Induction of PC12 cell differentiation by flavonoids is dependent upon extracellular signal-regulated kinase activation

Authors


Address correspondence and reprint requests to Pamela Maher, Department of Cell Biology, The Scripps Research Institute, 10550 N. Torrey Pines Road, La Jolla, CA 92037, USA.
E-mail: pmaher@scripps.edu

Abstract

Many of the physiological benefits attributed to flavonoids are thought to stem from their potent antioxidant and free radical scavenging properties. Recently, it was shown that flavonoids protect nerve cells from oxidative stress by multiple mechanisms, only one of which is directly related to their antioxidant activity, suggesting that specific flavonoids may have other properties that could make them useful in the treatment of conditions that lead to nerve cell death. In particular, it was asked if any flavonoid could mimic neurotrophic proteins. To examine this possibility, we looked at the ability of flavonoids to induce nerve cell differentiation using PC12 cells. PC12 cells were treated with a variety of flavonoids to determine if there was a correlation between their neuroprotective activity and their neurite outgrowth-promoting activity. In addition, the signaling pathways required for flavonoid-induced differentiation were examined. We found that only a small subset of the flavonoids that were neuroprotective could induce neurite outgrowth by an extracellular signal-regulated kinase-dependent process. There was a strong correlation between the concentrations of the flavonoids that were neuroprotective and the concentrations that induced differentiation. These results suggest that the consumption of specific flavonoids could have further beneficial effects on nerve cells following injury, in pathological conditions or in normal aging.

Abbreviations used
CHX

cycloheximide

ERK

extracellular signal-regulated kinase

FGF

fibroblast growth factor

FGFR1

fibroblast growth factor receptor-1

GPCR

G protein-coupled receptor

MEK

mitogen-activated protein kinase/ERK kinase

NAC

N-acetylcysteine

NGF

nerve growth factor

PBN

α-phenyl-N-tert-butylnitron

PI3K

phosphatidylinositol 3-kinase

PKA

protein kinase A

PKC

protein kinase C

SDS

sodium dodecyl sulfate

SDS–PAGE

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

TPA

tetradecanoylphorbol acetate

Flavonoids are a family of plant-derived polyphenolic compounds widely distributed in fruits and vegetables and therefore regularly consumed in the human diet (for review see Middleton et al. 2000). A number of physiological benefits have been attributed to flavonoids, including protection from cardiovascular disease and cancer, and many of these effects are thought to stem from their potent antioxidant and free radical scavenging properties. One of us recently showed that flavonoids protect nerve cells from oxidative stress caused by GSH depletion (oxytosis) by three distinct mechanisms, only one of which is directly related to their antioxidant activity (Ishige et al. 2001). In addition to preventing the accumulation of reactive oxygen species following the exposure of nerve cells to glutamate or other inducers of oxidative stress, flavonoids can also prevent the early loss of GSH and the late influx of Ca2+. These data, as well as the literature (Middleton et al. 2000), suggest that specific flavonoids may have other important properties that could make them therapeutically useful in the treatment of conditions that lead to nerve cell death. Indeed, anthocyanidins, a class of flavonoids, are implicated in the positive effects of a blueberry-rich diet on age-related deficits in brain function (Galli et al. 2002).

To test the idea that flavonoids might have additional actions on nerve cells that could promote the recovery of higher neuronal functions, we looked at the ability of flavonoids to induce neurite outgrowth in PC12 cells, a well-studied model system of neuronal differentiation. In response to neurotrophic factors such as nerve growth factor (NGF), PC12 cells undergo a series of physiological changes culminating in a phenotype resembling that of sympathetic neurons (for review see Keegan and Halegoua 1993). These changes are the result of the activation of a coordinated series of signaling pathways and include the cessation of cell division, the expression of genes encoding nerve cell-specific proteins and the extension of neuritic processes. The signaling pathways underlying these changes have been well characterized and both the Ras–extracellular signal-regulated kinase (ERK) cascade and phosphatidylinositol 3-kinase (PI3K) have been shown to play critical roles in PC12 cell differentiation.

To determine if flavonoids could promote nerve cell differentiation, we tested a variety of flavonoids, including several that did not protect against oxidative stress, to determine if there was any correlation between the neuronal survival-promoting activity and the neurite outgrowth-promoting activity. In addition, we examined the signaling pathways that are activated by the flavonoids and are required for differentiation. We show here that no flavonoids that were not neuroprotective could induce neurite outgrowth but a small subset of the flavonoids that were neuroprotective did induce neurite outgrowth by an ERK-dependent process. However, unlike neurotrophic factors, which activate ERK rapidly after addition to cells, ERK activation by flavonoids was only seen after several hours of treatment. Furthermore, these flavonoids activated the complete Ras–ERK cascade by a novel mechanism that appears to be distinct from any known activation pathways.

Experimental procedures

Materials

Flavonoids were purchased from Alexis (San Diego, CA, USA) or Sigma/Aldrich (St. Louis, MO, USA) and solubilized in dimethylsulfoxide. PD98059 was obtained from Biomol (Plymouth Meeting, PA, USA) and solubilized in dimethylsulfoxide. U0126 was obtained from Promega (Madison, WI, USA) and solubilized in dimethylsulfoxide. GF109203X, Go6983, PD184352 and SU6656 were obtained from Calbiochem (San Diego, CA, USA) and solubilized in dimethylsulfoxide. Ascorbic acid, cycloheximide (CHX), forskolin, N-acetylcysteine (NAC), LY294002 and tetradecanoylphorbol acetate (TPA) were obtained from Sigma. NGF was obtained from Upstate Biotechnology (Lake Placid, NY, USA). All other chemicals were reagent grade.

Cell culture

PC12 cells (Greene and Tischler 1976) were obtained from D. Schubert (Salk Institute, La Jolla, CA, USA) and maintained in Dulbecco's modified Eagle's medium (Gibco, Carlsbad, CA, USA) supplemented with 10% horse serum (Hyclone, Logan, UT, USA), 5% fetal calf serum (Hyclone) and antibiotics. To examine the effects of flavonoids and other agents on the PC12 cells, the culture medium was removed and replaced by the chemically defined N2 medium (Gibco) 18 h prior to the start of the experiment. PC12 cells expressing an inducible, dominant negative form of Ras were obtained from S. Halegoua (SUNY, Stony Brook, NY, USA) and cultured in the same medium as the wild-type PC12 cells except for the addition of geneticin (200 µg/mL) (Thomas et al. 1992). To induce expression of dominant negative Ras, the cells were treated for 18 h with 1 µg/mL dexamethasone in N2 medium prior to the addition of fisetin. The PKA-deficient PC12 cells (A126-1B2 cells) were obtained from J. A. Wagner (Harvard Medical School, Boston, MA, USA) and cultured in the same medium as wild-type PC12 cells (Van Buskirk et al. 1985).

Stimulation with flavonoids

PC12 cells in N2 medium were treated with the flavonoids and other agents as described in the figure legends and, after the indicated time periods, the cells were solubilized in sodium dodecyl sulfate (SDS) sample buffer containing 0.1 mm Na3VO4 and 1 mm phenylmethylsulfonyl fluoride, boiled for 5 min and either analyzed immediately or stored frozen at −70°C.

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and immunoblotting

Proteins were separated on 10% SDS–polyacrylamide gels and transferred to nitrocellulose. Equal loading and transfer of the samples was confirmed by staining the nitrocellulose with Ponceau-S. Transfers were blocked for 2 h at room temperature with 5% non-fat milk in Tris-buffered saline/0.1% Tween 20 and then incubated overnight at 4°C in the primary antibody diluted in 5% bovine serum albumin in Tris-buffered saline/0.05% Tween 20. The primary antibodies used were: phospho-p44/42 MAP kinase antibody (#9101, 1/1000), phospho-MEK1/2 antibody (#9121, 1/1000) and phospho-Akt antibody (#9275, 1/1000) from Cell Signaling (Beverly, MA, USA); pan ERK antibody (#E17120, 1/10000) and anti-Ras (#R02120, 1/500) from Transduction Laboratories (San Diego, CA, USA). The transfers were rinsed with Tris-buffered saline/0.05% Tween 20 and incubated for 1 h at room temperature in horseradish peroxidase-goat anti-rabbit or goat anti-mouse (Biorad, Hercules, CA, USA) diluted 1/5000 in 5% non-fat milk in Tris-buffered saline/0.1% Tween 20. The immunoblots were developed with the Super Signal reagent (Pierce, Rockford, IL, USA).

Differentiation assay

PC12 cells in N2 medium were treated with flavonoids, NGF and/or other agents for 24 h at which time the cells were scored for the presence of neurites. PC12 cells produce neurites much more rapidly when treated in N2 medium than when treated in regular growth medium (Kimura and Schubert 1992). For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length were observed.

Microscopy

A light microscope (Inverted Microscope Diaphot-TMD; Nikon, Tokyo, Japan) equipped with a phase-contrast condenser (Phase contrast-2 ELWD 0.3; Nikon), 10 × objective lens and a digital camera (Coolpix 990; Nikon) was used to capture the images with the manual setting.

Protein synthesis

PC12 cells in N2 in 35-mm dishes were labeled for 30 min with 10 µCi [3H]leucine (Perkin Elmer, Boston, MA, USA). Following the labeling, the cells were rinsed with cold phosphate-buffered saline and lysed with 1 mL 10% trichloroacetic acid containing 1 mm dithiothreitol and 1 mm cold leucine. The protein precipitate was collected by centrifugation at 8000 × g for 10 min at 4°C and the pellet resuspended in 200 µL of 0.2 N NaOH. Then 100 µL of each sample was counted in a liquid scintillation counter.

Ras activity assay

The Ras activation assay kit from Upstate Biotechnology was used to assay Ras activity per the manufacturer's instructions.

GTPase assay

The ability of fisetin or other flavonoids to directly activate Ras was tested using pure recombinant wild-type p21ras (Calbiochem) in combination with a GTPase assay (Lander et al. 1993). Briefly, to 100 µL of assay buffer containing 100 mm NaCl, 12.5 mm Tris, pH 7.4, 5 mm MgCl2, 0.25 mm EGTA, 2 mm dithiothreitol, 1 mm ATP, 1 mm adenosine 5′-(β, γ-imido) triphosphate, 10 mm creatine phosphate and 5 U creatine kinase was added 1 µm wild-type p21ras, 100 000 cpm of GTP-γ-32P (5000 Ci/mmol; Perkin Elmer) and 1 µL of diluted flavonoid or dimethylsulfoxide. The samples were incubated for 10 min to 2 h at 37°C with shaking and the reaction was stopped by the addition of 0.8 mL acid charcoal [1 N HCl containing 10% (v/v) Norit A]. The samples were mixed well and then spun for 3 min at 14 000 × g and 450 µL of the supernatants were counted in a liquid scintillation counter.

Results

To determine if flavonoids can induce the differentiation of nerve cells, we utilized PC12 cells, a well-characterized model of neuronal differentiation. The cells were treated with different flavonoids using a range of concentrations and neurite outgrowth determined 24 h later. NGF was used as a positive control. As shown in Fig. 1, certain flavonoids such as fisetin were very effective at inducing neurite outgrowth giving results almost indistinguishable from those obtained with NGF. We tested the differentiation-inducing ability of a variety of flavonoids, including all of those that we had previously found to protect central nervous system-derived nerve cells (HT22 cells) from oxidative stress-induced death (Ishige et al. 2001), as well as a number of flavonoids that did not protect against oxidative stress. As shown in Table 1, we found that no flavonoids that were not protective against oxidative stress induced neurite outgrowth, but only a subset of those that were protective induced neurite outgrowth. Fisetin was the most effective flavonoid at inducing neurite outgrowth, but several other flavonoids, including luteolin, quercetin and isorhamnetin, also induced neurites in at least 50% of the cells. These data suggest certain structural requirements for the induction of neurite outgrowth, which will be addressed in the Discussion.

Figure 1.

Fisetin induces PC12 cell differentiation. PC12 cells in N2 medium were untreated or treated with 1, 5, 10 or 20 µm fisetin or NGF (50 ng/mL). The cells were examined 24 h later and photographed by phase-contrast microscopy and scored for the presence of neurites (b). For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of 10 independent experiments.

Table 1.  Induction of PC12 cell differentiation by flavonoids
FlavonoidFree hydroxyl
positions
Common
name
Differentiation
EC50m)
HT22 protection
EC50m)a
  • Half maximal effective concentrations (EC50) for differentiation were determined by exposing PC12 cells to different doses of each flavonoid and assessing differentiation as described in Materials and methods.

  • a Data for HT22 cells taken from Ishige et al. (2001).

  • b No: these compounds did not induce differentiation at 100 µm or lower concentrations.

  • c No: these compounds did not protect cells significantly from 5 mm glutamate at 50 µm or lower concentrations.

  • d

    Rutin is quercetin with the substitution of the C3 hydroxyl group.

  • e

    At 6 h, toxic with 24 h treatment.

Flavone and flavonol  nobnoc
3flavonolno6
3-methoxy nono
6 nono
6-methoxy nono
7 nono
3, 6 no6
3, 7 no6
5, 7chrysinnono
3, 5, 7galanginno6
5, 6, 7baicaleinno1
4′, 5, 7apigeninnono
3, 4′, 5, 7kaempferolno10
3′, 4′, 5, 7luteolin105
3, 3′, 4′, 7fisetin 53
3, 3′, 4′, 5, 7quercetin103
3, 3′ methoxy, 4′, 5, 7isorhamnetin107.5
3, 3′, 4′, 5, 7drutinnono
2′, 3, 4′, 5, 7morinno70
3, 3′, 4′, 5, 5′, 7myricetinnono
Isoflavone4′, 5, 7genistein50e170
Flavanone  nono
6 nono
4′, 5, 7naringeninnono
3, 3′, 4′, 5, 7taxifolinno50
Flavanol3, 3′, 4′, 5, 7catechinnono
3, 3′, 4′, 5, 7epicatechinnono
Anthocyanidin3, 3′, 4, 4′, 5, 7cyanidinnono

Several of the signaling pathways required for neurotrophic factor-induced neurite outgrowth have been elucidated. A number of studies have shown that both NGF- and fibroblast growth factor (FGF)-induced neurite outgrowth requires the sustained activation of the Ras–ERK cascade (Cowley et al. 1994; Pang et al. 1995). To determine if the flavonoids could also be working through this signaling pathway, PC12 cells were treated with 10 µm fisetin for times ranging from 5 min to 8 h and ERK activation determined by immunoblotting with an anti-phosphoERK antibody. As shown in Fig. 2(b), fisetin induced ERK phosphorylation but with a time course much different from that of NGF (Fig. 2a). Maximal activation was seen at 5–7 h after the addition of fisetin rather than the typical peak time of2.5–5 min seen with NGF. The dose dependence for ERK activation by fisetin closely parallels the dose dependence for neurite outgrowth with maximal activity seen at 10–20 µm (Fig. 2c). Luteolin, quercetin and isorhamnetin also induced ERK activation with a similar time course to fisetin, although in keeping with their reduced effect on neurite outgrowth, the level of ERK activation was significantly less than that seen with fisetin (data not shown). Only flavonoids that induced neurite outgrowth also stimulated ERK activation (data not shown). All further studies were carried out with fisetin because this was the most effective of the flavonoids at inducing both ERK activation and PC12 cell differentiation.

Figure 2.

(a) Time course of NGF-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 50 ng/mL NGF for 1 min−8 h. The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in two independent experiments. (b) Fisetin induces ERK activation. PC12 cells were untreated (ct) or treated with 10 µm fisetin for 5 min to 8 h. The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in five independent experiments. (c) Dose dependence of fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 1, 5, 10, 20, 30, 40 or 50 µm fisetin for 6 h. The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments.

To determine if ERK activation by fisetin is required for neurite outgrowth, three structurally unrelated inhibitors of mitogen-activated protein kinase/ERK kinase (MEK), the kinase immediately upstream of ERK, were used. As shown in Fig. 3(a), all of these inhibitors were very effective at blocking ERK activation by fisetin. The same concentrations were then used in the differentiation assay and, as shown in Fig. 3(b), all inhibited fisetin-induced neurite outgrowth, indicating a requirement for this pathway in fisetin-induced differentiation.

Figure 3.

(a) MEK inhibitors block fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h in the absence or presence of 50 µm PD98059 (98), 10 µm U0126 (U) or 1 µm PD184352 (18) (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (b) MEK inhibitors block fisetin-induced PC12 cell differentiation. PC12 cells in N2 medium were untreated or treated with 10 µm fisetin in the absence or presence of 50 µm PD98059, 10 µm U0126 or 1 µm PD184352 (1 h pre-treatment). After 24 h, the cells were scored for the presence of neurites. For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of three independent experiments.

To determine whether fisetin activated the entire Ras–ERK cascade or only the later part, the effect of fisetin on the activation of the upstream components of this cascade was determined. As shown in Fig. 4, fisetin induced MEK1 phosphorylation as determined by immunoblotting with a phospho-specific antibody (Fig. 4a) and Ras activation as determined by Raf-1 Ras binding domain binding (Fig. 4b). To confirm a role for Ras activation in fisetin-induced differentiation, cells were treated with the highly specific farnesyltransferase inhibitor FTI277, which blocks the farnesylation of Ras that is required for its membrane localization and activation (McGuire et al. 1996). Treatment of the PC12 cells with 10 µm FTI277 blocked fisetin-induced ERK activation (Fig. 4c) and significantly reduced fisetin-induced differentiation (Fig. 4d), indicating that Ras activation is required for the effects of fisetin on PC12 cells. In agreement with these data, expression of a dominant negative Ras mutant in PC12 cells also significantly reduced fisetin-induced ERK activation and differentiation (Figs 4e and f).

Figure 4.

(a) Fisetin induces MEK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin for 5 min−8 h. The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated MEK (anti-phospho MEK) and an antibody that detects phosphorylated and unphosphorylated MEK1 (anti-MEK). Similar results were obtained in three independent experiments. (b) Fisetin induces Ras activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin for 10 min−6 h. The cells were scraped into ice-cold lysis buffer and activated Ras was precipitated from the clarified supernatant with the Ras binding domain from Raf-1. The precipitates were analyzed by SDS–PAGE and immunoblotting with an antibody specific for Ras. (c) Ras activation is required for fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h in the absence or presence of 10 µm FTI277 (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (d) Ras activation is required for fisetin-induced PC12 cell differentiation. PC12 cells in N2 medium were untreated or treated with 10 µm fisetin in the absence or presence of 10 µm FTI277 (1 h pre-treatment). After 24 h, the cells were scored for the presence of neurites. For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of three independent experiments. (e) Dominant negative Ras inhibits ERK activation by fisetin. PC12 cells in N2 medium were untreated (– dex) or treated for 18 h with 1 µg/mL dexamethasone (+ dex) to induce the dominant negative Ras. The cells were then either left untreated (ct) or treated with 10 or 20 µm fisetin for 6 h. The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (f) Dominant negative Ras blocks fisetin-induced differentiation. PC12 cells in N2 medium were untreated (– dex) or treated for 18 h with 1 µg/mL dexamethasone (+ dex) to induce the dominant negative Ras. The PC12 cells were then either left untreated (ct) or treated with 10 µm fisetin. After 24 h, the cells were scored for the presence of neurites. For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of two independent experiments.

One possible explanation for the delayed activation of the Ras–ERK cascade by fisetin is that fisetin induces the synthesis of a protein that is required for the activation of the cascade and the accumulation of this protein up to a level that is sufficient to activate ERK takes several hours. To test this idea, PC12 cells were pre-treated with two different concentrations of CHX prior to the addition of fisetin and then ERK activation was determined 6 h later. The lower concentration (1 µg/mL) of CHX reduced protein synthesis by 88 ± 1% and the higher concentration of CHX (10 µg/mL) reduced protein synthesis by 93 ± 1%. As shown in Fig. 5(a), neither concentration of CHX had any effect on fisetin-induced ERK activation, indicating that protein synthesis is not required for the activation of this pathway by fisetin. A second possibility is that fisetin causes the release of a neurotrophic factor from the cells, which then results in the activation of the receptor for this factor and subsequent signaling to the Ras–ERK cascade. Several different growth factors can induce the differentiation of PC12 cells, including NGF and FGFs. To test the idea that FGF release was responsible for the activation of the Ras–ERK pathway by fisetin, we utilized SU5402, an inhibitor of fibroblast growth factor receptor-1 (FGFR1) signaling (Mohammadi et al. 1997). Since FGFR1 is the only FGF receptor expressed on our PC12 cells (Maher 1999) and is the FGF receptor required for PC12 cell differentiation (Lin et al. 1996), the use of this agent eliminated the need to identify which, if any, of the 20 + members of the FGF family might be released by fisetin-treated PC12 cells. Interestingly, we also found SU5402 to be a very effective inhibitor of the NGF receptor (Fig. 5b) so that we were able to test for a possible role of NGF release in the effects of fisetin on PC12 cells at the same time. PC12 cells were pre-treated with 20 µm SU5402 and then stimulated with FGF-2, NGF or fisetin for the indicated time periods and ERK activation was assessed by immunoblotting. As shown in Fig. 5(b), SU5402 efficiently blocked ERK activation by both FGF-2 and NGF but had no effect on ERK activation by fisetin. We were unable to test the effect of SU5402 on fisetin-induced differentiation because overnight treatment of the cells with this drug resulted in their death. Together, these data indicate that neither FGFs, NGF nor their receptors are required for fisetin-induced ERK activation or differentiation. In addition, the effect of fisetin on the Ras–ERK cascade requires the continuous presence of the flavonoid. A short treatment with fisetin does not result in ERK activation if the cells are harvested 6 h later (data not shown).

Figure 5.

(a) Cycloheximide does not inhibit fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h in the absence or presence of either 1 µg/mL or 10 µg/mL cycloheximide (CHX) (30 min pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (b) Inhibition of neurotrophic receptor activation does not inhibit fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h or NGF (50 ng/mL) or FGF-2 (FGF: 25 ng/mL) for 10 min in the absence or presence of 20 µg/mL SU5402 (SU) (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in two independent experiments.

A number of studies have demonstrated that protein kinase C (PKC) activation can lead to ERK activation (e.g. Schonwasser et al. 1998). In addition, the activity of both the classical and novel PKC isoforms can be modulated by certain natural products that bear some structural resemblance to the active flavonoids (Mochly-Rosen and Kauvar 1998). Furthermore, the spin trap, α-phenyl-N-tert-butylnitron (PBN), was recently shown to induce PC12 cell differentiation through an ERK-dependent pathway that required PKC activity (Tsuji et al. 2001). These results suggested the possibility that fisetin induces the Ras–ERK cascade and PC12 cell differentiation by activating one or more PKC isozymes. To test this idea, we used two different PKC inhibitors: GF109203X, a general PKC inhibitor which blocked both ERK activation and neurite outgrowth induced by PBN (Tsuji et al. 2001), and Go6983, an inhibitor of all PKC isozymes except PKCµ (Zeidman et al. 1999). Neither of these inhibitors significantly reduced fisetin-induced ERK activation (Fig. 6a and b), although the same concentrations effectively blocked TPA-induced ERK activation in the PC12 cells. In agreement with the minimal effect on fisetin-induced ERK activation, neither of these inhibitors had a significant effect on fisetin-induced neurite outgrowth (Fig. 6c). These data indicate that fisetin does not induce differentiation by activating any of the known PKC isozymes.

Figure 6.

(a, b) PKC is not required for fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h or 100 ng/mL TPA for 1 h in the absence or presence of 5 µm GF109203X (GF) or 1 µm Go6983 (Go) (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (c) PKC is not required for fisetin-induced PC12 cell differentiation. PC12 cells in N2 medium were untreated or treated with 10 µm fisetin in the absence or presence of 5 µm GF109203X or 1 µm Go6983 (1 h pre-treatment). After 24 h, the cells were scored for the presence of neurites. For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of three independent experiments.

In addition to PKC inhibitors, antioxidants also were shown to block PBN-induced ERK activation and differentiation in PC12 cells (Tsuji et al. 2001). Antioxidants were also reported to inhibit NGF-induced ERK activation and differentiation in PC12 cells (Kamata et al. 1996). Furthermore, antioxidants were found to block ERK activation induced by the phenolic antioxidants butylated hydroxyanisole and tert-butylhydroquinone, which suggested that oxidative intermediates derived from the phenolic antioxidants were responsible for their effects on the Ras–ERK cascade (Yu et al. 1997). Under certain conditions, some flavonoids or their metabolites have also been reported to exhibit pro-oxidant behavior (Cao et al. 1997; Metodiewa et al. 1999). To determine if the effects of fisetin on PC12 cells are due to a similar type of pro-oxidant behavior, cells were pre-treated with NAC, which blocked ERK activation by PBN (Tsuji et al. 2001), t-HBQ and butylated hydroxyanisole (Yu et al. 1997). In contrast to the results with these other agents, NAC failed to block fisetin-induced ERK activation or differentiation (Fig. 7). Similar results were obtained with ascorbic acid (Fig. 7), as well as with several additional antioxidants (not shown) which had proven effective against ERK activation and/or PC12 cell differentiation induced by other agents. Thus, fisetin does not appear to induce ERK activation through a pathway involving pro-oxidant behavior.

Figure 7.

(a) Antioxidants do not inhibit fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h in the absence or presence of 10 mm N-acetylcysteine (NC) or 200 µm ascorbic acid (AA) (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (b) Antioxidants do not inhibit fisetin-induced PC12 cell differentiation. PC12 cells in N2 medium were untreated or treated with 10 µm fisetin in the absence or presence of 10 mm N-acetylcysteine (NAC) or 200 µm ascorbic acid (1 h pre-treatment). After 24 h, the cells were scored for the presence of neurites. For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of three independent experiments.

The tyrosine kinase pp60c–src (Src) has also been implicated in PC12 cell differentiation (Keegan and Halegoua 1993) and Src activity can lead to ERK activation (e.g. Thandi et al. 2002). To determine if Src activation plays a role in either ERK activation or differentiation induced by treatment with fisetin, the Src kinase inhibitor SU6656 was used (Blake et al. 2000). As shown in Fig. 8, 2 µm SU6656 completely blocked both ERK activation and neurite outgrowth induced by NGF but had little or no effect on either ERK activation or differentiation induced by fisetin, indicating that Src activation does not play a role in fisetin-induced differentiation.

Figure 8.

(a) Src is not required for fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h or 50 ng/mL NGF for 10 min in the absence or presence of 2 µm SU6656 (SU) (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (b) Src is not required for fisetin-induced PC12 cell differentiation. PC12 cells in N2 medium were untreated or treated with 10 µm fisetin or 50 ng/mL NGF in the absence or presence of 2 µm SU6656 (1 h pre-treatment). After 24 h, the cells were scored for the presence of neurites. For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of three independent experiments. (c) Phase contrast photomicrographs of cells treated as in (b).

PI3K has been implicated in NGF-stimulated PC12 cell differentiation (Kimura et al. 1994; Jackson et al. 1996) and PI3K activity can lead to activation of the Ras–ERK cascade (e.g. Cussac et al. 1999). To test for a role for PI3K in fisetin-induced ERK activation and differentiation in PC12 cells, the PI3K inhibitor LY294002 was used. As shown in Fig. 9, LY294002 had little or no effect on fisetin-induced ERK activation or neurite outgrowth (Fig. 9a), although the same concentration completely blocked Akt activation by NGF and FGF-2 (Fig. 9c). These data indicate that PI3K is not involved in fisetin-induced PC12 cell differentiation.

Figure 9.

(a) PI3K activity is not required for fisetin-induced ERK activation. PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin (fis) for 6 h in the absence or presence of 10 µm LY294002 (LY) (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments. (b) PI3K activity is not required for fisetin-induced PC12 cell differentiation. PC12 cells in N2 medium were untreated or treated with 10 µm fisetin in the absence or presence of 10 µm LY294002 (1 h pre-treatment). After 24 h, the cells were scored for the presence of neurites. For each treatment, 100 cells in each of three separate fields were counted. Cells were scored positive if one or more neurites > 1 cell body diameter in length was observed. The results presented are the average ± SD of three independent experiments. (c) Inhibition of Akt activation by LY294002. PC12 cells in N2 medium were untreated (ct) or treated with 50 ng/mL NGF in the absence or presence of 10 µm LY294002 (LY) (1 h pre-treatment). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated Akt (anti-phospho Akt) and an antibody that detects phosphorylated ERKs (anti-phospho ERK). Similar results were obtained in two independent experiments. (d) PKA activity is not required for fisetin-stimulated ERK activation. A126-1B2 cells deficient in PKA activity or wild-type PC12 cells in N2 medium were untreated (ct) or treated with 10 µm fisetin for 6 h (fis), 10 µm forskolin for 10 min (fsk) or 100 ng/mL TPA for 60 min (TPA). The cells were scraped into sample buffer and analyzed by SDS–PAGE and immunoblotting with an antibody specific for phosphorylated ERKs (anti-phospho ERK) and an antibody that detects phosphorylated and unphosphorylated ERKs (anti-ERK). Similar results were obtained in three independent experiments.

Finally, protein kinase A (PKA) has also been implicated in PC12 differentiation (Schubert et al. 1978). Furthermore, some studies have suggested that that the Ras superfamily member Rap1 and the protein kinase B-Raf can link PKA stimulation to MEK1 activation in PC12 cells (Vossler et al. 1997). To address the role of PKA in fisetin-stimulated ERK activation and differentiation, we first tested the ability of several different PKA inhibitors to inhibit fisetin-stimulated ERK activation using inhibition of forskolin-stimulated ERK activation as a positive control. The results from these studies were ambiguous so we turned to the A126-1B2 variant of PC12 cells, which are deficient in PKA activity (Van Buskirk et al. 1985). These cells have been used by many laboratories to address the role of PKA activity in the responses of PC12 cells to various stimulants (e.g. Cremins et al. 1986; Bouschet et al. 2003). As shown in Fig. 9(d), we found that ERK activation by all stimuli, including TPA, was somewhat reduced in the A126-1B2 cells. However, whereas induction of ERK activation by forskolin was completely eliminated in these cells, the stimulation of ERK activation by fisetin was only slightly reduced, suggesting that PKA activation does not play a significant role in fisetin-induced ERK activation. These results are consistent with the inability of fisetin to stimulate the production of cAMP or to induce Rap1 activation (data not shown). In our hands, the A126-1B2 cells exhibited a considerable degree of differentiation, even when grown under basal conditions, so we were not able to accurately assess the ability of fisetin to induce differentiation in these cells.

Taken together, these data suggested that fisetin, as well as the other active flavonoids, could be activating the Ras–ERK cascade through a direct effect on Ras activity. Earlier studies showed that small molecules such as NO can directly activate Ras (Lander et al. 1993). To test the idea that fisetin induces the Ras–ERK cascade by directly activating Ras, the GTPase activity of untreated and flavonoid-treated wild-type recombinant p21ras was compared. As shown in Fig. 10, although the GTPase activity of wild-type p21ras increased linearly with time as reported previously (Lander et al. 1993), neither fisetin nor luteolin (data not shown) had any effect on the GTPase activity of Ras, indicating that the mechanism of activation of the Ras–ERK cascade by flavonoids does not involve a direct effect on Ras activity. It is therefore likely that fisetin activates the Ras–ERK cascade by a novel mechanism.

Figure 10.

Flavonoids cannot directly activate Ras. The GTPase activity of untreated and flavonoid-treated wild-type recombinant p21ras was measured as described in Materials and Methods. Results are from a single experiment with each point run in duplicate and are presented as the average ± SD. Similar results were obtained in three independent experiments.

Discussion

The data presented in this study demonstrate that certain flavonoids can induce the differentiation of PC12 cells and that this effect is mediated by the activation of the Ras–ERK cascade. Previously, one of us had shown that certain flavonoids can protect both a nerve cell line (HT22 cells) and primary cultures of cortical neurons from oxidative stress-induced death (Ishige et al. 2001). Several of these same flavonoids are also capable of inducing PC12 cell differentiation. However, many of the other flavonoids that were very effective at protecting the HT22 cells from oxidative stress did not induce PC12 cell differentiation, suggesting that the two activities are not directly related (Table 1). On the other hand, no flavonoids that were not neuroprotective induced differentiation. Thus, although there is no direct correlation between the two activities, an absence of effect in the oxidative stress assay appears to be indicative of an absence of effect in the differentiation assay.

What structural features distinguish those flavonoids that can induce differentiation from those that can not? Four structural determinants for differentiation activity can be deduced from the results shown in Table 1. All of the active flavonoids possess a catechol structure on the B ring (C3′ and C4′), an unsaturated C ring with an oxygen group on C4, and a hydroxyl group on C7 in the A ring. More hydroxyl groups, particularly on the B ring, appear to counteract the differentiation-inducing effect. Whether this is due to a direct effect on function or to the indirect consequences of an increase in hydrophilicity is not clear. Although the structural requirements for differentiation are similar to those for neuroprotection, they are distinct in that neuroprotection is not as dependent on a catechol structure on the B ring and a C7 hydroxyl in the A ring. The narrower specificity of structural requirements for differentiation suggests that the flavonoids may all induce differentiation through a single mechanism as opposed to the multiple mechanisms underlying protection from oxidative stress (Ishige et al. 2001).

A number of pieces of evidence indicate that the Ras–ERK cascade is essential for flavonoid-induced PC12 cell differentiation. First, all of the flavonoids that induce differentiation stimulate ERK activation in a dose-dependent manner. Second, three different MEK inhibitors block both flavonoid-induced ERK activation and differentiation. Although these same inhibitors have also been shown to block the activation of ERK5 (Mody et al. 2001), it is unlikely that ERK5 is involved in the effects of the flavonoids because PD184352 only blocks ERK5 activation at concentrations of 10 µm (Mody et al. 2001), whereas 1 µm PD184352 effectively blocked flavonoid-induced ERK activation and differentiation. Furthermore, immunoblotting with an antibody to dually phosphorylated and therefore active ERK5 revealed little or no activation by fisetin but excellent activation by NGF (data not shown). Third, both flavonoid-induced ERK activation and differentiation are blocked by both a dominant negative form of Ras and by the farnesyltransferase inhibitor FTI277, which also inhibits Ras activation, indicating that the entire pathway from Ras to ERK is required. Furthermore, since FTI277 is highly specific for farnesyl transferase, these data suggest that other small G proteins, such as Rap1 (McGuire et al. 1996; Vossler et al. 1997), which have been implicated in PC12 cell differentiation, are unlikely to be involved in the effects of the flavonoids on these cells since Rap1 must be geranylgeranylated for proper cellular localization and function (McGuire et al. 1996).

One of the more interesting characteristics of the activation of the Ras–ERK cascade by flavonoids is the lengthy treatment time required for activation. In contrast to neurotrophic factors such as NGF or FGF-2, maximal ERK activation by the flavonoids requires hours rather than minutes. These results are also in contrast to those obtained with other small molecules that have been shown to induce PC12 differentiation in an ERK-dependent manner (e.g. Satoh et al. 2001; Tsuji et al. 2001; Obara et al. 2002). In all cases, these other small molecules induce ERK activation with time courses similar to those seen with neurotrophic factors. However, the lengthy treatment time required for ERK activation by flavonoids also results in a very extended time frame for ERK activation and prolonged ERK activation is associated with differentiation (Cowley et al. 1994).

Despite the clear evidence for the Ras–ERK cascade in flavonoid-induced differentiation, how the flavonoids activate this cascade remains unclear. Several studies have shown that activation of this pathway by PKC activators requires Ras (Marais et al. 1998; Hamilton et al. 2001) and several different PKC isozymes have been implicated in PC12 differentiation (Brodie et al. 1999; Wooten et al. 1999). However, two different PKC inhibitors had no effect on flavonoid-induced ERK activation and differentiation in PC12 cells at concentrations that completely blocked TPA-stimulated ERK activation (Fig. 6). Although this is not a direct test of the ability of the inhibitors to block the activation of atypical PKCs, it does indicate that the inhibitors are active in the PC12 cells. Furthermore, concentrations of Go6983 as low as 1 µm can block atypical PKC-induced ERK activation (Cussac et al. 1999). PI3K is also implicated in both ERK activation and PC12 differentiation (Kimura et al. 1994; Jackson et al. 1996; Cussac et al. 1999). However, the PI3K inhibitor LY294002 had no effect on flavonoid-induced ERK activation or differentiation at a concentration that completely blocked NGF-induced Akt activation (Fig. 9). Furthermore, the active flavonoids did not induce Akt activation in the PC12 cells (data not shown). Similarly, Src is implicated in NGF-dependent ERK activation and differentiation in PC12 cells (Keegan and Halegoua 1993). Using the Src kinase inhibitor SU6656, we confirmed these data with NGF but did not observe any effect on fisetin-induced ERK activation or differentiation, indicating that Src is not involved in the action of fisetin on the PC12 cells (Fig. 8). Finally, PKA is also implicated in both ERK activation and PC12 differentiation (Schubert et al. 1978; Vossler et al. 1997) in PC12 cells. However, ERK activation by fisetin is only slightly reduced in a variant PC12 cell line deficient in PKA, whereas ERK activation by forskolin is completely eliminated. In addition, fisetin does not stimulate the production of cAMP, the biological activator of protein kinase A (data not shown). Furthermore, although Ras requires farnesylation for proper cellular localization and function, Rap1, which is thought to link PKA stimulation to MEK1 activation in PC12 cells (Vossler et al. 1997), is geranylgeranylated. Thus, in contrast to the results shown in Fig. 4, the farnesyl transferase inhibitor FTI277 would not be expected to block ERK activation by fisetin if Rap1 played a significant role in this activation.

Several other small molecules are known to induce PC12 differentiation, including the spin trap PBN (Tsuji et al. 2001), the sequiterpenoid β-eudesmol (Obara et al. 2002), several cyclopentenone prostaglandin derivatives (Satoh et al. 2001) and some tyrosine kinase inhibitors including the isoflavone, genistein (Bixby and Jhabvala 1992). Except for genistein, these compounds have little or no structural similarity to the active flavonoids identified in this study. Furthermore, although both PBN and β-eudesmol were both shown to induce differentiation in an ERK-dependent manner, the time course of ERK activation as well as the proposed mechanisms underlying its activation were very distinct from the results obtained with fisetin and the other active flavonoids (Tsuji et al. 2001; Obara et al. 2002). Despite the relative similarities in structure between genistein and the active flavonoids, it is unlikely that they induce nerve cell differentiation by the same mechanisms. First, genistein is a known tyrosine kinase inhibitor and the concentrations that induced differentiation also greatly reduced overall levels of protein tyrosine phosphorylation (Bixby and Jhabvala 1992). In contrast, fisetin had little or no effect on protein tyrosine phosphorylation (data not shown). Second, genistein lacks the catechol group on the B ring, which appears to be necessary for the differentiation-promoting activity of fisetin and the other flavonoids (see above).

Thus, although the mechanism underlying activation of the Ras–ERK cascade by fisetin and the other active flavonoids still remains to be identified, the data presented here clearly eliminate a number of explanations suggested by the scientific literature. Instead, all of the data suggest that fisetin activates the Ras–ERK cascade through a novel mechanism. This mechanism could involve the activation of a G protein-coupled receptor (GPCR) either directly by fisetin or indirectly through the fisetin-stimulated release of a GPCR-activating small molecule from the PC12 cells. Ras dependent activation of ERK by GPCRs has been described (for review see Gutkind 2000) and involves multiple upstream signaling pathways. Although this is an intriguing possibility, the complexity of GPCRs and their downstream signaling pathways (e.g. Thandi et al. 2002) makes exploration of this possibility beyond the scope of this study.

In summary, we have identified a select group of flavonoids that are not only neuroprotective but also promote the differentiation of nerve cells. Thus, these compounds (or derivatives thereof) might be of particular benefit in the treatment of neurological conditions that involve the loss of connections between nerve cells, thereby promoting the recovery of higher neuronal function. Furthermore, since flavonoids are low molecular weight compounds that can cross the blood–brain barrier (Abd El Mohsen et al. 2002), they could be especially useful as therapeutic agents. Although the precise molecular target of the action of flavonoids with respect of their differentiation-promoting activity remains to be determined, the Ras–ERK cascade is clearly a critical site of action.

Acknowledgements

This work was supported by National Institutes of Health grant NS28121 and funding from the Mericos-TSRI Neurobiology and Vision Science Research Program. The authors thank Dr David Schubert for critical reading of the manuscript and helpful discussions and Dr Hideo Kimura for SU5402. The authors also thank Dr Simon Halegoua for the PC12 cells expressing an inducible, dominant negative form of Ras and Dr John Wagner for the A126-1B2 cells.

Ancillary