Fig 2: Secretory proteome profiling reveals novel regulatory proteins upon LPS administration. A Experimental workflow for human epithelial cell culture, secretome profiling, repair-related screen and lung-related screen. BEAS-2B cells were treated with LPS (0.1 ug/mL, n = 3) or medium alone (n = 3) for 24 hours in DMEM. B Gene Ontology (GO) enrichment of upregulated secreted proteins, prioritized by gene ratio (number of differentially expressed proteins associated with a GO term divided by the total proteins in that term) and FDR (Benjamini-Hochberg; p < 0.05). Top processes areshown. C Immunoblotting verifies secreted protein expression of RCN3, TNC, YKL-40 and BAT3 after LPS exposure (0.1 ug/mL for 24 hours) in BEAS-2B cells. CM: conditioned medium; CL: cell lysate. The ratios of CM/ (CL+CM) are presented in a scatter plot (n = 4, ** p < 0.01 versus control group). D RCN3 concentration in human BALF is measured by ELISA: organizing pneumonia (OP, n = 8), idiopathic pulmonary fibrosis (IPF, n = 18) and control (n = 9); *** p < 0.001

Fig 3: Full-length RCN3 is secreted via the ER-Golgi pathway in an N140-glycosylation-dependent manner. A Top panel: immunoblot shows RCN3 bands in cell lysate (CL) and conditioned medium (CM) from wildtype mouse embryonic fibroblasts (MEFs), but not from RCN3 knockout MEFs (KO). Bottom panel: RCN3 in CL and CM from BEAS-2B (2B) and human lung fibroblasts (LF, CP-H011). The tubulin expression and Ponceau S staining are shown as loading controls. B Double-tagged RCN3 (N-terminal Flag after the signal peptide; C-terminal Myc before HDEL) confirms secretion of full-length RCN3. Following transfection of 293T cells, Flag immunoprecipitates from CL and CM were blotted with anti-RCN3, anti-Myc, anti-Flag, and anti-HDEL antibodies. Intact Flag and Myc in CM indicate full-length secreted RCN3. C Flag immunoprecipitates of CM from flag-RCN3 transfected 293T cells were separated by SDS-PAGE. The PAGE gel was stained by Coomassie blue Staining (top panel) and the region in the red square was subjected to mass spectrometry (MS) analyses. The amino acid fragments of RCN3 captured by mass spectrometry are underlined (bottom panel). MS: mass spectrometry. D Deletion of HDEL increases RCN3 secretion. Quantified fractions of RCN3 are shown (mean ± SD (two-tailed t test, n = 4 independent biological replicates, ** p < 0.01). E Brefeldin A (BFA) inhibits RCN3 secretion and increases GRP78. GAPDH and Ponceau S are loading controls. F Immunofluorescence staining of Flag (green) and Calnexin (red, ER marker) or Golgi97 (red) in Hela cells transfected with flag-RCN3 wildtype or without HDEL. Flag-RCN3 without HDEL was barely detected in the cytoplasm. DAPI was used for nuclear staining, scale bar: 5um. G Schematic diagram of the glycosylation site of RCN3 (top panel). LC-MS/MS identified a fragment deamidated at N140 (NATYGHYAPGEEFHDVEDAETYK) from CL and CM RCN3; the MS/MS fragment ion spectrum of the deamidated N-glycopeptide is shown in bottom panel. H PNGase F treatment and mutation at the glycosylation site (N140Q, N140A and T142A) lower the apparent molecular mass and reduce secretion. Quantified fractions of RCN3 are shown (mean SD; two-tailed t test, n = 4 independent biological replicates, ** p < 0.01). PNGF: PNGase F. I Immunofluorescence staining of Flag (green) and Calnexin (red, ER marker) or Golgi97 (red) in Hela cells transfected with flag-RCN3 wildtype or with N140Q mutation. Flag-RCN3 with N140Q mutation largely remained in the cytoplasm and partially colocalized with Calnexin. DAPI is used for nuclear staining, scale bar: 5um. J Immunoblotting shows that N140Q mutant pregulated GRP78 expression. The ratios to GAPDH are presented in bar graphs as values relative to vector, and data presented as mean ± SD (two-tailed t test, n = 4, *p < 0.05 versus vector control)

Fig 4: Schematic summary of the proposed profibrotic effects of secretory RCN3. During the initiation phase of ALI, injured epithelial cells secrete RCN3 in a regulated manner; elevated alveolar RCN3 activates fibroblasts via engaging the TGFβR1–Smad3. Activated myofibroblast further release fibrotic mediators including TGFβ1, leading to increased excessive fibroblast activation and extracellular matrix (ECM) deposits. Alveolar RCN3 peaks earlier than TGFβ1 secretion post ALI, suggesting that RCN3 acts as an early epithelial signal triggering the imbalance toward pulmonary fibrosis, and targeting RCN3 in BALF could be a potential early-window therapeutic strategy for pulmonary fibrosis post-ALI

Fig 5: RCN3 activates fibroblasts via canonical TGF‑β signaling and engages TGFβR1. A RNA-Sequencing analysis of total RNA from CP-H011 treated with vehicle or 0.1 µg/mL rhRCN3 (n = 3). A total of 543 differentially-expressed genes, (DEGs) with the foldchange (FC) > 1.5 and FDR < 0.05 were identified. Hierarchical clustering presented the gene expression profiles segregated based on RCN3 treatment. B GO enrichment of these DEGs, prioritized by gene ratio and BH-corrected FDR (p < 0.05); top processes are shown. C The KEGG pathway analysis of the differentially expressed genes. D Immunoblot indicates that the phosphorylation of Smad3 was upregulated, while the phosphorylation of AKT, Erk1/2 and JNK remained unchanged. The ratios of phosphorylated/total proteins are presented in scatter plots as values relative to control (n = 4, **p < 0.01 versus vehicle). E The SBE (SMAD binding element) reporter assay validates that rhRCN3 activates the TGFβ signaling pathway (n = 5, ***p < 0.001). F Immunoprecipitation assay shows that his-RCN3 interacts with TGFβR1 but not TGFβR2. The His-immunoprecipitates and total lysate (input) were analyzed by immunoblot using anti-TGFβR1 and anti-TGFβR2 antibodies. G Immunoblot of Smad3 in the fibroblast cells treated with 0.1 µg/mL rhRCN3 in response to TGFβ receptor inhibitors, LY364947 and LY2109761. H Knockdown of TGFβR2 did not inhibit the phosphorylation of Smad3 induced by rhRCN3 (0.1 µg/mL). I Knockdown of TGFβR1 inhibited the phosphorylation of Smad3 induced by rhRCN3 (0.1 µg/mL). J Biolayer interferometry (BLI) studies using rhRCN3 and TGFβR1 proteins indicate the interaction between RCN3 and TGFβR1 exhibiting a clear concentration gradient with an affinity (KD) of 26.91nM. BLI binding representative of at least 4 independent traces. K The best potential direct interaction models of RCN3 with the extracellular domain of the TGFβR1 predicted by the protein-protein docking tool ClusPro server. The predicted structures of RCN3 obtained from the UniProt by AlphaFold model (AF-Q15910 and AF-Q96D15) and the solution structure of the extracellular domain of the TGFβR1 (PDB: 2L5S) were submitted to the ClusPro web-based server to analyze their interaction (http://cluspro.bu.edu/). The model shows that residues 60–70 and 80–93 of RCN3 fold and form a docking surface that inserts into the cavity of extracellular domain of TGFβR1. Abbreviations: DEG, differentially expressed gene; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; SBE, SMAD‑binding element; IP, immunoprecipitation; BLI, biolayer interferometry; FC, fold change; FDR, false discovery rate