ARTICLE IN PRESS

0952-3278/$ - se

doi:10.1016/j.pl

�Correspondfax: +2712 321

E-mail addr

Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–45

www.elsevier.com/locate/plefa

Effects of arachidonic acid, docosahexaenoic acid and prostaglandinE2 on cell proliferation and morphology of MG-63 and MC3T3-E1

osteoblast-like cells

M. Coetzeea,�, M. Haaga, A.M. Jouberta, M.C. Krugerb

aDepartment of Physiology, University of Pretoria, PO Box 2034, Pretoria, 0001, South AfricabInstitute of Food, Nutrition and Human Health, Massey University, Private Bag 11222, Palmerston North, New Zealand

Received 2 August 2006; accepted 1 October 2006

Abstract

During bone remodelling bone is resorbed by osteoclasts and replaced again by osteoblasts through the process of bone

formation. Clinical trials and in vivo animal studies suggest that specific polyunsaturated fatty acids (PUFAs) might benefit bone

health. As the number of functional osteoblasts is important for bone formation the effects of specific PUFAs on in vitro

osteoblastic cell proliferation were investigated. Morphological studies were conducted to determine whether exposure of the cells to

these agents caused structural damage to the cells thereby yielding invalid results. Results from this study showed that arachidonic

acid (AA) and docosahexaenoic acid (DHA) both inhibit cell growth significantly at high concentrations. The anti-mitotic effect of

AA is possibly independent of PGE2 production, as PGE2 per se had little effect on proliferation. Further study is required to

determine whether reduced proliferation due to fatty acids could be due to increased differentiation of osteoblasts to the mature

mineralising osteoblastic phenotype.

r 2006 Elsevier Ltd. All rights reserved.

1. Introduction

The mature skeleton is a metabolically active organthat undergoes continuous remodelling by a process thatreplaces old bone with new bone. During the remodel-ling cycle bone is resorbed by osteoclasts and replacedagain by osteoblasts through the process of boneformation [1]. The bone formation rate in vivo is largelydetermined by the number of mature functioningosteoblasts, which in turn is determined by the rate ofreplication of osteoblastic progenitors and the life-spanof mature osteoblasts [2,3]. Agents stimulating cellproliferation do this by binding to receptors withintrinsic tyrosine kinase activity [4]. These receptorsshare a common signal transduction pathway that, via a

e front matter r 2006 Elsevier Ltd. All rights reserved.

efa.2006.10.001

ing author. Tel.: +2712 319 2445;

1679.

ess: magdalena.coetzee@up.ac.za (M. Coetzee).

complex kinase cascade, leads to cell proliferation. Oneof the components involved in the cascade is mitogen-activated protein (MAP) kinase, which has been shownto be essential in the proliferative response of several celltypes [5]. Stimulation of cell proliferation dependson the activity of the cell cycle. Cyclins and cyclin-dependent kinases (cdks) regulate the progressionthrough each stage of the cell cycle [6].

Osteoblasts originate from bone marrow stromalprecursor cells that then differentiate into matureosteoblasts [7]. Once the osteoblast has differentiatedand completed its cycle of matrix synthesis, it can eitherbecome a flattened lining cell on the bone surface, beburied in bone as an osteocyte, or undergo apoptosis(programmed cell death), which is a biological processthat eliminates unwanted or damaged cells [7]. It hasbeen shown that the majority of osteoblasts willeventually undergo apoptosis [2]. Apoptosis is an activeprocess that is controlled from within the cell by a large

ARTICLE IN PRESSM. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–4536

number of regulatory factors, but can be induced orinhibited by external factors through receptor-mediatedmechanisms [8,9].

Studies conducted over the past decade showed thatbone active hormones such as oestrogen (E2) [10,11] andparathyroid hormone (PTH) [12,13] are beneficial tobone. Supplementation of diets with polyunsaturatedfatty acids (PUFAs) also showed promising effectson bone [10,14,15]. It has been shown that PUFAsupplementation increases bone formation in animals[10,16] and an anti-resorptive effect has been observed inelderly women after 3 years of PUFA supplementation[17,18]. Results from animal studies suggest that somen-3 PUFAs could reduce the risk of osteoporosis andfractures [10,15]. It has also been shown that a reductionof the n-6/n-3 PUFA ratio could result in increased bonestrength in animals [16] and in humans [19].

Although results from clinical trials and in vivoanimal studies suggest that specific PUFAs might benefitbone health, the cellular mechanisms of differentPUFAs have not been clarified and need to beinvestigated. Changes in dietary PUFAs are reflectedin the composition of various tissues, including bonecells such as the osteoblasts [20,21]. The cellularpresence of specific PUFAs, therefore, could affectosteoblastic functioning via modulation of the synthesisof fatty acid products (e.g. prostaglandins), prolifera-tion, differentiation and synthesis of proteins e.g.receptor activator of nuclear factor-kb-ligand andosteoprotegerin. To determine whether n-3 and n-6PUFAs affect osteoblast cell proliferation in vitro in asimilar manner, MG-63 and MC3T3-E1 osteoblastswere exposed to arachidonic acid (AA) (representativeof the n-6 PUFA family), and docosahexaenoic acid(DHA) (representative of the n-3 PUFA family).Prostaglandin E2 (PGE2) a product of AA metabolismin osteoblasts and previously implicated in bone home-ostasis [22,23], was included in this study. Apart fromproliferation studies, morphological studies were con-ducted to determine whether exposing the cells toPUFAs and PGE2 at the concentrations applied in thisstudy, caused structural damage to the cells therebyyielding invalid results.

2. Materials and methods

2.1. Reagents and materials

Sigma Chemical Co (St. Louis, MO, USA) suppliedL-glutamine, crystal violet, trypan blue, AA, DHA,PGE2, propidium iodide and Hoechst no 33342. Heat-inactivated fetal calf serum (FCS) was obtained fromHighveld Biological (Pty) Ltd (Sandringham, SouthAfrica). DMEM was obtained from Sterilab Services(Kempton Park, SA). Gentamycin was supplied by

Gibco (Invitrogen Corp., Carlsbad, CA, USA). Allother chemicals were of analytical grade and suppliedby Sigma Chemical Co (St. Louis, MO, USA). Glasscoverslips and sterile cell cluster plates were supplied byLASEC (Johannesburg, South Africa).

2.2. Cell cultures and maintenance

MG-63 (human osteoblast-like, osteosarcoma-de-rived) cells were purchased from the American TypeCulture Collection (ATCC), Rockville, MD, USA.Nontransformed MC3T3-E1 mouse calvaria fibroblasts(established from the calvaria of an embryo/fetusC57BL/6 mouse) described to differentiate to osteo-blasts [24], were purchased from Deutsche Sammlungvon Mikroorganismen und Zellkulturen (DSMZ),Braunschweig, Germany. Cell cultures were maintainedin DMEM (with 10% heat-inactivated FCS) at 37 1C ina humidified atmosphere of 95% air and 5% CO2. Allcell cultures were supplemented with 2mM L-glutamineand gentamycin (25 mg/ml). Fatty acid stock solutionswere stored in small aliquots at �70 1C and the workingsolutions freshly prepared each time prior to their use.The final ethanol concentration in the culture mediumdid not exceed 0.2%. Previous studies in our laboratoryshowed no toxic effects of the ethanol vehicle at thisconcentration.

2.3. Cell culture media for proliferation studies

Ascorbic acid has been shown to stimulate prolifera-tion of MC3T3-E1 cells. This effect appears to bemediated through the stimulatory effect of ascorbic acidon collagen synthesis [25]. Dulbecco’s modified Eagle’smedium (DMEM), which is ascorbic acid free, wastherefore used for all experiments. In our experimentalconditions both cell lines tolerated DMEM well. Fetalcalf serum contains various growth factors, whichreportedly also affect cell proliferation [25,26]. To limitthe proliferative effects of high FCS levels, FCS contentin the culture media was limited to 5%.

2.4. Proliferation studies

Proliferation can best be evaluated over an extendedperiod of time; it was therefore decided to evaluate theeffects of the different agents over a period of 72 h.Longer periods are not suitable as cells then often reachconfluency, which causes contact inhibition. MG-63 andMC3T3-E1 cells were seeded in sterile 96-well cultureplates at a density of 3000 cells/well after trypan blueexclusion. After cells had attached firmly for a period of24 h, culture medium was replaced with DMEMcontaining 5% FCS. Vehicle (ethanol, 0.2%), PUFAs(AA and DHA) ranging from 2.5 to 20 mg/ml or PGE2

(concentrations ranging from 10�10 to 10�6M) were

ARTICLE IN PRESSM. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–45 37

then added. After 72 h, the experiment was terminatedby replacing growth medium with 1% glutaraldehyde inPBS for 15min. For determination of proliferation, anadaptation of the crystal violet staining procedure [27]was applied as follows: crystal violet solution (1%, inPBS) was added to the fixed cells for 30min, thereafterthe plates were immersed in running tap water for15min. After the plates had dried, 200 ml 0.2% TritonX-100 was added to each well and incubated at roomtemperature for 90min, and 100 ml of the liquid contentsubsequently transferred to 96-well microtiter plates.Absorbance (O.D.) was read on an ELX800 UniversalMicroplate Reader (Bio-Tek Instruments Inc., Analy-tical Diagnostic Products, Weltevreden, South Africa) ata wavelength of 570 nm. Triton X-100 (0.2% in water)was used as a blank. Crystal violet is a basic dye, whichstains cell nuclei [28]. Spectrophotometer readings ofcolour intensity are therefore an indication of DNAcontent and therefore cell numbers. Results are pre-sented as percentage relative to control. Three indepen-dent experiments were conducted (n ¼ 8).

2.5. Morphology study: haematoxylin and eosin (H&E)

cell staining

MG-63 and MC3T3-E1cells (200 000/well) wereseeded aseptically onto heat-sterilised coverslips in six-well culture plates. Growth medium with vehicle (0.2%ethanol), PUFAs (AA and DHA) ranging from 2.5 to20 mg/ml or PGE2 (concentrations ranging from 10�10 to10�6M) were then added to near confluent monolayersfor 48 h. At the end of culture, the experiment wasterminated by removing the coverslips from the clusterplates, inserting them into coverslip holders followed byexposure to Bouin’s fixative (AccustanTM Bouin’ssolution, Sigma) for 30min. Thereafter cells werestained by standard haematoxylin and eosin (H&E)staining procedures [29]. Photographs were taken with400 ASA film using a Nikon camera attached to a NikonOpthiphot microscope (Nikon, Tokyo, Japan).

2.6. Detection of apoptosis by Hoechst 33342 and

propidium iodide staining (HOE/PI staining)

Previous studies reported observation of apoptoticeffects within 24–48 h of exposing cells to apoptoticagents [2]. Thus, exposure of the cells to the test agentswere limited to 48 h. Apoptosis and oncosis wereassessed by viability staining as follows: 40 000 cellswere seeded in DMEM with 5% FCS onto heat-sterilised coverslips in sterile 24-well cluster plates andleft to adhere for 24 h. Culture medium containingvehicles, AA and DHA (ranging from 2.5 to 20 mg/ml)or PGE2 (ranging from 10�10 to 10�6M) was then addedto subconfluent cells for 48 h. DMEM without FCS wasemployed as positive control for apoptosis [2]. At the

end of culture, after treatment with vehicle and testagents, the growth medium was discarded and the cellsgently rinsed with PBS. Thereafter 500 ml/well HoechstNo 33342 (HOE) (0.5 mg/ml in PBS) was added to eachwell. After 30min of incubation at 37 1C, 125 ml of a0.5mg/ml PI solution was added directly to each well.Within 5min, coverslips were mounted on microscopeslides with mounting fluid (90% glycerol, 4% N-propyl-gallate, 6% PBS). Photographs were taken with 400ASA film on a Nikon Optiphot microscope (Nikon,Tokyo, Japan) with UV-light and a blue filter. While allcells take up HOE (blue), only cells with intact cellmembranes exclude propidium iodide (red). Viable cells’nuclei therefore stain blue, while the nuclei of cells withdecreased membrane integrity become bright red [30].Although apoptotic cells may have an aberrant appear-ance they stain blue, indicating that these cells still havefunctional cell membranes capable of excluding PI.

2.7. Statistics

For each of the agents tested (AA, DHA, and PGE2)three separate proliferation experiments were conducted(n ¼ 8). Data were expressed as mean7SD. Statisticalanalysis was performed using Statistics for Windowssoftware (version 2, Tallahassee, Florida, USA). Theresults were analysed with one way ANOVA followedby Bonferroni’s post-hoc testing. Po0.05 was consid-ered to be significant.

3. Results

3.1. Proliferation studies

3.1.1. Effects of arachidonic acid on the proliferation of

MG-63 cells and MC3T3-E1 cells after 72 h exposure

The effects of the n-6 PUFA AA on the proliferationof MG-63 and MC3T3-E1 cells are shown in Fig. 1.In the MG-63 cell line, compared to control, AAat concentrations of 2.5–5 mg/ml had no effect oncell proliferation, but AA at higher concentrations(10–20 mg/ml) inhibited proliferation significantly. Inthe MC3T3-E1 cell line, compared to control, AAinhibited cell proliferation at all concentrations. Thehighest concentration of AA (20 mg/ml) caused a 40–50% inhibition in both cell lines tested.

3.1.2. Effects of docosahexaenoic acid on the

proliferation of MG-63 cells and MC3T3-E1 cells after

72 h exposure

The effects of the n-3 PUFA DHA on the proliferationof MG-63 and MC3T3-E1 cells are depicted in Fig. 2. Inthe MG-63 cell line, DHA dose-dependently inhibited cellproliferation at concentrations of 5–20mg/ml. In theMC3T3-E1 cell line, DHA inhibited cell proliferation

ARTICLE IN PRESS

0

20

40

60

80

100

120

control 2.5 5 10 20

AA concentration (µg/ml)

Pro

lifer

atio

n (

% o

f co

ntr

ol)

*

*

MG-63

0

20

40

60

80

100

120

control 2.5 5 10 20

AA concentration (µg/ml)

Pro

lifer

atio

n (

% o

f co

ntr

ol)

MC3T3-E1

* **

*

Fig. 1. Effects of arachidonic acid on MG-63 and MC3T3-E1 cell proliferation. Cells were seeded at 3 000 per well in 96-well plates in DMEM with

5% FCS, preincubated for 24 h and treated for 72 h with vehicle (0.2% ethanol) (control) or AA (2.5–20mg/ml). Cell number was determined by

crystal violet staining as described in Materials and Methods and is presented as a percentage relative to control. Results shown are the mean7SD,

n ¼ 8. *Significant difference from control, Po0.05. The experiment was repeated three times, each experiment yielding comparable data.

0

20

40

60

80

100

120

control 2.5 5 10 20

DHA concentration (µg/ml)

Pro

lifer

atio

n (

% o

f co

ntr

ol)

*

*

*

MG-63

0

20

40

60

80

100

120

control 2.5 5 10 20

DHA concentration (µg/ml)

Pro

lifer

atio

n (

% o

f co

ntr

ol)

MC3T3-E1

* *

Fig. 2. Effects of docosahexaenoic acid on MG-63 and MC3T3-E1 cell proliferation. Cells were seeded at 3 000 per well in 96-well plates in DMEM

with 5% FCS, pre-incubated for 24 h and treated for 72 h with vehicle (0.2% ethanol) (control) or DHA (2.5–20mg/ml). Cell number was determined

by crystal violet staining as described in Materials and Methods and is presented as a percentage relative to control. Results shown are the

mean7SD, n ¼ 8. *Significant difference from control, Po0.05. The experiment was repeated three times, each experiment yielding comparable

data.

M. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–4538

only at higher concentrations of 10–20mg/ml. Comparedto the MC3T3-E1 cell line, the inhibitory effect of DHAon the MG-63 cell line was much more severe, causing aninhibition of more than 80% at 20mg/ml.

3.2. Effects of prostaglandin E2 on the proliferation of

MG-63 cells and MC3T3-E1 cells after 72 h exposure

The effects of PGE2 (10�10–10�6M) on the proliferation

of MG-63 and MC3T3-E1 cells are depicted in Fig. 3.Compared to control, PGE2 exposure caused a slightinhibition of cell proliferation in both cell lines. In the MG-63 cell line the greatest effect was observed at a relativelylow concentration of 10�10M (12% inhibition), while theproliferation of MC3T3-E1 cells were affected mostly athigher concentrations (10�8–10�7M) (8% inhibition).

4. Morphology study: haematoxylin and eosin (H&E) cell

staining

Haematoxylin and eosin staining was conducted todetermine whether exposing the cells to PUFAs andPGE2 caused structural damage to the cells.

4.1. MG-63 cells

Fig. 4 shows the effects of PUFAs on the morphologyof MG-63 cells. Exposing the cells to vehicle (0.2%ethanol) had no effect on cell morphology (Fig. 4A andB). Although 20 mg/ml of AA inhibited cell proliferationconsiderably (Fig. 1) no structural damage to the cellswas observed at this concentration (Fig. 4C and D).DHA had major effects on MG-63 cell proliferation asshown in Fig. 2. The photographed field of these cellsclearly shows that 48 h DHA exposure affected thegrowth pattern of these cells (Fig. 4E). Although normaldividing cells are visible, the cells are less confluent thanthe control cells (Fig. 4A). Apoptotic cells are visible athigher magnification (Fig. 4F). No damage or abnor-malities were detected in the MG-63 cells after exposureto PGE2 (results not shown).

4.2. MC3T3-E1 cells

Exposing the cells to vehicle (0.2% ethanol) or AAhad no effect on cell morphology (Fig. 5A–D).Although 20 mg/ml DHA inhibited MC3T3-E1 cellproliferation by 40% (Fig. 2), no morphological effects

ARTICLE IN PRESS

Fig. 4. Photomicrographs of haematoxylin and eosin (H&E) stained MG-63 cells after 48 h polyunsaturated fatty acid exposure. Cells were pre-

incubated in DMEM with 5% FCS for 24 h and subsequently exposed to vehicle (0.2% ethanol) (control) and PUFAs (AA and DHA) for 48 h. H&E

staining was then performed as described in Materials and Methods. (A) Control cells (exposed to vehicle only); (B) control cells (black arrow

indicates cells in anaphase); (C) cells exposed to 20 mg/ml AA; (D) cells exposed to 20 mg/ml AA (arrows indicate mitotic cells, An-anaphase,

M-metaphase); (E) cells exposed to 20mg/ml DHA; and (F) cells exposed to 20mg/ml DHA (arrow indicates an apoptotic cell). (A, C and E: original

magnification 100� ) (B, D and F: original magnification 400� ).

0

20

40

60

80

100

control 10-10 10-9 10-8 10-7 10-6

PGE2 concentration (M)

Pro

lifer

atio

n (

% o

f co

ntr

ol)

0

20

40

60

80

100

control 10-10 10-9 10-8 10-7 10-6

PGE2 concentration (M)

Pro

lifer

atio

n (

% o

f co

ntr

ol)* * *

MG-63 MC3T3-E1

**

Fig. 3. Effects of prostaglandin E2 on MG-63 and MC3T3-E1 cell proliferation. Cells were seeded at 3 000 per well in 96-well plates in DMEM with

5% FCS, pre-incubated for 24 h and treated for 72 h with vehicle (0.2% ethanol) (control) or PGE2 (10�10–10�6M). Cell number was determined by

crystal violet staining as described in Materials and Methods and presented as a percentage relative to control. Results shown are the mean7SD,

n ¼ 8. *Significant difference from control, Po0.05. The experiment was repeated three times, each experiment yielding comparable data.

M. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–45 39

of this PUFA on these cells were observed (Fig. 5E andF). DHA exposure did not cause apoptosis of these cellsand normal dividing cells were evident (Fig. 5F). No

morphological damage or abnormalities were detectedin the MC3T3-E1 cells after 48 h exposure to PGE2

(results not shown).

ARTICLE IN PRESSM. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–4540

4.3. Hoechst 33342 and propidium iodide (HOE/PI)

staining for detection of apoptosis

HOE/PI staining was performed to investigate theeffects of AA, DHA and PGE2 on the viability of MG-63andMC3T3-E1 cells. Exposure to vehicle (0.2% ethanol)did not compromise the membrane integrity of eitherMG-63 cells (Fig. 6A) or MC3T3-E1 cells (Fig. 7A) asevident by the cells’ nuclei staining blue only.

Previous studies showed that depriving cells of FCS for aperiod of 24–48h causes apotosis [2]. Our study confirmedthis observation as multinuclear apoptotic MG-63 andMC3T3-E1 cells could be detected after 48h of FCS-deprivation (Figs. 6B and 7B). Although AA inhibited cellproliferation in both cell lines (Fig. 1), no apoptotic cellswere detected after exposing these cells to 20mg/ml AA(Figs. 6C and 7C) thereby confirming the results obtainedfrom H&E staining of these cells (Figs. 4D and 5D).

Exposing MG-63 and MC3T3-E1 osteoblasts toDHA inhibited cell proliferation significantly. To

Fig. 5. Photomicrographs of haematoxylin and eosin (H&E) stained MC3T3

incubated in DMEM with 5% FCS for 24 h and subsequently exposed to vehi

staining was then performed as described in Materials and Methods. (A) Con

a mitotic cell in metaphase); (C) cells exposed to 20mg/ml AA; (D) cells expo

cells exposed to 20 mg/ml DHA; and (F) cells exposed to 20mg/ml DHA

magnification 100� ) (B, D and F: original magnification 400� ).

determine whether this observation could be attributedto apoptotic effects of DHA on these cells, HOE/PIviability staining was performed. In MC3T3-E1 cells noapoptotic cells were detected after 48 h exposure to20 mg/ml DHA (Fig. 7D). However, exposing MG-63cells to similar DHA concentrations resulted in theformation of multi-nucleated cells showing extensivenuclear blebbing (Fig. 6D), which may be an early stepin apoptosis [31], thereby confirming results obtained byH&E staining (Fig. 4F). PGE2 did not affect MG-63 orMC3T3-E1 cell viability (results not shown).

5. Discussion and conclusions

5.1. Proliferation studies

In vivo studies have shown that dietary PUFAs couldhave beneficial effects on bone [10,14–16]. The cellulareffects of PUFAs have, however, not been extensively

-E1 cells after 48 h polyunsaturated fatty acid exposure. Cells were pre-

cle (0.2% ethanol)(control) and PUFAs (AA and DHA) for 48 h. H&E

trol cells (exposed to vehicle only); (B) control cells (the arrow indicates

sed to 20 mg/ml AA (the arrow indicates a mitotic cell in telophase); (E)

(arrows indicate mitotic cells in metaphase). (A, C and E: original

ARTICLE IN PRESS

Fig. 6. Photomicrographs of MG-63 cells after Hoechst and propidium iodide (HOE/PI) fluorescent staining for detection of apoptosis. Cells were

pre-incubated in DMEM with 5% FCS for 24 h and subsequently exposed to vehicle (0.2% ethanol) (control) and PUFAs (AA and DHA) for 48 h.

HOE/PI fluorescent staining was then performed and photomicrographs taken as described in Materials and Methods. (A) Control cells (exposed to

vehicle only); (B) multinucleated cell formation after 48 h of FCS deprivation; (C) cells exposed to 20mg/ml AA for 48 h; and (D) cells exposed to

20 mg/ml DHA for 48 h (arrow indicates an apoptotic cell) (Original magnification of photomicrographs: 400� )

Fig. 7. Photomicrographs of MC3T3-E1 cells after Hoechst and propidium iodide fluorescent (HOE/PI) staining for detection of apoptosis. Cells

were pre-incubated in DMEM with 5% FCS for 24 h and subsequently exposed to vehicle (0.2% ethanol) (control) and PUFAs (AA and DHA) for

48 h. HOE/PI staining was then performed and photomicrographs taken as described in Materials and Methods. (A) Control cells (exposed to vehicle

only); (B) cells after being deprived of FCS for 48 h (arrow indicates a multinucleated apototic cell); (C) cells exposed to 20mg/ml AA for 48 h; and

(D) cells exposed to 20mg/ml DHA for 48 h. (Original magnification of photomicrographs: 400� .)

M. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–45 41

investigated. As the number of functional osteoblasts isimportant for bone formation, the effects of PUFAs aswell as PGE2 on osteoblast proliferation were investi-gated in the present study.

5.2. Polyunsaturated fatty acids and prostaglandin E2

Depending on the cell type, culture conditions andconcentrations PUFAs have been reported to either

ARTICLE IN PRESSM. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–4542

stimulate [32] or inhibit [27,31,33,34] the proliferation ofvarious cell types. The n-3 PUFAs eicosapentaenoic acid(EPA) and DHA have generally been described asinhibitors of cell proliferation [34–38], while the n-6PUFA AA has shown various effects depending on theorigin of the cell type and experimental conditions[37,39].

In our experimental model, AA as well as DHAinhibited proliferation in both osteoblastic celllines (Figs. 1 and 2). DHA inhibited proliferation ofthe MG-63 osteosarcoma-derived cell line to a higherdegree than that of the murine MC3T3-E1 osteo-blastic cells (80% inhibition versus 20% inhibition at20 mg/ml) (Fig. 2). The difference in response of thesetwo cell lines may be explained by the fact that theMG-63 osteoblast cell line is osteosarcoma-derived,while the MC3T3-E1 osteoblastic cell line is a normalcell line. Others have confirmed the anti-prolifera-tive effect of DHA in a variety of cancer cell lines[35,38].

The anti-proliferative effect of the PUFAs has beenattributed to inhibition of the G1 to S phase transitionof the cell cycle [31,36,40]. It was speculated that thiseffect may be due to inhibition of the expression oractivity of some cyclins or cyclin-dependent kinasesrelated to cell cycle progression [36]. Further mechan-isms whereby PUFAs may inhibit cell growth to beconsidered, may include the following: Incorporation ofPUFAs in the cell membrane with subsequent modifica-tion in fluidity and permeability has been proposedwhich may induce some changes in cell behaviour [41].PUFAs may also act as second messengers by promo-ting the transfer of signals from the cell surface to thenucleus, thereby affecting signaling mechanisms in-volved in cell proliferation [27]. Since tyrosine kinaseactivity is considered to be important for stimulation ofcell growth [4], one of the mechanisms whereby PUFAsinhibits cell growth might be through inhibition oftyrosine kinase activity. Joubert et al. [27] showed anunexpected increase in tyrosine kinase activity in twooesophageal cancer lines after PUFA exposure. How-ever, PUFAs may regulate tyrosine kinase activity inosteoblastic cells differently and this needs to beinvestigated.

As PGE2 is synthesised from AA [42], AA mayexert its cellular effects via prostaglandin synthesis[36,43]. Results from our laboratory showed that bothMG-63 and MC3T3-E1 cells produce considerableamounts of PGE2 when exposed to AA [44]. Todetermine whether the anti-mitotic effect of AA onthese cell lines was mediated via PGE2, the effect ofPGE2 exposure on cell proliferation was investigated.Compared to control, PGE2 caused a slight inhibitionof cell proliferation in both cell lines (Fig. 3). In theMG-63 cells the greatest effect was observed ata relatively low concentration of 10�10M (12%

inhibition), while the proliferation of MC3T3-E1cells was inhibited mostly at higher concentrations(10�8–10�7M) (8% inhibition). Others confirmed theanti-proliferative effect of PGE2 on osteoblast-like cells[45–47]. Although the mechanism by which prostaglan-dins inhibit cell proliferation is not fully understood, ithas been shown that prostaglandins exert their effectsthrough specific prostaglandin receptors (EP) located oncell membranes [48]. It is not known which EP receptorsMG-63 cells express but MC3T3-E1 cells predominantlyexpress EP1 and EP4 receptors [49]. We suggest that inour model, the inhibitory effect of AA on osteoblast cellproliferation is possibly independent of PGE2 produc-tion, as PGE2 per se had little effect on proliferation inthe cell lines tested (Fig. 3). Joubert et al. [27] reportedAA-mediated inhibition of cell proliferation in oeso-phageal carcinoma cells. In their study addition of thecyclo-oxygenase blocker indomethacin did not abolishthe inhibitory effect of AA on cell proliferation, therebydemonstrating that AA per se had an effect on thesecells, thereby supporting our observation.

The inhibitory effect of PUFAs on cell proliferationin our osteoblast-like cell system could be due to theformation of PUFA peroxidation products, such asmalondialdehyde [50], in the culture media. PUFAsare particularly susceptible to oxidation and areassociated with the lipid peroxidation chain reaction[51]. Lipid peroxidation products can react with othermolecules, such as proteins and DNA that are harmfulto the cell. These peroxidation products could causemembrane damage changing signal transduction andcell metabolism [51]. Although not in osteoblast-likecells, Shiina et al. [34] reported inhibition of vascularsmooth muscle cell proliferation by eicosapentaenoicacid (EPA) (n-3) that was reversed by the additionof antioxidants. This observation was confirmed byDommels et al. [50] who demonstrated that antioxi-dants such as vitamin E and vitamin C could partiallyreverse the AA- and EPA-induced decrease in prolifera-tion in human coleorectal carcinoma cells. Anotherstudy, however, showed that DHA had a dose-dependent inhibitory effect on rat uterine stromal cellproliferation which was independent of lipid peroxida-tion, since it was not reversed by the addition ofantioxidants [52].

Fatty acids and their products are also importantligands for the peroxisome proliferator activatedreceptors PPARg [53,54]. DHA in particular has beenshown to activate PPARg [54]. Activation ofPPARg and PPARd has been implicated in bothinhibition and induction of cell proliferation. Therefore,with both of these subtypes being expressed inosteoblastic cell lines and in rat calvaria cells [36],the effects of PUFAs on osteoblasts may bemediated by the activation of PPARg and needs to beinvestigated.

ARTICLE IN PRESSM. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–45 43

6. Morphological studies

Morphological studies were conducted to determinewhether exposing the osteoblastic cells to PUFAs andPGE2 caused structural damage to the cells therebyyielding invalid results. Previous studies showed thatPUFAs at high concentrations could be toxic to cells[31,33,51]. The PUFA concentrations (2.5–20 mg/ml)applied in this study, however, is regarded to be withinthe physiological ranges of serum free fatty acidsreported for humans and mice [55].

In our model, no morphological damage or abnor-malities were detected after exposing MG-63 andMC3T3-E1 cells to AA (2.5–20 mg/ml) or PGE2 (rangingfrom 10�10 to 10�6M). DHA exposure (2.5–20 mg/ml) toMC3T3-E1 cells also did not harm these cells andnormal dividing cells were evident. However, DHA atconcentrations of 10–20 mg/ml affected MG-63 celldensity considerably and a number of apoptotic cellscould be seen at high magnification (Fig. 4F). Hoechstand propidium iodide fluorescent staining confirmed thepresence of multi-nucleated cells with extensive nuclearblebbing (Fig. 6D) that may be indicative of an earlystep in apoptosis [31]. As DHA is highly unsaturated,apoptosis may be due to the formation of DHAperoxidation products in the culture media and per-oxides are known to enhance apoptosis [56]. MG-63cells, being osteosarcoma-derived, could be moresusceptible to DHA (and its peroxidation products)than normal cells such as MC3T3-E1 osteoblasts.Others confirmed the apoptotic effects of DHA oncancer cell lines [35,38]. Apart from cancer cell lines,DHA has been shown to induce apoptosis in normalproliferating human endothelial cells [57].

A well-established strategy for inducing apoptosisin cell culture is to remove serum from the growthmedium [2]. Serum provides components such asproteins, amino acids, lipids, growth factors, vitamins,hormones and attachment factors. It also acts as a pHbuffer, and provides protease inhibitors [26]. FreePUFAs are easily oxidised in culture media andperoxide levels may increase to cytotoxic levels in themedium [34]. Binding of PUFAs to albumin in serumprotects cells from the cytotoxic effects of highconcentrations of free fatty acids [36,38,51]. It has beendemonstrated that albumin not only prevents thecytotoxic action of PUFAs, but also interferes withboth the uptake of free unesterified fatty acids [58] andfree radical generation in tumour cells [59]. The relativelow fetal calf serum concentration of 5% in the culturemedia of our model may cause a decrease in DHA-binding to serum proteins which could result in anincreased amount of free DHA entering the cells [58]thereby causing harmful effects to these cells. It maytherefore be necessary to use higher FCS concentrationsin the culture medium in order to protect cells from

possible harmful effects of DHA and its oxidationproducts.

Although AA [31,51,58] as well as PGE2 [60]have been reported to be inducers of apoptosis, wedid not observe apoptotic cells in either MG-63 orMC3T3-E1 cell lines after 24–48 h exposure to AA orPGE2 at the concentrations used. Kim et al. [57] alsowere unable to demonstrate AA-induced apoptosis inhuman endothelial cells, thereby confirming ourresults. During apoptosis cell fragments ‘pinch off’ asseparate small membrane-bound vesicles known asapoptotic bodies that contain the condensed cytoplas-mic proteins and intact organelles with nuclearfragments [61]. Adjacent cells recognise the apoptoticbodies and rapidly eliminate them through phagocytosisthereby avoiding an inflammatory response. Apoptosisoccurs quickly and cells undergoing this form ofdeath disappear within hours without causing damageto surrounding cells or tissues [61]. Under in vitrocell culture conditions apoptotic bodies accumulatein the culture medium, since they cannot all beremoved physiologically through phagocytosis byneighbouring cells [62]. These floating apoptotic celldebris may not be detected when cell preparationsare fixed and stained for morphological studiesresulting in the underestimation of apoptosis in this cellmodel.

MG-63 cells, which are widely used as a model forhuman osteoblasts, are derived from human osteosar-coma tissue. Cancer cells are reported to be moresusceptible to PUFAs than normal cell lines [35,38].Compared to the normal murine MC3T3-E1 cell line,in our study, the MG-63 cell line was more susceptible toanti-proliferative effects of PUFAs and apoptosis.Osteosarcoma-derived cells have undergone an extendedperiod of abnormal growth in vivo; therefore, the cellregulatory mechanisms of these cells might differ fromthose in normal cells. The MG-63 cells may thus exhibita deregulated proliferation/differentiation relationship,which might affect their response to various bone activeagents [63]. Our results suggest that the MG-63 cell linemight not be a suitable model for investigating normalosteoblastic cell proliferation.

In summary, in this in vitro cell-based study, AAand DHA inhibited proliferation of the two cell linestested, thereby reducing the number of osteoblasts.However, osteoblasts in culture can differentiateinto mature mineralising osteoblasts when stimulatedwith osteogenic agents [64]. As there is a reciprocalrelationship between reduced proliferation and subse-quent induction of cell differentiation in vitro [64,65],further research should be conducted to investi-gate whether inhibition of cell proliferation byPUFAs in our model could be due to increaseddifferentiation to the mature mineralising osteoblasticphenotype.

ARTICLE IN PRESSM. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–4544

Acknowledgements

This material is based upon work supported by theResearch Development Programme of the University ofPretoria and the National Research Foundation, SouthAfrica under Grant number 2053854.

References

[1] M. Gowen, Cytokines and cellular interactions in the control of

bone remodelling, in: J.N.M. Heersche, J.A. Kanis (Eds.), Bone

and Mineral Research/8, Elsevier Science, Amsterdam, 1994,

pp. 77–114.

[2] R.L. Jilka, R.S. Weinstein, T. Bellido, A.M. Parfitt, S.C.

Manologas, Osteoblast programmed cell death (apoptosis):

modulation by growth factors and cytokines, J. Bone Miner.

Res. 13 (1998) 793–802.

[3] R.L. Jilka, R.S. Weinstein, T. Bellido, P. Roberson, A.M. Parfitt,

S.C. Manologas, Increased bone formation by prevention of

osteoblast apoptosis with parathyroid hormone, J. Clin. Invest.

104 (1999) 439–446.

[4] P. Van der Geer, T. Hunter, R.A. Lindberg, Receptor protein-

tyrosine kinases and their signal transduction pathways, Annu.

Rev. Cell Biol. 10 (1994) 251–337.

[5] M.H.G. Verheijen, L.H.K. Defize, Parathyroid hormone inhibits

mitogen-activated protein kinase activation in osteosarcoma cells

via a protein kinase A-dependent pathway, Endocrinology 136

(1995) 3331–3337.

[6] T. Onishi, W. Zhang, X. Cao, K. Hruska, The mitogenic effect of

parathyroid hormone is associated with E2F-dependent activa-

tion of cyclin-dependant kinase 1 (cdc2) in osteoblast precursors,

J. Bone Miner. Res. 12 (1997) 1596–1605.

[7] L.G. Raisz, Physiology and pathophysiology of bone remodelling,

Clin. Chem. 45 (1999) 1353–1358.

[8] D.E. Hughes, B.F. Boyce, Apoptosis in bone physiology and

disease, Mol. Pathol. 50 (1997) 132–137.

[9] A. Leach, Apoptosis: molecular mechanism for physiologic cell

death, Clin. Lab. Sci. 11 (1998) 346–349.

[10] C.K. Schlemmer, H. Coetzer, N. Claassen, M.C. Kruger,

Oestrogen and essential fatty acid supplementation corrects bone

loss due to ovariectomy in the female Sprague Dawley rat,

Prostaglandins Leukot. Essent. Fatty Acids 61 (1999) 381–390.

[11] H.M. Heshmati, S. Khosla, S.P. Robins, W.M. O’Fallon, L.J.

Melton, B.L. Riggs, Role of low levels of endogenous estrogen in

regulation of bone resorption in late postmenopausal women,

J. Bone Miner. Res. 17 (2002) 172–178.

[12] R.M. Locklin, S. Khosla, R.T. Turner, B.L. Riggs, Mediators of

the biphasic responses of bone to intermittent and continuously

adminstered parathyroid hormone, J. Cell Biochem. 89 (2003)

180–190.

[13] Y. Jiang, J.J. Zhao, B.H. Mitlak, O. Wang, H.K. Genant, E.F.

Eriksen, Recombinant human parathyroid hormone (1–34)

[Teriparatide] improves both cortical and cancellous bone

structure, J. Bone Miner. Res. 10 (2003) 1932–1941.

[14] G. Fernandes, R. Lawrence, D. Sun, Protective role of n-3 lipids

and soy protein in osteoporosis, Prostaglandins Leukot. Essent.

Fatty Acids 68 (2003) 361–372.

[15] D. Sun, A. Krishnan, K. Zaman, R. Lawrence, A. Bhattacharya,

G. Fernandes, Dietary n-3 fatty acids decrease osteoclastogenesis

and loss of bone mass in ovariectomized mice, J. Bone Miner.

Res. 18 (2003) 1206–1216.

[16] N. Claassen, H.C. Potgieter, M. Seppa, et al., Supplemented

gamma-linolenic acid and eicosapentaenoic acid influence bone

status in young male rats: effects on free urinary collagen

crosslinks, total urinary hydroxyproline, and bone calcium

content, Bone 16 (1995) 385S–392S.

[17] D.H. Van Papendorp, H. Coetzer, M.C. Kruger, Biochemical

profile of osteoporotic patients on essential fatty acid supple-

mentation, Nutr. Res. 15 (1995) 325–334.

[18] M.C. Kruger, H. Coetzer, R. de Winter, G. Gericke, D.H. van

Papendorp, Calcium, gamma-linolenic acid and eicosapentaenoic

acid supplementation in senile osteoporosis, Aging Clin. Exp.

Res. 10 (1998) 385–394.

[19] L.A. Weiss, E. Barrett-Connor, D. von Muhlen, Ratio of n-6 to

n-3 fatty acids and bone mineral density in older adults: the

Rancho Bernardo stud, Am. J. Clin. Nutr. 81 (2005) 934–938.

[20] M.A. Moyad, An introduction to dietary/supplemental omega-3

fatty acids for general health and prevention: Part I, Urol. Oncol.

23 (2005) 28–35.

[21] T.G. Atkinson, H.J. Barker, K.A. Meckling-Gill, Incorporation

of long-chain n-3 fatty acids in tissues and enhanced bone marrow

cellularity with docosahexaenoic acid feeding in post-weanling

Fischer 344 rats, Lipids 32 (1997) 293–302.

[22] H. Kawaguchi, C.C. Pilbeam, J.R. Harrison, L.G. Raisz, The role

of prostaglandins in the regulation of bone metabolism, Clin.

Ortop. 31 (1995) 36–46.

[23] W.S.S. Jee, Y.F. Ma, The in vivo anabolic actions of prostaglan-

dins in bone, Bone 21 (1997) 297–304.

[24] H. Sudo, H.-A. Kodama, Y. Amagai, S. Yamamoto, S. Kasai, In

vitro differentiation and calcification in a new clonal osteogenic

cell line derived from newborn mouse calvaria, J. Cell Biol. 96

(1983) 191–198.

[25] S.-I. Harada, T. Matsumoto, E. Ogata, Role of ascorbic acid in

the regulation of proliferation in osteoblast-like MC3T3-E1 cells,

J. Bone Miner. Res. 6 (1991) 903–908.

[26] D.P. Lennon, S.E. Haynesworth, R.G. Young, J.E. Dennis, A.I.

Caplan, A chemically defined medium supports in vitro prolifera-

tion and maintains the osteochondral potential of rat marrow-

derived mesenchymal stem cells, Exp. Cell Res. 219 (1995)

211–222.

[27] A.M. Joubert, A. Panzer, F. Joubert, M.-L. Lottering, P.C.

Bianchi, J.C. Seegers, Comparative study of the effects of

polyunsaturated fatty acids and their metabolites on cell growth

and tyrosine kinase activity in oesophageal carcinoma cells,

Prostaglandins Leukot. Essent. Fatty Acids 61 (1999) 171–182.

[28] R.J. Gillies, N. Didier, M. Denton, Determination of cell number

in monolayer cultures, Anal. Biochem. 159 (1986) 109–113.

[29] J. Kiernan, Histological and Histochemical Methods, Pergamon

Press, London, 1990, pp. 96–97.

[30] G. Ciancio, A. Pollack, M.A. Taupier, N.L. Block, G.L. Irvin III,

Measurement of cell-cycle phase-specific cell death using Hoechst

33342 and propidium iodide: preservation by ethanol fixation,

J. Histochem. Cytochem. 36 (1988) 1147–1152.

[31] J.C. Seegers, M. de Kock, M.-L. Lottering, et al., Effects of

gamma-linolenic acid and arachidonic acid on cell cycle progres-

sion and apoptosis induction in normal and transformed cells,

Prostaglandins Leukot. Essent. Fatty Acids 56 (1997) 271–280.

[32] C. Tessier, J.-M. Fayard, H. Cohen, J.-F. Pageaux, M. Lagarde,

C. Laugier, Docosahexaenoic acid is a potent inhibitor of rat

uterine stromal cell proliferation, Biochem. Biophys. Res.

Commun. 207 (1995) 1015–1021.

[33] M. De Kock, M.-L. Lottering, J.C. Seegers, Differential cytotoxic

effects of gamma-linolenic acid on MG-63 and HeLa cells,

Prostaglandins Leukot. Essent. Fatty Acids 51 (1994) 109–120.

[34] T. Shiina, T. Terano, J. Saito, Y. Tamura, S. Yoshida,

Eicosapentaenoic acid and docosahexaenopic acid suppress the

proliferation of vascular smooth muscle cells, Atherosclerosis 104

(1993) 95–103.

[35] A.P. Albino, G. Juan, F. Traganos, et al., Cell cycle arrest and

apoptosis of melanoma cells by docosahexaenoic acid: association

ARTICLE IN PRESSM. Coetzee et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 76 (2007) 35–45 45

with decreased pRb phosphorylation, Cancer Res. 60 (2000)

4139–4145.

[36] A.C. Maurin, P.M. Chavassieux, E. Vericel, P.J. Meunier, Role of

polyunsaturated fatty acids in the inhibitory effect of human

adipocytes on osteoblastic proliferation, Bone 31 (2002) 260–266.

[37] U. Danesch, P.C. Weber, A. Sellmayer, Differential effects of n-6

and n-3 polyunsaturated fatty acids on cell growth and early gene

expression in Swiss 3T3 fibroblasts, J. Cell Physiol. 168 (1996)

618–624.

[38] R.A. Siddiqui, L.J. Jenski, K. Neff, K. Harvey, R.J. Kovacs, W.

Stillwell, Docosahexaenoic acid induces apoptosis in Jurkat cells

by a protein phosphatase-mediated process, Biochim. Biophys.

Acta 1499 (2001) 265–275.

[39] T. Hori, Y. Yamanaka, M. Hayakawa, S. Shibamoto, N. Oku, F.

Ito, Growth inhibition of human fibroblasts by epidermal growth

factor in the presence of arachidonic acid, Biochem. Biophys. Res.

Commun. 169 (1990) 959–965.

[40] T. Terano, T. Tanaka, Y. Tamura, et al., Eicosapentaenoic acid

and docosahexaenoic acid inhibit vascular smooth muscle cell

proliferation by inhibiting phosphorylation of Cdk2-cyclinE

complex, Biochem. Biophys. Res. Commun. 254 (1999) 502–506.

[41] W. Stillwell, W. Ehringer, L.J. Jenski, Docosahexaenoic acid

increases permeability of lipid vesicles and tumor cells, Lipids 28

(1993) 103–108.

[42] W.L. Smith, Prostanoid biosynthesis and mechanism of action,

Am. J. Physiol. 263 (1992) F181–F191.

[43] J. Klein-Nulend, P.N. Bowers, L.G. Raisz, Evidence that

adenosine 3050-monophosphate mediates hormonal stimulation

of prostaglandin production in cultured mouse parietal bones,

Endocrinology 126 (1990) 1070–1075.

[44] M. Coetzee, M. Haag, N. Claassen, M.C. Kruger, Stimulation of

prostaglandin E2 (PGE2) production by arachidonic acid,

oestrogen and parathyroid hormone in MG-63 and MC3T3-E1

osteoblast-like cells, Prostaglandins Leukot. Essent. Fatty Acids

76 (2005) 423–430.

[45] M. Centrella, S. Casinghino, T.L. McCarthy, Differential actions

of prostaglandins in separate cell populations from fetal rat bone,

Endocrinology 135 (1994) 1611–1620.

[46] M.A. Fang, D.A. Kujubu, T.J. Hahn, The effects of prostaglan-

din E2, parathyroid hormone, and epidermal growth factor

on mitogenesis, signaling, and primary response genes in UMR

106-01 osteoblast-like cells, Endocrinology 131 (1992) 2113–2119.

[47] M. Ho, J.-K. Chang, L. Chuang, H. Hsu, G. Wang, Effects of

nonsteroidal anti-inflammatory drugs and prostaglandins on

osteoblastic functions, Biochem. Pharmacol. 58 (1999) 983–990.

[48] S. Narumiya, G.A. FitzGerald, Genetic and pharmacological

analysis of prostanoid receptor function, J. Clin. Invest. 108

(2001) 25–30.

[49] M. Suda, K. Tanaka, K. Natsui, et al., Prostaglandin E receptor

subtypes in mouse osteoblastic cell line, Endocrinology 137 (1996)

1698–1705.

[50] Y.E.M. Dommels, M.M.G. Haring, N.G.M. Keestra, G.M.

Alink, P.J. van Bladeren, B. van Ommen, The role of

cyclooxygenase in n-6 and n-3 polyunsaturated fatty acid

mediated effects on cell proliferation, PGE2 synthesis and

cytotoxicity in human colorectal carcinoma cell lines, Carcino-

genesis 24 (2003) 385–392.

[51] C. Pompeia, T. Lima, R. Curi, Arachidonic acid cytotoxity: can

arachidonic acid be a physiological mediator of cell death?, Cell

Biochem. Funct. 21 (2003) 97–104.

[52] C. Tessier, J.-M. Fayard, H. Cohen, J.-F. Pageaux, M. Lagarde,

C. Laugier, Docosahexaenoic acid is a potent inhibitor of rat

uterine stromal cell proliferation, Biochem. Biophys. Res.

Commun. 207 (1995) 1015–1021.

[53] B.M. Forman, J. Chen, R.M. Evans, Hypolipidemic drugs,

polyunsaturated fatty acids, and eicosanoids are ligands for

peroxisome proliferator-activated receptors a and d, Poc. Natl.

Acad. Sci. USA 94 (1997) 4312–4317.

[54] K. Yu, W. Bayona, C.B. Kallen, et al., Differential activation of

peroxisome proliferator-activated receptors by eicosanoids,

J. Biol. Chem. 270 (1995) 23975–23983.

[55] E.A. Nunez, Free fatty acids as modulators of the steroid

hormone message, Prostaglandins Leukot. Essent. Fatty Acids 48

(1993) 63–70.

[56] C.B. Thompson, Apoptosis in the pathogenesis and treatment of

disease, Science 267 (1995) 1456–1462.

[57] H.J. Kim, C.A. Vosseler, P.C. Weber, W. Erl, Docosahexaenoic

acid induces apoptosis in proliferating human endothelial cells,

J. Cell Physiol. 204 (2005) 881–888.

[58] A.M. Monjazeb, K.P. High, C. Koumenis, F.H. Chilton,

Inhibitors of arachidonic acid metabolism act synergistically to

signal apoptosis in neoplastic cells, Prostaglandins Leukot.

Essent. Fatty Acids 73 (2005) 463–474.

[59] G. Ramesh, U.N. Das, R. Koratkar, M. Padma, P.S. Sagar,

Effect of essential fatty acids on tumor cells, Nutrition 8 (1992)

343–347.

[60] D.M. Brown, G.L. Warner, J.E. Ales-Martınez, D.W. Scott, R.P.

Phipps, Prostaglandin E2 induces apoptosis in immature normal

and malignant B lymphocytes, Clin. Immunol. Immunopathol. 63

(1992) 221–229.

[61] A. Leach, Prostaglandin E2 induces apoptosis in immature

normal and malignant B lymphocytes, Apoptosis: molecular

mechanism for physiologic cell death, Clin. Lab. Sci. 11 (1998)

346–349.

[62] T.G. Cotter, S.V. Lennon, J.M. Glynn, D.R. Green, Microfila-

ment-disrupting agents prevent the formation of apoptotic bodies

in tumor cells undergoing apoptosis, Cancer Res. 52 (1992)

997–1005.

[63] P.V.N. Bodine, M. Trailsmith, B.S. Komm, Development

characterization of a conditionally transformed adult human

osteoblastic cell line, J. Bone Miner. Res. 11 (1996) 806–819.

[64] T.A. Owen, M. Aronow, V. Shalhoub, et al., Progressive

development of the rat osteoblast phenotype in vitro: reciprocal

relationships in expression of genes associated with osteoblast

proliferation and differentiation during formation of the bone

extracellular matrix, J. Cell Physiol. 143 (1990) 420–430.

[65] L.D. Quarles, D.A. Yohay, L.W. Lever, R. Caton, R.J.

Wenstrup, Distinct proliferative and differentiated stages of

murine MC3T3-E1 cells in culture: an in vitro model of osteoclast

development, J. Bone Miner. Res. 7 (1992) 683–692.