Assessing the influence of curcumin in sex-specific oxidative stress, survival and behavior in Drosophila melanogaster

Oxidative stress, which is the accumulation of reactive oxygen and nitrogen species (RONS), occurs as a result of a mismatch of endogenous and exogenous antioxidant pathway activities and oxidative load. Over-abundance of intracellular RONS creates a toxic environment, ultimately leading to denaturation and degradation of proteins and DNA (Liu et al., 2012), lipid peroxidation, tissue damage, and, consequently, cellular apoptosis and cytotoxicity (Nguyen et al., 2018). The oxidative stress hypothesis suggests that accumulation of RONS underlies functional losses associated with aging, such as sarcopenia and neuronal loss (reviewed in Sohal et al., 2002). The high lipid content of the brain makes this organ particularly vulnerable to lipid peroxidation, one of the primary damaging pathways of RONS.

Sexual dimorphism of oxidative stress and endogenous antioxidants may underlie sex differences in aging and age-related pathologies (Camus et al., 2012; Ruszkiewicz et al., 2019). Human females exhibit a more robust antioxidant response, consistent with greater longevity and decreased risk of age-related diseases (Bellanti et al., 2013; Signorelli et al., 2006; Viña and Borrás, 2010). In Drosophila melanogaster and mice, preservation of mitochondrial functionality and longevity as an adaptive response to oxidative stress is female specific (Pomatto et al., 2017; Valle et al., 2007). Drosophila provide a well-established model for investigating the sex-specific impacts of antioxidant treatment on longevity, oxidative stress and neural function. Sex differences in multiple physiological processes, including metabolic and neural functions, are well documented in Drosophila (Chen et al., 2017; Lnenicka et al., 2006; Pomatto et al., 2017).

Curcumin, a yellow pigment derived from turmeric, demonstrates promise as a neuroprotectant due to its ability to alleviate oxidative stress through RONS scavenging and interaction with central nervous system cells and processes in multiple model systems (Han et al., 2012; Jaroonwitchawan et al., 2017). Recent evidence suggests extensive involvement in regulation of endogenous antioxidant pathways by manipulation of genetic elements (Han et al., 2012; Huang et al., 2018). Pre-treatment with curcumin decreases many biomarkers of oxidative damage after exposure of Drosophila and mice to oxidative stressors (Chen et al., 2017; Seong et al., 2015; Strong et al., 2013). Administering sub-toxic doses of curcumin in neuronal and somatic human cell lines alleviates the effects of paraquat-induced oxidative stress by activation of autophagy-related pathways (Daverey and Agrawal, 2016; Han et al., 2012).

Here, we assessed the sex-specific role of curcuminoids (curcumin and its derivatives demethoxycurcumin and bisdemethoxycurcumin) in oxidative stress, development and behavior in wild-type Drosophila. We confirmed that treatment with dietary curcuminoids reduced peroxidated lipids, extended longevity and demonstrated sex-specific differences in the impact of curcuminoids on motility and oxidative stress in Drosophila.

All Drosophila melanogaster were maintained at 25°C in a 12 h:12 h light:dark cycle at 60% humidity. All experiments were conducted on the Oregon R strain, which was reared on Nutri-Fly German formulation (GF) medium, prepared according to the manufacturer's directions (Genessee Scientific, San Diego, CA, USA). Curcuminoid treatments were accomplished by inclusion of the appropriate amount of curcuminoids, derived from Curcuma longa (C1386, Sigma-Aldrich, St Louis, MO, USA; 98% pure standard consisting of curcumin, demethoxycurcumin and bisdemethoxycurcumin and referred to as curcuminoids hereafter) to generate media with 0.25, 2.5 and 25 mmol l⁻¹ curcuminoids. No standard concentration of curcumin or curcuminoids has been developed in Drosophila, so this broad range of curcuminoid concentrations was chosen to encompass the association properties of many potential binding partners in diverse tissues.

The rate of gut clearance of third instar larvae was determined as previously described with minor modifications (Shell et al., 2018). Briefly, Oregon R embryos were reared to eclosion on standard Nutri-Fly GF medium and on medium supplemented with Blue 1 dye. Wandering third instar larvae from each medium were collected while minimizing the transfer of media components and soaked in phosphate-buffered saline (PBS) for 30 min intervals up to 1.5 h. After each interval, five groups of ten larvae were removed from the PBS, blotted dry on Kimwipes to remove liquid and any debris, and each group of ten larvae was saved in a 1.5 ml Eppendorf tube. Aliquots of 100 μl PBS were added to each tube and the larvae were homogenized with a Kontes motor and pellet pestle (Kimble, Rockwood, TN, USA). The homogenate was centrifuged at maximum speed for 10 min (∼14,000 g), the supernatant was filtered through a 0.22 μm filter, and the absorbance at 625 nm was measured on a BioSpectrometer kinetic photometer (Eppendorf, Hamburg, Germany). Each set of five homogenates was measured in triplicate.

Liquid chromatography grade water, acetonitrile, methanol and high-purity formic acid were purchased from Sigma-Aldrich and used without further purification. The ratio of curcumin, demethoxycurcumin and bisdemethoxycurcumin in the purchased standard was found to be 67:29:4, respectively, based on high-resolution liquid chromatography–mass spectrometry (HR LC-MS) peak areas determined as described below. For compound analysis, a 5 mg ml⁻¹ curcuminoid standard stock was prepared in methanol, passed through a 0.22 μm filter, and stored at room temperature in an amber vial for no more than 2 weeks before use.

Groups of 10 third instar larvae were collected, rinsed and visually inspected to ensure that no food or curcuminoid remained on the body surface. Larvae were soaked in PBS for 1 h and prepared as described above for gut motility assessment, transferred to pre-weighed 1.5 ml Eppendorf tubes, larval masses determined, and tubes frozen at −20°C for fewer than 4 days before extraction. For the tissue specificity studies, individuals were soaked in 1× PBS solution for 1 h and dissected and divided into brain, gut and compositional tissue as previously described (Brent et al., 2009). Gut tissue samples were also collected from larvae in which the 1 h PBS soak was omitted to further assess gut clearance. All tissue components were individually collected with forceps and dissecting scissors, which were rinsed in ethanol and then water between each use to avoid potential chemical carryover, and larvae were rinsed with PBS between the collection of each component. All samples were collected in pre-weighed 1.5 ml Eppendorf tubes with 40 of each larval component individually per tube (N=5 tubes each). Tissue masses were determined, and tubes were frozen at −80°C for less than a week before extraction.

Aliquots of 200 μl methanol were added to each tube and the larvae were ground for roughly 1 min using a motorized Kontes pellet pestle rod system. To limit carryover, fresh pestles were used for each tube of larvae. Ground larvae in methanol were then sonicated on ice using a Qsonica Q125 sonicator (Melville, NY, USA) with a 0.25-inch probe for three pulses of 10 s at 20% power with about 10 s between each pulse. Tubes were then centrifuged at maximum speed (∼14,000 g) for 2 min, the methanol was transferred into new 1.5 ml Eppendorf tubes, and 200 μl of methanol was added to resuspend the larval pellets. The process of sonication, centrifugation and decanting methanol was repeated for a total of three rounds, and the total methanol extract from each tube of larvae was combined. The final combined methanol extracts were centrifuged at maximum speed for 4 min, the extract was transferred into a syringe while care was taken not to disturb the pellet, and the extract was filtered through a 0.22 μm filter directly into LC-MS vials for analysis. For brain and compositional tissue samples, the clarified extracts were dried in vacuo, water was removed via lyophilization, and the samples were resuspended in either 90 or 200 μl of methanol, respectively, before injection. Extraction efficiency was assessed by spiking a 2 μg aliquot of curcuminoid standard into each tube containing 10 larvae (N=10 tubes). The tubes were extracted as above and direct HR LC-MS determination and quantitation were conducted as described below. Extraction efficiency was determined to be 90.5±7.5%, and all data were scaled relative to this measurement.

High-resolution (HR) liquid chromatography and electrospray ionization quadrupole time-of-flight mass spectrometry (LC-ESI-TOF-MS) analysis was performed using an Agilent 1260 high performance liquid chromatography (HPLC) system equipped with a quaternary pump, an autosampler and a 1260-diode array detector upstream of a 6230 TOF mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) using a Dual ESI source in positive ionization mode. A 10 μl aliquot from each sample was injected and separation was achieved using an Acclaim RSLC 120 C18 column (3.0×110 mm, 2.2 μm, 120 Å; Thermo Scientific, Waltham, MA, USA) at 30°C with a flow rate of 0.75 ml min⁻¹ where line A was water+0.1% (v/v) formic acid, and line B was acetonitrile+0.1% (v/v) formic acid with the following program. The column was pre-equilibrated in 60% A–40% B and this composition was held for 5 min post-injection (Chao et al., 2018). The composition of mobile phase was then changed to 42% A–58% B over 15 min, and then brought to 0% A–100% B, with both transitions utilizing a linear gradient. This composition was held for 5 min followed by changing back to 60% A–40% B over 4 min utilizing a linear gradient. The column was equilibrated in 60% A–40% B for 5 min prior to the next injection. Under these chromatographic conditions, bisdemethoxycurcumin, demethoxycurcumin and curcumin eluted at roughly 7.9, 8.8 and 9.7 min, respectively. A standard curve was created via dilution of the 5 mg ml⁻¹ stock of curcuminoid standard such that 10–400 ng of the compounds was applied to the column and analysed in triplicate via the HR LC-MS program described above. The peak areas for the curcuminoids were summed to account for the presence of three compounds in the masses of the standard applied to column. The mass spectrometer had the following settings: mass range, 100–3000 m/z in profile mode utilizing the extended dynamic range (2 GHz); gas temperature, 325°C; drying gas, 8 liters min⁻¹; nebulizer, 25 psig; capillary voltage, 3500 V; fragmentor, 150 V; skimmer, 65 V; OCT 1 RF Vpp, 750 V; acquisition rate, 1 spectrum s^–1; time, 1000 ms per spectrum; transients per spectrum, 9982. Data were processed with MassHunter Workstation software, version B.06.00, build 6.0.633.0 (Agilent Technologies).

HPLC analysis was performed using an Agilent 1100 HPLC system equipped with a binary pump, degasser, an autosampler and a diode array detector. A 50 μl aliquot from each sample was injected and separation was achieved using a Zorbax Eclipse XDB-C8 guard column (4.6×150 mm, 5 μm, 80 Å; Hewlett Packard, Palo Alto, CA, USA) upstream of an Ultra C18 column (4.6×250 mm, 5 μm, 10 Å; Restek, Bellefonte, PA, USA) at 30°C with a flow rate of 1.0 ml min⁻¹ where line A was water+0.1% (v/v) formic acid, and line B was acetonitrile+0.1% (v/v) formic acid with the following program. The column was pre-equilibrated in 70% A–30% B and this composition was held for 3 min post-injection. The composition of mobile phase was then changed to 100% B over 15 min utilizing a linear gradient. This composition was held for 3 min followed by changing back to 70% A–30% B over 2 min utilizing a linear gradient. The column was equilibrated in 70% A–30% B for 3 min prior to the next injection. Under these chromatographic conditions, bisdemethoxycurcumin, demethoxycurcumin and curcumin eluted at roughly 13.7, 14.0 and 14.4 min, respectively, by measuring absorbance at 425 nm. A standard curve was created via dilution of the 5 mg ml⁻¹ stock of curcuminoid standard such that 1–1000 ng of the compounds was applied to the column and analysed in triplicate via the HPLC program described above. The peak areas for the curcuminoids were summed to account for the presence of the three compounds in the masses of the standard applied to the column.

Oregon R embryos were reared to eclosion on four different concentrations of curcuminoids (0, 0.25, 2.5 and 25 mmol l⁻¹). Wandering third instar larvae were rinsed in PBS, visually inspected to ensure that no media or curcuminoid remained on the surface, and sexed. For the acute oxidative stress challenge, larvae were soaked for 1 h in either 3% H₂O₂ as an oxidative stressor, or PBS. Groups of 10 larvae of each sex were saved in pre-weighed 1.5 ml Eppendorf tubes, and larval weight was recorded. Tubes of larvae were stored for less than a week at −80°C before testing, and all manipulation was conducted on ice unless otherwise specified. The thiobarbituric acid reactive substances (TBARS) protocol was conducted essentially as previously described with minor modifications (Ohkawa et al., 1979). Briefly, 150 μl of 20% trichloroacetic acid (TCA) was added to each tube, larvae were ground for roughly 1 min using a motorized Kontes pellet pestle rod system, and the homogenate was incubated on ice for 10 min. The homogenate was centrifuged for 5 min at maximum speed (∼14,000 g) and the clarified supernatant was transferred to a new 1.5 ml Eppendorf tube taking care not to disturb the pellet. A 150 μl aliquot of 0.5% thiobarbituric acid (TBA) in 20% TCA was added to each tube, vortexed briefly to mix, and incubated at 90°C for 15 min. After incubation, the tubes were centrifuged for 5 min at maximum speed, the absorbances at 532 nm (malondialdehyde; MDA) and 600 nm (non-specific turbidity) were measured in triplicate on an Eppendorf BioSpectrometer kinetic photometer, and the MDA–TBA product concentration was determined by comparison with a standard curve made with tetramethoxypropane as previously described (Ohkawa et al., 1979).

Oregon R embryos were reared to eclosion on four different concentrations of curcuminoids (0, 0.25, 2.5 and 25 mmol l⁻¹). Within 24 h of eclosion, adult flies were split by sex and grouped into vials containing medium with 0, 0.25, 2.5 or 25 mmol l⁻¹ of curcuminoids, with no more than 10 individuals per vial to prevent overcrowding. Flies were transitioned to fresh medium every 4–8 days as needed to prevent F2 eclosion and maintain medium integrity. For the survival analysis, deaths were scored every 48 h and any escapees were excluded from the analysis. Negative geotaxis was performed as previously described (Le Bourg and Lints, 1992). Briefly, flies were tapped to the bottom of a vial and any that passed a 5 cm mark within 18 s were recorded as a passing event. Three replicates were collected for each vial for each time point.

Oregon R embryos were reared to eclosion on four different concentrations of curcuminoids (0, 0.25, 2.5 and 25 mmol l⁻¹). Within 24 h of eclosion, 5–25 adult flies per vial were split by sex and grouped into vials containing media with 0, 0.25, 2.5 or 25 mmol l⁻¹ of curcuminoids. At 3 days after emergence (DAE), flies were starved on water-soaked filter paper for 3–6 h, and then transferred to a 3% H₂O₂ and 5% sucrose solution. Death events were visually scored every 12 h until completion of the assay.

Oregon R larvae were reared from hatching on four different concentrations of curcuminoids (0, 0.25, 2.5 and 25 mmol l⁻¹). Wandering third instar larvae were sexed and groups of 5–7 larvae were placed on a 9×13-inch agar dish and allowed 30 s of acclimation. After the acclimation period, the agar dish was centred on a light box and motility was recorded for 3–4 min. The camera was placed approximately 16 inches above the agar plate, and an image of a ruler was used to scale the video in the analysis software. The videos taken were reduced to 180 s, cropped to frame and converted to MP4 file format using iMovie (https://www.apple.com/uk/imovie/). The MP4 files were subsequently imported into Adobe Photoshop and exported as an image sequence at one image per second. The image sequence was then uploaded to ImageJ (https://imagej.net/Welcome) and converted into an image stack (image>stack>images to stack). The stack was inverted in color, converted to an 8-bit type, and the threshold was set at maximum intensity to ensure the larvae were the only objects highlighted in color. The wrMTrck plug-in (http://www.phage.dk/plugins/wrmtrck.html) was utilized to measure distance, length, average speed, maximum speed and body length per second. Parameters and settings for wrMTrck were used in accordance with a previous study focused on optimization of wrMTrck for measuring Drosophila larval motility (Brooks et al., 2016). The number of contractions, contraction rate and distance per contraction were measured manually.

Kaplan–Meier survival analyses were used to assess both the longevity and acute hydrogen peroxide adult data, and significance was determined with Mantel–Cox and Wilcoxon tests. For the negative geotaxis assay, pass rates from 2 to 16 DAE cohorts were analysed by three-way ANOVA to reveal the effects of age, curcumin concentration and sex on nervous system function. All analyses were completed in R (https://www.r-project.org/). Graphical preparation of data was conducted using GraphPad Prism software (GraphPad, San Diego, CA, USA).

A primary focus of this work was to determine the influence of curcumin on various developmental and phenotypic parameters of fruit flies as a means of establishing a baseline upon which future research may build. To achieve such outcomes, media were supplemented with varying concentrations of a curcuminoid standard and adult flies were allowed to breed and lay eggs in these media. The eggs and larvae would spend the duration of development in the presence of the curcuminoids and thus the influence of these compounds could be compared with flies reared in media lacking curcuminoids. A 98% pure curcuminoid standard derived from the C. longa plant was utilized for supplementation, which contained curcumin, demethoxycurcumin and bisdemethoxycurcumin in a 67:29:4 ratio. The underlying assumption of this work is that the larvae ingest or are permeable to curcuminoids in the medium such that some level of these compounds is present in tissues to modulate developmental and functional events.

To determine whether Drosophila larvae accumulate curcuminoids in their tissues, third instar larvae reared on media supplemented with 25 mmol l⁻¹ of curcuminoids were harvested in sets of 10 larvae per tube (N=8 tubes), extracted with methanol, and the amount of extracted curcuminoids was quantified using HR LC-MS. Larvae were incubated in PBS for 1 h prior to methanol extraction to both rinse their exteriors and allow the contents of their gut to pass. Supplementation of growth medium with Blue 1 dye, soaking for up to 1.5 h in 30 min increments, homogenization, and spectrophotometric analysis of soluble extracts indicated that the blue dye consumed in the medium passed out of the larval guts within an hour, consistent with previous findings on gut motility (Fig. S1; Shell et al., 2018).

Analysis of the pure curcuminoid standard with HR LC-MS found three peaks with retention times of roughly 7.9, 8.8 and 9.7 min, and masses corresponding to bisdemethoxycurcumin (calcd [M+H]+=309.1121, 2.6 ppm error), demethoxycurcumin (calcd [M+H]+=339.1227, 0.6 ppm error) and curcumin (calcd [M+H]+=369.1333, 0.6 ppm error), respectively (Fig. 1; Fig. S2). Extracts from third instar larvae reared on medium without curcuminoids lacked peaks with any of the curcuminoid masses at these retention times (Fig. 1). Spiking the curcuminoid standard into these control extracts revealed two new peaks, which correspond to curcumin and demethoxycurcumin in both masses and retention times. Three peaks were not observed because bisdemethoxycurcumin is a minor constituent of the standard mix and is only observed during analysis of high standard concentrations. Extracts from third instar larvae reared on media supplemented with 25 mmol l⁻¹ curcuminoids displayed two peaks, which are identical to curcumin and demethoxycurcumin in both masses and retention times. Spiking the curcuminoid standard into these larval extracts did not create new peaks for curcumin and demethoxycurcumin, no new masses were observed for these compounds, and the integrated area of the existing curcuminoid peaks increased. Integration of curcuminoid peak areas and comparison with a standard curve determined that larvae harbored 103±42 ng curcuminoids mg⁻¹ of larval tissue. Taken together, these data indicate that fruit fly larvae accumulate curcuminoids in their tissues, and that the HR LC-MS methodology used here is sufficient to unambiguously detect and quantify these compounds from larval extracts.

Although HR LC-MS is a sufficient method to detect curcuminoids in larval extracts, it is not cost effective or readily available to many research groups. Informed by the HR LC-MS results, an HPLC method was utilized that was also able to separate the three curcuminoids with retention times of roughly 13.7, 14.0 and 14.4 min corresponding to bisdemethoxycurcumin, demethoxycurcumin and curcumin, respectively (Fig. S3). Like the HR LC-MS results, two peaks corresponding to curcumin and demethoxycurcumin were observed in extracts from third instar larvae reared on media supplemented with 25 mmol l⁻¹ of curcuminoids, and no peaks were present in larvae reared on media without curcuminoids. Similarly, spiking the curcuminoid standard into negative controls resulted in new peaks with retention times like those of the standard and did not produce new peaks in larval extracts displaying curcuminoid peaks. Integration of curcuminoid peak areas and comparison with a standard curve determined that third instar larvae reared on media supplemented with 25 mmol l⁻¹ curcuminoids harbored 139±70 ng curcuminoids mg⁻¹ of larval tissue or 257±130 ng curcuminoids per larva (N=10 tubes), which is not significantly different from the value obtained from HR LC-MS analysis (t-test, P=0.21). Assessment of the curcuminoids harbored in the tissue of larvae raised on 2.5 or 0.25 mmol l⁻¹ of curcuminoids found 44±5 and 2±1 ng curcuminoids mg⁻¹ of larval tissue (N=10 tubes each), respectively. These data collectively demonstrate that HR LC-MS and HPLC can identify and quantify curcuminoids in larval extracts and that all of the larvae used in the following experiments contain curcuminoids in their tissues.

The results presented above indicate that the supplemented dietary curcuminoids are present in the larvae, but their tissue-specific accumulation remained a question. Due to the lipophilic nature of curcuminoids, it is possible that uptake is poor, they are not incorporated into cells, or never leave the gut, which is why studies in medicinal chemistry are working to derivatize curcumin for improved uptake kinetics. To determine the levels of curcuminoids present in the brain, gut and compositional tissues, 40 larvae reared on media supplemented with 25 mmol l⁻¹ curcuminoids were dissected, their tissues saved individually, and extraction and HPLC analysis conducted as described above. Only one curcuminoid supplementation condition was used for analysis because quantification from smaller amounts of sample, rather than whole larvae, would only provide results within the detectable range if derived from the highest accumulating larvae. Integration of curcuminoid peak areas and comparison with a standard curve determined that third instar larvae reared on media supplemented with 25 mmol l⁻¹ curcuminoids harbored 44±29 ng curcuminoids per larva in their compositional tissue, 85±53 ng curcuminoids per larva in their gut, and curcuminoids were not observed above the detectable limit from brain samples (N=5 tubes each). The combination of the composition tissue and gut curcuminoids is 129±81 ng curcuminoids per larva, which is not significantly different from the value obtained from the whole-larvae extractions (t-test, P=0.12). As a verification of the blue dye gut clearance assay described above, the guts of larvae not incubated in PBS for 1 h prior to extraction yielded 225±133 ng curcuminoids per larva. A visual inspection of the guts of larvae directly off media displayed far more contents than the largely emptied guts of larvae incubated in PBS for 1 h. These data indicate that curcuminoids remain in the gut after the majority of contents clear, accumulate in the compositional tissues, but were not observed in brain samples.

Previous research has indicated that treatment with curcuminoids modulates organismal responses to oxidative stress via radical scavenging abilities (Daverey and Agrawal, 2016; Jaroonwitchawan et al., 2017). To validate the presumed protective role of dietary curcuminoids, larval and adult fruit flies were assessed for their responses to both acute and chronic oxidative stress. Oregon R Drosophila were reared on a range of curcuminoid concentrations, third instar larvae were sexed, and then exposed to a 1 h acute oxidative challenge by soaking in 3% hydrogen peroxide. Larvae were homogenized, their peroxidated lipids were quantified using a TBARS assay, and the results were analysed using a three-way ANOVA to determine the interactions of sex, curcuminoid pre-treatment and hydrogen peroxide challenge on the amount of peroxidated lipids. Curcuminoid pre-treatment had an interaction with the hydrogen peroxide challenge, confirming the antioxidant properties of curcuminoids to an acute oxidative stressor in third instar larvae (F=1.27, P=0.029; Fig. 2). Sex has a significant impact on the amount of peroxidated lipids (F=37.98, P=1.38×10⁻⁸; Fig. 2), which is consistent with previous literature (Ruszkiewicz et al., 2019). Based on these data, the longevity of Drosophila adults reared on curcuminoids from hatching to adulthood was determined as a proxy for assessing lifetime oxidative stress levels.

To test whether curcuminoid pre-treatment resulted in an improvement in resistance of adult Drosophila to an acute oxidative challenge, we assessed the survival of adults reared on various concentrations of curcuminoid when challenged with 3% H₂O₂. Rearing of females on all curcuminoid concentrations tested resulted in a higher survival rate than females reared without curcuminoids (Fig. 3A). Interestingly, males reared on 0.25 mmol l⁻¹ curcuminoids did not exhibit a higher survival rate than males reared on 0 mmol l⁻¹ curcuminoids (Mantel–Cox, P>0.05; Fig. 3B), although higher curcuminoid concentrations resulted in higher survival rates than 0 mmol l⁻¹ curcuminoid-reared males (Mantel–Cox, P<0.0001, for both 2.5 and 25 mmol l⁻¹ curcuminoids; Fig. 3). Consistent with previous studies, we found that females had a higher median survival at every curcuminoid pre-treatment concentration (Fig. 3C; Pomatto et al., 2017). Taken together, these results indicate that dietary curcuminoids have a protective effect against an acute oxidative stress challenge, and extend the life of flies exposed to the chronic effects of oxidative stress.

Adult Drosophila reared from hatching on all concentrations of curcuminoids exhibit a longer lifespan and median survival compared with those reared on untreated media (Kaplan–Meier test, P<0.0001; Fig. 4). Interestingly, we observed a sexual dimorphism in the median survival of flies reared at the lower curcuminoid concentrations tested. Males flies reared on 0.25 mmol l⁻¹ curcuminoids exhibited greater survival compared with males reared on 0 mmol l⁻¹ curcuminoids (Mantel–Cox test, P<0.0001), but a lower median survival than females from the same 0.25 mmol l⁻¹ curcuminoid treatment (Mantel–Cox test, P=0.0010; Fig. 4). Similarly, males reared on 2.5 mmol l⁻¹ curcuminoids exhibited a lower survival than females also reared on 2.5 mmol l⁻¹ curcuminoids (Mantel–Cox test, P=0.0001), but a greater survival than males reared on 0 mmol l⁻¹ curcuminoids (Mantel–Cox test, P<0.0001; Fig. 4). Females reared on the highest curcuminoid concentration examined (25 mmol l⁻¹) did not exhibit a greater median survival than males reared on the same concentration (Mantel–Cox test, P>0.05), although both survived longer than flies of the matched sex without curcuminoid treatment (Fig. 4).

Brain tissue has the second highest abundance of lipids in the body after adipose tissue, and oxidation of brain lipids is associated with neural dysfunction and death (Butterfield et al., 2010). Analysis of larval locomotion provides a rapid, accessible assessment of several aspects of in vivo function of pharmacological treatments, including impacts on nervous system function (Kohsaka et al., 2017). Third instar crawling larvae exploring an open field have a bipartite locomotion pattern consisting of runs and reorientation phases, and genetic or pharmacological manipulations can independently alter each component of locomotion (Berni, 2015; Günther et al., 2016). To determine if curcuminoid treatment changed either component of larval locomotion, we analysed the exploratory behavior of freely crawling third instar larvae in an open field. Third instar larvae reared on the indicated curcuminoid concentrations were recorded exploring an agarose open field. Parameters of their locomotion were analysed with wrMTrk, and two-way ANOVAs were used to compare the effects of sex and curcuminoid concentration on velocity, distance moved per contraction, and reorientation of larvae in the field.

Curcuminoid treatment increases the turn rate in freely exploring larvae (ANOVA, F=8.7, P=8.9×10⁻⁵). Although several curcuminoid treatment concentrations resulted in increased male turn rates compared with females from the same treatment concentration, a robust interaction between sex and curcuminoids on the turn rate of freely exploring larvae was not detected (F=1.16, P=0.334; Fig. 5). Treatment with curcuminoids did not alter distance traveled or velocity of freely moving larvae on an open field (Fig. 5). No sexual dimorphism was evident in either distance traveled or velocity in any treatment group (Fig. 5).

To examine whether curcuminoids have a sex-dependent effect on neural function in Drosophila, the reflexive locomotor response of negative geotaxis in curcuminoid-treated flies was tested as they aged. This assay has frequently been used to assess neural function in aging and neurodegenerative models (Le Bourg and Lints, 1992; Prasad and Muralidhara, 2014; Rhodenizer et al., 2008; Yang et al., 2012). We found age to be the strongest predictor of performance in the negative geotaxis assay (ANOVA, F=89.59, P=3.12×10⁻⁴⁶), consistent with previous studies (Le Bourg and Lints, 1992; Rhodenizer et al., 2008). Curcuminoids interacted with the sex component (F=9.36, P=1.22×10⁻⁵), indicating a sexual dimorphism in the effect of curcuminoids on negative geotaxis in the aging flies (Fig. 6). It is important to note that sex alone did not affect the locomotor function (F=0.4192, P=0.519), nor was there an interaction between sex and age with regard to neural function (F=1.547, P=.0.157). The three-way interaction between all variables (curcuminoid concentration, sex and age) was significant (F=5.42, P=5.07×10⁻¹⁰), indicating that the interaction between curcuminoids and sex altered the rate of loss of locomotor function. Treatment of females with 0.25 mmol l⁻¹ curcuminoids accelerated the loss of neural function compared with 0 mmol l⁻¹ in females (Fig. 6A; green versus brown line). In contrast, the 0.25 mmol l⁻¹ curcuminoid concentration slowed the rate of neural function loss in males compared with males without curcuminoids (Fig. 6B; green versus brown line).

The purpose of this work was to determine the role of curcuminoids in phenotypic and developmental processes in an in vivo model system. Absorption of all curcuminoid components was confirmed with HR LC-MS (Fig. 1) prior to the assessment of phenotypic parameters. Curcuminoid treatment resulted in protection from an acute hydrogen peroxide challenge (Figs 2 and 3). Fewer peroxidated lipids were detected in two groups: females compared with males and curcuminoid-treated versus non-supplemented third instar larvae after an acute oxidative challenge (Fig. 2). Curcuminoid treatment also resulted in increased longevity in both male and female Drosophila, although some concentrations resulted in less protection for males than females (Fig. 4A–C). Freely crawling third instar larvae treated with curcuminoids demonstrated no changes in either velocity or distance moved per contraction, parameters frequently controlled by muscle or motor neuron function (Fig. 5A,B). In contrast, an increase in the rate of turning was observed in curcuminoid-treated larvae (Fig. 5C), a process controlled by the subesophageal ganglia (SEG) of the brain (Kendroud et al., 2018; Tastekin et al., 2015). Taken together, these data suggest that chronic treatment with curcuminoids has sex-dependent impacts on oxidative stress and brain function.

Our data show curcuminoid presence in Drosophila tissues (Fig. 1), consistent with curcuminoid absorption via the gut, and additionally our data also support the hypothesis that curcuminoids have an impact on neural function. Curcuminoid treatment resulted in changes to negative geotaxis (Fig. 6) and larval reorientation rate (Fig. 5), behaviors centrally controlled by mushroom body dopaminergic afferent neurons and the SEG, respectively (Sun et al., 2018; Tastekin et al., 2015). Previous studies have demonstrated that treatment with dietary curcuminoids can ameliorate metal and chemical-induced neurotoxicity in rat brains, as measured by HPLC and TBARS (Candan and Tuzmen, 2008), and influence central nervous system (CNS) cell composition and molecular expression (Sandur et al., 2007). Furthermore, curcuminoids are detectable following oral administration in mice and rat brain tissues, despite limited bioavailability (Schiborr et al., 2010; Suresh and Srinivasan, 2007; Tsai et al., 2011). Although curcuminoids were not observed in brain tissue in this study, this might be a function of the much smaller brain size and need for a much larger sample, which is the subject of further investigations by our group. Taken together, we speculate that curcuminoid treatment may influence neural function by modulating changes to the aforementioned locomotive associated CNS structures. However, whether these changes are by way of direct or indirect alterations and, additionally, the potential involvement of the gut–brain axis, requires further investigation.

Curcuminoid treatment was not associated with changes in larval speed, larval distance per contraction, or adult negative geotaxis at 2 DAE (Fig. 5A–C). These metrics arise largely from the function of motor neurons and muscles (Camuglia et al., 2018; Rhodenizer et al., 2008), which indicates that either curcuminoids have no effect on these components or that the effect was below assay detection level. The complex behavior of freely crawling larvae allowed the analysis of other locomotory components, including turning frequency. Curcuminoid treatment resulted in an increased turning frequency (Fig. 5C), which is a behavior that frequently follows reorientation processing within the CNS (Günther et al., 2016), and data suggest that these reorientation phases are controlled by the SEG (Tastekin et al., 2015). The SEG is a brain region responsible for incorporation of sensory cues (Kendroud et al., 2018), and recent evidence has demonstrated sexual dimorphism in the Drosophila SEG (Ye et al., 2012). In light of the findings described above, the data presented here suggest that curcuminoid treatment could modulate or exacerbate sex-dependent differences in the SEG, leading to the observed phenotypes.

We are grateful to Nancy Shough (Southern Oregon University, SOU), David K. Marshall (SOU) for technical support, and Lauren Millman (SOU) for administrative support.