Infants who develop BPD have an airway endotype defined by vimentin expression and ciliary loss

2 Rationale: Bronchopulmonary Dysplasia (BPD) results from abnormal lung 3 development after preterm birth, with structural deficits at every respiratory tree level. 4 BPD with lower airway disease is emerging as a clinically significant phenotype with 5 increased mortality, and there is a significant knowledge gap in the molecular 6 mechanisms whereby preterm birth disrupts normal airway development. 7

To develop a human model of lower airway disease after preterm birth and 8 to characterize a molecular endotype of evolving BPD (eBPD) at baseline and in 9 response to injury. 10

We used a combination of an ex vivo organotypic Airway Epithelial Cell (AEC 11 models) and well-characterized pathologic and transcriptomic patient samples for 12 quantitative immunohistochemistry and RNA sequencing analyses. 13 Measurements and Main Results: Compared to AECs from healthy patients, eBPD-14 derived AECs have a molecular endotype of reduced proliferation, impaired 15 differentiation to ciliated epithelium, and an expanded vimentin-positive population with 16 a transcriptional shift toward stromal cell-associated genes. With hyperoxia exposure, 17 eBPD-derived AECs exhibited a pronounced vimentin response ex vivo, which parallels 18 the increased vimentin expression of airway cells observed in lung tissue from human 19 infants with BPD. 20

In this organotypic model of neonatal airway differentiation, we find that 21 infants with eBPD have impaired differentiation, increased expression of vimentin, and 22 concomitant loss of cilia, with an exaggerated increase in vimentin expression after 23 hyperoxia injury, findings that mimic the effects of prematurity in airway cells in human 24 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 2 patients. These data provide a foundation for future mechanistic studies interrogating 25 the role of intermediate filaments in epithelial differentiation and repair. 26 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 3

27 Bronchopulmonary Dysplasia (BPD) is a severe chronic lung disease of infants 28 born extremely premature, characterized by aberrant postnatal development and 29 superimposed injury of the respiratory tree (1, 2). Historically, BPD research has 30 focused on the aspects of the disease with arrest of alveologenesis, with the airway 31 pathology being relatively understudied and underrepresented in animal models of BPD. 32 BPD diagnostic criteria stratify the disease by clinical severity, and patients requiring 33 invasive respiratory support at 36 weeks corrected gestational age (CGA) are 34 diagnosed with Severe BPD (sBPD) Type 2 or Grade 3 BPD (2, 3). However, this 35 classification does not fully capture the anatomical heterogeneity of the disease. While 36 arrested alveologenesis is a hallmark pathologic feature of BPD, lower airway disease is 37 also prevalent, with up to 90% of patients with sBPD having some degree of obstructive 38 lung disease on infant pulmonary function testing (4). In addition, many patients have 39 multilevel airway obstruction with gas-trapping and need for prolonged positive pressure 40 support to assist with patent airways during exhalation; indeed, severe airway 41 obstruction remains a relatively understudied aspect of BPD pathology. Here, we seek 42 to tackle the knowledge gap in understanding the molecular and cellular factors 43 associated with airway disease in patients with sBPD (4). 44 The airway epithelium serves many roles in the lung, including mucociliary 45 clearance, barrier functions, and innate immune responses to pathogens. A well-46 differentiated pseudostratified airway epithelium in the airway may have more than 15 47 distinct cell types as identified across multiple existing lung cell atlases, including 48 ciliated, goblet, and secretory cells that arise from (multiple subtypes of) progenitor 49 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 4 basal cells (5-8). BPD and BPD-related airway epithelia have been underrepresented in 50 most published tissue and single-cell analyses and while apical cilia, vital to movement 51 of airway secretions and entrapment of pathogens prior to cough (9), have been shown 52 to be structurally abnormal and dysfunctional in premature infants with evolving BPD 53 (10), the mechanisms whereby abnormalities in airway epithelium relate to clinical 54 respiratory disease in BPD have not been explored. 55 To this end, we used minimally invasive tracheal aspirates to characterize lower 56 airway epithelial cells from patients with evolving BPD under normal and hyperoxia-57 exposed conditions to mimic the exposures of prematurity. We found abnormal structure 58 and function in BPD-derived AECs and a failure of the normal airway differentiation 59 program that associates with vimentin expression under organotypic culture conditions 60 and in single-cell analysis of infant lung tissue. 61 62

63 Detailed methods are provided in the data supplement. 64 65 Airway Epithelial Cell Collection and Isolation 66 AECs were isolated from tracheal aspirates (TAs) collected as part of the routine 67 respiratory care of endotracheally intubated premature infants born = <28 weeks 68 gestation and who were less than 30 days of age in the University of Washington NICU 69 (total N=27). BPD diagnosis and severity was assigned by LCE at 36 weeks CGA after 70 chart review, per 2001 NIH guidelines and Abman et al., 2017 (11). Control AECs were 71 derived from airway brushings collected from N=22 healthy children aged 2-16 years 72 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 5 intubated for elective surgeries at Seattle Children’s Hospital. Both cohorts were 73 obtained after consents per Seattle Children’s Hospital IRB-approved protocols (LCE 74 STUDY00000263 and JSD STUDY#00001596). Parents of subjects provided written 75 consent and children over 7 years of age provided assent. Details about AEC isolation 76 and culture are provided in the supplement. 77 78 Airway Epithelial Cell Organotypic Culture 79 Isolated AECs were differentiated for three weeks at air-liquid interface (ALI) with 80 PneumaCult-ALI media in the basolateral compartment, as previously described (12-81 15). Quality control was performed in both control and BPD AEC cohorts to ensure they 82 produced an organotypic differentiated epithelial cell culture with mucociliary 83 morphology and without leak indicative of epithelial barrier. TEER measurements were 84 also used on occasion to confirm barrier function, with additional details in the 85 supplement. 86 87 Hyperoxia-induced Epithelial injury of Organotypic AEC Cultures 88 Fully differentiated organotypic AEC cultures were exposed to 0.85 FiO2 for 24, 89 48, 72, or 96 hours in a hyperoxia chamber placed in the incubator (BioSpherix, Parish, 90 NY), with 0.21 FiO2 exposed cultures as controls. Hyperoxia-induced epithelial injury 91 was evaluated by visual inspection of cell morphology, immunostaining, and RNA 92 sequencing. Increased expression of GPX2 (16) and NQO1 (17, 18) in the FiO2 0.85-93 exposed group was used to validate hyperoxia exposure through known responses to 94 oxidative stress. 95 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 6 RNA Sequencing and Analysis 96 Bulk RNA sequencing (RNA-seq) was performed on organotypic AEC cultures, 97 with support from the Benaroya Research Institute Genomics and Bioinformatics Cores, 98 with additional details in the supplement. RNA-seq datasets presented in this article will 99 be submitted to the National Center for Biotechnology Information Gene Expression 100 Omnibus by the time of publication. 101 102 Immunostaining and Confocal Imaging 103 Proliferative or fully differentiated AEC cultures were washed with PBS, fixed 104 with 10% formalin for one hour at room temperature, and labeled by wholemount 105 immunofluorescence staining prior to quantitative confocal imaging, with additional 106 details in the supplement. 107 108 Analysis of Human Pathological Specimens 109 Pathological specimens from premature infants born at less than 26 weeks 110 gestation who died from pulmonary/evolving BPD (N=4) or non-pulmonary (N=4) 111 causes, were analyzed (IRB-approved protocol (GHD, Study00003664; Figure 7A). 112 Those patients in the group noted to have non-pulmonary deaths were supported with 113 oxygen and/or noninvasive respiratory support at birth/ DOL1 with milder lung disease, 114 whereas those pulmonary causes of death required invasive respiratory support from 115 birth/DOL1 on and had severe lung disease on pathological analysis. Formalin-fixed 116 paraffin embedded 5-µm serial sections of trachea taken at the level of the thyroid were 117 stained after antigen retrieval with anti-tubulin (Sigma, 1:10,000), anti- pancytokeratin 118 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 7 (Dako, 1:100), anti-EPCAM (Invitrogen, 1:800), anti-vimentin (Biogenic, 1:100), and 119 anti-TP63 (Biocare, 1:200). Staining was quantified in a blinded manner in three 120 randomly selected areas of tracheal epithelium at 20X magnification and captured with 121 a digital camera mounted on a Nikon Eclipse 80i microscope using NIS-Elements 122 Advanced Research Software v4.60. Care was taken to avoid the denuded areas near 123 the endotracheal tube. 124 125 Data Analysis 126 Analysis of quantitative immunohistochemistry and total cell numbers was 127 conducted with GraphPad Prism 8 (GraphPad Software Inc, La Jolla CA). Mann-128 Whitney or Student’s t-test were performed when appropriate to compare two treatment 129 groups, depending on distribution of the data. Asterisks indicate statistical significance, 130 *p<0.05, **p < 0.005, ***p < 0.0005 and **** p<0.0001. 131 132 Single-cell RNA Sequencing Analysis 133 We interrogated a recently released single cell RNA-seq (scRNA-seq) atlas of 134 human infant lung injury (8) from infants who died with BPD/BPD+pulmonary 135 hypertension (N=4) and term infant controls (N=2) to look for differentially expressed 136 genes in AECs. Full code used in the analysis of these data is available at 137 https://github.com/SucreLab/HumanNeonatal. A table with the clinical characteristics of 138 infants whose lungs were sequenced is in the supplement for the manuscript. In 139 addition, raw data are available on NCBI GEO database (accession # pending) and raw 140 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 8 and processed data are available at the LungMAP Consortium and as an interactive 141 data explorer at www.sucrelab.org/lungcells. 142 143 144

145 Identification of Deficient AEC Growth and Differentiation in Evolving BPD 146 We established an ex vivo system from tracheal aspirate-derived AECs (Figure 147 1A) to examine AEC proliferation and differentiation in evolving BPD (eBPD). We 148 defined eBPD as infants born premature (<28 weeks), less than 30d old, who were later 149 diagnosed with BPD per the standard definitions. Despite identical seeding densities, 150 eBPD AECs grew more slowly than healthy control AECs, and most eBPD AECs (7/12, 151 58%) had features of failed differentiation including lack of apical 152 polarization/topography, evidence of impaired barrier function with apical leak, and 153 altered cell morphology. By comparison, only 1/6 control AEC organotypic cultures had 154 any features of poor differentiation. eBPD AECs stained for apical cilia and actin 155 displayed disorganized morphology with stretched actin network, decreased ALI 156 thickness, and scant apical cilia relative to controls (Figure 1B). 157 Bulk RNA-seq of control and eBPD AECs was performed in both proliferative 158 (day 7 of submerged condition) and differentiated cultures. There were 5235 159 differentially expressed genes (DEG) between control proliferative and differentiated 160 cultures, 979 DEG between eBPD proliferative and differentiated cultures, and 1842 161 DEGs between control and eBPD differentiated cultures (Figure 1C, Table 2). In Gene 162 Enrichment analyses, the 500 most significantly upregulated genes in eBPD samples 163 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 9 were associated with extracellular matrix (ECM) gene expression and epithelial 164 mesenchymal transition (EMT), as detailed in Figure 1C. Expression of traditional 165 hallmark genes of epithelial-mesenchymal transition was increased, including CDH2 (9-166 fold) and VIM (6-fold) in eBPD cultures in comparison with controls. 167 We next attempted to “rescue” eBPD cultures by using a more supportive 168 medium that was becoming more widely used as these experiments evolved (see 169 methods). With this new medium, the failure rate of differentiation in eBPD cultures 170 decreased to 36% but still was greater than controls (<10%). Despite healthier 171 organotypic AEC cultures with increased proportion of ciliated cells in both groups 172 cultured in this medium, eBPD cultures had persistently decreased ALI thickness and 173 apical cilia compared with control cultures (Figure E1). 174 We next compared the growth potentials of eBPD and control AECs by 175 quantifying P63+ basal cells in proliferative cultures and total cell numbers in fully 176 differentiated cultures. Proliferative AECs from eBPD samples had fewer P63+ basal 177 cells at day one, three, and seven days after plating on transwells - and fewer cells 178 overall in fully differentiated cultures - despite consistent plating density (Figure E2). 179 Bulk RNA-seq comparing gene expression in control and eBPD AECs in proliferative 180 (Supplementary Tables E1 and E2) and fully differentiated ALI cultures (Supplementary 181 Table E3) was performed. TP63 and Ki67 expression were decreased in eBPD AECs 182 from all timepoints, compared with controls. 183 Taken together, these data demonstrate that AECs from premature infants with 184 eBPD do not differentiate normally and upregulate ECM gene expression modules that 185 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 10 may be associated with extracellular remodeling and/or other changes in AEC function 186 specific to BPD pathophysiology. 187 188 Identification of Vimentin-expressing AECs in eBPD 189 Building on the finding of increased EMT-associated gene expression, we 190 interrogated the mesenchymal transcriptional signature of cells in AEC cultures from 191 patients with eBPD and healthy controls and found that AECs from patients with eBPD 192 had increased expression of vimentin by quantitative immunostaining and bulk RNA-seq 193 (Figure 2A). Vimentin-positive cells were present in similar numbers of control and 194 eBPD AECs after 3 days of proliferation. In contrast, there was an increase in the 195 number of vimentin-positive AECs in eBPD AECs after 7 days in proliferative conditions 196 with a further increase in fully differentiated AECs, compared with controls (Figure 2B). 197 Bulk RNA-seq in well-differentiated control (N=5) and eBPD (N=5) cultures 198 demonstrated 22-fold greater vimentin expression in eBPD cultures (Figure 2C). We 199 used several methods to distinguish poor AEC differentiation in eBPD associated with 200 vimentin expression from stromal contamination of the seeding population of cells 201 (Figure E3). 202 203 Additive Increases in Vimentin Expression after Hyperoxic AEC Injury 204 We employed hyperoxia exposure to model the airway disease associated with 205 BPD by treating fully differentiated AECs with 0.85 FiO2 or normoxia for 24, 48, 72, or 206 96 hours and investigating structural and molecular changes after hyperoxic injury. We 207 found a dose-dependent response in loss of apical cilia, with increased ciliary loss after 208 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 11 subsequent days of hyperoxia in control cultures (Figure 3B). At baseline (normoxia), 209 there was an increased number of VIM+ AECs in eBPD cultures compared with controls 210 (Figure 4A) that was consistent with mRNA expression differences (Figure 2C). Each 211 day of hyperoxia exposure further increased the number of VIM+ AECs in both control 212 and eBPD groups (Figure 4B). 213 214 Increased Vimentin Expression in BPD AECs at the Single Cell Level 215 We next explored cell-specific vimentin expression in BPD using a publicly 216 available scRNA-seq atlas (8). The authors sequenced 43,607 single cells, and 217 clustering analyses with cell-type annotation identified seven epithelial clusters from 218 N=2 term controls and N=4 BPD samples isolated from human lung tissue after autopsy 219 (Figure 5A). Five distal lung AEC populations were identified, including basal, 220 multiciliated, RASC, Secretory SCGB3A1/SCGB3A2, and Secretory MUC5B cells 221 (Figure 5B). Vimentin was localized to RASC and Secretory SCGB3A1/SCGB3A2 222 AECs, and vimentin expression was increased in these subpopulations in BPD samples 223 compared with term controls (Figure 5C). In the 763 non-ciliated respiratory epithelial 224 cells, but not other airway epithelial subtypes, vimentin expression was increased in the 225 BPD samples, compared with unchanged expression of EPCAM between control and 226 BPD cells (Figure 5D), supporting the possibility that increased vimentin expression is 227 linked to airway injury and maladaptive repair after preterm birth. Interestingly, vimentin 228 expression in BPD mesenchymal cells is also slightly increased compared with control 229 mesenchyme (data not shown). 230 231 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 12 Aberrant AEC Differentiation and Increased Vimentin Expression in Pathologic 232 Samples from Patients with eBPD 233 We sought to confirm our findings by investigating tracheal epithelial cells in 234 pathologic samples from premature infants with and without eBPD. Controls were born 235 extremely premature, did not require intubation on DOL1, had lower oxygen 236 requirements in the 48-72h prior to demise, and had minimal lung disease on pathology 237 (blinded analysis by GHD, Figure 6). Patients with eBPD were born extremely 238 premature, required invasive mechanical ventilation on DOL1, had high oxygen 239 requirements prior to demise, and had confirmed lung pathology including patchy 240 lobular hyperinflation, airway mucostasis and hypertensive changes of the pulmonary 241 arteries. Concordant with airway mucostasis in the evolving BPD cases, tracheal 242 epithelia had severely reduced area of apical cilia compared with age-matched 243 premature infants (***p<0.0005, Figure 6B). Additionally, we confirmed increased 244 vimentin expression in tracheal AECs from patients with evolving BPD (*p<0.05). 245 Tracheal AECs from patients with evolving BPD also had a trend toward decreased 246 EPCAM (p=0.15) and increased basal cell TP63 (*p<0.05) expression. 247 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 13

248 In a combination of orthogonal techniques with primary neonatal human data, we 249 demonstrate that AECs from premature infants who develop severe BPD exhibit 250 deficient growth and differentiation when compared with AECs from pediatric healthy 251 controls. As sampling the lower airway in premature infants is inherently challenging, we 252 developed an ex vivo organotypic AEC model and demonstrated that AECs from 253 patients with evolving sBPD fail to undergo the normal differentiation program to the 254 polarized, pseudostratified respiratory epithelium observed in AECs from healthy 255 pediatric donors. Further, these sBPD AEC cultures expressed more vimentin RNA and 256 protein. When modeling the hyperoxia-induced injury associated with preterm birth by 257 exposure of a subset of ex vivo cultures to 0.85 FiO2, we observed an increase in non-258 ciliated AEC-specific expression of vimentin in both control and sBPD AECs, a finding 259 we confirmed in postmortem infant airways and a recently generated single cell atlas of 260 neonatal lung injury. Together, these data highlight a new role for vimentin in evolving 261 airway injury in preterm infants that correlates with response to hyperoxia and risk of 262 BPD severity. 263 There is a growing appreciation for the different anatomic manifestations of BPD, 264 and many survivors of prematurity have life-long lower airway disease (19-23). Small 265 airway obstruction in these patients is multifactorial, a convergence of congenitally 266 narrowed airways, dysanaptic growth, and repeated airway injury and repair. How the 267 airways of patients born prematurely grow and remodel in the neonatal period is not well 268 understood due to an absence of airway function data in during the neonatal period. 269 Moreover, there are not good predictors for which infants with evolving BPD will develop 270 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 14 lower airway obstruction and hyperinflation. Modeling airway epithelial differentiation 271 using human samples from well characterized donors with and without BPD has the 272 potential to greatly advance understanding of mechanisms driving AEC growth, 273 differentiation, repair, and function in BPD and identify potential novel therapeutic 274 targets. In contrast to recent reports deriving airway basal stem cells from human 275 pluripotent stem cells (iPSCs)(24) and basal stem cells from tracheal aspirates using 276 EpCAM+ selection and/or mTOR inhibition (25, 26), our organotypic AEC model derives 277 cells from unmodified tracheal aspirates. A strength of this approach is the preservation 278 of the cellular diversity in the airway niche, replicating more of the cellular and molecular 279 deficits that accompany eBPD. In our ex vivo model of AEC growth and differentiation, 280 we identified a primary differentiation defect in AECs from patients with eBPD. Also, 281 organotypic AEC cultures from patients with evolving BPD demonstrated distinct 282 transcriptomic signatures with increased expression of matrix associated genes, a 283 finding replicated in tissue and single-cell transcriptomics from human patients. In 284 addition to baseline differences, we observed marked changes in the hyperoxia 285 response, with decreased apical cilia in AECs from BPD patients. This replicates a 286 known marked loss of AEC apical cilia and mucociliary clearance in BPD (10), an 287 underappreciated driver of airway disease in this population that is not faithfully 288 mimicked in most animal models (27, 28). Data from our organotypic cultures suggest 289 that the loss of ciliated epithelia dovetails with an apparent expansion of vimentin-290 expressing AECs; future work to discover how this reorganization of AEC subtypes 291 affects airway structure and function and the lineage and spatial relationships of 292 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 15 epithelial cells in the premature airway will be crucial to a mechanistic understanding of 293 airway development and BPD. 294 Vimentin is an intermediate filament protein with a known role promoting alveolar 295 wound repair in adult lungs and driving cellular migration in the wound scratch assay 296 (29). From our data, we conclude that the AEC shift to vimentin-positive cells is an 297 indicator of AEC plasticity, but whether it is both a driver or merely an indicator of injury 298 remains undetermined. Vimentin is likely helpful and necessary for the remodeling and 299 repair of injured lower airway epithelia, though potentially at the expense of mucociliary 300 clearance. In alveolar epithelial repair, there is growing appreciation of alveolar type 2 301 cell expression of intermediate filament keratin-8 as both a marker and mediator of 302 fibrosis after injury, with genetic knockout of keratin-8 preventing fibrosis in mouse 303 models (30). Whether these changes are transient or sustained is an important 304 question, especially in light of long-term airway obstruction in survivors of BPD (19-23). 305 With larger patient cohorts, and a longitudinal analysis of cultures from patients with 306 evolving and progressive BPD, we plan to determine the trajectory of airway epithelial 307 changes – including vimentin and other mediators- in the continuum of injury to repair. 308 There are several limitations to this study. AECs from healthy children and 309 patients with evolving BPD have different sources (bronchial brushings versus tracheal 310 aspirates), but airway brushings were not feasible in the first 30d of life. Also, the 311 number of AECs in tracheal aspirates was smaller and variable (200-1000), and 312 therefore these cells likely underwent more doublings in establishing ALI cultures. To 313 mitigate this, we prioritized complementary and concurrent analyses with pathological 314 samples from patients with evolving BPD. In addition, as being intubated is required for 315 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 16 obtaining a tracheal aspirate, our analyses include a subset of patients with more 316 severe early disease. Despite these limitations, the strengths of this study include 317 careful phenotyping and rich clinical metadata associated with evolving BPD patients 318 and the development of an organotypic ex vivo culture system that replicates features of 319 airway disease in preterm infants. 320 321 In summary, these results support an early, aberrant airway epithelial cell 322 differentiation deficit that associates with progression to BPD. We propose that this 323 “vimentin-high” airway molecular endotype of BPD may be a driver of the physiological 324 consequences of lower airway obstruction. Leveraging this platform to test for AEC 325 plasticity, maladaptive repair, and extracellular matrix expression may allow for early 326 identification of patients at highest risk for lower airway disease, as well as provide a 327 quantifiable system for screening and testing needed targeted therapies. 328 329 330 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 17 Table 1 331 Demographics eBPD Healthy Controls N 27.0 22 sex (% male) 66.7 45.5 age (days for BPD, years for control) 13.9 +/- 8.9 10.3 +/- 4.7 gestational age (weeks) 24.9 +/-1.9 term race- white (%) 70.4 63.6 race- Black (%) 11.1 9.1 race- asian 11.1 13.6 race- unknown/not reported or other (%) 7.4 13.6 ethnicity- Hispanic or Latino/a or Latinx (%) 29.6 18.2 ethnicity- Non-Hispanic or Latino/a or Latinx (%) 63.0 81.8 ethnicity- unknown/not reported (%) 7.4 0.0 BW (g) 716.7 +/- 215.4 n/a maternal chorioamnionitis (%) 40.7 n/a maternal chorioamnionitis unknown (%) 7.4 n/a antenatal steroids (%) 74.1 n/a postnatal steroids (%) 96.3 n/a Clinical Data eBPD Healthy Controls FiO2 0.36 +/- 0.11 n/a PEEP 7.9 +/- 1.8 n/a jet ventilator (%) 63.0 n/a HFOV (%) 7.4 n/a ACVG (%) 25.9 n/a SIMV (%) 3.7 n/a Mild/Moderate BPD (%) 35.3 n/a Severe I (%) 40.7 n/a Severe Type 2 BPD or death (%) 37.0 n/a 332 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 18 Figure Legends 333 Figure 1. Establishment and RNA Sequencing of Organotypic Airway Epithelial 334 Cell Cultures from eBPD Tracheal Aspirates. (A) Organotypic AEC cultures were 335 generated from unmodified tracheal aspirates isolated from intubated premature infants, 336 grown in submerged culture until confluent, and then differentiated at air-liquid interface 337 (ALI). (B) 3D morphological assessment of AEC differentiation, which includes 338 quantitative immunostaining of apical cilia, polarization, and ALI thickness. Decreased 339 ALI thickness, % polarized ciliated epithelium, and correlation between these two 340 measures, in demonstrated in organotypic cultures from eBPD patients (N=6), 341 compared with controls (N=6). (C, left) Heat map with transcriptomes of healthy 342 pediatric controls and eBPD patients in submerged and ALI conditions. (C, right) Gene 343 Enrichment analysis of top 500 genes upregulated in eBPD and downregulated in 344 control AECs after differentiation at ALI, demonstrating differential expression of genes 345 controlling extracellular matrix across multiple databases with an emphasis on 346 epithelial-mesenchymal transition in Hallmark Gene Analysis. 347 (https://maayanlab.cloud/Enrichr). (Green = anti-TubA4A, Red = phalloidin, Blue = 348 DAPI. ****p< 0.0001, Mann-Whitney test, magnification bar = 100 µm) 349 350 Figure 2. Increased Vimentin-expressing AECs from Patients with eBPD. (A) 3D 351 projections of vimentin immunohistochemistry on organotypic AEC cultures from eBPD 352 and control patients. (B) Quantification of VIM+ cells during proliferation phases (Day 3 353 and Day 7 submerged cultures) and at ALI after 3 weeks of differentiation, with 354 increased VIM+ cells in eBPD samples at day 7 and at ALI. One dot = one image, with 355 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 19 N=2-3 images per condition, from total of N=5 control donors at day 3, N=4 eBPD at day 356 3, N=3 controls at day 7, N=6 eBPD at day 7. (C) Bulk RNA Sequencing on N=5 Control 357 and N=5 eBPD differentiated cultures, showing 22-fold upregulation of vimentin 358 expression in eBPD AECs. (Green = anti-vimentin, Red = anti-EPCAM, Blue = DAPI. 359 *p<0.05, ****p< 0.0001, Mann-Whitney test, magnification bar= 100 µm) 360 361 Figure 3. Hyperoxia-induced AEC Injury Ex Vivo With Loss of Cilia. (A) 362 Immunostaining of organotypic AECs from healthy controls in room air (RA) or after 363 exposure to hyperoxia for 24, 48, 72, or 96 hours. Increased loss of apical cilia (green) 364 with additional days of exposure to 0.85 FiO2. (B) Hyperoxia-induced dose-dependent 365 ciliary loss, with % of ciliated culture quantified in N=2-3 images per condition in N=4-8 366 donors per condition as specified. (Green = anti-TubA4A, Red = phalloidin, Blue = 367 DAPI. ****p<0.0001, Mann-Whitney test, magnification bar = 100µm) 368 369 Figure 4. Increased VIM-expressing AECs in BPD at Baseline and After Hyperoxia-370 induced Epithelial Injury Ex Vivo. (A) Immunostaining of organotypic AEC showing 371 increased vimentin expression (green) in BPD cultures at baseline (room air, 0.21 FiO2) 372 and after 96h of exposure to 0.85 FiO2. (B) Quantification of VIM+ cells in room air or 373 hyperoxia-exposed conditions, with comparisons of control and eBPD organotypic 374 cultures as indicated. Green line indicates baseline (room air) differences of VIM+ cells 375 between control and eBPD AECs. One dot = one image with N=2-3 images per N=3-8 376 donors per condition. (Green = anti-vimentin, Red = anti-EPCAM, Blue = DAPI. 377 **p<0.005, ****p<0.0001, Mann-Whitney test, magnification bar = 100µm) 378 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 20 379 Figure 5. Increased Vimentin Expression in Non-ciliated, Secretory AECs Isolated 380 from Human Lung tissue from Patients with Established BPD. (A) Single cell RNA 381 Sequencing library created from lungs of patients with BPD (N=4) compared with term 382 controls (N=2). (B) Five distal lung AEC populations were identified by cell-specific 383 markers, including basal, multiciliated, RASC, and Secretory SCGB3A1/SCGB3A2, and 384 Secretory MUC5B cells. (C) Localization of vimentin expression to RASC and Secretory 385 SCGB3A1/SCGB3A2 AECs in BPD samples. (D) Vimentin, but not EPCAM, expression 386 was increased in 763 non-ciliated airway epithelial cells isolated from pathological 387 specimens from patients with BPD compared with term controls. (***p<0.0005 Mann-388 Whitney test). 389 390 Figure 6. Correlating AEC Differentiation Defects Ex vivo with Complementary 391 Analyses in Pathologic Samples from Patients with eBPD. (A) Clinical data from 392 premature control (N=4) and eBPD (N=4) patients whose autopsies were reviewed. 393 Control patients were born premature but supported with non-invasive respiratory 394 support at birth and had milder lung injury on pathological analyses. eBPD patients 395 required invasive respiratory support at birth with higher FiO2 and had severe lung 396 injury on pathology. Tracheal AECs from premature infants (GA 23-26 weeks) who died 397 from eBPD have decreased apical cilia (stained with anti-TUBA4+), increased VIM+ 398 cells, and decreased expression of the cell adhesion marker EPCAM, compared with 399 premature controls who died from non-pulmonary causes. (Values are mean +/- SD, 400 *p<0.05, Mann-Whitney test, magnification bar = 20µm). 401 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 21 402 403 404 405 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 22

1. Thebaud B, Goss KN, Laughon M, Whitsett JA, Abman SH, Steinhorn RH, Aschner 406 JL, Davis PG, McGrath-Morrow SA, Soll RF, Jobe AH. Bronchopulmonary 407 dysplasia. Nat Rev Dis Primers 2019; 5: 78. 408 2. Stoll BJ, Hansen NI, Bell EF, Walsh MC, Carlo WA, Shankaran S, Laptook AR, 409 Sanchez PJ, Van Meurs KP, Wyckoff M, Das A, Hale EC, Ball MB, Newman NS, 410 Schibler K, Poindexter BB, Kennedy KA, Cotten CM, Watterberg KL, D'Angio CT, 411 DeMauro SB, Truog WE, Devaskar U, Higgins RD, Eunice Kennedy Shriver 412 National Institute of Child H, Human Development Neonatal Research N. Trends 413 in Care Practices, Morbidity, and Mortality of Extremely Preterm Neonates, 1993-414 2012. JAMA 2015; 314: 1039-1051. 415 3. Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit Care Med 416 2001; 163: 1723-1729. 417 4. Shepherd EG, Clouse BJ, Hasenstab KA, Sitaram S, Malleske DT, Nelin LD, 418 Jadcherla SR. Infant Pulmonary Function Testing and Phenotypes in Severe 419 Bronchopulmonary Dysplasia. Pediatrics 2018; 141. 420 5. Whitsett JA. Airway Epithelial Differentiation and Mucociliary Clearance. Ann Am 421 Thorac Soc 2018; 15: S143-S148. 422 6. Travaglini KJ, Nabhan AN, Penland L, Sinha R, Gillich A, Sit RV, Chang S, Conley 423 SD, Mori Y, Seita J, Berry GJ, Shrager JB, Metzger RJ, Kuo CS, Neff N, 424 Weissman IL, Quake SR, Krasnow MA. A molecular cell atlas of the human lung 425 from single-cell RNA sequencing. Nature 2020; 587: 619-625. 426 7. Sikkema L, Ramírez-Suástegui C, Strobl DC, Gillett TE, Zappia L, Madissoon E, 427 Markov NS, Zaragosi LE, Ji Y, Ansari M, Arguel MJ, Apperloo L, Banchero M, 428 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 23 Bécavin C, Berg M, Chichelnitskiy E, Chung MI, Collin A, Gay ACA, Gote-429 Schniering J, Hooshiar Kashani B, Inecik K, Jain M, Kapellos TS, Kole TM, Leroy 430 S, Mayr CH, Oliver AJ, von Papen M, Peter L, Taylor CJ, Walzthoeni T, Xu C, 431 Bui LT, De Donno C, Dony L, Faiz A, Guo M, Gutierrez AJ, Heumos L, Huang N, 432 Ibarra IL, Jackson ND, Kadur Lakshminarasimha Murthy P, Lotfollahi M, Tabib T, 433 Talavera-López C, Travaglini KJ, Wilbrey-Clark A, Worlock KB, Yoshida M, van 434 den Berge M, Bossé Y, Desai TJ, Eickelberg O, Kaminski N, Krasnow MA, 435 Lafyatis R, Nikolic MZ, Powell JE, Rajagopal J, Rojas M, Rozenblatt-Rosen O, 436 Seibold MA, Sheppard D, Shepherd DP, Sin DD, Timens W, Tsankov AM, 437 Whitsett J, Xu Y, Banovich NE, Barbry P, Duong TE, Falk CS, Meyer KB, Kropski 438 JA, Pe'er D, Schiller HB, Tata PR, Schultze JL, Teichmann SA, Misharin AV, 439 Nawijn MC, Luecken MD, Theis FJ. An integrated cell atlas of the lung in health 440 and disease. Nat Med 2023; 29: 1563-1577. 441 8. Shirazi SP, Negretti NM, Jetter CS, Sharkey AL, Garg S, Kapp ME, Wilkins D, Fortier 442 G, Mallapragada S, Banovich NE, Eldredge LC, Deutsch GH, Wright CVE, Frank 443 DB, Kropski JA, Sucre JMS. Bronchopulmonary dysplasia with pulmonary 444 hypertension associates with semaphorin signaling loss and functionally 445 decreased FOXF1 expression. Nat Commun 2025; 16: 5004. 446 9. Legendre M, Zaragosi LE, Mitchison HM. Motile cilia and airway disease. Semin Cell 447 Dev Biol 2021; 110: 19-33. 448 10. Lee RMKW, Rossman CM, O'Brodovich H, Forrest JB, Newhouse MT. Ciliary 449 Defects Associated with the Development of Bronchopulmonary Dysplasia. 450 American Review of Respiratory Disease 1984; 129: 190-193. 451 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 24 11. Wenger TL, Hopper RA, Rosen A, Tully HM, Cunningham ML, Lee A. A genotype-452 specific surgical approach for patients with Pfeiffer syndrome due to W290C 453 pathogenic variant in FGFR2 is associated with improved developmental 454 outcomes and reduced mortality. Genet Med 2019; 21: 471-476. 455 12. Lopez-Guisa JM, Powers C, File D, Cochrane E, Jimenez N, Debley JS. Airway 456 epithelial cells from asthmatic children differentially express proremodeling 457 factors. J Allergy Clin Immunol 2012; 129: 990-997 e996. 458 13. Reeves SR, Kolstad T, Lien TY, Elliott M, Ziegler SF, Wight TN, Debley JS. 459 Asthmatic airway epithelial cells differentially regulate fibroblast expression of 460 extracellular matrix components. J Allergy Clin Immunol 2014; 134: 663-670 461 e661. 462 14. Altman MC, Reeves SR, Parker AR, Whalen E, Misura KM, Barrow KA, James RG, 463 Hallstrand TS, Ziegler SF, Debley JS. Interferon response to respiratory syncytial 464 virus by bronchial epithelium from children with asthma is inversely correlated 465 with pulmonary function. J Allergy Clin Immunol 2018; 142: 451-459. 466 15. Doni Jayavelu N, Altman MC, Benson B, Dufort MJ, Vanderwall ER, Rich LM, White 467 MP, Becker PM, Togias A, Jackson DJ, Debley JS. Type 2 inflammation reduces 468 SARS-CoV-2 replication in the airway epithelium in allergic asthma through 469 functional alteration of ciliated epithelial cells. J Allergy Clin Immunol 2023; 152: 470 56-67. 471 16. McGrath-Morrow S, Lauer T, Yee M, Neptune E, Podowski M, Thimmulappa RK, 472 O'Reilly M, Biswal S. Nrf2 increases survival and attenuates alveolar growth 473 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 25 inhibition in neonatal mice exposed to hyperoxia. Am J Physiol Lung Cell Mol 474 Physiol 2009; 296: L565-573. 475 17. Burke R, Chu C, Zhou GD, Putluri V, Putluri N, Stading RE, Couroucli X, Lingappan 476 K, Moorthy B. Role of Human NADPH Quinone Oxidoreductase (NQO1) in 477 Oxygen-Mediated Cellular Injury and Oxidative DNA Damage in Human 478 Pulmonary Cells. Oxid Med Cell Longev 2021; 2021: 5544600. 479 18. Sucre JM, Wilkinson D, Vijayaraj P, Paul M, Dunn B, Alva-Ornelas JA, Gomperts 480 BN. A three-dimensional human model of the fibroblast activation that 481 accompanies bronchopulmonary dysplasia identifies Notch-mediated 482 pathophysiology. Am J Physiol Lung Cell Mol Physiol 2016; 310: L889-898. 483 19. Doyle LW, Andersson S, Bush A, Cheong JLY, Clemm H, Evensen KAI, Gough A, 484 Halvorsen T, Hovi P, Kajantie E, Lee KJ, McGarvey L, Narang I, Nasanen-485 Gilmore P, Steinshamn S, Vollsaeter M, Vrijlandt E, Adults born Preterm 486 International C. Expiratory airflow in late adolescence and early adulthood in 487 individuals born very preterm or with very low birthweight compared with controls 488 born at term or with normal birthweight: a meta-analysis of individual participant 489 data. Lancet Respir Med 2019; 7: 677-686. 490 20. Allen JL. Airway function throughout the lifespan: Pediatric origins of adult 491 respiratory disease. Pediatr Investig 2019; 3: 236-244. 492 21. Levin JC, Sheils CA, Gaffin JM, Hersh CP, Rhein LM, Hayden LP. Lung function 493 trajectories in children with post-prematurity respiratory disease: identifying risk 494 factors for abnormal growth. Respir Res 2021; 22: 143. 495 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 26 22. Cousins M, Hart K, Kotecha SJ, Henderson AJ, Watkins WJ, Bush A, Kotecha S. 496 Characterising airway obstructive, dysanaptic and PRISm phenotypes of 497 prematurity-associated lung disease. Thorax 2023; 78: 895-903. 498 23. Doyle LW, Irving L, Haikerwal A, Lee K, Ranganathan S, Cheong J. Airway 499 obstruction in young adults born extremely preterm or extremely low birth weight 500 in the postsurfactant era. Thorax 2019; 74: 1147-1153. 501 24. Hawkins FJ, Suzuki S, Beermann ML, Barilla C, Wang R, Villacorta-Martin C, 502 Berical A, Jean JC, Le Suer J, Matte T, Simone-Roach C, Tang Y, Schlaeger 503 TM, Crane AM, Matthias N, Huang SXL, Randell SH, Wu J, Spence JR, Carraro 504 G, Stripp BR, Rab A, Sorsher EJ, Horani A, Brody SL, Davis BR, Kotton DN. 505 Derivation of Airway Basal Stem Cells from Human Pluripotent Stem Cells. Cell 506 Stem Cell 2021; 28: 79-95 e78. 507 25. Amonkar GM, Wagner R, Bankoti K, Shui JE, Ai X, Lerou PH. Primary culture of 508 tracheal aspirate-derived human airway basal stem cells. STAR Protoc 2022; 3: 509 101390. 510 26. Lu J, Zhu X, Shui JE, Xiong L, Gierahn T, Zhang C, Wood M, Hally S, Love JC, Li H, 511 Crawford BC, Mou H, Lerou PH. Rho/SMAD/mTOR triple inhibition enables long-512 term expansion of human neonatal tracheal aspirate-derived airway basal cell-513 like cells. Pediatr Res 2021; 89: 502-509. 514 27. Morty Rory E. Using Experimental Models to Identify Pathogenic Pathways and 515 Putative Disease Management Targets in Bronchopulmonary Dysplasia. 516 Neonatology 2020; 117: 233-239. 517 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 27 28. Nardiello C, Mižíková I, Morty RE. Looking ahead: where to next for animal models 518 of bronchopulmonary dysplasia? Cell Tissue Res 2017; 367: 457-468. 519 29. Rogel MR, Soni PN, Troken JR, Sitikov A, Trejo HE, Ridge KM. Vimentin is 520 sufficient and required for wound repair and remodeling in alveolar epithelial 521 cells. FASEB J 2011; 25: 3873-3883. 522 30. Wang F, Ting C, Riemondy KA, Douglas M, Foster K, Patel N, Kaku N, Linsalata A, 523 Nemzek J, Varisco BM, Cohen E, Wilson JA, Riches DWH, Redente EF, Toivola 524 DM, Zhou X, Moore BB, Coulombe PA, Omary MB, Zemans RL. Regulation of 525 epithelial transitional states in murine and human pulmonary fibrosis. The Journal 526 of Clinical Investigation 2023; 133. 527 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint A B Healthy s BPD sBPD Submerged ALI Healthy C Control BPD Control BPD 0 10 20 30 40 50 60 70μm ALI Thickness Control BPD 0 20 40 60 80 100% % Ciliated Cells 0 20 40 60 0 50 100 ALI thickness (μm) % polarized ciliated cells Correlation R2 = 0.3 p<0.0001 TubA4A Phalloidin DAPI Database Term p value FDR GOBP 2025 Extracellular Matrix Organization (GO:0030198) 2. 55E-24 5. 31E-21 GO Cellular Component 2025 Co llagen-Containing Extracellular Matrix (GO:0062023) 2. 16E-33 3. 77E-31 Jensen Tissues Mesenchyme 2.89E-19 2. 09E-16 Jensen Compartments Protei naceous extrace llular matrix 3.90E-42 4. 82E-39 Reactome Pathways 2024 Extrace llular Matrix Organization 8.28E-37 6. 38E-34 MSigDB Hallmark 2020 Epithelial Mesenchymal Tr ansition 1.09E-62 4.03E-61 Gene Enrichment Analysis - Top 500 Genes Up in eBPD ALI Compared with Control Day 0-7 Proliferation Day 7 Air-liquid Interface Day 28 Fully Differentiated Tracheal Aspirate Figure 1 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint A B DAPI VIM ControlBPDControlBPDControlBPD 0 20 40 60 80 100Vimentin positive cells (%) VIM C D3Sub D7Sub ALI UP in BPDDOWN in BPD EPCAM Control BPD Figure 2 log2 FC -log10 FDR (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint RA24h48h72h96h TubA4A Phalloidin DAPI A B RA 24h 48h 72h 96h 0 20 40 60 80 Cilia percentage % 8 5 5 5 5N = Figure 3 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint Control BPD A Vimentin EPCAM DAPI B Ctr RA Ctr 24h Ctr 48h Ctr 72h Ctr 96h BPD RABPD 24hBPD 48hBPD 72hBPD 96h 0 10 20 30 40 50 60 70 80 % Vimentin Positive Cells RAO2 - 96h Quantitative Immunostaining Vimentin Figure 4 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint VIM and EPCAM Expression in Non-ciliated AECsVIM Expression in Secretory Cells and RASCs Five AEC Populations Identified by sc-Seq Fraction of cells in group (%) 0 1 2 3 4 Term infant BPD Expression Level in Airway Epithelium VIM 0.0 0.5 1.0 1.5 2.0 Term infant BPD Expression Level in Airway Epithelium EPCAM A UMAP1 UMAP2 Systemic venous EC Alveolar FB Activated FB Alveolar MyoFB Ductal MyoFB Pericyte VSMC T Cell Alveolar macrophage pDC cDC Mast cell Plasma cell Neutrophil Mesenchyme Immune MUC5B SCGB1A1 SCGB3A1 SCGB3A2 FOXJ1 KRT5 Secretory MUC5B Secretory -3A1,-3A2 RASC Multiciliated Basal Endothelium Epithelium Epithelium Endothelium Mesenchyme *** Immune AT1 AT2 RASC Secretory -3A1, -3A2 Secretory MUC5B Basal Multiciliated gCap aCap abCap Arterial EC Lymphatic Pulmonary venous EC Adventitial FB NK Cell NKT Cell Monocyte B Cell Basophil B n.s. UMAP1 Figure 5 DC umap1 00 3 44 3 22 11 1 umap1 BPD- VIMTerm Control - VIM umap1 umap2 umap2 umap2 cell type Mean Expression in Group 0.0 2.5 5.0 20 40 60 80 100 Secretory MUC5B Secretory -3A1,-3A2 RASC Multiciliated Basal (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint Control BPD VIMEPCAM TUBA 0 2000 4000 6000 8000 10000μm 2 (+/20x field) 0 20000 40000 60000 80000μ m 2 (+/20x field) ns p=0.15 0 1000 2000 3000 4000 5000 6000μ m 2 (+/20x field) * * A B Clinical Features Control (N=4) BPD (N=4) GA (weeks) 25.4+/-1 24.5+/-1 Male Sex (%) 75 75 Age at demise (days) 14.5 +/- 7.5 11.8 +/- 3.9 Intubated at DOL1 (%) 0 100 FiO2 DOL1 0.23 +/- 0.02 0.49 +/- 0.35 FiO2 48-72h before demise 0.32 +/- 0.16 0.78 +/- 0.26 Mucostasis on pathology (% cases) 25 100 Control BPD Control BPD Control BPD Figure 6 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 1 Supplement 1 2 Detailed Methods 3 4 Airway Epithelial Cell Organotypic Culture 5 Unprocessed tracheal aspirates were placed in warm PneumaCult Ex or Ex plus 6 medium, transported to the lab, and then cultured in collagen-coated T25 flasks. These 7 cultures were allowed to proliferate into islands, passaged once more into T25s, then 8 plated at P2 or P3 at 200,000 cells/ 24 well transwell for one week or until confluent. 9 AECs were then differentiated for three weeks at air liquid interface with PneumaCult-10 ALI media in the basolateral compartment. In parallel with neonatal AECs, control 11 bronchial AECs from healthy school-age children were collected while underdoing 12 sedated procedures. 4mm Harrell unsheathed bronchoscope cytology brushes 13 (CONMED®) were inserted through an endotracheal tube as we have previously 14 described prior to proliferation of P2 AECs in submerged culture followed by 15 differentiation of P3 AECs at ALI, as previously described (1-4). Quality control was 16 performed in both control and BPD AEC cohorts to ensure they produced an 17 organotypic differentiated epithelial cell culture with mucociliary morphology and without 18 leak indicative of epithelial barrier. TEER measurements were also used on occasion to 19 confirm barrier function. 20 21 Immunostaining of Organotypic Cultures 22 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 2 AEC cultures were washed with PBS and fixed with 10% formalin for 1h at the 23 room temperature (RT) and labeled by wholemount immunofluorescence staining. After 24 permeation with 0.5% Triton X-100, cells were then blocked with 10% goat serum in 25 PBS for 1h at RT. Incubation with primary antibodies were performed overnight at 4oC, 26 followed by recovery to room temperature and extensive washing with PBS. AECs were 27 incubated with fluorescently labeled stains or secondary antibodies for one hour at room 28 temperature. Thereby, cells were washed with PBS three times before mounting of 29 transwell with DAPI mounting medium on slides. For each AEC sample, two or three 30 locations were imaged. If the cilia stained with R-TUBA distributed evenly on the 31 samples, the images were randomly taken. If the cilia distributed unevenly, the images 32 included both low-density and high-density regions of TUBA+ staining, which were 33 averaged. 34 The following primary antibodies or stains were used: R-TUBA (Invitrogen, 35 1:200), R-Vimentin (Invitrogen, 1:100), R-TP63 (Invitrogen,1:100) and Phalloidin 36 (Invitrogen, 1:100). The following secondary antibodies were used: goat anti rabbit 37 Alexa Fluor-488 (Invitrogen, 1:1000) and goat anti rabbit Alexa Fluor-647 (Invitrogen, 38 1:1000). Phalloidin stain was applied with the incubation of secondary antibodies. 39 40 Quantitative AEC Differentiation Analysis (High Content Image Analysis) 41 Leica SP5 and Zeiss LSM 980 confocal microscopes with 10x and 40x oil 42

lens was used. For super resolution microscopy, sequential stacks of images 43 were taken along the z-axis at optimal intervals, 30-90 slices were imaged at 0.5 um 44 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 3 step size at resolution 1024x1024 pixels. Acquired images were processed using Fiji 45 [19] and Volocity (Quorum technologies) softwares. 46 Images from Leica SP5 confocal microscope were imported in Fiji/ImageJ2 47 v2.3.2/1.53q using Bio-format importer. ALI thickness was calculated in ImageJ by 48 counting the number of Z-stacks containing staining and multiplying by the thickness of 49 each layer (0.5 microns). Maximum intensity of every Nuclei Z-stack was projected in 50 order to count total number of cells per image, using manual counting with ImageJ cell 51 counter. Projections of maximum intensity of cilia and actin staining of the apical layer 52 were used as composite images to count the number of apical cells containing cilia 53 using manual counting with ImageJ cell counter. Exported data were analyzed in 54 GraphPad prism v9.4.1. 55 Cilia per cell analyses included the following procedure for data in Figure 1. 56 Maximum intensity of Z-stack containing cilia signal was projected and exported in tiff. 57 Those images were then imported in CellProfiler 4.2.1 (5). First illumination was 58 corrected using background correction and gaussian smoothing. Cilia objects were 59 identified from corrected images using Global two classes Otsu thresholding method, 60 with intensity de-clumping. For every image, we reported the total number of objects. 61 Subsequent cilia quantification included measurement of total TUBA+ positive area in 2-62 3 areas per transwell; total cell number and ALI thickness were also measured for each 63 sample. 64 65 RNA Sequencing and Analysis 66 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 4 Bulk RNA sequencing (RNA-seq) was performed on organotypic AEC cultures, 67 with support from the Benaroya Research Institute Genomics and Bioinformatics Cores, 68 with additional details in the supplement. RNA was isolated using RNAqueous™ Total 69 RNA Isolation Kit (Invitrogen) prior to reverse transcription and cDNA generation 70 (SMART-Seq, Takara Bio). We constructed sequencing libraries with the NexteraXT 71 DNA sample preparation kit (Illumina) to generate Illumina-compatible barcoded 72 libraries. Libraries were pooled and quantified using a Qubit® Fluorometer (Life 73 Technologies). Dual-index, single-read sequencing of the pooled libraries was carried 74 out on a HiSeq2500 sequencer (Illumina) with 58-base reads, using HiSeq v4 Cluster 75 and SBS kits (Illumina) with a target depth of 5 million reads per sample. FASTQs were 76 aligned to the human reference genome GRCh38.77 using TopHat (v1.4.1) and gene 77 counts were generated using htseq-count. QC and metrics analysis was performed 78 using the Picard family of tools (v1.134). Low-quality libraries (median CV of coverage > 79 1, total reads < 5 million) were excluded. Filtered and normalized gene counts were 80 generated from raw counts by trimmed-mean of M values (TMM) normalization and 81 filtered for genes that are expressed with at least 1 count per million total reads in at 82 least 10% of the total number of libraries. Data analysis was performed using the R 83 language using limma (6). Significance in volcano plots was determined by an adjusted 84 p-value cutoff ≤ 0.05 and a log2 fold change ≥ 1 between conditions. Analysis of gene 85 expression programs was performed with Enrichr databases (7-9). Gene set enrichment 86 analysis was also performed using the HALLMARK gene sets (10, 11). 87 88 89 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 5 Supplemental Figure Legends 90 91 92 Figure E1. Decreased Apical Cilia and ALI thickness in eBPD Cultures Grown in 93 PneumaCult Ex-Plus Supportive Media. (A) Decreased apical cilia in eBPD cultures 94 identified by immunostaining for TubA4A and actin. (B) Quantification showing 95 decreased ALI thickness and decreased % ciliated cells in organotypic cultures from 96 patients with eBPD (N=6), compared with controls (N=5). One dot is one image, 2-3 97 confocal images per donor (Green= anti-TubA4A, Red = phalloidin, Blue = DAPI. 98 *p<0.05, ****p<0.0001, Mann Whitney test, magnification bar = 100µm.) 99 100 Figure E2. Decreased Proliferative Capacity of eBPD AECs. (A) TP63 101 immunostaining of proliferative AECs from control and eBPD patients, on day 1 after 102 plating on transwell membranes. Quantification of (B) TP63+ cells in proliferative 103 conditions and (C) total cell numbers in fully differentiated Control and BPD AEC 104 cultures. One dot is one image, with N=2-3 images per condition. Control donor 105 numbers: Day 1=4 donors, Day 3=5, Day 7=6, ALI=9. eBPD donor numbers: Day 1=3 106 donors, Day 3=3, Day 7=5, ALI=11. (Green= anti-TP63, Blue= DAPI. *p< 0.05, ****p< 107 0.0001, Mann-Whitney test, magnification bar = 100mm) 108 109 Figure E3. Increased Vimentin-positive Stromal Cells in Organotypic AEC 110 Cultures from Patients with eBPD. (Top) Well-differentiated AEC culture from control 111 patient, stained with DAPI, phalloidin, vimentin, and EPCAM antibodies. (Middle) AEC 112 cultures from a patient with eBPD, showing a well-differentiated and polarized culture 113 with increased vimentin and abundant EPCAM staining. (Bottom) AEC from a patient 114 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 6 with eBPD, showing a poorly differentiated culture with increased vimentin staining of 115 cells resembling contaminating fibroblasts. (Blue = DAPI, Red = phalloidin, Green = 116 anti-vimentin, purple = anti-EPCAM, magnification bar = 100µm) 117 118 119 120 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 7 Supplement References 1. Lopez-Guisa JM, Powers C, File D, Cochrane E, Jimenez N, Debley JS. Airway epithelial cells from asthmatic children differentially express proremodeling factors. J Allergy Clin Immunol 2012; 129: 990-997 e996. 2. Reeves SR, Kolstad T, Lien TY, Elliott M, Ziegler SF, Wight TN, Debley JS. Asthmatic airway epithelial cells differentially regulate fibroblast expression of extracellular matrix components. J Allergy Clin Immunol 2014; 134: 663-670 e661. 3. Altman MC, Reeves SR, Parker AR, Whalen E, Misura KM, Barrow KA, James RG, Hallstrand TS, Ziegler SF, Debley JS. Interferon response to respiratory syncytial virus by bronchial epithelium from children with asthma is inversely correlated with pulmonary function. J Allergy Clin Immunol 2018; 142: 451-459. 4. Doni Jayavelu N, Altman MC, Benson B, Dufort MJ, Vanderwall ER, Rich LM, White MP, Becker PM, Togias A, Jackson DJ, Debley JS. Type 2 inflammation reduces SARS-CoV-2 replication in the airway epithelium in allergic asthma through functional alteration of ciliated epithelial cells. J Allergy Clin Immunol 2023; 152: 56-67. 5. McQuin C, Goodman A, Chernyshev V, Kamentsky L, Cimini BA, Karhohs KW, Doan M, Ding L, Rafelski SM, Thirstrup D, Wiegraebe W, Singh S, Becker T, Caicedo JC, Carpenter AE. CellProfiler 3.0: Next-generation image processing for biology. PLoS Biol 2018; 16: e2005970. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 8 6. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015; 43: e47. 7. Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV, Clark NR, Ma'ayan A. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics 2013; 14: 128. 8. Kuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, Koplev S, Jenkins SL, Jagodnik KM, Lachmann A, McDermott MG, Monteiro CD, Gundersen GW, Ma'ayan A. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res 2016; 44: W90-97. 9. Xie Z, Bailey A, Kuleshov MV, Clarke DJB, Evangelista JE, Jenkins SL, Lachmann A, Wojciechowicz ML, Kropiwnicki E, Jagodnik KM, Jeon M, Ma'ayan A. Gene Set Knowledge Discovery with Enrichr. Curr Protoc 2021; 1: e90. 10. Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, Puigserver P, Carlsson E, Ridderstrale M, Laurila E, Houstis N, Daly MJ, Patterson N, Mesirov JP, Golub TR, Tamayo P, Spiegelman B, Lander ES, Hirschhorn JN, Altshuler D, Groop LC. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet 2003; 34: 267-273. 11. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005; 102: 15545-15550. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint 9 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint A B C Control BPD TubA4A Phalloidin DAPI Control BPD 0 10 20 30 40 50 60 70 80% Ciliated Control BPD 0 10 20 30 40 50 60 70μM ALI Thickness Ciliated Cells Supplementary Figure E1 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint A Control BPD 0 1000 2000 3000Cell number after 21-day ALI Control BPD 0 20 40 60 80% TP63 Positive Cells Day 3 Total Cell Number Differentiated TP63+ Cells Day 3 Proliferative B C Control BPD Supplementary Figure E2 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint DAPI Vimentin EpCamActin 100um ControlBPD-1BPD-2 Supplementary Figure E3 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.16.676660doi: bioRxiv preprint