Tissue damage-induced intestinal stem cell division in Drosophila

Abstract

Stem cell division is essential for tissue integrity during growth, aging, and pathogenic assaults. Adult gastrointestinal tract encounters numerous stimulations, and impaired tissue regeneration may lead to inflammatory diseases and cancer. Intestinal stem cells in adult Drosophila have recently been identified and shown to replenish the various cell types within the midgut. However, it is not known whether these intestinal stem cells can respond to environmental challenges. By feeding dextran sulfate sodium and bleomycin to flies and by expressing apoptotic proteins, we show that Drosophila intestinal stem cells can increase the rate of division in response to tissue damage. Moreover, if tissue damage results in epithelial cell loss, the newly formed enteroblasts can differentiate into mature epithelial cells. By using this newly established system of intestinal stem cell proliferation and tissue regeneration, we find that the insulin receptor signaling pathway is required for intestinal stem cell division.

Figure 1. DSS feeding causes mortality and increases enteroblast number

(A) Various amounts of DSS as indicated were included in the sucrose feeding medium and the percentage of flies alive after each day is expressed as survival rate. The inclusion of DSS causes increased mortality. (B) 3% of each of the indicated polysaccharides was included in the feeding medium and the flies counted each day. High molecular weight DSS and dextran do not change significantly the viability of the flies, while low molecular weight DSS does. (C) An illustration of the cell types and differentiation pathways in adult midgut. (D–F) Cell proliferation effect revealed by the esg-Gal4/UAS-CD8GFP marker. The type of polysaccharide used is indicated in the lower right corner of each panel. The scale bar shown in panel C is 20μm. (G) Quantification of esg-Gal4/UAS-CD8GFP positive cells after feeding with various polysaccharides as indicated. The number of GFP positive cells were counted in multiple images for each experiment and normalized by 100 unstained cells as revealed by DAPI staining. The error bar is standard deviation. (H–O) The blue color in these panels and all other figures is DAPI staining of DNA. The red or green staining is for the antigens indicated to the left. The left columns are control gut staining, and the right columns are stained guts from DSS fed flies. The DSS feeding experiments were performed with 3% of 40 kDa DSS for 2 days at 29°C unless specified. The arrow in panel H indicates an ISC. (P) Quantification of the positively stained cells for the various markers as indicated. The relative number is presented as stain-positive cells per 100 unstained cells revealed by DAPI staining.

Figure 2. DSS increases ISC division and does not affect cell fate determination

(A–B) Immunofluorescent staining using antibody against phospho-H3 is shown (green). DSS increases the number of phospho-H3-positive cells. (C) Phospho-H3-positive cell count in whole gut, from anterior midgut next to cardia to posterior midgut before malpighian tubules. The average number per gut and standard deviation is shown. (D–I) Green staining is anti-β–galactosidase for enteroblasts, red staining is anti-Delta for ISC, and blue is DAPI for DNA. 3% DSS feeding was carried out for 2 days. The arrows in control and DSS panels point to ISC with positive Delta staining but no β–galactosidase staining. The inserts are enlargements of the cells indicated by the arrows. (J–O) Double staining for phospho-H3 (green) and Delta (red). All phospho-H3 positive cells have Delta staining (arrow), with or without DSS feeding. (P–U) Each cluster of GFP positive cells represent one lineage originated from one parental ISC after MARCM. All MARCM feeding experiments were performed for 14 days in 18°C. The control clones contain 1 to 2 cells. The arrow points to an ISC with Delta staining and GFP. In DSS fed flies of similar age, the clone has grown to 7 cells in this one confocal plane. Within each cluster, almost always only one cell is Delta positive (red, indicated by arrowhead). Therefore, ISC asymmetry and renewal after DSS feeding is normal.

Figure 3. DSS disrupts basement membrane organization and causes enteroblast accumulation

(A–F) MARCM clones are GFP positive, indicated by the arrowheads. In the DSS fed sample, the cluster of GFP positive cells stay close to the basal side and are not closely associated with the phalloidin staining at the apical side of bigger enterocytes (arrow). Therefore, these accumulated GFP positive cells are enteroblasts but not mature enterocytes. The orientation of these sagittal images is apical to the top. (G, H) Cross-sections of gut tissue from control and DSS treated flies. The tissue sections were stained with toluidine blue. The enterocyte nuclei have dark blue staining. The lumens are indicated on the panels. The arrows point to the basal layer, which is smoother in control and discontinuous in DSS samples. (I, J) Surface views of 3-D reconstructed confocal images of guts dissected from a collagenIV-GFP fly line. In control gut, the longitudinal filaments (top to bottom direction) are connected by transverse filaments (left to right direction). This scaffolding pattern probably represents part of the basement membrane structure. In gut from DSS fed fly, this GFP pattern is less organized, suggesting a disruption of basement membrane structure. (K, L) Confocal sagittal views of control and DSS treated guts from collagenIV-GFP flies. There are clear differences between the two samples: the two layers of longitudinal filaments seen in control gut become more rounded in DSS treated gut. The orientation of the images is apical to the top.

Figure 4. Bleomycin feeding induces damage specifically in epithelial layer

(A) Bleomycin ranges from 0.25 to 250 μg/ml or 3% dextran was included in the feeding medium. Feeding of bleomycin causes dose-dependant increase in mortality in adult flies. (B–G) DNA strand breaks are detected by anti-H2AvD staining (red). In control guts, low level of H2AvD punctate staining is co-localized with DAPI staining. The fluorescent signal is highly increased in bleomycin treated guts, is co-localized with DAPI staining and is only present in the bigger enterocyte nuclei (arrow). The small nuclei, marked by esg-Gal4/UASGFP expression (arrowhead), have no detectable H2AvD staining. (H–K) Bleomycin causes epithelial layer disorganization. Spectrin (green) has higher level expression at the apical side of enterocytes. The orientation of all sagittal images is apical to the top. The open arrows indicate enterocytes that have come off the epithelial layer after bleomycin feeding. (L, M) Plastic tissue sections show that bleomycin feeding causes falling off of some enterocytes (open arrow). (N) Feeding experiments were performed using the indicated amount of bleomycin in the feeding medium. The number of phospho-H3-positive cells in the whole midgut was counted after 2 days of feeding. There is approximately 10 fold increase in ISC division after 25 μg/ml bleomycin feeding. (O, P) Double staining for Delta and Su(H)Gbe-lacZ after bleomycin feeding, 25 μg/ml for 2 days. For control see Fig. 2D. The arrowheads indicate an ISC that has high level of punctate Delta (red) and no β-galactosidase (green). (Q, R) Clonal analysis by MARCM shows that within one lineage only one Delta positive cell is present after bleomycin feeding. For control see Fig. 2P. The arrowheads point to one Delta positive cell (red) within a GFP positive cluster. The arrows indicate a bigger cell resembling an enterocyte within the clone.

Figure 5. Epithelial cell loss is coincident with ISC division and enteroblast differentiation

(A–L) All images are sagittal views showing MARCM GFP clusters after feeding with tissue damaging agents as indicated. In control, the cluster contains two small GFP positive cells, representing ISC and enteroblast (arrowhead). With bleomycin feeding (25 μg/ml), more GFP positive cells are present in each cluster and some GFP positive cells are bigger and have tightly associated phalloidin staining (red) at the apical side, suggesting that they are mature enterocytes (arrow). DSS feeding causes accumulation of enteroblasts that are not associated with phalloidin staining (arrowhead). DSS and bleomycin together causes accumulation of GFP positive cells as well as bigger cells that have tightly associated phalloidin staining (arrow). (M) Quantification of the number of MARCM GFP clones that also contained mature cells. Isolated clones were counted from whole guts and only those that contained mature cells are presented in the graph. The result indicates that bleomycin but not DSS causes facilitated differentiation. (N–Q) Confocal images of NPC1b-Gal4/UAS-CD8GFP expression in midgut. N and P are surface views; O and Q are sagittal views. The arrow indicates lack of GFP signal in an enterocyte nucleus, but enterocyte cytoplasm has strong GFP signal. All small cells do not show GFP signal in both nuclei and cytoplasm (arrowhead). (R) Transgenic expression of viral anti-apoptotic protein p35 in epithelial cells by NPC1b-Gal4 suppresses approximately 50% of the bleomycin-induced ISC division based on phospho-H3 count. (S) Transgenic expression of apoptotic proteins Hid and Reaper in epithelia cells by NPC1b-Gal4 increases ISC division. Two lines of Reaper, on X and 2^nd chromosomes, were used. Phospho-histone 3 positive cells were counted in guts from 3–5 day old flies. (*) Note that the NPC-Gal4 driven GFP expression was detected only in the posterior portion of midguts and we counted phospho-H3 positive cells only in this regions rather than in the whole gut, and thus the overall number of phospho-H3 positive cells was lower than that in other experiments.

Figure 6. Insulin receptor signaling is required for ISC division

(A) Wild type, InR^E19/P05545 transheterozygotes, chico¹ and Armadillo-Gal4/UAS-FOXO^™ flies were set up for feeding with 3% DSS or 25 μg/ml bleomycin for 2 days. The guts were stained for phospho-H3 and the number of positively stained cells presented. These three fly lines of insulin-signaling components showed significantly reduced ISC division. (B–G) In wild type guts, Armadillo/β-catenin staining (red) reveals cell membranes of smaller ISC and enteroblasts and of bigger enterocytes. After DSS and bleomycin feeding, the wild type guts have more small cells and are less organized. InR mutant guts after DSS and bleomycin feeding have fewer small cells and the organization is more similar to that of control guts. (H) Quantification of wild type and InR mutant MARCM GFP clones. The solid portion of each bar represents the percentage of clones with 3 or more cells, and the open portion represents the percentage of clones with 2 or 1 cells. The majority of InR mutant clones had 2 or 1 cells, showing that they had much reduced proliferation. (I, J) Representative images of InR mutant and wild type MARCM GFP positive clones after DSS feeding. (K) Quantification of phospho-H3 positive cells in control and InR^A1325D expressing flies. Armadillo-Gal4, NPC1b-Gal4, esg-Gal4, and heat shock-Gal4 were used to cross with UAS-InR^A1325D. 3–5 day old flies from the crosses and various controls were used for gut dissection and phospho-H3 staining. (L–Q) Cellular phenotypes in midgut induced by esg-Gal4/UAS-InR^A1325D. Armadillo/β-catenin staining shows that there are more small cells, and the overall arrangement appears more disorganized. The Delta staining is more prominent in ISC of esg-Gal4/UAS-InR^A1325D flies. There are also more esg-Gal4/UAS-CD8GFP positive cells. (R–U) Colocalization of phospho-H3 positive cells with the ISC marker Delta in the gut of esg-Gal4/UAS-InR^A1325D flies. 93.0% (n=499) of the phospho-H3 positive cells also contained Delta staining.

Figure 7. Systemic insulin level can regulate ISC division

(A) Quantification of phospho-H3 positive cells in gut after insulin neurosecretory cell ablation. Two different transgenic lines of UAS-Reaper (Rpr, on X and 2^nd chromosome) were crossed with DILP2-Gal4. These Reaper expressing flies showed approximately 10 days delay of hatching. These flies were used for 3% DSS and 25 ug/ml bleomycin feeding experiments for 2 days. Control was 5% sucrose alone. There was approximately 85% reduction for DSS- and 50% reduction for bleomycin-induced ISC division. (B–G) Armadillo/β-catenin staining of DILP2-Gal4/UAS-Reaper fly guts. The wild type guts have more small cells and disorganized after DSS and bleomycin feeding, while the Dilp2-Rpr fly guts have fewer cells and appear more organized, similar to that of InR mutant. (H) Quantification of dilp2 mRNA expression in heads and guts. The RT-PCR cycle numbers at log phase of dilp2 and rp49 from wild type and DILP2-Gal4/UAS-Reaper were used to calculate the relative expression level 1/(2^cy#a-cy#rp49). The expression level of rp49 is 1. (I) DILP2-, Armadillo- and esg-Gal4 driver lines were crossed with UAS-DILP2, and the parental UAS-DILP2 and DILP2-Gal4 were included as controls. 3–5 day old flies from the crosses were assayed for phospho-H3 staining in the guts. Transgenic expression of DILP2 in various tissues causes a similar increase in ISC division. (J) A model for regulation of ISC division. The systemic level of insulin can regulate InR signaling in ISC to control proliferation. Damaged epithelium may release other signals, or change cell-cell contacts, that cooperate with insulin signaling to increase ISC division and subsequent enteroblast differentiation for tissue repair.