In vitro studies in Tubularia: a morphological role for digestive cells

Cnidarians have often provided the experimental system for studying the morphological role of cell layers, as well as control mechanisms involved in the expression of a differentiated state (Tardent, 1963). Previous studies, however, have been restricted by a single limitation, the absence of in vitro investigations. This limitation has resulted in incomplete descriptions of cellular migration in these organisms, the morphological role(s) of the cell layers, and the differentiation sequence of the pluripotent interstitial cells.

Extended maintenance of cnidarian cells in a totally defined medium has not been previously reported. Sanyal & Mookerjee (1960) studied Hydra cells in vitro using pond water as a culture medium. The cells, however, lived only a few hours. Phillips (1961) using edamine, an enzymic digest of lactalbumin, succeeded in keeping Anemone cells viable for at least 12 days. In this system, however, all the cells appeared morphologically identical.

Martin & Tardent (1963), employing ‘slow-to-coagulate’ lobster blood, successfully cultured Tubularia cells. The major drawbacks to this study were the complex culture methodology and the use of lobster blood. Since no serum is found in situ in cnidarians, investigators of cellular differentiation or the morphological role of cell layers normally desire a medium without serum. An additional advantage to a medium lacking serum is the opportunity to develop a system to test chemical parameters of differentiation directly, without the possibility of serum interacting with the normal differentiation process.

Burnett, Ruffing, Zongker & Necco (1968) employed a chemically defined medium for Tubularia cells. They observed mitotic figures in stained preparations and what appeared to be differentiation in interstitial cells. Unfortunately, to date, these studies have not been successfully repeated or extended.

In this study a chemically defined in vitro system for Tubularia cells has been developed. Time-lapse microcinematography has been used to detail the initial establishment of cultures, cellular migration and cell differentiation. From analyses of these films and direct light-microscope observations, a morphological role for gastrodermal digestive cells as a substrate for other cell types has been hypothesized.

Tubularia spectabilis were routinely cultured at Instant Ocean (IO) (Aquarium Systems, Inc., Wickliffe, Ohio, U.S.A.) at 12 °C. They were fed Artemia salina nauplii twice weekly and cleaned 4 h after feeding. All experimental animals were starved 24 h prior to use during which time they were placed in a beaker of IO containing 0·25% kanamycin.

Animals were cut 2 cm from the hydranth and the excised section was placed in filtered IO which had been titrated to pH 3 with HC1. This process relaxed the animals and allowed the coenosarc to remain intact. The hydranth was then excised and discarded. The proximal end of the hydrocaulus was held with forceps while coenosarc was extruded from the perisarc with a second pair of forceps. Isolated coenosarc was transferred to a sterile physiological salt solution known as Necco’s solution (Burnett et al. 1968). This solution consisted of (g/1): NaCl, 25·800; KC1,0·644; CaCl₂,0·644; MgCl₂.6H₂O, 3·080; KH₂PO₄, 0·020; Na₂HPO₄, 0·080; NaHCO₃, 0·200; Na₂SO₄, 0·920; pyruvic acid, 1·500; fumaric acid, 1·050; L glutamine, 1·090; phenol red, 0·020.

Using cataract knives, the cylindrical coenosarc was cut longitudinally yielding a flat sheet of cells. This flat sheet of cells was then transferred to a Petri dish which contained the final culture medium. This medium consisted of 3 parts Necco’s solution, 1 part glass distilled H₂O, 0·4 parts Eagle’s MEM (Grand Island Biological Comp.)

The following antibiotics were added to the final culture medium: 0·3 g/1 kanamycin, 0·25 g/1 penicillin, 0·25 g/1 streptomycin, 0·25 g/1 fumagillin B.

After 5 min in the Petri dish the coenosarc was transferred into fresh final medium on a substrate. Substrates used in this study included coverslips, autoclaved 0·22 μ m Millipore filters, 1 μ m Nucleopore filter and collagen (rat tail) coated coverslips. After the coenosarc was maneuvered to lay flat on the substrate, excess liquid was withdrawn, leaving only a small amount of fluid surrounding the cells. The substrate with the explant was placed in a Petri dish and the Petri dish placed in a desiccator. All cultures were incubated at 12°C. After the initial 3 days, fresh medium was added to the cultures. New medium was then added every 2 days. Change of medium was accomplished by pipetting fresh medium, maintained at 12°C, on to the explant for 15 min and then withdrawing the excess liquid.

The final culture medium was aliquoted and stored at - 20°C. All procedures were carried out under sterile conditions. All solutions were filtered through 3 GS sintered glass filters.

A Bolex camera with a model 500 Sage instrument panel was used to observe the cells in culture. A Zeiss photomicroscope was employed for observing the cultures. Cinematographic observations were recorded at 4frames/min on Kodak Tri X reversal film.

Coverslip cultures were incubated in media containing 25 μ Ci/ml [³H]thymidine (6·7 c/ml) for 24 h. These cultures were washed two times for 10 min each with cold thymidine solution (concentration ten times that of labeled solution) and fixed 24 h later for autoradiography. All cultures were fixed in Bouin’s fixative. The coverslip cultures were dipped in a mixture of equal parts liquefied gel form nuclear track emulsion (Ilford K-5) and distilled water, allowed to dry, and placed in light tight boxes with a desiccant for 3-4 weeks at 2°C. After the incubation period the coverslips were developed in Kodak D-19 and fixed with Kodak Rapid fix. After water washing the cultures were stained with 0·01% aqueous toluidine blue (pH 8), dehydrated, cleared in toluene and mounted. The cellular incorporation of [³H]thymidine was compared in three regions of the cultures. All labeled nuclei in a 1 mm² area of five randomly chosen fields on each of 11 explants were counted. The labeled nuclei were then separated into the regions where they were found. These regions consisted of the central explant; the periphery of the explant, an area the width of four cells; and the monolayer of digestive cells, which extended beyond the periphery of the explant.

Initially, criteria were established to determine the time at which Tubularia cells had successfully gone into culture. The principal criteria included the attachment of gastrodermal digestive cells to the substrate, cellular migration and the absence of discharged nematocysts.

While preparing and testing over 3500 cultures, it was also observed that cultures had to be maintained at 11-12°C. When the temperature exceeded 15°C, gastrodermal digestive cells often retracted from the substrate. In this investigation cultures were maintained for only 4-week periods. During the initial testing period, however, cultures were maintained for up to 8 weeks.

At 24 h after explantation on to coverslips, large pigment-filled (carotenoid) gastrodermal digestive cells (Cohen, 1952) attached to the substrate. These cells also retained attachments to the explant. The cells displayed a fibroblastlike appearance during the attachment process. Furthermore, a large fanshaped extension was exhibited at their point of attachment to the substrate (Fig. 1). Cinematographic evidence revealed that this extension ruffled. The fan-like membrane occurred not only at the free edge of the gastrodermal digestive cells but also extended around the sides of the cells (Fig. 2). Attachment was not overtly polarized, yet a radial outgrowth of the digestive cells was often observed.

Between 24-48 h of culture, the digestive cells became vacuolized. During this time period time-lapse photography revealed two migration patterns in the digestive cells. Initially, the digestive cells migrated individually or retained a thin membranous attachment to the tissue explant. Cinematographic evidence showed that these cells migrated by a gliding motion. When individually migrating digestive cells came in contact with each other, they did not pile up but formed a sheet or monolayer of digestive cells. Once a monolayer was formed, the digestive cells moved as a sheet of cells with the most peripheral cells exhibiting a fan-shaped membrane (Fig. 3). The fan-shaped membranes of these peripheral digestive cells acted as a leading edge and appeared to pull the other cells of the monolayer. These membranes also appeared to be the main source of attachment of the sheet to the coverslip for when these membranes detached from the surface, the entire sheet retracted to the explant.

At 60 h after explanting, time-lapse studies showed the monolayer of gastrodermal digestive cells was infiltrated by gland cells, nematocysts, epitheliomuscular cells and interstitial cells (Fig. 4). Gastrodermal digestive cell migration still occurred but at a reduced rate. As epithelio-muscular cells and gland cells migrated on to the monolayer of gastrodermal cells, the digestive cells underwent a process of vacuolation leaving areas of what seemed to be extremely thin cytoplasm. This event was observed consistently and may be essential for further development of the culture.

From time-lapse observations, interstitial cells appeared to extend small filopodial processes as they migrated. These cells moved either between digestive cell attachments or on the monolayer of the vacuolated digestive cells. Nematocysts per se did not migrate actively over the monolayer but were passively pushed by the migratory movements of the epithelio-muscular cells. Gland cells were observed to migrate by an amoeboid motion between vacuolated digestive cell attachments.

Autoradiographic studies of 48-72 h cultures treated with [³H]thymidine showed nuclear incorporation of label in digestive cells, epithelio-muscular cells, interstitial cells, cnidoblasts and gland cells (Fig. 5 A, B). Approximately 84% of the total nuclear incorporation occurred in the explant; 16% of the incorporation occurred in the monolayer. Approximately 58% of the nuclear incorporation in the explant was concentrated at the periphery of the explant (Table 1). These observations are consistent with the suggestion that these cells were undergoing mitosis. Although no grain counts were made, heavy nuclear incorporation allowed relative ease in judging differences in amounts of [³H]thymidine incorporation. Reduced amounts of silver grains in interstitial cells and digestive cells in the monolayer portion of the tissue culture were observed.

Various substrates were employed to test their effect on the establishment and outgrowth of cultures. Tissues were routinely explanted on to glass coverslips. These cultures were therefore used as a standard for substrate comparisons.

Explants were cultured on 0·22 μ m Millipore filters and 1 μ m Nucleopore filters (Table 2). Explants on neither substrate exhibited digestive cell attachments. Forty-eight hours after explantation it appeared that gland, interstitial and epithelio-muscular cells migrated directly on to the filter and attached to it (Fig. 6 A, B).

Explants were also placed on collagen-coated coverslips (Fig. 7 A, B). Within 12 h digestive cells attached to the substrate. Twelve hours later the migrated area was invaded by gland, epithelio-muscular and interstitial cells. Therefore, the rate of digestive cell attachment occurred twice as rapidly on collagen-coated coverslips as it did on uncoated glass coverslips. Also, a greater concentration of digestive cells per unit area migrated from the explant on to the collagencoated coverslips. Finally, migration and attachment of vacuolated digestive cells occurred farther away from the explants than in control cultures. Migration and outgrowth of gland, epithelio-muscular and interstitial cells also occurred twice as rapidly in the collagen substrate cultures (Table 2).

A labile material has been isolated from hydranth extracts of Tubularia. This material presumably is derived from nerve cells found in the hydranth region (Brooks, Lesh-Laurie & Glickman, 1971). Preliminary tissue culture studies with this material were performed with Tate summer’ animals. Only 10-15% of the explants from coenosarc of Tate summer’ animals normally entered culture (Table 2). After 24 and 48 h these cultures normally showed reduced amounts of digestive cell outgrowth. When the culture medium was supplemented with the hydranth extract (1 part extract:70 parts medium), 100% of the explants went into culture. In addition digestive cells attached to the substrate within 12 h after explantation. Attachment and migration of the digestive cells therefore occurred at twice the rate found in ‘normal’ coverslip cultures (i.e. employing coenosarc from any other time of the year) (Table 2).

At no time did bacterial contamination occur or increase with the addition of the hydranth extract. Also, no osmotic swelling of cells was observed.

Several unique effects on cell populations were observed between 48-60 h culture after using the hydranth extract as a culture supplement. Interstitial cell nest population increased 2-3 times. The relative frequency of interstitial cell derivatives (e.g. nerves, cnidoblasts and nematocysts) also increased (Fig. 7A, B).

This isolated material has also been added to Tate summer’ explants 24 and 48 h after the coenosarc was placed in culture. Before adding the isolated material, these explants displayed none of the criteria for successful cultures. Twenty-four hours after the addition of the isolated material, 100% of these explants showed digestive cell attachment and migration. The normal sequence of tissue culture events then followed.

A chemically defined in vitro system has been established for Tubularia. Twenty-four hours after explantation, attachment, migration and outgrowth of digestive cells occur. In 48-60 h cultures, interstitial cells, epithelio-muscular cells and gland cells migrate on to the monolayer of gastrodermal digestive cells. All cells in vitro bear considerable similarity cytologically to the cells in situ. From the results of this investigation it is postulated that the elongate, vacuolated, digestive cells play an important role in the establishment and maintenance of cnidarian tissue cultures. This role may involve the digestive cells’ migratory activity and their capacity to serve as a substrate for other cells.

The digestive cells were the first cells to attach to the substrate and migrate farthest in culture. When grown on glass coverslips, the digestive cells attached to the substratum by a large fan-like membrane, similar to that observed in vertebrate embryonic cells (Abercrombie, 1965). Time-lapse revealed that this fan-like membrane acted as a leading edge for the gliding movement of the cells. The digestive cells moved either individually or in sheets. Digestive cells moving individually did not pile up when coming into contact with other digestive cells but formed a monolayer of cells. This behavior is analogous to ‘contact inhibition’ described by Abercrombie (1967).

Explants grown on collagen-coated coverslips exhibited a more rapid attachment and accelerated migration. The effects observed with collagen-coated coverslips indicated that the digestive cells adhered more quickly to collagen than to glass. An hypothesis explaining this increased rate of attachment and migration could center around collagen acting as a mesogleal mimic. Burnett & Hausman (1969) offered a theory that digestive cells and epidermal cells use the collagen-elastin mesoglea as a point of attachment for migratory activity. In vitro observations are consistent with this suggestion.

Time-lapse and cytological observations revealed that epithelio-muscular cells, interstitial cells and gland cells normally wandered only over a digestive cell monolayer. When the explanted tissue was placed on a substrate (e.g. a Millipore filter) interstitial cells, gland cells and epithelio-muscular cells were found several hundred micra away from the initial tissue mass. No digestive cell monolayer attached nor formed in the presence of the filters. These observations indicate that a smooth or collagen-like substrate may be necessary for digestive cell attachment. Interstitial, gland and epithelio-muscular cells may require a substratum with grooves; a substratum that may be similar to a digestive cell monolayer. The depressions and/or grooves may serve as anchoring points for the cells’ processes.

The morphological importance of digestive cells in situ can be postulated from the results of this in vitro investigation. In situ, when a hydranth is excised, the mesoglea in the animal breaks down (Tardent, 1962). At the wound surface, digestive cells are the first cells to close the wound (Glickman, unpublished observation). Epidermal cells then migrate over the digestive cells. Therefore, all of the in situ evidence supports the hypothesis that digestive cells serve as a substrate for other cells similar to that which occurs in vitro. Without a suitable substrate (e.g. the surface of digestive cells) epithelio-muscular cells may be incapable of migratory movement. Tardent & Erymann (1959) showed that during hydranth histogenesis, digestive cells shifted distally while the epitheliomuscular cells remained attached to the perisarc. From their reports one could postulate that it is only after digestive cells have been established as a substrate that the other cell types may then migrate into the primordial region.

Autoradiographic studies using [³H]thymidine have shown nuclear incorporation of the label in digestive cells, epithelio-muscular cells, interstitial cells, gland cells and cnidoblasts. Fifty-eight per cent of the nuclear incorporation in the expiant occurred in cells at the periphery. This observation is consistent with the idea that the periphery of the explant may act as a ‘feeder’ for the migrated area.

In situ, Tardent (1963) reported that the number of mitotic figures in regenerating and non-regenerating Tubularia was extremely low. Campbell (1967) found mitotic activity in the hydranth region but proximal to the hydranth, few mitoses were observed. The in vitro system may exhibit an increase in the rate of mitotic activity as compared to the in situ situation. However, detailed quantitative analyses must be performed in order to compare both systems accurately.

Without the perisarc which is normally impermeable to drugs (Burnett et al. 1968), cells in an in vitro system can be treated directly to note cellular effects from any isolatable hydranth materials. Burnett (1966) has hypothesized that a quantitative change in an inducer may result in a qualitative effect on cnidarian cellular differentiation. Lesh (1970) showed that in regenerating Hydra, the relative frequency of interstitial cells and interstitial cell derivatives varied with varying concentrations of an isolated inductive material. Since interstitial cells are present in tubularian tissue cultures, it may be possible to produce qualitative differences in the direction of interstitial cell differentiation by supplementing the culture media.

Several events were observed when a material isolated from Tubularia hydranths was supplemented to the culture media. ‘Late summer’ animal expiants exhibited digestive cell attachment and migration. Interstitial cell populations increased 2-to 3-fold. Interstitial cell derivatives (e.g. nerves, cnidoblasts) were also more abundant in these cultures. Experiments are in progress to quantitate these observations and to detect any changes in mitotic activity occurring with the hydranth supplement.

M. G. wishes to express his sincere gratitude to Dr Allison L. Burnett for his stimulating introduction to cnidarian development. This work was supported by a grant from Research Corporation and by a USPHS Biomedical grant to Case Western Reserve University. A portion of this work was completed while M.G. was a NSF Undergraduate Research Fellowship recipient.