Animals

Mice were housed within a temperature-controlled room (21–22 °C) under a 12 h light/dark cycle and allowed free access to food and water.

Animal models

Four different mouse transplant models were used: (1) C57BL/6 CD45.2⁺ donors into C57BL/6, CD45.1⁺ recipients; (2) LDL receptor knockout donors expressing GFP^+/− (GFP^+/−LDLR^−/−) into GFP^−/− LDLR^−/− recipients; (3) miR-106b^−/− C57BL6 CD45.2⁺ donors into C57BL/6 CD45.1⁺ recipients; and (4) Vav1Cre^+/−or Vav1Cre^−/−Jarid2^fl/fl mice. To generate fetal liver hematopoietic stem cell (FL-HSC) donors, we obtained C57BL/6 CD45.2⁺ (Jax/Lab 002014), GFP^+/− C57BL/6 (Jax/Lab 004353), LDLR^−/− (Jax/Lab 002207), miR-106b^−/− CD45.2⁺ (Jax/Lab 008460) mice, Vav1Cre CD45.2⁺ (Jax/Lab 00861) and Jarid2^fl/fl CD45.2⁺ (Jax/Lab 031141). We generated GFP^+/− LDLR^−/− mice by crossing GFP^+/− C57BL/6 with LDLR^−/− mice. LDLR^−/− were initially crossed with mice that constitutively express the green fluorescent protein (GFP^+/−) as donors to facilitate the identification of GFP-positive embryos by UV lamp and assessment of donor bone marrow chimerism in GFP^−/− recipients. C57BL/6 CD45.2⁺ were used as donors, and C57BL/6 CD45.1⁺ as recipients to facilitate the determination of engraftment without the influence of GFP. There were no differences in insulin resistance phenotype regardless of the engraftment verification model. To generate vitamin D-deficient or -sufficient [VD(−) or VD( + )] FL-HSC donors, we transitioned the dam’s diet 4 weeks prior to pregnancy to either vitamin D-deficient (Harlan TD.87095) or sufficient (Harlan TD.96348) diet⁶⁷. Females were mated with vitamin D-sufficient males to prevent the effects of vitamin D deficiency on male fertility. At gestational day 13.5, VD(−) or VD( + ) FL-HSCs from males and females were harvested for transplantation. A subgroup of pregnant dams was allowed to progress to term. Pups born both vitamin D-deficient and sufficient were weaned to vitamin D-deficient or sufficient diets for 8 weeks. Glucose and insulin tolerance test were performed, and peritoneal macrophages were obtained for RNA expression. All experiments included male and female animals. Mice were euthanized using cervical dislocation according to institutional guidelines. E13.5 embryos are not viable after c-section. Protocols were approved by the Washington University Institutional Animal Care and Use Committee (Protocol 21-0127) and complied with ethical regulations for laboratory animal studies.

Primary FL-HSC transplants and secondary BM transplants

Fetal liver cells at embryonic day 13.5 include HSCs with a high proliferative capacity, increasing donor engraftment by 10-fold compared to BM stem cell donors⁶⁸. For fetal liver transplantation, vitamin D-deficient or sufficient C57BL6 or LDLR^−/− pregnant mice were sacrificed at 13.5 days gestation, with the vaginal plug counted as day 0.5⁶⁹. For the GFP model, positive embryos were selected by UV lamp before dissecting each fetal liver. Fetal livers were then rinsed in sterile saline, followed by trypsinization for 15 min at 37 °C. Fetal liver cells were resuspended in cold DMEM with 5% fetal bovine serum (FBS, Gibco #16000044), filtered through a 70‐µm filter (BD #352350), centrifuged at 125×g for 10 min, re-filtered through a 40‐µm filter, and centrifuged at 125×g for 5 min. The cells were then rinsed in 15 mL of cold phosphate‐buffered saline (PBS), pelleted, resuspended in 1 mL of PBS, counted using a hemocytometer and adjusted to 10⁵ cells per µL. Cells were genotyped by PCR for the Sry sex-determining region of chr Y to create mixed male and female pools for donation. Fetal liver cells, including HSCs, were injected intravenously within 8 h into 8-week-old vitamin D-sufficient recipient male and female mice following lethal irradiation with 10 Gy from a ¹³⁷Cs gamma irradiator source. For bone marrow transplantation, BM cells were isolated from 24-week-post-primarily transplanted recipients of VD(−) or VD(+) HSCs by flushing the femurs and tibias with ice-cold PBS²⁸. Total BM was washed, triturated using a 24-gauge needle (Benson Dickson), collected by centrifugation at 300 g for 4 min, and diluted with PBS. After lysis of erythrocytes using Red Blood Cells Lysis buffer (Roche#11814389001), cells were counted. Eight-week-old vitamin D-sufficient C57BL6 (CD45.1⁺) or GFP^−/− LDLR^−/− male and female recipient mice were lethally irradiated with 10 Gy from a ¹³⁷Cs gamma irradiator source. Within 6 h after irradiation, recipient BM was reconstituted with ~5 × 10⁶ donor marrow cells via a single injection. Eight weeks after either type of transplantation and reconstitution, recipient engraftment was evaluated by flow cytometry quantification of the percentage of CD45.2 or GFP positivity in peripheral leukocytes of recipients, with only those animals >87% chimeric used in experiments^28,33.

Metabolic assessment

Blood samples: Fasting serum glucose, cholesterol, triglycerides, and FFA were measured after 6 h fasting using commercially available kits. For glucose and insulin tolerance tests, transplant recipient mice were evaluated at 8 weeks or 24 weeks post-transplant. Mice fasted for 6 h before peritoneal injection with 10% D-glucose (1 g/kg, Pfizer, # 00409-7517-16) or insulin (0.75 U/kg, Humulin R # U-100). For both studies, tail vein blood glucose was assayed using a glucometer at baseline and 30, 60, and 120 min after injection⁷⁰. Plasma insulin was assessed at 30 min during GTT (by electrochemiluminescence immunoassay). Hyperinsulinemic-euglycemic clamps and in vivo 2-deoxyglucose uptake (2-DG) assays were performed five days after double-lumen catheters were placed. Animals fasted overnight, and glucose turnover was measured in the basal state and during the clamp at 12 weeks post-transplant in conscious mice, as previously described^28,70,71. Immediately after euthanasia, hind limb muscles and perigonadal fat were harvested, washed with PBS, and placed in liquid nitrogen until pending analysis. Frozen tissue samples were ground, boiled, and centrifuged. Accumulated 2-DG in the supernatant was separated by ion exchange chromatography using a Dowex 1-X8 (100–200 mesh) anion exchange column. Data are expressed as µmol/100 grams of tissue/min [(2-DG x mean blood glucose)/area under the curve]. For insulin-stimulated adipocyte 2-DG uptake, differentiated 3T3-L1 adipocytes or primary adipocytes were co-cultured for 72 h with peritoneal or stromal vascular macrophages or exposed to described conditioned media. For co-culture, macrophages (0.3 × 10⁶) were placed on inserts in transwell plates with 3T3-L1 adipocytes or primary isolated perigonadal adipocytes in the bottom chamber. After macrophage co-culture, adipocytes were serum-starved for 3 h, washed, and incubated with or without insulin (10 nM) for 30 min, then incubated for 10 min with radioactive 2-DG (PerkinElmer NEC 720A250UC). Cells were washed in cold Krebs–Ringer phosphate HEPES (KRPH), and after lysis, [¹⁴C] was determined by scintillation counting to measure 2-DG uptake⁷². Cytochalasin B, a glucose transport inhibitor (50 µM), was used to correct for non-specific background uptake. Data is presented as a ratio of 2-DG uptake after insulin stimulation to that of non-insulin-stimulated cells. Immunoblots for phospho-AKT (Ser 473) (Cell Signaling #4058 dilution 1 µg/mL), AKT (Cell Signaling #112580 dilution 0.5 µg/mL), and β-Actin (Cell Signaling #8457 dilution 0.5 µg/mL) were performed in homogenized perigonadal adipose tissue from HSC recipients with or without insulin stimulation (Humulin R) 1 µg/mL for 5 min. For skeletal muscle 2-DG uptake, primarily transplanted mice were fasted for 6 h, then paired soleus and extensor digitorum longus (EDL) muscles from anesthetized mice were excised and incubated using a 2-step incubation protocol⁷¹. For all incubation steps, vials were continuously gassed with 95% O₂/5% CO₂ and shaken in a heated water bath, and one muscle from each mouse was incubated in solution supplemented with 100 μU/ml of insulin (Humulin R) while the contralateral muscle was incubated in solution without insulin (basal) followed by three incubation steps as previously described⁷¹. After incubation with 2-DG for 15 min, muscles were rapidly blotted on filter paper dampened with incubation medium, trimmed, freeze-clamped, and stored at −80 °C. Muscles were homogenized, and 2-DG uptake was determined by scintillation counting.

Isolation of peritoneal and adipose tissue macrophages and primary adipocytes for macrophage phenotype characterization and adipose 2DG uptake

Peritoneal macrophages

Unstimulated macrophages were collected following peritoneal PBS injection and placed in 12-well transwell inserts (Costar polycarbonate filters, 3-µm pore size) for co-culture with adipocytes or in 12-well plates for 6–8 h for media collection as previously described^28,67. Stromal vascular fraction isolation. Mice were perfused through the right ventricle with 25 mL of ice-cold PBS. Epididymal fat pads, were excised and minced in PBS with 0.5% BSA. Collagenase I (1 mg/ml, Millipore Sigma SCR 103) was added before incubation with shaking/rotating, and digestion was stopped with pre-warmed KHB (Millipore Sigma). The cell suspension was filtered through a 250-μm filter and spun at 500 × g × 5 min to separate floating adipocytes from the stromal vascular fraction (SVF) pellet. The SVF pellet was resuspended in FACS buffer and incubated with anti-CD14 magnetic microbeads (Miltenyi Biotec 130-050-201) and F4/80 (Miltenyi Biotec, 130-110-443) to isolate ATM, which were then placed in 12-well plates with transwell inserts for adipocyte culture or in collagen-coated plates with DMEM plus 10% exosome-depleted FBS for 6 h for media miRNA expression and cytokine analysis⁷³.

Primary adipocytes

Perigonadal fat pads were initially treated with collagenase as described above, but then floating adipocytes were spun at 100 × g × 1  min. Buffer and SVF pellet underneath the floating adipocytes were removed, and cells were washed with pre-warmed KRPH. Adipocytes were transferred into collagen-I-coated 12-well plates (0.5 × 10⁶ cells/plate) and incubated at 37 °C for 60 min before starting the 2-DG protocol.

3T3-L1 adipocyte differentiation

Murine 3T3-L1 pre-adipocytes (American Type Culture Collection CL-173) were grown, maintained, and induced to differentiate using a standard protocol⁷⁴. Fully differentiated adipocytes (12 days post differentiation induction) were maintained in DMEM supplemented with 10% FBS (Millipore Sigma) until two days before experimentation when cells were fed with 10% calf serum (Millipore Sigma). Prior to experiments, the media was changed to serum-starved DMEM (low glucose) for 2–3 h.

Macrophage and adipocyte co-cultures

Transwell chambers were utilized (Costar polycarbonate filters, 3-µm pore size) as previously described²⁸ for co-cultures. Membranes and 12-well plates were coated with fibronectin (Sigma Catalog # 341631) overnight at 4 °C. Peritoneal or SVF macrophages were cultured in DMEM plus 10% exosome-depleted FBS for 6 h, and media was tested for miRNA expression and cytokine levels prior to co-culture. Differentiated 3T3-L1 adipocytes were grown to 80–90% confluence in 12-well plates, with a maximum of 0.5 × 10⁶ cells/well. Macrophages (0.3 × 10⁵ cells/well) were added to the transwell upper chamber, with adipocytes in the lower chamber. Cells were co-cultured for 72 h in DMEM/F12 media with 10% exosome-depleted FBS with 2-DG quantification as described above. Some 3T3-L1 cells were cultured only with macrophage-conditioned media but additionally incubated with 0.2 µg/ml TNFα-, IL-1β-, or IL-6-neutralizing antibodies (R&D Biosystems MAB4101) for 72 h, with 2-DG quantification as described above⁷¹.

Analysis of monocytes and adipose tissue macrophages (ATMs) and HSCs in the fetal liver by conventional flow cytometry

Monocyte and eWAT SVF macrophage cell surface marker analysis was performed using a FACStar Plus as previously described^75,76. After isolation, including CD11b selection with microbeads (Catalog #130-049-601) cells were resuspended in flow cytometry buffer (BD Bioscience # 51-2091KZ), and >10⁵ cells were analyzed for each sample. Monocytes were incubated with 10 µg/ml of anti-mouse APC-CD45.1 (BioLegend, #110714) or 10 µg/ml of anti-mouse PE-CD45.2 (BioLegend #109808) for 15 min on ice, then washed before flow cytometry with utilization of 5 µg/ml APC mouse IgG2a/K (BioLegend #400219) and 5 µg/ml of PE mouse IgG2a/K (BioLegend #400211) isotype controls. EGFP-positive cells were excited at 488 nm and measured by flow cytometry at 530 nm. To determine the percentage of ATM cells, CD11b+ SVF from epididymal fat were incubated with 10 µg/ml of anti-mouse PE-F4/80 (BioLegend, #12-4801-82) with the utilization of 5 µg/ml PE rat IgG2a K isotype Control (BioLegend#12-4321-80). To further characterize ATM cells, CD11b+ SVF were stained with 4 µg/ml anti-CCR7 APC-eFluor780 (#47-0271-82 eBioscience) and 4 µg/ml anti-CD86 eFluor450 (#48-0862-80 eBioscience) for M1 macrophage marker expression and 20 µg/ml anti-CD163-Cy5.5 (# M130 Bioss USA) and 4 µg/ml anti-CD206-Alexa700 (#FAB2535N R&D Systems) for M2 macrophage marker expression. Fetal liver cells were incubated with antibody cocktails for long-term hematopoietic stem cell (HSC) antibodies: 10 µg/ml Alexa Fluor® 700 or PE anti-mouse lineage cocktails (#79923 or #78035 BioLegend), 4 µg/ml Sca1- Ly-6A-Alexa Fluor® 700 (#56-5981-82 eBioscience), 4 µg/ml CD117-APC (#17-1172-81 eBioscience), 4 µg/ml CD150-APC 780, (#47-1502-82 eBioscience); Short-term HSCs antibodies: 4 µg/ml CD34 – PE-Cy7 (#25-0349-42 eBioscience), 5 µg/ml Flt3 – PE-Cy5 (#15-1351-82 eBioscience); 5 µg/ml pro B cell: B220-efluor450 (#48-0452-6B2 eBioscience), 4 µg/ml CD43-PE (#12-0431-82 eBioscience); pro T cells: 4 µg/ml CD25-PE (#12-0259-80 eBioscience), CD44-APC (#17-0441-82 eBioscience); 8 µg/ml CMP/GMP: CD16/32-eFluor450 (#48-0161-82 eBioscience); 5 µg/ml MDP/CDP: CD115 APC-eFluor780 (# 47-1152-82 eBioscience).

Cells were acquired using FACStar Plus flow cytometer. Cell aggregates, dead, and cellular debris were excluded based on FSC/SSC. Batch analysis by Flow Jo version 9.6.2 was used for gating consistency and selection of positive populations. Unstained samples were used to control background autofluorescence signals. Flow cytometry data are presented as the percentage of fluorophore- or GFP-positive live cells or as the ratio of M1/M2 percentage of fluorophore-positive live cells in detailed ATM characterization. The gating strategy of the HSC analysis is presented in Supplemental Fig. 2.

Analysis of immune cell populations in tissues by spectral flow cytometry

Epidydimal adipose tissue (eWAT), inguinal adipose tissue (SubCu) and liver tissues were collected. Lymph nodes were excised from SubCu. Tissues were washed with PBS, minced to small pieces with scissors and digested. Adipose tissues were digested during 30 min at 37 °C in phenol red-free DMEM + 0.5% BSA + collagenase D (1 mg/mL; #11088882001, Roche). Upon digestion, cells were sieved through 100-μm cell strainers and spun down at 500 × g, 10 min, 4 °C without breaking. Supernatants were carefully aspirated and the pellet containing (SVF) was incubated with red blood cell lysis buffer (#A1049201, Gibco) buffer for 3 min. 10 ml of cold DMEM was added and cells were spun at 500 × g, 10 min, 4 °C. Cells were then resuspended in PBS and counted to obtain total SVF counts in both fat pads from each mouse. Liver samples were processed according to ref. ⁷⁷. Briefly, minced tissue was digested in phenol red-free DMEM + 0.5% BSA + collagenase D (1 mg/mL; #11088882001 Roche) + DnaseI (5 mg/mL; #10104159001 Roche) for 30 min at 37 °C. Cells were then sieved through 70-µm cell strainers and centrifuged at 50×g for 3 min at 4 °C to initially separate hepatocytes (pellet) and non-parenchymal cells (NPCs) (supernatant). The supernatant containing NPCs was collected and NPC suspension was centrifuged at 163 g for 7 min at 4 °C to pellet the NPCs. NPC pellet was resuspended in red blood cell lysis buffer 1 mL and incubated for 5 min, then washed and centrifuged at 163×g for 7 min at 4 °C to re-pellet the NPCs. In total, 2–3 × 10⁵ cells were incubated with Live-or-Dye 665/685 viability dye (1000× diluted according to manufacturer’s instruction; #32013, Biotium) in PBS for 15 min at 4 °C, in 96-well V-bottom plate. Cells were washed and resuspended in FACS buffer (DPBS, 2% FBS, 2 mM EDTA). TruStain FcX PLUS (2.5 μg/mL; #156604 Biolegend) was added to block non-specific Fc receptor binding, and cells were incubated for 10 min. Antibody cocktail of 1 μg/ml CD45.2—BV750 (#747251 BD Biosiences), 1 μg/ml CD11b—BV605 (#101257 Biolegend), 1 μg/ml CD11c—PE-Cy5 (#117316 Biolegend), 1 μg/ml F4/80—PE-Cy7 (#123114 Biolegend), 1 μg/ml CD3—BV650 (#100229 Biolegend), 1 μg/ml CD4—PE-Dazzle594 (#100566 Biolegend), 1 μg/ml CD8—PE (#100708 Biolegend), 1 μg/ml CD19 -BV711 (#115555 Biolegend), 2 μg/ml NK1.1—BV480 (#746265 BD Biosciences), 0.5 μg/ml Ly-6C—BV570 (#128029 Biolegend), 2.5 μg/ml Ly-6G—FITC (#127606 Biolegend), 1 μg/ml Siglec-F—APC-Cy7 (#565527 BD Biosciences), 2.5 μg/ml FcεRIα—Alexa Fluor® 700 (#134324 Biolegend), 1 μg/ml CD45.1—BV421 (#110732 Biolegend) was added, and cells were incubated for 30 min at 4 °C. Cells were washed 3× with FACS buffer, resuspended in ice-cold PBS and acquired on Cytek Nothern Lights 3-laser spectral cytometer.

An unmixing of the spectral data was performed using SpectroFlo software (Cytek). For unbiased analysis, FCS files were processed using OMIQ analysis platform. Doublets, debris, and dead cells were excluded, and live CD45+ were downsampled to an equal number of cells from each mouse. Data from independent experiments were pooled (15,472 events per condition for eWAT, 8940 events per condition for SubCu, 4148 events per condition for liver), and uniform manifold approximation and projection (UMAP)⁷⁸ was used to visualize the data. FlowSOM algorithm was used to generate filters, which were then manually inspected and adjusted in some cases. The total number of cells in identified filters were calculated from total SVF cell counts and CD45 percentages in each sample (mouse).

F4/80 immunofluorescence staining

For tissue sections, ketamine/xylazine-anesthetized mice were perfused for 10 min with 4% paraformaldehyde before perigonadal fat pads were collected and immersed in the same fixative for 12–15 h at 4 °C. After rinsing and PBS wash, the tissue was dehydrated with gradual steps of EtOH and paraffin-embedded. Adipose tissue was cut into 3–4 μm sections, then slides were deparaffinized, rehydrated, and blocked for endogenous peroxidase activity (1% H₂O₂ in TBST). Following the manufacturer’s recommendations, the slides were stained with F4/80-specific antibodies (1:300 Abcam ab6640). F4/80-positive cells were counted in 15 fields with a ×20 objective. A similar protocol was used for F4/80 immunohistochemistry in 5-μm paraffin sections of brown adipose tissue and muscle, using a 1:100 dilution of the Abcam ab6640 antibody and hematoxylin counterstaining. Adipocyte size was measured from a minimum of 200 adipocytes per mouse using ImageJ software.

Monocyte adhesion and migration

Briefly, 3 × 10⁵ CD14⁺eWAT SVF macrophages were added to fibronectin-coated plates to assess adhesion and incubated for 4 hr at 37 °C before adhered cells were stained crystal violet and absorbance was quantified. Transwell migration assays were performed (Costar polycarbonate filters, 5-μm pore size) as previously described.⁷⁹ Membranes and 12-well plates were coated with fibronectin (5 μl/mL; Life Technologies) overnight at 4 °C. CD14⁺eWAT SVF macrophages (0.3 × 10⁵ cells/well) were added to the upper chamber, and MCP-1 (100 ng/well; Sigma) in 0.8% agarose solution was added to the lower chamber to stimulate migration. Cells migrating into the lower chamber after 8 h of incubation were manually counted and presented as a percentage of cells migrated.

Microarray and bioinformatics analysis

Purified DNA-free RNA from primary recipients’ BM cells was quantified, and quality was assessed using a Nanodrop ND‐100 spectrophotometer and hybridized to Affymetrix GeneChip miRNA 4.0 microarrays by Washington University’s Genome Technology Access Center. The array signal data was processed with Partek Genomics Suite 6.6 (Partek, St Louis, MO). Upregulated genes (ratio >1.19 and nominal P < 0.05) and downregulated genes (ratio <0.94 and nominal P < 0.05) in BM of VD- FL-HSC recipients were used as input for enrichment pathway analysis using Enrichr, Embryonic Stem Cells Atlas of Pluripotency Evidence (ESCAPE), gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Wiki-pathway databases. RNA sequences from bone marrow array data (GSE158763) have been deposited in the NCBI GEO repository.

Methylation analysis

Genomic DNA from bone marrow and 6-month-old eWAT ATM was obtained. Next-generation sequencing methylation analyses were performed by EpigenDx, Inc. (Hopkinton, Massachusetts, United States). 500 ng of extracted DNA samples were bisulfite-modified using the EZ-96 DNA Methylation-Direct KitTM (ZymoResearch; Irvine, CA; cat# D5023) per the manufacturer’s protocol. All bisulfite-modified DNA samples were amplified using separate multiplex or simplex PCRs. PCR products from the same sample were pooled, and libraries were prepared using a custom Library Preparation method created by EpigenDx. Library molecules were purified using Agencourt AMPure XP beads (Beckman Coulter; Brea, CA; cat# A63882). Barcoded samples were then pooled in an equimolar fashion before template preparation and enrichment were performed on the Ion ChefTM system using Ion 520TM & Ion 530TM ExT Chef reagents (Thermo Fisher; Waltham, MA; cat# A30670). Following this, enriched, template-positive library molecules were sequenced on the Ion S5TM sequencer using an Ion 530TM sequencing chip (cat# A27764). FASTQ files from the Ion Torrent S5 server were aligned to the local reference database using open-source Bismark Bisulfite Read Mapper with the Bowtie2 alignment algorithm. Methylation levels were calculated in Bismark by dividing the number of methylated reads by the total number of reads. The % methylation increase was calculated as CpG methylation of recipients of HSC VD(−) – CpG methylation of HSC VD(+) recipients/CpG methylation of HSC VD(+) recipients.

Exosome isolation

Media was centrifuged at 2000×g for 30 min. The supernatant was transferred to a new tube and centrifuged at 100,000×g for 18 h. Density-gradient-based isolation was then performed using the Total Exosome Isolation kit (Invitrogen)⁸⁰. The purity of exosome vesicles was confirmed by western blot showing the presence of EV marker protein expression as previously described^33,80.

MicroRNA and mRNA expression via RT-qPCR

miRNA purification and isolation from 3T3-L1 cells were performed using the mirVana miRNA kit (Invitrogen Ambion) and anti-miR (AM10067). RTq-PCR was performed using the TaqMan reagent kit, with the relative expression of miRNA calculated by the comparative threshold cycle method relative to miR-39 from C. elegans. For media assays, we spiked in miR-39 as an exogenous housekeeping miRNA control prior to extraction (Qiagen #219610). TaqMan primers were obtained from Life Technologies for miRNAs 106b-5p (#000442), 106b-3p (#002380), let7g-5p (#002282), 142-3p (#000464), 330-3p(#001062), 19b-5p (#002425), 125a-5p (#002198), 331-3p (#000545), U6 (#4427975) and 39 (#467942). mRNA RTq-PCR was performed with the GeneAmp 7000 Sequence Detection System using the SYBR® Green reagent kit (Applied Biosystems)³³. We used the following mouse oligonucleotides; for Jarid2 forward 5′-GCGGTAAATGGGCTTCTTGG-3’; Jarid2 reverse 5’- TGCTAGTAGAGGACACTTGGGA-3’; Ppargc1a forward 5′-AGCCTCTTTGCCCAGATCTT-3’; Ppargc1a reverse 5′-GGCAATCCGTCTTCATCCAC-3’, Mef2b forward 5′-GACCGTGTGCTGCTGAAGTA-3; Mef2b reverse 5′-AGCGT CTCGAGGATGTCAGT-3’; Nrf1 Forward 5′-GTACAAGAGCATGATCCTGGA-3’; Nrf1-reverse 5′-GCTCTTCTGTGCGGACATC-3’; Atp5g forward 5’-AGTTGGTGTGGCTGGATCA-3’; Mrpl32 forward, 5’-AAGCGAAACTGGCGGAAAC-3’; Mrpl32 reverse, 5’-GATCTGGCCCTTGAACCTTCT-3’; PDPK1 forward, 5’-CCACTGAGGAAGATCGACAGAC-3’; PDPK1 reverse, TCTGCTCAGCTTCACCGCATTC-3’; PIK3CA forward, 5’-CACCTGAACAGACAAGTAGAGGC-3’; PIK3CA reverse, 5’-GCAAAGCATCCATGAAGTCTGGC-3’; PIK3R1 forward, 5’-CAAACCACCCAAGCCCACTACT-3’; PIK3R1 reverse, 5’- CCATCAGCAGTGTCTCGGAGTT-3’. All assays were done in triplicate, with data expressed as relative expression of mRNA normalized to mouse ribosomal protein Mrpl32.

Plasmids and small interfering RNA transfection

3T3-L1 cells were transfected with mirVana miR-106b (LifeTechnology MC10067), let 7g-5p (LifeTechnology MC11758), 330-3p (LifeTechnology MC10732), 19b-5p (LifeTechnology MC13042), 142-3p (LifeTechnology M C10398),125a-5p (LifeTechnology MC12561), 331-3p (LifeTechnology MC10881) mimics or miR-106b or Let7g antagomirs (AM10067, MH11758 respectively). 72 h after transfection, 3T3-L1 cells were evaluated for 2-DG uptake and mRNA expression in lysates by RT-qPCR. 3T3-L1 cells were also transfected with pre-miR-106b siRNA sense oligonucleotides 5’-CCU AAU GAC CCU CAA GCC GUU-3 and antisense 5’-CGG CUU GAG GGU CAU UAG GUU-3’ then exposed to VD( + ) HSC or VD(-) HSC-recipient peritoneal macrophage media for 72 h before assessment of pre-miR-106-5p or mature miR-106b-5p. Peritoneal macrophages from VD( + ) HSC or VD(-) HSC recipients were transduced with lentivirus containing sense either Jarid2-siRNA (LifeTechnology 4390771), Ppargc1a-siRNA (LifeTechnology, AM16708), or control-siRNA (LifeTechnology AM461) for 48 h and mRNA expression was determined as previously described⁸¹.

Western blot analysis

3T3-L1 adipocytes were homogenized in RIPA lysis buffer containing protease and phosphatase inhibitors. Lysates were clarified, centrifuged, and resolved by SDS-PAGE. Samples were transferred to PVDF membranes that were subsequently probed with the following antibodies for protein and phosphoprotein detection: 9 µg/ml anti-PIK3CA (ab40776, Abcam), 4 µg/ml anti-PI3KR1 (ab191606 Abcam), 3 µg/ml anti-PDPK1 (ab52893, Abcam,) 1:1000 dilution ant-AKT (9272 S, Cell Signaling Technology), 1:500 dilution anti-pAKT (9271L, Cell Signaling Technology) and 1:2000 dilution β-actin (3700S, Cell Signaling Technology). Protein levels were quantified using Image Studio Lite Ver 5.2 (LI-COR) and normalized to β-actin protein levels.

Human samples

Specimens & data were obtained in a de-identified manner from the Women and Infant Specimen Consortium (WIHSC) ICTS/CTSA ULI PR000448 IRB#201013004 consisting of 30 women with singleton pregnancies resulting in full-term vaginal or C-section delivery that fulfill the inclusion criteria were recruited during the day 7 am to 4 pm at Barnes and Jewish hospital by the Women and Infant Specimen Consortium (WIHSC) from 06/26/2017 to 05/15/2018. Sex from singleton was specified in Supplementary Table 4. Racial identity was provided by the Women and Infant Specimen Consortium (WIHSC) based on skin color. Intrauterine growth retardation, small/large for gestational age, diabetes (gestational, type 1, and type 2 DM), preeclampsia, chorioamnionitis, acute infection (fever or active herpes), and moderate or severe alcohol or drug abuse during pregnancy were excluded. Venous cord blood plasma was collected for 25-hydroxyvitamin D levels (LC-MS/MS)^75,76 and quantification of miRNA expression from plasma exosomes^80,82. Cord blood monocytes were isolated as previously described⁸¹. Monocytes were stabilized for 2 h in 100% serum from the original patient to mimic in vivo conditions for experiments^75,76. Isolated monocytes were co-cultured with 3T3-L1 adipocytes in transwell chambers for insulin-stimulated 2-DG studies. Monocyte mRNA was also assessed by qPCR⁸¹.

Statistical analysis

Experiments were carried out in duplicate or triplicate, with “n” referring to the number of distinct samples. Gaussian distribution was verified by Kolmogorov–Smirnov distance. Parametric data are expressed as mean ± SEM and analyzed by two-sided t tests, paired or unpaired as appropriate, or by one-way ANOVA and Tukey’s post-test for more than two groups. Statistical analysis was carried out using GraphPad Prism version 8.4.3.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.