A miRNA/CXCR4 signaling axis impairs monopoiesis and angiogenesis in diabetic critical limb ischemia

Patients with peripheral artery disease (PAD) and diabetes have the highest risk of critical limb ischemia (CLI) and amputation, yet the underlying mechanisms remain incompletely understood. MicroRNA (miRNA) sequencing of plasma from diabetic patients with or without CLI was compared to diabetic mice with acute or subacute limb ischemia to identify conserved miRNAs. miRNA-KO mice on high-fat diet were generated to explore the impact on CLI. Comparison of dysregulated miRNAs from diabetic individuals with PAD and diabetic mice with limb ischemia revealed conserved miR-181 family members. High-fat–fed, diabetic Mir181a2b2-KO mice had impaired revascularization in limbs due to abrogation of circulating Ly6Chi monocytes, with reduced accumulation in ischemic skeletal muscles. M2-like KO macrophages under diabetic conditions failed to produce proangiogenic cytokines. Single-cell transcriptomics of the bone marrow niche revealed that the reduced monocytosis in diabetic KO mice was a result of impaired hematopoiesis, with increased CXCR4 signaling in bone marrow Lineage–Sca1+Kit+ (LSK) cells. Exogenous Ly6Chi monocytes from nondiabetic KO mice rescued the impaired revascularization in ischemic limbs of diabetic KO mice. Increased Cxcr4 expression was mediated by the miR-181 target, Plac8. Taken together, our results show that MiR-181a/b is a putative mediator of diabetic CLI and contributes to changes in hematopoiesis, monocytosis, and macrophage polarization.

Gys) followed by injection of bone marrow donor cells (4 x 10 6 cells) by tail vein injection, followed by recovery for 6 weeks.
In some experiments, mice were placed on a high-fat sucrose-containing (HFSC) diet consisting of 58 kcal% fat and 28 kcal% carbohydrates for 4 weeks prior to surgery (D09071704, Research Diets). Insulin tolerance testing (ITT) and glucose tolerance testing (GTT) were performed after 4 weeks of HFSC diet feeding. Briefly, for GTT, mice were fasted for 12 h and then injected i.p. with D-glucose (G7201, Sigma, 1 g/kg).
ITT was performed on mice after 6 h fasting and injected i.p. with recombinant human regular insulin (0.75 U/kg). Blood glucose levels were measured before injection and at 15, 30, 60, 90, and 120 min after glucose or insulin injection using glucometer.
The animals will be sacrificed at the end of each experimental point as described above. If an animal appears to be sick or suffering, it will be euthanized by CO2 asphyxiation. These methods are consistent with recommendations from the panel on Euthanasia of the American Veterinary Medical Association.

Hindlimb ischemia mouse models
Mice were subjected to two different surgeries to replicate critical limb ischemia: 1) femoral artery ligation, which causes immediate cessation of blood flow hence (Acute ischemia); and 2) ameroid constrictors which gradually expands from fluid adsorption inducing artery occlusion (Sub-acute ischemia). Briefly, mice were injected i.p. with 150 μl of 20% ketamine/5% xylazine in 0.9% saline. Once anesthetized, the right medial thigh to the suprapubic area was treated with a commercial emollient to remove fur and sterilized with Povidone iodine. Skin and fascia were dissected away to the femoral bed. In the acute hindlimb ischemia model, femoral artery and surrounding tissue was proximally and distally ligated with 7-0 Prolene sutures. The arterial bed in between sutures was cauterized. Abrogation of blood flow compared to the contralateral limb (<10%) was confirmed using a laser Doppler imager (Moor Instruments, UK). Sub-acute hindlimb ischemia model was performed using ameroid constrictors, which induce gradual femoral artery occlusion over 1-3 days. Two ameroid constrictors were placed on the femoral artery, one proximal to the lateral circumflex femoral artery and the second proximal to the bifurcation of the popliteal and saphenous arteries. Both constrictors were positioned with the slot facing up, ensuring proper setting of the artery within the constrictor. Mice were sutured closed at the level of the fascia and subsequently, the skin. Sham-treated mice were treated the same way except once the femoral artery was visualized, the incision was closed without ligation of the femoral artery or ameroid constrictor addition. Percent blood flow recovery was calculated by comparing a ratio of ischemic paw to contralateral paw Doppler count profiles and normalized blood flow recovery was calculated by comparing the ratio of ischemic to contralateral paw Doppler count profiles to day 0 post-operative percent blood flow.

Endothelial Cell Isolation
Gastrocnemius muscles or lungs were grinded with scissors and digested by using 1 mg/ml Collagenase type 2 (Worthington Biochemical LS004177) and 1 mg/ml Dispase II (Roche, 04942078001) and incubated at 37°C for 40 minutes. Digestion was neutralized with DMEM/F12 medium containing 10% FBS, followed by centrifugation at 500g for 10 min at 4℃. The slurry was passed through cell strainers (Corning Falcon/Westnet). After centrifugation, the cell pellet was re-suspended in incubation buffer (PBS pH 7.2, 0.1% BSA, 2mM EDTA, 0.5% FBS). Endothelial cells were captured using magnetic Dynabeads (sheep antirat IgG, Invitrogen, 00412289) conjugated with rat anti-mouse CD31 antibody (BD Biosciences, 557355) at a ratio of 5:1 Dynabeads/antibody and allowed to tumble at 4°C for 20 minutes. The slurry of lysate and Dynabead/antibody mixture was bound on a Dynamag-2 Magnet (Invitrogen) for 1 minute and the supernatant was collected as a non-endothelial cell fraction. The beads containing bound endothelial cells were then washed on the Dynamag-2 Magnet five times using wash buffer (PBS pH 7.2, 0.1% BSA) and the resultant pellet was collected as an endothelial cell fraction.

Plasma miRNA sequencing
The EdgeSeq miRNA Whole Transcriptome Assay from HTG Molecular Diagnostics, Inc. (AZ, USA) was used to measure miRNA expression in plasma from human donors and mice. The HTG EdgeSeq system combines quantitative nuclease protection assay chemistry with a next-generation sequencing platform to enable the semi-quantitative analysis of 2,083 human miRNA transcripts in a single assay. Fifteen microliters of plasma were used for extraction-free sample processing and quantitative nuclease protection assay using the EdgeSeq processor (HTG Molecular Diagnostics, Inc.). The libraries were sequenced using Illumina NextSeq, and data were parsed through HTG EdgeSeq before count data were assessed for quality and analyzed using R.

Bulk RNA-Seq analysis
RNA-Seq analysis was performed after ribodepletion and standard library construction using Illumina HiSeq2500 V4 2x150 PE (Genewiz). All samples were processed using

Cell Culture and Transfection
Human umbilical vein endothelial cells (HUVECs; Lonza) were cultured in endothelial cell growth medium EGM-2 (Lonza, CC-3162). Cells that were utilized for experiments were passaged no more than six times. Stimulation to M2-like macrophages with 20 ng/ml interleukin-4 for 48 hours.

RNA Isolation and real-time quantitative PCR
Total RNA was extracted by using Trizol reagent following the manufacturer's protocol (Invitrogen, 15596-026). The concentration and quality control of RNA was examined using NanoDrop 2000 (ThermoFisher). Isolated RNA was reverse transcribed using miScript reverse transcription kit from Qiagen (218061)

Cytokine analysis
Mouse plasma and supernatant from cultured BMDM were collected and subjected to ELISA using Mouse Cytokine Discovery Assay (Eve Technologies).

Mononuclear cell preparation and flow cytometry
Peripheral blood, bone marrow and gastrocnemius muscle were used for characterization of leukocyte cell populations by flow cytometry. Peripheral blood was drawn from the right ventricle to EDTA contained tube, and cells were filtering through 70μm strainers after lysis of erythrocytes by using RBC Lysis Buffer (eBioscience, 00-4333-57). For bone marrow, one femur was crushed with a mortar and pestle and homogenized by passing through a 40μm strainer. For gastrocnemius muscle, after being washed with pre-cold 1x PBS, tissue was placed in the digestion buffer (1 mg/ml collagenase type 2) (Worthington, LS004177), 1 mg/ml dispase II (Roche, 04942078001) and minced with scissors. Then, tissues in digestion solution were incubated at 37°C for 30 min at a speed of 200 rpm shaking. The solution was passed through 70μm strainers and then centrifuged at 500xg for 10 minutes at 4°C. After that, samples were resuspended to obtain single cell suspensions for next step.
After preparing the single cell suspension according to the above methods, the samples from murine PBMC, bone marrow, and gastrocnemius sequentially filtered The antibodies for flow cytometry were attached in table S4. All the flow data analysis were analyzed by FlowJo 10.7.1.

Ly6C hi monocyte purification and infusion
After collecting bone marrow by crushing the femur with a mortar and passing through 40μm strainers, monocytes from miR-181 KO and WT mice were enriched by using monocyte isolation kit (BM) (miltenyibiotec, 130-100-629). Purified monocytes were stained with anti-CD45, anti-CD115 and anti-Ly6C antibodies for sorting. Live CD45 + CD115 + Ly6C hi cells were sorted by FACS Sorter Aria II.
One day post-ligation, 3x10 5 sorted Ly6C hi monocytes from WT and KO mice were injected by tail vein to HFSC diet-fed KO mice. After 3, 7 and 14 days, blood flow was detected by using a laser Doppler imager and the tissue samples were collected to analyze further after 14 days.

Single-Cell RNA Sequencing
All

Single-Cell RNA Sequencing Data Analysis
The sequenced data were processed into expression matrices with the Cell Ranger Single-cell software 6.1 (https://support.10xgenomics.com/single-cell-geneexpression/software/pipelines/latest/using/count) on the Harvard Medical School O2 cluster. Raw base-call files from HiSeq4000 sequencer were demultiplexed to first generate FASTQ files using the cellranger mkfastq pipeline. Subsequently, the reads were aligned to the mouse transcriptome (mm 10-3.0.0), cell barcodes and unique molecular identifiers (UMI) were filtered and corrected using the cellranger count pipeline. The final output filtered expression matrices were imported into the Seurat package in R and built into Seurat objects using the CreateSeuratObject function.
DoubletFinder was used for removing potential doublets in the single-cell data.
Filtering during this step included only genes detected in >3 cells, cells with >500 distinct genes and >500 UMI. Cells with >10% mitochondrial percentage were excluded. Cell types was identified based on gene markers for each type of cells (Table 2). Data normalization, scaling, and regression by mitochondrial content were then performed using the SCTransform command under default settings in Seurat.

Principal component analysis and nonlinear dimensional reduction using Uniform
Manifold Approximation and Projection (UMAP) was performed. Cell clustering was then assessed across a range of predetermined resolution scales to ensure separation of known major BM cell types without excessively sub-clustering. The enriched cell marker for each cluster of BM cell were shown in UMAP plots ( fig. S6, table S5). The FindAllMarkers function in Seurat was applied to performs parallel differential expression testing of all cells within a cluster versus all other cells in the data set via nonparametric Wilcoxon rank-sum test using default parameters.

Data availability
All relevant data are available from the authors. The RNA-seq data are accessible at: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE196566. Source data are provided with this paper.

Supplemental Figures
Supplemental Figure 1. Endothelial miR-181b kinetics and regulation of proliferation and apoptosis.
A. Expression of miR-181b normalized to U6 in a variety of mouse (3T3 embryonic fibroblast; BMDM macrophages; VSMC vascular smooth muscle cell; SM EC skeletal muscle endothelial cell; B.End3 brain EC) and human (THP1 monocytes; LHMVEC lung microvascular EC; HUVEC umbilical vein EC; CASMC coronary artery smooth muscle cell; DHMVEC dermal microvascular EC) cell lines. HUVECs transfected with 100nM NS-i or miR-181-i and treated with ± high-glucose and ± hypoxia were subjected to B. apoptosis assay (n=6) and C. BrDU proliferation assay (n=6). D.