Obesity is a pre-disposing condition for chronic obstructive pulmonary disease, asthma, and pulmonary arterial hypertension. Accumulating evidence suggests that metabolic influences during development can determine chronic lung diseases (CLD). We demonstrate that maternal obesity causes early metabolic disorder in the offspring. Here, interleukin-6 induced bronchial and microvascular smooth muscle cell (SMC) hyperproliferation and increased airway and pulmonary vascular resistance. The key anti-proliferative transcription factor FoxO1 was inactivated via nuclear exclusion. These findings were confirmed using primary SMC treated with interleukin-6 and pharmacological FoxO1 inhibition as well as genetic FoxO1 ablation and constitutive activation. In vivo, we reproduced the structural and functional alterations in offspring of obese dams via the SMC-specific ablation of FoxO1. The reconstitution of FoxO1 using IL-6-deficient mice and pharmacological treatment did not protect against metabolic disorder but prevented SMC hyperproliferation. In human observational studies, childhood obesity was associated with reduced forced expiratory volume in 1s/forced vital capacity ratio Z-score (used as proxy for lung function) and asthma. We conclude that the interleukin-6-FoxO1 pathway in SMC is a molecular mechanism by which perinatal obesity programs the bronchial and vascular structures and functions, thereby driving CLD development. Thus, FoxO1 reconstitution provides a potential therapeutic option for preventing this metabolic programming of CLD.

Animal model of metabolic programming 

C57BL/6 (WT) were studied. Virgin female WT mice from our own colony served as future dams. After weaning [postnatal day 21 (P21)], they either received a high fat diet (HFD^mat; modified #C1057, Altromin, Lage, Germany) for induction of obesity or standard diet (SD, SD^mat; ssniff #R/M-H, V1534-0) for 7 weeks. At the end of the 7 weeks of HFD or SD feeding, female mice received an intraperitoneal (ip) glucose tolerance test (GTT). Subsequently, HFD^mat and SD^mat females were time-mated with SD-fed males. After birth, the litter size of all dams was normalized to six for each litter. Water and chow were available ad libitum and feed was withdrawn only for experimental reasons. After weaning at P21, male offspring of both groups were directly analyzed or were fed SD until P70, defining two groups

Laser-Microdissection 

Microdissection of bronchi and lung vessels was performed at P21 with PFA-fixed and paraffin-embedded lung tissue sections (13µm) on MembraneSlides (Leica #11600288, Wetzlar, Germany). Tissues were deparaffinized in isopropanol (Roth, #6752, Karlsruhe, Germany) and rehydrated in a graded ethanol series (100%, 96%, 80% and 70% for one minute each) to DEPC-H2O. Sections of bronchi and vessels were dissected with PALMProbe software (Zeiss, Oberkochen, Germany) at PALMMicroBeam microscope (Zeiss) through laser-microdissection with laser settings at speed 73, energy 56 µJ and focus 88. Laser-microdisseceted bronchi and vessels were collected in AdhesiveCap 500 opaque tubes (Zeiss, #415190-9201-000). After collecting the dissectates, samples were incubated with protein kinase K (Fermentas, Thermo Fisher Scientific, #EO0491, Waltham, Massachusetts, USA) overnight to dissolve bronchi and vessels out of the adhesive cap. Afterwards, RNA was extracted according to an established protocol with TRI reagent (Sigma Aldrich, #T9424-200ml, St. Louis, MO, USA). Laser-microdissection of bronchi and vessels at P70 was performed with cryosections from lung tissues on glass slides (Leica, PEN Membrane Slides #115051 2,0, Wetzlar, Germany) and stained with hemalaun. Bronchi, vessels, and adjacent tissues were microdissected with Laser Microbeam System (P.A.L.M., Bernried, Germany). RNA was extracted with the micro RNeasy kit (Qiagen, Valencia, CA, USA) afterwards.

RNA-Seq Analysis

Raw fastq files were trimmed using BBduk and mapped to the Mus musculus genome (GhCm38, using gene annotation v97) with the STAR aligner according to Lexogene recommendations. Sequence reads were assigned to genomic features using feature counts and DeSeq2 was used for statistical evaluation of significant differentially expressed genes between SD and HFD mice with a q-value cut-off of 0.05. Plots were created using pheatmap and heatmaply and manhattanly for interactive volcano plots, for gene ontology enrichment analysis, and pathway discovery we used clusterProfiler, gene annotation was performed with BioMart. Promotors were defined based on proximity with transcription start site (TSS) described for the genes. For the analysis, 2000bp region upstream of the TSS was selected. Distal enhancers were not included.