Alveolar stem cells transdifferentiate to drive bronchiolar regeneration

Chronic and acute pulmonary diseases (such as chronic obstructive pulmonary disease, asthma, pneumonia and bronchiectasis) involve bronchial epithelial damage that impairs airway regeneration and worsens progression^{5,6,7,8,9,10,11}. Understanding the cellular and molecular mechanisms of the epithelial repair is therefore critical for targeted therapies. The lung epithelium comprises distinct compartments—trachea, bronchioles and alveoli—each with specialized progenitors that maintain homeostasis and repair^5,12,13,14. In the trachea, basal cells serve as multipotent stem cells^{15,16,17,18,19,20,21} that sustain the epithelial lining. The bronchioles contain club cells and neuroendocrine (NE) cells that drive local regeneration^22,23,24,25. The alveolar compartment relies on alveolar type II (AT2) cells that self-renew and differentiate into AT1 cells to maintain gas exchange^{26,27,28,29,30,31}. Bronchoalveolar stem cells (BASCs), which are located at the bronchioalveolar-duct junction (BADJ) and co-express Scgb1a1 and Sftpc, show multipotency across both bronchiolar and alveolar compartments^32,33,34,35. Recent studies also reveal club cell plasticity in alveolar repair^1,2,3,4.

While airway progenitors contribute to alveolar repair, the reciprocal capacity of alveolar cells to participate in bronchiolar regeneration remains unclear. In vitro studies suggest that human AT2 cells can differentiate into airway basal cells³⁶ and transdifferentiate into bronchiolar secretory cells, a phenomenon that is also observed in non-human primate injury models³⁷. However, genetic manipulation limitations in these systems have prevented definitive in vivo evidence for this transcompartmental repair. In mouse models, conventional Sftpc-creER lineage tracing lacks the specificity to accurately track AT2 cell fate after bronchiolar injury, because both AT2 and BASCs express Sftpc³², resulting in simultaneous labelling. Given that BASCs actively contribute to bronchiolar repair^33,34,35, this co-labelling confounds interpretation of AT2 cell-specific contributions. Thus, new genetic tools for selective AT2 tracing are needed to conclusively determine whether alveolar cells contribute to bronchiolar regeneration in vivo.

In this study, we established an AT2-specific in vivo labelling system using a dual-gene-marker-mediated lineage tracing strategy. Using this approach in bronchiolar injury models, we found that AT2 cells can migrate across anatomical boundaries and transdifferentiate into functional club and ciliated cells within bronchioles, contributing to cross-compartmental epithelial repair. Mechanistically, this process depends on integrin-β1-mediated migration and Notch-dependent fate specification. After bronchial injury, infiltrating immune cells secrete SPP1 to direct AT2 cell migration towards injured bronchioles, where residual ciliated cells activate Notch signalling to drive AT2-to-club transdifferentiation.

We first identified the origin of bronchiolar regeneration after naphthalene injury. We observed SFTPC⁺ AT2 cells within the bronchiolar epithelium after injury, suggesting their potential involvement in repair (Extended Data Fig. 1a,b). To track AT2 cells, we crossed Sftpc-creER (ref. ³⁸) with R26-tdTomato (R26-tdT) mice. At 4 weeks after injury, SFTPC-labelled cells migrated into bronchioles and transdifferentiate into club and AT1 cells (Extended Data Fig. 1c–e). However, SFTPC⁺ cells comprise two distinct cell types: SFTPC⁺ AT2 cells and SFTPC⁺SCGB1A1⁺ BASCs (Extended Data Fig. 1f,g). This distinction is crucial because BASCs, which are naphthalene resistant, can transdifferentiate into club cells after injury^33,34,35 (Extended Data Fig. 1h–j). Thus, single-recombinase systems like Sftpc-creER cannot distinguish these populations, potentially confounding AT2-specific interpretation.

To determine AT2 cell-specific contributions, we first generated a Scgb1a1-DTR-GFP knock-in line by inserting diphtheria toxin receptor (DTR) and GFP sequence into the Scgb1a1 locus (Extended Data Fig. 1k). In this system, all Scgb1a1-expressing cells (club cells and BASCs) should co-express GFP and DTR, enabling diphtheria toxin (DT)-mediated ablation. Immunostaining confirmed GFP/DTR expression in all SCGB1A1⁺ populations in vehicle-treated mice, while DT treatment effectively eliminated GFP⁺ cells, validating system functionality (Extended Data Fig. 1l–p). We next generated triple transgenic mice: Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP (Fig. 1a,b). As predicted, tamoxifen treatment labelled AT2 cells (tdT⁺ only), BASCs (tdT⁺GFP⁺) and club cells (GFP⁺ only). We refer to this system as AT2-tracer1. These mice were treated with tamoxifen and DT, and lung tissue was collected for analysis (Fig. 1c).

DT administration selectively eliminated GFP⁺ populations, leaving only tdT⁺ AT2 cells in alveoli (Fig. 1d,e). After 4 weeks of recovery, lineage tracing revealed that tdT-labelled AT2 cells migrated to bronchioles and generated tdT⁺GFP⁺ club cells (Fig. 1f). These AT2-derived cells transdifferentiated exclusively into E-cadherin (ECAD)-positive epithelial cells, including club cells, AT1 cells, ciliated cells and undifferentiated AT2 cells (Fig. 1g). Quantification showed that 77.42 ± 4.71% of new club cells originated from AT2 cells (Fig. 1h). AT2-derived cells in bronchioles comprised 38.18 ± 4.55% club cells, 14.61 ± 2.62% undifferentiated AT2 cells, 18.90 ± 4.48% AT1 cells and 0.27 ± 0.07% ciliated cells (Fig. 1i). AT2-derived tdT⁺GFP⁺ club cells often localized near residual ciliated cells, whereas ciliated-free areas were covered by non-club cells, including tdT⁺ flattened AT1-like cells (Extended Data Fig. 1q–t). We also detected rare AT2-derived GFP⁺SFTPC⁺ BASC-like cells at days 7, 14 and 28 after DT (0.87 ± 0.15%, 0.70 ± 0.16% and 0.33 ± 0.08%, respectively; Extended Data Fig. 1u,v). These data provide conclusive evidence that AT2 cells actively migrate and participate in bronchiolar epithelial repair.

We further identified two migratory routes for tdT⁺ AT2-derived cells towards injured bronchioles: through the BADJ (route 1) and through bronchiolar wall channels (route 2) (Fig. 1j and Extended Data Fig. 2a–g). Immunostaining demonstrated that these channels were gaps in the bronchiolar wall, which were lined by tdT⁺ AT2-derived AT1 cells and connected the airway lumen to the alveolar spaces (Extended Data Fig. 2f–h). tdT⁺ coverage of bronchiolar epithelium increased over time (Extended Data Fig. 2e), and both routes remained observable throughout the time (Extended Data Fig. 2f,g). Quantitative analysis revealed that route 1 events increased rapidly from day 3 to 7, whereas route 2 events, though detectable at day 3, showed only a modest increase by day 7 (Fig. 1k). These data suggest that AT2 cells contribute mainly through route 1, with route 2 serving as a supplementary pathway (Fig. 1l).

To investigate whether AT2 cells contribute at the single-cell level, we generated Sftpc-creER;R26-Confetti;Scgb1a1-DTR-GFP mice (Extended Data Fig. 2i–k). Our results showed that individually labelled AT2 cells possess multipotent transdifferentiation capacity during bronchiolar repair, including club and AT1 cells (Extended Data Fig. 2k,l). Quantitative analysis showed AT2-derived bronchiolar populations comprised 33.69 ± 7.90% club, 17.17 ± 3.42% undifferentiated AT2 and 24.24 ± 4.78% AT1 cells (Extended Data Fig. 2m). Thus, single AT2 cells can transdifferentiate into multiple lineages during DT-induced bronchiolar repair (Extended Data Fig. 2n).

Basal cells, which serve as airway epithelial stem cells, are located primarily in the trachea, with minimal presence in the distal bronchioles, and have a critical role in epithelial regeneration^15,16,17,18. To trace basal cell fate, we used p63-dreER;R26-RSR-tdT (R26-rox-stop-rox-tdTomato) mice, in which proximal basal cells were efficiently labelled with tdT. Rare tdT⁺ basal cells were also detected in the distal bronchioles (Extended Data Fig. 3a–c). We then used p63-dreER;R26-RSR-tdT;Scgb1a1-DTR-GFP mice to simultaneously trace basal cells (tdT⁺) and ablate club cells (GFP⁺; Extended Data Fig. 3d). After tamoxifen induction and DT treatment, we analysed lungs after a 4-week repair period (Extended Data Fig. 3e). Immunostaining revealed that nearly all regenerated GFP⁺ club cells in the trachea were tdT⁺, confirming their basal cell origin. Moreover, rare basal cells in the distal bronchioles underwent clonal expansion and contributed a small number of club cells (Extended Data Fig. 3f,g). These findings indicate that basal cells primarily regenerate proximal airways. Furthermore, we detected SCGB3A2⁺SCGB1A1⁻ intralobular serous (IS) cells in the bronchioles, consistent with previous reports³⁹ (Extended Data Fig. 3h,i). These cells were rare in number but proliferated after DT injury, suggesting that they may also participate in club cell regeneration (Extended Data Fig. 3j–m).

As NE cells have been implicated in club cell regeneration^24,25, we tested their contribution after DT injury using Ascl1-creER;R26-tdT;Scgb1a1-DTR-GFP mice, which specifically label CGRP⁺ NE cells (tdT⁺) while ablating club cells (GFP⁺; Extended Data Fig. 3n–p). At 4 weeks after injury, immunostaining and quantification detected a small population of tdT⁺GFP⁺ club cells, accounting for 4.59 ± 1.36% of distal club cells (Extended Data Fig. 3q–s). Thus, NE cells contribute a small subset of club cells after injury (Extended Data Fig. 3t).

Together with previous findings, these results demonstrate that club cell regeneration is compartment specific, driven primarily by AT2 cells and basal cell, with minor contributions from NE cells or possible IS cells.

To study AT2-mediated regeneration physiologically, we used naphthalene instead of DT ablation. As naphthalene-resistant cells (such as BASCs) persist after injury^22,33,40, conventional single-recombinase AT2 tracing cannot be used due to potential confounding labelling. To overcome this, we used a dual-recombinase system with the nested reporter (R26-NR)⁴¹, in which Dre-rox and Cre-loxP recombinations activate ZsGreen (ZsG) and tdT, respectively⁴¹ (Fig. 2a). We generated Scgb1a1-creER;Sftpc-dreER;R26-NR mice to trace AT2 cells independently of BASCs. Here Sftpc-dreER labels all SCGB1A1^–SFTPC⁺ AT2 cells with ZsG (termed AT2-tracer2), while Scgb1a1-creER marks SCGB1A1⁺SFTPC^– club cells and SCGB1A1⁺SFTPC⁺ BASCs with tdT (Fig. 2a,b). Tamoxifen induced efficient and specific ZsG labelling of AT2 cells, while BASCs converted to tdT (Extended Data Fig. 4a–c), validating system specificity.

We then administered tamoxifen and naphthalene to mice and analysed lungs at 4 weeks (short term) and 6 months (long term) after injury (Fig. 2c). Immunostaining revealed that 42.93 ± 5.60% of regenerated club cells were ZsG⁺ at 4 weeks, confirming AT2-tracer2 origin (Fig. 2d,e). Besides club cells, these ZsG⁺ cells also contributed to AT2, AT1 and ciliated cells within the bronchiolar epithelium (Fig. 2f and Extended Data Fig. 4d), consistent with AT2-tracer1 findings (Fig. 1h,i). AT2-tracer2-derived bronchiolar cells were mainly club (62.31 ± 4.98%), with fewer AT2 (9.45 ± 1.43%), AT1 (6.63 ± 1.91%) and ciliated (0.79 ± 0.13%) cells (Fig. 2g). After naphthalene injury, no channel structures formed in the bronchiolar wall; instead, AT2 cells contributed through BADJ, with their epithelial coverage increasing over repair time (Extended Data Fig. 4e,f). At 6 months after injury, most AT2-derived bronchiolar cells were club or ciliated, although some still expressed SFTPC and AGER (Extended Data Fig. 4g). After secondary naphthalene injury, AT2-derived club cells proliferated and transdifferentiated into ciliated cells, contributing to bronchiolar repair (Extended Data Fig. 4h–l). Single-cell multi-omics analysis further revealed that AT2-derived club cells converged with native airway cells at both transcriptional and epigenetic levels, and acquired corresponding functional features (Fig. 2h–l and Supplementary Fig. 1a). The AT2-specific epigenetic signature rapidly declined after transdifferentiation and remained low from 4 weeks to 6 months after injury, whereas the club-cell-specific epigenetic and secretory/immune signatures increased progressively during repair (Fig. 2j–l). Collectively, these data show that AT2 cells participate in bronchial epithelial repair after naphthalene injury (Fig. 2b).

To evaluate the relative contributions of club cell, AT2 cell and BASC progenitors to airway regeneration in a single mouse model, we used the R26-TLR³ dual-recombinase system, as regenerated club cells may originate from club cell expansion, BASCs or AT2 cell differentiation. Using Scgb1a1-creER;Sftpc-dreER;R26-TLR mice, we established three distinct tracing populations: club-tracer (tdT⁺ZsG^– club cells), AT2-tracer3 (tdT^–ZsG⁺ AT2 cells) and BASC-tracer (tdT⁺ZsG⁺ BASCs) (Fig. 2m,n). Immunostaining confirmed specific labelling (Extended Data Fig. 4m–o).

We analysed lungs at 4 weeks (short term) and 6 months (long term) after injury (Fig. 2o). At 4 weeks, quantitative analysis revealed that AT2-tracer3 contributed the majority (52.35 ± 7.09%) of regenerated club cells, followed by BASC-tracer (26.66 ± 6.15%) and club-tracer (14.89 ± 3.72%; Fig. 2p,q). Notably, AT2-tracer3-derived cells undergo cell proliferation and exhibit multiple differentiation potential, giving rise to club cells (61.30 ± 5.92%), AT2 cells (5.09 ± 1.19%), AT1 cells (5.10 ± 1.12%) and ciliated cells (0.87 ± 0.14%), but not mesenchymal cells (Fig. 2r and Extended Data Fig. 4p,q), consistent with our AT2-tracer2 data (Fig. 2g). At 6 months, these cell types remained within the bronchiolar epithelium (Extended Data Fig. 4r). These results establish AT2 cells as a major cellular source for distal bronchiolar epithelial repair (Fig. 2s).

We next performed single-cell RNA-sequencing (scRNA-seq) analysis of AT2-tracer3, BASC-tracer and club-tracer cells from Scgb1a1-creER;Sftpc-dreER;R26-TLR mice at 4 weeks after naphthalene injury (Fig. 3a and Extended Data Fig. 5a). Analysis of AT2-tracer3 cells revealed five cell clusters: cluster 1 (mature AT2), cluster 2 (transient), cluster 3 (club-like), cluster 4 (pre-AT1) and cluster 5 (AT1) (Fig. 3b and Extended Data Fig. 5b,c). Pseudotime analysis revealed a bifurcating AT2-derived pathway through a CLDN4⁺ intermediate to club cells or pre-AT1 to AT1, confirming bipotency (Fig. 3b). Club cell specification correlated with Sftpc downregulation and Scgb1a1 upregulation, with peak Cldn4 in the transient state (Fig. 3c,d). Functional enrichment differed: transient cells upregulated wound response, apoptosis and migration (Fig. 3e), whereas club cells enriched xenobiotic metabolism, inflammation and Notch signalling (Fig. 3f).

scRNA-seq analysis of BASC-tracer cells revealed four distinct clusters: cluster 1 (canonical BASCs), cluster 2 (transient intermediates), cluster 3 (club cell) and cluster 4 (ciliated cells) (Extended Data Fig. 5d–f). Pseudotime analysis showed a unidirectional trajectory: BASCs to transient cells to club cells to ciliated cells (Extended Data Fig. 5e). Transcriptomically, the BASC-to-club-cell transition involved injury response, migration, naphthalene catabolism and Notch signalling, with glutathione metabolism additionally upregulated in mature club cells (Extended Data Fig. 5g,h). Transcriptomic comparison revealed key differences in AT2- versus BASC-to-club transdifferentiation (Extended Data Fig. 5i–l and Supplementary Table 1). At initiation, BASCs showed stronger club cell and weaker AT2 cell signatures compared with AT2 cells. In the transient state, BASCs maintained a higher club cell signature, but the Notch and PATS²⁹ signatures did not differ from AT2-tracer3 cells. Although both generated club cells, AT2-derived club cells retained AT2 and PATS expression signature, in contrast to BASC-derived counterparts. To assess origin-dependent long-term fate, we performed scRNA-seq analysis of lineage-traced cells at 6 months after injury (Fig. 3a). AT2-tracer3 yielded five clusters (residual AT2, regenerated club, ciliated, pre-AT1 and AT1), whereas BASC-tracer and club-tracer populations each gave three clusters (club, ciliated and goblet cells) (Extended Data Fig. 5m). Transient progenitors from acute repair were no longer detectable. Although AT2-derived club cells retained distinct signatures at 4 weeks, they converged with BASC- and club-cell-derived club cells by 6 months, losing residual PATS and alveolar traits (Extended Data Fig. 5n,o). Thus, prolonged repair resolves lineage-specific differences through maturation to a common end point, consistent with previous epigenetic analyses (Fig. 2h–l).

Our findings suggest that the CLDN4⁺ intermediate state can transdifferentiate into club and AT1 cells after injury (Fig. 3d). To test this, we first generated Cldn4-LSL-dre (Cldn4-loxP-stop-loxP-dre) mice (Extended Data Fig. 6a). We also generated Lamp3-creER (AT2-tracer4) for efficient AT2 labelling (Extended Data Fig. 6b–e). We then generated Lamp3-creER;Scgb1a1-DTR-GFP;Cldn4-LSL-dre;R26-RSR-LSL-tdT mice to trace AT2-derived CLDN4⁺ transient cells (Fig. 3g,h). The reporter requires dual recombination: Cre excises LSL from both Cldn4-LSL-dre and the reporter and, subsequently, injury-induced Cldn4 gene activation drives expression of Dre to excise RSR cassette, thereby labelling AT2-derived CLDN4⁺ cells with tdT (Fig. 3g). Club cells derived from these transient cells co-express tdT and GFP (yellow). Four weeks after DT injury, control (PBS treated) mice showed minimal tdT signal, whereas injured mice exhibited abundant tdT⁺GFP⁺ club cells (Fig. 3i). We also detected tdT⁺AGER⁺ AT1 cells and residual tdT⁺SFTPC⁺ AT2 cells in the bronchiolar epithelium (Fig. 3j). Quantification showed that 62.10 ± 6.16% of new club cells originated from CLDN4⁺ transient cells (Fig. 3k). Among tdT⁺ lineages, club cells, AT1 cells and AT2 cells accounted for 56.55 ± 4.54%, 14.05 ± 3.13% and 7.48 ± 2.06%, respectively (Fig. 3l). To further validate the functional role of CLDN4 during transition, we generated Cldn4-flox mice and crossed them with Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP mice (Fig. 3m and Extended Data Fig. 6f). At 4 weeks after DT, Cldn4 knockout (KO) significantly reduced the proportion of AT2 cells contributing to club cells (control versus KO: 65.94 ± 6.50% versus 25.49 ± 3.54%) and AT1 cells (21.11 ± 2.74% versus 8.82 ± 1.56%) (Fig. 3m–p and Extended Data Fig. 6g). Collectively, these data suggest that CLDN4⁺ intermediate cells derived from AT2 cells not only contribute to the bronchiolar epithelium but also functionally promote AT2-cell-mediated epithelial repair.

We found that Notch signalling is upregulated in AT2-tracer3-derived club cells and BASC-tracer-derived transient cells, suggesting its role in fate transition (Fig. 3f and Extended Data Fig. 5h). To functionally validate this observation, we developed a system combining lineage tracing with conditional Rbpj KO (Rbpj^rox/+) using Dre-rox recombination in SFTPC⁺ cells (Extended Data Fig. 6h). Using Scgb1a1-creER;Sftpc-dreER;R26-TLR;Rbpj^rox/rox (mutant) and Scgb1a1-creER;Sftpc-dreER;R26-TLR;Rbpj^rox/+ (control) mice, we traced and disrupted Notch signalling simultaneously in AT2-tracer3 and BASC-tracer cells (Fig. 4a and Extended Data Fig. 6i,j). After tamoxifen induction and naphthalene injury, a 4 week repair period study revealed that Rbpj-KO cells migrated to the bronchioles but did not express club cell markers, while retaining alveolar markers instead (Fig. 4b and Extended Data Fig. 6k). Notably, the proportion of club cells derived from AT2-tracer3 and BASC-tracer was significantly reduced in mutants than that in the controls (AT2-tracer3: 1.86 ± 0.80% versus 38.09 ± 5.29%; BASC-tracer: 6.48 ± 2.08% versus 25.97 ± 4.97%; Fig. 4c). Furthermore, among the bronchiolar cells from AT2-tracer3, mutants exhibited significantly fewer club cells compared with the controls (3.16 ± 1.60% versus 55.89 ± 6.47%), but more alveolar cells (AT2 cell: 18.14 ± 2.70% versus 7.12 ± 1.62%; AT1 cell: 12.07 ± 3.09% versus 6.95 ± 0.96%; Fig. 4d). Independent AT2-tracer2 tracing confirmed that Rbpj-deficient AT2 cells did not generate club cells during bronchiolar repair (Extended Data Fig. 6l–n). The above results demonstrate that Notch signalling is essential for AT2-to-club cell conversion. Rbpj deletion not only blocks this transition but also leads to aberrant accumulation of alveolar cells in bronchioles (Fig. 4e).

To comprehensively assess Notch’s role in modulating cell fate, we developed complementary loss- and gain-of-function models using DT-induced injury with AT2-tracer1. DT-induced club cell depletion caused bronchial wall thickening, which subsides as the club cells were repaired (Extended Data Fig. 6o,p). In loss-of-function experiments (Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP;Rbpj^flox/flox mice), Rbpj deletion in AT2 cells significantly reduced club cell formation (mutant (3.95 ± 1.83%) versus control (31.54 ± 7.39%)) while increasing alveolar differentiation (AT2: mutant (38.92 ± 8.08%) versus control (19.50 ± 3.23%); AT1: mutant (29.46 ± 3.95%) versus control (20.26 ± 4.25%); Fig. 4f–h and Extended Data Fig. 6q,r). Conversely, Notch overexpression (Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP;R26-NICD) enhanced AT2-to-club-cell conversion (45.19 ± 4.69% in mutant) while suppressing alveolar fates (AT2: 10.07 ± 3.03% in mutant; AT1: 10.37 ± 2.94% in mutant; Fig. 4f–h and Extended Data Fig. 6q,r). These complementary approaches demonstrate that Notch signalling functions as a critical bidirectional regulator of AT2 cell fate decisions during bronchiolar repair (Fig. 4i). Rbpj KO increased bronchiolar wall thickness and reduced small airway ventilation (Extended Data Fig. 6s,t), indicating that Notch-mediated club cell regeneration is essential for structural and functional recovery. tdT⁺GFP⁻ AT2 cells sorted from DT-injured lungs did not transdifferentiate into club cells in club medium alone but did so substantially after Notch activation, indicating that achieving AT2-to-club conversion in vitro requires strong drivers such as Notch signalling (Extended Data Fig. 6u–z and Supplementary Fig. 1b).

scRNA-seq analysis revealed that migrating AT2 cells require local Notch signals for club cell transdifferentiation (Fig. 3f). After injury, ciliated cells were the dominant Notch-ligand source (high JAG1/2) among diverse epithelia (Extended Data Fig. 7a–c). Interaction analysis confirmed JAG1/2–Notch1/2 signalling between ciliated and AT2-derived cells (Fig. 4j), with similar networks in BASC lineages (Extended Data Fig. 7d–f), suggesting a conserved mechanism. To functionally validate these findings, we developed a genetic system that combines AT2 tracing (Sftpc-dreER;R26-RSR-tdT, designated as AT2-tracer5) with conditional ablation of both club cells/BASCs (through Scgb1a1-DTR-GFP) and ciliated cells (through Foxj1-creER;R26-iDTR; Extended Data Fig. 7g–i). We then established two mouse groups: Sftpc-dreER;R26-RSR-tdT;Scgb1a1-DTR-GFP (control mice) and Sftpc-dreER;R26-RSR-tdT;Scgb1a1-DTR-GFP;Foxj1-creER;R26-iDTR (experimental mice). This design achieved ablation of club cells and BASCs in both groups, selective ciliated cell ablation in experimental mice and AT2 cell tracing in all animals. At 4 weeks after injury, ciliated ablation substantially reduced AT2-to-club-cell conversion (experimental: 3.84 ± 2.25%; control: 25.71 ± 4.87%; Extended Data Fig. 7i–l). Spatial analysis provided further mechanistic insights, showing that successfully converted AT2-derived club cells (tdT⁺GFP⁺) predominantly localized adjacent to ciliated cells, while unconverted AT2-derived cells (tdT⁺GFP^–) typically lacked this close association (Extended Data Fig. 7m–o).

Given the strong JAG1–NOTCH2 association, we deleted this pair respectively in ciliated and AT2-derived cells to test whether ciliated cells regulate AT2 fate through Notch (Extended Data Fig. 7p). First, we generated experimental (Foxj1-creER;Jag1^flox/flox) and control (Foxj1-creER;Jag1^flox/+) mice that also carried Sftpc-dreER;R26-RSR-tdT;Scgb1a1-DTR-GFP (Fig. 4k). After tamoxifen induction, Jag1 was deleted in ciliated cells; AT2 cells were labelled with tdT, club cells with GFP and BASCs with both. DT was then administered to ablate club cells and BASCs (Fig. 4k). Immunostaining and quantification at 4 weeks after DT treatment showed that Jag1 deletion significantly reduced AT2-derived club cells in the bronchioles (Fig. 4k–n and Extended Data Fig. 7q–s), demonstrating that ciliated cells regulate AT2-to-club-cell transdifferentiation through the Notch ligand Jag1. To investigate Notch2 function, we generated conditional Notch2-flox mice (Extended Data Fig. 7t,u). For specific deletion in AT2-derived cells (including AT2 cells and transient cells), we produced experimental Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP;Notch2^flox/flox mice and controls with Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP;Notch2^flox/+ (Fig. 4o). Immunostaining at 4 weeks after DT treatment showed that Notch2 deletion significantly inhibited AT2-to-club-cell transition (Fig. 4o–r and Extended Data Fig. 7v), consistent with the Rbpj-deletion results (Fig. 4h). Collectively, these findings demonstrate that Notch signalling from ciliated cells critically regulates AT2-to-club-cell transition (Fig. 4s).

To investigate how AT2 cells migrate in response to bronchial injury, we performed scRNA-seq analysis of PBS- and DT-treated Scgb1a1-DTR-GFP lungs (Supplementary Fig. 1c). DT injury induced neutrophil infiltration and peribronchiolar fibroblast expansion, creating an inflammatory and stromal microenvironment (Extended Data Fig. 8a,b). CellChat analysis revealed global signalling increase after injury, with neutrophils, monocytes and peribronchiolar fibroblasts becoming major signal sources towards AT2 cells (Fig. 5a and Extended Data Fig. 8c). Among chemotactic pathways, SPP1 signalling from immune cells was robustly induced (Extended Data Fig. 8d). Peribronchiolar fibroblast-to-AT2 signals were dominated by structural, adhesion and morphogen pathways, rather than by generating migratory cues (Extended Data Fig. 8d–f). SPP1 is a secreted chemotactic factor that binds to integrins and CD44 to promote migration^42,43,44. DT injury significantly enhanced interactions between immune-derived SPP1 and integrin heterodimers on AT2 cells (Fig. 5b), correlating with upregulation of migration and S-phase gene signatures (Extended Data Fig. 8g,h and Supplementary Table 1). These data suggest that SPP1-mediated chemotaxis drives AT2 recruitment to injured bronchioles.

To gain spatial resolution, we performed spatial transcriptomics analysis of DT-injured lungs (at day 7) using wild-type mice as controls (Extended Data Fig. 8i). DT injury induced a localized approximately threefold increase in peribronchiolar immune infiltration (Fig. 5c). Hotspot analysis revealed that, under homeostatic conditions, no modules associated with leukocyte activation or chemotaxis were enriched near bronchioles, whereas DT-injured lungs exhibited such modules (for example, module 9) concentrated in the bronchiolar lumen and peribronchiolar areas (Fig. 5d,e and Extended Data Fig. 8j,k). Module 9 contained pro-inflammatory and chemokine genes (Spp1 and Ccl6). Spp1 was upregulated in peribronchiolar immune cells after DT injury (Fig. 5d), and chemotaxis-related gene sets were enriched in these cells (Fig. 5f). Together, these spatial data indicate that bronchiolar injury triggers localized immune cell accumulation and secretion of chemotactic factors like SPP1, potentially establishing a gradient to direct AT2 migration.

To test this hypothesis, we validated SPP1 expression in immune cells by immunostaining. After DT injury, many CD45⁺ immune cells infiltrated the bronchiolar lumen and showed high SPP1 expression (Fig. 5g,h). Infiltration peaked within the first 7 days (early response stage) and then declined (late stage; Fig. 5i and Extended Data Fig. 9a,b). Using AT2-tracer1 mice, tdT⁺ AT2 lineage cell migration began during the early stage and coverage increased over time (Fig. 5j (red line) and Extended Data Fig. 9c,d). Immune infiltration peaked during the early phase (blue line), suggesting that SPP1 from these cells induces AT2 migration (Fig. 5j). Injury also induced Mmp14 in AT2 cells, fibroblasts and SMCs, and Mmp9 in immune cells, potentially degrading matrix to facilitate AT2 cell migration (Extended Data Fig. 9e–g). Indeed, at day 3 after injury, SPP1⁺MMP9⁺ immune cells were localized near migrating AT2-derived cells at both BADJ and channel sites (Extended Data Fig. 9h–j). Similar dynamics were observed after naphthalene injury (Extended Data Fig. 9k–p). We next performed scratch and Transwell assays to test whether infiltrating SPP1^high immune cells promote AT2 migration. Recombinant SPP1 significantly accelerated wound closure and enhanced Transwell migration of AT2 cell line (Extended Data Fig. 10a–d). Adding immune cells to the lower chamber also increased migration, an effect that was blocked by an SPP1 inhibitor (Extended Data Fig. 10e–g and Supplementary Fig. 1c). We then validated SPP1 function in vivo. In AT2-tracer1 mice, SPP1 inhibitor treatment significantly reduced bronchiolar coverage of tdT⁺ cells compared with the controls (Fig. 5k–n). Similarly, in naphthalene-injured AT2-tracer2 mice, SPP1 inhibition decreased ZsG⁺ cell coverage at day 28 (Extended Data Fig. 10h–j). Collectively, these experiments show that immune-derived SPP1 acts as a chemoattractant guiding AT2 cell migration to injured bronchioles (Fig. 5n).

scRNA-seq analysis revealed a robust interaction between Spp1 and the integrin β1 (Itgb1) expressed on AT2 cells (Fig. 5b). To test whether ITGB1 regulates AT2 migration, we performed in vitro scratch and Transwell assays. Addition of an anti-ITGB1 antibody significantly inhibited wound closure and reduced cell migration (Extended Data Fig. 10k–n). To extend these findings in vivo, we generated AT2-specific Itgb1-KO mice (Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP;Itgb1^flox/flox). At 28 days after DT, Itgb1 KO significantly reduced AT2 cell migration towards the bronchioles compared with in the control mice (Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP;Itgb1^flox/+) (Fig. 5o–r). Moreover, Itgb1-KO mice showed increased bronchiolar wall thickness and impaired small airway ventilation (Extended Data Fig. 10o–s). These findings demonstrate that AT2 cells require ITGB1 to receive chemotactic cues from infiltrating immune cells, directing their migration to the bronchioles after injury (Fig. 5s).

The widely used Sftpc-creER system labels both AT2 cells and BASCs, confounding specific analysis of AT2 cell contributions to bronchiolar repair. To address this, we developed five distinct complementary AT2-tracer systems and showed that AT2 cells participate in bronchiolar repair through trans-compartmental migration and transdifferentiation. After injury, infiltrating immune cells secrete SPP1 to form a chemotactic gradient, probably acting through ITGB1 on AT2 cells to promote migration. Once at the injured site, they enter a transient intermediate state before transdifferentiating into club cells. This transdifferentiation depends on Notch signalling: AT2-derived cells sense Notch ligands from adjacent ciliated cells to activate club cell specification (Extended Data Fig. 10t). These findings underscore AT2 cell plasticity and expand their regenerative capacity beyond alveolar maintenance. Furthermore, KO of Itgb1 or Rbpj in AT2 cells suppresses functional recovery, highlighting the importance of club cell regeneration for lung function.

In the DT model, AT2 cells migrate into bronchioles through BADJ and bronchiolar wall channels—although classification is somewhat subjective. By contrast, naphthalene injury induces only BADJ-mediated migration. We found that matrix metallopeptidases secreted by immune cells, AT2 cells and peribronchiolar fibroblasts may promote channel formation after DT injury, but direct evidence is lacking. Further studies are needed to elucidate the microenvironmental differences between these two injury models. Moreover, alveolar and bronchiolar epithelial regeneration are known to be regulated by niche cells such as fibroblasts, endothelial cells and immune cells^{27,30,45,46,47,48,49}. Whether other cell types or signalling pathways mediate AT2-to-club-cell transition remains to be investigated.

Our finding that AT2 cells respond to bronchiolar injury aligns with a recent report showing that airway injury triggers myeloid-mediated lung alveolar responses⁵⁰. Although AT2 cells and BASCs show distinct transdifferentiation kinetics during short-term repair, these differences resolve over time, with AT2-derived club cells acquiring full functional maturity—including progenitor capacity and ciliation—and undergoing stable epigenetic reprogramming to airway-specific profiles that resemble native cells. Owing to technical limitations, several questions remain: whether AT2 cells adopt a BASC-like intermediate state during transdifferentiation, and whether IS cells contribute to club cell regeneration. Both of these questions require lineage-tracing studies. Transcriptome sequencing combined with in vitro/in vivo validation revealed that immune-secreted SPP1 regulates AT2 migration. We propose that fibroblasts remodel ECM and provide structural support, potentially complementary to immune-driven chemotaxis.

Our findings have important implications for lung regeneration therapies^5,13. While current clinical approaches primarily target airway-resident stem cells (such as basal cells and secretory cells) for airway repair^5,6,12, our research suggests that AT2 cells could represent an alternative cellular source for airway regeneration, particularly given their ability to fully mature into functional club cells⁵. Moreover, the identification of Notch signalling as a key regulator of AT2-to-club-cell transdifferentiation reveals a potential therapeutic target for enhancing alveolar-to-bronchiolar repair. We found that ectopic activation of Notch signalling in AT2 cells promotes AT2-to-club-cell transdifferentiation in vitro, providing a cellular model for studying AT2 cell lineage conversion. However, mouse–human lung structural differences preclude direct extrapolation, leaving the existence of a conserved AT2-to-club program in humans unresolved^37,51,52,53.

Mice

All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC, protocol SIBCB-S374-1702-001-C1) at the Center for Excellence in Molecular Cell Science, Chinese Academy of Sciences, and were performed in accordance with its guidelines. Sftpc-creER, Sftpc-dreER, Scgb1a1-creER, Ager-creER, Hopx-2A-dreER, p63-dreER, Foxj1-creER, R26-NR, R26-TLR, R26-tdT, R26-RSR-tdT, R26-RSR-LSL-tdT, R26-Confetti, R26-iDTR, R26-NICD, Rbpj^flox/+ and Jag1^flox/+ mouse lines were described previously^{3,22,33,38,41,54,55,56,57,58,59,60,61}. The Scgb1a1-DTR-GFP, Ascl1-creER, Itgb1^flox/+, Rbpj^rox/+, Lamp3-creER, Cldn4-LSL-dre, Cldn4^flox/+ and Notch2^flox/+ mouse lines were generated by Shanghai Model Organisms Center (SMOC). For the Scgb1a1-DTR-GFP mouse line, the P2A-DTR-GFP sequence was inserted into the Scgb1a1 locus by homologous recombination. For the Ascl1-creER mouse line, the creER^T2 sequence was inserted into the Ascl1 locus by homologous recombination. For the Itgb1^flox/+ mouse line, two loxP sites were added on either side of exon 3 of Itgb1 by homologous recombination. For the Rbpj^rox/+ mouse line, two rox sites were added on either side of exon 6 and exon 7 of Rbpj by homologous recombination. For the Lamp3-creER mouse line, the 2A-creER sequence was inserted into the Lamp3 locus by homologous recombination. For the Cldn4-LSL-dre mouse line, the loxP-stop-loxp-dresequence was inserted into the Cldn4 locus by homologous recombination. For the Cldn4^flox/+ mouse line, two loxP sites were added on either side of exon 1 of Cldn4 by homologous recombination. For the Notch2^flox/+ mouse line, two loxP sites were added on either side of exon 3 of Notch2 by homologous recombination. All mice were housed and bred in the specific pathogen free (SPF) animal facility of the Center for Excellence in Molecular Cell Science, under controlled conditions at 22 ± 2 °C, 50 ± 10% relative humidity and a 12 h–12 h light–dark cycle, with ad libitum access to food and water. Depending on the experimental cohort, mice were used at either 7–20 weeks of age or 6 months of age, and all animals were maintained on the C57BL6/ICR mixed background. Both male and female mice were randomly used for experiments. The mice were administrated 0.2 mg g⁻¹ tamoxifen (Sigma-Aldrich, T5648; ABCONE, T56488) dissolved in corn oil through oral gavage at the indicated timepoints. For clonal analysis, the dose of tamoxifen was reduced to 0.02 mg g⁻¹. The mice were injected intraperitoneally 10 ng g⁻¹ DT (Sigma-Aldrich, D0564-1MG) dissolved in PBS (Invitrogen, C10010500BT) for five times at indicated time to induce cell ablation.

Genomic PCR

Genomic DNA was prepared from the mouse tissues as previously described³. Mouse tissues were dissociated with lysis buffer (100 mM Tris HCl, pH 7.8, 5 mM EDTA, 0.2% SDS, 200 mM NaCl and 100 mg ml⁻¹ proteinase K) for 6–12 h at 55 °C, and then the supernatants with genomic DNA were collected to obtain DNA precipitation using isopropanol after centrifugation at maximum speed (21,130g) for 4 min. The DNA precipitation was then washed in 70% ethanol and air-dried, followed by dissolved with deionized water for PCR amplification.

Lung tissue collection and immunofluorescence staining

The lung tissues were collected as previously reported³. Mice were euthanized and then the blood within the lung was eliminated by perfusing 10–15 ml PBS through the right ventricle. Subsequently, the lungs were infused with 1.5 ml of 4% paraformaldehyde (PFA) through the trachea, and the collected tissue was fixed in 4% PFA for 1 h (h) at 4 °C, followed by washing three times in PBS. The lungs were then dehydrated overnight at 4 °C using 30% sucrose and then embedded in optimal cutting temperature (OCT) compound (Sakura). The lung tissues embedded with OCT could be stored at −80 °C for several months. For immunostaining, frozen sections (9–12 µm) were first prepared on negatively charged slides, which were air-dried at room temperature and subsequently washed several times with PBS. The slides were then incubated with 5% PBS-stabilized serum and 0.2% Triton X-100 (PBSST buffer) for 30 min at room temperature, followed by overnight incubation with primary antibodies at 4 °C. Primary antibodies used were as follows: E-cadherin (R&D system, AF748, 1:500), SCGB1A1 (Santa Cruz, SC-9772, 1:200), SCGB1A1 (Abcam, ab213203, 1:300), SCGB3A2 (R&D system, AF3545-SP, 1:500), tdTomato (Rockland, 600-401-379, 1:1000), tdTomato (Rockland, 200-101-379, 1:500), GFP (Abcam, ab6662, 1:500), collagen IV (Millipore, Ab769, 1:500), GFP (Nacalai Tesque, 04404-84, 1:500), GFP (Invitrogen, A11122, 1:500), AGER (R&D system, MAB1179-100, 1:200), SFTPC (Millipore, AB3786, 1:200), ZsGreen (Clontech, 632474, 1:1,000), Ki-67 (Abcam, ab15580-100ug, 1:400), DTR (R&D system, AF-259-NA, 1:500), Ace-Tub (Sigma-Aldrich, T7451, 1:300), Krt5 (BioLegend, 905504, 1:500), CGRP (Sigma-Aldrich, C8198, 1:500), PDGFRα (R&D system, AF1062, 1:500), LAMP3 (NovusBio, 1006F7.05, 1:300), FOXJ1 (Abcam, ab235445, 1:300), T1A/PDPN (DSHB, 8.1.1, 1:100), HOPX (Abcam, ab307670, 1:100), JAGGED1 (Abcam, ab7771, 1:100), MMP9 (Thermo, MA5-15886, 1:200), SPP1/OPN (R&D system, AF808, 1:400), CD45 (eBioscience, 17-0451-82, 1:200).

The sections were then incubated with Alexa-conjugated secondary antibodies for 30 min at room temperature and nuclei were counterstained with 4′6-diamidino-2-phenylindole (DAPI, Vector lab). Second antibodies were listed as follows: Alexa donkey anti-rabbit 555 (Invitrogen, A31572, 1:1,000), Alexa donkey anti-rabbit 488 (Invitrogen, A21206, 1:1,000), Alexa donkey anti-rabbit 647 (Invitrogen, A31573, 1:1,000), Alexa donkey anti-goat 488 (Invitrogen, A11055, 1:1,000), Alexa donkey anti-goat 555 (Invitrogen, A21432, 1:1,000), Alexa donkey anti-goat 647 (Invitrogen, A21447, 1:1,000), Alexa donkey anti-mouse 647 (Invitrogen, A31571; 1:1,000), Alexa donkey anti-mouse 488 (Invitrogen, A21202, 1:1,000), Alexa donkey anti-rat 647 (Invitrogen, A21247, 1:1,000), Alexa donkey anti-rat 488 (Invitrogen, A21208, 1:1,000). For weak signals, the horseradish-peroxidase-conjugated secondary antibodies were used for signal amplification using a Tyramide signal amplification kit (PerkinElmer). Finally, the slides washed in PBS followed by mounting with medium. Images were obtained using the Olympus FV1200, FV4000 confocal system and the Zeiss 880 confocal system. Images were analysed using ImageJ (NIH, v.2.3.0/1.53q).

Lung tissue dissociation and flow cytometry

Lung epithelium was isolated according to previously established protocols³. The mice were euthanized, and the lungs were perfused with cold PBS through the right ventricle to eliminate blood cells, followed by the injection of protease solution consisting of collagenase type I (500 U ml⁻¹; Gibco 17100–017), elastase (4 U ml⁻¹; Worthington Biochemical Corporation LS002279), dispase (5 U ml⁻¹; BD Biosciences 354235) and DNase I (0.33 U ml⁻¹; Worthington Biochemical, LS002139) in DMEM (Gibco, 11965092) through the trachea. Subsequently, the lungs were minced and incubated in 4 ml of protease solution for 30 min at 37 °C with continuous agitation. The tissue fragments were then dissociated by pipetting during the digestion process, followed by adding equal volume of the DMEM/F12 with 10% FBS to terminate the digestion. After being centrifuged at 1,000g for 5 min at 4 °C and washing with DMEM, the precipitate was resuspended in 3 ml enzyme mixture of 0.1% Trypsin-EDTA (Gibco, 25200072) containing DNase I and incubated for 20 min at 37 °C with agitation.

The digestion was then terminated by addition of DMEM/F12 with 10% FBS, after which the cells were sequentially filtered through 100 µm strainers and centrifuged at 1000g for 5 min at 4 °C. The red blood cells were lysed by incubation of red blood cell lysis buffer (2–5 ml, eBioscience, 00-4333-57) at room temperature for 5 min. The process was ended by adding 8–10 ml PBS and then centrifuging at 1,000g for 5 min at 4 °C. The precipitate was resuspended in block solution (eBioscience, 14–0161, 1:100) for 5 min to block nonspecific antigens before further staining. The cells were then incubated with primary antibodies at 4 °C for 30 min. The following antibodies were used: CD45 APC-eFluor 780 (eBioscience, 47-0451-82, 1:200), Ter-119 APC-eFluor 780 (eBioscience, 47-5921-82, 1:200). Next, the cells were washed with PBS, followed by centrifugation. Subsequently, the cells were resuspended in PBS supplemented with DAPI and DNase I.

Bulk AT2 cells were excluded using Lyso-Tracker by staining the lysosomes in AT2 cells. For Lyso-Tracker staining, the cells were resuspended in DMEM containing LysoTracker (Invitrogen, Lyso-Tracker Deep Red, L12492) and incubated at 37 °C for 10 to 15 min. The cells were next centrifuged at 1,000g for 5 min at 4 °C, and then washed with PBS. Subsequently, the cell pellets were resuspended in PBS supplemented with DAPI and DNase I. For scRNA-seq, the AT2-tracer3, BASC-tracer and club-tracer cells were sorted from Scgb1a1-creER;Sftpc-dreER;R26-TLR lungs at 4 weeks and 6 months after naphthalene injury. For scATAC–seq, the AT2-tracer2 cells were sorted from Scgb1a1-creER;Sftpc-dreER;R26-NR lungs at 4 weeks and 6 months after naphthalene injury. All cells in the single-cell suspension were negatively selected for CD45 and Ter119, and then gated the tdT⁺ZsG^– cells as club-tracer, tdT⁺ZsG⁺ cells as BASC-tracer and tdT^–ZsG⁺Lyso-Tracker^low cells as AT2-tracer3. The Sony MA900 and BD FACS Aria systems were used for flow cell sorting. Fluorescence-activated cell sorting (FACS) data were analysed using FlowJo v.10.4 (FlowJo). Representative gating strategies used in this study are shown in Supplementary Fig. 1.

In vitro organoid culture

Mouse AT2 or club organoid experiments was performed as previously described^2,62. tdT⁺GFP^–AT2 cells or tdT^–GFP⁺ club cells were sorted from Sftpc-creER;R26-tdT;Scgb1a1-DTR-GFP mice. Freshly sorted cells were resuspended in AdDMEM/F12 (Gibco, 12634010) after centrifugation and counted to calculate the cell number. For each droplet, 8,000–12,000 cells in 30 μl AdDMEM/F12 were mixed gently with Matrigel (Corning, 354230) at a ratio of 1:1 by pipetting. Then, 60 μl of the mixture was placed at the bottom of the each well of a 24-well culture plate, or 30 μl was placed into each well of LabTek 8 chamber slides (Thermo Fisher Scientific, 177445), respectively. Next, the plates or 8-chamber slides were placed in a 37 °C incubator for 20–30 min to solidify the Matrigel, followed by addition of 600 µl culture medium per well. AT2 culture medium was prepared as previously reported⁶²: Advanced DMEM/F12 supplemented with N2 (Gibco, 17502048), B27 (Gibco, 17504044), antibiotic–antimycotic (Gibco,15240062), 15 mM HEPES (Gibco, 15630080), GlutaMАХ (Gibco, 35050061), 1.25 mM N-acetyl-l-cysteine and 10 µM SB431542 (ABCAM, Ab120163), 3 µM CHIR99021 (StemCell, 72052), 1 µM BIRB796 (StemCell, 72682), insulin-transferrin-selenium (Thermo Fisher Scientific, 41400045), 50 ng ml⁻¹ human EGF (Thermo Fisher Scientific, PHG0313), 10 ng ml⁻¹ mouse FGF10 (R&D, 6224-FG-025), 10 ng ml⁻¹ mouse noggin (Thermo Fisher Scientific, 250-38) and 5 µg ml⁻¹ heparin. For the first 4 days, 10 µM Y-27632 (Selleck, S1049) and 10 ng ml⁻¹ mouse IL-1β (BioLegend, 575104) was additionally added. Club cell culture medium was prepared according to manufacturer’s instructions: bioGenous Mouse Airway Organoid Basal Medium A supplemented with Supplement B and Supplement C (bioGenous, K2203-MA). The medium was changed every other day. Organoids were captured using the Olympus CKX53 microscope and collected at day 7–10. Analysis of colony-forming efficiency and size of organoids was performed at 7 days after plating.

Whole-organoid immunostaining

Organoids were cultured in LabTek 8 chamber slides (Thermo Fisher Scientific, 177445) and fixed with 500 μl of 4% PFA for 30 min at room temperature. Wells were gently washed three times with PBS for 5 min each at room temperature. For immunostaining, organoids were blocked with 0.5% Triton X-100 in PBS containing 5% NDS for 30 min at room temperature. Primary antibodies diluted in 0.5% Triton X-100 in PBS were added to each well and incubated at 4 °C overnight. After three washes with PBS, secondary antibody incubation was performed for 2 h at room temperature. For organoid fixation and sectioning, organoids were fixed with 4% PFA at 4 °C for 2 h or at room temperature for 1 h, respectively, followed by washing with PBS. For cryosectioning, organoids from submersion cultures were either embedded directly in OCT or first immersed in 1% agarose and then embedded in OCT before sectioning (8–10 μm).

Wound healing assay

Migration assay was performed and developed as previously reported⁶³. In brief, MLE-12 (Sunncell, SNL-414) were seeded into 24-well cell culture plates and cultured until they reached 80% confluence. A sterile 200 μl pipette tip was used to gently scratch the cell monolayer in a straight line across the centre of each well, with the long axis of the tip held perpendicular to the well surface to ensure a uniform gap width. After scratching, wells were washed gently with PBS and replenished with fresh medium with or without indicated treatments: the SPP1 groups were supplemented with SPP1 (1 μg ml⁻¹, MedChemExpress, HY-P78358, resolved in 1% BSA to obtain 500 ng ml⁻¹ solution) and the anti-ITGB1 groups were supplemented with ITGB1 antibody (1:100, Millipore, MAB1997), and control groups were supplemented with an equivalent volume of 1% BSA. Images of the same scratch area were captured at 0 h and 24 h (end point) using the Olympus CKX53 microscope. The wound area (px²) was measured using ImageJ software (NIH) at 0 h (A₀) and at the end point (A_t). The percentage of wound closure was calculated using the formula: wound closure rate (%) = (1 – A_t/A₀) × 100%. Each condition was tested in triplicate, and five independent experiments were performed. Results are presented as mean ± s.d.

Transwell assay

The migratory capacity of MLE-12 was assessed using 24-well Transwell inserts with 8.0-μm-pore-size polycarbonate membranes (Corning, 353097).

For migration assays, 2–3 × 10⁵ epithelial cells were resuspended in 150 μl of DMEM/F12 containing with 1% FBS and seeded into the upper chamber. The lower chamber was filled with 600 μl of complete medium (DMEM/F12 with 10% FBS) with or without the indicated treatments (recombinant SPP1 or anti-ITGB1 antibody).

For co-culture experiments, primary CD45⁺ lung immune cells were sorted using FACS and 2 × 10⁵ cells were seeded into the bottom of 24-well plate in complete medium: RPMI-1640 (Thermo Fisher Scientific, 22400089) with 10% FBS (with or without SPP1 inhibitor (100 μM, R&D, AF808)). Cells were incubated for 2–6 h before the addition of the Transwell insert to allow macrophage adherence. Subsequently, 2–3 × 10⁵ epithelial cells were resuspended in 150 μl of RPMI-1640 containing 1% FBS and seeded into the 24-well Transwell insert.

After incubation at 37 °C in 5% CO₂ for 16 h, non-migrated cells on the upper surface of the membrane were gently removed with a cotton swab. Migrated cells that attached to the lower surface were fixed with 4% PFA for 15–30 min at room temperature and stained with 0.5% crystal violet. For quantification, five randomly selected fields per membrane were imaged using the Olympus IX73 microscope. The area of migrated cells was measured using ImageJ software (NIH), and the results were expressed as the relative coverage rate, defined as the migrated area in the experimental group normalized to that in the control group (control = 1). Each condition was tested in triplicate, and at least five independent experiments were performed. The results are presented as the mean ± s.d. of migrated cells per field.

3D reconstruction

Three-dimensional image data processing was performed using Imaris software (v.9.5.1) (Bitplane, http://www.bitplane.com/imaris/imaris). The Surface module was used for 3D reconstruction. Reconstructed surfaces were visualized and analysed for morphological features as indicated.

RNA preparation and RT–qPCR

The total RNA of isolated cells sorted by FACS was extracted using Trizol reagent (Invitrogen, 15596018) according to the manufacturer’s protocol. The Prime Script RT kit (Takara, RR047A) was used to generate cDNA using the reverse transcription (RT) procedure. Quantitative PCR (qPCR) was performed with SYBR Green qPCR master mix (Thermo Fisher Scientific, 4367659) and QuantStudio 6 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific). The Gapdh expression level was used as the internal control in this paper.

The primer sequences for qPCR were as follows: forward primer of Rbpj: ACTGTTCAATCGCCTTCGGT; reverse primer of Rbpj: TGACAGTCTGCCCGTAATGG; forward primer of Itgb1: CAAGCAGGGCCAAATTGTGG; reverse primer of Itgb1: GAAGCAGCTGTTGTGGTTGG; forward primer of Notch2: GGACTGCCAATACTCCACCT; reverse primer of Notch2: GCTTCCATTTTCGCAGGGAT; forward primer of Cldn4: CTTCATCGGCAGCAACATCG; reverse primer of Cldn4: GATGACCATAAGGGCTCGGG.

Naphthalene-induced lung injury

Bronchiolar injury was induced by naphthalene as previously described³³. Naphthalene (Sigma-Aldrich, 84679) was freshly dissolved in sterile corn oil (25 mg ml⁻¹). For severe bronchiolar injury, the adult mice were treated with 500 mg per kg naphthalene or vehicle (corn oil) by intraperitoneal injection. The lung tissues were collected at the indicated time.

In vivo SPP1 inhibition

For in vivo inhibition of SPP1, the SPP1/OPN expression inhibitor (HY-146064; MedChemExpress) was administered to mice through intraperitoneal injection at a dose of 5 mg per kg body weight once every other day for 4 consecutive weeks. The compound was first dissolved in DMSO (Sigma-Aldrich, D2650) and then diluted with a vehicle comprising PEG300 (HY-Y0873; MedChemExpress), Tween-80 (HY-Y1891; MedChemExpress) and PBS to a final concentration of 2.5 mg ml⁻¹. Control mice received an equal volume of vehicle alone.

Pulmonary function and airway responsiveness measurements

All animal procedures were performed in accordance with institutional guidelines and approved protocols. Mice were deeply anesthetized with an intraperitoneal injection of tribromoethanol (Thermo Fisher Scientific, T48402-25G, 250 mg per kg) to ensure adequate anaesthesia throughout the procedure. After confirmation of loss of pedal reflex, a tracheotomy was performed, and the mice were placed into a whole-body plethysmograph (EMMS) connected to a flow measurement system for baseline lung function assessment. Functional residual capacity, forced vital capacity, as well as chord compliance were measured. For airway responsiveness measurements, under the same deep anaesthesia, mice were connected to a resistance and compliance system (EMMS) through the plethysmograph. Mice were challenged with increasing doses of aerosolized acetylcholine (Sigma-Aldrich, A6625): 0, 8, 16, 32, 64, 128 and 256 mg ml⁻¹, each for 3 min. Lung resistance was recorded and expressed as the percentage change from the baseline lung resistance measured during nebulization of PBS. Data acquisition was performed using eDacq (v.1.9.9) software. At the end of the experiments, animals were euthanized under deep anaesthesia by an approved method.

Single-cell library preparation

For single-cell 5′ transcriptomic profiling, FACS-purified cell suspensions were processed using the Chromium Next GEM Single Cell 5′ Reagent Kit (v2 or v3 Chemistry). Specifically, cell suspensions were loaded into the 10x Chromium controller for gel bead-in-emulsion (GEM) generation. Reverse transcription, cDNA amplification and library construction were conducted in accordance with the manufacturer’s instructions.

For single-cell multiome analysis, nuclei were isolated from FACS-sorted cells according to the manufacturer’s instructions (10x Genomics). Joint profiling of chromatin accessibility (ATAC) and gene expression (RNA) was performed on the Chromium platform. Specifically, nuclei were subjected to an in-bulk transposition reaction using the Chromium Next GEM Single Cell Multiome ATAC+Gene Expression Kit. After transposition, transposed nuclei were loaded for GEM generation. This system allows for the simultaneous capture of ATAC fragments and poly-adenylated mRNA from the same individual nucleus through unique barcoding. ATAC and GEX (cDNA) libraries were independently amplified and indexed as per the standard 10x Genomics protocol.

Libraries were pooled and sequenced on the Illumina NovaSeq 6000 platform in paired-end 150 bp mode. Each library was sequenced to a total depth of approximately 200 GB of raw data, ensuring an average sequencing depth exceeding 50,000 read pairs per cell/nucleus.

Stereo-seq library construction

Fresh lung tissue was collected following intratracheal instillation of pre-chilled 20% Tissue-Tek OCT (Sakura, 4583) in sterile 1× PBS, then washed with 4 °C PBS. Lung lobes were dissected and immediately frozen on dry ice. The tissue was subsequently transferred to and stored in a −80 °C freezer. Before sectioning, the lung tissue was rewarmed to −20 °C. The commercialized Stereo-seq protocol was performed as previously described (201ST13114, Stereo-seq Kit for FF, v.1.3, STOmics)⁶⁴. In brief, 10 µm cryosections were adhered to Stereo-seq ChipT and incubated at 37 °C for 5 min. The sections were then fixed in methanol (Sigma-Aldrich, 34860-1L-R) for 30 min, followed by permeabilization at 37 °C for 12 min. After washing with 0.1× SSC (Thermo Fisher Scientific, Invitrogen, AM9770) supplemented with 2 U μl⁻¹ RNase inhibitor (Thermo Fisher Scientific, EO0382), the released RNA was captured by oligo probes and subjected to reverse transcription for 2 h. cDNA was then released from the chip by incubation at 55 °C for 10 min, denatured, amplified and purified using 0.8× SPRIselect (Beckman, B23318), and quantified with the Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific, Invitrogen, Q32854). The Stereo-seq Library Preparation Kit v1.1 was used for library construction according to the manufacturer’s instructions. A total of 100 ng cDNA was used as template. After cDNA multiple displacement amplification and single-stranded DNA (ssDNA) purification, the ssDNA concentration was quantified using the Qubit ssDNA Assay Kit (Thermo Fisher Scientific, Invitrogen, Q32856). After purification, the product underwent a second round of amplification. Subsequently, the PCR product was subjected to double-sided magnetic bead selection using room-temperature equilibrated beads at ratios of 1:0.55 and 1:0.15. Sequencing was performed on the MGI Tech (DNBSEQ-T7) platform.

Single-cell transcriptomic data processing and analysis

Raw sequencing reads were trimmed using Trim Galore (v0.6.7) with the parameters ‘--quality 20 --illumina --stringency 13 --length 150 --paired’ to remove adapter sequences and filter low-quality reads. The cleaned reads were then aligned to a customized mm10 mouse reference genome using Cell Ranger (v.6.1.1) ‘cellranger count’ with the default parameters to generate gene expression matrices.

Downstream analysis was conducted using Seurat (v.4.1.2). To ensure high-quality transcriptomic profiles, we excluded cells with fewer than 1,000 detected genes and a mitochondrial gene content exceeding 10%. For lineage-specific analysis, fluorescence-labelled populations, comprising AT2-tracer3, BASC-tracer and club-tracer cells, were identified and extracted based on reporter transcript levels.

The filtered data were normalized to adjust for sequencing depth variations, after which highly variable genes were systematically identified. To mitigate technical biases, scaled expression matrices were generated by regressing out potential technical confounders, including cell cycle effects, mitochondrial content and sequencing depth. Dimensionality reduction was performed using principal component analysis (PCA) based on highly variable genes. The cells were then projected into UMAP space with the top significant principal components (PCs) as input. Subsequently, cell clusters were identified through unsupervised clustering analysis. Potential outlier cells were identified and subsequently exclude from downstream analyses using DBSCAN (v1.1-10) with empirically determined parameters. Cluster-specific marker genes were determined through differential expression analysis, followed by cell type annotation using known lung epithelial gene signatures. To integrate multiple datasets, we used a canonical correlation analysis-based integration to harmonize cross-batch variations while preserving biological heterogeneity.

Cell differentiation trajectories were reconstructed by applying diffusion map analysis (destiny v.3.20.0) to infer pseudotemporal ordering. For functional enrichment analyses, we performed Gene Ontology (GO) and KEGG terms enrichment analyses using R package limma (v.3.62.1) and clusterProfiler (v.4.14.4). Cell type signatures were derived from published datasets: club/AT2 cells³³, pre-AT1 transitional cells (PATS)^29,65 and Notch signalling genes (GO:0007219). Enrichment scores were calculated as the mean of z-score normalized expression of signature genes.

Intercellular signalling networks were systematically inferred using CellChat (v.1.6.1) following the standard pipeline with default parameter. In brief, we used the standard mouse signalling database to compute communication probabilities and identify significant signalling hubs. Separate CellChat objects were constructed for different experimental conditions and subsequently merged to enable comparative analysis.

Single-cell multiome sequencing and analysis

After sequencing, raw fastq files were processed using Cell Ranger ARC (v.2.1.0) against the customed mm39 reference genome. Initial quality control was performed using Signac (v.1.16.0) and Seurat (v.5.3.0). Nuclei were filtered based on standard quality control metrics for downstream analysis. Specifically, we retained only those nuclei with a total fragment count between 1,000 and 150,000 and total read count ranging from 800 to 75,000. Furthermore, the analysis was restricted to nuclei exhibiting a mitochondrial transcript proportion of less than 20%, and a transcription start site (TSS) enrichment score exceeding 4.

To resolve cell identities across homeostatic and injured conditions, we performed cross-modality integration. For the RNA modality, data were normalized using the SCTransform function, and integration anchors were identified in the PCA space. For the ATAC modality, peak calling was performed on each sample, and a unified peak set was generated by merging and reducing overlapping genomic ranges. Term frequency-inverse document frequency normalization and latent semantic indexing were applied for ATAC dimensionality reduction. We used WNN analysis to construct a multimodal graph based on both integrated SCT and integrated latent semantic indexing embeddings, followed by UMAP visualization and clustering.

Lineage-specific epigenetic signatures were defined by identifying differentially accessible regions between homeostatic AT2 cells and native club cells using the FindMarkers function. Single-cell epigenetic signature scores were computed using the AddChromatinModule function, which aggregates normalized signals from lineage-specific peaks while controlling for technical biases.

Transcriptional maturation of AT2-derived cells was assessed using the AddModuleScore function. Representative gene sets were curated for native club cell functions: secretory and immune (such as Scgb1a1, Scgb3a2 and Muc5b).

Spatial transcriptomic analysis

Spatial transcriptomic libraries were generated using the Stereo-seq platform as indicated above. Raw data were processed to generate coordinate-annotated gene expression matrices at cellular resolution using SAW (v.8.2.2) and downstream analysis was primarily conducted using Stereopy (v.1.6.2). To ensure high-quality data, cells were filtered based on the following criteria: a minimum of two hundred total counts, at least ten detected genes and a mitochondrial content of no more than 6%. Cells characterized by high haemoglobin levels, specifically those with a haemoglobin content reaching or exceeding 5%, were excluded from the analysis to minimize potential contamination or artefacts. Gene-level filtration was performed using pyBioMart to retain only protein-coding genes, while removing genes located on X and Y chromosomes to minimize sex-linked bias.

Filtered count matrices were normalized using sctransform function. Highly variable genes were identified, and PCA was performed. A neighbourhood graph was constructed in the PCA space (dim = 50), followed by UMAP visualization and Leiden clustering. Cell identities were manually annotated by comparing cluster-specific marker genes with canonical lineage markers.

To investigate spatially defined microenvironments, we utilized a Lasso strategy to capture specific anatomical regions. Genomic coordinates were mapped back to the cellular bins to categorize nuclei into bronchiolar regions and alveolar regions. For robust comparative analysis, we generated pseudobulk expression profiles by aggregating raw counts from immune cells localized within the bronchiolar regions. Differential expression analysis was performed using the DESeq2 (v.1.46.0). GSEA and GO enrichment were conducted to identify functional pathways associated with spatially restricted injury responses.

To identify gene expression patterns that are spatially informative independent of predefined clusters, we used the Hotspot algorithm⁶⁶. This approach defines informative genes as those of which the transcriptional variation aligns with the spatial proximity of cells. We assessed spatial autocorrelation to identify genes with non-random spatial distributions across the tissue. Subsequently, significant spatial modules were defined by grouping genes that exhibited high local correlations, representing coordinated expression variation among spatially neighbouring cells. These module scores were then projected onto spatial coordinates to visualize regional functional heterogeneity.

Statistical analyses

Mice were randomly assigned to experimental groups, with no blinding of sample processing or data collection. Mice that died in the experiments were not used for analysis. For quantification, all five lung lobes from each mouse were analysed. From each mouse, we examined at least ten non-consecutive lung tissue sections. Within each section, a minimum of five 10× fields containing airways was analysed, resulting in the assessment of at least 50 fields per mouse. Representative images are shown from at least three independent biological replicates, which yielded similar results. For statistical analysis, the data from all fields for each mouse were first averaged to generate a single composite value per animal (for example, the percentage of lineage-positive cells). Thus, in the lineage tracing study, the statistical sample size (n) refers to the number of mice. For single-cell and spatial transcriptomics analyses, n refers to the number of cells or fields, respectively, as indicated in the corresponding figure legends. Normality (Shapiro–Wilk test) and homogeneity of variance (F-test) were assessed before analyses. Data from ≥3 independent experiments are presented as mean ± s.d. Initially, data were subjected to comparison across groups by conducting a Shapiro–Wilk normality test and an equal variance test. Two-group comparisons were performed using unpaired two-tailed Student’s t-tests (equal variance) or Welch’s t-tests (unequal variance); non-normal data were analysed using Mann–Whitney U-tests. Multigroup comparisons were performed using one-way ANOVA with Tukey’s test (normal data) or Kruskal–Wallis test with Dunn’s test (non-normal data). All data were acquired in this study from at least three independent experiments and presented as mean values ± s.d. as indicated in figure legends. Specific quantification methods and P values are indicated in the figures and legends; P < 0.05 was considered to be significant. All analyses used GraphPad Prism (v.9.3.1), R (v.4.4.2) and Python (v.3.8.18). All illustrations in this study were created by Photoline software (v.23.02).

Reporting summary

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

The codes executing the analysis workflow are available at Zenodo⁶⁷ with persistent access at https://doi.org/10.5281/zenodo.21768825.

1. Kathiriya, J. J., Brumwell, A. N., Jackson, J. R., Tang, X. & Chapman, H. A. Distinct airway epithelial stem cells hide among club cells but mobilize to promote alveolar regeneration. Cell Stem Cell 26, 346–358 https://doi.org/10.1016/j.stem.2019.12.014 (2020).

   Article  PubMed  PubMed Central  CAS  Google Scholar 

2. Choi, J. et al. Release of Notch activity coordinated by IL-1beta signalling confers differentiation plasticity of airway progenitors via Fosl2 during alveolar regeneration. Nat. Cell Biol. 23, 953–966 https://doi.org/10.1038/s41556-021-00742-6 (2021).

   Article  PubMed  PubMed Central  CAS  Google Scholar 

3. Liu, K. et al. Tracing the origin of alveolar stem cells in lung repair and regeneration. Cell 187, 2428–2445 https://doi.org/10.1016/j.cell.2024.03.010 (2024).

   Article  PubMed  CAS  Google Scholar 

4. Lv, Z. et al. Alveolar regeneration by airway secretory-cell-derived p63⁺ progenitors. Cell Stem Cell 31, 1685–1700 https://doi.org/10.1016/j.stem.2024.08.005 (2024).

   Article  PubMed  CAS  Google Scholar 

5. Basil, M. C., Alysandratos, K. D., Kotton, D. N. & Morrisey, E. E. Lung repair and regeneration: Advanced models and insights into human disease. Cell Stem Cell 31, 439–454 https://doi.org/10.1016/j.stem.2024.02.009 (2024).

   Article  PubMed  PubMed Central  CAS  Google Scholar 

6. Basil, M. C. et al. The cellular and physiological basis for lung repair and regeneration: past, present, and future. Cell Stem Cell 26, 482–502 https://doi.org/10.1016/j.stem.2020.03.009 (2020).

   Article  PubMed  PubMed Central  CAS  Google Scholar 

7. Swindle, E. J., Collins, J. E. & Davies, D. E. Breakdown in epithelial barrier function in patients with asthma: identification of novel therapeutic approaches. J. Allergy Clin. Immunol. 124, 23–34 https://doi.org/10.1016/j.jaci.2009.05.037 (2009).

   Article  PubMed  CAS  Google Scholar 

8. Holgate, S. T. The airway epithelium is central to the pathogenesis of asthma. Allergol. Int. 57, 1–10 https://doi.org/10.2332/allergolint.R-07-154 (2008).

   Article  PubMed  CAS  Google Scholar 

9. Peng, Y. et al. Angiotensin-converting enzyme 2 in peripheral lung club cells modulates the susceptibility to SARS-CoV-2 in chronic obstructive pulmonary disease. Am. J. Physiol. Lung Cell. Mol. Physiol. 322, L712–L721 https://doi.org/10.1152/ajplung.00305.2021 (2022).

   Article  PubMed  PubMed Central  Google Scholar 

10. Peng, Y. et al. Aberrant epithelial cell proliferation in peripheral airways in bronchiectasis. Front. Cell Dev. Biol. 8, 88 https://doi.org/10.3389/fcell.2020.00088 (2020).

    Article  ADS  PubMed  PubMed Central  Google Scholar 

11. Reynaud, P. et al. Club cell loss as a feature of bronchiolization in ILD. Front. Immunol. 12, 630096 https://doi.org/10.3389/fimmu.2021.630096 (2021).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

12. Rock, J. R. & Hogan, B. L. M. Epithelial progenitor cells in lung development, maintenance, repair, and disease. Annu. Rev. Cell Dev. Biol. 27, 493–512 https://doi.org/10.1146/annurev-cellbio-100109-104040 (2011).

    Article  PubMed  CAS  Google Scholar 

13. Lee, J. H. & Rawlins, E. L. Developmental mechanisms and adult stem cells for therapeutic lung regeneration. Dev. Biol. 433, 166–176 https://doi.org/10.1016/j.ydbio.2017.09.016 (2018).

    Article  PubMed  CAS  Google Scholar 

14. Rawlins, E. L. & Hogan, B. L. Epithelial stem cells of the lung: privileged few or opportunities for many? Development 133, 2455–2465 https://doi.org/10.1242/dev.02407 (2006).

    Article  PubMed  CAS  Google Scholar 

15. Hong, K. U., Reynolds, S. D., Watkins, S., Fuchs, E. & Stripp, B. R. Basal cells are a multipotent progenitor capable of renewing the bronchial epithelium. Am. J. Pathol. 164, 577–588 https://doi.org/10.1016/S0002-9440(10)63147-1 (2004).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

16. Hong, K. U., Reynolds, S. D., Watkins, S., Fuchs, E. & Stripp, B. R. In vivo differentiation potential of tracheal basal cells: evidence for multipotent and unipotent subpopulations. Am. J. Physiol. Lung C 286, L643–L649 https://doi.org/10.1152/ajplung.00155.2003 (2004).

    Article  CAS  Google Scholar 

17. Rock, J. R. et al. Basal cells as stem cells of the mouse trachea and human airway epithelium. Proc. Natl Acad. Sci. USA 106, 12771–12775 https://doi.org/10.1073/pnas.0906850106 (2009).

    Article  ADS  PubMed  PubMed Central  Google Scholar 

18. Yang, Y. et al. Spatial-temporal lineage restrictions of embryonic p63⁺ progenitors establish distinct stem cell pools in adult airways. Dev. Cell 44, 752–761 https://doi.org/10.1016/j.devcel.2018.03.001 (2018).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

19. Lin, B. et al. Airway hillocks are injury-resistant reservoirs of unique plastic stem cells. Nature 629, 869–877 https://doi.org/10.1038/s41586-024-07377-1 (2024).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

20. Tata, P. R. et al. Dedifferentiation of committed epithelial cells into stem cells in vivo. Nature 503, 218–223 https://doi.org/10.1038/nature12777 (2013).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

21. Watson, J. K. et al. Clonal dynamics reveal two distinct populations of basal cells in slow-turnover airway epithelium. Cell Rep. 12, 90–101 https://doi.org/10.1016/j.celrep.2015.06.011 (2015).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

22. Rawlins, E. L. et al. The role of Scgb1a1⁺ Clara cells in the long-term maintenance and repair of lung airway, but not alveolar, epithelium. Cell Stem Cell 4, 525–534 https://doi.org/10.1016/j.stem.2009.04.002 (2009).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

23. Pardo-Saganta, A. et al. Parent stem cells can serve as niches for their daughter cells. Nature 523, 597–601 https://doi.org/10.1038/nature14553 (2015).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

24. Ouadah, Y. et al. Rare pulmonary neuroendocrine cells are stem cells regulated by Rb, p53, and Notch. Cell 179, 403–416 https://doi.org/10.1016/j.cell.2019.09.010 (2019).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

25. Song, H. et al. Functional characterization of pulmonary neuroendocrine cells in lung development, injury, and tumorigenesis. Proc. Natl Acad. Sci. USA 109, 17531–17536 https://doi.org/10.1073/pnas.1207238109 (2012).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

26. Barkauskas, C. E. et al. Type 2 alveolar cells are stem cells in adult lung. J. Clin. Invest. 123, 3025–3036 https://doi.org/10.1172/JCI68782 (2013).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

27. Nabhan, A. N., Brownfield, D. G., Harbury, P. B., Krasnow, M. A. & Desai, T. J. Single-cell Wnt signaling niches maintain stemness of alveolar type 2 cells. Science 359, 1118–1123 https://doi.org/10.1126/science.aam6603 (2018).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

28. Zacharias, W. J. et al. Regeneration of the lung alveolus by an evolutionarily conserved epithelial progenitor. Nature 555, 251–255 https://doi.org/10.1038/nature25786 (2018).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

29. Kobayashi, Y. et al. Persistence of a regeneration-associated, transitional alveolar epithelial cell state in pulmonary fibrosis. Nat. Cell Biol. 22, 934–946 https://doi.org/10.1038/s41556-020-0542-8 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

30. Choi, J. et al. Inflammatory signals induce at2 cell-derived damage-associated transient progenitors that mediate alveolar regeneration. Cell Stem Cell 27, 366–382 https://doi.org/10.1016/j.stem.2020.06.020 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

31. Wang, Z. et al. Enhanced glycolysis-mediated energy production in alveolar stem cells is required for alveolar regeneration. Cell Stem Cell 30, 1028–1042 https://doi.org/10.1016/j.stem.2023.07.007 (2023).

    Article  PubMed  CAS  Google Scholar 

32. Kim, C. F. et al. Identification of bronchioalveolar stem cells in normal lung and lung cancer. Cell 121, 823–835 https://doi.org/10.1016/j.cell.2005.03.032 (2005).

    Article  ADS  PubMed  CAS  Google Scholar 

33. Liu, Q. et al. Lung regeneration by multipotent stem cells residing at the bronchioalveolar-duct junction. Nat. Genet. 51, 728–738 https://doi.org/10.1038/s41588-019-0346-6 (2019).

    Article  PubMed  CAS  Google Scholar 

34. Liu, K. et al. Bi-directional differentiation of single bronchioalveolar stem cells during lung repair. Cell Discov. 6, 1 https://doi.org/10.1038/s41421-019-0132-8 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

35. Salwig, I. et al. Bronchioalveolar stem cells are a main source for regeneration of distal lung epithelia in vivo. EMBO J. 38, e102099 https://doi.org/10.15252/embj.2019102099 (2019).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

36. Kathiriya, J. J. et al. Human alveolar type 2 epithelium transdifferentiates into metaplastic KRT5⁺ basal cells. Nat. Cell Biol. 24, 10–23 https://doi.org/10.1038/s41556-021-00809-4 (2022).

    Article  PubMed  CAS  Google Scholar 

37. Kadur Lakshminarasimha Murthy, P. et al. Human distal lung maps and lineage hierarchies reveal a bipotent progenitor. Nature 604, 111–119 https://doi.org/10.1038/s41586-022-04541-3 (2022).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

38. 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 https://doi.org/10.1073/pnas.1117988108 (2011).

    Article  PubMed  PubMed Central  Google Scholar 

39. Beppu, A. K. et al. Epithelial plasticity and innate immune activation promote lung tissue remodeling following respiratory viral infection. Nat. Commun. 14, 5814 https://doi.org/10.1038/s41467-023-41387-3 (2023).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

40. Giangreco, A., Reynolds, S. D. & Stripp, B. R. Terminal bronchioles harbor a unique airway stem cell population that localizes to the bronchoalveolar duct junction. Am. J. Pathol. 161, 173–182 https://doi.org/10.1016/s0002-9440(10)64169-7 (2002).

    Article  PubMed  PubMed Central  Google Scholar 

41. He, L. J. et al. Enhancing the precision of genetic lineage tracing using dual recombinases. Nat. Med. 23, 1488–1498 https://doi.org/10.1038/nm.4437 (2017).

    Article  PubMed  CAS  Google Scholar 

42. Yim, A., Smith, C. & Brown, A. M. Osteopontin/secreted phosphoprotein-1 harnesses glial-, immune-, and neuronal cell ligand-receptor interactions to sense and regulate acute and chronic neuroinflammation. Immunol. Rev. 311, 224–233 https://doi.org/10.1111/imr.13081 (2022).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

43. Pang, X. et al. Osteopontin as a multifaceted driver of bone metastasis and drug resistance. Pharmacol. Res. 144, 235–244 https://doi.org/10.1016/j.phrs.2019.04.030 (2019).

    Article  PubMed  CAS  Google Scholar 

44. Li, X., Lian, J. & Lu, H. The role of SPP1⁺ TAMs in cancer: impact on patient prognosis and future therapeutic targets. Int. J. Cancer 157, 1763–1771 https://doi.org/10.1002/ijc.70018 (2025).

    Article  PubMed  CAS  Google Scholar 

45. Moiseenko, A. et al. Identification of a repair-supportive mesenchymal cell population during airway epithelial regeneration. Cell Rep. 33, 108549 https://doi.org/10.1016/j.celrep.2020.108549 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

46. Ding, B. S. et al. Endothelial-derived angiocrine signals induce and sustain regenerative lung alveolarization. Cell 147, 539–553 https://doi.org/10.1016/j.cell.2011.10.003 (2011).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

47. Desai, T. J., Brownfield, D. G. & Krasnow, M. A. Alveolar progenitor and stem cells in lung development, renewal and cancer. Nature 507, 190–194 https://doi.org/10.1038/nature12930 (2014).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

48. Cassandras, M. et al. Gli1⁺ mesenchymal stromal cells form a pathological niche to promote airway progenitor metaplasia in the fibrotic lung. Nat. Cell Biol. 22, 1295–1306 https://doi.org/10.1038/s41556-020-00591-9 (2020).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

49. Reyes, N. S. et al. Sentinel p16^INK4a+ cells in the basement membrane form a reparative niche in the lung. Science 378, 192–201 https://doi.org/10.1126/science.abf3326 (2022).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

50. Wong, I. G. et al. Airway injury induces alveolar epithelial responses mediated by macrophages. Cell Rep. 45, 116881 https://doi.org/10.1016/j.celrep.2025.116881 (2026).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

51. Pezet, M. G. et al. Human respiratory airway progenitors derived from pluripotent cells generate alveolar epithelial cells and model pulmonary fibrosis. Nat. Biotechnol. https://doi.org/10.1038/s41587-025-02569-0 (2025).

    Article  PubMed  PubMed Central  Google Scholar 

52. 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 https://doi.org/10.1126/sciadv.aba1972 (2020).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

53. Basil, M. C. et al. Human distal airways contain a multipotent secretory cell that can regenerate alveoli. Nature 604, 120–126 https://doi.org/10.1038/s41586-022-04552-0 (2022).

    Article  ADS  MathSciNet  PubMed  PubMed Central  CAS  Google Scholar 

54. Chung, M. I. & Hogan, B. L. M. Ager-CreER^T2: a new genetic tool for studying lung alveolar development, homeostasis, and repair. Am. J. Respir. Cell Mol. Biol. 59, 706–712 https://doi.org/10.1165/rcmb.2018-0125OC (2018).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

55. Han, X. et al. A suite of new Dre recombinase drivers markedly expands the ability to perform intersectional genetic targeting. Cell Stem Cell 28, 1160–1176 https://doi.org/10.1016/j.stem.2021.01.007 (2021).

    Article  PubMed  CAS  Google Scholar 

56. Liu, K. et al. Triple-cell lineage tracing by a dual reporter on a single allele. J. Biol. Chem. 295, 690–700 https://doi.org/10.1074/jbc.RA119.011349 (2020).

    Article  PubMed  Google Scholar 

57. Snippert, H. J. et al. Intestinal crypt homeostasis results from neutral competition between symmetrically dividing Lgr5 stem cells. Cell 143, 134–144 https://doi.org/10.1016/j.cell.2010.09.016 (2010).

    Article  PubMed  CAS  Google Scholar 

58. Murtaugh, L. C., Stanger, B. Z., Kwan, K. M. & Melton, D. A. Notch signaling controls multiple steps of pancreatic differentiation. Proc. Natl Acad. Sci. USA 100, 14920–14925 https://doi.org/10.1073/pnas.2436557100 (2003).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

59. Liu, K. et al. An inducible hACE2 transgenic mouse model recapitulates SARS-CoV-2 infection and pathogenesis in vivo. Proc. Natl Acad. Sci. USA 120, e2207210120 https://doi.org/10.1073/pnas.2207210120 (2023).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

60. Buch, T. et al. A Cre-inducible diphtheria toxin receptor mediates cell lineage ablation after toxin administration. Nat. Methods 2, 419–426 https://doi.org/10.1038/nmeth762 (2005).

    Article  PubMed  CAS  Google Scholar 

61. Kiernan, A. E., Xu, J. & Gridley, T. The Notch ligand JAG1 is required for sensory progenitor development in the mammalian inner ear. PLoS Genet. 2, e4 https://doi.org/10.1371/journal.pgen.0020004 (2006).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

62. 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 https://doi.org/10.1016/j.xpro.2022.101447 (2022).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

63. Lampugnani, M. G. Cell migration into a wounded area in vitro. Methods Mol. Biol. 96, 177–182 https://doi.org/10.1385/1-59259-258-9:177 (1999).

    Article  PubMed  CAS  Google Scholar 

64. Chen, A. et al. Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays. Cell 185, 1777–1792 https://doi.org/10.1016/j.cell.2022.04.003 (2022).

    Article  ADS  PubMed  CAS  Google Scholar 

65. Strunz, M. et al. Alveolar regeneration through a Krt8⁺ transitional stem cell state that persists in human lung fibrosis. Nat. Commun. 11, 3559 https://doi.org/10.1038/s41467-020-17358-3 (2020).

    Article  ADS  PubMed  PubMed Central  CAS  Google Scholar 

66. DeTomaso, D. & Yosef, N. Hotspot identifies informative gene modules across modalities of single-cell genomics. Cell Syst. 12, 446–456 https://doi.org/10.1016/j.cels.2021.04.005 (2021).

    Article  PubMed  CAS  Google Scholar 

67. Liu, K. et al. Transcompartmental AT2 cell plasticity drives bronchiolar epithelial regeneration. Zenodo https://doi.org/10.5281/zenodo.21768825 (2026).

Download references

Author notes

1. These authors contributed equally: Kuo Liu, Zixin Liu, Xinfeng Meng, Muxue Tang

Authors and Affiliations

1. Zhejiang Key Laboratory of Signaling Mechanisms in Tumor Progression and Intervention, School of Life Science, Hangzhou Institute for Advanced Study, Hangzhou, University of Chinese Academy of Sciences, Hangzhou, China

   Kuo Liu, Zhongxiao Wang, Xueying Yang, Xufeng Li & Bin Zhou

2. CAS CEMCS-CUHK Joint Laboratory, New Cornerstone Science Laboratory, Key Laboratory of Multi-Cell Systems, Shanghai Institute of Biochemistry and Cell Biology, Center for Excellence in Molecular Cell Science, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Shanghai, China

   Kuo Liu, Zixin Liu, Muxue Tang, Shan Yang, Yanli Zhang, Zan Lv, Hengwei Jin, Wenjuan Pu, Huan Zhao, Pengfei Sui & Bin Zhou

3. School of Life Science and Technology, ShanghaiTech University, Shanghai, China

   Xinfeng Meng, Fanglin Di & Bin Zhou

4. Department of Pulmonary and Critical Care Medicine, Shanghai Chest Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China

   Chenfei Li & Feng Li

5. Shanghai Key Laboratory of Lung Inflammation and Injury, Shanghai, China

   Pengfei Sui

6. CAS CEMCS-CUHK Joint Laboratory, Department of Chemical Pathology, Li Ka Shing Institute of Health Sciences, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China

   Bin Zhou

Contributions

K.L. and B.Z. conceived and designed the study and wrote the manuscript; K.L., Z. Liu., X.M., M.T., H.J., W.P. and H.Z. performed the experiments and contributed to data interpretation. K.L. and X.M. performed the in vitro cellular experiments. Z. Liu, M.T., Z.W. and X.Y. performed single-cell and spatial transcriptomics analyses. K.L., F.D. and C.L. conducted pulmonary function measurements. S.Y., Y.Z., Z. Lv, X.L., H.J., W.P. and H.Z. managed mouse breeding, genotyping and colony maintenance. F.L. and P.S. provided expert oversight and contributed to the writing of the manuscript. B.Z. supervised the project, coordinated collaboration among authors and assumed overall responsibility for the study. All of the authors approved the final manuscript.

Corresponding author

Correspondence to Bin Zhou.

Competing interests

The authors declare no competing interests.

Peer review information

Nature thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available.

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