(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

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



Erebosis, a new cell death mechanism during homeostatic turnover of gut enterocytes

['Hanna M. Ciesielski', 'Division Of Developmental Biology', 'Regenerative Medicine', 'Kobe University', 'Kobe', 'Physiological Genetics Laboratory', 'Riken Cpr', 'Hiroshi Nishida', 'Division Of Cell Physiology', 'Tomomi Takano']

Date: 2022-05

Many adult tissues are composed of differentiated cells and stem cells, each working in a coordinated manner to maintain tissue homeostasis during physiological cell turnover. Old differentiated cells are believed to typically die by apoptosis. Here, we discovered a previously uncharacterized, new phenomenon, which we name erebosis based on the ancient Greek word erebos (“complete darkness”), in the gut enterocytes of adult Drosophila. Cells that undergo erebosis lose cytoskeleton, cell adhesion, organelles and fluorescent proteins, but accumulate Angiotensin-converting enzyme (Ance). Their nuclei become flat and occasionally difficult to detect. Erebotic cells do not have characteristic features of apoptosis, necrosis, or autophagic cell death. Inhibition of apoptosis prevents neither the gut cell turnover nor erebosis. We hypothesize that erebosis is a cell death mechanism for the enterocyte flux to mediate tissue homeostasis in the gut.

Funding: This work was supported by AMED-PRIME (17939907) and the JSPS KAKENHI (JP16H06220) to S.K.Y.; Cooperative Study Program (20-225) of National Institute for Physiological Sciences to S.K.Y. and M.F.; Joint research program of the Institute for Molecular and Cellular Regulation (Gunma University) to S.K.Y. and T.N. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, we report the discovery of a previously unrecognized phenomenon in enterocytes, which are characterized by accumulation of Ance and loss of indispensable molecules and organelles. Ance is a Drosophila homolog of the mammalian Angiotensin-converting enzyme [ 10 – 13 ]. Although Angiotensin-converting enzyme is well known to convert angiotensin 1 to angiotensin 2 and to inactivate bradykinin, its substrate specificity is relatively low [ 14 ]. Drosophila Ance is secreted, and its substrates remain to be determined. Its expression has been described in imaginal cells (imaginal discs, abdominal histoblasts, gut imaginal cells, and imaginal salivary gland) [ 11 ], hemocyte progenitors [ 15 ], male accessory glands [ 16 ], and the testis [ 12 ]. Recently, Ance was shown to regulate imaginal disc development downstream of Dpp signaling [ 17 ].

Epithelial tissues are in a state of flux [ 1 ]. In the Drosophila gut, enterocytes, which absorb nutrition, are the major differentiated cell type [ 2 ]. The turnover rate under the physiological condition has been estimated from 4 days to 3 weeks [ 3 – 5 ]. Enterocytes have been described to die by apoptosis and to be apically extruded to the lumen during physiological homeostasis as well as under stress conditions such as infection and tissue damage [ 3 , 4 , 6 – 8 ]. In response to loss of enterocytes, intestinal stem cells (ISCs) proliferate to maintain the gut homeostasis [ 3 ]. On the other hand, it has been difficult to detect apoptosis reliably in the gut; thus, vital dyes such as SYTOX have been often used to detect dying cells [ 3 , 9 ], implying that alternative mechanisms might exist to regulate the gut cell turnover.

Results

We serendipitously discovered that immunostaining of Ance demonstrates patchy labeling of gut enterocytes while looking at its expression patterns in diverse organs (Fig 1A and 1B). We detected Ance mainly in R1, R2, and R4 regions of the adult midgut (Fig 1A and 1B). Cell turnover is the most active in R4 (also known as P1 to P3) [2,3,18,19], and most research on the Drosophila midgut turnover focused on the R4 region or on the posterior midgut [3–5]. We also decided to focus on the R4 region in this study. Ance staining was specific because 2 Ance mutants [17] demonstrated no Ance immunostaining (S1A–S1D Fig). In addition to the patchy Ance staining within cells, we also noticed that there is a weak signal of Ance outside of the cell at the apical, luminal side, which reflects a secreted one (S1E Fig). During aging up to 3 weeks after eclosion, the number of Ance-positive cells remained constant (S1F Fig). We also confirmed that Ance-positive cells exist in the midgut of 2-month-old flies (S1G Fig). We observed similar expression patterns of Ance in both males and females (S1H Fig); thus, we used female flies for technical simplicity in this study.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. An Ance+ enterocyte subpopulation exhibits unique features. (A) Cells in R2 (arrowhead) and R4 (white line) express Ance. (B) An enterocyte subpopulation displays Ance protein. (C) Nuclei of Ance+ cells (arrowhead) are flattened and lie in close apposition to the basal, visceral muscle. (D-E) Quantification of nuclear height (D) and nuclear area (E). (F) Some Ance+ enterocytes show weak (white arrowhead) or absent (yellow arrowhead) genome staining by DAPI. (G) Immunostaining of nuclear LaminB1 is reduced in Ance+ enterocytes (arrowheads). (H) Ance+ enterocytes decrease F-actin indicated by the absence of phalloidin staining. (I-K) Ance+ enterocytes exhibit less cell adhesion components. Immunostaining for Dlg labeling septate junctions (I) and β-catenin/Arm labeling adherens junctions (J) was reduced in Ance+ cells (arrowheads). Likewise, cadherin GFP knock-in flies demonstrate less adherens junctions in Ance+ enterocytes (arrowhead) (K). Statistical significance was determined by using a two-tailed unpaired t test (D and E). S1 Data provides the source data used for all graphs and statistical analyses. Scale bars, 200 μm (A), 20 μm (B-C, F-K). Arm, Armadillo; Dlg, Discs large. https://doi.org/10.1371/journal.pbio.3001586.g001

We noticed that cells that have Ance possess several unusual characteristics. They are usually flat and lie in close apposition to the basal, visceral muscle layer (Fig 1C and 1D). The nuclear diameter tends to be larger than surrounding enterocyte neighbors (Fig 1E). Occasionally, genome staining with DAPI or Hoechst was weak, or disappeared, in an extreme case (Figs 1F and S1I), suggesting that they either lose a DNA content or adopt a closed chromatin structure that makes the dyes inaccessible. The nuclear lamin is also reduced in Ance-positive cells (Figs 1G and S1J). Cells with Ance have a low amount of actin filament (Figs 1H and S1K). They also have less amounts of septate junctions and adherens junctions (Figs 1I, 1J, S1L and S1M). This was not due to an artifact of immunostaining, since cad-GFP also demonstrates a reduced amount of adherens junctions in cells with Ance (Figs 1K and S1N).

Myo1D-Gal4 has been extensively used as a pan enterocyte driver in the field. We examined whether Myo1D-Gal4 promotes expression of GFP in Ance-positive cells.

Intriguingly, we did not detect GFP signals in all enterocytes with Myo1D-Gal4 (Fig 2A). We found that Ance and Myo1D-Gal4–driven GFP demonstrate a complementary pattern, that is, an inverse relationship (Fig 2A and 2B). This suggests that either Myo1D-Gal4 does not drive GFP expression in Ance-positive cells or that GFP is denatured, degraded, or somehow lost. Ance-positive cells could not be labeled with GFP driven by actin-Gal4, tub-Gal4, or another enterocyte driver Npc1b-Gal4 (S2A–S2C Fig).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Ance+ enterocytes undergo erebosis. (A) Ance+ immunostaining displays an inverse relationship to Myo1D-driven GFP fluorescence. Erebotic cells are labeled by Ance or loss of GFP signals. (B) A correlation analysis of median fluorescence intensity of Ance and GFP in the cell cytoplasm indicates that enterocytes exhibiting high Ance signals show low GFP signals and vice versa. Pearson’ correlation coefficient (R) was calculated: R = −0.6062, R2 = 0.3675, P < 0.0001. (C) Myo1D-driven GFP and nuclear RFP reveal a gradual process of erebosis. At early erebosis (1), cells lose cytoplasmic GFP but retain nuclear GFP and RFP. At intermediate erebosis (2), cells do not have signals of GFP but still retain nuclear RFP. At late erebosis (3), cells lose both GFP and RFP signals. Erebosis-induced protrusion can be observed in enterocytes near erebotic cells (yellow arrowheads). Small cells adjacent to erebotic cells are frequently visible (orange arrowheads). (D) GFP protein was not detectable by a GFP nanobody in erebotic enterocytes lacking Myo1D-driven GFP fluorescence. (E) Time-lapse imaging demonstrates that loss of GFP occurs in a live condition. The yellow arrowhead indicates a cell that had already undergone erebosis at the time of imaging. The orange arrowhead indicates a cell that lost cytoplasmic GFP during imaging. The loss of GFP occurred within 5 minutes. S1 Data provides the source data used for all graphs and statistical analyses. Scale bars, 20 μm. https://doi.org/10.1371/journal.pbio.3001586.g002

At this moment, due to the peculiarity of Ance-positive cells (weak nuclear staining, flat and large nucleus, loss of F-actin and junction components, loss of GFP signals), we named this phenomenon erebosis, based on the ancient Greek word έρεβος [erebos] meaning deep darkness.

Intriguingly, when we expressed both GFP and nlsRFP using Myo1D-Gal4, we noticed that erebotic cells (cytoplasmic GFP-negative cells) usually have nlsRFP signals (Fig 2C). This indicates that Myo1D does induce expression of Gal4 in erebotic cells but that erebotic cells lose GFP signals. Close examination showed that there are 3 types of erebotic cells (cytoplasmic GFP negative) with different amounts of nuclear GFP and nuclear RFP. The first one retains nuclear RFP and faint signals of nuclear GFP. The second one has only nuclear RFP without nuclear GFP. The third one has lost both nuclear RFP and nuclear GFP. Assuming that losing fluorescence is a gradual process, we reason that, during erebosis, cells first lose cytoplasmic GFP, then nuclear GFP, followed by loss of nuclear RFP. Longer retention of nuclear RFP signals compared to nuclear GFP is likely due to the stability difference of the fluorescent proteins. For example, RFP is more resistant to low pH than GFP [20]. Similar to the case of nuclear GFP and RFP, cytoplasmic RFP became reduced but persisted longer than cytoplasmic GFP in erebotic cells (S2D and S2E Fig). As a typical feature of erebosis, we also noticed that enterocytes surrounding the erebotic cells often protrude toward erebotic cells (Fig 2C).

Since GFP signals are lost in erebotic cells, we investigated whether loss of GFP signals is due to protein denaturing, for example, in an acidic environment within erebotic cells, or due to bona fide loss of protein. If denaturing leads to loss of GFP signals, a GFP antibody should still be able to detect GFP in erebotic cells. We could not detect GFP in erebotic cells even with a GFP nanobody (Fig 2D), suggesting that GFP is likely lost rather than being denatured in erebotic cells.

To clarify that loss of GFP is not an artifact of fixation, we visualized erebotic cells and the moment of erebosis performing live imaging. We readily observed erebotic cells in a live condition (Fig 2E and S1 Movie). Time-lapse live imaging revealed that loss of cytoplasmic GFP occurs fast (within 5 minutes) (Fig 2E and S1 Movie). The nuclear signals of GFP and RFP persisted relatively long time, at least 12 hours after the loss of cytoplasmic GFP. Live imaging definitively demonstrates that loss of GFP is not an artifact of fixation, as occurring in a live setting.

Taking advantage of GFP loss in erebotic cells, we observed the ultrastructure of erebotic cells by performing immuno-electron microscopy (immuno-EM) with a GFP antibody in the gut where Myo1D-Gal4 drives GFP (Fig 3A–3E). Cells with a low amount of GFP show a reduced number of mitochondria and a reduced cytoplasmic content above the nucleus (Fig 3C and 3D). We also observed that these cells have shorter microvilli (Fig 3E). The immuno-EM data prompted us to examine organelles in erebotic cells with fluorescent microscopy. Consistent with the immuno-EM data, we found that erebotic cells have a reduced number of mitochondria (Fig 3F). They also have reduced amounts of ER and Golgi (Fig 3G).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Erebotic cells lose organelles. (A) Low magnification transmission electron micrograph of GFP-negative (GFP−) cells. Squares indicate the regions shown in high-magnification images. (A’) The enlarged image of the basal region of GFP− cells shows sparse labeling by anti-GFP (arrowheads). (A”). The enlarged image of the apical region of GFP− cells shows short microvilli, a reduced number of mitochondria, and reduced apical cytoplasm. (B) Low magnification transmission electron micrograph of GFP-positive (GFP+) cells. Squares indicate the regions shown in high-magnification images. (B’) The enlarged image of the basal region of GFP+ cells shows numerous labeling by anti-GFP (arrowheads). (B”) The enlarged image of the apical region of GFP+ cells shows well-developed microvilli and the apical cytoplasm rich with mitochondria. (C) Quantitation of apical mitochondria area fraction shows fewer apical mitochondria in GFP− cells. (D) Quantitation of the ratio of apical membrane–nucleus distance and cell height shows reduced apical cytoplasmic content in GFP− cells. (E) Quantitation of microvilli length shows shorter microvilli in GFP− cells. (F) Erebotic cells exhibit a reduced number of mitochondria. (G) Immunostaining for the cis-Golgi marker GM130 and for the ER marker KDEL shows that erebotic cells have very few of these organelles. S1 Data provides the source data used for all graphs. Scale bars, 5 μm (A, B), 1 μm (A’, A”, B’, B”), 20 μm (F, G). https://doi.org/10.1371/journal.pbio.3001586.g003

What is the function of erebosis? The loss of fundamental cellular components such as organelles, cytoskeleton, junctions, the nuclear membrane, and, possibly, DNA made us to imagine that this is some type of cell death. At least, it is difficult to imagine that erebotic cells are maintaining life with active metabolism. In general, there are 3 pathologically categorized types of cell death: apoptosis, necrosis, and autophagic cell death. We could not detect caspase activation in erebotic cells based on cleaved DCP1(Fig 4A). Caspase inhibition by p35, miRNAs for rpr, hid, and grim or dpf mutation did not suppress erebosis either (Figs 4B–4D and S3A–S3E). Clonal analyses of homozygous H99 or the Dronc null mutation, both of which suppress apoptosis [21–23], demonstrated that erebosis occurs in the mutant clones (S3F–S3H Fig). Erebotic cells did not have any feature that occurs with infection-induced cell shedding, such as up-regulation of upd2, or involvement of JNK or IMD pathways [6] (S4A–S4F Fig). Erebosis was also observed in a sterile condition, where flies were bleached when they were embryos and cultured with antibiotics (S4G–S4J Fig). A necrosis marker, propidium iodide, which can enter cells when the plasma membrane is breached, did not enter erebotic cells (Fig 4E). An autophagy marker (mCherry-Atg8a) did not label erebotic cells either (Fig 4F). Autophagy inhibition through knockdown of Atg genes did not suppress erebosis (S5 Fig).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Erebosis is unprecedented cell death. (A) Immunostaining for cleaved caspase 1 (cDCP1) does not show any increased signal in erebotic cells marked by absence of GFP fluorescence. (B, C) Erebotic cells indicated by absence of Myo1D-driven GFP and presence of Ance staining are visible even with suppression of apoptosis by expression of miRNA against rpr, hid, and grim. (D) Quantification of the percentage of erebotic cells with/without miRNA for rpr, hid, and grim. (E) Live imaging after PI feeding demonstrates that PI cannot enter erebotic cells. (F) Autophagosomes labeled by mCherry::Atg8a are reduced in erebotic (Ance+) cells. (G) Erebotic cells with a lower nuclear height and less GFP signals are labeled by TUNEL staining (arrowhead). (H) Correlation analysis of the nuclear height and nuclear GFP intensity demonstrates that cells at late erebosis (shorter nuclear height and weaker nuclear GFP signals) tend to be TUNEL positive. Pearson’ correlation coefficient (R) was calculated: R = 0.8098, R2 = 0.6558, P < 0.0001. (I) Erebotic cells have reduced amounts of cellular ATP labeled by BioTracker ATP-red, indicating their reduced metabolic activity. Note that the punctate signals of BioTracker ATP-red likely represent ATP in mitochondria. (J) Correlation analysis of ATP detected by BioTracker ATP-red and GFP intensity. Pearson’ correlation coefficient (R) was calculated: R = 0.6588, R2 = 0.4340, P < 0.0001. (K) Progenitor cells labeled by escargot-driven GFP (esg>GFP) are present in close proximity to Ance+ enterocytes. (L) Distance between erebotic cells and progenitor cells. (M) Immunostaining for Ance shows that a cytoplasmic legacy of an erebotic cell (arrowhead) is being pushed up by 2 enterocytes. The 2 enterocytes are considered to be relatively young because they have only GFP expression, not nlsRFP, and maturation of GFP is faster than of RFP. This captures a potential moment of the cell replacement. (N) Schematic of the erebosis process. Erebotic cells lose cytoskeleton, adhesion components, and organelles. Erebotic cells eventually become TUNEL positive, die, and are replaced by new enterocytes. Statistical significance was determined by using a two-tailed unpaired t test (D). S1 Data provides the source data used for all graphs and statistical analyses. Scale bars, 20 μm. EB, enteroblast; EC, enterocyte; ISC, intestinal stem cell; PI, propidium iodide. https://doi.org/10.1371/journal.pbio.3001586.g004

We investigated whether a general cell death marker TUNEL, which detects the DNA nick, labels erebotic cells. TUNEL is known to label not only apoptotic cells but also other cell death such as necrotic cells [24–26]. TUNEL did not label all erebotic cells, but labeled ones with the shorter nucleus height and lower GFP signals, which we found correlate (Fig 4G and 4H), suggesting that TUNEL labels cells at late erebosis. Inhibition of caspase-activated DNAse affected neither TUNEL nor erebosis (S6A–S6E Fig), consistent with the idea that erebosis is a process that is distinct from apoptosis. These observations suggest that erebosis is a gradual process toward cell death, which eventually leads to nicking and degradation of DNA. To further investigate the idea that erebosis is a mechanism toward cell death, we examined indispensable “house-keeping” molecules in addition to the features we already characterized in erebotic cells. We found that erebotic cells lose tubulin in addition to F-actin, and Fibrillarin, an important nucleolus component (S6F and S6G Fig). They also have a reduced amount of cellular ATP (Fig 4I and 4J). In sum, erebosis does not have any traditional feature of apoptosis, necrosis, or autophagy, and we speculate that it is a process toward cell death.

Since erebotic cells were only positively labeled with Ance, we investigated the role for Ance in erebosis. We confirmed that 2 RNAis driven by Myo1D-Gal4 decrease signals of Ance (S7A–S7D Fig). Interestingly, knockdown of Ance with these RNAis did not affect erebosis (S7A–S7C and S7E Fig). Ance mutation did not affect the gut cell turnover detected by the mitotic activity of ISCs (S7F Fig). Ance overexpression did not affect erebosis (S7G–S7J Fig). We found that Ance inhibition in enterocytes affects the amount of excreted waste that is labeled intensely with bromophenol blue (S7K–S7M Fig). Since Ance does not affect erebosis itself, we interpret that the effect of Ance inhibition on excretion is not due to erebosis, but due to another mechanism that Ance itself mediates. One interesting note on Ance is that its transcription detected by Ance-Gal4–mediated GFP was also observed in the cells surrounding erebotic cells (S7N and S7O Fig). Thus, non-erebotic, normal enterocytes also express some amount of Ance. This constitutive expression of Ance likely contributes to the weak signals of Ance at the apical side (S1E Fig). Thus, erebotic cells do not acutely start to transcribe Ance; rather, they accumulate Ance nontranscriptionally. We hypothesized that cells may incorporate secreted Ance during erebosis. To test this hypothesis, we made a signal peptide mutant of Ance, and in this mutant, we did not observe erebotic cells that accumulate Ance (S7P Fig). This supports the idea that accumulation of Ance in erebotic cells is not mediated by its up-regulation but is likely mediated by incorporation of secreted one. Nontranscriptional regulation of Ance in erebotic cells is consistent with the correlation between GFP loss and Ance accumulation, where GFP loss occurs very rapidly, too rapidly to correlate with a transcriptional change.

A fundamental question is what role erebosis plays in the gut tissue homeostasis. While we were examining features of erebotic cells, we noticed that there are often small, stem cell–like cells beneath (basal to the erebotic cells) or near erebotic cells (Fig 2C). We confirmed that progenitor cells labeled by esg>GFP often reside close to erebotic cells (Fig 4K and 4L). Careful examination of confocal images also showed that occasionally differentiated enterocytes labeled by Myo1D>GFP crawl beneath erebotic cells. In one example, there was, above GFP-positive cells, strong Ance staining without DAPI, which reflects a cytoplasmic legacy without DNA after erebosis (Fig 4M). Importantly, GFP-positive cells in these observations are relatively young enterocytes because of their low expression of RFP (nlsDsRed), which takes longer maturation time than GFP [27,28]. These findings lead us to speculate that erebosis might be involved in the cell turnover mechanism in the gut. In fact, contrary to what has been believed, global inhibition of apoptosis in enterocytes with either p35 or miRNAs for rpr, hid, and grim did not affect ISC mitosis based on PH3 (S8A and S8B Fig). p35 did not affect EdU staining in progenitors or the number of cells lineage-traced from ISCs (S8C–S8E Fig). dpf mutants did not show any change of the PH3-positive cell number either (S8F Fig). Taken together, these data imply that apoptosis may not play a major role in the gut cell turnover under the physiological condition.

[END]

[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001586

(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/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.