AdipoRon hydrochloride was from Tocris (Minneapolis, MN). Adiponectin was from R&D (Minneapolis, MN). Anti-CD31 antibody (553370) was from BD Pharmingen (San Jose, CA). Anti-SPB (ab40876), -AQP5 (ab78486), -TWIST1 (ab50887) and -ERG (ab92513) antibodies were from Abcam (Cambridge, MA). Anti-VEGF164 (AF-493-NA) and -RAGE (MAB1179) antibodies were from R&D. Anti-VEGFR2 antibody (2479) was from Cell Signaling (Danvers, MA). Anti-adiponectin antibody (MA1-054) was from Thermo Fisher Scientific (Waltham, MA). Anti-actin antibody (A5441) was from Sigma (St. Louis, MO).

Quantitative reverse transcription (qRT)-PCR was performed with the iScript reverse transcription and iTaq SYBR Green qPCR kit (BioRad, Hercules, CA) using the BioRad real time PCR system. Cyclophilin and beta-2-microglobulin controlled for overall cDNA content. The primers for mouse Vegf, Vegfr2, Twist1 and cyclophilin and human TWIST1, VEGFR2, and beta-2-microglobulin were previously described (18, 35). The primers for mouse adiponectin forward; 5′-TGTTCCTCTTAATCCTGCCCA-3′ and reverse; 5′-CCAACCTGCACAAGTTCCCTT-3′, Adipor1 forward; 5’-AGACAACGACTACCTGCTACA-3’ and reverse; 5’-GTGGATGCGGAAGATGCTCT-3’; Adipor2 forward; 5’-GCCAAACACCGATTGGGGT-3’ and reverse; 5’-GGCTCCAAATCTCCTTGGTAGTT-3’. The protein levels of mouse VEGF and adiponectin were measured using ELISA (R&D systems). Immunoblotting was performed as we previously reported (19, 36). Gene knockdown was performed using the RNA interference technique as we previously reported (18, 35). In brief, we used siLentfect transfection reagent (BioRad) with siRNA (10 nM) following manufacturer instruction. The siRNA for human TWIST1 was previously described (18). As a control, siRNA duplex with an irrelevant sequence (QIAGEN, Hilden, Germany) was used.

C57BL6 (stock# 664), adiponectin knockout (B6;129-Adipoq^tm1Chan/J; Adipoq^−/−, stock# 8195) and control B6129SF2/J (stock# 101045) mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Adiponectin mRNA expression decreased by 82% in lungs isolated from Adipoq^−/− mice compared with those from control B6129SF2/J mice (Figure 3A). Heterozygote B6.V-Lepob/J mice (stock# 632, Lep^ob/+) were obtained from the Jackson Laboratory and bred to obtain homozygote (Lep^ob/ob). Lep^ob/+ mice were maintained with standard diet (LabDiet 5LOD, 4.5% fat). For the PNX experiments, Lep^ob/+ and Lep^ob/ob mice were fed with LabDiet 5K20 (10% fat) for 8 weeks, starting at 4 weeks of age. Human lung tissues were obtained as discarded surgical specimens from patients [Medical College of Wisconsin (MCW) tissue bank; Table 1]. De-identified patient demographic data were collected using the Generic Clinical Research Database at MCW. All protocols are approved by the Institutional Review Board (IRB) of MCW and Froedtert Hospital and ECs isolated from de-identified human lungs are determined as non-human subjects (PRO00047689).

ECs from mouse lungs and human lung ECs were isolated using anti-CD31 conjugated magnetic beads as previously described (17, 19, 25, 36). Briefly, lung tissue was cut into small pieces using small scissors and treated with collagenase A (1 mg/ml) for 30 min at 37°C. The tissue suspension was filtered through a 40 mm cell strainer (Falcon) to remove the undigested cell clumps and separate single cells. Cells were centrifuged (1,000 rpm, 5 min) at room temperature (RT) and the pellet was resuspended into 0.5 ml RBC Lysis Buffer (sigma, 1 min, RT). The lysis reaction was stopped by adding 10 ml 10% FBS/DMEM, and centrifuged (1,000 rpm, 5 min, RT). For mouse lung EC isolation, the pellet was resuspended into 0.5 ml 4% FBS/PBS with APC anti-mouse CD31 antibody (Biolegend, San Diego, CA, 1/100), incubated (20 min, on ice) and washed three times with 4% FBS/PBS. Cells were centrifuged (1,000 rpm, 5 min, RT) and resuspended into 0.1 ml 4% FBS/PBS with anti-APC conjugated microbeads (Miltenyl Biotec, Bergisch Gladbach, Germany), incubated (10 min, on ice) and washed three times with 4% FBS/PBS. The cells were then resuspended in 0.5 ml 4% FBS/PBS and CD31-positive mouse ECs were magnetically separated using MACS column (Miltenyl Biotec) according to the manufacturer's instruction. To increase the purity of the magnetically separated fraction, the eluted fraction was enriched over a second new MACS column. For human lung EC isolation, the cell pellet was resuspended into 1 ml 4% FBS/PBS with CD31-conjugated Dynabeads (Invitrogen/Thermo Fisher), incubated (30 min, 4°C), washed three times with 4% FBS/PBS, and magnetically separated using a magnetic stand. FACS analysis confirmed that more than 89% of isolated ECs cells are CD31⁺ (Supplementary Figure S1A) (17, 19, 25, 36).

Human lung ECs were seeded (1 × 10⁵ cells/35 mm dish) and DNA synthesis was measured using the Click-iT EdU Cell Proliferation Kit (ThermoFisher). Cells were imaged using a Nikon A1R confocal laser scanning microscope and quantification was performed using ImageJ software (NIH) (18, 25, 36). EC migration was analyzed by seeding human lung ECs (1 × 10⁵ cells/100 μl) on a trans-well chamber (Corning Costar) coated with 0.5% gelatin. ECs migrating to 5% FBS for 16 h were stained with Wright Giemsa solution (Fisher Scientific) and counted (18, 25, 36).

The in vivo animal study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocols were reviewed and approved by the Institutional Animal Care and Use Committee of MCW (AUA 5598). Unilateral PNX was performed as described (17–19). Briefly, mice (C57BL6, Lep^ob/+, Lep^ob/ob, Adipoq^−/−, or B6129SF2/J, 12–15 week old) were anesthetized with Ketamine (100 mg/kg)/Xylazine (10 mg/kg, intraperitoneal injection), intubated and mechanically ventilated using a rodent ventilator (MiniVent, Harvard Apparatus, Holliston, MA). After ensuring adequate anesthesia, thoracotomy was performed, and the left lung was lifted through the incision and a 5-0 silk suture was passed around the hilum and tied. The hilum was then transected distal to the tie. The remaining portions of the hilum and tie were returned to the thoracic cavity. Sham-operated mice underwent thoracotomy without PNX.

RNA was collected from ECs isolated from C57BL6 mouse lungs 7 days after PNX and sham-operated mouse lungs using the RNeasy mini kit (Qiagen). The quantity and quality of RNA isolated from mouse lung ECs (n = 3 per group, each n was pooled from 2 mice) were measured by Agilent 2200 TapeStation, and all samples have an RNA integrity number >9.3. Total RNA samples were submitted to the Institute for Systems Biology Molecular and Cell Core (Seattle, WA) for RNA sequencing. Library preparation was employed using the Illumina TruSeq Stranded mRNA kit. Sequencing was performed using the Illumina NextSeq500. Paired-end sequencing was performed on a high output 150 cycle kit v2.5. The RNA sequencing reads were aligned to the mouse genome (mm10 reference genome) and read counting and differential gene expression analysis were performed with Basepair Tech using the Deseq2 pipeline. 903 significantly differentially expressed genes defined as having a |log2 fold change| >1, and a p-adjusted value with the FDR cutoff of 0.01 calculated by the Benjamini-Hochberg adjustment and filtered to <0.01 were defined (Supplementary Table S1). Gene ontology (GO) analysis of significant targets was done via The Database for Annotation, Visualization and Integrated Discovery (DAVID) v 6.8 using the Functional Annotation Chart tool. Charts were filtered by Biological Processes Gene Ontology (BP GO) Terms and sorted by p-value (Supplementary Table S2). The 903 significantly differentially expressed genes generated 262 BP GO Term categories (Supplementary Table S2), which are further categorized into two different GO Term charts; 117 GO Term categories related to angiogenesis (Supplementary Table S3) and 141 GO Term related to metabolic process detected as appearing on a master list comprised of Gene Card (Supplementary Table S5). These GO Terms categories were color-coded into groups encompassing adhesion/migration, cell cycle/apoptosis, inflammatory/immune response, and cell signaling (Supplementary Tables S4, S6).

Network generation was performed on all genes from the top 50 GO Term categories by p-value related to angiogenic factors or metabolic process with Ingenuity Pathway Analysis (IPA) software (QIAGEN). The network of angiogenic genes was constructed by starting with the shortest connections between those connected to Twist1 and all others, and adding the shortest connections between all genes connected to Twist1 and the remaining unconnected genes. Genes were eliminated if they were connected to less than 3 other genes. The network of metabolic genes was constructed and interaction with adiponectin was analyzed as described for Twist1. The resulting IPA networks from the angiogenic factors and metabolic processes were combined to create a network illustrating the overlapping genes between the angiogenic factor and metabolic processes categories. RNAseq results are available in NCBI Geo (GSE179227).

All phenotypic analysis was performed by masked observers unaware of the identity of experimental groups. Error bars (SEM) and p values were determined from the results of three or more independent experiments. Student's t-test was used for statistical significance for two groups. For more than two groups, one-way ANOVA with a post-hoc analysis using the Bonferroni test was conducted.

Angiogenic signaling is necessary for lung vascular and alveolar regeneration after unilateral PNX (13, 15, 17–19, 23). Angiogenesis is inhibited in obese adipose tissues (25). The effects of obesity on regenerative ability of the lungs remain unknown. When we performed unilateral PNX on C57BL6 mice under normal chaw (4.5% fat, 5LOD) or on Lep^ob/+ or Lep^ob/ob mice fed with 10% fat diet (5K20), there was a significant increase in the ratio of the weight of right lung lobe to mouse body length 7 days after left unilateral PNX in C57BL6 or Lep^ob/+ mice as consistent with previous reports (13, 15, 17–19); the ratio of lung weight to mouse body length was 14.2 × 10⁻³ (g/cm) in the sham-operated control mice, while the ratio increased by 1.5-fold in the lungs 7 days after PNX (Figure 1A). Although the lung weight was not significantly different among C57BL6, Lep^ob/+ and Lep^ob/ob sham-operated mice, the increases in the lung weight after PNX were significantly reduced in Lep^ob/ob mice 7 days after PNX compared to those in C57BL6 or Lep^ob/+ mice (Figure 1A). We also analyzed the effects of obesity on blood vessel density in the lung using EC marker, ETS-related gene (ERG) staining; ERG-positive EC density was 2.2- or 2.1- times higher in the C57BL6 or Lep^ob/+ mouse lungs after PNX compared to those in the sham-operated control mouse lungs, while these effects were suppressed in Lep^ob/ob mice after PNX (Figures 1B,C). These findings suggest that post-PNX lung growth and vascular regeneration are inhibited in obese mice.

A major angiogenic factor, VEGF is necessary for post-PNX lung growth and vascular regeneration (13). The protein levels of VEGF in the mouse lungs and the serum in C57BL6 mice and Lep^ob/+ mice increased 7 days after PNX, while these effects were attenuated in Lep^ob/ob mice after PNX (Figure 1D). Immunohistochemical (IHC) analysis confirmed that VEGF expression in AQP5-positive alveolar type1 (AT1) cells (37) increased in the post-PNX C57BL6 and Lep^ob/+ mouse lungs. In contrast, the post-PNX increases in the VEGF expression were suppressed in Lep^ob/ob mice after PNX (Figures 1B,C). Similarly, post-PNX increases in the VEGFR2 expression were inhibited in Lep^ob/ob mice (Figures 1B,C; Supplementary Figure S1B). Among three VEGF isoforms (VEGF120, 164, and 188), the mRNA levels of Vegf164 increased after PNX in control Lep^ob/+ mouse lungs, but not in the Lep^ob/ob mouse lungs (Supplementary Figure S1C). The levels of Vegf120 and 188 did not significantly change in the post-PNX mouse lungs (not shown). We have reported that angiopoietin (Ang)-Tie2 signaling is stimulated after PNX to control regenerative lung growth (17, 19, 38). Thus, we also examined the expression of Ang-Tie2 in the post-PNX mouse lungs. Ang2 and Tie2 expression significantly increased in the control Lep^ob/+ mouse lungs after PNX; in contrast, the effects were inhibited in Lep^ob/ob mouse lungs (Supplementary Figure S1D).

Adiponectin induces angiogenesis (26, 27) and stimulates tissue regeneration [muscle (28), liver (29)]. The levels of adiponectin decrease in obese animals (26). The protein levels of adiponectin in lungs and serum increased 7 days after PNX in Lep^ob/+ mice; in contrast, the effects were suppressed in post-PNX Lep^ob/ob mice (Figure 2A). IHC analysis confirmed that increases in the levels of adiponectin in the post-PNX Lep^ob/+ mouse lungs were attenuated in Lep^ob/ob mouse lungs (Figure 2B). Expression of the adiponectin receptor, Adipor1 and Adipor2 did not change after PNX as well as Lep^ob/+ vs. Lep^ob/ob mouse lungs (Supplementary Figure S1E). These results suggest that obesity suppresses the levels of adiponectin in the post-PNX mouse lungs.

Transcription factor Twist1 controls expression of angiogenic factors to regulate angiogenesis (18, 32–34). We have reported that the levels of TWIST1 are lower in obese human subcutaneous adipose tissue ECs compared to that in lean adipose tissues, which inhibits angiogenesis (25). Knockdown of endothelial Twist1 inhibited post-PNX lung growth (18). In the bulk RNAseq analysis (GSE179227, Supplementary Table S1), the levels of Twist1 significantly increased in ECs isolated from the post-PNX mouse lungs (7 days, Figure 2C). IPA network analysis demonstrated that genes listed in the top 50 BP GO terms of each category (129 genes in angiogenic and 131 genes in metabolic processes) interacted with Twist1, and 88 genes were present in both categories relating to angiogenic and metabolic processes; 27 genes interacted directly (e.g., ADAM12, CXCL1, MMP2, MMP13, COL3A, WNT5A) and 61 genes interacted indirectly (e.g., LOXL2, COL6A, FOXM1, WT1, THBS2, CXCL5) with adiponectin (Figure 2D).

Consistent with the RNAseq data, TWIST1 expression increased in the post-PNX Lep^ob/+ mouse lungs, but not in the Lep^ob/ob mouse lungs (Figure 2E). Twist1 mRNA expression was also 2.2-times higher in ECs isolated from Lep^ob/+ mouse lungs 7 days after PNX, while the levels of Twist1 did not significantly change in post-PNX Lep^ob/ob mouse lung ECs (Figure 2F), suggesting that obesity inhibits post-PNX induction of Twist1 expression.

To analyze the effects of adiponectin on regenerative lung growth, we performed unilateral PNX on Adipoq^−/− mice, in which the mRNA levels of adiponectin in the mouse lungs were lower by 82% compared to those in the background matched wild type (WT) B6129SF2/J mouse lungs (Figure 3A). While the weight of the remaining right lung lobe increased by 1.3-times in control WT B6129SF2/J mice 7 days after PNX, the increase in the post-PNX lung weight was suppressed in the Adipoq^−/− mouse lungs (Figure 3B). Post-PNX increases in the number of surfactant protein B (SPB)-positive alveolar type-II (AT2) cells and ERG-positive ECs in the WT mouse lungs were suppressed in the Adipoq^−/− mouse lungs (Figure 3C). Increases in the levels of VEGF, VEGFR2, and TWIST1 after PNX were also attenuated in the Adipoq^−/− mouse lungs (Figures 3C,D; Supplementary Figure S1F).

To further examine the effects of adiponectin-TWIST1 signaling on EC behaviors, we isolated ECs from lean (BMI < 30) or obese (BMI > 30) human lungs as we reported (17, 19, 25, 36) (Table 1), treated ECs with adiponectin (300 ng/ml) or in combination with TWIST1 siRNA and investigated the effects. Adiponectin increased the mRNA and protein levels of VEGFR2, as well as stimulated DNA synthesis and migration in lean lung ECs; in contrast siRNA-based knockdown of TWIST1 attenuated the effects (Figures 4A–C). Adiponectin also stimulated DNA synthesis and migration in the obese lung ECs, which was inhibited by TWIST1 knockdown; however, adiponectin failed to increase VEGFR2 expression in the obese lung ECs (Figures 4A–C).

We then investigated whether stimulation of adiponectin signaling induces angiogenesis and regenerative lung growth in obese mice by treating mice with AdipoRon, an agonist of adiponectin receptors AdipoR1 and AdipoR2 (39). AdipoRon (oral gavage, 150 μg/mouse) stimulated post-PNX lung growth and increased ERG⁺ EC density and expression of VEGF and VEGFR2 in Lep^ob/ob obese mouse lungs after PNX when analyzed by IHC (Figures 5A,B). AdipoRon also increased the levels of VEGF and VEGFR2 in the post-PNX Lep^ob/ob mouse lungs compared to that without treatment (Figure 5C). These results suggest that adiponectin signaling increases VEGF-VEGFR2 expression and restores lung vascular and alveolar regeneration in obese mice.