A human lung alveolus-on-a-chip model of acute radiation-induced lung injury

Abstract

Acute exposure to high-dose gamma radiation due to radiological disasters or cancer radiotherapy can result in radiation-induced lung injury (RILI), characterized by acute pneumonitis and subsequent lung fibrosis. A microfluidic organ-on-a-chip lined by human lung alveolar epithelium interfaced with pulmonary endothelium (Lung Alveolus Chip) is used to model acute RILI in vitro. Both lung epithelium and endothelium exhibit DNA damage, cellular hypertrophy, upregulation of inflammatory cytokines, and loss of barrier function within 6 h of radiation exposure, although greater damage is observed in the endothelium. The radiation dose sensitivity observed on-chip is more like the human lung than animal preclinical models. The Alveolus Chip is also used to evaluate the potential ability of two drugs - lovastatin and prednisolone - to suppress the effects of acute RILI. These data demonstrate that the Lung Alveolus Chip provides a human relevant alternative for studying the molecular basis of acute RILI and may be useful for evaluation of new radiation countermeasure therapeutics.

Fig. 1. Human Lung Alveolus Chip recapitulates hallmark features of RILI.

a Schematic of the alveolus-on chip model (created with BioRender.com), showing the confocal z-stack illustrating endothelial tube formation. Scale bar = 100 μm. b Line diagram showing experimental plan, Dex, dexamethasone; ALI, air-liquid interface. c 53bp1 immunostaining (green) for double-stranded DNA breaks 2 h after radiation. DAPI counterstaining is shown in white. Scale bar = 20 μm. d Formation of 53BP1 foci was quantified per nucleus and showed a dose-dependent increase for both alveolar epithelial and endothelial cells. Data shown are mean +/− S.D. (analyzed nuclei per group n = 50 nuclei; one-way ANOVA DF = 5, F = 525, p < 0.05. e Immunostaining with ZO-1 (epithelial cells; green) and VE-cadherin junctions (endothelial cells; magenta) post irradiation, showed that junction disruption required a minimum dosage of 16 Gy. Scale bar = 20 μm. f Barrier function assay showed a 7-fold increase in the apparent permeability co-efficient (Papp) at 6 h post radiation exposure to 16 Gy, but no difference in response to 12 Gy. One-way ANOVA: n = 4 chips in each condition, DF = 2, F = 80.6, p < 0.05. g Representative comparison of cytokine levels, 24 h post-radiation shows an inflammatory response to 16 Gy radiation, in the presence of PBMCs. One-way ANOVA, n = 4 chips for each condition, Data shown are mean +/− S.D. F(IL-6) = 40.05, F(IL-8) = 7.74, F (ICAM-1) = 34.07, p < 0.05. h Heatmap showing cytokine response to radiation at 6 h, 24 h, 48 h and 7 d post radiation exposure, in the presence of PBMCs, n = 3 chips for each condition, p < 0.05. i % change in barrier integrity normalized to 0 Gy control over 7 days post-radiation exposure, in the presence of PBMCs. 2-way ANOVA, n = 4 chips in each condition, DF = 3, F = 19.3, Data shown are mean +/− S.D. Significance at p < 0.05.

Fig. 2. Radiation injury causes ROS, cellular hypertrophy and increased recruitment of PBMCs over time on the alveolus chip.

a 16 Gy radiation induces increase in the ROS levels, 2 h after radiation injury, n = 8 chips, F = 2.48, DF = 7, Data shown are mean +/− S.D. p < 0.001. b IF images showing the effect of radiation on cellular hypertrophy in the epithelium and endothelium 6 h and 7 d post radiation exposure, Scale bar = 20 μm. Corresponding quantification of cellular area in the epithelium and endothelium in c, t-test with Welch’s correction, where n = 50 cells in each condition, DF = 49, F (epi_6h) = 7.7, p < 0.05, F(epi_7d) = 5.7, F(endo_6h) = 9.8, F(endo_7d) = 8.3, Data shown are mean +/− S.D. d Quantification of PBMC recruitment to the vascular endothelium at 6 h and 7 d post-radiation exposure, n = 10 images taken from 3 biological replicate chips in each condition, 2-tailed unpaired t-test with F(6 h) = 104.4, F (7 d) = 2326, DF = 9, p < 0.0001. e Heatmaps comparing gene expression profiles in Transwells and chip (n = 3 transwells or chips), p < 0.05. The chip shows susceptibility to radiation that is evidenced by the overexpression of several markers. f Heatmap showing cytokine profiles of transwells at 6 h and 24 h post-radiation. At 6 h, the transwells exhibit an upregulation of IL-6, IL-8 and ICAM-1 but at 24 h, no effect of radiation exposure is observed. n = 3 transwells. g Quantification of % nuclei expressing nuclear foci in response to radiation in transwells vs chips. Data shown are mean +/− S.D. Two-tailed Student’s t test with Welch’s correction, n = 6 frames from 2 biological replicates, F = 1.88, DF = 5, p < 0.0001.

Fig. 3. Transcriptomic analyses shows that DNA damage and cell cycle arrest are early effects of radiation.

a Volcano plot of differentially expressed genes (DEGs) of alveolar epithelium, 6 h after radiation exposure, for adjusted p < 0.01. b Volcano plot of differentially expressed genes (DEGs) of endothelium, 6 h after radiation exposure, for adjusted p < 0.01. Adjusted p-values were obtained by applying Bonferroni correction for multiple comparison. c GSEA of Hallmark gene sets on the epithelium showing downregulation of hallmark E2F targets, mitotic spindle, and G2M checkpoints and upregulation of EMT, with q < 0.05. d GSEA of the endothelium showing downregulation of hallmark E2F targets, G2M checkpoints, and mitotic spindle, and upregulation of inflammatory responses. Plots created with Pluto (https://pluto.bio).

Fig. 4. Progressive inflammation and higher endothelial susceptibility to radiation-injury 7d post radiation.

a Volcano plot of differentially expressed genes (DEGs) of alveolar epithelium, 7 d after radiation exposure signaling, for adjusted p < 0.01. b Volcano plot of differentially expressed genes (DEGs) of the endothelium 7 d after radiation exposure, for adjusted p < 0.01. Adjusted p-values were obtained by applying Bonferroni correction for multiple comparison. c GSEA of Hallmark gene sets of the epithelium showing upregulation of hallmark interferon response, cytosolic DNA sensing pathways and senescence associated secretory pathways, heatmaps showing top genes affected in the upregulation of cytosolic DNA sensing and chemokine. d GSEA of Hallmark gene sets on the endothelium showing upregulation of inflammatory response, TNF-a signaling, complement activation. GSEA was also performed for gene sets of Reactome, KEGG and GO: biological process and showed upregulation of the cytokine expression and NF-kB signaling. Representative enrichment plots showing increased cytokine-cytokine receptor interaction, lymphocyte mediated immunity are shown here. Plots created with Pluto (https://pluto.bio).

Fig. 5. Identification of HO-1 as a therapeutic target for RILI and testing the effect of lovastatin and prednisolone on RILI.

a Gene expression levels of HMOX1 in the epithelium and endothelium at 6 h and 7d post radiation, Two-tailed Student’s t test with Welch’s correction n = 4 chips, F(epi 6 h) = 8.8, F(endo_6h) = 3.9, F(epi_7d) = 3.3, F(endo_7d) = 17.4, dF = 3, p < 0.001. b Pathway showing mechanism of action of HMOX1 towards Bilirubin formation accompanied by the release of Fe²⁺ and CO (created with BioRender.com). c Effect of lovastatin and prednisolone of HMOX1 RNA level, 6 h after radiation exposure, one-way ANOVA, n = 4 chips, F = 1.46, DF = 3, p < 0.05 and (d) on HO-1 protein, 24 h after radiation exposure. One-way ANOVA, n = 4 chips, F = 8.56, DF = 3, p < 0.05. e Quantification of 53BP1 foci/nucleus in the epithelium and (f) endothelium. n = 23 nuclei analyzed in 3 chips, Ordinary one-way ANOVA, significance at p < 0.0001, F (epi) = 580.5, F(endo) = 122.3, dF = 3. g Effect of lovastatin and prednisolone on cellular hypertrophy in the alveolar epithelium, n = 25 cells for each condition, Ordinary one-way ANOVA with Tukey’s test, significance at p < 0.05. h Effect of lovastatin and prednisolone on the cytokine response induced by radiation injury 24 h post-exposure. Representative panel showing the cytokines that were affected by the drug treatments. One-way ANOVA, n = 4 chips, DF = 3, F(IL-6) = 10.7, F(IL-8) = 10.3, F(ICAM-1) = 28.3 and F(TNF-a) = 12.0. i ROS assay 2 h after radiation exposure showing that presence of lovastatin decreases ROS levels in radiation exposed samples, n = 8 chips, one-way ANOVA, F = 2677, DF = 3, p < 0.05. j HMOX1 expression, 7 d post-radiation by qPCR, n = 4 chips, one-way ANOVA, F(epi) = 1.7, F(endo) = 282.3, DF = 3, p < 0.0001 and (k) HO-1 expression by ELISA assay, 7 post-radiation, indicating that HMOX1 remains upregulated in both epithelium and endothelium in the presence of lovastatin, n = 4 chips, one-way ANOVA, F(epi) = 418.9, F(endo) = 277.3, DF = 3, p < 0.0001. l Changes in barrier permeability over time from 6 h to 7d post-radiation. m HMOX1 expression by qPCR 7 d after treating the endothelium with control scrambled (control dsiRNA) or HMOX1 siRNA, n = 3 chips, one-way ANOVA, F = 45.0, DF = 2, p < 0.05. n Number of g-H2AX foci per nucleus, showing that knockdown of HMOX1 shows lower number of foci, n = 58 nuclei, One-way ANOVA, F = 43.7, DF = 3, p < 0.05. o Cytokine profile shows that some inflammatory cytokines are downregulated post-HMOX1 knockdown. One-way ANOVA, n = 3 chips, DF = 2, F(MIP-1a) = 21.0, F(IL-8) = 8.9, F(MCP-1) = 20.8 and F(IL-1b) = 20.1. All Data shown are mean +/− S.D.