It was reported that fisetin at low concentrations provides neuroprotection in several models [25,26,27,28,29], while at high concentrations it induces reactive oxygen species (ROS) production, ER stress and cell death in some tumor cells [30,31,32]. We found that treatment of PC12 cells with fisetin (up to 40 µM) alone for 16 h did not alter cell viability (Figure 1b). It is known that Tm blocks N-glycosylation of proteins, thus leading to ER stress [3] and ultimately apoptosis in PC12 cells [33,34]. Figure 1c shows that Tm (1–5 µg/mL) caused 30–40% PC12 cell death after 16 h. Treatment of PC12 cells with fisetin (5–20 µM) dose-dependently reversed 1 µg/mL Tm-mediated cell death (Figure 1d). The cytoprotective effects of fisetin were reconfirmed by Calcein AM viability dye staining (Figure S1) [35].

Poly (ADP-ribose) polymerase-1 (PARP-1) is specifically proteolysed by caspases to a 24 and a 89 kDa fragment during the execution of the apoptotic program [36]. Figure 2a shows that treatment of PC12 cells with fisetin (10–20 µM) in the absence of Tm for 16 h decreased PARP-1 processing in a dose-dependent manner, although it did not significantly change cell viability (Figure 1b). Tm (1 µg/mL) treatment increased PARP-1 cleavage as compared with no Tm control, and co-treatment with fisetin counteracted this reaction (Figure 2b). Band quantification shows that Tm induced a significant increase in the ratio of cleaved/full length PARP-1, and fisetin (10‒20 µM) dose-dependently attenuated Tm-mediated cleavage. This result indicates that fisetin serves as an anti-apoptotic agent in PC12 cells.

Autophagy and apoptosis are important and interconnected stress response mechanisms. Microtubule-associated protein 1 light chain 3 (LC3) is the mammalian ortholog of yeast Atg8, and is required for the formation of autophagosome membrane. The conversion of LC3β from LC3β-I (free form, 18 kDa) to LC3β-II (phosphatidylethanolamine-conjugated form, 16 kDa) is an initiating step in autophagy in mammals [37]. PC12 cells cultured in the absence of Tm did not cause the conversion of LC3β (Figure 2a). In comparison, Figure 2c shows that an increase in the LC3β-II in PC12 cells was observed in the presence of Tm (1 µg/mL), confirming that autophagy was induced by Tm. Co-treatment of cells with 10 and 20 µM fisetin dose-dependently reduced Tm-mediated increase in the ratio of LC3β-II/LC3β-I. The formation of Atg12–Atg5 conjugate is a key factor in regulating the formation of autophagosome [38]. In parallel to the results found for LC3β conversion, Tm treatment for 16 h also enhanced Atg12–Atg5 conjugate formation and co-treatment of fisetin (10 and 20 µM) blocked its formation. This result suggests that fisetin represses Tm-mediated autophagy in PC12 cells.

We further investigated the effect of fisetin on Tm-mediated ER stress response. The first target was X-box-binding protein 1 (XBP1) mRNA splicing, a critical signal transducer in the unfolded protein response. The levels of the unspliced XBP1 (XBP1u) and active spliced XBP1 (XBP1s) mRNA were measured by RT-PCR and subsequent polyacrylamide electrophoresis. It was found that Tm (1 µg/mL) treatment significantly increased XBP1 splicing, but co-treatment with fisetin (5–20 µM) did not change the relative level of XBP1s to that of XBP1u (Figure 3a). A similar phenomenon was also observed for eIF2α phosphorylation, another ER stress signal transducer upstream of ATF4. Tm (1 µg/mL) treatment markedly induced eIF2α phosphorylation, while fisetin (10–20 µM) did not change its level (Figure 3b). These results indicate that fisetin did not affect Tm-activated IRE1-XBP1 or PERK-eIF2 pathway.

We then investigated whether fisetin affected Tm-mediated ER-target gene expression. GRP78 and CHOP are canonically upregulated during apoptosis induced by ER stress. It was found that Tm treatment caused a strong increase in GRP78 mRNA expression, up by ~20-fold, while co-treatment with fisetin (10–20 µM) dose-dependently reduced its expression and 90% inhibition was noted at 20 µM (Figure 3c).

Figure 3d shows that Tm induced CHOP mRNA by ~22-fold and co-treatment with 15 and 20 µM fisetin could dose-dependently attenuate the upregulation by 77% and 88%, respectively. It has been reported that tribbles-related protein 3 (TRB3), a target gene of CHOP, was responsible for Tm-mediated PC12 apoptosis [39]. In parallel to the results found for CHOP mRNA expression, Figure 3e shows that treatment of PC12 cells with 1 µg/mL Tm increased TRB3 mRNA expression by 24.7-fold, and 15 and 20 µM fisetin could dose-dependently attenuate its upregulation. In conclusion, fisetin (10–20 µM) inhibited Tm-upregulated UPR mRNA expression in a dose-dependent manner.

Western blotting reveals that treatment of PC12 cells with Tm modestly increased GRP78 and CHOP protein expression. Fisetin dose-dependently attenuated GRP78 expression; while its effect on Tm-mediated CHOP protein expression was less prominent (Figure 3f). The less obvious inhibition on CHOP protein expression could be due to the poor detection sensitivity and specificity of antibodies against rat CHOP.

It has been reported that Tm induces apoptosis in leukemia U937 cells through ROS generation [40]. Figure 4a shows that treatment of PC12 cells with fisetin (5–20 µM) alone for 16 h could decrease ROS level by about 15–20%. Treatment of PC12 cells with Tm (1 µg/mL) for 16 h modestly increased intracellular ROS production by about 11% (Figure 4b). Fisetin (10–20 µM) and N-acetyl-cysteine (NAC) (1 mM), a well-known antioxidant, completely abolished Tm-mediated ROS overproduction to a level lower than that of vehicle control. Current data support the notion that fisetin serves as a ROS scavenger in PC12 cells [28].

Nrf2 (nuclear E2 related factor 2)-ARE (antioxidant response element) is a primary sensor and oxidative stress regulator [23]. Fisetin (10 µM) was reported to cause activation of Nrf2-ARE and upregulation of heme oxygenase-1 (HO-1), cystine/glutamate transporter (xCT/SLC7A11) and glutamate cysteine ligase (GCL) in HT22 cells [25,41]. Figure 5a shows that fisetin (5–20 µM) treatment for 6 h significantly induced HO-1 mRNA expression in an inverted U curve, and the greatest increase (22-fold) was found for 10 µM. Consequently, HO-1 protein expression was significantly stimulated after fisetin treatment (Figure 5b). Band quantification shows that a hormetic dose–response curve was noted after 8 h treatment and ~4.5-fold induction was caused by 10 and 15 µM fisetin (Figure 5c). Treatment with fisetin for 16 h induced less HO-1 protein expression than treatment for 8 h, and plateau induction (~3.7-fold) was reached by doses equal to or higher than 15 µM.

Parallel induction effects were also observed for Nrf2-driven mRNA expression of glutathione metabolism related genes, such as GCLC (catalytic subunit of GCL), GCLM (modifier subunit of GCL) and xCT (Figure 5d–f). These results suggest that fisetin induced a hormetic effect on Nrf2-driven mRNA expression of Phase II antioxidant enzymes.

We further investigated how fisetin affected Nrf2-driven gene expression in the presence of Tm. Figure 6a shows that Tm (1 µg/mL) only slightly stimulated HO-1 mRNA expression by 1.84-fold, and co-treatment with fisetin (10–20 µM) enhanced HO-1 upregulation dose-dependently. Figure 6b shows that in parallel to the results observed in mRNA expression, treatment of PC12 cells with Tm (1 µg/mL) for 16 h slightly upregulated HO-1 protein expression and fisetin enhanced HO-1 protein expression dose-dependently.

To further investigate whether the increased survival rate seen in fisetin-treated cells was dependent on HO-1 activity, PC12 cells were treated with ZnPP (zinc protoporphyrin IX), a potent competitive inhibitor of HO enzyme activity, for 30 min, followed by various concentrations of fisetin for 30 min before exposure to Tm (1 µg/mL) for 16 h. Figure 6c shows that the addition of ZnPP (0.25 and 1 µM) caused cytotoxicity on PC12 cells in all groups, and attenuated the fisetin-mediated cytoprotective effect dose-dependently. These results indicate that HO-1 activity may provide a survival advantage in PC12 cells.

We then turned to investigate how Tm affects GSH-related gene expression. The expressions of GCLC and GCLM were not affected by Tm (1 µg/mL) (Figure 6d,e). Fisetin (5–20 µM) induced GCLM expression dose-dependently (Figure 6e), while GCLC expression was impacted following a hormetic dose–response curve (Figure 6d).

It was reported that xCT expression was upregulated by both Nrf2 and ATF4 [42]. We found that, different from other Nrf2-driven genes, Tm strongly stimulated xCT expression by 10-fold. In parallel to the results observed in the absence of Tm, fisetin (5–20 µM) enhanced xCT overexpression in an inverted U response curve in the presence of Tm (Figure 6f).

Induction of Nrf2 by flavonoids has been associated with the activation of various members of the MAPK family [43]. To investigate how fisetin affects MAPK activation, PC12 cells were treated with 10, 15 and 20 μM fisetin for 30 min prior to Tm (1 µg/mL) exposure for 2 and 4 h, and cell lysates were immunoblotted with specific antibodies. Figure 7a,b show that treatment of 1 µg/mL Tm for 2 and 4 h induced ERK and JNK activation, but not p38 MAPK activation. Co-treatment with fisetin (10–20 µM) significantly augmented ERK, JNK and p38 MAPK activation with different patterns: enhancement of ERK and JNK activation was noted during the test period, while p38 MAPK phosphorylation was only found 2 h after Tm insult.

To study the cytoprotective role of the MAPK pathways, PC12 cells were pre-incubated for 30 min with 5 µM inhibitor for each pathway, U0126 (ERK), SP600125 (JNK) and SB203580 (p38 MAPK), and 5–20 µM fisetin was then added 30 min prior to Tm (1 µM) exposure for 16 h. Figure 7c,d show that addition of JNK inhibitor SP600125 and p38 MAPK specific inhibitor SB203580 significantly exacerbated Tm-mediated cytotoxicity and attenuated the cytoprotective effects of fisetin. On the other hand, MEK inhibitor U0126 did not exert any effect on cell viability in the presence of Tm (Figure S2). These results indicate that JNK and p38 MAPK signaling pathways, but not ERK, may be associated with the cytoprotective effect of fisetin.

We further investigated whether JNK and p38 MAPK are involved in fisetin-mediated oxidative stress response gene expression. Figure 8a,b show that SB203580 significantly attenuated fisetin-enhanced HO-1 and xCT expression. This implies that fisetin-activated p38 MAPK signaling may be involved in upregulation of Nrf2-driven gene expression.

Furthermore, we investigated the role of JNK and p38 MAPK on ER stress responsive gene expression. Figure 8c,d show that Tm-mediated CHOP expression was increased by both SP600125 and SB203580, but GRP78 was only enhanced by SP600125. Furthermore, the 20 µM fisetin-inhibited expressions of CHOP and GRP78 were reversed by both SP600125 and SB203580. This indicates that both fisetin-mediated JNK and p38 MAPK activation may participate in down-regulation of ER stress. This result is in agreement with a previous finding that Tm-induced p38 MAPK activation may serve as an upstream negative regulator of ER stress, and confer adaptive cytoprotection against Tm-mediated cell injury [35].

Fisetin is a SIRT1 activator that extends lifespan in lower organisms [44]. Fisetin also induces SIRT1 expression and inhibits early adipogenesis in 3T3-L1 cells [45]. We found here that Tm (1 µg/mL) treatment for 16 h significantly decreased SIRT1 expression, but co-treatment with fisetin (5–20 µM) reversed the reduction (Figure 9a,b).

To further examine the cytoprotective role of SIRT1, we examined whether the increased survival rate seen in fisetin-treated cells was dependent on SIRT1 activity. We pre-treated cells with sirtinol (SIRT1 activation inhibitor) for 30 min, followed by fisetin for 30 min before exposure to Tm (1 µg/mL) for 16 h. Figure 9c shows that the addition of sirtinol (15 µM) did not significantly cause cytotoxicity in PC12 cells in the absence of Tm. However, sirtinol completely blocked the fisetin-mediated cytoprotective effect against Tm. These results indicate that SIRT1 activity is involved in fisetin-induced cytoprotection.

Fisetin (≥98%), tunicamycin (Tm), MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), RPMI-1640 medium, and other chemicals were from Sigma-Aldrich Co. (St. Louis, MO, USA), unless otherwise indicated. 1,4-Diamino-2,3-dicyano-1,4-bis(o-aminophenylmercapto) butadiene (U0126), a selective and potent inhibitor of MEK activity and activation of ERK1/2, and 4-(4′-fluorophenyl)-2-(4′-methylsulfinylphenyl)-5-(4′-pyridyl)-imidazole (SB 203580), a p38 MAP kinase inhibitor, were purchased from Promega (Madison, WI, USA). SP600125 (1,9-pyrazoloanthron), an inhibitor of JNK, was from Calbiochem (San Diego, CA, USA). Sirtinol, a specific inhibitor of SIRT1 and SIRT2, was from Santa Cruz (Dallas, TX, USA).

The rat adrenal pheochromocytoma cell line PC12 was obtained from the Bioresource Collection and Research Center (Hsinchu, Taiwan) and maintained in RPMI-1640 medium, which contains 2 mM glutamine, 1.5 g/L sodium bicarbonate, 4.5 g/L glucose, 10 mM HEPES, 1 mM sodium pyruvate, 100 U/mL penicillin and streptomycin, supplemented with 10% heat-inactivated horse serum (Hyclone, Logan, UT, USA) and 5% fetal bovine serum (Invitrogen, Carlsbad, CA, USA) in 5% CO₂ incubator at 37 °C.

For experiments testing the ability of tunicamycin (Tm) to induce cytotoxicity, cells were incubated in serum-free RPMI medium in the presence of Tm [34]. PC12 cells (5 × 10⁵/mL) were seeded in 24-well plates and pretreated with the indicated concentration of fisetin or an equivalent volume of DMSO vehicle control (final concentration of 0.1%) for 30 min, followed by Tm treatment for 16 h.

To measure the cell viability of adherent PC12 cells that had undergone Tm-induced damage, PC12 cells (5 × 10⁵/mL) were seeded on poly-l-lysine-coated 6-well plates in low serum (0.5% fetal bovine serum and 1% horse serum) medium for 16 h. Fisetin or an equivalent volume of DMSO vehicle (final concentration of 0.1%) was then added and incubated for 30 min prior to Tm treatment for an additional 16 h.

A cell-free blank with medium and tested reagent was employed in parallel. Cell viability was assessed by the mitochondrial-dependent reduction of MTT to purple formazan [58]. The cell viability was calculated by subtracting the OD₅₅₀ of cell-free blank from OD₅₅₀ of each sample and was expressed as percentage of the control (100%).

An esterase-dependent cell viability analysis Calcein AM (Invitrogen) was used for reconfirmation of MTT data [59]. Briefly, cells were incubated with 5 µM Calcein AM for 30 min at 37 °C, and the fluorescent signal was monitored using 485 nm excitation and 530 nm emission wavelengths.

RIPA buffer (Thermo Fisher Scientific, Inc., Rockford, IL, USA) was used for preparation of whole cell extracts and the protein concentration was measured by the Bradford method (Bio-Rad Laboratories, Hercules, CA, USA). Equal amounts of cell lysates were separated on SDS-PAGE and then transferred onto Hybond-P PVDF (GE Healthcare, Buckinghamshire, UK) using a CAPS transfer buffer at 30 mA overnight at 4 °C. The membranes were blocked in a freshly made blocking buffer (5% skim milk in PBS with 0.05% Tween 20, pH 7.4, PBS-T) for 6 h at room temperature. After washing with PBS-T, the membranes were incubated with an appropriate dilution (1:1000‒1:5000) of primary antibody (Table 1) overnight at 4 °C on a rocking platform. The membranes were then washed and incubated with suitable horseradish peroxidase-conjugated secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA, at a dilution of 1:10,000‒1:25,000) for 1 h at room temperature. The blots of were incubated by ECL Prime (GE Healthcare), and the chemiluminescent signals were then visualized with X-ray film. Densitometry of the bands was analyzed by ImageJ software version 1.50 (National Institutes of Health, Bethesda, MD, USA).

Illustra RNAspin Mini RNA Isolation Kit (GE Healthcare) was used for preparation of total RNA. High-Capacity cDNA Archive Kit (Applied Biosystems, Waltham, MA, USA) was used to prepare cDNA from 1 µg RNA. Real-time PCR (StepOne Real-Time PCR System, Applied Biosystems) was performed with 2 µL of the cDNA, 200 nM primers (Table 2) and Fast SYBR Green Master Mix (Applied Biosystems) in 25 µL reaction mixture. The amplification conditions were as follows: 95 °C for 2 min, 40 cycles at 94 °C for 15 s, and 60 °C for 60 s. Target gene expression was measured and normalized to the respective β-actin expression level. The identity and purity of the amplified product was checked through analysis of the melting curve carried out at the end of amplification. Relative expression was evaluated with the ΔΔC_t method.

XBP1s and XBP1u mRNA levels were measured using regular PCR as described in our previous publication [60].

Cellular reactive oxygen species were analyzed using the fluorescence probe 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) (Invitrogen), which passively diffuses into the cell and is cleaved and oxidized to 2′,7′-dichlorofluorescein (DCF). PC12 cells were stained with 20 µM H₂DCFDA at 37 °C for 30 min in the dark and then washed once in PBS. Fluorescence dye loaded cells were seeded in black 96-well plates (2 × 10⁵/well) and treated with fisetin or in combination with Tm for 16 h. The signals were then read at EX485 nm/Em535 nm using a fluorometer.

All experiments were repeated at least three times. The results were analyzed using Kruskal–Wallis H Test by SPSS version 18. If the Kruskal–Wallis H Test shows a significant difference between the groups, then pairwise comparisons were employed by Mann–Whitney U Tests, and a p-value of <0.05 was taken to be significant.