Impaired Fetal Lung Development can be Rescued by Administration of Extracellular Vesicles Derived from Amniotic Fluid Stem Cells

Incomplete lung development, also known as pulmonary hypoplasia, is a recognized cause of neonatal death and poor outcome for survivors. To date, there is no effective treatment that promotes fetal lung growth and maturation. Herein, we describe a novel stem cell-based approach that enhances fetal lung development via the administration of extracellular vesicles (EVs) derived from amniotic fluid stem cells (AFSCs). In experimental models of pulmonary hypoplasia, administration of AFSC-EVs promoted lung branching morphogenesis and alveolarization, and stimulated pulmonary epithelial cell and fibroblast differentiation. This regenerative ability was confirmed in two models of injured human lung cells, where human AFSC-EVs obtained following good manufacturing practices restored pulmonary epithelial homeostasis. AFSC-EV beneficial effects were exerted via the release of RNA cargo, primarily miRNAs, that regulate the expression of genes involved in fetal lung development. Our findings suggest that AFSC-EVs hold regenerative ability for underdeveloped fetal lungs, demonstrating potential for therapeutic application. One Sentence Summary Fetal lung regeneration via administration of extracellular vesicles derived from amniotic fluid stem cells


Introduction
Fetal lung development is a crucial step during embryogenesis, which if disrupted leads to a condition called pulmonary hypoplasia. Hypoplastic lungs have a reduced number of bronchiolar divisions, enlargement of airspaces, defective alveolarization, and impaired tissue maturation (1).
Pulmonary hypoplasia can be idiopathic or secondary to associated malformations, the most common of which is congenital diaphragmatic hernia (CDH) (1,2). CDH is a birth defect characterized by an incomplete closure of the diaphragm that leads to the herniation of intraabdominal organs into the chest, resulting in pulmonary hypoplasia (1,2). Pulmonary hypoplasia secondary to CDH has a mortality rate of 40% with most babies dying within the first days of life (3), and with 60% of survivors suffering from long-term morbidity (4,5). There is an unmet clinical need for an effective treatment that would rescue lung growth and maturation, but to date, none of the therapies tested has been successful (6). As pulmonary hypoplasia can be diagnosed as early as at the anatomy scan (18-20 weeks of gestational age), the paradigm of treatment in the last decades has focused on promoting lung growth and maturation before birth (6). Lung development is a complex process that is regulated by a network of signaling molecules, including small RNA species. In particular, some miRNAs are known to control biological processes that are important for lung development, such as branching morphogenesis and epithelial and mesenchymal differentiation (7)(8)(9), and have been reported to be missing or dysregulated in human and animal hypoplastic lungs (10)(11)(12)(13). Correcting the dysregulated network of signaling molecules would be beneficial to promote lung regeneration in fetuses with pulmonary hypoplasia.
A promising strategy to deliver a heterogeneous population of small RNA species is by administering extracellular vesicles (EVs) (14)(15)(16). EVs are small, subcellular, biological 4 membrane-bound nanoparticles that carry cargo in the form of genetic material and bioactive proteins (17)(18)(19). EVs are recognized key mediators of stem cell paracrine signaling and have been shown to promote tissue maturation and regeneration (20)(21)(22). Amniotic fluid stem cells (AFSCs) could be the ideal source of EVs to promote lung regeneration as AFSCs can integrate and differentiate into epithelial lung lineages (23), reduce lung fibrosis (24), repair damaged alveolar epithelial cells (25), and promote lung growth in a model of pulmonary hypoplasia secondary to CDH (26,27). AFSCs confer regenerative ability despite a low engraftment rate, thus suggesting a paracrine effect (26)(27)(28), which could, at least in part, be mediated by EVs.
Recently, EVs derived from AFSCs have been reported to hold regenerative potential in several animal models, including lung, kidney, and muscle injury (29).
In the present study, we have investigated whether administration of AFSC-EVs to various models of pulmonary hypoplasia could promote growth and maturation of underdeveloped fetal lungs.

AFSC-EV administration promotes growth and maturation in fetal hypoplastic lungs
The most robust experimental model of pulmonary hypoplasia relies on nitrofen administration to pregnant rats at embryonic day (E) 9.5 (30,31), which mainly targets retinoic acid synthesis (32). In this model, the whole litter has an impairment in lung development that is analogous to that of human fetuses (2,(30)(31)(32)(33). In utero nitrofen exposure causes a reduction in bronchiolar divisions and in the number of airspaces compared to lungs from unexposed fetuses (Fig. 1, A to C) (30). Administration of rat AFSC-EVs (mean size 140±5nm) to hypoplastic lung explants harvested during the pseudoglandular stage (E14.5) resulted in an increase in terminal branching morphogenesis (Fig. 1, A to C, fig. S1). Particularly, the total bud count and mean surface area 5 of lung explants treated with AFSC-EVs was increased compared to untreated hypoplastic lungs and similar to control lungs. This effect was specifically due to AFSC-EVs, as administration of AFSC conditioned medium (CM) or EV-depleted AFSC-CM did not rescue the impaired terminal branching. As larger AFSC-EVs (mean size 363±17nm) did not restore branching morphogenesis ( fig. S1F), in all the experiments performed hereafter we used small AFSC-EVs (mean size 140±5nm). We also tested whether the effects obtained with AFSC-EVs could be replicated by another source of EVs. There is growing evidence that EVs from mesenchymal stromal cells (MSCs) ameliorate experimental bronchopulmonary dysplasia (34), a neonatal lung condition similar to pulmonary hypoplasia secondary to CDH. However, administration of MSC-EVs to hypoplastic lung explants did not restore normal terminal branching (Fig. 1, A to C, fig.   S1). We also observed that AFSC-EVs did not affect branching morphogenesis in uninjured control lungs ( fig. S2A and B). When we investigated pathways responsible for lung branching morphogenesis, we observed that hypoplastic lungs had lower expression levels of Fgf10, Vegfα and its receptors (Flt1 and Kdr), compared to untreated lungs (Fig. 1D) (35,36). AFSC-EV administration improved the expression levels of these factors and receptors (Fig. 1D).
In addition to compromised fetal lung growth, pulmonary hypoplasia is characterized by impaired lung maturation, a feature replicated with in utero nitrofen exposure (2,30).
Hypoplastic lungs have decreased cell proliferation and delayed epithelial cell differentiation, as demonstrated by an increased number of distal progenitor cells (SOX9+) and by a reduced expression of surfactant protein C (SPC) (Fig. 1E to L, fig. S3A and B) (32,37,38). AFSC-EV administration rescued cell proliferation back to control levels and improved epithelial cell differentiation, as evidenced by a reduction of SOX9+ progenitor cell density (Fig 1E to F) and by an increased expression of SPC ( Fig. 1E-L). Hypoplastic lungs had similar levels of apoptosis 6 throughout the explant compared to control, as previously reported (33,39), and AFSC-EV administration did not alter this phenotype (Fig. 1M-N, fig. S3C).

AFSC-EV administration restores homeostasis and stimulates differentiation of lung epithelial cells
A hallmark of pulmonary hypoplasia secondary to CDH is the impairment in the homeostasis of the respiratory epithelium (2,40). Administration of AFSC-EVs to primary lung epithelial cells isolated from hypoplastic lungs of nitrofen exposed fetuses increased proliferation and reduced cell death back to control levels (Fig. 2,A and B). We confirmed that these effects were specific to AFSC-EVs, as they were not reproduced by the administration of either AFSC-CM, or EVdepleted AFSC-CM. Interestingly, MSC-EVs improved proliferation of primary lung epithelial cells, but failed to reduce cell death back to control levels ( fig. S3D and E). We also observed that administration of AFSC-EVs to uninjured control cells did not change their proliferation or cell death rates ( fig. S2C and D).
To study the effect of AFSC-EVs on respiratory epithelial cell differentiation, we generated fetal lung organoids (Fig. 2C, fig. S3F). The cell proliferation rate of organoids treated with AFSC-EVs was similar to that of organoids derived from control lungs, but higher than that of untreated organoids derived from hypoplastic lungs (Fig. 2D, fig. S3G). Moreover, AFSC-EV treated organoids had a more differentiated respiratory epithelium than untreated organoids, as demonstrated by higher expression levels of SPC (marker of early distal epithelium and alveolar type II cells) and CC10 (marker of club cells) (Fig. 2,E and F, fig. S3H and I).
AFSC-EV cargo content and its effect on lung epithelium 7 The beneficial effects of AFSC-EVs on nitrofen exposed hypoplastic lungs and on the respiratory epithelium were associated with the transfer of the EV cargo, which was detected throughout the lung parenchyma (Movie S1) and primary lung epithelial cells ( To test the role of the RNA cargo on fetal lung development, we performed an RNase-A enzymatic digestion of AFSC-EVs. We verified that RNase-A degraded the RNA cargo content and was captured by AFSC-EVs, as shown with immune-electron microscopy by its colocalization with TSG101 inside the AFSC-EVs (Fig. 3C-H). When we added RNase-pretreated AFSC-EVs to hypoplastic lung explants or to primary epithelial cells derived from hypoplastic lungs, we did not observe an increase in lung terminal branching and surface area, or an increase in cell proliferation and decrease in cell death, as seen with AFSC-EVs (Fig. 3, I-L). We confirmed that this effect was not due to a carry-over effect secondary to the transfer of RNase-A to the epithelial cells ( fig. S4B). This suggested that the delivery of AFSC-EV RNA cargo was a mediator of their biological effects on fetal lung development. We next used RNA-sequencing to identify and compare the small RNA cargo between AFSC-EVs and MSC-EVs. We found both AFSC-EV and MSC-EV cargos contained mRNA, tRNA, miRNA, and piRNA (Fig. 3M, fig.   S4C). Among all, miRNAs were the RNA species most proportionally different between the 8 cargos of the two populations. Compared to MSC-EVs, AFSC-EVs were enriched for miRNAs that are critical for lung development, such as the miR17~92 cluster and its paralogues (miRs-93, -106, -250, and -363; fold enrichment ranging from 5.1 -9.36; table S2, Fig. 3N). Moreover, AFSC-EVs contained miRNAs that have previously been reported as dysregulated in hypoplastic lungs, such as miR-33 and miR-200 (table S3) (10,12).
To identify the regulatory pathways affected by AFSC-EVs, we used mRNA-sequencing to compare gene expression between primary lung epithelial cells from nitrofen-exposed lungs treated with AFSC-EVs or MSC-EVs. We also applied the same treatment on primary lung epithelial cells from normal lungs. Nitrofen-exposure significantly altered the gene expression profile of primary lung epithelial cells compared to uninjured normal control epithelial cells ( fig.   S4D). Using gene set enrichment analysis, we found that AFSC-EV administration to nitrofen exposed primary epithelial cells altered the expression of genes related to epithelial differentiation and homeostasis maintenance, whereas MSC-EV administration altered the expression of genes involved with cell cycle regulation and nuclear organization ( Fig. 4A and B, fig. S5A and B). To understand the effects of AFSC-EVs on fetal lung epithelial cells, we manually queried these genes and identified those that are critical for lung development (table   S4). We then asked if the miRNAs identified in the AFSC-EV cargo were part of predicted regulatory networks with the mRNAs that were down-regulated in the target epithelial cells (Fig.   4C). The network that resulted from this analysis showed that there were genes important for lung development and that were regulated by the miR17~92 cluster and its paralogues. Small RNA-sequencing of the nitrofen exposed primary epithelial lung cells treated with AFSC-EVs showed that most of the miRNAs in the network had higher expression levels compared to the nitrofen only group (Fig. 4C). Lastly, we correlated the miRNA cargo content with the miRNA 9 target cell content, and their validated mRNAs in the target cells ( fig. S6). This triple analysis identified likely miRNA-mRNAs pairs that might be responsible for the phenotype observed.
This analysis provides indirect evidence that the miRNA cargo content of AFSC-EVs was transferred to the target cells after EV conditioning.

Towards the clinical translation of AFSC-EVs as a treatment for fetal lung regeneration
To test the effects of AFSC-EVs in vivo, we used a surgical model of pulmonary hypoplasia and CDH in fetal rabbits, as it allowed us to topically deliver our treatment ( . Lastly, when we tested human AFSC-EVs on the in vivo rabbit model, we found similar effects on alveolar wall thickness, as found with rat AFSC-EV administration ( fig. S7D to F).

Discussion
In this study, we have shown for the first time that administration of AFSC-EVs to various models of pulmonary hypoplasia promotes fetal lung regeneration. Specifically, AFSC-EVs administered to fetal hypoplastic lungs rescued branching morphogenesis and alveolarization by promoting epithelial and mesenchymal tissue growth and maturation and by re-establishing cellular homeostasis (Fig. 6). These beneficial effects were obtained through the release of AFSC-EV cargo. In particular, we identified miRNAs contained in AFSC-EVs that regulate lung maturation processes.
EVs are emerging as a successful strategy to promote tissue maturation and regeneration in various models. Their regenerative potential that we have observed in our models of fetal pulmonary hypoplasia has been observed in other models of tissue regeneration using either AFSC-EVs (29) or EVs from other stem cell sources (14-16, 20, 43, 44). In fact, there is increasing evidence that stem cell secreted EVs carry cargo that stimulates typical stem cells-like paracrine functions on target cells such as renewal, differentiation, and maturation (45)(46)(47). In this study, we confirmed that AFSC-EVs have a similar effect on hypoplastic lung explant and on lung epithelium to that exerted by their parent cells, AFSCs (fig. S8A to D). For this reason and because they are considered immunologically innocuous, EVs provide an advantageous and safer cell-free alternative compared to a stem cell-based therapy (48)(49)(50). Although conditioned medium could also be obtained in a GMP fashion and be a potential therapy in humans (51)(52)(53), in our study the EV fraction of the CM was more potent than the whole CM or than the EV- 11 depleted CM fraction, as also observed by others (54,55). Possibly, this effect is due to the fact that EVs carry active molecules that are more concentrated than in their parent cells ( fig. S8E). This is in line with our previous findings reporting that AFSC-EV concentration is the most important parameter responsible for their regenerative potential (29). Likewise, in the present study, we have shown that the impact of AFSC-EVs is dependent on the concentration of EVs isolated from the CM, as well as on the size of the vesicles that enter the target cells ( fig. S1D-G, Movie S2-3, and S5). How size contributes to biological function of EVs remains unknown (56).
Nonetheless, small EVs, also called exosomes (17), have traditionally been considered the EV subpopulation with more potent, protective, and pathological functions than larger vesicles, and therefore with more potential as diagnostic or therapeutic tool (29).
We used MSCs as an alternative source of stem cell EVs, as these cells are currently being tested in several clinical trials for the treatment of bronchopulmonary dysplasia (57), the postnatal lung condition that is most comparable to pulmonary hypoplasia. In our study, MSC-EV administration to our pulmonary hypoplasia models did not have similar beneficial effects as AFSC-EV administration, despite entering the primary lung cells ( fig. S2-3, Movie S6). The different effects obtained with the two populations of EVs may be due to differences in the disease pathogenesis, where pulmonary hypoplasia is mainly the result of abnormal and delayed lung development, and bronchopulmonary dysplasia is a chronic postnatal lung disease with substantial pulmonary inflammatory response (58). Moreover, our analysis of the EV cargo identified profile differences between the two EV populations, which could explain the outcome differences with MSC-EV administration.
In our models of pulmonary hypoplasia, we identified that the RNA cargo of AFSC-EVs was key to regenerate hypoplastic fetal lungs. This finding is in line with the knowledge that EV-12 mediated effects occur through the transfer of RNA species (14)(15)(16)(17). We found that miRNAs were the RNA species most differentially represented in the cargo between AFSC-EVs and MSC-EVs. When considering the AFSC-EV specific genes in the context of miRNA cargo enriched in AFSC-EVs, we found miRNA species that were important for the phenotypes that we observed, including branching morphogenesis, alveolarization, homeostasis, and cell differentiation (table S2). Specifically, a family of miRNAs that is enriched in AFSC-EVs compared to MSC-EVs is the miRNA 17~92 cluster. This cluster is key for lung branching morphogenesis (59), and when knocked out causes severe fetal pulmonary hypoplasia, making this a candidate mechanism that warrants further investigation (9). Moreover, the small RNA sequencing analysis revealed that members of this cluster were up-regulated in the nitrofen exposed primary lung epithelial cells following treatment with AFSC-EVs ( fig. S8F).
An improvement in fetal lung maturation was observed not just when the same species of AFSC-EVs were administered on the same species of target cells (i.e. rat AFSC-EVs on rat lung tissue or human AFSC-EVs on human lung epithelial cells), but also when we tested rat AFSC-EVs or human AFSC-EVs on the in vivo rabbit model ( Fig. 4A to F, fig. S7D to F). We speculate that the improvement in alveolarization observed in rabbit fetuses can be explained by the fact that some miRNAs and their targets are conserved across species (60). For instance, from the top 50 miRNAs that were enriched in the human AFSC-EVs, 13 are evolutionarily conserved between human and rabbit species, including one member of the miR17~92 cluster member (table S5).
Our study provides insights into the potential use of AFSC-EVs as a therapy for fetal pulmonary hypoplasia. Using GMP-grade human AFSC-EVs, we have confirmed similar beneficial effects on damaged epithelial cells derived from a fetal lung at the gestational age when pulmonary hypoplasia and CDH are typically detected. Further steps are needed for AFSC-EVs to be used 13 as a therapy in humans. One of the challenges will be identifying the most effective and safest route of administration to fetuses. In this study, we have employed topical administration via intra-tracheal injection in rabbit fetuses. This route could be further explored, also in clinical settings, as it is currently used to occlude the trachea of human fetuses with CDH that have the worst cases of pulmonary hypoplasia (42).
We acknowledge that our study has some limitations. Our findings are mainly based on the use of animal models and human lung epithelial cells to study a complex human condition with an unknown etiology. However, obtaining human lung tissues from babies with pulmonary hypoplasia is not considered ethically acceptable and nor has it been reported. Moreover, in this study we have focused mainly on the analysis of the fetal lung epithelium and mesenchyme, but we have not examined other lung tissue types potentially affected by AFSC-EV administration, such as pulmonary vessels, which are known to undergo vascular remodeling in hypoplastic lungs. Nonetheless, we have observed improvements in pathways important for lung vascular development, such as VEGF and FGF10 that suggest the opportunity for further studies. Another limitation is that little is known about factors that alter lung development and the mechanisms that are dysregulated in pulmonary hypoplasia. Similarly, it remains also unknown how exactly the EV RNA cargo species function. It is hoped that our transcriptomic profiling of both hypoplastic lungs and EVs will increase the understanding on the pathogenesis of pulmonary hypoplasia and on EV function. 14

Study Design
The objective of this study was to evaluate the ability of AFSC-EV administration to promote fetal lung growth and maturation in pulmonary hypoplasia. As obtaining human lung tissues from babies with CDH is not considered ethically acceptable, part of this study was conducted using animal models that closely resemble the degree of pulmonary hypoplasia that is observed in human fetuses. To advance towards a translational therapy, we obtained EVs from GMP-grade human AFSCs isolated from donated amniotic fluid during amniocentesis. We tested these human AFSC-EVs first onto a validated model of lung injury with the use of A549 alveolar epithelial cells (29,61). To more closely replicate the conditions of a human fetal lung, we also investigated the effects of GMP-grade human AFSC-EVs on nitrofen-exposed human pulmonary alveolar epithelial cells obtained from a healthy fetus at 21-weeks of gestation. As detailed below, experimental models and sample collections were approved by the appropriate regulatory EVs from rat and human AFSCs or bone marrow MSCs were isolated by ultracentrifugation from conditioned medium of cells that were treated with exosome-depleted FBS for 18 h, as previously described (29). In accordance with the International Society for Extracellular Vesicles guidelines, AFSC-EVs and MSC-EVs were characterized for size using Nanoparticle Tracking Analysis, morphology by transmission electron microscopy, and expression of canonical EVrelated protein markers by Western blot analysis, as previously described (29). To track EV migration into primary lung epithelial cells and lung explants, AFSC-EV and/or MSC-EV cargo was fluorescently labelled for RNA and protein using Exo-Glow™, and for lipid membrane using PKH26 red fluorescent cell linker. RNase enzymatic digestion of AFSC-EVs was conducted in a subset of experiments using AFSC-EVs to determine the role of RNA in fetal lung explants and cell cultures. A bioanalyzer was used to confirm the efficacy of RNase digestion. Co-labelled TSG101 and RNase were visualized by immuno-electron microscopy.

Experimental models of pulmonary hypoplasia
Ex vivo -In fetal rats, pulmonary hypoplasia was induced as previously described (30,40) with the administration of nitrofen to pregnant dams (100mg) by oral gavage on E 9.5. At E14.5, the dam was euthanized, and fetal lungs were micro-dissected. Lungs were grown as explants on nanofilter membranes for 72 h in culture medium alone (DMEM), AFSC-CM, AFSC-EVs (10% by volume), or MSC-EVs (10% by volume). Fetal lungs from dams that received olive oil (no nitrofen) at E9.5 served as control.
In vitro -1) For primary epithelial cell experiments in rats, a single cell suspension was obtained at E14.5 from pooled lungs from either control or nitrofen exposed rat fetuses by trypsinization for 20 minutes. Cells were subjected to three serial depletions of fibroblasts by incubation for 1h each following an established protocol (40,62). 2) For organoid studies, cells were seeded in a 16 ratio of 60:40 semi-solid Matrigel to medium, as previously described (63). Cells from nitrofentreated fetuses were cultured for 10 days with medium alone or with medium supplemented with 10% v/v AFSC-EVs or MSC-EVs. Lung organoids from fetuses whose mothers had not received nitrofen served as control. 3) A549 cells were treated for 24h with nitrofen (40µM), and treated with medium alone, human AFSC-EVs (10% v/v), or human MSC-EVs (10% v/v). Uninjured and untreated A549 cells served as control. 4) Human pulmonary alveolar epithelial cells (HPAEpiC) were obtained from the lungs of a healthy fetus at 21 weeks of gestation and used at first passage. Cells were stressed with 400µM nitrofen for 24h, and then treated with medium alone, or medium supplemented with 10% v/v human AFSC-EVs or human MSC-EVs.
In vivo -In fetal rabbits, pulmonary hypoplasia was induced secondary to surgical creation of a diaphragmatic hernia at E25 (41) in New Zealand rabbits. At E27, tracheal ligation was performed either alone or after intra-tracheally administration of a 50µL bolus containing either rat AFSC-EVs, rat MSC-EVs, or human AFSC-EVs (Movie S4). Lungs were harvested at E31 and immediately frozen for RNA extractions or fixed in 4% paraformaldehyde and embedded in paraffin.

Outcome measures
For lung morphometry, rat fetal lung explants were imaged by differential interference contrast microscopy and independently assessed by two blinded researchers for terminal bud density and surface area using ImageJ, as previously described (27). Rabbit fetal lungs were blindly evaluated with histology (hematoxylin and eosin staining) to assess the number of alveoli and the thickness of the alveolar wall, as described (64,65). For RNA expression, factors involved in rat or rabbit lung branching morphogenesis were assessed with quantitative polymerase chain reaction (RT-qPCR). For studies on lung tissue homeostasis on lung explants and organoids, 5- correlation studies, a Pearson coefficient was reported as r (confidence interval). P value was considered significant when p<0.05. All statistical analyses were produced using GraphPad Prism ® software version 6.0.
mRNA and miRNA sequencing analyses in lung epithelial cells were performed in R (version 3.6.0). Package "edgeR" (version 3.26.5) was used for differential analyses between two conditions with FDR < 0.1 considered as significant for both mRNA and small RNA-seq. 19 Table S3. miRNAs known to be involved in pulmonary hypoplasia and present in AFSC-EVs. Table S4. Genes differentially expressed in nitrofen-exposed lung epithelial cells.