Cell Components and Features of Planula Larvae
At room temperature (about 25 °C), fertilized eggs of Acropora tenuis cleave and develop into prawn chip-shaped blastulae (15 h after fertilization), donut-like gastrulae (1 day after fertilization), pear-shaped embryos (3.5 days after fertilization; Fig. 1b), and planula larvae (5.5 days after fertilization; Fig. 1c). We maintained planula larvae more than 1 month in the laboratory (Fig. 1d).
In order to learn more about basic cellular components and morphogenesis of A. tenuis planula larvae, histological sections were prepared from 24-day-old and 57-day-old planula larvae (Fig. 2). Twenty-four-day-old larvae are beginning to form the oral diverticulum or pharynx while 57-day-old larvae have completed pharynx formation. A transverse section of a 24-day-old larva shows that the ectoderm consists of especially elongated cells (Fig. 2b, c). Endoderm was not observed beneath the ectoderm at this stage (Fig. 2c). A longitudinally cut surface shows a diverticulum hanging from the oral concavity (Fig. 2a). The diverticulum is continuous with invaginated ectoderm (Fig. 2d, f, g). Cells in the diverticulum are loosely associated (Fig. 2e). Some 24-day-old planula larvae were beginning to open the oral aperture (Fig. 2f, g). The diverticulum has a cavity at the center (Fig. 2g) and constituent cells of the diverticulum have approached the wall of the ectoderm to form an endodermal cell layer (Fig. 2g). Histological sections of 67-day-old planula larvae show that ectoderm is already underlain by endoderm, the presumptive gastroderm of a forthcoming polyp (Fig. 2i, j).
Cell Dissociation and the Primary Cell Culture
Forty to 50 larvae were collected 3, 5, 10, 12, 14, 16, and 24 days after fertilization. They were treated with a mixture of trypsin, EDTA, and collagenase (TEC). TEC treatment for 1–2 h at 25–28 °C did not complete cell dissociation (Fig. 3a, b, arrowheads), but after 3–4 h, only single cells without debris were found in the culture dish (Fig. 3c). Irrespective of larval ages, TEC efficiently dissociated cells, whereas TE was less effective, and collagenase alone had no apparent effect on cell dissociation (data not shown). The dissociated cell population contained several types of cells (Fig. 3c, d), including brilliant, brown-colored cells approximately 10 μm in diameter (Fig. 3c), translucent cells 8–10 μm in diameter (Fig. 3d), and small, pale blue cells (2–4 μm in diameter) that may be cell fragments rather than intact cells (Fig. 3d). Elongated cells characteristic of live ectoderm were scarcely found in the culture dish (Fig. 3a–d). We performed cell dissociation on 57-day-old and 61-day-old larvae as well. Cell populations dissociated from these stages contained a very few brown cells (Fig. 3e, f), as will be discussed later.
Dissociated cell suspension (0.5 mL) was dispensed in each well of a 24-well plate. Small cell aggregates appeared 4 to 7 days later (Fig. 3g). Cell aggregates increased in number and size until 2 weeks of incubation (Fig. 3h) and then stopped growing without any signs of cell death (Fig. 3i).
Effect of Plasmin on Cell Growth
Modular proteases were applied to the primary cell culture in the basal growth medium immediately after cell dissociation, because modular proteases such as tunicate retinoic acid-inducible modular protease (TRAMP) (Ohashi et al. 1999) and enterokinase (Kawamura et al. 1999) sometimes exhibit cell proliferation activity in invertebrate cell cultures. Enterokinase had no apparent effects on planula cell growth. In contrast, plasmin (2 μg/mL) was effective in maintaining dissociated cells in culture for more than 2 weeks (Fig. 4a, b), during which a large number of a new type of cell appeared in the culture dish (Fig. 4a, b). The new type of cell looked dark and had an amorphous shape (Fig. 4c, yellow arrowhead). After 2 to 3 weeks of culture, brown cells extended filopodia (Fig. 4c, white arrowheads) and dark cells developed lamellipodia (Fig. 4e, black arrowhead). The intermediate type of cell having brown-colored cell body (Fig. 4d, thick arrow) and elongated lamellipodium (Fig. 4d, thin arrow) was often observed, suggesting that brown cells differentiate in vitro into dark cells. Brown cells formed loose cell aggregates similar to the oral diverticulum of planula larvae (Fig. 4f). The aggregate is easily dissociated into single cells by pipetting.
For subsequent cell culture, aliquots of polyclonal cell population were harvested from the primary 24-well culture plate, diluted fivefold with growth medium containing plasmin, and dispensed into 96-well plates (100 μL/well). Cells in clumps (Fig. 5a) proliferated with a doubling time of 2–3 days (Fig. 5b–d). For quantitative analysis of cell growth, the scattering type of cell was inoculated in a multiplate at a cell density of 1.5 × 104 cells/mL (Fig. 5e). Following an initial stagnation of cell growth, cells approximately doubled their original cell density every 2 days from 4 to 8 days of cell culture (Fig. 5f, g). Then, cell growth slowed, and they doubled in 4 days from 8 to 12 days of cell culture (Fig. 5g). In all, cells proliferated eightfold after 12 days of culture (Fig. 5f, g).
Making Cell Lines
From 3 weeks onward, some cell aggregates formed cell sheets (Fig. 6a, f). Other cell aggregates formed spheres (Fig. 6g) detached from the substratum. They resembled blastulas (Fig. 6g), and in the most extreme cases, gastrula-like aggregates appeared (Fig. 6h). Blastula-like and gastrula-like aggregates were returned to seawater, but they did not swim and never developed into planula larvae.
Therefore, so far, we have attempted to make several A. tenuis cell lines with different features. Eight representative lines and their features are listed in Table 1 and Fig. 6. Lines IVB6 and IIC5 are brilliant brown cells, which exist as single cells (Fig. 6b) or clusters (Fig. 6a). Lines IVB4 (Fig. 6c), IVC4 (Fig. 6d), and IVD1 (Fig. 6e) are dark, amorphous, flattened cells. IVC6 (Fig. 6f) has a propensity to form cell sheets. As mentioned above, IIID5 forms blastula-like cell clusters (Fig. 6h) and IIIB6 develops into gastrula-like cell clusters (Fig. 6g). These cell lines have been maintained for more than 8 months by replating them more than eight times (Table 1). In addition, the cells are cryo-preservable in liquid nitrogen. The cell lines are distributable to academic researchers upon request to Dr. Kaz Kawamura at Kochi University.
Immunochemical Proof of A. tenuis Cell Lines
We were confident that the established cell-lines originated from Acropora tenuis planula larvae. However, there was still a possibility that these cells represented contamination from unknown sources. To examine this possibility, we made two antibodies, anti-AtSnail and anti-AtFat1 (see Methods) and carried out Western blotting and immunohistochemistry using these antibodies. We first carried out Western blotting analysis against total proteins isolated from planula larvae or cultured cells and confirmed the specificity of the antibodies to A. tenuis. The analysis using anti-AtSnail antibody resulted in a band of approximately 30 kDa protein (Supplementary Fig. S1a, lanes 1, 2, asterisk), corresponding to the expected MW of AtSnail (29,813 Da). AtFat1 is expected to be an extraordinarily large protein of approximately 400 kDa. It is impossible to analyze this large protein by means of our SDS-PAGE system. In the present study, the anti-AtFat1 antibody reacted with a protein of approximately 200 kDa and with several minor polypeptides extracted from planula larvae (Supplementary Fig. S1b, lane 1, asterisk). The antibody did not stain any specific bands extracted from cultured cells (Supplementary Fig. S1b, lane 2). The results of anti-AtSnail antibody indicate that the antibody is specific for A. tenuis proteins.
Then, we carried out immunocytochemistry of 24-day-old and 67-day-old planula larvae. Nuclei of both ectoderm and endoderm were stained with anti-AtSnail antibody (Fig. 2e, h, k), and anti-AtFat1 stained cell bodies (data not shown).
In vitro cultured cells (Fig. 7a) were treated with TE for 5 min. All cells were dissociated from the culture dish and assumed spherical shapes (Fig. 7b). When the spherical cells were inoculated again into cell growth medium containing plasmin, they adhered to the substratum within 12 h (Fig. 7c). This feature of cells was utilized to stick them to coverslips. Cultured cells on the coverslip were stained with anti-AtSnail antibody (Fig. 7d–f). Immunostaining was not restricted to the nucleus (Fig. 7e) but also occurred in the cytoplasm. Anti-AtFat1 antibody was also used to stain cell aggregates mentioned above (Fig. 7f). The fluorescent signal appeared to emit from the whole cell aggregate (Fig. 7f). These results indicate that cells are derived from Acropora tenuis.
Molecular Characterization of Cell Lines
We carried out RNA-seq analyses of gene expression profiles of the eight cell lines listed in Table 1 and Fig. 6. Single-cell RNA-seq was performed on lines IVB6, IIC5, IVB4, IVC4, and IVD1 or a single-cell cluster of lines IVC6, IIID5, and IIIB6. For each cell line, the Illumina Novaseq platform yielded ~ 30 Gb of reads. Genome and transcriptome data of A. tenuis were downloaded from https://marinegenomics.oist.jp/acropora_tenuis/viewer/info?project_id=97. Single-cell RNA-seq reads were mapped to A. tenuis mRNA data using Salmon software (version 0.8.2) (Pertea et al. 2016). Data from this analysis are shown in Supplementary Table S1. Raw data of RNA-seq were deposited in GenBank and are accessible under BioProject ID, PSUB014043.
As a result, 902 mRNAs with corresponding gene models were identified, 676 of which were annotated with sequence similarities to those of other metazoans, while the remaining 226 resulted in no annotation. This confirmed the results of immunocytochemistry that the in vitro cell lines belong to Acropora tenuis and did not result from contamination.
Of 676 genes, 36 were expressed in all eight cell lines and 640 genes were expressed preferentially and/or specifically in a certain line. Further characterization showed that 54 genes were specifically expressed in IVB6, 71 gene in IIC5, 61 in IVB4, 120 in IVC4, 53 in IVD1, 146 in IVC6, 38 in IIID5, and 31 to IIIB6 (Supplementary Table 2). This indicates that these cell lines are not identical as far as can be judged from gene expression profiles.
Gene Ontology Profiles
Gene Ontology was applied to infer the properties of the in vitro cells. Four hundred sixty-eight of 676 genes were categorized under “cellular component,” “biological function,” and “molecular function,” and properties of the lines were inferred from these profiles. All eight lines exhibited similar GO patterns. That is, genes with molecular function were most abundant, genes with biological function next, and genes with cellular component were least abundant (Supplementary Fig. 2). Partly due to the single-cell-level analysis, the GO profile analysis did not always yield useful information to characterize the properties of each cell line.
Characterizing Properties of Cell Lines with Marker Genes
In Nematostella vectensis, another cnidarian, Sebe-Pedros et al. (2017) characterized nine cell types of larvae using 28 marker genes, including collagen and ferritin for markers of gastroderm; plac8 and zona pellucida for the epidermis; synaptotagmin, elav, rpamide, synapsin, and eag ion channel for neurons; Shaw ion channel for larval neurons; FGF1a and Frizzled for larval apical organ; nanos1, myc, neuroD, histone H1, CENP, soxB2a and septin for progenitors/undifferentiated cells; aquaporin for Dig filaments; carboxypeptidase A, trypsin, disintegrin, mucin, and chitin deacetylase for glandular/secretory cells; and spinalin, minicollagen, and nematogalectin for cnidocytes. Employing these features as criteria, we examined whether the eight lines of in vitro cells express marker genes, and if so, which lines express which markers. We found specific and shared expression profiles of several marker genes in the cell lines (Table 2; Table S2).
First, IVB6 and IIC, brilliant brown cells, expressed marker genes for gastroderm and those for glandular/secretory cells (Table 2). However, they likely have different properties since the former expressed markers for neurons, while the latter produced markers for progenitors and epidermis. Expression of two neuron-related genes, neuronal calcium sensor 1 and neurogenic locus Notch, was detected in IVB6 (Table S2).
Second, dark cell lines, IVB4, IVC4, and IVD1 expressed markers of gastroderm and glandular/secretory cells at high levels (Table 2). In addition, IVB4 cells also expressed myc (undifferentiation marker) (Table 2) and a photopigment, melanopsin-A abundantly (Table S2). On the other hand, IVC4 expressed neuronal markers (Table 1) and melanotransferrin (Table S2), whereas IVD1 expressed an epidermis marker (Table 1). Therefore, although all three lines have properties of gastroderm and glandular/secretory cells, each line has different properties as well.
Third, a gene for Shaw ion channel (larval neuron marker) was expressed only in IVC6 (cell-sheet) (Table 2). IVC6 cells also expressed synaptotagmin (neuron marker), frizzled (larval apical organ marker), and Notch (Table S2), suggesting that this cell line has neuronal properties. On the other hand, IVC6 expressed other genes, including those for collagen (gastroderm marker), disintegrin and trypsin (secretory cell markers), and also myc (undifferentiated cell marker). These results indicate that IVC6 comprises various cell types, but among the eight lines, these are the only cells with neuronal markers.
Fourth, Plac8, a novel placenta-specific gene (GenBank Accession, DC643839) that is used as an epidermis marker in Nematostella, is expressed only in IIIB6 (Table 2). IIIB6 is a gastrula-like cell cluster, in which an epidermis-like outer layer covers the cluster. The expression of plac8 in the line indicates that the outer-layer cells of IIIB6 have properties of epidermis. IIIB6 also expresses a very low level of collagen genes, suggesting properties of gastroderm (Table 2). On the other hand, marker genes expressed in IIID5 were different from those of IIIB6. IIID5 expressed a marker (collagen) of gastroderm, markers (trypsin, disintegrin and mucin) of glandular/secretory cells, and myc of undifferentiated cells (Table 2). Therefore, IIID5 is distinct from IIIB6, and has the potential to form several cell types. All cell properties mentioned above were stably maintained throughout successive cell cultures.
Although further characterization is required for each of the eight lines, gene expression profiles at the single-cell level indicate that each has its own properties.