Sequential Upregulation of Superoxide Dismutase 2 and Heme Oxygenase 1 by tert-Butylhydroquinone Protects Mitochondria during Oxidative Stress

Cell Viability Assay.

HT-22 cells were seeded in 96-well or 6-well plates (Corning) and were incubated overnight. After respective treatments, medium was removed and cells were incubated with 1 μM calcein–acetoxymethyl ester (calcein-AM) (Molecular Probes, Grand Island, NY) in phosphate-buffered saline (PBS) for 15 minutes at 37°C. Calcein-AM, a nonfluorescent dye, is converted to a green fluorescent calcein by intracellular esterases. Fluorescence was measured using the BioTek Synergy H1 Hybrid plate reader (BioTek, Winooski, VT; excitation, 495 nm; emission, 516 nm). Calcein-AM led to the detachment of cells from the bottom of the wells after small interfering RNA (siRNA) transfection, an effect that compromises the assay. As such, to measure cell viability of siRNA-transfected cells, we used morphologic changes of cells after respective exposures as observed microscopically. For each well, pictures were randomly taken in three different fields. Based on the morphology, cells in each photomicrograph were counted and categorized into live cells and dead cells by UVP (Upland, CA) imaging software. The average cell number from three different pictures was calculated to represent cell viability. For the primary cortical neurons, viability of the cells was assessed using calcein-AM and imaged using fluorescent microscopy.

ROS Detection.

Changes in intracellular ROS were measured by the ROS-reactive fluorescent indicator 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) (Molecular Probes). The nonfluorescent H₂DCFDA is converted to the highly fluorescent dichlorofluorescein (DCF) by ROS (Fukui et al., 2010; Tobaben et al., 2011; Kang et al., 2014). Briefly, HT-22 cells were plated overnight at a density of 5000 cells/well in a 96-well plate. After respective exposures, the medium was removed and the cells were washed once with PBS and then incubated with 10 μM H₂DCFDA for 30 minutes at 37°C. Mean fluorescence intensity of DCF was measured using the BioTek Synergy H1 Hybrid plate reader (excitation, 485 nm; emission, 530 nm). DCF fluorescence was standardized based on cell viability.

Mitochondrial Superoxide Measurement.

MitoSOX Red (Molecular Probes) is a fluorogenic dye targeted to mitochondria and generates red fluorescence after oxidation by superoxide (Mukhopadhyay et al., 2007; Fukui and Zhu, 2010; Pfeiffer et al., 2014). The fluorescence signal was measured by fluorescence-activated cell sorting (FACS) analysis. Cells were harvested, washed once with ice-cold PBS, and stained with 5 μM MitoSOX Red in Hanks' balanced salt solution for 10 minutes at 37°C. Cells were then washed twice with PBS before the red fluorescence intensity was analyzed using a flow cytometer (BD FACSCalibur; BD Biosciences, San Jose, CA). In each analysis, 10,000 events were recorded.

Mitochondrial Membrane Potential Analysis.

HT-22 cells were plated at a density of 5000 cells/well and exposed to glutamate alone or in combination with tBHQ. The medium was then removed, and cells were incubated in PBS containing 1 μM nonylacridine orange (NAO) (Molecular Probes) and 1 μM tetramethylrhodamine, ethyl ester (TMRE) (Sigma-Aldrich) for 20 minutes at 37°C. Under normal mitochondrial intermembrane potential, TMRE enters into mitochondria and quenches NAO. Collapse of mitochondrial membrane potential promotes NAO fluorescence. NAO fluorescence was measured using the BioTek Synergy H1 Hybrid plate reader (excitation, 485 nm; emission, 530 nm) and standardized based on cell viability.

Mitochondrial Ca²⁺ Detection.

Mitochondrial Ca²⁺ was measured using Rhod-2 AM, a fluorogenic dye specifically targeted to mitochondrial Rhod-2 AM (Molecular Probes), which exhibits fluorescence upon binding Ca²⁺. Cells were incubated with 2 μM Rhod-2 AM for 15 minutes at 37°C, washed with PBS twice, and analyzed immediately by flow cytometry (BD FACSCalibur). In each analysis, 10,000 events were recorded.

Mitochondrial Respiration Measurement.

HT-22 cells were plated at a density of 15,000/well in an XF^e96 plate (Seahorse Bioscience, North Billerica, MA). After respective exposures, the medium was exchanged 1 hour prior to the assay with XF assay medium (Seahorse Bioscience). Oligomycin (1 μM), carbonilcyanide p-triflouromethoxyphenylhydrazone (FCCP; 0.5 μM), and antimycin and rotenone mixture (1 μM) (Sigma-Aldrich) were diluted into XF^e96 medium and loaded into the accompanying cartridge. Injections of the components into the wells occurred at the time points specified. Oxygen consumption rate (OCR) was monitored using a Seahorse Bioscience XF^e96 Extracellular Flux Analyzer.

Caspase-3/7 Activity.

Caspase-3 and -7 activities were measured using a luminescence-based assay, Caspase-Glo 3/7 Assay (Promega, Madison, WI). According to the manufacturer’s protocol, cells were incubated with proluminescent caspase-3/7 substrate for 1 hour at room temperature. Following caspase cleavage, a substrate for luciferase is released, resulting in the luciferase reaction and the production of light. Luminescence was measured with the BioTek Synergy H1 Hybrid plate reader.

Immunocytochemistry.

HT-22 cells were fixed with 4% paraformaldehyde (Sigma-Aldrich) in PBS for 15 minutes after respective exposures. The cells were permeabilized with 0.25% Triton X-100 for 10 minutes and were then incubated in 10% serum (Sigma-Aldrich) blocking solution containing 0.3 M glycine for 30 minutes. Cells were exposed to anti-AIF antibody or anti-Nrf2 antibody (1:100 in blocking solution; Santa Cruz Biotechnology, Dallas, TX) overnight at 4°C, followed by incubation with appropriate fluorescence-conjugated secondary antibodies (Molecular Probes) for 1 hour. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (Molecular Probes). Images were acquired using a fluorescence confocal microscope (Zeiss Violet Confocal; Zeiss, Oberkochen, Germany) with a 40× objective.

Nuclear Isolation.

Cells were collected and resuspended in lysis buffer A (10 mM HEPES, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol, 0.05% NP40; pH 7.9). Samples were left on ice for 10 minutes and centrifuged (Allegra 64R centrifuge; Beckman Coulter, Irving, TX) at 4°C at 3000 rpm for 10 minutes. Supernatant was removed; cell pellets were resuspended in lysis buffer B (5 mM HEPES, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM dithiothreitol, 26% glycerol, 300 mM NaCl; pH 7.9) and homogenized with 20 full strokes in a glass homogenizer on ice. All the chemicals were ordered from Sigma-Aldrich. After the 30-minute incubation on ice, samples were centrifuged at 24,000g for 20 minutes at 4°C. Supernatants were collected as the nuclear fraction.

Western Blot.

For whole-cell lysis, cells were lysed in radioimmunoprecipitation assay buffer with cocktail protease inhibitors (EMD Millipore, Billerica, MA). Briefly, blots were probed with anti-AIF antibody (1:1000 dilution; Santa Cruz Biotechnology), anti-SOD2 antibody (1:1000 dilution; Santa Cruz Biotechnology), or anti–HO-1 antibody (1:1000 dilution; Abcam, Cambridge, MA) at 4°C overnight. Membranes were then exposed to the appropriate horseradish peroxidase–conjugated secondary antibodies (Santa Cruz Biotechnology), followed by chemiluminescence detection (Fisher, Waltham, MA) of antibody binding. Equal protein loading was controlled by reprobing the membrane with anti–β-actin antibody or an anti–histone deacetylase 1 antibody (1:1000 dilution; Santa Cruz Biotechnology). Chemiluminescence was detected using the UVP ChemiDoc-It TS2 Imager, and UVP software was used for quantification of Western blot signals.

GSH Measurement.

GSH level was measured using a luminescence-based assay, GSH-Glo Glutathione Assay (Promega). The assay was based on the conversion of a luciferin derivative into luciferin in the presence of GSH, catalyzed by glutathione S-transferase. According to the manufacturer’s protocol, cells were incubated with GSH-Glo reagent for 30 minutes at room temperature, followed by a 15-minute incubation with luciferin detection reagent, and luminescence was measured with the BioTek Synergy H1 Hybrid plate reader.

Transfection.

HT-22 cells were seeded at 4 × 10⁴ cells/well in 6-well plates at the time of transfection. siRNA selectively targeting mouse SOD2 (Santa Cruz Biotechnology) was used for transfection, and a scrambled nontargeting siRNA was used as the control. Transfections of the siRNA targeting SOD2 (50 pmol) or the scrambled control siRNA were performed using a siRNA transfection reagent (Santa Cruz Biotechnology) based on the protocol provided by the manufacturer. Transfection efficiency was observed microscopically, and SOD2 protein expression was determined by Western blot analysis after 24- or 36-hour transfection.

tBHQ Prevents Glutamate-Induced ROS and Mitochondrial Superoxide Generation.

Because ROS accumulation is a hallmark of glutamate-induced cell death in HT-22 cells (Tan et al., 1998), we determined whether tBHQ inhibited glutamate-induced intracellular ROS accumulation using an ROS-sensitive fluorescence indicator, H₂DCFDA. The accumulation of ROS was elevated by 2-fold after a 10-hour exposure to glutamate. Application of tBHQ at 0 and 6 hours after glutamate abrogated this ROS accumulation (Fig. 2A). A previous study revealed that glutamate also induced an increase in superoxide level in the mitochondria (Fukui et al., 2012). We then measured the mitochondrial superoxide level using the mitochondria-specific superoxide indicator MitoSOX. Using FACS analysis, we observed that glutamate induced a 2-fold increase of mitochondrial superoxide production. Even with a 6-hour exposure delay, tBHQ attenuated the accumulation of mitochondrial superoxide (Fig. 2, B and C).

Open in a separate window

Fig. 2.

tBHQ prevents glutamate-induced ROS and mitochondrial superoxide in HT-22 cells. (A) HT-22 cells were treated with glutamate (Glut) (5 mM) for 10 hours. tBHQ (10 μM) was applied at 0 or 6 hours after glutamate exposure. ROS levels were detected by H₂DCF, and fluorescence was measured by plate reader (n = 8). MFI, mean fluorescence intensity. (B) tBHQ (10 μM) was applied at 0 or 6 hours after glutamate (5 mM) exposure. Mitochondrial ROS was measured by MitoSOX after an 11-hour treatment and quantified by FACS analysis (n = 3). (C) Overlay of the FACS tracings for HT-22 cells stained with MitoSOX. All experiments were repeated three times, and the results indicate the mean ± S.E.M. *P < 0.05; **P < 0.01; ***P < 0.001 compared with glutamate-treated cells (one-way analysis of variance, Tukey’s test).

tBHQ Prevents Mitochondrial Membrane Potential Disruption and Mitochondrial Calcium Overload Induced by Glutamate.

It is well known that glutamate-induced oxidative damage causes the impairment of mitochondrial membrane potential and an increase of mitochondrial Ca²⁺ influx (Chen et al., 2003; Fukui et al., 2009). Thus, we evaluated the role of tBHQ in glutamate-compromised mitochondrial membrane potential using the NAO/TMRE fluorescence resonance energy transfer assay, and we measured mitochondrial calcium levels using the fluorescent indicator Rhod-2 AM. Glutamate exposure induced a collapse of the mitochondrial intermembrane potential (Fig. 3A) and a mitochondrial Ca²⁺ overload (Fig. 3, B and C). Both simultaneous and delayed tBHQ exposures attenuated glutamate-induced mitochondrial membrane potential reduction and the increase in mitochondrial calcium levels. tBHQ alone did not have an appreciable effect on mitochondrial Ca²⁺ dynamic.

Open in a separate window

Fig. 3.

tBHQ prevents glutamate-induced mitochondrial membrane potential disruption and mitochondrial Ca²⁺ overload in HT-22 cells. (A) Cells were treated with 5 mM glutamate (Glut). tBHQ (10 μM) was applied simultaneously or 6 hours after glutamate treatment. After an 11-hour treatment with glutamate, mitochondrial membrane potential was measured by NAO/TMRE assay. Fluorescence of NAO was measured by a plate reader (n = 8). (B) Cells were treated with glutamate (5 mM), and tBHQ (10 μM) was added at either 0 or 6 hours after glutamate. Mitochondrial Ca²⁺ was measured by Rhod-2 AM after an 11-hour treatment and quantified by FACS analysis (n = 3). (C) Histogram overlay representing the Rhod-2 levels in HT-22 cells. All experiments were repeated three times, and the results indicate the mean ± S.E.M. **P < 0.01; ***P < 0.001 compared with glutamate-treated cells (one-way analysis of variance, Tukey’s test).

tBHQ Attenuates the Exacerbation of Mitochondrial Respiration under Glutamate Toxicity.

Mitochondrial respiration deficiency is a key index of mitochondrial failure (Lin and Beal, 2006). To determine the effect of glutamate and tBHQ exposure on mitochondrial respiration, we measured OCR using a Seahorse XF^e96 analyzer (Fig. 4A). Based on the OCR after application of stimuli, four parameters were calculated to evaluate mitochondrial respiration. Glutamate alone led to a 60% reduction in ATP production–linked respiration and fully abolished maximal respiration and spare capacity; cotreatment with tBHQ ameliorated these effects of glutamate (Fig. 4, B–D). Proton leak was not affected by glutamate or tBHQ (Fig. 4E).

Open in a separate window

Fig. 4.

tBHQ prevents the impairment of mitochondrial metabolism induced by glutamate. After a 12-hour treatment with 10 mM glutamate (Glut) and 10 μM tBHQ, OCR was recorded by a Seahorse XF^e96 flux analyzer (n = 8). (A) OCR recording at baseline and subsequent to treatment with 1 μM oligomycin, 0.5 μM FCCP, and a 1 μM rotenone and antimycin mixture. ATP production (B), spare capacity (C), maximum respiration (D), and proton leak (E) were calculated. All experiments were repeated three times, and the results indicate the mean ± S.E.M. ***P < 0.001 compared with glutamate-treated cells (one-way analysis of variance, Tukey’s test).

tBHQ Blocks Mitochondria-Mediated Apoptosis under Glutamate Toxicity.

Glutamate-induced oxidative stress triggers mitochondria-mediated apoptosis by activating caspase-3/7 and AIF translocation to the nucleus (Landshamer et al., 2008; Fukui et al., 2009). We explored the regulation by tBHQ of glutamate-induced mitochondria-mediated apoptosis. Caspase-3/7 was not activated upon exposure of HT-22 cells to glutamate (Fig. 5A), whereas the AIF level in the nuclear fraction was increased by almost 3-fold after 16 hours of exposure to glutamate. tBHQ coexposure inhibited glutamate-induced AIF translocation to the nucleus (Fig. 5B). These findings were confirmed by immunocytochemistry. There was a clear translocation of AIF (Fig. 5C, red) into the nucleus (Fig. 5C, blue) at 12 hours after glutamate exposure; cotreatment with tBHQ prevented AIF translocation and preserved cell morphology.

Open in a separate window

Fig. 5.

tBHQ inhibits glutamate-induced, mitochondrial AIF–mediated apoptosis. (A) After an 11-hour treatment with glutamate (Glut) (5 and 10 mM) and tBHQ (10 μM), caspase-3/7 activities were measured (n = 3). (B) After a 16-hour treatment with 15 mM glutamate and 10 μM tBHQ, AIF level in nuclear fraction was measured by Western blot. Quantitation of AIF was normalized to histone deacetylase 1 (HDAC₁). Bars represent normalized relative densities plotted as mean ± S.E.M. calculated from four independent experiments. (C) Immunocytochemistry for AIF (red) and 4′,6-diamidino-2-phenylindole (DAPI; blue). Images were captured at a 12-hour exposure time to 5 mM glutamate and 10 μM tBHQ by confocal microscopy. All experiments were repeated at least three times, and the results indicate the mean ± S.E.M. ***P < 0.001 compared with glutamate-treated cells (one-way analysis of variance, Tukey’s test).

tBHQ Is Comparatively Ineffective against Electron Transport Chain Blockage–Induced Toxicity.

In view of the ability of tBHQ to prevent mitochondrial dysfunction in response to glutamate, we determined if tBHQ protected cells from mitochondria-specific toxins. Exposure to oligomycin (20 μM), an ATP synthase inhibitor, increased cell death by ∼40%, and this effect was partially rescued by tBHQ exposure (Fig. 6A). FCCP, a protonophore, disrupts the mitochondrial proton gradient by transporting protons across the membrane. Exposure of cells to FCCP (10 μM) alone reduced cell viability to 20% of control 24 hours after exposure; coexposure with tBHQ exerted moderate protection (Fig. 6B). Rotenone reduces oxidative phosphorylation by inhibition of mitochondrial complex I activity. Exposure of cells to rotenone (10 μM) induced 20% cell death, but tBHQ did not protect cells from rotenone-induced cytotoxicity in HT-22 cells (Fig. 6C). Compared with the efficacy of tBHQ’s protection against glutamate-induced cytotoxicity, we conclude that tBHQ is comparatively ineffective against electron transport chain (ETC) blockage–induced toxicity in HT-22 cells.

Open in a separate window

Fig. 6.

tBHQ is comparatively ineffective against mitochondrial ETC blockage–induced toxicity. HT-22 cells were treated with 10 μM oligomycin (Oligo) (A), 10 μM FCCP (B), or 10 μM rotenone (Rote) (C), plus 1–10 μM tBHQ for 24 hours. Cell viability was measured by calcein-AM assay. Each experiment was repeated at least three times. Results are reported as mean ± S.E.M. *P < 0.05; ***P < 0.001 compared with ETC stimuli (oligomycin, FCCP, rotenone) treatment–only group (one-way analysis of variance, Tukey’s test).

tBHQ Causes a Rapid Upregulation of SOD2 Expression and a Delayed Upregulation of HO-1 Expression.

The neuroprotective treatment of tBHQ can be delayed up to 8 hours, the time before cells lose normal cell morphology, and we observed cell death after 11 hours of glutamate exposure (Supplemental Fig. 2). These data indicate that the protection of tBHQ is initiated within 3 hours after application. Next, we investigated the underlying mechanisms accounting for the time course of this protection. Because GSH depletion is the primary cause of glutamate-induced toxicity, we first measured the GSH level after tBHQ and glutamate exposure. Glutamate induced a profound reduction in GSH synthesis, which was not ameliorated by tBHQ. However, tBHQ alone increased the intracellular GSH level by 20% (Fig. 7A). tBHQ, a phenolic compound, has been reported to be a free radical scavenger (Fig. 7B) (Alamed et al., 2009). We then determined if tBHQ served as an antioxidant by scavenging ROS. Within 30 minutes of exposure, tBHQ prevented H₂O₂-induced intracellular ROS accumulation (Fig. 7C). These data indicate that tBHQ does not prevent the reduction of GSH level induced by glutamate, but does function as a free radical scavenger to eliminate glutamate-induced oxidative stress. Further, we examined if tBHQ's protective effect worked through its free radical–scavenging activity. HT-22 cells were pretreated with tBHQ for 12 hours, and then cultures were incubated with glutamate in fresh medium without tBHQ. After 24-hour exposure to treatments, cell viability was measured. As shown in Fig. 7D, the protective effect was observed 24 hours after removal of tBHQ. This indicates that the protection of tBHQ is not due to a radical-scavenging effect for this chemical. Morphologic changes after tBHQ and glutamate exposure were observed microscopically (Fig. 7E).

Open in a separate window

Fig. 7.

tBHQ fails to block glutamate-induced GHS depletion, but scavenges intracellular ROS. (A) GSH level was detected at 8 hours after treatment with 5 mM glutamate (Glut) and 10 μM tBHQ (n = 5). (B) Chemical structure of tBHQ. (C) HT-22 cells were treated with H₂O₂ (12.5–100 μM) and tBHQ (10 μM) from 30 minutes. ROS levels were detected by H₂DCF, and fluorescence was measured by plate reader (n = 8). MFI, mean fluorescence intensity. (D) HT-22 cells were pretreated with tBHQ for 12 hours, and then cultures were incubated in glutamate in fresh medium without tBHQ. After 24 hours of exposure to treatments, cell viability was measured by calcein-AM assay (n = 8). Simultaneous treatment with tBHQ was used as a positive control. (E) Morphologic changes of cells were observed microscopically. Results are reported as mean ± S.E.M. **P < 0.01; ***P < 0.001 compared with glutamate treatment–only group (one-way analysis of variance, Tukey’s test).

Next, we investigated if activation of Nrf2 and its regulated gene expression by tBHQ contributes to its protection. There was a clear translocation of Nrf2 (Fig. 8A, red) into the nucleus (Fig. 8A, blue) at 1.5 hours after tBHQ exposure; this colocalization was diminished with prolonged tBHQ exposure. HO-1 has been reported as an important target to prevent glutamate-induced oxidative damage in HT-22 cells (Satoh et al., 2003; Rössler et al., 2004). Therefore, we monitored the time-dependent change in HO-1 protein expression following tBHQ exposure. There was no significant change in HO-1 level within 3 hours of exposure of tBHQ; however, a 30-fold increase in HO-1 expression was observed at 12 hours after tBHQ treatment (Fig. 8B). Previous studies have demonstrated that the Nrf2-ARE signaling pathway regulates SOD2 expression (Dong et al., 2008; Piantadosi et al., 2008; Yan et al., 2010), and SOD2 plays a critical role in protecting HT-22 cells against glutamate-mediated cytotoxicity (Stocker et al., 1987). Our data showed a 2-fold increase in SOD2 expression after 3 hours of exposure to tBHQ (Fig. 8C), while SOD2 level returned to normal with prolonged tBHQ exposure (unpublished data). As such, we asked if this rapid upregulation of SOD2 expression is a key factor contributing to the protection offered by tBHQ.

Open in a separate window

Fig. 8.

tBHQ induces a rapid increase of SOD2 expression followed by a delayed upregulation of HO-1 expression. HT-22 cells were treated with 10 μM tBHQ. (A) Immunocytochemistry for Nrf2 (red) and 4′,6-diamidino-2-phenylindole (DAPI; blue). Images were captured at 1.5, 3, 6, 9, and 12 hours of exposure to tBHQ by confocal microscopy. Samples were also collected at 1.5, 3, 12, and 24 hours after tBHQ exposure. Cell extracts were subjected to immunoblot with antibodies specific for HO-1 (B) and SOD2 (C). Quantitation of HO-1 and SOD2 was normalized to β-actin. Bars represent normalized relative densities plotted as mean ± S.E.M. calculated from four independent blots. **P < 0.01; ***P < 0.001 compared with control group (one-way analysis of variance, Tukey’s test).

Delayed Treatment with tBHQ Fails To Prevent Glutamate-Induced Cell Death in SOD2-Knockdown HT-22 Cells.

To characterize the role of SOD2 in tBHQ-mediated protection, we transfected HT-22 cells with siRNA targeting SOD2. After 24- or 36-hour exposure to siRNA, a transfection efficiency of >90% was achieved. SOD2 expression was reduced by 45% after 24-hour transfection (Supplemental Fig. 3). As shown in Fig. 9A, SOD2 protein level was reduced by 65% at 36 hours after transfection, the time we selected for the following study. We compared the protection efficacy of tBHQ in scrambled siRNA–transfected and SOD2-knockdown HT-22 cells. After 18 hours of treatment with glutamate, cell morphologic changes were observed (Fig. 9B). In SOD2-knockdown HT-22 cells, simultaneous treatment with tBHQ was still able to protect cells from glutamate toxicity. However, silencing SOD2 attenuated the protective effect of delayed tBHQ exposure (Fig. 9C).

Open in a separate window

Fig. 9.

Delayed treatment with tBHQ fails to overcome glutamate-induced cell death in a SOD2-knockdown HT-22 cell line. (A) HT-22 cells were transfected with scrambled and SOD2 siRNA for 36 hours. Transfection efficiency was measured by Western blot with antibody specific for SOD2. β-Actin was used to normalize loading. Bars represent normalized relative densities plotted as mean ± S.D. calculated from three independent blots (one-way analysis of variance [ANOVA], Tukey’s test). (B) Both scrambled and SOD2 siRNA–transfected HT-22 cells were treated with 5 mM glutamate (Glut). tBHQ was applied either simultaneously or 8 hours after glutamate exposure. Morphologic changes of cells after respective treatments were observed microscopically. (C) Based on the morphology, cells in each photomicrograph were counted and calculated to represent cell viability. Experiments were repeated three times independently. Results are reported as mean ± S.E.M. ***P < 0.001 compared with glutamate treatment–only group (two-way ANOVA, Bonferroni’s test).