Human Lung Organoid Model of Bronchiolitis Obliterans Syndrome Following Hematopoietic Cell Transplantation

Bronchiolitis obliterans syndrome (BOS), a manifestation of lung chronic graft-versus-host disease (cGVHD) following hematopoietic cell transplantation (HCT), is poorly understood because it is rare, has an insidious presentation, involves a protected site, and confers high mortality. Further, animal models incompletely recapitulate clinical disease, likely due to differences between human and murine bronchioles. We hypothesized that we could generate a new human lung organoid model for BOS investigations. This was feasible because our prior work established a unique core-shell architecture, wherein airway epithelial cells face outward, and fibroblasts are located internally, enabling direct interactions between the host lung epithelial and adoptively transferred donor immune cells.

We established an inverted co-culture system consisting of human primary airway epithelial cells and fibroblasts to recapitulate pre-HCT recipient airways, based on the minimal Matrigel scaffolding method described by Lee et al., (2024). Next, human leukocyte antigen (HLA)-mismatched mobilized peripheral blood stem cells (PBSCs) isolated from two healthy donors were introduced to the co-culture organoids in escalating doses (300, 3,000, and 30,000 cells). Organoids were then imaged using immunofluorescence microscopy for characterization. Bulk- and single-cell RNA sequencing (scRNAseq) were performed on cell lysates and live cell suspensions, respectively. DESeq2 package was used for differential gene expression (DGE) analysis, where the normalized counts were compared using the Wald test with Benjamini-Hochberg procedure to control false discovery rates. Graph-based unsupervised clustering was used to generate cell clusters using Pipseeker v3.2.

Advanced microscopy techniques confirmed a fibrous network of fibroblasts within the core of the organoids, with externally facing epithelia. Upon exposure to escalating doses of HLA-mismatched PBSCs, a dose-dependent epithelial disruption was observed. Subsequently, fibroblasts migrated to the external aspect of the organoid, followed by noticeable contraction, as confirmed by microscopy. Quantitative analysis of organoid diameter demonstrated a reduction post-PBSC exposure, with organoid sizes decreasing by 10.7 ± 4.5% (p<0.001) and 3.9 ± 4.5% (p<0.01) by day 2 after the addition of PBMCs from completely and partially HLA-mismatched PBSC donors, respectively. DGE analysis revealed the downregulation of Integrin beta-3 (ITGβ3) with increasing PBSC dose (p<0.0001). ITGβ3 inhibits fibroblast autophagy; its inhibition has been associated with fibroblast survival and proliferation. Additionally, matrix metalloproteinase-3 (MMP3) and MMP9, previously implicated in clinical BOS, were upregulated following PBSC exposure, showing increases of 2-3-fold and 10-15-fold, respectively, directly correlated with the PBSC dose (p<0.0001). Lastly, the unsupervised clustering of scRNAseq results successfully categorized key subtypes of airway epithelial cells (basal, secretory, and ciliated cells), fibroblasts, as well as PBSCs, validating the physiological complexity of the model.

Our human organoid model demonstrated PBSC dose-dependent epithelial lysis, and fibroblast extrusion and confluence in the epithelial deficit. Bulk-RNAseq analysis revealed that the previously identified BOS biomarkers of MMP3 and MMP9 were shown to be upregulated with increasing PBSC dose, while the marker constraining fibroblast proliferation (ITGβ3) exhibited an inverse correlation with PBSC dose. Our data support that our in vitro human lung co-culture organoid with PBSC exposure may serve as a model for BOS after HCT. However, further studies are needed to elucidate the ways in which this model mimics the clinical disease, with the goal of potentially interrogating disease pathogenesis and testing new targeted therapies.