Tissue homeostasis in sponges: quantitative analysis of cell proliferation and apoptosis

Background Tissues of multicellular animals are maintained due to a tight balance between cell proliferation and programmed cell death. Phylum Porifera is an early branching group of metazoans essential to understanding the key mechanisms of tissue homeostasis. This paper is dedicated to the comparative analysis of proliferation and apoptosis in intact tissues of two sponges belonging to distinct Porifera lineages, Halisarca dujardinii (class Demospongiae) and Leucosolenia variabilis (class Calcarea). Results Labeled nucleotides EdU and anti-phosphorylated histone 3 antibodies reveal a considerable number of cycling cells in intact tissues of both species. The main type of cycling cells are choanocytes - flagellated cells of the aquiferous system. The rate of proliferation remains constant in areas containing choanoderm. Cell cycle distribution assessed by the quantitative DNA stain reveals the classic cell cycle distribution curve. During EdU pulse-chase experiments conducted in H. dujardinii, the contribution of the choanocytes to the total amount of EdU-positive cells decreases, while contribution of the mesohyl cells increases. These findings could indicate that the proliferation of the choanocytes is not solely limited to the renewal of the choanoderm, and that choanocytes may participate in the general cell turnover through migration. The number of apoptotic cells in intact tissues of both species is insignificant. In vivo studies in both species with TMRE and CellEvent Caspase-3/7 indicate that apoptosis might be independent of mitochondrial outer membrane permeabilization. Conclusions A combination of confocal laser scanning microscopy and flow cytometry provides a quantitative description of cell turnover in intact sponge tissues. Intact tissues of H. dujardinii (Demospongiae) and L. variabilis (Calcarea) are highly proliferative, indicating either high rates of growth or cell turnover. Although the number of apoptotic cells is low, apoptosis could still be involved in the regular cell turnover.

Morphogenetic processes underlying the development, growth and regeneration of multicellular 3 animals have long been issues of scientific interest. These phenomena seem to be closely related to a 4 process of physiological regeneration, also known as cell turnover 1,2 . Cell turnover consists of three key 5 elements: 1) the elimination of old, non-functional cells by programmed cell death; 2) the proliferation 6 of somatic stem cells; and 3) the differentiation of newly generated cells along with their integration with 7 preexisting tissue 3 . Physiological regeneration through tissue renewal is an essential process for 8 multicellular animals, as it allows reparation of minor defects and tissue damage occurring during 9 normal functioning 1,4 . Abnormalities of tissue homeostasis are associated with numerous malfunctions, 10 including oncological and degenerative disorders 4,5 .

11
Sponges (phylum Porifera) are an early-branching metazoan group comprising aquatic, mostly 12 filter-feeding, sedentary organisms 6 . These animals differ substantially from other metazoans in their 13 body organization, lacking tissues and organ systems characteristic of other animals 7, 8 . The central 14 anatomical structure in a typical sponge body is its aquiferous system, a complex network of canals and 15 chambers through which the sponge pumps surrounding water for food and oxygen uptake 9 . The rest 16 of the sponge body is composed of the mesohyl, a tissue consisting of an extensive extracellular matrix 17 and numerous wandering cells of different type and function 9 . 18 Sponge tissues appear to be less specialized than tissues of Eumetazoa. In particular, sponge 19 tissues are multifunctional and highly plastic, as their cells show broad capabilities of transdifferentiation 20 into other cell types 10,11 . The tissue plasticity underlays many aspects of sponge biology, participating 21 in constant reorganization of the aquiferous system 10 , sexual and asexual reproduction 8 , 22 regeneration [12][13][14][15][16][17] and even movement 18-20 . 23 At the same time, sponges lack a single category of somatic stem cells. Choanocytes 24 (flagellated cells of the aquiferous system) and some amoeboid cells of the mesohyl (e.g., archaeocytes 25 of Demospongiae) are both considered to be stem cells in sponges 8,21,22 . It has been shown for several 26 species that choanocytes and archaeocytes proliferate in intact tissues, (trans)differentiate into other 27 SubG1 population consisted of small and highly granulated objects -possibly, dying cells and debris 1 ( Figure 8C). 2 CMFSW-E suspensions of L. variabilis did not provide reliable data: they contained too many 3 SubG1 and too few G2/M cells to obtain an accurate cell cycle distribution histogram and were ultimately 4 excluded from further analysis. We therefore used suspensions made from pre-fixed tissue of L. 5 variabilis. This method provided improved resolution: suspensions contained fewer SubG1 cells and 6 showed well-distinguishable G0/G1 and G2/M peaks. The general appearance of the histogram was 7 similar to H. dujardinii ( Figure 8В). The proportion of G0/G1 cells was 53.88±2.59%, 12.58±0.76% for 8 S cells and 4.63±1.79% for G2/M cells (Table 3). 9  (Table 4). 16

Investigation of apoptotic activity
Both choanocytes and mesohyl cells were present among the apoptotic cells. The endosome of H. 17 dujardinii showed even less apoptotic activity; for every 6-9 thousand normal cells, there were usually 18 1-2 CellEvent-positive ones (Table 4). 19 We also studied suspensions obtained via mechanical dissociation in CMFSW-E. They were 20 stained with CellEvent, fixed, and analyzed via a flow cytometer. Cells of H. dujardinii formed a single 21 peak with low/negative CellEvent fluorescence ( Figure 10А, B). The fraction of apoptotic cells was low, 22 with CellEvent-positive cells accounting only for 1±1.15% of the total population. In contrast, cells of L. 23 variabilis were clearly divided into two subpopulations by the level of CellEvent fluorescence ( Figure  24 10C, D). CellEvent-positive cells with elevated autofluorescence (as evidenced by increased short-25 wave fluorescence in the PacificBlue channel) comprised 60.8±13.46% of the total cell population 26 ( Figure 10C, D). We consider this result as a sampling artifact (see "Discussion" section) 27 Table 5). All CellEvent-positive cells (Q2 and Q3) were small and highly granulated, as indicated by 21 light scattering parameters ( Figure 11B). Population of TMRE-CE-cells was located in Q4 and 22 accounted for 1.67±0.35% in H. dujardinii and 13.70±3.25% in L. variabilis ( Figure 11C, D; Table 5). 23 While these cells were smaller than TMRE+ CE-cells, they tended to separate from TMRE+ CE+ cells 24 on scatter plots ( Figure 11B). 25

26
In multicellular animals, the long-term maintenance of tissue homeostasis is achieved by cell 27 turnover 1,4 . Cell turnover is composed of three complementary processes: cell proliferation, cell 28 differentiation and programmed cell death 3 . Sponges are basal metazoans well known for their unique 1 tissue dynamics. In this study, we mainly focused on cell proliferation and apoptosis in intact tissues of 2 sponges, providing promising data on cell turnover in sponges. 3 proportion of pH3-positive cells remained nearly the same. This body region also differed in the 1 proportion of choanocytes among the pH3-positive cells, with nearly half of the G2/M cells belonging to 2 the mesohyl/pinacoderm layer. This could be evidence of potential growth since new sclerocytes, 3 pinacocytes and porocytes are obviously needed to lengthen the diverticula. In the oscular rim, which 4 is devoid of choanocytes, the intensity of cell proliferation was very low. There are, however, a few 5 labeled cells, indicating either slow cell turnover or growth. 6

Cell proliferation in intact tissues of Halisarca dujardinii and
This study provides the first spatial analysis of cell proliferation across different body regions of 7 sponges representing distinct phylogenetic lines and types of aquiferous systems. In general, 8 proliferation activity appears to be evenly distributed throughout the tissues of sponges (excluding 9 regions that do not contain choanocytes). The proportion of cycling cells in H. dujardinii is lower than 10 that of other Demosponges, in which this parameter reaches 20-25% on average 34,38 . Among 11 calcareous sponges, the proportion of cycling cells in L. variabilis is slightly higher than in Sycon 12 coactum 38 . It should also be noted that during our regeneration studies, the proportion of EdU-labeled 13 cells in L. variabilis was overestimated when compared to current data, and is likely an artifact of the 14 counting technique 12 . Similar to the previous studies, we detected that DNA synthesis occurs primarily 15 in the choanocytes; moreover, we show that the distribution of mitotic cells in adult sponges have an 16 overall similar pattern. The proportion of proliferating cells in H. dujardinii and L. variabilis strongly 17 resembles those animals characterized by fast cell turnover, including platyhelminthes, cnidarians and 18 bivalves 39-42 . 19 The contribution of choanocytes to homeostasis 20 The physiological role of proliferation in intact tissues is either growth or renewal 3 . In order to 21 understand the general dynamics of sponge tissue, we need to figure out the fate of proliferating cells . 22 In intact sponges, cell proliferation takes place in the choanoderm. Since the choanoderm is 23 metabolically active and directly exposed to the environment, we should expect higher rates of 24 choanocyte turnover 4 . Similar to feeding structures of other suspension feeders, constant replacement 25 of old and damaged cells is crucial for the long-term maintenance of the choanoderm [42][43][44][45] . Despite 26 constantly facing a potentially hostile environment (e.g., toxic and mutagenic factors dissolved in 27 seawater), choanocytes undergo constant proliferation and represent the main source of new cells in 28 the choanoderm. It should be noted here that the proliferation of morphologically differentiated cells is 1 fairly common among metazoans, including vertebrates 3,46-49 . Cnidarians are particularly interesting 2 given that ecto-and endodermal epithelial cells act not only as barrier and contractile cells, but also 3 retain multipotency and the ability to proliferate 50 . 4 In fact, the boundary position of choanocytes might be advantageous since cell division requires 5 significant amount of resources. As the primary food-entrapping sponge cells, choanocytes are provided 6 with sufficient energy for constant proliferation and are able to respond appropriately to environmental 7 changes. Moreover, since the choanoderm is obviously enriched by cell-cell interaction, and at least 8 partially compartmentalized by extracellular matrix 18 , it could maintain the proliferative potential of 9 choanocytes through some signaling system resembling that of the stem cell niche. In some animals 10 (e.g., Hydra and planarians), the concept of 'niche' may be interpreted in a rather broad sense and 11 include a whole organ or tissue as a stem cell niche 69,70 . The choanoderm in sponges might be just an 12 example of such a "tissue niche". 13 Since the flagellum is resorbed during cell division, the filtration process becomes impaired. To Therefore, choanocytes are apparently the major part of cycling cells not only in the calcareous 5 sponge L. variabilis, but also in the demosponge H. dujardinii. It seems that the proliferation of 6 choanocytes is not only limited to the renewal of the choanoderm, as it has been previously shown that 7 choanocytes contribute to gametogenesis 8,53-55 , regeneration 12,14,16 and can express certain stem cell 8 markers (at least in demosponges and homoscleromphs) 26,28,56,57 . Additionally, it seems that 9 choanocytes may also participate in cell turnover through cell migration. Together, this data suggests 10 that the role of choanocytes in the stem cell system of sponges is underestimated. Choanocytes might 11 act as a major source of new cells (including cells of mesohyl) in both calcareous sponges and 12 demosponges during regular cell turnover or growth. 13 14 Under the steady-state condition, cell proliferation is counterbalanced by the process of cell 15 elimination. Apoptosis is considered to be the most common way of cell death in animals 3,32,58 . It is a 16 tightly regulated and complex process resulting in the fragmentation of a dying cell into apoptotic 17 bodies 31 . Once these apoptotic bodies are engulfed by other cells, neat cell death is reached. It should 18 be noted that tissues exposed to adverse environments usually display high rates of physiological 19 turnover (e.g., mammalian epidermis and intestine; gills, intestine and mantles of mollusks; epithelia of 20 ascidians and cnidarians) 42,58-62 . Given that choanocytes are highly proliferative, we would expect a 21 significant amount of dying cells. 22

Apoptotic activity in intact tissues of sponges
CLSM studies using CellEvent staining indicated that apoptotic rate is rather low in intact 23 tissues of H. dujardinii and L. variabilis, with less than 1% of cells containing active effector caspases. 24 We consider this result to be the most relevant, as it fits well with other known studies on sponges 16,33,34 . 25 Insignificant number of apoptotic cells was initially thought to indicate the possibility of caspase-26 independent cell death in intact sponges. However, low number of apoptotic cells is also characteristic 27 of tissues of other animals under the steady-state condition 63,64 . Thus, we suggest that apoptosis is 1 involved in regular tissue renewal in sponges. This situation is typical for mammalian cells since the permeabilization of the inner mitochondrial 10 membrane is shown to be tightly associated with the mitochondrial outer membrane permeabilization 11 (MOMP) 63 . In turn, MOMP is the critical event during the apoptotic response, resulting in the release of 12 some pro-apoptotic proteins from the mitochondrial intermembrane space into the cytoplasm. The most 13 crucial pro-apoptotic protein is cytochrome c, which is able to bind to cytoplasmic protein APAF1, variabilis maintained functional mitochondria during apoptosis. Thus, there is a possibility that the 29 apoptosis in sponge cells is MOMP-independent. Therefore, and in contrast to mammalian cells, the 1 TMRE-CE-population seen in these two sponges may contain not early apoptotic cells as originally 2 thought, but represents a population of damaged or dead cells and cell debris. Rare TMRE-CE+ cells 3 might be interpreted as non-viable, late apoptotic cells with non-functional mitochondria. 4 Histological studies of intact tissues 8 For histological studies, sponge tissues were fixed for 2 hours at 4 °C with 2.5% glutaraldehyde 9

Conclusions
(Electron Microscopy Science, 16020) and post-fixed for 1 hour at room temperature (RT) with 1% OsO4 10 (Electron Microscopy Science, 19100). Fixation and post-fixation were done in modified 0.1M Na-11 Cacodylate buffer (0.1M Na-Cacodylate, 85.55 mM NaCl, 5 mM CaCl2, 5 mM MgCl2; рН 7.0-7.5). After 12 the post-fixation, specimens were dehydrated and embedded in Epon/Araldite epoxy embedding media 13 (Electron Microscopy Science, 13940) according to the standard protocol previously described 12,75 . 14 Semi-thin sections (1 µm) were cut using LKB V and Leica UC6 ultramicrotomes. Sections were 15 stained with 1% toluidine blue -0.2% methylene blue mixture for 1-1.5 min at 60 °C and examined with 16 a Carl Zeiss Axioplan 2 microscope (Carl Zeiss) equipped with an AxioCam HRm (Carl Zeiss) digital 17 camera and AxioVision 3.1 (Carl Zeiss) software. 18 19 We  23 Intact H. dujardinii individuals were incubated in 5 mL of FSW with 200 μМ EdU for 12 hours 24 and then transferred into a running seawater aquarium. Sponges incubated in 5 mL of FSW with 100 25 µL of DMSO were used as negative controls. Sponges were fixed with 4% PFA PBS on day 1, 5 and 7 26 after being transferred into aquariums (3, 3 and 5 individuals, respectively). Individuals fixed 27 immediately after the EdU incubation served as the initial time point, day 0. All specimens were treated 1 for CLSM studies as described above (see "Study of cell proliferation with CLSM" section), but no 2 antibody staining was performed in this experiment. We studied areas of the endosome at some 3 distance from an oscular tube. One randomly chosen field of view within the region was analyzed in 4 each individual. For subsequent analyses, a Z-stack was obtained from each field of view. We estimated 5 the fractions of choanocytes and mesohyl cells among all EdU-positive cells (as a proportion of EdU-6 positive choanocytes/mesohyl cells to all EdU-positive cells). As a result, the main parameter of the 7 analysis was the mean proportion of choanocytes and mesohyl cells in the total number of EdU-positive 8 cells. Details of data acquisition, nuclei counting and further analysis have been described above (see 9 "Study of cell proliferation with CLSM" section). In this experiment, choanocyte chambers were identified 10 by their characteristic shape visualized by DAPI staining. 11 12 We combined calcium and magnesium-free seawater (20 mM KCl, 300 mM NaCl, 10 mM Tris-13

Evaluation of cell cycle distribution
HCl and 15 mM EDTA in distilled water) (CMFSW-E) and mechanical tissue dissociation to obtain         Each proportion is given as mean value and SD. 4 Each proportion is given as mean value and SD. Three individuals per species have been examined; n 6 is the number of analyzed stacks. 7 The fraction of each population is represented by mean value and SD. N is the number of analyzed 1 individuals. 2