PGRP-SC2 promotes gut immune homeostasis to limit commensal dysbiosis and extend lifespan

Abstract

Interactions between commensals and the host impact the metabolic and immune status of metazoans. Their deregulation is associated with age-related pathologies like chronic inflammation and cancer, especially in barrier epithelia. Maintaining a healthy commensal population by preserving innate immune homeostasis in such epithelia thus promises to promote health and longevity. Here, we show that, in the aging intestine of Drosophila, chronic activation of the transcription factor Foxo reduces expression of peptidoglycan recognition protein SC2 (PGRP-SC2), a negative regulator of IMD/Relish innate immune signaling, and homolog of the anti-inflammatory molecules PGLYRP1-4. This repression causes deregulation of Rel/NFkB activity, resulting in commensal dysbiosis, stem cell hyperproliferation, and epithelial dysplasia. Restoring PGRP-SC2 expression in enterocytes of the intestinal epithelium, in turn, prevents dysbiosis, promotes tissue homeostasis, and extends lifespan. Our results highlight the importance of commensal control for lifespan of metazoans and identify SC-class PGRPs as longevity-promoting factors.

Figure 1. Dysbiosis, ISC hyperproliferation, and changes in gene expression in axenically and conventionally aging intestines

A) Colony-forming units (CFUs) in intestinal extracts of NP1∷Gal4 transgenic lines in y¹w¹ and w¹¹¹⁸ background. See Fig. S1A for OreR flies. Midgut homogenates from flies at 5 days or 30 days of age were plated on nutrient rich medium (NR), or on selective plates allowing growth of Lactobacilli (MRS Agar), Acetobacteria or Enterobacteria. B) Quantification of mitotic figures (cells positive for phosphorylated Histone H3; pH3+) in guts of the same genotypes as in A. C) Quantification of mitotic figures in OreR flies of indicated ages, aged axenically (with antibiotics) or conventionally. See Fig. S1B for flies maintained on sterile food without antibiotics. D) Venn diagram comparing the number of genes significantly induced in conventionally or axenically aging flies as identified by RNASeq. Genes were classified as induced when expression levels were are least 3 times higher than at 2 days of age in at least three older timepoints. Only genes with combined expression values above 10 RPKM were considered. Outtakes of hierarchical clustering highlighting examples of genes induced in both axenic and conventional conditions (left), or induced selectively in conventional conditions (right). Hierarchical clustering was performed on log2 values of gene expression ratios between indicated ages and 2d old samples. E) GO enrichment analysis using Flymine.org of 67 genes expressed at least 3 times higher in guts of conventionally reared animals at at least 3 timepoints (upper chart). GO enrichment analysis of 112 genes induced more than 3 fold in both conditions compared to 2 day old samples at all four timepoints (lower chart). See also Figure S1 and Tables S1, S2, and S3.

Figure 2. Oxidative stress response, IMD/Rel activity and innate immune dysfunction in aging intestines

A) Oxidative stress response genes, as well as JNK target genes (mys, zip, kay) are induced in conventionally, but not in axenically aging intestines. Gene expression trajectories from RNAseq using RPKM values normalized to 2 day old samples. B) qRT-PCR detecting duox expression relative to actin5C. Averages and standard deviations (N=3; * p<0.05, Student’s Ttest). C–E) Scatter plots relating CFUs with age, Duox expression with CFUs, and mitotic figures with Duox expression in intestines of wild-type (OreR) flies. Fitted curves with R² values are indicated. The dotted red line indicates a threshold value for Duox expression at which mitotic figures increase strongly. F) Expression trajectories of innate immune response genes in axenically and conventionally reared flies. Log scale. G) qRT-PCR detecting Dpt expression relative to actin5C. Averages and standard deviations (N=3; ** p<0.01, * p<0.05; Student’s Ttest). H) S. Marcescens clearance: Bacterial load (CFU/gut) determined after feeding axenically reared young (5 day old) or old (45 day old) animals S.Marcescens (flies were fed 700µl of 5% sucrose containing a 100 fold concentrated suspension of bacteria grown to OD₆₀₀=1) for 1 day, then allowing a 3 day recovery period. Mock treated animals were fed 5% sucrose solution. See also Figure S2.

Figure 3. Foxo activation in aging intestines activates IMD/Rel signaling

A) Expression of Foxo target genes Lip4 and InR in axenically and conventionally reared animals as determined by RNASeq (lines) and qRT-PCR (for InR, relative to actin5C, bars). Averages and standard deviations (N=3; ** p<0.01, Student’s Ttest). B) X-Gal staining showing activation of thor-lacZ in aging intestines. C) qRT-PCR detecting Rel and Dpt expression relative to actin5C in wild-type (OreR) or foxo^W24 aged axenically. Averages and standard deviations (N>=3; p values from Student’s Ttest). D) qRT-PCR detecting Dpt expression relative to actin5C in animals expressing Foxo^RNAi using NP1∷Gal4 in w¹¹¹⁸ (left) or y¹w¹ (right) backgrounds and aged conventionally. Averages and standard deviations (N>=3; p values from Student’s Ttest). See also Figure S3.

Figure 4. Foxo mediates age-related repression of PGRP-SC2 expression and impairs innate immune responses

A) Age-related repression of the three PGRP-SC genes in the gut of axenic flies (from RNASeq; normalized to 2 day olds; PGRP-SC2 is the most highly expressed PGRP-SC). B) qRT-PCR detecting PGRP-SC2 relative to actin5C in OreR or foxo^W24 aged axenically (green) or conventionally (red). Averages and SEM (N>=8; p values from Student’s Ttest). C) qRT-PCR analysis as described above. D) PGRP-SC2 expression in animals over-expressing Foxo in ECs. Animals were reared at 18°C and Foxo expression was induced in adults by shifting animals to 29°C for 5 days. qRT-PCR analysis as described above. E) Induction of PGRP-SC2 by Erwinia Carotovora Carotovora (ATCC #15359) infection in animals over-expressing Foxo in ECs. Animals were reared at 18°C and Foxo expression was induced in adults by shifting animals to 29°C for 5 days. qRT-PCR analysis as described above. F) Induction of Dpt in response to Ecc infection in animals over-expressing Foxo in ECs with or without the Relish inhibitor Pyrrolidine dithiocarbamate (PDTC). Culture conditions and qRT-PCR analysis as describe above. G, H) Clearance of Serratia marcescens (SM, G) or Ecc (H) from guts of animals over-expressing Foxo in ECs. See also Figure S4.

Figure 5. Knockdown Foxo in Enterocytes reduces dysbiosis and ISC overproliferation

A) Age-related changes in gut commensal numbers in foxo loss of function conditions. Intestinal CFUs (averages and SEM) quantified in wild-type, foxo^W24 hererozygous, or foxo^W24 homozygous sibling animals of the indicated ages. foxo^W24 was backcrossed for 5 generations into the y¹w¹ background (**p<0.01 relative to wild-type). B–C) Intestinal CFUs (B) and mitotic ISCs (C) quantified from animals of the indicated ages in which Foxo (or Foxo and PGRP-SC2) was knocked down specifically in ECs (using NP1∷Gal4 backcrossed into w¹¹¹⁸). Averages and SEM are shown (N>17 for each data point). D–E) As above, using NP1∷Gal4 backcrossed into the y¹w¹ background (*p<0.05; **p<0.01). See also Figure S5.

Figure 6. Over-expression of PGRP-SC2 in ECs reduces dysbiosis, ISC overproliferation, and extends lifespan

A, B) Age-related changes in gut commensal numbers (A) and ISC proliferation (B) in PGRP-SC2 gain of function conditions. Intestinal CFUs (averages and SEM) quantified in wild-type, or PGRP-SC2 over-expressing animals (using NP1∷Gal4 in the w¹¹¹⁸ background). Two independent UAS∷PGRP-SC2 transgenic lines were used. Averages and SEM are shown (*p<0.05; **p<0.01). C) Quantification of intestinal CFUs (left; averages and SEM, NR plates) and ISC proliferation (right, averages and SEM) in 50 day old intestines of animals over-expressing PGRP-SC2 in ECs using the RU486-inducible 5966GS driver. D–F) Mortality trajectories of females over-expressing PGRP-SC2 in ECs using the RU486-inducible 5966GS driver in conventional (E) or axenic (F) conditions. Axenic flies were kept on sterile food in a laminar flow hood without addition of antibiotics. Note that RU486 treatment does not affect lifespan in wild-type conditions (F). G) Table summarizing parameters and statistics of the demographies shown in D–F, as well as of flies over-expressing PGRP-SC2 using NP1∷Gal4. PGRP-SC2 lines were backcrossed 7 generations into the w1118 background and crossed to similarly backcrossed heterozygous NP1∷Gal4 (see graphs in Fig. S6). Mortality of sibling populations with or without the NP1∷Gal4 transgene was compared. See also Figure S6.

Figure 7. Model for the age-related deregulation of innate immune and epithelial homeostasis in the Drosophila intestine

A) Model depicting the relationship between age, commensal dysbiosis, and intestinal dysplasia. Our results indicate that when intestinal CFUs and associated Duox expression reach a threshold, ISCs overproliferate, causing dysplasia. Over-expression of PGRP-SC2 delays the accumulation of commensals, resulting in delayed acquisition of dysplasia, extending lifespan. B, C) Proposed mechanism causing immune dysfunction and commensal dysbiosis in the fly intestine. In young animals, PGRP-SC2 maintains immune homeostasis by limiting IMD/Rel activity. Foxo-mediated suppression of PGRP-SC2 allows adjusting responsiveness and basal activity of IMD/Rel. Age-associated chronic activation of Foxo in ECs, in turn, represses PGRP-SC2 expression chronically, de-regulating IMD/Rel signaling. The associated immunosenescence causes commensal dysbiosis, which chronically activates Duox – mediated ROS production, triggering ISC over-proliferation and dysplasia.