Activation of innate immune signalling during development predisposes to inflammatory intestine and shortened lifespan

Early-life inflammatory response is associated with risks of age-related pathologies. How transient immune signalling activity during animal development influences life-long fitness is not well understood. Using Drosophila as a model, we find that activation of innate immune pathway IMD signalling in the developing larvae increases adult starvation resistance, decreases food intake, and shortens organismal lifespan. Interestingly, lifespan is shortened by the IMD activation in the larval gut and fat body, while starvation resistance and food intake are altered by that in neurons. The adult flies developed with IMD activation show sustained IMD activity in the gut, despite complete tissue renewal during metamorphosis. The inflammatory adult gut is associated with a greater amount of Gluconobacter sp., characteristic gut microbiota increased in response to immune activation. Removing gut microbiota by antibiotics attenuates the increase of IMD activity and rescues the shortened lifespan. This study demonstrates a tissue-specific programming effect of early-life immune activation on the adult physiology and organismal lifespan.


40
Immunity needs to be tightly controlled as both shortages and excesses of immune 41 activation are detrimental to organisms. Chronic, and often systemic, inflammatory 42 response occurs during ageing, which can increase the risk of various age-related 43 diseases [1]. Drosophila melanogaster is a genetically tractable model for studying 44 how immune pathways are involved in the ageing process. The immune deficiency 45 (IMD) pathway is an evolutionally conserved immune regulator in Drosophila, which 46 is a counterpart of the tumour necrosis factor receptor (TNFR) pathway in mammals 47 [2]. The IMD pathway is activated upon infection with bacteria possessing DAP-type 48 peptidoglycan, and is known to be spontaneously activated in aged animals, at least 49 partly in a gut microbiota-dependent manner. Removing microbiota or overexpressing 50 negative regulators for the IMD pathway in the adult gut attenuates age-related IMD 51 activation and concomitantly extends lifespan [3][4][5]. Activation of the IMD pathway in 52 gut progenitor cells induces hyperproliferation of intestinal stem cells [6]. A chronic 53 inflammatory condition in aged flies triggers neurodegeneration and shortens lifespan, 54 which can be rescued by inhibiting IMD signalling in glia cells [7]. Age-related 55 activation of the IMD pathway in the renal (Malpighian) tubules induced by a 56 commensal Acetobacter persici triggers age-dependent metabolic shifts, including 57 purine metabolism [8]. These studies have revealed that age-related immune activation 58 in various tissues leads to organismal ageing.  Various stressors regarded as risk factors for age-related diseases, such as 69 malnutrition, irradiation, chemical exposures, smoke, alcohol, or even mental stress, 70 commonly lead to inflammatory response. A longitudinal cohort study suggested that 71 childhood infection is correlated with the incidence of cardiovascular diseases in 40- 72 year old humans [14]. This and many other epidemiological studies have implied that early-life inflammation is associated with inflammatory diseases and mortality in 74 adulthood [15,16], however few studies directly test the causal relationship. Irradiation 75 during development increases cell death in the adult brain and decreases locomotive 76 ability and organismal lifespan in Drosophila [17]. In this condition, persistent 77 immune activation is observed in adult flies [18]. On the other hand, oral infection of 78 Erwinia carotovora (Ecc15) in larvae does not affect the lifespan of adult flies [19]. 79 Genetic manipulation is also useful to test how early-life signalling activity impacts  In this study, we attempted to test whether immune activation in a larval stage-91 restricted manner can alter adult fitness and lifespan. We found, in adult flies with 92 larval IMD activation, that immune and metabolic alteration occurs and shortens 93 organismal lifespan.

98
Establishment of mild immune activation during development 99 We used the Gene Switch (GS) system to achieve precise control of gene manipulation 100 [22]. GS is useful to induce any gene of interest by an inducer RU486 in a dose-101 dependent manner. For activation of the innate immune IMD pathway in larvae, we 102 overexpressed constitutive active form of IMD (IMD CA ) using a ubiquitous driver 103 Daughterless GS (Da GS ). IMD CA lacks the N-terminal inhibitory domain and therefore 104 is active in the absence of bacterial stimulation [6]. We put embryos of Da GS >IMD CA 105 and its negative control Da GS >LacZ on top of the standard Drosophila diets containing 106 various doses of RU486, and allowed them to develop into adult flies (Fig. 1A).

107
Feeding 200 μM of RU486, the concentration frequently used for adult flies, caused 108 higher developmental lethality even for the control flies in our laboratory condition.

109
Decreasing the RU486 concentration to 5 μM resulted in little effect on viability and 110 adult body weight for the control (Da GS >LacZ) animals, but strong lethality and 111 decreased body weight for Da GS >IMD CA , suggesting that IMD activation impairs 112 larval growth (Fig. 1B,C). When the concentration of RU486 was decreased to 1 μM, 113 adult Da GS >IMD CA flies showed normal body weight (Fig. 1C). At this concentration, 114 we observed mild developmental delay compared with the control (no RU486 115 treatment), but this was rather due to a side effect of RU486, as the phenotype was also 116 obvious for Da GS >LacZ flies (Fig. S1A). 117 We first confirmed that gene expression was indeed induced with as low as 1 118 μM of RU486, as visualised by GFP expression (Fig. S1B). The driver activity is 119 detected in the larval brain, the fat body, the gut and the Malpighian tubules (Fig. 120 S1C). To quantify the level of IMD activation, we performed quantitative RT-PCR 121 analysis for antimicrobial peptide (AMP) genes regulated by the IMD signalling 122 pathway. IMD target genes Diptericin A (DptA) and Drosocin (Dro) were upregulated 123 in the whole body of Da GS >IMD CA third instar larvae (Fig. 1D). These genes were 124 upregulated mildly in various tissues such as the brain, the gut, and the fat body (Fig. 125 S2A). We further performed a transcriptomic profiling by 3'mRNA-sequencing 126 analysis using the gut tissue. AMPs predominantly regulated by the IMD pathway 127 were all upregulated, while those regulated by other immune pathways were not, 128 suggesting that IMD signalling was specifically activated in this experimental setting 129 (Fig. 1E, Table S1). The list of differentially-expressed genes did not contain typical 130 damage-responsive genes such as upd3, which is known to be massively increased in 131 the larval gut upon oral infection of Ecc15 [19]. The AMP induction was already 132 suppressed in the whole body of young adult male flies (Fig. S2B). Therefore, in this 133 experimental setting, we can increase the IMD signalling pathway mildly in a juvenile-134 restricted manner.  despite the decrease in food intake, they had increased starvation resistance (Fig. 2E).

157
Larval IMD activation did not alter paraquat (oxidant) resistance, and it did induce 158 hyper susceptibility to high salt stress (Fig. S3).These phenotypes invalidated the 159 possibility that the shortened lifespan by larval IMD activation was simply due to the  To identify which tissue(s) shortens lifespan upon IMD activation, we overexpressed 167 IMD CA in a tissue-specific manner. We used Gene Switch drivers for neurons (Elav GS ), 168 the gut and the fat body (TI GS ), and the Malpighian tubules (Uro GS ) (Fig. 3A, Fig. S4). 169 We observed that overexpression of IMD CA only by TI GS decreased lifespan, 170 suggesting that IMD activation in the larval gut and/or fat body induces shortened 171 lifespan (Fig. 3B). Interestingly, however, starvation resistance was not elevated in 172 TI GS >IMD CA flies, but this phenotype was rather observed in Elav GS >IMD CA flies (Fig. 173 3C). Therefore, shortened lifespan and starvation resistance are distinctive phenotypes 174 triggered by the IMD activity from the different tissues. Similarly, the decreased food 175 intake was induced only when Elav GS was used to drive IMD activation (Fig. 3D). The 176 data indicate that food intake and starvation resistance have a correlation, while the 177 lifespan shortening occurred in parallel. In this study, we focused on the lifespan 178 phenotype. As the phenotype of lifespan shortening by TI GS >IMD CA is often weaker 179 than Da GS >IMD CA , we used the Da GS >IMD CA for the rest of the study.  Table   194 S2). Unlike the larval gut, where IMD target genes were specifically upregulated, the 195 adult gut showed increased antimicrobial peptides regulated by Toll-or JAK/STAT as 196 well (Fig. 4C). This observation suggested that increased antimicrobial peptides were not due to the sustained RU486 in their gut, but rather to the general inflammatory 198 response of the tissue. Although we could not deny the possibility that overexpressed 199 IMD CA protein in the larval gut remained in the adult gut, it was less likely considering 200 that the larval gut is completely degenerated and replaced by the newly-generated adult 201 gut [23][24][25]. 202 We assumed that an experience of larval immune activation augments 203 immunity as an adaptive response to prepare for future infection in the adult flies. As 204 the increased DptA expression is restricted in the gut, we asked whether the flies were 205 resistant to oral infection that could be influenced by AMPs [26]. Unexpectedly, the 206 larval IMD induction was not beneficial for adult flies against Pseudomonas 207 entomophila infection, but rather it increased susceptibility (Fig. S5). Therefore, higher 208 IMD activity in the gut is thought to be pathological and implies accelerated tissue 209 ageing. 210 We also noticed that intestinal alkaline phosphatase (IAP) Alp9 and Alp10 were  (Table S1). Alp10 was intriguingly listed in the upregulated genes in the   The result did not delineate a huge difference in the microbial composition (Fig. 5A).

241
The total number of live bacteria assessed by colony forming unit assay was also not 242 significantly changed (Fig. 5B). However, when we quantified the number of bacteria 243 by quantitative PCR using a set of primer detecting genera Acetobacter or 244 Gluconobacter, we noticed that Gluconobacter was significantly increased (Fig. 5C).

245
Although we cannot distinguish whether this dysbiosis would be a consequence or a 246 cause of the immune response, the change of the gut microbiome is another evidence 247 that the flies suffer from the inflammatory intestine.

248
To discover whether the gut microbiota are involved in the gut inflammatory 249 response and eventually shorten lifespan, we fed adult flies that experienced larval 250 IMD activation with antibiotics (Rifampicin, Tetracycline, and Ampicillin) to 251 eliminate gut microbiota. We found that IMD upregulation in the adult gut was 252 abolished by the antibiotic treatment (Fig. 5D). In this condition, larval IMD activation 253 did not shorten lifespan (Fig. 5E). These data suggest that gut microbiota contributed 254 to the pathological phenotypes. Together, larval IMD activation promotes the 255 dysregulated gut homeostasis including both host immune response and the gut 256 microbiota (Fig. 5F).  [47,48]. Antibiotic treatment during 295 development can also induce long-term changes in cytokine production in the brain 296 and associated behavioural alteration [49]. Whether immune signalling per se in the 297 developmental stage provokes inflammatory diseases and affects organismal lifespan, 298 as we observed in flies, needs to be tested in mammals.

300
The important question raised by the present study is how neuronal immune activation 301 in the larval stage leads to starvation resistance and decreased food intake in adults.

418
Ten male flies were placed in each vial containing 1% agar to avoid desiccation stress.

419
The level of the food was marked, and the vials were laid in a container with wet towels 420 to prevent water evaporation. The container was incubated at 25°C. After 24 hours, the 421 amount of the food remained in the capillaries was recorded. The vial without flies was 422 also included in the container to subtract evaporation.