All methods were carried out in accordance with relevant guidelines and regulations, and are reported in accordance with ARRIVE guidelines.
Patient samples and data
Clinical outcome- and flow cytometry data used for the retrospective correlation analyses herein, were obtained in a previously described clinical phase 1 safety study for treating erectile dysfunction (ED)2,3. Clinical efficacy was addressed by evaluation of IIEF-5 and EHS scores, and the ADRC phenotype assessed by flow cytometry using markers for CD31, CD34, CD73 and CD903.
Lipoaspirates for ADRC isolation (n = 10) were obtained from placebo patients enrolled in two separate phase 2 RCTs using ADRCs for cell therapy of ED (unpublished; awaiting final data analyses) and lymphedema (submitted). All patients gave written informed consent before participation.
Regulatory approvals
The ED-trials followed ATMP guidelines and were approved by The Danish Health and Medicines Authority [EUdra-CT nos. 2013–004220–11 (Phase 1) and 2015–005140-33 (Phase 2)], the Danish National Ethics Committee [nos. 37054 (Phase 1) and 51,658 (Phase 2)] as well as the Danish Data Protection Agency [nos. 2008–58–0035 (Phase 1 and 16/2816 (Phase 2)]. The Phase 1 ED study was registered at ClinicalTrials.gov (NCT02240823). All studies were performed in accordance with the Declaration of Helsinki and ICH-GCP guidelines.
Animals
8-week-old C57BL6 mice (male and female) were obtained from Janvier Labs (Le Genest-Saint-Isle, France), and kept in an animal facility, with controlled temperatures, a 12-h light/dark cycle. For collection of penile and aortic tissue, animals were sacrificed with CO2. All experimental protocols were approved by The Regional Committees on Health Research Ethics for Southern Denmark.
Isolation of adipose derived regenerative cells (ADRC)
ADRCs were isolated from human lipoaspirates as described in detail previously3,72.
Magnetic cell sorting enrichment of CD31+ ADRCs
Enrichment of CD31 positive and -negative ADRC subpopulations was accomplished by magnetic cell sorting (MACS) using MS separation columns (Miltenyi Biotech) under RNase-free conditions. Initially, the isolated ADRC was subjected to red cell lysis (RBC buffer, Miltenyi Biotech, cat no.130-094-183). Magnetic labelling was performed using a CD31 microbead kit (Miltenyi Biotech) after which cells were sorted on an OctoMACS™ Separator (Miltenyi Biotech). Hereby, CD31+ and CD31− ADRCs were obtained based on a positive or negative selection for the CD31 marker.
Flow cytometry
CD31-enrichment and -depletion of ADRCs was verified by flow cytometry using an anti-CD31 antibody (CD31-Vioblue antibody, Miltenyi Biotech cat. no. 130-106-503). Sample acquisition was performed on a BD™ LSRII flow cytometer and analysed using the FACSDiva™ software v8.0.1 and FlowJo v10.
ADRC flow cytometry data from the phase 1 ED clinical trial employed herein, were originally analysed using antibodies against CD34 (PECF594, clone 581), CD31 (Alexa Fluor® 647, clone WM59), CD73 (APC, clone AD2), CD90 (APC, clone 5E10), and appropriate isotype controls.
Conditioned media
ADRCs, CD31+ and CD31− ADRCs were cultured in endothelial cell basal medium (PromoCell, cat no. C-22210) with 1% penicillin/streptomycin (PS) (= EBM) supplemented with 2.5% fetal bovine serum (FBS) for 15 days without media exchange. Conditioned media (CM) was centrifuged at 2500×g for 15 min. After media collection, RNA was isolated from the ADRC, CD31+ and CD31− ADRCs as described below.
Corpus cavernosum explant co-culture assay
Mouse corpus cavernosum was prepared according to Ghatak et al.73. Briefly, following termination, the penile tissue from 6 to 8-week-old C57BL6 mice (Janvier, France) was dissected. For the ex vivo assay, the tissue was cut into three/four 1 mm pieces, and the explants were plated on growth factor reduced Matrigel (Corning, cat no. 356231). Following polymerization, explants were supplemented with 1 ml EBM and 2.5% FBS and incubated at 37 °C with 5% CO248.
After 24 h, media was replaced with 1.5 ml fresh EBM with 2.5% FBS. Next, Nunc™ polycarbonate Cell Culture Inserts with 0.4 µm pore size (Thermo Scientific catalog no.140620) were placed above the explant, and 20,000 cells (ADRCs, CD31− or CD31+) were seeded in 500 µl EBM with 2.5% FBS (3 wells pr. cell type). The plate was then incubated at 37 °C with 5% CO2 for 15 days without medium exchange.
Structurally distinct regions of tubular sprouting from corpus cavernosum explants were comparable to the regions previously described for the aortic ring assay37. The tubular network was quantified in defined, fixed areas (see Supplementary Fig. 2c) using the ImageJ software (Fiji) with an angiogenesis plugin. The angiogenesis evaluation was based on number of nodes, total length of tubes, number of meshes/mm2 and mesh coverage.
Aortic ring ex vivo assay
The aortic ring assay was set up as previously described74. Briefly, the thoracic aorta was dissected from 8 to 12 weeks old C57BL6 mice into 0.5–0.7 mm wide rings, and serum-starved in EBM for 24 h. For indirect co-culture assay, one aortic ring per well of a 24-well Nunc™ Carrier Plate was embedded between two 40 µl drops of Matrigel with 700 µl EBM containing 2.5% FBS. Next, Nunc™ polycarbonate Cell Culture Inserts with 0.4 µm pore size were placed above the aortic rings and 50,000 cells per insert (ADRCs, CD31+ or CD31− ADRCs) were seeded in 500 µl EBM with 2.5% FBS (8 wells pr. cell type). To test CM effects, Matrigel embedded aortic rings were cultured in 500 µl CM from ADRC’s, CD31+ or CD31− ADRCs. The plate was cultured at 37 °C for 8 days, at which time, pictures were acquired with phase contrast microscopy (5 × magnification) and analyzed using the angiogenesis plug-in in ImageJ.
RNA from the ADRCs, CD31− and CD31+ ADRCs cultured in the inserts was collected using Tri Reagent® (ThermoFisher Scientific, cat no. AM9738).
Stable isotope labeling of protein samples with TMT-10 plex
Proteins from CM were isolated by transferring the supernatant to five equivalents of ice-cold acetone. Proteins were reduced using 5 mM dithiothreitol (DTT), followed by 15 mM iodoacetamide blocking before trypsination overnight at 37 °C at a protein:trypsin (Promega, Madison, WI, USA) ratio of 50:1 w/w. 10 µg of the tryptic digest was labeled with a 10-plex TMT-kit (Thermo Scientific), resuspended in anhydrous ethanol, and a 40 µg sample was labeled according to the scheme in Supplementary Table 1. Labeled samples were pooled in equal ratios, dried in a vacuum centrifuge, re-dissolved in 50 µl trifluoroacetic acid solution (0.1%), purified, loaded on a reverse phase microcolumn (equal w/w amounts of Poros R2 and Oligo R3 material) and fractionated by high pH liquid chromatography as described75.
The Fractions were analyzed by RP‐nanoLC‐MS/MS on an Orbitrap Eclipse mass spectrometer (Thermo Fisher Scientific) equipped with a nano HPLC interface (Dionex UltiMate 3000 nano HPLC) as described75. Raw data files were quantified using Proteome Discoverer version 2.4 (Thermo Scientific) as previously described using human and bovine database searches76.
Isolation of primary murine cavernous pericytes
Penile tissue from mice was prepared and embedded in Matrigel as described for the Corpus Cavernosum explant assay. The Matrigel drop and the tissue was plated in a 60 mm petri-dish. After Matrigel polymerization, the petridish was supplemented with EBM containing 20% FBS. The medium was changed every 4 days. Cells sprouted from the corpus cavernosum explants (4 per dish), and after approximately 2 weeks, they became confluent. Only cells migrating out of the Matrigel onto the plastic surface of the petri-dish, were further sub-cultivated. These migrated pericyte cells were trypsinized and seeded at 10,000 cells/cm2 density in EBM with 20% FBS for further experiments.
Migration in vitro assays (with conditioned media)
To test the angiogenic effect of conditioned medium from CD31+ ADRCs, we adopted the wound healing assay using pericyte migration. Primary pericytes were seeded in culture inserts (ibidi culture-insert 2 well, ibidi GmbH, Martinsried, Germany) at a density of 25,000 cells per well. After allowing cells to attach overnight, we removed the culture inserts creating a cell-free gap and washed the cells with sterile PBS to remove non-adherent cells. We then provided 300 μl of CM from ADRC, CD31+ , or CD31− ADRCs, or EBM with 2.5% FBS as a control. Images of cell-free gaps were taken immediately after removing inserts with a bright field microscope at 5 × magnification. We monitored the gap for 24 h after culturing cells in respective CM, at which point images of the gaps were captured again. The cells migration ability was evaluated by the area of the gap they had covered in 24 h using ImageJ.
Immunofluorescence staining
Cultured pericytes and pericytes sprouting from corpus cavernous explant in Matrigel, were fixed with 4% neutral buffered formalin (NBF) for 15 min at room temperature and permeabilized with 0.1% Triton X-100/0.1% Na-Citrate/PBS for 10 min on ice. Cells were blocked with 2% BSA in PBS and stained with rabbit anti-NG2 (Millipore #AB5320, 1:200), mouse anti-PDGFRß (Novus biologicals#NBP1-43349, 1:200), rabbit anti-αSMA (abcam#GR195159, 1:100) or mouse anti-CD31 (Novus Biologicals#NB600-562, 1:200) at 4 °C overnight. Following a 90 min. incubation with donkey secondary antibodies (Alexa 488 or Alexa 555 labelled, Invitrogen. 1:200), the nuclei were stained with DAPI (VWR, cat no. 172867). Finally, images were captured at a 20 × magnification using a Leica DMI4000B instrument with a Leica DFC340 FX Digital Camera.
RNA purification and RT-qPCR
RNA was extracted from samples of ADRCs, CD31+ and CD31− ADRCs, which were obtained from the aortic ring co-culture assay, the CM production set-up, and from the in vivo Matrigel cell culture. The samples were homogenized, and extraction of the total RNA was performed using the Tri Reagent® protocol (ThermoFisher Scientific). The RNA quantity and purity was assessed by nanodrop measurements (Nanodrop® Technologies). The mRNA was reversed transcribed using a High-Capacity cDNA kit (Applied Biosystems, Thermo Fisher Scientific) and the RT-qPCR reaction was performed using Power SYBR® Green PCR kit (Applied Biosystems). Primers specific for the following target genes were used: PECAM1 (encoding CD31), DKK3 (Dickkopf 3), ANGPT2 (Angiopoietin 2), ANXA2 (Annexin 2) and VIM (Vimentin) (Integrated DNA technologies, USA) (Supplementary Table 2). The RT-qPCR analysis was performed on a QuantStudio 7 instrument (Applied Biosystems). The data from the RT-qPCR was analyzed using the qBase + software (Biogazelle, Belgium) and normalized against multiple housekeeping genes, chosen based on the geNorm analysis in qBase+.
Single-cell RNA sequencing
To enable single-cell RNA sequencing (scRNA-seq), at least 106 CD31+ ADRCs from each of four subjects (one male and three females) were methanol-fixed and stored at – 80 °C essentially as previously described77. Briefly, the cells were resuspended in 500 µl PBS with 1% BSA and 1 U/µl RNAsin PLUS RNase Inhibitor (Promega, Cat.no. N2615) and passed through a 40 µm Flowmi® Cell strainer (VWR, cat.no. 734-5950) to create a single-cell suspension. The cells were fixed for 30 min at – 20 °C and subsequently stored at – 80 °C until use.
Before sequencing, fixed cells were thawed and rehydrated (3 × saline sodium citrate, 0.04% BSA, 1 U/µl RNAsin Plus, 40 mM Dithiothreitol). Immediately following rehydration, 8000 cells were loaded onto the 10 × Genomics Chromium controller (10 × Genomics, PN110203). Libraries were prepared according to the instructions of the manufacturer using the 10 × Genomics Single-Cell 3′ v3, Chromium Single Cell B Chip Kit, 48 runs (10 × Genomics, 10 × Genomics, PN-1000073) and sequenced on an Illumina NovaSeq 6000 System (10 × Genomics, 20012850). Cell samples were kept for RNA purification, and RIN-values were measured using RNA 6000 Nano kit (Agilent, cat no. 5067-1512). The RIN-values were 7.25, 6.80, 5.00, and 7.53 for the four single-cell samples Pt. 1–4, respectively.
scRNA-seq data analysis
Using Cell Ranger software (v. 3.1.0), Illumina raw sequencing output files were demultiplexed and aligned to the GRCh38 human reference genome and human transcriptome reference (ENSEMBL version 102) to create a single-cell feature count matrix for each sample. The output metrics generated by Cell Ranger are listed in Supplementary Table 5. Overall, the four patient samples (Pt. 1–4) had an estimated total cell number of 25,928 and were sequenced with a coverage of around 220 million total reads per sample (250,368,339, 210,296,540, 223,244,205, and 195,915,808 for Pt. 1–4, respectively) corresponding to a mean of around 34,000 reads per cell (specifically 30,663, 21,555, 49,205, and 56,459 for Pt. 1–4, respectively) (Supplementary Table 5). All subsequent analyses were performed using R Version 4.1.0 and the Seurat R software package (Version 4.0.2). First, the data was converted to single-cell Seurat objects retaining genes expressed in at least five cells, each dataset was filtered, excluding low-quality cells with mitochondrial RNA contents (percent.mt) over 10%, or number of genes detected per cell (nFeature_RNA) below 200. We also excluded cells with nFeature_RNA above an upper threshold, which was set individually for each of the four samples: 4600, 4700, 4650, and 4450 for Pt. 1–4, respectively. A total of 24,403 cells passed these quality filtering steps. Next, each dataset was subjected to normalization with a scale factor of 10,000 using the “LogNormalize” method, identification of the top 2000 most variable genes using the “FindVariableFeatures”-function, and scaling with the following variables regressed: “nFeature_RNA” and “percent.mt”. Next, the 4 datasets were integrated using the “FindIntegrationAnchors” and “IntegrateData” commands in Seurat78. The integrated dataset was scaled, and principal components were identified. After doing a Jackstraw plot, in which a p-value is assigned to each principal component, uniform manifold approximation and projection (UMAP) was run and cell “neighbors” were identified (both with the reduction dimensions set to 54), and finally cells were clustered and visualized with the resolution parameter set to 1.
Differential gene expression analysis was performed for each individual cluster using the “FindMarkers”-function in Seurat, which compares the gene expression levels in a specific cluster with the corresponding genes in the cells of all other clusters using a Wilcoxon Rank Sum test with an adjusted p-value cutoff of 0.05 based on Bonferroni correction. All significantly differentially expressed genes are listed in Supplementary Table 7. Subsequently, the differentially expressed genes for each cluster were analyzed for significantly enriched gene ontology (GO) terms using Database for Annotation, Visualization, and Integrated Discovery (DAVID version 6.8; https://david.ncifcrf.gov/). All significantly enriched GO terms are listed in Supplementary Table 8.
Statistics
Quantitative data are expressed as mean ± standard deviation (SD) and statistical analyses were performed using the GraphPad Prism software version 9. Statistical significance was set at p < 0.05. and in each case determined using the appropriate statistical method following normality testing, as denoted in the corresponding figure legend.