The human fetal lung tissues used in this study were collected under IRB approval (The Lundquist Institute 18CR-32223-01) provided to the lab by the University of Washington Birth Defects Research Laboratory. They are de-identified and the only information collected were gestational age and known lung pathologies. Informed consent was provided for each lung collected and used in this study.

Fresh human fetal lung tissues between 12 and 21 weeks gestation (Table 1) were mechanically and enzymatically dissociated by treatment with DNAse I (8U/mL, 79256, Quiagen, GE) and collagenase IV (2 mg/mL, 17104-019, Gibco, MA, USA) to obtain primary lung fibroblasts. Fibroblasts were cultured by differential adhesion in Dulbecco’s Modified Eagle’s Medium: nutrient mix F-12 (D-MEM/F-12) with 1% FBS. Culture media was changed every 48 h and cells were kept in a cell culture incubator at 37°C, 5% CO2. Cells were plated in 12 wells plates or 4 wells chamber slides (respectively at 6 × 10⁴ and 3.5 × 10⁴ cells per well) until 60% confluency, then starved during 24 h before any treatment or assay. For all experiments, each “n” indicates a biological replicate.

siRNA pool for FGFR2, FGFR3 and FGFR4 (Dharmacon, CO, USA), detailed in Table 2, were incubated with lipofectamine (Lipofectamine RNAimax, Invitrogen, CA, USA) in D-MEM/F-12 with no serum or antibiotics and allowed to complex for 5 min at room temperature. The complex was added to the cell culture media at a final siRNA pool concentration of 10 nM for each and incubated for 24 or 48 h in the absence or presence of rhFGF18 (100 ng/mL, # 8988-F18, R&D, MN, USA).

RNA was extracted using the iNtRon Biotechnology, Inc. Easy-Spin™ Total RNA Extraction Kit (Burlington, MA, USA). RNA was reverse transcribed into cDNA using the Tetro cDNA Synthesis Kit (Bioline, Taunton, MA, USA) according to the manufacturer’s instructions. PCR products were amplified using specific TaqMan gene expression assays (listed in Table 3), Applied Biosystems, Foster City, CA, USA) and the TaqMan Universal PCR Master Mix II (Applied Biosystems). PCR products were detected using the StepONE Plus Real Time PCR System (Applied Biosystems). Each sample was run in triplicate.

Cells were washed with PBS, fixed in 4% PFA at room temperature (RT) for 30 min and blocked in 3% bovine serum albumine/5% Normal Goat Serum/0.1% Triton-X100 in TBS for 2 h. Primary antibodies (detailed in Table 4) were added overnight at 4°C. The next day slides were washed in PBS, stained with species appropriate secondary antibodies for 1 h at room temperature, and counterstained with DAPI.

Cells from chamber slides were immunostained for KI67 (Table 4) and imaged using a ×40 objective. Ten images were captured and quantified per sample. Cells were quantified using HALO® Image Analysis Platform (version 3.4.2986, Highplex FL and Area quantification module, Indica Labs, Inc; Albuquerque, NM).

Human fetal lung fibroblasts were plated in six well plates and incubated to reach confluence before being transfected with different siRNA in the presence or absence of rhFGF18 treatment. A straight line was drawn across the plate with a pipette tip. Images of the scratched cells were taken after 0- and 24-h using microscope (Leica DMi1, Leica Biosystems, Visa, CA). The migration ability by the cells was analyzed using ImageJ (Schindelin et al., 2012). Briefly, the closure area was calculated using the following formula: closure (%) = [(T₀ - T₂₄)/T₀] x 100, where T₀ represents the initial wound area and T₂₄ represents the remaining area at 24 h.

Following the different treatments, cells were lysed on ice in RIPA (Radioimmunoprecipitation assay) buffer supplemented with halt protease (Thermofisher) and phosphatase cocktail inhibitors (P5726 and P0044, Sigma). Proteins were quantified using DC protein assay kit (Biorad, Hercule, CA, USA) according to the manufacturer’s instructions. 20 μg of protein were loaded on an 8% Bis-Tris Plus precast polyacrylamide gel and run in 1X Bolt MES SDS running buffer (Invitrogen) using the Bolt System (Invitrogen Waltham, MA, USA). Transfers were performed using the iBlot 2 System (Invitrogen) and nitrocellulose gel transfer stacks. After the transfer, membranes were blocked in a 50:50 Odyssey Blocking (LI-COR, Lincoln, NE, USA): TBST (TBS and 0.1% Tween) solution at RT for at least 1 h and incubated with primary antibodies for 24 h at 4°C with continuous agitation (Table 4). The following day, membranes were washed with TBST and incubated with fluorescent secondary antibodies diluted in blocking buffer for 1 h at RT. Final detection was obtained by enhanced fluorescence with Chemidoc MP imaging system (Biorad, Hercules, CA, USA). Densitometry was analyzed using ImageLab software (Version 6.1, Biorad).

Statistical analyses were performed using GraphPad Prism (GraphPad Software Inc. La Jolla, CA, USA). Normality was assessed for each group using the Shapiro-Wilk test. If a group passed assumptions for parametric testing, a one-way ANOVA test was employed to compare each experiment to the control. The resulting p-values were corrected using the Dunnett test. If data did not pass the Shapiro-Wilk test, a Kruskal–Wallis test was conducted. The resulting p-values were adjusted using Dunn’s test. To compare two non-control groups (e.g., SiFGFR2 to SIFGFR2-FGF18), a t-test was performed if the data passed parametric assumptions. Furthermore, a Mann-Whitney test was used to compare two non-control samples when parametric testing was not applicable. The results were considered significant if p ≤ 0.05.

During human lung development, proliferation of mesenchymal cells is critical for branching morphogenesis and alveologenesis (Herriges and Morrisey, 2014) as defective proliferation can lead to congenital lung malformations and respiratory distress in newborns (Li et al., 2018). We first isolated and cultured human fetal lung fibroblasts (HFLF) from pseudoglandular (12–13 weeks) and canalicular (19–21 weeks) gestation lungs to evaluate the role of the FGF18/FGFR binding on cell proliferation. We previously demonstrated that FGF18 increases cell proliferation mainly in the epithelium of human explant cultures (Danopoulos et al., 2023). To determine whether FGF18 plays a role in cell proliferation in isolated lung mesenchymal cells at different gestational times, we treated HFLF with recombinant human protein FGF18 (rhFGF18). IF staining for KI67 demonstrated that FGF18 increased proliferation in HFLF independently of the age (pseudoglandular: 76.5% ± 4.9% vs. 47.25% ± 8%, p = 0.02, n = 4, Figure 1B; canalicular: 61% ± 8.2% vs. 35% ± 4.1%, p = 0.02, n = 4; Figure 1D). To identify the receptor through which FGF18 regulates HFLF proliferation, we first sought to investigate the expression of the different FGFRs in HFLF. We performed RT-qPCR to compare the baseline expression levels of the different FGFRs in both stages HFLF (Supplementary Figures S1A–C). The results showed no significant difference in the expression of FGFR2, 3, and 4 between pseudoglandular and canalicular stage HFLF. We next performed an siRNA assay for FGFR2, 3, and 4 to determine the ligand-receptor complex associated with proliferation. We confirmed successful downregulation of FGFR2 (siFGFR2, p = 0.012 in pseudoglandular and p = 0.014 in canalicular stage; n = 4, Supplementary Figure S1A), FGFR3 (siFGFR3, p = 0.003 in pseudoglandular and p = 0.018 in canalicular stage; n = 4, Supplementary Figure S1B) and FGFR4 (siFGFR4, p = 0.049 in pseudoglandular and p = 0.025 in canalicular stage; n = 4, Supplementary Figure S1C) by RT-qPCR as compared to the scrambled control (CTL SCR). The addition of rhFGF18 (100 ng/mL) did not alter the silencing effect (Supplementary Figures S1A–C). The pseudoglandular stage HFLF were treated with the different receptor siRNA (100 nM) in presence or absence of rhFGF18 for 24 h (Figures 1A, B) and stained with KI67 to assess proliferation. siRNA treatment for FGFR2 and FGFR4 did not affect proliferation, however, siRNA for FGFR3 significantly decreased the proliferation as compared to control (25% ± 3.6% vs. 47.25% ± 8%, p = 0.02; n = 4, Figure 1B). The addition of rhFGF18 following siRNA FGFR3 treatment rescued the decrease in cell proliferation to baseline control (56.7% ± 4.1% vs. 25% ± 3.6%, p = 0.01; n = 4, Figure 1B). However, the combination of siRNA for FGFR2 or FGFR4 treatments with rhFGF18 did not produce any significant rescue in proliferation compared to baseline control. The results suggested a role of rhFGF18 in pseudoglandular stage HFLF through FGFR2 and FGFR4. Canalicular stage HFLF were treated with the same conditions (Figures 1C, D). siRNA treatment for FGFR2, 3 or 4 significantly reduced HFLF proliferation (p = 0.03, p = 0.04 and p = 0.04 respectively; n = 4; Figure 1D). The addition of rhFGF18 only partially rescued the proliferation rate to baseline control when combined with siRNA targeting FGFR2 (33% ± 9% vs. 14.7% ± 3.5%, p = 0.04; n = 4; Figure 1D). These results suggested that rhFGF18 may have a role in promoting HFLF proliferation during the canalicular stage through FGFR3 and FGFR4. RT-qPCR analysis were performed to evaluate the expression levels of ETV4 and ETV5 transcription factors (Figures 1E–H), downstream targets of the FGF pathway that are frequently associated with cell proliferation (Herriges et al., 2015). Treatment with rhFGF18 led to a significant increase in the expression levels of ETV4 and ETV5 in both stage HFLF as compared to CTL SCR (respectively for pseudoglandular p = 0.04 and p = 0.02; and for canalicular p = 0.04 and p = 0.02; n = 3; Figures 1E–H). The expression of these transcription factors was not affected by the siRNA treatment; however, in pseudoglandular stage we noted that none of the siRNA rescue the expression of ETV4 and ETV5 after rhFGF18 treatment. Conversely, canalicular stage HFLF treated with siRNA for FGFR2 or FGFR3 and supplemented with rhFGF18, showed restored expression of ETV4 and ETV5 (respectively 0.03 ± 0.004 vs. 0.02 ± 0.004, p = 0.04 and 0.06 ± 0.008 vs. 0.04 ± 0.008, p = 0.02; n = 3; Figures 1G, H). No change was observed for FGFR4 siRNA. Finally, we investigated the protein levels of phospho-ERK through Western blot analysis (p-ERK, Supplementary Figure S1D–F). ERK signaling acts as a downstream target of the FGF pathway, promoting proliferation through the activation of ETVs transcription factors (Xiao et al., 2012; Joannes et al., 2016; Garg et al., 2020). Our findings demonstrate a significant increase in relative p-ERK following FGF18 treatment in both pseudoglandular and canalicular stages as compared to control (respectively 0.189 ± 0.014 vs. 0.105 ± 0.034, p = 0.048; Supplementary Figure S1E and 1.26 ± 0.16 vs. 0.92 ± 1.88, p = 0.005; Supplementary Figure S1F; n = 3). The addition of rhFGF18 after FGFR2 siRNA failed to induce ERK phosphorylation in pseudoglandular stage HFLF (Supplementary Figure S1E). A similar pattern was observed in the canalicular stage HFLF for FGFR4 (Supplementary Figure S1F). These results suggest that rhFGF18 enhances proliferation in both stages HFLF by upregulating ETVs transcription factor and ERK phosphorylation, through distinct FGFRs.

During human fetal lung development, mesenchymal cells migrate to provide structural support to the developing lung (Lee et al., 2017; Nasri et al., 2021). The migration of these cells is regulated by several factors, including FGF signaling (Joannes et al., 2016; El Agha et al., 2017). To identify the FGF18/FGFR complex(es) associated with this biological process, we performed scratch assays (Figure 2; Supplementary Figures S1G–I). We demonstrated that rhFGF18 significantly enhanced cell migration in both pseudoglandular and canalicular stage HFLF (respectively 55.2% ± 3.3% vs. 38.7% ± 6%; p = 0.01; and 29.1% ± 3.7% vs. 14.7% ± 2.3%; p = 0.04; n = 3; Supplementary Figures S1G,H). We noted that pseudoglandular stage HFLF presented a higher percentage of scratch closure at 24 h as compared to canalicular stage (p = 0.02). In addition, treatment with siRNA resulted in a significant reduction in migratory capacity of cells from both stages HFLF, as shown in Figures 2A, C, E. The percentage of wound closure was decreased by approximately 50% following siRNA treatment for all receptors (Figures 2B, D, F). In the pseudoglandular stage, the combination of FGFR2 siRNA and rhFGF18 did not restore HFLF migration (p = 0.51; n = 4; Figure 2B), while in the canalicular stage, HFLF migration was partially rescued (37% ± 2.4% vs. 22.2% ± 3.5%; p = 0.03; n = 4; Figure 2B). Interestingly, cells treated with FGFR3 siRNA and rhFGF18 failed to rescue the migration in both stages HFLF (Figure 2C). Finally, FGFR4 siRNA combined with rhFGF18 partially rescued the migratory capacities of pseudoglandular stage HFLF (38% ± 2.3% vs. 23.7% ± 6.1%, p = 0.03; n = 4; Figure 2F), whereas canalicular stage HFLF were unresponsive. The data indicated that rhFGF18 treatment enhanced cell migration in both pseudoglandular and canalicular stage HFLF through different receptors.

We recently demonstrated that FGF18 promotes human lung branching morphogenesis through regulation of mesenchymal progenitor cells (Danopoulos et al., 2023), but the underlying mechanisms remain unclear. Previous studies showed that during early embryonic mouse lung development, mesenchymal progenitor cells differentiate into myofibroblasts, while in later stages, they differentiate into lipofibroblasts (Al Alam et al., 2015; Riccetti et al., 2020; Liu et al., 2021). Therefore, we investigated the effect of the FGF18/FGFR pathway on mesenchymal cell fate. For this purpose, we assessed myofibroblast (ACTA2, COL1A1) and lipofibroblast (ADRP, PPARγ) markers in pseudoglandular and canalicular HFLF treated with rhFGF18 (Figure 3). RT-qPCR analysis revealed that treatment with rhFGF18 resulted in decreased ACTA2 expression in both pseudoglandular and canalicular stage HFLF (0.23 ± 0.07 vs. 0.4 ± 0.05, p = 0.006 and 0.16 ± 0.01 vs. 0.35 ± 0.03, p = 0.01 respectively, n = 4; Figures 3A, B). Furthermore, treatments with FGFR2, FGFR3 or FGFR4 siRNA did not alter ACTA2 expression from baseline expression (Figures 3A, B). Whereas the use of FGFR2 or FGFR3 siRNA combined with rhFGF18 significantly decreased ACTA2 gene expression in pseudoglandular stage HFLF (0.23 ± 0.06 vs. 0.42 ± 0.07, p = 0.02 and 0.16 ± 0.04 vs. 0.33 ± 0.03, p = 0.009 respectively; n = 4; Figure 3A), the combination of rhFGF18 with FGFR4 siRNA had no effect on ACTA2 expression. A decrease in ACTA2 protein levels was also observed by WB (Figure 3C, top panel, and Supplementary Figures S2A,B) (Figures 3D, E). Similar to what was observed in pseudoglandular stage, canalicular stage HFLF displayed a decreasing trend of ACTA2 expression when treated with FGFR2 siRNA and rhFGF18 (p = 0.08; Figure 2B), whereas FGFR3 siRNA combined with rhFGF18 showed no change. These results were also observed at the protein level, as demonstrated by WB and IF staining (Figure 3C, bottom panel; Figures 3F, G). In contrast to what was observed in pseudoglandular stage, HFLF treated with FGFR4 siRNA in conjunction with rhFGF18 showed a significant decrease in ACTA2 expression (0.15 ± 0.03 vs. 0.28 ± 0.04; n = 4, p = 0.02; Figure 3B), which was confirmed by WB (Figure 3C, bottom panel) and IF staining quantification (Figures 3F, G). The regulation of COL1A1, another myofibroblast marker, was also investigated (Supplementary Figures S2C,D). COL1A1 expression was significantly reduced after rhFGF18 treatment in both stages (3.42 ± 0.4 vs. 6.17 ± 0.6, p = 0.03, n = 4 in pseudoglandular stage and 2.06 ± 0.3 vs. 5.45 ± 0.4 in canalicular stage HFLF, p = 0.003, n = 3 Supplementary Figures S2C,D). However, the addition of rhFGF18 to the siRNA treatment did not lead to a similar decrease in both stages.

We then assessed the expression of ADRP (lipofibroblast marker) in the same conditions (Figures 3H–J). rhFGF18 significantly increased ADRP expression solely in pseudoglandular stage HFLF (0.04 ± 0.05 vs. 0.02 ± 0.004; n = 4; p = 0.001; Figures 3H, I). Although no significant differences were observed following treatment with siRNA and rhFGF18, we did observe a trend of upregulation in the pseudoglandular stage in the presence of FGFR2 and FGFR4 siRNA (Figure 3H). These results were confirmed by WB (Figure 3J and S2E and F). Moreover, PPARγ (lipofibroblast marker) expression was significantly upregulated in both stages HFLF by rhFGF18 (respectively 0.005 ± 0.0004 vs. 0.002 ± 0.0004; p = 0.01, and 0.001 ± 0.0001 vs. 0.008 ± 0.00005, p = 0.02; n = 3; Supplementary Figures S2G,H). PPARγ gene expression demonstrated an increasing trend in both pseudoglandular and canalicular stage cells treated with FGFR2 siRNA supplemented by rhFGF18, whereas a significant upregulation was observed for FGFR3 siRNA treated with rhFGF18 in the pseudoglandular stage only (0.003 ± 0.0005 vs. 0.002 ± 0.004; p = 0.008; n = 3; Supplementary Figure S2H). Combining FGFR4 siRNA with rhFGF18 did not produce the same results as rhFGF18 alone (Supplementary Figures S2G,H). Taken together, our data suggest that FGF18/FGFR(s) plays a role in promoting the differentiation of mesenchymal progenitors by reducing the expression of myofibroblast markers and increasing lipofibroblast markers. However, the effects seem to vary depending on the stage and the specific ligand/receptor interaction.