Novel netrin-1–Unc5b–Klf4 axis balances AT2 survival in experimental bronchopulmonary dysplasia

Abstract

Alveolar growth and repair are central processes in development and chronic lung disease, such as bronchopulmonary dysplasia (BPD), a neonatal lung disease without curative therapy. Alveolar epithelial type 2 (AT2) cells are the endogenous progenitor pool giving rise to alveolar epithelial type 1 cells and promoting alveolar repair. Since netrin-1, a regulator of cell homeostasis and stemness, has been linked to lung diseases, we now investigated its signaling and function in AT2 cells in a hyperoxia-based model of BPD and in lungs of infants with BPD. First, we demonstrated that prolonged hyperoxia reduced both netrin-1 and its receptor Unc5b in neonatal mouse lungs and in primary AT2 cells. Second, ex vivo studies using precision-cut lung slices (PCLS) and primary murine AT2 cell culture showed that netrin-1 regulates AT2 cell survival and the expression of Krüppel-like factor 4 (Klf4) through Unc5b, a transcription factor regulating cell survival. Third, single-cell and bulk transcriptomic analysis, as well as proximity-dependent biotin identification assay, showed Klf4 to be upregulated in AT2 cells during alveolarization, downstream of netrin-1, and to regulate AT2 cell survival. In vivo, Klf4 gene expression and protein abundance was significantly reduced in total lung homogenates and in AT2 cells of neonatal mice exposed to hyperoxia. Finally, KLF4⁺ cells, KLF4⁺ epithelial, and specifically KLF4⁺ AT2 cells were reduced in clinical BPD. In summary, our data identify a novel netrin-1–Unc5b–Klf4 axis in AT2 cells that is disrupted in BPD and could offer a novel target for endogenous alveolar repair.

Introduction

Chronic lung diseases (CLDs) are a leading cause of morbidity and mortality worldwide according to the World Health Organization. In preterm infants, bronchopulmonary dysplasia (BPD) is the most common CLD, characterized by failed alveolar formation and perturbed extracellular matrix remodeling.^1-3 Despite extensive clinical and experimental studies, current therapies are mainly supportive. In particular, causative molecular mechanisms and curative treatments to reverse the pathology are lacking and structural and functional sequelae persist into adulthood, highlighting the clinical relevance of this study. Thus, identification of novel endogenous molecular mechanisms may serve as a new promising target to regenerate or rebuild the lung in preterm infants and in adults with CLD.

Alveolarization is orchestrated by the precise spatial-temporal interplay of growth factors, cytokines, and transcription factors. Disruption of these processes by premature birth, mechanical ventilation, and/or prolonged hyperoxia (HYX) leads to an arrest of alveolarization with a reduced gas exchange surface area that extends beyond infancy.^4,5 These long-term sequelae include an emphysema-like lung structure and indicate a reduced endogenous regenerative capacity of the lung. Beyond the understanding of the disease mechanisms, the activation of endogenous regenerative resources of the lung is important. Since alveolarization continues into early adulthood, this early phase appears to be the optimal window to promote repair.^6,7

The regenerative capacity of the lung depends on resident progenitor cells in the distal lung that can self-renew and generate differentiated progeny.^8-10 The epithelial alveolar lining is predominantly composed of alveolar epithelial type 1 (AT1) and type 2 (AT2) cells.¹¹ AT1 cells cover about 95% of the alveolar surface and replenish the alveolar epithelial barrier during repair after lung injury. In contrast, AT2 cells give rise to AT1 cells under physiological conditions and following injury, thereby serving as progenitor cells for the alveolar regenerative niche.^11-13 Prolonged exposure of newborns to HYX adversely affects AT2 cell homeostasis and lung regeneration.¹⁴ So far, the molecular mechanisms maintaining AT2 progenitor cell capacity and survival remain elusive.

Netrin-1 (Ntn1) is a secreted laminin-related molecule that was initially described as an axon guidance cue and more recently reported as a pleiotropic protein implicated in cardiopulmonary diseases.^15-17 Interestingly, it also plays a role in the regulation of stemness, somatic cell function, and the protection of pluripotent cells from apoptosis.¹⁸ The Krüppel-like factors (Klf) comprise a family of transcription factors that is involved in regulating cell homeostasis and repair, including differentiation, survival, and stemness.¹⁹ In particular, Klf4 is crucial in the induction and maintenance of pluripotency.²⁰ Klf4 also regulates various central biological processes, eg, epithelial cell survival and differentiation.²¹ While a recent study suggests a stromal Ntn1–Klf4 axis that regulates stromal differentiation and renal blood vessel formation,²² the role of netrin-1 and Klf4 in AT2 cells and aberrant alveolarization remains elusive. Here, we now provide evidence of a novel netrin-1–Unc5b–Klf4 axis in AT2 cells, which is central in survival. Our study demonstrates that Klf4 expression is reduced in lung epithelium and AT2 cells of infants with BPD and is pivotal for AT2 cell homeostasis in a murine model of BPD.

Materials and methods

Human lung tissue

Postnatal de-identified human lung tissue, both BPD and non-BPD (control), were obtained from the LungMap tissue biorepository held at the University of Rochester’s BioRepository for INvestigation of Diseases of the Lung (BRINDL). Tissue is considered exempt from human subject regulation, as specified by the University of Rochester Research Subjects Review Board protocol (RSRB00056775). Informed consent was provided for each sample collected and used in this study. To determine a combined diagnosis based on clinical history and pathology of the lung tissue (ClinPathDx, Table S4, previously published²³), clinical metadata and histopathology was assessed for each sample.

Immunohistochemistry of human lung tissue

Paraffin-embedded BPD and control (nondiseased) postnatal human lungs were processed and stained as previously described.²⁴ For a detailed description, see the Supplementary material.

Animal studies

Single-cell RNA sequencing data processing and analyses

To study expression changes in epithelial cells at the single-cell level, a previously published, publicly available single-cell RNA sequencing (scRNA-seq) dataset from newborn mice was reanalyzed.²⁵ A single AT2 cell population was considered for this study. This was generated by merging the AT2 and AT2-Lyz1⁺ cluster identities. For a detailed description of the differential expression analysis and gene set enrichment analysis (GSEA), see the Supplementary material.

Experimental setup for lung development studies and hyperoxia-induced neonatal lung injury

All animal procedures were performed in accordance with German regulations and legal requirements and were approved by the local government authorities (Landesamt für Natur, Umwelt und Verbraucherschutz (LANUV), Nordrhein Westphalia (NRW), Germany; AZ.2010.A372; AZ.50.15.015; AZ.40.14.013; AZ.2015.A142; AZ.2015.A120; Stanford University Administrative Panel on Laboratory Animal Care protocol 19087). Male and female C57BL/6N (wild-type [WT]) mice were allowed food and water ad libitum and were housed in humidity- and temperature-controlled rooms exposed to a 12-hour dark/light cycle. Pups were sacrificed at embryonic day 15 (E15), E17, and E19 and postnatal day 1 (P1), P3, P7, P14, P21, and P28. For experimental setup for HYX-induced neonatal lung injury, pups from 2 to 8 litters were randomized, pooled, and equally distributed between dams on the day of birth (born within 12 hours of each other). Litters were either exposed to 85% (vol/vol) O₂ (HYX) or maintained at room air (21% [vol/vol] O₂; normoxia [NOX]) as previously described.²⁶ For a detailed description, see the Supplementary material.

Processing of lung tissue for histology

For histological studies, lungs were pressure-fixed intratracheally at 20 cm H₂O with phosphate-buffered 4% paraformaldehyde (PFA) (#P0873, Roth, Karlsruhe, Germany) overnight, followed by paraffin embedding for isotropic uniform random sectioning as previously described.²⁷ For a detailed description, see the Supplementary material.

Quantitative histomorphometric analysis

Hematoxylin and eosin–stained 3-μm sections were used for quantitative histomorphometry as previously described.²⁷ The radial alveolar count (RAC) method was used to count the number of alveoli across a terminal bronchus.²⁸ ImageJ-win64 (National Institutes of Health [NIH]) was used to measure the average surface of a single alveolus (alveolar surface area).²⁹ For a detailed description, see the Supplementary material.

Immunofluorescence staining

For Klf4 immunofluorescence (IF) staining, lung sections were deparaffinized and rehydrated. For a detailed description, see the Supplementary material. In brief, sections were incubated with anti-rabbit KLF4 diluted in antibody diluent (#4038, 1:150, Cell Signaling, Danvers, MA, USA), mouse anti-ATP binding cassette subfamily A member 3 (ABCA3) (#WMAB-ABCA3-17, 1:900, Seven Hills Bioreagents, Cincinnati, OH, USA), rabbit anti-mouse pro-surfactant protein C (SFTPC) (#AP3786, 1:500, Millipore, Burlington, MA, USA), chicken anti-mouse NTN1 (#NB100-1605, 1:300, Novus Biologicals, Centennial, CO, USA), rabbit anti-mouse UNC5B (#bs-11495R, 1:50, Bioss Antibodies, Woburn, MA, USA), rat anti-mouse Ki67 (#14-5698-82, 1:150, Invitrogen, Carlsbad, CA, USA), and rabbit anti-mouse Caspase 3 (#AF856, R&D Systems, 1:200, Minneapolis, MN, USA); cell nuclei were stained with DAPI (#D9542, Sigma-Aldrich, St Louis, MO, USA). The total number of positive as well as the number of double-positive cells were counted manually from 8 to 10 images of different areas, with the size of the tissue area 14.50 × 10.83 µm (1392 × 1040) per lung section using ImageJ-Fiji-win64 (NIH).³⁰ Fluorescence intensity was measured using ImageJ-Fiji-win64 out of images taken by Leica K5 wide-field fluorescence microscope with 63×/1.40 Numerical Aperture (NA) oil immersion objective, and upon background signal subtraction. Intensity was measured in AT2 (ABCA3 or SFTPC)–positive cells. AT2-positive cells were selected manually using the DAPI channel as a mask.

Immunocytochemistry

Cytospins were obtained by centrifugation of freshly isolated primary mouse AT2 cells (10 minutes at 300g, 100 000 cells/spin). Cells were fixed with 4% PFA for staining using the following antibodies: anti-rabbit KLF4 (#4038, 1:150, Cell Signaling), anti-rabbit SFTPC (#AP3786, 1:500, Millipore), chicken anti-mouse NTN1 (#NB100-1605, 1:300, Novus Biologicals), rabbit anti-mouse UNC5B (#bs-11495R, 1:50, Bioss Antibodies), or mouse anti-ABCA3 (#WMAB-ABCA3-17, 1:900, Seven Hills Bioreagents); cell nuclei were stained with DAPI (#D9542, Sigma-Aldrich). Cytospins were analyzed using a microscope (Olympus BX43, Olympus, Tokyo, Japan).

Immunoblotting

Immunoblotting was performed on protein isolated as previously described.³¹ Blots were incubated with the following antibodies: rabbit anti-mouse KLF4 (#4038, 1:100, Cell Signaling) or rabbit anti-mouse CDH1 (#3195, 1:2000, Cell Signaling), with mouse anti-mouse β-actin (#3700, 1:5000, Cell Signaling) serving as the loading control. Horse anti-mouse IgG, HRP-linked (#7076, Cell Signaling) and goat anti-rabbit IgG, HRP-linked (#7074, Cell Signaling) were used as secondary antibodies.

RNA extraction and real-time quantitative PCR

Total RNA was isolated from lung tissue using TRIzol and qRT-PCR was performed as described previously, using the 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA).³¹ The relative amount of the specific mRNA was normalized to Gapdh or Actb (β-actin). Primers and TaqMan probes were designed using Primer Express Software and are listed in Table S5.

Precision-cut lung slices

Murine lungs were isolated from 21-day-old mice. For a detailed description of the generation of precision-cut lung slices (PCLS), see the Supplementary material. The lung sections were then treated with 3 µg of recombinant netrin-1 (preincubated for 30 minutes at room temperature), 3 µg netrin-1 + 30 µg Fc control recombinant protein, or 30 µg recombinant UNC5B-Fc; PBS was used as vehicle for 48 hours. Afterward, PCLS were collected for RNA isolation or for fixation with 4% PFA and paraffin embedding. Next, sections of 3 µm thickness were produced and IF staining for anti-SFTPC (#WRAB-9337, Seven Hills Bioreagents) was combined with anti-KLF4 (#AF3158, R&D Systems). The KLF4 signal was amplified with the Alexa Fluor 594 Tyramide SuperBoost Kit, goat anti-mouse IgG (#B40915, Thermo Scientific, Waltham, MA, USA). For quantification, photomicrographs of representative areas were taken using a Leica K5 wide-field fluorescence microscope with 63×/1.40 NA oil immersion objective, (Leica, Wetzlar, Germany). For representative images, we used a confocal laser scanning microscope Olympus/Evident FV3000, with 40×/1.25 NA silicon oil immersion objective (Olympus/Evident) using CellSens Dimension software (Olympus). The total number of KLF4⁺ nuclei (cells) as well as the number of KLF4⁺ and SFTPC⁺ cells (double-positive) were counted manually from 8 to 10 images of different areas, with the size of the tissue area 14.50 × 10.83 µm (1392 × 1040) per lung section using ImageJ-Fiji-win64 (NIH)³⁰; lungs of 4 different mice were used.

Cell culture studies

Murine AT2 cells

Primary murine AT2 cells were isolated from C57BL/6N mice as described previously.³²

For a detailed description, see the Supplementary material. The AT2 cells were then treated with 0.5 µg of recombinant netrin-1 or preincubated for 30 minutes at room temperature with 0.5 µg netrin-1 + 5 µg recombinant UNC5B-Fc; PBS was used as vehicle for 48 hours.

Cell lines

Mouse lung epithelial cells

Mouse lung epithelial cells (MLE12; ATCC, CRL-2110; Manassas, VA, USA) were cultured according to the ATCC recommendations and as previously described.²³ For a detailed description, see the Supplementary material.

Production of recombinant proteins

For recombinant protein production, stable HEK293 EBNA cell lines were generated employing the Sleeping Beauty transposon system.³³ For a detailed description, see the Supplementary material.

3′ RNA sequencing analyses

RNA sequencing (RNA-seq) data of MLE12 cells were processed through the QuickNGS pipeline from Wagle  et al., Ensembl version 93(mm10).³⁴ For a detailed description, see the Supplementary material. Promoters were defined based on proximity with transcription start site (TSS) described for the genes. For the analysis, the 2000 bp region upstream of the TSS was selected and CiiiDER software was used.^35,36 Distal enhancers were not included.

RNA-seq data of primary AT2 cells treated with netrin-1 were analyzed by the genomic facility of the Institute for Lung Health (ILH) at the Justus-Liebig University (JLU), Giessen, Germany. For genome-wide analysis of gene expression, RNA-seq libraries from isolated mRNA were generated and sequenced by the ILH–Genomics and Bioinformatics at JLU. For a detailed description, see the Supplementary material.

BioID2 experiments

The DNA sequence of BioID2 HA tag was fused to the C-terminus of murine Klf4 (NP_034767.2, aa: 1-483) and cloned into Sleeping Beauty transposon vectors. A control plasmid was generated by cloning of BioID2 HA tag into Sleeping Beauty vectors. MLE12 cells were stably transfected with FuGENE HD transfection reagent (#E2311, Promega, Mannheim, Germany). The protein expression was then induced with 1 µg/mL doxycycline (#D9891, Sigma-Aldrich) for 24 hours. For a detailed description, see the Supplementary material.

Overexpression of Klf4, netrin-1, and mutated netrin-1 using a Sleeping Beauty transposon system

Murine Klf4 (NP_034767.2, aa: 1-483) and Ntn1 (NP_032770.2, aa: 25-604) containing a C-terminal HA tag or a mutated form of Ntn1 (R347A; R348A), which is not able to bind to its receptor Unc5b, were cloned into a commercially available plasmid pSBbi-Pur containing a Sleeping Beauty transposon (#60523, Addgene, Watertown, MA, USA) (Plasmid ID KV2620), which was a gift from Eric Kowarz.³³ For a detailed description, see the Supplementary material.

Generation of a Klf4 mutant murine lung epithelial cell line using CRISPR/Cas9 engineering

A Klf4^del MLE12 cell line was generated using CRISPR/Cas9 engineering as previously described.³⁷ For a detailed description, see the Supplementary material.

Bromodeoxyuridine (BrdU) incorporation assay

Cells were seeded into 96-well plates and grown for 24 hours. BrdU Cell Proliferation Assay Kit (#11647229001, Roche, Basel, Switzerland) was used according to the manufacturer’s instructions.

Caspase-3/7-Glo activity assay

The caspase-3/7 activity assay (#G8090, Promega) was performed according to the manufacturer’s protocol.

Statistical analysis

For statistical analyses, a Mann–Whitney U test was performed, when comparing 2 groups. For comparison of 4 groups (comparing 2 independent variables), a one-way or 2-way ANOVA with posttest (column matching) and a Mann–Whitney U test were conducted. Significances were indicated in the graph representing the mean value ± standard error of the mean (SEM). Data were graphically represented using GraphPad Prism 7 software (GraphPad Software, LLC).

Results

Acute and prolonged exposure to hyperoxia causes depletion of SFTPC⁺ cells and reduction of netrin-1–Unc5b in lungs during alveolarization

To test if the impaired alveolar regenerative capacity is linked to a loss of AT2 cells, we first performed IF staining of neonatal mouse lungs after exposure to HYX at postnatal day 7 (P7; early alveolarization), P14 (peak of alveolarization), and P28 (mature lungs) for SFTPC. The quantification of SFTPC⁺ cells showed a marked progressive loss of AT2 cells from P7 to P28 of more than 50%. Complimentary quantitative histomorphometry revealed that the loss of AT2 cells was followed by a reduced alveolar growth after HYX (Figure 1A-E).

Figure 1.

Hyperoxia (HYX) causes loss of alveolar epithelial type 2 (AT2) cells, arrest of alveolarization, and reduction of netrin-1–Unc5b expression in AT2 cells. Wild-type mice were exposed to normoxia (NOX; 21% O₂) or HYX (85% O₂) from postnatal day 1 (P1) to P28. (A) Representative immunofluorescence (IF) stainings for surfactant protein C (SFTPC, green) as an indicator of AT2 cells in lungs at P7, P14, and P28. DAPI was used for nuclear staining; scale bars: 30 µm. (B) Quantitative data summary of SFTPC⁺ cells per field of view. (C) Representative images of hematoxylin and eosin–stained lung sections at P7, P14, and P28; scale bars: 100 µm. (D, E) Summary data of quantitative histomorphometric analyses: radial alveolar counts (RAC) per 100 µm (D) and average surface of a single alveolus (E). (F) Heatmap of the transcriptomic analysis of netrin signaling molecules of lungs of newborn mice exposed to NOX or HYX for 14 days (n = 6 per group). The heatmap contains only genes with an average log-fold change (FC) < 2 and false discovery rate–corrected P value < .01. (G, H) Gene expression of Ntn1 and the dependence receptor Unc5b in lungs of newborn mice exposed to NOX or HYX at P7 and P14 using qRT-PCR. (I-K) Representative IF stainings for SFTPC (red) as an indicator of AT2 cells, co-localized with NTN1 (green) in lungs of mice exposed to NOX or HYX at P7, P14, and P28; DAPI was used for nuclear staining. White arrowheads depict SFTPC⁺NTN1⁺ cells; scale bars: 10 µm (I). Quantitative data summary of NTN1⁺ cells per field of view (J), and summary of SFPTC⁺ and NTN1⁺ double-positive relative to all SFTPC⁺ cells per field of view (K). (L-N) Representative images of IF staining for ABCA3 (green) as an indicator of AT2 cells, co-localized with UNC5B (red) in lungs exposed to NOX or HYX at P7, P14, and P28, DAPI was used for nuclear staining. White arrowheads depict ABCA3⁺ UNC5B⁺ cells; scale bars: 30 µm (L). Quantitative data summary of UNC5B⁺ cells per field of view (M) and quantitative data summary of ABCA3⁺ and UNC5B⁺ cells relative to all ABCA3⁺ cells per field of view (N). (O, P) Representative IF staining of NTN1 (green) and UNC5 (green) with SFTPC (red) and ABCA3 (red) in primary cultured murine AT2 cells, respectively (O). Gene expression of Ntn1 and Unc5b in primary cultured murine AT2 cells exposed to NOX or HYX using qRT-PCR (P). In representative IF stainings, field of view corresponds to 0.040 mm². Data are presented as mean ± SEM. *P < .05, **P < .01, ***P < .001, ****P < .0001 (2-way ANOVA with Sidak posttest); ^#  P < .05, ^##  P < .01 (Mann–Whitney U test); n = 4-7 per group.

Netrin-1 regulates stemness and regenerative processes¹⁸ but has not been investigated extensively in lung development or BPD to date. We found an up to 8-fold induction of Ntn1 mRNA during the transition from saccular to alveolar stage (Figure S1A). To determine whether netrin-1 is regulated by acute postnatal injury, we investigated lungs of newborn WT mice exposed to prolonged HYX from birth until P14 using a publicly available RNA-seq dataset from our group.²³ Genes of the netrin signaling machinery were significantly downregulated in neonatal lungs after HYX when compared to NOX; notable among those genes, Ntn1, netrin G2 (Ntng2), netrin-3 (Ntn3), and netrin-1 receptors such as Uncoordinated-5 (Unc5) (Figure 1F). These data were confirmed for Ntn1 and Unc5b using qRT-PCR at P7 and P14 (Figure 1G, H). Additional IF staining for NTN1 and UNC5B protein showed a co-localization with AT2 cells (SFTPC or ABCA3, respectively). The loss of AT2 cells during neonatal lung injury caused by prolonged HYX was related to a marked reduction of NTN1⁺ and UNC5B⁺ cells at P7, P14, and P28 by up to 80%. Likewise, the quantification of the co-IF staining with AT2 markers (SFTPC and ABCA3) showed a marked loss of NTN1⁺ and UNC5B⁺ AT2 cells (Figure 1I-N) as well as reduced abundance of NTN1 and UNCB5 in AT2 cells in lungs after HYX (Figure S1B). In addition, cultured primary murine AT2 cells showed positive IF staining for NTN1 and UNC5B, confirming expression of both in AT2 cells; exposure to HYX, however, reduced the expression of Ntn1 and Unc5b in cultured primary murine AT2 cells by more than 50%. (Figure 1O, P).

Netrin-1–Unc5b axis regulates AT2 cell homeostasis

Based on the previous findings, we next investigated the functional role of netrin-1–Unc5b axis in AT2 cells in vitro. To this end, we isolated primary AT2 cells from 21-day-old mice and treated the cells with either netrin-1 or vehicle (Figure 2A-C). First, we analyzed the transcriptomic profile using bulk RNA-seq and defined 324 differentially expressed genes (DEGs; P  < .05) when compared to vehicle-treated AT2 cells; notable among those were genes that regulate cell proliferation and differentiation (eg, Afap1l1, Bex1, and Gata4)^38-40 (Figure 2B, Figure S2A). Interestingly, Dlk1, a regulator of AT2 to AT1 differentiation was markedly reduced by netrin-1 (Figure 2B). Pathway annotation for DEGs identified processes involved in cell cycle regulation and development (Figure 2C). Further Metascape analysis of DEGs revealed a regulation of biological processes involved in cell survival, differentiation, and development, eg, cellular response to transforming growth factor beta (TGF-β) stimulus, focal adhesion PI3K Akt mTOR signaling pathway, and embryonic morphogenesis (Figure S2B). Collectively, the transcriptomic analysis points out an important role for netrin-1 signaling in AT2 cell homeostasis and epithelial development. To further identify transcription factors that could be involved in netrin-1–dependent regulation of AT2 cells, we used CiiiDER software, a tool for predicting transcription factor binding sites, to perform a gene promotor analysis of the top 100 up- and downregulated genes with P values <.05, using a deficit value of 0.1.³⁵ We found Klf4 among the transcription factors that had a binding site at most DEGs (91.69%; Figure 2D).

Figure 2.

Investigating a netrin-1–Unc5b–Klf4 axis in primary murine alveolar epithelial type 2 (AT2) cells. (A-D) The functional role of netrin-1 was analyzed by 3′ RNA-seq in primary murine AT2 cells (n = 4-5 independent samples per group). The volcano plot shows differentially expressed genes (DEGs) in the transcriptome of netrin-1–treated primary murine AT2 cells. Genes with log₂ fold change (FC) value >0.5 and P value <.05 were considered DEGs. The x- and y-axes show average log₂ FC and -log₁₀  P value, respectively (A, B). Top 10 pathways regulated by significantly differentially expressed genes in netrin-1–treated AT2 cells were identified using String DB, Biological Processes analysis. The adjusted P value for each pathway is shown in brackets (C). Pie chart depicting the putative binding sites for the transcription factor Krüppel-like factor 4 (Klf4) in the promoter regions of the top 100 upregulated and downregulated genes with adjusted P values <.05 in netrin-1–treated AT2 cells (D). (E-G) Primary murine AT2 cells were treated with recombinant netrin-1 for 48 h, followed by assessment of cell proliferation (E), apoptosis (F), and Klf4 gene expression (G) using BrdU assay, caspase-3/7 Glo assay, and qRT-PCR, respectively. (H-K) Scheme illustrating the treatment of primary murine AT2 cells with recombinant netrin-1 combined with Unc5b-Fc (netrin-1 + Unc5b-Fc) or Fc (control recombinant protein) for 48 h (H), followed by assessment of cell proliferation (I), apoptosis (J), and Klf4 gene expression (K) using BrdU assay, caspase-3/7 Glo assay, and qRT-PCR, respectively. (L-N) Scheme illustrating the treatment of MLE12 cell line, wild-type, and CRISPR/Cas9 Klf4 mutant with recombinant netrin-1 for 24 h (L), followed by assessment of cell proliferation (M, N). Actb served as a housekeeping gene. Data are presented as mean ± SEM. Mann–Whitney U test (E-G, I-K), Wilcoxon matched-pairs signed-rank test (M, N): *P < .05, **P < .01, ***P < .001. “Created in BioRender. Alejandre Alcázar, M. (2026) https://BioRender.com/wk6udld”.

To confirm a netrin-1–Unc5b–Klf4 axis in lung epithelial cells, we used 2 approaches: First, we isolated primary AT2 cells from 21-day-old mice and treated the cells with netrin-1. Netrin-1 induced an almost 3-fold increase of AT2 cell proliferation and a >15-fold increase of Klf4 gene expression, an effect that is completely prevented by blockade of Unc5b receptor using Unc5b-Fc; in contrast, apoptosis was not altered (Figure 2E-K). Similarly, netrin-1 significantly increased the viability of MLE12 cells. In contrast, this positive effect of netrin-1 was not detected in MLE-12 cells with deletion Klf4 (Klf4del), indicating that netrin-1 in part requires Klf4 for prosurvival cell signaling. Since MLE12 cells are immortalized murine lung epithelial cells, the effect of netrin-1 that we observed in primary murine AT2 cells is not as pronounced. Second, Ntn1 was overexpressed in MLE12 (Ntn1) using a Sleeping Beauty transposon system and maintained under either NOX or HYX. Netrin-1 significantly increased KLF4 and CDH1 protein along with SftpC expression under normoxic and hyperoxic conditions. To determine if these changes are mediated via the netrin-1 receptor Unc5b, MLE12 cells overexpressing a mutated form of Ntn1, which is unable to bind to Unc5b, were generated (m.Ntn1). While CDH1 protein was significantly increased, KLF4 protein and SftpC gene expression remained unchanged, suggesting that CDH1 is controlled by netrin-1 receptors other than Unc5b, whereas changes in KLF4 protein and SftpC are mediated via Unc5b (Figure S3A-F). In summary, our data demonstrate a netrin-1–Unc5b axis that regulates expression of Klf4 and homeostasis of AT2 cells, an important process in alveolar formation and repair.

Netrin-1–Unc5b axis regulates Klf4 in AT2 cells in vivo and impacts alveolar growth

To further confirm that the netrin-1–Unc5b axis regulates Klf4 in AT2 cells, we used PCLS of 21-day-old mice. Treatment of PCLS with netrin-1 for 48 hours induced Klf4 gene expression. In addition, we performed double IF staining of PCLS for KLF4 and ABCA3 (marker of AT2 cells). Quantitative analysis confirmed that netrin-1 increased the abundance of KLF4⁺ cells (nuclear staining) as well as KLF4 and ABCA3 double-positive cells in PCLS (Figure 3A-D). Treatment with Unc5b-Fc, however, blocked the netrin-1–mediated Klf4 gene expression and the increase in KLF4⁺ nuclei as well as AT2 cells with KLF4⁺ nuclei, supporting a netrin-1–Unc5b–Klf4 axis in AT2 cells (Figure 3E-H). In addition, we assessed the effect of netrin-1 on proliferation and apoptosis of AT2 cells in PCLS using co-IF staining of SFTPC with Ki67 or ABCA3 with active caspase-3, respectively. We avoided edge effects at the cut surfaces by deep sectioning of the PCLS. We detected increased proliferation of AT2 cells through netrin-1–Unc5b signaling (Figure 3I, J), whereas apoptosis was not affected (Figure S4A, B).

Figure 3.

Netrin-1 induces Krüppel-like factor 4 (Klf4) through Unc5b in alveolar epithelial type 2 (AT2) cells and regulates alveolar formation. (A-H) Representative images of co-immunofluorescence (IF) staining for ATP binding cassette subfamily A member 3 (ABCA3, red) as a marker for AT2 cells and KLF4 (green, nuclear staining) in precision-cut lung slices (PCLS) from 21- to 22-day-old mice. PCLS were treated with recombinant netrin-1 vs PBS (vehicle) (A) or netrin-1 with Fc (control recombinant protein) vs netrin-1 with Unc5b-Fc (E) for 48 h. DAPI was used for nuclear staining. White arrowheads denote ABCA⁺KLF4⁺ cells; scale bars: 30 µm. Assessment of Klf4 gene expression using qRT-PCR; Actb served as a reference gene (B, F). Quantitative data summary of KLF4⁺ (C, G) or KLF4⁺ABCA3⁺ (AT2) cells (D, H) per field of view. (I, J) Representative images of IF staining for surfactant protein C (SFTPC, red) as an indicator for AT2 co-localized with Ki67 (green; proliferation marker) in PCLS from 21- to 22-day-old mice. PCLS were treated with recombinant netrin-1 vs PBS (vehicle) (I) or netrin-1 with Fc (control recombinant protein) vs netrin-1 with Unc5b-Fc (J) for 48 h. DAPI was used for nuclear staining. White arrowheads denote SFTPC⁺Ki67⁺ cells; scale bars: 10 µm. Quantitative data summary of Ki67⁺ or SFTPC⁺Ki67⁺ AT2 cells per field of view. (K-N) Primary murine AT2 cells were treated with recombinant netrin-1 vs PBS (vehicle) or netrin-1 with Fc (control recombinant protein) vs netrin-1 with Unc5b-Fc for 24 hours. Assessment of gene expression using qRT-PCR: inflammatory markers Il1b (K), Il6 (L), Tnfa (M), and Cxcl10 (N). Data are presented as mean + SEM. ns, not significant; *P < .05, **P < .01; Mann–Whitney U test.

Since netrin-1 has an immunomodulatory function, we treated cultured primary murine AT2 cells with netrin-1 alone or in combination with Unc5b-Fc. Interestingly, netrin-1 alone induced an Unc5b-dependent expression of Il6, whereas other inflammatory genes (Il1b, Tnfa, and Cxcl10) were not induced by netrin-1 (Figure 3K-N).

Klf4 is regulated during alveolarization and markedly reduced in AT2 cells after neonatal lung injury

Previous studies have shown a regulation of Klf4 by oxygen after birth and its contribution to homeostasis of mesenchymal cells and fibrotic processes.^41,42 However, the role of the transcription factor Klf4 in alveolarization, injury, and repair remains elusive. Hence, we first investigated the spatio-temporal expression of Klf4 in WT mice (C57BL/6) during the pseudoglandular/canalicular (E15, E17), saccular (E19, P1, P3), and alveolar (P7, P21) stages. On a transcriptional level, we found a 40-fold increase of Klf4 mRNA during the saccular and alveolar stages when normalized to E15. On a protein level, we determined a >2-fold increase in KLF4 protein during the alveolar stage of development (P7 and P21) compared to the saccular stage (P3) and mature lungs (P28) (Figure 4A-C).

Figure 4.

Regulation of Krüppel-like factor 4 (Klf4) in neonatal lungs and alveolar epithelial type 2 (AT2) cells in normal development and after hyperoxia. (A; PG: pseudoglandular; CA: canalicular) Gene expression of Klf4 in total lung homogenates using qRT-PCR at embryonic day 15 (E15), E17, and E19 and at postnatal day 1 (P1), P3, P7, and P21; Gapdh served as reference gene. E15 was set to 1 (Kruskal–Wallis test followed by Dunn posttest). (B-D) KLF4 protein abundance during late lung development (B, C; Kruskal–Wallis test followed by Dunn posttest) and in lungs of mice exposed to normoxia (NOX; 21% O₂) or hyperoxia (HYX; 85% O₂) at P28 (D; Mann–Whitney test). GAPDH or β-actin served as a loading control. (E, F) Uniform manifold approximation and projection (UMAP) plot depicting the epithelial cell clusters in developing mouse lungs, including a single AT2 cell population. In the dataset reanalyzed here, mice were exposed to NOX or HYX from P1 onward and lungs were harvested at P3, P7, or P14 (E). UMAP plot showing the expression of Klf4 mRNA within the developing lung epithelium (F). (G) Dot plot showing the expression of Klf4 within the individual epithelial populations from NOX- or HYX-exposed lungs. (H–J) Dot plots depicting the expression of Klf4 within the lung epithelial populations at P3 (H), P7 (I), and P14 (J) in mice exposed to NOX or HYX. (K) Identification of signaling pathways downregulated in the AT2 cluster of lungs after HYX using gene set enrichment analysis (GSEA). (L) Top signaling pathways significantly downregulated in the AT2 cluster following HYX exposure in which decreased Klf4 was determined. All terms are significantly enriched (adjusted P < .05). Normalized enrichment score values were computed by GSEA on fold change–ranked genes. Expression levels in dot plots and UMAP plots are presented as log(TP10k + 1) values. Log(TP10k + 1) corresponds to log-transformed unique molecular identifiers per 10k. (M-O) Representative images of immunofluorescence (IF) staining for ABCA3 (red; AT2) co-localized with KLF4 (green) in lungs at P7, P14, and P28 after exposure to NOX or HYX. DAPI was used for nuclear staining; scale bars: 30 µm (M); white arrowheads depict KLF4⁺ and ABCA3⁺ cells. Quantitative data summary of KLF4⁺ cells per field of view (N), and ABCA3⁺ and KLF4⁺ related to all ABCA3⁺ cells per field of view (O). The field of view corresponds to 0.040 mm²; 2-way ANOVA followed by Sidak posttest. (P) IF staining of KLF4 and SFTPC in primary murine AT2 cells. Klf4 gene expression in primary murine AT2 cells after HYX or NOX for 24 h using qRT-PCR. 18s rRNA served as reference gene; Mann–Whitney t-test. Data are presented as mean ± SEM. **P < .01, ****P < .0001 (Kruskal–Wallis test followed by Dunn posttest; 2-way ANOVA followed by Sidak posttest); ^##P < .01 (Mann–Whitney U test).

Since we showed that netrin-1–Unc5b regulates Klf4 in AT2 cells, we next investigated Klf4 in acute neonatal lung injury and found that KLF4 protein abundance was significantly reduced by 50% in lungs of newborn mice exposed to prolonged HYX (Figure 4D). To investigate the spatio-temporal transcriptomic profile of Klf4 mRNA expression and protein abundance in AT2 cells during development and neonatal HYX, we pursued 3 approaches: First, we studied Klf4 expression patterns in the developing lung on the single-cell level. To this end, we used a publicly available mouse lung scRNA-seq dataset.²⁵ Within this dataset, the authors analyzed 10 060 epithelial cells from lungs of newborn mice exposed to NOX (21% O₂) or HYX (85% O₂) at P3, P7, and P14, comprising 5 clusters, including 2 AT2 populations.²⁵ For the purposes of this study, we have considered both AT2 clusters as a single AT2 population (Figure 4E). Most Klf4-expressing cells were identified as AT2 and AT1 cells, with highest expression levels present within the AT2 population (Figure 4F, G). Notably, Klf4 gene expression mainly increased within the AT2 cells in normally developing mice (Figure S5A). However, this increase of Klf4 mRNA during alveolarization was not evident to the same extent in AT2 cells from neonatal mice exposed to HYX (Figure S5B), resulting in decreased numbers of Klf4⁺ AT2 cells in HYX-exposed mice at all investigated timepoints (Figure 4H-J). Next, we analyzed the transcriptomic profile of AT2 cells after prolonged neonatal HYX employing differential state analysis, followed by GSEA (Figure 4K-L) (Tables S1 and S2). GSEA in the AT2 population revealed the regulation of multiple pathways involved in epithelial biology (eg, epithelium cell development, epithelium cell differentiation, and cell polarity), but also in angiogenesis (eg, vascular development, endothelium development, and artery morphogenesis). Both epithelial homeostasis and angiogenesis are central in alveolar formation and regeneration, indicating a cell-autonomous and non-cell-autonomous function of HYX on AT2 cells (Figure 4K). We next screened the signaling pathways downregulated in the AT2 cluster of lungs after HYX (using GSEA) for involvement of Klf4. The pathways in AT2 cells in which Klf4 is regulated are shown. Notable among those pathways were Wnt signaling, DNA binding, cell surface receptor signaling, and axon development—signaling cascades important in alveolar formation (Figure 4L). Second, we quantified KLF4 and SFTPC double-positive cells in lungs at P7, P14, and P28 after HYX. Similar to the scRNA-seq analysis, we found a significantly lower number of KLF4⁺ cells overall, and KLF4⁺ AT2 cells in particular (Figure 4M-O). Third, we confirmed expression of KLF4 in primary murine AT2 cells using IF staining (Figure 4P). Moreover, we exposed cultured primary murine AT2 cells to HYX. Consistent with the previous findings, HYX diminished Klf4 expression significantly in HYX-exposed AT2 cells when compared to normoxic conditions (Figure 4P).

In summary, our data demonstrate a marked upregulation of the transcription factor Klf4 in lungs, predominantly AT2 cells, during normal alveolarization. In contrast, acute injury such as HYX causes a significant loss of Klf4 in AT2 cells with an effect on cell-autonomous function on AT2 cell homeostasis in HYX-induced lung injury.

The interactome of Klf4 in AT2 cells regulates key pathways of cell survival and differentiation

Having shown the regulatory function of Klf4 in AT2 homeostasis, we next elucidated the interactome of Klf4 in lung epithelial cells. To this end, we pursued an integrative approach in which we used data from a proximity-dependent biotin identification assay (BioID) and RNA-seq data (Figure 5A): We overexpressed Klf4–biotin ligase fusion protein in MLE12 cells, followed by BioID with a biotin-based protein-proximity labeling in living cells. Two hundred eleven candidate interactors were found comparing Klf4 (BirA*-Klf4) with the control (BirA*-Egr-1, and BirA*-biotin ligase) (Figure 5B, C). Among these putative Klf4 interacting proteins, 26 transcription factors were identified. The heatmap in Figure 5D depicts the number of DEGs identified by RNA-seq of Klf4 overexpressing MLE12 that are predicted to be regulated by KLF4 interacting transcription factors. Subsequent analysis of those DEGs using the Kyoto Encyclopedia of Genes and Genomes database revealed significantly regulated pathways involved in the maintenance of stemness, cell survival, and differentiation (eg, TGF-β signaling pathway, PI3K-Akt signaling pathway, and Wnt signaling), which are listed in Figure 5E. A complete list of regulated pathways is provided in Table S3.

Figure 5.

Regulation of proliferation and stemness pathways as well as of markers of alveolar epithelial type 2 (AT2) cells in lung epithelial cells by Krüppel-like factor 4 (Klf4). (A) schematic of the experimental setup, elucidating the Klf4 protein interactome in MLE12 cells using proximity-dependent biotin identification assay (BioID) and RNA-seq data. (B-E) The functional role of Klf4 was analyzed by mass spectrometry–driven proteomics in MLE12 cells (n = 3 independent samples per group). (B) Heatmap shows proteins identified with BirA*-Klf4, BirA*-Egr-1, and BirA*-biotin ligase in MLE12 in FPKM (fragments per kilobase of exon model per million reads mapped). (C) Dot plot depicting 211 candidate interactors (red) using a log fold change (FC) difference <2 and a q-value <0.05 comparing BirA*-Klf4 with BirA*-biotin ligase in MLE12 cells. (D, E) Twenty-six candidate interacting transcription factors were identified and analyzed using the software StringDB with an interaction score >0.4. Heatmap shows differentially expressed genes (DEGs) identified by RNA-seq that are predicted to be regulated by Klf4 interacting transcription factors (D). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis for the DEGs identified by RNA-seq using the predicted network of the 26 Klf4 interacting transcription factors (E). (F-I) Transcriptomic profile of MLE12 with a CRISPR/Cas-induced deletion of Klf4 (Klf4^del) by 3′ RNA-seq (n = 3 independent samples per group). Volcano plot showing DEGs in the transcriptome of Klf4^del. Genes with average log-FC >0.5 and false discovery rate (FDR)–corrected P value <.05 were considered as DEGs. Negative log₁₀ adjusted P values are plotted against the log₂-FC. Selected genes are labeled. P values are calculated using Wald test and P-adjusted values using the FDR/Benjamini–Hochberg approach (F). Heatmap indicating DEG patterns for alveolar epithelial type 1 (AT1) cells (eg, Igfbp2, Rtkn2), or for AT2 cells such as SFTPC (G). Assessment of gene expression using qRT-PCR: Akap5, Hck, Igfbp2, Rtkn2, Sftpb, and Sftpc. 18s rRNA served as reference gene; Mann–Whitney t-test. Data are presented as mean ± SEM. *P < .05 (H). Pathway analyses were performed using KEGG analysis. Displayed are the top 10 pathways regulated by DEGs in Klf4^del (I). “Created in BioRender. Alejandre Alcázar, M. (2026) https://BioRender.com/wk6udld”.

To further substantiate that a loss of Klf4 directs alveolar epithelial cell survival and phenotype, MLE-12 with deletion of Klf4 (Klf4^del) were compared to their respective controls. Loss of Klf4 in MLEs dysregulated 1368 genes compared to control cells. The volcano plot and the heatmap confirmed down- and upregulation of genes characteristic for AT2 and AT1 cells, respectively, supporting an AT1-favoring transcriptomic profile after loss of Klf4 (Figure 5F, G). In addition, assessment of gene expression using qRT-PCR confirmed a significant upregulation of markers of AT1 cells (eg, Igfbp2, Rtkn2), whereas Sfptb was rather downregulated in MLE12 cells with deletion of Klf4 (Figure 5H). On the contrary, we did not detect marked changes in the expression of inflammatory cytokines (eg, Il6 and Tnfa) or Unc5b (Figure S5). Moreover, pathway analysis revealed enrichment of pathways important in cell survival and differentiation, eg, regulation of MAPK cascade (Figure 5I). When interpreting the data, it needs to be taken into account that MLE12 cells are immortalized cells and not primary murine AT2 cells.

Netrin-1 and Klf4 are decreased in lungs of infants with BPD

Finally, we analyzed postnatal human BPD lungs and age-matched control lungs from infants (Table S4). First, we performed IF co-staining for CDH1 (epithelial marker) and KLF4 protein in human tissue samples from non-BPD (control) and BPD lungs representing different matching time points from 3.55 to 68.21 months (representative images of 3 age-matched pairs are shown in Figure 6A). Quantification demonstrated that the absolute number of KLF4⁺ cells and the number of CDH1⁺KLF4⁺ cells per field of view were significantly decreased in lungs of BPD patients when compared to age-matched controls (Figure 6A, B). Second, we performed co-IF staining of KLF4 and SFTPC, followed by quantification of cells positive for both KLF4 and SFTPC. The analysis revealed a reduction of KLF4⁺ AT2 cells in lungs of infants with BPD when compared to non-BPD (Figure 6C, D).

Figure 6.

Reduced abundance of Krüppel-like factor 4 (Klf4) in alveolar epithelium and alveolar epithelial type 2 (AT2) cells of infants with and without bronchopulmonary dysplasia (BPD). (A, B) Representative images of co-immunofluorescence (IF) staining for CDH1 (marker of epithelial cells, green) and KLF4 (red) in age-matched BPD and control (Ctrl) lungs (pairs 1-3); white arrowheads indicate KLF4⁺CDH1⁺ cells (A). The analysis of KLF4⁺ as well as KLF4⁺CDH1⁺ cells in up to 9 fields of view is shown below the images (B). DAPI was used for nuclear staining. (C, D) Representative images for co-IF staining for surfactant protein C (SFTPC; marker of AT2, green) and KLF4 (red) in age-matched BPD and Ctrl lungs; white arrowheads indicate KLF4⁺SFPTC⁺ cells (C). The analysis of KLF4 and SFTPC double-positive cells and their percentage of all SFTPC⁺ cells are shown below the images (D); DAPI was used for nuclear staining. Field of view corresponds to 0.040 mm². Data are presented as mean ± SEM. *P <.05, **P <.01 (Mann–Whitney U test).

Discussion

The present study identifies a netrin-1–Klf4 axis that is essential for AT2 homeostasis and that is inhibited in neonatal mouse lungs after prolonged HYX and in lungs of infants with BPD. Our data provide evidence that Klf4 in AT2 cells is regulated by netrin-1–Unc5b and directs the function of AT2 cells, and in doing so could be relevant in alveolar growth and repair. This netrin-1–Unc5b–Klf4 mechanism in AT2 cells during HYX-induced alveolar damage is based on the following key findings: First, netrin-1 is markedly upregulated during alveolarization, whereas HYX-induced alveolar injury causes a reduction of NTN1⁺ and UNC5B⁺ AT2 cells. Second, netrin-1 regulates AT2 cell survival, pathways critical in maintenance of AT2 homeostasis as well as processes involved in epithelium development. Third, we describe the transcription factor Klf4 to be involved in netrin-1–regulated pathways critical in AT2 survival. Fourth, Klf4 is markedly upregulated and localized in AT2 cells during normal alveolar development and regulates pathways central in AT2 function. Fifth, the investigation of the interactome of Klf4 and deletion of Klf4 using BioID and transcriptomic profiling, respectively, identify Klf4 as a regulator of pathways that are central in maintenance of progenitor cell function and survival. Finally, the percentage of Klf4⁺ alveolar epithelial and AT2 cells is reduced in clinical and experimental BPD. These data identify a netrin-1–Unc5b–Klf4 axis in AT2 that regulates AT2 homeostasis and is reduced in aberrant lung structural maturation.

AT2 cells are alveolar progenitors cells and important for regeneration after lung injury by self-renewal and differentiation into AT1 cells.⁴³ Clinical and experimental studies have revealed a disruption of AT2 homeostasis in chronic lung pathologies resulting in reduced lung repair and regeneration.^44-47 Similarly, dysfunction and loss of AT2 cells as a consequence of genetic as well as acquired disorders are linked to clinical and experimental acute lung injury (ALI) and acute respiratory distress syndrome.^48-50 Moreover, acute exposure to HYX induces alveolar repair via expansion and differentiation of AT2 cells,^14,51,52 whereas we demonstrate here that prolonged HYX-induced injury causes a loss of AT2. Given the importance of AT2 cells in alveolar formation and repair, the identification of key integrators of AT2 homeostasis is of great importance. Defining signaling pathways and molecular nodes that converge in central developmental pathways in AT2 homeostasis such as Klf4 could offer a new avenues to promote regeneration.

In the present study, we demonstrate netrin-1 as a matrix protein that regulates AT2 function and is markedly reduced in HYX-based BPD. Netrin-1 is a secreted laminin-like protein with pleiotropic functions during embryonic development and in adult tissue homeostasis, and regulates axonal guidance, immune cell migration, and angiogenesis.^53,54 Most importantly, netrin-1 is a key modulator of embryonic stem cell function by protecting against apoptosis¹⁸ and attenuating inflammation in ALI.^17,55 Interestingly, in a model of LPS-induced lung injury, the inflammatory response (eg, increased IL-1b), was associated with reduced concentrations of netrin-1. On the contrary, treatment with netrin-1 mitigated the inflammation by inhibiting the Toll-like receptor 4 (TLR4)–NF-κB p65 axis and suppressing IL-1b expression.⁵⁶ Similarly, we show here that blocking netrin-1 signaling through Unc5b increases Il-1b expression, supporting the notion that reduction of netrin-1 in HYX-based BPD could favor an inflammatory response and contribute to lung growth arrest. Signal transduction of this secreted laminin-like protein is mediated through ligand binding to its main receptors, including Unc5b. Studies have shown that netrin-1 co-regulates Wnt and MAPK pathways through Unc5b in both mouse and human embryonic stem cells as well as in endothelial cells, thereby identifying netrin-1 as an integrator of cell pluripotency and self-renewal.^57,58 Here, we found a marked increase in netrin-1 expression during alveolarization, whereas HYX markedly reduced netrin-1 and Unc5b in total lungs, but also specifically in AT2 cells. The suppression of netrin-1 and Unc5b signaling in AT2 cells was related to regulation of genes that are central in cell proliferation and differentiation (eg, Afap1l1 or Bex1),³⁹ but also of Dlk1, a regulator of AT2 to AT1 differentiation.⁵⁹ In contrast, treatment of AT2 cells with netrin-1 promoted proliferation and AT2 survival. Netrin-1 is important in the regulation and maintenance of stemness, protecting established pluripotent cells from apoptosis. Moreover, Klf4 and Oct4 repress netrin-1 transcription, an essential mechanism required in the stage of generation of inducible pluripotent stem cells,¹⁸ establishing an interplay of netrin-1 signaling and the transcription factor Klf4 in stemness. In the present study, we demonstrate that netrin-1 induced Klf4 in AT2 cells through Unc5b receptor. These findings are in accordance with prior studies indicating the importance of netrin-1 signaling in cell homeostasis,^18,57,58 and indicate that it could serve as a therapeutic target for the promotion of alveolar repair.

Klf4 belongs to a family of zinc finger–containing transcription factors and regulates cell survival, cell differentiation, and stemness.¹⁹ Conventional Klf4 knockout mice die within 16 hours after birth as a result of an impaired skin barrier function.⁶⁰ Klf4 is physiologically upregulated at birth and expressed in neonatal lungs.⁴² Our present data now describe localization of Klf4 in AT2 cells along with an increased expression of Klf4 in neonatal lungs during the discrete window of alveolarization (P7-P21). However, similar to a murine model of ventilation-induced neonatal lung injury,⁶¹ we now demonstrate a marked downregulation of Klf4 in neonatal lungs with HYX-induced injury, and specifically also in AT2 cells. Moreover, the present data indicate an important role of Klf4 in cellular processes related to epithelial cell function and control cell differentiation as well as progenitor cell function (eg, Wnt signaling, DNA binding, cell surface receptor signaling, and axon development). This signaling pathways are essential not only in development, but also in the pathogenesis of CLDs.^{45,  62,63} In addition, deletion of Klf4 in lung epithelial cells identified an AT1-like, but not AT2-like, phenotype, possibly indicating that loss of Klf4 induces AT2 to AT1 transition. However, a AT2-cell specific transgenic reporter mouse is needed to confirm that loss of Klf4 leads to a reduction of AT2 via transition to AT1.

The transcription factor Klf4 directs cell differentiation and function depending on tissue and cell type^19,64,65 and contributes to lung pathology such as arterial remodeling or lung fibrosis.^41,66 However, the functional profile and the interactome of Klf4 in alveolar epithelial cells such as AT2 remain uncertain. Here, a bulk RNA-seq approach along with the integration of proximity ligation assay provide novel insight in the functional network of Klf4 in AT2 cells. The analysis identified the interactome of Klf4, including a transcription factor network that is involved in pathways that regulate the maintenance of progenitor cell function, for example, PI3K-Akt signaling pathway and Wnt signaling.

In the present study, prolonged HYX causes a depletion of AT2 cells in an experimental model for BPD, a process that is linked to a reduced abundance of Klf4. Similarly, in lungs of infants with BPD, we found a progressive loss of AT2 cells²³ along with a loss of Klf4 as a possible regulator of AT2 maintenance. Klf4 in AT2 cells could regulate alveolar repair mechanisms, both via cell-autonomous and non-cell-autonomous routes. Since developmental pathways might offer new treatment strategies to reverse lung pathologies and rebuild the lung structure,^45,67 the targeting of the developmentally significant netrin-1–Unc5b–Klf4 pathway could offer an interesting new perspective for the proregenerative treatment of BPD.

Despite the advancement in understanding the pathogenic mechanisms, BPD remains the most common cause of serious morbidity and mortality in premature infants. To date, the progression of lung disease is often inevitable and effective therapies to promote lung growth, repair, and regeneration are urgently needed. The present study demonstrates that the netrin-1–Unc5b–Klf4 axis could be a central hub in the regulation of alveolar progenitor cell survival (see graphical abstract and Figure 7), offering a possible new avenue to promote lung repair and regeneration.

Figure 7.

Employment of transcriptomic and protein profiling (multi-omics) together with ex vivo precision-cut lung slices (PCLS) and in vitro studies identified a netrin-1–Unc5b–Klf4 in alveolar epithelial type 2 (AT2) cells. Reduced activation of netrin-1–Unc5b–Klf4 in a hyperoxia-based murine model is linked to a depletion of AT2 cells and arrest of neonatal alveolar growth. In clinical and experimental bronchopulmonary dysplasia (BPD), Klf4 was significantly reduced, indicating a possible mechanism in AT2 cell homeostasis in alveolar repair after severe injury. “Created in BioRender. Alejandre Alcazar, M. (2026) https://BioRender.com/wk6udld”.

Author contributions

O.K., J.M., C.K.-M., S.D., I.M., L.K.E., B.S., D.A.A., M.K., and M.A.A.A. conceived and designed research. O.K., J.M., S.D., C.K.-M., S.S., M.A.M., D.R., T.I., S.A., L.K.E., C.V., D.H., M.B., T.P., K.G., I.A., T.G., M.K., and M.A.A.A. performed experiments. O.K., J.M., S.D., I.M., C.K.-M., S.S., M.A.M., D.R., M.D., S.M., D.H., M.B., T.P., C.B., T.G., D.A.A., M.K., and M.A.A.A. analyzed data. O.K., J.M., S.D., I.M., C.K.-M., C.M.A., S.M., D.H., M.B., T.P., M.O., P.K., S.B., G.S.P., S.S.P., W.S., D.A.A., J.D., M.K., and M.A.A.A. interpreted results of experiments. O.K., J.M., S.D., I.M., C.K.-M., S.S., M.D., S.M., D.H., C.B., T.G., and M.A.A.A. prepared figures. O.K., J.M., C.K.-M., C.M.A., D.A.A., M.K., and M.A.A.A. drafted the manuscript. O.K., J.M., C.K.-M., S.D., C.M.A., S.S., S.S.P., W.S., D.A.A., M.K., and M.A.A.A. edited and revised the manuscript. All authors approved the final version of the manuscript.

Supplementary material

Supplementary material is available at American Journal of Respiratory Cell and Molecular Biology online.

Conflicts of interest

Please see the ICMJE disclosure forms, which have been provided as supplementary material. The authors declare no competing interests.

Funding

This work was supported by Deutsche Forschungsgemeinschaft (1636/2-1; 1636/5-1, M.A.A.A.); Marga und Walter Boll Stiftung (210-09.01-18; 210-01.03-21; 210-07.03-25; M.A.A.A.); German Center of Lung Research (project number 82 DZLS85C1, project number 82 DZL005C1); Institute for Lung Health (project number 82 DZL 005C4); Cardiopulmonary Institute (CPI; project number 82 DZL C55B1); Center for Molecular Medicine Cologne (CAP10 [M.A.A.A.] and C01 [M.A.A.A. and M.K.); Deutsche Forschungsgemeinschaft - FOR 2722 (M.K.); National Institutes of Health, National Heart, Lung, and Blood Institute (NIH/NHLBI R01HL141856; D.A.); NIH/NHLBI and the Office of the Director (R01HL155104; S.D.); and NHLBI LungMAP Research Center Biorepository (1U01HL122700, G.P.) and Data Coordinating Center (1U01HL122638, G.P.).

Data availability

The scRNA-seq data used in this study, including fastq sequencing files, gene expression matrices, and cell metadata, were previously published²⁵ and deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus (GEO) database (accession code GSE151974). Code used for the analysis of scRNA-seq data pertinent to this study will be made available at the public GitHub repository and is available by request to the corresponding author. Bulk RNA-seq data used in this study from neonatal mice exposed to prolonged hyperoxia were previously published²³ and deposited in Dryad (doi:10.5061/dryad.rr4xgxd8m).