(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Differential gene expression analysis identified determinants of cell fate plasticity during radiation-induced regeneration in Drosophila

['Michelle Ledru', 'Department Of Molecular', 'Cellular', 'Developmental Biology', 'University Of Colorado', 'Boulder', 'Colorado', 'United States Of America', 'Caitlin A. Clark', 'Jeremy Brown']

Date: 2022-03 

Ionizing radiation (IR) is used to treat half of all cancer patients because of its ability to kill cells. IR, however, can induce stem cell-like properties in non-stem cancer cells, potentiating tumor regrowth and reduced therapeutic success. We identified previously a subpopulation of cells in Drosophila larval wing discs that exhibit IR-induced stem cell-like properties. These cells reside in the future wing hinge, are resistant to IR-induced apoptosis, and are capable of translocating, changing fate, and participating in regenerating the pouch that suffers more IR-induced apoptosis. We used here a combination of lineage tracing, FACS-sorting of cells that change fate, genome-wide RNAseq, and functional testing of 42 genes, to identify two key changes that are required cell-autonomously for IR-induced hinge-to-pouch fate change: (1) repression of hinge determinants Wg (Drosophila Wnt1) and conserved zinc-finger transcription factor Zfh2 and (2) upregulation of three ribosome biogenesis factors. Additional data indicate a role for Myc, a transcriptional activator of ribosome biogenesis genes, in the process. These results provide a molecular understanding of IR-induced cell fate plasticity that may be leveraged to improve radiation therapy.

Ionizing radiation (IR) is used to treat half of all cancer patients because of its ability to kill cells but treatment failures are common because tumors grow back (regenerate). Here, we asked which changes in the properties of cells facilitate regeneration in Drosophila (fruit flies) after exposure to radiation. We identified six genes whose products increase or decrease the regenerative potential of cells. These results help us understand how tissues regenerate after IR damage and will aid in designing better therapies that involve radiation.

Funding: TTS was supported by a National Institutes of Health grant R35 GM130374. AG and SF were supported by the University of Colorado Cancer Center’s Biostatistics and Bioinformatics Shared Resource that is funded by a National Institutes of Health grant P30 CA046934. The URL for the funder is https://www.nih.gov . The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. TTS, AG and SF received partial salary support from their respective funders.

Here, we performed a genome-wide analysis of gene expression changes during regeneration of the Drosophila larval wing disc. By dissociating the discs into single cells and using Fluorescence-Activated Cell Sorting (FACS) we monitored the hinge versus the rest of the disc, to identify genes that are differentially expressed in the two cell populations during regeneration and fate change. Functional testing of 42 candidate regulators identified six genes whose experimental manipulation compromised hinge-to-pouch conversion. These six genes encode hinge determinants (Zfh2 and Wg) or ribosome biogenesis factors (RpI135, Rs1, Tsr1 and Myc), suggesting that cell fate plasticity during regeneration in irradiated wing discs requires downregulation of transcripts for hinge determinants and concomitant upregulation of translation capacity. These results illustrate that cell fate change requires much more than simply transcribing genes for the new fate; suppression of old fate determinants and post-transcriptional mechanisms that generate new fate determinants appear equally important.

The hinge region of the wing disc displays unique features. It experiences a combination of high Wingless (Wg, Drosophila Wnt1) and Stat92E (Drosophila STAT3/5) signals, which act together to promote the growth and differentiation of the region during normal development into a structure that connects the wing blade to the body wall in the adult [ 14 ]. In the larval wing disc, hinge cells show different cytoskeletal and extracellular matrix organization than the rest of the disc and are highly tumorigenic; they undergo neoplastic transformation under conditions that have little effect on the rest of the wing disc such as mutations in tumor suppressors lgl and scrib, leading to the hinge being called a ‘tumor hotspot’ [ 15 ]. When the adjacent pouch region is genetically ablated by localized ectopic expression of pro-apoptotic genes, the cells of the hinge change fate to regenerate the pouch [ 16 ]. We have shown that hinge cells are resistant to IR-induced apoptosis and that this protection depends on Wg and Stat92E [ 14 ]. Importantly, irradiated hinge cells lose the hinge fate as seen by the loss of expression of hinge markers, gain pouch fate as seen by the gain of expression of a pouch marker, translocate to the pouch and help regenerate the latter. Candidate testing identified apoptotic caspases and cytoskeletal proteins as important for hinge-to-pouch conversion [ 12 , 17 ], but we lack a comprehensive understanding of this fate change process.

Regeneration of Drosophila larval organs called imaginal discs occurs without a dedicated stem cell pool. We identified a previously unknown mode of regeneration in Drosophila larval wing discs whereby epithelial cells acquire stem cell-like properties after irradiation [ 12 – 14 ]. These properties include the ability to change cell fate and translocate to areas of the disc with greater need for cell replenishment. The ability to behave like stem cells, we found, is induced by IR and is limited to a specific subset of cells within the wing disc, those of the future hinge that connects the wing blade to the body.

More than half of cancer patients receive ionizing radiation (IR), alone or as a component of their treatment ( www.cancer.org ). IR induces DNA damage to kill cells. Surviving cancer cells could, however, regenerate the tumor, leading to treatment failure. Regeneration of tumors is attributed to cancer cells with stem cell-like properties [ 1 , 2 ]. These cells, distinguished by cancer type-specific markers such as CD44 and ALDH for Head and Neck Squamous Cell Carcinoma, survive treatment and show elevated capacity to initiate tumors compared to their counterparts that lack the markers [ 3 , 4 ]. Identifying and eliminating cancer stem-like cells is a goal for improved therapy. A body of literature shows that the proportion of cancer cells with stem cell markers increases after treatment [ 5 ], and that in some cases, the treatment itself such as IR is found to induce stemness [ 6 – 9 ]. There is also evidence for non-stem cells acquiring stem cell-like properties in the context of regeneration of normal tissues after radiation damage. In irradiated mouse intestine, Paneth cells de-differentiate to populate the stem cell compartment [ 10 ]; forced activation of Notch signaling recapitulates this process in the absence of irradiation. In irradiated mouse salivary glands, acinar and duct cells adopt plasticity to regenerate acinar cells [ 11 ]. Thus, there is mounting evidence that IR can induce cell fate plasticity, but the mechanisms remain to be fully understood.

Results

Characterization of hinge-specific 30A-GAL4 driver To lineage-trace hinge cells, we used a hinge-specific 30A-GAL4 driver, whose expression is entirely within the hinge region as defined by antibody staining for the hinge determinant Zfh2 [12]. We mapped 30A-GAL4 by plasmid rescue to 21E2, at genomic location 2L:685363. 685363[+] (Release 5.28 of D. melanogaster genome), within an intron of the dachsous gene that encodes a member of the cadherin family ([18] gene nomenclature follows FlyBase). ds is expressed more broadly in the wing disc than 30A-GAL4 [18], suggesting that 30A recapitulates only a subset of ds expression. More important, expression of the dual-color G-trace reporter in which RFP is the real-time maker and GFP is the lineage tracer [19] showed that cell fates are stable in the 30A-GAL4 domain as indicated by RFP/GFP overlap ([12–14]; reproduced in Fig 1B and 1B’). After irradiation, however, some hinge cells marked with GFP lose hinge-specific 30A-GAL4 activity (become RFP-) and translocate to the pouch region ([12–14]; Fig 1F and 1F’ arrows). We interpret RFP+GFP+ cells as those in the hinge and RFP-GFP+ cells as those that originated in the hinge but lost the hinge identity. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Lineage tracing to monitor cell fate changes after irradiation. Larvae expressing G-trace under the control of 30A-GAL4 were irradiated with 0 or 4000R of X-rays. The discs were dissected at various times shown after irradiation, fixed, stained for DNA and imaged for DNA, RFP and GFP. The larvae were of the genotype w1118/+ or Y; 30A-GAL4>UAS-G-trace/+ produced by a cross between w1118 and 30A-GAL4>UAS-G-trace/SM5 and sorted for RFP/GFP. Scale bar = 100 microns. (A) The protocol used to generate the samples for RNAseq. (B-F) Representative discs from various time points. (B-B’) show discs at the time of irradiation. Note the overlap of RFP and GFP, indicating little fate plasticity, and the absence of GFP+ cells in the pouch within the yellow line in B. DNA images are used to discern the different regions of the disc; n = notum, h = hinge, p = pouch. (C-C’) show an unirradiated disc at 24h after mock irradiation, the time point for -IR RNAseq samples. (D-D’) show an irradiated disc at 24h after IR, showing little fate change. The IR+24h RNAseq samples were from such discs. (E-E’) show an irradiated disc at 48h after IR, showing the first indications of fate change as seen by RFP-GFP+ cells (arrows). Similar results are seen also in larvae that were irradiated 3 days from the end of egg collection. The IR+48h RNAseq samples included larvae irradiated 3 or 4 days from the end of egg collection. (F-F’) show an example of a disc with fate change as indicated by RFP-GFP+ cells (arrows) in the pouch (yellow circle). Such fate change was seen at the IR+72h time point when larvae were irradiated 3d (shown here) or 4d from the end of egg collection [14]. (G) The temperature shift and irradiation protocol used in the functional tests. The temperature shift inactivates GAL80ts that was present in the crosses for the functional tests. https://doi.org/10.1371/journal.pgen.1009989.g001 We have shown previously that fate change and translocation commence at about 48h after exposure to 4000R of X-rays [14]. We reproduce this result here. RFP-GFP+ cells are scarce at 24h after irradiation (Fig 1D’) but appear by 48h after irradiation (Fig 1E’ arrows). Fate change is maximal at 72h after irradiation such that RFP-GFP+ cells are found in the pouch (Fig 1F’ arrows; yellow line indicates the pouch). We lose the larvae to pupariation at later times after IR. This schedule of hinge-to-pouch conversion was observed whether we irradiated larvae 3 days from the end of egg collection [13] or 4 days from the end of egg collection [14]. We showed previously that RFP-GFP+ cells in the pouch have not only lost the hinge fate (became RFP-) but also have acquired the pouch fate as detected by VgE-lacZ expression [12]. To understand the molecular changes that regulate fate change and translocation, we analyzed gene expression at 24 and 48h after irradiation and at 24h without irradiation. These earlier time points, we reasoned, are more likely to capture changes that cause fate change/translocation by 72h after IR. The 72h time point was used in functional testing of candidate regulators of fate change described later, with a temperature shift in the protocol to activate GAL4 conditionally (Fig 1G).

IR-induced gene expression changes in the hinge versus the rest of the disc To capture IR-induced gene expression changes specifically in the hinge, the discs were dissociated into single cells and sorted into RFP+GFP+ (hinge) and RFP-GFP- pools (FACS profile in S1 Fig). The latter group would include cells of the pouch and the notum. We detected RFP-GFP+ cells as well, but these were not included in the analysis for two reasons. First, we reasoned that they have already changed fate and therefore were beyond the state of interest. Second, there were too few RFP-GFP+ cells at the time points of interest, 24 and 48h after IR, for meaningful analysis. Thus, we compared 6 samples: RFP+GFP+ (POS or pos) and RFP-GFP- (NEG or neg) for each of the three time points, -IR, IR+24h and IR+48h. Each sample was acquired in duplicate to achieve two biological replicates of the whole experiment. To assess the accuracy of cell sorting procedures, genes known to be differentially expressed in the hinge (8 genes), the pouch (9 genes) or the notum (10 genes) were identified from the literature [20–31], and their basal (-IR) expression was compared between POS and NEG samples (Fig 2A). POS cells showed higher expression of hinge-expressed genes such as zfh2 (log 2 = 3.3) [31], upd1 (log 2 = 3.6) [26], and Sox15 (log 2 = 4.1)[23]. Expression of ds, whose sequences drive G-trace, was also ~3 fold higher (log 2 = 1.8) in the POS cells than in NEG cells. Conversely, NEG cells, which would include both the pouch and the notum, showed higher levels of transcripts known to be primarily or exclusively expressed in the notum such as zfh1 (log 2 = 4.9), pnr (log 2 = 2.2), and tup (log 2 = 1.8) and the pouch such as ana (log 2 = 2.9), cyp310a1 (log 2 = 1.6), dve (log 2 = 1.4) and vg (log 2 = 1.2) [20,21]. We note that zfh1 is not expressed in the disc proper but in muscle precursors that are associated with the notum [32]. Gene set enrichment analysis [33] showed statistically significant enrichment of hinge gene expression in POS cells and enrichment of notum and pouch gene expression in NEG cells (Fig 2B, top three rows). One representative example each of a gene differentially expressed in POS (zfh2) or NEG cells (pnr) is shown in Fig 2C and 2D. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Differential gene expression in POS/pos (RFP+GFP+) versus NEG/neg (RFP-GFP-) cells without irradiation. FlyBase nomenclature for gene names is used throughout the manuscript (https://flybase.org). (A) The heatmap shows the expression in POS and NEG cells of genes that are known from the literature to show elevated expression in the hinge, notum, or pouch. (B) Aggregate gene set enrichment scores of the indicated gene sets. Each tick mark represents a gene in the pathway. The barplot shows the normalized enrichment scores (NES, the degree to which a gene set is overrepresented, normalized for gene set size) and p-values adjusted for False Discovery Rate (FDR, estimated probability that a gene set with a given NES represents a false positive finding). See www.gsea-msigdb.org for further definitions. The plot at the bottom shows the fold-changes of genes when comparing POS versus NEG samples in the same order as the pathway data above. Significant (FDR<0.05) genes are highlighted in red. (C-D) The expression in counts per million (CPM) in POS and NEG cells of zfh2 and pnr, representative transcripts with high expression in the hinge and the notum, respectively. Pairs of dots represent data from two biological replicates. https://doi.org/10.1371/journal.pgen.1009989.g002 As described above, the hinge region of the wing disc differs from the rest of the disc in terms of cellular architecture and regenerative potential [14,15]. To identify additional differences, we asked what functional groups show differential expression between the hinge and the pouch/notum. To our surprise, genes involved in ribosome biogenesis were among the most significant pathways (Fig 2B, 4th row from the top). These are expressed at lower levels in POS cells compared to the NEG cells. A key transcriptional regulator of ribosome biogenesis is the proto-oncogene myc [34–36]. Consistent with this, transcriptional targets of Myc are expressed at significantly lower levels in POS cells compared to the NEG cells (Fig 2B, bottom row).

Gene expression changes following irradiation in POS and NEG cells We have shown before that the hinge and the rest of the disc show similar incidence of cells in S and M phases without IR; that is, there are no inherent differences in cell proliferation between these disc regions [14]. Likewise, the hinge and the rest of the disc suffer a similar level of DNA damage after irradiation as detected by γ-H2Av (Drosophila γ-H2Ax) staining and show similar kinetics of DNA repair as seen by a time course of disappearance of γ-H2Av [14]. We published IR-induced genome-wide gene expression changes in the whole wing disc and identified changes in genes with roles in DNA damage responses such as DNA repair [37]. We expect these changes to be shared by cells in different parts of the wing disc. The present study is focused not on shared DNA damage responses, which typically occur within 24h after irradiation, but on cell fate changes that occur at later times. Specifically, we are interested in differences in the IR-induced behavior between the hinge and the rest of the disc. To this end, we performed a time course analysis using maSigPro to determine genes whose expression differs between POS and NEG cells following irradiation [38]. This analysis identified 821 genes significantly different across the time course (S1 Table). Gene clustering identified 7 clusters with similar expression patterns (Figs 3A and S2). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Time series and ORA analysis. (A) Heatmap of 821 significant genes from the time series analysis. Expression data is in S1 Table. Gene names are in S2 Table. (B) Median expression (counts per million) of the genes in Cluster 1 at various treatment and time points. Gene names are in S3 Table. Pairs of circles represent data from two biological replicates. (C) Top five most significant gene sets following ORA for genes in Cluster 1 are shown with FDR-adjusted p-values. Gene names are in S3 Table. (D) Median expression (counts per million) of the genes in Cluster 2 at various treatment and time points. Pairs of circles represent data from two biological replicates. (E) Top five most significant gene sets following ORA for genes in Cluster 2 are shown with FDR-adjusted p-values. Gene names are in S3 Table. https://doi.org/10.1371/journal.pgen.1009989.g003

Over-Representation Analysis identified enriched pathways in Clusters 1 and 2 Clustering of the significant genes revealed two clusters of particular interest. Cluster 1 included genes that show lower basal expression in POS cells compared to NEG cells but became more similar after irradiation (Fig 3B). We hypothesize that these genes could encode factors that are upregulated in hinge cells (POS) as they fate-change into pouch cells (NEG), thus making POS cells more like NEG cells. Over Representation Analysis (ORA) of these genes identified pathways involved in ribosome biogenesis and Myc targets (Fig 3C). Cluster 2 included genes that show higher basal expression in POS cells than in NEG cells but became more similar after irradiation (Fig 3D). We hypothesize these genes could encode factors that are downregulated in hinge cells (POS) as they fate-changed into pouch cells (NEG), thus making POS cells more like NEG cells. ORA of these genes identified pathways involved in differentiation-related processes of regionalization, development and specification (Fig 3E). All gene sets with an adjusted p-value <0.05 are shown in S3 Table. Based on these results and other considerations described in S4 Table, 42 genes were tested for functional importance in IR-induced fate change, using overexpression or knock-down approaches. Gene expression changes could drive fate change (drivers) or simply accompany fate change (passengers); therefore, functional tests are needed to tell these apart. Manipulation of six of these genes produced positive data and are described below.

Functional test of ORA groups; Myc An ORA group in Cluster 1 includes Myc targets (Fig 6A, see also Fig 3C and S3 Table). Myc is a key transcriptional regulator of ribosome biogenesis in Drosophila and vertebrates [34–36]. RpI135 is a Myc transcriptional target in the Drosophila wing disc [47] and is needed for IR-induced fate change (Fig 5I). Therefore, we asked if Myc is also required for IR-induced fate change. We expressed three different Myc RNAi constructs in the hinge conditionally with the 30A-GAL4 driver and quantified the effect on fate change. myc is an essential gene; broad and continuous depletion of Myc is expected to cause lethality. Therefore, as for RPI135, we expressed the RNAi constructs broadly and constitutively (with en-GAL4 driver, without GAL80 or IR) and quantified organismal lethality as an independent measure of RNAi efficacy (Fig 6B). Only two RNAi constructs against Myc produced lethality, with one (v2947) showing complete lethality and the other (BL25783) partial lethality. These two are also the constructs that depleted Myc protein (S4 Fig); the third construct, BL36123, had no effect on Myc protein level or lethality (Figs 6B and S4). In fate change assessment, the RNAi construct that produced complete lethality was also the one that inhibited fate change with statistical significance (v2947, Fig 6C–6E). We conclude that effective depletion of Myc inhibited IR-induced fate change. The effect of Myc or ribosome biogenesis on fate change could be because they are needed for fate-changing hinge cells to grow and proliferate as they translocate into the pouch. We do not favor this possibility for the following reason. Although constitutive loss of ribosome function or Myc is expected to inhibit growth and proliferation, conditional depletion of these functions using the protocol in Fig 1G does not appear to affect growth or proliferation; quantification of the RFP+ hinge area showed no significant difference between controls and RNAi discs (Fig 6F).

[END]

[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1009989

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/