Ctenophore immune cells undergo ETosis in response to pathogen exposure
To assess cellular immune responses to microbial challenge in vitro, we isolated live cells from whole ctenophore preparations (Fig. 1B). After incubation with fluorescent Escherichia coli, we observed motile, stellate cells competent for phagocytosing large amounts of bacteria (Supp. Video 1; Supp. Figure 1). We further observed that some stellate cells changed their morphology dramatically by retracting their processes, undergoing nuclear rotation, and subsequently rapidly extruding nuclear material (Fig. 1C; Supp. Video 2). This behavior is remarkably similar to that of vertebrate monocytes during the cytoskeletal rearrangements preceding extracellular DNA trap (ET) formation 10,24. This led us to speculate that some ctenophore immune cells were producing ETs in response to the presence of microbes.
We examined microbially-challenged Mnemiopsis cells using confocal microscopy and observed cells with decondensed DNA cast in large areas surrounding individual cell bodies. These networks of extracellular DNA were closely associated with individual E. coli (Fig. 1D). Three-dimensional rendering of confocal z-stacks revealed that E. coli bacteria were entangled in the extruded Mnemiopsis DNA (Fig. 1E; Supp. Video 3). To further characterize the extruded DNA using immunofluorescence, we stained cells with an antibody that recognizes an array of histone proteins (H1, H2A, H2B, H3, H4). We observed that Mnemiopsis immune cell ETs are composed of chromatin, typical of ETs described in other taxa (Fig. 1F). ETotic cells show diffuse staining over a large area, whereas non-ETotic cells display intact nuclei with stereotypical complements of concentrated DNA and histone labeling (Fig. 1D-F; 9). These data identify specific ctenophore cell types that produce bona fide extracellular chromatin traps when exposed to a microbial signature.
Diverse microbial signatures induce ET formation in Mnemiopsis leidyi independently of Non-ETotic cell death
To simultaneously and accurately quantify ETosis and non-ETotic cell death we developed a semi-automated imaging pipeline using a combination of two DNA dyes: Hoechst, a membrane-permeable stain which labels all cell nuclei, and SytoxGreen, a non-permeable stain, which selectively labels nuclei of dying or dead cells with compromised cell membranes. Both dyes label extracellular DNA nets (Fig. 2). Importantly, the nuclear envelopes of necrotic and apoptotic cells remain relatively intact, displaying concentrated fluorescent labeling 25. Using our novel image analysis pipeline, we performed automated comparisons of thousands of images to accurately calculate percentages of live, dead, or ETotic cells. Critically, the development of this pipeline allowed us to accurately identify and discriminate between ETosis and non-ETotic cell death following treatments.
After exposing isolated Mnemiopsis cells to E. coli (Fig. 2A, B), we analyzed the stained nuclei in each image. Using image segmentation analysis, we defined individual cell masks based on Hoechst signal and measured relative fluorescence intensities of Hoechst and SytoxGreen inside each cell mask (Fig. 2C). Using those measurements, we defined 3 distinct cell populations: live cells, ETotic cells and non-ETotic cell death (Fig. 2D). Live cells are negative for SytoxGreen fluorescence with no dispersion of Hoechst signals (Hoechsthigh/SytoxGreenlow) (Fig. 2B, 2D). Cells that have ETotic nuclei exhibit high dispersion of Hoechst fluorescence associated with extracellular DNA net formation (Hoechstlow). In contrast, cells that have condensed Hoechst fluorescence and high SytoxGreen fluorescent signals (Hoechsthigh/SytoxGreenhigh) are dying from non-ETotic cell death processes.This approach allowed us to accurately identify and quantify large numbers of cells, including those undergoing ETosis after incubation with E. coli.
We then expanded our analyses of ET production to include responses to additional microbial stimuli ( https://github.com/carolinestefani/ETosis-and-death-automated-pipeline). We examined whether Mnemiopsis immune cells undergo ETosis when exposed to a diverse array of microbes, including the following: heat-killed gram-negative bacteria E. coli, heat-killed gram-positive bacteria Staphylococcus aureus, and cell wall extract from yeast Saccharomyces cerevisiae (zymosan). ETotic cells display diffuse Hoechst signal characteristic of filamentous DNA nets following exposure to E. coli, S. aureus, or zymosan (Fig. 3A). In contrast non-ETotic cells maintain intact nuclear material with concentrated Hoechst labeling. We analyzed both ETosis and total cell death in Mnemiopsis cells after four hours of microbial exposure (N = 24 individual animals). Contour plots representing Hoechst and SytoxGreen signal intensities show three distinct clusters: ETotic cells, dying cells, and living cells (Fig. 3B).
ETosis events significantly increased in Mnemiopsis cells when exposed to zymosan or E. coli in vitro, with mean increases in detected ETotic events of 17% and 7% over untreated cells (Fig. 3C). Incubation with S. aureus also elicited a strong ETotic response in Mnemiopsis cells, with a mean increase in detected ETotic events of 69% over untreated cells. Notably, non-ETotic cell death increased significantly after incubation with E. coli but did not change after exposure to zymosan or S. aureus (Fig. 3C). Our data demonstrate that a panel of microbes stimulates ETosis in ctenophore cells and that we can accurately and efficiently measure Etosis and non-ETotic cell death events simultaneously.
Extracellular trap formation in Mnemiopsis is induced by classic pharmacological stimuli
ETosis can be induced in vertebrate granulocytes and leukocytes with pharmacological agents that activate production of ROS via discrete intracellular signaling pathways 10,11,16. Though ctenophore genomes appear to lack predicted gene homologs to many classic innate immunity pathway cell surface receptors and signaling intermediaries, many components of metazoan stress response pathways, as well as secondary messenger machinery, are present 20,26,27. We hypothesized that ETosis in Mnemiopsis immune cells could be stimulated by classic chemical agents broadly used in studies of vertebrate ETs.
We observed a significant induction of ETosis in Mnemiopsis after a four-hour exposure to PMA, with a mean increase of 4.3% over untreated control cells (Fig. 3D-F). Nigericin, a potassium ionophore, induces ET formation in vertebrates by initiating ROS release from mitochondria 10,28. Initiation of a calcium influx following stimulation by calcium ionophore A23817 (calcimycin) can also induce ET formation independent of NOX 16. We observed a significant induction of ETosis following incubation of Mnemiopsis cells with both the potassium ionophore nigericin and the calcium ionophore A23187, with a mean increase of 33% and 20% over untreated control cells, respectively (Fig. 3D-F). Nigericin exposure elicited the highest amount of ET formation (Fig. 3F; Supp. Video 4; Supp. Figure 2).
While we did not observe a significant change in levels of non-ETotic cell death following PMA stimulus, our results did indicate a significant increase in non-ETotic cell death following nigericin and A23187 treatments (Fig. 3F). Exposure of Mnemiopsis cells to all three classical pharmacological stimuli – PMA, the potassium ionophore nigericin, and the calcium ionophore A23187 – induced significant ET formation. These results implicate diverse signaling pathways involved in production of ETs that also retain deep evolutionary conservation across extant metazoan phyla.
ETosis in the model bivalve, Crassostrea gigas, is induced by diverse microbial and chemical stimuli
We sought another non-vertebrate model to assess conservation of ET formation across phyla because ET stimulation in Mnemiopsis immune cells is intriguingly similar to ETosis in mammalian systems. The formation of invertebrate extracellular DNA traps has been characterized most extensively in molluscs 3–5, 13,29. Bivalve molluscs, like the Pacific oyster Crassostrea gigas, have blood-like circulatory cells, collectively called hemocytes, that have immune functions 30. However, prior attempts to stimulate production of ETs in oyster via exposure to pathogen signatures and PMA has varied between studies 5. For example previous studies observed a robust ETotic response following challenge with Vibrio, a virulent bivalve pathogen, in oyster hemocytes 3, however the induction of ETosis with other microbes has been only modest or not observed in previous studies 5. Thus, conservation of ET induction pathways in bivalves remains unclear 3–5.
We measured ETs and cell death in Crassostrea hemocytes after exposure to zymosan, E. coli, and S. aureus (Fig. 4A-D). We found that ETosis is significantly stimulated in hemocytes exposed to all three pathogens (Fig. 4D; Supp. Figure 3), with mean increases of 13%, 19%, and 22% over untreated hemocytes, respectively. Hemocyte non-ETotic cell death (apoptosis, necrosis) increased significantly with E. coli exposure but not for zymosan challenge (Fig. 4D).
We also assayed Crassostrea hemocytes for ET production after exposure to pharmacological reagents that engage distinct signaling pathways in vertebrate monocytes. We analyzed the behavior of Crassostrea hemocytes after four-hour incubation with PMA, the potassium ionophore nigericin, and the calcium ionophore A23187 (Fig. 4E). Exposure to PMA and the calcium ionophore A23187 induced significant ET production in isolated Crassostrea hemocytes (Fig. 4F, 4G). However, the potassium ionophore nigericin showed no significant induction (Fig. 4F, 4G). In contrast, a large proportion of Crassostrea hemocytes produce ETs after 4 hours of exposure to A23187, similar to prior studies 5.