Spen limits intestinal stem cell self-renewal

Abstract

Precise regulation of stem cell self-renewal and differentiation properties is essential for tissue homeostasis. Using the adult Drosophila intestine to study molecular mechanisms controlling stem cell properties, we identify the gene split-ends (spen) in a genetic screen as a novel regulator of intestinal stem cell fate (ISC). Spen family genes encode conserved RNA recognition motif-containing proteins that are reported to have roles in RNA splicing and transcriptional regulation. We demonstrate that spen acts at multiple points in the ISC lineage with an ISC-intrinsic function in controlling early commitment events of the stem cells and functions in terminally differentiated cells to further limit the proliferation of ISCs. Using two-color cell sorting of stem cells and their daughters, we characterize spen-dependent changes in RNA abundance and exon usage and find potential key regulators downstream of spen. Our work identifies spen as an important regulator of adult stem cells in the Drosophila intestine, provides new insight to Spen-family protein functions, and may also shed light on Spen's mode of action in other developmental contexts.

Fig 1. spen inactivation leads to an accumulation of intestinal stem cell-like cells in the adult midgut.

(A) Intestinal stem cells (ISCs) produce post-mitotic progenitor cells, called enteroblasts (EBs) that directly terminally differentiate. Notch activated EBs (EB^EC) produce enterocytes (ECs) whereas enteroendocrine cells (EEs) are derived from ISCs in a Notch-independent manner via an EE progenitor (EEP) intermediate, which is thought to divide once. (B) spen gene encodes a protein characterized by 3 RNA recognition motifs (RRM) in the N-terminus (RED), and a SPOC (Spen Paralog Ortholog C-terminus) domain in the C-terminus (GREEN), both included in the different spen isoforms. The three mutant alleles spen^9E34, spen^9C95 and spen⁵ have premature stop codons, which are frequently targeted by nonsense-mediated decay. The spen^9E34 allele has a C>T substitution (2L:185541). The spen^9C95 allele has a C>T substitution (2L:190065). The spen⁵ allele has nucleotide deletion (2L:192716–192769). (C) wild-type (WT), (D) spen⁵, (E) spen^9E34 and (F) spen^9C95 mutant clones 5d AHS marked by GFP (GREEN), Dl, (ISCs, RED), DAPI (BLUE). Insets show RED channel only. (G) A transgenic BAC construct containing the spen wild-type genomic region suppressed spen⁵ phenotypes in 5 days MARCM clones. (H) spen^RNAi (VDRC#KK-108828) in MARCM clones (GFP, GREEN; DAPI, BLUE), 5 days. (I) Quantification of Dl+ cells per clone, 5 days AHS. (J) wild-type and (K) spen⁵ mutant clones had ISCs (mitotic cells, arrow heads) expressing the Notch receptor (RED in J, K). (L, L’) wild-type clones or (M, M’) spen⁵ mutant clones contained ISCs co-labeling with Dl (RED) and Sanpodo (Spdo, GREEN, L’, M’). A single focal plane is shown. An additional 2X zoom is shown in L’, M’. (N-N”) spen^7RNAi expression in esg+ cells. A single focal plane is shown. (Delta, RED; cytoplasmic GFP, GREEN; DAPI, BLUE), 10 days. (O, O’) wild-type clones or (P, P’) spen⁵ mutant clones contained ISCs co-labeling with Dl (RED) and Rab5::YFP. (GREEN, I’, J’). A single focal plane is shown. (Q) Quantification of the % of wild-type and spen⁵ mutant clones with mitotic cells (PH3), 5 days AHS. Fisher’s test. (R) Quantification of the number of PH3+ Dl+ cells per Dl+ cells. (S) Quantification of clone size in wild-type and spen⁵ mutant clones. (T, T’) Enterocyte (EC) and enteroendocrine (EE, Pros, nuclear RED) cells were produced in wild-type and (U, U’) spen⁵ mutant clones, clones marked by GFP (GREEN); EE (Pros, nuclear red); ISC (Delta, membrane RED); DAPI (BLUE). (V) Quantifications of EE and ECs per clone. When not specified, a non-parametric Mann-Whitney test was performed. Error bars represent the Standard Error of the Mean (sem). p<0.05, *. p<0.01, **. p<0.001, ***. p<0.0001, ****. Scale bar:10μm.

Fig 2. Spen has stem cell autonomous and non-autonomous functions to control ISC numbers and proliferation state.

(A-D) Quantification of E-N’ (A) the number of mitotic cell (PH3+ cells) or (B) relative increase in proportion per posterior midgut (PMG) in the indicated genotypes. (C) the density of Dl+ ISCs or (D) relative increase. (E, E’, G, G’, I, I’, K, K’, M, M’) wild-type UAS-GFP control and (F,F’, H, H’, J, J’,L, L’, N, N’) UAS-spen^RNAi expressed for 10 days at 29°C. An increased number of PH3+ cells and Dl+ ISCs was detected upon expression of UAS-spen^RNAi in ISCs and EBs (E- F’), ISCs only (G-H’), and EBs only (I- J’; using esg^ts, esg^ts combined with NRE-GAL80 –to block expression in Notch active EB cells, or NRE^ts, respectively). (K-N’) An increase in PH3+ cells was also found upon expression of UAS-spen^RNAi in Enteroendocrine cells (pros^voila.ts) or in Enterocytes using myo1A^ts. GFP in GREEN marks cell type expression, Delta+ ISCs, RED, mitotic cells (PH3+, Gray), DNA (DAPI, BLUE). Scale bar: 20μm.

Fig 3. spen controls an early stem cell commitment event.

(A, A’) Wild-type and (B, B’) spen^9C95 5 day AHS MARCM clones showed activation of the Notch pathway, as detected by NRE-lacZ (Notch Responsive Element, also called Su(H)-GBE-LacZ) [57], ßGal, Green) (GFP, GREEN; DAPI, BLUE). (C) Quantification of cell type distribution, ISCs (Dl+ cells) in RED; EBs (EC precursors, NRE-LacZ+, in GREEN), and combined other cells representing mostly EEs and ECs, in BLUE. While the production of EB (NRE+) cells per clone was not affected upon spen inactivation, the proportion of Dl+ ISC-like cells significantly increased spen^9C95 clones. (D) Upon spen inactivation, ISC-like cells were more produced compare to EB (NRE+) cells. (E) Quantification of the number of GFP+ cells per clone or (F) Dl+ cells per clone for G-J (below). Single cell clones have been included in these quantifications (E-F). (G) Control wild-type clones produce multi-cell clones, whereas (H) the expression of an activated form of the Notch receptor (UAS-N^act) promoted differentiation. (I) Similarly, spen mutant ISCs produced multi-cell clones whereas (J) the expression of UAS-N^act drove differentiation and resulted in single-cell clones. (Dl+, RED; clones marked by GFP in GREEN; DAPI in BLUE (G-J). (K) Scheme of experimental set-up: Mitotic clones were first generated, then UAS-N^Act was expressed after 10 days in these clones by releasing Gal80ts inhibition at 29°C during 4 days. (L) spen⁵; UAS-N^Act genotype at 18°C where Gal80ts was active therefore UAS-N^Act was not induced. (M-P) At 29°C, Gal80ts became inactive in (M) wild-type clones, (N) those expressing UAS-N^Act, (O) spen⁵ clones, and (P) spen⁵; UAS-N^Act. Delta+ in RED; clones GFP in GREEN; DAPI in BLUE. (Q) Neither spen⁵ nor (R) spen^9C94 mutant clones showed defects in sensory organ specification, lineage decisions, or Dl protein accumulation at 18h after pupation (Q) or 24h after pupal formation (APF) (R). (S) Germline and (T) follicular cell clones of spen^9C94 (lack of GFP, GREEN) did not present visible phenotypes akin to defects in Notch signaling. (U, U’) Control or (V, V’) UAS-Delta overexpressing clones. Spdo+ in RED; Delta in BLUE; clone GFP in GREEN. (W) Quantification of number of cells per clone U-V’. (X) Quantification of number of Spdo+ cells per clone. For all quantifications a non-parametric Mann-Whitney Two-Way ANOVA test was performed. Error bars represent the Standard Error of the Mean (sem). p<0.01, **. p<0.001, ***. p<0.0001, ****. Scale bars: 10μm, except in T where is 20μm.

Fig 4. Spen mutants are less sensitive to lowered EGFR signaling, but are suppressed by lowering Akt/Insulin signaling.

(A-D) Genetic interaction between spen and EGF receptor (EGFR). (A) Wild-type clones, (B), EGFR^DN clones, (C), spen⁵ clones, and (D) spen⁵;EGFR^DN clones, 10d after heat shock (AHS). (Delta+, RED; GFP, GREEN; DAPI, BLUE,) (E) Quantification of: cells per clone, (F) Dl+ cells per clone, and (G) Dl cell proportion per clone in A-D. (H-K) Genetic interaction between spen and AKT (H) Wild-type clones, (I), AKT^RNAi clones, (J), spen⁵ clones, and (K) spen⁵;AKT^RNAi clones, 10d after heat shock (AHS). (Delta+, RED; GFP, GREEN; DAPI, BLUE,) (L) Quantification of: cells per clone, (M) Dl+ cells per clone, and (N) Dl cell proportion per clone in A-D. p<0.01, **. p<0.001, ***. p<0.0001, ****. Mann-Whitney Two-Way ANOVA test. One cell clones were excluded in this quantification as only stem cell clones were analyzed. Error bars represent the Standard Error of the Mean (sem). Scale bar: 25μm.

Fig 5. spen-dependent transcript identification by RNAseq analysis from two-color FACS sorted ISC and EB.

(A) ISCs and EBs were sorted using a two-color FACS sorting approach following 2 day expression using esgGAL4, Gal80ts to drive UAS-RFP expression in both cells, and NRE-Venus (GREEN) to mark only EB cells (EC precursors). Wild-type controls (top panels) showed vesicular Dl staining in esg+ NRE- cells, spen^RNAi (bottom panels) showed membrane accumulation of Dl in esg+ NRE- cells. Scale bar: 10μm. (B) Flow cytometry analysis of ISC and EB sorting. The scatterplot revealed two distinct populations: RFP positive (ISCs) and RFP and GFP positive EBs cells. (C) Principal component analysis (PCA) based on the 300 top genes with the highest variance clustered our samples in to four distinct groups. (D) Differential gene expression comparison between WT ISCs and WT EBs revealed 366 differentially expressed genes, FDR of 0.05. Genes significantly enriched in ISC (RED), genes significantly enriched in EB (GREEN), known ISC markers (BLUE), cell-cycle related-genes (YELLOW). (E) Differential gene expression comparison between WT vs. spen^RNAi expressing ISCs or (F) EBs; FDR of 0.05. (G-I) RT-qPCR validation of spen-dependent genes from whole midguts that expressed ubiquitinously spen^RNAi during 2days (tubGS), differentially expressed in RNAseq data from (H) ISCs and (I) in EBs. (J) Comparison of spen-dependent differentially expressed (DE) genes between ISCs and EBs. (K) spen-dependent differentially used exons. Arrowheads represent exons that are out of the axis range. (L, M) Overlap between differentially expressed (DE) genes and genes with alternative exon usage (AEU) in absence of spen in ISCs, or (L) in EBs (M).

Fig 6. Examples of types of altered exons upon spen knockdown.

Visualization of the relative exon usage, in a subset of example transcripts (DEXseq plot): (A) The relative exon usage of spenito was affected by spen^RNAi in EBs (B) Ribosomal protein encoding transcript RpS21 and (C) RpL19 had exon usage alteration upon spen^RNAi in both ISCs and EBs. (D) RpS11, (E) mRpL16, (F) RpL3 and (G) RpS7 showed differential exon usage in spen knock-down EBs. Additional examples of splicing could be found upon spen^RNAi expression in ISCs: within (H) RpL17 and (I) Ubi-p63E transcripts. On top, each line corresponds the relative exon usage of a single “counting bin” generated as defined in the Materiels & Methods. The relative exon usage is shown on top for each sample groups independently: ISC-WT (BLUE), ISC-SPEN (RED), EB-WT (GREEN), EB-SPEN (YELLOW). Significant differentially used exons (FDR of 0.05) are represented in PINK and non-expressed exons in both samples are represented in WHITE. Below, are shown the different annotated isoforms from each transcript, with the coding sequence (CDS) in LIGHT GRAY and the untranslated regions (UTRs) in DARK GRAY.

Fig 7. Model. Autonomous and non-autonomous roles of spen in the control of intestinal stem cells.

In the intestinal stem cell (ISC), spen acts autonomously to limit their proliferation and stem cell self-renewal. Its inhibition leads to an accumulation of the Delta protein at the membrane. This is likely due to an effect on Dl trafficking, which may be direct or indirect. spen also acts in a non-autonomous manner in the Enteroblast (EB) to limit ISC numbers. Furthermore, spen inactivation in EBs, Enteroendocrine cells (EE) and Enterocytes (EC) leads to enhanced ISC proliferation. Further studies will be required to better understand Spen’s molecular functions.