Ketogenic diet reduces early mortality following traumatic brain injury in Drosophila via the PPARγ ortholog Eip75B

Traumatic brain injury (TBI) is a common neurological disorder whose outcomes vary widely depending on a variety of environmental factors, including diet. Using a Drosophila melanogaster TBI model that reproduces key aspects of TBI in humans, we previously found that the diet consumed immediately following a primary brain injury has a substantial effect on the incidence of mortality within 24 h (early mortality). Flies that receive equivalent primary injuries have a higher incidence of early mortality when fed high-carbohydrate diets versus water. Here, we report that flies fed high-fat ketogenic diet (KD) following TBI exhibited early mortality that was equivalent to that of flies fed water and that flies protected from early mortality by KD continued to show survival benefits weeks later. KD also has beneficial effects in mammalian TBI models, indicating that the mechanism of action of KD is evolutionarily conserved. To probe the mechanism, we examined the effect of KD in flies mutant for Eip75B, an ortholog of the transcription factor PPARγ (peroxisome proliferator-activated receptor gamma) that contributes to the mechanism of action of KD and has neuroprotective effects in mammalian TBI models. We found that the incidence of early mortality of Eip75B mutant flies was higher when they were fed KD than when they were fed water following TBI. These data indicate that Eip75B/PPARγ is necessary for the beneficial effects of KD following TBI. In summary, this work provides the first evidence that KD activates PPARγ to reduce deleterious outcomes of TBI and it demonstrates the utility of the fly TBI model for dissecting molecular pathways that contribute to heterogeneity in TBI outcomes.


Introduction
Traumatic brain injury (TBI) is a major health issue worldwide [1]. It is a leading cause of disability and death, and its clinical management is challenging because the physical, behavioral, cognitive, and emotional sequelae are highly variable. Variation in sequelae among TBI patients stems from heterogeneity of primary injuries to the brain as well as heterogeneity of Eip75B, an ortholog of PPARγ, was necessary to mediate the beneficial effect of KD on the MI 24 .

Fly lines and culturing
Flies were maintained in humidified incubators at 25˚C on solid CMYD. DGRP lines and Eip75B mutant fly lines were obtained from the Bloomington Stock Center (Indiana University).

MI 24 and lifespan assays
Flies were injured using a HIT device as described by Katzenberger et al. [26,27]. Vials containing 60 flies at 0-7 days old were injured by 4

Ketogenic diet following TBI reduces the incidence of early mortality
We previously found that the diet consumed directly after TBI in Drosophila substantially affects the incidence of early mortality [28,29]. Flies fed cornmeal-molasses-yeast diet (a standard laboratory fly diet) or simple carbohydrates (i.e., sucrose, glucose, and fructose) during the 24 h following TBI have a significantly higher MI 24 than flies fed water. To further explore the effect of diet on the MI 24 , we examined different concentrations of cornmeal-molassesyeast diet (CMYD), yeast diet (YD, S. cerevisiae), and ketogenic diet (KD, a commercial mouse ketogenic diet). Based on caloric content, CMYD is high in carbohydrate and low in protein and fat, YD is high in carbohydrate and protein and low in fat, and KD is high in fat and low in carbohydrate and protein (Table 1). Diets were dissolved in water at 0.5 g/ml, serially diluted by 2-fold increments in water down to 0.0625 g/ml, and 200 μl was absorbed onto a filter paper disc that was placed at the bottom of a vial. 0-7 day old, mixed sex w 1118 flies cultured on solid CMYD (i.e., undiluted CMYD), were subjected to four strikes from the HIT device with 5 min between strikes and transferred to vials with diets at different concentrations. We found that the MI 24 increased to a similar extent with increasing concentrations of CMYD and YD ( Fig 1A). In contrast, the MI 24 was not affected by increasing concentrations of KD. We also examined the effect of CMYD, YD, and KD as well as glucose and sucrose at approximately the same caloric content (0.3 cal/200 μl) ( Table 1). MI 24 values were similar for flies fed CMYD, YD, glucose, or sucrose and they were significantly higher than the MI 24 of flies fed water ( Fig 1B). In contrast, the MI 24 of flies fed KD was the same as that of flies fed water. These data support our prior finding that ingestion of carbohydrate after TBI increases the MI 24 and demonstrate that ingestion of fat after TBI does not increase the MI 24 compared with water.
An alternative interpretation of the data in Fig 1A and 1B is that flies did not consume KD, suggesting that starvation and water have equivalent effects on the MI 24 . To test this possibility, we determined the MI 24 of 0-7 day old, mixed sex w 1118 flies that were starved by placing them in vials with a dry filter paper disc following TBI. In contrast with flies fed KD, the MI 24 of starved flies was substantially higher than that of flies fed water (Fig 1C), demonstrating that consuming KD rather than starvation was beneficial. As an additional approach to test if flies consumed KD, we examined the lifespan of uninjured, mixed sex w 1118 flies cultured on solid CMYD to 0-7 days old and thereafter on water or 0.3 cal/200 μl CMYD or KD. The median and maximum lifespans of flies cultured on KD (14.7 ± 1.1 days and 37 days, respectively) were longer than those of flies cultured on water (10.6 ± 0.1 days and 13 days, respectively) and shorter than those of flies cultured on CMYD (23.4 ± 0.6 days and 74 days, respectively), indicating that flies examined in Fig 1A and 1B consumed KD ( Fig 1D). Further support for this conclusion is provided in Fig 4B.

Ketogenic diet is similarly beneficial following TBI in females versus males and in different genetic backgrounds
To investigate whether KD has sex-specific effects on TBI outcomes, we compared effects of KD, water, and solid CMYD on the MI 24 of 0-7 day old female, male, and mixed sex w 1118 flies. In each case, solid CMYD resulted in a significantly higher MI 24 than both water and KD, and water and KD had equivalent MI 24 values (Fig 2A). Moreover, comparisons of male, female, and mixed sex flies, revealed that KD as well as solid CMYD and water had similar effects on the MI 24 . Therefore, sex does not alter the effects of KD, water, and solid CMYD on secondary injury mechanisms that cause early mortality following TBI.

PLOS ONE
We previously found that the MI 24 of flies fed solid CMYD varied significantly among fly lines with different genetic backgrounds, including inbred fly lines from the Drosophila Genetic Reference Panel (DGRP) [26,28,29,34]. To determine the extent to which genetic background affects the MI 24 of flies fed KD following TBI, we examined three lines from the DGRP that have different MI 24 values when fed solid CMYD following TBI [28,29]. We fed 0-7 day old, mixed sex flies 0.3 cal/200 μl KD or CMYD for 24 h following TBI and determined the MI 24 of each line. For all three DGRP lines, the MI 24 of flies fed KD was lower than the MI 24 of flies fed CMYD (Fig 2B). Moreover, the w 1118 line and the DGRP lines fed KD had comparable MI 24 values, whereas these values varied among the same fly lines when fed CMYD (Figs 1B and 2B). These results indicate that the beneficial effect of KD on the MI 24 does not depend on the starting value of the MI 24 (on CMYD) in different fly lines, leading to uniformly low MI 24 values for flies fed KD. However, it does appear that the beneficial effect of KD has a limiting threshold beyond which it cannot act further, resulting in a proportionally greater rescuing effect for lines with higher MI 24 values on CMYD.

Ketogenic diet following TBI has beneficial long-term effects on lifespan
For both humans and flies, individuals that survive TBI manifest a variety of long-term consequences, including reduced lifespan, as a result of secondary injuries triggered by primary injuries to the brain. The exact connection between primary injuries and secondary injuries is complex, and the details are still poorly understood. The fact that the MI 24 is reduced in flies fed KD immediately after TBI, raises the question of whether the beneficial effects of KD extend to longer-term pathological consequences of TBI. We examined this possibility using lifespan as a readout. Lifespan was determined for 0-7 day old, mixed sex w 1118 flies fed 0.3 cal/200 μl KD or CMYD for 24 h following TBI with surviving flies subsequently cultured on solid CMYD. As we reported previously, flies fed CMYD for 24 h after injury had a reduced lifespan relative to uninjured controls (Kaplan-Meier Fisher's Exact Test, p = 4.1X10 -9 at 50%) (Fig 3) [26,29]. The same was true for flies fed KD (Kaplan-Meier Fisher's Exact Test, p = 1.7X10 -7 at 50%). However, notably, injured flies fed KD rather than CMYD for 24 h after injury had a significantly longer median lifespan (40.3 ± 0.2 days vs. 37.8 ± 0.28 days, Kaplan-Meier Fisher's Exact Test, p = 1.0X10 -4 at 50%). Moreover, the difference in median lifespan between injured flies and uninjured controls was much narrower for KD-fed than for CMYDfed flies (40.3 ± 0.2 days vs. 44.2 ± 0.8 days for KD; 37.8 ± 0.28 days vs. 48.2 ± 1.0 days for CMYD). Thus, flies that avoid early mortality following TBI because of the beneficial effects of KD during a 24 h window after primary injuries continue to manifest long-term benefits of this diet weeks later.

Beneficial effects of KD on early mortality are mediated by the PPARγ ortholog Eip75B
In mammals, the mechanism of action of KD is mediated by the transcription factor PPARγ, which has neuroprotective effects in a number of progressive neurological disorders, including TBI [19,20]. Because KD exerts a protective effect following TBI in flies as well as mammals, we hypothesized that the underlying mechanism is conserved as well. If so, the protective effect in flies should depend on the transcription factor Eip75B (ecdysone-induced protein 75B), the Drosophila ortholog of PPARγ. The orthologous relationship is inferred both from amino acid sequence identity (i.e., Eip75B is the most significant match to human PPARγ in a BLAST search of the Drosophila proteome) and from common activation by the PPARγ agonist pioglitazone [37][38][39]. Under this hypothesis, mutational loss of Eip75B should result in loss of the beneficial effect of KD. Thus, we examined the effect of water and 0.3 cal/200 μl KD on the MI 24 of 0-7 day old, mixed sex Eip75B mutant flies. Three hypomorphic alleles of Eip75B (Eip75B MI04895 , Eip75B KG04491 , and Eip75B BG02576 ) containing transposon insertions within the transcribed region were examined ( Fig 4A). As in Fig 1, the MI 24 was comparably low in control w 1118 flies fed either water or KD (ordinary one-way ANOVA with Dunnett's Multiple Comparison test, p = 0.835) (Fig 4B). In contrast, although water-fed Eip75B KG04491 , Eip75B BG02576 , and Eip75B MI04895 /Eip75B BG02576 flies still had low MI 24 values comparable to that of water-fed w 1118 controls, MI 24 values were higher in KD-fed Eip75B mutants (ordinary one-way ANOVA with Dunnett's Multiple Comparison test, p = 0.078, p = 0.033, and p = 0.006, respectively), indicating that the beneficial effect of KD was impaired in these mutants. Furthermore, higher MI 24 values in KD-fed versus water-fed mutants provides further evidence that flies consumed KD. The beneficial effect of KD was, however, retained in Eip75B MI04895 homozygotes (ordinary one-way ANOVA with Dunnett's Multiple Comparison test, p = 0.999), which we attribute to a presumptive weaker loss of function of Eip75B caused by this mutation. Eip75B MI04895 only disrupts three of the seven Eip75B pre-mRNA isoforms, whereas Eip75B KG04491 and Eip75B BG02576 disrupt four and five isoforms, respectively (Fig 4A). Thus, while it remains possible that differences in genetic background underlie differences in MI 24 values for Eip75B mutant flies fed water versus KD, the data support the conclusion that activation of Eip75B/PPARγ by KD triggers mechanisms that reduce early mortality following TBI.

Discussion
TBI patients face a spectrum of neurobehavioral sequelae initiated by primary injuries to the brain and mediated by the interplay of genetic and environmental factors that control pathophysiological cascades. Here, we discovered that an interaction between the genetic factor Eip75B/PPARγ and the environmental factor KD affects early mortality in a fly TBI model. In particular, whereas flies fed high-carbohydrate CMYD or YD exhibited a dose-dependent increase in early mortality compared with flies fed water, flies fed high-fat KD showed no increase in early mortality (Fig 1A and 1B). The beneficial effect of KD on early mortality was equivalent in males and females, conserved in different genetic backgrounds, and had longterm beneficial effects on lifespan as well (Figs 2 and 3). However, the beneficial effect of KD on early mortality was diminished in flies mutant for Eip75B, a transcription factor orthologous to mammalian PPARγ, suggesting that KD exerts its effect through Eip75B (Fig 4B). These data provide a mechanistic link between KD and PPARγ in modifying TBI outcomes and demonstrate the utility of the fly TBI model for dissecting interactions between genetic and environmental factors that affect TBI outcomes.

The KD-Eip75B/PPARγ pathway may reduce early mortality following TBI by inhibiting Relish/NF-κB
Our data suggest that KD reduces early mortality following TBI by activating Eip75B/PPARγ. However, it remains to be determined what occurs downstream of Eip75B/PPARγ to exert this effect. One possibility is that Eip75B/PPARγ controls expression of genes involved in inflammation. In mammals, activation of PPARγ by dietary fatty acids mitigates neuroinflammation by inhibiting NF-κB, a transcriptional activator of cytokine, chemokine, and adhesion genes downstream of Toll-like receptor (TLR)/Interleukin-1 receptor (IL-1R) and Tumor necrosis factor-α receptor (TNFR) innate immune response signaling pathways [19]. PPARγ inhibits NF-κB by a variety of mechanisms, including ubiquitination and degradation, export from the nucleus, competition for cofactors, and steric inhibition of DNA binding [40]. In mammalian TBI models, reduced NF-κB activity resulting from treatment with the PPARγ agonist pioglitazone or other pharmacological agents improves outcomes [22][23][24][25][41][42][43][44][45][46]. Reducing NF-κB activity also improves TBI outcomes in flies [33]. Heterozygosity for a null mutation of Relish, one of three NF-κB genes in Drosophila, reduces early mortality and increases lifespan following TBI. Relish functions in the Immune-deficiency (Imd) pathway that is homologous to the TNFR pathway in mammals and controls transcription of numerous antimicrobial peptide genes (AMPs) that produce resistance to infection [47,48]. A confirmed transcriptional target of Relish in TBI is the AMP gene Metchnikowin (Mtk), which when mutated reduces early mortality and increases lifespan following TBI [32]. Thus, KD-mediated activation of Eip75B/ PPARγ may reduce early mortality following TBI by inhibiting Relish/NF-κB. This could be tested in the fly TBI model by examining effects of KD and pioglitazone on the MI 24 , lifespan, and expression of AMP genes in wild type as well as Relish and Mtk mutant flies.

KD and water appear to reduce early mortality following TBI by different mechanisms
Genetically diverse fly lines fed KD or water following TBI consistently had a lower incidence of early mortality relative to flies fed high-carbohydrate diets (Figs 1A and 1B and 2) [29]. These data suggest that KD and water might activate the same protective pathways. Water is a fasting condition where the amount of available carbohydrate is decreased, forcing a switch to the use of fatty acids as a nutrient supply through beta-oxidation and ketogenesis [49]. Ketogenesis converts acetyl-CoA to ketone bodies (e.g., β-hydroxybutyrate, acetone, and acetoacetate) that are used by the brain and other tissues to produce energy. Flies with impaired mitochondrial ATP synthase activity produce elevated amounts of β-hydroxybutyrate, indicating that ketogenesis operates in flies [50]. Additionally, aggressive behaviors and early mortality induced by TBI in flies are reduced when flies are raised on high-carbohydrate diet supplemented with β-hydroxybutyrate relative to high-carbohydrate diet alone, indicating that β-hydroxybutyrate operates in the fly TBI model [35]. Nonetheless, several lines of evidence argue that KD and water act by distinct mechanisms to reduce early mortality following TBI. First, the MI 24 of Eip75B mutants differed depending on whether they were fed KD or water, indicating that water acts independently of Eip75B/PPARγ (Fig 4B). Second, we previously found that flies fed water versus CMYD exhibited increased expression of AMP genes following TBI, suggesting that the beneficial effects of water are not mediated by inhibition of Relish, which activates the transcription of AMP genes [29]. Third, while heterozygosity for a mutation of Relish reduced the incidence of early mortality for flies fed CMYD following TBI, it did not affect the incidence of early mortality for flies fed water following TBI, suggesting that water functions either downstream or independently of Relish [33].
In conclusion, our observations indicate that KD signals through PPARγ to improve TBI outcomes in flies. Thus, the fly TBI model offers considerable potential for understanding the cellular and molecular mechanisms that underlie the beneficial effects of KD and may ultimately facilitate development of therapeutic intervention for TBI in humans.