Download Visual function in regenerating teleost retina following

Visual Neuroscience ~2007!, 24, 299–307. Printed in the USA.
Copyright © 2007 Cambridge University Press 0952-5238007 $25.00
DOI: 10.10170S0952523807070265
Visual function in regenerating teleost retina following
surgical lesioning
ALLEN F. MENSINGER 1,2 and MAUREEN K. POWERS 1,3
1
Vision Research Center and Department of Psychology, Vanderbilt University, Nashville, Tennessee
Department of Biology, University of Minnesota, Duluth, Minnesota
3
Gemstone Vision Research Institute, Rodeo, California
2
(Received October 12, 2006; Accepted March 9, 2007!
Abstract
Regeneration of the teleost retina following surgical extirpation of 25% to 100% of the neural retina was
investigated in goldfish ~Carrasius auratus! and sunfish ~Lepomis cyanellus!. The retina will regenerate following
removal of up to 95% of the neural retina, however complete extirpation prevented regeneration. Visual sensitivity
was assessed by examining components of the electroretinogram ~ERG! and the dorsal light reflex ~DLR! during
regeneration. B-wave amplitudes in the experimental eyes increased throughout the study and central connections
were reestablished as indicated by the progressive improvement in the dorsal light reflex. The recovery of visual
function was closely correlated with retinal regeneration. Visual recovery progressed more slowly than following
complete cytotoxic destruction of the mature retina ~Mensinger & Powers, 1999! because the surgery removed a
large number of the pluripotent cell population and restricted the number and distribution of regenerating foci.
Keywords: Regeneration, Retina, Vision, Electroretinogram, Photoreceptors
The plasticity of the teleost nervous system provides an excellent model for observing developmental processes in an adult
animal. The teleost optic nerve has been the focus of numerous
studies investigating regeneration and synaptogenesis. Sperry ~1949!
was the first to document the restoration of function following
optic nerve transection. More recent studies have demonstrated the
capacity of severed optic nerve fibers to reform highly topographical projections and restore visual function ~Meyer, 1977, 1980;
Schmidt & Edwards, 1983; Northmore & Masino, 1984; Meyer
et al., 1985; Callahan & Mensinger, 2007!. Retinal regeneration
provides a more complex problem, because it involves the genesis
of entire classes of retinal neurons, including new central projections to the tectum. Although, the regenerative potential of the
teleost retina has been well documented ~Raymond et al., 1988;
Hitchcock et al., 1992; Cameron & Easter, 1995; Raymond &
Hitchcock, 1997; Hitchcock & Raymond, 2004!, little work has
been done to assess the visual capability of the regenerated tissue.
It is imperative to assess the physiology of regenerating systems in
addition to the morphology, to ascertain the success of the
regeneration.
A previous study ~Mensinger & Powers, 1999! demonstrated
that up to 50% of the b-wave amplitude of the ERG was restored
in the teleost retina following complete cytotoxic lesioning of the
mature teleost retina. Unfortunately, optical side effects caused by
the cytotoxin, such as shrinkage of pupil diameter and cataract
formation, precluded more accurate visual assessment of the regenerated tissue. Retinal regeneration has been documented following surgical removal of small patches of retina without
Introduction
The teleost retina is a dynamic tissue in which neurogenesis occurs
at the cililiary marginal zone ~CMZ! and throughout the central
retina during the life of the fish. Multipotent retinal stem cells in
the marginal zone will generate all classes of retinal neurons;
however, the stem cells that persist throughout the central retina
normally will develop only into rod photoreceptors ~Johns &
Easter, 1977; Johns & Fernald, 1981; Johns, 1982; Raymond &
Hitchcock, 2000; Hitchcock & Raymond, 2004; Raymond et al.,
2006!. However, following cytotoxic destruction ~Maier & Wolburg, 1978; Kurz-Isler & Wolburg, 1982; Raymond et al., 1988! or
surgical excision ~Hitchcock & Raymond, 1992; Hitchcock et al.,
1992; Cameron & Easter, 1995; Raymond & Hitchcock, 1997! of
the central retina, the stem cells become pluripotent and differentiate into various classes of retinal neurons in the central retina of
mature fish ~Raymond et al., 1988; Hitchcock et al., 1992, Cameron & Easter, 1995; Hitchcock & Raymond, 2004!. The replacement of retinal neurons in adult goldfish results in complete
anatomical regeneration of the retina ~Raymond et al., 1988!,
establishment of normal synaptic connections ~Hitchcock & Cirenza, 1994!, and restoration of axonal contacts with the brain
~Stuermer et al., 1985!.
Address correspondence and reprint requests to: Dr. Allen F. Mensinger, Department of Biology, University of Minnesota, Duluth, 1035
Kirby Drive, Duluth, MN 55812. E-mail: [email protected]
299
300
compromising the system’s optics ~Hitchcock et al., 1992!, and
therefore offered an opportunity of more accurate assessment of
visual function in the regenerating retina.
Materials and methods
Animal husbandry
Goldfish ~Carassius auratus! were obtained from Ozark fisheries
~Richland, MO! and sunfish ~Lepomis cyanellus) were seined from
ponds on the McCoy Farm ~Dickson, TN!. Fish ranged from 4 to
6 cm standard length. Fish were housed in 40–120 L aquaria
equipped with mechanical, chemical and biological filtration, and
maintained at 208C on a 12:12 LD cycle. Fish were fed a diet of
commercial fish food ~TetraMin! supplemented with frozen Artemia sp. All animal care and surgical procedures conformed to
institutional, American Physiological Society and Society for Neuroscience animal care guidelines.
Surgical extirpation
Fish were dark-adapted overnight and transferred to opaque containers under dim red light before light onset. Fish were anesthetized in a 0.01% solution of tricaine methane-sulfonate ~MS-222,
Sigma, St. Louis, MO!, removed from their containers, and placed
between moist sponges under a dissecting microscope. Retinas
were removed within 15 min of light exposure by one of two
methods: ~1! patch extirpation—the retina was accessed transsclerally, and a section of the dorsal retina extirpated; ~2! aspiration—
the retina was accessed through the front of the eye, and varying
amounts of retinal tissue aspirated.
Patch extirpation
The conjunctiva was removed from the eye, and the eye pivoted
ventrally in its socket, revealing the dorsal sclera. A small incision
was made with a microblade, parallel and posterior to the limbus
at approximately the 12 o’clock position. Trans-scleral incisions
were made with an iris scissors parallel to the limbus approximately 458 from vertical, followed by a short medial incision. The
resulting tissue flap was retracted from the sclera, and inverted to
reveal the neural retina and pigment epithelium. The neural retina
and the CMZ were teased away from the underlying pigment
epithelium, and a minimum of 15% to 25% of the total retinal area
was removed. The scleral flap was repositioned and the wound
closed with cyanoacrylate glue ~Pacer, Rancho Cucamonga, CA!.
Patch extirpation was performed on goldfish ~N ⫽ 12! and sunfish
~N ⫽ 5!.
Aspiration
A small incision was made near the limbus at the 12 o’clock
position. Lateral cuts were made with an iris scissors along the
limbus, approximately 908 in each direction. The resulting eye flap
was pulled anterior and downward, displacing the lens anterior,
and exposing the interior of the eye. The eye was visually divided
into quadrants using the cross hairs of an ocular micrometer, with
the optic nerve head serving as an approximate center. Varying
amounts ~50% to100%! of the retina were aspirated using a suction
pipette ~tip diameter 1 mm!, attached to a vacuum aspirator. The
CMZ was removed concurrently with the neural retina at the
margins of the aspirated areas. Because some tissue was also
removed from the border of adjacent quadrants, the tissue extir-
A.F. Mensinger and M.K. Powers
pation estimates should be considered conservative. Presurgical
dark adaptation rendered the neural retina amendable to aspiration,
limiting ~range 0% to 5% of total aspirated area! the amount of PE
that was damaged or dislodged. Following tissue removal, the eye
was filled with sterile saline ~0.9%!, and the flap reattached to the
limbus with cyanoacrylate. Seventy goldfish and 10 sunfish contributed to this portion of the experiment.
The contralateral eye of each experimental fish was left intact,
and served as an intrafish control. Several fish underwent either
patch ~N ⫽ 4 goldfish; N ⫽ 2 sunfish! or aspiration ~N ⫽ 4
goldfish; N ⫽ 2 sunfish! sham surgeries, which mimicked the
surgical procedures except for retina removal. Following all surgeries, fish were revived by fresh water circulation over the gills,
returned to home aquaria, and isolated from tankmates until normal equilibrium was restored. In later experiments, post-operative
sunfish were housed separately.
Electroretinogram
Methodology was identical to a previous investigation ~Mensinger
& Powers, 1999!. Briefly, fish were anesthetized in a 0.001%
solution of tricaine methane-sulfonate ~MS-222, Sigma!, placed
between sponges in a small Plexiglas aquarium, and the anesthetic
solution continuously recirculated over the gills via an intraoral
tube. An AgCl2 recording electrode was inserted into an incision at
the limbus of the control or experimental eye, and a reference
electrode placed into the ipsilateral nare. Light from a regulated
tungsten-halogen source ~Ealing model 227-1403! was focused on
a shutter ~Uniblitz model 325b!, collimated, and refocused on a
308 ' fiber optic light pipe. The fiber optic was placed 5 cm from
the eye, which remained just above water level. The eyes were
covered with moist pieces of filter paper. The light beam was
adjusted to be perpendicular to the eye and at this distance; the
light from the fiber optic covered the entire surface of the eye. The
irradiance of the light was maintained at 20 uW cm⫺2 s⫺1 as
measured by a UDT 40⫻ radiometer. Fish were dark adapted for
15 min, and presented with five consecutive 200 ms flashes of
broad band light at 30 s intervals to either the experimental or
control eye. The resulting waveforms were averaged by computer.
The entire procedure was repeated for the other eye ~trial sequence
alternated between control and experimental eye!, and then the fish
was revived with fresh water, and returned to its home aquarium.
Monochromatic stimuli for the spectral sensitivity experiment
were provided by placing interference filters ~Melles-Griot, half
width at half height 8 nm! in the collimated portion of the beam of
the apparatus. Intensity response functions were generated via an
ascending method of limits, starting with a subthreshold stimulus
and increasing intensity until the criterion response ~15 µV! was
obtained.
Dorsal light reflex
Methodology was similar to the dorsal light reflex used to assess
visual function following cytotoxic lesioning ~Mensinger and Powers, 1999!. Fish were placed individually in a 40-L aquarium
equipped with a 15 W overhead florescent light. The top of the
aquarium was covered with several sheets of white paper to diffuse
the light. An opaque barrier with a small port for the videocamera
surrounded the aquarium. Fish were dark adapted for one hour, and
then videotaped for five minutes following light onset. The fish
was filmed from the front of the arena and a 1-cm ⫻ 1-cm grid was
placed on the back of the aquarium for reference. The DLR was
Visual function in regenerating teleost retina
301
determined by analyzing single frames of the videotape and calculating the degree of tilt from the fish’s normal, vertical upright
position. Five measurements were made at one minute intervals,
and the results averaged to calculate the DLR.
Histology
Fish were sacrificed with an overdose of MS-222. Both eyes were
removed and fixed overnight in Bouins solution @picric acid;
formaldehyde ~37% v0v!; glacial acetic acid; 15:5:1#, the lens
removed, and eyecup cut into dorsal0ventral or nasal0temporal
halves. This tissue then was dehydrated in an alcohol series,
embedded in paraffin, serially sectioned at 8 mm, and stained with
Lee stain ~0.025% Methylene Blue and 0.025% Basic fuchsin!.
Ocular and retinal morphology was quantified with Neurolucida ~Microbrightfield, Colchester VT!. The retinas ~patch N ⫽ 7;
aspiration N ⫽ 10! of the experimental and control eyes were
reconstructed, and divided into laminar and non-laminar categories, based on morphological examination. The laminar tissue
consisted of retinal regions with mature photoreceptors and normal
retinal stratification and included: ~1! normal or original tissue; ~2!
advanced regenerated tissue; and ~3! mature CMZ generated tissue. The non-laminar categories consisted of retinal regions that
appeared either abnormal or immature and were divided into three
categories: ~1! immature—containing some retinal stratification
but lacking mature photoreceptors; ~2! abnormal—multiple and0or
inverted retinal layers; and ~3! missing—devoid of retinal tissue
~Figs. 1, 2!.
A minimum of five representative radial sections, spaced at
least 100 mm apart, from each half of each eyecup were reconstructed. Radial sections from the contralateral control eye were
reconstructed from approximately the same position as the experimental eye. The retinal outline was traced and the length ~ µm! that
each morphological category occupied on the radial section was
calculated with the Neurolucida program MORPH ~Microbrightfield!. The CMZ contribution for the experimental eye was calculated by measuring the retinal length, distal to the last transretinal
fusion ~defined in results; Fig. 1a!. Because control eyes lacked a
distinct morphological marker such as a transretinal fusion, the
margin was calculated by subtracting the average retinal lengths of
control eyes at the beginning of the study, from the average lengths
Fig. 2. Light micrographs of radial cross-section through a retina 240 days
following retinal aspiration. ~A! The arrows indicate the terminal fusion
marker separating regenerated retina ~left!, and the newly established
marginal growth zone. ~B! Laminar fusions ~arrows! from the central
section of the retina. ~C! Abnormal regenerated retina. The arrows indicate
a circular cluster of photoreceptors ~rosette!. Scale bars ⫽ 25 mm.
during the course of the study. Light micrographs were made with
a Kodak DSC-200 digital camera ~Rochester, NY!, imported into
Photoshop ~Adobe, San Jose, CA! and printed with a Kodak
XLS-8300 printer without alteration.
Results
Surgery
Fig. 1. Light micrograph montage of a radial cross-section through a dorsal
retina 270 days following patch extirpation. The arrows indicate laminar
fusions. The tissue origin is as follows: margin generated—left of ~a!;
regenerated—bordered by ~a! and ~c!; original—right of ~c!. Scale bar ⫽
50 mm.
Both surgical methods proved successful at avoiding the optical
side effects previously experienced with cytotoxic lesioning, and
pupil diameter and lens clarity were unaffected. Only a few fish
~3%! manifested post-operative complications such as infection or
fungal growth around the incision and were excluded from the
study. Some experimental eyes did experience post-operative shrinkage, primarily the eyes with greater percentages of retina aspirated
~ⱖ75%!. However, average eye diameter remained within 10% of
the control eye.
Surgery did not adversely affect goldfish health, because average survival time of the experimental fish paralleled both unoperated and sham control groups, with several control and experimental
302
A.F. Mensinger and M.K. Powers
goldfish surviving until the conclusion of the study ~580 days!.
However, surgery negatively impacted the sunfish because postoperative fish exhibited increased aggressive behavior. This resulted in a high number of post-operative deaths, and necessitated
that surviving fish be maintained separately for the duration of the
study. Even following complete healing of the incisions, sunfish
with diminished visual function that had previously cohabited,
could not be placed in the same aquarium without fatal implications. This contrasted with sham operated or control groups that
could be routinely maintained in schools of five. These postoperative complications resulted in lower numbers of experimental
sunfish than goldfish.
Morphology
Regenerated tissue was morphologically distinct from control tissue by slight “imperfections” in the retinal anatomy. The most
noticeable difference was small, laminar fusions ~Raymond et al.,
1988; Hitchcock et al., 1992! that transversed the inner nuclear
layer and ganglion cell layers. These areas may denote the union of
disparate tissue such as regenerating tissues joining from two foci,
merging with original tissue, or with the CMZ. Tissue that was
similar in appearance to controls, free of laminar fusions, situated
near the retinal border, and bounded medially by a terminal fusion
marker, was considered margin generated. These CMZ areas could
be only accurately distinguished in the experimental eye when
contiguous with regenerating areas ~Figs. 1, 2!. Table 1 summarizes the morphometrics of the reconstructed retinas.
Patch extirpation
Regenerated retina following patch extirpation ~⬎120 days!
showed normal retinal stratification interspersed with laminar fusions at the junction of the surgical incisions. Marginal growth was
only observed in areas contiguous with original or regenerated
Table 1. Morphometric analysis of regenerating retina
Sample size
Time ~days!
Laminar tissue
Original tissue
Regenerated tissue
CMZ tissue
Total laminar
Non-laminar tissue
Immature tissue
Abnormal tissue
Missing tissue
Total non-laminar
Retinal area ratio
~experimental0control!
Putative functional retina
area 2
B-wave amplitude
~experimental0control!
Patch
Aspiration
Ouabain 1
7
304 6 47
10
272 6 31
9
270 6 34
81.4 6 3.8
13.2 6 2.0
1.7 6 0.8
96.3 6 2.9
31.0 6 5.1
42.3 6 14.7
7.862.4
81.1 6 3.1
None
67.6 6 6.5
12.9 6 2.6
80.5 6 4.6
1.2 6 0.7
2.3 6 1.7
none
3.5 6 2.9
98.1 6 3.8
7.3 6 2.0
9.6 6 3.2
2.0 6 1.3
18.9 6 3.3
85.1 6 6.7
8.5 6 4.8
10.3 6 5.8
0.7 6 0.6
19.5 6 9.6
84.5 6 7.3
94.9 6 5.1
67.8 6 6.0
65.0 6 5.3
88.7 6 3.7
55.4 6 7.1
45.8 6 7.3
Values are mean percentages 6 SE.
1
~Mensinger & Powers, 1999!
2
~experimental retinal area0control retinal area! * laminar percentage of
experimental retina
tissue indicating the CMZ was removed with the neural retina.
Retinal regeneration into the extirpated zones was completed by
day 240, and the regenerated areas appeared free of abnormalities.
Total retinal area was within 2% of controls with the regenerated
area consisting of 13.2 6 2.0% of the area of the experimental
retina ~Table 1!. Fig. 1 shows a light micrograph montage of a
radial cross-section through a regenerated area 270 days following
patch removal. The section was taken from the dorsal portion of
the nasal hemisection, close to the nasal-temporal border. Margin
generated tissue is visible to the left of the terminal laminar fusion
~arrow a!. The regenerated area exhibited normal laminar organization, and contained only a single laminar fusion ~arrow b! within
its borders. The ventral fusion denoted the merger of the regenerating retina with the original retinal tissue, which is visible to the
right of the arrow ~arrow c!.
Aspiration
In contrast to the orderly appearance of the regenerated retina
following patch extirpation, the morphology of the advanced stage
aspirated retina displayed greater variation, and was characterized
by retinal tissue in several stages of development ~Fig. 2!. In
reconstructed aspirated retinas ~N ⫽ 10, average time 272 days
post-surgery!, approximately 2.0% of the retina was devoid of
neural tissue, and an additional 16.9% of the retina was incomplete
or abnormal. The incomplete areas were characterized by inner
layers lacking mature photoreceptors, whereas the abnormal retinas included circular clusters of photoreceptors termed “rosettes”
~Fig. 2C; del Cerro et al., 1992!, significantly thickened portions
of neural retina, or detached areas. Thus, an average of 18.9% of
the experimental retina was lacking mature photoreceptors or
normal stratification and was considered unlikely to contribute to
the b-wave of the ERG.
Areas characterized by normal retinal stratification with mature
photoreceptors were observed contiguous and expanding outward
from the non-extirpated tissue, with 42.3 6 14.7% of the experimental eye composed of regenerated tissue. CMZ growth was
limited to areas contiguous with original or regenerated tissue
indicating the CMZ was removed with the neural retina. The
average amount of functional retina contributed by the CMZ was
7.8 6 2.4%. Total functional area ~original, regenerated, and
margin tissue! comprised 81.1% of the area of the experimental
eye, and total retinal ~laminar and non-laminar! area of the experimental eye averaged 85.1 6 6.7% of controls ~Table 1!.
Electrophysiology
Patch extirpation
ERGs were detectable in excised retinas ~N ⫽ 8! three weeks
following surgery ~earlier recordings were not attempted to allow
the incision to heal! ~Fig. 3!. The initial goldfish ERGs were less
than expected based on tissue extirpation estimates ~15 to 25%!,
because b-wave amplitudes were approximately 64.2 6 6.1% of
controls three weeks after surgery. The b-wave amplitude showed
a moderate increase through day 108, before increasing to 91.0 6
2.8% of control amplitude by day 225. Continual improvement
was noted through day 570 with the experimental eyes in surviving
fish ~N ⫽ 2! indistinguishable from controls ~98.0 6 2.0!. Patch
extirpation also was used to remove approximately 20% of the
sunfish retina ~N ⫽ 5!. Sunfish b-wave amplitude was reduced to
59.8 6 7.9% of controls, 21 days following surgery, before increasing to 84.3 6 6.0% by 35 days. At the 175-day mark, the
Visual function in regenerating teleost retina
Fig. 3. The percent recovery ~experimental0control! of the b-wave is
plotted versus the time in days following removal of retinal patches in
goldfish ~circle! and sunfish ~square! eyes. Data represent mean b-wave
amplitude percentages and sample sizes range from 4–8 for intervals
⬍500 days and 2 for data ⬎500 days. Error bars ⫽ 1 SE.
b-wave amplitude of the experimental eye had recovered to 93.7 6
3.1% of the control eye, which was comparable to goldfish recovery ~91%! at the same stage ~Fig. 3!.
303
to 40.2 6 7.9% by day 102, and 64.0 6 12.7% of controls by
nine months. Extirpation of 50% of the goldfish retina ~N ⫽ 10!
initially reduced b-wave amplitude to approximately a third of
controls. These retinas displayed steady and continued improvement throughout the study, attaining 70.1 6 3.0% of control
amplitudes at 270 days. Removal of half the sunfish retina
~N ⫽ 5! reduced b-wave amplitude to 40.4 6 10.0% of the
control eye, before recovering to 74.7 6 2.8% of controls by
210 days ~Fig. 4!.
Fig. 5 plots the b-wave amplitude ~ µV! of the control and
experimental eye versus time ~days! following 75% retinal removal. The mean amplitude of the control b-waves increased
approximately 20% during the experiment. Retinal aspiration initially reduced b-wave amplitude to 9.9 6 2.1 µV or approximately
15% of controls. The mean b-wave amplitude of the experimental
eyes increased to 68.0 6 6.9 µV by day 240, but remained below
pre-surgical level.
The percentage of laminar ~original, regenerated, and CMZ
tissue! and nonlaminar ~partial, abnormal, and missing tissue!
retina in the experimental eye was calculated, and compared to
total retinal area of controls. These values were then plotted
against the percent recovery of the experimental b-wave amplitude. There was a linear correlation ~r 2 ⫽ 0.79! between the
percentage of laminar retinal area and b-wave amplitude ~Fig. 6!.
Spectral sensitivity
Retinal aspiration
Fig. 4 plots the percent recovery of b-wave amplitude of the
experimental eye compared to its intrafish control eye, versus
time ~days! after aspiration of goldfish ~50% to 100%! and
sunfish ~50%! retina. Regeneration was not observed in goldfish
~N ⫽ 15! following complete removal of the retina during the
study. Goldfish ~N ⫽ 8! with 95% of the neural retina removed,
displayed very small b-waves at day 52, with amplitudes increasing to 11.1 6 4.4% of controls by 210 days and averaging
26.0 6 1.9% of controls by day 270. Removal of 75% of the
goldfish retina ~N ⫽ 12! initially reduced b-wave amplitude to
13.0 6 6.9% of controls, however b-wave amplitude increased
The dark-adapted spectral sensitivity of several ~N ⫽ 4! fish with
75% retinal aspiration following recovery of least 60% ~range 60
to 72%! of b-wave amplitude ~at 550 nm! was tested. The irradiance ~photons s⫺1 cm⫺2 ! required to induce a 15 µV criterion
response for each wavelength was plotted versus the wavelength
~nm! ~Fig. 7!. The experimental eyes were approximately 40% less
sensitive at 550 nm than controls which correlates with the percentage of b-wave recovery. The composition of the regenerated
retina was similar to controls.
Dorsal light reflex
Post-operative fish swam with the control eye positioned away
from downwelling light. The deviation from the fish’s normal
upright position was correlated with the b-wave recovery in the
Fig. 4. The percent recovery ~experimental0control! of the b-wave is
plotted versus the time in days following aspiration of sunfish ~SF: open
circles—50% aspiration! and goldfish ~GF: triangles—50%; squares—
75%; diamonds—95%; solid circles—100%! retina. The data represent
mean b-wave amplitude percentages and sample sizes range from 4 to 15.
Error bars ⫽ 1 SE.
Fig. 5. The average b-wave amplitude ~ µV! from the experimental ~square!
and control ~circle! eyes is plotted versus time ~days! following 75%
retinal aspiration in the goldfish. Sample sizes range from 5 to 12. Error
bars ⫽ 1 SE.
304
A.F. Mensinger and M.K. Powers
Fig. 6. The percent recovery ~experimental0control! of the b-wave is
plotted versus the percentage of laminar retinal area ~experimental0control!
in the goldfish retina. Symbols ~open circles—aspiration; filled circles—
patch! represent individual goldfish. The solid line is a linear regression
through the combined data y ⫽ 3.86 ⫻ ⫹ 0.81, r 2 ⫽ 0.79.
Fig. 8. ~A! The dorsal light reflex ~degrees! is plotted versus the percentage
recovery of the b-wave ~experimental0control! for patch ~solid circle! and
aspirated goldfish ~open circle!. ~B! The dorsal light reflex ~degrees! is
plotted versus the percentage recovery of the b-wave ~experimental0
control! for patch ~solid circle!, and aspirated sunfish ~open circle!. Each
point represents one animal.
108. Recovery of 70% to 80% of b-wave amplitude virtually
eliminated the tilt ~Fig. 8B!.
Fig. 7. The irradiance ~photons s⫺1 cm⫺2 ! necessary to achieve the criterion 15 µV response in the experimental ~circle! and control ~square!
goldfish eye following 75% aspiration is plotted versus wavelength ~nm!.
Tests were conducted from 210 to 280 days post surgery. Each data point
represents the average values from four fish.
experimental eye. Fig. 8A plots the DLR in degrees of vertical
deviation versus percent recovery of the b-wave amplitude. Goldfish exhibited a pronounced 188 to 308 deviation following complete aspiration of the retina, and this behavior persisted throughout
the study. The degree of recovery of the b-wave amplitude was
correlated to the degree of inclination of the DLR. Following
aspiration of 75% of the retina, recovery of 20% to 30% of b-wave
amplitude reduced the inclination to 10.7 6 1.28. B-wave recovery
of 40% to 60% produced an average tilt of 4.3 6 0.88, whereas
60% to 80% recovery was correlated with a vertical deviation of
3.3 6 1.28. The patch excised goldfish initially showed only a
minimum deviation ~⬍58, 28 days!, which was eliminated after
b-wave recovery of 95% of controls. Retinal extirpation had a
much more profound effect on sunfish equilibrium. Aspiration of
50% of the neural retina caused a DLR of up to 388. This
exaggerated behavior persisted until b-wave function reached approximately 40% of controls when the inclination dropped below
Discussion
The goal of this study was to determine the functional recovery of
regenerated retinal tissue following surgical lesioning, and circumvent the cytotoxic effects of eye shrinkage and cataract formation.
Because cytotoxically lesioned retina displayed comprehensive
morphological regeneration, but regained only 40% to 50% of
b-wave amplitude, it was hypothesized that lens occlusion and
tissue shrinkage were partly responsible for reduced sensitivity of
the regenerated retina ~Mensinger & Powers, 1999!. Greater visual
sensitivity ~b-wave amplitude! was achieved than following cytotoxic lesioning. However, surgical extirpation fell unexpectedly
short of achieving parity with control eyes during the 280 day
window established for regeneration following complete cytotoxic
lesioning ~Raymond et al., 1988!. The following analysis explains
the mechanisms behind the unexpected slower recovery rates and
associated reduced visual sensitivities following surgical lesioning.
ERG analysis
The ERG was chosen to assess visual function, because it allowed
repeated sampling over time, and analysis of the establishment of
photoreceptor synapses with the inner retina ~b-wave! ~Mensinger
& Powers, 1999!. It had been predicted that regeneration would
transpire between the 140 days required for small patch regeneration ~Hitchcock et al., 1992!, and the 280 days necessary for
Visual function in regenerating teleost retina
regeneration following complete cytotoxic lesioning ~Raymond
et al., 1988; Mensinger & Powers, 1999!. However, retinal area
following patch extirpation did not become statistically indistinguishable ~t-test; P ⬎ 0.20! from controls, until approximately
270 days after surgery. Regeneration lagged further behind in eyes
subjected to 75% aspiration, because these retinas continued to
contain incomplete areas, and attained only 85% of control retinal
area by 270 days.
The slower regeneration rates were caused by the different
retinal responses after injury. Previous studies ~Hitchcock et al.,
1992! demonstrated that after surgical removal, neuroepitheliallike dividing cells cluster at the wound edge, and form a “blastema.” It was postulated that the rod precursor or stem cells on
the edge of the wound are stimulated to become pluripotent, and
generate all classes of retinal neurons ~Raymond & Hitchcock,
2000; Otteson et al., 2001; Wu et al., 2001!. The blastema at the
wound edge contains this “embryonic zone,” and it slowly
migrates into the wound from the edges, leaving behind a trail
of regenerated tissue ~Cameron & Easter, 1995!. In the present
study, histological examination confirmed that following surgery,
only the wound edge was actively regenerating. The absence of
regenerating foci in surgically excised or aspirated areas noncontiguous with the migrating blastema indicated that the pluripotent cell population had been removed along with the neural
retina. Additionally, both surgical techniques either removed or
damaged the CMZ, because retinal tissue was not observed in
areas noncontiguous with other regenerating tissue. Therefore,
regeneration was limited to the rate that the blastema would
migrate into the excised retina. This contrasted with the multiple
foci of the cytotoxically lesioned retina, which because of their
greater number and wide spread distribution, allowed regeneration to proceed more rapidly.
The current model for fish retinal regeneration is that stem cells
in the inner nuclear layer ~INL!, initiate regeneration ~Julian et al.,
1998; Otteson et al., 2001; Wu et al., 2001; reviewed in Hitchcock
et al., 2004!. However, the cellular origins of the regeneration still
remain to be confirmed. Recent studies have suggested that Müller
glial cells may serve as stem or precursor cells for retina regeneration in the fish ~Yurco & Cameron, 2005! and the chicken
~Fischer & Reh, 2001!. Whatever the mechanism, surgical removal
of the neural retina limits regeneration to the edge of the wound
margin. Although transdifferention of retinal pigmented epithelium
~RPE! can result in retinal regeneration in amphibians and recent
studies have detailed the process in the newt ~Mitsuda et al., 2005!
and frog ~Yoshii et al., 2007!, there has been no evidence of
transdifferation in fish. The absence of regenerating loci that was
not in contact with original or regenerating tissue indicates that the
RPE was not the source of retinal regeneration in the fish.
Origin of b-wave amplitude increase
Following 75% aspiration, regenerated and original tissue accounted for approximately 80% of total retinal area of the experimental eye at day 270. The b-wave amplitude in control eyes
increased 20 µV ~20%! during the study. This was attributed to
natural margin addition and central neurogenesis of rod photoreceptors. However, experimental b-wave amplitudes displayed a
six-fold gain to approximately 60 µV during this period. Thus,
margin addition and central neurogenesis in the intact retina was
insufficient to account for the gains in the experimental eye and the
regenerated retina was the main source for the increase in b-wave
amplitude.
305
However, despite the steady increase in absolute sensitivity,
experimental b-waves were only 55.4% of control b-waves at
270 days, despite 81% of the experimental eye containing normal,
stratified retina. Although there may be several explanations, such
as the morphological differences in the regenerated tissue, we
believe that the reduced sensitivity may be attributed to our
methodology. The intrafish b-wave comparison proved to be an
effective method in the cytotoxic lesioning study ~Mensinger &
Powers, 1999!, because the marginal growth zone continued to
expand equally in the two eyes, canceling its contribution to the
ERG, and allowing a direct comparison of original and regenerated
tissue. In the present study, margin addition was restricted in the
experimental eye to areas contiguous to original or regenerated
tissue. Thus removal of the marginal growth zones in the experimental eye, contributed to lower than expected experimental0
control ratios. To more accurately assess the b-wave sensitivity
of the regenerated retina, the relative retina area ~experimental0
control! was multiplied by the laminar, and presumably functional area of the experimental eye. The predicted maximum
b-wave amplitude of the experiment eye was 58%, which is
approximately the b-wave amplitude ~54%! determined in the
experimental eye ~Table 1!. Therefore it appears that the sensitivity
of the regenerated tissue is close to normal and the regenerated
retina is functional.
Other mechanisms
Other factors may have also contributed to the reduced sensitivity of the experimental eye. Although presurgical dark adaptation facilitated separation of the retina from the underlying
pigment epithelium, small areas of the pigment epithelium may
have been removed or damaged, impeding regeneration into
these areas. This is consistent with previous studies, which
showed that intact PE was necessary for teleost retinal regeneration ~Schmidt et al., 1978; Knight & Raymond, 1994!. Additionally, abnormalities such as areas of detached retina, abnormally
enlarged layers and circles or “rosettes” of photoreceptors ~del
Cerro et al., 1992! were noted in the aspirated eyes. Although
several of these anomalies had well defined photoreceptors, it
was unlikely they possessed the proper neural connections
to contribute to the b-wave. Our methodology also did not
quantify other morphometric features that could have also accounted for reduction in b-wave amplitude of the experimental
eyes such as: ~1! lower bipolar cell density; ~2! lower number of
synapses; ~3! delay in synapse formation.
Spectral sensitivity
Previous studies have shown that the regenerated retinas can
contain an abnormal photoreceptor mosaic ~Cameron & Easter,
1995! and that cell-type-specific tangential arrays are not restored
especially for cone photoreceptors ~Vihtelic & Hyde, 2000; Cameron & Carney, 2004; Stenkamp et al., 2001; Stenkamp & Cameron 2002!. Although the current study did not address the spatial
re-organization of the photoreceptors, the spectral sensitivity of the
regenerated retina was determined. The regenerated retina was
reduced in sensitivity because of lower percentages of functional
tissue; however its spectral sensitivity was similar to controls
indicating that the relative ratio of photoceptors and visual pigments was restored. It has been hypothesized that visual pigment
distribution after regeneration enables restoration of normal spectral sensitivity ~Cameron et al., 1997!.
306
Behavior
The ERGs indicated that the regenerated retina was functional,
however the question remained whether central connections were
established, and if the new tissue could mediate normal visual
function. The dorsal light reflex was used to assess behavioral
function; it provided a simple, natural noninvasive test for the
establishment of central connections ~Silver, 1974; Powers, 1978;
Lin & Yazulla, 1994!. The DLR has evolved as a camouflage
mechanism for pelagic fish acting as a form of counter illumination ~Walls, 1967!. Fish will tilt their vertical axis in reference to
solar position throughout the day, to align the lighter underside
with downwelling light to minimize their silhouette to ventrally
situated predators. When the visual input to one eye is interrupted
or constricted, the fish will adjust its vertical position accordingly,
in an attempt to balance the light input to both eyes. The DLR
proved an accurate measure, as morphological differences in retinal area as small as 10% were detectable. As regeneration progressed, the postural imbalance was reduced indicating that central
connections were being reestablished. This reduction was a product of synaptogenesis and not habituation, because the tilt persisted
in animals with permanent retinal damage.
In summary, the retina will regenerate following surgical removal of up to 95% of the neural retina. Regeneration commences
at the edge of the wound, and slowly continues over the extirpated
area. Visual function was restored, and the spectral sensitivity of
the retina matched the control eye. The fish could mediate at least
rudimentary visual behavior with the regenerated tissue.
Acknowledgments
The authors thank P. Raymond and P. Hitchcock for detailing the surgical
procedures and R. Wall for performing the histology; J. McCoy for
assisting with sunfish collection and the McCoy family for allowing access
to their property. Supported by National Institute of Health grants F32EY06348, P30-EY06348, and RO1-EY08256.
References
Callahan, M.P. & Mensinger, A.F. ~2007!. Restoration of Visual Function Following Optic Nerve Regeneration in Bluegill ~Lepomis Macrochirus! X Pumpkinseed ~Lepomis Gibbosus! Hybrid Sunfish. Visual
Neuroscience 24, 309–317.
Cameron, D.A. & Easter, S.S., Jr. ~1995!. Cone photoreceptor regeneration in adult fish retina: Phenotypic determination and mosaic pattern
formation. Journal of Neuroscience 15, 2255–2271.
Cameron, D.A. & Carney, L.H. ~2004!. Cellular patterns in the inner
retina of adult zebrafish: Quantitative analyses and a computational
model of their formation. Journal of Comparative Neurology 471,
11–25.
Cameron, D.A., Cornwall, M.C. & MacNichol, E.F., Jr. ~1997!. Visual
pigment assignments in regenerated retina. Journal of Neuroscience 17,
917–923.
Del Cerro, M., Notter, M.F., Seigel, G., Lazar, E., Chader, G. & del
Cerro, C. ~1992!. Intraretinal xenografts of differentiated human retinoblastoma cells integrate with the host retina. Journal of Neuroscience 15, 2255–2271.
Fischer, A.J. & Reh, T.A. ~2001!. Muller glia are a potential source of
neural regeneration in the postnatal chicken retina. Nature Neuroscience 4, 247–252.
Hitchcock, P., Ochocinska, M., Sieh, A. & Otteson, D. ~2004!. Persistent and injury-induced neurogenesis in the vertebrate retina. Progress
in Retinal and Eye Research 23, 183–194.
Hitchcock, P.F. & Cirenza, P. ~1994!. Synaptic organization of regenerated retina in the goldfish. Journal of Comparative Neurology 343,
609– 616.
Hitchcock, P.F. & Raymond, P.A. ~1992!. Retinal regeneration. Trends in
Neuroscience 15, 103–108.
Hitchcock, P.F. & Raymond, P.A. ~2004!. The teleost retina as a model
for development and regeneration biology. Zebrafish 1, 257–271.
A.F. Mensinger and M.K. Powers
Hitchcock, P.F., Lindsey Myhr, K.J., Easter, S.S., Mangione-Smith,
R. & Dwyer Jones, D. ~1992!. Local regeneration in the retina of the
goldfish. Journal of Neurobiology 23, 187–203.
Johns, P.R. & Easter, S.S., Jr. ~1977!. Growth of the adult goldfish eye.
II Increase in retinal cell number. Journal of Comparative Neurology
176, 331–342.
Johns, P.R. ~1982!. The formation of photoreceptors in the growing retinas
of larval and adult goldfish. Journal of Neuroscience 2, 179–198.
Johns, P.R. & Fernald, R.D. ~1981!. Genesis of rods in teleost fish retina.
Nature 293, 141–142.
Julian, D., Ennis, K. & Korenbrot, J.I. ~1998!. Birth and fate of
proliferative cells in the inner nuclear layer of the mature fish retina.
Journal of Comparative Neurology 394, 271–282.
Knight, J.K. & Raymond, P.A. ~1994!. Retinal pigmented epithelium
does not transdifferentiate in adult goldfish. Journal of Neurobiology
27, 447– 456.
Kurz-Isler, G. & Wolburg, H. ~1982!. Morphological study on the
regeneration of the retina in the rainbow trout after ouabain damage:
Evidence of dedifferentiation of photoreceptors. Cell Tissue Research
225, 165–178.
Lin, Z-S. & Yazulla, S. ~1994!. Depletion of retinal dopamine increases
brightness perception in goldfish. Visual Neuroscience 1, 683– 693.
Maier, W. & Wolburg, H. ~1978!. Regeneration of the goldfish retina after
exposure to different doses of ouabain. Cell Tissue Research 202, 99–118.
Mensinger, A.F. & Powers, M.K. ~1999!. Visual function in regenerating
teleost retina following cytotoxic lesioning. Visual Neuroscience 16,
241–251.
Meyer, R.L. ~1977!. Eye-in-water electrophysiology mapping of goldfish
with and without tectal lesions. Experimental Neurology 56, 23– 41.
Meyer, R.L. ~1980!. Mapping the normal and regenerating retinotectal
projection of goldfish with autoradiographic methods. Journal of Comparative Neurology 189, 273–289.
Meyer, R.L., Sakurai, K. & Schauwecker, E. ~1985!. Topography of
regenerating optic nerve fibers in goldfish traced with local wheat germ
injections into retina: Evidence for discontinuous microtopography in
the retinotectal projection. Journal of Comparative Neurology 239,
27– 43.
Mitsuda, S., Yoshii, C., Ikegami, Y. & Araki, M. ~2005!. Tissue interaction between the retinal pigment epithelium and the choroid triggers
retinal regeneration of the newt Cynops pyrrhogaster. Developmental
Biology 280, 122–132.
Northmore, D.P.M. & Masino, T. ~1984!. Recovery of vision in fish after
optic nerve crush: A behavioral and electrophysiological study. Experimental Neurology 84, 109–125.
Otteson, D.C., D’Costa, A.R. & Hitchcock, P.F. ~2001!. Putative stem
cells and the lineage of rod photoreceptors in the mature retina of the
goldfish. Developmental Biology 232, 62–76.
Powers, M.K. ~1978!. Light-adapted spectral sensitivity of the goldfish: A
reflex measure. Vision Research 18, 1131–1136.
Raymond, P.A, Barthel, L.K., Bernardos, R.L. & Perkowski, J.J.
~2006!. Molecular characterization of retinal stem cells and their niches
in adult zebrafish. BMC Developmental Biology 6, 36.
Raymond, P.A. & Hitchcock, P.F. ~1997!. Retinal regeneration: Common
principles but a diversity of mechanisms. Advances in Neurology 72,
171–184.
Raymond, P.A. & Hitchcock, P.F. ~2000!. How the neural retina regenerates. In Vertebrate Eye Development. Volume 31, ed. Fini, M.E.,
pp. 197–218. Berlin, Germany: Springer Verlag.
Raymond, P.A., Reifler, M.J. & Rivlin, P.K. ~1988!. Regeneration of
goldfish retina: Rod precursors are a likely source of regenerating cells.
Journal of Neurobiology 19, 431– 463.
Schmidt, J.S., Cicerone, C.M. & Easter, S.S. ~1978!. Expansion of half
retinal projection to tectum in goldfish-electrophysiological and anatomical study. Journal of Comparative Neurology 177, 257–277.
Schmidt, J.T. & Edwards, D.L. ~1983!. Activity sharpens the map during
regeneration of the retinotectal projection in goldfish. Brain Research
269, 29–39.
Silver, P.H. ~1974!. Photopic spectral sensitivity of the neon tetra ~Paracheirodon innesi ~Myers!! found by the use of a dorsal light reaction.
Vision Research 14, 329–334.
Sperry, R.W. ~1949!. Reimplantation of eyes in fishes ~Bathygobius
soporator! with recovery of vision. Society of Experimental Biology
Medical Proceedings 71, 80–81.
Stenkamp, D.L. & Cameron, D.A. ~2002!. Cellular pattern formation in
the retina: Retinal regeneration as a model system. Molecular Vision 8,
280–293.
Visual function in regenerating teleost retina
Stenkamp, D.L., Powers, M.K., Carney, L.H. & Cameron, D.A. ~2001!.
Evidence for two distinct mechanisms of neurogenesis and cellular
pattern formation in regenerated goldfish retinas. Journal of Comparative Neurology 431, 363–381.
Stuermer, C.A.O., Niepenberg, A. & Wolburg, H. ~1985!. Aberrant
axonal paths in regenerated goldfish retina and tectum following intraocular injection of ouabain. Neuroscience Letters 58, 333–338.
Vihtelic, T.S. & Hyde, D.R. ~2000!. Light-induced rod and cone cell
death and regeneration in the adult albino zebrafish ~Danio rerio!
retina. Journal of Neurobiology 44, 289–307.
Walls, G.L. ~1967!. The Vertebrate Eye and its Adaptive Radiation. New
York: Hafner.
307
Wu, D.M., Schneiderman, T., Burgett, J., Gokhale, P., Barthel, L. &
Raymond, P.A. ~2001!. Cones regenerate from retinal stem cells
sequestered in the inner nuclear layer of adult goldfish retina. Investigative Ophthalmology and Visual Science 42, 2115–2124.
Yoshii, C., Ueda, Y., Okamoto, M. & Araki, M. ~2007!. Neural retinal
regeneration in the anuran amphibian Xenopus laevis postmetamorphosis: Transdifferentiation of retinal pigmented epithelium
regenerates the neural retina. Developmental Biology 303, 45–56.
Yurco, P. & Cameron, D.A. ~2005!. Responses of Muller glia to retinal
injury in adult zebrafish. Vision Research 45, 991–1002.