Ret receptor tyrosine kinase sustains proliferation and tissue maturation in intestinal epithelia

Abstract

Expression of the Ret receptor tyrosine kinase is a defining feature of enteric neurons. Its importance is underscored by the effects of its mutation in Hirschsprung disease, leading to absence of gut innervation and severe gastrointestinal symptoms. We report a new and physiologically significant site of Ret expression in the intestine: the intestinal epithelium. Experiments in Drosophila indicate that Ret is expressed both by enteric neurons and adult intestinal epithelial progenitors, which require Ret to sustain their proliferation. Mechanistically, Ret is engaged in a positive feedback loop with Wnt/Wingless signalling, modulated by Src and Fak kinases. We find that Ret is also expressed by the developing intestinal epithelium of mice, where its expression is maintained into the adult stage in a subset of enteroendocrine/enterochromaffin cells. Mouse organoid experiments point to an intrinsic role for Ret in promoting epithelial maturation and regulating Wnt signalling. Our findings reveal evolutionary conservation of the positive Ret/Wnt signalling feedback in both developmental and homeostatic contexts. They also suggest an epithelial contribution to Ret loss-of-function disorders such as Hirschsprung disease.

Keywords: Drosophila; Ret; enteroendocrine; intestine; stem cell.

Figure 1. Ret is expressed in the Drosophila adult midgut

1. Cartoon summarising different cell types in the adult midgut and the immunohistochemical markers used to identify them. Ret‐expressing cells are highlighted in green and include enteric neurons (the nuclei of which are embryonic lethal abnormal vision (Elav)‐positive) and two types of adult intestinal progenitors: stem cells (ISCs, escargot (esg)‐positive, Suppressor of Hairless (Su(H))‐negative) and enteroblasts (EBs, esg‐ and Su(H)‐positive). Ret is absent from visceral muscle (Phalloidin‐positive) and from most of the two other epithelial cell types: enteroendocrine cells (EEs, Prospero (Pros)‐positive) and enterocytes (ECs, labelled as Armadillo (Arm)‐positive cells with large endoreplicating nuclei, visualised with DAPI).

2. Expression of Ret in axons [labelled with a Ret antibody in red and a Ret‐Gal4‐driven membrane‐tagged (CD8) GFP reporter Flybow 1.1 (Hadjieconomou et al, 2011)] innervating the anterior midgut.

3. Ret‐Gal4‐driven expression of a nuclear GFP reporter [Stinger GFP (Barolo et al, 2000)] indicates that some of the anterior midgut innervation emanates from neuronal cell bodies (co‐labelled with anti‐Elav) located in one of the enteric ganglia: the hypocerebral ganglion.

4. Expression of Ret protein (in red) and the same reporter as in (B) in axons of hindgut‐innervating neurons.

5. All Ret‐positive epithelial cells co‐express the ISC/EB marker esg‐Gal4.

6. Co‐staining of the Ret‐Gal4 reporter with a Su(H)‐LacZ reporter indicates that the doublets of small Ret‐positive cells contain one Su(H)‐negative ISC and one Su(H)‐positive EB.

7. Co‐staining of the Ret‐Gal4 reporter with the cell membrane marker Arm and the EE nuclear marker Pros indicates that neither EEs (Arm, Pros⁺) nor ECs (Arm⁺ cells with large DAPI nuclei) express Ret, although very low levels of Ret can be detected in a few ECs (data not shown).

Data information: In panels (E–G), DAPI is used to visualise all nuclei. For full genotypes, see the Appendix.

Figure 2. Ret levels modulate adult ISC proliferation

1. Representative images (left) and quantifications (right) of the number of intestinal progenitor cells in control midguts or midguts in which Ret has been downregulated from adult ISCs/EBs [achieved by esg‐Gal4, tub‐Gal80 ^TS‐driven Ret‐RNAi, enhanced by UAS‐Dicer2 (Dcr2) co‐expression (Dietzl et al, 2007)] for 4, 10 or 20 days.

2. MARCM clone size quantifications (graph) and representative images (clones labelled in green with GFP) reveal that clones lacking Ret (Ret ^KO) or expressing Ret‐RNAi are smaller than control clones 10 days after clone induction.

3. Quantifications of mitoses (pH3‐positive cells, graph) and visualisation of intestinal progenitors (using esg‐driven GFP, image panels) in midguts of flies with the same genotypes as in (A). The regenerative response triggered by damage‐inducing DSS in control flies is reduced following Ret downregulation from ISC/EBs.

4. pH3 quantifications of DSS‐damaged midguts of wild‐type control, Ret heterozygous (Ret ^KO/+) and Ret mutant (Ret ^KO/KO) flies.

5. Representative images of the number of intestinal progenitors (left) and pH3 quantifications (right) in control midguts or midguts in which Ret has been over‐expressed from adult ISCs/EBs (achieved by esg‐Gal4, tub‐Gal80 ^TS‐driven Ret misexpression) for 10 days. In both image panels, intestinal progenitors (ISC/EBs) are labelled with esg‐Gal4‐driven GFP.

6. Quantifications of mitoses (pH3‐positive cells, graph) and visualisation of intestinal progenitors (using esg‐driven GFP, image panels) in midguts of flies with the same genotypes as in (A), aged for 20 days.

7. Midgut pH3 quantifications of wild‐type control, Ret heterozygous (Ret ^KO/+) and Ret mutant (Ret ^KO/KO) flies aged for 20 days.

8. MARCM clone size quantifications (graph) and representative images (clones labelled in green with GFP) reveal that clones over‐expressing Ret (UAS‐Ret) are larger than control clones 8 days after clone induction.

Data information: In all image panels, DAPI is used as a generic nuclear marker, and intestinal progenitors (ISC/EBs) are labelled with esg‐Gal4‐driven GFP. For full genotypes, see the Appendix. Values are presented as average ± standard error of the mean (SEM). P‐values from Mann–Whitney–Wilcoxon test (ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001).

Figure EV1. Cell fate and Ret expression analyses in Ret mutants/knockdowns

1. Like cells of control clones, cells of MARCM clones expressing a Ret‐RNAi transgene (Ret ^RNAi MARCM) or entirely lacking Ret expression (Ret ^KO MARCM) are positive for a Su(H)‐LacZ (an EB marker), Pros (an EE marker) and Pdm1 (an EC marker), indicating that loss of Ret does not compromise the ability of intestinal progenitors to differentiate.

2. Ret immunostainings of adult midguts indicate that adult‐specific downregulation of Ret in intestinal progenitors (achieved by esg‐Gal4, tub‐Gal80TS expression of a Ret‐RNAi transgene) effectively downregulates Ret protein in these cells (labelled with esg‐Gal4‐driven GFP).

3. Ret immunostaining is also absent from the adult intestinal progenitors of Ret ^KO mutants [labelled with anti‐horseradish peroxidase (HRP)].

4. Quantifications of different midgut epithelial cell types (based on the markers used in previous panels) in 4‐day‐old control and Ret ^KO mutants. No obvious differences are apparent. Values are presented as average ± standard error of the mean (SEM). See Materials and Methods for quantification details. ns, P > 0.05 (Mann–Whitney–Wilcoxon test).

Data information: In all image panels, arrowhead points at cells positive for the relevant marker inside a clone. For full genotypes, see the Appendix.

Figure 3. A positive, autoregulatory Ret/Wg feedback loop sustains intestinal stem cell proliferation

1. pH3 quantifications indicate that upd1, sspitz, wg or sgg ^DN expression in adult intestinal progenitors (achieved using esg‐Gal4, tub‐Gal80 ^TS) all promote their proliferation. Simultaneous downregulation of Ret (by co‐expression of one of the above genes together with Ret‐RNAi and UAS‐Dcr2) significantly reduces the wg‐ or sgg ^DN‐triggered proliferation increase, but not that triggered by other mitogens. Image panels (right) show representative images of the number of intestinal progenitor cells (labelled with esg‐Gal4‐driven GFP) in midguts exposed to ectopic wg and sgg ^DN, with and without co‐downregulation of Ret.

2. Representative images of Wg ligand levels (left, labelled using an anti‐Wg antibody) and Wg‐positive area quantifications (right) in control midguts or midguts in which Ret has been over‐expressed from adult ISCs/EBs (achieved by esg‐Gal4, tub‐Gal80 ^TS‐driven expression of a UAS‐Ret transgene) for 10 days. In both image panels, intestinal progenitors (ISC/EBs) are labelled with esg‐Gal4‐driven GFP.

3. pH3 quantifications indicate that the proliferation increase resulting from Ret over‐expression in intestinal progenitors (achieved by esg‐Gal4, tub‐Gal80 ^TS‐driven expression of a UAS‐Ret transgene) is significantly reduced by co‐expression of dominant‐negative versions of Dsh or Pan (dsh ^DN and pan ^DN, respectively). Image panels (right) show representative images of the number of intestinal progenitor cells (labelled with esg‐Gal4‐driven GFP) in midguts in which Ret has been over‐expressed in these cells, and its reduction when Ret is over‐expressed together with dsh ^DN or pan ^DN.

4. wg transcript levels relative to alphaTub84B transcript levels in dissected midguts of control flies or flies expressing wg, Ret or Ret‐RNAi from adult intestinal progenitors. Transcript levels were assessed 10 days after Gal80 transgene induction.

5. Images to the left show lack of GFP expression from wg ^KO ‐Gal4 (abbreviated as wg^TS>) in the midgut epithelium, both in control homeostatic conditions (top) and upon expression of Ret in adult intestinal progenitors (bottom panel, 10 days after Gal80 transgene induction). The right graph shows a quantification of the number of pH3‐positive cells in the posterior midgut of flies following 10 days of adult‐restricted expression of Ret from wg ^KO ‐Gal4 (abbreviated as wg^TS>) relative to control flies. The number of pH3‐positive cells is not significantly different between the two groups of flies.

Data information: In all image panels, DAPI is used as generic nuclear marker. For full genotypes, see the Appendix. Values are presented as average ± standard error of the mean (SEM). P‐values from Mann–Whitney–Wilcoxon test (ns, P > 0.05; **P < 0.01; ***P < 0.001).

Figure 4. Ret‐induced stem cell proliferation requires Src42A and is blocked by Fak

1. (Left) Representative images of the number of intestinal progenitors (labelled with esg‐Gal4‐driven GFP) and the efficiency of Ret misexpression (assessed by Ret (top) or V5 (bottom) immunostaining in red) in midguts in which Ret (control, top) or a V5‐tagged, kinase‐dead Ret (RetK805M, bottom) has been over‐expressed from adult intestinal progenitors using esg‐Gal4, tub‐Gal80 ^TS. (Right) pH3 quantifications of the proliferation increase resulting from Ret over‐expression in adult intestinal progenitors (control experiment), and its absence following over‐expression of the kinase‐dead Ret identical conditions.

2. Induction of pSrc (visualised using a pSrc antibody in red) resulting from esg‐Gal4, tub‐Gal80 ^TS‐ driven over‐expression of Ret in adult intestinal progenitors (labelled with esg‐Gal4‐driven GFP).

3. (Left) Representative images of Wg ligand levels (labelled using an anti‐Wg antibody in red) and the number of intestinal progenitors (labelled with esg‐Gal4‐driven GFP) in midguts in which Ret has been over‐expressed from adult intestinal progenitors using esg‐Gal4, tub‐Gal80 ^TS, alone or together with a Src42A‐RNAi transgene. (Right) pH3 quantifications of the proliferation increase resulting from Ret over‐expression in adult intestinal progenitors, and its suppression by Src42A, but not Src64B, downregulation using RNAi transgenes.

4. Representative images of Wg ligand levels (visualised using an anti‐Wg antibody) in control midguts or midguts in which a constitutively active form of Src42A has been over‐expressed from adult ISCs/EBs (achieved by esg‐Gal4, tub‐Gal80 ^TS‐driven Src42A ^CA misexpression) for 10 days.

5. pH3 quantifications of ISC proliferation in midguts in which a constitutively active form of Src42A has been over‐expressed from adult ISCs/EBs using esg‐Gal4, tub‐Gal80 ^TS for 10 days, alone or together with a Ret‐RNAi transgene. Ret downregulation significantly reduces the proliferation increase resulting from Src42A ^CA expression.

6. Induction of pFak (visualised using a pFak antibody in red) resulting from esg‐Gal4, tub‐Gal80 ^TS‐driven over‐expression of Ret in adult intestinal progenitors (labelled with esg‐Gal4‐driven GFP).

7. (Left) Representative images of the number of intestinal progenitors (labelled with esg‐Gal4‐driven GFP) in midguts in which Ret has been over‐expressed from adult intestinal progenitors using esg‐Gal4, tub‐Gal80 ^TS, alone or together with Fak. (Right) pH3 quantifications of the proliferation increase resulting from Ret over‐expression in adult intestinal progenitors, and its suppression by Fak co‐expression.

8. Representative images of Src phosphorylation (labelled using an anti‐pSrc antibody in red) and the number of intestinal progenitors (labelled with esg‐Gal4‐driven GFP) in midguts in which Ret has been over‐expressed from adult intestinal progenitors using esg‐Gal4, tub‐Gal80 ^TS, alone or together with Fak.

Data information: In all image panels, DAPI is used as generic nuclear marker. For full genotypes, see the Appendix. Values are presented as average ± standard error of the mean (SEM). P‐values from Mann–Whitney–Wilcoxon test (ns, P > 0.05; ***P < 0.001).

Figure 5. Ret expression in the mouse intestinal epithelium

1. Analysis of Ret transcript expression in adult small intestine by RT–qPCR. Ret transcript is detected in RNA samples from adult duodenum, jejunum and ileum.

2. Ret transcript is detected in RNA samples prepared from E16.5 FEnS cultures and organoid cultures derived from adult small intestine.

3. In situ hybridisation of Ret in P1 cross sections of small intestine. The higher magnification image to the right shows Ret transcript expression in scattered cells of a villus and intervillus.

4. Adult ileum sections co‐stained with Ret (green) and E‐cadherin (E‐cad, in blue). In addition to the abundant neuronal fibres at the bottom of the image, Ret‐positive cells can be observed amongst the epithelial cells, positive for E‐cad (arrows).

5. Same section labelled with antibodies against Ret (green) and chromogranin‐A (CgA, in red). Ret‐positive cells are CgA‐positive, but not all CgA‐positive cells are Ret‐positive.

6. Organoid cultures co‐stained with Ret (green) and E‐cad (blue). Ret‐ and E‐cad‐positive cells can be observed (arrows).

7. Single confocal section of organoid shown in (F) labelled with antibodies against Ret (green) and CgA (red). The Ret‐ and E‐cad‐positive cells are also CgA‐positive.

Data information: Values are presented as averages, and each dot corresponds to an independent sample. In both (A and B), Ret transcript levels were normalised relative to Actb levels.

Figure EV2. Validation of Ret antibody and co‐stainings with epithelial cell type‐specific markers

1. Adult small intestinal tissue (ileum) stained with an anti‐Ret antibody. Both neuronal fibres (at the bottom of the image) and scattered epithelial cells (arrows) are labelled.

2. Neuronal and epithelial signals are absent from adult small intestinal tissue processed in parallel and subject to the same protocol as in for (A) except for incubation with the primary antibody. Only background, non‐epithelial staining remains.

3. Adult ileum section labelled with antibodies against Ret (green) and the Paneth cell marker lysozyme (Lyz, in red). Ret‐positive cells (arrow) are not Lyz‐positive.

4. Adult ileum section labelled with antibodies against Ret (green) and the goblet cell marker mucin 2 (Muc2, red). Ret‐positive cells (arrow) are not mucin 2‐positive.

Data information: In (C and D), bottom panels are higher magnification images of the image regions boxed in the top panels.

Figure EV3. Ret isoform usage in the intestine

1. In RNAseq samples obtained from adult small intestine, 24% of reads are assigned to the Ret51 isoform (Ensemble ID ENSMUST00000032201). The rest of reads (76%) map to two different transcript annotations which give rise to the same Ret9 isoform but differ slightly in their 3′UTR. Ret9 corresponds to the recently updated mm10 gene model for ENSMUST00000088790 (ENSMUST00000088790_mm10Update) whereas Ret9 variant corresponds to the previous mm9 gene model for ENSMUST00000088790.

2. In RNAseq data obtained from epithelial organoids derived from adult small intestine, a similar isoform ratio is observed, but only one of the two Ret9 transcripts (Ret9, not Ret9 variant) is observed. See Materials and Methods for further details.

Figure 6. Ret control of epithelial maturation in mouse

1. Ret transcript levels relative to Actb transcript levels in RNA prepared from epithelial cultures derived from neonatal small intestinal tissue of Ret51 mice or their control littermates.

2. Representative images of FEnS/organoid cultures derived from neonatal Ret51 mice and their control littermates.

3. Quantifications of FEnS/mini‐gut number in organoid cultures at passage 2. A higher proportion of FEnS is apparent in organoid cultures derived from Ret51 mice.

4. Branching quantifications of organoids derived from Ret51 mice and control littermates (see Materials and Methods for details). Organoids derived from Ret51 mice are significantly less branched.

5. Axin2 transcript levels relative to Gapdh transcript levels in RNA prepared from epithelial cultures derived from neonatal small intestinal tissue of Ret51 mice or their control littermates.

6. Representative images of FEnS cultures derived from wild‐type E16.5 mice, grown in control medium (ENR, top) or control medium supplemented with GDNF (ENR + GDNF, bottom).

7. Branching quantifications of the GDNF supplementation experiments in (F).

8. Axin2 transcript levels relative to Gapdh transcript levels in RNA prepared from FEnS epithelial cultures grown in control medium (ENR) or control medium supplemented with GDNF (ENR + GDNF).

9. Representative images of lysozyme expression in small intestinal tissue. Lysozyme expression in intervilli appears to be less abundant in Ret51 mice. Panels to the right are close‐ups of lysozyme‐positive intervilli belonging to the left panels. The right graph shows quantification of the numbers of lysozyme‐positive cells in intervillar regions. Ret51 mice have significantly fewer lysozyme‐positive cells.

Data information: Values are presented as average ± standard error of the mean (SEM), and each dot corresponds to an independent biological replicate (A, E, H) or an independent sample (I). P‐values from Mann–Whitney–Wilcoxon test (*P < 0.05; **P < 0.01; ***P < 0.001).

Figure EV4. Assessment of cell cycle state and viability of control and Ret51 cells derived from epithelial cultures

1. Representative cell cycle histograms of cells dissociated from control and Ret51 FEnS/organoid cultures.

2. Quantifications of the percentage of cells in G0/G1 relative to S/G2/M in both genotypes. No obvious differences are apparent.

3. Quantifications of the percentage of dead cells or cells undergoing apoptosis in both genotypes. No obvious differences are apparent. See Materials and Methods for quantification details.

4. Transcript levels of epithelial cell differentiation markers relative to Gapdh transcript levels in RNA prepared from epithelial cultures derived from neonatal small intestinal tissue of Ret51 mice or their control littermates, quantified in parallel to the Axin2 transcript in Fig 6H.

Data information: Values are presented as averages, and each dot corresponds to an independent biological replicate.