1
STEM CELLS AND DEVELOPMENTVolume 19, Number 10, 2010© Mary Ann Liebert, Inc.DOI: 10.1089/scd.2010.0070
In addition to hematopoietic stem cells, cord blood (CB) also contains different nonhematopoietic CD45 − , CD34 − adherent cell populations: cord blood mesenchymal stromal cells (CB MSC) that behave almost like MSC from bone marrow (BM MSC) and unrestricted somatic stem cells (USSC) that differentiate into cells of all 3 germ layers. Distinguishing between these populations is diffi cult due to overlapping features such as the immuno-phenotype or the osteogenic and chondrogenic differentiation pathway. Functional differences in the differen-tiation potential suggest different developmental stages or different cell populations. Here we demonstrate that the expression of genes and the differentiation toward the adipogenic lineage can discriminate between these 2 populations. USSC, including clonal-derived cells lacking adipogenic differentiation, strongly expressed δ-like 1/preadipocyte factor 1 (DLK-1/PREF1) correlating with high proliferative potential, while CB MSC were charac-terized by a strong differentiation toward adipocytes correlating with a weak or negative DLK-1/PREF1 expres-sion. Constitutive overexpression of DLK-1/PREF1 in CB MSC resulted in a reduced adipogenic differentiation, whereas silencing of DLK-1 in USSC resulted in adipogenic differentiation.
Introduction
Over the last years, the clinical use of allogenic cord blood (CB) for hematopoietic stem cell transplanta-
tion has increased dramatically. In comparison with other stem cell sources, well-characterized CB grafts are immedi-ately available in numerous CB banks worldwide [ 1–3 ]. The presence of primitive nonhematopoietic stem/progenitor cells in CB was reported by our group [ 4 ] and confi rmed by others [ 5–8 ]. Although cells derived from CB possess sev-eral overlapping features with mesenchymal stromal cells (MSCs) derived from bone marrow (BM), such as immuno-phenotype, osteogenic, and chondrogenic in vitro and in vivo differentiation potential [ 9 ]; they differ from BM MSC with regard to their immunological behavior [ 10 , 11 ], their transcriptome [ 12 ], and their neural differentiation potential [ 4 , 13 ]. In 2004, we were able to show that adipogenic differ-entiation was observed in CB-derived cells [ 4 ]. Since then, we characterized a large number of CB-derived cell lines and respective clones assessing their adipogenic, chondro-genic, and osteogenic differentiation potential. We were able
to characterize unrestricted somatic stem cells (USSC) from CB that have unique proliferation capacities and can be dif-ferentiated in vitro into the mesodermal, endodermal, and ectodermal lineages [ 4 , 13–16 ]. USSC provide a supportive cell layer for hematopoietic cells [ 5 , 17 ], a function they share with BM MSC. CB-derived cells that exhibit a restricted dif-ferentiation potential toward the neural lineage [ 18 ] and a high adipogenic differentiation potential were identifi ed as MSC/progenitor cells [ 6 ]. Markers capable of distinguishing between USSC and MSC in CB are still unknown. Therefore, distinguishing unrestricted from MSC determined cells in CB solely relies on determining their functional differenti-ation potential. An important established player for adipo-genesis is the δ-like 1/preadipocyte factor 1 (DLK-1/PREF-1) protein, a member of the epidermal growth factor (EGF)-like family that is homologous to members of the Notch/Delta/Serrate family, which however lacks the characteristic DSL motif [ 19 , 20 ]. DLK-1 exists in 2 forms, as transmembrane and secreted proteins [ 21 ], and is expressed in multiple embry-onic tissues [ 22 ]. After birth, DLK-1 is down-regulated in
DLK-1 as a Marker to Distinguish Unrestricted Somatic Stem Cells and Mesenchymal Stromal Cells in Cord Blood
Simone Maria Kluth , 1 Anja Buchheiser, 1 Amelie Pia Houben , 1 Stefanie Geyh , 1 Thomas Krenz, 1 Teja Falk Radke , 1 Constanze Wiek , 2 Helmut Hanenberg, 2,3 Petra Reinecke, 4 Peter Wernet , 1 and Gesine Kögler 1
1 Institute for Transplantation Diagnostics and Cell Therapeutics , 2 Department of Pediatric Hematology, Oncology and Clinical Immunology, Children’s Hospital , and 4 Institute of Pathology, Heinrich Heine University Medical Center, Duesseldorf, Germany.
3 Department of Pediatrics, Wells Center for Pediatric Research, Riley Hospital for Children, Indiana University School of Medicine, Indianapolis, Indiana.
KLUTH ET AL. 2
(DMEM) low glucose (Cambrex, Charles City, IA) with 30% FCS (Perbio, Woburn, MA), 10 −7 M dexamethasone (Sigma-Aldrich, St. Louis, MO), penicillin/streptomycin and l-glu-tamine (PSG; Cambrex). When colonies were detected, cells were expanded without dexamethasone in a closed system applying cell stacks (Corning, Corning, NY). CB-derived cells (USSC and CB MSC) were incubated at 37°C in 5% CO 2 in a humidifi ed atmosphere. Reaching 80% confl uence, cells were detached with 0.25% trypsin (Cambrex) and replated 1:3.
Generation of cell clones employing the AVISO CellCelector ™
Clonal populations were obtained from established cell lines and, as additional approach, already during generation of cell lines by applying special cloning cylinders with sili-cone grease at the end of the cylinder (Chemicon, Billerica, MA) ( Fig. 1B ). In this case, cell lines were generated as de-scribed before and if distinct, separate, colonies were ob-served, a cloning cylinder was attached on a single colony and cells were trypsinated according to the standard pro-tocol. Cells of one colony were subsequently plated at low density into 6-well cell culture plates and single cells were picked employing the AVISO CellCelector ™ . Remaining cells were expanded as a whole bulk culture and referred to as initial cell line. By this approach, clonal lines were established from the youngest cells. The isolation of sepa-rate colonies by cloning cylinders allowed to examine and compare clonal cells derived from different colonies of the same CB. Clonal populations were then obtained applying the AVISO CellCelector ™ (Aviso, Greiz, Germany). This com-bined system consists of a robotic arm with an application-dependent tool, an integrated microscope, and a specialized computer software ( Fig. 1B ). For the purpose of generating clonal cell populations, a thin glass capillary with a diam-eter of 800 μm was used as application tool. In brief, cells were plated at low density (166 cells/cm 2 ) in 6-well cell cul-ture plates and after allowing the cells to get adherent again, distinct single cells were selected, picked, and transported to a defi ned destination well of a 96-well cell culture plate
most cells of the body, except in preadipocytes, pancreatic β cells, thymocytes, and cells in the adrenal gland. DLK-1 ex-pression decreases in preadipocytes under fetal calf serum (FCS) conditions in cell culture [ 23 ]. Importantly, DLK-1 con-trols lineage commitment and differentiation of MSC to adi-pocytes [ 24 ], and prolonged expression has been shown to inhibit the differentiation of preadipocytes to mature adipo-cytes [ 20 ]. DLK-1 knockout mice display growth retardation, skeletal malformation, scoliosis, hypotonicity, and obesity [ 25 ]. In fetal liver, isolation of hepatoblasts is currently based on the expression of DLK-1 as cellular marker [ 26 ]. DLK-1 expression can be detected in the cardiac mesoderm during early embryogenesis [ 27 ] and is expressed on the fetal liver cell line AFT024 that is capable of promoting the formation of “cobblestone areas” of proliferation [ 28 ]. We hypothesized that assessing and regulating the DLK-1 expression might offer an important means to distinguish between USSC and CB MSC and to infl uence the differentiation pattern of both cell types.
In this study, we demonstrate that different nonhematopoi-etic stem/progenitor cell populations exist in CB that can be distinguished by their adipogenic differentiation poten-tial and by their DLK-1 expression profi le. Overexpression of DLK-1 in initially negative CB MSC results in a marked blunting of their ability to undergo adipogenic differentia-tion, whereas silencing of high DLK-1 expressing USSC led to the formation of an adipogenic phenotype.
Materials and Methods
Generation and expansion of CB-derived cells
USSC and CB MSC were generated by the same method. Classifi cation of the adherent cells into USSC and CB MSC was only possible after generation by determining the adipo-genic differentiation potential and DLK-1 expression. CB was collected from umbilical cord vein with informed consent of the mother. Mononuclear cells (MNC) were obtained by fi coll (Biochrom, density 1.077 g/cm 3 ) gradient separation followed by ammonium chloride lysis of RBCs. The 5–7 ×10 6 CB MNC/mL were cultured in Dulbecco’s modifi ed Eagle’s medium
FIG. 1. Generation, growth kinetics, and age-related assessment of CB-derived cells and corresponding clonal popula-tions. ( A ) Unrestricted somatic stem cells (USSC) and cord blood mesenchymal stromal cells (CB MSC) generation was indicated by colonies (red ring) after 6–25 days. These cells grew into monolayer of fi broblastic, spindle-shaped type with a median generation effi ciency of 43% ( n = 370 cell lines) of which 37% ( n = 137) reached >6% and 10% ( n = 37) >9 passages. Scale bar = 200 μm. ( B ) A single cell colony was trypsinized applying cloning cylinders (Chemicon), plated into 6-well plates (Corning) at low density and clones were isolated applying the AVISO CellCelector ™ . Selected cells were automati-cally picked by a robotic arm and transported to the destination plate. After 14 days, the single cell-derived colonies, plated on 96-well plates containing conditioned medium, were trypsinized and expanded. Scale bar = 200 μm. ( C ) Employing the AVISO CellCelector ™ , clonal populations were generated with a median generation frequency of 33% ( n = 205). Single cells were picked and deposited in destination wells with photo documentation before and after selection. Following isolation, 38% ( n = 78) of the clonal populations could be further expanded for 6 or more and 21% ( n = 43) for 9 or more passages. Scale bar = 200 μm. ( D ) The median generation effi ciency of clonal populations of established USSC bulk cultures was higher (38%) than the generation effi ciency of clonal populations of established CB MSC cultures (26%). ( E ) Cumulative population doublings (CPD) of all CB MSC, CB MSC-derived clonal populations, USSC and USSC-derived clonal populations, and BM MSC were compared with each other under standard cell culture conditions. In P5, USSC reached 32.4 ± 2.2 CPD, USSC clonal populations 37.3 ± 4.1 CPD, CB MSC 29.8 ± 2.8 CPD, CB MSC clonal populations 40.5 ± 0.8 CPD, and BM MSC 24 ± 3.8 CPD. In P9, USSC reached 43.1 ± 4.1 CPD, CB MSC 37.8 ± 4.1 CPD, BM MSC 26.2 ± 4.7 CPD, USSC clonal populations 50 ± 2.4 CPD, and CB MSC clonal populations 46.7 ± 3.2 CPD. ( F ) Southern blot membrane: Comparison of telomere length of different cell lines ( n = 21) in a minimum of 4 different passages. USSC have longer telomeres (10.7 ± 4 kbp n = 11) compared with CB MSC (9.5 ± 0.98 kbp n = 10) and clonal populations (6.2 ± 0.83 kbp n = 8).
DLK-1 DISTINGUISHES USSC FROM MSC IN CB 3
(TRFs) were visualized. Finally, the size distribution of the TRFs was compared with a DNA length standard and evalu-ated utilizing ImageQuant ™ TL and TELRUN (Version 1.4).
Senescence assay . See Supplementary Data, available online at www.liebertonline.com/scd.
Immunophenotyping of USSC and CB MSC . See Supple-men tary Data.
In vitro differentiation into osteoblasts, adipocytes, and chon-droblasts . Differentiation into osteoblasts, adipocytes, chon-droblasts, and neural cells was performed as described previously [ 4 , 13 ]. Adipocytes were induced applying DMEM high glucose, 10% FCS, PSG, 10 −6 M dexamethasone, 0.2 mM
and cultured with preconditioned medium. For verifi cation, pictures were taken before and after each picking process to document successful single cell selection ( Fig. 1B and 1C ).
Telomere length detection and senescence assay
Telomere length assay . The Telo TAGGG Telomere Length Assay (Roche Diagnostics, Germany) was used for sensitive detection of telomeric DNA from cell cultures and for determi-nation of the telomere length. The genomic DNA was digested applying frequently cutting restriction enzymes ( Hin fI/ Rsa I). After digestion, the DNA fragments were separated by gel electrophoresis, blotted, and telomere restriction fragments
A
B
C
D
E
F
10%
21%
6%
10%
53%
P 1/2
P 3/4
P 5/6
P 7/8
P 9+
21%
8%
9%
13%
CB MSCDerivedClones
USSCDerivedClones
0
0USSC USSC Clones
P5P9
CB MSC CB MSC Clones
10
20
30
Mean C
PD
40
50
60
10 20 30 40 50
49%
P 1/2
P 3/4
P 5/6
P 7/8P 9+
KLUTH ET AL. 4
Affymetrix chip analysis . Five micrograms of total RNA were labeled as described in the Affymetrix Expression Manual Version 2. The Affymetrix Gene Chip protocol (Version 2) was followed for hybridization, washing and staining on the GC Scanner 3000 with G7 update. Images were analyzed with Affymetrix MAS5.0 and “global scal-ing” was employed for normalization.
Statistics
Statistics were evaluated applying the GraphPadInStat software. A 2-tail unpaired t -test was used.
Results
Generation of CB-derived cells
Over the last years, we initiated adherent cell cultures from 860 CB samples, from which 43% ( n = 370) gave rise to an average of 1–11 colonies per CB ( Fig. 1A ). After trypsiniza-tion, these spindle-shaped cells grew into monolayer within 2–3 weeks ( Fig. 1A ). Cell lines that did not reach >2 passages ( n = 196; 53%) were not characterized. Once established, 10% ( n = 37) of these cell lines reached passage (P) 3–4, 6% ( n = 22) P5–6, 21% ( n = 77), and P7–8 10% ( n = 37) yielded >9 passages ( Fig. 1A ). Already in P4, 1.5 × 10 9 cells could be obtained after expansion under GMP grade conditions [ 29 ]. If cells reached P9 ( n = 37), they could be further expanded to >20 passages, theoretically yielding up to 10 15 cells. USSC and CB MSC were always generated from fresh CB (<36 h after delivery). At the moment, there is no evidence that cel-lular parameters infl uence the cloning effi ciency of USSC and CB MSC. As previously shown by our group, there was no correlation between generation of CB-derived cell lines and gestational age, CB volume, the number of nucleated cells in the CB collections, hours after elapse, or the number of MNC in the CB after gradient separation [ 17 ].
Growth kinetics and age-related assessment of cells
Overall, 623 single cells from 12 different adherently growing CB-derived cell lines were isolated applying the AVISO CellCelector ™ ( Fig. 1B ) with a median cloning effi -ciency of 33% ( n = 205, Fig. 1C ). Out of these 205 clonal cell populations, 38% ( n = 78) could be expanded for further 6 passages and 21% ( n = 43) for further 9 passages.
The median generation effi ciency of clonal populations of established USSC bulk cultures was higher (38%) than the generation effi ciency of clonal populations of established CB MSC cultures (26%) ( Fig. 1D ).
Cumulative population doublings (CPD) were assessed from all cell lines tested, including BM MSC. USSC could be cultured for up to maximal 63 CPD, CB MSC reached up to maximal 53 CPD, and BM MSC up to 35 CPD in culture. Clonal populations yielded up to 57 CPD ( Fig. 1E ). Comparing the CPDs in P5, there were no differences between USSC (32.4 ± 2.2 CPD) and CB MSC (29.8 ± 2.8 CPD) ( P = 0.1239), while the CPDs in P9 (USSC 43.1 ± 4.1 CPD, CB MSC 37.8 ± 4.1 CPD) were signifi cantly different ( P = 0.0486). USSC-derived clonal populations (37.3 ± 4.1 CPD) reached higher CPDs at the same passage compared with USSC (32.4 ± 2.2 CPD) in P5 ( P =0.0214), as well as in P9 (USSC clonal populations 50 ± 2.4 CPD, USSC 43.1 ± 4.1) ( P = 0.0001). In P5, CB MSC-
indomethacine, 0.1 mg/mL insulin, and 1 mM 3-isobutyl-methylxanthine. Media was changed twice a week and lipid vacuoles stained with Oil Red O after 21 days. Size and area of chondrogenic pellets were determined applying the AVISO CellCelector ™ .
RNA isolation and reverse transcription polymerase chain reaction . Total RNA of cells was isolated utilizing the RNeasy Mini Kit (Qiagen, Valencia, CA). RNA of differentiated cells was isolated using TRI Reagent ® (Sigma-Aldrich) following the instruction protocol. One microgram RNA was reverse-transcribed applying SuperScriptIII (Invitrogen, Carlsbad, CA) according to the manual.
Reverse transcription polymerase chain reaction and real-time polymerase chain reaction . For reverse transcription poly-merase chain reaction (RT-PCR), 2 min 94°C, 35 cycles of 30 s 94°C, 30 s 60°C and 30 s 72°C, and 5 min 72°C were per-formed. Real-time PCR was performed with SYBR ® Green PCR Mastermix (Applied Biosystems, Foster City, CA; primer see Supplementary Table 1; Supplementary mate-rials are available online at www.liebertonline.com/scd). Evaluation of TaqMan Gene Expression Assays (Applied Biosystems) was performed utilizing the SDS2.3 program and expression normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH).
Immunohistochemistry . The 10,000 cells/cm 2 were plated on chamber slides (Lek-Tek, Chamber Slide ™ , 2-Well Glass Slide). Cells were fi xed with −20°C acetone for 2 min or 4% paraformaldehyde (PFA). Primary antibody (Supplementary Table 2; Supplementary materials are available online at www.liebertonline.com/scd) incubation was performed at 4°C overnight. Second antibody (Supplementary Table 2) was applied in 1:400 dilutions. Nuclei were stained with ProLong ® Gold antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen).
Western blot analysis . Total protein was analyzed in west-ern blot analysis (NuPAGE ® System, Invitrogen) and ECL PLUS Western Blotting Detection Reagents (GE Healthcare, Piscataway, NJ) applied for detection (see Supplementary Data).
Overexpression of DLK-1 . Full-length human DLK-1 was am-plifi ed using Phusion ® Taq Polymerase (New England Biolabs, Beverly, MA). Insert and vector (pCL7Egwo; Supplementary Fig. 6, available online at www.liebertonline.com/scd) were digested employing Fast Digest Eco RI and Fast Digest Xho I (both Fermentas). The eGFP gene was replaced by the DLK-1 gene. Products were electrophoresed on 2% agarose gels and then purifi ed using QIAquick ® Gel Extraction Kit (Qiagen). Insert and vector were ligated applying T4 DNA Ligase (New England Biolabs) in a 3:1 ratio. Transformed electro-competent Top 10 bacteria (Invitrogen) were plated on LB agar containing ampicillin (500 μg/L) overnight. DLK-1 + col-onies were expanded in 100 mL LB media and vectors were isolated applying MaxiPrep ® Kit (Qiagen). 293T cells were transfected with pCL7Egwo_DLK-1, Galv ™ , and CD/NL-BH utilizing FuGene HD Transfection Reagent. CB MSC were transfected with the supernatant.
Silencing of DLK-1 in USSC . DLK-1 silencing was performed using human DLK-1 shRNA (SHCLNG-NM_003836; Sigma-Aldrich; see Supplementary Fig. 5; available online at www.liebertonline.com/scd) following manufacturer’s instructions. Evaluation was performed analyzing DLK-1 expression by real-time PCR and the adipogenic differentiation potential.
DLK-1 DISTINGUISHES USSC FROM MSC IN CB 5
www.liebertonline.com/scd). BM MSC reached signifi cantly less CPD in P5 as well as in P9 compared with USSC (P5 P = 0.0028; P9 P = 0.0013), USSC-derived clonal populations (P5 P < 0.0001; P9 P < 0.0001),CB MSC (P5 P = 0.0122; P9 P = 0.0142), and CB MSC-derived clonal populations (P5 P < 0.0001; P9 P = 0.0001).
derived clonal populations reached 40.5 ± 0.8 CPD, while CB MSC reached only 29.8 ± 2.8 CPD ( P < 0.0001). In P9, growth kinetics of CB MSC-derived clonal populations (46.7 ± 3.2 CPD) and CB MSC (37.8 ± 4.1 CPD) were still signif-icantly different ( P = 0.0022) ( Fig. 1E and Supplementary Fig. 1A; Supplementary materials are available online at
A
B
C
USSC 1
USSC
00
200
400
600
800
1,000
1,200
1 × 104
2 × 104
3 × 104
4 × 104
5 × 104
6 × 104
7 × 104
8 × 104
9 × 104
CB MSC
Are
a [
μm
2]
Dia
mete
r [μ
m]
BM MSC
USSC 2
Acia
n B
lue-P
AS
Acia
n B
lue-P
AS
HE
Sta
inin
g
CB MSC 1 BM MSC 1
USSC CB MSC BM MSC
Contr
ol
Oil
Red 0
Aliz
arin R
ed
USSC 1 Clone A3 CB MSC 1 CB MSC 1 Clone B7 BM MSC2 Preadipo Cytes
FIG. 2. In vitro mesodermal differentiation. ( A ) Osteogenic differentiation was evaluated in 2–3 independent experiments. While no differences in the osteogenic differentiation potential between unrestricted somatic stem cells (USSC) ( n = 36) and cord blood mesenchymal stromal cells (CB MSC) ( n = 27) could be detected by Alizarin Red staining after 14 days of induc-tion, the osteogenic differentiation potential of clonal populations ( n = 39), BM MSC ( n = 8, positive control), and preadipo-cytes (positive control) was restricted. Scale bar = 100 μm. Adipogenic differentiation was confi rmed by Oil red O staining 14 days after adipogenic induction in 2–4 independent experiments. A high adipogenic differentiation potential could be illustrated for CB MSC ( n = 28), CB MSC-derived clonal populations ( n = 18), BM MSC ( n = 8, positive control), and human preadipocytes (positive control), whereas USSC ( n = 35) and USSC-derived clonal populations ( n = 16) never possessed the adipogenic phenotype. Noninduced control population. Scale bar = 200 μm. ( B ) Chondrogenic differentiation was evalu-ated analyzing diameter (in μm) and area (in μm 2 ) of chondrogenic pellets. USSC ( n = 5) formed slightly bigger pellets com-pared with CB MSC ( n = 5) and BM MSC ( n = 15). ( C ) Hematoxylin–eosin staining did not reveal any differences between the analyzed sections of differentiated USSC, CB MSC, and BM MSC after 21 days of induction. Alcianophilia staining of USSC, typical for cartilage, was slightly stronger and little less diffuse compared with CB MSC and BM MSC (Alcian blue–periodic acid-Schiff (PAS) staining overview and in detail).
KLUTH ET AL. 6
observed for the clonal populations, BM MSC and preadi-pocytes (all in P4–6). Adipogenic differentiation was docu-mented by Oil Red O staining of lipid vacuoles ( Fig. 2A ). On Day 14 after induction, 43.1% out of 65 CB cell lines and 46% out of 39 cell clones showed the adipogenic phe-notype, while no lipid vacuoles were ever detected in the noninduced control cells ( Fig. 2A ). The osteogenic as well as the adipogenic differentiation potential of USSC and CB MSC was weaker in P10. Corresponding clonal popula-tions in P9–10, if they could be expanded so far, differed in their differentiation capability (Supplementary Fig. 4A; Supplementary materials are available online at www.liebertonline.com/scd). However, compared with the ini-tial cell lines both the osteogenic and the adipogenic differ-entiation potential was restricted.
These results imply that the adipogenic differentiation potential could be a distinguishing feature between the dif-ferent adherently growing CB-derived cells. We therefore defi ned CB-derived cells with high adipogenic differentia-tion potential as CB MSC ( n = 28) and the ones without as USSC ( n = 35).
After 21 days of chondrogenic induction, no signifi cant differences could be detected analyzing diameter and area or hematoxylin–eosin staining of chondrogenic pellets formed by USSC ( n = 5), CB MSC ( n = 5), and BM MSC ( n = 15) ( Fig. 2B ). Alcian blue–periodic acid-Schiff (PAS) staining did not reveal clear differences between USSC compared with CB MSC and BM MSC ( Fig. 2C ). While USSC exhibited a slightly stronger, cartilage specifi c alcianophilia morphology in the central region of the chondrogenic pellet, Alcian blue–PAS staining in CB MSC and BM MSC indicated a weaker, more diffuse alcianophilia morphology. No remarkable differ-ences between CB MSC and BM MSC were observed.
Screening for a marker linked to adipogenic differentiation revealed DLK-1 as a potential key player to inhibit the adipogenic pathway
Peroxisome proliferator activator γ ( PPARγ ), CCAAT enhancer-binding protein α ( CEBPα ), adiponectin ( ADIPOQ ), fatty acid-binding protein 4 ( FABP4 ), and perilipin ( PLIN ) are specifi c genes that are up-regulated during adipogen-esis, while DLK-1 , one of the key players in adipogenesis, is
To determine more precisely the “biological age” of tested cell populations, a telomere length assay ( Fig. 1F ) was performed in P9. USSC telomeres are in average 10.7 ± 1.4 kbp long. CB MSC had 9.5 ± 0.98 kbp long telomeres and in BM MSC telomere length was shortest with 8.8 ± 1.3 kbp ( Fig. 1F ). USSC-derived clonal populations had shorter telomeres 6.6 ± 0.8 kbp com-pared with their initial cell line in P9 (data not shown).
In addition, senescence was analyzed at least in 2 differ-ent passages by histochemical detection of the senescence-associated β-galactosidase. USSC exhibited a senescent rate of 30%–45% whereas already 70%–80% CB MSC were se-nescent after 9 passages. For comparison, almost 95% of BM MSC in P9 were senescent (Supplementary Fig. 1B, available online at www.liebertonline.com/scd).
Expression of specifi c surface antigens was analyzed in fl ow cytometry. All lines tested were negative for CD31, CD34, CD45, CD56, CD106, AC133 (CD133/1), CD184, and HLA-DR. They expressed high levels of CD13, CD29, CD44, CD71, CD73, CD105, CD146, CD166, and HLA-ABC but low levels of NG2, PDGFRα, and PDGFRβ. Fluorescence-activated cell sorting (FACS) analysis did not reveal any signifi cant differ-ences between the different cell lines and CB-derived clones (Supplementary Fig. 2; Supplementary materials are avail-able online at www.liebertonline.com/scd).
RT-PCR, real-time PCR, and immunohistochemistry were performed to determine the expression analysis of OCT4A [ 30 , 31 ], NANOG, SOX2, KLF4, and c-MYC uti-lized for the induction of pluripotent stem cells from adult human fi broblasts [ 32 ]. There was no expression of OCT4A, NANOG, and SOX2 in any of the CB-derived cell popula-tions (Supplementary Fig. 3; Supplementary materials are available online at www.liebertonline.com/scd).
In vitro adipogenic differentiation as a distinguishing feature between cells derived from CB
BM MSC ( n = 8) and cultured human preadipocytes (Cell Line Service) ( n = 2) were used as positive controls for osteogenic, chondrogenic, and adipogenic differentia-tion. Whereas no differences in the formation of osteoblasts between the different CB-derived cell lines in the analyzed passages (P5 and 9) could be observed by Alizarin Red staining ( Fig. 2A ), a restricted differentiation potential was
FIG. 3. Screening for a marker linked to adipogenic differentiation revealed DLK-1 as a potential key player to inhibit the adipogenic pathway ( A ) CB-derived colonies were lysated and DLK-1 expression determined related to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression. Two out of 8 colonies revealed a strong DLK-1 expression, while 6 out of 8 colonies showed a low or no DLK-1 expression. ( B ) While analyzing DLK-1 expression of cord blood mesenchymal stromal cells (CB MSC), unrestricted somatic stem cells (USSC), and corresponding clonal populations, a very high expres-sion could be determined for USSC and most of the USSC-derived clonal populations whereas CB MSC and almost all CB MSC-derived clonal populations were DLK-1 − or very weakly positive. DLK-1 expression was clearly down-regulated in some USSC-derived clonal populations in subsequent passages under standard culture conditions compared with the initial cell line by real-time PCR. nTERA-2 was used as positive control and dermal fi broblasts as negative control. ( C ) After 5 days of osteogenic induction, DLK-1 was clearly down-regulated in USSC, while RUNX2 expression was up-regulated, indicating an early osteogenic differentiation. ( D ) During adipogenic induction, DLK-1 was down-regulated in USSC, however not completely absent. PPARγ2 expression was clearly up-regulated in CB MSC but never in USSC during adipogenic induction. ( E ) Immunohistochemical staining of USSC, CB MSC, and BM MSC after 21 days of adipogenic induction and noninduced controls with antibodies to human PLIN. CB MSC, BM MSC, and preadipocytes were PLIN + , USSC were always PLIN − . Scale bar = 200 μm. ( F ) Immunohistochemical staining of USSC, CB MSC, and BM MSC after 21 days of adipogenic induction and noninduced controls with antibodies to human DLK-1 confi rmed the data of RT-PCR and western blot analysis. Human preadipocytes: positive control for adipogenic differentiation; nTERA-2: positive control for DLK-1. Scale bar = 200 μm.
DLK-1 DISTINGUISHES USSC FROM MSC IN CB 7
down-regulated as schematically shown in Supplementary Figure 4B [ 23 ]. DLK-1 expression of CB-derived colonies displayed a heterogeneous expression profi le. Two out of 8 colonies were DLK-1 + , while 5 colonies were weakly posi-tive and 1 colony was DLK-1 − as detected by real-time PCR
( Fig. 3A ). DLK-1 expression could be verifi ed by RT-PCR in 16 out of 32 CB-derived cell lines and 11 out of 23 cell clones. These cells never differentiated toward the adipogenic lineage and were therefore termed, as mentioned before, USSC. CB MSC ( n = 16), and cultured preadipocytes, were
A
C
E F
D
B
Expre
ssio
n R
ela
ted t
o G
AP
DH
Expre
ssio
n R
ela
ted to G
AP
DH
Expre
ssio
n R
ela
ted to G
AP
DH
Expre
ssio
n R
ela
ted t
o G
AP
DH
CB
Colo
ny 1
CB
MS
C 1
CB
MS
C 1
Clo
ne A
10
CB
MS
C 1
Clo
ne B
1
CB
MS
C 1
Clo
ne B
4
CB
MS
C 1
Clo
ne C
1
CB
MS
C 1
Clo
ne D
8.2
CB
MS
C 1
Clo
ne D
9
US
SC
1 P
6
US
SC
1 C
lone A
2 P
8
US
SC
1 C
lone A
2 P
9
US
SC
1 C
lone A
2 P
10
US
SC
2
US
SC
2 C
lone B
4 P
7
US
SC
2 C
lone B
7 P
7
nT
ER
A2 P
7
Derm
al F
ibro
bla
ts
CB
Colo
ny 2
CB
Colo
ny 3
CB
Colo
ny 4
CB
Colo
ny 5
DLK1RUNX2
DLK1
PPARg
CB
Colo
ny 6
CB
Colo
ny 7
CB
Colo
ny 8
US
SC
2 d
0
CB
MS
C2 P
5 d
0
CB
MS
C2 P
5 d
7
CB
MS
C2 P
5 d
14
CB
MS
C2 P
5 d
21
US
SC
4 P
5 d
0
US
SC
4 P
5 d
7
US
SC
4 P
5 d
14
US
SC
4 P
5 d
21
0.0
US
SC
2 d
5
US
SC
2 d
13
AdipogenicInduction
CB
MS
C1
BM
MS
C1
US
SC
1P
rea
dip
ocyte
s
CB
MS
C1
BM
MS
C1
US
SC
1P
read
ipocyte
s
WithoutInduction
AdipogenicInduction
WithoutInduction
CB
MS
C 1
CB
MS
C 1
d5
CB
MS
C 1
d13
0.0
2.0 × 10–4
1.0 × 10–4
0.0
4.0 × 10–4
1.0 × 10–4
1.5 × 10–4
2.0 × 10–4
2.5 × 10–4
3.0 × 10–4
8.0 × 10–4
1.2 × 10–5
1.6 × 10–5
2.0 × 10–5
2.0 × 10–4
3.0 × 10–4
4.0 × 10–4
5.0 × 10–4
6.0 × 10–4
7.0 × 10–4
8.0 × 10–4
9.0 × 10–4
1.0 × 10–5
0.0
2.0 × 10–4
4.0 × 10–4
4.0 × 10–3
8.0 × 10–3
1.2 × 10–1
1.6 × 10–1
4.0 × 10–4
6.0 × 10–4
8.0 × 10–4
1.0 × 10–5
KLUTH ET AL. 8
( n = 9) were differentiated into the neural lineage. USSC and USSC-derived clones revealed a neural morphology demonstrated by immunohistochemistry for human β-3-tubulin and neurofi lament and by real-time PCR detecting neurofi lament light, medium, and heavy polypeptides and β-3-tubulin (Supplementary Fig. 5A and 5B) as described by Greschat et al. [ 13 ]. No neural phenotype was ever detected for CB MSC or corresponding clonal populations (Supplementary Fig. 5A).
DLK-1 overexpression in CB MSC
To test whether DLK-1 overexpression of CB MSC might lead to a broader differentiation capacity, CB MSC were transduced with a lentiviral vector (pCL7Egwo_DLK-1), expressing the full-length human DLK-1 cDNA of an in-ternal constitutive promotor and with the control vector pCL7Egwo, containing eGFP instead of DLK-1. Effectiveness of transduction was evaluated by FACS analysis and fl uo-rescence microscopy (Supplementary Fig. 6A). Transduced cells revealed no morphological changes or differences in growth kinetics compared with the nontransduced cells. Concentrating of virus resulted in cell death while virus dilution of 1:10 did not affect the adipogenic differentiation potential of CB MSC (data not shown). Following transfec-tion, high amounts of DLK-1 could be detected in real-time PCR ( Fig. 4A ), confi rmed by immunohistochemistry ( Fig. 4C ). Correlating with the high DLK-1 expression, these DLK-1 overexpressing cells (CB MSC DLK-1+ ) revealed a di-minished adipogenic differentiation potential compared with the initial cell line ( Fig. 4C ).
Silencing of DLK-1 in high expressing USSC
DLK-1 expression was repressed in high express-ing USSC (USSC DLK-1− ) employing a lentiviral plasmid (pLKO.1-puro; SHCLNG-NM_003836, Sigma-Aldrich) for stable transfection. After puromycin selection, cells were expanded and DLK-1 expression was compared with the initial USSC. In real-time PCR, 2 out of 5 plasmids (NM_003836.3-873s1c1: CCGGAGGTCTCACCTGTGTCAAG A AC T CG AG T T C T T G AC AC AG G T G AG ACC T TTTTTG; NM_003836.2-1227s1c1: CCGGCCTG GCCGTCAAC ATCATCTTCTCGAGAA GATGATGTTGACGGCCA GGTTTTTG) resulted in a 10-fold decrease of DLK-1 expression com-pared with the initial cell line ( Fig. 4A ) and no DLK-1 stain-ing in immunohistochemistry ( Fig. 4C ), accompanied by a 20-fold up-regulation of PPARγ2 expression after 21 days of adipogenic induction compared with the initial USSC ( Fig. 4B ) and separate lipid vacuoles stained by Oil Red O ( Fig. 4C ).
Discussion
In addition to current clinical use, CB is under intense experimental investigation in preclinical models of patho-physiology and it is anticipated that widespread use of CB for nonhematopoietic tissue regeneration will increase. USSC and MSC from CB could have distinct biological advantages compared with their adult BM counterpart. During the last years, MSC from adult BM [ 34–37 ] has drawn attention to therapeutic applications. MSC lines or MSC like lines have been isolated from adult, fetal,
DLK-1 − or very weakly positive but exhibited a strong adipo-genic differentiation capacity. BM MSC from young donors were weakly DLK-1 + and BM MSC from older donors were DLK-1 − , correlating with a weaker adipogenic differentia-tion potential of BM MSC from young donors than BM MSC from older donors. After analyzing DLK-1 expression of USSC and CB MSC and corresponding clonal populations, a very low expression could be detected for CB MSC-derived clonal populations ( n = 46), while the DLK-1 expression of USSC-derived clonal populations ( n = 43) varied from very high expression to very low expression ( Fig. 3B ). Some USSC-derived clonal populations lost their ability to express DLK-1 with subsequent passages (P6–10) as illustrated by real-time PCR ( Fig. 3B ). This indicates a loss of DLK-1 with “biolog-ical” aging in culture. Analyzing 65 cell lines, we also found 2 out of 65 bulk cell lines that expressed high levels of DLK-1 and revealed no adipogenic differentiation in P5. With sub-sequent passages DLK-1 was down-regulated and an adi-pogenic phenotype could be detected (in P10) by Oil Red O staining (Supplementary Fig. 4C).
We classifi ed this cell population between USSC and CB MSC (differentiation potential P5 vs. P10) as intermediate ( n = 2).
After osteogenic as well as adipogenic induction ( Fig. 3C and 3D ), DLK-1 expression was down-regulated in USSC, however not completely absent. During osteogenic differentiation, DLK-1 down-regulation in USSC was ac-companied by an up-regulation of RUNX2 expression ( Fig. 3C ). PPARγ2 expression in adipogenic differentiation was up-regulated in CB MSC during adipogenic differentia-tion, while no signifi cant up-regulation could be detected in USSC after adipogenic induction ( Fig. 3D ). Fourteen days after adipogenic induction, expression of the adipo-genic markers PPARγ , FABP4 , and PLIN could be detected in all BM MSC (positive control, n = 3) and also in all CB MSC ( n = 7). After adipogenic induction, ADIPOQ expres-sion was hardly detectable in CB MSC. No expression of adipogenic genes could ever be detected in USSC ( n = 6) and in the respective noninduced controls (Supplementary Fig. 4D).
RT-PCR results were confi rmed by western blot analysis (Supplementary Fig. 4E and 4F) and immunohistochemical staining for PLIN, a marker for fi nal adipogenesis [ 33 ], and DLK-1 21 days after adipogenic induction ( Fig. 3E and 3F ). USSC were tested negative for PLIN after adipogenic dif-ferentiation while CB MSC were positive and the highest amount of PLIN was detected in differentiated BM MSC and in human preadipocytes (positive controls) ( Fig. 3E and Supplementary Fig. 4E). No PLIN could be detected in the noninduced controls. The adipogenic-induced USSC, CB MSC, and BM MSC (negative control) and noninduced cells were also analyzed for DLK-1 expression on the pro-tein level. In accordance to the failed adipogenic differen-tiation, DLK-1 was highly expressed in USSC detected by immunohistochemistry ( Fig. 3F ) and western blot analysis (Supplementary Fig. 4F).
USSC and CB MSC differ in their neural differentiation potential
USSC ( n = 16), USSC-derived clonal populations ( n = 6), CB MSC ( n = 4), and CB MSC-derived clonal populations
DLK-1 DISTINGUISHES USSC FROM MSC IN CB 9
MSC. As clearly presented in the work by Crisan et al. [ 41 ], MSC derived from pericytes exhibit a very strong adipogenic differentiation even on a clonal level. Due to the restricted adipogenic differentiation potential, we do not presume a pericyte but a fetal or fetal liver origin of USSC.
However, inconsistent data on adipogenic differenti-ation potential of CB-derived cell populations have been published [ 6 , 7 , 42 ]. Here we demonstrate that USSC and CB MSC can be distinguished by their adipogenic differentia-tion potential and the inverse correlation with expression of DLK-1 . High DLK-1 expressing USSC never exhibited the
and embryonic human and animal tissue; however, the true MSC progenitors had never been exactly defi ned [ 38 ]. Takashima et al. [ 39 ] were the fi rst to describe neu-roepithelial cells to supply the initial transient wave of multipotent MSC differentiation toward both the neural and the mesodermal lineage. Nevertheless, data from Takashima et al. demonstrated that it is unlikely that MSC established from adult BM are of neural crest origin. It is widely believed that BM MSC are derived from mesoderm [ 40 ]. During fetal development, MSC or MSC progenitors play a vital role in tissue remodeling and differentiation. Recently, it was proposed that pericytes are the origin of
A
C
B
0 0.0
0.2
0.4
0.6
0.8
1.0
1.2
2
4
6
100
150
200
250
300
350
Rela
tive E
xpre
ssio
n o
f D
LK
-1 R
ela
ted t
o n
TE
RA
-2
Rela
tive E
xpre
ssio
n o
f P
PA
Rg R
ela
ted
to H
um
an A
dip
ocyte
CB
-MS
C 1
US
SC
3
US
SC
3 C
onstr
uct1
US
SC
3 C
onstr
uct2
US
SC
3 N
on Induced
US
SC
3
US
SC
3 C
onstr
uct1
US
SC
3 C
onstr
uct2
Hum
an A
dip
ocyte
s
Derm
al F
ibro
bla
sts
nT
ER
A2
Derm
al F
ibro
bla
sts
CB
-MS
C+
DLK
-1
CB MSC1
Adip
ogenic
Diffe
rentiation
DLK
-1
CB MSC1DLK1+
USSC3 USSC3DLK1– + Control
FIG. 4. DLK-1 overexpression in non-DLK-1 expressing cord blood mesenchymal stromal cells (CB MSC) and DLK-1 si-lencing in high expressing unrestricted somatic stem cells (USSC). ( A ) DLK-1 expression of CB MSC, CB MSC transduced with pCLEGwo7 containing full-length DLK-1 (CB MSCDLK-1 + ), USSC, and USSC transduced with pKLO.1-puro was analyzed by real-time PCR, utilizing nTERA-2 as positive control. Successful transfection of CB MSC with human DLK-1 was indicated by real-time PCR in comparison with the initial cell line. The transduced cell population expressed 230 times more DLK-1 than the positive control nTERA-2. DLK-1 silencing in USSC resulted in a 10-fold down-regulation of USSC compared with the initial cell line. Dermal fi broblasts were used as negative control. ( B ) After 21 days of adipogenic induction, the expres-sion of PPARγ2 in USSC3DLK-1 − and USSC3 was compared with the noninduced controls, dermal fi broblasts, and human adipocytes. USSC3DLK-1 expressed 20-folds more PPARγ2 than the initial cell line. ( C ) Comparison of CB MSC1DLK-1 + with CB MSC1 and USSC3DLK-1 − to USSC3. Immunohistochemical staining with an antibody to human DLK-1 and Oil Red O staining after 14 days of adipogenic induction was performed with the transfected cell population compared with the origin cell line. No DLK-1 was observed in CB MSC1, but CB MSC1DLK-1 + were strongly DLK-1 + . In contrast, USSC3DLK-1 − were DLK-1 − in immunohistochemistry, while the initial USSC3 revealed a strong DLK-1 signal. In comparison with the initial CB MSC1, the adipogenic differentiation potential of CB MSC1DLK-1 + was restricted, while USSC3DLK-1 − formed separate lipid vacuoles, and USSC3 were negative in Oil Red O staining. As positive controls, nTERA-2 was employed for DLK-1 staining, and differentiated human preadipocytes were used for adipogenic differentiation. Scale bar = 200 μm.
KLUTH ET AL. 10
observation that newborns have adipose tissue only in the toes [ 54 ].
Since a hallmark of stem and progenitor cells is their ability to proliferate and give rise to functional progeny on a single cell level, the analysis here was based on both USSC and CB MSC lines including clonal-derived cells. While analyzing clonal populations from CB MSC, as well as clonal populations from USSC, heterogeneous DLK-1 expression profi le could be detected ( Fig. 3B ). Particular single cell clones from DLK-1 expressing USSC cell lines (isolated applying the AVISO CellCelector ™ ) down-reg-ulated DLK-1 expression with increasing age (P6–10, Fig. 3B ). Overexpression of DLK-1 applying lentiviral vector pCL7Egwo_DLK-1 in nonexpressing CB MSC changed the cells toward a more undifferentiated status—here a re-stricted adipogenic differentiation potential compared with the initial cell line was detected ( Fig. 4 ). Silencing of DLK-1 in high expressing USSC resulted in a low adipogenic dif-ferentiation potential, indicated by up-regulation of PPARγ expression and separate lipid vacuoles stained by Oil Red O ( Fig. 4B and 4C ). These data are consistent with previous published data of Wang and Sul who demonstrated that mouse embryonic fi broblast (MEF) cells overexpressing DLK-1 were not able to form adipocytes, while silencing of DLK-1 resulted in an improved adipogenic differentiation potential [ 55 ].
Osteogenic differentiation was identical in all the CB-derived lines as shown by Alizarin Red staining (USSC and CB MSC) ( Fig. 2A ). Microarray analysis showed a down-regulation of DLK-1 (7,276.6-fold higher expression before induction) and SOX9 (26,202.4-fold higher expression before induction) in CB-derived cells after osteogenic induction explaining why, in contrast to previous published data in murine multipotent stem cells (C3H10T1/2) and 3T3-L1 cells [ 55 ], no differences in the osteogenic differentiation potential between USSC and CB MSC could be observed. With regard to the endodermal differentiation [ 52 ], our data are consis-tent with the data from Tanimizu et al. [ 26 ], who expressed DLK-1 in the fetal hepatocyte primary culture and demon-strated that the time course of expression of hepatic differ-entiation marker genes was not altered. The fact that DLK-1 is specifi cally expressed in fetal liver suggests that DLK-1 is implicated in proliferation and/or differentiation of hepato-cytes. Another possibility is its involvement in hematopoi-esis. Moore et al. [ 28 ] described the isolation of DLK-1 expressing fetal murine stroma lines (eg, AFT024) that were able to support hematopoiesis. The authors have demon-strated that not all generated cells from fetal liver expressed DLK-1, and some of them did not exhibit hematopoietic sup-porting activity.
Several groups demonstrated DLK-1 expression in the embryo, which marks the growing branches of organs that develop through the process of branching morphogenesis [ 47 ]. In addition to high proliferating hematopoietic cells, CB MNC can easily form osteoclasts in vitro (unpublished observation by the authors). As reviewed by Gimble et al. [ 56 ], preadipocytes, which are not well defi ned biologically, can create a microenvironment conductive to osteoclast formation and resorptive activity through direct cell con-tact and release of soluble factors. Since tissue/organs must be designed and remodeled during fetal development, the presence of such cells like USSC or MSC, simultaneously
adipogenic phenotype, or expression of adipogenic genes after adipogenic induction. The inverse correlation of adi-pogenic differentiation and DLK-1 expression is well known and established in detail [ 20 , 21 , 24 , 43–45 ]. Our fi ndings are based on these data, which showed that high DLK-1 expres-sion prevents adipogenic differentiation in murine 3T3-L1 cells and human MSC.
It has been reported that DLK-1 expression keeps the cells in an undifferentiated status [ 46–48 ] and that DLK-1 is highly expressed during embryonic development [ 22 ]. Here we are able to show that DLK-1 expression prevents the forma-tion of lipid vacuoles (Supplementary Fig. 4). Cell lines that expressed high levels of DLK-1 in early passages were not able to differentiate toward the adipogenic lineage, whereas the loss of DLK-1 expression with subsequent passages led to the development of adipocytes. Moreover, DLK-1 express-ing USSC are able to differentiate toward neural cells in vitro, whereas MSC behave different [ 18 ]. The data demon-strate that DLK-1 plays an essential role in maintaining the undifferentiated/proliferative status of USSC in CB and may have potential to determine the cell fate. Low expression of DLK-1 in BM MSC can be explained by the age of the donors (44 ± 1.5-year-old healthy donors), while preadipocytes lose DLK-1 expression during cultivation with FCS [ 23 ].
This function of DLK-1 seems to be consistent with pre-viously reported data in other cellular systems. DLK-1 has an important function in the differentiation of hepatocytes, osteoblasts, neurons, pancreas, and skeletal muscle during human fetal development [ 22 , 49 , 50 ]. In addition, there is evidence that DLK-1 is a negative regulator of mesodermal differentiation [ 51 ]. In extraembryonic tissues, DLK-1 expres-sion could be detected in the visceral endodermal cells sur-rounding the blood island of the yolk sac and in the stromal cells of the placental villi [ 22 ].
USSC and CB MSC could already be isolated in the pri-mary cultures based on DLK-1 expression in primary colo-nies, as shown in Figure 3A , and followed further by applying real-time PCR. High DLK-1 expression in USSC defi ned cells with an extensive proliferation and a broader differentia-tion potential as indicated here ( Fig. 1D , 2 ). USSC could be expanded up to 63 CPD ( Fig. 1D ) and be differentiated into neural or endodermal cells in vitro (Supplementary Fig. 5) [ 13 , 15 , 52 ]. The high expression of DLK-1 correlated with an inhibition of adipogenic differentiation. No expression of any adipogenic marker ( FABP4 , PPARγ , PLIN , and ADIPOQ ) could be detected in USSC applying adipogenic culture con-ditions (Supplementary Fig. 4B). Although the lack of the adipogenic differentiation potential seems to be contradic-tory to a multipotent cell, it was also shown for embryonic stem cells that no spontaneous adipogenic differentiation could be observed. It was demonstrated that embryonic stem cell-derived adipocytes can be generated only by spe-cial protocols applying retinoic acid. In addition, according to the HOX code USSC cluster rather to embryonic stem cells than to fi broblasts [ 53 ]. Analyzing the less potent CB MSC, a weak DLK-1 expression and a very strong adipogenic dif-ferentiation were observed ( Fig. 2A ). These CB MSC also revealed slightly shorter telomeres, a higher percentage of senescent cells compared with USSC and revealed a reduced proliferative capacity ( Fig. 1E and Supplementary Fig. 1B). When compared with adult BM, the reduced level of cells in CB to generate adipocytes correlates with the biological
DLK-1 DISTINGUISHES USSC FROM MSC IN CB 11
References
1. Eapen M , P Rubinstein , MJ Zhang , C Stevens , J Kurtzberg , A
Scaradavou , FR Loberiza , RE Champlin , JP Klein , MM Horowitz
and JE Wagner . ( 2007 ). Outcomes of transplantation of unrelated
donor umbilical cord blood and bone marrow in children with
acute leukaemia: a comparison study . Lancet 369 : 1947 – 1954 .
2. Gluckman E and V Rocha ; EBMT Paediatric, Acute Leukemia
Working Parties; Eurocord . ( 2008 ). Indications and results of
cord blood transplant in children with leukemia . Bone Marrow
Transplant 41(Suppl. 2) : S80 – S82 .
3. Rocha V and E Gluckman . ( 2007 ). Outcomes of transplantation
in children with acute leukaemia . Lancet 369 : 1906 – 1908 .
4. Kögler G , S Sensken , JA Airey , T Trapp , M Müschen , N Feldhahn ,
S Liedtke , RV Sorg , J Fischer , C Rosenbaum , S Greschat , A
Knipper , J Bender , O Degistirici , J Gao , AI Caplan , EJ Colletti ,
G Almeida-Porada , HW Müller , E Zanjani and P Wernet . ( 2004 ).
A new human somatic stem cell from placental cord blood
with intrinsic pluripotent differentiation potential . J Exp Med
200 : 123 – 135 .
5. Chan SL , M Choi , S Wnendt , M Kraus , E Teng , HF Leong , and
S Merchav . ( 2007 ). Enhanced in vivo homing of uncultured and
selectively amplifi ed cord blood CD34+ cells by cotransplanta-
tion with cord blood-derived unrestricted somatic stem cells .
Stem Cells 25 : 529 – 536 .
6. Chang YJ , DT Shih , CP Tseng , TB Hsieh , DC Lee and SM
Hwang . ( 2006 ). Disparate mesenchyme-lineage tendencies in
mesenchymal stem cells from human bone marrow and umbil-
ical cord blood . Stem Cells 24 : 679 – 685 .
7. Kern S , H Eichler , J Stoeve , H Klüter and K Bieback . ( 2006 ).
Comparative analysis of mesenchymal stem cells from bone
marrow, umbilical cord blood, or adipose tissue . Stem Cells
24 : 1294 – 1301 .
8. Kim BO , H Tian , K Prasongsukarn , J Wu , D Angoulvant ,
S Wnendt , A Muhs , D Spitkovsky and RK Li . ( 2005 ). Cell
transplantation improves ventricular function after a myocar-
dial infarction: a preclinical study of human unrestricted so-
matic stem cells in a porcine model . Circulation 112 : I96 – 104 .
9. Kögler G , P Critser , T Trapp and M Yoder . ( 2009 ). Future of
cord blood for non-oncology uses . Bone Marrow Transplant
44 : 683 – 697 .
10. van den Berk LC , BJ Jansen , KG Siebers-Vermeulen , MG Netea ,
T Latuhihin , S Bergevoet , RA Raymakers , G Kogler , CC Figdor ,
GJ Adema and R Torensma . ( 2009 ). Toll-like receptor trigger-
ing in cord blood mesenchymal stem cells . J Cell Mol Med
13 : 3415 – 3426 .
11. Winter M , XN Wang , W Daubener , A Eyking , M Rae , AM
Dickinson , P Wernet , G Kogler and RV Sorg . ( 2008 ). Suppression
of cellular immunity by cord blood-derived unrestricted somatic
stem cells is cytokine dependent . J Cell Mol Med 13 : 2465 – 2475 .
12. Jansen BJ , C Gilissen , H Roelofs , A Schaap-Oziemlak , J Veltman ,
RA Raymakers , J Jansen , G Kogler , CG Figdor , R Torensma and
GJ Adema . ( 2009 ). Functional differences between mesenchymal
stem cell populations are refl ected by their transcriptome . Stem
Cells Dev , 19 : 481 – 490 .
13. Greschat S , J Schira , P Küry , C Rosenbaum , MA de Souza Silva ,
G Kögler , P Wernet and HW Müller . ( 2008 ). Unrestricted somatic
stem cells from human umbilical cord blood can be differenti-
ated into neurons with a dopaminergic phenotype . Stem Cells
Dev 17 : 221 – 232 .
14. Kogler G , S Sensken and P Wernet . ( 2006 ). Comparative gener-
ation and characterization of pluripotent unrestricted somatic
stem cells with mesenchymal stem cells from human cord
blood . Exp Hematol 34 : 1589 – 1595 .
15. Sensken S , S Waclawczyk , AS Knaupp , T Trapp , J Enczmann , P
Wernet and G Kogler . ( 2007 ). In vitro differentiation of human
cord blood-derived unrestricted somatic stem cells towards an
endodermal pathway . Cytotherapy 9 : 362 – 378 .
infl uencing hematopoiesis [ 28 ], hepatocytic, and osteogenic can be mandatory to fulfi ll this function.
DLK-1 is a member of the EGF-like family that is homolo-gous to members of the Notch/Delta/Serrate family, which however lacks the characteristic DSL motif [ 19 , 20 ]. Baladron et al. were the fi rst to describe the negative regulation of Notch1 activation through dlk in murine 3T3-L1 and C3H10T1/2 cells [ 57 ]. It is known that Notch1 expression is required for adipogenesis [ 58 ] and that Notch ligands inhibit the Notch receptors present on the same cell, as well as acti-vating receptors on adjacent cells [ 59 , 60 ]. Binding of DLK-1 to Notch1 seems to inhibit the capability of cells to differentiate into adipocytes in vitro. Bray et al. were able to describe the functional interaction of Notch1 and DLK-1 despite lacking the DSL motif in Drosophila [ 61 ].
Notch2 and DLK-1 seem to be coexpressed in hepatoblasts [ 62 ] but no direct interactions of DLK-1 and Notch2–4 have been described so far. Whether the negative effect of DLK-1 expression on adipogenic differentiation of CB-derived cells is Notch-mediated needs to be further defi ned.
Therefore, we assume that DLK-1 is one of the genes refl ect-ing the undifferentiated status of USSC. Although the differ-ent nonhematopoietic CB-derived populations do not differ in the cell surface, the expression profi le is different. Data of S. Liedtke of our group document very clearly that USSC and MSC from CB are completely different in their HOX code [ 53 ]. CB MSC cluster strongly together with BM MSC, whereas USSC cluster strongly together with human embryonic stem cells. Although we could defi ne 2 out of 36 cell lines that behaved like USSC in P5 (high DLK-1 expression, no adiogenic differentiation) and turned more toward CB MSC in P10 (low DLK-1 expression, formation of adipocytes) (Supplementary Fig. 4), at the moment it is speculative whether there is a hierar-chy or precursor daughter relationship between USSC and CB MSC or if they are derived from different origins (for instance, close to fetal liver/fetal bone marrow). These results, together with the different migrating capacity based on c-MET expres-sion between the CB adherent cells [ 16 ] and immunological differences to BM MSC [ 10 , 11 ], strongly suggest that there are at least 2 distinct nonhematopoietic populations in CB that can be distinguished by their differentiation capability and by their DLK-1 expression profi le.
Acknowledgments
First of all, we would like to thank Prof. P. Bianco (Sapenzia University, Rome) for constructive discussion of the data. Thanks to Aurélie Lefort and Daniela Stapelkamp for their excellent technical support. Thanks to Valentina Bart for histological stainings. Prof. J. Eckel kindly provided human preadipocytes and adipocytes from the Deutsches Diabetes Zentrum, Duesseldorf. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) FOR 717 project Ko2119/6-1, the bilateral project Ko2119/8-1, and the German José Carreras Leukemia Foundation grant DJCLS-R07/05v to Prof. Kögler and by grants from the BMBF net-work for Inherited Bone Marrow Failure Syndromes and the DFG SPP1230 (HH) to Prof. Hanenberg.
Author Disclosure Statement
The authors do not have any commercial associations to disclose.
KLUTH ET AL. 12
34. Bianco P and PG Robey . ( 2001 ). Stem cells in tissue engineering .
Nature 414 : 118 – 121 .
35. Caplan AI . ( 2009 ). Why are MSCs therapeutic? New data: new
insight . J Pathol 217 : 318 – 324 .
36. Pittenger MF , AM Mackay , SC Beck , RK Jaiswal , R Douglas , JD
Mosca , MA Moorman , DW Simonetti , S Craig and DR Marshak .
( 1999 ). Multilineage potential of adult human mesenchymal
stem cells . Science 284 : 143 – 147 .
37. Prockop DJ . ( 1997 ). Marrow stromal cells as stem cells for non-
hematopoietic tissues . Science 276 : 71 – 74 .
38. da Silva Meirelles L , AI Caplan and NB Nardi . ( 2008 ). In search
of the in vivo identity of mesenchymal stem cells . Stem Cells
26 : 2287 – 2299 .
39. Takashima Y , T Era , K Nakao , S Kondo , M Kasuga , AG Smith
and S Nishikawa . ( 2007 ). Neuroepithelial cells supply an initial
transient wave of MSC differentiation . Cell 129 : 1377 – 1388 .
40. Dennis JE and P Charbord . ( 2002 ). Origin and differentiation of
human and murine stroma . Stem Cells 20 : 205 – 214 .
41. Crisan M , S Yap , L Casteilla , CW Chen , M Corselli , TS Park , G
Andriolo , B Sun , B Zheng , L Zhang , C Norotte , PN Teng , J Traas ,
R Schugar , BM Deasy , S Badylak , HJ Buhring , JP Giacobino , L
Lazzari , J Huard and B Peault . ( 2008 ). A perivascular origin for
mesenchymal stem cells in multiple human organs . Cell Stem
Cell 3 : 301 – 313 .
42. Manca MF , I Zwart , J Beo , R Palasingham , LS Jen , R Navarrete , J
Girdlestone and CV Navarrete . ( 2008 ). Characterization of mes-
enchymal stromal cells derived from full-term umbilical cord
blood . Cytotherapy 10 : 54 – 68 .
43. Smas CM and HS Sul . ( 1997 ). Molecular mechanisms of adi-
pocyte differentiation and inhibitory action of pref-1 . Crit Rev
Eukaryot Gene Expr 7 : 281 – 298 .
44. Kim KA , JH Kim , Y Wang and HS Sul . ( 2007 ). Pref-1 (preadipo-
cyte factor 1) activates the MEK/extracellular signal-regulated
kinase pathway to inhibit adipocyte differentiation . Mol Cell
Biol 27 : 2294 – 2308 .
45. Lee K , JA Villena , YS Moon , KH Kim , S Lee , C Kang and HS
Sul . ( 2003 ). Inhibition of adipogenesis and development of glu-
cose intolerance by soluble preadipocyte factor-1 (Pref-1) . J Clin
Invest 111 : 453 – 461 .
46. Samulewicz SJ , A Seitz , L Clark and E Heber-Katz . ( 2002 ).
Expression of preadipocyte factor-1(Pref-1), a delta-like protein,
in healing mouse ears . Wound Repair Regen 10 : 215 – 221 .
47. Yevtodiyenko A and JV Schmidt . ( 2006 ). Dlk1 expression marks
developing endothelium and sites of branching morphogenesis
in the mouse embryo and placenta . Dev Dyn 235 : 1115 – 1123 .
48. Li L , SJ Forman and R Bhatia . ( 2005 ). Expression of DLK1 in
hematopoietic cells results in inhibition of differentiation and
proliferation . Oncogene 24 : 4472 – 4476 .
49. Jensen CH , B Teisner , P Højrup , HB Rasmussen , OD Madsen ,
B Nielsen and K Skjødt . ( 1993 ). Studies on the isolation, struc-
tural analysis and tissue localization of fetal antigen 1 and its
relation to a human adrenal-specifi c cDNA, pG2 . Hum Reprod
8 : 635 – 641 .
50. Larsen JB , CH Jensen , HD Schrøder , B Teisner , P Bjerre and C
Hagen . ( 1996 ). Fetal antigen 1 and growth hormone in pituitary
somatotroph cells . Lancet 347 : 191 .
51. White JD , T Vuocolo , M McDonagh , MD Grounds , GS Harper ,
NE Cockett and R Tellam . ( 2008 ). Analysis of the callipyge phe-
notype through skeletal muscle development; association of
Dlk1 with muscle precursor cells . Differentiation 76 : 283 – 298 .
52. Waclawczyk S , S Geyh , M Kluth , U Floegel , A Buchheiser and
G Kögler . ( 2009 ). Hepatic differentiation of unrestricted somatic
stem cells (USSC) . Hum Gene Ther 20 : 1400 – 1401 .
53. Liedtke S , A Buchheiser, J Bosch , F Bosse , F Kruse , X Zhao , S
Santourlidis and G Kögler . ( 2010 ). The HOX code as a “biolog-
ical fi ngerprint” to distinguish functionally distinct stem cell
populations derived from cord blood . Stem Cell Res 5 : 40 – 50 .
54. Charbord P , M Tavian , L Humeau and B Péault . ( 1996 ).
Early ontogeny of the human marrow from long bones: an
16. Trapp T , G Kögler , A El-Khattouti , RV Sorg , M Besselmann ,
M Föcking , CP Bührle , I Trompeter , JC Fischer and P Wernet .
( 2008 ). Hepatocyte growth factor/c-MET axis-mediated tropism
of cord blood-derived unrestricted somatic stem cells for neu-
ronal injury . J Biol Chem 283 : 32244 – 32253 .
17. Kögler G , TF Radke , A Lefort , S Sensken , J Fischer , RV Sorg
and P Wernet . ( 2005 ). Cytokine production and hematopoiesis
supporting activity of cord blood-derived unrestricted somatic
stem cells . Exp Hematol 33 : 573 – 583 .
18. Hill AJ , I Zwart , HH Tam , J Chan , C Navarrete , LS Jen and R
Navarrete . ( 2009 ). Human umbilical cord blood-derived mesen-
chymal stem cells do not differentiate into neural cell types or
integrate into the retina after intravitreal grafting in neonatal
rats . Stem Cells Dev 18 : 399 – 409 .
19. Laborda J , EA Sausville , T Hoffman and V Notario . ( 1993 ). dlk, a
putative mammalian homeotic gene differentially expressed in
small cell lung carcinoma and neuroendocrine tumor cell line .
J Biol Chem 268 : 3817 – 3820 .
20. Smas CM and HS Sul . ( 1993 ). Pref-1, a protein containing EGF-
like repeats, inhibits adipocyte differentiation . Cell 73 : 725 – 734 .
21. Smas CM , L Chen and HS Sul . ( 1997 ). Cleavage of membrane-
associated pref-1 generates a soluble inhibitor of adipocyte dif-
ferentiation . Mol Cell Biol 17 : 977 – 988 .
22. Floridon C , CH Jensen , P Thorsen , O Nielsen , L Sunde , JG
Westergaard , SG Thomsen and B Teisner . ( 2000 ). Does fetal
antigen 1 (FA1) identify cells with regenerative, endocrine and
neuroendocrine potentials? A study of FA1 in embryonic, fetal,
and placental tissue and in maternal circulation . Differentiation
66 : 49 – 59 .
23. Gregoire FM , CM Smas and HS Sul . ( 1998 ). Understanding adi-
pocyte differentiation . Physiol Rev 78 : 783 – 809 .
24. Abdallah BM , CH Jensen , G Gutierrez , RG Leslie , TG
Jensen and M Kassem . ( 2004 ). Regulation of human skeletal
stem cells differentiation by Dlk1/Pref-1 . J Bone Miner Res
19 : 841 – 852 .
25. Moon YS , CM Smas , K Lee , JA Villena , KH Kim , EJ Yun and
HS Sul . ( 2002 ). Mice lacking paternally expressed Pref-1/Dlk1
display growth retardation and accelerated adiposity . Mol Cell
Biol 22 : 5585 – 5592 .
26. Tanimizu N , M Nishikawa , H Saito , T Tsujimura and A
Miyajima . ( 2003 ). Isolation of hepatoblasts based on the expres-
sion of Dlk/Pref-1 . J Cell Sci 116 ( Pt 9 ): 1775 – 1786 .
27. Anisimov SV , KV Tarasov , D Tweedie , MD Stern , AM Wobus
and KR Boheler . ( 2002 ). SAGE identifi cation of gene transcripts
with profi les unique to pluripotent mouse R1 embryonic stem
cells . Genomics 79 : 169 – 176 .
28. Moore KA , B Pytowski , L Witte , D Hicklin and IR Lemischka .
( 1997 ). Hematopoietic activity of a stromal cell transmembrane
protein containing epidermal growth factor-like repeat motifs .
Proc Natl Acad Sci USA 94 : 4011 – 4016 .
29. Radke TF , A Buchheiser , A Lefort , M Maleki , P Wernet and
G Kogler . ( 2007 ). GMP-conform generation and cultivation of
unrestricted somatic stem cells (USSC) from cord blood using
the SEPAX-separation method and a closed culture system
applying cell stacks . Blood 110 : 367 .
30. Liedtke S , J Enczmann , S Waclawczyk , P Wernet and G Kogler .
( 2007 ). Oct4 and its pseudogenes confuse stem cell research .
Cell Stem Cell 1 : 364 – 366 .
31. Liedtke S , M Stephan and G Kögler . ( 2008 ). Oct4 expression
revisited: potential pitfalls for data misinterpretation in stem
cell research . Biol Chem 389 : 845 – 850 .
32. Takahashi K , K Tanabe , M Ohnuki , M Narita , T Ichisaka , K
Tomoda and S Yamanaka . ( 2007 ). Induction of pluripotent stem
cells from adult human fi broblasts by defi ned factors . Cell
131 : 861 – 872 .
33. Greenberg AS , JJ Egan , SA Wek , NB Garty , EJ Blanchette-Mackie
and C Londos . ( 1991 ). Perilipin, a major hormonally regulated
adipocyte-specifi c phosphoprotein associated with the periph-
ery of lipid storage droplets . J Biol Chem 266 : 11341 – 11346 .
DLK-1 DISTINGUISHES USSC FROM MSC IN CB 13
61. Bray SJ , S Takada , E Harrison , SC Shen , and AC Ferguson-Smith .
( 2008 ). The atypical mammalian ligand Delta-like homologue
1 (Dlk1) can regulate Notch signalling in Drosophila . BMC Dev
Biol 8 : 11 .
62. Tanimizu N and A Miyajima . ( 2004 ). Notch signaling con-
trols hepatoblast differentiation by altering the expression of
liver-enriched transcription factors . J Cell Sci 117 ( Pt
15 ): 3165 – 3174 .
immunohistochemical study of hematopoiesis and its microen-
vironment . Blood 87 : 4109 – 4119 .
55. Wang Y and HS Sul . ( 2009 ). Pref-1 regulates mesenchymal cell
commitment and differentiation through Sox9 . Cell Metab
9 : 287 – 302 .
56. Gimble JM , CE Robinson , X Wu and KA Kelly . ( 1996 ). The func-
tion of adipocytes in the bone marrow stroma: an update . Bone
19 : 421 – 428 .
57. Baladrón V , MJ Ruiz-Hidalgo , ML Nueda , MJ Díaz-Guerra , JJ
García-Ramírez , E Bonvini , E Gubina and J Laborda . ( 2005 ). dlk
acts as a negative regulator of Notch1 activation through interac-
tions with specifi c EGF-like repeats . Exp Cell Res 303 : 343 – 359 .
58. Garcés C , MJ Ruiz-Hidalgo , J Font de Mora , C Park , L Miele , J
Goldstein , E Bonvini , A Porrás and J Laborda . ( 1997 ). Notch-1
controls the expression of fatty acid-activated transcrip-
tion factors and is required for adipogenesis . J Biol Chem
272 : 29729 – 29734 .
59. Micchelli CA , EJ Rulifson and SS Blair . ( 1997 ). The function and
regulation of cut expression on the wing margin of Drosophila :
Notch, Wingless and a dominant negative role for Delta and
Serrate . Development 124 : 1485 – 1495 .
60. de Celis JF and SJ Bray . ( 2000 ). The Abruptex domain of Notch
regulates negative interactions between Notch, its ligands and
Fringe . Development 127 : 1291 – 1302 .
Received for publication February 7, 2010 Accepted after revision March 23, 2010
Prepublished on Liebert Instant Online March 23, 2010
Address correspondence to: Prof. Dr. Gesine Kögler
Institute for Transplantation Diagnostics and Cell Therapeutics University of Duesseldorf Medical Center
Moorenstrasse 5 D-40225 Duesseldorf
Germany
E-mail : [email protected]
Supplementary Data
Supplementary Materials and Methods
Senescence assay
The senescence-associated β-galactosidase (SA-β-Gal) is histochemically detectable in cytosol of senescent cells at pH 6. Cells were incubated with 5-bromo-4-chloro-3-indolyl-β-d-galactoside (X-Gal), a substrate of SA-β-Gal. Senescence was detected in single cells by X-Gal. For positive control, the lysosomal galactosidase, active at about pH 4, was applied. Therefore, the cells were treated with a pH 4 solution. The 1 × 104 cells were plated on chamber slides (Lek-Tek, Chamber Slide™, 2-Well Glass Slide) and after 3 days fi xed in 4% form-aldehyde. Cells were covered with 1,000 μL senescence-asso-ciated staining solution, control cells were covered with 1,000 μL positive control solution. All cells were incubated over-night at 37°C. Finally, cells were stained with Nuclear Fast Red (Certistain, Merck). Evaluation was performed by count-ing the senescent cells using a light optical microscope.
Immunophenotyping
Flow cytometry was performed on BD FACSCanto fl ow cytometer employing the FACSDIVA software (Version 5.0.3., BD biosciences) for recording and WinMDI 2.8 for
analysis. The 1 × 105 cells in 100 μL phosphate-buffered saline (PBS) were incubated for 30 min with 5 μL antibody (Supplementary Table 2) and washed with PBS; for fi xation 4% PFA was used.
For intracytoplasmic staining, cells were fi xed and per-meabilized utilizing the BD Cytofi x/Cytoperm kit (BD Biosciences).
Western blot
Total protein was isolated with RL-lysis buffer (20 mM Tris, pH 7.4, 140 mM NaCl, 10 mM NaF, 10 mM Na pyro-phosphate, 1% Triton X-100, 1 mM ethylenediaminetetra-acetic acid (EDTA), 1 mM ethylene glycol tetraacetic acid (EGTA), 1.6 mM Na vanadate, 20 mM β-glycerol phosphate, Complete Mini Protease Inhibitor Cocktail). Western blot was performed utilizing the NuPAGE® System (Invitrogen, Carlsbad, CA). Gel electrophoresis was performed utilizing 4%–12% Bis–Tris gels and 1× 2-(N-morpholino)ethanesul-fonic acid (MES) running buffer. Proteins were blotted on nitrocellulose membranes and blocked with 5% bovine serum albumin (BSA) in phosphate-buffered saline Tween 20 (PBST). The primary antibody (Supplementary Table 2) was applied in a 1:500 dilutions. The second antibody (Supplementary Table 2) was diluted 1:10,000. Proteins were detected applying ECL PLUS Western Blotting Detection Reagents (GE Healthcare).
SUPPLEMENTARY FIG. 1. (A) Unrestricted somatic stem cells (USSC)-derived clonal populations (37.3 ± 4.1 CPD) reached higher CPDs at the same passage compared with USSC (32.4 ± 2.2 CPD) in P5 (P = 0.0214), as well as in P9 (USSC clonal populations 50 2.4 CPD, USSC 43.14 ± 4.1) (P = 0.0001). In P5 cord blood mesenchymal stromal cells (CB MSC)-derived clonal populations reached 40 0.78 CPD, while CB MSC reached only 29 ± 2.8 CPD (P < 0.0001). In P.9 growth kinetics of CB MSC-derived clonal populations (46.72 ± 3.2 CPD) and CB MSC (37.8 ± 4.1 CPD) were signifi cantly differ-ent (P = 0.0022). (B) β-Galactosidase staining indicating se-nescent cells was exemplifi ed for USSC (n = 8), CB MSC (n = 3), and BM MSC (n = 6) after P9 (43.6.4 CPD). In contrast to BM MSC, which contains 95% senescent cells in P9 (26.2.7 CPD), CB MSC and USSC exhibited fewer senescent cells. As posi-tive control, the activity of the lysosomal β-galactosidase was shown at pH 4.0. Scale bar = 100 μm.
A
B
Growth Kinetics of USSC and Corresponding Clones
Growth Kinetics of CB MSC and Corresponding Clones
Senescence
US
SC
2C
B M
SC
3B
M M
SC
4
Positive Control
0
0
0 20 40 60 80
Days
CP
D
100 120 140
10
20
30
40
50
60
0
21
29
37
46
51
59
Days of Culture
Initial USSC
Initial CB MSC
Clone A2
Clone B3
Clone B1
Clone A1
Clone A8
Clone A5
Clone A4
Clone B4
Clone C1
Clone B5
Clone D9
Clone A10
Clone C2
Clone B1
66
75
82
90
10
0
10
6
10
20
30
40
CP
C
50
60
70
SUPPLEMENTARY FIG. 2. CB-derived cells were analyzed for representative antigens: unrestricted somatic stem cells (USSC), cord blood mesenchymal stromal cells (CB MSC), and clonal USSC, MSC populations were negative for the antigens CD31, CD34, CD45, CD56, CD106, CD133/2, CD144, and HLA-DR (A). The expression of CD13, CD29, CD44, CD 71, CD73, CD105, CD146, CD166, and HLA-ABC was always positive in the different CB-derived populations (B). Cells were labeled with the mAb specifi c for the molecules indicated (fi lled histograms) or isotype controls (open histograms).
A
B
USSC
CD31
CD13 CD105
CD29 CD146
CD44 CD166
CD71 HLA-ABC
CD73
CD34
CD45
CD56
Negative E
xpre
ssio
nP
ositiv
e E
xpre
ssio
n
CD106
CD133/2
CD144
HLA-DR
CB MSC USSCClones
CB MSCClones
USSC CB MSC USSCClones
CB MSCClones
USSC CB MSC USSCClones
CB MSCClones
USSC CB MSC USSCClones
CB MSCClones
SUPPLEMENTARY FIG. 3. (A) Expression of the stem cell markers OCT4A, NANOG, SOX2, and the transcription factors KLF4 and c-MYC. RT-PCR analysis of different cell lines (n = 8 unrestricted somatic stem cells (USSC), n = 7 cord blood mesenchymal stromal cells (CB MSC)) was performed in a minimum of 2 different passages. USSC (Lanes 1–7), BM MSC (Lanes 8–14), and nTERA-2 (Lane 15), RT− (Lane 16). OCT4, NANOG, and SOX2 could never be detected in any of the cell populations tested. (B) To increase the detection level for OCT4A expression, comparative real-time PCR was performed. Selected CB-derived cells (3 USSC, 2 CB MSC) expressed the same amount of OCT4A as human dermal fi broblasts (NHDF). (C) Immunohistochemical staining of OCT4A and NANOG in representative samples of USSC (n = 4), BM MSC (n = 1, con-trol), and CB MSC (n = 2). As a positive control, the teratocarcinoma cell line nTERA-2 was used. OCT4 and NANOG were not expressed in USSC and BM MSC. The embryonic carcinoma cell line nTERA-2 revealed positive nuclear staining for both markers. Scale bar = 200 μm.
A C
B
1OCT4
OCT4
US
SC
1C
B M
SC
1U
SS
C1 C
lone A
1B
M M
SC
2N
TE
RA
2
Nanog
456bp
Nanog
495bp
SOX2
466bp
cMYC
472bp
KLF4
409bp
GAPDH
228bp
0.00000
0.00025
0.00050
0.00075
Rela
tive O
CT
4 E
xpre
ssio
n
0.00100
0.5
1.0
USSC
Line1
USSC
Line2
USSC
Line3
USSC
Line4
USSC
Line5
BM-M
SC L
ine1
BM-M
SC L
ine2
NHDF
nTER
A
ESC H
8.1
5 10 15
SUPPLEMENTARY FIG. 4. (A) Osteogenic and adipogenic differentiation of unrestricted somatic stem cells (USSC), cord blood mesenchymal stromal cells (CB MSC), and clonal populations. (B) The differentiation potential of bulk cell lines in P10 was weaker than in P5/P6. Compared with the initial cell lines, the differentiation potential of the clonal populations was even more restricted. (C) Scheme of adipogenic differentiation. Preadipocytes are derived from a mesendodermal/ec-todermal precursor. These cells express high amounts of DLK-1. DLK-1 down-regulation is accompanied by up-regulation of PPARs and CEBPs. These transcription factors regulate the expression of adipogenic specifi c genes like ADIPOQ, FABP4, and PLIN. Premature white adipocytes are characterized by various lipid vacuoles, whereas mature white adipocytes con-tain just one single lipid vacuole. (D) RT-PCR of PPARγ, CEBPα, ADIPOQ, FABP4, PLIN, and GAPDH after 14 days of adipo-genic induction. Positive signals for PPARγ, FABP4, and PLIN could be demonstrated in induced CB MSC (n = 7 tested, red frame), BM MSC (positive control, yellow frame) and in human preadipocytes (positive control). ADIPOQ expression was hardly detectable in CB MSC. CEBPα was equally expressed in all cell lines tested. No expression of any adipogenic specifi c genes was found in USSC (n = 6 tested, blue frame), dermal fi broblasts, or in the noninduced controls. “+” induced samples, “−“ noninduced samples. (E) Western blot analysis detecting human PLIN on protein level after 21 days of adipogenic in-duction. PLIN could be detected in CB MSC (red frame) and BM MSC (yellow frame), preadipocytes, adipocytes (as positive controls), but not in the noninduced controls or in USSC (blue frame). As negative control, human dermal fi broblasts (NHDF) were used. β-Actin was used as loading control for all samples. “+” induced samples, “−“ noninduced samples. (F) Western blot analysis detecting human DLK-1. After 21 days of adipogenic induction, high DLK-1 expression could be detected in noninduced USSC. After adipogenic induction, it was down-regulated in USSC (blue frame). CB MSC (red frame) and BM MSC (yellow frame) were DLK-1− or weakly positive. β-Actin was applied as loading control for all samples. “+” induced samples, “−“ noninduced samples.
A
C
E
F
D
BUSSC1
P9USSC1
Clone A5 P9
Adipogenesis
US
SC
1 +
P1
0
PPARy 351 bp
CEBPα 184 bp
ADIPOQ 362 bp
FABP4 280 bp
PLIN 488 bp
GAPDH 228 bp
CB
MS
C 1
+ P
10
US
SC
1 –
P1
0
CB
MS
C 1
– P
10
BM
MS
C 2
+ P
7
BM
MS
C 2
– P
7
Ad
ipo
cyte
s
Pre
ad
ipo
cyte
s+ n
TE
RA
-2
US
SC
1 –
US
SC
1 +
BM
MS
C1
+
BM
MS
C1
–
CB
MS
C1
–
Pre
ad
ipo
cyte
s –
Pre
ad
ipo
cyte
s +
CB
MS
C1
+
Ad
ipo
cyte
s
Skin
Fib
rob
lasts
PLIN
β-Actin
DLK-1
β-Actin
Pre
ad
ipo
cyte
s–
CB
MS
C1
+
US
SC
1 –
US
SC
1 +
BM
MS
C1
+
CB
MS
C1
–
BM
MS
C1
–
Skin
Fib
rob
lasts
ES
Mesendodermal/Ectodermal Precursor
Hypothetical
Neural Crest(Takashimaet al. 2007)
DLK-1-Fetal Liver
Preadipocyte
DLK-1
PPAR, CEBP
ADIPOQ
FABP4
PLINAdipocyte
PrematureAdipocyte
MSC
?
MSC
?
CB MSC 4P10
CB MSC 1 CloneB4 P9 P5 P10 Undifferentiated
SUPPLEMENTARY FIG. 5. (A) Immunohistochemical staining of neurofi lament and β-3-tubulin in unrestricted somatic stem cells (USSC) (n = 16) and cord blood mesenchymal stromal cells (CB MSC) (n = 4) after 14 days of neuronal induction. Neural morphology could be detected in USSC but never in CB MSC. (B) Real-time PCR detecting neurofi lament light, me-dium and heavy chain, as well as β-3-tubulin in USSC after 14 days of neural induction. A clear up-regulation of all 4 genes could be detected for USSC.
A
B
Neurofilament
d02 (
∆ C
tNF
-L/ ∆
CtG
AP
DH
)2 (
∆ C
tNF
-H/
∆C
tGA
PD
H)
2 (
∆ C
t β3
/ ∆
Ct
GA
PD
H)
2 (
∆ C
tNF
-M/ ∆
CtG
AP
DH
)
Neurofilament, Light Polypeptide
Neurofilament, Heavy Polypeptide β-3-Tubulin
Neurofilament, Medium Polypeptide
0 0
1
2
50
100
0
5
10
0
2
4
d7 d14
d0 d7 d14 d0 d7 d14
d0 d7 d14
US
SC
5C
B M
SC
1
β-3 Tubulin
200 μm 200 μm
200 μm 200 μm
SUPPLEMENTARY FIG. 6. (A) Successful transfection of pCL7Egwo_DLK-1 was evaluated by transfection with the pCL7Egwo plasmid containing eGFP instead of DLK-1. A high eGFP signal could be observed in the cells transduced with the pCL7Egwo plasmids, while no eGFP signal could be detected in the cells transduced with pCL7Egwo_DLK-1 analyzed by fl uorescence microscopy and fl uorescence-activated cell sorting (FACS) analysis. Scale bar = 200 μm. (B) Vector map of pCL7Egwo. The eGPF gene was replaced by the full-length DLK-1 gene. (C) Vector map of pKLO.1-puro, containing a puro-mycin resistance gene. Transfected cells were selected by puromycin.
A
B C
CB MSC1 + pCL7EGwo
101
0
20
40
60
80
100
102
Comp-FITC-A
Sample NameEOFP_2fca
EOFP_Non EOFP
% o
f M
ax
103
104
105
CB MSC1 + pCL7EGwo + DLK-1
pCL7EGwo8378 bp
Sense Strand
Sense Strand
Antisense Strand
Antisense Strand
pLKO.1-puro7086 bp
hPGK
puroR
SIN/3′ LTR
f1 ori
ampR
pUC ori
RSV/5′ LTR
(Ψ) Psi
cpptU6
RRE
5′-3′- UU
Su
pplem
en
tar
y T
abl
e 1
.
Pri
mer
Gen
eSe
quen
ceP
rim
er fo
rwar
d 5′
–3′
Pri
mer
rev
erse
5′–
3′P
rodu
ct
RT-
PC
R
PPA
Rγ
NM
_01
58
69.3
PPA
Rg
_fo
r: G
CT
GT
TA
TG
GG
TG
AA
AC
TC
TG
PPA
RG
_re
v:
AT
AA
GG
TG
GA
GA
TG
CA
GG
CT
C3
51
PLI
NN
M_
0026
66.3
PL
IN_
for:
CT
CA
CC
TT
GC
TG
GA
TG
GA
GA
PL
IN_
rev
: C
GA
GT
GT
TG
GC
AG
CA
AA
TT
C4
88
FAB
P4N
M_
001
44
2.1
FAB
P4
_fo
r: G
CT
TT
GC
CA
CC
AG
GA
AA
GT
GFA
BP
4_
rev
: A
TG
AC
GC
AT
TC
CA
CC
AC
CA
G2
80
AD
IPO
QN
M_
00
479
7.2
AD
IPO
Q_
for:
TT
CT
GA
TT
CC
AT
AC
CA
GA
GG
AD
IPO
Q_
rev
: G
GT
AT
AC
AT
AG
GC
AC
CT
TC
TC
362
DLK
-1N
M_
00
38
36.4
DL
K1_
for:
AC
GG
GG
AG
CT
CT
GT
GA
TA
GA
DL
K1_
rev
: G
CT
TG
CA
CA
GA
CA
CT
CG
TA
G4
68
CE
BPa
NM
_0
04
36
4.2
C
EB
Pa
_fo
r: G
AG
TC
AC
AC
CA
GA
AA
GC
TA
GC
EB
Pa
_re
v: G
AT
GG
AC
TG
AT
CG
TG
CT
TC
184
Oct
4 N
M_
0027
01O
CT
4_
for:
AG
CC
CT
CA
TT
TC
AC
CA
GG
CC
OC
T4
_re
v: T
GG
GA
CT
CC
TC
CG
GG
TT
TT
G4
56
Nan
ogN
M_
024
86
5N
AN
_fo
r: T
GA
GA
TG
CC
TC
AC
AC
GG
AG
NA
N_
rev
: T
TG
CT
CC
AG
GT
TG
AA
TT
GT
TC
495
SOX
2N
M_
00
310
6S
OX
2_
f: A
GA
AC
CC
CA
AG
AT
GC
AC
AA
CS
OX
2_
r: A
TG
TA
GG
TC
TG
CG
AG
CT
GG
T4
66
KLF
4N
M_
00
42
35
KL
F4
_f:
AG
AA
GG
AT
CT
CG
GC
CA
AT
TT
KL
F4
_r:
GG
TC
TC
TC
TC
CG
AG
GT
AG
GG
40
9
c-M
YC
NM
_0
024
67M
yc_
for1
: G
AG
GC
TA
TT
CT
GC
CC
AT
TT
My
c_re
v1:
GA
AA
CT
CT
GG
TT
CA
CC
AT
GT
C47
2
Hnf
4aN
M_
00
04
57, N
M_1
788
50,
NM
_178
849
,
NM
_0
010
30
00
4, N
M_1
759
14,
NM
_0
010
30
00
3
TA
C T
CC
TG
C A
GA
TT
T A
GC
CG
GT
C A
TT
GC
C T
AG
GA
G C
AG
C4
65
Hnf
1aN
M_
00
05
45
CA
G A
GC
CA
T G
TG
AC
C C
AG
AG
TG
A G
GT
GA
A G
AC
CT
G C
TT
GG
195
HSA
NM
_0
00
47
7.5
AA
A G
CC
TT
G G
TG
TT
G A
TT
GC
GT
C A
GC
CA
T T
TC
AC
C A
TA
GG
210
GY
S2N
M_
021
957
.3G
TT
AT
A C
TC
CA
G C
TG
AA
T G
CA
CA
T G
CT
GG
T A
AT
AT
C T
GC
CT
A2
85
FBP
1N
M_
00
05
07.
3T
GC
CG
T C
AC
TG
A G
TA
CA
T C
CG
CT
AA
C A
AG
AA
G A
GC
CC
C A
A15
3
AR
G1
NM
_0
00
04
5.2
GG
A A
AC
TT
G C
AT
GG
A C
AA
CC
CC
A T
CA
CC
T T
GC
CA
A T
TC
C2
26
PPA
Rγ
NM
_01
58
69.3
PPA
Rg
_fo
rRT
: T
CC
AT
GC
TG
TT
AT
GG
GT
GA
AP
PA
Rg
_re
vR
T: T
CA
AA
GG
AG
TG
GG
AG
TG
GT
C19
3
GA
PD
HN
M_
0024
6G
AP
DH
_fo
rRT
: G
AG
TC
AA
CG
GA
TT
TG
GT
CG
TG
AP
DH
_re
vR
T: T
TG
AT
TT
TG
GA
GG
GA
TC
TC
G2
28
Ov
erex
pre
ssio
n
DLK
-1N
M_
00
38
36.4
DL
K-K
lon
f:
CC
GC
CA
CT
CG
AG
GC
CA
CC
AT
G
AC
CG
CG
AC
CG
AA
GC
CC
TC
DL
K-K
lon
r: T
AG
CT
TG
AA
TT
CT
TA
GA
TC
T
CC
TC
GT
CG
CC
GG
CC
1,15
1 b
p
Taq
Ma
n A
ssay
s
DLK
-1S
up
erA
rrra
yP
PH
0074
5B-2
00
GA
PD
HS
up
erA
rrra
yP
PH
001
50E
-20
0
NF,
lig
ht
po
lyp
epti
de
Su
per
Arr
ray
HS
001
962
45
-ml
NF,
med
ium
po
lyp
epti
de
Su
per
Arr
ray
HS
001
9357
2-m
l
NF,
hea
vy
po
lyp
epti
de
Su
per
Arr
ray
HS
00
60
6024
-ml
β-3-
Tubu
linS
up
erA
rrra
yH
S0
08
0139
0-m
l
Supplementary Table 2. Antibodies Used in this Study
Fluorescence-activated cell sorting (FACS) analysis
Conjugated Clone Company
CD13 PE L138 BDCD29 FITC K20 ImmunotechCD31 FITC 5.6E ImmunotechCD34 FITC 581 ImmunotechCD44 FITC J.173 ImmunotechCD45 FITC 2D1 BDCD56 PE NKH-1 ImmunotechCD71 FITC YDJ1.2.2 ImmunotechCD73 PE AD2 BDCD105 PE 166707 R&DCD106 FITC BBIG-V3 (IE10) BDCD133/2 PE 51-10C9 MiltenyiCD166 PE 3A6 BDCD184 PE 12G5 BDHLA-ABC FITC B9.12.1 ImmunotechHLA-DR PE G46-6 BDUnconjugated Clone CompanyCD146 128018 R&DSecondary antibody CompanyFITC-conjugated goat anti-mouse Beckman/CoulterImmunhistochemistry and western blotPrimary antibody Clone CompanyDLK-1/PREF-1 LC-12 Santa CruzNANOG MAB1997 R&DOCT4 C-10 Santa CruzHSA HSA-11 Sigma AldrichPLIN Polyclonal Novus Biologicalsβ-3-Tubulin MAB 1637 ChemiconNF neurofi lament NA 1297-0100 BiotrendSecondary antibody CompanyRhodamine Red™-X-conjugated goat anti-mouse IgG DianovaFITC-conjugated goat anti-mouse DianovaFITC-conjugated goat anti-rabbit Dianovagoat anti-mouse IgG_HRP Santa Cruzgoat anti-rabbit IgG-HRP Santa Cruz