IMR-90 normal human lung fibroblasts, a commercially available primary line of lung fibroblasts isolated from a clinically normal 16-week female fetus (cat no: I90-83; RRID : CVCL_0347), were purchased from Coriell Institute. Fibroblasts at passage numbers 5-9 were used in this study. Cell authentication of fibroblasts lines were verified by morphology and examination of type I collagen and (α-SMA) as described previously (37) by our lab and Coriell Institute. Fibroblasts were cultured in DMEM (cat no: 15-013-CM; Corning) supplemented with 10% fetal bovine serum (FBS) (cat no: FP-0500-A; Atlas), 1X Penicillin-Streptomycin (cat no: 30-002-CI; Corning), and 1X GlutaMAX (cat no: 35050-061; Thermo Fisher Scientific) in a CO₂ (5%) incubator at 37°C.

Primary human IPF and non-IPF fibroblasts/tissues were obtained from University of Alabama at Birmingham (UAB) Tissue Procurement Facility, derived from de-identified lung tissues, which protocol was approved by the UAB Institutional Review Board and conformed to the Declaration of Helsinki. The diagnosis of IPF was made by a multidisciplinary approach according to American Thoracic Society/European Respiratory Society guidelines (38). Primary HLFs were cultured following the same protocol as IMR-90s.

Aged 18-month-old (72 weeks old or greater) C57BL/6J (RRID#IMSR_JAX:000664) male mice were obtained from the Jackson Laboratories (Bar Harbor, Maine, USA). Male mice were used for the study since they develop a more severe fibrotic response from a single-dose of bleomycin compared to female mice (39). Animals were housed under pathogen-free conditions with food and water ad libitum. Littermates of the same sex were randomly assigned to experimental groups. All experiments and procedures were approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham (Birmingham, AL).

Drosophila melanogaster stocks were raised with a standard regular diet: agar 11 g/L, active dry yeast 30 g/L, yellow cornmeal 55g/L, molasses 72mL/L, 10% nipagen (in ethanol), propionic acid 6mL/L. To evaluate OGT regulation of pericardin, transgenic stocks were obtained from Bloomington Drosophila Stock Center (BDSC) and Vienna Drosophila Resource Center (VDRC): w¹¹¹⁸ (BDSC: 3605), UAS-Sxc (BDSC: 83725), UAS-Sxc-RNAi (BDSC: 82460), control RNAi (BDSC: 36303), and UAS-GFP (BDSC: 5431), r4-Gal4 (BDSC: 33832), and Ubi-Gal4 (BDSC: 32551). Act88F-Gal4 was obtained from the laboratory of Richard Cripps (40). Flies were housed at 25 °C, 50% humidity in a 12 h light/12 h dark cycle. Flies were transferred onto fresh media every 3 days.

E. coli strain DH5α (cat no: C2987H; New England Biolabs), Turbo (cat no: C2984H; New England Biolabs) and BL21 (DE3) (cat no: C2527H; New England Biolabs) were used for cloning and protein expression, respectively. Wild-type E. coli isolates were grown at 37°C in Luria Broth (LB) (cat no: IB49020; IBI Scientific) medium. Engineered strains with pET-Duet-hOGT-hSMAD3 were grown in LB plus ampicillin (100 mg/mL) at 37°C. Protein expression was induced with 0.2 mM isopropyl β-D-1thiogalactopyranoside (IPTG) (cat no: 34060; Thermo Fisher Scientific) overnight at 25°C in a shaker rotating at 250 rpm.

Following a single dose of normal saline or bleomycin administration between day 19 and 21, mice were given 8 doses of sham RNA (siCTRL) or siOGT (Horizon Discovery) diluted in 1X PBS every other day via oropharyngeal instillation to randomized groups until euthanized at day 40. Mice were anesthetized for each siRNA treatment as described above. Post treatment care was carried out as described above.

Lungs were separated by lobes into RINO tubes (cat no: NAVYR1-RNA; Next Advance) and filled with ice cold milliQ water. Lungs were then homogenized using a bullet blender (Bullet Blender 24; Next Advance) at max speed for 1 minute at 4°C followed by placing samples on ice for a 2-minute cool down, then repeat the process for 2-4 more cycles. Samples were centrifuged and inhibitors were added to maintain sample integrity: 1X Halt (1:100, cat no: 78442, Thermo Fisher Scientific), 1X Z-PUGNAc (1:1000, cat no: 3384, Tocris), 1X Thiamet G (1:1000, cat no: 13237; Cayman Chemical) and 1X Benzonase (1:1000, E1014; Millapore Sigma). Homogenized lungs were split for protein, RNA, and hydroxyproline assays.

Histological procedures were performed as previously described (43). Briefly, mice were euthanized and lungs were perfused with 5 mL of PBS. Lungs were then inflated and fixed in 10% neutral-buffered formalin (cat no: 57225, Thermo Fisher Scientific) for 24 hours and then switched to 70% ethanol for dehydration and storage. The UAB Histology core processed, sectioned, and stained the lung tissue sections. Lung tissue sections were stained for Masson’s trichrome to detect collagen and O-GlcNAc (RL2; cat no: NB300-524; Novus Biologicals). Images were taken at 4x and 20x magnification on a Nikon ECLIPSE Ts2 Inverted microscope with an image analysis software (NIS-Elements D).

We crossed Gal4 virgin females to RNAi/GFP/w1118/Sxc-RNAi/UAS-Sxc males to obtain F1 adult flies for experiments. F1 adult flies were collected upon exclusion and then separated into male and female flies in groups of 25-30 on day 3. Flies were transferred onto fresh media every 3 days. All experiments were performed on 3-week-old female flies. Gal4 drivers used include Act88F-Gal4 (40) which drives largely in indirect flight muscle (IFM) from 24h after puparium formation (APF), r4-Gal4, which drives expression in the fat body from late embryo through adult stages (44), and Ubi-Gal4, which is a ubiquitous driver (45).

Total collagen was measured using a hydroxyproline assay kit (cat no: QZBTOTCOL1; QuickZyme Biosciences), and the experiment was performed according to the manufacturer’s instructions. Briefly, mouse lung homogenates were weighed and transferred into screw capped tubes and an equal volume of 12M HCl (cat no: 258148-500ML-GL; Millipore Sigma) was added (final concentration is 6M HCl). Samples were incubated for 20 hours at 95°C.The following day, samples were cooled down to room temperature and centrifuged for 10 minutes at 13,000 x g and supernatant was transferred to a new tube. Hydrolyzed samples were diluted with water down to a final concentration of 4M HCl. Samples were loaded on a 96-well plate with assay buffer and incubated at room temperature for 20 minutes on a plate shaker. Detection reagent was added to each well and the plate was briefly mixed and placed in a 60°C oven for 60 minutes. Plate was cooled on ice for 5 minutes and hydroxyproline concentration was measured at 570 nm using a BioTek plate reader (Agilent, United States).

Recombinant proteins from E. coli cells were size filtered through Amicon^® Ultra-4 Centrifugal Filter unit (cat no: UFC8050; Millipore Sigma). Additionally, this step serves as a buffer exchange to store proteins in 20 mM Tris HCl pH 8, 150 mM NaCl, 1 mM dithiothreitol (DTT). Following size exclusion purification, recombinant proteins were subjected to S-protein Agarose (cat no: 697043; Millipore Sigma) affinity purification to capture recombinant human (rh) Smad3. S-tagged protein purification was performed according to the manufacturer’s instructions using Disposable Plastic Columns (cat no: 29922; Thermo Fisher Scientific). Briefly, S-protein Agarose was added to recombinant proteins and allowed to bind overnight at 4°C. rhSmad3 was eluted using 3 M MgCl₂ and followed by buffer exchange to store purified rhSmad3.

O-GlcNAc enzymatic labeling and detection or azide modified protein capturing was performed as described in the manufacturer’s protocol (cat no: C33368, C33372, and C10416; Thermo Fisher Scientific). Briefly, 200 ug of proteins isolated from human lung fibroblasts were precipitated using the chloroform/methanol method. Proteins were then enzymatically labeled with or without the Gal-T1 (Y289L) overnight at 4°C. Following overnight labeling, the samples were precipitated again using the chloroform/methanol method and azide-labeled proteins underwent Click-iT biotin labeling for western blot applications or alkyne agarose bead enrichment for mass spectrometry analysis.

Statistical analysis description for each experiment can be found in the figure legends. Western blot images were quantified using FIJI and analyzed relative to β-actin. Each qPCR experiment was performed in duplicates or triplicates and analyzed for relative expression by calculating the 2^-ΔCT relative to GAPDH expression. Immunofluorescence images were quantified by measuring signal intensity from select regions from each photo. Statistical analyses were performed in Prism 10 (GraphPad). Error bars represent the mean ± standard error of the mean (SEM). Outliers were determined using the 1.5xIQR rule. A classic student t-test or one-way ANOVA with Tukey’s multiple comparison test was performed to assess significance. Statistical significance was always represented as follows: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

A previous genomic study reported increased expression of OGT in IPF human lung tissues compared to lung tissues from healthy controls (48); however, no other characterization of OGT or other proteins involved in the HBP were investigated. Moreover, the specific cell types demonstrating increased OGT expression were not identified. To determine the relevance of the HBP and O-GlcNAc in IPF, we analyzed a previously published dataset containing human lung cell scRNA-seq (single-cell RNA-sequencing) data (GSE147066) from IPF (n=32) and non-IPF donors (n=29) for genes in the HBP ( Figure 1A ) (41). Cell populations were stratified based on expression of CTHRC1 (collagen triple helix repeat containing 1) by the original authors and four subpopulations were identified ( Figure 1B ). Based on the spatial mapping, the fibroblast and myofibroblast cell populations had higher expression of collagens, COL1A1 and COL3A1, compared to the pericyte and smooth muscle cell populations ( Supplementary Figure 1 ). In addition, spatial cluster maps of GFPT1 (GFAT; the rate-limiting step in the HBP) (49), OGT and MGEA5 (gene encoding OGA) showed more expression in the selected IPF fibroblast and myofibroblast cell populations that had increased collagen expression (COL1A1 and COL3A1) and other fibrotic markers (ACTA2 and TGFB1) compared to the non-IPF controls ( Figure 1C ). Consistent with the public scRNA-seq data, our mRNA expressions of OGT, MGEA5, GFPT1, COL1A1, COL3A1, ACTA2, and TGFB1 from IPF human lung fibroblasts (HLFs) were all significantly increased compared to non-IPF donors ( Figures 1D-J ). Collectively, these data indicate that increased expression of HBP markers (OGT, MGEA5, and GFAT1) is observed in IPF fibroblast and myofibroblast cell populations with augmented collagen expression.

To determine the levels of the O-GlcNAc modification in IPF lung tissue, non-IPF and IPF lung tissues were subjected to IHC staining with antisera against O-GlcNAc ( Figure 2A ). Increased HRP staining of O-GlcNAc was observed in fibrotic regions of the parenchyma in IPF lung tissue compared to non-IPF controls. Increased O-GlcNAc staining was also observed following bleomycin-induced lung fibrosis in mice compared to saline controls (PBS), a shared feature with our findings in the human tissue ( Figure 2B ). These data indicate that global O-GlcNAc levels are increased in fibrotic regions of lung parenchyma in both IPF and an experimental model of pulmonary fibrosis.

To investigate the effect of gain- and loss-of-function of OGT on collagen accumulation, we performed overexpression (OE)/knockdown (KD) of the Drosophila homolog of Ogt, Sxc (referred to as dOgt), in Drosophila melanogaster and examined the consequences on a type IV collagen-like protein, pericardin, expression. Since the flies do not have lungs, we examined the expression of pericardin in the fat bodies of Drosophila, which are liver- and adipose-like tissue in the abdomen, to objectively determine the role of OGT on collagen accumulation in vivo (50, 51). We conducted whole-body (Ubi-Gal4), indirect flight muscle (Act88F-Gal4), and fat body-specific (r4-Gal4) KD or OE of dOgt ( Supplementary Figure 2A ; Figure 3A ). Global O-GlcNAc levels and dOGT protein expression from whole-body and thorax upon respective manipulations showed that dOgt KD or OE can reduce or increase O-GlcNAc and dOGT expression, respectively ( Supplementary Figures 2B-G ). Furthermore, confocal microscopy imaging demonstrated that pericardin deposition in the fat bodies was significantly decreased following dOgt KD, but significantly increased upon dOgt OE compared to the control flies ( Figures 3B, C ). Altogether, these data suggest that OGT regulates collagen-like protein accumulation in vivo, and may be a critical component of the ECM remodeling and fibrotic process in higher order species.

Since we observed increased O-GlcNAc in human and mouse pulmonary fibrosis and regulation of collagen expression by OGT in D. melanogaster, we wanted to determine the effects of reducing O-GlcNAc on fibrosis resolution. Aged mice (18 months), an established model for non-resolving pulmonary fibrosis (52), were administered 8 doses of OGT siRNA (siOGT), via oropharyngeal instillation, to knockdown OGT following peak fibrosis or 3 weeks post-bleomycin treatment ( Figure 4A ) (53). Collagen deposition/fibrosis 6-weeks post bleomycin (BLEO+siCTRL) was increased compared to controls (siCTRL). Strikingly, oropharyngeal-delivered KD of OGT post-peak fibrosis resulted in significant reduction in collagen deposition as determined by Masson’s trichrome staining and quantification of lung collagen content by measurement of total hydroxyproline from acid hydrolyzed lung homogenate ( Figures 4B, C ). Immunoblots for pan collagen, O-GlcNAc, OGT, phosphorylated Smad3 (p-Smad3), and total Smad3 validated our findings and fibrotic markers were elevated in the BLEO+siCTRL group compared to the BLEO+siOGT group, which demonstrated similar levels to baseline (siCTRL) ( Figures 4D-H ). Transcript analyses of COL1A1, COL3A1, OGT, and ACTA2 showed a similar upregulation trend with BLEO+siCTRL treatment and downregulation of each gene after siOGT treatment ( Figures 4I-L ). Transcript analyses of the collagenases, MMP2 and MMP9, exhibited a contrasting trend to ACTA2 and collagen suggesting a suppression of MMP2,-9 expression during fibrosis that is upregulated following OGT knockdown ( Figures 4M, N ). Furthermore, TIMP3 (tissue inhibitor of metalloprotease 3) increased during fibrosis and was reduced following OGT KD ( Figure 4O ). These data indicate that OGT regulates collagen expression and fibrosis resolution in a pre-clinical model of pulmonary fibrosis.

The TGF-β pathway is a master regulator of fibrosis and the TGF-β1-induced fibroblast-to-myofibroblast transition (FMT) has been documented in fibrosis and used as an in vitro method to mechanistically study fibrogenesis (54, 55). To determine the role of O-GlcNAc in ECM remodeling and FMT, HLFs were treated with TGF-β1 in the presence and absence of OGT inhibitor (OSMI-4). TGF-β1 administration increased global O-GlcNAc levels in HLFs similar to the levels observed in primary IPF lung fibroblasts ( Figures 5A-D ). OGT inhibitors (OSMI-4, OSMI-1, Ac₄5S-GlcNAc) abrogated the effects of TGF-β1 induction on O-GlcNAc levels and upregulated OGT expression while concurrently downregulating OGA expression as a feedback response to the inhibition ( Figures 5C, D ; Supplementary Figures 3A-E, H, I ). Blocking O-GlcNAc transfer attenuated TGF-β1 induction of FMT as determined by the expression of pan collagen ( Figures 5C, E ), Col1α1 ( Figures 5C, F ; Supplementary Figure 3D, F, H, J ), Col3α1 ( Figures 5C, G ), and α-SMA ( Figures 5C, H ; Supplementary Figure 3D, G, H, K ), all of which were significantly reduced. Validation of these observations were shown by immunofluorescence staining of Col1α1 and α-SMA ( Figure 5I ). Conversely, the opposite result was observed following pharmacological inhibition of OGA [Thiamet G, (TMG); increases O-GlcNAc] where O-GlcNAc, Col1α1, Col3α1, and α-SMA protein expression were significantly increased compared to vehicle control, and congruous with TGF-β1 induction ( Figures 6A-E ). Furthermore, immunofluorescence images validate our prior observations ( Figure 6F ). Collectively, these data demonstrate that the OGT/O-GlcNAc axis regulates collagen (and α-SMA) expression in vitro and may be a key mediator in the FMT, a process central to pulmonary fibrosis.

Our data demonstrates a role for OGT and the O-GlcNAc modification in FMT and collagen expression/accumulation, and fibrosis resolution. However, specific proteins modified by O-GlcNAc have not been determined. Therefore, we subjected HLFs to Click-iT™ O-GlcNAc Enzymatic Labeling and Detection/Enrichment and mass spectrometry (MS) analysis ( Figures 7A, B ). MS analysis identified over 2,500 O-GlcNAc modified proteins ( Supplementary Table 1 ). Furthermore, several proteins enriched from the O-GlcNAc purification were associated with TGF-β signaling including TGF-β1 and α-SMA ( Table 1 ). Smad4 and Smad5, which have been documented as O-GlcNAc modified proteins (56, 57), were also identified in the analysis. A novel finding was that Smad3 contained the O-GlcNAc modification. To validate this finding, we inserted a SMAD3 gene into a pET-Duet bacterial plasmid (flanked by an affinity S-tag) containing human OGT. Co-expression of OGT and Smad3 in BL21 DE3 E. coli confirmed that Smad3 is modified by OGT by immunoprecipitation ( Figure 7C ). These findings are the first to demonstrate Smad3 as an O-GlcNAc modified protein, which may have implications in Smad activation and downstream targets associated with FMT and fibrosis.

TGF-β1 signaling through the Smad3 pathway is well documented in fibrosis (67–71). However, the regulation of Smad3 phosphorylation by O-GlcNAc modification has not been determined. Therefore, we treated HLFs with TGF-β1 in the presence or absence of OGT inhibition. Phosphorylation of Smad3 (p-Smad3; Ser-423/425) was significantly reduced following TGF-β1 administration in primary HLFs when OGT was inhibited ( Figures 7D, E ). Reciprocally, when OGA was blocked to increase O-GlcNAc levels, p-Smad3 levels were enhanced with TGF-β1 stimulation compared to TGF-β1 stimulation alone ( Figures 7F, G ). Taken together, these findings suggest that the O-GlcNAc modification of Smad3 enhances its phosphorylation and may be critical for control of downstream effectors regulated by TGF-β1/Smad3.