Polyamines Modulate Nitric Oxide Production and Cox-2 Gene Expression in Response to Mechanical Loading in Human Adipose Tissue-Derived Mesenchymal Stem Cells

Authors

  • Geuranne S. Tjabringa,

    1. Department of Oral Cell Biology, Academic Center of Dentistry Amsterdam, Universiteit van Amsterdam, and Vrije Universiteit Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Peter S. Vezeridis,

    1. Department of Oral Cell Biology, Academic Center of Dentistry Amsterdam, Universiteit van Amsterdam, and Vrije Universiteit Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Behrouz Zandieh-Doulabi,

    1. Department of Oral Cell Biology, Academic Center of Dentistry Amsterdam, Universiteit van Amsterdam, and Vrije Universiteit Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Marco N. Helder,

    1. Department of Oral Cell Biology, Academic Center of Dentistry Amsterdam, Universiteit van Amsterdam, and Vrije Universiteit Center, Amsterdam, The Netherlands
    2. Department of Orthopaedic Surgery, Vrije Universiteit University Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Paul I.J.M. Wuisman,

    1. Department of Orthopaedic Surgery, Vrije Universiteit University Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Jenneke Klein-Nulend Ph.D.

    Corresponding author
    1. Department of Oral Cell Biology, Academic Center of Dentistry Amsterdam, Universiteit van Amsterdam, and Vrije Universiteit Center, Amsterdam, The Netherlands
    • ACTA-Vrije Universiteit, Department of Oral Cell Biology, Van der Boechorststraat 7, NL-1081 BT Amsterdam, The Netherlands. Telephone: +31-20 444 8660; Fax: +31-20 444 8683
    Search for more papers by this author

Abstract

For bone tissue engineering, it is important that mesenchymal stem cells (MSCs) display a bone cell-like response to mechanical loading. We have shown earlier that this response includes increased nitric oxide (NO) production and cyclooxygenase-2 (COX-2) gene expression, both of which are intimately involved in mechanical adaptation of bone. COX-2 gene expression is likely regulated by polyamines, which are organic cations implicated in cell proliferation and differentiation. This has led to the hypothesis that polyamines may play a role in the response of adipose tissue-derived MSCs (AT-MSCs) to mechanical loading. The aim of this study was to investigate whether genes involved in polyamine metabolism are regulated by mechanical loading and to study whether polyamines modulate mechanical loading-induced NO production and COX-2 gene expression in human AT-MSCs. Human AT-MSCs displayed a bone cell-like response to mechanical loading applied by pulsating fluid flow (PFF), as demonstrated by increased NO production and increased gene expression of COX-2. Furthermore, PFF increased gene expression of spermidine/spermine N (1)-acetyltransferase, which is involved in polyamine catabolism, suggesting that mechanical loading modulates polyamine levels. Finally, the polyamine spermine was shown to inhibit both PFF-induced NO production and COX-2 gene expression, suggesting that polyamines modulate the response of human AT-MSCs to mechanical loading. In conclusion, this is the first study implicating polyamines in the response of human AT-MSCs to mechanical loading, creating opportunities for the use of polyamines in tissue engineering approaches targeting skeletal defects.

Introduction

Mesenchymal stem cells (MSCs) constitute a promising tool for tissue engineering approaches. These stem cells are characterized by self-renewal capacity, long life span, and the potential to differentiate into several lineages, including bone, cartilage, muscle, and fat [1]. Although numerous studies have focused on bone marrow-derived MSCs, recent studies have reported on the presence of 100–1,000-fold higher numbers of MSCs per volume in adipose tissue [2, 3]. Adipose tissue can be harvested from patients by minimally invasive methods and is in general abundantly available, creating opportunities for the use of adipose tissue-derived MSCs (AT-MSCs) in tissue engineering approaches targeting skeletal defects [4].

For bone tissue engineering it is crucial that MSCs differentiate into osteoblast-like cells and display a bone cell-like response to mechanical loading [5]. Mechanosensitivity is a characteristic of bone cells that regulates adaptation of bone mass and shape according to the mechanical demands [6]. Mechanical loading of bone induces bone deformation, resulting in an interstitial flow of fluid on the ostocytes forming a network in the calcified bone matrix [7, [8]–9]. This fluid flow evokes a cellular response resulting in bone adaptation [10, 11]. Osteocytes appear to be the main mechanosensitive cells in bone, whereas osteoblasts and periosteal fibroblasts are less responsive to mechanical stress by fluid flow [10, 11]. These mechanosensitive cells respond to fluid flow-exerted shear stress by producing signaling molecules such as nitric oxide (NO) and prostaglandins and are thought to signal to osteoblasts and osteoclasts to regulate bone remodeling [12, 13]. Prostaglandins play a key role in the functional adaptation of bone to mechanical loading [14, 15], and prostaglandin production starts with the conversion of arachidonic acid, which is released from phospholipids present in the cell membrane, into prostaglandin G2 [16]. Subsequently, the conversion of prostaglandin G2 results in the formation of prostaglandin H2 [16]. These reactions are mediated by cyclooxygenase (COX) enzymes [16, 17]. Isomerisation of prostaglandin H2 results in the biological active prostanoids, including prostaglandin E2 [16]. Two isoforms of COX have been described, the constitutive COX-1 and the inducible COX-2 [16, 18]. COX-2 has been demonstrated to be the mechanosensitive isoform [17, 19, 20] that is crucially involved in the response of bone tissue to mechanical loading [14]. Recent studies in our laboratory have demonstrated that goat-derived AT-MSCs, which have been stimulated in their osteogenic differentiation by 1,25-dihydroxyvitamin-D3, are mechanosensitive and show a bone cell-like response, including increased NO production and COX-2 gene expression, to mechanical loading by pulsating fluid flow (PFF) [5]. These mechanosensitive AT-MSCs may constitute an interesting target for bone tissue engineering approaches.

Interestingly, polyamines, which are organic cations derived from amino acids present in all mammalian cell types, have also been implicated in the regulation of COX-2 gene expression [21]. Depletion of polyamines by difluoromethylornithine treatment caused an induction of COX-2 gene expression in cells from the human adenocarcinoma cell line Caco-2. Furthermore, an association between polyamines and COX-2 was supported by a study demonstrating that treatment of Caco-2 cells with nonsteroidal anti-inflammatory drugs, which inhibit COX-1 and COX-2, increased spermidine/spermine N (1)-acetyltransferase (SSAT) gene expression and decreased polyamine levels in colon cancer cells [22]. Intracellular levels of polyamines are regulated by a complex network of factors that affect polyamine biosynthesis and catabolism via different enzymatic steps. SSAT regulates polyamine catabolism, whereas polyamine modulating factor 1 (PMF-1) modulates SSAT gene expression [23]. Polyamines are involved in various cellular processes, including proliferation, differentiation, apoptosis, and carcinogenesis [24, [25], [26]–27], and have been implicated in bone growth and development [21, 28, [29]–30]. Recent studies in our laboratory suggest a role for polyamines in the osteogenic differentiation of AT-MSCs [31], indicating that polyamines may constitute an interesting tool for bone tissue engineering approaches targeting skeletal defects. Because COX-2 is intimately involved in bone adaptation to mechanical loading [14] and polyamines modulate COX-2 gene expression [21], polyamines may also play a role in the response of AT-MSCs to mechanical loading.

To our knowledge, no studies have reported on the role of polyamines in the response of human AT-MSCs to mechanical loading. We hypothesize that genes involved in polyamine metabolism are mechanoresponsive in human AT-MSCs and that polyamines modulate the response of human AT-MSCs to mechanical loading. To test this hypothesis, human AT-MSCs were subjected to PFF, and the effect of PFF on gene expression of SSAT, which is involved in polyamine catabolism, was determined. Furthermore, we studied whether mechanical loading induced a bone cell-like response in human AT-MSCs, including increased NO production and COX-2 gene expression. Finally, we investigated whether polyamines modulate this mechanical loading-induced response in human AT-MSCs.

Materials and Methods

Resection and Tumescent Liposuction

Human subcutaneous adipose tissue samples were obtained from five female donors (mean age 36 years, range 34–38) as waste material after elective surgery and donated upon informed consent of the patients by the Departments of Plastic Surgery from various clinics in The Netherlands. Resection and tumescent liposuction were used to obtain adipose tissue from the abdomen and hip. For resection, the material was cut out en-block under general anesthesia. For tumescent liposuction, a hollow blunt-tipped cannula was introduced into the subcutaneous space through a small incision. To minimize blood loss and tissue contamination by peripheral blood cells prior to aspiration, the vasoconstrictive agents epinephrine (1 mg/ml; Centrafarm B.V., Etten-Leur, The Netherlands, http://www.centrafarm.nl), prilocaïne HCl (1.5%; Pharmacy, Vrije Universiteit [VU] University Medical Center, Amsterdam, The Netherlands, http://www.vumc.nl/english), and kenacort-A10 (10 mg/l; Bristol-Myers Squibb, New York, http://www.bms.com) were added to a 0.65% NaCl saline solution (Baxter, Utrecht, The Netherlands, http://www.baxter.com), which was infused into the adipose compartment. The adipose tissue was mechanically disrupted by moving the cannula through the adipose compartment, and the liposuction material was aspirated by gentle suction.

AT-MSC Isolation

To isolate the stromal vascular fraction, resected adipose tissue was chopped into small pieces, and both the lipoaspirate and the chopped resected adipose tissue were washed extensively with phosphate-buffered saline (PBS). Adipose tissue was digested with 0.1% collagenase A (Roche Diagnostics GmbH, Mannheim, Germany, http://www.roche.de) in PBS containing 1% bovine serum albumin (BSA) for 45 minutes at 37°C under intermittent shaking. After washing with Dulbecco's modified Eagle's medium (DMEM-glucose; Cambrex Bio Science Verviers S.p.r.l., Verviers, Belgium, http://www.cambrex.com) containing 10% fetal bovine serum (FBS; HyClone, Logan, UT, http://www.hyclone.com) and centrifugation for 10 minutes at 600g, the cell pellet was resuspended in PBS and passed through a 200-μm mesh (Beldico S.A., Marche-en-Famenne, Belgium, http://www.beldico.be) to remove debris. Subsequently, cells were subjected to ficoll density centrifugation (lymphoprep, ρ = 1.077 g/ml, osmolarity 280 ± 15 mOsm; Axis-Shield, Oslo, Norway, http://www.axis-shield.com) to remove contaminating erythrocytes. The cell-containing interface was harvested and washed with DMEM containing 10% FBS. Viability of the cells was assessed by the trypan blue exclusion assay using light microscopy. Three to 10 × 106 cells were resuspended in a mixture of DMEM containing 10% FBS and Cryoprotective medium (1:1; Freezing Medium; Cambrex Bio Science Verviers S.p.r.l.), frozen using a Kryosave (HCI Cryogenics BV, Hedel, The Netherlands, http://www.linde-gascryoservices.com), and stored in liquid nitrogen.

AT-MSC Culture

Upon thawing, human AT-MSCs were cultured in culture medium consisting of DMEM (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 10% FBS (HyClone), 500 μg/ml streptomycin sulfate (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), 600 μg/ml penicillin (Sigma-Aldrich), 50 μg/ml gentamycin (Invitrogen), and 2.5 μg/ml fungizone (Invitrogen). After reaching confluency, cells were passaged using 0.25% trypsin and 0.1% EDTA in PBS.

Human AT-MSC Characterization

Human AT-MSCs at passage one were analyzed by fluorescence-activated cell sorting (FACS) for the expression of the MSC surface markers CD105/endoglin and CD166/activated leukocyte cell adhesion molecule (ALCAM). Cells were washed in washing buffer (PBS/0.1% BSA) and blocked for 10 minutes with washing buffer containing 5% goat serum and 0.1% BSA. After washing, cells were incubated for 15 minutes with phycoerythrin-labeled anti-CD166/ALCAM (BD Biosciences, PharMingen, San Diego, http://www.bdbiosciences.com) monoclonal antibodies. To analyze CD105/endoglin expression, cells were incubated with anti-CD105/ALCAM monoclonal antibodies (Abcam plc, Cambridge, UK, http://www.abcam.com), washed in washing buffer, and incubated with a fluorescein isothiocyanate-labeled secondary antibody (Biotrend, Cologne, Germany, http://www.biotrend.com). Control cells were incubated with the same concentration isotype-matched control antibodies of irrelevant specificity (Dako Denmark A/S, Glostrup, Denmark, http://www.dako.dk). After washing in washing buffer, expression of the indicated markers was assessed with a FACScan (BD Biosciences, PharMingen).

Pulsating Fluid Flow

Human AT-MSCs (up to passage three) were seeded on poly-lysine-coated (50 ug/ml; poly(l-lysine) hydrobromide, molecular weight 15–30 × 104; Sigma-Aldrich) glass slides (size 2.5 × 6.5 cm) at 2 × 105 cells per glass slide and cultured overnight in a Petri dish. The next day, cells were subjected to 1 hour of PFF as described previously [10], in culture medium containing only 2% FBS. Briefly, PFF was generated by pumping 13 ml of medium through a parallel-plate flow chamber containing the AT-MSCs. The cells were subjected to a 5-Hz pulse with a mean shear stress of 0.6 Pa, a pulse amplitude of 0.3 Pa, and a peak shear stress rate of 8.4 Pa/second. Static control cultures were kept in a Petri dish under conditions similar to those of the experimental cultures (i.e., 37°C in a humidified atmosphere of 5% CO2 in air). To study the effect of polyamines on the response of human AT-MSCs to PFF, cells were preincubated for 30 minutes and subjected to 1 hour of PFF or static control treatment in medium containing 2% FBS with or without the polyamine spermine at 1 or 10 μM. Medium samples for measurement of NO concentrations were taken at 0, 10, 30, and 60 minutes of PFF or static control treatment. RNA was isolated from both PFF-treated and static control cells using Trizol reagent (Invitrogen) either directly or after postincubation without PFF for the indicated time periods in fresh medium. Subsequently, gene expression of SSAT, COX-1, COX-2, runx-2, osteopontin, and PMF-1 was determined.

Nitric Oxide

NO production was measured as nitrite (NO2) accumulation in conditioned medium using Griess reagent containing 1% sulfanilamide, 0.1% naphtylethelene-diamine-dihydrochloride, and 2.5 M H3PO4. Serial NaNO2 dilutions in nonconditioned medium were used as standard curve. The absorbance was measured at 540 nm with a microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA, http://www.bio-rad.com).

Real-Time Polymerase Chain Reaction

Total RNA was extracted using Trizol reagent (Invitrogen) according to the manufacturer's instructions. To increase RNA yield, 5 μg of glycogen (Roche Diagnostics GmbH) was added to RNA in isopropanol prior to centrifugation. Total RNA (500–750 ng) was reverse-transcribed using 250 U/ml transcriptor reverse transcriptase (Roche Diagnostics GmbH), 0.08 A260 units random primers (Roche Diagnostics GmbH), and 1 mM each dNTP (Invitrogen) in transcriptor reverse transcriptase reaction buffer for 30 minutes at 55°C followed by 5 minutes of inactivation of reverse transcriptase at 85°C. The cDNA was diluted (10×), and 2 μl of cDNA was used per reaction for real-time polymerase chain reaction (PCR) using a SYBRGreen reaction kit for 18S (3 mM MgCl2) or Fast start plus SYBRGreen kit (both from Roche Diagnostics GmbH) for SSAT, COX-1, COX-2, runx-2, osteopontin, and PMF-1 in a LightCycler (Roche Diagnostics GmbH). To avoid contamination with genomic DNA, intron-spanning primers (except for 18S) were designed (Table 1) using Clone manager suite software program, version 6 (Scientific & Educational Software, Cary, NC, http://www.scied.com), and used at a concentration of 1 μM.

Table Table 1.. Primers used for real-time polymerase chain reaction
original image

PCR conditions were as follows. For 18S: 1 minute of preincubation at 95°C, followed by 30 cycles of amplification at 95°C for 5 seconds, 56°C for 10 seconds, 72°C for 15 seconds, and 82°C for 5 seconds, followed by melting curve analysis; for SSAT: 10 minutes of preincubation at 95°C, followed by 35 cycles of amplification at 95°C for 5 seconds, 57°C for 5 seconds, 72°C for 15 seconds, and 82°C for 5 seconds followed by melting curve analysis; for COX-1: 10 minutes of preincubation at 95°C, followed by 35 cycles of amplification at 95°C for 10 seconds, 57°C for 8 seconds, 72°C for 10 seconds, and 82°C for 5 seconds, followed by melting curve analysis; for COX-2: 10 minutes of preincubation at 95°C, followed by 40 cycles of amplification at 95°C for 5 seconds, 57°C for 10 seconds, 72°C for 15 seconds, and 82°C for 5 seconds followed by melting curve analysis; for runx-2: 10 minutes of preincubation at 95°C, followed by 35 cycles of amplification at 95°C for 2 seconds, 55°C for 8 seconds, 72°C for 10 seconds, and 82°C for 5 seconds, followed by melting curve analysis; for osteopontin: 10 minutes of preincubation at 95°C, followed by 35 cycles of amplification at 95°C for 2 seconds, 57°C for 8 seconds, 72°C for 10 seconds, and 82°C for 5 seconds, followed by melting curve analysis; for PMF-1: 7 minutes of preincubation at 95°C, followed by 44 cycles of amplification at 95°C for 10 seconds, 58°C for 5 seconds, 72°C for 10 seconds, and 85°C for 5 seconds, followed by melting curve analysis.

With the LightCycler software, the crossing points were assessed and plotted versus the serial dilution of known concentrations of the standards derived from each gene. PCR efficiency (E) was obtained by using the formula E = 10−1/slope. Data were used only if the PCR efficiency calculated was between 1.85 and 2.0.

Data analysis was performed using the LightCycler software. After normalization for 18S housekeeping gene expression, relative target gene expression was determined.

Cell Viability Assay

Cells were seeded in 96-well plates and incubated for 90 minutes with various concentrations of spermine in culture medium containing 2% serum. After postincubation for 3 hours, 1 day, or 4 days in culture medium containing 10% serum, medium was replaced with medium containing (tetrazolium salt [WST-1]; Roche Diagnostics GmbH) reagent (1:10). The assay is based on a reaction between mitochondrial dehydrogenase released from viable cells and tetrazolium salt of WST-1. Absorbance at 450 nm was measured at different time periods, and cell viability was determined as the percentage of absorption as compared with cells not stimulated with spermine.

Statistical Analysis

Data on NO production were obtained from four separate experiments using four different donors, data on real-time PCR data were obtained from four to eight separate experiments using four different donors, and data on NO production and gene expression after spermine treatment were obtained from three separate experiments using three different donors. Data are expressed as mean ± SEM, and statistical analysis was performed using the Student's t test.

Results

Human AT-MSCs were characterized by FACS analysis using monoclonal antibodies directed against the MSC markers CD105/endoglin and CD166/ALCAM. Approximately 58% of human AT-MSCs at passage 1 were positive for CD105/endoglin (Fig. 1A), and 59% of the human AT-MSCs were positive for CD166/ALCAM (Fig. 1B). The expression of these stem cell markers in human AT-MSCs was similar at passage 3 and at passage 1.

Figure Figure 1..

Flow cytometric analysis of human adipose tissue-derived mesenchymal stem cells (AT-MSCs). After one passage of culturing, human AT-MSCs were stained with monoclonal antibodies directed against either (A) CD105/endoglin (MSC marker) or (B) CD166/ALCAM (activated leukocyte cell adhesion molecule, MSC marker). An isotype-matched monoclonal antibody served as a control, which is indicated as gray area. The signals are shown as white area.

For bone tissue engineering applications, it is important that human AT-MSCs differentiate along the osteogenic pathway and display a bone cell-like response to mechanical stimuli. This response to mechanical loading includes increases in NO production and in COX-2 gene expression, which mediates prostaglandin production by bone cells and is intimately involved in the response of bone tissue to mechanical loading. The effect of mechanical loading on NO production by human AT-MSCs was studied by subjecting cells to 1 hour of PFF or keeping cells under static control conditions. NO production was significantly increased (by threefold) after 60 minutes of PFF treatment as compared with the static control cultures (Fig. 2).

Figure Figure 2..

Effect of pulsating fluid flow (PFF) on nitric oxide (NO) production by human adipose tissue-derived mesenchymal stem cells. Cells were subjected to PFF (closed circles) for 1 hour or kept under static control conditions (open circles), and NO production in medium samples taken at various time points was determined. Data are expressed as mean ± SEM. Statistical analysis was performed using the Student's t test. Asterisk indicates significant effect of PFF, p < .05.

To study the effect of mechanical loading on COX-2 gene expression in human AT-MSCs, cells were subjected to 1 hour of PFF or kept under static control conditions. After postincubation for various time periods, gene expression of COX-2 was determined. PFF increased COX-2 gene expression by sixfold after 3 hours of postincubation and by fivefold after 6 hours of postincubation as compared with the static control cultures (Fig. 3A). In contrast to COX-2 gene expression, that of COX-1 was not affected by PFF (Fig. 3B). Prostaglandin production by human AT-MSCs was not affected during the 1 hour of PFF treatment as compared with static control treatment (data not shown). This may be explained by the time point of measurement of prostaglandin production which precedes upregulation of COX-2.

Figure Figure 3..

Response of human adipose tissue-derived mesenchymal stem cells to PFF. Cells were subjected to PFF for 1 hour or kept under static control conditions, and RNA was isolated after various postincubation periods. Gene expression of COX-2(A), COX-1(B), and SSAT(C) was determined by real-time polymerase chain reaction, and data were normalized to 18S gene expression. Data are presented as the ratio of PFF-treated control over static control and as mean ± SEM. Statistical analysis was performed using the Student's t test. Asterisk indicates significant effect of PFF, p < .05. Abbreviations: COX, cyclooxygenase; PFF, pulsating fluid flow; SSAT, spermidine/spermine N (1)-acetyltransferase.

In addition to increasing COX-2 gene expression, 1 hour of PFF increased gene expression of runx-2, which is a transcription factor involved in early stages of osteogenic differentiation, 3 hours after PFF treatment by 1.3-fold (data not shown). In contrast to runx-2 gene expression, that of osteopontin, a bone matrix molecule involved in later stages of osteogenic differentiation, was not significantly affected by PFF within the time frame of the experiment (data not shown).

To study whether mechanical loading regulates expression of polyamine-related genes, the effect of PFF on gene expression of SSAT, which regulates polyamine catabolism, was determined (Fig. 3C). Cells were subjected to PFF for 1 hour or kept under static control conditions, and gene expression of SSAT was determined by real-time PCR at various postincubation periods after PFF treatment. One hour of PFF treatment increased SSAT gene expression at 3 hours, but not immediately after PFF treatment, nor after 6 hours of postincubation. In addition, the effect of PFF on gene expression of PMF-1, which regulates SSAT gene expression, was determined. One hour of PFF did not affect PMF-1 gene expression at 3 hours after PFF treatment (data not shown). To test polyamine toxicity, the effect of polyamines on cell viability of human AT-MSCs was studied (Fig. 4A). Cells were treated for 90 minutes with different concentrations of spermine, and cell viability was determined after various postincubation periods. Spermine (0.1–1,000 μM) did not affect cell viability 3 hours after treatment. After 1 and 4 days of postincubation, only high concentrations of spermine (1,000 μM) affected cell viability. Therefore, in the following experiments, spermine was used at nontoxic concentrations of 1 and 10 μM.

Figure Figure 4..

Effect of the polyamine spermine on the response of human AT-MSCs to PFF. To determine the effect of spermine on cell viability, cells were seeded in 96-well plates and incubated for 90 minutes with various concentrations of spermine. After postincubation for 3 hours (closed squares), 1 day (open circles), or 4 days (closed circles), medium was replaced with medium containing WST reagent (1:10). (A): Cell viability was determined by measuring absorbance at 450 nm, and data are presented as the percentage of absorption as compared with cells not stimulated with spermine. To determine the effect of spermine on PFF-induced NO production and COX-2 gene expression, human AT-MSCs were preincubated for 30 minutes and then subjected to PFF for 1 hour or kept under static control conditions (stat) in medium with or without the polyamine spermine. (B): NO production in medium samples taken after 60 minutes PFF was determined. (C):COX-2 gene expression was determined 3 hours after PFF treatment. Data on COX-2 gene expression are presented as ratio of PFF-treated over time-matched static control, whereby the static control was set at 1. Data are presented as mean ± SEM, and statistical analysis was performed using the Student's t test. Asterisk indicates significant effect of spermine on PFF-induced NO production and COX-2 gene expression, p < .05. Abbreviations: AT-MSC, adipose tissue-derived mesenchymal stem cell; COX, cyclooxygenase; NO, nitric oxide; PFF, pulsating fluid flow; WST, tetrazolium salt.

To study whether polyamines modulate PFF-induced NO production and COX-2 gene expression, human AT-MSCs were preincubated in medium with or without the polyamine spermine (1 or 10 μM) for 30 minutes and subsequently subjected to PFF for 1 hour in medium with or without spermine (1 or 10 μM). Spermine was shown to decrease PFF-induced NO production (Fig. 4B). Furthermore, spermine decreased PFF-induced COX-2 (Fig. 4C) gene expression as measured at 3 hours after PFF treatment. Spermine did not affect COX-2 gene expression in static control cells (data not shown).

To study the effect of polyamines on gene expression of SSAT and PMF-1, human AT-MSCs were preincubated in medium with or without the polyamine spermine (1 or 10 μM) for 30 minutes and subsequently subjected to PFF or kept under static control conditions for 1 hour in medium with or without spermine (1 or 10 μM). In static control cultures, spermine at both 1 and 10 μM increased SSAT gene expression (Fig. 5A), whereas PMF-1 gene expression was increased by 1 μM spermine (Fig. 5B). In flow-treated cells, spermine did not affect gene expression of SSAT and PMF-1 (data not shown).

Figure Figure 5..

Effect of the polyamine spermine on gene expression of SSAT(A) and PMF-1(B) in static control cells. Human adipose tissue-derived mesenchymal stem cells were incubated for 90 minutes in medium with the polyamine spermine. After 3 hours of incubation in fresh medium without spermine, gene expression of SSAT and PMF-1 was determined. Data are presented as ratio of spermine-treated over nontreated control cells, whereby the nontreated control was set at 1. Data are presented as mean ± SEM, and statistical analysis was performed using the Student's t test. Asterisk indicates significant effect of spermine on SSAT and PMF-1 gene expression in static control cells, p < .05. Abbreviations: PMF-1, polyamine modulating factor 1; SSAT, spermidine/spermine N (1)-acetyltransferase.

Discussion

In the present study, we demonstrate that gene expression of SSAT, which regulates polyamine catabolism, is responsive to PFF, suggesting that mechanical loading regulates polyamine levels in human AT-MSCs. Furthermore, polyamines were found to decrease PFF-induced NO production and COX-2 gene expression in human AT-MSCs, suggesting that polyamines modulate the response of human AT-MSCs to mechanical loading (Fig. 6). Our findings, for the first time, implicate polyamines in the response of AT-MSCs to mechanical loading, which may open opportunities for the use of polyamines in bone tissue engineering approaches using AT-MSCs.

Figure Figure 6..

Proposed model of the response of human AT-MSCs to mechanical loading. In this model, mechanical loading increases NO production and gene expression of COX-2, both of which are implicated in the mechanical adaptation of bone. In addition, gene expression of runx-2, which is a transcription factor expressed during early stages of osteogenic differentiation, is increased. Furthermore, mechanical loading increases gene expression of the polyamine-related gene SSAT, resulting in regulation of polyamine levels. Regulation of polyamine levels subsequently results in modulation of the response of human AT-MSCs to mechanical loading. Finally, spermine also modulates SSAT gene expression, resulting in regulation of polyamine levels. Abbreviations: AT-MSC, adipose tissue-derived mesenchymal stem cell; COX, cyclooxygenase; NO, nitric oxide; SSAT, spermidine/spermine N (1)-acetyltransferase.

We found that PFF increases gene expression of SSAT, which regulates polyamine catabolism. Furthermore, gene expression of PMF-1, which regulates SSAT gene expression, was not affected 3 hours after PFF treatment. This suggests that PMF-1 is not involved in the response of human AT-MSCs to mechanical loading or that modulation of PMF-1 gene expression by PFF occurs at a time point different from that of SSAT. Intracellular levels of polyamines are regulated by a complex network of factors that affect polyamine biosynthesis and catabolism via different enzymatic steps [23]. Biosynthesis is initiated by conversion of the amino acid ornithine to the polyamine putrescine, and this polyamine can be converted to the polyamines spermidine and spermine. Catabolism involves acetylation of polyamines by SSAT, which targets them for cellular excretion [32, 33]. Therefore, the finding that PFF increases SSAT gene expression in human AT-MSCs suggests that PFF decreases polyamine levels in these cells. However, PFF may also initially increase polyamine levels. To restore polyamine homeostasis in the cells, a feedback mechanism may be initiated resulting in increased SSAT gene expression. Interestingly, spermine was shown to increase gene expression of both SSAT and PMF-1 in static control cultures, suggesting that polyamine homeostasis is regulated via a feedback mechanism.

Polyamines are involved in a variety of cellular processes, including proliferation and differentiation, and several studies have implicated a role for polyamines in bone growth and development [28, 34]. Polyamines have also been suggested to play a role in the calcification of preosseous cartilage [28]. In epiphyseal cartilage from calf scapulas, polyamines were more abundantly present in the ossifying area as compared with the resting region [28]. In addition, alkaline phosphatase activity, which is a marker for bone cell differentiation, was enhanced by the addition of polyamines [28]. Further support for a role of polyamines in bone is provided by the observation that parathyroid hormone, which is involved in bone remodeling, regulates the activity of enzymes involved in polyamine metabolism as well as polyamine levels in rabbit costal chondrocytes [34]. In addition to suggesting a role in bone growth and development, recent studies in our laboratory suggest a role for polyamines in the osteogenic differentiation of AT-MSCs [31]. In the present study, we demonstrate that polyamines also modulate PFF-induced NO production and COX-2 gene expression. This suggests that in addition to regulating differentiation of AT-MSCs, polyamines may modulate the response of human AT-MSCs to mechanical loading.

For bone tissue engineering, it is important that MSCs differentiate into osteoblast-like cells and display a bone cell-like response to mechanical loading. Several studies have demonstrated that MSCs are mechanosensitive [35, [36], [37]–38]. In vivo, unloading hindlimbs using a rat tail suspension model resulted in reduced mRNA levels for growth factors and osteopontin in marrow stromal cells [37]. In vitro, mechanical loading increased alkaline phosphatase levels and the number of cells expressing alkaline phosphatase in rat bone marrow stromal cells [35]. Furthermore, matrix mineralization in human bone marrow-derived MSCs was enhanced in response to cyclic strain, suggesting that bone marrow-derived MSCs are stimulated in their osteogenic differentiation by mechanical loading [38]. In contrast to the present study, these in vitro studies were performed in medium supplemented with factors inducing osteogenic differentiation [35, 36, 38]. Recently, Knippenberg et al. [5] compared mechanoresponsiveness of goat-derived AT-MSCs in medium supplemented with osteogenic factors (including 1,25-dihydroxyvitamin-D3) with that of AT-MSCs in medium without osteogenic factors. Only AT-MSCs that were stimulated toward osteogenic differentiation showed a bone-cell like response to PFF (i.e., increased NO production and upregulated COX-2 gene expression) [5]. In the present study, we demonstrate that human AT-MSCs are also responsive to mechanical loading by PFF, even without stimulation of osteogenic differentiation, as demonstrated by increased NO production and COX-2 gene expression. However, NO production was increased relatively late in human AT-MSCs (after 60 minutes of PFF treatment) as compared with goat-derived AT-MSCs that had been stimulated in their osteogenic differentiation (after 5 minutes of PFF treatment) and mature bone cells. Therefore, we can only speculate that the unstimulated AT-MSCs display a bone cell-like response to mechanical loading, which is delayed as compared with osteogenically stimulated AT-MSCs and mature bone cells. In addition, future studies are needed to determine whether human AT-MSCs that are stimulated in their osteogenic differentiation show an earlier NO response to PFF as compared with unstimulated human AT-MSCs.

In addition to regulation of NO production and COX-2 gene expression, we demonstrate that gene expression of runx-2, which is a bone-related transcription factor expressed during early stages of osteogenic differentiation, is increased by PFF. Different studies have demonstrated that mechanical loading affects osteogenic differentiation of MSCs [38, 39]. Fluid flow was shown to increase gene expression of late phenotypic markers of osteoblastic differentiation, including osteopontin and bone sialoprotein in bone marrow stromal cells that were exposed to medium containing osteogenic factors [39]. Furthermore, cyclic strain has been shown to induce matrix mineralization in bone marrow-derived human MSCs [38]. We show that 1 hour of PFF increased runx-2 gene expression in unstimulated human AT-MSCs after 3 hours postincubation. Gene expression of osteopontin, which is a bone matrix molecule involved in later stages of osteogenic differentiation, was not significantly affected at this time point. This suggests that 1 hour of PFF affects early differentiation markers in human AT-MSCs and that stimulation with a combination of mechanical and biological factors may result in enhanced osteogenic differentiation of these stem cells. The signal transduction pathway mediating cellular responses to PFF may include mitogen-activated protein kinase (MAPK). Preliminary studies in our laboratory have shown that PFF induces activation of the MAPK extracellular signal-regulated kinase (ERK)1/2 in human AT-MSCs, suggesting that MAPKs may mediate the response of human AT-MSCs to PFF. This assumption is supported by other studies demonstrating involvement of ERK1/2 in the response of human MSCs to mechanical loading [38, 40].

Conclusion

This is the first study demonstrating that PFF regulates expression of the polyamine-related gene SSAT in human AT-MSCs. Furthermore, polyamines modulate PFF-induced NO production and COX-2 gene expression. This study suggests that mechanical loading may regulate polyamine levels in human AT-MSCs and that polyamines modulate the response of human AT-MSCs to mechanical loading. For bone tissue engineering, it is important that MSCs differentiate into bone-like cells and display a bone cell-like response to mechanical loading. Our finding that polyamines modulate the response of human AT-MSCs to mechanical loading creates opportunities for the use of polyamines in tissue engineering approaches targeting skeletal defects.

Disclosures

The authors indicate no potential conflicts of interest.

Acknowledgements

This study was supported by the VU interfaculty research initiative MOVE. The work of Peter Vezeridis was supported by a grant from the U.S. Fulbright Fellowship Program. The Technology Foundation STW supported the work of Behrouz Zandieh-Doulabi (STW grant no. VPG.5935). We thank Gerrit Jan Schuurhuis and Guus Westra (Department of Hematology, VU University Medical Center) for help in FACS analysis, Florine van Milligen (Department of Pathology, VU University Medical Center) for kindly donating human AT-MSCs, and Ruud Bank for help in polyamine analysis.

Ancillary