Phytochemistry 98 (2014) 164–173
Contents lists available at ScienceDirect
Phytochemistry journal homepage: www.elsevier.com/locate/phytochem
The stilbenes resveratrol, pterostilbene and piceid affect growth and stress resistance in mammalian cells via a mechanism requiring estrogen receptor beta and the induction of Mn-superoxide dismutase Ellen L. Robb 1, Jeffrey A. Stuart ⇑ Department of Biological Sciences and Cold Climate Oenology and Viticulture Institute, Brock University, St. Catharines, ON L2S 3A1, Canada
a r t i c l e
i n f o
Article history: Received 15 June 2013 Received in revised form 25 November 2013 Available online 19 December 2013 Keywords: Stilbene Superoxide dismutase Mitochondria Reactive oxygen species Estrogen receptor beta Cell growth Stress resistance
a b s t r a c t The mitochondrial antioxidant enzyme, Mn superoxide dismutase (MnSOD), has been shown to confer cytoprotection and to regulate cell cycle progression. Resveratrol, a phytoestrogen found in red wines and other foods, has been previously reported to increase MnSOD protein levels and activity both in vitro and in vivo. Numerous structural analogues of resveratrol produced via the same stilbene synthesis pathway (e.g. pterostilbene and piceid) and also present in foods and red wine may be capable of eliciting the same effects. Furthermore, in humans resveratrol is rapidly metabolized to resveratrol-40 sulfate, resveratrol-3-glucuronide and other metabolites in vivo. Although these metabolites may accumulate to relatively high levels in plasma and tissues, little is known about their biological activities. Here the activities were compared of these stilbenes and stilbene metabolites in mammalian cells. Two key cellular activities associated with resveratrol were examined: inhibition of proliferative growth and increased stress resistance (important anti-cancer and cell protective activities, respectively). While resveratrol-40 -sulfate and resveratrol-3-glucuronide had no effect on either cell growth or stress resistance, both pterostilbene and piceid were at least as effective as resveratrol. Using pharmacological and genetic approaches, it was found that the effects of pterostilbene and piceid required an induction of the mitochondrial enzyme MnSOD and intact mitochondrial respiration. In addition, using estrogen receptor beta (ERbeta) knockout mouse myoblasts, it was demonstrated that the effects of stilbene compounds on cell growth and stress resistance all require ERbeta. Taken together, these results indicate that resveratrol, pterostilbene and piceid all activate the same mitochondrial response in mammalian cells, and therefore these latter two molecules might be as effective as resveratrol in eliciting positive health outcomes in vivo. Ó 2013 Elsevier Ltd. All rights reserved.
Introduction The mitochondrial matrix antioxidant enzyme, Mn-superoxide dismutase (MnSOD), confers protection against cell death (e.g., Venkataraman et al., 2004; Silva et al., 2005; Kowluru et al., 2006; Fisher and Goswami, 2008; Dumont et al., 2009). Interestingly, in addition to this cytoprotective ability, MnSOD exerts effects on cell replication (e.g. Li et al., 1995; Liu et al., 1997; Yan et al., 1996; Behrend et al., 2005; Sarsour et al., 2008) and differentiation (e.g. Velarde et al., 2012; Schneider et al., 2011). The ability of transgenic MnSOD over-expression to slow the growth rate of cancerous cells is of particular interest. Many cancerous cell lines are characterized by abnormally low levels of MnSOD expression,
⇑ Corresponding author. Tel.: +1 905 688 5550x4814; fax: +1 905 688 1855. E-mail address: [email protected] (J.A. Stuart). Current address: Mitochondrial Biology Unit, Medical Research Council, Cambridge, UK. 1
0031-9422/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.phytochem.2013.11.019
and re-establishing MnSOD expression to normal levels can slow the growth of these cells (Li et al., 1995; Liu et al., 1997; Yan et al., 1996; Behrend et al., 2005). Interestingly, the only enzymatic activity attributed to MnSOD is the conversion of the superoxide anion to hydrogen peroxide in the mitochondrial matrix, which suggests that its effect on cell proliferation and stress resistance is directly related to its influence on the redox environment of mitochondria, the mechanistic details of which are unclear. Identifying a means of inducing MnSOD expression in cancerous cells through dietary or pharmacological interventions may have therapeutic potential. A robust upregulation of MnSOD was identified that was elicited by treatment of human and mouse cells with resveratrol (1) (see Fig. 1) (Robb et al., 2008a; Robb and Stuart, 2011), a compound found in red wines, mulberries and peanuts, that is widely associated with inhibitory effects on the growth of both cancerous and normal cell lines. This observation has now been made for a collection of diverse cell types that includes cardiomyocytes, neuroblastomas (SK-N-BE), a hippocampal
165
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
neuronal cell line (HT22), coronary arterial endothelial cells, and pheochromocytoma cells (PC6.3) (Movahed et al., 2012; Albani et al., 2009; Fukui et al., 2010; Ungvari et al., 2009; Kairisalo et al., 2011). Using siRNA, it was demonstrated that the growth inhibitory effect of resveratrol (1) is dependent upon MnSOD induction (Robb and Stuart, 2011). Similarly, Fukui et al. (2010) showed that an induction of MnSOD is necessary for the cytoprotective effects of resveratrol (1) in cultured neuronal cells. Its ability to stimulate MnSOD expressionis not exclusive to cells in culture. MnSOD upregulation in whole brain tissue, skeletal muscle, hematopoietic stem cell populations, and dopaminergic neurons can be achieved in vivo via dietary supplementation (Robb et al., 2008b; Jackson et al., 2011; Zhang et al., 2013; Mudò et al., 2011). Thus, MnSOD is a target of resveratrol (1) both in vitro and in vivo that has been linked to important biological outcomes. Resveratrol (1) belongs to the stilbene family of phytoalexins, which are produced by grapevine tissues (Vitis vinifera) in response to a variety of biotic and abiotic stresses. In the biosynthetic pathway, resveratrol is an intermediary, which undergoes subsequent modification to pterostilbene (2) and piceid (3) (Fig. 1). Recent studies have shown that pterostilbene (2) is at least as effective as resveratrol at inhibiting cancer cell growth in vitro and slowing tumour growth in vivo (Lin et al., 2012; Moon et al., 2013; McCormack et al., 2012). A limited amount of available data suggests that piceid (3) may exert similar effects (Su et al., 2013). Thus, both the structural similarities and the observed parallels in their biological activities in mammals suggest that resveratrol (1), pterostilbene (2), and piceid (3) might all work via the same basic mechanism. In a similar vein, two of the major metabolites of resveratrol (1) in vivo are resveratrol-40 -Sulfate (RES-S) (4) andresveratrol-3O-b-D-glucuronide (RES-G) (5), which both retain basic structural
similarity to resveratrol (1) (Fig. 1). The effects of these two metabolites on cell proliferation have not been well explored. Among the earliest biological activities attributed to resveratrol (1) was its estrogenic effect in mammalian cells. It is a phytoestrogen capable of binding to estrogen receptor alpha and beta (Gehm et al., 1997; Bowers et al., 2000; Salah et al., 2013). Its effect on MnSOD expression is inhibited by the estrogen receptor (ER) antagonist ICI 182,780, and can be phenocopied using the specific ERbeta agonist diarylpropionitrile (DPN), but not the ERalpha agonist propyl pyrazoletriol (PPT), suggesting the importance of specifically ERbeta in resveratrol’s cellular effects (Robb and Stuart, 2011). ERbeta agonists are generally associated with anti-proliferative effects (Warner and Gustafsson, 2010). A relatively wide range of structurally related molecules is predicted to selectively bind ERbeta, suggesting that multiple small molecules could target this estrogen receptor in vivo. Here a variety of molecular and cellular approaches were used to characterize the effects of resveratrol (1) derived stilbeneson cell growth and cytoprotection. siRNA MnSOD knockdown and MnSOD-null cells are used to demonstrate the requirement of this enzyme for the growth inhibitory effects of resveratrol (1), pterostilbene (2), and piceid (3). The only known activity of MnSOD is the dismutation of superoxide in the mitochondrial matrix (derived primarily from respiration) to hydrogen peroxide. Therefore, respiratory deficient rho0 PC3 prostate cancer cells were utilized to show that respiratory superoxide production is necessary for the observed effect on growth (Hoffmann et al., 2004; Chandel et al., 1998). Finally, using ERbeta-null cells the involvement of ERbeta in the induction of MnSOD expression and subsequent effects on cell growth and cytoprotection elicited by resveratrol (1) derived stilbenes were demonstrated. Together, these experiments reveal
H3C
O
O
CH3
HO
OH
pterostilbene (2)
OH
HO OH trans-resveratrol (1)
O HO O
HO
OH
piceid (3)
HO OH
O
OH
OH
O HO HO
O
O
HO OH HO OH
resveratrol-3-O-β-D-glucuronide (5)
OH
S O
O
trans-resveratrol-4’-sulfate (4)
Fig. 1. Molecules included in this study: trans-resveratrol (1), pterostilbene (2), piceid (3), trans-resveratrol-40 -sulfate (RES-S) (4), resveratrol-3-O-b-D-glucuronide (RES-G) (5), and their relationship to trans-resveratrol.
166
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
a pathway in which ERbeta agonists including resveratrol (1), pterostilbene (2), and piceid (3) stimulate the expression of MnSOD, which via the modulation of mitochondrial ROS metabolism regulates cell growth and cytoprotection. We suggest that this shared mechanism of action underlies the observation that all three molecules are capable of eliciting similar effects on mammalian cells.
resveratrol (1) (Fig. 2A and B). In contrast, however, neither RESS (4) nor RES-G (5) affected MnSOD expression. In the same experiment, both pterostilbene (2) (20 lM) and piceid (3) (50 lM) significantly inhibited the proliferative growth of C2C12 cells (Fig. 2C), whereas again neither RES-S (4) nor RES-G (5) had any effect at the concentrations tested (up to 50 lM). Similar results were observed in cultured human lung fibroblasts (MRC5; data not shown). Since resveratrol (1) has cytoprotective properties and pretreatment of C2C12 cells with resveratrol (1) elicits a cytoprotective phenotype (e.g. Robb and Stuart, 2011), it can be predicted that pterostilbene (2) and piceid (3) would be similarly cytoprotective. All of these compounds are chemical antioxidants, and so were removed immediately prior to the experiment by washing and refreshing cells with new media 1 h before exposure to hydrogen peroxide (30–160 lM) for 2 h. Treatment with resveratrol (1), pterostilbene (2), or piceid (3) for 48 h prior to hydrogen peroxide exposure significantly increased the LD50 in C2C12 cells (Fig. 2D). The cytoprotective properties of these molecules appeared not to be mediated by broad increases in antioxidant enzymes generally, since virtually no induction of CuZnSOD, catalase, glutathione peroxidase was observed (Table 1). Similarly, the increased stress resistance elicited by these compounds was not associated with increased heat shock protein 60, or heat shock protein 70 protein levels (Table 1).
Results Effects of resveratrol analogues on C2C12 myoblasts
MnSOD’s anti-proliferative and cytoprotective properties have been well documented. To isolate the putative role of stilbene-
MnSOD protein level/Cs activity
MW
MnSOD
Memcode protein stain
DMSO EtOH Resveratrol (1) RES-S (4) RES-G (5) Piceid (2) Pterostilbene (3)
0.15 0.12
* * *
0.09 0.06 0.03 0
D
C 25 20 15 10 5 0
200 *
*
*
Hydrogen Peroxide LD50 iin uM
30 Population Doubling Time (h)
Growth inhibition and cytoprotection in the absence of MnSODor mitochondrial ROS
B
Piceid (3)
Pterostilbene (2)
Resveratrol (1)
RES-S (4)
EtOH
DMSO
A
RES-G (5)
Induction of MnSOD expression appears to be a critical event underlying two key biological activities of resveratrol (1): inhibition of cell growth and stimulation of cellular stress resistance (Robb and Stuart, 2011). It can therefore, be hypothesized that pterostilbene (2) and piceid (3), as resveratrol (1) derivatives with a high degree of structural similarity (Fig. 1), would work via the same mechanism. In addition, trans-resveratrol-40 -sulfate (RES-S) (4) and resveratrol-3-O-b-D-glucuronide (RES-G) (5) (Fig. 1) are two by products of resveratrol (1) metabolism in humans that are detectable at high levels in plasma and tissue of individuals consuming resveratrol (1) (Wenzel and Somoza, 2005). Since these metabolites also retain some structural similarity and might also be biologically active, they were included in this study. C2C12 myoblasts were treated for 48 h with resveratrol (1) (range of concentrations up to 50 lM), piceid (3) (up to 50 lM), pterostilbene (2) (up to 20 lM), RES-S (4) (up to 50 lM), or RES-G (5) (up to 50 lM). Cells were then harvested to determine the effects of these compounds on MnSOD expression. As predicted pterostilbene (2) and piceid (3) elicited a similar induction of MnSOD as did
* 160
*
*
120 80 40 0
Fig. 2. Piceid and pterostilbene, but not RES-S or RES-G, increase population doubling time, stress resistance, and MnSOD expression in C2C12 myoblasts. (A) Representative Western blot and corresponding protein stain of MnSOD in C212 myoblasts. (B) Relative MnSOD protein levels in C2C12 myoblasts treated with DMSO, ethanol, resveratrol (1) (25 lM), RES-S (4) (50 lM), RES-G (5) (50 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). Since phytoestrogens including resveratrol can stimulate a general increase in mitochondrial abundance in many cell types, to accurately evaluate changes in MnSOD levels citrate synthase activity was used as a proxy of mitochondrial abundance. (C) Average population doubling time in C2C12 myoblasts treated with DMSO, ethanol, resveratrol (1) (25 lM) RES-S (50 lM), RES-G (50 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (D) Hydrogen peroxide LD50 of C2C12 myoblasts treated with DMSO, ethanol, resveratrol (1) (25 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). Data shown represent the means of 3–5 independent trials. Error bars represent SEM. ⁄p < 0.05 compared to vehicle control.
167
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
Table 1 Effects of resveratrol analogues on antioxidant enzymes and heat shock proteins. catalase activity, glutathione peroxidase activity and CuZnSOD, Hsp60 and Hsp70 protein level in C2C12 myoblasts. Treatment
Catalase activity (mmol/min/mg cellular protein)
Glutathione peroxidase activity (mmol/min/mg cellular protein
CuZn superoxide dismutase protein level relative to internal control
Hsp60 protein level/Cs activity relative to internal control
Hsp70 protein level relative to internal control
DMSO EtOH Resveratrol (1) Pterostilbene (2) Piceid (3) RES-S (4) RES-G (5)
15.51 ± 0.24 15.80 ± 0.47 15.42 ± 0.07 16.32 ± 0.34 19.18 ± 0.50* N.D. N.D.
67.32 ± 5.27 66.52 ± 8.14 51.41 ± 5.04 68.41 ± 7.58 51.22 ± 3.07* N.D. N.D.
0.96 ± 0.12 0.99 ± 0.08 1.25 ± 0.43 1.20 ± 0.16 1.26 ± 0.19 1.06 ± 0.17 1.11 ± 0.13
0.089 ± 0.020 0.097 ± 0.018 0.102 ± 0.010 0.095 ± 0.014 0.106 ± 0.018 0.100 ± 0.015 0.067 ± 0.011
1.05 ± 0.22 0.91 ± 0.36 1.12 ± 0.18 1.21 ± 0.09 1.34 ± 0.17 1.25 ± 0.13 1.27 ± 0.15
Data represents the mean of duplicate measurements for 3 independent trials. N.D. = not determined. p < 0.05.
*
mediated MnSOD induction in cell growth inhibition and cytoprotection, siRNA against MnSOD was used to prevent its induction during 48 h treatment with resveratrol (1), piceid (3), or pterostilbene (2) (Fig. 3A). Using this approach, none of the stilbenes had any effect on proliferative growth (Fig. 3B) or stress resistance (Fig. 3C) in C2C12 myoblasts. Similar effects were observed in MRC5 fibroblasts (data not shown). To gain additional insight into the role of MnSOD in the effects of stilbenes in mammalian cells, these experiments were repeated (but without siRNA) using mouse embryonic fibroblast lines (MEFs) derived from MnSODnull and wildtype mice. In agreement with our siRNA experiments, the growth inhibitory effects of resveratrol (1), pterostilbene (2) and piceid (3) were absent in MnSOD-null MEFs (Fig. 3D). Stress resistance experiments were not conducted in MnSOD-null MEFs. The only known activity reported for MnSOD is the dismutation of superoxide to hydrogen peroxide in mitochondria. Within the mitochondrial matrix, superoxide is produced by the single electron reduction of molecular oxygen during respiration. To determine whether the growth inhibitory effects of stilbenes were dependent upon mitochondrial ROS, respiration deficient cells were created. PC3 prostate cancer cells were used for this experiment, since they are highly responsive to stilbenes and an existing protocol is available for rendering the cells mtDNA-free (rho0; see Experimental section). Using normal and rho0 cells, we investigated whether: (1) the stilbene-mediated induction of MnSOD would occur even in the absence of respiration, and (2) if MnSOD was induced, would its effect on cell growth still be evident. It was found that the induction of MnSOD activity by resveratrol (1), pterostilbene (2) or piceid (3) was virtually identical in normal and rho0 cells (Fig. 4A), indicating that this was independent of respiratory superoxide production. While the induction of MnSOD protein levels was paralleled by increases in MnSOD activity in the normal PC3 cells, in rho0 PC3 cells MnSOD protein levels, but not activities, were increased (Fig. 4C,D). Also, the induction of MnSOD expression in rho0 cells was not accompanied by effects on cell growth (Fig. 4B). Estrogen receptor beta is required for the effects of resveratrol, pterostilbene and piceid Resveratrol (1) is an ER agonist (Gehm et al., 1997; Bowers et al., 2000). Its effects on MnSOD expression, cell proliferation and stress resistance are abolished by an ER antagonist (ICI 182, 780) and can be reproduced using the specific ERbeta agonist DPN (Robb and Stuart, 2011). To determine if the observed effects of stilbenes were mediated through ERbeta two approaches were used. Firstly, the experiments were repeated in the presence of the ER antagonist ICI182,780, and secondly they were repeated with myoblasts cell lines established from ERbeta-null mice (see Methods).
ICI182,780 prevented the effects of all three molecules on MnSOD, cell growth, and stress resistance (Table 2). Similarly, the absence of ERbeta diminished the effects of resveratrol (1), piceid (3) and pterostilbene (2) on MnSOD protein levels (Fig 5A), population doubling time (Fig. 5B) and hydrogen peroxide LD50 (Fig. 5C). Together, these results suggest a role for ERbeta in the effects of stilbenes on these cellular activities. Discussion Resveratrol (1) is a naturally occurring stilbene phytoalexin produced by V. vinifera in response to stress that has been much studied for its cytoprotective and growth inhibitory effects. Surprisingly, the structurally related molecules pterostilbene (2) and piceid (3), produced from resveratrol (1) in V. vinifera under the same conditions by the same biosynthetic pathway and which are also present in grapes skins and red wines, have received far less attention. Quite recently, the growth inhibitory effects of pterostilbene (2) in vitro (Lin et al., 2012; Moon et al., 2012) and in vivo (McCormack et al., 2012) have been reported, and a limited amount of data describing a similar effect is available for piceid (Su et al., 2013). Comparable to resveratrol (1), pterostilbene (2) induces a strong upregulation of MnSOD that is concurrent with reduced proliferative growth in breast and pancreatic cancer cells (Moon et al., 2012; McCormack et al., 2012). Here it is shown that all three compounds, resveratrol (1), pterostilbene (2) and piceid (3), inhibit the proliferative growth of mammalian cells via a mechanism involving the ERbeta-mediated induction of MnSOD in mitochondria. MnSOD is a cell cycle regulator in mammalian cells, via poorly understood redox modifications of key regulatory proteins that are not yet fully elucidated, though they continue to be actively studied (Kim et al., 2010; Sarsour et al., 2008; Menon et al., 2007). In many cells, reduced MnSOD expression is associated with increased replication rates, which can be slowed by restoration of MnSOD to normal levels (Li et al., 1995; Liu et al., 1997; Yan et al., 1996; Behrend et al., 2005). The inhibitory effects of resveratrol, pterostilbene (2), and piceid (3) on cell growth measured in this study are dependent upon their ability to upregulate MnSOD expression and activity. This was demonstrated using two complementary approaches: (1) preventing the induction of MnSOD expression using siRNA, and (2) eliminating MnSOD by using MnSOD-null MEFs. None of the three compounds inhibited proliferative growth in the absence of the MnSOD induction. MnSOD catalyzes the dismutation of mitochondrial matrix superoxide to hydrogen peroxide, and in so doing may affect the concentrations of several ROS and reactive nitrogen species (RNS) including superoxide, hydrogen peroxide (Liochev and Fridovich, 2007), nitric oxide, and peroxynitrite as a result of secondary reac-
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173 Control siRNA
Piceid (3)
Pterostilbene (2)
DMSO
*
Control siRNA
2
DMSO EtOH Piceid (3) Pterostilbene (2)
*
EtOH
Relative MnSOD protein level
2.5
DMSO
A
Piceid (3)
Pterostilbene (2)
MnSOD siRNA
EtOH
168
MnSOD
1.5 1 0.5
Ponceau S protein stain
0 Control siRNA
MnSOD siRNA
C Hydrogen Peroxide LD50 (uM)
B Population Doubling Time (h)
35 30
* *
25 20 15 10 5 0
200 * 160 *
DMSO EtOH Resveratrol (1) Piceid (3) Pterostilbene (2)
120
80
40
0 Control siRNA
MnSOD siRNA
Control siRNA
MnSOD siRNA
MEF Population Doubling Time (h)
D 80
*
70 *
*
60 50 40 30 20 10 0 Wild Type MEF
MnSOD -/- MEF
Fig. 3. MnSOD is essential for piceid and pterostilbene to increase population doubling time and stress resistance in C2C12 myoblasts. (A) Relative MnSOD protein level, representative western blot and corresponding protein stain in C2C12 myoblasts treated with control siRNA or MnSOD siRNA ± DMSO, ethanol, piceid (3) (50 lM) prpterostilbene (2) (20 lM in ethanol). (B) Average population doubling time in C2C12 myoblasts treated with control siRNA or MnSOD siRNA ± DMSO, ethanol, piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (C) Hydrogen Peroxide LD50 of C2C12 myoblasts treated with control siRNA or MnSOD siRNA ± DMSO, ethanol, piceid (50 lM) or pterostilbene (20 lM in ethanol). (D) Average population doubling time in wildtype and MnSOD null mouse embryonic fibroblasts treated with DMSO, ethanol, resveratrol (15 lM), piceid (25 lM) or pterostilbene (10 lM in ethanol). Data shown represents the mean of 3 independent trials. Error bars represent SEM. ⁄p < 0.05 compared to vehicle control.
tions (Keller et al., 1998). An important observation made in this study was that the ability of MnSOD to inhibit growth was present at physiological oxygen levels (3%, necessary for the culture of primary myoblasts generated from wildtype and ERbeta null mice) and was not therefore an artefact of high oxygen levels in culture (typically 18–20%). It is possible that MnSOD’s effect on cell growth could be related to an as yet uncharacterized activity of the protein beyond ROS metabolism. To verify that the observed effects of MnSOD on proliferative cell growth were indeed mediated by matrix superoxide metabolising activity we used respiration-deficient rho0 PC3 cells. Rho0 cells have been productively used to study a variety of phenomena that are modulated by mitochondrial ROS metabolism (Schauen et al., 2006; Chandel and Schumacker, 1999; Zhang and Gutterman, 2007). PC3 cells are an appropriate cell type for this investigation since they express high levels of
ERbeta (McPherson et al., 2010) and the growth of wildtype PC3 cells is strongly inhibited by ERbeta agonists, including DPN and resveratrol (1). It was found that, while the stilbenes and DPN all strongly upregulated MnSOD protein levels in rho0 PC3 cells despite the absence of respiration, there was no concomitant inhibition of growth. This indicates that mitochondrial respiration is not necessary for the ERbeta mediated induction of MnSOD. This observation is important because resveratrol (1) has been shown to inhibit mitochondrial respiration (McPherson et al., 2010; Zini et al., 1999) and some respiratory inhibitors promote superoxide production (reviewed by Lambert and Brand, 2009). The fact that MnSOD protein levels were almost identically upregulated in the absence of respiration thus indicates a more direct mechanism of upregulating MnSOD expression. Interestingly, although MnSOD protein levels were significantly higher in the rho0 cells there
169
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
A
4.0 * 3.0
*
*
3.5 *
*
*
2.5 2.0 1.5 1.0 0.5 0.0 rho0 PC3
Wildtype PC3
*
70
*
60
*
50 40 30 20 10 0 rho0 PC3
MnSOD Activity/CS Activity
0.10
DMSO
Resveratrol (1)
Pterostilbene (2)
Piceid (3)
EtOH
DMSO
Resveratrol (1)
Piceid (3)
80
Wildtype PC3
D
rho0 PC3
Wildtype PC3
EtOH
C Pterostilbene (2)
DMSO EtOH Resveratrol (1) Piceid (3) Pterostilbene (2)
B
4.5
PC3 Population Doubling Time (h)
MnSOD protein level/CS Activity
5.0
*
0.09 0.08
*
*
0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00
Wildtype PC3
rho0 PC3
Fig. 4. Resveratrol, piceid and pterostilbene require mitochondrial ROS production to affect cell growth. (A) Average MnSOD protein levels normalized to citrate synthase activity in wildtype and rho0 PC3 cells treated with DMSO, ethanol, resveratrol (1) (25 lM), RES-S (4) (50 lM), RES-G (5) (50 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (B) Average population doubling time in wildtype and rho0 PC3 cells treated with DMSO, ethanol, resveratrol (1) (25 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (C) Representative MnSOD activity gel in wildtype and rho0 PC3 cells. (D) MnSOD activity normalized to citrate synthase activity in wildtype and rho0 PC3 cells treated with DMSO, ethanol, resveratrol (1) (25 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). ⁄p < 0.05 compared to vehicle control of the same genotype.
Table 2 Effects of resveratrol analogues on population doubling time, resistance to hydrogen peroxide induced cell death, and fold change in MnSOD protein level in the absence and presence of ICI182,780 in C2C12 myoblasts. Treatment
DMSO EtOH Resveratrol (1) Pterostilbene (2) Piceid (3)
Population Doubling Time (h)
H2O2 LD50 (lM)
No antagonist
Antagonist
No antagonist
Antagonist
No antagonist
Antagonist
14.98 ± 1.16 15.01 ± 0.94 26.34 ± 1.25* 33.61 ± 1.08* 27.14 ± 1.11*
15.29 ± 0.11 14.87 ± 0.16 17.23 ± 1.22 22.19 ± 2.17$ 21.24 ± 1.05#
92 ± 11 84 ± 9 144 ± 19* 166 ± 18* 149 ± 16*
82 ± 7 93 ± 6 96 ± 11 110 ± 9 89 ± 17
1.06 ± 0.15 1.10 ± 0.18 2.10 ± 0.24* 2.22 ± 0.29* 1.84 ± 0.17*
0.94 ± 0.12 1.04 ± 0.08 0.82 ± 0.16 1.19 ± 0.18 0.97 ± 0.10
MnSOD protein level relative to internal control
Antagonist = ICI182,780. Data represents the mean of duplicate measurements for 3 independent trials. * p < 0.05 when compared to vehicle control. # p = 0.056. $ p = 0.052.
was no concomitant increase in MnSOD activity. MnSOD activity is modulated by a variety of post-translational modifications including acetylation (Ozden et al., 2011) and methylation (Sarsour et al., 2012). It is, therefore, hypothesized that these signalling pathways may be disrupted in rho0 cells, but at this time the mechanism underlying this observation were not known. Importantly, there was also no effect of MnSOD induction on cell growth in the absence of respiration. Thus, it seems that the ability of stilbenes to modulate cellular activities via MnSOD rely upon an active respiratory chain, though further work is required to fully characterize this. Endogenous estrogens like 17beta-estradiol exert a variety of effects on mitochondrial function, including an upregulation of MnSOD in vitro and in vivo. In cultured vascular smooth muscle cells estrogen treatment significantly increases MnSOD activity, and this is accompanied by a reduction in proliferative cell growth
(Sivritas et al., 2011). In rats, estrogen treatment increases MnSOD levels in mitochondria isolated from brain tissue (Razmara et al., 2007), and a downregulation of MnSOD is observed in vascular tissue of ovariectomized mice (Strehlow et al., 2003). The involvement of ERbeta signalling in the pathway upstream of MnSOD is supported here by the ability of the ERbeta-specific agonist DPN to reproduce the phenotype, and the absence of effects of resveratrol (1) and its derivatives in cells lacking ERbeta. Here data is shown for myoblasts, but virtually identical results using ERbetanull fibroblasts derived from the same mice were found (not shown). ERbeta is predicted to bind quite a wide range of dietary phytoestrogens including resveratrol (1) (Yuan et al., 2011), which has been demonstrated in vitro (Bowers et al., 2000; Zhu et al., 2006). In agreement with the anti-proliferative effects of resveratrol (1) and its derivatives, activation of ERbeta using pharmacological agonists is generally associated with cell cycle
170
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
B 40 0.25 0.20
*
Population Doubling Time (h)
Myoblast MnSOD protein level/ Cs activity
A
*
* *
0.15 0.10 0.05 0
ERBeta +/+
30
*
*
*
25 20 15 10 5 0
ERbeta -/-
*
35
DMSO EtOH Resveratrol (1) Piceid (3) Pterostilbene (2)
ERBeta+/+
ERbeta-/-
Hydrogen peroxide LD50 (µM)
C 125 100
*
*
* *
75 50 25 0 ERBeta+/+
ERbeta-/-
Fig. 5. Involvement of ERbeta in resveratrol, piceid and pterostilbene effects on population doubling time, stress resistance and MnSOD protein level. Relative MnSOD protein levels normalized to citrate synthase activity over control group in wildtype and ERbeta null myoblasts treated with DMSO, DPN (10 lM), resveratrol (1) (25 lM), piceid (3) (50 lM) and pterostilbene (2) (10 lM). (B) Average population doubling time in wildtype and ERbeta null mouse myoblasts treated with DMSO, DPN (10 lM), resveratrol (1) (25 lM), piceid (3) (50 lM) and pterostilbene (2) (10 lM). (C) Hydrogen Peroxide LD50 over vehicle control in wildtype and ERbeta null myoblasts treated with DMSO, DPN (10 lM), resveratrol (1) (25 lM), piceid (3) (50 lM) and pterostilbene (2) (10 lM). Data shown represents the mean of duplicate measures from cell lines established from 5 individuals of each genotype. Error bars represent SEM. ⁄p < 0.05 compared to vehicle control of the same genotype.
inhibition and anti-proliferative effects (reviewed in Warner and Gustafsson, 2010; Nilsson et al., 2011). The ERbeta agonist diarylpropionitrile (DPN) slows growth of the colon cancer cell line MC38 (Motylewska et al., 2009), and lentivirus transfection of ERbeta into a cancerous colon cell line significantly reduces cell proliferation (Hartman et al., 2009). It is concluded that ERbeta is mediating the effects of resveratrol (1), pterostilbene (2), and piceid (3) on MnSOD, which in turn affects a reduction in growth rate and enhanced cellular stress resistance. Indeed it is noted that these are two phenotypes which have been described for ERbeta agonists generally (Nilsson et al., 2011). Two of the most abundant metabolites of resveratrol (1) in vivo, RES-S (4) and RES-G (5) (Wenzel and Somoza, 2005), had no significant effect on MnSOD expression, proliferative growth, or stress resistance, suggesting that these common metabolites of resveratrol in vivo do not contribute to the inhibition of cell growth. In agreement with these results, the proliferative rate of the cancerous mammary cell line MCF7 has also been shown to be unaffected by the 30 -sulfate and 4-sulfate resveratrol metabolites (Miksits et al., 2009). However, sulfate metabolites of resveratrol (1) do apparently maintain other properties, and were shown to share the ability of resveratrol (1) to inhibit inducible nitric oxide synthase activity in cultured macrophage cells (Hoshino et al., 2010).
Conclusions In conclusion, evidence is provided that the ability of resveratrol (1), pterostilbene (2), and piceid (3) to slow cell growth, which has been observed in multiple cell types, is dependent upon the induction of MnSOD expression in the context of normal mitochondrial respiration. This induction of MnSOD is not caused
by a resveratrol-mediated augmentation of mitochondrial ROS production, since it occurs in the absence of respiration. Resveratrol (1) is a demonstrated low-affinity ERbeta agonist, and the data herein indicate that the induction of MnSOD requires the presence of ERbeta. The effects can be phenocopied by the highly specific ERbeta agonist DPN, abolished using an ER agonist (ICI 182,780), and are absent in ERbeta-null cells. Importantly, the effects we and others have observed with resveratrol (1) and pterostilbene (2) in cultured cells can be achieved in vivo via dietary supplementation, which suggests that both molecules, and perhaps also piceid (3), would be capable of modulating MnSOD expression in a variety of human health contexts that might benefit from this.
Experimental General experimental procedures All estrogens and phytoestrogens used in this study were purchased from commercial sources and were of the highest purity available. trans-Resveratrol (1) was obtained from A.G. Scientific (San Diego, CA). trans-Piceid (3) was purchased from Sigma (St. Louis, MO). Pterostilbene (2) was obtained from Cayman Chemical (Ann Arbor, MI). trans-Resveratrol-40 -Sulfate (RES-S) (4) sodium salt and trans-Resveratrol-3-O-b-D-glucuronide (RES-G) (5) were purchased from Toronto Research Chemicals (Toronto, Canada). ICI 182780 and DPN were purchased from Tocris Biosciences (Ellisville, MO). All polyphenols purchased were of greater than 95% purity. Ethidium bromide, pyruvate and uridine were purchased from Bioshop (Burlington, Canada). Experiments with cultured cells were performed on cells grown at 37 °C and 5% CO2, and either 3% O2 (for primary cell lines:
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
MnSOD null, ERbeta null) or 18% O2 (for all other cell lines). The established C2C12 mouse myoblast and PC3 human prostate cancer cell lines were cultured in accordance with the distributor’s protocol (American Type Culture Collection; Manassas, VA) and subcultured as required in Dulbecco’s modified Eagle medium with high glucose (C2C12) or minimum essential medium (PC3) with 10% (v/v) FBS, 1% (v/v) penicillin streptomycin solution, and 2% (v/v) non-essential amino acids (Hyclone; Logan, UT). MnSOD null and wild-type mouse embryonic fibroblasts were generously provided by Dr. Prabhat Goswami of the University of Iowa. MnSOD null and wild type mouse embryonic fibroblasts were cultured at 37 °C, 5% CO2, 3% O2in DMEM high glucose supplemented with 10% (v/v) FBS, 1% (v/v) penicillin streptomycin solution, and 2% (v/v) non essential amino acids. ERbeta null and control myoblast cell lines were generated from 5 individuals each of wildtype and ERbeta/ mice on a C57/BL6 background (Taconic Farms, Hudson, NY, USA) (Krege et al., 1998). Mice were euthanized using isofluorane and rapid cervical dislocation following a protocol approved by the animal care committee at Brock University, and in accordance with the guidelines of the Canadian Council on Animal Care. Briefly, to isolate primary myoblasts hind limb skeletal muscle was excised, finely minced and suspended in pronase solution (2 mg/mL) at 37 °C for 1 h. The suspension was titurated by repeated pipetting and then filtered through sterile cheesecloth and centrifuged at 240 g for 5 min. The resulting supernatant was then centrifuged at 500 g for 5 min. The pellet was resuspended in growth media (Ham’s F10 nutrient mixture, 20% FBS, nonessential amino acids, penicillin, streptomycin, gentamicin, amphotericin B, basic human growth factor), plated on to collagen coated tissue culture dishes and incubated at 37 °C, 3%O2 and 5%CO2. RPMI-1640 Gentamicin, amphotericin B, collagen, and pronase were obtained from Sigma–Aldrich (St. Louis, MO). Basic human growth factor was purchased from Invitrogen (Burlington, Canada). An additional equal volume of media was added 48 h after plating, and the full volume of media was exchanged 96 h after plating. Myoblast outgrowth was evident at 96 h. Myoblasts were subcultured at a ratio of 1:3 when they reached 70% confluence. Rho0 PC3 prostate cancer cells were generated as outlined in Joshi et al. (1999). Briefly, PC3 cells were incubated in DMEM high glucose containing 10% (v/v) FBS, 1% (v/v) penicillin streptomycin solution, 2% (v/v) non-essential amino acids, and supplemented with ethidium bromide (200 ng/mL), pyruvate (100 lg/mL), uridine (50 lg/mL) for eight weeks. Culture media was refreshed daily. Ethidium bromide, pyruvate and uridine were added directly to the media each day. The ethidium bromide treatment was stopped after eight weeks. Rho0 status was verified by the absence of oxygen consumption and TFAM signal (not shown). Rho0 PC3 cells were stored as frozen stocks. For subsequent experiments, frozen cell stocks were thawed and propagated for several passages at 37 °C, 5% CO2, 18% O2 then treated as described. Chemicals used in enzyme assays were purchased from Sigma– Aldrich (St. Louis, MO) or Bioshop (Burlington, ON, Canada) unless otherwise indicated. Resveratrol analogue and estrogen antagonist treatments Each compound (resveratrol (1), pterostilbene (2), piceid (3), RES-S (4), RES-G (5)) was tested at 1, 10, 20 or 25, and 50 lM to evaluate possible toxicity and to determine the concentration required for maximal effects on population doubling time (data not shown). The lowest concentration giving a maximal effect was selected for further experiments. All compounds were dissolved in DMSO, with the exception of pterostilbene (2), which was dissolved in EtOH–H2O (95:5, v/v). In all cases, media was refreshed daily with freshly prepared compounds. All treatments
171
were of 48 h duration. Comparisons were made to the appropriate vehicle control for each compound and the data presented for each compound correspond to the concentration that yielded maximal effects on population doubling time. Where the estrogen receptor antagonist ICI 182,780 was used, it was added directly to the culture media at a final concentration of 10 lM 24 h prior to polyphenol treatment. This concentration was maintained throughout the 48 h incubation period and was also refreshed daily. siRNA treatment siRNA to MnSOD, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), a scrambled control sequence, and NeoFx Transfection Agent™ were purchased from Ambion (Austin, TX). C2C12 myoblasts were plated with NeoFx Transfection Agent™ and 2–10 nM of commercially available siRNA to MnSOD, GAPDH or a scrambled control sequence at a density of 1.0 106 cells/mL. siRNA containing medium was replaced 24 h after transfection. Phytoestrogen treatments of the knock-down cells were performed 6–72 h following transfection. Knock down efficiency was determined via western blot. Stress resistance and death experiments To evaluate stress resistance, myoblasts were washed three times with PBS and incubated with serum free medium containing 2% BSA with or without hydrogen peroxide (30–160 lM; 2 h). Following the incubation time, the media was removed, cells were rinsed with PBS and then incubated with serum containing medium for 18 h. Cell number and viability were calculated relative to the appropriate vehicle control at the indicated time points following the stressor exposure via trypan blue exclusion and cell counting. Samples of culture medium were taken to measure the release of lactate dehydrogenase (LDH) into the medium from dead cells. Lactate dehydrogenase activity LDH activity in the culture medium was measured in a solution containing 20 mM HEPES buffer (pH 7.3), 0.2 mM NADH and culture medium (200 lL). LDH activity was followed spectrophotometrically by the rate conversion of NADH to NAD+ at 340 nm. In RNAi experiments, LDH release was measured using the Cytotoxicity Detection Kit™ (Roche Applied Science, Laval, Canada) per the manufacturer’s instructions, as this kit is more accurate at the very low volumes used in these experiments. Preparation of whole cell lysates Cells for western blot or activity assay measurements were lysed by incubation for one hour in ice cold lysis buffer (10 mM Tris pH 8.0, 150 mM NaCl, 2 mM EDTA, 2 mM dithiothreitol, 0.4 mM PMSF, 40% (v/v) glycerol, 0.5% (v/v) NP40) with periodic sonication (Ultrasonic Inc., Sonicator W-375; setting 3). After incubation, cells were centrifuged at 10,000 g at 4 °C for 10 min (Thermo Scientific, IEC Micromax/Micromax RF). The protein concentration of the resulting supernatant was determined by the Bradford method using a BioRad protein determination kit (Hercules, CA). Whole cell lysates were stored in aliquots at 80 °C. Quantitative Western blotting Equal amounts of whole cell lysate (15 lg) were separated by SDS–PAGE (5% stacking, 12% resolving gels) and electroblotted onto a polyvinylidene fluoride membrane. Prestained broad range protein marker (Frogga Bioscience, Toronto, Canada) was included
172
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
on each gel. Memcode Reversible Protein Stain Kit™ (Thermo Fisher Scientific, Mississauga, Canada), or Ponceau S staining was used in accordance with the manufacturer’s instructions to verify equal protein loading/transfer. Following blocking, membranes were incubated with an antibody to MnSOD (1:5000 dilution; Enzo Life Sciences, Brockville, Canada), CuZnSOD (1:1000 dilution; Stressgen, Victoria, Canada), hsp70 (1:1000 dilution; Abcam, Cambridge, MA) or hsp60 (1:2000 dilution; Stressgen, Victoria, Canada). The membranes were visualized using the Odyssey infrared imaging system from LI-COR Biosciences, with IR-linked secondary antibodies to rabbit or mouse (1:5000 dilution; Rockland Immunochemicals, Gilbertsville, PA). Western blot quantitative analyses were performed using Odyssey imaging software 1.0. ERbeta null and control mice were purchased from Taconic Farms (Germantown, NY). Bovine superoxide dismutase was purchased from Sigma (St. Louis, MO). The C2C12 and PC3 cell lines were purchased from the American Type Culture Collection. All other chemicals and purified enzymes were obtained either from Sigma–Aldrich (St. Louis, MO), BioShop (Burlington, Canada) or Fisher Scientific (Mississauga, Canada) unless otherwise stated. Citrate synthase activity Citrate synthase (CS) activity was measured in a buffer containing 50 mM Tris pH 8.0, 0.5 mM 5,50 -dithiobis(2-nitrobenzoic acid), 0.1 mM acetyl-coenzyme A, 0.05% Triton X-100 and 5 lg protein. The reaction was initiated by the addition of 0.5 mM oxaloacetate, and absorbance was followed at 412 nm using a Bio-Tek KC4 plate reader. Chemicals used in the assays were purchased from Sigma– Aldrich (St. Louis, MO) or Bioshop (Burlington, ON). Antioxidant enzyme activities Catalase and glutathione peroxidase activity was measured as outlined in Robb et al. (2008b). Briefly, for catalase activity, the disappearance of hydrogen peroxide following the addition of whole cell lysate (50 lg) was monitored at 240 nm in a solution of 25 mM phosphate buffer pH 7.0 at 30 °C. Glutathione peroxidase activity was monitored by following the consumption of NADPH at 340 nm in a solution of 50 mM phosphate buffer pH 7.0, 0.4 mM EDTA, 0.15 mM NADPH, 1 mM reduced glutathione and 1 unit glutathione reductase. The reaction was initiated following the addition of 0.0007% (v/v) hydrogen peroxide. MnSOD activity was measured using an in-gel assay as described in Robb et al. (2008a,b). Whole cell lysate protein (100– 120 lg) was resolved by native-PAGE at 20 mA at 4 °C for 2 h. Gels were then incubated in 1.23 mM nitroblue tetrazolium for 15 minutes at room temperature. Gels were rinsed with deionised H2O and were transferred to a solution of 28 mM TEMED and 46 lM riboflavin in potassium phosphate buffer (pH 7.8) for 15 min at room temperature and protected from light. Gels were rinsed with deionised H2O and placed under fluorescent light until the background achieved a uniform violet colour. 5 mM KCN was added to the staining solution to isolate MnSOD activity. SOD activity was extrapolated from an in-gel standard curve constructed using bovine superoxide dismutase. Chemicals used in the assays were purchased from Sigma–Aldrich (St. Louis, MO) or Bioshop (Burlington, ON). Statistical analysis Data for three or more experimental groups were analyzed by repeated measures two-factor ANOVA. Post-hoc comparisons between means were done by Tukey’s test using Systat v.12. Data comparing two experimental groups were analyzed using the
student’s t-test. All data are presented as means ± standard error of the mean (SEM). A p-value of
Contents lists available at ScienceDirect
Phytochemistry journal homepage: www.elsevier.com/locate/phytochem
The stilbenes resveratrol, pterostilbene and piceid affect growth and stress resistance in mammalian cells via a mechanism requiring estrogen receptor beta and the induction of Mn-superoxide dismutase Ellen L. Robb 1, Jeffrey A. Stuart ⇑ Department of Biological Sciences and Cold Climate Oenology and Viticulture Institute, Brock University, St. Catharines, ON L2S 3A1, Canada
a r t i c l e
i n f o
Article history: Received 15 June 2013 Received in revised form 25 November 2013 Available online 19 December 2013 Keywords: Stilbene Superoxide dismutase Mitochondria Reactive oxygen species Estrogen receptor beta Cell growth Stress resistance
a b s t r a c t The mitochondrial antioxidant enzyme, Mn superoxide dismutase (MnSOD), has been shown to confer cytoprotection and to regulate cell cycle progression. Resveratrol, a phytoestrogen found in red wines and other foods, has been previously reported to increase MnSOD protein levels and activity both in vitro and in vivo. Numerous structural analogues of resveratrol produced via the same stilbene synthesis pathway (e.g. pterostilbene and piceid) and also present in foods and red wine may be capable of eliciting the same effects. Furthermore, in humans resveratrol is rapidly metabolized to resveratrol-40 sulfate, resveratrol-3-glucuronide and other metabolites in vivo. Although these metabolites may accumulate to relatively high levels in plasma and tissues, little is known about their biological activities. Here the activities were compared of these stilbenes and stilbene metabolites in mammalian cells. Two key cellular activities associated with resveratrol were examined: inhibition of proliferative growth and increased stress resistance (important anti-cancer and cell protective activities, respectively). While resveratrol-40 -sulfate and resveratrol-3-glucuronide had no effect on either cell growth or stress resistance, both pterostilbene and piceid were at least as effective as resveratrol. Using pharmacological and genetic approaches, it was found that the effects of pterostilbene and piceid required an induction of the mitochondrial enzyme MnSOD and intact mitochondrial respiration. In addition, using estrogen receptor beta (ERbeta) knockout mouse myoblasts, it was demonstrated that the effects of stilbene compounds on cell growth and stress resistance all require ERbeta. Taken together, these results indicate that resveratrol, pterostilbene and piceid all activate the same mitochondrial response in mammalian cells, and therefore these latter two molecules might be as effective as resveratrol in eliciting positive health outcomes in vivo. Ó 2013 Elsevier Ltd. All rights reserved.
Introduction The mitochondrial matrix antioxidant enzyme, Mn-superoxide dismutase (MnSOD), confers protection against cell death (e.g., Venkataraman et al., 2004; Silva et al., 2005; Kowluru et al., 2006; Fisher and Goswami, 2008; Dumont et al., 2009). Interestingly, in addition to this cytoprotective ability, MnSOD exerts effects on cell replication (e.g. Li et al., 1995; Liu et al., 1997; Yan et al., 1996; Behrend et al., 2005; Sarsour et al., 2008) and differentiation (e.g. Velarde et al., 2012; Schneider et al., 2011). The ability of transgenic MnSOD over-expression to slow the growth rate of cancerous cells is of particular interest. Many cancerous cell lines are characterized by abnormally low levels of MnSOD expression,
⇑ Corresponding author. Tel.: +1 905 688 5550x4814; fax: +1 905 688 1855. E-mail address: [email protected] (J.A. Stuart). Current address: Mitochondrial Biology Unit, Medical Research Council, Cambridge, UK. 1
0031-9422/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.phytochem.2013.11.019
and re-establishing MnSOD expression to normal levels can slow the growth of these cells (Li et al., 1995; Liu et al., 1997; Yan et al., 1996; Behrend et al., 2005). Interestingly, the only enzymatic activity attributed to MnSOD is the conversion of the superoxide anion to hydrogen peroxide in the mitochondrial matrix, which suggests that its effect on cell proliferation and stress resistance is directly related to its influence on the redox environment of mitochondria, the mechanistic details of which are unclear. Identifying a means of inducing MnSOD expression in cancerous cells through dietary or pharmacological interventions may have therapeutic potential. A robust upregulation of MnSOD was identified that was elicited by treatment of human and mouse cells with resveratrol (1) (see Fig. 1) (Robb et al., 2008a; Robb and Stuart, 2011), a compound found in red wines, mulberries and peanuts, that is widely associated with inhibitory effects on the growth of both cancerous and normal cell lines. This observation has now been made for a collection of diverse cell types that includes cardiomyocytes, neuroblastomas (SK-N-BE), a hippocampal
165
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
neuronal cell line (HT22), coronary arterial endothelial cells, and pheochromocytoma cells (PC6.3) (Movahed et al., 2012; Albani et al., 2009; Fukui et al., 2010; Ungvari et al., 2009; Kairisalo et al., 2011). Using siRNA, it was demonstrated that the growth inhibitory effect of resveratrol (1) is dependent upon MnSOD induction (Robb and Stuart, 2011). Similarly, Fukui et al. (2010) showed that an induction of MnSOD is necessary for the cytoprotective effects of resveratrol (1) in cultured neuronal cells. Its ability to stimulate MnSOD expressionis not exclusive to cells in culture. MnSOD upregulation in whole brain tissue, skeletal muscle, hematopoietic stem cell populations, and dopaminergic neurons can be achieved in vivo via dietary supplementation (Robb et al., 2008b; Jackson et al., 2011; Zhang et al., 2013; Mudò et al., 2011). Thus, MnSOD is a target of resveratrol (1) both in vitro and in vivo that has been linked to important biological outcomes. Resveratrol (1) belongs to the stilbene family of phytoalexins, which are produced by grapevine tissues (Vitis vinifera) in response to a variety of biotic and abiotic stresses. In the biosynthetic pathway, resveratrol is an intermediary, which undergoes subsequent modification to pterostilbene (2) and piceid (3) (Fig. 1). Recent studies have shown that pterostilbene (2) is at least as effective as resveratrol at inhibiting cancer cell growth in vitro and slowing tumour growth in vivo (Lin et al., 2012; Moon et al., 2013; McCormack et al., 2012). A limited amount of available data suggests that piceid (3) may exert similar effects (Su et al., 2013). Thus, both the structural similarities and the observed parallels in their biological activities in mammals suggest that resveratrol (1), pterostilbene (2), and piceid (3) might all work via the same basic mechanism. In a similar vein, two of the major metabolites of resveratrol (1) in vivo are resveratrol-40 -Sulfate (RES-S) (4) andresveratrol-3O-b-D-glucuronide (RES-G) (5), which both retain basic structural
similarity to resveratrol (1) (Fig. 1). The effects of these two metabolites on cell proliferation have not been well explored. Among the earliest biological activities attributed to resveratrol (1) was its estrogenic effect in mammalian cells. It is a phytoestrogen capable of binding to estrogen receptor alpha and beta (Gehm et al., 1997; Bowers et al., 2000; Salah et al., 2013). Its effect on MnSOD expression is inhibited by the estrogen receptor (ER) antagonist ICI 182,780, and can be phenocopied using the specific ERbeta agonist diarylpropionitrile (DPN), but not the ERalpha agonist propyl pyrazoletriol (PPT), suggesting the importance of specifically ERbeta in resveratrol’s cellular effects (Robb and Stuart, 2011). ERbeta agonists are generally associated with anti-proliferative effects (Warner and Gustafsson, 2010). A relatively wide range of structurally related molecules is predicted to selectively bind ERbeta, suggesting that multiple small molecules could target this estrogen receptor in vivo. Here a variety of molecular and cellular approaches were used to characterize the effects of resveratrol (1) derived stilbeneson cell growth and cytoprotection. siRNA MnSOD knockdown and MnSOD-null cells are used to demonstrate the requirement of this enzyme for the growth inhibitory effects of resveratrol (1), pterostilbene (2), and piceid (3). The only known activity of MnSOD is the dismutation of superoxide in the mitochondrial matrix (derived primarily from respiration) to hydrogen peroxide. Therefore, respiratory deficient rho0 PC3 prostate cancer cells were utilized to show that respiratory superoxide production is necessary for the observed effect on growth (Hoffmann et al., 2004; Chandel et al., 1998). Finally, using ERbeta-null cells the involvement of ERbeta in the induction of MnSOD expression and subsequent effects on cell growth and cytoprotection elicited by resveratrol (1) derived stilbenes were demonstrated. Together, these experiments reveal
H3C
O
O
CH3
HO
OH
pterostilbene (2)
OH
HO OH trans-resveratrol (1)
O HO O
HO
OH
piceid (3)
HO OH
O
OH
OH
O HO HO
O
O
HO OH HO OH
resveratrol-3-O-β-D-glucuronide (5)
OH
S O
O
trans-resveratrol-4’-sulfate (4)
Fig. 1. Molecules included in this study: trans-resveratrol (1), pterostilbene (2), piceid (3), trans-resveratrol-40 -sulfate (RES-S) (4), resveratrol-3-O-b-D-glucuronide (RES-G) (5), and their relationship to trans-resveratrol.
166
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
a pathway in which ERbeta agonists including resveratrol (1), pterostilbene (2), and piceid (3) stimulate the expression of MnSOD, which via the modulation of mitochondrial ROS metabolism regulates cell growth and cytoprotection. We suggest that this shared mechanism of action underlies the observation that all three molecules are capable of eliciting similar effects on mammalian cells.
resveratrol (1) (Fig. 2A and B). In contrast, however, neither RESS (4) nor RES-G (5) affected MnSOD expression. In the same experiment, both pterostilbene (2) (20 lM) and piceid (3) (50 lM) significantly inhibited the proliferative growth of C2C12 cells (Fig. 2C), whereas again neither RES-S (4) nor RES-G (5) had any effect at the concentrations tested (up to 50 lM). Similar results were observed in cultured human lung fibroblasts (MRC5; data not shown). Since resveratrol (1) has cytoprotective properties and pretreatment of C2C12 cells with resveratrol (1) elicits a cytoprotective phenotype (e.g. Robb and Stuart, 2011), it can be predicted that pterostilbene (2) and piceid (3) would be similarly cytoprotective. All of these compounds are chemical antioxidants, and so were removed immediately prior to the experiment by washing and refreshing cells with new media 1 h before exposure to hydrogen peroxide (30–160 lM) for 2 h. Treatment with resveratrol (1), pterostilbene (2), or piceid (3) for 48 h prior to hydrogen peroxide exposure significantly increased the LD50 in C2C12 cells (Fig. 2D). The cytoprotective properties of these molecules appeared not to be mediated by broad increases in antioxidant enzymes generally, since virtually no induction of CuZnSOD, catalase, glutathione peroxidase was observed (Table 1). Similarly, the increased stress resistance elicited by these compounds was not associated with increased heat shock protein 60, or heat shock protein 70 protein levels (Table 1).
Results Effects of resveratrol analogues on C2C12 myoblasts
MnSOD’s anti-proliferative and cytoprotective properties have been well documented. To isolate the putative role of stilbene-
MnSOD protein level/Cs activity
MW
MnSOD
Memcode protein stain
DMSO EtOH Resveratrol (1) RES-S (4) RES-G (5) Piceid (2) Pterostilbene (3)
0.15 0.12
* * *
0.09 0.06 0.03 0
D
C 25 20 15 10 5 0
200 *
*
*
Hydrogen Peroxide LD50 iin uM
30 Population Doubling Time (h)
Growth inhibition and cytoprotection in the absence of MnSODor mitochondrial ROS
B
Piceid (3)
Pterostilbene (2)
Resveratrol (1)
RES-S (4)
EtOH
DMSO
A
RES-G (5)
Induction of MnSOD expression appears to be a critical event underlying two key biological activities of resveratrol (1): inhibition of cell growth and stimulation of cellular stress resistance (Robb and Stuart, 2011). It can therefore, be hypothesized that pterostilbene (2) and piceid (3), as resveratrol (1) derivatives with a high degree of structural similarity (Fig. 1), would work via the same mechanism. In addition, trans-resveratrol-40 -sulfate (RES-S) (4) and resveratrol-3-O-b-D-glucuronide (RES-G) (5) (Fig. 1) are two by products of resveratrol (1) metabolism in humans that are detectable at high levels in plasma and tissue of individuals consuming resveratrol (1) (Wenzel and Somoza, 2005). Since these metabolites also retain some structural similarity and might also be biologically active, they were included in this study. C2C12 myoblasts were treated for 48 h with resveratrol (1) (range of concentrations up to 50 lM), piceid (3) (up to 50 lM), pterostilbene (2) (up to 20 lM), RES-S (4) (up to 50 lM), or RES-G (5) (up to 50 lM). Cells were then harvested to determine the effects of these compounds on MnSOD expression. As predicted pterostilbene (2) and piceid (3) elicited a similar induction of MnSOD as did
* 160
*
*
120 80 40 0
Fig. 2. Piceid and pterostilbene, but not RES-S or RES-G, increase population doubling time, stress resistance, and MnSOD expression in C2C12 myoblasts. (A) Representative Western blot and corresponding protein stain of MnSOD in C212 myoblasts. (B) Relative MnSOD protein levels in C2C12 myoblasts treated with DMSO, ethanol, resveratrol (1) (25 lM), RES-S (4) (50 lM), RES-G (5) (50 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). Since phytoestrogens including resveratrol can stimulate a general increase in mitochondrial abundance in many cell types, to accurately evaluate changes in MnSOD levels citrate synthase activity was used as a proxy of mitochondrial abundance. (C) Average population doubling time in C2C12 myoblasts treated with DMSO, ethanol, resveratrol (1) (25 lM) RES-S (50 lM), RES-G (50 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (D) Hydrogen peroxide LD50 of C2C12 myoblasts treated with DMSO, ethanol, resveratrol (1) (25 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). Data shown represent the means of 3–5 independent trials. Error bars represent SEM. ⁄p < 0.05 compared to vehicle control.
167
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
Table 1 Effects of resveratrol analogues on antioxidant enzymes and heat shock proteins. catalase activity, glutathione peroxidase activity and CuZnSOD, Hsp60 and Hsp70 protein level in C2C12 myoblasts. Treatment
Catalase activity (mmol/min/mg cellular protein)
Glutathione peroxidase activity (mmol/min/mg cellular protein
CuZn superoxide dismutase protein level relative to internal control
Hsp60 protein level/Cs activity relative to internal control
Hsp70 protein level relative to internal control
DMSO EtOH Resveratrol (1) Pterostilbene (2) Piceid (3) RES-S (4) RES-G (5)
15.51 ± 0.24 15.80 ± 0.47 15.42 ± 0.07 16.32 ± 0.34 19.18 ± 0.50* N.D. N.D.
67.32 ± 5.27 66.52 ± 8.14 51.41 ± 5.04 68.41 ± 7.58 51.22 ± 3.07* N.D. N.D.
0.96 ± 0.12 0.99 ± 0.08 1.25 ± 0.43 1.20 ± 0.16 1.26 ± 0.19 1.06 ± 0.17 1.11 ± 0.13
0.089 ± 0.020 0.097 ± 0.018 0.102 ± 0.010 0.095 ± 0.014 0.106 ± 0.018 0.100 ± 0.015 0.067 ± 0.011
1.05 ± 0.22 0.91 ± 0.36 1.12 ± 0.18 1.21 ± 0.09 1.34 ± 0.17 1.25 ± 0.13 1.27 ± 0.15
Data represents the mean of duplicate measurements for 3 independent trials. N.D. = not determined. p < 0.05.
*
mediated MnSOD induction in cell growth inhibition and cytoprotection, siRNA against MnSOD was used to prevent its induction during 48 h treatment with resveratrol (1), piceid (3), or pterostilbene (2) (Fig. 3A). Using this approach, none of the stilbenes had any effect on proliferative growth (Fig. 3B) or stress resistance (Fig. 3C) in C2C12 myoblasts. Similar effects were observed in MRC5 fibroblasts (data not shown). To gain additional insight into the role of MnSOD in the effects of stilbenes in mammalian cells, these experiments were repeated (but without siRNA) using mouse embryonic fibroblast lines (MEFs) derived from MnSODnull and wildtype mice. In agreement with our siRNA experiments, the growth inhibitory effects of resveratrol (1), pterostilbene (2) and piceid (3) were absent in MnSOD-null MEFs (Fig. 3D). Stress resistance experiments were not conducted in MnSOD-null MEFs. The only known activity reported for MnSOD is the dismutation of superoxide to hydrogen peroxide in mitochondria. Within the mitochondrial matrix, superoxide is produced by the single electron reduction of molecular oxygen during respiration. To determine whether the growth inhibitory effects of stilbenes were dependent upon mitochondrial ROS, respiration deficient cells were created. PC3 prostate cancer cells were used for this experiment, since they are highly responsive to stilbenes and an existing protocol is available for rendering the cells mtDNA-free (rho0; see Experimental section). Using normal and rho0 cells, we investigated whether: (1) the stilbene-mediated induction of MnSOD would occur even in the absence of respiration, and (2) if MnSOD was induced, would its effect on cell growth still be evident. It was found that the induction of MnSOD activity by resveratrol (1), pterostilbene (2) or piceid (3) was virtually identical in normal and rho0 cells (Fig. 4A), indicating that this was independent of respiratory superoxide production. While the induction of MnSOD protein levels was paralleled by increases in MnSOD activity in the normal PC3 cells, in rho0 PC3 cells MnSOD protein levels, but not activities, were increased (Fig. 4C,D). Also, the induction of MnSOD expression in rho0 cells was not accompanied by effects on cell growth (Fig. 4B). Estrogen receptor beta is required for the effects of resveratrol, pterostilbene and piceid Resveratrol (1) is an ER agonist (Gehm et al., 1997; Bowers et al., 2000). Its effects on MnSOD expression, cell proliferation and stress resistance are abolished by an ER antagonist (ICI 182, 780) and can be reproduced using the specific ERbeta agonist DPN (Robb and Stuart, 2011). To determine if the observed effects of stilbenes were mediated through ERbeta two approaches were used. Firstly, the experiments were repeated in the presence of the ER antagonist ICI182,780, and secondly they were repeated with myoblasts cell lines established from ERbeta-null mice (see Methods).
ICI182,780 prevented the effects of all three molecules on MnSOD, cell growth, and stress resistance (Table 2). Similarly, the absence of ERbeta diminished the effects of resveratrol (1), piceid (3) and pterostilbene (2) on MnSOD protein levels (Fig 5A), population doubling time (Fig. 5B) and hydrogen peroxide LD50 (Fig. 5C). Together, these results suggest a role for ERbeta in the effects of stilbenes on these cellular activities. Discussion Resveratrol (1) is a naturally occurring stilbene phytoalexin produced by V. vinifera in response to stress that has been much studied for its cytoprotective and growth inhibitory effects. Surprisingly, the structurally related molecules pterostilbene (2) and piceid (3), produced from resveratrol (1) in V. vinifera under the same conditions by the same biosynthetic pathway and which are also present in grapes skins and red wines, have received far less attention. Quite recently, the growth inhibitory effects of pterostilbene (2) in vitro (Lin et al., 2012; Moon et al., 2012) and in vivo (McCormack et al., 2012) have been reported, and a limited amount of data describing a similar effect is available for piceid (Su et al., 2013). Comparable to resveratrol (1), pterostilbene (2) induces a strong upregulation of MnSOD that is concurrent with reduced proliferative growth in breast and pancreatic cancer cells (Moon et al., 2012; McCormack et al., 2012). Here it is shown that all three compounds, resveratrol (1), pterostilbene (2) and piceid (3), inhibit the proliferative growth of mammalian cells via a mechanism involving the ERbeta-mediated induction of MnSOD in mitochondria. MnSOD is a cell cycle regulator in mammalian cells, via poorly understood redox modifications of key regulatory proteins that are not yet fully elucidated, though they continue to be actively studied (Kim et al., 2010; Sarsour et al., 2008; Menon et al., 2007). In many cells, reduced MnSOD expression is associated with increased replication rates, which can be slowed by restoration of MnSOD to normal levels (Li et al., 1995; Liu et al., 1997; Yan et al., 1996; Behrend et al., 2005). The inhibitory effects of resveratrol, pterostilbene (2), and piceid (3) on cell growth measured in this study are dependent upon their ability to upregulate MnSOD expression and activity. This was demonstrated using two complementary approaches: (1) preventing the induction of MnSOD expression using siRNA, and (2) eliminating MnSOD by using MnSOD-null MEFs. None of the three compounds inhibited proliferative growth in the absence of the MnSOD induction. MnSOD catalyzes the dismutation of mitochondrial matrix superoxide to hydrogen peroxide, and in so doing may affect the concentrations of several ROS and reactive nitrogen species (RNS) including superoxide, hydrogen peroxide (Liochev and Fridovich, 2007), nitric oxide, and peroxynitrite as a result of secondary reac-
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173 Control siRNA
Piceid (3)
Pterostilbene (2)
DMSO
*
Control siRNA
2
DMSO EtOH Piceid (3) Pterostilbene (2)
*
EtOH
Relative MnSOD protein level
2.5
DMSO
A
Piceid (3)
Pterostilbene (2)
MnSOD siRNA
EtOH
168
MnSOD
1.5 1 0.5
Ponceau S protein stain
0 Control siRNA
MnSOD siRNA
C Hydrogen Peroxide LD50 (uM)
B Population Doubling Time (h)
35 30
* *
25 20 15 10 5 0
200 * 160 *
DMSO EtOH Resveratrol (1) Piceid (3) Pterostilbene (2)
120
80
40
0 Control siRNA
MnSOD siRNA
Control siRNA
MnSOD siRNA
MEF Population Doubling Time (h)
D 80
*
70 *
*
60 50 40 30 20 10 0 Wild Type MEF
MnSOD -/- MEF
Fig. 3. MnSOD is essential for piceid and pterostilbene to increase population doubling time and stress resistance in C2C12 myoblasts. (A) Relative MnSOD protein level, representative western blot and corresponding protein stain in C2C12 myoblasts treated with control siRNA or MnSOD siRNA ± DMSO, ethanol, piceid (3) (50 lM) prpterostilbene (2) (20 lM in ethanol). (B) Average population doubling time in C2C12 myoblasts treated with control siRNA or MnSOD siRNA ± DMSO, ethanol, piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (C) Hydrogen Peroxide LD50 of C2C12 myoblasts treated with control siRNA or MnSOD siRNA ± DMSO, ethanol, piceid (50 lM) or pterostilbene (20 lM in ethanol). (D) Average population doubling time in wildtype and MnSOD null mouse embryonic fibroblasts treated with DMSO, ethanol, resveratrol (15 lM), piceid (25 lM) or pterostilbene (10 lM in ethanol). Data shown represents the mean of 3 independent trials. Error bars represent SEM. ⁄p < 0.05 compared to vehicle control.
tions (Keller et al., 1998). An important observation made in this study was that the ability of MnSOD to inhibit growth was present at physiological oxygen levels (3%, necessary for the culture of primary myoblasts generated from wildtype and ERbeta null mice) and was not therefore an artefact of high oxygen levels in culture (typically 18–20%). It is possible that MnSOD’s effect on cell growth could be related to an as yet uncharacterized activity of the protein beyond ROS metabolism. To verify that the observed effects of MnSOD on proliferative cell growth were indeed mediated by matrix superoxide metabolising activity we used respiration-deficient rho0 PC3 cells. Rho0 cells have been productively used to study a variety of phenomena that are modulated by mitochondrial ROS metabolism (Schauen et al., 2006; Chandel and Schumacker, 1999; Zhang and Gutterman, 2007). PC3 cells are an appropriate cell type for this investigation since they express high levels of
ERbeta (McPherson et al., 2010) and the growth of wildtype PC3 cells is strongly inhibited by ERbeta agonists, including DPN and resveratrol (1). It was found that, while the stilbenes and DPN all strongly upregulated MnSOD protein levels in rho0 PC3 cells despite the absence of respiration, there was no concomitant inhibition of growth. This indicates that mitochondrial respiration is not necessary for the ERbeta mediated induction of MnSOD. This observation is important because resveratrol (1) has been shown to inhibit mitochondrial respiration (McPherson et al., 2010; Zini et al., 1999) and some respiratory inhibitors promote superoxide production (reviewed by Lambert and Brand, 2009). The fact that MnSOD protein levels were almost identically upregulated in the absence of respiration thus indicates a more direct mechanism of upregulating MnSOD expression. Interestingly, although MnSOD protein levels were significantly higher in the rho0 cells there
169
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
A
4.0 * 3.0
*
*
3.5 *
*
*
2.5 2.0 1.5 1.0 0.5 0.0 rho0 PC3
Wildtype PC3
*
70
*
60
*
50 40 30 20 10 0 rho0 PC3
MnSOD Activity/CS Activity
0.10
DMSO
Resveratrol (1)
Pterostilbene (2)
Piceid (3)
EtOH
DMSO
Resveratrol (1)
Piceid (3)
80
Wildtype PC3
D
rho0 PC3
Wildtype PC3
EtOH
C Pterostilbene (2)
DMSO EtOH Resveratrol (1) Piceid (3) Pterostilbene (2)
B
4.5
PC3 Population Doubling Time (h)
MnSOD protein level/CS Activity
5.0
*
0.09 0.08
*
*
0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00
Wildtype PC3
rho0 PC3
Fig. 4. Resveratrol, piceid and pterostilbene require mitochondrial ROS production to affect cell growth. (A) Average MnSOD protein levels normalized to citrate synthase activity in wildtype and rho0 PC3 cells treated with DMSO, ethanol, resveratrol (1) (25 lM), RES-S (4) (50 lM), RES-G (5) (50 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (B) Average population doubling time in wildtype and rho0 PC3 cells treated with DMSO, ethanol, resveratrol (1) (25 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). (C) Representative MnSOD activity gel in wildtype and rho0 PC3 cells. (D) MnSOD activity normalized to citrate synthase activity in wildtype and rho0 PC3 cells treated with DMSO, ethanol, resveratrol (1) (25 lM), piceid (3) (50 lM) or pterostilbene (2) (20 lM in ethanol). ⁄p < 0.05 compared to vehicle control of the same genotype.
Table 2 Effects of resveratrol analogues on population doubling time, resistance to hydrogen peroxide induced cell death, and fold change in MnSOD protein level in the absence and presence of ICI182,780 in C2C12 myoblasts. Treatment
DMSO EtOH Resveratrol (1) Pterostilbene (2) Piceid (3)
Population Doubling Time (h)
H2O2 LD50 (lM)
No antagonist
Antagonist
No antagonist
Antagonist
No antagonist
Antagonist
14.98 ± 1.16 15.01 ± 0.94 26.34 ± 1.25* 33.61 ± 1.08* 27.14 ± 1.11*
15.29 ± 0.11 14.87 ± 0.16 17.23 ± 1.22 22.19 ± 2.17$ 21.24 ± 1.05#
92 ± 11 84 ± 9 144 ± 19* 166 ± 18* 149 ± 16*
82 ± 7 93 ± 6 96 ± 11 110 ± 9 89 ± 17
1.06 ± 0.15 1.10 ± 0.18 2.10 ± 0.24* 2.22 ± 0.29* 1.84 ± 0.17*
0.94 ± 0.12 1.04 ± 0.08 0.82 ± 0.16 1.19 ± 0.18 0.97 ± 0.10
MnSOD protein level relative to internal control
Antagonist = ICI182,780. Data represents the mean of duplicate measurements for 3 independent trials. * p < 0.05 when compared to vehicle control. # p = 0.056. $ p = 0.052.
was no concomitant increase in MnSOD activity. MnSOD activity is modulated by a variety of post-translational modifications including acetylation (Ozden et al., 2011) and methylation (Sarsour et al., 2012). It is, therefore, hypothesized that these signalling pathways may be disrupted in rho0 cells, but at this time the mechanism underlying this observation were not known. Importantly, there was also no effect of MnSOD induction on cell growth in the absence of respiration. Thus, it seems that the ability of stilbenes to modulate cellular activities via MnSOD rely upon an active respiratory chain, though further work is required to fully characterize this. Endogenous estrogens like 17beta-estradiol exert a variety of effects on mitochondrial function, including an upregulation of MnSOD in vitro and in vivo. In cultured vascular smooth muscle cells estrogen treatment significantly increases MnSOD activity, and this is accompanied by a reduction in proliferative cell growth
(Sivritas et al., 2011). In rats, estrogen treatment increases MnSOD levels in mitochondria isolated from brain tissue (Razmara et al., 2007), and a downregulation of MnSOD is observed in vascular tissue of ovariectomized mice (Strehlow et al., 2003). The involvement of ERbeta signalling in the pathway upstream of MnSOD is supported here by the ability of the ERbeta-specific agonist DPN to reproduce the phenotype, and the absence of effects of resveratrol (1) and its derivatives in cells lacking ERbeta. Here data is shown for myoblasts, but virtually identical results using ERbetanull fibroblasts derived from the same mice were found (not shown). ERbeta is predicted to bind quite a wide range of dietary phytoestrogens including resveratrol (1) (Yuan et al., 2011), which has been demonstrated in vitro (Bowers et al., 2000; Zhu et al., 2006). In agreement with the anti-proliferative effects of resveratrol (1) and its derivatives, activation of ERbeta using pharmacological agonists is generally associated with cell cycle
170
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
B 40 0.25 0.20
*
Population Doubling Time (h)
Myoblast MnSOD protein level/ Cs activity
A
*
* *
0.15 0.10 0.05 0
ERBeta +/+
30
*
*
*
25 20 15 10 5 0
ERbeta -/-
*
35
DMSO EtOH Resveratrol (1) Piceid (3) Pterostilbene (2)
ERBeta+/+
ERbeta-/-
Hydrogen peroxide LD50 (µM)
C 125 100
*
*
* *
75 50 25 0 ERBeta+/+
ERbeta-/-
Fig. 5. Involvement of ERbeta in resveratrol, piceid and pterostilbene effects on population doubling time, stress resistance and MnSOD protein level. Relative MnSOD protein levels normalized to citrate synthase activity over control group in wildtype and ERbeta null myoblasts treated with DMSO, DPN (10 lM), resveratrol (1) (25 lM), piceid (3) (50 lM) and pterostilbene (2) (10 lM). (B) Average population doubling time in wildtype and ERbeta null mouse myoblasts treated with DMSO, DPN (10 lM), resveratrol (1) (25 lM), piceid (3) (50 lM) and pterostilbene (2) (10 lM). (C) Hydrogen Peroxide LD50 over vehicle control in wildtype and ERbeta null myoblasts treated with DMSO, DPN (10 lM), resveratrol (1) (25 lM), piceid (3) (50 lM) and pterostilbene (2) (10 lM). Data shown represents the mean of duplicate measures from cell lines established from 5 individuals of each genotype. Error bars represent SEM. ⁄p < 0.05 compared to vehicle control of the same genotype.
inhibition and anti-proliferative effects (reviewed in Warner and Gustafsson, 2010; Nilsson et al., 2011). The ERbeta agonist diarylpropionitrile (DPN) slows growth of the colon cancer cell line MC38 (Motylewska et al., 2009), and lentivirus transfection of ERbeta into a cancerous colon cell line significantly reduces cell proliferation (Hartman et al., 2009). It is concluded that ERbeta is mediating the effects of resveratrol (1), pterostilbene (2), and piceid (3) on MnSOD, which in turn affects a reduction in growth rate and enhanced cellular stress resistance. Indeed it is noted that these are two phenotypes which have been described for ERbeta agonists generally (Nilsson et al., 2011). Two of the most abundant metabolites of resveratrol (1) in vivo, RES-S (4) and RES-G (5) (Wenzel and Somoza, 2005), had no significant effect on MnSOD expression, proliferative growth, or stress resistance, suggesting that these common metabolites of resveratrol in vivo do not contribute to the inhibition of cell growth. In agreement with these results, the proliferative rate of the cancerous mammary cell line MCF7 has also been shown to be unaffected by the 30 -sulfate and 4-sulfate resveratrol metabolites (Miksits et al., 2009). However, sulfate metabolites of resveratrol (1) do apparently maintain other properties, and were shown to share the ability of resveratrol (1) to inhibit inducible nitric oxide synthase activity in cultured macrophage cells (Hoshino et al., 2010).
Conclusions In conclusion, evidence is provided that the ability of resveratrol (1), pterostilbene (2), and piceid (3) to slow cell growth, which has been observed in multiple cell types, is dependent upon the induction of MnSOD expression in the context of normal mitochondrial respiration. This induction of MnSOD is not caused
by a resveratrol-mediated augmentation of mitochondrial ROS production, since it occurs in the absence of respiration. Resveratrol (1) is a demonstrated low-affinity ERbeta agonist, and the data herein indicate that the induction of MnSOD requires the presence of ERbeta. The effects can be phenocopied by the highly specific ERbeta agonist DPN, abolished using an ER agonist (ICI 182,780), and are absent in ERbeta-null cells. Importantly, the effects we and others have observed with resveratrol (1) and pterostilbene (2) in cultured cells can be achieved in vivo via dietary supplementation, which suggests that both molecules, and perhaps also piceid (3), would be capable of modulating MnSOD expression in a variety of human health contexts that might benefit from this.
Experimental General experimental procedures All estrogens and phytoestrogens used in this study were purchased from commercial sources and were of the highest purity available. trans-Resveratrol (1) was obtained from A.G. Scientific (San Diego, CA). trans-Piceid (3) was purchased from Sigma (St. Louis, MO). Pterostilbene (2) was obtained from Cayman Chemical (Ann Arbor, MI). trans-Resveratrol-40 -Sulfate (RES-S) (4) sodium salt and trans-Resveratrol-3-O-b-D-glucuronide (RES-G) (5) were purchased from Toronto Research Chemicals (Toronto, Canada). ICI 182780 and DPN were purchased from Tocris Biosciences (Ellisville, MO). All polyphenols purchased were of greater than 95% purity. Ethidium bromide, pyruvate and uridine were purchased from Bioshop (Burlington, Canada). Experiments with cultured cells were performed on cells grown at 37 °C and 5% CO2, and either 3% O2 (for primary cell lines:
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
MnSOD null, ERbeta null) or 18% O2 (for all other cell lines). The established C2C12 mouse myoblast and PC3 human prostate cancer cell lines were cultured in accordance with the distributor’s protocol (American Type Culture Collection; Manassas, VA) and subcultured as required in Dulbecco’s modified Eagle medium with high glucose (C2C12) or minimum essential medium (PC3) with 10% (v/v) FBS, 1% (v/v) penicillin streptomycin solution, and 2% (v/v) non-essential amino acids (Hyclone; Logan, UT). MnSOD null and wild-type mouse embryonic fibroblasts were generously provided by Dr. Prabhat Goswami of the University of Iowa. MnSOD null and wild type mouse embryonic fibroblasts were cultured at 37 °C, 5% CO2, 3% O2in DMEM high glucose supplemented with 10% (v/v) FBS, 1% (v/v) penicillin streptomycin solution, and 2% (v/v) non essential amino acids. ERbeta null and control myoblast cell lines were generated from 5 individuals each of wildtype and ERbeta/ mice on a C57/BL6 background (Taconic Farms, Hudson, NY, USA) (Krege et al., 1998). Mice were euthanized using isofluorane and rapid cervical dislocation following a protocol approved by the animal care committee at Brock University, and in accordance with the guidelines of the Canadian Council on Animal Care. Briefly, to isolate primary myoblasts hind limb skeletal muscle was excised, finely minced and suspended in pronase solution (2 mg/mL) at 37 °C for 1 h. The suspension was titurated by repeated pipetting and then filtered through sterile cheesecloth and centrifuged at 240 g for 5 min. The resulting supernatant was then centrifuged at 500 g for 5 min. The pellet was resuspended in growth media (Ham’s F10 nutrient mixture, 20% FBS, nonessential amino acids, penicillin, streptomycin, gentamicin, amphotericin B, basic human growth factor), plated on to collagen coated tissue culture dishes and incubated at 37 °C, 3%O2 and 5%CO2. RPMI-1640 Gentamicin, amphotericin B, collagen, and pronase were obtained from Sigma–Aldrich (St. Louis, MO). Basic human growth factor was purchased from Invitrogen (Burlington, Canada). An additional equal volume of media was added 48 h after plating, and the full volume of media was exchanged 96 h after plating. Myoblast outgrowth was evident at 96 h. Myoblasts were subcultured at a ratio of 1:3 when they reached 70% confluence. Rho0 PC3 prostate cancer cells were generated as outlined in Joshi et al. (1999). Briefly, PC3 cells were incubated in DMEM high glucose containing 10% (v/v) FBS, 1% (v/v) penicillin streptomycin solution, 2% (v/v) non-essential amino acids, and supplemented with ethidium bromide (200 ng/mL), pyruvate (100 lg/mL), uridine (50 lg/mL) for eight weeks. Culture media was refreshed daily. Ethidium bromide, pyruvate and uridine were added directly to the media each day. The ethidium bromide treatment was stopped after eight weeks. Rho0 status was verified by the absence of oxygen consumption and TFAM signal (not shown). Rho0 PC3 cells were stored as frozen stocks. For subsequent experiments, frozen cell stocks were thawed and propagated for several passages at 37 °C, 5% CO2, 18% O2 then treated as described. Chemicals used in enzyme assays were purchased from Sigma– Aldrich (St. Louis, MO) or Bioshop (Burlington, ON, Canada) unless otherwise indicated. Resveratrol analogue and estrogen antagonist treatments Each compound (resveratrol (1), pterostilbene (2), piceid (3), RES-S (4), RES-G (5)) was tested at 1, 10, 20 or 25, and 50 lM to evaluate possible toxicity and to determine the concentration required for maximal effects on population doubling time (data not shown). The lowest concentration giving a maximal effect was selected for further experiments. All compounds were dissolved in DMSO, with the exception of pterostilbene (2), which was dissolved in EtOH–H2O (95:5, v/v). In all cases, media was refreshed daily with freshly prepared compounds. All treatments
171
were of 48 h duration. Comparisons were made to the appropriate vehicle control for each compound and the data presented for each compound correspond to the concentration that yielded maximal effects on population doubling time. Where the estrogen receptor antagonist ICI 182,780 was used, it was added directly to the culture media at a final concentration of 10 lM 24 h prior to polyphenol treatment. This concentration was maintained throughout the 48 h incubation period and was also refreshed daily. siRNA treatment siRNA to MnSOD, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), a scrambled control sequence, and NeoFx Transfection Agent™ were purchased from Ambion (Austin, TX). C2C12 myoblasts were plated with NeoFx Transfection Agent™ and 2–10 nM of commercially available siRNA to MnSOD, GAPDH or a scrambled control sequence at a density of 1.0 106 cells/mL. siRNA containing medium was replaced 24 h after transfection. Phytoestrogen treatments of the knock-down cells were performed 6–72 h following transfection. Knock down efficiency was determined via western blot. Stress resistance and death experiments To evaluate stress resistance, myoblasts were washed three times with PBS and incubated with serum free medium containing 2% BSA with or without hydrogen peroxide (30–160 lM; 2 h). Following the incubation time, the media was removed, cells were rinsed with PBS and then incubated with serum containing medium for 18 h. Cell number and viability were calculated relative to the appropriate vehicle control at the indicated time points following the stressor exposure via trypan blue exclusion and cell counting. Samples of culture medium were taken to measure the release of lactate dehydrogenase (LDH) into the medium from dead cells. Lactate dehydrogenase activity LDH activity in the culture medium was measured in a solution containing 20 mM HEPES buffer (pH 7.3), 0.2 mM NADH and culture medium (200 lL). LDH activity was followed spectrophotometrically by the rate conversion of NADH to NAD+ at 340 nm. In RNAi experiments, LDH release was measured using the Cytotoxicity Detection Kit™ (Roche Applied Science, Laval, Canada) per the manufacturer’s instructions, as this kit is more accurate at the very low volumes used in these experiments. Preparation of whole cell lysates Cells for western blot or activity assay measurements were lysed by incubation for one hour in ice cold lysis buffer (10 mM Tris pH 8.0, 150 mM NaCl, 2 mM EDTA, 2 mM dithiothreitol, 0.4 mM PMSF, 40% (v/v) glycerol, 0.5% (v/v) NP40) with periodic sonication (Ultrasonic Inc., Sonicator W-375; setting 3). After incubation, cells were centrifuged at 10,000 g at 4 °C for 10 min (Thermo Scientific, IEC Micromax/Micromax RF). The protein concentration of the resulting supernatant was determined by the Bradford method using a BioRad protein determination kit (Hercules, CA). Whole cell lysates were stored in aliquots at 80 °C. Quantitative Western blotting Equal amounts of whole cell lysate (15 lg) were separated by SDS–PAGE (5% stacking, 12% resolving gels) and electroblotted onto a polyvinylidene fluoride membrane. Prestained broad range protein marker (Frogga Bioscience, Toronto, Canada) was included
172
E.L. Robb, J.A. Stuart / Phytochemistry 98 (2014) 164–173
on each gel. Memcode Reversible Protein Stain Kit™ (Thermo Fisher Scientific, Mississauga, Canada), or Ponceau S staining was used in accordance with the manufacturer’s instructions to verify equal protein loading/transfer. Following blocking, membranes were incubated with an antibody to MnSOD (1:5000 dilution; Enzo Life Sciences, Brockville, Canada), CuZnSOD (1:1000 dilution; Stressgen, Victoria, Canada), hsp70 (1:1000 dilution; Abcam, Cambridge, MA) or hsp60 (1:2000 dilution; Stressgen, Victoria, Canada). The membranes were visualized using the Odyssey infrared imaging system from LI-COR Biosciences, with IR-linked secondary antibodies to rabbit or mouse (1:5000 dilution; Rockland Immunochemicals, Gilbertsville, PA). Western blot quantitative analyses were performed using Odyssey imaging software 1.0. ERbeta null and control mice were purchased from Taconic Farms (Germantown, NY). Bovine superoxide dismutase was purchased from Sigma (St. Louis, MO). The C2C12 and PC3 cell lines were purchased from the American Type Culture Collection. All other chemicals and purified enzymes were obtained either from Sigma–Aldrich (St. Louis, MO), BioShop (Burlington, Canada) or Fisher Scientific (Mississauga, Canada) unless otherwise stated. Citrate synthase activity Citrate synthase (CS) activity was measured in a buffer containing 50 mM Tris pH 8.0, 0.5 mM 5,50 -dithiobis(2-nitrobenzoic acid), 0.1 mM acetyl-coenzyme A, 0.05% Triton X-100 and 5 lg protein. The reaction was initiated by the addition of 0.5 mM oxaloacetate, and absorbance was followed at 412 nm using a Bio-Tek KC4 plate reader. Chemicals used in the assays were purchased from Sigma– Aldrich (St. Louis, MO) or Bioshop (Burlington, ON). Antioxidant enzyme activities Catalase and glutathione peroxidase activity was measured as outlined in Robb et al. (2008b). Briefly, for catalase activity, the disappearance of hydrogen peroxide following the addition of whole cell lysate (50 lg) was monitored at 240 nm in a solution of 25 mM phosphate buffer pH 7.0 at 30 °C. Glutathione peroxidase activity was monitored by following the consumption of NADPH at 340 nm in a solution of 50 mM phosphate buffer pH 7.0, 0.4 mM EDTA, 0.15 mM NADPH, 1 mM reduced glutathione and 1 unit glutathione reductase. The reaction was initiated following the addition of 0.0007% (v/v) hydrogen peroxide. MnSOD activity was measured using an in-gel assay as described in Robb et al. (2008a,b). Whole cell lysate protein (100– 120 lg) was resolved by native-PAGE at 20 mA at 4 °C for 2 h. Gels were then incubated in 1.23 mM nitroblue tetrazolium for 15 minutes at room temperature. Gels were rinsed with deionised H2O and were transferred to a solution of 28 mM TEMED and 46 lM riboflavin in potassium phosphate buffer (pH 7.8) for 15 min at room temperature and protected from light. Gels were rinsed with deionised H2O and placed under fluorescent light until the background achieved a uniform violet colour. 5 mM KCN was added to the staining solution to isolate MnSOD activity. SOD activity was extrapolated from an in-gel standard curve constructed using bovine superoxide dismutase. Chemicals used in the assays were purchased from Sigma–Aldrich (St. Louis, MO) or Bioshop (Burlington, ON). Statistical analysis Data for three or more experimental groups were analyzed by repeated measures two-factor ANOVA. Post-hoc comparisons between means were done by Tukey’s test using Systat v.12. Data comparing two experimental groups were analyzed using the
student’s t-test. All data are presented as means ± standard error of the mean (SEM). A p-value of
