TRAUMATIC brain injury (TBI) affects 1.7 million people in the USA annually and is one of the leading causes of death and disability worldwide (Dixon 2017; Maas et al. 2017). Despite its prevalence, treatments have not yet been developed that effectively reduce consequences of TBI, which can include cognitive impairment, emotional lability, post-raumatic epilepsy, and death (Zaloshnja et al. 2008; Bazarian et al. 2009; Diamond et al. 2015; Tortella 2016; Juengst et al. 2017; DeWitt et al. 2018; Ng and Lee 2019). Neurologic damage following TBI results from both primary and secondary injuries. Primary injures are inflicted by mechanical forces on the brain that compress, stretch, displace, or tear blood vessels, neurons, and glia. In contrast, secondary injuries result from activation of cellular and molecular pathways in the minutes to years following the initial physical insult and can occur at and beyond the site of the primary injury (Pearn et al. 2017; Ng and Lee 2019). Primary injuries are mainly mitigated through preventive measures such as helmets, but it remains to be determined how best to ameliorate secondary injuries (Hatton 2001; Somayaji et al. 2018).

Neuroinflammation is an immediate cellular response to TBI that plays both pathological and protective roles (Simon et al. 2017). This dichotomy is illustrated by studies of transcription factors of the nuclear factor-κB (NF-κB) family, which are major transcriptional activators of inflammatory genes. Direct inhibition of NF-κB in neurons through genetic means increases acute mortality and neurological deficits following TBI in mice (Mettang et al. 2018). In contrast, indirect inhibition of NF-κB by a variety of pharmacological agents, including metformin, ghrelin, resveratrol, ethylpyruvate, curcumin, allyl isothiocyanate, omega-3 polyunsaturated fatty acid, and pioglitazone, improves outcomes following TBI in mice or rats (Laird et al. 2010; Su et al. 2011; Feng et al. 2016; Chen et al. 2018; Tao et al. 2018; Caglayan et al. 2019; Deng et al. 2020; Shao et al. 2020). Adding further complexity to the neuroinflammatory response, individual transcriptional targets of NF-κB, including cytokines and chemokines, differentially contribute to TBI outcomes (Shohami et al. 1997; Scherbel et al. 1999; Sullivan et al. 1999; Ziebell and Morganti-Kossmann 2010; Di Battista et al. 2016). Therefore, NF-κB is a prime candidate target for TBI therapies, but development of maximally effective therapies will require further investigation of the functional relationship between NF-κB activation and other factors that affect progression of secondary injuries.

Drosophila melanogaster encodes three NF-κB homologs, Dorsal (Dl), Dorsal-related immunity factor (Dif), and Relish (Rel) (Lemaitre and Hoffmann 2007; Ganesan et al. 2011). Dl functions in formation of dorsal-ventral polarity in embryos, whereas Dif and Rel function in cellular and humoral innate immunity in defense against pathogen infection. Dif functions in the Toll pathway and Rel functions in the Immune-deficiency (Imd) pathway to activate transcription of genes such as antimicrobial peptide (AMP) genes that produce resistance to infection. While the Toll and Imd pathways have beneficial effects in the context of infection, chronic activation of either pathway can promote neurodegeneration (Tan et al. 2008; Chinchore et al. 2012; Petersen et al. 2012; Petersen and Wassarman 2012; Cao et al. 2013; Kounatidis et al. 2017). Additionally, chronic activation of the Imd pathway through a variety of means reduces lifespan, whereas mutation of Rel increases lifespan and suppresses neurodegeneration (Cao et al. 2013; Kounatidis et al. 2017). In mammals, Toll-like receptor (TLR)/Interleukin-1 receptor (IL-1R) pathways and the Tumor necrosis factor-α receptor (TNFR) pathway are homologous to the Toll and Imd pathways, respectively (Tanji and Ip 2005). Studies of TLR and TNFR pathways in mammalian TBI models provide an equally complex picture as studies of NF-κB. Knockout of TLR2 or TLR4 in mice improves neurological function following TBI, whereas knockout of TNFR in mice increases lesion volume following TBI (Sullivan et al. 1999; Zu and Zha 2012; Laird et al. 2014; Jiang et al. 2018; Shi et al. 2019). Because innate immune pathways are evolutionarily conserved between flies and mammals, studies of flies, which have a simpler immune system and a more extensive experimental toolbox than rodents, may provide novel insights into roles that NF-κB-mediated TLR and TNFR pathways play in determining TBI outcomes (Dhankhar et al. 2020).

We and others have developed fly TBI models to investigate the pathways that control development of secondary injuries following TBI (Katzenberger et al. 2013, 2015a, 2015b; Barekat et al. 2016; Sen et al. 2017; Lee et al. 2019; Putnam et al. 2019; Sanuki et al. 2019; Shah et al. 2019; Saikumar et al. 2020). Our model uses a spring-based instrument called a High-impact trauma (HIT) device to inflict physical trauma by rapid acceleration–deceleration forces (Katzenberger et al. 2013, 2015a,b,c, 2016; Fischer et al. 2018; Swanson et al. 2020). Strikes from the HIT device affect both the fly head and body, but the main pathologies appear to be primarily driven by injuries to the brain. Flies that have sustained a TBI share behavioral and physiologic characteristics with mammals, including temporary incapacitation, ataxia, transient hyper-glycemia, intestinal barrier dysfunction, progressive neurodegeneration, and reduced lifespan. Additionally, like mammals, the innate immune response in flies is rapidly and persistently activated following TBI (Katzenberger et al. 2013, 2015a,b; Barekat et al. 2016; Sanuki et al. 2019). Expression of AMP genes substantially increases within 30 min after TBI and is sustained for >24 hr (Katzenberger et al. 2016). Furthermore, AMP gene expression contributes to TBI outcomes, as mutation of some individual AMP genes can suppress or enhance outcomes (Swanson et al. 2020).

Here, we show that Rel is a dose-dependent modifier of TBI outcomes. Proteomics analysis revealed that the relative abundance of Rel protein increased in fly heads shortly after a primary injury. Genetic analysis revealed that heterozygosity, but not homozygosity, for a null mutation of Rel (Reldel) reduced detrimental consequences of TBI. Finally, gene expression analysis differentiated transcriptional targets of the Toll and Imd pathways following TBI and identified gene expression changes that may underlie the beneficial effects of heterozygosity for Reldel.

Materials and MethodsFly lines and culturing

Flies were maintained at 25° on cornmeal molasses food containing 30 g Difco granulated agar (Becton-Dickinson, Sparks, MD), 44 g YSC-1 yeast (Sigma, St. Louis, MO), 328 g cornmeal (Lab Scientific, Highlands, NJ), 400 ml unsulfured Grandma’s molasses (Lab Scientific), 3.6 L water, 40 ml propionic acid (Sigma), and tegosept (8 g methyl 4-hydroxybenzoate in 75 ml of 95% ethanol) (Sigma). In Figure 4, flies were fed water by placing 200 µ liters of water on a filter paper disk at the bottom of a vial. w1118 flies used for proteomics analyses were also used in our prior analyses of TBI (Katzenberger et al. 2013; 2015a; 2016). RelE20 flies and w1118 flies to which RelE20 flies were backcrossed, were provided by Stanislava Chtarbanova (University of Alabama, Alabama). Drosophila Genetic Reference Panel (DGRP) lines were obtained from the Bloomington Stock Center (Mackay et al. 2012).

Fly head collection for proteomics analysis

For proteomics analysis of fly heads after TBI, 0- to 7-day-old male w1118 flies were subjected to four strikes from the HIT device, with 5 min between strikes. Following injury, flies were transferred to cornmeal molasses food at 25° until the time of collection: immediately (0), 0.5, 1, 2, 4, 6, 8, 12, 16, and 24 hr. As controls, 0- to 7-day-old w1118 male uninjured flies were cultured under the same conditions and collected at the same time points. Fly heads were removed from bodies by freezing flies in liquid nitrogen, vortexing the frozen flies, and separating heads from bodies by passing them through a sieve. Head samples from injured and control uninjured flies were stored at −80° until all samples were ready for analysis. Each sample contained ∼200 heads, and three independent samples were prepared for each condition.

Preparation of fly heads for proteomics analysis

Fly head samples were resuspended in 200 µl of 6 M guanidine hydrochloride in 100 mM tris(hydroxymethyl)aminomethane (Tris) pH 8.0 and maintained at 4°. Rupturing of heads was performed by bead-beating using 2.8-mm metal beads and four rounds of beating at 30 Hz for 5 min with 1 min of rest between each round at 4°. Cell lysis occurred with two cycles of heating at 100° for 5 min followed by 5 min of cooling at ambient temperature. The protein concentration of each sample was determined by the BCA Protein Assay Kit (Pierce, Rockford, IL). Proteins were precipitated by addition of methanol up to a final concentration of 90% by volume followed by centrifugation for 20 min at 15,000 × g. Supernatant was removed, and protein pellets were air-dried at room temperature.

Protein digestion followed previously developed protocols from Hebert et al. (2014) and Richards et al. (2015) with the following specific changes. Protein pellets were resuspended in 50 µl of lysis buffer containing 8 M urea, 100 mM Tris (pH 8.0), 10 mM TCEP, and 40 mM chloroacetamide to denature, reduce, and alkylate proteins. Sonication for 10 min ensured that proteins were in solution. The resuspended protein solution was diluted to 1.5 M urea with 100 mM Tris (pH 8.0) and vortexed for 30 sec. Endoproteinase LysC was added at a 100:1 ratio, trypsin was added at a 50:1 protein to enzyme mass ratio, and samples were incubated for 12 hr at room temperature. After incubation, each sample was prepared for desalting using a 96-well Strata polymeric reversed phase 10-mg SPE (styrene divinylbenzene) cartridge. Cartridge wells were primed with 1 ml of acetonitrile (ACN) then followed by 1 ml of 0.1% trifluoroacetic acid (TFA). Each sample was acidified with TFA to a final pH of 2.0 or less and then centrifuged for 15 min at 2000 X g, spinning all nonprotein material to the bottom of the vial. Acidified samples were loaded on to the cartridge, washed with 1 ml of 0.1% TFA, and eluted with 600 µl of 80% ACN, 0.1% TFA into a clean 96-well plate to be dried. Eluted peptides were vacuum dried and resuspended in 0.2% formic acid. Peptide mass was assayed with the NanoDrop Microvolume UV-Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA).

LC-MS/MS analysis

For each analysis, 1 µg of peptides was loaded onto a 75 × 360 µm (ID × OD) 30-cm-long column with an integrated electrospray emitter packed with a 1.7-m-particle-size C18 BEH column (Waters, Milford, MA) (Shishkova et al. 2018). The mobile phases used were as follows: phase A consisted of 0.2% formic acid, and phase B consisted of 0.2% formic acid in 70% ACN. Peptides were eluted with a gradient of acetonitrile increasing from 0 to 55% B over 100 min followed by a 1-min increase to 100% B, 2 min wash at 100% B, 3 min descend to 0% B and a final 10 min of equilibration in 100% A for a total of a 120-min analysis.

Eluted peptides were analyzed with an Orbitrap Fusion Lumos Tribrid Mass Spectrometer (Thermo Fischer Scientific, Waltham, MA). Survey scans were performed at a resolution of 240,000 in the range of m/z 300 to 1,350 and 1e+06 automatic gain control (AGC) target using quadrupole isolation of 0.7 Da for data-dependent tandem MS (MS/MS) scans. MS/MS scans were collected using the top-speed mode with 1-sec cycle and dynamic exclusion of 15 sec on precursors with charge states 2 to 4. Isolated precursors were fragmented by higher-energy collisional dissociation with a normalized collisional energy of 25%. Mass analysis of product ions was performed in the ion trap using the “turbo” scan rate over the mass range of 150–1400 m/z with a maximum inject time of 17 ms and the normalized AGC target of 3e+4 (Trujillo et al. 2019).

Analysis of proteomics data

Thermo RAW files were processed using MaxQuant (version 1.6.0.16) (Tyanova et al. 2016). Searches were performed against a target decoy database of D. melanogaster, including isoforms (UniProt; downloaded August 8, 2019), using the Andromeda search engine (Tyanova et al. 2016). Quantitation was performed as label-free quantitation (LFQ) with an LFQ minimum ratio of 1. “Match between runs” option was enabled. Search parameters included fixed modification for carbamidomethylation of cysteine residues, variable modification for oxidation of methionine, N-terminal acetylation, and a maximum of two missed cleavages. The peptide spectral match false-discovery rate (FDR) and protein FDR were both set to 1%.

Using RStudio (v 1.2.5033), LFQ values and protein intensities for each sample were log2 transformed and filtered to retain proteins that had no missing values across three replicates and fulfilled a >50% cutoff of measurements across all the samples. A total of 277,020 protein measurements were obtained from the fly heads, resulting in 4860 unique proteins with an interreplicate coefficient of variation of 14.49% for 20 different conditions that differed in time as well as TBI status. Normalization across the three batches was performed by transforming each protein by the mean average of each respective protein at time zero by condition. The average and standard error of the mean (SEM) were calculated from normalized protein LFQ values across the three replicates. Due to biological variability among the fly replicates, we did not correct P-value calculations and instead looked for the proteins that had a clear change in abundance between uninjured and injured flies across the time points. Fold changes were determined for each time point and are relative to control time zero. We employed a partial least square discriminant analysis using R (pls package) to reveal the top 25 discriminating proteins between control and TBI samples regardless of time point.

Generation of Reldel by CRISPR/Cas9 gene editing

A Rel null allele was generated by CRISPR/Cas9 homology directed repair using the pHD-DsRed-attP vector (Plasmid #51019; Addgene) as the donor template (Gratz et al. 2014). One-kilobase left and right homology arms were generated immediately flanking the Rel cleavage sites (5′ cleavage site: −99 bp; 3′ cleavage site: +4219 bp, relative to the transcription start site). The homology arms were amplified via PCR with 20-bp overhangs homologous to the pHD-DsRed-attP vector, and the vector was amplified via PCR with 20-bp overhangs homologous to the homology arms. PCR products were purified with the Wizard SV Gel and PCR cleanup system (Promega, Madison, WI). The editing plasmid was constructed by Gibson assembly of the four DNA fragments (Gibson et al. 2009). Synthetic guide RNA (sgRNA) plasmids were generated using the pU6-3 gRNA vector generously provided by the O’Connor-Giles lab (Brown University). Guide sites were selected using flyCRISPR optimal target finder (Gratz et al. 2014). Both sgRNAs had no predicted off-targets. sgRNA plasmids were constructed by amplifying sgRNAs and the pU6-3 gRNA vector by PCR with 20 bp overhangs. PCR products were purified with the Wizard SV Gel and PCR cleanup system (Promega). sgRNA plasmids were assembled using the Kinase Ligase Dpn (KLD) mix (New England Biolabs, Ipswich, MA). sgRNA plasmid sequences were confirmed by sequencing. Plasmid DNA was prepared for injection by a mini- or midi-prep kit (Qiagen, Hilden, Germany) and was injected into vas-Cas9(II) embryos (Stock #56552; Bloomington Drosophila Stock Center) by BestGene. Genetically identical control lines were developed from the injection stock. Positive transformants were identified through the fluorescent DsRed eye marker. Cas9 was removed from the stocks by backcrossing to the injection stock that did not contain Cas9.

Mortality and lifespan assays

Flies were injured using a HIT device as described by Katzenberger et al. (2013). Flies were injured by four strikes at 10-min intervals with the spring deflected to 90°. Except where stated otherwise, TBI was inflicted using this protocol. All vials contained 60 flies (30 males and 30 females) at 1–7 days old. We used a 1- to 7-day age range, as opposed to a smaller age range, because it was easier to collect a sufficient number of flies for the experiments, and we previously found that sample-to-sample variability of the MI24 of 1- to 7-day-old flies was similar to that of smaller age ranges (Katzenberger et al. 2013). The average mortality at 24 hr for each uninjured fly line did not exceed 2% on food and 5% on water, in line with previous studies using this model (Katzenberger et al. 2013, 2015a, 2016; Fischer et al. 2018). The exception was the RelE20 line, which had an average uninjured mortality of 4.3% on food and 8.7% on water.

Lifespan was determined using 2- to 8-day-old flies that survived 24 hr following TBI. At least four vials containing 20 flies (10 males and 10 females) were examined per condition, and each condition was independently repeated three times (n ≥ 120 per sex). Vials were maintained at 25° and scored for survival every 3 days. Percent survival was averaged among vials for each condition.

qRT-PCR analysis

Total RNA was extracted from 20 uninjured or injured whole male flies recovered on food for 0, 2, 4, 6, or 8 hr following TBI. RNA extraction was performed using Trizol (Invitrogen, Carlsbad, CA), according to a modified protocol described by Bogart and Andrews (2006). RNA purification was performed using the RNeasy Mini Kit and RNase-Free DNase (Qiagen, Hilden, Germany). For each sample, 1 µg of RNA was reverse transcribed using the iSCRIPT complementary DNA synthesis kit (Bio Rad, Hercules, CA). Quantitative PCR was performed using iTaq Universal SYBR Green SuperMix (Bio Rad) and the Bio Rad CFX96 Real-Time PCR Detection System. Biological replicates of each condition were performed in triplicate, and technical replicates were performed in duplicate. Primer sequences are shown in Supplemental Material, Table S6.

Statistical analyses

All data are presented as means ± SEM. A two-way ANOVA with Tukey’s post hoc analysis was used to compare MI24 outcomes of Rel mutants to controls on food and water (Figures 4, A and B). A two-tailed Student’s unpaired t-test was used to compare MI24 outcomes of RAL/+ and RAL/Reldel flies (Figure 4C). Statistical differences in survival were quantified using a log-rank test (Figure 5 and Figure S2). For comparisons across multiple groups for qRT-PCR a two-way ANOVA followed by a Tukey’s post hoc test was applied (α = 0.05, number of comparisons = 2) (Figure 3C and Figure 6). Statistical analysis of the qRT-PCR time course of control expression compared to +/Reldel flies was performed using a Student’s unpaired t-test for each time point (Figure 7). All statistical analyses were performed using GraphPad Prism 8.

RNA-seq

mRNA isolation from whole flies, construction of mRNA libraries, high-throughput sequencing, and analysis of RNA-seq data were performed as described in Katzenberger et al. (2016).

Data availability

All flies and reagents used in the study will be made publicly available. Proteomics data sets are available at ProteomeXchange via the PRIDE database (accession number PXD021869). RNA-seq data sets are available at GEO repository (accession number GSE157102). Figure S1 shows Multidimensional scaling (MDS) analysis of RNA-seq data sets described in Figure 8 and Table 1. Table S2 contains statistical analyses of data in Figure 5 that are summarized in Table S1. Table S3 lists genes categorized in Figure 8, A and B. Table S4 lists genes categorized in Figure 8, C and D. Table S5 lists genes categorized in Figure 8, E and F. Table S6 contains the sequence of primers used for qRT-PCR analysis. The authors affirm that all other data necessary for confirming the conclusions of the article are present within the article, figures, and tables. Supplemental material available at figshare: https://doi.org/10.25386/genetics.13072988.

Table 1 Gene targets of innate immune response pathways in injured and uninjured fliesResultsProteomics analysis identifies proteins whose expression discriminates between uninjured and injured fly heads

We used bottom-up proteomics analysis to investigate longitudinal changes in fly head protein expression following TBI. In this study, time and injury were the two factors taken into consideration for identifying biological processes that are perturbed during repair and degeneration associated with TBI. Whole fly heads from 0- to 7-day-old male w1118 flies were collected across 10 time points ranging from immediately following injury to 24 hr after injury. In addition, all time points included uninjured control flies of the same age, sex, and genotype, providing a proteomics abundance baseline for comparison. The Material and Methods section supplies a detailed description of how flies were injured, heads were collected, and proteins were extracted and analyzed. In brief, flies were injured, fly heads were collected from uninjured and injured flies by freezing them in liquid nitrogen and decapitation by vortexing; head tissues and cells were disrupted by bead-beating and heating, and proteins were extracted with methanol and enzymatically digested to peptides. Peptides were identified using nanoflow liquid chromatography coupled to tandem mass spectrometry (nLC-MS/MS), and quantified using LFQ (Cox et al. 2014).

A partial least squares discrimination analysis (PLS-DA) model revealed which proteins were responsible for the distinction between samples of uninjured and injured fly heads, regardless of time after TBI (Figure 1A). The list of proteins that drove separation between the two groups was enriched for those belonging to mitochondrial bioenergetics processes, neurosecretory cells, neuronal injury, markers of oxidative stress, and innate immune response pathways. The top 25 proteins in Component 1 of the PLS-DA included Rel and two other innate immune response proteins: Drs (Drosomysin) – an AMP; and GNBP-like3 (gram-negative bacteria binding protein-like3) – a pattern recognition receptor (Figure 1B). Thus, the proteomics analysis identified a protein expression signature that distinguishes uninjured from injured heads during the 24 hr after TBI.

Figure 1

Proteomics analysis of fly heads following TBI. (A) PLS-DA of protein expression in head of uninjured (control) and injured (TBI) 0- to 7-day-old male w1118 flies. Each dot represents an independent sample at 0, 0.5, 1, 2, 4, 6, 8, 12, 16, or 24 hr after TBI. (B) The top 25 genes that drove Component 1 separation. Volcano plots of protein abundance differences and P-values between uninjured and injured flies at (C) 4, (D) 6, and (E) 8 hr after TBI, respectively. Blue dots indicate proteins that met the P < 0.05 significance cutoff, green dots indicate proteins that met both the P < 0.05 significance cutoff and the twofold change cutoff, and gray dots indicate proteins that met neither cutoff.

Volcano plots of the data showed that the amount of only a few proteins was significantly different between uninjured and injured heads (Figure 1, C–E). We defined significance as having a log2 fold change >1 and P-value < 0.05. At 4, 6, and 8 hr after TBI, amounts of two, three, and nine proteins, respectively, were significantly altered between uninjured and injured fly heads. Rel was the only protein whose abundance was significantly changed at all three of these time points. Thus, TBI is associated with small, transient changes in protein expression in the head. Analysis of Rel expression over the 24-hr time course showed that its expression rapidly increased between 1 and 2 hr, peaked at 6 hr, and gradually declined thereafter (Figure 2). This time course is similar to the time course of expression of AMP genes as well as secondary injuries that lead to early mortality following TBI, suggesting that Rel is necessary for these events (Katzenberger et al. 2015a, 2016).

Figure 2

Rel protein expression increases in heads at 4–8 hr following TBI. Average expression of Rel protein in uninjured (control) and injured (TBI) flies at times after TBI normalized to time 0. Error bars indicate the SEM of three independent samples. *P < 0.05.

Generation of a null allele of Rel

To assess the necessity of Rel for secondary injuries following TBI, we used CRISPR/Cas9 gene editing by homology directed repair to generate a null allele, Reldel (Figure 3A). The rationale behind generating a new Rel mutant line was to maintain a uniform genetic background between control and Rel mutant flies, since genetic background significantly influences outcomes of TBI (Katzenberger et al. 2013; 2015a, 2016). We deleted the coding region of Rel in a control fly line and replaced it with DsRed, which encodes a red fluorescent protein that we used to monitor the genotype of Reldel flies. We confirmed the genotypes of +/Reldel and Reldel/Reldel flies by PCR analysis of genomic DNA extracted from whole flies (Figure 3B). In addition, qRT-PCR analysis of RNA extracted from whole flies showed that Rel mRNA expression was reduced in +/Reldel flies, although it did not reach statistical significance (P = 0.11), and eliminated in Reldel/Reldel flies that were either uninjured or injured (P < 0.0001) (Figure 3C). These data indicate that Reldel is a null allele. Expression of Rel was significantly increased in injured relative to uninjured +/+ and +/Reldel flies at 4 hr after injury, based on a Student’s t-test (P = 0.012 and 0.017, respectively). However, when this comparison was analyzed together with Rel expression from uninjured and injured +/Reldel and Reldel/Reldel flies using a two-way ANOVA and Tukey’s post hoc test, it did not meet the threshold for significance (Figure 3C). Nevertheless, these findings are in agreement with the proteomics data collected from fly heads that demonstrated an increase in the amount of Rel protein 4–8 hr after injury (Figures 1 and 2).

Figure 3

Generation of a null allele of Rel by CRISPR/Cas9 gene editing. (A) Diagrams of the wild-type Rel genomic locus and the Rel replacement donor that was used to generate the Reldel allele. Horizontal arrows indicate transcription start sites, vertical arrows indicate sites targeted by sgRNAs for cleavage by Cas9, numbered arrowheads indicate the direction and identity of primers used for PCR in panel B, and dotted lines indicate ∼1-kb homology arms on either side of the DsRed gene that were used for homology directed repair. The DsRed gene is described in more detail in Gratz et al. (2014). (B) PCR analysis of the Rel locus using genomic DNA from +/+, +/Reldel, and Reldel/Reldel flies. Primer sets refer to primers indicated in panel A. (C) qRT-PCR analysis of Rel mRNA from +/+, +/Reldel, and Reldel/Reldel flies that were collected 4 hr following TBI (+) or not (−) subjected to TBI. The amount of Rel mRNA was normalized to the amount of RpL32 mRNA. Significance was determined from ΔCt values by a two-way ANOVA with Tukey’s post hoc test. Error bars indicate the SEM of three independent samples. ****P < 0.0001.

+/Reldel flies are resistant to TBI-induced mortality

To investigate a role for Rel in acute outcomes of TBI, we measured the mortality index at 24 hr (MI24), which represents the percent mortality of injured flies normalized to the percent mortality of uninjured flies 24 hr following TBI. The MI24 is a reproducible measure, and it correlates with other consequences of TBI, including incapacitation and intestinal barrier permeability as well as the median lifespan of uninjured flies (Katzenberger et al. 2013, 2015a; Fischer et al. 2018). While mortality following TBI in flies should not be equated with mortality following TBI in mammals, the molecular and cellular events that lead to mortality in flies are likely to be conserved in mammals and to contribute to the pathophysiology of TBI in mammals.

We examined the MI24 of injured 1- to 7-day-old mixed-sex control flies (+/+), homozygous Rel mutant flies (Reldel/Reldel), and heterozygous Rel mutant flies (+/Reldel) that were generated by crossing +/+ and Reldel/Reldel flies. Flies were injured and incubated for 24 hr at 25° with cornmeal molasses food (hereafter referred to as food). +/+ and Reldel/Reldel flies had a similar MI24, whereas +/Reldel flies had a significantly lower MI24 than +/+ flies (Figure 4A). The same effects on the MI24 were observed with another Rel allele RelE20, which is likely to be a null allele because it lacks a translation start codon due to imprecise excision of a P element (Figure 4B) (Hedengren et al. 1999). Relative to control w1118 flies, the MI24 was not affected for homozygous RelE20/RelE20 flies and was significantly lower for heterozygous +/RelE20 flies. These data indicate that Rel is a dose-dependent enhancer of secondary injuries that promote early mortality following TBI – loss of a single copy of Rel is sufficient to reduce this mortality. To determine the effect of diet on the ability of heterozygosity for Rel to reduce mortality following TBI, we repeated the MI24 assay with flies fed water (i.e., fasted) instead of food following TBI. In accord with our prior findings, for every genotype (+/+, +/Rel, and Rel/Rel) and both Rel alleles (Reldel and RelE20), flies fed water had a lower MI24 than flies fed food (Katzenberger et al. 2015a, 2016). However, for both Reldel and RelE20, neither homozygous nor heterozygous flies fed water had a significantly altered MI24 relative to their respective control flies (Figure 4, A and B). These data suggest that a diet of food following TBI induces a Rel-dependent secondary injury mechanism, but water (i.e., starvation) does not. To determine if the ability of heterozygosity for Reldel to reduce the MI24 extends to genetic backgrounds beyond the one in which the Reldel mutation was generated, we crossed the Reldel mutation into different genetic backgrounds. We crossed females from +/+ or Reldel/Reldel lines to males from nine genetically diverse RAL lines from the Drosophila Genetic Reference Panel (DGRP) collection of ∼200 inbred, fully sequenced fly lines from a natural population (Mackay et al. 2012). The nine lines were selected because they represent the range of MI24s of the DGRP collection (Katzenberger et al. 2015a, 2016). We injured 1- to 7-day-old mixed-sex F1 progeny and determined the MI24. The nine RAL/+ control flies had MI24s that ranged from 41 to 58, which is not as large as the 7 to 58 range for the nine parental RAL/RAL lines, most likely because rather than having distinct genomes, half of each RAL/+ genome is shared, resulting also in loss of homozygosity for any unique recessive alleles affecting the MI24 in each RAL line (Katzenberger et al. 2015a). Moreover, five of the nine RAL/Reldel flies had a significantly lower MI24 than their respective RAL/+ control flies. There was no correlation between the MI24 and the ability of heterozygosity for Reldel to suppress mortality. The fact that heterozygosity for Reldel reduced early mortality following TBI to different extents in different genetic backgrounds indicates that the protective effect of the Reldel mutation is modified by polymorphisms in the genetic backgrounds.

Figure 4

Survival following TBI is increased in +/Reldel heterozygotes. (A) The MI24 of 1- to 7-day-old, mixed sex +/+, +/Reldel, and Reldel/Reldel flies fed cornmeal molasses food (food) or water for 24 hr following TBI. (B) The MI24 of 1- to 7-day-old, mixed sex +/+, +/RelE20, and RelE20/RelE20 flies fed food or water for 24 hr following TBI. (C) The MI24 of 1- to 7-day-old mixed sex F1 progeny from crosses between the indicated RAL lines and control flies (RAL/+) or Reldel/Reldel flies (RAL/Reldel) fed food for 24 hr following TBI. All bars represent the average of ≥3 biological replicates of ≥3 technical replicates. Each dot represents an independent sample of 60 flies. Error bars represent the SEM. Significance for panels A and B was determined using a two-way ANOVA with Tukey’s post hoc test. Significance for panel C was determined using a Student’s t-test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns = not significant.

+/Reldel flies have improved lifespan following TBI

To investigate a role for Rel in chronic outcomes of TBI, we determined the lifespan of +/+, +/Reldel, and Reldel/Reldel flies that were injured at 1 to 7 days old and survived 24 hr following TBI on a food diet. Female and male flies were injured and maintained together in vials but were scored separately for survival. In all cases, the lifespan of injured flies was significantly shorter than the lifespan of uninjured flies of the same sex and genotype (Figure 5 and Tables S1 and S2). In the absence of injury, female but not male +/Reldel flies had a small but significant increase in lifespan compared to +/+ flies (Figure 5A and Table S2). In contrast, male but not female Reldel/Reldel flies had a small but significant decrease in lifespan compared to +/+ flies. Moreover, following TBI, both female and male +/Reldel flies but not Reldel/Reldel flies had a significantly longer lifespan than +/+ flies (Figure 5B and Table S2). The median lifespan of injured female and male flies increased 32% and 15%, respectively (Table S1). These data indicate that in addition to promoting secondary injuries leading to early mortality after TBI in a dose-dependent manner, Rel also contributes to long-term consequences of TBI such as reduced lifespan in a dose-dependent manner. Thus, loss of one copy of Rel is sufficient to improve long-term survival following TBI.