Search
Close this search box.

Cell-fate conversion of intestinal cells in adult Drosophila midgut by depleting a single transcription factor – Nature Communications

Depleting Ttk in ECs initiates EC-to-EE transdifferentiation

To investigate the potential role of Ttk in ECs, we used an EC-specific GAL4 driver Myo1A-GAL4ts to express ttk-RNAi (VDRC#GD4414 or BDSC#TRiP.JF02088) and deplete Ttk in ECs for seven days. Strikingly, we observed that depleting Ttk caused approximately 19.5 ± 2.1% (n = 17, totally 1419 ECs counted) of differentiated ECs within the posterior midgut to express the EE signature protein Pros (Fig. 1b–d; Supplementary Fig. 1a, b). The acquired EE-like identity in Ttk-depleted ECs was accompanied by the loss of EC identity, which was reflected by reductions in EC-specific markers including Myo1A-GAL4, UAS-GFP (or Myo1A>GFP hereafter, for simplicity) (Fig. 1b–d), the brush border marker A142-GFP (Fig. 1e, f), and the EC-specific TF Pdm1 (Fig. 1g, h). Collectively, these data suggest that the acquisition of EE identity in Ttk-depleted EC is accompanied by the loss of EC identity.

To determine if the transdifferentiation of ECs to EEs occurred directly through a switch in transcriptional programs or indirectly through dedifferentiation into a progenitor cell state before adopting a different cell fate, we analyzed the expression dynamics of Pdm1 and Pros following Ttk depletion in ECs. We observed that both the expression levels of the EC-specific TF Pdm1 and the EE-specific TF Pros exhibited an inverse correlation in each individual Ttk-depleted EC. Occasionally some ECs expressed both Pdm1 and Pros at low levels (Fig. 1i, arrowheads), and in general, the appearance of Pros expression seemed to occur immediately after the loss Pdm1 (Fig. 1g–i). Furthermore, we conducted staining for the progenitor cell marker esg-lacZ during the reprogramming process. However, no discernible signal was observed in ECs depleted of Ttk (Supplementary Fig. 2). These observations indicate that the transdifferentiation is direct and not through a progenitor cell state.

To address the concern that late EBs might already exhibit Myo1A-GAL4 expression, potentially leading to a shift in the differentiation pathway instead of transdifferentiation in differentiated ECs, we conducted co-staining experiments using the EB-specific marker NRE-lacZ along with Myo1A>GFP. Interestingly, we observed co-localization of lacZ in several GFP+ cells with smaller nuclear size (Supplementary Fig. 3a, indicated by arrowheads), suggesting the presence of Myo1A-Gal4 expression in EBs. We therefore employed Mex1-GAL4, an alternative EC driver23, to determine if Ttk depletion still induced transdifferentiation. Mex1-Gal4 displayed more specificity in ECs, as no colocalization was observed with the EB marker NRE-lacZ (Supplementary Fig. 3b). Importantly, knocking down Ttk using Mex1-Gal4 similarly activated Pros in differentiated cells, leading to the loss of Mex1>GFP expression in these transformed cells (Supplementary Fig. 4a, b, indicated by arrowheads).

To further exclude the possibility of an altered differentiation pathway during the EB stage leading to the emergence of Pros+ polyploid cells (cells with a nuclear size greater than 20 μm2), we employed NRE-GAL4 to deplete Ttk specifically in EBs. Remarkably, this approach resulted in the transdifferentiation of EBs into entirely diploid Pros+ cells, with no presence of polyploid Pros+ cells observed (Supplementary Fig. 3d, e). Taken together, these findings provide compelling evidence supporting the conclusion that Ttk depletion has the capacity to directly induce transdifferentiation of cell fate from ECs to EE-like cells.

The transdifferentiated cells exhibit functional EE characteristics

EEs are responsible for regulating various physiological processes by secreting a range of peptide hormones24. In Drosophila, class I and II EE subtypes secrete Tachykinin (Tk) and Allatostatin C (AstC), respectively21. We found that a small subset of the Pros+ ECs that underwent EC-to-EE transdifferentiation expressed Tk, AstC, and Rab3, suggesting that these cells have acquired the hormone-producing function of EEs (Fig. 2a–f and Supplementary Figs. 1c, d, 4c–f). Interestingly, these transdifferentiated cells exhibit a similar pattern to normal EEs in terms of mutually exclusive expression of Tk and AstC, indicating that the subtype feature of EEs is maintained in the transdifferentiated cells (Supplementary Fig. 5a, b). To assess the stability of the transformed cell population induced by Myo1A>ttk-RNAi, we subjected the flies back to the permissive temperature of 18 °C for a duration of 7 days. Interestingly, during this period, the Tk+ polyploid cells persisted, indicating the presence of a stable transformed cell population rather than a transient activation of Ttk-target gene expression (Supplementary Fig. 5c).

Fig. 2: The transdifferentiated cells exhibit functional EE characteristics.
figure 2

af In normal guts, peptide hormones Tk and AstC are exclusively expressed in diploid EEs (a, d). However, upon Ttk knockdown in ECs, Tk and AstC expression are observed in polyploid cells (b, c, and e, f, indicated by yellow arrowheads). g, h Transmission electron microscopy (TEM) images of normal and ttk-depleted guts. Abundant lipid droplets are present in normal ECs (g, red arrow), while they are diminished in Ttk-depleted ECs, with small secretory granules appearing instead (h, yellow arrow). i Transcriptome profiles of normal ECs, EEs, and Ttk-depletion-induced EE-like cells. The transdifferentiated EE-like cells, induced by Ttk depletion in ECs, exhibit a general shift in their transcriptome towards an EE-like state. j Gene ontology analysis of upregulated (left panel) and downregulated genes (right panel) upon Ttk depletion in ECs. Statistical test of p value: modified Fisher’s exact test (EASE score). Scale bars, 20 μm.

Additionally, our electron microscopy analysis showed that the putative transdifferentiated cells contained numerous secretory granules (Fig. 2h, arrowhead). This characteristic is typically associated with secretory cells, whereas normal ECs are characterized by the presence of lipid droplets (Fig. 2g, arrow). Taken together, these observations suggest that the depletion of Ttk leads to a conversion of ECs into cells that molecularly and functionally resemble EEs.

Genome-wide RNA-seq analysis reveals a complete EC-to-EE transcriptome switch in Tk+ polyploid cells

To determine to what extent transdifferentiation occurs in order for hormone-producing function to be obtained, we sorted TK+ polyploid cells using TKEn-GFP, a GFP reporter driven by the TK gene enhancer21, and performed RNA sequencing analysis.

We found that EC-derived TK+ polyploid cells had virtually identical gene expression patterns to EEs, with highly expressed programs such as neurotransmitter signaling, vesicle secretion, and neuropeptide signaling (Fig. 2g, h left panel, and Supplementary Data 1). Conversely, many EC-specific programs, including cytoskeleton organization, DNA replication, metabolic and energy production, and responses to stress, were downregulated in TK+ ECs (Fig. 2h, right panel). The analysis of a comprehensive collection of gene sets associated with EC and EE functions further validated the observed transcriptome shifts (Supplementary Fig. 1e, f). Based on these observations, we propose that, following Ttk ablation in ECs, a small subset of Pros+ cells undergoes a complete switch in cell identity from ECs to EEs in terms of both gene expression and function, despite the cells still remaining polyploid. We therefore speculate that the acquisition of hormone-producing function of EEs may require a complete transcriptional reprogramming, that is, a genome-wide transcriptome shift from that of EC to EE.

Pros mediates EC-to-EE transdifferentiation

Ttk suppresses the specification of EE fate by repressing achaete-scute genes, whose expression promotes EE fate specification by inducing Pros expression16. We recently showed that Pros acts as a selector factor for EE fate15, and this may explain why transdifferentiation towards EE occurs following Ttk depletion in ECs. We found that overexpression of Ttk in EEs was able to eliminate Pros expression in EEs (Fig. 3a, b), supporting a role for Ttk in suppressing Pros expression. Knockdown of Pros completely abolished Ttk depletion-induced EC-to-EE transdifferentiation, and the expression of EC-specific markers (Myo1A>GFP and Pdm1) failed to be downregulated (Fig. 3c, e–h), with no activation of peptide hormones observed (Fig. 3d). Therefore, Pros is a key mediator for Ttk depletion-induced EC-to-EE transdifferentiation.

Fig. 3: Ttk suppresses pros to prevent EC-to-EE transdifferentiation.
figure 3

a, b Overexpression of Ttk69 in EEs leads to the suppression of Pros expression and loss of peptide hormones. c, d Depleting Pros expression in Ttk-depleted ECs using pros-RNAi prevents the expression of AstC. eh Depleting Pros expression in Ttk-depleted ECs using pros-RNAi abolishes the suppression of Myo1A-GFP (e) and Pdm1 expression (f). Conversely, overexpressing Pros alone is sufficient to suppress Myo1A-GFP (g) and Pdm1 expression (h) in ECs. i A heatmap displays the expression patterns of differentially expressed genes between ECs and EEs in EC>Pros-OE and EC>ttk-RNAi & pros-RNAi cells. A general transcriptome reprogramming towards an EE identity is observed in EC>Pros-OE cells, while co-depleting Pros prevents the expression of EE identity genes induced by Ttk-RNAi. j Another heatmap illustrates the expression of selected EE signature genes in normal EEs, ECs, and EC>Pros-OE cells. An overall activation of EE identity genes, including peptide hormones, secretory factors, and EE subtype-regulating TFs, is observed in ECs overexpressing Pros. Scale bars, 20 μm.

We previously showed that the ectopic expression of Pros in ECs can induce AstC expression in some ECs, indicating that Pros may be sufficient to induce EC-to-EE transdifferentiation15. In fact, overexpression of Pros robustly suppressed the expression of EC marker genes, such as Myo1A>GFP and Pdm1, indicating a cell identity switch (Fig. 3g, h). To analyze the role of Pros in EC-to-EE transdifferentiation systemically, we isolated ECs from Pros overexpressed guts (Myo1Ats>pros), and Ttk & Pros co-depleted guts (Myo1Ats>ttk-RNAi & pros-RNAi), respectively, and performed transcriptome analysis by RNA sequencing. By comparing them to the transcriptome of normal EEs and ECs, we found that overexpression of Pros alone in ECs was able to globally upregulate EE-specfic genes and downregulate a significant portion of EC-specific genes (Fig. 3i, j and Supplementary Data 2), although the expression levels of the up- or down-regulated genes were not comparable to the levels in normal EEs or ECs, likely due to incomplete or non-occurring transdifferentiation in many ECs that were sorted out for the analysis. Importantly, depletion of Pros virtually prevented Ttk depletion-induced transcriptome switch from EC to EE (Fig. 3i). These observations demonstrate that Pros is the key downstream factor that mediates EC-to-EE transdifferentiation.

The ability of Pros to suppress EC-specific markers suggests that this selector protein not only activates EE-specific transcriptional programs but also inhibits EC-specific transcriptional programs. This dual function may underlie its ability to promote EC-to-EE transdifferentiation.

Depletion of pros in EEs causes cell dedifferentiation into intestinal progenitor cells

Given the high degree of cell plasticity observed in differentiated ECs, we investigated whether differentiated EEs possess similar properties. Our recent study showed that depleting Pros in EEs leads to a nearly complete loss of EE-specific transcriptional programs15. Here we revisited the transcriptome data of Pros-depleted EEs and compared it with that of normal EEs. We found that, apart from the 1051 downregulated genes that belong to the EE signature gene sets, approximately 1700 genes were significantly upregulated (Supplementary Data 3). GO analysis of these upregulated genes indicated enrichment in genes involved in mitosis (Fig. 4a). This raised the possibility that these cells had acquired characteristics of intestinal progenitor cells, as EEs and ECs are typically post-mitotic cells. To test this possibility, we examined the expression of the top 250 progenitor cell-specific genes, and the Gene Set Enrichment Analysis (GSEA) plot showed significant upregulation of this gene set in Pros-depleted EEs (Fig. 4b, Supplementary Data 4).

Fig. 4: EE-to-ISC dedifferentiation following the depletion of pros in EEs.
figure 4

a GO analysis of genes upregulated upon Pros depletion in EEs. Statistical test of p value: modified Fisher’s exact test (EASE score). b GSEA plot of the top 250 progenitor identity genes in transcriptome alterations induced by Pros depletion. Statistical test of p value: empirical phenotype-based permutation test. c Heatmap displaying the expression of a set of progenitor signature genes in progenitor cells, control EEs, and Pros-depleted EEs. d, e Pros depletion leads to the expression of esg in prosV1-GAL4 > GFP+ cells outside of the R3 region (yellow arrows in e), while rarely detected in the normal gut (white arrows in d). f, g Sox100B is not detected in normal EEs (f), but is ectopically activated in Pros-depleted EEs (g, yellow arrow). h, i Pros depletion leads to the activation of Dl in prosV1-GAL4 > GFP+ cells (yellow arrows in i), while rarely detected in GFP+ EEs of the normal gut (white arrows in k). j, k Quantification of the percentage of esg+ (j) or Dl+ (k) cells in GFP+ EEs outsize R3. Error bars represent Mean ± SEM, **p < 0.01; ****p < 0.0001 (two-tailed Student’s t test); “n” indicate the number of guts used for quantification; source data are provided as a Source Data file. l, m Pros depletion leads to the detection of mitotic PH3 signal in prosV1-GAL4 > GFP+ cells (yellow arrows in m), in addition to GFP PH3+ cells (white arrows in l and m). nr Lineage tracing of normal and Pros-depleted EEs after tracing for 3 days (no) and 8 days (pr). Knocking down Pros induces Dl expression in many lineage-labeled cells (o, q, yellow arrow). Pdm1+ polyploid cells can be found in the lineage-labeled cells after tracing for 8 days (q, r, green arrow). (s-u) Lineage tracing of normal (s) and Pros-depleted Tk+ EEs(t, u) using Tk-GAL4 driver. After tracing for 3 days, Pros depletion leads to lineage+ small cell clusters, and some of them regained Dl expression (t, yellow arrow). After tracing for 7 days, Pdm1+ polyploid cells could be observed in lineage-labeled cells (u, green arrow). Scale bars, 50 μm.

We further analyzed the transcription levels of the previously-defined top 32 progenitor cell-specific TFs25, including esg, sox100B, fkh, and klu, and additional progenitor cell enriched genes including cell cycle regulators, and cell polarity and cell adhesion-related genes, and found that all of these genes were upregulated to some extent following Pros depletion (Fig. 4c). Immunostaining of the gut confirmed that the expression of esg was activated in about 36.5% ± 5.6% Pros-depleted EEs (Fig. 4d, e, j, and Supplementary Fig. 6a). Similarly, the expression of Sox100B, which is typically present in ISCs and EBs26, was activated in Pros-depleted EEs (Fig. 4f, g). Interestingly, about 30.5% ± 3.7% of the ProsV1>GFP+ cells regained Dl expression following Pros depletion, whereas Dl was rarely expressed in control ProsV1>GFP+ cells (Fig. 4h, i, k). In addition, consistent with the derepression of cell cycle genes, the mitotic marker PH3 was detected in the Pros-depleted ProsV1>GFP+ cells, suggesting that Pros depletion enables these EEs to re-enter mitotic cell cycle (Fig. 4l, m). Collectively, these observations indicate that Pros-depleted EEs can become ISCs by undergoing dedifferentiation.

The dedifferentiated cells can divide and differentiate into ECs

We next performed cell lineage tracing of Pros-depleted EEs using the UAS-flp and flp-out cassette system to test our hypothesis. This system allows for labeling cells of interest and their progeny over time. On day 3 after inducing lineage labeling, we found that in control samples, only Pros+ cells, but not Dl+ were present in the lineage, and it was extremely rare to see cells in the lineage not being Pros+ (Fig. 4n). In contrast, many diploid cells in the lineage of Pros-depleted EEs showed Dl expression, suggesting acquisition of ISC identity. In addition, there were apparently more lineage-labeled cells in the EE>pros-RNAi gut epithelium, indicating cell proliferation may have occurred (Fig. 4o, yellow arrow). On day 7 after inducing lineage labeling, we found that again in control samples, only Pros+ cells, not Dl+ cells, were present in the labeled lineage (Fig. 4p, white arrow). In contrast, lineage cells of Pros-depleted EEs became diversified, with some remaining diploid with or without Dl expression, and others differentiating into ECs as evidenced by polyploid nuclei and Pdm1 expression (Fig. 4q–r, green arrow).

To address the concern that dedifferentiation might occur in early progenitor EEPs rather than in differentiated EEs, we employed an alternative EE-specific Gal4 driver named Tk-GAL4, which is expressed specifically in class II EE subtypes, and performed cell lineage tracing experiments to track the fate of Tk+ EEs after depleting Pros. Interestingly, we observed the emergence of Dl+ diploid cells and Pdm1+ polyploid cells within the Tk lineage cells, while lineage-negative EEs retained their Pros+ cell identity (Fig. 4t–u). These findings provide further evidence that the dedifferentiation process occurs in differentiated EEs rather than in early progenitor EEPs.

These results demonstrate that the depletion of Pros in differentiated EEs triggers robust cell dedifferentiation leading to the acquisition of ISC identity. These findings highlight a dual function of Pros in EEs. One is to act as a selector protein to establish and maintain EE-specific transcriptional programs. Another is to concurrently suppress stem cell-specific transcriptional programs. Indeed, transient overexpression of Pros in ISCs promotes their exit from the cell cycle and differentiation into EEs16. Moreover, Pros is known to have a dual role in the neuroblast-neuron lineage. It directly suppresses the transcription of cell cycle genes and neuroblast-specific genes, thereby inhibiting neuroblast self-renewal. At the same time, it directly promotes the transcription of neuron-related genes, leading to neuron specification27,28. This dual function of Pros may explain the EE dedifferentiation phenotype upon depletion of Pros.

A shared chromatin accessibility profile among EC, EE, esg+ progenitor cells and transdifferentiated cells

The above observations of transdifferentiation or dedifferentiation suggest high cell plasticity for both types of differentiated intestinal cells. In mammalian ISC lineage, secretory progenitor cells can differentiate into enterocytes upon atoh1 depletion, and chromatin accessibility analysis shows that both secretory and absorptive progenitor cells share a genome-wide similarity in their profiles, rendering their transdifferentiation potential29. We wondered whether ECs and EEs in the Drosophila midgut, despite being terminally differentiated cells, shared a similar chromatin accessibility profile.

To explore this, we performed ATAC-seq (Assay for Transposase-Accessible Chromatin using sequencing) analysis of FACS-sorted esg+ progenitor cells, EEs, and ECs. We first compared our ATAC-seq data to previously published ATAC-seq data on ISCs30, as well as Elav+ larval neurons, germline cells and eye imaginal disc cells31,32,33. As shown in a PCA plot, cells from different lineages were all distantly separated (Fig. 5a), which is consistent with the idea that lineage-restricted genes become separately accessible during embryo development from a fertilized egg to specify distinct germ layers and organs34. Interestingly, all intestinal cells, including progenitor cells and terminally differentiated cells, were clustered together in the PCA plot (Fig. 5a). Despite ISCs, ECs and EEs having very different transcriptome profiles, with the Pearson correlation scores ranging from 0.6 to 0.84 (Fig. 5b), their chromatin accessibility profiles showed higher similarity, with the correlation scores ranging from 0.78 to 0.88 (Fig. 5c).

Fig. 5: Broadly similar chromatin accessibility landscape among different intestinal cell types.
figure 5

a Principal Component Analysis (PCA) plot displaying the clustering of ATAC-seq profiles of intestinal epithelium cells, including profiles of ECs, EEs, progenitor cells, as well as pros-depleted EEs and Ttk-depleted ECs generated in this study, with published ATAC-seq profiles of intestinal cells and cells from other organs including ovary, testis, imaginal disc, and brain. b Pearson correlation coefficient analysis of RNA sequencing profiles in progenitor cells (esg+), ECs, and EEs. c Pearson correlation coefficient analysis of chromatin accessibility profiles (ATAC-seq profiles) in progenitor cells (esg+), ECs, and EEs. d Integrated map displaying a heatmap of expression levels of the top 1000 identity genes for progenitor cells, EEs, and ECs (right), along with the corresponding chromatin accessibility profiles of these identity genes in esg+ progenitor cells, EEs, and ECs (left). e Chromatin accessibility profiles (left) and transcription levels (right) of Esg, Pros, and Pdm1, which are representative TFs of progenitor cells, EEs, and ECs, respectively. f Integrated map showing a heatmap of significantly altered genes between normal and pros-depleted EEs (left), along with the chromatin accessibility profiles of these differentially expressed genes in normal and pros-depleted EEs (right). g Integrated map displaying a heatmap of significantly altered genes between normal and Ttk-depleted ECs (left), along with the chromatin accessibility profiles of these differentially expressed genes in normal ECs and Ttk-depleted ECs (right).

To investigate the chromatin loci directly relevant to cell type-specific physiological functions, we selected the top 1000 transcribed genes specific to EC, EE, and progenitor cell and analyzed the chromatin accessibility in these gene loci (Fig. 5d). Again, despite significant differences in transcriptional activity, these gene loci exhibited strikingly similar ATAC-seq profiles across all the cell types analyzed.

We further investigated the chromatin loci of esg, pros, and nub, which are representative TFs for progenitor cells, EEs, and ECs, respectively. esg is specifically expressed in intestinal progenitor cells and is essential in preventing cell differentiation22,35. As previously mentioned, pros is EE-specific and a selector gene for the establishment and maintenance of EE identity15, while nub is specifically expressed in ECs and regulates EC differentiation22. Despite pronounced differences in their transcriptional activities among the different cell types, all three genes exhibited open chromatin status in all three cell types, although the pros locus manifested unique accessible chromatin in EEs (Fig. 5e).

The observed similarity in chromatin accessibility among intestinal lineages may account for the cell fate plasticity observed among different intestinal cell types. It is therefore conceivable that depletion of cell identity maintenance TFs leads to transcriptome shifts without affecting the original chromatin accessibility landscape. To confirm this, we compared the chromatin accessibility profiles of the differentially-expressed genes in normal EEs and Pros-depletion-induced progenitor-like cells, or normal ECs and Ttkdepletion-induced EE-like cells. The Pearson correlation scores revealed a significant positive correlation in the ATAC-seq profiles between cell groups before and after Pros (0.89) or Ttk (0.65) depletion. These data indicate that despite the global changes in the transcriptome, the chromatin accessibility profile remained largely unchanged following the cell fate switches (Fig. 5f, g). This suggests that the underlying chromatin landscape remains relatively stable during the process of cell fate reprogramming.

Age factor as a potential barrier to the EC-to-EE transdifferentiation

The discovery of EC-to-EE transdifferentiation following Ttk depletion in Drosophila provides an in vivo genetic system to understand the molecular mechanisms underlying cell plasticity. As mentioned earlier, about 20% of polyploid cells successfully turns on Pros expression following Ttk depletion, and the percentage of cells successfully turning on peptide hormone expression is further limited, indicating heterogeneity in cell plasticity among ECs.

One potential explanation for the variation in cell plasticity among enterocytes (ECs) is that newly formed ECs exhibit high plasticity, which gradually decreases with age (or with increased maturation or polyploidy). To investigate this hypothesis, we can employ an approach that involves identifying specific markers for young/ newly-formed and aged/pre-existing ECs, and then comparing the efficiency of EC-to-EE transdifferentiation following Ttk depletion in these different age groups. The TF Sox21a has been shown to play a critical role in promoting EC differentiation36,37. It is expressed at low levels in progenitor cells but is highly upregulated in differentiating and early ECs36. Therefore, Sox21a could serve as a marker to distinguish between young and aged ECs.

Given that the turnover rate of the intestinal epithelium is typically slow under normal conditions, we induced epithelial renewal by subjecting the flies to heat shock treatment, so that more differentiating ECs with Sox21a expression could be observed. We then examined Sox21a expression in two known EC-specific GAL4 lines, Myo1A-GAL4 and mex1-GAL4. In the unstressed gut, Myo1A-GAL4 and mex1-GAL4 showed largely similar expression patterns, except that the expression level of mex1>GFP was more variable among ECs (Fig. 6a, b). In the stressed guts, however, many Myo1A-GAL4 > GFP+ cells exhibited high levels of Sox21a expression, whereas only a few Mex1-GAL4 > GFP+ cells displayed high Sox21a expression, indicating Myo1A-GAL4 > GFP, but not Mex1>GFP frequently marks early-differentiating ECs (Fig. 6d, e, arrowheads). Additionally, we screened several enhancer trap GAL4 lines from different sources and identified GMR23G10-GAL4 as another EC-specific driver. Interestingly, in the unstressed gut, GMR23G10-GAL4 was expressed only in a subset of ECs in the posterior midgut, displaying a mosaic expression pattern (Fig. 6c). Co-staining with anti-Sox21a in the stressed gut revealed that GMR23G10-GAL4 was exclusively expressed in Sox21a ECs, and many polyploid cells that still retained low levels of Sox21a, presumably the newly differentiated ECs or young ECs, are negative for GMR23G10 > GFP expression. This observation indicates that GMR23G10-GAL4 marks elder ECs only (Fig. 6f). Based on these observations, we consider Myo1A-GAL4 as a driver for both young and aged ECs, mex1-GAL4 for aged/ pre-existing ECs predominantly, and GMR23G10-GAL4 for aged/ pre-existing ECs specifically.

Fig. 6: Age factor as a potential barrier to the EC-to-EE transdifferentiation.
figure 6

af The expression patterns of myo1A>nGFP, Mex1>nGFP, and GMR23G10>nGFP in normal (ac) and stressed (df) guts. The stressed gut is co-stained with anti-Sox21a (in white). Arrowheads indicate Sox21ahigh (a, b) and Sox21alow cells (c). g Depleting Ttk in ECs using mex1-GAL4 leads to Pros (arrowheads) or Tk (arrows) in a subset of polyploid cells. h Depleting Ttk in ECs using GMR23G10-GAL4 leads to Pros (arrowheads) but not Tk expression in a subset of polyploid cells. i, j Quantification of the percentage of Pros+ (i) or the number of Tk+ (j) polyploid cells in the posterior midgut upon Ttk-depletion driven by myo1A-GAL4; mex1-GAL4, and GMR23G10-GAL4, respectively; Error bars represent Mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed Student’s t test). “n” indicate the number of guts used for quantification; source data are provided as a Source Data file. Scale bars, 20 μm.

We next analyzed and compared the EC-to-EE transdifferentiation efficiency in these different EC populations. Knocking down Ttk in Mex1-GAL4 > GFP+ ECs caused approximately 14.1% ± 1.3% (60/426) of the cells to turn on Pros expression, and only a small number of cells (0.58 ± 0.19 cells per gut, n = 12 guts) showed Tk expression, suggesting that these cells have a lower transdifferentiation efficiency than the Myo1A>GFP+ cells (Fig. 6g, i, j). Interestingly, knocking down Ttk in GMR23G10 > GFP+ ECs caused approximately 8.9 ± 1.8% (60/727) of cells to turn on Pros expression (Fig. 6h, i), but none of these cells (totally 727 cells examined) showed Tk expression (Fig. 6h, j), indicating that GMR23G10-GAL4 > GFP+ ECs have the lowest transdifferentiation efficiency. These data support the idea that cell plasticity is high in young or newly differentiated ECs and declines rapidly with maturation/ age.

The process of EC differentiation, maturation, and aging is often accompanied by an increase in endoreplication, which involves DNA replication without cell division, leading to polyploidy38,39. This raises the question of whether polyploidy can hinder cell transdifferentiation. To investigate this, we employed RNAi to knock down E2F1, a TF essential for driving EC endoreplication. As a result, we observed a significant reduction in nuclear sizes of myo1A>GFP+ cells (Supplementary Fig. 7a, b), indicating successful interference with the endoreplication process. Interestingly, when E2F1 was co-depleted with Ttk, we observed a mild but significant increase in Pros+ cells within the myo1A>GFP+ cell population (Supplementary Fig. 7c, d). This finding suggests that polyploidy might act as a barrier during the process of EC-to-EE transdifferentiation.

Lineage confliction as a barrier to the EC-to-EE transdifferentiation

Conflicting alternative lineages have been indicated as potential barriers against reprogramming in mammalian studies40,41, and our previous work showed that Ttk depletion in ISCs caused the derepression of several neuroblast-specific TFs, including Deadpan (Dpn), which is a neural stem cell factor that, when activated in ISCs, drives self-propagation and prevents EE differentiation42. We speculated that if Ttk depletion in ECs led to the derepression of neuroblast genes, lineage confliction could potentially occur and prevent EC-to-EE transdifferentiation. Although the RNA-seq data of the transdifferentiated cells did not capture any expression of the neuroblast genes, we detected Dpn protein expression, by surprise, in many ECs in Myo1A>ttk-RNAi guts. By co-staining with Pros, we found that the expression of Dpn and Pros in ECs were mutually exclusive (Supplementary Fig. 8a, b).

To further examine the inverse relationship between Dpn and Pros expression, we generated a pros-mCherry knock-in line which yields a Pros Protein fused with mCherry at its C-terminal, and used it to sort out Pros-activated and Pros-not-activated polyploid cells in Myo1Ats>ttk-RNAi guts and performed quantitative PCR analysis. There was indeed a significant upregulation of Dpn expression in Pros-not-activated cells compared to that in Pros-activated cells (Supplementary Fig. 8c).

To investigate whether Dpn depression acts as a barrier to EC-to-EE transdifferentiation, we simultaneously depleted Dpn and Ttk in ECs, and observed significantly increased percentage of polyploid cells turning on Pros, Tk or AstC: approximately 65.3 ± 7.1% (n = 5 guts) of polyploid cells turned on Pros expression, and on average 8.6 ± 1.7 (n = 7 guts) polyploid cells turned on AstC expression (Supplementary Fig. 8d–g), showing approximately 3–5 fold increase in EC-to-EE transdifferentiation efficiency.

These observations suggest that the derepression of neural stem cell factors and EE-promoting factors following Ttk depletion can cause a lineage-conflict scenario in which Dpn exerts a negative effect on Pros activation, thereby preventing EC-to-EE transdifferentiation.

A genetic screen identified NuRD complex as a barrier to the EC-to-EE transdifferentiation

The establishment of this in vivo EC-to-EE transdifferentiation system in Drosophila provided an opportunity to identify new regulators involved in reprogramming through genetic screens. In this regard, we screened about 120 chromatin-related regulators, including polycomb group genes, regulators for heterochromatin formation, chromatin remodeling factors, and histone modification enzymes (Supplementary Data 5), and aimed to identify those whose depletion could increase transdifferentiation efficiency following Ttk depletion (Fig. 7a). From this screen, we identified MEP-1 and Mi-2, both encode essential components of the NuRD nucleosome remodeling complex43,44. We found that simultaneous knockdown of MEP-1 and Ttk (Myo1Ats>MEP-1-RNAi & ttk-RNAi) or dMi-2 and Ttk in ECs (Myo1Ats> dMi-2-RNAi & ttk-RNAi) led to a significant increase in Pros+ ECs (approximately 85.6 ± 1.8% and 81.3 ± 2.04% ECs turned on Pros, respectively) or TK+ ECs (5.7 ± 0.4 and 5.8 ± 0.5 ECs per R5 region turned on Tk, respectively) (Fig. 7b–g), indicating enhanced transdifferentiation. As controls, knockdown of MEP-1 or dMi-2 alone did not produce any obvious phenotype (Supplementary Fig. 9). As the NuRD complex possesses histone deacetylase activity for the establishment of local repressive transcription, and has been implicated in Ttk and other transcriptional factor-mediated transcriptional repression44,45, it may provide an epigenetic memory of local repressive transcriptional status in ECs, thereby acting as a barrier to transdifferentiation. Interestingly, NuRD has been implicated as an important regulator of reprogramming in mammalian cells46.

Fig. 7: Genetic screen identifying NuRD complex as a barrier to EC-to-EE transdifferentiation.
figure 7

a Screening strategy for identifying potent suppressors of EC-to-EE trans-differentiation upon Ttk-depletion. Candidate TFs or epigenetic regulators were individually or co-depleted with Ttk-RNAi using Myo1A-GAL4. The percentage/number of polyploid cells expressing Pros or Tk was calculated and compared. b, c Compared to the control (b), knocking down ttk in ECs leads to the activation of Pros and Tk expression in a subset of these cells (c). d, e Compared to the control of Ttk knockdown alone (c), simultaneously knocking down Ttk with either MI2 (d) or MEP1 (e) significantly increases the proportion of polyploid cells expressing Pros and Tk. f Quantification of percentages of Pros+ cells in polyploid cells. Error bars represent Mean ± SEM; ***p < 0.001 (two-tailed Student’s t test). “n” indicates the number of guts used for quantification; source data are provided as a Source Data file. g Quantification of the number of polyploid cells expressing Tk in each gut. Error bars represent Mean ± SEM ***p < 0.001 (two-tailed Student’s t test). “n” indicates the number of guts used for quantification; source data are provided as a Source Data file. h Schematic explaining cell-fate plasticity in the Drosophila ISC lineage. Despite having distinct transcriptome profiles, ISCs, EEs, and ECs exhibit a remarkably similar chromatin accessibility profile on loci associated with stem cell identity genes, EC identity genes, and EE identity genes. The permissive chromatin landscape allows for cell identity switches within the ISC lineage: Depletion of Ttk in ECs initiates EC-to-EE transdifferentiation, which is mediated by the master regulator Pros. Depletion of Pros in EEs initiates EE-to-ISC dedifferentiation due to the derepression of ISC identity genes. However, the reprogramming process is hindered by various barriers, including NuRD chromatin remodeling complex-mediated epigenetic memory, lineage confliction, polyploidy, and age, which limit fate-conversion efficiency. Scale bars, 20 μm.