Some sea squirts (ascidians1) can develop to form a normal adult body, complete with a heart, gut, and pharynx by at least three different developmental pathways . These include development from an egg to a larva which undergoes metamorphosis, asexual budding from existing zooids, and regeneration from adult tissues. In principle there could be three distinct sets of morphogenetic processes that end up reaching the same end state. However, it would be more parsimonious if the final tissue arrangement (or some intermediate state) is an attractor for some general process shared among all three development pathways.
There are several hypothetical mechanisms by which different cell types could consistently sort out from each other so that one tissue is always outside, and another inside (or next to each other) . One of these is the Differential Adhesion Hypothesis (DAH) . In this hypothesis, different tissues are assumed to act as liquids with different surface tensions leading them to sort out like water and oil. Depending on their mutual adhesive behaviors (and therefore surface tensions) two tissues may completely separate, one tissue may partially engulf the other, or one tissue may fully engulf the other. Although the physics of the original DAH is problematic , it does explain many observations in cell sorting experiments in an elegantly simple way , and updated versions are much more physically realistic . In addition, several distinct physical mechanisms (e.g. differential contractility or differential adhesion mobility) could result in similar surface-tension-like properties and cell sorting . Other mechanisms (e.g. chemotaxis) could also drive cell sorting.
Do any of these models help explain the capacity for some sea squirts to form complex structures from very different starting geometries?
Planning and preliminary attempts
To investigate this we want to dis-aggregate cells from adult zooids and observe their re-aggregation behavior. Do different cell types separate out from each other? If so, do they approximate the normal tissue arrangement? Dissociated starfish embryos (as late as gastrulation) can reaggregate to form normal larvae, although large aggregates form larvae with duplicated structures ). Some sea squirts can rebuild zooids from aggregates of stem cells . And, one report suggests that minced ascidian zooids can reorganize and regenerate to form remarkably normal zooids (although with some duplicated parts) . So it is conceivable – although highly unlikely – that dissociated adult zooids (which would mostly be differentiated cells) could reconstruct themselves into some semblance of their normal structure.
If the cells do sort out, even partially, we could then begin to test some of the hypothetical driving mechanisms mentioned above, e.g. by using micropipette aspiration to test whether cell aggregates are better described by a viscoelastic solid model (as in whole frog embryos ), a liquid-with-surface tension model (predicted by the DAH and some of its alternatives), or something else.
The first step is to choose a study organism. We need a species that is locally abundant and has at least two developmental modes. For now, a species tentatively identified as Clavelina oblonga (Fig. 1), seems the best choice. It is seasonally quite common (at this time of this year it is one of the two most common sea squirts here), and forms colonies with large zooids (~2 - 3 cm long), so it has at least two developmental modes but also has a lot of tissue per zooid.
The second, and more difficult, step is figuring out how to dissociate the cells without killing them. So far, we've tried cutting the zooid tissues out of the tunic, and placing them in Calcium- Magnesium- free artificial seawater (CMF-ASW)  for about 1.5 hrs, while mechanically disrupting them by repeatedly pipetting the tissues with a Pasteur pipette, and then centrifuging them to concentrate the cells, and adding back regular seawater (with Ca2+ an Mg2+). CMF-ASW treatment did result in more loose cells than doing this in plain seawater (only an n of 2 in each treatment), but we only obtained a few cells, and most of those were not motile (checked with time lapse videos).
Next we want to try speeding this up so that the zooids only spend 10 - 15 min in CMF-ASW to try to improve cell survival. We also want to compare whole zooids treated with CMF-ASW to whole control zooids to make sure that the cells can survive CMF-ASW treatment. After that, we're considering adding trypsin to the CMF-ASW to try to break down the extracellular matrix (ECM) further.
First Successful Attempt: Disaggregating Zooids and Observing Aggregation 10/3/14
Fig. 2: Time lapse video of dis-aggregated sea squirt (C. oblonga) cells. 13 hrs total; the image frame is ~350 µm wide (video taken by CG).
Four zooids were first incubated in 1:10 seawater in 0.505M NaCl for 20 minutes. During this incubation period, their tissues were extracted from the tunic, sliced, and placed into four centrifuge tubes. The tubes were centrifuged four times at 3000 rpm for 10 seconds, and the supernatant was replaced with CMF-ASW each time, and the tissues were mechanically agitated by pipetting them in and out. After the fourth centrifugation, the supernatant was replaced with seawater to reintroduce the calcium and magnesium ions, and the contents were centrifuged for a final time. The tissues were in CMF-ASW for a total of 15 minutes. The tissues were transferred to four separate wells in a 24-well plate. The biggest chunks of tissues and any eggs/larvae were removed.
All four wells had a fairly abundant distribution of live cells, a much higher cell survival rate than previous attempts. Timelapse videos were conducted on each well within an hour after the procedure, with 30 frames for each well over the course of 15 minutes. Well 1 was selected to undergo a one hour timelapse video, with a frame every 30 seconds, and then a 13-hour timelapse video with a frame every 5 minutes (Fig. 2).
The 15-minute timelapse videos shortly after the experiment showed that each well had mixture of disaggregated tissue cells from many cell types, and that many of them were alive. The 13-hour timelapse of Well 1 had an interesting result. The cells partially aggregated into several small clumps at a total of six hours after the experiment. The aggregates, however, seemed to shrink and slow down in movement, and stop altogether at nine hours after the experiment. Another 10-minute timelapse, at 18 hours after the experiment, with 30 second frame intervals showed signs of life in a couple of cells that were not aggregated, but no clear sign of movement or life in the aggregates themselves.
Fig. 3: Kymographs and stills from timelapse video in Fig. 2. Top: Kymograph. A re-slice of the movie frames to show an image of a single line across one clump over time: horizontal is the spatial axis (~110 µm); vertical is time (13 hrs). Features in the image leave dark bands that shift left or right if they move along the line over time. At the beginning they shift back and forth quite a bit; in the middle, bands are parallel to the time axis (vertical); at the end they converge. Middle and Bottom: Still frames from the beginning and end of the time lapse respectively. (Video taken by CG; Kymograph by MV).
One interesting feature of the longer time lapse is that there appear to be three phases (see kymograph: Fig. 3, top): for approximately four hours the cells move back and forth without obvious direction; then they are fairly still for about four hours; finally, clumps began to contract. During the last phase, there appeared to be fine filaments (perhaps fibers of matrix or cell extensions?) pulled between some cells and clumps (see full size version of Fig. 3, bottom).
The cells may not have aggregated more completely because they stuck to the bottom of the well plate (which was not coated in agar) instead of being able to pull together. On the flip side, agar might impede the ability for cells to move towards each other, since they could be using the bottom of the glass as a medium to "crawl" along. In the next experiment, we shall repeat this protocol, but put some dis-aggregated cells in agar-coated wells and others in uncoated wells see which will improve cell survival and/or aggregation rate.
Quantification of aggregation:
How can we quantify cell aggregation? Some obvious things (measuring the size of the largest clump, or the width of the largest clear space) seem problematic because they would depend on the initial cell density and the presence of initial clumps (we cannot completely disaggregate the cells at this point). Other methods attempted include using spatial autocorrelation2, using band pass and low pass filters3, and filtering with a variance filter4.
Final Experiment 2014
In order to test the theories of cell sorting in colonial ascidians and observe cell sorting behaviors, a cell dissociation method of tissues in adult zooids was developed using calcium and magnesium-free artificial seawater. Since calcium and magnesium are required for cells to adhere to each other, removal of these ions allows dissociation. Then, an experiment was performed to see whether dissociated cells were attracted to chunk of tissue containing the neural ganglion since many studies suggest the ganglion’s high regenerative potential (17)(18).
Materials and Methods
Clavelina oblonga colonies were collected off the docks at the Duke Marine Lab in Beaufort, North Carolina. These colonies were kept in a seawater table that was constantly filtered with seawater from just off the island. The water temperature, therefore, matched the temperature of the North Carolina waters. The experiments for developing a cell dissociation method were conducted between September and the end of October when water temperatures varied between 16 to 27 °C. The experiments for testing dissociated cell attraction to tissue with ganglion were conducted within the month of November, when water temperatures varied between 18 to 13 °C.
Seawater with Antibiotics
In a small beaker, 140 µL of an antibiotic mixture with 21.9 mg/mL of sodium penicillin G and 36.5 mg/mL of streptomycin sulphate was added to 50 mL of seawater.
Calcium-Magnesium-Free Artificial Seawater
In an Erlenmeyer flask, 13.11 g NaCl, 0.335 g KCl, 2.31 g Na2SO4, and 0.185 g EDTA were added to 500 mL of DI water (11). The pH was adjusted to around 8 by gradually adding 1 M NaOH.
A small colony was separated from a larger colony, and five zooids that were fully developed and close to 2 cm were cut from the colony and placed in a small dish with 1:10 seawater in 0.505 M NaCl to anesthetize the tissues prior to dissection. A small nick was made in the tunic near the top of each zooid with a razor blade to allow the solution to access the tissues. The amount of time spent in this solution varied between twenty to thirty minutes. In the last ten minutes of this time frame, a slice was made along the tunic, and tissues were extracted with tweezers. The tissues were then sliced many times using a razor blade, while attempting to remove any feces or larvae from the tissue. The tissues were then transferred to three centrifuge tubes using a pipette, with two tissues in two of the tubes and one tissue in one tube. Since each zooid is the same genotype, tissues did not need to be separated based on zooid.
Tissues were centrifuged at 3000 rpm for 15 seconds, and supernatant was removed with a pipette and calcium-magnesium-free artificial seawater (CMF-ASW) was added. This process was repeated three more times, and after each addition of CMF-ASW, the contents in the tube were pipetted in and out ~10 times. After the fourth centrifugation and addition of CMF-ASW, contents of all tubes were pipetted into one small dish with a layer of CMF-ASW. Any larvae or tissue chunks that were not dissociated were removed with fine tweezers. The dissociated tissues were then pipetted back into the three centrifuge tubes, centrifuged at 3000 rpm for 15 seconds, and the supernatant was replaced with seawater with antibiotics. Half of a centrifuge tube (0.5 mL) of dissociated cell mixture was added to each agar-coated well that required dissociated cells. Overall, the tissue cells were subjected to CMF-ASW for around 20 minutes.
Nine wells in a 24-well plate were coated with 1% agar, and an equal amount of seawater with antibiotic was added to each well. Three wells were for tissue chunk with dissociated cells, another three were for dissociated cells only, and another three were for tissue chunk only. A tiny marker dot was madein the center at the bottom of each well.
Six fully developed zooids close to 2 cm in length were separated from a colony and placed in 1:10 seawater in 0.505 M NaCl and were given a nick near the top of their tunic. After ten minutes, a razor blade was used to slice off the top of the zooid in the pharynx region just beneath the siphons. The tissue chunk was extracted from the sliced-off top with tweezers, and the two siphon tubes were cut off the chunk. Six tissue chunks containing pharynx and ganglion were pipetted into six wells designated for tissue chunk treatment.
The protocol for cell dissociation was carried out for five zooids, and 0.5 mL of the dissociated cell solution was added to each of the six wells dedicated for dissociated cell treatment.
This entire procedure was repeated to make a total of four trials, all using different colonies of C. Oblonga.
Photos and Time-Lapse Movies
Each well was photographed around 30 minutes after the dissociated cells were added to the wells to allow cells to settle. The microscope used was Leitz Labovert inverted microscope attached to an Olympus DP71 camera, and the objective used was 4x. All images are black and white. For wells with tissue chunk, the image was framed so that the chunk was placed in the bottom left corner. For the wells with only dissociated cells, the image was framed so that the marker dot was in the bottom left corner.
After ~14 hours, each well was photographed for a second time with the same camera framing and objective magnification. In the 14 hours between taking the before and after photos, a well with dissociated cells and tissue chunk was selected for a 14 hour time-lapse movie at 210 frames with a four minute interval at the 4x objective.
Cell Density Analysis
In order to measure cell density in the before and after pictures, an algorithm was developed using Image J. Photos were imported into Image J, and the chunk or marker dot in the bottom left corner in each image was selected using polygon selection and inverted to exclude. A high variance filter with radius 7 and then a threshold using Triangle were applied. The image was then inverted to make the cells black and the background white. Measurements were set to mean gray value only and a scale was removed before black areas were measured and divided by 255.
A two-way analysis of variance test was conducted using log-transformed data of “after” cell density minus “before” cell density using the variables of the three different treatments and four different trials in order to see if there is at least one difference in cell change among treatments. The data was log-transformed in order to normalize it.
Results and Discussion
Preliminary tests of my cell dissociation method and all 14-hour time-lapses of the wells with both dissociated cells and tissue chunk showed dissociated cells that were active and alive. Both with and without the tissue chunk, dissociated cells generally moved around actively before aggregating with nearby dissociated cells into many small clumps. Once the dissociated cells aggregated into small groups, their movement slowed, although some cells did sometimes leave their clump and become active again. This active movement of cells and high dissociation proves that the methodology to dissociate cells in this experiment was mostly successful, although some undissociated clumps and eggs sometimes did remain in the solution. The addition of antibiotics to the seawater successfully prevented the growth of bacteria without being harmful to the tissue cells. Overall, the cell dissociation method yielded high viability of cells with high levels of dissociation, but the mixture was not 100% dissociated.
The algorithm to measure cell density does work, as the wells with only tissue chunk had a significantly lower cell density than wells with dissociated cells. This is what was expected since the algorithm only measures cell percentage in the area excluding the tissue chunk. The algorithm, however, is not perfect. For one of the pictures in the fourth trial, the algorithm did not accurately calculate cell density, so this trial was not included in the analysis. FTT bandpass filtering reduces the “halo” around cell aggregations, and may allow a better measurement of images with particularly dense cell counts than the high variance filter used in the current algorithm. For future analyses, this alternative highlighting technique may provide a better algorithm.
The purpose of the experiment was to see if the tissue chunk attracted dissociated adult tissue cells, which were mostly differentiated. If the tissue chunk did attract cells over a long range, a larger increase in cell density in the area surrounding the tissue would be expected compared to the change in cell density in the well with only dissociated cells. The well with only the tissue chunk served as a control to see if cells spewed from chunk, which did happen at a small extent.
The two-way ANOVA revealed that there was not a significant difference in the logarithmic change in cell density among treatments, trials, or their interaction. A boxplot comparing cell change among treatments actually revealed that the logarithmic cell change for the dissociated cells with the tissue chunk was slightly lower than that of dissociated cells only (Fig. 2). This phenomenon could be explained by how some aggregations of dissociated cells tended to stick to the tissue chunk, as seen in the 14-hour time-lapse, which would then be excluded from the cell density measurement in the “after” pictures. The 14-hour time-lapses revealed that cells clumped into small groups, and was very similar to the behavior of cells in wells with only dissociated cells seen in preliminary experiments. No significant movement of dissociated cells towards the tissue chunk was observed, except large nearby clumps did tend to stick to the chunk. The tissue chunk, interestingly, was active and moved around, most likely due to the beating cilia from pieces of pharynx.
From the evidence in this experiment, it can be concluded that the tissue chunk containing ganglion and some pharynx did not attract dissociated cells over long distances. The dissociation method, however, was successful, and will be useful in future experiments testing cell sorting and reaggregation in colonial ascidians. A possible future experiment is encouraging a larger group of reaggregation, perhaps through water movement, and observing and measuring the cell sorting behavior. Other possible experiments include testing dissociated cell attraction to tissues from other parts of the body, or dissociating and reaggregating certain areas of the body with known high stem cell count such as the heart (3). Regardless, the study of cell sorting mechanisms is the fundamental basis for studies in tissue engineering, regenerative medicine, and stem cell research, and studying how animals with high regenerative potential can perform these feats in nature is the first step to discovering if these abilities can be harnessed for use in medicine in the future.
I thank Michelangelo von Dassow, my advisor who guided me through the entirety of the project. I also thank Dan Rittschoff, who provided me with a lab space to conduct all my experiments, and Beatriz Orihuela, who helped me care for my ascidians. I thank Maria Wise for helping me find specimens and orienting me with water tables. I also thank Tara, Jim Hench, and the Duke Marine Lab for their support in this project.