Mechanisms of damage prevention , signalling , and repair impact 1 the ability of Drosophila to tolerate enteric bacterial infection 2 3 4

Many insects thrive on decomposing and decaying organic matter containing a large diversity of both commensal and pathogenic microorganisms. The insect gut is therefore frequently exposed to pathogenic threats and must be able not only to detect and clear these potential infections, but also be able to repair the resulting damage to gut tissues in order to tolerate relatively high microbe loads. In contrast to the mechanisms that eliminate pathogens, we currently know less about the mechanisms of disease tolerance, and most of this knowledge stems from systemic infections. Here we investigated how well-described mechanisms that either prevent, signal, control, or repair tissue damage during infection contribute to the phenotype of disease tolerance during gut infection. We orally infected adult Drosophila melanogaster flies with the bacterial pathogen Pseudomonas entomophila in several loss-of-function mutants lacking epithelial responses including damage preventing dcy (drosocrystallin - a major component of the peritrophic matrix), damage signalling upd3 (unpaired protein, a cytokine-like molecule), damage controlling irc (immune-regulated catalase, a negative regulator of reactive oxygen species) and tissue damage repairing egfr1 (epidermal growth factor receptor). Overall, we detect effects of all these mechanisms on disease tolerance. The deterioration of the peritrophic matrix in dcy mutants resulted in the highest loss of tolerance, while loss of function of either irc or upd3 also reduced tolerance in both sexes. The absence of tissue damage repair signalling (egfr1) resulted in a severe loss in tolerance in male flies but had no substantial effect on the ability of female flies to tolerate P. entomophila infection, despite carrying greater microbe loads than males. Together, our findings provide empirical evidence for the role of damage limitation mechanisms in disease tolerance and highlight how sex differences in these mechanisms could generate sexual dimorphism in immunity.

The mechanisms of pathogen clearance are well-described in many animal species (Cooper 2018), 70 but we currently know less about the mechanisms underlying disease tolerance. Given that 71 tolerance reflects a distinct ability to maintain health independently of pathogen clearance, we 72 might expect these mechanisms to be related to processes such as detoxification, reduction of 73 inflammation, or tissue damage control and cellular renewal ( To investigate how these mechanisms of damage prevention (dcy), signalling (upd3) control (irc) and 129 renewal (egfr 1 ) contribute to disease tolerance during gut infections we employed oral infections in 130 Drosophila lines carrying loss-of-function mutations in each of these genes on a common genetic 131 background (w 1118 ). We orally challenged these flies with a range of infection doses of Pseudomonas 132 entomophila and then quantified their effects on survival, pathogen loads and disease tolerance 133 responses during period of peak infection burden.

Materials and methods 155
Fly strains: The following fly stocks were obtained from VDRC (Vienna Drosophila Resource 156 Centre) and Bloomington Stock Centre, Indiana: dcy (BL26106), irc (BL29191), egfr 1 (BL2079), upd3 157 (BL19355). All the mutants were subsequently backcrossed into the wild type w 1118 (VDRC stock# 158 60000) for at least 10 generations. All fly lines were maintained in plastic vials (12 ml) on a standard 159 cornmeal diet (Lewis's medium) at a constant temperature of 25°C (±2°C). 1. Survival following oral infection: We analysed survival data using a mixed effects Cox model 210 using the R package 'coxme' (Therneau 2015). We specified the model as: survival ~ fly 211 line * treatment * sex * (1|vials/block), with 'fly line', 'treatment' and 'sex' and their 212 interactions as fixed effects, and 'vials' nested in 'block' as a random effect for wild type 213 w 1118 and mutant flies. 214 215 2. Internal bacterial load: We found that residuals of bacterial load data were non-normally 216 distributed when tested using Shapiro-Wilks's test. Hence, we first log-transformed the 217 data and then confirmed that the log-transformed residuals were still non-normally 218 distributed. Subsequently, we used a non-parametric one-way ANOVA (Kruskal-Wallis 219 test) to test the effects of each fly line, that is, wild-type w 1118 and mutant flies on males and 220 females separately following oral P. entomophila infection.

249
Results 250 1. Flies lacking damage-preventing dcy are more susceptible to oral P. entomophila infections 251 than those lacking components that minimise, signal or repair tissue damage 252 Following oral infection with three different doses of Pseudomonas entomophila, we found that flies 253 lacking major components of gut immunity such as tissue damage -preventing dcy, signalling upd3, 254 repairing egfr 1 and controlling irc, were all significantly more susceptible to oral P. entomophila 255 infections than the wild type w 1118 flies ( Fig. 3B and 1A     The higher susceptibility of mutants to oral bacterial exposure could either be caused by their 279 inability to supress the bacterial growth or due to their inability to tolerate the damage inflicted 280 during oral infection. To distinguish between these mechanisms, we first quantified internal 281 bacterial loads across several time points, that is, 0-15 minutes, 24-hours and 96-hours following 282 oral exposure to P. entomophila. All fly lines showed sex-differences at 24-hours peak load, though 283 by 96-hours this sex difference was no longer present in flies lacking damage sensing (upd3) and 284 tissue renewal mechanisms (egfr) (Fig. 4, Table S2). Flies lacking the negative regulator of ROS irc 285 always exhibited lower levels of bacterial load compared to wild type flies (Fig. 4B, Table S2). 286

Female Male
Mean bacterial load (log 10 ) While some of the variation in survival between mutants (Fig. 3) may be explained by variation in 299 resistance [that is, their ability to clear infection (Fig. 4)], some of that variation may also arise due 300 to differences in tolerance. We were therefore interested in measuring disease tolerance in these 301 lines using the reaction norm of survival relative to bacterial loads, where the slope of the linear 302 relationship reflects the degree of tolerance: steep negative slopes indicate a rapid mortality with 303 increases in pathogen loads (low tolerance), while less steep or flat slopes reflect relatively more showed reduced tolerance to oral bacterial P. entomophila infections compared to wild type w 1118 flies 312 (Fig. 5, Table 3). Here, the differences in tolerance between wild type w 1118 and mutant flies are 313 indicated by a significant interaction between the fly line bacterial load for survival, which reflects 314 the overall rate at which fly health (survival) changes with bacterial load (tolerance) between fly 315 lines in both males and females (Fig. 5. Table 2). Males lacking egfr 1 show lower tolerance to 316 bacterial infection than females (Fig. 5, Table 2), suggesting sexual dimorphism in gut cell renewal. 317 However, both males and females lacking major component of the peritrophic matrix that is, dcy 318 showed significantly reduced tolerance (Fig. 5), indicating the general importance of tissue damage 319 preventing epithelial barrier in defence against oral infections (Kuraishi, Hori, and Kurata 2013). 320

351
Discussion 352 In the present work we tested how mechanisms of tissue damage prevention (dcy), signalling (upd3) 353 control (irc) and renewal (egfr 1 ) contribute to disease tolerance during enteric infection. We present 354 evidence that all of these mechanisms contribute to disease tolerance, and that some of these 355 Mean bacterial load (log 10 )