- Email: [email protected]
Oxygen consumption in undifferentiated versus differentiated adipogenic mesenchymal precursor cells
Respiratory Physiology & Neurobiology 146 (2005) 107–116
Oxygen consumption in undifferentiated versus differentiated adipogenic mesenchymal precursor cells Dennis von Heimburg a, ∗, 1 , Karsten Hemmrich a, 1 , Sascha Zachariah a , Harald Staiger b , Norbert Pallua a a
b
Department of Plastic Surgery and Hand Surgery–Burn Center, University Hospital of the Aachen University of Technology, Pauwelsstr. 30, 52057 Aachen, Germany Department of Internal Medicine IV, Eberhard-Karls-University T¨ubingen, Otfried-M¨uller-Street 10, 72076 T¨ubingen, Germany Accepted 6 December 2004
Abstract To date, no adequate implant material for the correction of soft tissue defects such as after extensive deep burns, tumor resections or in congenital defects is available. A biohybrid composed of viable adipose precursor cells and an optimised matrix could help towards a solution. Morphologically, preadipocytes resemble fibroblasts and have not yet built a large cytoplasmic lipid droplet as found in differentiated adipocytes. Additionally, preadipocytes are smaller than mature adipocytes allowing a quicker revascularization after transplantation. Furthermore, transplanted preadipocytes can form adipose tissue in vivo whereas the transplantation of mature adipocytes often gives poor results, i.e. oil cysts or shrinkage of the transplant. Since these observations point to differences in metabolic activity between preadipocytes and adipocytes, we investigated the oxygen consumption of preadipocytes stimulated to undergo differentiation, and fibroblasts, by measuring the respiration with a Clark-type oxygen electrode. Preadipocytes had a significantly lower oxygen consumption than mature adipocytes. This advantage in respiration and the better revascularization of undifferentiated adipose tissue cells allow the development of innovative transplants and point to preadipocytes as promising tool to improve transplantations in adipose tissue reconstruction. © 2004 Elsevier B.V. All rights reserved. Keywords: Cell culture; Respiration; Oxygen consumption
1. Introduction ∗
Corresponding author. Present address: Praxisklinik Kaiserplatz, Kaiserstr. 14, 60311 Frankfurt, Germany. Tel.: +49 69 92884747; fax: +49 69 92884744. E-mail address: [email protected] (D.v. Heimburg). 1 First and second author are both first authors. 1569-9048/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.resp.2004.12.013
The reconstruction of soft tissue defects is a problem since an ideal filler material for the correction of congenital deformities and cancer defects still has to be found. Autologous mature adipose tissue has been used as free graft for the reconstruction of soft tissue defects for more than 100 years (Neuber, 1893) and is still in
108
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
use in lack of a better alternative, although the results are poor and unpredictable (Billings and May, 1989). The transplants are largely absorbed and replaced by fibrous tissue and oil cysts (Peer, 1956). These poor outcomes of free fat autotransplantation are thought to be due to the mature fat cells’ low tolerance of ischemia, hypoxia, and the slow rate of revascularization of the grafts (Smahel et al., 1990). Oxygen and nutrient supply are limiting factors for the survival of transplanted cells. The critical time span during which a non-adequate blood supply (tolerance of ischemia) and an insufficient concentration of oxygen (hypoxia) are tolerated is highly crucial in transplantation of adipose tissue. Mature fat cells which do not get in contact with surrounding capillaries by inosculation within the first 4 days will die (Smahel et al., 1990; von Heimburg and Pallua, 2001). In contrast to adipocytes, the transplantation of isolated and expanded autologous preadipocytes, located between mature fat cells in adipose tissue, represents a possible and highly promising source for soft tissue engineering (Smahel et al., 1990; von Heimburg et al., 2001a). Preadipocytes demonstrate in vitro differentiation and dedifferentiation under different conditions (Bj¨orntorp et al., 1980) and are a potential material for soft tissue engineering (Smahel, 1989) due to their ability to proliferate and differentiate into adipose tissue after transplantation (von Heimburg et al., 2003). It has been found that preadipocytes, compared to mature fat cells, survive a trauma of transplantation much more undamaged showing good attachment and re-accumulation of fat at the new location (Billings and May, 1989). In earlier studies, we were able to show that adipocytes which had been delipidated by mechanical in vivo-compression showed much better results after transplantation than did control grafts (von Heimburg et al., 1994). This is most likely the result of geometrical changes, i.e. easier revascularization and smaller distance between capillaries and mitochondria. However, it is likely that the differences in outcome between preadipocyte and adipocyte transplantation also depend on differing metabolic activities. We therefore here tested the hypothesis that mature adipocytes have a higher oxygen consumption than undifferentiated preadipocytes. This question has not been addressed in any way in the literature yet. In order to verify this assumption, we analyzed cellular metabolic activities in various adipose tissue cells by measuring
the oxygen consumption of undifferentiated and differentiated primary human preadipocytes, undifferentiated and differentiated 3T3-L1-preadipocytes, as well as primary human fibroblasts with a Clark-type oxygen electrode. Our results might help to improve the understanding of metabolism and resistance to hypoxia in different adipose tissue cells planned for transplantation. 2. Methods 2.1. Cell isolation and cell culture Preadipocytes were isolated out of freshly excised human subcutaneous adipose tissue of young adults (age: 18–29 years) at the Department of Plastic Surgery and Hand Surgery–Burn Center who underwent elective operations (e.g. abdominoplasty). Fibrous tissue and visible blood vessels were removed, the adipose tissue was minced and digested by collagenase CLS TypI 0.2% (Biochrom Berlin, Germany) at 37 ◦ C for 45 min under constant shaking. The relation tissue to enzyme was 1:1. Digestion was stopped by adding two volumes of Dulbecco’s modified Eagle medium (DMEM) 10% fetal calf serum (FCS). After filtration (250 m), the fat layer was removed and the cell suspension was centrifuged again. After resuspension of the pelleted cells in DMEM/F12 (3:1) 10% FCS (added 100 U ml−1 penicillin, 100 g ml−1 streptomycin), cells were seeded on tissue culture dishes (63.6 cm2 , Greiner, Solingen, Germany) with a seeding density of 3 × 104 cells/cm2 (Hauner et al., 1995). Pools of three different donors were cultured together. Cells were cultured at 37 ◦ C at 5% CO2 , medium was changed on day 2. Cellular growth, expansion, and differentiation of preadipocytes was carried out following two different established culturing protocols (Fig. 1). 2.1.1. Protocol 1 Primary preadipocytes and cell line 3T3L1-preadipocytes were grown to confluency in DMEM/F12 (3:1) medium supplemented with EGF (epidermal growth factor, 10 ng ml−1 , Sigma) (Hauner et al., 1995), 100 U ml−1 penicillin, and 100 g ml−1 streptomycin (Fig. 1A and B). Medium contained 10% FCS for primary human preadipocytes and 5% FCS for 3T3-L1-preadipocytes. Medium was
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
109
Fig. 1. Culturing protocols. 3T3-L1-preadipocytes and primary human preadipocytes were cultured as described in Section 2 following two different protocols. Serum-free DMEM/F12 (SFM) was supplemented with various substances as indicated. Shown in A and B is protocol 1 which only differs in the amount of FCS used, i.e. 5% for 3T3-L1-preadipocytes and 10% for primary human preadipocytes.
changed every second day until confluence was reached. Adipogenic conversion was then promoted for 14 days by changing medium to DMEM/F12 (3:1) 5% FCS with 100 U ml−1 penicillin, 100 g ml−1
streptomycin, 1 M insulin, 0.1 M cortisol, and 0.5 mM isomethylbutylxanthine (IBMX). After 1 week of incubation, medium was used as before but without cortisol and IBMX for another week.
110
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
2.1.2. Protocol 2 Primary human preadipocytes were grown in serumfree DMEM/F12 (3:1) added 100 U ml−1 penicillin, 100 g ml−1 streptomycin supplemented with 1 M biotin, 17 M pantothenat, 5 g ml−1 transferrin, and 1 M insulin (Fig. 1C). Starting on day 2, 200 M ascorbate and basic fibroblast growth factor (bFGF) were added to this medium, replaced by serum-free DMEM/F12 with 200 M ascorbate, 0.1 M cortisol, and 0.5 mM 3-isobutyl-1-methylxanthine (IBMX) from days 8 to 14. Starting on day 15, serum-free DMEM/F12 was supplemented with only 200 M ascorbate. Finally, measurements were carried out on day 25 after seeding. As controls, primary cells of the same donor (preadipocytes and fibroblasts) and cell line cells (3T3-L1-preadipocytes) were cultured in growth medium. Fibroblasts were isolated out of freshly excised human dermis of the same adipose tissue donors. Dermis was treated with collagenase CLS TypI 0.2% (Biochrom Berlin, Germany) at 37 ◦ C for 60 min under constant shaking. After filtration (250 m) and centrifugation (700 × g at 17 ◦ C for 7 min), the resuspended cells (DMEM/F12 (3:1) 10% FCS, 100 U ml−1 penicillin, 100 g ml−1 streptomycin) were plated and grown to confluency in DMEM/F12 (3:1) medium (containing 10% FCS, 100 U ml−1 penicillin, 100 g ml−1 streptomycin, 10 ng ml−1 endothelial growth factor (EGF), and used for respiration analyses.
the electrode unit, a 2 cm × 2 cm piece of cigarette paper was saturated with a 50% KCl-solution and uncreasedly placed on the electrodes. The depression in which the silver electrode was located was filled
2.2. Morphological cell analysis Monolayer cultures of all cell types were analyzed using a phase-contrast microscope at a 100× magnification. Cells were photographed prior to detachment. Main criteria for differentiation were an increasing number and size of visible lipid droplets as well as a change in morphology from elongated contours to a round shape (compare Fig. 2). 2.3. Measurements of respiratory rate by Clark-type oxygen electrode Respiration was measured using a Clark-type electrode which consisted of a platinum cathode and a silver anode both embedded into plastic. To prepare
Fig. 2. Morphology of primary human preadipocytes in non-differentiated and differentiated state. Shown are human preadipocytes before (A) differentiation and 8 days (B) and 14 days (C) after adipogenic conversion has been induced by differentiation medium (see Section 2, protocol 1). Magnification 50×.
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
111
Fig. 3. Clark-type electrode and measuring system in detail (A) Shown is the Clark-type electrode. The probe determines O2 concentration in aqueous solutions over a period of time. The electrode itself is located at the bottom of the chamber. The platinum cathode is located on top of a dome-like center which is surrounded by the anode of silver metal. These two electrodes are connected with each other by a thin layer of electrolytes (50% saturated KCl-solution). Directly placed on top of the dome, there is an oxygen-permeable teflon foil. If current is given to the electrodes, the platinum electrode is negatively charged whereas the silver electrode is positively charged. Oxygen diffuses through the teflon membrane to the platinum electrode and is reduced there. At the same time, the silver electrode is oxidised and accumulates silver chloride. The current which is produced by the electrode is proportional to the oxygen tension in the solution. (B) This graphic shows the setting of the measuring system with the Clark-type electrode being embedded in a chamber. The currencies detected by the measuring unit are sent to an analog–digital-converter, processed and forwarded to a computer in digital format. The incoming signals are then detected by a software from Hansatech and converted into graphs on a computer.
with additional electrolyte solution to maintain a sufficient reservoir of fluid (compare Fig. 3A). Thus, the cigarette paper established a uniform connection between the two electrodes. On top of this layer, a 2 cm × 2 cm teflon membrane was stretched without any folds to have tight contact to the electrode. Afterwards, a sealing rubber was firmly fastened around the narrowing underneath the platinum electrode in the dome-like protrusion of the electrode (Fig. 3A). The electrode unit was then placed and fixed on the bottom of the measuring chamber (Fig. 3B). A magnetic stirrer was positioned on top in the measuring chamber which had a fixed total volume of 1 ml. To maintain a
stable chamber temperature, a surrounding additional chamber was filled with water and connected with a water circulation system. This system included a water tub (temperature set to 37 ◦ C) and a pump (Fig. 3B). The currencies detected by the measuring unit were sent to an analog–digital-converter, processed and forwarded to a computer in digital format. The incoming signals were detected by a software from Hansatech and converted into graphs. Next, the measuring system was calibrated. Since the magnetic stirrer caused an electromagnetic interference on the measurements, it was used constantly during calibration and during all measurements.
112
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
Calibration of the electrode to zero was performed by using natriumdithionit to completely reduce the oxygen in the measuring solution. Calibration for the maximum value was carried out by measuring 1 ml of air-saturated PBS in the glass-made and sealed chamber. Oxygen tension of air-saturated medium was calculated by the formula of Truesdale and Downing (1954).
Determined was the gradient of the linear regression line in the interval between 10 and 25 min. The data detected by the software had the unit mV/reading und had to be converted into nmol O2 /min. Data acquisition was set to 5 readings/s, 1000 mV corresponded to an oxygen consumption of 0.241 mol/ml and the
Cs = (14.16 − 0.3943T + 0.007714T 2 1 −0.0000646T 3 ) 32
Cs = oxygen(µmol/ml),
T = temperature(◦ C) In PBS, an oxygen concentration of 0.214 mol/ml was calculated at a temperature of 37 ◦ C. Respiration of primary human differentiated preadipocytes, fibroblasts and 3T3-L1-cells (after adipogenic stimulation) and the non-differentiated matching controls of the same donors was measured. In the first series, a suitable concentration of cells in the chamber was determined. This was performed by using human dermal fibroblasts. Fibroblasts were detached from plates by trypsin and centrifuged (200 × g at 17 ◦ C for 10 min). After removal of the supernatant, cells were diluted and counted in a Neubauer’s chamber. In a volume of 0.1 ml of PBS, cell suspensions of 105 , 2 × 105 , 4 × 105 , 6 × 105 , 8 × 105 , 1 × 106 , 1.2 × 106 1.4 × 106 , 1.6 × 106 , 1.8 × 106 , 2 × 106 , and 3 × 106 cells were placed in the Clark-type electrode chamber and measured (in triplicate for each cell number). All investigations of the oxygen consumption in mesenchymal cells were carried out following the same procedure as described for human dermal fibroblasts with 4 × 105 cells in 0.1 ml PBS (n = 10). To determine the oxygen consumption of the system during the measuring period, PBS instead of cell suspensions was measured as blank (n = 10). Oxygen consumption analysis of the cell suspensions was started 1 min before cells were injected into the measuring chamber with a Hamilton syringe. Data were then collected over a period of 27 min. At the end of each cycle, the chamber was prepared for the next measurement by removing the cell suspension and rinsing with distilled water followed by PBS washing. First analyzed were the measurements of the cellnumber dependent oxygen consumption in fibroblasts, afterwards the cell type-specific consumption kinetics.
Fig. 4. Oxygen consumption of human dermal fibroblasts. An adequate cell concentration for the measuring chamber was determined by using human dermal fibroblasts that were detached from the plates, centrifuged and diluted in 1 ml of PBS to gain cell suspensions of 2 × 105 , 4 × 105 , 6 × 105 , 8 × 105 , 1 × 106 , 1.2 × 106 1.4 × 106 , 1.6 × 106 , 1.8 × 106 , 2 × 106 , and 3 × 106 cells. The cells were placed in the measuring chamber and measured. (A) Shown is the cell-number dependent oxygen consumption of fibroblasts at cell densities as indicated. Data are based on three individual experiments per every tested cell concentration. (B) Since increase in oxygen consumption was linear in the interval between 2 × 105 and 6 × 105 cells/ml, additional experiments were performed in this range analyzing the respiration of human dermal fibroblasts in the three concentrations 2 × 105 , 4 × 105 , and 6 × 105 cells/ml. Data are based on 10 individual experiments. Given is the initial regression line in the interval from 2 × 105 to 6 × 105 (dotted line) and the corrected graph subtracting the oxygen consumption by the detecting probe (dashed line). * p < 0.01.
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
measuring chamber had a total volume of 1 ml. Therefore, the loss in currency per reading was converted into nmol O2 /min by multiplication with the factor 0.0723. 2.4. Statistical analysis SigmaStat® 2.03 was used for statistical analyses. The measurements to define a suitable cell number for the electrode chamber were evaluated by calculating the gradient of the linear regression line (Fig. 4B). Data on fibroblasts in Fig. 4 were analyzed by T-test for disjunct samples. Differences at p < 0.01 were considered significant (* ). Data on oxygen consumption of various cell types in differentiated versus undifferentiated state (Fig. 5) were analyzed for significance by Mann–Whitney Utest. Differences at p < 0.05 were considered significant (* ) and differences at p < 0.01 highly significant (** ).
113
3. Results 3.1. Respiration in human dermal fibroblasts To establish the measuring system, cell suspensions of 2 × 105 to 3 × 106 were used. Initially, the system showed a line parallel to the x-axis proving that the oxygen content of the measuring chamber remained constant. When injecting cells into the chamber, which initially caused artefacts, we found a line with negative gradient which was constant for the interval between 10 and 25 min. Our analyses showed that the increase in oxygen consumption due to increasing cell numbers was linear in the interval between 2 × 105 and 6 × 105 cells/ml giving stable results of respiration (Fig. 4A). Focussing on this interval, we performed additional experiments analyzing the respiration of human dermal fibroblasts in the three concentrations 2 × 105 , 4 × 105 , and 6 × 105 cells (n = 10) (Fig. 4B). The results of the respiratory rate revealed an increasing oxygen consumption in a linear manner from 0.641 × 10−4 ± 0.212 × 10−4 nmol/min in 2 × 105 cells to 0.904 × 10−4 ± 0.196 × 10−4 nmol/min in 4 × 105 cells to 1.274 × 10−4 ± 0.192 × 10−4 nmol/ min in 6 × 105 cells (Fig. 4B). Calculating the regression line from these measurements, we found a linear correlation between cell number and oxygen consumption: O2 -consumption = 0.31 nmol O2 / min +1.58 ×10−6 nmol O2 / min ×cell number
Fig. 5. Oxygen consumption of mesenchymal cells. Fibroblasts were cultured to subconfluence in growth medium containing 5% FCS as described in Section 2. Human primary preadipocytes were either cultured in growth medium containing 10% FCS until confluence (lane 2) and then stimulated to differentiate in the presence of serum (lane 3), or kept in serum-free medium supplemented with differentiation-inducing factors (lane 4). 3T3-L1-preadipocytes were cultured in growth medium containing 5% FCS (lane 5, nonstimulated). Some cultures were then induced to differentiate by differentiation-inducing factors (lane 6, stimulated). Shown here is the oxygen consumption of the cultured cells in differentiated vs. non-differentiated state. Data are based on 10 individual experiments. For details on cultivation and differentiation media see Fig. 1B and C). * p < 0.05 and ** p < 0.01.
From this formula, we could calculate an oxygen consumption of 0.31 nmol/l in the absence of cells. This was caused by the currency detecting probe and was also confirmed by our experiments: when determining oxygen levels in the absence of cells in pure PBS, we found an oxygen consumption of 0.278 ± 0.115 nmol O2 /min. Since this value was significantly different from 0, it had to be subtracted from the oxygen consumption levels detected with cells. Hence, the corrected equation for the regression line was O2 -consumption = 0.03 nmol O2 / min +1.58 ×10−6 nmol O2 / min ×cell number The oxygen consumption of 2 × 105 fibroblasts was therefore 0.363 ± 0.212 nmol O2 /min. For 4 × 105
114
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
and 6 × 105 fibroblasts, we calculated values of 0.626 ± 0.196 nmol O2 /min and 0.996 ± 0.192 nmol O2 /min, respectively. Since both, results from 2 × 105 and 4 × 105 fibroblasts and the measurements from 4 × 105 and 6 × 105 fibroblasts, differed significantly from each other (p < 0.01), a cell number of 4 × 105 was chosen for the analyses in mesenchymal cells with and without stimulation of adipogenic differentiation. 3.2. Respiration in stimulated and non-stimulated mesenchymal cells Considering the corrected equation for the determination of oxygen consumption, unstimulated human primary preadipocytes revealed an oxygen consumption of 0.591 ± 0.302 nmol/min (n = 10) which was not significantly different from the results obtained with unstimulated fibroblasts (0.626 ± 0.196 nmol O2 /min). After a stimulation period of 14 days with differentiation medium in the presence of 10% FCS (Fig. 1B, protocol 1), the preadipocytes showed no morphological signs of differentiation and oxygen consumption was not significantly elevated (0.591 ± 0.302 nmol O2 /min, n = 10) (Fig. 5). In contrast to that, respiration rate was markedly increased in differentiated preadipocytes cultured in serum-free medium over the whole period (Fig. 1C, protocol 2). These cells showed all signs of differentiation, i.e. change in morphology, accumulation of lipid droplets and had an increased oxygen consumption of 2.865 ± 0.219 nmol O2 /min (n = 10) which was significantly elevated (p = 0.01) compared to an equal number of unstimulated preadipocytes taken from the same donor after the same time in vitro. Similar observations were made when examining the respiration of 3T3-L1-preadipocytes (Fig. 5, bars 5 and 6). When non-differentiated, these cells showed an oxygen consumption of 1.231 ± 0.332 nmol O2 /min (n = 10) which was markedly lower than in stimulated, morphologically differentiated 3T3-L1 cells (2.886 ± 0.69 nmol O2 /min, n = 10, p < 0.001) (Fig. 5). 4. Discussion The survival of transplanted cells depends on the supply with oxygen and nutrients. Tolerance to ischemia differs between various cell types and between
a differentiated and an undifferentiated state. Peer (1956) has found in histological studies that adipocytes tolerate an ischemic period of 4 days at maximum after transplantation of adipose tissue. Therefore, adipocytes inevitably undergo necrosis if not sufficiently connected to vasculature within 4 days. Rossatti transplanted 4 cm × 1 cm large pieces of adipose tissue onto the ear of rabbits and found that the marginal areas had established adequate vascular supply within the first 2–4 days. These areas survived whereas all central parts became necrotic due to delayed vascularization (Rossatti, 1960). Lentrodt transplanted free 3 cm × 3 cm × 1.5 cm large pieces of adipose tissue on minipigs. By using vital dye (disulphine-blue) which stains vascularised tissue, he showed that only peripheral and not central parts of the transplant were sufficiently connected to surrounding vessels within 4 days. After a period of 8 days, small vessels in central areas also showed staining. However, these central parts finally became necrotic (Lentrodt, 1971). Since the critical period of time during which adipose tissue tolerates ischemia is limited to 4 days, it is a compelling aim to extent this span of time for adipose tissue reconstruction. Eppley et al. analyzed rat adipose tissue transplants and discovered that the application of bFGF-dextran pearls (“drug delivery”) extended the survival time of the transplants. These effects were explained by the activation of preadipocytes within the adipose tissue transplant (Eppley et al., 1992). In earlier studies we showed that tissue expanders which had been implanted into rats and subsequently filled, caused a delipidation of adipocytes by mechanical in vivo-compression. Free transfer of these “delipidated” transplants resulted in better survival rates and less necrosis compared to control grafts (von Heimburg et al., 1994). This effect is supported by the results of the present study: differentiated, lipid containing preadipocytes have a higher oxygen consumption than immature preadipocytes. Klaus et al. (1995) confirmed that in brown adipose tissue, which generally has a significantly higher activity of metabolism, oxygen consumption of brown adipocytes is lower in vitro if differentiation is suppressed. However, a comparison of respiration in adipocytes versus preadipocytes of human white adipose tissue has not yet been reported. It is known that the enzymes lipoproteine lipase (LPL) or glycerophosphate dehydrogenase (GPDH)
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
are not expressed in preadipocytes but highly expressed in mature adipocytes and in differentiated preadipocytes (Smas and Sul, 1995; Rosen and Spiegelman, 2000). Since these enzymes are key regulators of lipid metabolism and highly expressed in mature fat cells, they confirm an elevated metabolic activity of these cells. Hence, this study aimed at confirming the differences in expression of metabolic enzymes by applying respirometric methods. In microrespirometric experiments, the determination of a suitable cell number is an important precondition since it has to be considered that the oxygen consumption rate per cell decreases with increasing cell density. This is especially relevant if different cell types are compared to each other as performed in these experiments. It must be considered that a cell number that has proven to be suitable for one cell type is not necessarily adequate in others. To eliminate this problem for the fibroblasts and stimulated preadipocytes as used here, we tested all parameters in rather low cell densities (see Section 2). The measurements of oxygen consumption with the Clark-type electrode reveal a significantly lower respiration rate in undifferentiated preadipocytes. The results of differentiated preadipocytes grown in serumfree medium show significantly higher respiration rates than in undifferentiated preadipocytes of the same patient. Preadipocytes that had been cultured and stimulated in serum-containing medium showed no signs of adipose conversion and had an unchanged oxygen consumption. The results from comparing differentiated and undifferentiated 3T3-L1-preadipocytes also confirm that cellular differentiation of 3T3-L1-adipocytes leads to a significantly higher respiration rate. Taking our experimental setting, oxygen consumption rate is most likely even higher in mature adipocytes than in converted preadipocytes. This is due to the fact that stimulated preadipocytes can only be properly analyzed as long as they remain attached to the culture dish. During the final steps of adipogenic maturation, cells detach from the dishes and the multiple intracellular vacuoles unite into one big central lipid droplet. Since lipid metabolism is maximal in these cell types, the contrast between preadipocytes and adipocytes is most likely even more prominent than shown by our results. We found that only morphologically differentiated preadipocytes that contained lipid droplets as
115
confirmed by light microscopy showed a significantly increased respiration. This effect was observed with primary human preadipocytes as well as cell line preadipocytes. Two different protocols which especially differed in the concentration or use of fetal calf serum (FCS), respectively, were evaluated in this study (compare Fig. 1). Our results demonstrate that preadipocytes which had been stimulated in the presence of 10% FCS did not commit adipogenic conversion. This is most likely due to differentiation-inhibiting effects of proteins in the FCS which especially become prominent at high concentrations of FCS (10%) (compare Fig. 5, lane 3). Therefore, oxygen consumption was not elevated in human preadipocytes differentiated in the presence of 10% FCS. Lower amounts of FCS (5%), however, as used for 3T3-L1-preadipocyte differentiation did not show inhibitory influences on adipogenic conversion (Fig. 5, lane 6). We conclude from our findings that undifferentiated preadipocytes have a higher tolerance to ischemia than mature adipocytes. In general, a reduced oxygen consumption of undifferentiated cells confirms a higher acceptance of ischemia which allows free transplantation with a high rate of surviving cells. Furthermore, preadipocytes have a 20-times smaller volume compared to the mature cells since the major volume of adult adipocytes is caused by the large cytoplasmic lipid inclusion consisting of up to 95% triglycerides. This difference in volume allows a quicker revascularization after transplantation and thus offers the development of transplants with significantly higher resistance. Ideally, differentiation first starts after successful re-establishment of the circulation in the transplanted tissue at the new location. Hormonal and nutritional signals affect adipocyte differentiation in a positive or negative manner (Ailhaud et al., 1992), and components involved in cell–cell or cell–matrix interactions are also pivotal in regulating the differentiation process (Gregoire et al., 1998). The different findings according to the varying concentrations of FCS further underline the central role of serum factor concentrations on the extend of adipogenic conversion. An understanding of these regulative elements allows to control and influence differentiation of preadipocytes, to inhibit maturation and thereby keep energy consumption on a low level.
116
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
In conclusion, we consider the transplantation of preadipocytes as a promising way to improve survival and enlarge the size of transplants in adipose tissue reconstruction. Finally, it might be the combination of both, the use of undifferentiated preadipocytes and the inhibition of differentiation in the transplant by adding maturation-inhibiting factors, which leads to the development of innovative strategies in adipose tissue transplantation. Successful differentiation of transplanted human preadipocytes to mature adipose tissue could be demonstrated in several in vivo studies (von Heimburg et al., 2001a, 2001b, 2003).
Acknowledgements We gratefully acknowledge the advice of Ms S. Klaus in establishing the methods.
References Ailhaud, G., Grimaldi, P., N´egrel, R., 1992. Cellular and molecular aspects of adipose tissue development. Ann. Rev. Nutr. 12, 207–233. Billings Jr., E., May Jr., J.W., 1989. Historical review and present status of free fat graft autotransplantation in plastic and reconstructive surgery. Plast. Reconstr. Surg. 83, 368–381. Bj¨orntorp, P., Karlsson, M., Petterson, P., Sypniewska, G., 1980. Differentiation and function of rat adipocyte precursor cells in primary culture. J. Lipid Res. 21, 714–723. Eppley, B., Snyders, R.J., Winkelmann, T., Delfino, J., 1992. Autologous facial fat transplantation: improved graft maintenance by microbead bioactivation. J. Oral Maxillofac. Surg. 50, 477–482. Gregoire, F.M., Smas, C.M., Sul, H.S., 1998. Understanding adipocyte differentiation. Physiol. Rev. 78, 783–809. Hauner, H., Rohrig, K., Petruschke, T., 1995. Effects of epidermal growth factor (EGF), platelet-derived growth factor (PDGF) and
fibroblast growth factor (FGF) on human adipocyte development and function. Eur. J. Clin. Invest. 25, 90–96. Klaus, S., Ely, M., Encke, D., Heldmaier, G., 1995. Functional assessment of white and brown adipocyte development and energy metabolism in cell culture: Dissociation of terminal differentiation and thermogenesis in brown adipocytes. J. Cell. Sci. 108, 3171–3180. Lentrodt, J., 1971. Experimentelle Befunde u¨ ber die Vaskularisation von freien Fettgewebstransplantaten. Dt. Zahn-, Mund- und Kieferheilkunde 57, 15–20. Neuber, G., 1893. Fetttransplantation. Verh. Dtsch Ges Chir. 66. Peer, L., 1956. The neclected free fat graft. Am. J. Surg. 18, 233– 250. Rosen, E.D., Spiegelman, B.M., 2000. Molecular regulation of adipogenesis. Ann. Rev. Cell Dev. Biol. 16, 145–171. Rossatti, B., 1960. Revascularisation and phagocytosis in free fat autografts: an experimental study. Br. J. Plast. Surg. 13, 35–41. Smahel, J., 1989. Experimental implantation of adipose tissue fragments. Br. J. Plast. Surg. 42, 207–211. Smahel, J., Meyer, V., Sch¨utz, K., 1990. Vascular augmentation of free adipose tissue grafts. Eur. J. Plast. Surg. 13, 163–168. Smas, C.M., Sul, S.H., 1995. Control of adipocyte differentiation. Biochem. J. 309, 697–710. Truesdale, G.A., Downing, A.L., 1954. Solubility of oxygen in water. Nature 173, 1236. von Heimburg, D., Kuberka, M., Rendchen, R., Hemmrich, K., Rau, G., Pallua, N., 2003. Preadipocyte-loaded collagen scaffolds with enlarged pore size for improved soft tissue engineering. Int. J. Artif. Organs 26, 1064–1076. von Heimburg, D., Lemperle, G., Dippe, B., Kruger, S., 1994. Free transplantation of fat autografts expanded by tissue expanders in rats. Br. J. Plast. Surg. 47, 470–476. von Heimburg, D., Pallua, N., 2001. Two-year histological outcome of facial lipofilling. Ann. Plast. Surg. 46, 644–646. von Heimburg, D., Ulrich, D., Hemmrich, K., Pallua, N., 2001a. Soft tissue engineering by implantation of autologous adipose precursor cells into the rabbit ear-pathophysiology in adipose tissue transplantation. Clin. Exp. Plast. Surg. 33, 127–132. von Heimburg, D., Zachariah, S., Low, A., Pallua, N., 2001b. Influence of different biodegradable carriers on the in vivo behavior of human adipose precursor cells. Plast. Reconstr. Surg. 108, 411–420.
Oxygen consumption in undifferentiated versus differentiated adipogenic mesenchymal precursor cells Dennis von Heimburg a, ∗, 1 , Karsten Hemmrich a, 1 , Sascha Zachariah a , Harald Staiger b , Norbert Pallua a a
b
Department of Plastic Surgery and Hand Surgery–Burn Center, University Hospital of the Aachen University of Technology, Pauwelsstr. 30, 52057 Aachen, Germany Department of Internal Medicine IV, Eberhard-Karls-University T¨ubingen, Otfried-M¨uller-Street 10, 72076 T¨ubingen, Germany Accepted 6 December 2004
Abstract To date, no adequate implant material for the correction of soft tissue defects such as after extensive deep burns, tumor resections or in congenital defects is available. A biohybrid composed of viable adipose precursor cells and an optimised matrix could help towards a solution. Morphologically, preadipocytes resemble fibroblasts and have not yet built a large cytoplasmic lipid droplet as found in differentiated adipocytes. Additionally, preadipocytes are smaller than mature adipocytes allowing a quicker revascularization after transplantation. Furthermore, transplanted preadipocytes can form adipose tissue in vivo whereas the transplantation of mature adipocytes often gives poor results, i.e. oil cysts or shrinkage of the transplant. Since these observations point to differences in metabolic activity between preadipocytes and adipocytes, we investigated the oxygen consumption of preadipocytes stimulated to undergo differentiation, and fibroblasts, by measuring the respiration with a Clark-type oxygen electrode. Preadipocytes had a significantly lower oxygen consumption than mature adipocytes. This advantage in respiration and the better revascularization of undifferentiated adipose tissue cells allow the development of innovative transplants and point to preadipocytes as promising tool to improve transplantations in adipose tissue reconstruction. © 2004 Elsevier B.V. All rights reserved. Keywords: Cell culture; Respiration; Oxygen consumption
1. Introduction ∗
Corresponding author. Present address: Praxisklinik Kaiserplatz, Kaiserstr. 14, 60311 Frankfurt, Germany. Tel.: +49 69 92884747; fax: +49 69 92884744. E-mail address: [email protected] (D.v. Heimburg). 1 First and second author are both first authors. 1569-9048/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.resp.2004.12.013
The reconstruction of soft tissue defects is a problem since an ideal filler material for the correction of congenital deformities and cancer defects still has to be found. Autologous mature adipose tissue has been used as free graft for the reconstruction of soft tissue defects for more than 100 years (Neuber, 1893) and is still in
108
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
use in lack of a better alternative, although the results are poor and unpredictable (Billings and May, 1989). The transplants are largely absorbed and replaced by fibrous tissue and oil cysts (Peer, 1956). These poor outcomes of free fat autotransplantation are thought to be due to the mature fat cells’ low tolerance of ischemia, hypoxia, and the slow rate of revascularization of the grafts (Smahel et al., 1990). Oxygen and nutrient supply are limiting factors for the survival of transplanted cells. The critical time span during which a non-adequate blood supply (tolerance of ischemia) and an insufficient concentration of oxygen (hypoxia) are tolerated is highly crucial in transplantation of adipose tissue. Mature fat cells which do not get in contact with surrounding capillaries by inosculation within the first 4 days will die (Smahel et al., 1990; von Heimburg and Pallua, 2001). In contrast to adipocytes, the transplantation of isolated and expanded autologous preadipocytes, located between mature fat cells in adipose tissue, represents a possible and highly promising source for soft tissue engineering (Smahel et al., 1990; von Heimburg et al., 2001a). Preadipocytes demonstrate in vitro differentiation and dedifferentiation under different conditions (Bj¨orntorp et al., 1980) and are a potential material for soft tissue engineering (Smahel, 1989) due to their ability to proliferate and differentiate into adipose tissue after transplantation (von Heimburg et al., 2003). It has been found that preadipocytes, compared to mature fat cells, survive a trauma of transplantation much more undamaged showing good attachment and re-accumulation of fat at the new location (Billings and May, 1989). In earlier studies, we were able to show that adipocytes which had been delipidated by mechanical in vivo-compression showed much better results after transplantation than did control grafts (von Heimburg et al., 1994). This is most likely the result of geometrical changes, i.e. easier revascularization and smaller distance between capillaries and mitochondria. However, it is likely that the differences in outcome between preadipocyte and adipocyte transplantation also depend on differing metabolic activities. We therefore here tested the hypothesis that mature adipocytes have a higher oxygen consumption than undifferentiated preadipocytes. This question has not been addressed in any way in the literature yet. In order to verify this assumption, we analyzed cellular metabolic activities in various adipose tissue cells by measuring
the oxygen consumption of undifferentiated and differentiated primary human preadipocytes, undifferentiated and differentiated 3T3-L1-preadipocytes, as well as primary human fibroblasts with a Clark-type oxygen electrode. Our results might help to improve the understanding of metabolism and resistance to hypoxia in different adipose tissue cells planned for transplantation. 2. Methods 2.1. Cell isolation and cell culture Preadipocytes were isolated out of freshly excised human subcutaneous adipose tissue of young adults (age: 18–29 years) at the Department of Plastic Surgery and Hand Surgery–Burn Center who underwent elective operations (e.g. abdominoplasty). Fibrous tissue and visible blood vessels were removed, the adipose tissue was minced and digested by collagenase CLS TypI 0.2% (Biochrom Berlin, Germany) at 37 ◦ C for 45 min under constant shaking. The relation tissue to enzyme was 1:1. Digestion was stopped by adding two volumes of Dulbecco’s modified Eagle medium (DMEM) 10% fetal calf serum (FCS). After filtration (250 m), the fat layer was removed and the cell suspension was centrifuged again. After resuspension of the pelleted cells in DMEM/F12 (3:1) 10% FCS (added 100 U ml−1 penicillin, 100 g ml−1 streptomycin), cells were seeded on tissue culture dishes (63.6 cm2 , Greiner, Solingen, Germany) with a seeding density of 3 × 104 cells/cm2 (Hauner et al., 1995). Pools of three different donors were cultured together. Cells were cultured at 37 ◦ C at 5% CO2 , medium was changed on day 2. Cellular growth, expansion, and differentiation of preadipocytes was carried out following two different established culturing protocols (Fig. 1). 2.1.1. Protocol 1 Primary preadipocytes and cell line 3T3L1-preadipocytes were grown to confluency in DMEM/F12 (3:1) medium supplemented with EGF (epidermal growth factor, 10 ng ml−1 , Sigma) (Hauner et al., 1995), 100 U ml−1 penicillin, and 100 g ml−1 streptomycin (Fig. 1A and B). Medium contained 10% FCS for primary human preadipocytes and 5% FCS for 3T3-L1-preadipocytes. Medium was
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
109
Fig. 1. Culturing protocols. 3T3-L1-preadipocytes and primary human preadipocytes were cultured as described in Section 2 following two different protocols. Serum-free DMEM/F12 (SFM) was supplemented with various substances as indicated. Shown in A and B is protocol 1 which only differs in the amount of FCS used, i.e. 5% for 3T3-L1-preadipocytes and 10% for primary human preadipocytes.
changed every second day until confluence was reached. Adipogenic conversion was then promoted for 14 days by changing medium to DMEM/F12 (3:1) 5% FCS with 100 U ml−1 penicillin, 100 g ml−1
streptomycin, 1 M insulin, 0.1 M cortisol, and 0.5 mM isomethylbutylxanthine (IBMX). After 1 week of incubation, medium was used as before but without cortisol and IBMX for another week.
110
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
2.1.2. Protocol 2 Primary human preadipocytes were grown in serumfree DMEM/F12 (3:1) added 100 U ml−1 penicillin, 100 g ml−1 streptomycin supplemented with 1 M biotin, 17 M pantothenat, 5 g ml−1 transferrin, and 1 M insulin (Fig. 1C). Starting on day 2, 200 M ascorbate and basic fibroblast growth factor (bFGF) were added to this medium, replaced by serum-free DMEM/F12 with 200 M ascorbate, 0.1 M cortisol, and 0.5 mM 3-isobutyl-1-methylxanthine (IBMX) from days 8 to 14. Starting on day 15, serum-free DMEM/F12 was supplemented with only 200 M ascorbate. Finally, measurements were carried out on day 25 after seeding. As controls, primary cells of the same donor (preadipocytes and fibroblasts) and cell line cells (3T3-L1-preadipocytes) were cultured in growth medium. Fibroblasts were isolated out of freshly excised human dermis of the same adipose tissue donors. Dermis was treated with collagenase CLS TypI 0.2% (Biochrom Berlin, Germany) at 37 ◦ C for 60 min under constant shaking. After filtration (250 m) and centrifugation (700 × g at 17 ◦ C for 7 min), the resuspended cells (DMEM/F12 (3:1) 10% FCS, 100 U ml−1 penicillin, 100 g ml−1 streptomycin) were plated and grown to confluency in DMEM/F12 (3:1) medium (containing 10% FCS, 100 U ml−1 penicillin, 100 g ml−1 streptomycin, 10 ng ml−1 endothelial growth factor (EGF), and used for respiration analyses.
the electrode unit, a 2 cm × 2 cm piece of cigarette paper was saturated with a 50% KCl-solution and uncreasedly placed on the electrodes. The depression in which the silver electrode was located was filled
2.2. Morphological cell analysis Monolayer cultures of all cell types were analyzed using a phase-contrast microscope at a 100× magnification. Cells were photographed prior to detachment. Main criteria for differentiation were an increasing number and size of visible lipid droplets as well as a change in morphology from elongated contours to a round shape (compare Fig. 2). 2.3. Measurements of respiratory rate by Clark-type oxygen electrode Respiration was measured using a Clark-type electrode which consisted of a platinum cathode and a silver anode both embedded into plastic. To prepare
Fig. 2. Morphology of primary human preadipocytes in non-differentiated and differentiated state. Shown are human preadipocytes before (A) differentiation and 8 days (B) and 14 days (C) after adipogenic conversion has been induced by differentiation medium (see Section 2, protocol 1). Magnification 50×.
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
111
Fig. 3. Clark-type electrode and measuring system in detail (A) Shown is the Clark-type electrode. The probe determines O2 concentration in aqueous solutions over a period of time. The electrode itself is located at the bottom of the chamber. The platinum cathode is located on top of a dome-like center which is surrounded by the anode of silver metal. These two electrodes are connected with each other by a thin layer of electrolytes (50% saturated KCl-solution). Directly placed on top of the dome, there is an oxygen-permeable teflon foil. If current is given to the electrodes, the platinum electrode is negatively charged whereas the silver electrode is positively charged. Oxygen diffuses through the teflon membrane to the platinum electrode and is reduced there. At the same time, the silver electrode is oxidised and accumulates silver chloride. The current which is produced by the electrode is proportional to the oxygen tension in the solution. (B) This graphic shows the setting of the measuring system with the Clark-type electrode being embedded in a chamber. The currencies detected by the measuring unit are sent to an analog–digital-converter, processed and forwarded to a computer in digital format. The incoming signals are then detected by a software from Hansatech and converted into graphs on a computer.
with additional electrolyte solution to maintain a sufficient reservoir of fluid (compare Fig. 3A). Thus, the cigarette paper established a uniform connection between the two electrodes. On top of this layer, a 2 cm × 2 cm teflon membrane was stretched without any folds to have tight contact to the electrode. Afterwards, a sealing rubber was firmly fastened around the narrowing underneath the platinum electrode in the dome-like protrusion of the electrode (Fig. 3A). The electrode unit was then placed and fixed on the bottom of the measuring chamber (Fig. 3B). A magnetic stirrer was positioned on top in the measuring chamber which had a fixed total volume of 1 ml. To maintain a
stable chamber temperature, a surrounding additional chamber was filled with water and connected with a water circulation system. This system included a water tub (temperature set to 37 ◦ C) and a pump (Fig. 3B). The currencies detected by the measuring unit were sent to an analog–digital-converter, processed and forwarded to a computer in digital format. The incoming signals were detected by a software from Hansatech and converted into graphs. Next, the measuring system was calibrated. Since the magnetic stirrer caused an electromagnetic interference on the measurements, it was used constantly during calibration and during all measurements.
112
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
Calibration of the electrode to zero was performed by using natriumdithionit to completely reduce the oxygen in the measuring solution. Calibration for the maximum value was carried out by measuring 1 ml of air-saturated PBS in the glass-made and sealed chamber. Oxygen tension of air-saturated medium was calculated by the formula of Truesdale and Downing (1954).
Determined was the gradient of the linear regression line in the interval between 10 and 25 min. The data detected by the software had the unit mV/reading und had to be converted into nmol O2 /min. Data acquisition was set to 5 readings/s, 1000 mV corresponded to an oxygen consumption of 0.241 mol/ml and the
Cs = (14.16 − 0.3943T + 0.007714T 2 1 −0.0000646T 3 ) 32
Cs = oxygen(µmol/ml),
T = temperature(◦ C) In PBS, an oxygen concentration of 0.214 mol/ml was calculated at a temperature of 37 ◦ C. Respiration of primary human differentiated preadipocytes, fibroblasts and 3T3-L1-cells (after adipogenic stimulation) and the non-differentiated matching controls of the same donors was measured. In the first series, a suitable concentration of cells in the chamber was determined. This was performed by using human dermal fibroblasts. Fibroblasts were detached from plates by trypsin and centrifuged (200 × g at 17 ◦ C for 10 min). After removal of the supernatant, cells were diluted and counted in a Neubauer’s chamber. In a volume of 0.1 ml of PBS, cell suspensions of 105 , 2 × 105 , 4 × 105 , 6 × 105 , 8 × 105 , 1 × 106 , 1.2 × 106 1.4 × 106 , 1.6 × 106 , 1.8 × 106 , 2 × 106 , and 3 × 106 cells were placed in the Clark-type electrode chamber and measured (in triplicate for each cell number). All investigations of the oxygen consumption in mesenchymal cells were carried out following the same procedure as described for human dermal fibroblasts with 4 × 105 cells in 0.1 ml PBS (n = 10). To determine the oxygen consumption of the system during the measuring period, PBS instead of cell suspensions was measured as blank (n = 10). Oxygen consumption analysis of the cell suspensions was started 1 min before cells were injected into the measuring chamber with a Hamilton syringe. Data were then collected over a period of 27 min. At the end of each cycle, the chamber was prepared for the next measurement by removing the cell suspension and rinsing with distilled water followed by PBS washing. First analyzed were the measurements of the cellnumber dependent oxygen consumption in fibroblasts, afterwards the cell type-specific consumption kinetics.
Fig. 4. Oxygen consumption of human dermal fibroblasts. An adequate cell concentration for the measuring chamber was determined by using human dermal fibroblasts that were detached from the plates, centrifuged and diluted in 1 ml of PBS to gain cell suspensions of 2 × 105 , 4 × 105 , 6 × 105 , 8 × 105 , 1 × 106 , 1.2 × 106 1.4 × 106 , 1.6 × 106 , 1.8 × 106 , 2 × 106 , and 3 × 106 cells. The cells were placed in the measuring chamber and measured. (A) Shown is the cell-number dependent oxygen consumption of fibroblasts at cell densities as indicated. Data are based on three individual experiments per every tested cell concentration. (B) Since increase in oxygen consumption was linear in the interval between 2 × 105 and 6 × 105 cells/ml, additional experiments were performed in this range analyzing the respiration of human dermal fibroblasts in the three concentrations 2 × 105 , 4 × 105 , and 6 × 105 cells/ml. Data are based on 10 individual experiments. Given is the initial regression line in the interval from 2 × 105 to 6 × 105 (dotted line) and the corrected graph subtracting the oxygen consumption by the detecting probe (dashed line). * p < 0.01.
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
measuring chamber had a total volume of 1 ml. Therefore, the loss in currency per reading was converted into nmol O2 /min by multiplication with the factor 0.0723. 2.4. Statistical analysis SigmaStat® 2.03 was used for statistical analyses. The measurements to define a suitable cell number for the electrode chamber were evaluated by calculating the gradient of the linear regression line (Fig. 4B). Data on fibroblasts in Fig. 4 were analyzed by T-test for disjunct samples. Differences at p < 0.01 were considered significant (* ). Data on oxygen consumption of various cell types in differentiated versus undifferentiated state (Fig. 5) were analyzed for significance by Mann–Whitney Utest. Differences at p < 0.05 were considered significant (* ) and differences at p < 0.01 highly significant (** ).
113
3. Results 3.1. Respiration in human dermal fibroblasts To establish the measuring system, cell suspensions of 2 × 105 to 3 × 106 were used. Initially, the system showed a line parallel to the x-axis proving that the oxygen content of the measuring chamber remained constant. When injecting cells into the chamber, which initially caused artefacts, we found a line with negative gradient which was constant for the interval between 10 and 25 min. Our analyses showed that the increase in oxygen consumption due to increasing cell numbers was linear in the interval between 2 × 105 and 6 × 105 cells/ml giving stable results of respiration (Fig. 4A). Focussing on this interval, we performed additional experiments analyzing the respiration of human dermal fibroblasts in the three concentrations 2 × 105 , 4 × 105 , and 6 × 105 cells (n = 10) (Fig. 4B). The results of the respiratory rate revealed an increasing oxygen consumption in a linear manner from 0.641 × 10−4 ± 0.212 × 10−4 nmol/min in 2 × 105 cells to 0.904 × 10−4 ± 0.196 × 10−4 nmol/min in 4 × 105 cells to 1.274 × 10−4 ± 0.192 × 10−4 nmol/ min in 6 × 105 cells (Fig. 4B). Calculating the regression line from these measurements, we found a linear correlation between cell number and oxygen consumption: O2 -consumption = 0.31 nmol O2 / min +1.58 ×10−6 nmol O2 / min ×cell number
Fig. 5. Oxygen consumption of mesenchymal cells. Fibroblasts were cultured to subconfluence in growth medium containing 5% FCS as described in Section 2. Human primary preadipocytes were either cultured in growth medium containing 10% FCS until confluence (lane 2) and then stimulated to differentiate in the presence of serum (lane 3), or kept in serum-free medium supplemented with differentiation-inducing factors (lane 4). 3T3-L1-preadipocytes were cultured in growth medium containing 5% FCS (lane 5, nonstimulated). Some cultures were then induced to differentiate by differentiation-inducing factors (lane 6, stimulated). Shown here is the oxygen consumption of the cultured cells in differentiated vs. non-differentiated state. Data are based on 10 individual experiments. For details on cultivation and differentiation media see Fig. 1B and C). * p < 0.05 and ** p < 0.01.
From this formula, we could calculate an oxygen consumption of 0.31 nmol/l in the absence of cells. This was caused by the currency detecting probe and was also confirmed by our experiments: when determining oxygen levels in the absence of cells in pure PBS, we found an oxygen consumption of 0.278 ± 0.115 nmol O2 /min. Since this value was significantly different from 0, it had to be subtracted from the oxygen consumption levels detected with cells. Hence, the corrected equation for the regression line was O2 -consumption = 0.03 nmol O2 / min +1.58 ×10−6 nmol O2 / min ×cell number The oxygen consumption of 2 × 105 fibroblasts was therefore 0.363 ± 0.212 nmol O2 /min. For 4 × 105
114
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
and 6 × 105 fibroblasts, we calculated values of 0.626 ± 0.196 nmol O2 /min and 0.996 ± 0.192 nmol O2 /min, respectively. Since both, results from 2 × 105 and 4 × 105 fibroblasts and the measurements from 4 × 105 and 6 × 105 fibroblasts, differed significantly from each other (p < 0.01), a cell number of 4 × 105 was chosen for the analyses in mesenchymal cells with and without stimulation of adipogenic differentiation. 3.2. Respiration in stimulated and non-stimulated mesenchymal cells Considering the corrected equation for the determination of oxygen consumption, unstimulated human primary preadipocytes revealed an oxygen consumption of 0.591 ± 0.302 nmol/min (n = 10) which was not significantly different from the results obtained with unstimulated fibroblasts (0.626 ± 0.196 nmol O2 /min). After a stimulation period of 14 days with differentiation medium in the presence of 10% FCS (Fig. 1B, protocol 1), the preadipocytes showed no morphological signs of differentiation and oxygen consumption was not significantly elevated (0.591 ± 0.302 nmol O2 /min, n = 10) (Fig. 5). In contrast to that, respiration rate was markedly increased in differentiated preadipocytes cultured in serum-free medium over the whole period (Fig. 1C, protocol 2). These cells showed all signs of differentiation, i.e. change in morphology, accumulation of lipid droplets and had an increased oxygen consumption of 2.865 ± 0.219 nmol O2 /min (n = 10) which was significantly elevated (p = 0.01) compared to an equal number of unstimulated preadipocytes taken from the same donor after the same time in vitro. Similar observations were made when examining the respiration of 3T3-L1-preadipocytes (Fig. 5, bars 5 and 6). When non-differentiated, these cells showed an oxygen consumption of 1.231 ± 0.332 nmol O2 /min (n = 10) which was markedly lower than in stimulated, morphologically differentiated 3T3-L1 cells (2.886 ± 0.69 nmol O2 /min, n = 10, p < 0.001) (Fig. 5). 4. Discussion The survival of transplanted cells depends on the supply with oxygen and nutrients. Tolerance to ischemia differs between various cell types and between
a differentiated and an undifferentiated state. Peer (1956) has found in histological studies that adipocytes tolerate an ischemic period of 4 days at maximum after transplantation of adipose tissue. Therefore, adipocytes inevitably undergo necrosis if not sufficiently connected to vasculature within 4 days. Rossatti transplanted 4 cm × 1 cm large pieces of adipose tissue onto the ear of rabbits and found that the marginal areas had established adequate vascular supply within the first 2–4 days. These areas survived whereas all central parts became necrotic due to delayed vascularization (Rossatti, 1960). Lentrodt transplanted free 3 cm × 3 cm × 1.5 cm large pieces of adipose tissue on minipigs. By using vital dye (disulphine-blue) which stains vascularised tissue, he showed that only peripheral and not central parts of the transplant were sufficiently connected to surrounding vessels within 4 days. After a period of 8 days, small vessels in central areas also showed staining. However, these central parts finally became necrotic (Lentrodt, 1971). Since the critical period of time during which adipose tissue tolerates ischemia is limited to 4 days, it is a compelling aim to extent this span of time for adipose tissue reconstruction. Eppley et al. analyzed rat adipose tissue transplants and discovered that the application of bFGF-dextran pearls (“drug delivery”) extended the survival time of the transplants. These effects were explained by the activation of preadipocytes within the adipose tissue transplant (Eppley et al., 1992). In earlier studies we showed that tissue expanders which had been implanted into rats and subsequently filled, caused a delipidation of adipocytes by mechanical in vivo-compression. Free transfer of these “delipidated” transplants resulted in better survival rates and less necrosis compared to control grafts (von Heimburg et al., 1994). This effect is supported by the results of the present study: differentiated, lipid containing preadipocytes have a higher oxygen consumption than immature preadipocytes. Klaus et al. (1995) confirmed that in brown adipose tissue, which generally has a significantly higher activity of metabolism, oxygen consumption of brown adipocytes is lower in vitro if differentiation is suppressed. However, a comparison of respiration in adipocytes versus preadipocytes of human white adipose tissue has not yet been reported. It is known that the enzymes lipoproteine lipase (LPL) or glycerophosphate dehydrogenase (GPDH)
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
are not expressed in preadipocytes but highly expressed in mature adipocytes and in differentiated preadipocytes (Smas and Sul, 1995; Rosen and Spiegelman, 2000). Since these enzymes are key regulators of lipid metabolism and highly expressed in mature fat cells, they confirm an elevated metabolic activity of these cells. Hence, this study aimed at confirming the differences in expression of metabolic enzymes by applying respirometric methods. In microrespirometric experiments, the determination of a suitable cell number is an important precondition since it has to be considered that the oxygen consumption rate per cell decreases with increasing cell density. This is especially relevant if different cell types are compared to each other as performed in these experiments. It must be considered that a cell number that has proven to be suitable for one cell type is not necessarily adequate in others. To eliminate this problem for the fibroblasts and stimulated preadipocytes as used here, we tested all parameters in rather low cell densities (see Section 2). The measurements of oxygen consumption with the Clark-type electrode reveal a significantly lower respiration rate in undifferentiated preadipocytes. The results of differentiated preadipocytes grown in serumfree medium show significantly higher respiration rates than in undifferentiated preadipocytes of the same patient. Preadipocytes that had been cultured and stimulated in serum-containing medium showed no signs of adipose conversion and had an unchanged oxygen consumption. The results from comparing differentiated and undifferentiated 3T3-L1-preadipocytes also confirm that cellular differentiation of 3T3-L1-adipocytes leads to a significantly higher respiration rate. Taking our experimental setting, oxygen consumption rate is most likely even higher in mature adipocytes than in converted preadipocytes. This is due to the fact that stimulated preadipocytes can only be properly analyzed as long as they remain attached to the culture dish. During the final steps of adipogenic maturation, cells detach from the dishes and the multiple intracellular vacuoles unite into one big central lipid droplet. Since lipid metabolism is maximal in these cell types, the contrast between preadipocytes and adipocytes is most likely even more prominent than shown by our results. We found that only morphologically differentiated preadipocytes that contained lipid droplets as
115
confirmed by light microscopy showed a significantly increased respiration. This effect was observed with primary human preadipocytes as well as cell line preadipocytes. Two different protocols which especially differed in the concentration or use of fetal calf serum (FCS), respectively, were evaluated in this study (compare Fig. 1). Our results demonstrate that preadipocytes which had been stimulated in the presence of 10% FCS did not commit adipogenic conversion. This is most likely due to differentiation-inhibiting effects of proteins in the FCS which especially become prominent at high concentrations of FCS (10%) (compare Fig. 5, lane 3). Therefore, oxygen consumption was not elevated in human preadipocytes differentiated in the presence of 10% FCS. Lower amounts of FCS (5%), however, as used for 3T3-L1-preadipocyte differentiation did not show inhibitory influences on adipogenic conversion (Fig. 5, lane 6). We conclude from our findings that undifferentiated preadipocytes have a higher tolerance to ischemia than mature adipocytes. In general, a reduced oxygen consumption of undifferentiated cells confirms a higher acceptance of ischemia which allows free transplantation with a high rate of surviving cells. Furthermore, preadipocytes have a 20-times smaller volume compared to the mature cells since the major volume of adult adipocytes is caused by the large cytoplasmic lipid inclusion consisting of up to 95% triglycerides. This difference in volume allows a quicker revascularization after transplantation and thus offers the development of transplants with significantly higher resistance. Ideally, differentiation first starts after successful re-establishment of the circulation in the transplanted tissue at the new location. Hormonal and nutritional signals affect adipocyte differentiation in a positive or negative manner (Ailhaud et al., 1992), and components involved in cell–cell or cell–matrix interactions are also pivotal in regulating the differentiation process (Gregoire et al., 1998). The different findings according to the varying concentrations of FCS further underline the central role of serum factor concentrations on the extend of adipogenic conversion. An understanding of these regulative elements allows to control and influence differentiation of preadipocytes, to inhibit maturation and thereby keep energy consumption on a low level.
116
D. von Heimburg et al. / Respiratory Physiology & Neurobiology 146 (2005) 107–116
In conclusion, we consider the transplantation of preadipocytes as a promising way to improve survival and enlarge the size of transplants in adipose tissue reconstruction. Finally, it might be the combination of both, the use of undifferentiated preadipocytes and the inhibition of differentiation in the transplant by adding maturation-inhibiting factors, which leads to the development of innovative strategies in adipose tissue transplantation. Successful differentiation of transplanted human preadipocytes to mature adipose tissue could be demonstrated in several in vivo studies (von Heimburg et al., 2001a, 2001b, 2003).
Acknowledgements We gratefully acknowledge the advice of Ms S. Klaus in establishing the methods.
References Ailhaud, G., Grimaldi, P., N´egrel, R., 1992. Cellular and molecular aspects of adipose tissue development. Ann. Rev. Nutr. 12, 207–233. Billings Jr., E., May Jr., J.W., 1989. Historical review and present status of free fat graft autotransplantation in plastic and reconstructive surgery. Plast. Reconstr. Surg. 83, 368–381. Bj¨orntorp, P., Karlsson, M., Petterson, P., Sypniewska, G., 1980. Differentiation and function of rat adipocyte precursor cells in primary culture. J. Lipid Res. 21, 714–723. Eppley, B., Snyders, R.J., Winkelmann, T., Delfino, J., 1992. Autologous facial fat transplantation: improved graft maintenance by microbead bioactivation. J. Oral Maxillofac. Surg. 50, 477–482. Gregoire, F.M., Smas, C.M., Sul, H.S., 1998. Understanding adipocyte differentiation. Physiol. Rev. 78, 783–809. Hauner, H., Rohrig, K., Petruschke, T., 1995. Effects of epidermal growth factor (EGF), platelet-derived growth factor (PDGF) and
fibroblast growth factor (FGF) on human adipocyte development and function. Eur. J. Clin. Invest. 25, 90–96. Klaus, S., Ely, M., Encke, D., Heldmaier, G., 1995. Functional assessment of white and brown adipocyte development and energy metabolism in cell culture: Dissociation of terminal differentiation and thermogenesis in brown adipocytes. J. Cell. Sci. 108, 3171–3180. Lentrodt, J., 1971. Experimentelle Befunde u¨ ber die Vaskularisation von freien Fettgewebstransplantaten. Dt. Zahn-, Mund- und Kieferheilkunde 57, 15–20. Neuber, G., 1893. Fetttransplantation. Verh. Dtsch Ges Chir. 66. Peer, L., 1956. The neclected free fat graft. Am. J. Surg. 18, 233– 250. Rosen, E.D., Spiegelman, B.M., 2000. Molecular regulation of adipogenesis. Ann. Rev. Cell Dev. Biol. 16, 145–171. Rossatti, B., 1960. Revascularisation and phagocytosis in free fat autografts: an experimental study. Br. J. Plast. Surg. 13, 35–41. Smahel, J., 1989. Experimental implantation of adipose tissue fragments. Br. J. Plast. Surg. 42, 207–211. Smahel, J., Meyer, V., Sch¨utz, K., 1990. Vascular augmentation of free adipose tissue grafts. Eur. J. Plast. Surg. 13, 163–168. Smas, C.M., Sul, S.H., 1995. Control of adipocyte differentiation. Biochem. J. 309, 697–710. Truesdale, G.A., Downing, A.L., 1954. Solubility of oxygen in water. Nature 173, 1236. von Heimburg, D., Kuberka, M., Rendchen, R., Hemmrich, K., Rau, G., Pallua, N., 2003. Preadipocyte-loaded collagen scaffolds with enlarged pore size for improved soft tissue engineering. Int. J. Artif. Organs 26, 1064–1076. von Heimburg, D., Lemperle, G., Dippe, B., Kruger, S., 1994. Free transplantation of fat autografts expanded by tissue expanders in rats. Br. J. Plast. Surg. 47, 470–476. von Heimburg, D., Pallua, N., 2001. Two-year histological outcome of facial lipofilling. Ann. Plast. Surg. 46, 644–646. von Heimburg, D., Ulrich, D., Hemmrich, K., Pallua, N., 2001a. Soft tissue engineering by implantation of autologous adipose precursor cells into the rabbit ear-pathophysiology in adipose tissue transplantation. Clin. Exp. Plast. Surg. 33, 127–132. von Heimburg, D., Zachariah, S., Low, A., Pallua, N., 2001b. Influence of different biodegradable carriers on the in vivo behavior of human adipose precursor cells. Plast. Reconstr. Surg. 108, 411–420.