SEARCH

SEARCH BY CITATION

Keywords:

  • Adipogenesis;
  • Adult bone marrow stem cells;
  • Differentiation;
  • Hypoxia

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References

Human mesenchymal stem cells (hMSCs) have the capacity to differentiate along several pathways to form bone, cartilage, tendon, muscle, and adipose tissues. The adult hMSCs reside in vivo in the bone marrow in niches where oxygen concentration is far below the ambient air, which is the most commonly encountered laboratory condition. The study reported here was designed to determine whether oxygen has a role in the differentiation of hMSCs into adipocytes. Indeed, when exposed to atmosphere containing only 1% of oxygen, the formation of adipocyte-like phenotype with cytoplasmic lipid inclusions was observed. The effect of hypoxia on the expression of adipocyte-specific genes was determined by real-time reverse transcription polymerase chain reaction. Interestingly, neither of the two central regulators of adipogenesis—the transcription factors peroxisome proliferator-activated receptor γ2 (PPAR-γ2) and ADD1/SREBP1c—was induced. Furthermore, hypoxia did not have any effect on the transcription of early (lipoprotein lipase) or late (aP2) marker genes. By the same token, neither of the mature adipocyte-specific genes—leptin and adipophilin—was found responsive to the treatment. High level of induction, however, was observed with the PPAR-γ–induced angiopoietin-related gene, PGAR. The lack of an adipocyte-specific transcription pattern thus indicates that despite accumulation of the lipid, true adipogenic differentiation did not take place. In conclusion, hypoxia appears to exert a potent lipogenic effect independent of PPAR-γ2 maturation pathway.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References

In everyday practice, the in vitro experiments are routinely being set up at an oxygen concentration in the gas phase that corresponds to that of the ambient air, and it is being overlooked that such conditions provide for oxygen levels that several-fold surpass those found under physiological conditions. Hypoxia has previously been shown to have a critical role in the development of placenta [1]; moreover, early human embryogenesis takes place in an environment with a very low oxygen tension [2]. Adult stem cells residing in the bone marrow are similarly exposed to low oxygen tension. The mean oxygen concentration of the bone marrow has been reported to be approximately 7% [3]. While, at present, it is not possible to measure spatial concentration of oxygen, mathematical models have indicated that there is a gradient across the marrow from the rather well-oxygenated sinuses to the relatively hypoxic areas near the bone endosteum [4]. It has been proposed that the hematopoietic stem cells (HSCs) are maintained in an undifferentiated state in hypoxic niches in the bone marrow and that the expansion and differentiation of the stem cells occurs along an oxygen gradient [5]. Recently, Danet et al. [6] demonstrated that culturing human HSCs under reduced oxygen conditions (1.5%) improved their survival and increased the number of bone marrow–repopulating cells. Furthermore, hypoxia has been shown to stimulate the formation of granulocytic-monocytic progenitors [7] and profoundly alter several characteristics of macrophages [8].

Bone marrow is also a place of residence of mesenchymal stem cells (MSCs), which are pluripotent cells with the potential for differentiating into several cell types, including adipocytes, chondrocytes, myocytes, and osteoblasts [9]. It is noteworthy that oxygen has been shown to modulate the differentiation in at least some of these cell types, specifically osteoblasts and adipocytes. In particular, culturing rat MSCs at reduced oxygen tension (5%) increased their bone-forming potential [10], and in a study of adipogenesis by Yun et al. [11], hypoxia inhibited the conversion of mouse pre-adipocytes to adipocytes upon induction with a classical adipogenic cocktail. The initial data thus seem to indicate that oxygen has a regulatory role in at least some differentiation pathways.

To obtain a better understanding of the effect of hypoxia on adipogenic conversion of MSCs, we subjected a human telomerase reverse transcriptase–immortalized human MSC (hMSC) line, hMSC-TERT [12], to oxygen concentrations ranging from 1%–21%, and we followed the accumulation of intracellular lipid and expression of marker genes associated with early and late stages of adipogenesis.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References

Cell Cultures

In this study we used the hMSC-TERT cell line, which represents a lineage of human bone marrow stromal cells that was immortalized through the expression of the catalytic subunit of human telomerase [12]. The cell line retains characteristics that are typical of MSCs. The cells were maintained in Eagle's minimal essential medium (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 10% fetal calf serum (FCS) (Helena Bioscience, Sunderland, U.K., http://www.helenachemical.com), 100 IU/mL penicillin (Invitrogen), and 100 μg/mL streptomycin (Invitrogen). For the experimental procedures, the hMSC-TERT cells were seeded in 2 mL medium in six-well plates (Corning, Costar, Acton, MA, http://www.corning.com) at a concentration of 2 × 105 cells per well. The cells were incubated in a humidified atmosphere with 20% oxygen and 5% CO2 at 370°C for 72 hours to achieve confluence, after which the concentration of FCS in the medium was lowered to 5%. Five days after the seeding, the cells were subjected to either hypoxic treatment or treatment with the peroxisome proliferator-activated receptor γ2 (PPAR-γ2) ligand rosiglitazone (BRL 49653, Novo Nordisk, Bagsvaerd, Denmark, http://novonordisk.com).

Hypoxic Treatment

For the hypoxia treatment, one half of the cultures was transferred to an In Vivo 400 hypoxic workstation (Maltec, Aarhus, Denmark, http://www.maltec.nymbursko.com), where, for up to 3 days, the incubation proceeded in atmosphere containing 5% CO2; 1%, 2%, 3%, 4%, or 6% O2; and a balance of nitrogen. Duplicate wells from both the experimental hypoxic cultures and the control ambient cultures were processed at 2, 4, 8, 24, 48, and 72 hours after the start of hypoxic treatment. The experiments were carried out twice, in an identical fashion.

Induction of Adipogenesis by Rosiglitazone

For induction experiments with rosiglitazone, the medium was changed to Dulbecco's Modified Eagle's Media (Invitrogen) supplemented with 10% FCS; 1 μM dexamethasone (Sigma-Aldrich, Copenhagen, Denmark; http://www.sigma-aldrich.com); 0.45 mM isobutyl methylxanthine (IBMX) (Sigma-Aldrich); 170 nM insulin (Invitrogen); 0.2 mM indomethacin (Sigma-Aldrich); 1 μM rosiglitazone; 100 IU/mL penicillin; and 100 μg/mL streptomycin. The cells were maintained in this medium for 2 weeks with medium changes at 3- to 4-day intervals.

Measurement of Pericellular Oxygen Tension

The pericellular oxygen tension was measured directly in the plates by an OX500 oxygen minielectrode (Unisense, Aarhus, Denmark, http://www.unisense.com). Measurements were performed in four separate cultures at 1-hour intervals for up to 4 hours after the equilibrium had been reached.

Oil Red O Staining

Accumulation of triglycerides was visualized by staining with Oil red O (Sigma-Aldrich), essentially as described previously [13]. Briefly, we prepared a stock solution of 0.5% (w/v) Oil red O in isopropanol. To achieve a working concentration, 6 mL of stock solution was mixed with 4 mL of water. The cells were directly fixed in situ with 10% buffered formaldehyde at 40°C for 1 hour, washed with phosphate-buffered saline (PBS), stained with Oil red O working solution for 15 minutes, and finally washed with water. The triglyceride accumulation was assessed microscopically with a IX-70 inverted microscope (Olympus, Hamburg, Germany, http://www.olympus-global.com) using a Hoffman modulation contrasting mode that made possible simultaneous visualization of cell morphology. Images were captured by the Spot RT camera (Diagnostic Instruments, Sterling Heights, MI, http://www.diaginc.com) through AnalySIS software (Olympus).

Analysis of Triglyceride Accumulation by Fluorescence Microscopy

For fluorescent labeling of triglyceride vesicles, the cells were grown in Lab-Tek II chamber slides (Nalge Nunc International, Biotech Line, Slangerup, Denmark, http://www.nalgenunc.com) at a concentration of 20 × 104 cells per cm2 and subjected to a hypoxic treatment or rosiglitazone-mediated adipogenic induction according to protocols described above. One hour before the end of the incubation period, the BODIPY reagent (Molecular Probes, Leiden, The Netherlands, http://www.probes.com) and the Hoechst 33342 dye (Molecular Probes) were added in final concentrations of 10 and 2 μg/ml, respectively. After staining, the cells were washed, fixed with 4% buffered formaldehyde for 15 minutes, washed again, and mounted in a DAKO fluorescent mounting medium (DAKO, Copenhagen, Denmark, http://www.dako.com). The preparations were evaluated with Axiovert 200M inverted microscope (Carl Zeiss, Göttingen, Germany, http://www.zeiss.de) using a filter set with bandpass excitation and emission from 450 to 490 and from 515 to 555 nm, respectively. A series of Z-sections was taken by Axio Cam MRm camera (Zeiss) controlled through Axio Vision software (Zeiss). The individual image planes were further processed using a 3D deconvolution based on regularized inverse filter (Axio Vision; Zeiss), and the X-Y maximum intensity projection of the original image volume was constructed with Meta Morph software package (Universal Imaging Corporation, Downingtown, PA, http://www.image1.com).

Nile Red Staining and Flow Cytometric Analysis

To obtain a quantitative measure of the intracellular triacyl-glycerol accumulation, the cells were stained with Nile red dye [14]. The stock solution was made by dissolving Nile red (Sigma-Aldrich) to 1 mg/mL in dimethyl sulfoxide, and the working solution was prepared fresh immediately before use by diluting 100-fold in PBS. Prior to the staining, the cells were dislodged by trypsinization, washed, and fixed at 40°C for 15 minutes in 2% buffered formaldehyde. After fixation, the cells were washed, resuspended in 400 mL PBS, and stained after addition of 44 μL Nile red working solution on ice for 30 minutes. For flow cytometry, the cells were washed in PBS, and the analysis was done in the FACStar Plus (Becton, Dickinson, Broendby, Denmark, http://www.bd.com) using a 488-nm band for excitation and measuring emission at 585 nm. The analysis and presentation of the data were done using the WinMIDI software package (http://facs.scripps.edu/software.html).

Real-Time Reverse Transcription Polymerase Chain Reaction (RT-PCR)

RNA was isolated from the cells by means of the GenElute total RNA kit (Sigma-Aldrich) and treated with DNase I (Roche Molecular Biochemicals, Mannheim, Germany, http://www.roche.com). As a positive control for adipocyte-specific gene expression, we used total RNA isolated from human adipose tissue (a gift from J. M. Bruun, Aarhus Amtssygehus). For cDNA synthesis, approximately 5 μg of RNA served as a template for the Moloney murine leukemia virus reverse transcriptase (Sigma-Aldrich) with random decamer primers (Invitrogen). The levels of transcripts were determined by real-time RT-PCR. For each sample and gene, 25-μL reactions were set up in duplicate, each with 12.5 μL of Sybr Green supermix (Bio-Rad, Herlev, Denmark, http://www.bio-rad.com), 0.5 pmol of each primer, and 0.1μL of cDNA. The following primers were used (official gene symbol is given in square brackets):

  • ADD1/SREBP1c [SREBF1]

    • forward: 5′-ggagccatggattgcactttc-3′

    • reverse: 5′-atcttcaatggagtgggtgcag-3′

  • PPAR-γ2 [PPARG2]

    • forward: 5′-tcaggtttgggcggatgc-3′

    • reverse: 5′-tcagcgggaaggactttatgtatg-3′

  • adipophilin [ADPF]

    • forward: 5′-cgctgtcactggggcaaaaga-3′

    • reverse: 5′-atccgactccccaagactgtgtta-3′

  • perilipin [PLIN]

    • forward: 5′-gaattcgctctcgggctccatcag-3′

    • reverse: 5′-gcccccgctgtctcctcaacca-3′

  • aP2 [FABP4]

    • forward: 5′-atgggatggaaaatcaacca-3′

    • reverse: 5′-gtggaagtgacgcctttcat-3′

  • lipoprotein lipase [LPL]

    • forward: 5′-gaaccgctgcaacaatctgggctatga-3′

    • reverse: 5′-tgctgcttcttttggctctgactttattga-3′

  • adiponectin [ACDC]

    • forward: 5′-tctgtttcccacctcacctga-3′

    • reverse: 5′-caggacgtcatcatagaaccactt-3′

  • leptin [LEP]

    • forward: 5′-ggagtgacggtcccacactg-3′

    • reverse: 5′-ccatgtaataaacccagagtgtgatc-3′

  • peroxisome proliferator-activated receptor γ-induced angiopoietin-related gene (PGAR) [ANGPTL4]

    • forward: 5′-tctccgtacccttctccact-3′

    • reverse: 5′-agtactggccgttgaggttg-3′

  • glyceraldehyde 3-phosphate dehydrogenase (GAPDH) [GAPD]

    • forward: 5′-gacccccaccacactgaatctcc-3′

    • reverse: 5′-tgacaaggtgcggctcccatag-3′

To normalize for input loads of cDNA, 18S rRNA (forward: aggaccgcggttctattttgttgg, reverse: cccccggccgtcc-ctctta) was used as an internal standard. The amplification reaction for 18S was essentially as described above for the other genes, with the only exception that 0.025 μL cDNA was used per reaction. The amplifications were performed in an iCycler (Bio-Rad), using two-temperature cycling consisting of denaturation step at 950°C for 15 seconds, and an annealing/extension step at 680°C (600°C for ADD1/ SREPB1c and adipophilin) for 30 seconds.

A relative sensitivity of the real-time RT-PCR assays was determined individually for each gene. To this end, a 10-fold dilution series was made of the cDNA prepared on the basis of fat tissue RNA and amplified as described above. Based on these results, it was possible to determine the lower detection limit relative to the expression in fat tissue for all studied genes.

Western Blot Analysis

For immunoblotting, the cells were lysed in sodium dodecyl sulfate (SDS) sample buffer, and the concentration of protein was determined by the BCA Protein Assay Kit (Pierce, Rock-ford, IL, http://www.piercenet.com). Aliquots corresponding to 20 μg were loaded onto a 10% SDS-polyacrylamide gel, and after electrophoresis, the proteins were electroblotted onto a polyvinylidene diflouride membrane (Millipore, Bedford, MA, http://www.millipore.com). To confirm equal loading and transfer of samples, the proteins were stained with SYPRO Ruby (Molecular Probes), after which detection was carried out in the Fluor-S MultiImager (Bio-Rad) and the analysis was performed using the TotalLab software package (Phoretix, Newcastle, U.K., http://www.nonlinear.com). As primary antibodies in the immunodetection, we used a 20,000-fold dilution of rabbit whole antiserum to GAPDH (Ab9485; Abcam, Cambridge, U.K., http://www.abcam.com) and a 2,000-fold dilution of affinity purified rabbit antibody to PPAR-γ2 (Ab8935; Abcam). As a secondary antibody, we used a 50,000-fold dilution of biotinylated donkey anti-rabbit immunoglobulins (RPN1004; Amersham Biosciences, Hillerød, Denmark, http://www4.amershambiosciences.com), and streptavidin conjugated with horse-radish peroxidase (RPN1051; Amersham Biosciences) was used in 60,000-fold dilution as a label. Visualization was accomplished using an ECL Plus reagent (Amersham Biosciences), and the detection was carried out in the Fluor-S MultiImager.

Statistical Analysis

The significance of decreasing oxygen partial tension for lipid accumulation was determined using the Jonckheere-Terpstra test based on a one-tailed alternative for each time interval. When found statistically significant at the level of 5%, pairwise comparisons at the given time point were done using the Kolmogorov-Smirnov nonparametric statistics. The comparisons of GAPDH and PGAR mRNA levels in hypoxia with ambient controls were made using a two-tailed t-test. All calculations were done with the aid of the exact tests package of SPSS 11.0 software (SPSS, Chicago, http://www.spss.com).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References

Lipid Accumulation

To evaluate the effect of hypoxia on adipogenesis, hMSC-TERT cells were incubated at oxygen tensions covering the range from 1%– 6% for up to 72 hours. Only when they were cultured at oxygen tension corresponding to 1% did the cells respond with a distinct accumulation of lipid droplets. The changes of the phenotype became visible microscopically, as well as macroscopically, as soon as 24 hours after the initiation of treatment, as indicated by Oil red O staining (Fig. 1). Incubation of the cells at 2% oxygen and above (4% and 6% data not shown) did not appear to have any significant effect on lipid accumulation even after 72 hours of hypoxic exposure (Fig. 1). To obtain more information about the distribution of lipid load at the population level and the kinetics of lipid accumulation, the cells were stained with Nile red and subjected to flow cytometric analysis. As evidenced in Figure 2A, the lipid accumulation occurred in a symmetric fashion, corresponding to a normal distribution across the whole population of cells, at all time intervals. Furthermore, the quantitation of lipid accumulation in time by flow cytometry confirmed that when growing in 1% oxygen atmosphere, the onset of triglyceride accumulation was evident already after 24 hours and became statistically significant after 48 hours (Fig. 2B).

thumbnail image

Figure Figure 1.. Effect of oxygen tension on lipid droplet formation in hMSC-TERT cells. The hMSC-TERT cells were cultured at ambient (21%) and hypoxic (2% and 1%) oxygen concentrations for up to 3 days. Lipid droplet accumulation was examined by a Hoffman modulation contrast microscopy after staining with Oil red O.

Download figure to PowerPoint

thumbnail image

Figure Figure 2.. Flow cytometric analysis of lipid accumulation and distribution in hMSC-TERT cells incubated at varying oxygen concentrations. The hMSC-TERT cells were incubated at 1%, 2%, and 21% oxygen concentrations for up to 3 days. At time 0, 8, 24, 48, and 72 hours of treatment, cells were fixed, stained with Nile red, and analyzed in a flow cytometer. (A): Results from a representative experiment of cells cultured for up to 72 hours at 1% oxygen concentrations. The shaded area represents cells prior to hypoxic treatment; open areas represent treated cells. (B): Lipid accumulation represented as the geometric mean obtained from flow cytometric analysis as a function of time and oxygen concentration. Each point represents the average of at least two experiments, each in duplicate, in which error bars denote standard error of mean. Asterisks indicate p < .05 compared with control cultures.

Download figure to PowerPoint

To obtain a better understanding of the differences between hypoxia-induced and rosiglitazone-induced lipid accumulation, the cells were incubated either at 1% oxygen partial pressure for 48 hours or with rosiglitazone-containing induction medium for 14 days, after which the triglyceride accumulation was visualized with Oil red O and fluorescent marker BODIPY (Fig. 3). The Oil red O staining revealed that the cells grown in hypoxia accumulated lipids in a remarkably different fashion from those incubated with rosiglitazone. In hypoxia, the lipids were found to form a diffuse pattern, whereas, in rosiglitazone-treated cells, the lipids were typically confined to discrete areas. The close-up micrographs of the BODIPY-stained cells further confirmed this observation. In hypoxia, all cells accumulated a small number of little lipid droplets, but in the foci of lipid accumulation, the rosiglitazone-treated cells contained an abundance of lipid inclusions that varied greatly in size, yet were for the most part significantly larger than those in hypoxia.

thumbnail image

Figure Figure 3.. Differential distribution of neutral lipids in hMSC-TERT cultures exposed to low oxygen tension or undergoing adipogenic differentiation. (A):The hMSC-TERT cells were cultured under ambient and hypoxic conditions, corresponding to 21% and 1% oxygen concentrations, respectively. After 72 hours, the cultures were stained with Oil red O and examined using a Hoffman modulation contrast microscopy or stained with BODIPY and Hoechst 33342 and subjected to an epifluorescence examination. (B): The hMSC-TERT cells were cultured under standard conditions or induced into adipogenesis by dexamethasone, isobutyl methylxanthine, insulin, indomethacin, and rosiglitazone. After 14 days, the cells were stained with Oil red O and fluorescent dyes.

Download figure to PowerPoint

Pericellular Oxygen Tension

The finding that the lipid accumulation in the hMSC-TERT cells was apparently controlled over a very narrow oxygen gradient warranted further investigation. Hence, to determine accurately at which point the lipid accumulation occurred, the pericellular oxygen tension was measured. The oxygen tensions in close vicinity of the cells were found almost 10-fold lower than above the medium (Table 1). Thus the lipid accumulation was induced at levels of pericellular oxygen concentration corresponding to 0.14% but, strikingly, the lipogenic process was efficiently abrogated by an increase in oxygen concentration as little as 0.1 percentage points. It is also noteworthy that when the cultures were transferred from ambient oxygen tension to hypoxia, the equilibrium between oxygen in the gas and liquid phases was reached after as soon as 1 hour (data not shown).

Table Table 1.. Correlation between atmospheric oxygen concentration and pericellular oxygen tension
  • a

    aThe numbers are given as the mean of 16 different measurements ± standard error of mean.

Atmospheric concentration (%)Pericellular concentration (%)a
20.915.1± 0.281
20.23 ± 0.008
10.14 ± 0.011

Adipocyte-Specific Gene Expression

To determine the transcriptional phenotype of the cells grown in hypoxia and compare it with that of control cultures and genuine adipose tissue, a comprehensive group of adipocyte-specific genes was analyzed by a sensitive real-time RT-PCR (Table 2). First, the expression of the adipogenesis-related transcription factors ADD1/SREPB1c and PPAR-γ2 was evaluated. Surprisingly, no expression of these genes was detected in spite of the distinct lipid accumulation. Thereafter, we examined whether the lipid accumulation could be attributed to changes in the expression of some of the genes involved in lipid transport, storage, or metabolism. However, adipophilin, aP2, perilipin, and lipoprotein lipase were not expressed in the cells. Finally, when assessing the expression of adipocytokines, including adiponectin, leptin, and PGAR, we found that PGAR was upregulated 12-fold in 1% O2 (12-fold). The expression of the other two adipocytokines, adiponectin and leptin, however, could not be detected.

Table Table 2.. Expression of adipocyte-specific genes in hMSCs after exposure to 72 hours of hypoxia or 14 days of rosiglitazone treatment
  1. a

    Expression levels are arbitrarily indicated with ++ in adipose tissue. The + and +++ indicate expression levels that are lower (and still detectable) or higher than that of adipose tissue, respectively. An expression level below detection limit is indicated by —. Numbers in parentheses indicate the detection limit of the gene expression relative to that of adipose tissue.

  2. b

    Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; hMSC, human mesenchymal stem cell.

GeneFunctionExpression in adipose tissueExpression in hMSC (21% O2)Expression in hMSC (1% O2)Expression in hMSC (ros)
ADD1/SREBP1cTranscription factor++— (10−3)++
PPAR-γ2Transcription factor++— (10−2)+
AdipophilinLipid transport/storage++— (10−3)++
aP2Lipid transport/metabolism++— (10−3)++
Lipoprotein lipaseLipoprotein fatty acid metabolism++— (10−2)++
AdiponectinAdipocytokine++— (10−3)
LeptinAdipocytokine++— (10−4)
PGARAdipocytokine++++++++
GAPDHGlucose metabolism++++++++

To confirm that the cell line could express the adipocyte-specific markers under appropriate culture conditions, the rosiglitazone-treated cells were subjected to the same analysis. As indicated in Table 2, the results from real-time RT-PCR analysis corroborated the expression of all adipocyte-specific genes, except for adiponectin and leptin.

Although the hypoxia-induced transcription of PGAR decreased with increasing oxygen concentrations, the expression was still statistically significantly higher than the expression at 21% for all hypoxic conditions (Fig. 4). The transcription of GAPDH, which is responsive to hypoxia in several cell types, was also statistically significantly increased, but only in an atmosphere with less than 4% O2 (Fig. 4). This finding indicates that HIF-1 is stabilized in the hMSC-TERT cells below a specific level of oxygen concentration.

thumbnail image

Figure Figure 4.. Transcriptional regulation of GAPDH and PPAR-γ–induced angiopoietin-related gene (PGAR) by low oxygen. The hMSC-TERT cells were incubated at 1%, 2%, 3%, 4%, and 6% oxygen for 72 hours, after which RNA was harvested and analyzed by real-time reverse transcription polymerase chain reaction. The graph shows the fold regulation of GAPDH (filled bars) and PGAR (open bars) in hypoxic cells compared with cells cultured at ambient oxygen tensions (21%). The bars indicate the average of at least two independent experiments, each in duplicate (n = 4). Error bars designate standard error of the mean. Asterisks indicate p < .05 compared with control cultures at ambient air. Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PPAR-γ, peroxisome proliferator-activated receptor γ.

Download figure to PowerPoint

Hypoxia and Rosiglitazone-Induced Expression of GAPDH and PPAR-γ2

The effect of hypoxia and rosiglitazone treatment on protein levels of GAPDH and PPAR-γ2 in hMSC-TERT cells was determined by Western blotting (Fig. 5). The analysis revealed that treatment with 1% oxygen for 72 hours resulted in accumulation of GAPDH, whereas treatment with rosiglitazone did not increase GAPDH protein levels above the basal expression. Furthermore, only cells treated with rosiglitazone accumulated PPAR-γ2, thus confirming the findings by real-time RT-PCR.

thumbnail image

Figure Figure 5.. Induction of GAPDH and PPAR-γ2 proteins by low oxygen and rosiglitazone. The hMSC-TERT cells were cultured in normal growth medium at 21% or 1% oxygen for 72 hours, after which the whole cell lysates were analyzed by immunoblotting. For adipogenic induction, the normal growth medium was supplemented with dexamethasone, isobutyl methylxanthine, insulin, indomethacin, and rosiglitazone, and the analysis was carried out after 14 days. Abbreviations: GAPDH, glycer-aldehyde-3-phosphate dehydrogenase; PPAR-γ2, peroxisome proliferator-activated receptor γ2.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References

At present, the mechanisms underlying the commitment of the stem cells to any given differentiation pathway in vivo are not clear. In vitro, as a rule, the course of adipogenic conversion of uncommitted stem cells and pre-adipocytes is controlled by the addition of selected supplements and growth factors. The induction protocols differ widely, depending on the cell lines used. For instance, adipogenesis is induced in the rodent cell line Ob17 by incubation with fatty acids, whereas 3T3-L1 and 3T3-F442A cells require incubation with cyclic AMP, dexamethasone, and insulin [15]. Likewise, human pre-adipocytes isolated from subcutaneous fat can be readily induced to undergo adipose conversion upon stimulation with thiazolidinediones, while pre-adipocytes from omental fat cannot [16]. It has also been established that incubating hMSCs with a mixture of insulin and dexamethasone supplemented with either rabbit serum or the synthetic PPAR-γ2 ligand, rosiglitazone, will induce the conversion of uncommitted precursor cells to adipocytes [17]. In addition to all the approaches that are based on chemical induction, it appears plausible that the adipogenic pathway can also be modulated by oxygen [11].

In this investigation, we studied whether the phenotypic responses of an hMSC line during hypoxia are congruent with changes associated with adipogenesis. We found that culturing cells in an atmosphere of 1% oxygen led to a rapid generation of cytoplasmic lipid vacuoles and that this accumulation was evident in as little as 24 hours. It is intriguing that the distribution of lipid was in a disseminated pattern across the whole cell population and that the lipids accumulated as small droplets only. This is in contrast with the rosiglitazone-mediated adipogenesis in which only a sub-population of cells was found to accumulate lipids in a focal manner and, importantly, the changes took place only after several weeks of treatment. The cells thus responded to the induction by rosiglitazone in a fashion comparable with primary MSCs [18, 19].

The atypical nature of lipid accumulation after hypoxic exposure in our study was further highlighted by the analysis of adipocyte-specific gene expression. Somewhat unexpectedly, we found that the accumulation of lipids was not accompanied by changes in any of the marker genes characteristic of mature adipocytes. The analysis revealed that neither of the adipocyte-specific transcription factors ADD1/SREBP1c and PPAR-γ2, which are very early markers of adipogenesis, were transcribed. Recently Torii et al. [20] described a pre-adipose cell line, which was refractory to PPAR-γ2 stimulation while susceptible to adipogenesis upon induction with IBMX, hydrocortisone, and indomethacin. This indicates that adipogenic conversion is possible through both PPARγ2–dependent and –independent pathways, contingent upon the permissiveness of the progenitor. The absence of the adipocyte-like expression profile in our MSCs as a result of hypoxia, though, suggests that the true adipogenic conversion did not take place. Here, it is important to note that a possibility of inherent unresponsiveness of the cells could be ruled out since unambiguous adipogenic differentiation was regularly achievable using rosiglitazone.

We found that PGAR was strongly induced by even moderate levels of hypoxia. However, while PGAR was first identified as a PPAR-γ2–regulated gene, which was predominantly expressed in adipose and placental tissue [21], other studies have revealed that the protein is also present in hepatocytes [22], as well as in hypoxia-treated cardiomyocytes [23] and ischemic endothelial cells [24]. It thus appears that PGAR is rather universally expressed as a result of hypoxia and may not be not suitable as a marker of adipogenesis. In addition, PGAR was expressed in mild hypoxic conditions (up to 6% oxygen) without a concomitant accumulation of lipid droplets. This suggests that the lipid accumulation and PGAR induction are unrelated events.

Based on our observations, it can be concluded that hypoxia is capable of driving MSCs into an adipocyte-mimicking phenotype. However, it remains unclear why the adipocyte-specific, hypoxia-responsive genes, the adipose hormone leptin and lipid-binding protein adipophilin [25, 26], were not upregulated during our hypoxic treatment. Using an analogy—for example, erythropoietin, which requires the hepatic nuclear factor 4 as a co-factor for the hypoxia-induced transcription [27]—we hypothesize that it was the lack of specific co-factors that prevented the expression of two genes. It is worth exploring whether leptin and adipophilin are subject to modulation by oxygen only against the background of mature adipocytes or also less mature pre-adipocytes.

Among possible explanations for the hypoxia-induced lipid accumulation is upregulation of C/EBP-α, which has been shown to be sufficient for induction of adipogenesis in 3T3-L1 pre-adipocytes [28]. The expression of C/EBP-α, however, was also shown to lead to rapid expression of aP2, which was not seen in this study. It is thus likely that expression of C/EBP-α does not play a part in lipid accumulation in the hypoxia-treated cells. Another explanation for the increased lipid accumulation under hypoxic conditions could be a modulation of the transport of fatty acids. Of special interest is to look into the involvement of the fatty acid translocase FAT/CD36, which has previously been shown to be involved in the acute regulation of fatty acid uptake [29].

Another important outcome of our investigation was the finding that lipid accumulation in the MSCs is controlled within a very narrow range of pericellular oxygen tension, which corresponds to a change by mere 0.1 percentage points, from 0.14%–0.23%. This is a significantly smaller difference than that measured in the experimental atmosphere, an increase by 1 percentage point, from 1%–2%. It is important to realize that while the atmospheric oxygen tension is relatively straightforward to control, the actual and effective pericellular oxygen tension varies with cell density, cell metabolism, height of the medium, and other factors. Thus it is desirable in the future to use the direct pericellular values instead of the surrogate data reflecting the local atmosphere, since such an approach would eliminate a major source of error and make possible a better reproducibility of experiments between laboratories.

Summary

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References

hMSCs are capable of differentiating into various cell types, including adipocytes, myocytes, chondrocytes, and osteocytes. We have observed that hMSCs incubated under hypoxic conditions (1% oxygen) form adipocyte-like phenotype with cytoplasmic accumulation of lipid. In spite of increased levels of the PPAR-γ–induced angiopoietin-related gene (PGAR) transcripts, the accumulation of lipids was not accompanied by increased transcription of adipocyte-specific genes such as ADD1/SREBP1c, PPAR-γ2, lipoprotein lipase, aP2, leptin, perilipin, and adipophilin. In conclusion, it appears that under specific hypoxic conditions hMSCs may acquire adipocyte-mimicking morphology in the absence of true adipogenic conversion.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References

We thank Dr. J.M. Bruun, Aarhus Amtssygehus, Denmark, for kindly providing RNA isolated from adipose tissue. Also, we appreciate the technical assistance of Anna Bay Nielsen. This work was supported by Danish Medical Research Council grant no. 2052-01-0045 and EU grant no. LSHC-CT-2003-502932 and Novo Nordisk Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Summary
  8. Acknowledgements
  9. References