Development of the optic tectum in larval zebrafish (Danio rerio)
Bonita Song
Adviser: Stuart Thompson
Experimental neurobiology, Spring 2004
Abstract
Zebrafish (Danio rerio) begin reacting to visual stimuli barely after hatching; by seven days,
their saccade reflex (quickly flicking the head and eyes towards a stimulus) is fully developed.
Therefore, the brain region responsible for these activities - the optic tectum - must undergo
enormous transformation within this time period. In order to observe the morphology of this
transformation, I examined 3-day old, 5-day old, and 7-day old larvae, accessing a variety of
techniques. Specifically, I looked at: stained serial sections of whole fish using a conventional light
microscope; confocal images of whole and dissected fish after injection with lipophilic FM 1-43FX dye;
and scanning electron microscope images of FM 1-43FX-injected fish, which were produced on a
recently-invented, serial-sectioning scanning electron microscope. Qualitative analysis reveals that
structural change in the tectum is subtle but apparent, with cells proliferating most rapidly along the
bilateral axis of the tectal lobes, and along the outermost tips of the distal ventricular branches.
Introduction
Fast developing and reproducing, the zebrafish Danio rerio undergoes enormous development
in the first few days of its life. Its visual capabilities grow especially rapidly, enabling zebrafish larvae
to pursue moving prey by its second day after hatching. Early researchers had presumed this prey-
pursuit milestone to mark the official beginning of zebrafish visual behavior (Clark, 1981); however,
more recent findings have revealed even earlier signs. Barely-hatched fish, for example, may exhibit
a visually-evoked startle response (as opposed to touch-evoked response, which begins even earlier).
Next, tracking eye movements often kick in a few hours later. (Easter and Nicola, 1996) Ultimately, a
vigorous, adult-like saccade reflex may begin to develop in larvae as early as 3 days after hatching.
(Easter and Nicola, 1997)
Necessarily, the brain region largely responsible for these behaviors - the optic tectum (or, in
mammals, the superior colliculus) - develops at a hectic pace during these first few days. Comprising
two conspicuous, layered structures atop the teleost brain, this complex integrative center is
responsible for numerous functions, including the execution of precise visuomotor capabilities (Salas,
et al. 1997). The optic tectum is also the home of a retinotopic map: the superficial layer of each
lobe of the tectum receives input from the contralateral retina in a precise, organized fashion.
Energetic formation of this map begins just 3 days after hatching, as retinal ganglion cells (RGCS)
project axons into the outermost reaches of the tectum.
In this project, I recorded the morphological development of the zebrafish brain, examining
fish at 3 days, 5 days, and 7 days after hatching. Because the brain's 3-D layout is so integral to its
function, I used methods which would elaborate this layout. Namely, I approached the task through
three different sets of 3-D images: serial sectioning; confocal microscopy; and scanning electron
microscopy.
To begin, I serially sectioned one complete brain for each age, ultimately producing for each a
cataloque of consecutive, 3-micron-thick sections for examination under a conventional light
microscope. These catalogues provided me with a comprehensive view of neuroblast proliferation
and migration through the first week.
I also helped prepare confocal images of whole and dissected larvae, which were injected with
the lipophilic, fixable tracer dye FM 1-43FX. Not only did the confocal microscope reveal the brain
with matchless clarity and detail, but the use of FM 1-43FX dye afforded particular advantages. For,
FM 1-43FX and its analogs are presumed to intensely stain cell membranes and especially synaptic
vesicles. Therefore, if current understanding is accurate, FM 1-43FX actually pinpoints the synapses
which were fired between the times of injection and fixation.
Finally, I looked at electron microscope serial-section images of the neuropil region in the
tectum of a 5-day old fish. Like the fish examined under the confocal microscope, this fish had also
been injected with FM 1-43FX. Both the sections and images were produced by a serial-sectioning,
scanning electron microscope recently invented by Winfried Denk and Heinz Horstmann. (2004)
Materials and Methods
Aquaculture
Adult zebrafish were purchased online and kept in single-sex tanks of aerated, 28.5 °C water.
The fish were maintained on a 24-h cycle of 14 h light and 10 h darkness, and fed alternately with
dechorionated brine shrimp eggs (www.brineshrimpdirect.com) and TetraMin Flakes (TetraWerke,
Melle, Germany). Fish were bred according to the "Simple Method for Steady, Low-level Embryo
Production" as described in The Zebrafish Book (Westerfield, 2000) and the fertilized eggs were
gathered the next morning by siphon and counted. After letting the eggs sit in a weak bleach
solution (.Iml 5% sodium hypochlorite in 170ml water) for 2x5 minutes, eggs were distributed into
six-well plates containing 10% Hanks, .14 mM HEPES and a trace of methylene blue. Thereupon they
were reared in this same "system water" solution, which was changed fresh daily. Well-plates were
placed in a 28.5°C incubator set to a 14h/10h light cycle, and thereafter inspected daily for bacteria
and microciliates. Eventually fish were segregated according to hatching date, and once they were
four days old (post-hatching age) they were fed baby Artemia (brine shrimp) every other day.
Serial sectioning
Fixation and dehydration. Larvae were anesthetized with [?2%1 tricaine methanesulfonate
(Sigma, St. Louis, MO) and fixed whole overnight at room temperature on a slow rotator, in a solution
comprising .5% glutaraldehyde in 75% Hanks + 10mM HEPES. The next morning they were rinsed in
(non-glutaraldehyde-containing) 75% Hanks + 10mM HEPES for 2x10 minutes, and then processed
for 10 minutes in each of 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, and 100% EtOH solutions,
all stages of which occurred upon a slow rotator.
Embedding. The samples were then embedded in JB-4 Plus plastic (Embedding Kit Catt 18570
from Polysciences Inc.). To this I used a novel protocol, because when I had followed the
manufacturer's directions with the resources available, I had difficulty obtaining proper orientation of
the samples. So, first I infiltrated the samples according to the directions: they sat at room
temperature and over slow rotation in a solution of JB-4 Plus + benzoyl peroxide catalyst for 2x1 hour
and then overnight. Then I embedded the samples in large disk-shaped molds in non-airtight
conditions at 4°C in the directed solution of JB-4 Plus + benzoyl peroxide + accelerator for 2 hours.
Finally I let the samples harden in vacuum conditions (Denton Vacuum Model DV-502) at room
temperature for several hours. The results of this procedure were not perfect, but they were vastly
improved over those of standard procedure, wherein the fish became misaligned for sectioning, or the
plastic did not even harden.
Histology and Microscopy. The samples were cut out of their disks into blocks, according to
the parameters needed to achieve rostrocaudal orientation in the microtome. Approximately the first
third of each fish was then sliced into Zu-thick sections, using a Dupont-Sorval Model MT-2B Porter
Blum Ultramicrotome. The ultramicrotome used glass knives, which I manufactured and replaced as
needed using a standard knife cutter (LKB Knife Maker Model 7801). One at a time, each section was
dropped into a droplet of water on a microscope slide, manipulated and flattened with forceps. They
were arranged serially in order of right-to-left, top-to-bottom (like words on a page).
Image Processing. Photographs of all sections were taken at 20x with white light under a
Olympus Model BH-2 Microscope with Plan Objectives, using a Nikon Coolpix 995 digital camera [3.2
megapixel] attached to the top of microscope with an adapter (Martin Microscope MMCOOL adapter).
Any extremely blurry pictures, pictures of mangled sections, or otherwise completely unusable images
were discarded, and the remaining were enhanced, aligned and compiled into .avi movies using
Irfanview software (www.irfanview.com).
Confocal microscopy
FM 1-43FX. (wWW. probes.com, Catt F
5355) This fluorescent lipophilic dye is presumed to
brightly stain the outer leaflet of cell membranes. Moreover, it is believed to become internalized into
synaptic vesicles, causing nerve terminals to fluoresce intensely. Although it is water-soluble, it will
not automatically fluoresce in aqueous medium; it must be taken up by cell membranes.
Injection. Each larva was placed in a Petri dish and submerged in 1.5% agarose gel, and then
anesthetized with [??%1 tricaine methanesulfonate (Sigma, St. Louis, MO). The dish was then placed
under a dissection microscope, and the larva was pressure-injected with FM 1-43FX dye using a
microelectrode produced on a microcapillary tube puller. In general, the microelectrode was inserted
twice: once on each side of the bilateral axis, along the imaginary line connecting the outer back
edges of larva's eyes. Following injection, the fish was returned to its system water in order to let the
dye circulate through their ventricular system and into the cell membranes.
Microscopy. Stacks and single images were produced along the dorsal-ventral and
rostrocaudal axes, respectively. Magnifications used were 20x, 60x, and 100x. In each case Kalman
4 averaging was applied (i.e., each image was produced by taking 4 pictures of one plane and
averaging them together) to reduce noise.
Serial-Sectioning Scanning Electron Microscopy
Electron microscope stacks were produced in Heidelberg, Germany, on a microscope recently
invented by Winfried Denk and Heinz Horstmann. (Max Planck Institute) Samples were produced
through the collaboration of Denk, Horstmann, JoAnn Buchanan, Stuart Thompson and Steven Smith,
For further information regarding methods and materials used for this part of the project, see Denk,
Horstmann, 2004.
Results
Serial Sectioning
Primitive analysis of the serial sectioning stacks reveal differences in the rate of cell migration
along the proximal-distal axis. As apparent in Figure 1, during the first week after hatching the
cellular region of the deep ventricular zone grows everywhere larger as cells proliferated. However,
this increase does not occur evenly: the greatest increase in cell number appears to occur along the
bilateral axis.
It is conceivable that this difference merely results from sampling error, i.e., that it simply
reflects natural variability between the fish that were sectioned. However, excluding natural variance,
it is not likely that the perceived difference is an artifact. To begin, these images were chosen to
reflect - as much as possible - the same region in each brain: each is of the front-most section past
the eyes. Moreover, it cannot be the result of an oblique cutting job. For, if the sectioning had
occurred at an oblique angle, the eyes would have appeared egg shaped in previous pictures; in fact
however, the cylindrical eyes appear uniformly wide from top to bottom.
In addition to the uneven thickening along the proximal edges of the tectal lobes, each end of
the cellular regions of each lobe becomes more "flared." In other words, the most superficial layer of
the tectum - which is where the topographic map of retinal information will eventually reside - is
filled in from the edges, more rapidly than the middle zones. This results in a neuropil-filled "pocket"
in each lobe that becomes increasingly enclosed by cell bodies migrating around it. Interestingly,
although overall cell proliferation apparently occurs most rapidly along the proximal edge of each lobe
(as mentioned earlier), flaring is more apparent along the distal ends of the tectum.
To be sure, just as with the accelerated proliferation along the bilateral axis, it is possible that
this flare-effect is due to sampling error. However, it cannot be some unremarkable consequence of
the inwardly curving shape of the brain. That is, it is not possible that the cells at the distal edges are
just getting "pushed" around the edges because there isn't enough room. After all, the
periventricular zone actually slopes down at its rightmost tip in the 7-day old fish (Figure 1b), which
would provide proliferating cells with ample extra room to distribute themselves more evenly.
Finally, besides comparing growth rates within the image plane, I also compared the rate of
growth across images, i.e., along the rostrocaudal axis. Apparently, in contrast to the non-uniform
growth pattern along the proximal-distal axis, growth along the rostrocaudal axis appears quite
uniform. (Figure 2) Based on rough inspection, rate of cell increase did not appear significantly
greater in the hindbrain than in the fore-regions, or vice versa.
Confocal and scanning electron microscopy
Confocal and electron microscope images displayed tiny, intense puncta (appearing white in
the confocal images; black in the SEM images), which were scattered among the cell bodies and in
the neuropil. Although information is not yet conclusive, the size of these dots suggest that they
could indeed be the remnants of synaptic vesicles. (See Figures 3, 4) Note: when looking at Figure
3b, large spots of brightness are probably not significant but rather artifacts of the injection process;
a portion of the FM 1-43FX dye tends to remain concentrated in the location of injection, whether or
not that region is especially active.
Discussion
My suggestion that cells most repidly fill the superficial layer of the tectum (i.e., that the
cellular region of the deep zones "flares"), concurs with previous studies on zebrafish brain
development. In particular, Murray explicitly suggests (1996) that the "neuropil-rich marginal zone"
develops at a considerably faster rate than the rest of the tectum. This idea also fits with current
information on the timing and location of retinal ganglion cell (RGC) projections. For, as mentioned in
the Introduction, RGCs begin to assemble retinofugal pathways through the tectum by the third day
after hatching, and continue to develop in a plastic fashion thereafter; thus the superficial layer must
undergo especially intense development from 3 to 7 days after hatching.
It is not apparent, however whether this flaring is due to cell proliferation or migration. To be
sure, at a glance it appears that cells are indeed proliferating faster at the tips than at the middle
regions. However, closer inspection reveals that the largest flare (Figure 1b, on the right) is
correlated with a large "dip" in the region immediately adjacent. This suggests that flaring is at least
partially the result of preferential migration, rather than just proliferation.
As with in-plane observations, my observation regarding across-plane growth is also supported
by previous studies. (Easter, Nicola 1997) To be sure, it stands in possible conflict with the
"behavioral encephalization" theory which was developed several years ago (Armstrong and Higgins,
1971) and was later supported by Kimmel et al. (1974) For, unlike my suggestion that growth is
uniform along the rostrocaudal plane, this theory suggests that neural control and development
progresses in a caudorostral fashion, from the spinal cord to the forebrain. Nevertheless, this theory
is far from standard. Since its formulation it has been refuted repeatedly (See Easter, Nicola 1997 for
more examples), by evidence that suggests a more or less even development along the rostrocaudal
axis.
Finally, regarding the images produced with FM 1-43FX, it will be interesting to see how brain
activity - including synapse formation - develops and is distributed over the layers of the tectum at
different ages. Based on present data, one might speculate that the area of greatest activity - and
thus the densest region of FM 1-43FX dyed puncta - might reside along the superficial margins of the
tectum in a 3-day or 4-day old fish. However, further research using confocal and electron
microscope techniques must be conducted before such a speculation can be assessed.
Acknowledgements
I would like to say that Stuart Thompson, Christian Reilly, and above all Chris Patton
completely shattered the limits of obligation in helping me with this project.
References
1. Armstrong PB, Higgins DC (1971) Behavioral encephalization in the bullhead embryo and its
neuroanatomical correlates. J Comp Neurol 143:371-384
2. Burrill JD, Easter Jr SS (1994) Development of the retinofugal projections in the embryonic and
larval zebrafish (Brachydanio rerio). J Comp Neurol 346:583-600.
3. Clark, DT (1981). Visual responses in developing zebrafish. Ph.D. Thesis, University of Oregon,
Eugene.
4. Denk Wand Horstmann H (2004) Unpublished work. Max-Planck Institute, Heidelberg, Germany.
5. Easter Jr SS, Nicola GN (1996) The development of vision in the zebrafish (Danio rerio) Dey Biol
180:646-663
6. Easter Jr SS, Nicola GN (1997) The development of eye movements in the zebrafish (Danio rerio).
Dey Psychobiol. 31(4):267-76.
7. Kimmel CB, Patterson J, Kimmel RO (1974) The development and behavioral characteristics of the
startle response in the zebra fish. Dey Psychobiol 7:47-60.
8. Murray, M.E. (1996) Development of the zebrafish optic tectum (apoptosis). Ph.D. Thesis, Ohio
University.
9. Roeser T, Baier H (2003) Visuomotor Behaviors in Larval Zebrafish after GFP-Guided Laser
Ablation of the Optic Tectum. J Neurosci 23(9):3726 -3734
10. Salas C, Herrero L, Rodriguez F, Torres B (1997) Tectal codification of eye movements in goldfish
studied by electrical microstimulation. Neurosci 78:271-288.
11. Westerfield M (2000) The zebrafish book. A guide for the laboratory use of zebrafish (Danio
rerio). 4" ed., U of Oregon Press, Eugene.
Figure 1. In-plane comparison of cell proliferation and migration between 3-day and 7-day fish.
Figure 1a. Three days after hatching, the cell bodies are evenly spread along the ventricular zone of tectum
with a slight outward flare at both edges of each lobe
Figure 1b. Seven days after hatching, the difference in cell volume between the proximal and distal sides of
each lobe is pronounced, with a much greater increase in cell volume along the bilateral axis. In addition, the
flare" of cells on either side of each lobe is larger, especially along the distal edges.
Figure 2. Across-plane comparison of cell proliferation and migration between 3-day and 7-day fish.
Figure 2b (i). 7-day tectum at mid-eye (above)
Figure 2a (i). 3-day tectum at mid-eye (above)
looks much like a horizontally compacted version of
looks much like a horizontally compacted version of
that in Figure 1b. (below)
that in Figure 1a. (below)
re for comparison)
igure
Figure 1a. (repeated here for comparison)

Figure 2b(ii). 7-day brain just past the otoliths is
Figure 2a (ii). 3-day tectum just past the otoliths
nearly filled with cell bodies.
is almost completely filled with cell bodies.
Figure 3. Confocal images of the optic tectum of a 5-day old fish.
Figure 3a. (right) cell body region at [100?X)
reveals bright puncta, suggesting that these were
spots of synaptic activity. Comparison with scale
bar of indicates that these puncta are the
appropriate size to be synapses.
Figure 3b. (below) cross-section of brain,
analogous to the cross-sections produced in
serial sections for examination under a
conventional light microscope (Figures 1, 2)
Figure 4. Serial-sectioning scanning electron microscope image of the neuropil region in the optic
tectum of a 5-day old fish.
It is assumed that the two very large circles are cell bodies, that the stringy masses below them are processes,
and that the oblong grey-and-black "pills" - such as the one just to the left of the cell bodies - are
mitochondria. Further research, however, must be done to establish the location and appearance of synapses,
and the exact connectivity behind them.