PERK Limits Drosophila Lifespan by Promoting Intestinal Stem Cell Proliferation in Response to ER Stress

Abstract

Intestinal homeostasis requires precise control of intestinal stem cell (ISC) proliferation. In Drosophila, this control declines with age largely due to chronic activation of stress signaling and associated chronic inflammatory conditions. An important contributor to this condition is the age-associated increase in endoplasmic reticulum (ER) stress. Here we show that the PKR-like ER kinase (PERK) integrates both cell-autonomous and non-autonomous ER stress stimuli to induce ISC proliferation. In addition to responding to cell-intrinsic ER stress, PERK is also specifically activated in ISCs by JAK/Stat signaling in response to ER stress in neighboring cells. The activation of PERK is required for homeostatic regeneration, as well as for acute regenerative responses, yet the chronic engagement of this response becomes deleterious in aging flies. Accordingly, knocking down PERK in ISCs is sufficient to promote intestinal homeostasis and extend lifespan. Our studies highlight the significance of the PERK branch of the unfolded protein response of the ER (UPRER) in intestinal homeostasis and provide a viable strategy to improve organismal health- and lifespan.

Fig 1. ROS-independent induction of proliferation of ero1l-deficient ISCs.

(A)Major branches of UPR^ER and ROS signaling pathways in Drosophila. (B)Model of Xbp1-Hrd1 branch of the UPR^ER and ROS signaling network regulating ISC proliferation. (C)Loss of Ero1L in ISCs promotes ISC proliferation. PH3+ cell frequency in intestines is quantified for wild-type flies and flies expressing Ero1L^RNAi both in ISCs/EBs (esg^ts) and specifically in ISCs (esg^ts,Su(H)Gbe::G80). Averages and SEM are shown. P values from Student’s T test, N>15. (D)DHE fluorescence in wild-type controls and Ero1L^RNAi expressing ISCs/EBs (esg^ts). GFP, green; DHE, red; DHE is shown as separate channel in white. Averages and SEM for relative DHE intensity are shown. P values from Student’s T test, N > 200 (from 5–8 guts). (E)Quantification of pH3⁺ cells in wild-type flies and in flies over-expressing Ero1L under the control of esg::Gal4, UAS-GFP; tubG80^ts exposed to mock (5% sucrose) and tunicamycin. Averages and SEM are shown. P values from Student’s Test. N = 10. (F)Larger clone sizes (MARCM) in Ero1LRNAi or Ero1L^335QRS condition compared with their own wild-type controls at 3 day after heat shock. Averages and SEM are shown. P values from Student’s T test. Number of clones analyzed: n = 121 (FRT40); n = 89 (Ero1LRNAi); n = 66 (FRT80B); n = 63 (Ero1L^335QRS). (G)eIF2α phosphorylation in Ero1L-deficient ISCs (knockdown using esg::Gal4, Su(H)Gbe::G80, tubG80ts). DAPI, blue; YFP, green; peIF2α or Dl shown as separate channels in white. Arrowheads for orientation. Quantification of phospho-eIF2α staining in ISCs relative to neighboring ECs in wild-type guts and in guts with Ero1L-deficient ISCs shown on the right. Averages and SEM are shown. P values from Student’s Test. (H)Increased eIF2α phosphorylation in Ero1L-deficient ISCs/EBs (esg^ts) is repressed by loss of PERK in ISCs/EBs (esg^ts). DAPI, blue; nlsGFP, green; peIF2α, white. Arrowheads for orientation. (I)Knockdown of PERK in ISCs/EBs (esg^ts) inhibits ISCs proliferation induced by loss of Ero1L. Frequency of pH3+ cells quantified for wild-type flies and flies expressing dsRNA against PERK only,dsRNA against Ero1L only, and for flies co-expressing dsRNA against PERK and dsRNA against Ero1L (combined with Dicer2 to enhance RNAi efficiency; shown on the left). Averages and SEM are shown. P values from Student’s T test, N = 10. Representative images are shown on the right. DAPI, blue; nlsGFP, green. See also S1 Fig.

Fig 2. PERK integrates local and tissue-wide ER stress responses to regulate ISC proliferation.

(A) Phosphorylation of eIF2α is induced in ISCs/EBs when young intestines are exposed to excessive ER stress. Guts of esg::Gal4, UAS::nlsGFP (nuclear-localized GFP) were immunostained with anti-peIF2α and anti-Armadillo antibodies in either the mock condition (5% sucrose) or after treatment with 50 μM Tunicamycin. (DNA: DAPI blue; GFP: green. Armadillo:red). peIF2α channel is shown separately on the right. (B) Xbp1 is spliced in ISCs/EBs when young intestines are exposed to excessive ER stress. GFP (GFP only expressed in Xbp1 splicing) and peIF2α channels are shown separately on the right. (C,D) Knockdown of PERK in ISCs/EBs (esg^ts, C) or specifically in ISCs (esg^ts/Su(H)Gbe::G80, D) prevents tunicamycin-induced ISC proliferation. Intestines of PERK^RNAi co-expressed with Dicer2 or wild-type (Dicer2 only) were examined. Averages and SEM are shown. P values from Student’s T test, N>10. (E) Phosphorylation of eIF2α in ISCs/EBs is not induced under ER stress in ISCs in which PERK is knocked down (esg^ts). Intestines were exposed to either mock conditions (5% sucrose) or to 50 μM tunicamycin (TM). DAPI, blue; GFP, green, peIF2α shown as separate channel in white. Arrowheads for orientation. Quantification of phospho-eIF2α staining in ISCs relative to neighboring ECs in wild-type guts and in guts expressing PERK^RNAi in ISCs/EBs (esg^ts) shown on the right. Averages and SEM are shown. P values from Student’s Test. (F) Quantification of MARCM clone sizes at 8 days and 13 days after heat shock for PERK loss-of-function (PERK^e01744, PERK^RNAi) conditions. Averages and SEM are shown. P values from Student’s T test. Number of clones examined: n = 455 (8 d FRT82B); n = 213 (8 d PERK^e01744); n = 389 (8 d PERK^RNAi); n = 200 (13 d FRT82B); n = 198 (13 d PERK^e01744). Representative images for PERK loss-of-function (PERK^e01744, PERK^RNAi) at 8 days and 13 days after heat shock are shown on the right. (DAPI, blue; GFP, green). See also S2 and S4 Figs.

Fig 3. PERK-induced transcriptome changes in ISCs.

(A, B) Cell sorting strategy to isolate ISCs from intestines expressing YFP in ISCs exclusively under the control of esg^ts, Su(H)Gbe::Gal80. Cells were gated based on forward and side scatter, YFP fluorescence (A), and cell size (B). Gate P3 (ISCs) was sorted and collected for subsequent RNAseq analysis. Cells in P3 are Dl+ (Sousa-Victor, personal communication). (C)FPKM values for selected genes across the four analyzed libraries (wild-type and PERK^RNAi expressing ISCs in both mock and tunicamycin-treated guts). A selected set of highly expressed genes is shown in the upper 5 rows, revealing no significant differences in expression across the four libraries. The lower rows show selected genes induced by Tunicamycin in wild-type ISCs, but not in PERK^RNAi-expressing ISCs. Total cell numbers of ISCs isolated for RNAseq analysis are: mock WT (119,000), mock PERK (111,000), TM WT (112,000) and TM PERK (70,000). (D) Venn diagram for TM-induced transcriptional changes by RNA-seq analysis in wild-type ISCs and ISCs expressing dsRNA against PERK. Genes were considered induced when FPKM differences were at least 3 fold and FPKM values were at least 5.0. (E) Induction of genes involved in antioxidant and detoxification responses and in cell cycle regulation and mitosis in response to tunicamycin. Data was acquired by RNA-seq analysis of ISCs isolated by FACS. Fold induction (TM/Mock) is shown for wild-type and PERK-deficient ISCs. See details in Materials and Methods. See also S3 Fig.

Fig 4. Non-autonomous activation of PERK in ISCs by JAK/Stat signaling.

(A) Loss of Xbp1 (Xbp1^RNAi: HMS03015) specifically in EBs (using Su(H)Gbe::Gal4,tubG80ts) increases the phosphorylation of eIF2α in ISCs/EBs. DAPI, blue; GFP, green, peIF2α red; peIF2α shown as separate channel in white. Arrowheads for orientation. (B) Increased phosphorylation of eIF2α in ISCs/EBs and activation of JAK/Stat signaling in the gut of flies in which Xbp1 is knocked down in ECs. 2XSTAT::GFP was used as a reporter for Stat activity. (DAPI, blue; GFP, green; Dl and peIF2α shown as separate channels in white). (C) Knockdown of Stat in ISCs (esg^ts,Su(H)Gbe::G80) inhibits eIF2α phosphorylation in ISCs under ER stress. Intestines were exposed to either mock conditions (5% sucrose) or to 50 μM tunicamycin (TM). DAPI, blue; YFP, green, peIF2α shown as separate channel in white. Arrowheads for orientation. (D) Knockdown of Stat or Upd3 in ISCs prevents tunicamycin-induced ISC proliferation. Averages and SEM are shown. P values from Student’s T test, N>10. (E) Knockdown of JNK (Bsk^RNAi) or Upd (Upd3^RNAi) inhibits ISC over-proliferation induced by loss of Xbp1 in EBs (using Su(H)Gbe::Gal4,tubG80ts). Averages and SEM are shown. P values from Student’s T test, N>10. (F) Activation of JAK/Stat pathway either Upd2 overexpression in ECs (using NP1^ts) or specifically in ISCs by over-exression of Hop^tuml (using esg^ts/Su(H)Gbe::G80) promotes the phosphorylation of eIF2α in ISCs. DAPI, blue; GFP, green peIF2α and Dl shown as separate channel in white. See also S2, S4 and S5 Figs.

Fig 5. Limiting PERK activity in ISCs promotes intestinal homeostasis.

(A) Young (5 day old) and old (30 day old) guts of esg::Gal4, UAS::nlsGFP flies immunostained with anti-peIF2α and anti-Dl antibodies. Enlarged imaged are shown on the right. White arrowheads point to individual ISCs (DNA: DAPI, blue; ISCs/EBs: GFP, green, peIF2α, red). peIF2α and Dl channels are shown separately on the right). (B) Intestines of wild-type flies stained with peIF2α antibodies at 5 days or 33 days of age and of old flies over-expressing Hrd1 or Xbp1 (Xbp1^d08698) in ISCs/EBs (esg::Gal4) Arrowheads point to ISCs/EBs (marked by GFP). Relative peIF2α intensity was determined by calculating the ratio of fluorescence intensity in GFP+ cells and nearby ECs. Averages and SEM are shown. P values from Student’s T test, N > 160 (from 4–5 guts each).

Fig 6. Knockdown of PERK in ISCs extends lifespan in flies.

(A) Knockdown of PERK in ISCs using RU486-inducible driver 5961GS leads to lifespan extension. 200–400 female flies of each genotype were assayed. (B) Summary of parameters and lifespan statistics for the flies in Fig 6A. See also S6 Fig.

Fig 7. Model of UPR^ER/ROS signaling network regulating ISC proliferation.

Left: Model of UPR^ER/ROS signaling network regulating ISC proliferation. PERK activation, either through cell-autonomous ER stress or in response to non-autonomous ER stress and JAK/Stat activation, is required for the induction of ISC proliferation. Inhibition of CncC activity by JNK in response to ER-induced ROS is further required to permit ISC regenerative responses. Right: Age-related loss of proliferative homeostasis as a consequence of increased ER stress.