription factors of activated inflammatory pathways in IPF might contribute to, or facilitate, fibrotic progression. Thus, targeting the pre-activated inflammatory signalling pathways in IPF may represent a novel therapeutic approach to attenuate disease progression. Supplementary Information Supplementary Material 1. Acknowledgements The authors would like to acknowledge Pamela Nicholson form the Next Generation Sequencing Platform at the University of Bern for her support with the Bulk RNA-sequencing, and Heidi Tschanz-Lischer from the Interfaculty Bioinformatics Unit (IBU) at the University of Bern, for her support with the data analysis. Abbreviations Alveolar epithelial cells Acute exacerbation Alveolar macrophage Bronchoalveolar lavage Bone morphogenetic protein Cyclic adenosine monophosphate Chemokine (C-C motif) ligand Chronic obstructive pulmonary disease Chemokine (C-X-C motif) ligand Differentially expressed genes Differential gene expression Diffusing capacity for carbon monoxide Dulbecco’s modified Eagle Medium Extracellular matrix Epithelial–mesenchymal transition Endoplasmic reticulum Extracellular signal-regulated kinases Fatty acid synthase Foetal bovine serum Formalin-fixed paraffin embedded Fibrotic IPF area Gene ontology G protein-coupled receptor Gene set enrichment analysis NF-κB inhibitor zeta Interleukin Interstitial lung diseases Interferon alpha Interferon gamma Idiopathic pulmonary fibrosis Interferon-regulatory factor Insulin-transferrin-sodium selenite Janus kinases Kyoto encyclopedia of genes and genomes Lipopolysaccharides Mitogen-activated protein kinase The molecular signatures database Myeloid differentiation primary response 88 Normal lung tissue Neutrophil extracellular traps Non-fibrotic IPF area Nuclear factor kappa B NOD-like receptor Peripheral blood mononuclear cells Principal component analysis Precision-cut lung slices Phosphodiesterase 4 Phosphoinositide 3-kinase Protein kinase A Proliferator-activated receptor Progressive pulmonary fibrosis Pattern recognition receptor RNA sequencing Single-cell RNA sequencing Spatial transcriptomic S100 calcium-binding protein A12 Systemic sclerosis Signal transducer and activator of transcription Transforming growth factor beta 1 Transition IPF area Toll-like receptors Tumour necrosis factor Unfolded protein response Weighted gene co-expression network analysis Authors’ contributions MFC and CM designed the experimental study; TMM, and PD contributed to normal lung tissue and clinical data acquisition; LN and CPM contributed to IPF lung tissue and clinical data acquisition; CM and MFC contributed to analysis, interpretation of the data and drafting of the manuscript; NK contribute to the interpretation of the data and drafting the manuscript. All the authors critically revised and approved the final version of the manuscript. Funding Bangerter Foundation (Switzerland) to MFC. Data availability The datasets used and/or analysed during the current study are available from the corresponding author upon request. Declarations Ethics approval and consent to participate The study and experimentation in human control lung tissue were approved by the local Bernese Ethics Commission Bern, Switzerland (approval number: KEK-BE_2024 − 01841). IPF lung tissue collection was approved by the local ethics committee at Hanover Medical School, Germany (8867_BO_K_2020). Human samples were collected and handle in accordance with the Declaration of Helsinki, and donors were informed prior to recruitment, and signed a written consent form. Consent for publication NA. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement The datasets used and/or analysed during the current study are available from the corresponding author upon request.