Patients

BM samples from 28 (17 male and 11 female) de novo acquired AA patients with a median age of 31 years (range 18–59 years) were analyzed after the signing of a written informed consent form in accordance with the Declaration of Helsinki. The study was approved by the Committee for Medical Care and Safety, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Science and Peking Union Medical College (ethical approval documents reference number KT2014005-EC-1). This cohort consisted of four patients with severe AA and 24 with nonsevere AA. The diagnosis and severity classification of AA was established by morphological examination of the BM and peripheral blood samples after excluding any other acquired BM failure syndromes, such as paroxysmal nocturnal hemoglobinuria, myelodysplastic syndrome, and congenital BM failure syndromes, according to international criteria [12]. Samples from 19 (14 male and 5 female) age-matched (range 20–56 years) healthy controls were obtained after they had signed the written informed consent form described above.

Animals

Our experimental research on NOD/SCID and nude mice followed internationally recognized guidelines. Ethical approval for the animal experiments was provided by the Ethical Committee of the Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Science and Peking Union Medical College. The ethical approval documents reference number is KT2012003-m-6.

Isolation and identification of BM-MSCs

BM-MSCs were isolated and cultured in Dulbecco’s modified Eagle’s medium [13]. BM-MSCs were identified by their surface markers with a panel of monoclonal antibodies against CD13 (WM15), CD29 (MAR4), CD44 (G44-26), CD49e (IIA1), CD73 (AD2), CD105 (266), CD166 (3A6), CD31 (WM59), CD34 (581), CD45 (HI30), CD90 (5E10), HLA-ABC (G46-2.6), HLA-DR (G46-6), CD14 (M5E2), CD40 (5C3), and CD11b (ICRF44), along with the appropriate isotype monoclonal antibodies using a FACScanflow cytometer (BD Biosciences, San Jose, CA, USA).

Illumina HiSeqTM 2000 sequencing

RNA samples were first treated with DNase I to degrade any possible DNA contamination. Next, mRNA was enriched using oligo (dT) magnetic beads for eukaryotes and fragmented into short fragments of approximately 200 bp. The first strand of cDNA was synthesized using a random hexamer-primer, and then buffer, dNTPs, RNase H, and DNA polymerase I were added to synthesize the second strand. Double-stranded cDNA was purified with magnetic beads, followed by end reparation and 3’-end single nucleotide A (adenine) addition. Finally, sequencing adaptors were ligated to the fragments, which were enriched using polymerase chain reaction (PCR) amplification. A sample library was qualified and quantified using an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA) and an ABI StepOnePlus Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA) during the quantitative-competitive (QC) step. Library products were ready for sequencing via Illumina HiSeqTM 2000 (Illumina, San Diego, CA, USA) or other sequencer when necessary. Next, quantitative real-time polymerase chain reaction (qRT-PCR) was performed to confirm the gene expression levels of RNA transcripts with sequence-specific oligonucleotide primers as described previously.

Separation of CD106+ MSCs and CD106 MSCs

MSCs were labeled with PE-conjugated anti-CD106 antibody (BD Biosciences). CD106+ MSCs and CD106 MSCs were separated using a CD106-positive selection magnetic-activated cell sorting (MACS) isolation kit (Miltenyi Biotech, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. CD106+ MSCs or CD106 MSCs (≥90% purity) were used for subsequent experiments.

Flow cytometry (FCM)

Cells were stained with antibodies along with the appropriate isotype controls (BD Biosciences) according to the manufacturer’s instructions. Data acquisition was performed using an LSR II flow cytometer (BD Biosciences) and analyzed with FlowJo 7.6 software (FlowJo, Ashland, OR, USA).

Hematoxylin and eosin (H&E) and histochemical staining analysis

H&E staining (Sigma-Aldrich) and histochemical staining (Abcam, Cambridge, UK) were performed according to the manufacturers’ instructions. Samples were photographed using a Nikon ElipseTi-U microscope (Nikon, Tokyo, Japan).

Immunofluorescence

The expression of cell surface molecules was assessed according to the manufacturer’s instructions. Normal MSCs (N MSCs) were stained for CD106 and NF-kB. N MSCs were first washed and fixed with 4% formaldehyde for 15 min and then blocked with blocking buffer (phosphate-buffered saline (PBS)/5% normal serum) for 60 min. N MSCs were labeled with mouse-anti-human CD106 overnight and then labeled with Alexa Fluor® 546-goat anti-mouse IgG conjugated secondary antibody for 60 min. N MSCs with 0.3% Triton™ X-100 for 60 min were labeled with rabbit-anti-human NF-κB overnight and then labeled with Alexa Fluor® 488-conjugated secondary antibody (donkey anti-rabbit) for 60 min. The nucleus was marked with DAPI.

Selected CD106+/CD106 MSCs were stained for NF-κB and nuclei to examine the NF-κB expression level. The expression of NF-κB was calculated as the mean of the fluorescence intensity in 6 continuous views. CD106+/CD106 MSCs were first washed and fixed with 4% formaldehyde for 15 min and then blocked with blocking buffer (PBS/5% normal serum/0.3% Triton™ X-100) for 60 min. Then, CD106+/CD106 MSCs were labeled with rabbit-anti-human NF-κB overnight and then labeled with Alexa Fluor® 488-conjugated secondary antibody or Alexa Fluor® 546-conjugated secondary antibody for 60 min. The nucleus was marked with DAPI.

Samples were photographed using an UltraVIEWVoX Confocal Imaging System (PerkinElmer, Waltham, MA, USA).

Western blotting

Western blotting procedures were performed according to the protocol described by Song et al. [14]. Briefly, BM-MSCs were collected, washed, and lysed with RIPA lysis buffer (Beyotime Institute of Biotechnology, Shanghai, China) supplemented with PMSF (Invitrogen, Carlsbad, CA, USA). Total protein was extracted and quantified by the BCA protein assay kit (Pierce, Woodland Hills, CA, USA). A total of 30 μg protein was denatured, separated by SDS-PAGE electrophoresis, and transferred to a PVDF membrane. The transferred membranes were blocked using 5% bovine serum albumin (BSA) in TBST, incubated with anti-human CD106 mouse monoclonal antibody (Abcam) overnight, and then incubated with the corresponding horseradish peroxidase (HRP)-conjugated secondary antibody at a dilution of 1:2000 for 2 h. Bands were visualized using enhanced chemiluminescence (ECL; Thermo-Fisher, Scientific, Waltham, MA, USA) detection reagents, and scanned images were quantified using Image J (https://imagej.nih.gov/ij/). The ratio of target gene to β-actin was used for the semiquantification and comparison between the two groups.

Formation of capillary-like structures

Wells in 96-well plates were covered with 50 μl of growth factor-reduced Matrigel (BD Biosciences). Aliquots of CD106+ MSCs and CD106 MSCs were seeded at a density of 10,000 cells/cm2 and cultured in a humidified atmosphere with 5% CO2 for 24 h. The formation of capillary-like structures was observed using an Olympus IX71 inverted microscope (Olympus, Tokyo, Japan), and pictures were taken at different time points using an Olympus DP71 camera (Olympus).

Matrigel plug assay

To confirm our in vitro data, we examined vasculogenesis in vivo by performing a Matrigel plug assay. Aliquots of 5 × 105 MSCs were resuspended in 500 μl of Matrigel (BD Biosciences) according to the manufacturer’s instructions and implanted into the back of 42-day-old nude mice (n = 6 in each group). Mice implanted with Matrigel only were used as negative controls. After 21 days, the Matrigel plugs were harvested, assayed for microvessels identified as luminal structures with red blood cells using H&E staining, and counted.

Detection of cytokine levels

Supernatants obtained from MSC-conditioned medium were used to detect vascular endothelial growth factor (VEGF) levels using an enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions.

Purification of CD34+ cells

CD34+ cells were freshly purified from umbilical cord blood (UCB) using a CD34/MACS isolation kit (MiltenyiBiotec, Bergisch, Gladbach, Germany) according to the manufacturer’s instructions. Cell fractions with 95 ± 5% CD34+ cell purity was used for subsequent experiments.

Co-culture of CD34+ cells with MSCs

A total of 5 × 104 CD34+ cells suspended in 1 ml of serum-free StemSpan™ H3000 (Stem Cell Technologies, Vancouver, Canada) culture medium was applied to feeder layers composed of 5 × 104 BM-MSCs, as described previously. Cocultures were incubated for 14 days, and the culture medium was replenished every 3.5 days. Nonadherent viable cells were stained for FCM analysis using the antibodies anti-CD34-APC, anti-CD61-PE, anti-CD41a-PE, and anti-CD42b-PE, along with the appropriate isotype controls (BD Biosciences) according to the manufacturer’s instructions. Data acquisition was performed using an LSR II flow cytometer (BD Biosciences) and analyzed with FlowJo7.6 software (FlowJo).

A total of 1 × 105 CD34+ cells suspended in serum-free StemSpan™ H3000 (Stem Cell Technologies) culture medium was applied to feeder layers composed of BM-MSCs, which were blocked by BAY-11-7082, as described previously. Cocultures were incubated for 10 days, and the culture medium was replenished every 3.5 days. Nonadherent viable cells were stained for FCM analysis using CD34-APC along with the appropriate isotype controls (BD Biosciences) according to the manufacturer’s instructions. Data acquisition was also performed using an LSR II flow cytometer (BD Biosciences) and analyzed with FlowJo7.6 software (FlowJo).

Assay of colony forming units (CFUs)

Cocultures of CD34+ cells with MSCs or blocked MSCs were incubated for 14 days or 10 days. Nonadherent viable cells (2.5 × 102 or 5 × 102 in each well of a 24-well-plate) were plated on 0.5 ml of methylcellulose medium (Stem Cell Technologies) to evaluate the in vitro effects of MSCs on CFU growth. Colonies of >50 cells were scored after 14 days of culture. Experiments were performed in triplicate.

CFU-megakaryocyte (CFU-MK) assay

Cocultures of CD34+ cells with MSCs were incubated for 14 days. Nonadherent viable cells (5 × 102 in each well of a 24-well-plate) were plated on semisolid Iscove’s modified Dulbecco’s medium (IMDM; Gibco) supplemented with 1% methylcellulose, 10% fetal bovine serum (FBS), 1% BSA, 10–4 M mercaptoethanol, 2 mM l-glutamine, and 100 ng/ml thrombopoietin. Cultures were incubated at 37 °C in a humidified atmosphere with 5% CO2. CFU-MK colonies were identified after 14 days of culture under an Olympus IX71 inverted microscope (Olympus), and typical colonies were selected for immunostaining. Cell smears were prepared using Cytospin (Thermo-Fisher Scientific), stained with anti-CD41a antibody, and observed using an UltraVIEWVoX Confocal Imaging System (PerkinElmer).

Co-transplantation of CD34+ cells and MSCs in NOD/SCID mice

Aliquots of cell preparations containing 2 × 105 CD106+ MSCs or CD106 MSCs and 1 × 105 UCB CD34+ cells in 15 μl of Roswell Park Memorial Institute (RPMI) 1640 medium were injected into the tail vein of 28- to 35-day-old NOD/SCID irradiated mice (n = 6 for each pair of cells) using a Hamilton syringe. Some low-dose irradiated mice (320 cGy) were sacrificed 42 days after xenotransplantation. CD45+ cells were analyzed after staining with human anti-CD45-APC antibody along with the appropriate isotype controls (BD Biosciences) according to the manufacturer’s instructions. Some high-dose irradiated mice (360 cGy) (n = 6 for each pair of cells) were observed until 56 days after xenotransplantation or death.

Blockade assay

BM-MSCs were blocked by incubation with 500 ng/ml of the CD106 blocking antibody ab47159 for 1 h at 4 °C. BM-MSCs were blocked by incubation with 10 nM of the NF-κB-specific inhibitor BAY-11-7082 for 0.5 h, 1 h, or 12 h.

NanoPro analysis for NF-κB

Cells were lysed in Bicine/CHAPS Lysis Buffer (ProteinSimple) supplemented with DMSO Inhibitor Mix (ProteinSimple) and Aqueous Inhibitor Mix (ProteinSimple) at 4 °C for 30 min, and the lysate was mixed with Premix G2 (pH 3–10) (ProteinSimple) and pI Standard Ladder 3 (44:1, ProteinSimple). The rabbit anti-NF-κB (p65) antibody (primary antibody, Cell Signaling Technology) was diluted 1:50 in antibody dilution buffer, and the anti-human IgG-HRP (secondary antibody, Protein Simple) was diluted 1:100 in antibody dilution buffer. Luminol/peroxide was mixed at a 1:1 ratio. The NanoPro 1000 (ProteinSimple) was loaded and run according to the manufacturer’s specifications. Emitted light was quantified for 30 s, 60 s, 120 s, 240 s, 480 s, and 960 s. Compass software 2.5.11 (ProteinSimple) was used to identify and quantify chemiluminescence peaks and optimize tracings.

TNF-α-stimulated MSCs and LPS-stimulated MSCs

CD106+ MSCs or CD106 MSCs at passage 4 were stimulated by incubation with TNF-α (PepTech, Burlington, MA, USA) 10 ng/ml for 1 h, 4 h, or 24 h, or 500 ng/ml LPS (Sigma-Aldrich) for 0.5 h, 1 h, 2 h, or 4 h.

Statistical analysis

Analysis of variance in conjunction with Student’s t test was performed to identify significant differences. All analyses were performed using GraphPad Prism 6.0 software (GraphPad, La Jolla, CA, USA).