Introduction

Brain tumor is a significant cause of morbidity and mortality worldwide that disproportionately affects children and young adults. The global mortality rate of brain cancer is about 2.5 per 100,000 individuals1. Current treatment regimens, such as chemotherapy, surgical interventions, and radiation therapy, employed to increase affected individuals’ survivability, do not guarantee the non-recurrence of tumors. They are often accompanied by limitations like reduced bioavailability, non-specific cytotoxicity, etc., thus imposing a substantial financial and emotional burden on the affected family. With the expansion of the therapeutic field, studies are being conducted to use plant-derived natural products for supplementing currently available treatments, which may not only eliminate the hassle of side effects but also provide a promising hope of efficiency in inhibiting tumorigenesis. One such drug is dihydroartemisinin (DHA), a semi-synthetic derivative of the established anti-malarial drug artemisinin that has been reported to deliver its anti-tumor effects in cervical, pancreatic, and gastric cancer cells and a variety of other cell lines2,3,4. When combined with other drugs like temozolomide and cisplatin, DHA effectively inhibited the growth of cancer cell lines and was non-toxic to healthy cells5. Assessment of its selective nature in an in vitro study elucidated the formation of complexes with iron and transferrin to kill cancer cells by regulating oxidative stress, leaving the normal cells unaffected6. However, its cell-specific cytotoxic efficacy in an in vivo system remains to be tested.

Drosophila has been studied as a model system for several decades to study the effects of various drugs7. In this study, the Drosophila brain tumor model used has a loss-of-function mutation in the tumor suppressor gene, lethal (2) giant larvae [l(2)gl], which codes for P127 protein that is responsible for apical-basal polarity8,9,10,11. Loss-of-function in l(2)gl leads to neoplasia in the brain hemispheres and imaginal discs in the Drosophila12. The human homolog of l(2)gl, known as LLGL-1 (formerly HUGL), plays a critical role in regulating cell polarity and proliferation13. Studies have shown that knockout of Lgl1 in mice leads to severe brain dysplasia, where neural progenitor cells overproliferate without undergoing differentiation14. Furthermore, decreased expression of LLGL-1 has been associated with the development of colorectal cancer15. Drosophila with homozygous recessive l(2)gl mutation also exhibit sluggishness and delayed pupation, and ultimately die at the third-instar larval or the pupal stage16.

Artemisinin is a sesquiterpene lactone with a 1,2,4-trioxane ring system extracted from the Chinese herb Qinghaosu (Artemisia annua or annual wormwood) and popular for its potent anti-tumor activity against various cancer cell lines17. Our previous work tested the anti-tumor activity of artemisinin and curcumin against brain tumors in a Drosophila model. Combining artemisinin with curcumin rescued 80% of the larval population that developed into late-stage pupae with significantly reduced tumor size. However, it failed to eradicate the brain tumor; therefore, no pupae developed into adult flies18.

This current follow-up study aimed to harness the properties of a non-toxic potential anti-cancer drug and treat malignant brain tumors in Drosophila larvae. We studied the effects of DHA on brain morphology and organization of the replicating neuroblast cells and other behavioral aspects such as locomotion and longevity of l(2)gl/l(2)gl mutant larvae. We also assessed the levels of reactive oxygen species (ROS) in treated larvae to decipher the probable mode of action of DHA. The present study examined the neoplastic inhibitory potential of DHA, highlighting the ROS-mediated therapeutic effect of the drug in the Drosophila glioma model.

Results

DHA rescues lethality of l(2)gl / l(2)gl larvae through effective inhibition of tumor growth

For an investigation of the anti-cancer activity of the candidate drug, 0–500 µM of DHA was selected for comparative tumor regression analysis in the late third-instar larvae of wild-type Oregon R+ and l(2)gl/l(2)gl mutants based on prior literature studies demonstrating the utility of broader micromolar ranges in Drosophila models for identifying effective therapeutic window during initial drug screening19. In our previous studies using the same l(2)gl/l(2)gl tumor model, we have taken a 0–750 µM range of artemisinin for testing its anti-cancer potential, where we found 500 µM to be the minimum effective concentration for eliciting tumor regression but did not rescue lethality18. In our current follow-up study, we have taken dihydroartemisinin as the candidate drug, which is the primary metabolite of artemisinin, and based on the previous findings18 we have narrowed down the range to 0–500 µM for our pilot studies. This strategy ensured robust dose–response characterization and aligns with canonical approaches employed in Drosophila-based therapeutic discovery20,21,22.

The treated l(2)gl/l(2)gl mutants showed maximum tumor regression from 250 µM DHA with a mean brain size of (102 ± 1.17)x10³ µm² against a brain size of (157 ± 1.52)x10³ µm² observed in the untreated mutant (Fig. 1A). Meanwhile, dosage response analysis with the wild-type larvae showed no major difference in the brain size with drug treatment. The overall brain size of the wild-type larvae was (86 ± 4.67)x10³ µm² even at a higher concentration of 500 µM similar to (84 ± 4.47)x10³ µm² at 0 µM, suggesting selectivity and minimal off-target effects (Fig. 1B). The mean brain size [(95 ± 1.83)x10³ µm²] of these 250 µM DHA-treated mutant larvae was comparable to that of wild-type larvae [(96 ± 1.27)x10³ µm²]; whereas, that of untreated l(2)gl/l(2)gl larvae was (138 ± 2.31)x10³ µm² (Fig. 1C). When cultured in 250 µM drug-treated food, the l(2)gl mutant larvae also showed a restoration of wing pouch area from (74 ± 1.67)x10³ µm² to (57 ± 1.61)x10³ µm². There was no significant difference in the wing pouch area between the treated larvae and the wild-type larvae [(58 ± 1.22)x10³ µm²] (Fig. 1D).

Fig. 1
figure 1

DHA restores brain and wing disc architecture, enhancing eclosion and survival of l(2)gl/l(2)gl mutants. (A) Upon DHA treatment, the dosage response curve in the third-instar tumorous larval brain with concentrations ranging from 0–500 µM showed 250 µM to be the optimal concentration for complete restoration of brain size (N = 20). (B) Upon DHA treatment, the dosage response curve in the third-instar wild-type larval brain with concentrations ranging from 0–500 µM showed no significant effect on the brain size (N = 20) (C) The image panel shows a representative dissected brain of wild-type, untreated l(2)gl/l(2)gl mutant, and treated mutant with a graph showing an effective brain tumor regression after treatment (Scale bar: 100 μm) (N = 20). (D) Image panel for dissected wing disc of wild-type, untreated l(2)gl/l(2)gl mutant, and treated mutant, with a graph showing an effective regression in tumorous growth in wing disc after treatment (Scale bar: 100 μm) (N = 20). (E) Longevity assay of wild-type and l(2)gl/l(2)gl larvae reared on the vehicle/drug-supplemented food revealed that 97% l(2)gl/l(2)gl larvae metamorphosed into the pupal stage when reared on 250 µM DHA in comparison with that seen in 500 µM artemisinin (78%) and 100 µM curcumin (67%) formulations on the 14th day (data for artemisinin and curcumin adapted from Das et al., 2014, with permission from IUBMB Life)18. On the 14th day of DHA treatment, the l(2)gl/l(2)gl flies eclosed with straight-wing phenotypes, which were not observed with artemisinin and curcumin. Data for artemisinin and curcumin were reused from our previously published study in IUBMB Life18 with permission obtained via RightsLink. New data for DHA was generated and analyzed as part of the current study. The final composite figure with all of the three drugs, artemisinin, curcumin and DHA, was plotted by Sushree Sulava, the first author of this manuscript, using GraphPad Prism software (version 5.01). (F) Graph showing no significant deviation in lethality in the wild-type strain in the wide drug-dosing bracket of (0-500) µM in food medium (N = 100; six trials were conducted for this assay) (*p < 0.05, **p < 0.01 and ***p < 0.001).

Full size image

The l(2)gl/l(2)gl mutants bearing brain tumors typically exhibit an extended third-instar larval stage, culminating in lethality, with only a minority reaching early pupation23. To assess the therapeutic efficiency of DHA, the drug-treated cultures were monitored for survival and developmental progression. In the untreated control, 89% of l(2)gl/l(2)gl mutants displayed larval death. In contrast, upon treatment with artemisinin and curcumin, as reported previously from our work, about 78% and 67% of larvae metamorphosed into the brown pupal stage, respectively. However, there was no eclosion of adult homozygous flies from cultures exposed to these two candidate drugs18. On the other hand, DHA formulation treatment yielded an improved developmental outcome with about 97% larvae reaching the brown pupal stage, from which eclosion of about 75% l(2)gl/l(2)gl adult flies with straight wings was observed (Fig. 1E). The eclosion rate was calculated through the ratio of successfully eclosed homozygous l(2)gl/l(2)gl adults to the total number of homozygous l(2)gl/l(2)gl larvae that initiated pupation. In addition, we have also scored for the percent lethality in the wild-type strain wherein we did not find any significant change in the percent lethality in the range of 0–500 µM (Fig. 1F). Hence, at 250 µM, DHA elicited potent suppression of neoplastic phenotype in the l(2)gl/l(2)gl model without inducing adverse effects on organismal viability in the mutant and the wild-type strain.

Immunostaining reveals the revival of concentric rings of replicating cells in the drug-treated third-instar larval l(2)gl/l(2)gl Drosophila brain

Bromodeoxyuridine (BrdU) is a thymidine analog incorporated in the newly replicated DNA of a live cell/tissue24. While distinct concentric ring patterns of newly replicated cells are observed in the outer proliferative center of neuroblasts in wild-type Oregon R+ third-instar larvae with about 29 BrdU-labelled cells (in 590 μm² 10X view-field area), this pattern was highly disrupted in l(2)gl/l(2)gl model with only about ten labelled cells in the same marked region (area shown in schematic figure Fig. 2A), owing to the rampant proliferative nature of the neuroblasts, which becomes prominent from the second-instar larval stage leading to a non-uniform distribution of the cells across the lobe instead of clustered arrangement in the proliferation center25,26,27. This was restored to 28 cells after treatment with DHA (Fig. 2B). Anatomically, these results proved that DHA at an optimal concentration of 250 µM can restore the wild-type phenotype in larvae genetically marked to have brain tumors.

Fig. 2
figure 2

DHA restores the cellular rearrangement of the tumor-bearing third-instar larval Drosophila brain. (A) An illustration indicating the outer proliferation center (OPC) in the optic lobe (OL) in Drosophila larval brain and the central brain (CB). The illustration was drawn by Sushree Sulava, the first author of this manuscript, using Microsoft Powerpoint (Office 2021). (B) A representative image of a BrdU-stained dissected third-instar larval brain reared on a vehicle or drug-supplemented food showed treatment-specific restoration of the characteristic concentric ring pattern of newly replicating cells. The top panel represents a representative figure showing the outer proliferation center and the brain morphology (White arrowheads indicate the uniform ring-like cellular arrangement in the proliferation center) (Scale bar: 20 μm) (N = 8) (*p < 0.05, **p < 0.01, and ***p < 0.001).

Full size image

DHA executes transient ROS surge in l(2)gl / l(2)gl mutant to implement anti-tumor effect

Levels of Reactive Oxygen species (ROS) were analyzed in treated and untreated larval brains to examine the mode of action of the anti-tumor property of DHA. ROS acts as a double-edged sword in cancer. While ROS levels kickstart some pro-tumorigenic signaling cascades, elevation beyond a certain threshold can prove fatal for cancer cells and thus inhibit their progression28,29,30. Brains from the 8th-10th day for l(2)gl/l(2)gl untreated control and treated larvae were dissected to check for fluorescence in early and late third-instar larval stages and compared with wild-type third-instar larvae (6th-7th day). The mean ROS production in the late third-instar wild-type larvae on the 7th day was (27.6 ± 0.83)x10³ RFU, whereas, in the untreated l(2)gl/l(2)gl tumorous larvae, ROS produced at their early (8th day) and late (10th day) third-instar stages were (39.4 ± 1.15)x10³ and (39.2 ± 1.15)x10³ RFU, respectively. When these mutant larvae were reared on DHA, ROS production was 1.4-fold higher in the early third-instar larval brain with (54.2 ± 1.42)x10³ RFU on the 8th day, which dropped to (25.9 ± 0.45)x10³ RFU upon continued drug treatment until the late third-instar stage (10th day). The latter ROS levels in the late third-instar stage of treated l(2)gl/l(2)gl mutant (10th day) were comparable to the late third-instar stage of wild-type larvae (7th day) (Fig. 3). Thus, the drug rescues the mutant strain’s morphological and anatomical aspects and delivers its anti-tumorigenic impact, specifically in the cancer cells only, through ROS generation.

Fig. 3
figure 3

DHA executes a transient ROS surge in l(2)gl/l(2)gl mutant to implement anti-tumor effect. Representative figure panel with illustration showing the ROS signals in the dissected brain samples (White arrows indicate some of the quantified regions). The illustration was drawn by Sushree Sulava, the first author of this manuscript using Microsoft Powerpoint (Office 2021). Quantitation of ROS levels in the larval brain in wild-type and tumorous strains with and without drug treatment showed elevated levels of ROS post-DHA treatment on the 8th day in the early third-instar stage of the l(2)gl/l(2)gl mutant, followed by a steep decrease in ROS activity on the 10th day (late third-instar of the l(2)gl/l(2)gl mutant), similar to levels seen on the 7th day for the late third-instar stage of the wild-type larvae (N = 6) (Scale bar: 20 μm) (*p < 0.05, **p < 0.01, and ***p < 0.001).

Full size image

DHA restores locomotor function in l(2)gl/l(2)gl mutant at 250 µM dose

Locomotion studies on l(2)gl-deficient third-instar larvae revealed their sluggish movement in 30 s (11.7 ± 0.56 mm) compared to wild-type Oregon-R+ (26 ± 0.23 mm) third-instar larvae. When these mutants were reared on DHA, they covered 25.8 ± 0.30 mm distance within the same period (Fig. 4A), thus restoring the locomotory phenotype. We conducted a negative geotaxis assay in the newly eclosed adult flies where we observed 17 out of 20 DHA-treated l(2)gl/l(2)gl flies crossed the predetermined 7.0 cm mark on the climbing apparatus in 30 s, which would be comparable to the wild-type (18 out of 20 flies) (Fig. 4B). These observations interestingly showed an active behavior of flies in the overall population, like the wild-type; thus, single fly analysis was performed to draw conclusive evidence. The time taken by individual flies to cross the boundary was recorded. DHA-treated l(2)gl/l(2)gl adult flies displayed an average of 6.17 s to cross the 7.0 cm mark, which was comparable to 5.9 s by wild-type (Fig. 4C). This result confirmed the reinstatement of the locomotory abilities of the l(2)gl/l(2)gl mutant exposed to drug treatment.

Fig. 4
figure 4

DHA ameliorates the locomotory defects in the l(2)gl/l(2)gl mutant at 250 µM. (A) Locomotion assay of third-instar larvae (wild-type, drug-treated, and untreated l(2)gl/l(2)gl mutants) showed rescue in the sluggish crawling movement (N = 20). (B) Age-related behavior assay (negative geotaxis) verified the rescue effect of DHA-treated newly eclosed adult flies with the mutation. Ten trials were conducted for this assay, with each trial noting the number of flies out of twenty flies that covered the required distance of 7.0 cm in 30 s. (C) Single fly behavior analysis done with age and sex-matched young adult flies (N = 10) with time as a parameter showed the almost similar time taken by the treated rescued flies and the wild-type flies to cross the predetermined 7.0 cm mark in the climbing apparatus (*p < 0.05, **p < 0.01, and ***p < 0.001).

Full size image

Early and continuous DHA administration is crucial for non-tumorigenic phenotype sustenance in the l(2)gl/l(2)gl mutant strain

To explore the therapeutic window, the l(2)gl/l(2)gl larvae with stage-based temporal exposure to the drugs were dissected at the late third-instar stage. We observed that early drug administration from the 8th day was sufficient to yield a regular brain size restoration [(105 ± 3.89)x10³ µm²], which was similar to that of the wild-type [(95 ± 1.38)x10³ µm²]. In contrast, late drug administration in the late third-instar stage (10th day) resulted in almost no such rescue in brain phenotype [(128 ± 2.6)x10³ µm²], emphasizing the timely effect of DHA in implementing its efficiency (Fig. 5A).

Fig. 5
figure 5

Continuous and early DHA administration maintains the non-tumorigenic phenotype in l(2)gl/l(2)gl mutant strain. (A) Plot showing the effect of differential time-based drug exposure in Drosophila third-instar larvae on days 8, 9, and 10 post-egg laying, demonstrating early exposure (day 8) is required to elicit a significant anti-tumorigenic effect. (B) Schematic representation of the drug withdrawal-based generation experiment till the fourth generation. (C) Plot represents drug withdrawal’s effect on the brain area in the rescued l(2)gl/l(2)gl mutants (WD: Withdrawal of DHA) (N = 20). (D) Plot represents the effect of drug withdrawal on the wing pouch area in the imaginal discs in the rescued l(2)gl/l(2)gl mutants (N = 20) (*p < 0.05, **p < 0.01, and ***p < 0.001).

Full size image

Moreover, to explore the role of DHA in tumor inhibition and survival versus tumor relapse, the rescued adult l(2)gl/l(2)gl mutants were allowed to mate and grow with and without DHA for four generations. Post-20th day of mating, there was an eclosion of l(2)gl/l(2)gl adult offspring with straight wings, with drug treatment. The DHA-rescued l(2)gl/l(2)gl flies (founder group, gray box) were split into cultures harboring either vehicle only (green box Gen-I in Fig. 5B) or DHA with vehicle (purple box Gen-I in Fig. 5B) and self-crossed, and the larval progeny were screened for tumor relapse. After the analysis of the target tissue size in larval population, the cultures were monitored for the eclosion of eclosed adult flies (l(2)gl/l(2)gl) from the drug withdrawal group (green box Gen-I) and the resulting adult flies were then split again into two cultures: the vehicle only (green box Gen-II) or DHA with the vehicle (purple box Gen-II) and the cultures were monitored for larval tissue dissection and subsequent adult l(2)gl/l(2)gl fly eclosion. This splitting procedure was followed till the fourth generation. Meanwhile, we also considered a control group of l(2)gl/l(2)gl rescued adult line flies which were exposed to DHA treatment (pink box in Fig. 5B) for each generation (Fig. 5B). The rescued homozygous mutants receiving continuous DHA treatment exhibited suppression of tumor growth in the brain and the wing disc tissues. However, when subjected to a drug-free medium for two generations, the rescued l(2)gl/l(2)gl group showed a pattern of tumor resurgence. The brain tumor reappears in the second generation of the rescued mutants subjected to drug withdrawal [(128.0 ± 2.6)x10³ µm2] (green box), and persists even in the subsequent generations, with the third generation withdrawal mutants bearing mean brain size of (129.2 ± 4.0)x10³ µm2 (green box) and the fourth generation withdrawal mutant showing brain tumor [(131.7 ± 2.2)x10³ µm2] (green box) comparable with that of the untreated l(2)gl/l(2)gl control larvae [(135.4 ± 2.2)x10³ µm2]. Interestingly, the l(2)gl/l(2)gl control group cultured in drug medium (purple box) showed an effective tumor regression [(93.7 ± 1.79)x10³ µm2] and rescue effect in the F2 generation, even though the F1 parental generation was cultured without the drug, which suggested that acute exposure to DHA during larval development is sufficient to reverse the tumor phenotype. This effect was seen with each generation, wherein the progeny collected from the drug-withdrawn parents were exposed to the drug medium again and monitored till eclosion. Three other controls were set in parallel (wild-type control, untreated control, and the treated group subjected to continuous drug exposure). The rescued group receiving continuous drug exposure maintained the normalcy of the brain tissue architecture with a mean value of (91.1 ± 1.75)x10³ µm2 similar to that of the wild-type [(95.3 ± 1.92)x10³ µm2] (Fig. 5C). This aligned with the eclosion rate that we observed for the adult l(2)gl/l(2)gl flies in each group. The eclosion rate for the drug withdrawal group (green box) shifted from 100% in Gen-I to 98% in Gen-II and kept decreasing to 91% in Gen-III and 86% in Gen-IV. In contrast, the subgroup of the drug withdrawal group subjected to immediate DHA treatment for one generation (purple box) showed 100% eclosion rate of l(2)gl/l(2)gl adult flies, which was similar to that seen in case of the group with continuous exposure to DHA (pink box) i.e., 100% eclosion. Remarkably, l(2)gl homozygous mutants originally rescued by drug treatment maintained high adult eclosion rates across four drug-free generations, suggesting a sustained, multi-generational effect of the pharmacological intervention on organismal development.

A similar tumor recurrence pattern was seen in the case of wing imaginal discs wherein the wing pouch size of the rescued mutants increased from (58.6 ± 1.23)x10³ µm2 to (69.7 ± 1.91)x10³ µm2 after being subjected to two generations of drug withdrawal (green box), which was close to the untreated control mutant wing disc size [(73.5 ± 1.35)x10³ µm2], which was prominent even in the fourth generation [(75.4 ± 2.14)x10³ µm2]. The wild-type control showed a size of (59.7 ± 1.25)x10³ µm2 (Fig. 5D). The l(2)gl/l(2)gl mutants notably displayed different anatomical pharmacology in the tumor tissues in chronic DHA treatment versus DHA withdrawal regimes, indicating that continuous drug supply is necessary for the rescued flies to maintain them in the non-tumorous state.

Discussion

Brain tumors comprise more than 90% of CNS-based tumors and have exhibited an increasing trend in mortality rate over the past few years31. Despite various advancements in treatment approaches in the field of cancer therapeutics, a range of side effects accompanying the treatment option’s efficacy limits its usage as an ideal oncotherapeutic agent32. Opting for natural product-derived drugs removes this limitation and provides additional benefits regarding abundance and tissue-specific non-cytotoxicity. Various in vitro and in vivo studies have revealed a range of potential candidates with therapeutic implications. Plumbagin reduces the expression of the E2F transcription factor and influences the cell cycle regulatory factors in brain cancer cells33. Resveratrol can induce cellular death and senescence and cause cell cycle arrest via p53, mTOR, Wnt, or STAT3 signaling processes. When administered with temozolomide, resveratrol can reduce the cancer cell resistance to the drug’s action through the O6-methylguanine methyltransferase34. Artemisinin, in addition to its anti-malarial activity, has been shown to induce cell cycle arrest and a caspase-3-dependent cellular death in glioma35. It has also demonstrated its promising anti-tumor activity in breast cancer cells by activating increased T-cell mediated immune signaling and repression of the regulatory T-cells in the tumor microenvironment36. Dihydroartemisinin (DHA), a derivative of artemisinin, has also been reported to deliver anti-cancer effects in vitro and in vivo in gastric cancer, hepatocellular carcinoma, pancreatic cancer, etc37,38,39. This study aimed to evaluate the efficiency of DHA as a potential anti-cancer agent in the Drosophila brain tumor model.

Drosophila, having homologs for about 75% of disease-causing genes in humans, confers a significant advantage as a screening model for cancer research40. The modern genetic tools available help the dissection of novel mechanisms of tumorigenesis. Drosophila strains having mutation-driven constitutional activation of PI3K and EGFR pathways have been shown to model glioblastoma and other human cancers41. A mutation in the l(2)gl gene, which otherwise is responsible for maintaining cell polarity, disrupts cell proliferation and polarity patterns, leading to neoplastic growth in the brain and imaginal discs with developmental arrest at the larval-pupal transition phase, resulting in lethality11. Gont et al., 2014 have reported PTEN loss (occurring in about 85–90% of glioblastomas) leading to lgl-1/HUGL-1 (human homolog of lgl) inactivation, resulting in the increased invasion of glioma cells, whereas Lgl3SA (the constitutively active form of lgl-1) decreased the glioma invasiveness both in vitro and in vivo in an immunocompromised intracerebral mouse model25,42. Liu et al., 2015 showed that HUGL-1, which is otherwise downregulated in glioma tissues, inhibited gliomagenesis in vitro and an intracranial nude mouse model26,43. Thus, the central features of lgl function as a tumor suppressor in Drosophila are conserved in human glioblastoma.

In this study, we have reported for the first time that 250 µM DHA concentration is sufficient in not only restoring the normal architecture of the brain and the imaginal discs and the locomotory aspects but also in inducing a significant rescue in lethality in the l(2)gl/l(2)gl mutants. About 97% of DHA-fed l(2)gl/l(2)gl larvae metamorphosed into the pupal stage, from which 75% of l(2)gl/l(2)gl flies eclosed with the straight-wing phenotype. This pronounced rescue highlights the DHA-induced pharmacological modulation of developmental processes disrupted by tumor suppressor deficiency. Many anti-cancer candidates show non-selective cytotoxicity, and chronic exposure may lead to adverse effects involving multi-system toxicity, thereby increasing mortality even after treatment44,45. In contrast, our study demonstrated that long-term exposure to DHA in the l(2)gl/l(2)gl mutants did not result in any observable deleterious effects. Treated mutants exhibited a persistent normal phenotype with standard brain and wing imaginal discs without eliciting any negative impact on the fecundity of the flies, thus corroborating our optimum dosage analysis. In addition, the studies carried out in the wild-type control showed consistent brain architecture in treated and untreated conditions. We also monitored wild-type flies across this dosage window (0–500 µM) and observed no significant deviation in lethality, affirming the compound’s low toxicity in a non-mutant background. DHA, at a tolerable dose of 250 µM, enabled robust suppression of neoplastic phenotypes without compromising organismal viability even under chronic exposure conditions. These findings suggested 250 µM to be a physiological concentration, highlighting its favorable safety profile that is divergent from the toxicity profiles typically reported for conventional anti-cancer candidates.

In contrast, the rescued mutant strain, when subjected to DHA withdrawal from the medium for at least two generations, demonstrated a resurrected tumorous growth pattern in both the target tissues, thus accentuating the effect of continuous exposure to DHA in generating l(2)gl/l(2)gl flies bearing wild-type phenotype with a regressed tumorous growth. The withdrawal experiment has therapeutic significance in illustrating the long-term plasticity of responses to DHA and highlights the protective effect of DHA within a single generation. An intriguing aspect of the study is that the withdrawal group of F1 larval progeny (green box, Gen-I) maintained a non-tumorous state of the brain [(96.6 ± 1.97)x10³ µm2] without any exposure to the drug medium wherein only the parents (F0 in larval stages) had prior exposure to the drug before culturing them (F0 eclosed adults) in a standard food medium. The observation that the l(2)gl/l(2)gl mutants, initially rescued with DHA treatment, continued to exhibit high eclosion rates up to the fourth generation under drug withdrawal conditions is particularly striking. Despite the eventual recurrence of neoplastic lesions in the larval brain, the ability to complete metamorphosis remains robust (> 80%). This sustained developmental stability suggests that the initial drug exposure may have induced inheritable epigenetic modifications capable of reshaping the developmental circuitry across generations. These epigenetic alterations likely shift gene expression thresholds in key pathways governing polarity, proliferation, or stress responses, thereby enhancing developmental resilience despite the absence of the drug. A plausible explanation would be a germline reprogramming induced in the Drosophila parents exposed to the drug, thereby altering gene expression patterns in the withdrawal progeny cultured in the standard medium through transient epigenetic repriming, which aligns with findings from several other studies46,47. However, the gradual decline in eclosion frequency (from 100 to 86% by Gen-IV) is suggestive of decay kinetics of transgenerational epigenetic memory in the mutant, where such marks persist for only a few generations before reverting. Thus, our findings suggest transgenerational epigenetic inheritance as a mechanism through which a single pharmacological intervention can exert long-lasting, though ultimately diminishing, benefits on organismal development and viability. Similarly, Kovalchuk et al., 2018 showed a variation in molecular gene expression signatures in mice exposed to chemotherapy, which was also apparent in the unexposed progeny, displaying changes such as variations in the levels of DNMT1 and MeCP248, thus indicating epigenetic transmission across generations. The exact role of epigenetics in the generational protective effects rendered by DHA warrants further research.

Due to impaired cell cycle regulation in tumor cells, the proliferation rate is magnified several times to yield a disruptive distribution pattern of the replicating cells in the proliferative center in the optic lobes25. In line with these findings, the vehicle-treated larvae in the current study exhibited marked overproliferation and disruption of the stereotypical concentric ring architecture in the outer proliferation center (OPC), consistent with the disorganized cellular distribution typically associated with uncontrolled growth. Notably, DHA administration revived the concentric ring pattern on the outer proliferation center on optic lobes, as revealed by the BrdU assay. This structural normalization suggests a re-establishment of spatially regulated proliferation, indicating that the drug may exert its effects by reinstating proper regulation of cell cycle control mechanisms and tissue organization. These findings not only support the functional efficacy of DHA in restraining cellular overproliferation but also highlight its capacity to rescue morphogenetic integrity in the developing brain, an aspect that has not been clearly addressed in prior studies. This offers a potential strategy for targeting early proliferative dysregulation in neural tissues.

While many cancer cells adapt to persistent intrinsic oxidative stress by upregulating their endogenous antioxidant systems49,50 this adaptation can be strategically targeted. The pharmacological abrogation of this redox homeostasis using pro-oxidant agents presents a promising therapeutic avenue, as it preferentially disrupts the survival of cancer cells. Several in vitro studies on DHA have provided a probable mechanistic approach to selective oxidative stress through ROS production in fibrosis cases51 and various cancers52,53,54. Cancer cells show greater iron uptake, as has been evident from the increased accumulation of endocytosed transferrin found in the lysosomal fractions of cancer cells, and artemisinin has been reported to form Fe(III)–artemisinin complex, thereby explaining its selectivity towards cancer cells55,56,57. These complexes can lead to carbon-centered radicals-mediated lysosomal membrane disruption, leading to escalated ROS generation and eventually activating apoptotic signaling in the cancer cells17. At the early third-instar stage (8th day), untreated l(2)gl/l(2)gl larvae exhibited about 1.40-fold elevation in ROS compared with the wild-type larvae at the equivalent developmental stage (6th day). This elevation likely reflects tumor onset and progression within the larval brain. Remarkably, the early third-instar l(2)gl/l(2)gl mutants treated with DHA displayed a further 1.40-fold ROS elevation over untreated mutants (8th day), amounting to about a 1.99-fold increase versus the early third-instar wild-type control larvae (6th day). ROS levels gradually decreased to basal levels in the late third-instar stage (10th day) and were comparable to the late third-instar stage of the wild-type control (7th day). This pronounced amplification in ROS levels in the treated mutants may overwhelm the endogenous antioxidant defenses of the tumor cells, driving them past a critical threshold of oxidative stress that is incompatible with cellular survival, thus triggering apoptosis29,58. This observed ROS-surge in the DHA-treated tumorous larvae supports the notion that DHA may function through a pro-oxidant mechanism with tumor-specific cytotoxic potential in the l(2)gl/l(2)gl Drosophila tumor model. Further, transcriptomic analysis may be conducted to understand the pathways involved in this mechanism. Drosophila glioma models harboring mutations in key pro-tumorigenic factors or tumor-suppressors may be exploited to further elucidate the pathways targeted by the drug. A comparative study of differential drug responses will illuminate the specific molecular interactions, thus helping us in identifying the crucial nodes within the signaling networks critical for tumorigenesis that are being targeted by DHA.

To assess the functional impact of DHA treatment on l(2)gl/l(2)gl larvae, we performed a quantitative analysis measuring locomotor performance through larval crawling assay. Untreated homozygous mutant larvae exhibited pronounced hypokinesia, whereas pharmacological intervention significantly restored locomotion efficiency in the treated l(2)gl/l(2)gl mutants in the same time frame of thirty seconds. The negative geotaxis assay also showed similar motor abilities between the wild-type strain and the drug-rescued l(2)gl/l(2)gl flies. These findings indicate that the compounds not only inhibit tumor growth but also suppress tumor-induced motor deficits, thereby enhancing the overall physiological state suggestive of improved organismal health.

In this study, DHA demonstrated significant anti-tumor efficacy in the l(2)gl/l(2)gl brain tumor model of Drosophila, highlighting its potential as a natural product-based therapeutic agent against glioma, capable of modulating multiple oncogenic networks. Studies in other Drosophila brain tumor models are in progress, which will provide novel insights into the anti-cancer activity of DHA. Demonstrating conserved phenotypic rescue across species would significantly bolster the therapeutic promise of these compounds in treating oncogenic growth while preserving organismal vitality. To fully realize its clinical potential, further research is required to decode the functional signaling cascades of DHA in glioma regression for it to be embedded in personalized oncology and next-generation therapeutic platforms.

Materials and methods

Fly husbandry and optimal dosage efficiency

Oregon R+ (wild-type) and l(2)gl (tumorous; genotype: y1 w67c23; l(2)gl4/CurlyO, y+; Bloomington stock number 9042) fly stocks were reared on standard food medium containing agar (Himedia, Thane, Maharashtra, India), maize powder, sugar, and yeast extract powder (Himedia), nepagin (anti-fungal, Himedia), and maintained under normal uncrowded conditions at 22 ± 1 °C with 80–90% relative humidity18. Previous studies conducted in our lab have used two other candidate drugs: artemisinin (Obtained as a gift from Ipca Laboratories Ltd.) and curcumin (C1386, Sigma). In our current study, the drug-formulated medium contained dihydroartemisinin (D3793, TCI chemicals, Chennai, Tamil Nadu, India) and Neelicol brilliant blue FCF dye (Neelicol food dyes, Roha, Maharashtra, India) (0.125 mg/ml) along with the standard food medium containing agar to confirm drug ingestion. For this purpose, the drugs were dissolved in 1.0 ml of absolute ethanol and then added to the food medium along with the FCF dye, followed by stirring for solubilization. Synchronous culture, a prerequisite condition, was maintained throughout the experiment18. Third-instar l(2)gl/l(2)gl larvae (w/w) exhibit unpigmented Malpighian tubules, while l(2)gl/CyO heterozygotes (bearing w/w⁺ allele) display yellow tubules due to partial white gene activity, enabling phenotypic distinction59,60. The l(2)gl/l(2)gl third-instar larvae (N = 20) reared in DHA (0–500 µM) were dissected in 1X PBS under a Nikon SMZ25 stereofluorescent microscope (Nikon, Tokyo, Japan) with immediate image acquisition. Larval brain lobes and the pouch region of the wing imaginal discs were traced in the images and quantified through the NIS-Elements BR (version 5.20.00) software. The minimum concentration that showed maximum tumor regression was used for further assessment.

Temporal drug exposure regimes in Drosophila larval brain

Synchronized cultures of wild-type, untreated, and treated mutants were set in standard food. Cohorts of l(2)gl/l(2)gl larvae were transferred to standard and drug-containing medium at specific developmental stages of the third-instar larval stage (day 8, 9, and 10 after egg-laying). Each group was maintained on the standard or drug-supplemented medium for the remainder of larval development. The brain tissues were dissected in each group at the late third-instar stage and subjected to image acquisition through the Nikon SMZ25 stereofluorescent microscope for further quantification.

Bromodeoxyuridine (BrdU) assay in Drosophila larval brain

Immunohistochemistry was performed to visualize bromodeoxyuridine in the Drosophila larval brain18,61. The dissected brains of third-instar homozygous larvae were incubated in 20 µM BrdU (Sigma, St. Louis, Missouri, USA) dissolved in Schneider’s insect medium (Invitrogen, Scotland, UK) for 3 h with subsequent fixation in 4.0% (w/v) paraformaldehyde (Himedia) for 20 min, permeabilization in 0.3% Triton X-100 (Sigma) in PBS. The samples were then subjected to DNA hydrolysis by incubation in 2 N hydrochloric acid for 20 min at room temperature with shaking, followed by a PBS wash. The samples were blocked in 10% normal horse serum (Invitrogen, Waltham, Massachusetts, USA) and then incubated in mouse anti-BrdU antibody [1:50; Sigma, (B8434)]. Using anti-mouse Alexa Fluor-488 secondary antibody [1:500; Invitrogen (A32723)] and DAPI (Sigma), the tissues were visualized under the Leica SP8 laser-scanning confocal microscope (Leica, Wetzlar, Hessen, Germany). Five random areas (590 µm2) within the larval brains’ outer proliferating center (OPC) in the same visual field were chosen in ImageJ analysis software (ImageJ version 1.54j software; https://imagej.net/ij/download.html) was used for image processing), and the BrdU-positive cell count was quantified in treated and untreated tissue sections. For each category, eight replicates were recorded and analyzed.

ROS assay using Dihydroethidium (DHE) in Drosophila larval brain

Dissected larval brain samples were incubated in 30 µM of dihydroethidium (DHE; Invitrogen) diluted with Schneider’s medium for 6 min in the dark at room temperature, followed by washes in Schneider’s medium. Mounting was performed immediately in the Vectashield mounting solution (Vector Laboratories, Newark, California, USA), with immediate image capture using the Leica SP8 laser-scanning confocal microscope62. The captured brain images were subjected to ImageJ analysis (ImageJ version 1.54j software; https://imagej.net/ij/download.html). Quantification is reported in the Relative Fluorescence Unit (RFU) = Mean intensity of brain lobes – (Area of cell X Mean intensity of background). The average RFU of N = 6 replicates (each with ten trials) was plotted on the 3D graph using OriginPro software (version 10.1.0.178).

Life span and locomotor defects post drug treatment in third-instar Drosophila larvae

Post-sixth day of the mating period, the third-instar l(2)gl/l(2)gl larvae were transferred for treatment, and the survival rate was recorded till the fourteenth day to evaluate drug efficacy. In our current study, 250 µM DHA was used. For artemisinin and curcumin data from our previous studies, 500 µM and 100 µM, respectively, were used18. For locomotion assay, 1.0% agarose (Himedia) plates were prepared in 60 mm X 15 mm petri dishes with the Neelicol brilliant blue FCF dye to visualize traces of crawling third-instar larvae63. The distance covered by the larvae in 30 s was recorded, and a comparative analysis was performed on the average distance covered by wild-type, untreated, and treated tumorous third-instar larvae.

Behavioral study of the l(2)gl/l(2)gl flies by negative geotaxis

Age and sex-matched wild-type (Oregon R+) and DHA-rescued l(2)gl/l(2)gl adult flies were placed in a polystyrene climbing apparatus with a mark of 7.0 cm distance and gently tapped to the bottom64. Climbing ability was calculated from (a) the number of flies crossing the 7.0 cm mark in 30 s, and (b) the time taken by a single fly to cross the 7.0 cm mark. The data was represented using GraphPad Prism software (version 5.01). For the former assay, twenty flies were tapped to the bottom, and the number of flies travelling upwards to cover the required distance of 7.0 cm in 30 s was noted. Ten trials were conducted for this assay. For the latter assay, individual flies were investigated to check the time taken to cross the preset mark of 7.0 cm.

The Cy (Curly) mutation causes an upward curvature of the wings, which suggests a functional impact on flight performance due to altered wing morphology. Hence, the l(2)gl/Cyo adult flies are not taken as controls for this study65,66,67.

Tumor recurrence analysis

Late 20th -day adult l(2)gl/l(2)gl flies were set up for synchronized culture, and a batch of five flies from each sex was transferred into a formulated drug (250 µM DHA) containing medium and control standard food for mating. The same batch of adult flies was made to transfer after the subsequent four days, and the number of transferred flies was noted for the life expectancy of these straight-wing phenotype flies for tumor recurrence. In the meantime, on the post-20th day of mating, the offspring eclosed, and any phenotypic change was noted to check for tumorigenesis to conclude whether the flies needed to be fed continuously with the drug to reduce the tumor size, or they could survive long once treated with 250 µM DHA. For this drug withdrawal assessment, an experimental setup was created. Adult homozygous l(2)gl/l(2)gl flies initially rescued via DHA treatment were used as the founder population (Gen-0, gray box)(refer to Fig. 5B). These rescued adults were divided into two experimental conditions: (i) drug withdrawal (− DHA; green boxes) and (ii) exposure to the drug (+ DHA; purple boxes). Both groups were allowed to lay eggs for 24–48 h. From each condition, the larval progeny were collected at the late third-instar stage (constituting the Gen-I larvae), and the brain tissue and the wing imaginal discs were dissected and processed for tumor quantification. In parallel, vials were monitored for the eclosion of Gen-I homozygous l(2)gl/l(2)gl adult flies. The presence and proportion of viable adults in each treatment condition were recorded to measure developmental rescue and survival. Subsequently, adult homozygotes eclosing from each drug-withdrawal group (green box) were again split into new + DHA and − DHA subcultures, initiating the next generation (Gen-I → Gen-II, and so on). There were three other control larval groups for this experiment: (i) Oregon R+ (wild-type), (ii) untreated l(2)gl/l(2)gl (generated from the self-cross of l(2)gl/Cyo flies in standard food medium) and, (iii) rescued l(2)gl/l(2)gl group with continuous exposure to DHA treatment (generated from the self-cross of l(2)gl/Cyo flies in DHA containing food medium). This generational cycle of dissection at the larval stage and survival assessment at the adult stage was carried through four consecutive generations (Gen-I to Gen-IV), under strictly parallel treatment conditions.

This generational paradigm enabled systematic evaluation of tumor relapse dynamics, drug dependence, and long-term maintenance of the rescued phenotype under conditions of drug withdrawal versus continuous exposure.

Statistical analysis

The data are represented as mean ± SEM. The two groups were compared using the Students’ two-tailed t-test. Statistical analysis was performed using GraphPad Prism software. A p-value < 0.05 was considered significant for all data sets (*p < 0.05, **p < 0.01, and *** p < 0.001).