Triggering AHR resolves TGF-β1 induced fibroblast activation and promotes AT1 cell regeneration in alveolar organoids

Hagan, Andrew S.; Williams, Scott; Mathison, Casey J. N.; Yan, Shanshan; Nguyen, Bao; Federe, Glenn C.; Kuzu, Guray; Siefert, Joseph C.; Hampton, Janice; Chichkov, Victor; Whitney Barnes, S.; King, Frederick J.; Taylor, Brandon L.; Walker, John R.; Zhao, Rui; Elliott, Jimmy; Phillips, Dean P.; Fang, Bin; Decker, Rebekah S.

doi:10.1038/s42003-025-08446-5

Download PDF

Article
Open access
Published: 09 July 2025

Triggering AHR resolves TGF-β1 induced fibroblast activation and promotes AT1 cell regeneration in alveolar organoids

Communications Biology volume 8, Article number: 1025 (2025) Cite this article

3288 Accesses
1 Altmetric
Metrics details

Subjects

Abstract

Regeneration of the alveolar epithelium is necessary to restore tissue architecture and gas exchange capabilities in chronic pulmonary diseases such as fibrosing interstitial lung disease. While it is known alveolar type 2 (AT2) cells give rise to alveolar type 1 (AT1) cells to repair the alveolar epithelium after injury, methods to promote this process under pathological settings are poorly understood. Here, using a complex 3D organoid culture with TGF-β1 dependent impaired AT1 spheroid formation, we performed a high-throughput screen (HTS) with ~16,800 compounds to identify small molecules that increase number of AT1 spheroids. Longitudinal single cell RNA sequencing (scRNA-seq) revealed that DB-11-BE87 increased AT1 regeneration by reducing TGF-β1 induced fibroblast activation, concurrently with AHR activation in those cells. These studies highlight a HTS system to identify factors that can promote AT1 differentiation and suggest AHR activation as a method to counteract pathological TGF-β1 signaling in pulmonary disease.

TRβ activation confers AT2-to-AT1 cell differentiation and anti-fibrosis during lung repair via KLF2 and CEBPA

Article Open access 07 October 2024

Opposing roles for TGFβ- and BMP-signaling during nascent alveolar differentiation in the developing human lung

Article Open access 09 September 2023

Autocrine TGF-β-positive feedback in profibrotic AT2-lineage cells plays a crucial role in non-inflammatory lung fibrogenesis

Article Open access 31 August 2023

Introduction

Efficient air-gas exchange is the fundamental function of the lung and occurs at the alveolar epithelium, the most distal end of the respiratory tree. The alveolar epithelium consists of the squamous alveolar type 1 (AT1) and the cuboidal alveolar type 2 (AT2) cells. Composing fewer absolute numbers than AT2s, AT1s comprise 95% of the alveolar surface due to their elongated morphology¹. During homeostasis and injury AT2 cells self-renew and give rise to differentiated AT1 cells to maintain the alveolar epithelium². Recent work has uncovered transitional cell states along the AT2 to AT1 differentiation trajectory that accumulate during models of pulmonary fibrosis and may have a pathogenic role in disease^3,4,5. Published single cell atlases of idiopathic pulmonary fibrosis (IPF) report a decrease in both AT2 and AT1 cells, confirming a depleted alveolar epithelium in diseased lung tissue^6,7. Increasing AT2 to AT1 differentiation and limiting the accumulation of transitional cell states could provide benefit to patients with fibrosing interstitial lung disease.

Alveolar niche fibroblasts populations under homeostasis support AT2 progenitor cell self-renewal⁸. Conversion of homeostatic fibroblasts to myofibroblasts marks progression of lung fibrosis and is a major contributor of excess extracellular matrix (ECM) deposition in fibrotic lung disease^9,10. Intermediate fibroblast to myofibroblast cell states are non-invasive, however they may respond to TGF-β1 stimulation to produce an invasive myofibroblast subpopulation with high ECM expression in the fibrotic lung^11,12. TGF-β1 stimulation also inhibits alveolar spheroid formation through fibroblast activation suggesting a decreased ability to support lung epithelial repair¹³. Decreasing pathogenic fibroblasts subtypes could improve epithelial niche support.

In the mouse, AT1s are marked by Hopx expression^14,15. For this study, we developed a high-content imaging (HCI) assay to quantify AT1-containing spheroids (Hopx-GFP+) derived from a Hopx-GFP mouse lung homogenate 3D organoid culture containing mesenchymal cells. To induce a fibrotic disease-like state in these cells, TGF-β1 was added to the culture and resulted in impaired AT1 spheroid formation. To identify compounds which promote Hopx-GFP spheroid formation and revert the TGF-β1 induced phenotype, we performed a high throughput screen (HTS) of 16,800 unique small molecules. We identified LA-13-ZS70 as a small molecule capable of promoting Hopx-GFP spheroid formation in the presence of TGF-β1. Single-cell RNA sequencing (scRNA-seq) of organoids treated with DB-11-BE87, a medicinal chemistry refinement of LA-13-ZS70, revealed compound dependent promotion of epithelial organoid formation with decreased aberrant transitional AT2-AT1 epithelial cells. Differentially expressed genes (DEGs) and cell-cell interaction analyses suggested aryl hydrocarbon receptor (AHR) in the mesenchyme as a likely target of DB-11-BE87. This was further supported by activation of the AHR pathway in reporter assays. DB-11-BE87 treatment reduced TGF-β1 induced fibroblast to myofibroblast transition, decreasing myofibroblasts and increasing fibroblast transitional states. This work suggests AHR activation in the pulmonary mesenchyme resolves TGF-β1 induced fibroblast activation and is a beneficial strategy to reduce fibroblast to myofibroblast transition and increase beneficial epithelial niche support.

Results

Development of a murine Hopx-GFP ^± assay

To develop a HCI assay for HTS, the alveolar organoid assay initially described by Barkauskas et al. was modified². Briefly, all distal lung cells from Hopx-GFP mice were isolated and seeded in optically clear 384-well plates in a mixture of culture medium and 50% Matrigel. After Matrigel solidification, media was added and replaced on day 4 or day 5. At endpoint (~day 12), spheroid cultures were imaged live to measure Hopx-GFP+ spheroids (Fig. 1A). Addition of TGF-β1 impaired spheroid formation at highest dose, while addition of galunisertib, a TGF-βR1 inhibitor, eliminated this effect and increased spheroid number (Fig. 1B, C). GFP signal imaged and quantified was specific to Hopx-GFP mouse. Galunisertib promoted spheroid formation in the absence of exogenous TGF-β1 suggesting endogenous TGFβ signaling in culture conditions. Gene expression analysis of canonical AT1 markers (Hopx, Ager, Aqp5) demonstrated concordance with GFP signal (Fig. 1D). Interestingly the Hopx-GFP allele behaves as a knockout allele with roughly 50% loss of expression of Hopx compared to C57BL/6 controls, however this did not impact other canonical AT1 markers (Ager, Aqp5). To identify molecules that reverted the impaired AT1 spheroid formation, a screen of ~16,800 compounds was performed on the alveolar spheroid assay in the presence of TGF-β1 (Fig. 1E)¹⁶. Positive (30 µM galunisertib, Fig. 1E, red) and negative (empty well, Fig. 1E, blue) controls displayed distinct effects on spheroid formation. 100 hits were called by being >3 standard deviations from median values by both Hopx-GFP+ spheroid number and total integrated signal intensity (Fig. 1E, green). LA-13-ZS70 (Fig. 1E, star) was selected as a candidate for additional validation due to attractive chemical properties and efficacy in other phenotypic assays. Its activity was indicated by concentration-response curve (EC₅₀ = 376 nM, count of Hopx-GFP+ spheroids) (Fig. 2A, B).

**Fig. 1: High-throughput screening on multi-cellular murine alveolosphere assay identifies small molecules that rescue TGFβ impaired alveolar type 1 spheroid formation.**

**Fig. 2: DB-11-BE87 increases AT1-AT2 differentiation with partial dependency on TGFB pathway activation.**

To enable a highly multiplexed gene expression readout of the alveolar organoids, we developed a lung-specific RNA-mediated oligonucleotide annealing, selection, and ligation with next-generation sequencing (RASL-seq) gene expression panel¹⁷. Literature mining was conducted to assemble canonical cell lineage markers for inclusion on the panel. Additionally, we performed bulk RNA sequencing to determine genes enriched in sorted AT1 and AT2 cells from genetically labeled animals. Sftpc^CreER/+; ROSA26^tdTomato/+; Hopx-GFP lungs were digested and sorted into AT1 fractions (Hopx-GFP+, Sftpc-tdTomato-, CD31-, CD45-) vs. AT2 fractions (Hopx-GFP-, Sftpc-tdTomato+, CD31-, CD45-) (Supplementary Fig. 1A). Probes for AT1 and AT2 specific markers were included into RASL-seq panel with canonical cell lineage markers to enable transcriptional profiling of compound effects (Supplementary Fig. 1B–Dand Supplementary Data 1). Higher expression of AT1 lineage markers, including Ager, Aqp5, and Gprc5a, and airway signaling pathway genes were observed upon treatment with LA-13-ZS70 compared to the DMSO control, confirming the improved alveolar organoid regeneration in the presence of LA-13-ZS70 (Fig. 2G, Supplementary Data 2).

Although LA-13-ZS70 improved spheroid formation, it displayed poly-pharmacology with a multi-kinase inhibition profile (BTK, EPHA4, EPHB4, KIT, LCK, MAPK14, PDGFRA). Subsequent medicinal chemistry efforts informed by structure-activity relationship generated DB-11-BE87, a structural analogue of LA-13-ZS70 (Fig. 2C, F). A concentration-response curve confirmed activity of DB-11-BE87 (DB-11-BE87: EC₅₀ = 151 nM, count of Hopx-GFP+ spheroids) (Fig. 2D, E). In addition, significant increases (log2 fold change >1 and FDR < 0.05) of canonical AT1 markers including Ager, Aqp5, Gprc5a, Hopx, and Igfbp2 were observed by RASL-seq (Fig. 2H and Supplementary Data 2). Both LA-13-ZS70 and DB-11-BE87 increased canonical AT2 markers as well including Sftpc and Sftpb, (Fig. 2G, H and Supplementary Data 2), though to a lesser extent than AT1 markers.

Spheroids treated with DMSO, LA-13-ZS70, or DB-11-BE87 were stained with cell type specific markers (Supplementary Fig. 2). Both LA-13-ZS70 and DB-11-BE87 increase AT1 markers (GFP, HOPX, and AQP5) confirming RASL-seq results at the protein level. The AT2 marker pro-SPC was decreased in the LA-13-ZS70 but not DB-11-BE87 treated samples compared to DMSO control. In all samples pro-SPC labeled cells in the same region as AT1 markers though colocalization could not be assessed in single-plex. Additionally, both LA-13-ZS70 and DB-11-BE87 decreased fibroblast and macrophage markers (COL1A1 and IBA1; respectively). Hopx expression displayed a similar concentration-response curve to the Hopx-GFP+ imaging readout in DB-11-BE87 treated samples (Fig. 2.E, I). Activity of DB-11-BE87 was attenuated when TGF-β1 was removed from culture suggesting dependency (Fig. 2J). Together, our HTS identified DB-11-BE87 that reverted TGF-β1 impaired AT1 spheroid formation.

Longitudinal scRNA-seq of alveolar organoid culture identifies pulmonary fibrosis associated Fn1+ aberrant epithelial cells

To understand the effect of DB-11-BE87 on mixed distal lung cell cultures, longitudinal scRNA-seq was performed at three different time points (day 1, 3, and 11) following TGF-β1 and compound treatment (DMSO, 10 µM DB-11-BE87, 10 µM QA-92-TQ17) (n = 3–4 per group) (Fig. 3A). Compound QA-92-TQ17 was chosen as a minimally active structural analogue of DB-11-BE87 (Supplementary Fig. 3). As a control, cells freshly dissociated from untreated mouse lung on day 0 were profiled by scRNA-seq. After QC, 117, 450 cells were obtained from all conditions. Cells from four major categories were identified using canonical markers and were clearly separated on uniform manifold approximation projection (UMAP), a dimensional-reduction technique for improved visualization and interpretation of single cell data (Fig. 3A)¹⁸.

**Fig. 3: Alveolosphere scRNA-seq detects *Fn1*+ aberrant epithelial cells.**

The proportion of the four cell categories in the cell culture changed over time, although significant differences in the composition of compound-treated samples and the DMSO controls were not observed at day 1 or day 3. By day 11, there was a marked increase in the population of epithelial cells in the DB-11-BE87-treated sample (p = 2.6e-6, student’s t test Fig. 3B); consistent with the increased Hopx-GFP observed in the HCI assay. To examine epithelial cell types in our samples, we integrated epithelial cells from each condition and identified cell clusters based on the transcriptome similarity. Cell type labels were assigned to each cluster according to previously established cell type markers (Secretory cells: Scgb3a1, Scgb3a2, Scgb1a1; Ciliated: Dynlrb2, Foxj1; AT2: Sftpa1, Sftpc; Lyz1 + AT2: Lyz1; AT1: Ager, Hopx, Pdpn; Proliferating cells: Top2a, Mki67)^3,4. As expected, major epithelial cell types were observed on day 0 before culture with few transitional or proliferating cells (Supplementary Fig. 4A). Notably, clusters resembling recently described transitional AT2-AT1 cell states that persist in mouse fibrosis models were captured in this in vitro assay (Trans AT2: Lgals3 high, Gclc high; Krt8+ alveolar differentiation intermediate (ADI): Krt8 high, Cldn4 high) (Fig. 3C, D)^3,4.

To determine lineage relationship between these populations, the ratio of spliced to unspliced transcriptomic reads was calculated to deduce the RNA velocities and predict differentiation trajectories across cells¹⁹. As expected, RNA velocity trajectory indicated a clear differentiation of AT2 cells into trans AT2 and then AT1 cells (Fig. 3E and Supplementary Fig. 5A). Interestingly, the trajectory also suggests a bifurcate fate of Krt8 + ADI cells, one part differentiated into AT1 while the rest transitioned into another terminal cell state featured by high expression of Fn1 (hereafter Fn1⁺ Krt8⁺ ADI, Fig. 3E, F). A similar result was also observed by diffusion pseudo-time analysis (Supplementary Fig. 5B, C)²⁰. Recent scRNA-seq studies identified aberrant basaloid cells, also known as KRT5^-/KRT17⁺ epithelial cells, in the distal lung of IPF patients, which exhibited high level of Fn1^6,7. To examine if the Fn1⁺ Krt8⁺ ADI cells in TGF-β1 treated mouse lung cell organoids are similar to the aberrant basaloid cell population in IPF patients, the transcriptomes of alveolar epithelial cells from the two species were compared. High similarity was observed between the Fn1⁺ Krt8⁺ ADI and the aberrant basaloid cells (Fig. 3G). Cell composition analysis indicates that DB-11-BE87 on day 11 significantly increased transAT2 (student’s t test p = 3e-4) and Krt8⁺ ADI (Fn1^-) (p = 6e-4) cells, decreased Krt8⁺ ADI (Fn1⁺) (p = 0.02) cells (Fig. 3H and Supplementary Fig. 4B) and resulted in lowered Fn1 expression in the culture on day 11 (Fig. 3I). This suggests a stronger and healthier AT2- > AT1 transition than DMSO controls. Together, our data suggest that TGF-β1 treated alveolar organoids contain a population similar to pulmonary fibrosis associated Fn1+ aberrant epithelial cells, which can be ameliorated by treatment with DB-11-BE87.

TGF-β1 induces fibroblast activation in alveolar organoids

TGF-β1 is a master regulator that stimulates fibroblast to myofibroblast differentiation and contributes to tissue remodeling in fibrotic disease²¹. To examine the effect of TGF-β1 on alveolar organoid formation, we inspected temporal transcriptional changes in the mesenchymal compartment of the alveolar organoid culture (Fig. 3A). Alveolar fibroblasts, peribronchial fibroblasts, pericytes, smooth muscle cells, and adventitial fibroblasts were identified in the freshly disassociated cells on day 0 (Supplementary Fig. 6A)¹¹. One day after growing in media, reduced expression of mesenchymal cell markers, e.g., Col1a1, Col3a1, Acta2, and Postn, was observed in all cell clusters (Supplementary Fig. 6C). Concurrently, cells became more dispersed on UMAP projection suggesting a rise in transcriptome diversity because of adaption to culture environment (Supplementary Fig. 6B).

On day 3, markers of mesenchymal cell types once again became prominent (Fig. 4A, B). Comparing to the cells on day 1, several newly emerging cell populations on day 3 were identified on the integrated UMAP, including clusters 0, 5, 6, 7, 8, and 10 (Fig. 4A, C). Cluster 5, 6, 7, and 8 showed a high level of Cthrc1 with cluster 5 and 8 also exhibiting a proliferative profile (Fig. 4A, F). By examining the RNA velocity trajectories, cluster 6 and 7 were likely derived from transitioning alveolar fibroblasts and adventitial fibroblasts, respectively (Fig. 4D, E). A recently published atlas of collagen-producing cells in mouse lung has identified a subpopulation of fibroblasts that express Cthrc1. This subpopulation emerges in the bleomycin-induced lung fibrosis model and has been found in patients with IPF and scleroderma¹¹. Comparison between the two datasets through integration suggests that Cthrc1⁺ fibroblasts observed in the two models are closely similar to each other (Supplementary Fig. 7A). Accordant with the reported fibrotic nature of these cells, we observed gene expression profiles consistent with fibrosis-related pathways such as the TGFβ, WNT, and ECM pathways (Fig. 4F).

**Fig. 4: TGFβ induces fibroblast activation in alveolosphere model.**

The newly emerged cell cluster 10 exhibited the highest level of Acta2 expression and its transcriptomic profile was distinct from smooth muscle cells (Fig. 4B). RNA velocity analysis indicated that they originated from the transitioning alveolar fibroblasts, suggesting a fibroblast to myofibroblast transition already underway by day 3 (Fig. 4E). Interestingly, neither DB-11-BE87 nor QA-92-TQ17 exhibited any influence on this transition, as composition of mesenchymal cells in the organoids was comparable across all treatments (Supplementary Fig. 7B).

DB-11-BE87 resolves TGF-β1 induced fibroblast activation in the alveolosphere model

Mesenchymal cells respond to physical and chemical stimuli to change cellular state²¹. The integrated UMAP shows that cells from day 11 formed distinct non-overlapping clusters with cells from day 3, suggesting that mesenchymal cells underwent a dramatic shift of cell state by day 11 (Fig. 5A). Most cells from day 11 accumulated into two unique clusters; cluster 3 and 5 (Fig. 5B). Cluster 3 represented myofibroblast cells expressing the highest level of Acta2 and exhibiting enrichment of muscle contraction and inflammation pathways (Fig. 5C–E). Cluster 5 displayed low expression of Acta2 but high levels of Sfrp1 and Spp1, two recently emerged markers of intermediate states between normal fibroblasts and the Cthrc1⁺ pathogenic fibroblasts that accumulate during fibrosis¹². In addition, these cells contained highly active adipogenesis genes, exemplified by Lpl, implying a lipo-fibroblast identity (Fig. 5C–E).

**Fig. 5: DB-11-BE87 activates AHR and resolves TGF-β1 induced fibrosis in the alveolosphere model.**

To examine the dynamics of the cell type transition, RNA velocity trajectories were projected onto the UMAP (Fig. 5B). Myofibroblast cells appeared to undergo a de-differentiation process, generating transitional cell populations that eventually became the Spp1⁺/Sfrp1⁺ fibroblasts (Fig. 5B, D). This de-differentiation process implies that TGF-β1 supplementation no longer sustains fibroblast activation on day 11. Notably DB-11-BE87 treated samples display a decrease in the proportion of myofibroblast cells (cluster 3; student’s t test p = 0.006) and increase in the proportion of transitional fibroblast cells (cluster 0 and cluster 1; p = 0.0058 and p = 0.0064, respectively) compared to other treatments (Fig. 5F). Together, these data suggest that DB-11-BE87 accelerates the resolution of TGF-β1 induced fibroblast activation in the mouse alveolosphere culture model.

In addition to exerting profibrogenic effect on fibroblasts, TGF-β1 also induces pulmonary fibrosis by directly acting on lung epithelial cells^22,23. To investigate the direct effects of TGF-β1 and BD-11-BE87 on epithelial cells we isolated EpCAM⁺ cells from Hopx-GFP mice and established an alveolosphere assay with only epithelial cells (Supplementary Fig. 8). To maintain epithelial cells in a stromal free culture we adapted the assay to using define culture conditions^24,25. In contrast to alveolosphere cultures containing mesenchymal cells (Fig. 2E, I), epithelial only cultures did not show an increased in GFP+ spheroids or Hopx expression with increasing concentrations of BD-11-BE87 (Supplementary Fig. 8B, C). Comparison of epithelial only cultures ± TGF-β1 by RASL-seq did not show many differentially expressed genes (Supplementary Fig. 8D; Supplementary Data 3). RASL-seq of epithelial only cultures treated with DB-11-BE87 vs. control did not display an increase of canonical AT1 markers (Supplementary Fig. 8E and Supplementary Data 3). These results are consistent with BD-11-BE87 and TGF-β1 acting indirectly on epithelial cells through stromal populations in the alveolosphere culture.

DB-11-BE87 activates AHR and rescues TGF-β1 impaired AT1 spheroid formation while AHR antagonists do not

To identify the target and mechanism of action (MoA) of DB-11-BE87, we compared DB-11-BE87-treated samples with the DMSO controls and identified differentially expressed genes (DEGs) across unsupervised cell clusters from all four cell categories. To minimize the impact of secondary effects, we investigated cells from the earliest time point (day 1) and identified signaling pathways and biological processes that were enriched in the DEGs from each cell cluster. As expected, most cell clusters did not exhibit significantly perturbed pathways due to the lack of DEGs at early time points. Nevertheless, several mesenchymal cell clusters exhibited shared and significantly perturbed pathways such as Aryl hydrocarbon receptor (AHR) pathway (Supplementary Fig. 9). By examining genes involved in those pathways, Cyp1a1 and Cyp1b1 were the most frequently observed genes up-regulated by DB-11-BE87 (Supplementary Fig. 9B). Since Cyp1a1 and Cyp1b1 are well-established downstream targets of AHR, we sought to examine the hypothesis that DB-11-BE87 functions as an AHR agonist. A list of canonical AHR downstream targets were assembled, including Cyp1a1, Cyp1b1, Cyp2a1, Ahrr, Tiparp, Ptgs1, Ptgs2, and Nqo1^26,27, and module activity of these genes was calculated for each mesenchymal cell cluster (methods). Significant linear correlations were observed between Ahr expression and AHR module activity in the DB-11-BE87-treated cells at all time points (Fig. 6A). In contrast, AHR module activities in DMSO controls remained below the background level across cell clusters and time points. Interestingly, although the minimally active analogue QA-92-TQ17 showed weak AHR activation on day 1 and day 3, its effect dropped to the background level on day 11. This is consistent with its minimal effect on AT1 regeneration in the Hopx-GFP alveolar organoid imaging assay (Supplementary Fig. 3).

**Fig. 6: AHR agonists rescue TGF-β1 impaired alveolar type 1 spheroid formation while AHR antagonists do not.**

To confirm effect of AHR activation in the murine Hopx-GFP spheroid assay we investigated three AHR agonists (TE-33-UQ34, EB-06-KR05, QB-43-PX71) and two antagonists (GD-70-TQ68, DC-78-LZ94) (Fig. 6B). Compounds were validated as AHR agonist or antagonists based on their demonstrated activity in a murine H1L1.1c2 cell line stably expressing an AHR response luciferase reporter (Fig. 6C). In the murine alveolosphere assay AHR agonists increased GFP adjusted count spheroids while AHR antagonists decreased this count (Fig. 6D). Together these data suggest that DB-11-BE87 resolves TGF-β1 induced fibroblast activation through behaving similar to an AHR agonist.

DB-11-BE87 induced AT2 to AT1 differentiation through IL-1β signaling

While fibroblast activation was attenuated with treatment of DB-11-BE87 we were unable to find a likely signal from the fibroblast compartment promoting the resolution of Fn1+ aberrant epithelial cells. It has been reported that HIF1A-mediated glycolysis controls AT2-AT1 conversion in the bleomycin-injured mouse lung, and that this process is promoted by IL-1β signaling via interstitial macrophages⁵. To interrogate if IL-1β could be promoting the resolution of Fn1+ aberrant epithelial cells in DB-11-BE87 treated samples we inspected the macrophage compartment of the alveolar organoid culture (Fig. 7A–C). In line with this model, we observed that DB-11-BE87 treatment resulted in an increase of Il1b expression by day 11 in the culture (Fig. 7A–C). Meanwhile, highly active hypoxia and glycolysis pathways were observed in the epithelial cells of the DB-11-BE87-treated organoids (Fig. 7D) suggesting the epithelium is displaying a signature consistent with previously reported IL-1β mediated enhanced glycolysis ⁵.

**Fig. 7: DB-11-BE87 induces hypoxia mediated glycolysis in AT2 on day 11 via IL-1β signaling.**

Since DB-11-BE87 activates AHR signaling in mesenchymal cells, we hypothesized that Il1b induction in interstitial macrophages was trigged by cell-cell communication between fibroblasts and macrophages. To characterize the niche environment in the organoids, we quantified communication probabilities between cell type pairs using CellChat²⁸. Results show that mesenchymal cells produced the most outgoing signals among four categories, suggesting that they are the primary drivers of the niche environment in the organoids (Supplementary Fig. 10A). Interestingly, Cthrc1+ fibroblasts and Spp1/Sfrp1⁺ fibroblasts are the two cell types that generate the most communication signals directed towards interstitial macrophages on day 3 and day 11, respectively (Fig. 7E). This coincides with the strongest Ahr expression and AHR module activity in the two cell types in DB-11-BE87-treated organoids (Fig. 7F). These results imply that Il1b production in interstitial macrophages is associated with AHR activation in fibroblast cells.

To identify AHR downstream targets that mediate communications between fibroblasts and macrophages, we inspected predicted ligand-receptor pairs between the two Ahr-expressing fibroblast subtypes and the interstitial macrophages from day 11 (Supplementary Fig. 10B). There are 31 ligands produced in both Cthrc1+ fibroblasts and Spp1/Sfrp1⁺ fibroblasts, mediating potential communications with interstitial macrophages. However, majority of them were also expressed in a wide range of cell types, arguing against their connection to the DB-11-BE87-induced AHR activation (Supplementary Fig. 10D). In addition, only one ligand Sema6d was commonly up-regulated by DB-11-BE87 in both fibroblast subtypes (Supplementary Fig. 10C). These results imply that AHR activation in fibroblasts may not be the direct trigger of Il1b induction in the interstitial macrophages. Rather, a more complex communication network among multiple cell types present in the culture may be involved in this process. Together, our study suggests that DB-11-BE87 induces AHR activation in fibrotic fibroblasts, which accelerates the resolution of TGF-β1 induced fibroblast activation and promotes AT2 to AT1 differentiation through an indirect induction of macrophage-mediated IL-1β signaling.

Discussion

Regeneration of the alveolus is a major goal in modern pulmonary biology. Central to alveolar regeneration is re-establishing the correct structure of the damaged epithelium including promoting the differentiation capacity of AT2 to AT1 cells that form the site of air-gas exchange. TGF-β1 and other signaling molecules from pathological mesenchyme direct AT2s into an abnormal intermediate cell state which accumulates in fibrosis^7,29. Using a complex multicellular 3D organoid culture with TGF-β1 dependent impaired AT1 spheroid formation, we identified the small molecule DB-11-BE87 as able to decrease Fn1⁺ Krt8⁺ ADI cells and rescue spheroid formation, a potentially high impact finding. Subsequent scRNA-seq and reporter assays suggests AHR activation in the mesenchyme reverts the deleterious effects of TGF-β1 signaling on AT2-AT1 cell regeneration. This line of inquiry emphasizes the strength of using multicellular cultures coupled with scRNA-seq to identify cell non-autonomous MoAs for early drug discovery.

AHR is a ligand activated transcription factor whose primary function is controlling xenobiotic metabolism²⁷. Following ligand binding, AHR translocates to the nucleus and induces transcription of target genes, including cytochrome p450 enzymes (CYP1A1, CYP1A2, CYP1B1), leading to hydrocarbon detoxication. Generation of Ahr null mice identified multiple physiologic roles for AHR outside of exogenous metabolism including regulation of the immune response, hepatic development, regeneration, and cancer^30,31. Recent epithelial-specific removal of AHR unveiled its function as an activator of differentiation factors and a repressor of stemness factors during epithelial injury in organoid culture³². The role of AHR in lung epithelial cell regeneration has not been well studied.

AHR controls complex transcriptional events in a ligand, cell-type, and context-specific manner through acting as a sensor for a wide variety of compounds^33,34. AHR has been extensively studied in the context of activation by polycyclic aromatic hydrocarbons found in cigarette smoke and immunomodulation in respiratory disease^35,36. Additional evidence suggest AHR modulates tissue fibrosis³⁷. In the murine model of bleomycin induced pulmonary fibrosis, removal of AHR has a heightened fibrotic response while activation of AHR has an attenuated response^38,39. Our results suggest a benefit to AHR activation in these rodent models could be through promoting AT1 cell regeneration, a characteristic of recovery after bleomycin injury⁴⁰. While currently approved drugs for idiopathic pulmonary fibrosis (Nintedanib and Pirfenidone) target fibroblast activation, it is unclear if these therapies promote AT1 cell regeneration suggesting AHR activation could have novelty.

Acquisition of a contractile and extracellular matrix synthesizing fibroblast phenotype is a defining feature of fibrosing interstitial lung disease^41,42. TGF-β1 is the dominant perturbation known to trigger this phenotypic conversion and has been studied extensively in fibrosis for over the past 30 years⁴³. TGF-β1 treated human fibroblasts have impaired ability to support mouse AT2 spheroid formation¹³. AHR ligands decrease expression of Tgfb1 and inhibit TGF-β1 induced myofibroblast formation^44,45,46. Crosstalk between the TGFβ and AHR pathways has been described, although the physiological importance of this interaction is not completely understood^47,48. Our results suggest that AHR activation in fibroblasts reverts the pathogenic effects of TGF-β1 and restores the innate epithelial support role of the alveolar stroma. Our results are consistent with AHR activation accelerating the de-differentiation of myofibroblasts however lineage tracing would be necessary to demonstrate this.

As AT1 cells are non-proliferative it is likely that TGF-β1 and identified DB-11-BE87 are affecting progenitor cells. While we observed a decrease in aberrant Fn1⁺ epithelial cells upon DB-11-BE87 treatment, suggesting promotion of AT1 differentiation, we also see an increase of AT2 markers. It is possible DB-11-BE87 is also increasing progenitor proliferation. It is difficult to determine what cell type a small molecule is acting on in the multicellular spheroid assay. While our scRNA-seq cell-chat analysis and sorted epithelial only cultures suggest it is the fibroblasts that are directly responding to DB-11-BE87, these are not definitive and further experiments are necessary.

The experiments in this manuscript support DB-11-BE87 accelerating the resolution of TGF-β1 induced fibroblast activation in vitro; treating samples at the same time with TGF-β1 and DB-11-BE87. It will be important in future studies to extend these results to in vivo models of pulmonary fibrosis such as the bleomycin model^49,50. Critical to these experiments will be to: 1) determine if DB-11-BE87 has beneficial effects on reducing pulmonary fibrosis and on increasing epithelial recovery; 2) determine if DB-11-BE87 can revert established fibrosis (dosed during the therapeutic window and not as a preventive treatment); 3) assess the most effective route of administration (systemic vs. local delivery); and 4) assess for any off-organ on-target toxicity as is seen with directly inhibiting the TGF-β1 pathway⁵¹.

There are interspecies differences between mouse and humans’ ability to respond to AHR ligands⁵². We have no evidence to suggest AHR activity is beneficial for human alveolar regeneration. DB-11-BE87 increases AHR target gene expression however it is unclear if DB-11-BE87 directly binds to AHR acting as a ligand or if it activates the AHR pathway though a disparate mechanism. While DB-11-BE87 increases the AHR pathway activation, the dependence of AHR for DB-11-BE87’s efficacy has not been established. Further work on testing the effect of DB-11-BE87 on human alveolar cultures, direct binding to AHR, and efficacy dependence of AHR will extend the clinical relevance of this work.

Target identification and MoA studies following on hits from phenotypic screening is notoriously difficult⁵³. Target identification is desired to define on-target limitations, design confirmatory follow-up experiments, and identify pharmacodynamic markers for efficacy in vivo⁵⁴. Lack of target identification has historically led to discontinuation of drug discovery programs hampering progression of promising molecules to the clinic. Here we leveraged scRNA-seq to identify the mesenchyme as the primary site of action of DB-11-BE87, identifying a cell-non autonomous effect of AHR activation on promoting alveolar regeneration. These insights would have been un-realized with traditional bulk transcriptomic approaches and highlight the need for understanding phenotypic screens at a single-cell resolution.

This study leveraged multiple technical approaches to identify a new MoA in AT1 regeneration. HTS with a HCI readout in a multicellular assay was used to probe a diverse chemical library to find modulators of a desired phenotype. RASLseq was performed to confirm hits and to inform subsequent medicinal chemistry. scRNA-seq was executed for MoA determination. We propose these approaches are replicable and present an avenue for identifying new MoAs in forward pharmacology.

Materials and methods

Mice

All mice (Mus musculus) were housed in a barrier facility free of pathogens. All studies were performed under a protocol approved by the Institutional Animal Care and Use Committee at Novartis-San Diego (Protocol No. 20-429). We have complied with all relevant ethical regulations for animal use. Mouse strain Hopx^3FlagGFP, referred to as Hopx-GFP, has been described⁵⁵. All mouse strains (Hopx^3FlagGFP, referred to as Hopx-GFP; Sftpc^CreER; ROSA26^tdTomato) have been described^40,55,56. All mouse strains were maintained on a C57BL/6 J background. Male mice only were used for all studies. Mice 8–13 weeks of age were used for all studies.

Hopx-GFP murine organoid assay

Adult Hopx-GFP+ mice were sacrificed with isoflurane, perfused with PBS through the right ventricle, and the lungs were extracted. Trachea, connective tissue, and upper bronchiole was removed under dissection microscope. Lungs were cut into small pieces (>2 mm²) and incubated in protease solution for 30 minutes at 37 °C in a shaking incubator. Protease solution contained: 450 U/mL Collagenase Type I (Gibco; 17100–017), 4 U/mL Elastase (Worthington Biochemical Company; LS002279), Dispase 5 U/mL (BD Biosciences; 354235), and 0.33 U/mL DNaseI (Roche; 10104159001) dissolved in DMEM/F12 (Gibco; 11320033). Digestion enzymes were inactivated with DMEM/F12 + 10% FBS, cells were centrifuged at 500 g and resuspended in Trypsin/DNase solution and incubated for 30 minutes at 37 °C in a shaking incubator. Trypsin/DNase solution contained: 0.1% Trypsin-EDTA (Gibco; 25200056) + 0.13 U/mL DNaseI (Roche; 10104159001) in DMEM/F12. Trypsin/DNase solution was inactivated with DMEM/F12 + 10% FBS, cells were filtered through a 70 µM cell strainer, centrifuged at 300 g (2×), resuspended in RBC cell lysis buffer (Sigma; R7757), and digestion enzymes inactivated with DMEM/F12 + 10% FBS. Cells were centrifuged at 500 g, resuspended in DMEM/F12 + 2% FBS, counted on a hemacytometer, and diluted to 1 × 10^⁶ cells per mL. Cells were filtered through a 40 µM cell strainer, resuspended 1:1 in Matrigel Growth Factor Reduced (Corning; 354230), and 20 µL of solution (10,000 cells) was plated into the bottom of 384-well CellCarrier Ultra tissue culture plate (PerkinElmer; 6057300) pre-spotted with compounds (10 µM final concentration). 30 µM of TGFβ-R1 inhibitor galunisertib was pre-spotted as a positive control. Once solidified 50 µL of ITS media was added per well and plates were incubated in a standard cell culture incubator. ITS media contained: DMEM/F12 + 5% FBS, 1/100 ITS-G (Gibco; 41400045), 1/100 Antibiotic-Antimycotic (Gibco; 15240096), 0.5 ng/mL TGF-β1 (R&D Systems; 240-B-002). Fresh ITS media was added on top after 4–5 days in culture. On day 8 60 µL of ITS media was removed and replaced with 60 µL of fresh ITS media. On day 10–11 ITS media was removed and fresh ITS media containing a 1:2500 dilution of Lysoview-405 (Biotium; 70066) was added. Plate was placed in cell culture incubator for ~24 h before live fluorescent imaging for Hopx^GFP was performed on high content confocal imaging platform: ImageXpress Micro (Molecular Devices; primary screen) or Opera Phenix (PerkinElmer; reconfirmation and subsequent experiments).

For epithelial only cultures ITS media was replaced by AT2 maintenance medium for murine culture^24,25.

HCI imaging, processing, and data normalization

The image z-stacks of the Hopx-GFP channel were compressed to maximum intensity projections (MIPs). Representative Hopx-GFP MIPs were then used to train an ilastik pixel classifier to differentiate between spheroids, artifacts, dim objects, and background⁵⁷. The spheroid probability images output from ilastik together with MIPs from Hopx-GFP channel were passed to a CellProfiler pipeline for object segmentation and numeric feature measurements⁵⁸. Spheroids were initially identified by thresholding the spheroid probability images (>0.5). Spheroid morphological and intensity features were then measured, and small (<500 pixels) and dim (Hopx-GFP upper quartile intensity < 0.0075) spheroids were filtered out. Mean aggregation of the spheroid features was utilized to generate well-level data, was written to CSV, and analyzed in Spotfire (Tibco). For adjusted measurements data was normalized by setting negative control wells to 1.

Small molecule compound deck

Primary small molecule screening deck consisted of ~16,800 compounds selected based on diversity and attractive chemical properties (molecular weight, cLogP, solubility, etc)¹⁶.

Single cell sample prep

Hopx-GFP murine organoid assays were established as above except scaled to plate 4–7 drops of 50 µL cells-medium/Matrigel mixture in each well of a 6-well plate. At designated times (day 0, day 1, day 3, and day 11) cells were removed from Matrigel for scRNA-seq. Briefly, media was removed and washed with ice cold PBS. TrypLE Express (Gibco; 12604013) was added to wells and incubated at 37 °C, followed by scraping of Matrigel bubbles to collect in a 15 mL conical tube. Cells were mechanically disrupted by pipetting ~10 times and incubated at 37 °C for an additional 10 min. Cells were mechanically disrupted again and enzyme was inactivated with DMEM/F12 + 10% FBS + 1/100 Penicillin-Streptomycin (Gibco; 15140122). Cells were centrifuged at 300 g at 4 °C, resuspended to 1000 cells/µL in DMEM/F12 + 10% FBS, and filtered through a Flowmi 40 µM cell strainer (Sigma; BAH136800040).

AHR agonist and antagonist assay

H1L1 1c2 cells stably expressing a dioxin response element (DRE) driven luciferase reporter were plated at 5000 cells/well in 40 μL of medium. Medium contained MEMa + 10% FBS, 1% Antibiotic-Antimycotic, 1% Non-essential amino acids, 1% HEPES, 1% Sodium pyruvate, and 200 μg/mL geneticin. Compounds were added at indicated molarity. Plates were centrifuged for 15 s at 1000 rpm and mixed with a plate mixer. Plates were placed in cell culture incubator for 18 h. 20 μL of luciferase reporter gene assay reagents was added (Bright Glo, Promega, E2610) and luminescence was on a multimodal plate reader. For antagonist assay 3 nM of dioxin was added at time of cell plating.

Single cell transcriptomics

The Chromium Next GEM Single Cell 3’ reagent kit v3.1 from 10X Genomics (protocol CG000204 Rev C) was used in scRNA-seq library preparation. Single-cells were partitioned into nanoliter-scale Gel Beads-in-emulsion (GEMs) using Chromium Controller. Upon cell lysis, the poly-A tails of mRNA are captured on the beads containing an Illumina read 1 sequencing primer, 16nt 10X barcode, 12nt UMI and 30nt poly (dT) sequence. Reverse transcription occurs followed by amplification of full-length cDNA from polyA mRNA. Enzymatic fragmentation and size-selection was used to optimize cDNA amplicon size. End repair, A-tailing, Adaptor ligation and PCR create the final sequencing libraries that contain the P5 and P7 primers used in Illumina bridge amplification. The libraries were sequenced using NextSeq 550 High Output Kit v2.5 (150 cycles).

Organoid histology and immunohistochemistry

Hopx-GFP murine organoid assay was scaled to plate 4–7 drops of 50 µL cells-medium/Matrigel mixture in each well of a 6-well plate. Samples were treated with either DMSO, LA-13-ZS70, or DB-11-BE87 at the start of culture. On day 10 media was removed and washed with ice cold PBS. Matrigel was dissolved using Cell Recovery Solution (Corning; 354253) at 4 °C for two hours on ice. Spheroids were collected in a 15 mL conical tube, centrifuged at 300 g at 4 °C, washed with ice cold PBS, and fixed for 30 minutes at room temperature in 4% paraformaldehyde (EMS; 15710). Samples were washed in PBS, and dehydrated into ethanol (30%, followed by 70%) before routine paraffin embedding and microtome sectioning. Serial sections were stained with hematoxylin and eosin or the following antibodies to determine cell composition: GFP (Lifespan Biosciences, LS-C67081, 2.2 µg/mL); HOPX (Novus, NBP1-92003; 1:200 dilution); AQP5 (Abcam, ab92320, 0.76 µg/mL); ProSPC (EMD Millipore, AB3786, 1:800); COL1A1 (Cell Signaling, 72026 s, 5 µg/mL); IBA1 (Wako, 019-19741, 1:2000, 2.5 µg/mL)

Flow sorting AT2 and AT1 populations

Adult Sftpc^CreER; ROSA26^tdTomato; Hopx-GFP mice were sacrificed with pentobarbital, perfused with PBS through the right ventricle, and the lungs were instilled with 1.5 mL of enzymatic cocktail. Enzymatic cocktail contained: 5 U/mL Dispase (Worthington; NPRO2), 4 U/mL Elastase (Worthington; 2294), 200 U/mL Collagenase Type 1 (Gibco, 10104159001), 0.33 U/mL DNaseI (Roche; 10104159001) in PBS. Lung was removed from chest and incubated in an additional 3.5 mL of enzymatic cocktail for 45 minutes at room temperature. Lung was transferred to dish containing 10 mL of DNase I solution and pulled apart completely using forceps. DNase I solution contained: 1 mg DNase I in 10 mL DMEM (11965118) + 10% FBS + 1% Penicillin-Streptomycin (Gibco; 15140122). Lung tissue in DNase I solution was gently rocked at low speed for 10 minutes at room temperature. Cell suspension was filtered through a 100 µM cell strainer followed by a 40 µM cell strainer, centrifuged at 300 g, resuspended in RBC cell lysis buffer (Sigma; R7757), lysis inactivated with PBS, centrifuged at 300 g, and resuspended in FACS buffer. Cells were stained with anti-CD45 conjugated to APC (1:100, eBioscience, 17-0451-83), PDPN conjugated to PE-cy7 (1:200, Biolegend, 127410) and 20 ng/mL DAPI. DAPI-; CD45-; PDPN-; GFP-; tdTomato+ cells were identified as AT2 cells and sorted to purity. DAPI-, CD45-; PDPN+; GFP + ; tdTomato- cells were identified as AT1 cells and sorted to purity (Supplementary Fig. 1A, E).

Bulk RNAseq of sorted AT2 and AT1 cells

RNA from sorted AT2 and AT1 cells was isolated using Direct-zol RNA miniprep (Zymo Research; R2050). RNA yield was quantified, libraries generated, and sequenced according to standard methods⁵⁹.

RASL-seq probe design and sample prep

The RASLseq assay was adapted from a previously published protocol⁶⁰. Briefly, Hopx-GFP murine organoids were plated in 384 well format in Matrigel (Corning) and incubated with compound for 12–13 days. After treatment, Matrigel was dissolved using Cell Recovery Solution (Corning) at 4 °C and washed in DPBS (Gibco). Cultures were then aspirated to 10 µL and an equal volume of Proteinase- K RASL lysis buffer was added to the well. Samples were then processed as published⁶⁰. For detection of mouse specific lung epithelial markers, a panel of 127 genes was used with 3 oligo pairs designed per gene (Supplementary Data 4). For balancing gene expression, the custom lung panel was run in conjunction with a 978 gene panel⁶¹.

RASL-seq data processing

Count matrix was obtained by mapping the raw reads to the probe sequences (Supplementary Data 1) as previously described⁶⁰, and samples with <15,000 mapped reads were removed. Next, read counts were normalized by the total mapped reads, i.e., RPM (Reads Per Million). Out of three probes designed for each gene, the one with the most consistent measurement to a previous benchmarking RNA-seq dataset (not shown) was chosen to serve as the expression level value of that gene. Consistency of measurement was evaluated by inter-replicate similarity determined by the Pearson correlation coefficient of log2(RPM + 1) of all the genes. Samples with maximum inter-replicate similarity <0.8, or >40% undetected probes compared to the best replicate, were considered as outliers, and excluded from downstream analysis. Batch effect across plates was removed using pyCombat⁶², then quantile normalization was applied to adjust log2(RPM + 1) across replicates. Differential expression analysis was performed using limma⁶³ in R⁶⁴.

scRNA-seq data analysis

Raw reads were mapped to the mm10 mouse reference genome and a per-cell count matrix was generated using CellRanger pipeline (10X Genomics, version 5.0.1)⁶⁵. The processed count matrix was used to filter out low quality cells with <700 detected genes or >10% mitochondrial reads. Doublets were identified using DoubletFinder and excluded from downstream analysis⁶⁶. UMI counts were normalized by Seurat sctransform⁶⁷ and top 2000 variable genes were selected for dimension reduction and 10-15 principal components were used for Leiden clustering at the downstream of Seurat v4 framework⁶⁸. To ensure identification of rare and transient cell types and states, we set resolution to 1.0 for initial cell clustering and merge clusters of the same type based on cluster markers identified by FindAllMarkers function in Seurat. Cells from replicates of each condition were pooled together and major cell categories including epithelial, endothelial, mesenchymal, and immune cells, were identified first using the described workflow. Cell types within each category were then identified using the same protocol.

Data integration and batch removal

Although cells were highly consistent across replicates of the same treatment, marked transcriptomic changes were observed in cells from different conditions. To identify matched cell types and states and remove batch effect across conditions, integrated cells across conditions using canonical correlation analysis in Seurat on 2000 most variable genes⁶⁹. Results of data integration were inspected on UMAP with cell type labels previously identified for each condition. For comparative analysis to bleomycin treated mouse lung, mesenchymal cells from GSE132771 were integrated with our data using the same method¹¹.

Differential expression and gene set enrichment analysis

To unbiasedly identify differentially expressed genes (DEGs) induced by compound treatment, we employed a cell label agnostic approach that searches DEGs from unsupervised cell clusters. Pseudobulk-based framework was used to reduce impact of dropout and technical variability attributed to single cell sequencing⁷⁰. Specifically, raw UMI counts from all cells of the same type were pooled together for each biological replicate and fed into pseudoBulkDGE function of Scran package in R Bioconductor.

Differentially expressed protein coding genes with FDR < 0.01 were ordered based on their log2 fold change. To identify enriched gene sets, we ran the fGSEA package v1.28.0⁷¹ with default setting while limiting the gene set sizes between 5 and 500 genes. Cell clusters with enriched gene sets (p < 0.05) in their DEGs were selected for further analysis.

To quantify pathway activity in cells, AddModuleScore function in Seurat was used to estimate a combined expression level of all pathway genes relative to randomly chosen genes. Cell cycle score was determined in a similar way using the CellCycleScoring function and default cell cycle genes in Seurat.

RNA velocity and trajectory analysis

Kallisto bustool pipeline (version 0.46.2)⁷² was used to align reads onto mouse genome (Ensembl assembly release 101, 3’ cDNA) using the option –fr-stranded. Spliced and unspliced reads were counted by Velocyto¹⁹. Data were then normalized by Scanpy⁷³ and RNA velocity was calculated by scVelo package⁷⁴. Moments were calculated using the scvelo.pp.moments() function with parameters pcs = 15 and neighbours = 10. Velocity estimation was performed using the scvelo.tl.velocity() function with default stochastic mode. To estimate trajectory of AT2 cells differentiating into Fn1+ and Fn1- Krt8 + ADI cells, diffusion map was calculated by Destiny package⁷⁵ with 2000 most variable genes and 15 pcs.

Cross species comparison

Similarity of human and mouse lung epithelial cell types was estimated using highly expressed genes. Specifically, single cell data from IPF patients (GSE135893) was reprocessed according to the published protocol⁷. Next, average gene expression for each cell type was calculated by AverageExpression function in Seurat and low expressing genes with max log expression <1 across all cell types were excluded. Finally, ortholog genes from two species with same symbols were selected and the Spearman correlation coefficient was calculated for each pair of cell types.

Statistics and reproducibility

EC₅₀ and IC₅₀ were calculated on log transformed compound concentrations using a three-parameter dose-response curve (Domatics, Graphpad Prism). Zero concentration controls were set as a suitably low dose to not adversely affect results. Goodness of fit was calculated by the coefficient of determination (R²) (Domatics, Graphpad Prism). Significant differences in mean values between two sets of data were calculated by using a two-tailed t-test. Significant differences in mean values between more than two sets of normally distributed data were calculated using a one-way ANOVA with Tukey’s HSD for pair-wise post-hoc analysis. Data are represented as mean ± s.d. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Number of replicates are indicated in Figure legends. Unless otherwise stated replicates are separate wells of a single experimental run. Phenotypic effect of selected compounds for followup was repeated in at least 2 independent experimental runs. scRNAseq experiment was performed one time. AHR agonist and antagonist experiments were performed one time. All graphical figures were created with BioRender.com. All figures were assembled in Illustrator 2023 (Adobe). Plots were generated in Prism 9 (Domatics, Graphpad Prism).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

scRNA-seq data generated in this study has been uploaded to the Gene Expression Omnibus (GEO; GSE269634). Numerical data for generating graphs is available in Supplementary Data 5.

References

Crapo, J. D., Barry, B. E., Gehr, P., Bachofen, M. & Weibel, E. R. Cell number and cell characteristics of the normal human lung. Am. Rev. Respir. Dis. 126, 332–337 (1982).
CAS PubMed Google Scholar
Barkauskas, C. E. et al. Type 2 alveolar cells are stem cells in adult lung. J. Clin. Investig. 123, 3025–3036 (2013).
Article CAS PubMed PubMed Central Google Scholar
Kobayashi, Y. et al. Persistence of a regeneration-associated, transitional alveolar epithelial cell state in pulmonary fibrosis. Nat. Cell Biol. 22, 934–946 (2020).
Article CAS PubMed PubMed Central Google Scholar
Strunz, M. et al. Alveolar regeneration through a Krt8+ transitional stem cell state that persists in human lung fibrosis. Nat. Commun. 11, 3559 (2020).
Choi, J. et al. Inflammatory signals induce AT2 cell-derived damage-associated transient progenitors that mediate alveolar regeneration. Cell Stem Cell 27, 366–382.e7 (2020).
Article CAS PubMed PubMed Central Google Scholar
Adams, T. S. et al. Single-cell RNA-seq reveals ectopic and aberrant lung-resident cell populations in idiopathic pulmonary fibrosis. Sci. Adv. 6, 1–17 (2020).
Article Google Scholar
Habermann, A. C. et al. Single-cell RNA sequencing reveals profibrotic roles of distinct epithelial and mesenchymal lineages in pulmonary fibrosis. Sci. Adv. 6, eaba1972 (2020).
Zepp, J. A. et al. Distinct mesenchymal lineages and niches promote epithelial self-renewal and myofibrogenesis in the lung. Cell 170, 1134–1148.e10 (2017).
Article CAS PubMed PubMed Central Google Scholar
El Agha, E. et al. Two-way conversion between lipogenic and myogenic fibroblastic phenotypes marks the progression and resolution of lung fibrosis. Cell Stem Cell 20, 261–273.e3 (2017).
Article CAS PubMed Google Scholar
Lederer, D. J. & Martinez, F. J. Idiopathic Pulmonary Fibrosis. N. Engl. J. Med. 378, 1811–1823 (2018).
Article CAS PubMed Google Scholar
Tsukui, T. et al. Collagen-producing lung cell atlas identifies multiple subsets with distinct localization and relevance to fibrosis. Nat. Commun. 11, 1–16 (2020).
Article Google Scholar
Mayr, C. H. et al. Sfrp1 inhibits lung fibroblast invasion during transition to injury-induced myofibroblasts. Eur. Respir. J. 63, 2301326 (2024).
Ng-Blichfeldt, J. P. et al. Tgf-β activation impairs fibroblast ability to support adult lung epithelial progenitor cell organoid formation. Am. J. Physiol. Lung Cell Mol. Physiol. 317, L14–L28 (2019).
Article CAS PubMed Google Scholar
Jain, R. et al. Plasticity of Hopx+ type I alveolar cells to regenerate type II cells in the lung. Nat. Commun. 6, 6727 (2015).
Van der Velden, J. L., Bertoncello, I. & McQualter, J. L. LysoTracker is a marker of differentiated alveolar type II cells. Respir. Res. 14, 1–7 (2013).
Google Scholar
Schuffenhauer, A. et al. Evolution of Novartis’ small molecule screening deck design. J. Med. Chem. 63, 14425–14447 (2020).
Article CAS PubMed Google Scholar
Li, H., Qiu, J. & Fu, X. D. RASL-seq for massively parallel and quantitative analysis of gene expression. Curr. Protoc. Mol. Biol. 1, 4–13(2012).
Becht, E. et al. Dimensionality reduction for visualizing single-cell data using UMAP. Nat. Biotechnol. 37, 38–47 (2019).
Article CAS Google Scholar
La Manno, G. et al. RNA velocity of single cells. Nature 560, 494–498 (2018).
Article PubMed PubMed Central Google Scholar
Haghverdi, L., Büttner, M., Wolf, F. A., Buettner, F. & Theis, F. J. Diffusion pseudotime robustly reconstructs lineage branching. Nat. Methods 13, 845–848 (2016).
Article CAS PubMed Google Scholar
Hinz, B. et al. The myofibroblast: one function, multiple origins. Am. J. Pathol. 170, 1807–1816 (2007).
Article CAS PubMed PubMed Central Google Scholar
Wu, H. et al. Progressive pulmonary fibrosis is caused by elevated mechanical tension on alveolar stem cells. Cell 180, 107–121.e17 (2020).
Article CAS PubMed Google Scholar
Enomoto, Y. et al. Autocrine TGF-β-positive feedback in profibrotic AT2-lineage cells plays a crucial role in non-inflammatory lung fibrogenesis. Nat. Commun. 14, 4956 (2023).
Article CAS PubMed PubMed Central Google Scholar
Konishi, S., Tata, A. & Tata, P. R. Defined conditions for long-term expansion of murine and human alveolar epithelial stem cells in three-dimensional cultures. STAR Protoc. 3, 101447 (2022).
Article CAS PubMed PubMed Central Google Scholar
Katsura, H. et al. Human lung stem cell-based alveolospheres provide insights into SARS-CoV-2-mediated interferon responses and pneumocyte dysfunction. Cell Stem Cell 27, 890–904.e8 (2020).
Article CAS PubMed PubMed Central Google Scholar
Torti, M. F., Giovannoni, F., Quintana, F. J. & García, C. C. The Aryl hydrocarbon receptor as a modulator of anti-viral immunity. Front. Immunol. 12, 624293 (2021)
Larigot, L., Juricek, L., Dairou, J. & Coumoul, X. AhR signaling pathways and regulatory functions. Biochimie Open 7, 1–9 (2018).
Jin, S., Plikus, M. V. & Nie, Q. CellChat for systematic analysis of cell-cell communication from single-cell and spatially resolved transcriptomics. Nat. Protoc. 20, 180–219 (2025)
Kathiriya, J. J. et al. Human alveolar type 2 epithelium transdifferentiates into metaplastic KRT5+ basal cells. Nat. Cell Biol. 24, 10–23 (2022).
Article CAS PubMed Google Scholar
Fernandez-Salguero, P. et al. Immune system impairment and hepatic fibrosis in mice lacking the dioxin-binding Ah receptor. Science 268, 722–726 (1995).
Article CAS PubMed Google Scholar
Schmidt, J. V., HUEI-TING Sutt, G., Reddy, J. K., CELESTE SIMONt, M. & Bradfield, C. A. Characterization of a Murine Ahr Null Allele: Involvement of the Ah Receptor in Hepatic Growth and Development (Dioxin/2,3,7,8-Tetrachlorodibenzo-p-Dioxin/Gene Targeting/Liver), Vol. 93. https://www.pnas.org (1996).
Shah, K. et al. Cell-intrinsic aryl hydrocarbon receptor signalling is required for the resolution of injury-induced colonic stem cells. Nat. Commun. 13, 1827 (2022).
Rothhammer, V. & Quintana, F. J. The aryl hydrocarbon receptor: an environmental sensor integrating immune responses in health and disease. Nat. Rev. Immunol. 19, 184–197 (2019).
Article CAS PubMed Google Scholar
Leclerc, D., Staats Pires, A. C., Guillemin, G. J. & Gilot, D. Detrimental activation of AhR pathway in cancer: an overview of therapeutic strategies. Curr. Opin. Immunol. 70, 15–26 (2021).
Article CAS PubMed Google Scholar
Moorthy, B., Chu, C. & Carlin, D. J. Polycyclic aromatic hydrocarbons: from metabolism to lung cancer. Toxicol. Sci. 145, 5–15 (2015).
Beamer, C. A. & Shepherd, D. M. Role of the aryl hydrocarbon receptor (AhR) in lung inflammation. Semin. Immunopathol. 35, 693–704 (2013).
Yang, C. E. et al. Aryl hydrocarbon receptor: from pathogenesis to therapeutic targets in aging-related tissue fibrosis. Ageing Res. Rev. 79, 101662 (2022).
Wu, S.-M. et al. Aggravation of pulmonary fibrosis after knocking down the aryl hydrocarbon receptor in the insulin-like growth factor 1 receptor pathway. Br. J. Pharm. 179, 3430–3451 (2022).
Article CAS Google Scholar
Takei, H., Yasuoka, H., Yoshimoto, K. & Takeuchi, T. Aryl hydrocarbon receptor signals attenuate lung fibrosis in the bleomycin-induced mouse model for pulmonary fibrosis through increase of regulatory T cells. Arthritis Res. Ther. 22, 20 (2020).
Article CAS PubMed PubMed Central Google Scholar
Rock, J. R. et al. Multiple stromal populations contribute to pulmonary fibrosis without evidence for epithelial to mesenchymal transition. Proc. Natl. Acad. Sci. USA 108, E1475–E1483 (2011).
Richeldi, L., Collard, H. R. & Jones, M. G. Idiopathic pulmonary fibrosis. Lancet 389, 1941–1952 (2017).
Article PubMed Google Scholar
King, T. E., Pardo, A. & Selman, M. Idiopathic pulmonary fibrosis. Lancet 378, 1949–1961 (2011).
Article PubMed Google Scholar
Roberts, A. B. et al. Transforming growth factor type β: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc. Natl. Acad. Sci. USA 83, 4167–4171 (1986).
Article CAS PubMed PubMed Central Google Scholar
Henry, E. C., Welle, S. L. & Gasiewicz, T. A. TCDD and a putative endogenous AhR Ligand, ITE, elicit the same immediate changes in gene expression in mouse lung fibroblasts. Toxicol. Sci. 114, 90–100 (2009).
Article PubMed PubMed Central Google Scholar
Lehmann, G. M. et al. The aryl hydrocarbon receptor ligand ITE inhibits TGFβ1-induced human myofibroblast differentiation. Am. J. Pathol. 178, 1556–1567 (2011).
Article CAS PubMed PubMed Central Google Scholar
Woeller, C. F., Roztocil, E., Hammond, C. L., Feldon, S. E. & Phipps, R. P. The aryl hydrocarbon receptor and its ligands inhibit myofibroblast formation and activation: implications for thyroid eye disease. Am. J. Pathol. 186, 3189–3202 (2016).
Article CAS PubMed PubMed Central Google Scholar
Gomez-Duran, A. et al. Fitting a xenobiotic receptor into cell homeostasis: how the dioxin receptor interacts with TGFβ signaling. Biochem. Pharm. 77, 700–712 (2009).
Article CAS PubMed Google Scholar
Haarmann-Stemmann, T., Bothe, H. & Abel, J. Growth factors, cytokines and their receptors as downstream targets of arylhydrocarbon receptor (AhR) signaling pathways. Biochem. Pharmacol. 77, 508–520 (2009).
Adamson, I. Y. & Bowden, D. H. The pathogenesis of bleomycin-induced pulmonary fibrosis in mice. Am. J. Pathol. 77, 185–197 (1974).
CAS PubMed PubMed Central Google Scholar
Moeller, A., Ask, K., Warburton, D., Gauldie, J. & Kolb, M. The bleomycin animal model: a useful tool to investigate treatment options for idiopathic pulmonary fibrosis?. Int. J. Biochem. Cell Biol. 40, 362–382 (2008).
Article CAS PubMed Google Scholar
Anderton, M. J. et al. Induction of heart valve lesions by small-molecule ALK5 inhibitors. Toxicol. Pathol. 39, 916–924 (2011).
Article CAS PubMed Google Scholar
Moriguchi, T. et al. Distinct Response to Dioxin in an Arylhydrocarbon Receptor (AHR)-Humanized Mouse. Proc. Natl. Acad. Sci. USA. 100, 5652–5657 (2003).
Schirle, M. & Jenkins, J. L. Identifying compound efficacy targets in phenotypic drug discovery. Drug Discov. Today 21, 82–89 (2016).
Article CAS PubMed Google Scholar
Schirle, M. & Jenkins, J. L. Contemporary techniques for target deconvolution and mode of action elucidation. in RSC Drug Discovery Series, Ch. 5, 83–103 (Royal Society of Chemistry, 2021).
Takeda, N. et al. Hopx expression defines a subset of multipotent hair follicle stem cells and a progenitor population primed to give rise to k6+ niche cells. Development 140, 1655–1664 (2013).
Article CAS PubMed PubMed Central Google Scholar
Madisen, L. et al. A robust and high-throughput Cre reporting and characterization system for the whole mouse brain. Nat. Neurosci. 13, 133–140 (2010).
Article CAS PubMed Google Scholar
Berg, S. et al. ilastik: interactive machine learning for (bio)image analysis. Nat. Methods 16, 1226–1232 (2019).
Article CAS PubMed Google Scholar
Carpenter, A. E. et al. CellProfiler: image analysis software for identifying and quantifying cell phenotypes. Genome Biol. 7, 1–11 (2006).
Beri, P. et al. A high-throughput cigarette smoke-treated bronchosphere model for disease-relevant phenotypic compound screening. PLoS ONE 18, e0287809 (2023).
Li, H., Qiu, J. & Fu, X. D. RASL-seq for massively parallel and quantitative analysis of gene expression. Curr. Protoc. Mol. Biol. 1, 4–13 (2012).
Godinez, W. J. et al. Compound activity prediction with dose-dependent transcriptomic profiles and deep learning. J. Chem. Inf. Model. (2023).
Behdenna, A., Haziza, J., Azencott, C.-A. & Nordor, A. pyComBat, a Python tool for batch effects correction in high-throughput molecular data using empirical Bayes methods. BMC Bioinformatics. 24, 459 (2023).
Ritchie, M. E. et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).
Article PubMed PubMed Central Google Scholar
R. Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. (2021).
Zheng, G. X. Y. et al. Massively parallel digital transcriptional profiling of single cells. Nat. Commun. 8, 14049 (2017).
McGinnis, C. S., Murrow, L. M. & Gartner, Z. J. DoubletFinder: doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 8, 329–337.e4 (2019).
Article CAS PubMed PubMed Central Google Scholar
Choudhary, S. & Satija, R. Comparison and evaluation of statistical error models for scRNA-seq. Genome Biol. 23, 27 (2022).
Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e29 (2021).
Article CAS PubMed PubMed Central Google Scholar
Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 36, 411–420 (2018).
Article CAS PubMed PubMed Central Google Scholar
Squair, J. W. et al. Confronting false discoveries in single-cell differential expression. Nat. Commun. 12, 5692 (2021).
Sergushichev, A. A. An algorithm for fast preranked gene set enrichment analysis using cumulative statistic calculation. https://doi.org/10.1101/060012.
Melsted, P. et al. Modular, efficient and constant-memory single-cell RNA-seq preprocessing. Nat. Biotechnol. 39, 813–818 (2021).
Article CAS PubMed Google Scholar
Wolf, F. A., Angerer, P. & Theis, F. J. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 1–5 (2018).
Bergen, V., Lange, M., Peidli, S., Wolf, F. A. & Theis, F. J. Generalizing RNA velocity to transient cell states through dynamical modeling. Nat. Biotechnol. 38, 1408–1414 (2020).
Article CAS PubMed Google Scholar
Haghverdi, L., Buettner, F. & Theis, F. J. Diffusion maps for high-dimensional single-cell analysis of differentiation data. Bioinformatics 31, 2989–2998 (2015).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

We thank D. Quackenbush and F. Lo for HTS imaging and analysis; M. Lindstrom and vivarium staff for animal caretaking; C. Trussell for flow sorting; and F. Luna for RNA sequencing. BioRender was used for generation of Fig. 1A (https://BioRender.com/lhzhpzr), Supplementary Fig. 1 (https://BioRender.com/dacbxlz) and Supplementary Fig. 8 (https://BioRender.com/k1x5zr0). This study was funded by Novartis Biomedical Research.

Author information

Rui Zhao
Present address: Pulmongene (Beijing) Ltd., Medical Technology Center, Beijing, China

Authors and Affiliations

Novartis Biomedical Research, San Diego, CA, USA
Andrew S. Hagan, Scott Williams, Casey J. N. Mathison, Shanshan Yan, Bao Nguyen, Glenn C. Federe, Guray Kuzu, Joseph C. Siefert, Janice Hampton, Victor Chichkov, S. Whitney Barnes, Frederick J. King, Brandon L. Taylor, John R. Walker, Rui Zhao, Jimmy Elliott, Dean P. Phillips, Bin Fang & Rebekah S. Decker

Authors

Andrew S. Hagan
View author publications
Search author on:PubMed Google Scholar
Scott Williams
View author publications
Search author on:PubMed Google Scholar
Casey J. N. Mathison
View author publications
Search author on:PubMed Google Scholar
Shanshan Yan
View author publications
Search author on:PubMed Google Scholar
Bao Nguyen
View author publications
Search author on:PubMed Google Scholar
Glenn C. Federe
View author publications
Search author on:PubMed Google Scholar
Guray Kuzu
View author publications
Search author on:PubMed Google Scholar
Joseph C. Siefert
View author publications
Search author on:PubMed Google Scholar
Janice Hampton
View author publications
Search author on:PubMed Google Scholar
Victor Chichkov
View author publications
Search author on:PubMed Google Scholar
S. Whitney Barnes
View author publications
Search author on:PubMed Google Scholar
Frederick J. King
View author publications
Search author on:PubMed Google Scholar
Brandon L. Taylor
View author publications
Search author on:PubMed Google Scholar
John R. Walker
View author publications
Search author on:PubMed Google Scholar
Rui Zhao
View author publications
Search author on:PubMed Google Scholar
Jimmy Elliott
View author publications
Search author on:PubMed Google Scholar
Dean P. Phillips
View author publications
Search author on:PubMed Google Scholar
Bin Fang
View author publications
Search author on:PubMed Google Scholar
Rebekah S. Decker
View author publications
Search author on:PubMed Google Scholar

Contributions

Conceptualization: D.P.P, B.F., R.S.D., and R.Z.; Investigation: S.W., C.J.N.M, S.Y., B.N., G.C.F., G.K., J.C.S., J.H., V.C., and S.W.B.; Methodology: F.J.K., B.L.T., J.R.W., D.P.P., B.F., R.S.D., and R.Z.; Formal analysis: S.W., C.J.N.M., S.Y., B.N, G.C.F., G.K., J.C.S., J.H., V.C., S.W.B., F.J.K., B.L.T., J.R.W., D.P.P., B.F., and R.S.D.; Visualization: A.S.H., B.F., and G.K.; Writing - original draft: A.S.H., B.F., and R.S.D.; Writing - review & editing: A.S.H., S.W., C.J.N.M., S.Y., B.N., G.C.F., G.K., J.C.S., J.H., V.C., S.W.B., F.J.K., J.R.W., J.E., D.P.P., B.F., and R.S.D.; Supervision: F.J.K., B.L.T., J.R.W., J.E., D.P.P., B.F., and R.S.D.

Corresponding authors

Correspondence to Bin Fang or Rebekah S. Decker.

Ethics declarations

Competing interests

The authors declare the following competing interests: All authors are current or former employees and shareholders of Novartis.

Peer review

Peer review information

Communications Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editor: Ophelia Bu. [A peer review file is available].

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Transparent Peer Review file

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Supplementary Data 5

Reporting Summary

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Hagan, A.S., Williams, S., Mathison, C.J.N. et al. Triggering AHR resolves TGF-β1 induced fibroblast activation and promotes AT1 cell regeneration in alveolar organoids. Commun Biol 8, 1025 (2025). https://doi.org/10.1038/s42003-025-08446-5

Download citation

Received: 11 July 2024
Accepted: 26 June 2025
Published: 09 July 2025
DOI: https://doi.org/10.1038/s42003-025-08446-5