Damage-induced regeneration of the intestinal stem cell pool through enteroblast mitosis in the Drosophila midgut.

Many adult tissues and organs including the intestine rely on resident stem cells to maintain homeostasis and regeneration. In mammals, the progenies of intestinal stem cells (ISCs) can dedifferentiate to generate ISCs upon ablation of resident stem cells. However, whether and how mature tissue cells generate ISCs under physiological conditions remains unknown. Here, we show that infection of the Drosophila melanogaster intestine with pathogenic bacteria induces entry of enteroblasts (EBs), which are ISC progenies, into the mitotic cycle through upregulation of epidermal growth factor receptor (EGFR)-Ras signaling. We also show that ectopic activation of EGFR-Ras signaling in EBs is sufficient to drive enteroblast mitosis cell autonomously. Furthermore, we find that the dividing enteroblasts do not gain ISC identity as a prerequisite to divide, and the regenerative ISCs are produced through EB mitosis. Taken together, our work uncovers a new role for EGFR-Ras signaling in driving EB mitosis and replenishing the ISC pool during fly intestinal regeneration, which may have important implications for tissue homeostasis and tumorigenesis in vertebrates.

Introduction

The Drosophila midgut is the functional equivalent of mammalian small intestine where food is digested and nutrients are absorbed, and the intestinal epithelium protects the internal gut milieu from the external environment (Sansonetti, 2004; Jiang & Edgar, 2012; Li & Jasper, 2016; Zwick et al, 2019). During normal homeostasis, both the Drosophila midgut and the mammalian small intestine undergo cellular turnover to maintain tissue integrity and function. In response to tissue damage such as feeding with dextran sodium sulfate (DSS) or bacteria, they mount regenerative programs to accelerate stem cell division and differentiation to effectively replenish damaged mature cells (Amcheslavsky et al, 2009; Biteau et al, 2011; Jiang & Edgar, 2012; Jiang et al, 2016). In both mammals and Drosophila, the intestinal stem cells (ISCs) divide continuously to give rise to stem cells and progenitor cells, which differentiate into enterocytes (ECs), enteroendocrine (EE) cells (Fig 1A), Paneth cells, tuft cells, or goblet cells (Micchelli & Perrimon, 2006; Ohlstein & Spradling, 2006; Barker et al, 2007; Jiang & Edgar, 2012; Beumer & Clevers, 2016; Hung et al, 2020). These progenitors from ISCs in mammals enter the trans‐amplifying compartment to rapidly divide before terminal differentiation, but the enteroblast (EB) progenitor cells in the Drosophila intestine, which undergo endoreplication to become ECs, are not able to proliferate (Micchelli & Perrimon, 2006; Ohlstein & Spradling, 2006, 2007; Beumer & Clevers, 2016). During intestinal regeneration in Drosophila, the EGFR‐Ras pathway and other signaling pathways, such as the Notch, Wnt, Hh, Insulin, Hippo, JAK–STAT, JNK, and BMP pathways, and the Par complex, are implicated in the regulation of ISC self‐renewal, proliferation, and/or differentiation (Micchelli & Perrimon, 2006; Ohlstein & Spradling, 2006, 2007; Biteau et al, 2008; Lin et al, 2008; Jiang et al, 2009, 2011; Buchon et al, 2010; Karpowicz et al, 2010; Ren et al, 2010; Shaw et al, 2010; Staley & Irvine, 2010; O'Brien et al, 2011; Goulas et al, 2012; Biteau & Jasper, 2014; Tian & Jiang, 2014; Beehler‐Evans & Micchelli, 2015; Tian et al, 2015, 2017; Zeng & Hou, 2015; Chen et al, 2018). As a result of activation of EGFR‐Ras signaling, ISCs enter mitosis quickly to speed up proliferation and regeneration (Buchon et al, 2010; Biteau & Jasper, 2011; Jiang et al, 2011). However, whether EGFR‐Ras signaling induces EBs to enter mitosis remains unexplored.

Figure 1

Bacterial infection can induce EB mitosis

ISC lineages in Drosophila adult midguts. Dl marks ISCs. Su(H)‐Gal4>UAS‐CD8:GFP marks EBs. Pdm1 and Prospero (Pros) are markers for EC and EE, respectively.

The Drosophila posterior midguts containing Su(H)‐Gal4>UAS‐CD8:GFP (Su(H)>GFP) with sucrose (Suc, control) (B–C′), P.e. for 36 h (D–F′), Ecc15 for 36 h (G–I′), and DSS for 36 h (J–K′) were immunostained for GFP (green), PH3 (red), and DAPI (blue). (C, C′, E, E′, H, H′ and K, K′) Magnification of selected areas containing PH3+GFP− cells from B, D, G, and J. The EB cell close to the PH3+ cell (white arrow) in C, C′ is out of focus. (F, F′, I and I′) Magnification of selected areas containing PH3+ EBs (GFP+) from D and G. White arrowheads and red arrows indicate PH3 in GFP− and GFP+ cells, respectively.

Quantification of PH3+GFP− ISCs (L) and PH3+GFP+ EBs (M) with given treatments. n = 21 guts for each treatment.

Data information: Three independent experiments were performed, and error bars are ±SEM. ***p < 0.001 (Student's t‐test) (L, M).

In the mammalian intestine, the Dll1+ secretory progenitor cells, the mature tuft cells, and Paneth cells undergo dedifferentiation to produce regenerative ISCs after ablation of resident ISCs (van Es et al, 2012; Tetteh et al, 2016; Schmitt et al, 2018; Yu et al, 2018; Jones et al, 2019; Murata et al, 2020). By contrast, in Drosophila midguts, the differentiating ECs were found to produce regenerative ISCs through amitosis in the re‐feeding condition in starved midguts (Lucchetta & Ohlstein, 2017). The EB progenitor cells, which normally enter endoreplication to become ECs, have been shown to re‐enter the mitotic cycle upon P.e. infection or with misexpression of cell cycle genes such as Cdc25 phosphatase string (stg) and CycE, or loss of a CDK1 inhibitory kinase Myt1 (Kohlmaier et al, 2015; Willms et al, 2020). However, the mechanisms through which EBs enter mitosis in response to P.e. infection and whether EB mitosis produces regenerative ISCs remain unclear.

In this study, we find that EBs enter the mitotic cycle during intestinal regeneration in response to bacterial infection through activation of EGFR‐Ras signaling. We show that activation of EGFR‐Ras signaling in EBs is sufficient to induce EB mitosis. Following the lineages of EBs with the Flp‐out and two‐color lineage tracing systems, we show that the regenerative stem cells are produced through EB mitosis. Furthermore, we provide evidence to show that these regenerative stem cells are multipotent.

Results

Bacterial infection induces EBs to enter mitosis

To determine whether EBs enter the mitotic cycle upon infection of multiple bacterial strains, adult female flies were fed with pathogenic bacteria (Pseudomonas entomophila, P.e., or Erwinia carotovora carotovora strain 15, Ecc15) for 36 h (h), and mitosis in their midguts was examined by immunostaining for the mitotic marker PH3 (phospho‐Ser10‐Histone H3). The EBs were marked by expression of UAS‐GFP driven by the EB‐specific Gal4 (Su(H)‐Gal4) (Zeng et al, 2010) (Su(H)‐Gal4>UAS‐CD8:GFP is referred to as Su(H)>GFP) (Fig 1A). Adult female flies fed with sucrose (Suc) were used as the control. In midguts of adult female flies fed with P.e. or Ecc15, the number of the mitotic cells (PH3+) without GFP expression (i.e., ISCs) (PH3+GFP− cells) (Fig 1D–E′ and G‐H′, white arrowheads) was significantly increased when compared with those in the control (Fig 1B–C′) (Fig 1L: 64.8 PH3/gut in P.e. infection vs. 59.95 PH3/gut in Ecc15 infection vs.5.5 PH3/gut in the control; n = 21), similar to what was reported previously (Buchon et al, 2009; Jiang et al, 2011). Interestingly, PH3 staining was present in some EBs (GFP positive) in P.e. infected midguts (PH3+GFP+, Fig 1D, F and F′, red arrows), as previously reported (Kohlmaier et al, 2015), or in Ecc15‐infected midguts (PH3+GFP+, Fig 1G, I and I′, red arrows), but not in control midguts (Fig 1B) (Fig 1M: 8.33 PH3/gut in P.e. vs. 5.3 PH3/gut in Ecc15 vs. 0 PH3/gut in the control, n = 21), suggesting that bacterial infection prevents some EBs to enter the endoreplication cycle, and induces them to enter the mitotic cycle. The entry into mitosis in EBs in response to P.e. or Ecc15 infection suggests that pathogenic bacterial infection has a general effect on triggering EB mitosis in the adult midguts. In contrast, in the midguts of flies fed with dextran sodium sulfate (DSS) or Bleomycin (Bleo), which induces ISC proliferation (Amcheslavsky et al, 2009; Ren et al, 2013; Tian et al, 2015), PH3 signal was found only in ISCs without GFP expression (DSS: 29.8 PH3/gut, n = 21, Fig 1J–L. Bleo: 55.2 PH3/gut, n = 21, Fig EV1A), but not in EBs with GFP expression (Figs 1J and M, and EV1A).

Figure EV1

Mitosis in the midguts in response to different treatments

Feeding with Bleo can induce mitosis in ISCs, but not in EBs, and quantification of PH3+GFP− ISCs and PH3+GFP+ EBs with given treatments. n = 21 guts for each treatment. Three independent experiments were performed, and error bars are ±SEM. ***p < 0.001 (Student's t‐test).

Knockdown of stg in EBs can reduce EB mitosis upon P.e. infection, and quantification of PH3+ in EBs (GFP+) in the Drosophila midguts expressing Su(H) >CD8:GFP with or without UAS‐stg‐RNAi upon P.e. infection. n = 10 guts for each genotype. Three independent experiments were performed, and error bars ±SEM. ***p < 0.001 (Student's t‐test).

Activation of Insulin signaling with InR in EBs noncell‐autonomously promotes ISC proliferation, and quantification of PH3+ in ISCs (GFP−) and EBs (GFP+) in the Drosophila midguts expressing Su(H) >GFP with or without InR for 4 days at 29°C. n = 12 guts for each genotype. Three independent experiments were performed, and error bars ±SEM. ***p < 0.001 (Student's t‐test).

To ascertain that EBs indeed underwent mitosis upon bacterial infection, we examined the expression of Delta (Dl), an ISC marker, and found that the dividing ISCs without GFP expression (PH3+GFP−) showed Dl expression (Fig 2A–A″ and C–C″, white arrows), but these dividing EBs (PH3+GFP+) did not gain Dl expression (GFP+PH3+Dl−, Fig 2B–B″, red arrows, 100%, n = 89), suggesting that both EBs and ISCs enter mitosis upon bacterial infection. We performed additional experiments to examine EB mitosis in which ISCs were marked by the expression of UAS‐GFP driven by esgGal4, su(H)‐Gal80. Adult female flies with esgGal4, su(H)‐Gal80>UAS‐GFP were fed with Suc, P.e. bacteria or Dss, and their intestines were dissected out for immunostaining with anti‐GFP, PH3, and Pros (the EE cell marker) antibodies. We found PH3 in GFP+ ISCs in all conditions (Fig 2D–F″, white arrows). However, PH3 expression was present in EBs marked by Pros− GFP− and small nuclei only upon P.e. infection (Fig 2E–E″, red arrows, G), but not upon Suc and Dss feeding. We also knocked down stg, the Cdc25 phosphatase that activates the mitotic kinase CDK1, in EBs with su(H) >UAS‐stg‐RNAi and examined whether EB mitosis was blocked in response to P.e. infection. The number of EBs with PH3 expression was significantly reduced compared with intestines without stg knockdown (Fig EV1B). These results confirmed that EBs can enter the mitotic cycle upon bacterial infection.

Figure 2

Dividing EBs do not gain Dl expression

The midguts from the adult female flies fed with Suc (control) (A–A″), P.e. (B–B″) or DSS (C–C″) were immunostained with anti‐PH3 and anti‐Dl antibodies. The white arrows indicate that GFP− cells with PH3 have Dl staining and the red arrows indicate that EBs with GFP and PH3 (GFP+PH3+) do not have Dl expression. The yellow arrows in B show that two ISCs with PH3 are out of focus.

When the ISCs were marked by esgGal4, su(H)‐Gal80>UAS‐GFP, PH3 expression was found in GFP+ ISCs with feeding of Suc (control) (D–D″), P.e. (E–E″) and Dss (F–F″). However, PH3 expression is present in EBs with GFP−Pros− and small nuclei upon P.e. infection (E–E″, red arrows). The white arrows indicate GFP+ cells with PH3 expression.

Quantification of PH3+GFP−Pros− EBs with given treatments. n = 11 guts for each treatment. Data information: Three independent experiments were performed, and error bars are ±SEM. ***p < 0.001 (Student's t‐test) (G).

In response to P.e. infection, both ISCs and EBs can enter mitosis, therefore we wonder whether EB mitosis and ISC mitosis are compensatory. To test this, we blocked ISC proliferation by expressing UAS‐stg‐RNAi with Dl‐Gal4 and examined EB mitosis upon P.e. infection. We found that expression of stg in ISCs significantly reduced overall PH3+ cell number (Fig 3A). Quantification of mitotic EBs (PH3+GFP−Pros−) revealed a slight reduction rather than increase in EB mitosis when ISC proliferation was blocked (Fig 3B), indicating that EB mitosis does not act as a compensatory mechanism for ISC mitosis upon injury.

Figure 3

The EB mitosis is not compensatory for ISC mitosis in response to P.e. infection

Quantification of PH3+ cells when the ISC mitosis was blocked by expressing stg‐RNAi in ISCs with feeding of Suc or P.e. n = 10 guts for each treatment.

Quantification of PH3+ EBs (GFP−Pros−) with or without stg expression in ISCs upon P.e. infection. n = 10 guts for each treatment.

Data information: Three independent experiments were performed.

EGFR‐Ras signaling in EBs is required and sufficient to induce EB mitosis

Bacterial infection activates several signaling pathways including the EGFR‐Ras pathway through up‐regulation of ligands to induce ISC proliferation (Jiang et al, 2011). To assess the expression level of ligands of the EGFR‐Ras pathway, we performed qRT‐PCR analysis and found that P.e. infection increased the expression of ligands vein (vn) and Keren (Krn) (Fig 4A). In addition, we examined the activity of mitogen‐activated protein kinase (MAPK) in EBs, which marks the activity of EGFR‐Ras signaling, by using an antibody against the diphosphorylated and active form of MAPK (dpERK) (Gabay et al, 1997). Without P.e. infection, the dpErk level was high in ISCs (GFP−, Fig 4B–B″, yellow arrowheads) but low in EBs (GFP+, Fig 4B–B″, white arrows) (Jiang et al, 2011). Upon P.e. infection, the level of dpERK was greatly increased in both EBs (Fig 4C–C″, white arrows) and ISCs (Fig 4C–C″, yellow arrowheads), suggesting that EGFR‐Ras signaling was up‐regulated in both EBs and ISCs upon bacterial infection.

Figure 4

Activation of EGFR‐Ras signaling in EBs is sufficient and required to induces cell‐autonomous EB mitosis

P.e. infection up‐regulates the expression of EGFR ligands (vn and Krn).

In the control, the dpErk level is high in ISCs (yellow arrowheads), but low in EBs (white arrows).

The level of dpErk is greatly up‐regulated in both EBs (white arrows) and ISCs (yellow arrowheads) upon bacterial infection.

The Drosophila midguts expressing Su(H) >GFP with control (D–E′), UAS‐Ras for 6 days (F–H′), EGFR for 4 days (I–K′), and UAS‐λTop for 5 days (L–N′) were immunostained for GFP (green), PH3 (red) and DAPI (blue). (E, E′, G, G′, J, J′, M, and M′) Magnification of selected areas containing PH3+GFP− cells from D, F, I and L. (H, H′, K, K′, N, and N′) Magnification of selected areas containing PH3+GFP+ cells from F, I, and L. White arrowheads and red arrows indicate PH3 in GFP− ISCs and GFP+ EBs, respectively.

Quantification of PH3+GFP− (P) and PH3+GFP+ (O) in midguts with indicated genotypes. n = 13 guts for each genotype.

Quantification of PH3+ EBs (GFP+) with the given genotypes in response to P.e. infection. n = 11 guts for each genotype.

Data information: Three independent experiments were performed, and error bars are ±SEM. **p < 0.01 (Student's t‐test) (A). ***p < 0.001 (Student's t‐test) (O−Q).

To determine whether activation of EGFR‐Ras signaling induces EBs to enter the mitotic cycle, we overexpressed the active form of Ras (UAS‐Ras ) (Karim & Rubin, 1998 ) with the EB‐specific inducible Gal4/Gal80ts system (Su(H)‐Gal4 Tub‐Gal80 ; referred to as Su(H) ) and examined the expression of the mitotic marker (PH3). In the experiments with Gal80ts, female flies expressing Su(H) >UAS‐CD8:GFP with or without UAS − gene were raised to adults at 18°C (Gal4 is ‘off’) and then shifted to 29°C to degrade Gal80ts (Gal4 is ‘on’) so that Su(H)‐Gal4 can drive expression of UAS‐gene. Our assay with expression of Ras in EBs showed that activation of Ras in EBs induced cell‐autonomous EB mitosis (GFP+ PH3+, Fig 4F, H and H′, red arrows; Fig 4O: 12.7 PH3/gut in Ras vs. 0 PH3/gut in the control (Fig 4D), n = 13). In addition, we found that activation of Ras in EBs promoted ISC proliferation noncell‐autonomously (PH3+ GFP−, Fig 4F–G′, arrowheads, Fig 4P: 27.8 PH3/gut in Ras vs. 5.4 PH3/gut in the control (Fig 4D–E′), n = 13), indicated by increased number of PH3 in GFP− ISCs. To confirm that EGFR‐Ras signaling induces EB mitosis, we used the same EB‐specific Gal4/Gal80 system (Su(H) ) to overexpress the active forms of EGFR (EGFR and λTop) (Queenan et al, 1997; Lesokhin et al, 1999) in EBs, and found that overexpression of either EGFR or λTop induced EB mitosis (PH3+GFP+; Fig 4I, K, K′, L, N and N′, red arrows, Fig 4O: 9.9 PH3/gut in EGFR vs. 6.6 PH3/gut in λTop vs. 0 PH3/gut in the control, n = 13) and increased ISC mitosis noncell‐autonomously (GFP− PH3+, Fig 4I–J′ and L–M′, arrowheads, Fig 4P). As a control, we used Su(H) to drive expression of an active form of InR (UAS‐InR ) in EBs and the PH3 signal was detected in ISCs but was not detected in EBs (Fig EV1C).

As P.e. infection up‐regulated EGFR‐Ras signaling and activation of EGFR‐Ras signaling induced EB mitosis, we asked whether EGFR‐Ras signaling was required for EB mitosis in response to P.e. infection. To this end, we inactivated EGFR‐Ras signaling in EBs by knocking down either EGFR or Ras with UAS‐EGFR‐RNAi (or UAS‐Ras‐RNAi) driven by Su(H) (referred to as Su(H) >EGFR‐RNAi (or Su(H) >Ras‐RNAi)) and analyzed the mitotic marker in the EBs upon P.e. infection. The knockdown of either EGFR or Ras decreased the frequency of EB mitosis, as indicated by the reduced number of EBs with PH3 signal (PH3+ GFP+, Fig 4Q: 8.9 PH3/gut in control vs. 4.5 PH3/gut in EGFR‐RNAi vs. 4.4 PH3/gut in Ras‐RNAi; n = 11), indicating that EGFR‐Ras signaling in EBs is required for EB mitosis upon bacterial infection.

To determine whether the GFP+ EBs gain ISC identity as a prerequisite to divide, we analyzed the expression of Dl, the ISC marker, in midguts with EGFR expression. In contrast to the ISCs that showed expression of both PH3+ and Dl (Fig 5A–A″, white arrows), the dividing EBs (GFP+) did not possess Dl expression (Fig 5B–B″, red arrows; 100%, n = 42), suggesting that these EBs do not gain ISC identity. In summary, activation of EGFR‐Ras signaling in EBs induces cell‐autonomous EB mitosis and promotes ISC proliferation noncell‐autonomously.

Figure 5

Activation of EGFR‐Ras signaling in EBs induces EB mitosis

PH3 and Dl staining in the control (A–A″) or midguts with EGFR overexpression (B–B″). The white arrows indicate that GFP− ISCs with PH3 have Dl staining (A–B″); the red arrows indicate that GFP+ EBs with PH3 (GFP+PH3+) do not have Dl expression (B–B″).

The different mitotic phases in ISCs with PH3 and the centrosome marker (Cnn‐GFP) with Dl‐Gal4>UAS‐Cnn‐GFP.

The different mitotic phases in EBs with PH3 and Cnn‐GFP when EGFR (F–H) or Ras (I–K) is expressed in EBs with Su(H)‐Gal4 or these flies were fed with P.e. (L–N).

To further characterize EB mitosis, we examined the phases of mitosis in EBs and ISCs with the mitotic marker PH3 and the centrosome marker Cnn‐GFP. Different phases of mitosis such as prophase, metaphase and anaphase were found in dividing EBs (Fig 5F–N), as were detected in ISCs (Fig 5C–E), suggesting normal mitosis in EBs, but difference from the amitosis (Lucchetta & Ohlstein, 2017).

P.e. infection induces EBs to generate Dl+ ISC‐like cells through EGFR‐Ras signaling

To determine the cell fate of the progenies of EB mitosis, we used the Flip‐out lineage tracing system (Fig 6A; Golic & Lindquist, 1989; Germani et al, 2018). In this system, Su(H)‐Gal4 driven flippase (Flp) induces the removal of the “stop cassette” in actP (FRT Stop FRT) LacZ (referred to as actP>stop>lacZ), thus LacZ is expressed to mark the EBs and their progenies. For lineage tracing of control flies fed with Suc, the GFP+ EBs were labeled with LacZ expression (Dl GFP + LacZ +, Fig 6B–B‴, white arrows), whereas LacZ expression was absent in the ISCs (Dl + GFP , Fig 6B–B‴, blue arrowheads). In contrast, after feeding with P.e., LacZ expression was not only found in GFP+ EBs (Fig 6C–C‴, white arrows), but also in some GFP− cells (Fig 6C–C‴, red arrowheads). The expression of LacZ but lack of GFP suggests that these cells are progenies of EBs but are different from the EB fate. In some of these LacZ + GFP − cells, Dl expression was detected (Fig 6C–C‴, red and blue arrowheads, Fig 6D), suggesting that these EB progeny cells may have adopted an ISC‐like fate upon P.e. infection. In addition, some EBs (GFP+) without LacZ expression had been found due to quick regeneration of EBs from ISCs (Fig 6C).

Figure 6

Dl ISC‐like cells are generated from EBs upon

The schematic drawing of the Flip‐out lineage tracing system.

Flies expressing UAS‐Flp; Su(H)ts UAS‐CD8:GFP; ActP>Stop>LacZ were fed with Suc (B, B‴) or P.e. (C, C‴) for 36 h and their midguts were dissected and immunostained with antibodies against GFP, β‐gal and Dl. Blue arrowheads indicate Dl+ ISCs; white arrows indicate GFP+ EBs; and red arrowheads indicate EB progeny cells (LacZ+GFP−).

Quantification of Dl+ ISC‐like cells from EBs (GFP−LacZ+Dl+) in the control, P.e. infected midguts, n = 50.

Quantification of Dl+ ISC‐like cells from EBs (GFP−LacZ+Dl+) in the control with Suc or P.e., or knockdown of EGFR with Suc or P.e. infection, n = 50.

Data information: Three independent experiments were performed, and error bars are ±SEM. ***p < 0.001 (Student's t‐test) (D, E).

To determine whether the generation of new Dl+ ISC‐like cells from EBs upon P.e. infection depends on EGFR‐Ras signaling, we knocked down EGFR in EBs using the Flp‐out system. We found that the total number of regenerative Dl+ cells from EBs in response to P.e. infection was reduced when EGFR was knocked down (total number of GFP LacZ + Dl + cells in five midguts: 28.5 in P.e. infection vs. 11.3 in EGFR‐RNAi with P.e. infection, Fig 6E), indicating that EGFR‐Ras signaling is required for generation of Dl+ ISC‐like cells from EBs.

Activation of EGFR‐Ras signaling in EBs is sufficient to drive EBs to produce dl+ ISC‐like cells

To determine whether activation of EGFR‐Ras signaling in EBs is sufficient to induce EBs to produce Dl+ ISC‐like cells, we used the same Su(H)‐Gal4‐driven Flip‐out lineage tracing system to express UAS‐Ras in EBs and compared it with the control (Fig 7A–A‴). We found that LacZ expression was detected in EB cells (GFP + LacZ +, Fig 7B–B‴, white arrows) and their progeny (GFP − LacZ +, Fig 7B–B‴, red arrowheads) upon Ras expression. In addition, we found that some EB progeny cells (LacZ + GFP ) exhibited Dl expression (GFP LacZ + Dl +) (Fig 7B–B‴, red and blue arrowheads, and Fig 7D), indicating that activation of EGFR‐Ras signaling in EBs is sufficient to induce EBs to generate Dl+ ISC‐like cells.

Figure 7

Activation of EGFR‐Ras signaling in EBs induces EBs to produce Dl

ISC‐like cells

The midguts from adult female flies expressing UAS‐Flp; Su(H) UAS‐CD8:GFP; ActP>Stop>LacZ with (B–B‴) or without (A–A‴) UAS‐Ras or with UAS‐Ras and UAS‐stg‐RNAi (C–C‴) were dissected and immunostained with antibodies against GFP, β‐gal and Dl. The bigger GFP− nucleus with LacZ expression (only red arrowheads) indicate differentiating or mature ECs. Blue arrowheads indicate Dl+ ISCs; white arrows indicate GFP+ EBs; and red arrowheads indicate EB progeny cells (LacZ+GFP−).

Quantification of Dl+ ISC‐like cells from EBs (GFP LacZ + Dl +) in the indicated genotypes, n = 50.

Data information: Three independent experiments were performed, and error bars are ±SEM. ***p < 0.001 (Student's t‐test) (D).

New ISCs are generated through EB division

To confirm that the EB progeny cells with Dl expression are generated from the mitotic divisions, we blocked EB mitosis by knocking down stg and examined whether the new Dl+ ISC‐like cells were still generated from EBs. To this end, we co‐expressed Ras and stg‐RNAi in EBs with the Su(H)‐Gal4‐driven Flip‐out system and examined the lineages from EBs. We found that the frequency of new Dl+ ISC‐like cells (LacZ+GFP cells) was significantly reduced (Fig 7C and D), suggesting that the generation of new Dl+ ISC‐like cells is suppressed by inhibiting EB mitosis.

Next, we directly marked two cells from one EB division and characterized the two cells by using the two‐color lineage tracing system (Fig 8A and B). This system is based on the mitotic division of cells, thus only cells from EB division will be marked by the single color (GFP or RFP). After the single‐color clones are produced and the number of cells in single‐color clones is counted, we can determine whether the cells from EBs are ISCs (Fig 8B). In this experiment, 3–5‐day‐old flies expressing UAS‐Flp; Su(H) ; FRT82BGFP/FRT82BRFP with or without UAS‐EGFR raised in 18°C were transferred to 29°C for 5 days and these pairs of single‐color clones with GFP or RFP in the midguts were analyzed. We found that pairs of single‐color clones were produced in midguts with activation of EGFR‐Ras signaling by UAS‐EGFR expression (Figs 8D–D″ and F, and EV2), but not in the control midguts (Fig 8C–C″ and F). This result indicates that EBs with EGFR expression underwent mitosis, but EBs in the control did not. Interestingly, after the number of cells in single‐color clones was counted, we found that the clones in most of pairs of single‐color clones had more than one cell (Fig 8D–D″ and F), indicating that EBs underwent symmetric cell division to produce two ISCs. In addition, we found low frequency of clones with asymmetric cell division, as indicated by one clone with one cell and the other clone with multiple cells in the pair of single‐color of clones (Figs EV2B–B″ and 8F), or with symmetric non‐ISC division, as marked by only one cell in each of the pair of single‐color of clones (Figs EV2A–A″ and 8F). Then, we examined the progenies of EBs and the division pattern of EBs in response to P.e. infection with the two‐color lineage tracing system. After feeding with P.e. for 36 h and normal food for 24 h, single‐color clones were found (Fig 8E–E″ and F), and there was more than one cell in most of the pairs of single‐color clones (Fig 8F). In summary, the results indicate that both activation of EGFR‐Ras signaling and bacterial infection induce EBs to generate ISCs through mitosis.

Figure 8

Activating EGFR‐Ras signaling in EBs or P.e. infection induces EB division to generate two ISCs

Schematic drawing of the cell division that produces differentially labeled twin‐spot cells (RFP+ GFP− and RFP−GFP+) through FRT‐mediated mitotic recombination. UAS‐Flp and transgenic overexpression is driven by Su(H) .

Schematic drawings of differentially labeled twin‐spot clones generated by FLP/FRT‐mediated mitotic recombination of dividing cells in EBs.

The representative twin clones from midguts when EGFR is overexpressed in EBs (D–D″, red and green arrows) or adult flies were fed with P.e. for 36 h (E–E″, red and green arrows), but no twin clone is found in the control (C–C″).

Quantification of different types of twin clones in the control midguts, or in the midguts with EGFR overexpression or P.e. infection. n = 10 guts for each genotype.

Data information: Three independent experiments were performed (F).

Figure EV2

Activation of EGFR‐Ras signaling in EBs induces an EB to generate one ISCs or non‐ISCs

Representative twin clones from midguts show symmetric division in EBs to produce two non‐ISCs (A–A″) or asymmetric division in EB to produce one ISC with multiple cells (green) and one non‐ISC (red) (B–B″).

The single‐color clones with more than one cell indicate that new cells from EB mitosis are ISCs which can self‐renew and divide. Our next question is whether the new ISCs from EBs can differentiate into mature cells (EE cells or ECs). To test this, we first examined whether the progenies from EBs can differentiate into EE cells or pre‐EE cells with an EE marker, Prospero (Pros). When we performed the Flp‐out lineage tracing experiments with activation of EGFR‐Ras signaling or P.e. infection and immunostaining with the antibody against Pros and compared them with the control (Fig 9A–A″), we found that some EB progeny cells without GFP expression (LacZ + GFP ) have Pros expression (Fig 9B–D), indicating that these ISCs from EBs are functional to generate pre‐EE or EE cells. Second, we examined whether the progenies from EBs can differentiate into ECs with an EC marker (Pdm1). Because the Flp‐out lineage tracing system cannot distinguish between ECs from original EBs and ECs from the regenerative ISC‐EBs, we examined single‐color clones with multiple cells upon EGFR expression or P.e. infection with the EC marker (Pdm1). We found that the Pdm1 signal were identified in the single‐color clones (Fig 9E–F″, 100% in single‐color clones with multiple cells, n = 31), indicating that the new ISCs from EB mitosis can differentiate into mature ECs. Taken together, these results suggest that the regenerative ISCs from EBs are multipotent and functional.

Figure 9

The regenerative ISCs from EBs can differentiate into mature cells

ISCs from EBs upon Ras expression or P.e. infection could differentiate to generate EE or pre‐EE cells (B‐C″, LacZ + Pros + GFP , blue arrowheads), but no EE or pre‐EE cells from EBs are found in the control (A–A″, white arrowheads).

Quantification of EE cells from progenies of EBs with indicated genotypes or treatments. n = 9 guts for each genotype or treatment.

ECs which were marked by Pdm1 expression were found in pairs of single‐color clones (white arrows) in midguts with EGFR expression or upon P.e. infection.

A model for bacterial infection and activation of EGFR‐Ras signaling to induce EB mitosis and ISC regeneration.

Data information: Three independent experiments were performed, and error bars are ±SEM. ***p < 0.001 (Student's t‐test) (D).

Discussion

Previous studies found that only resident ISCs in Drosophila midguts localized at the basal side of the gut epithelium undergo asymmetric cell division to produce renewed ISCs and EBs (Micchelli & Perrimon, 2006; Ohlstein & Spradling, 2006). Our studies show that EBs can generate functional stem cells (Fig 9G), and there are two possibilities about the process of regeneration of ISCs with mitosis. One could be that EBs directly revert to ISCs by direct reprogramming (Higa et al, 2022), like induced pluripotent stem cells (Takahashi & Yamanaka, 2006), and then start mitosis. The other one could be that EBs enter mitosis, and then produce ISCs. Our results from immunostaining with Dl support the latter statement, and the two‐color lineage tracing experiments further demonstrate that two regenerative ISCs are produced from one EB division, which is distinct from amitosis in the intestine (Lucchetta & Ohlstein, 2017). In addition, we examined expression of genes or activity of signaling pathways for ISC maintenance, such as esg (Korzelius et al, 2014), Notch signaling (Micchelli & Perrimon, 2006; Ohlstein & Spradling, 2006, 2007), and integrin (Lin et al, 2013), in dividing EBs and nondividing EBs. We found down‐regulation of esg expression (Fig EV3A–A‴), but no change in Notch signaling (Fig EV3B and B′) and integrins (Fig EV3D and D′), suggesting no direct reprogramming of EBs. Therefore, activation of EGFR‐Ras signaling in EBs which forces EBs to enter the mitotic cycle, together with mitosis, plays a role in the cell fate determination toward ISCs. The replenishment of stem cell pool in mammals is mediated by dedifferentiation, and regeneration of ISCs through mitosis could be another conserved mechanism from Drosophila to mammals.

Figure EV3

Expression of reporters in the dividing EBs

A representative image showed that esg expression with a reporter esg‐lacZ was down‐regulated in mitotic EBs upon P.e. infection (red arrows). White arrows indicate GFP+ EBs without PH3 expression.

The Notch reporter E(spl)mb‐CD2 was normally expressed in dividing EBs (red arrows). White arrows indicate GFP+ EBs without PH3 expression.

The integrin was normally expressed and localized in dividing EBs (D, D′, red arrows) compared with nondividing EBs (C, C′, white arrows).

The Su(H)‐Gal4 has been used to mark the EB daughter cells with elevated Notch signaling (Zeng et al, 2010). This Gal4 is also used in previous lineage tracing studies to show its expression only in EBs but not in ISCs, and this Gal4‐induced lineages include ECs but no EE cells, suggesting the specificity of this Gal4 in labeling EBs in the adult midguts (Biteau & Jasper, 2014; Zeng & Hou, 2015). Resident ISCs are unique in their expression of Dl, but Dl was not detected in the dividing EBs upon bacterial infection or activation of EGFR‐Ras signaling, suggesting that these dividing EBs are not ISCs. Therefore, these EBs are referred to as uncommitted EBs, which adopt a fate between the ISCs and EBs. In our lineage tracing experiments, EE cells which are derived from ISCs during normal intestinal development (Biteau & Jasper, 2014; Guo & Ohlstein, 2015; Zeng & Hou, 2015) were identified, indicating that these regenerative ISCs are multipotent. Although we cannot exclude the possibility that these EE cells are generated from EBs with activation of EGFR‐Ras signaling or P.e. infection, our further experiment with the production of mature ECs from the progenies of regenerative ISCs suggests that the regenerative ISCs from these uncommitted and dividing EBs are multipotent.

Infection of two different Gram‐negative bacteria induces the same EB mitosis phenotype, indicating that the immunity signaling pathways might be involved in EB mitosis and/or regeneration of ISCs. The previous studies showed that bacterial infection could induce the IMD pathway in the Drosophila intestine (Buchon et al, 2009; Zhai et al, 2018), but another study reported that Ras/MAPK pathway suppresses IMD in the intestine and fat body (Ragab et al, 2011). Therefore, it is unlikely that EGFR‐Ras signaling regulates EB mitosis through IMD pathway. Both bacterial infection and activation of EGFR‐Ras signaling can activate JAK–STAT pathway in ISCs through up‐regulation of ligands to stimulate ISC proliferation (Jiang et al, 2009, 2011; Zhai et al, 2018). In our studies, noncell‐autonomous ISC proliferation was observed when EGFR‐Ras signaling was activated in EBs, suggesting that the cytokines of the JAK–STAT pathway could be up‐regulated. Indeed, we examined expression of cytokines when EGFR‐Ras signaling was activated in EBs and found that upd3 was up‐regulated (Fig EV4). In addition, JAK–STAT signaling was found to be sufficient to induce dedifferentiation of spermatogonia into germline stem cells in the Drosophila testis (Brawley & Matunis, 2004). Thus, EGFR‐Ras signaling may regulate EB mitosis through the JAK–STAT pathway. To test how specific it is for EGFR‐Ras signaling to regulate EB mitosis, we examined several other signaling pathways such as Wnt, Hh, Hippo, and JNK, which can regulate ISC proliferation, we found that activation of Wnt, Hh, and JNK could not induce EB mitosis, and loss of wts can stimulate EB mitosis (Fig EV5). Further studies will aid our understanding of the mechanisms underlying intestinal regeneration.

Figure EV4

Activation of EGFR‐Ras signaling in EBs up‐regulates upd3 expression

Activation of EGFR‐Ras signaling in EBs with Su(H) >UAS‐EGFR up‐regulates the expression of upd3. Three independent experiments were performed, and error bars are ±SEM. ***p < 0.001 (Student's t‐test).

Figure EV5

Knockdown of wts in EBs promotes EB mitosis

Quantification of PH3+ in EBs (GFP+) in the midguts expressing Su(H) >CD8:GFP with or without UAS‐wts‐RNAi for 5 days at 29°C. n = 10 guts for each genotype. Three independent experiments were performed, and error bars ±SEM. ***p < 0.001 (Student's t‐test).

Dysplasia in the mammalian gastrointestinal tract, which is considered as a carcinoma precursor (Li & Jasper, 2016; Pulusu & Lawrance, 2017) and characterized by atypical cellular features, aberrant cell proliferation, and differentiation as well as disorganized architecture, can be mimicked in the Drosophila intestine (Micchelli & Perrimon, 2006; Ohlstein & Spradling, 2006; Biteau et al, 2008; Apidianakis et al, 2009; Jiang et al, 2009). Upon bacterial infection, the Drosophila intestine undergoes rapid regeneration to replace dying cells and this rapid regeneration can be subverted toward dysplasia (Apidianakis et al, 2009; Jiang et al, 2009). During regeneration, ISCs with increased proliferation and increased pool size produce more differentiated cells to replace the lost or damaged cells (Apidianakis et al, 2009; Jiang et al, 2009). Contrary to previous reports, our results indicate that both ISCs and EBs can enter the mitotic cycle, which could make regeneration more efficient. We also find that EBs can generate functional ISCs in response to injury or oncogenic pathway activation, which may contribute to the increased pool of ISCs. Previous studies showed that Kras with Apc and activation of NF‐ƙB (Schwitalla et al, 2013) acts as a driving force to promote tumorigenesis by dedifferentiation in mammalian colorectal cancer model (Janssen et al, 2006). Therefore, it would be interesting to determine whether the dedifferentiation is mediated by mitosis in mammalian intestines to contribute to tumorigenesis.

Material and Methods

Drosophila genetics and transgenes

Transgenic lines included UAS‐EGFR‐RNAi (VDRC43267), UAS‐EGFR (BL#9534), UAS‐Ras (Jiang et al, 2011), UAS‐Ras‐RNAi (BL#34619), UAS‐InR (BL#8440), UAS‐stg‐RNAi (BL#34831), UAS‐Flp (BL#55808), FRT82Bubi‐GFP/CyO, FRT82Bubi‐RFP/CyO, Su(H)‐Gal4, Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP, UAS‐FLP; Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP; actP>CD2>LacZ/+, esgGal4/UAS‐GFP, su(H)‐Gal80 and Dl‐Gal4, tub‐Gal80 UAS‐GFP. For feeding experiments with bacteria, Bleomycin (Bleo) or dextran sodium sulfate (DSS), 2–3 days‐old female adult flies were used. Flies were cultured in an empty vial containing a piece of chromatography paper (Fisher) wet with 5% (wt/vol) sucrose (Suc) solution as feeding medium (mock treatment) or with P.e. or Ecc15 or 5% DSS or Bleo (25 μg/ml) and 5% (wt/vol) sucrose for one or more days. For experiments with tubGal80ts, the cross with right genotypes were set up and cultured at 18°C to restrict Gal4 activity. Two‐ to three‐day‐old F1 adult flies were shifted to 29°C for the indicated periods of time to inactivate Gal80ts and allow Gal4 to activate UAS transgenes. For bacterial infection with knockdown of genes, flies bearing Su(H) >GFP or Su(H) >GFP + EGFR‐RNAi (or Ras‐RNAi) were raised to adults at 18°C, and these adult females were transferred to 29°C for 6 days and then fed with P.e. for 36 h. For the lineage tracing experiments with P.e. infection, flies bearing UAS‐Flp; Su(H) UAS‐GFP; actP>stop>LacZ (or with UAS‐EGFR‐RNAi) were raised to adults at 18°C and then were transferred to 29°C for 6 days and followed by feeding with Suc (control) or P.e. for 36 h and normal food for 24 h, and their midguts were immunostained with antibodies against GFP, β‐gal and Dl (ISC marker) or Pros (EE marker). For two‐color lineage tracing experiments, the cross with right genotypes were set up and cultured at 18°C, and then 3–5‐day‐old flies expressing UAS‐Flp; Su(H) ; FRT82BGFP/FRT82BRFP with or without UAS‐gene raised in 18°C were transferred to 29°C for the indicated time.

Immunostaining

Female flies were used for gut immunostaining in all experiments. The entire intestine was dissected out and fixed in 1× PBS plus 8% EM‐grade paraformaldehyde (Polysciences) for 1 h. Samples were washed and incubated with primary and secondary antibodies in a solution containing 1× PBS, 0.5% BSA, and 0.1% Triton X‐100. The following primary antibodies were used: mouse anti‐Dl (DSHB), 1:100; rabbit anti‐LacZ (MP Biomedicals), 1:1,000; rabbit and mouse anti‐PH3 (Millipore), 1:1,000; goat anti‐GFP (Abcam), 1:1,000; Mouse anti‐Pros (MR1A); Rabbit anti‐Pdm1 (gift from X. Yang, Institute of Molecular and Cell Biology, Singapore). Secondary antibodies conjugated to Alexa Fluor 546 donkey anti‐mouse and anti‐rabbit (Molecular Probes) and Alexa Fluor 633 donkey anti‐mouse and anti‐rabbit and 488 Donkey anti‐goat (Jackson immunoresearch) were used at 1:400. Fluorescently labeled samples were counterstained with DAPI for visualization of DNA. Images were captured with a Zeiss LSM 800 confocal microscope and assembled in Adobe Photoshop.

RT‐qPCR

Total RNA was extracted from 10 female guts using Rneasy Plus Mini Kit (74134; Qiagen), and cDNA was synthesized using the iScript cDNA synthesis kit (Bio‐Rad). RT‐qPCR was performed using iQ SYBR Green System (Bio‐Rad). Primer sequences used are: 5′‐TCACACATTTAGTGGTGGAAG‐3′ and 5′‐TTGTGATGCTTGAATTGGTAA‐3′ (for vn), 5′‐CGTGTTTGGCAACAACAAGT‐3′ and 5′‐TGTGGCAATGCAGTTTAAGG‐3′ (for Krn), 5′‐CGCCCAAGAATGAAAGAGAG‐3′ and 5‐AGGTATGCTGCTGGTGGAAC‐3′ (for Spi) and 5′‐GAGCACCAAGACTCTGGACA‐3′ and 5′‐CCAGTGCAACTTGATGTTGC‐3′ (for upd3). RpL11 was used as a normalization control. Relative quantification of mRNA levels was calculated using the comparative CT method.

Statistical analyses

Statistical analyses were performed with a two‐tailed unpaired t‐test. p‐Value is indicated by asterisks in the Figures: **p < 0.01; ***p < 0.001. Differences at p < 0.01 were considered significant.

Genotypes for flies in each figure

Figure 1. (B–K′) Su(H)‐Gal4 UAS‐CD8:GFP/+.

Figure 2. (A–C″) Su(H)‐Gal4 UAS‐CD8:GFP/+, (D–F″) esgGal4/UAS‐GFP, su(H)‐Gal80.

Figure 4. (B–C″) Su(H)‐Gal4 UAS‐CD8:GFP/+. (D–E′) Su(H)‐ Gal4 tub‐ Gal80 UAS‐CD8:GFP/+. (F–H′) Su(H)‐Gal4 tub‐ Gal80 UAS‐CD8:GFP /UAS‐Ras . (I–K′) Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP /UAS‐EGFR . (L–N′) Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP /UAS‐λTop.

Figure 5. (A–A″) Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP /+. (B–B″) Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP /UAS‐EGFR . (C–E) UAS‐Cnn‐GFP /+; Dl‐Gal4 tub‐Gal80 /+. (F–H) UAS‐Cnn‐GFP/+; Su(H)‐Gal4 tub‐Gal80 /UAS‐EGFR . (I–K) UAS‐Cnn‐GFP/+; Su(H)‐Gal4 tub‐Gal80 /UAS‐Ras . (L–N) UAS‐Cnn‐GFP/+; Su(H)‐Gal4 tub‐Gal80 /+.

Figure 6. (B–C‴) UAS‐Flp; Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP; actP>stop>LacZ.

Figure 7. (A–A″) UAS‐Flp; Su(H)‐Gal4 tub‐Gal80 ‐Gal4 tub‐Gal80 UAS‐CD8:GFP; actP>stop>LacZ. (B–B‴) UAS‐Flp; Su(H) UAS‐CD8:GFP/UAS‐Ras ; actP>stop>LacZ. (C–C‴) UAS‐Flp; Su(H) UAS‐CD8:GFP/UAS‐Ras ; actP>stop>LacZ/UAS‐stg‐RNAi.

Figure 8. (C–C″, E–E″) UAS‐Flp; Su(H)‐Gal4 tub‐Gal80 ; FRT82BGFP/FRT82BRFP. (D–D″) UAS‐Flp; Su(H)‐Gal4 tub‐Gal80 /UAS‐EGFR ; FRT82BGFP/ FRT82BRFP.

Figure 9. (A–A″, C–C″) UAS‐Flp; Su(H) ‐Gal4 tub‐Gal80 UAS‐CD8:GFP; actP>stop>LacZ. (B‐B″) UAS‐Flp; Su(H) ‐Gal4 tub‐Gal80 UAS‐CD8:GFP/UAS‐Ras ; actP>stop>LacZ. (E–E″) UAS‐Flp; Su(H)‐Gal4 tub‐Gal80 /UAS‐EGFR ; FRT82BGFP/FRT82BRFP. (F–F″) UAS‐Flp; Su(H)‐Gal4 tub‐Gal80 ; FRT82BGFP/FRT82BRFP.

Genotypes for flies in each supplementary figure

Figure EV1. (A) Su(H)‐Gal4 UAS‐CD8:GFP/+. (B) Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP/+; UAS‐stg‐RNAi/+. (C) Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP/+; UAS‐InR /+.

Figure EV2. UAS‐Flp; Su(H)‐Gal4 tub‐Gal80 /UAS‐EGFR ; FRT82BGFP/FRT82BRFP.

Figure EV3. (A–A‴) Su(H)‐Gal4/UAS‐GF P; esg‐lacZ. (B) E(spl)mb‐CD2; Su(H)‐Gal4‐UAS‐CD8:GFP. (C–D′) Su(H)‐Gal4‐UAS‐CD8:GFP.

Figure EV4. Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP/UAS‐EGFR .

Figure EV5. Su(H)‐Gal4 tub‐Gal80 UAS‐CD8:GFP; UAS‐wts‐RNAi.

Author contributions

Aiguo Tian: Conceptualization; data curation; formal analysis; supervision; funding acquisition; investigation; methodology; writing – original draft; project administration; writing – review and editing. Virginia Morejon: Methodology; writing – review and editing. Sarah Kohoutek: Methodology; writing – review and editing. Yi‐Chun Huang: Methodology. Wu‐Min Deng: writing – review and editing. Jin Jiang: writing – review and editing.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Expanded View Figures PDF

Click here for additional data file.

PDF+

Click here for additional data file.