henhaobio
Sign in

Science AdvancesBioengineeringPeer-reviewed journal article

Autologous human iPSC-derived Alveolus-on-Chip reveals early pathological events of M. tuberculosis infection

Chak Hon LukiD, Gabriel L. Conway, Kim Jee Goh, Antony FearnsiD, Irene Rodríguez Hernández, Nathan J. DayiD, Natalia AthanasiadiiD, Rocco D’AntuonoiD, Enrica PellegrinoiD, Janick D. Stucki, Nina HobiiD, Maximiliano Gabriel GutierreziD

PDFDOIPreprint
In-depth full-text digest available · Jump to digest ↓

Abstract

Immunocompetent and experimentally accessible alveolar systems to study human respiratory diseases are lacking. Here, we developed a single donor human induced pluripotent stem cell (iPSC)-derived Lung-on-Chip (iLoC) containing Type II and I alveolar epithelial cells, vascular endothelial cells, and macrophages in a microfluidic device that mimic lung 3D mechanical stretching and air-liquid interface. Imaging and scRNA-seq analysis revealed that the iLoC recapitulated cellular profiles present in the human distal lung. Infection of the iLoC with the human pathogen Mycobacterium tuberculosis (Mtb) showed that both macrophages and epithelial cells were infected and showed limited bacterial replication. Stochastically, large macrophage clusters containing necrotic core-like structure and Mtb replication were observed. A genetically engineered autophagy deficient iLoC revealed that after Mtb infection, macrophage necrosis was higher upon ATG14 deficiency without bacterial replication. Altogether, we report an autologous, genetically tractable human alveolar model to study lung diseases and therapies.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (17 Mar 2025) · CC BY 4.0

A single-donor iPSC-derived alveolar chip with epithelial, endothelial and macrophage cells models early Mycobacterium tuberculosis infection and shows ATG14-deficient macrophages die more without extra bacterial replication.

Immunocompetent, experimentally accessible human alveolar models built from one genetically tractable donor are scarce, and work combining iPSC with organ-on-chip for the distal lung is limited. This work assembles an autologous four-cell-type chip that can host isogenic gene-edited components.

Key results

  1. ShownThe chip forms a functional epithelial-endothelial barrier with iAT2 and iAT1 cells plus macrophages, and scRNA-seq annotation plus lung-atlas mapping supports alveolar-like identities, although key AT2/AT1 markers are expressed lower than in primary or human lung.

    TER ≥200 Ω·cm^2 with macrophages present; 26907 cells integrated across 4 conditions (n = 1 each); reference HLCA of over 584,000 cells and 61 cell identities; clustifyr Spearman correlation and Seurat label transfer. · Functional macrophages integrated into the…; The iLoC benchmarks human alveolar lung cells; Bioinformatic analysis of single-cell RNA…; Figure legends

  2. SuggestedAdding iVEC changed the scRNA-seq profile: macrophages resided more stably, epithelial cells showed more immune/antiviral pathway enrichment, and unique macrophage subclusters appeared. Breathing motion had a smaller effect on cell composition and lowered immune pathways in non-endothelial cells.

    iPSDM subclusters 2, 5, 6, 8 (iLoC-E) enriched only with iVEC; GSEA on Reactome (adjusted p < 0.05); 1 sample per condition, so no replicate-level variance is available. · The iLoC benchmarks human alveolar lung cells; Basal endothelial cells impact apical macrophage…; Bioinformatic analysis of single-cell RNA…; Figure legends

  3. ShownUnder a low-dose Mtb challenge, both macrophages and epithelial cells were infected but endothelium was not. Bacteria did not replicate appreciably in the first 48 h, whereas by 120 h total Mtb area rose and the barrier collapsed.

    MOI 0.01; infected iPSDM 2.82±1.78% and iAE 2.95±1.40% (Fig. 5d, n = 12-13 FOVs from 3 experiments); no significant replication by 48 h and a significant rise in total Mtb area at 120 h pi (Fig. 5f, n = 13-17 FOVs from 2-3 experiments); sharp TER decline at 120 h pi (Fig. 5h); one-way ANOVA. · The iLoC mimics early stages of infection with…; Figure legends

  4. ShownCytokine secretion during infection was polarized: apical IL-8, IP-10 and MCP-1 rose in the first 48 h with no basal increase, and no further rise occurred at 120 h, when cell death and barrier loss had set in.

    IL-8 elevated at 48 h pi vs uninfected; IP-10 and MCP-1 elevated at 48 h pi vs 2 h pi; Bio-Plex 48-plex panel; n = 2-6 independent experiments, one-way ANOVA. · The iLoC mimics early stages of infection with…; Cytokine measurements; Figure legends

  5. ShownIn a chip containing ATG14 knockout macrophages with wild-type epithelium and endothelium, macrophage death was higher, while intracellular Mtb load was comparable and no extra replication appeared, which suggests a bacterial-load-independent trigger of death.

    Higher basal and Mtb-induced Live/Dead-positive ATG14KO iPSDM at 2 h and 48 h pi; similar Mtb area per cell vs WT; larger infected macrophage fraction at 48 h; one KO line; n = 7-21 FOVs from 2-3 experiments. · A genetically engineered autophagy deficient…; Figure legends

Methods, data and reproducibility

Human iPSC were differentiated into lung progenitors (to iAT2/iAT1), endothelial cells and macrophages, then assembled on a stretchable membrane chip with air-liquid interface. Readouts were imaging, TER, cytokine panels, scRNA-seq and Mtb infection, with an ATG14KO macrophage variant.

  • iPSC-derived alveolar epithelial (iAT2/iAT1), vascular endothelial and macrophage cells on the AX12 chip; scRNA-seq of iAM and iLoC under static and breathing conditions · human · n = 26907 cells integrated; n = 1 per condition · GSE252601
  • Human iPSC lines: WTSIi018-B from a healthy donor was used for the figures; KOLF2 and an ATG14 knockout KOLF2 derivative are also described. · human · n = 3 iPSC lines used to date for iLoC generation
  • Reference healthy Human Lung Cell Atlas (Sikkema et al. 2023) for label transfer and correlation · human · n = over 584,000 cells, 61 cell identities
  • Mtb H37Rv infection of the iLoC from the apical side, 2 h, 48 h and 120 h post infection · Mycobacterium tuberculosis · n = n = 12-13 FOVs from 3 experiments (Fig. 5d)
Methods & tools
Cell Ranger v.6.0.1, Seurat v4, clustifyr v1.10.0, ClusterProfiler v.4.6.2, MSigDB C2-CP:REACTOME v2023.1, DESeq2 test (Seurat FindMarkers), Cellpose, btrack, napari, FIJI, QuPath 0.4.4, Prism v.10.0.1
Reagents & environment
R-4.2.0; Cell Ranger v.6.0.1 (GRCh38-2020-A); Seurat v4; clustifyr v1.10.0; ClusterProfiler v.4.6.2; MSigDB v2023.1; FlowJo v10.8.1; FIJI 2.9.0/1.54f; QuPath 0.4.4.
Reproducibility gaps
Code is said to be on GitHub, but no URL appears in the text (DOIs are listed in a key resources table). Microscopy data are available only on request from the lead contact.
Limitations (5)
  • The chip has mechanical stretch but no fluidic flow and lacks fibroblasts and neutrophils, which limits physiological and immune relevance. · stated by the authors
  • Chip assembly is complex and time-sensitive because no long-term storage or cryopreservation approach exists; key alveolar markers were expressed lower than in primary alveolar cells. · stated by the authors
  • At 120 h pi the ZO-1 barrier was disrupted, so single-cell segmentation was no longer reliable and a total Mtb area readout was used instead. · stated by the authors
  • The scRNA-seq comparisons of iVEC and breathing effects rest on one sample per condition, so replicate variance is not estimated and the effects remain preliminary. · noted in this summary
  • The ATG14 conclusion rests on one knockout line and a single gene; the figures use one healthy-donor iPSC line, so generality across donors is untested. · noted in this summary

Read this if you build or use iPSC-derived organ-on-chip models, or study early tuberculosis and autophagy in a human alveolar setting with scRNA-seq benchmarking.

Citations
10
FWCI
44.57
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISDeboosère '26Ektnitphong '25Thacker '20Yamagata '26Iakobachvil '21Pahari '23
80% semantic matchMicrobiology
Nathalie Deboosère, Yoel Dagan, Aurélie Burette+6 more·bioRxiv·Apr 30, 2026· 1 citations