STEM CELLS AND DEVELOPMENT 17:1095–1108 (2008)
© Mary Ann Liebert, Inc.
DOI: 10.1089/scd.2007.0154
Comparison of Human Placenta- and Bone Marrow–Derived Multipotent Mesenchymal Stem Cells
Sarah Barlow,1 Gary Brooke,1 Konica Chatterjee,1 Gareth Price,2 Rebecca Pelekanos,1 Tony Rossetti,1 Marylou Doody,3 Deon Venter,2,4,5 Scott Pain,6 Kristen Gilshenan,6 and Kerry Atkinson1,5
Bone marrow is the traditional source of human multipotent mesenchymal stem cells (MSCs), but placenta appears to be an alternative and more readily available source. This study comprehensively compared human placenta–derived MSC (hpMSC) and human bone marrow–derived MSC (hbmMSC) in terms of cell characteristics, op-timal growth conditions and in vivo safety specifi cally to determine if hpMSC could represent a source of human MSC for clinical trial. MSC were isolated from human placenta (hpMSC) and human bone marrow (hbmMSC) and expanded ex vivo using good manufacturing practice–compliant reagents. hpMSC and hbmMSC showed similar proliferation characteristics in different basal culture media types, fetal calf serum (FCS) concentrations, FCS heat-i
nactivation experiments, fl ask types and media replacement responsiveness. However, hpMSC and hbmMSC differed with respect to their proliferation capabilities at different seeding densities, with hbmMSC proliferating more slowly than hpMSC in every experiment. hpMSC had greater long-term growth ability than hbmMSC. MSC from both sources exhibited similar light microscopy morphology, size, cell surface phenotype, and mesodermal differentiation ability with the exception that hpMSC consistently appeared less able to differ-entiate to the adipogenic lineage. A comparison of both hbmMSC and hpMSC from early and medium passage cultures using single-nucleotide polymorphism (SNP) GeneChip analysis confi rmed GTG-banding data that no copy number changes had been acquired during sequential passaging. In three of three informative cases (in which the gender of the delivered baby was male), hpMSC were of maternal origin. Neither hpMSC nor hbmMSC caused any acute toxicity in normal mice when injected intravenously at the same, or higher, doses than those currently used in clinical trials of hbmMSC. This study suggests that human placenta is an acceptable alternative source for human MSC and their use is currently being evaluated in clinical trials.
Introduction
T he mesenchymal stem cell (MSC) is a stem cell located within the stroma of the bone marrow and
other organs including placenta. They have been phenotypically charac-terized using a variety of markers [1–3]. When isolated by plastic adherence and expanded ex vivo, these cells have been shown to differentiate into cell types of mesenchymal origin including chondrocytes, adipocytes, and osteocytes [1]. In the bone marrow they provide support for hematopoiesis [4]. In addition, they are able to differentiate into endothe-lial cells, form capillaries in vitro and secrete growth factors important in angiogenesis including vascular endothelial growth factor [1]. It has also been shown that MSC demon-strate plasticity beyond their traditional mesodermal lineage, in that they have been induced to generate tissues of both ectodermal (neurons) and endodermal (hepatocytes) nature [5,6]. In support of these observations, undifferentiated MSC express many lineage-specifi c genes other than those of mes-enchymal lineage [7]. Their ability to differentiate into a wide variety of cell types, together with their reproducibility of isolation, high expansion potential and capacity for useful modifi cation using molecular biological engineering tech-niques, make them good candidates for the repair and regen-eration of a large variety of tissues. They have been shown
1Adult Stem Cell Laboratory, Biotherapy Program, and 2Molecular Genetics Laboratory, Mater Medical Research Institute, Brisbane, Queensland, Australia.
3Cytogenetics, Mater Health Services Pathology, and 4Pathology, Mater Health Services, Brisbane,
Queensland, Australia.
5The University of Queensland, Brisbane, Queensland, Australia.
6Mater Research Support Centre, Brisbane, Queensland, Australia.
1095
BARLOW ET AL.
1096
biohazard cabinet the cord blood was drained and the
建筑钢模u mbilical cord and external membranes were removed. Placental tissue (including amnion, chorion and desidua basalis) was then dissected into pieces ~5 g in size (300 g in total). This tissue was placed in beakers and washed with 500 mL HBSS/100 g tissue. The pieces of placenta were then divided equally between 50 mL tubes at ~10 g/tube. Dulbecco’s modifi ed Eagle’s medium–Low G lucose (DMEM-LG; JRH Biosciences) with 100 U /mL Collagenase, type I (Worthington Biochemical Corporation) and 5 μg/mL DNase I (Roche) was added to each tube up to a total volume of 50 mL. T ubes were incubated on a shaker (220 rpm, 37°C, 2 h), then pulse c entrifuged at 540g to remove large partic-ulate matter and the cell suspensions were passed through 70 μm fi lters (Becton-Dickinson). The remaining tissue was washed with HBSS and the resulting cell suspension also fi ltered through 70 μm fi lters. The combined fi ltered cells were centrifuged (540g, 5 min, 20°C) and resuspended in 30 mL HBSS, and 12 mL 1.073 g/mL Percoll™ was underlayed. Samples were centrifuged (540g, 20°C, 20 min) and the inter-face removed and washed twice with HBSS (fi rst at 540g , 20°C, 10 min and second at 300g, 20°C, 5 min). The mononu-clear cells were then plated at 2–4 × 105 cells/cm 2 in tissue culture fl asks for ex vivo expansion.
Cell culture
Depending upon the experiment bone marrow or pla-cental cells were plated into either 25 or 75 cm 2 tissue cul-ture fl asks (Nunc). The standard tissue culture medium was basal media (e.g., DMEM-LG ) with 20 % (v /v) Australian-sourced FCS and 50 μg/mL gentamicin (Pharmacia). Initial experiments compared fi ve different lots of nonheat-inactivated FCS (Invitrogen; catalogue number: 10099-141; lots: 560339 and 1188431; Invitrogen; catalogue number: 10099-158; lot: 1180078; and JRH Biosciences; catalogue num-ber: 12003-500M; lot: 4J0046). One of these (12003-500M) was selected for all subsequent experiments. Cultures were incu-bated in humidifi ed 5 % CO 2 incubato
rs and the media was replaced twice a week. Cells were isolated by plastic adher-ence in expansion cultures and nonadherent cells were washed off the cultures after 3 days.
Cell cultures were passaged when 90–95% confl uent. For passaging, the fl asks were washed with HBSS and in-cubated with TrypLE™ Select (Invitrogen) for 5–10 min at 37°C. Dissociated cells were removed, then pelleted by centrifugation (540g, 5 min, 4°C). The supernatant was dis-carded and cells were resuspended in tissue culture media. At p assage 1, hbmMSC and hpMSC were seeded into three replicate 25 cm 2 fl asks per experiment. From passage 2, cells were seeded from one 25 cm 2 fl ask to another 25 cm 2 fl ask.For the experiments comparing basal media, FCS
c oncentration (20 or 10% v /v) an
d heat inactivation of FCS (60 min at 56°C), hpMSC and hbmMSC wer
e plated at 3–4 × 104 cells/fl ask (1,600 cells/cm 2) at every pas-sage in each experiment from passage 1 onwards. MSC were cultured in one o
f fi ve different basal culture media: DMEM-LG , Minimum Essential Medium, α-Modifi cation (α-MEM, JRH Biosciences; catalogue number: 51451-500M), DMEM-High G lucose (DMEM-HG ; Invitrogen; catalogue
in preclinical studies to improve myocardial function (after acute myocardial infarction), cerebral function (after cerebral infarction), liver and joint damage [8–11].
Importantly, MSC appear to have a major advantage over many other cell types for cellular therapy, in that they are immunologically privileged and even in large outbred animals can generally be transplanted across major histo-compatibility complex (MHC) barriers without the need for immune suppression [12]. The mechanism for this is not fully understood at present, but appears to be an active process that leads to suppression of T-cell function [13–15]. This has important implications for the therapeutic appli-cation of MSC, because MSC derived from healthy unre-lated volunteer donors can be cryopreserved, thus making them available in a timely manner for patients in a variety of acute and chronic clinical settings. MHC-identical, MHC-haploidentical, and MHC-unmatched MSC have been used successfully in the clinic [16–18].
Bone marrow is the traditional source of human MSC. However, human MSC have been generated from a wide variety of tissues and organs including placenta [19–21], cord blood [22], amnion [23], amniotic fl uid [24], fat [25], lung [26], and liver [27]. Most of these sources, including large volumes of normal bone marrow, are relatively diffi cult to access as a tissue source for the isolation of MSC. In contrast, placenta is readily and widely available. To determine whether human placenta–derived
MSCs (hpMSCs) represent an appropriate alternative source of human MSC for use in clinical trials, we optimized ex vivo expansion conditions for both types of MSC. The literature describes a number of differences in the culture conditions described to generate human MSC, including various types of basal media, concentrations of fetal calf serum (FCS), with and without heat inactivation of FCS, seeding densities, fl ask types, and medium replace-ment schedule. We explored all these variables for generat-ing hpMSC and bone marrow–derived MSC (hbmMSC). We also compared the following biological characteristics of the two types of MSC: morphology, size, cell surface phenotype, mesodermal differentiation ability, karyotype, single-nucle-otide polymorphism (SNP), and in vivo safety in murine recipients. Wherever possible we used good manufacturing practice–compliant reagents, so that the protocols were able to be directly translated to the clinical trial setting.
Materials and Methods Cell harvest
Bone marrow. Human bone marrow (3–5 mL) obtained from the iliac crest of healthy adult donors was diluted 1:5 with Hank’s Balanced Salt Solution (HBSS) (Invitrogen). The cell suspension was underlayed with 12 mL 1.073 g/mL Percoll™ (GE Healthcare) and centrifuged (540g , 20°C, 20 min) without a brake. Cells from the interface layer were washed twice with HBSS (fi rst at 540g , 20°C, 10 min, then 250g , 20°C, 5 min). The mononuclear cells were plated at a density of between 0.5 an
d 1 × 105 cells/cm 2 in tissue culture fl asks for ex vivo expansion.
Placenta. Human placentas were obtained from healthy mothers during routine Caesarian section births. In a
COMPARISON OF HUMAN PLACENTA AND BONE MARROW MSC 1097
30 s denaturation at 94°C, 30 s hybridization at 56°C, 1 min
e longation at 68°C. Product size and primer sequences used were, 5′–3′: Perilipin (PLIN adipogenic d ifferentiation marker) 123 bp (5′)AAACAGCATCAGCGTTCCCATC (3′)AG TG TTG G CAG CAAATTCCG , Runt Related Transcription Factor 2 (RUNX2 osteogenic differentiation marker) 114 bp (5′)C G CAAAACCACA G AACCACAA G T G C G (3′)GTTGG TCTC G G T G G CT G G TA G , Aggrecan (ACAN chondro-genic differentiation marker) 150 bp (5′)CGGGTCTCACTG CCCAACTACCC G (3′)GCCTTTCACCACGACTTCCAG, and Actin 458 bp (5′)ATCCTCACCCTGAAGTACC (3′)CTCCTTAATG TCACG CACG . Samples were analysed on 1.5% agarose gels and stained with SYBR ® Safe DNA gel stain (Invitrogen).
Flow cytometry
Cell surface phenotype and cell size estimation. Cells were detached from fl asks using TrypLE select (Invitrogen), washed and added to wells of a 96-well plate. Cells were in-cubated for 10 min at room temperature with unconjugated mouse anti-human IgG 1, IgG 2A , IgG 2B , CD29, CD34, CD44, CD45, CD49d, CD73, CD90, CD105, CD166, MHC I, and MHC II antibodies (BD Biosciences). Excess antibody was removed by washing wells with phosphate-buffered saline. Donkey anti-mouse phycoerythrin secondary antibody (Jackson Laboratory, Bar Harbor, ME) was added to each of these wells and cells incubated for 10 min at room temper-ature. After further washing, fl ow cytometry analysis was performed on a FACS Calibur (Becton-Dickinson) using FCS Express Version 3 software. For cell size estimation, 29.6 μm Spheroblank size calibration beads were used (Spherotech, Lake Forest, IL).
Genetic profi ling
Cytogenetic analysis of G TG -banded metaphases were performed on MSCs from fi ve placental and four bone mar-row samples. G enetic stability was further explored using G eneChip SNP-based human mapping analysis to evalu-ated genomic gains and losses.
Karyotype analysis. Flasks containing cells from fi ve hpMSC and four hbmMSC cultures at passag
e 3 (all sam-ples) and at passages 5 and 8 (two hbmMSC and all hpMSC samples) were processed using standard cytogenetic tech-niques. The fl asks were harvested 1–3 days after receipt. Colchicine was added to the fl asks for 3–18 h, after which the cells were released from the fl ask surface with trypsin/versene and treated with hypotonic KCl. The cells were fi xed with methanol/glacial acetic acid 3:1 by volume and spread onto glass slides. Metaphase cells were GTG-banded using trypsin and Geimsa stain. As many metaphases as possible were karyotyped up to a maximum of 20 metaphases per sample.
SNP microarray analysis of genome copy number. DNA was extracted from frozen cell pellets using the QIAamp DNA Blood Mini Kit (Qiagen). Cells were protease digested for 10 min at 56°C before mixing with 0.5 volumes of ethanol and concentrated on the QIAamp Spin Column. The column was washed and genomic DNA eluted and 250 ng labeled using
lrxnumber: 11960-044), Iscove’s Modifi ed Dulbecco’s Medium (IMDM; Invitrogen; catalogue number: 31980-030) and RPMI 1640 Medium (RPMI; Invitrogen; catalogue number: 21870-076).
For the experiment on seeding density, hpMSC and hbmMSC were plated at either 2.5 × 103 cells/fl ask
(100 cells/cm 2) or 2.5 × 104 cells/fl ask (1,000 cells/cm 2
).For the experiments on fl ask type and frequency of medium change (weekly vs. twice weekly), hbmMSC were plated at 5.6 × 103 cells/fl ask at every passage in each exper-iment, and hpMSC were plated 4 × 104 cells/fl ask. hpMSC and hbmMSC were cultured in one of three different 25 cm 2 fl ask types: Nunc (catalogue number: 156367), BD Falcon (BD Biosciences; catalogue number: 353108), and Corning (catalogue number: 3056).
Mesodermal lineage differentiation
Osteogenic lineage. Osteogenic differentiation was induced by culturing 90% confl uent MSC for 3 weeks in DMEM-HG, 10% FCS, 0.1 μM dexamethasone, (Mayne Pharma; Australia Register Number: 16375; Melbourne, Victoria, Australia), 50 μM L -ascorbic acid-2-phosphate (Sigma; catalogue number: A8960-5G; Castle Hill, New South Wales, Australia), 10 mM β-glycerol phosphate disodium salt pentahydrate (Sigma; catalogue number: 50020) and 0.3 mM inorganic (sodium) phosphate (Sigma) [28,29]. Osteogenic differentia-tion was assessed by staining cells in wells with Alizarin Red S.
Chond rogenic lineage. Chondrogenic differentiation was induced by culturing pellets of 5 × 105 MS
公交车
诗洁C for 3 weeks in DMEM-HG, 0.1 μM dexamethasone, 1 mM sodium p yruvate (Sigma; catalogue number: P5280-25G ), 50 μM l-ascorbic acid-2-phosphate, 35 mM l-proline (Sigma; catalogue num-ber: 81710), 10 ng/mL TG F β1 (R&D Systems; catalogue number: 243-B3) and 50 mg/mL ITS Premix (human recom-binant insulin, human transferrin, and selenious acid; BD Biosciences; catalogue number: 354351) [5]. Chondrogenic differentiation was assessed by staining frozen sections of the cell pellets with periodic acid Schiff.Ad ipogenic lineage. Adipogenic differentiation was induced by culturing 80% confl uent MSC for 3 weeks in DMEM-HG, 1 μM dexamethasone, 5 μg/mL insulin (Sigma), 60 μM indo-methacin (Sigma; catalogue number: 17378-5G), and 0.5 mM 3-Isobutyl-1-methylxanthine (IBMX; Sigma; catalogue num-ber: I5879) [28]. Adipogenic differentiation was assessed by staining cells in wells with Oil Red O. It should be noted that these differentiation protocols were originally optimised for use with bone marrow as a source of MSC.
RNA extraction and reverse transcriptase polymerase chain reaction for mesod ermal lineage markers. RNA was isolated from differentiated or undifferentiated MSC using an RNeasy kit (Qiagen) according to manufacturer’s guide-lines. Primers were designed for human messenger RNA (mRNA) and all RNA samples were DNase pretreated. The reverse transcriptase reaction was performed wit
h oligo-dT and Superscript III (Invitrogen) according to manufac-turer’s instructions. PCR was performed on complemen-tary DNA using Taq accuprime Supermix (Invitrogen) according to manufacturer’s instructions for 32 cycles:
BARLOW ET AL.
1098
change in cell number/change in time. Differences in cell proliferation (using cumulative cell numbers and popula-tion doublings) and cell viability were determined using nonparametric statistical tests. For each time point of each experiment, and for each experiment overall, the Wilcoxon–Mann–Whitney test (two-tailed) was used to determine whether signifi cant differences existed between two con-ditions. When more than two conditions were present, the Kruskal–Wallis test was used to determine whether sta-tistically signifi cant differences existed between at least two conditions. When signifi cant differences were found, the Posthoc Kruskal–Wallis multiple comparisons test or Wilcoxon–Mann–Whitney test (one-tailed) was used to determine which conditions were signifi cantly different from each other at each time point. Differences between variables were considered signifi cant when p ≤ 0.05.
Ethical approval
All experiments were approved by the Mater Health Services Human Research Ethics Committee and/or the Animal Ethics Committee of the University of Queensland.
Results
Growth optimization
Basal culture media. There was a signifi cant difference in hpMSC proliferation at specifi c multiple time points and over the entire experiment (p < 0.001) when cultured in fi ve different basal culture media (Fig. 1A) (*p < 0.05). By passage 2, hpMSC in RPMI had died, and so proliferation was signifi cantly less in this medium than the other four media. By passage 3, hpMSC in DMEM-HG had also died, so cell proliferation was signifi cantly less in this medium than DMEM-LG, α-MEM and IMDM. There was no signifi -cant difference in hpMSC proliferation between DMEM-LG, α-MEM, and IMDM. Viability was not different between passages (median 85%, range 75–100%).There was no signifi cant difference in hbmMSC prolifer-ation (Fig. 1B) or viability (median 85%, range 75–95%) when cultured in DMEM-LG or α-MEM. hbmMSC proliferated more slowly than hpMSC and it thus took hbmMSC a longer time to reach a given number of cells than hpMSC (
Fig. 1A and B).
FCS concentration. There was a signifi cant difference in hpMSC proliferation over the entire experiment (p < 0.001), with hpMSC growing better in 20% FCS compared to 10% FCS. hbmMSC proliferated more slowly than hpMSC and numbers of both hpMSC and hbmMSC decreased in culture over time in 10% (v /v) FCS compared to 20% (Fig. 1 C and D). There was no signifi cant difference in viability of hpMSC (median 90%) or hbmMSC (median 80%) between the two conditions from passage 1 through passage 3 (Fig. 2C and D).
FCS heat inactivation. There was no difference in hpMSC- or hbmMSC proliferation rate or viability when cells were cultured in medium with FCS that had been heat inactivated or not heat inactivated (data not shown).Seed ing d ensities. There was a signifi cant difference in hpMSC proliferation over the entire experiment (p = 0.042),
the Affymetrix XbaI SNP GeneChip™ Labelling kit, as per manufacturer’s instructions. In brief, DNA was digested with 1 U XbaI (NEB) for 2.5 h and ligated to Adaptor Xba oli-gomers using 250 U T4 DNA Ligase (NEB). Three PCR reac-tions were performed on a total of 75 ng of postligation DNA and purifi ed using MiniElute 96 UF PCR Purifi cation plates (Qiagen). Forty microgram of DNA was
fragmented and end-labeled with dUTP-biotin before a 16-h hybridization on the 50K XbaI Human Mapping GeneChip. The GeneChip was washed according to Mapping 100 Kv1_450 washing protocol and scanned on a GS3000 scanner.
Human Mapping XbaI GeneChip data was extracted and analyzed using GCOS v1.3 and GDAS v3.0.1 from Affymetrix. Copy number estimates were produced using Chromosome Copy Number Tool v2.0.0.9 (CCNT) (Affymetrix) and graphed using Copy Number Analyser for G eneChip v2.0 (CNAG) [30].
In vivo toxicity stud ies. Passage 3 and 5 hpMSC were injected intravenously via the tail vein into Balb/c mice (Animal Resources Centre, Perth, Australia) at three doses: 2 × 104 cells/mouse (1 × 106 cells/kg), 2 × 105 cells/mouse (1 × 107 cells/kg), and 2 × 106 cells/mouse (1 × 108 cells/kg). At the request of the University of Queensland Animal Ethics Committee, each dose was administered to only two female Balb/c mice (for minimization of animal utilization) in 200 μl sodium chloride 0.9% (Baxter Australia) with 4% (v /v) FCS (not heat-inactivated) through 26-gauge needles. Cells were not fi ltered.
Passage 5 hbmMSCs were administered to mice in the same way and at the same doses. Mice wer
e monitored closely for the fi rst 5 h after cell administration, and then once each day for 3 days thereafter. A mouse health score sheet was used to assess animals during this experiment: One point was given if animals showed ≥15% weight loss, hunched posture at rest, decreased activity or ruffl ed fur. Mice were sacrifi ced at day 3, or beforehand if they reached a score of 4.
Data analysis
Every experiment used four bone marrow and four pla-centa samples with three replicates per experiment and sample. All line and bar graphs display median values with interquartile ranges. To analyze growth kinetics, cumula-tive cell number was determined at each passage in every experiment.
Cumulative cell number at passage X equals the sum of the three confl uent cell cultures at passage X divided by the sum of the cells plated into the three passage X cultures multiplied by the cumulative cell number at passage X-1. Cumulative cell numbers were calculated from the number of cells seeded from passage 1.
The number of population doublings was determined in the long-term cultures at every passage. Two to the power of (population doublings at passage X) equals cumulative MSC number at passag
e X divided by MSC number at pas-sage 0. Therefore, population doublings at passage X equals log 10 (cumulative MSC number at passage X divided by MSC number at passage 0) divided by log 10 2.
The rate of cell proliferation is illustrated in each expan-sion kinetics graph by the gradient of each line, which is the
10C e l l N u m b e r
B
小学生文具盒
1.00E +071.00E +061.00E +051.00E +041.00E +03
10
2030
40
Time (days)
Time (days)
01020304050607080
Alpha-MEM DMEM-LG Alpha-MEM DMEM-LG
IMDM DMEM-HG RPMI
1110C e l l N u m b e r
D
1.00E +071.00E +061.00E +051.00E +041.00E +03
10
20
30
40
Time (days)0
10
10%20%20
30
40
10C e l l N u m b e r
F
1.00E +071.00E +061.00E +051.00E +041.00E +03
20
40
60
80
100
Time (days)
Time (days)0
10
100 cells/cm 21000 cells/cm 2100 cells/cm 1000 cells/cm 20
30
50
40
10C e l l N u m b e r
H
1.00E +071.00E +061.00E +051.00E +041.00E +03
20
40
60
80100
Time (days)
Time (days)
010Nunc Falcon Costar
203080
4050607010C e l l N u m b e r
J
1.00E +071.00E +061.00E +051.00E +041.00E +03
20
40
60
80
100
Time (days)
Time (days)0
10
Twice weekly Weekly 20
双层杯
30
40
Nunc Falcon Costar
Twice weekly Weekly
Human MSC growth optimization. Effect of basal medium on ex vivo expansion ability of (A ) hpMSC and (< 0.05 at indicated time points). Effect of fetal calf serum concentration on ex vivo expansion ability of (C ) hpMSC ) hbmMSC. Effect of seeding density at passage 1 on ex vivo expansion ability of (E ) hpMSC and (F ) hbmMSC (*p at indicated time points). Effect of fl ask type on ex vivo expansion ability of (G ) hpMSC and (H ) hbmMSC hpMSC (*p <indicated time points). Effect of medium replacement frequency on ex vivo expansion ability of (I ) hpMSC and (J ) hbmMSC.
豫公网安备41160202000603
站长QQ:729038198
关于我们
投诉建议