AEPA
United States
Environmental Protection
Agency
Environmental Research
Laboratory
Gulf Breeze FL 32561
Research and Development
EPA-600/S3-82-075 Sept. 1982
Project Summary
Usefulness of the
Self-Fertilizing Cyprinodontid
Fish, Rivulus marmoratus as an
Experimental Animal in Studies
Involving Carcinogenesis,
Teratogenesisand Mutagenesis
Christopher C. Koenig, Daniel C. Abel, Courtney W. Klingensmith, and
Michael B. Maddock
Rivulus marmoratus is a naturally
self-fertilizing cyprinodontid fish
inhabiting mangrove marshes through-
out the Caribbean. As a result of
internal self-fertilization, this oviparous
species is composed of a number of
isogenic, homozygous lines (clones),
several of which have been identified
by histocompatibility experiments and
maintained in laboratory culture for
over 3O years.
Simplified culture and handling
methods are given and data are
presented on the reproduction, growth
and development of rivulus under
laboratory culture as a prelude to the
evaluation of its potential as a bioassay
animal. Several types of bioassays
were run and evaluated using rivulus:
behavioral, carcinogenic, teratogenic,
toxic, and mutagenic. Advantages and
disadvantages of using rivulus for
such bioassays are discussed. Be-
haviorally, rivulus is capable of detect-
ing and avoiding water contaminated
with H2S. They respond (EC50 =
123.6 ppb H2S) by leaping from the
water and remaining emergent for
various periods of time while respiring
cutaneously. Hepatocellular carcinoma
among other pathologic changes were
observed in livers of rivulus a year after
exposure of adults and larvae to
diethylnitrosamine (45,30,15 ppm in
water) for 5 weeks and 12 weeks,
respectively. No pathologic changes
were found in embryos exposed
similarly. High rates of various skeletal
malformations resulted in offspring of
adults exposed to dibutyl phthalate
(DBP) and 2,3,4,6-tetrachlorophenol
(TECP) at concentrations of 20, 10
and 5% (DBP - 0.740.
This Project Summary was devel-
oped by EPA 's Environmental Research
Laboratory, Gulf Breeze. FL, to an-
nounce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Rivulus marmoratus (common name
= rivulus) is a self-fertilizing cyprino-
dontid fish inhabiting mangrove marshes
throughout the Caribbean. As a result of
natural internal self-fertilization, this
oviparous species produces isogenic,
homozygous lines (clones). Several
clones have been established for nearly
30 years by Dr. Robert W. Harrington,
-------
Jr. from single wild-caught hermaphro-
dites taken in Florida. Rivulus possesses
a number of attributes that make it
attractive as an experimental animal. It
is hardy, easy and inexpensive to
maintain in the laboratory and has a
wide salinity tolerance, short generation
time and desirable reproductive and
genetic aspects. Our main purpose was
to investigate the potential of rivulus as
an experimental animal useful in
bioassay studies and particularly in the
study of carcinogenicity, teratogenicity
and mutagenicity.
Methods
Laboratory Culture of Rivulus
marmoratus
Rivulus were held singly in stackable
glass culture dishes in 14 o/oo synthetic
seawater. Temperature is held at 25° ±
0.5°C and the photoperiod is set at 14
hr. Frozen brine shrimp (Artemia) are
fed to adult rivulus and live brine shrimp
nauplii arefedtotheyoung. Reproducing
adults are held singly in culture dishes
outfitted with teflon bottom screens to
prevent the fish from eating their eggs.
Fish are fed and eggs are removedthree
times a week and water is changed
about once a month. This system has
the advantages of simplicity and eco-
nomy. Also, chances of unwanted
contamination of the culture water are
minimal because the entire apparatus is
made of teflon and glass.
Reproduction, Growth and
Development of Rivulus
marmoratus
Because internal self-fertilization is
the natural mode of reproduction in
rivulus, other unusual characteristics
relating to reproduction result. Eggs
may be emitted at various times during
embryological development either singly
or in clusters. The average intraparental
incubation time is about 10 hr. A
relatively large proportion (22.5%) of
eggs are non-viable (assumed unfertil-
ized). Healthy adult rivulus during
reproductive periods produce about an
egg per day on an average.
Growth experiments indicated that
culture dishes 11.4cm in diameter were
optimum for rearing all life stages but
the 5.1 cm culture dishes were con-
venient for early life stages. Also, food
should be changed from live nauplii to
frozen brine shrimp at about 2 months
after hatching. The first eggs were laid
about 6 months after hatching.
The embryonic development of rivulus
(Table 1) is similar to that of other
cyprinodontid fishes. Although hatching
usually takes place in 12 to 15 days after
fertilization, a delay in hatching may
extend this up to 14 additional days. The
control of hatching or the cause of
delayed hatching in rivulus is not
known.
There are several advantages and
disadvantages to using rivulus embryos
for research purposes. The main ad-
vantage, aside from genetic, is that
under laboratory conditions eggs may
be obtained throughout the year.
However, because fertilization and the
early stages of development (often
through blastulation) are intraparental,
observations must be restricted to the
latter stages. Also, because eggs are not
fertilized synchronously it is difficult to
obtain a number of eggs in the same
stage of development.
Behavioral Bioassay using
Rivulus marmoratus
Rivulus has the capacity to leap from
the water and remain emergent for
extended periods of time while respiring
cutaneously. This behavior has been
observed in the laboratory in response
to contamination of culture water by
H2S, ammonia, formalin and rotenone.
This study was designed to evaluate
the use of rivulus as a test animal for
early warning systems in biological
monitoring.
Flow-through bioassays were carried
out in covered beakers outfitted with
Table 1. Summary of the Development Stages of Rivulus marmoratus
Stage
Time (hr)
Description
1 2.5 1 cell
2 3.5 2 cells
344 cells
4 4.5 8 cells
5 5.5 16 cells
6 6.5 32 cells
7 8 Early blastula
8 9.5 High blastula
9 10.5 Late or flat blastula
10 15 Gastrulation begins, expanding blastula
11 19 Epiblast covers 1/3 yolk
12 22 Epiblast covers 1/2 yolk, germ ring and embryonic
shield are forming
13 24 Epiblast covers 2/3 yolk
14 25.5 Large yolk plug, embryonic shield enlarging
15 31 Epiboly complete
16 34.5 Head and tail regions recognizable
17 36 Optic vesicles appear, somite formation begins
18 43.5 Optocoeles are prominent, auditory vesicles form
19 53 Optic cup and lens formation
20 55.5 Heart beats, no circulation
21 58 Body movement
22 62 Circulation
23 71 Increased vitelline circulation
24 73 Urinary bladder and otoliths first appear, pigmentation
on brain, trunk, and yolk near embryo
25 77 Pectoral fins appear, otoliths are prominent
26 90 Liver first appears, pigmentation on optic cup
27 105 Increased pigmentation and body movement
28 140 Pigmentation of the optic cup obscures the lens,
circulation in pectoral fin and liver, caudal fin
developing
29 180 Gas bladder and anal fin formation
30 211 Rays in caudal fin form, dorsal fin develops, the jaw
appears
31 240 Pectoral fin movement, circulation in the caudal fin, a
delay in hatching can occur
32 310 Hatching, rays in the dorsal and anal fin form
33 post-hatching Fish is actively swimming
-------
shelves onto which the fish could jump.
Each of five beakers were supplied by
individual head tanks containing either
test or control solutions. Dissolved
oxygen concentration was held at 2 ppm
± 0.05 ppm in all test containers and
one control container; the other control
container was kept at air-saturation.
Both position of beakers and assignment
of fish were random. Randomly selected
fish were allowed 1 hr acclimation
which was followed by 1 hr of observa-
tion. Each fish was used only once. HaS
and Qz tensions were monitored during
all 11 runs.
Results, analyzed by probit analysis,
indicate an EC50 (median effective
concentration) of 123.59 ppb H2S and
95% confidence limits were 63.68 and
181.97. Also, there was a positive
correlation (p < 0.05; Spearman rank
correlation) between HzS concentration
and length of time emergent.
Although more comprehensive testing
is required, the data indicate rivulus
shows promise in water quality man-
agement for the following reasons: 1)
the behavior is easily quantified; 2) the
response is rapid; 3) normal variation in
environmental parameters such as
salinity and temperature do not elicit the
response; 4) by the use of fish from the
same clone genetic variability can be
eliminated; and 5) rivulus adapts quickly
to laboratory conditions and large
numbers can be held at low cost.
Carconogenesis Bio assay
using Rivulus marmoratus
Liver cancer (hepatocellular carcinoma)
has been induced in a number of fish
species including rainbow trout and
several species of small aquarium fish.
Among the most active chemical agents
which cause liver cancer are the
aflatoxins and the nitrosamines. The
purpose of this study was to establish
the sensitivity of various life stages of
rivulus to diethylnitrosamine (DEN) so
that the potential of rivulus as a test
animal for carcinogenic studies may be
evaluated.
Adult rivulus (30 fish per concentration)
were exposed to DEN in the culture
water in concentrations of 45, 30, 15,
and 0 mg/l for 5 weeks. Larvae (< 1
month old; 30 per concentration) were
exposed to the same concentration
levels for 12 weeks. Embryos (56 per
concentration) were exposed to DEN in
the culture water in concentrations of
1000, 100, 32, 10 and 0 mg/l for 1
week. All surviving rivulus were held in
uncontaminated culture water for one
year, then killed in 10% buffered
formalin and prepared for histological
examination. Slides of liver sections
were examined and various pathologic
changes were described by Dr. John C.
Harshbarger (Registry of Tumors in
Lower Animals, Smithsonian Institution).
Results indicate that DEN induced
hepatocellular carcinoma, among other
pathologic changes, in rivulus as has
been shown for a number of other
experimental animals. Histological
analyses of the livers of adults and
larvae are summarized in Tables 2 and
3. Figure 1 compares normal livertissue
and solid pattern hepatocellular carcin-
oma from a fish exposed to 30 ppm DEN
for 12 weeks. No pathologic changes
were seen in rivulus exposed to DEN as
embryos. There were no obvious indi-
cations of tumors when whole livers
were examined grossly
Advantages of using rivulus in studies
involving induction of carcinoma include:
1) a variety of isogenic, homozygous
strains may be used; 2) tissue transplants
are possible within clones; 3) no
aeration is necessary in the simple
culture system, thus volatile chemicals
such as DEN are not lost; and 4) embryo
studies are possible throughout the
year. It is not known whether clear
determination of hepatocellular carci-
noma could be made with an induction
period shorter than one year.
Table 2. Pathologic Changes m Livers of Rivulus marmoratus Adults Exposed to Diethylnitrosamine for 5 Weeks
Incidence^
Nominal
Water
Cones, (mg/l)
0
15
30
45
No. fish
(Start)
30
30
30
30
No. Fish3
{End}
23
20
21
16
Fish
with
Neoplasmsh
0/23(0f
0/20(0}
7/21(33.31
5/16(31 3)
Hepatocellular
Carcinoma
0/23(0)
0/20(0)
6/21(28.6)
5/16(31.3)
Incipient
Neoplasms
0/23(0)
0/20(0)
2/21(9.5)
0/16(0)
Cholangioma
0/23(0)
0/20(0)
1/21(4.8)
0/16(0)
Adenofibrosis
0/23(0)
1/20(5.0)
13/21(61 9)
8/16(50.0)
Granuloma
0/23(0)
0/20(0)
5/21(23 8)
2/16(12.5)
"Fish were killed one year after end of exposure period.
hNo of fish with neoplasms (hepatocellular carcinoma, cholangioma, or incipient neoplasms)/'total no. of fish
c/Vo. of fish with pathologic change/total no of fish.
"Nos. in parentheses are percents
Table 3. Pathologic Changes in Livers of Rivulus marmoratus Larvae Exposed to Diethylnitrosamine for 12 Weeks
Incidence"
Nominal
Water
Cones (mg/l)
0
15
30
45
No Fish
(Start)
30
30
30
30
No. Fish
(Endf
23
18
14
17
Fish
with
Neoplasms*
0/23(0f
17/18(94.4)
12/14(857)
16/17(94.1)
Hepatocellular
Carcinoma
0/23(0)
12/18(66.7)
9/14(64.3)
13/17(76.5)
Incipient
Neoplasms
0/23(0)
7/18(38.9)
4/14(28.6)
4/17(23.5)
Cholangioma
0/23(0)
4/18(22 2)
1/14(7.1)
7/17/41 1)
Adenfibrosis
0/23(0)
16/18(88.9)
6/14(42.9)
14/17(824)
Granuloma
0/23(0)
11/18(61.1)
7/14(50.0)
7/17(41.2)
aFish were killed one year after end of exposure period.
^No. of fish with neoplasms (hepatocellular carcinoma, cholangioma or incipient neoplasms)/total no. of fish.
cNo. of fish with pathologic change/total no. fish.
aNos in parentheses are percents
-------
Figure 1. Liver section from rivulus exposed to 30 ppm DEN for 12 weeks. Border of
hepatocellular carcinoma (dark) and normal tissue (light) is distinct.
H & E x 400.
Teratogenicity Bioassays and
Chronic Full Life Cycle
Exposure of Rivulus
marmoratus to Sublethal
Levels of Selected Toxicants
Coastal and marine environments
have become contaminated with a
variety of organic and inorganic pollu-
tants. Many of these pollutants are
known teratogens, mutagens, and
carcinogens but there is little or no
information on the vast majority. Often,
effects such as tumors and various
abnormalities are seen in natural
populations from contaminated areas;
however, laboratory experiments on
affected populations are often difficult
or impossible because of problems
associated with maintaining laboratory
populations One way to gain insight
into the mechanisms of action of the
various pollutants is to expose species
which are conducive to laboratory
culture to suspect chemicals and
observe responses throughout the full
life cycle. This work was designed to
evaluate the potential of rivulusfor such
studies. Five priority chemicals were
selected for this study, among them the
known teratogen, dibutyl phthalate
(DBF). The other toxicants used were
pentachlorophenol (PCP), 2,3,4,6-tetra-
chlorophenol (TECP), 2,3,5-trichloro-
phenol (TRCP) and bromoform.
Sublethal water concentrations for
the full life cycle chronic study were
chosen as fractions (20, 10 and 5%) of
static 96-hr LC50 tests on newly
hatched larvae Randomly chosen
adults were exposed to the various
concentrations for eight months during
which time eggs were collected, incu-
bated and reared. Eggs for each fish
were randomly placed in either uncon-
taminated water or water contaminated
with the same chemicalandtothesame
degree as that of parental adults. This
was done to distinguish between pre-
hatchmg developmental effects and
post-hatching chronic exposure effects.
At the end of the exposure period all
offspring were killed, examined grossly
and group was cleared and stained for
the determination of skeletal abnormal-
ities.
The chronic exposure levels, based on
larval LC50's, were as follows. PCP -
0074, 0.037, 0.185; TECP - 0.220,
0.110, 0.055; TRCP - 0360, 0.180,
0.090; DBP - 0.740, 0.370, 0.185; and
bromoform - 8 40, 4 20, 2.10 mg/l.
Production of infertile eggs over all
exposures and controls averaged 22.6%.
The proportion of viable eggs that
hatched varied from about 70 to 90%
overall. The most common abnormality
among embryos was an edematous
pericardial cavity but this showed no
clear relationship to chemical or con-
centration. Mortality rates among post-
hatch offspring were variable but dose-
related in some cases, such as in
response to TECP exposure Gill and fin
erosion was clearly the result of
exposure to TECP and this undoubtedly
contributed to the dose-related mortali-
ties. Also, degree of damage to fins and
gills was dose-related
Skeletal abnormalities observed in
cleared and stained fish were divided
into six categories: vertebral fusion,
deformed centra, abnormal neural
and/or hemal spines, abnormal dorsal
fins, abnormal pectoral fins and ab-
normal pelvic fins. Rates of skeletal
abnormality in controls were about 30%
and this compares well with previous
work done by Dr. R. W. Harrington, Jr
on the same clones of rivulus. Prelim-
inary work on wild-caught fish from
Naples, Florida, indicate spontaneous
rates of skeletal abnormality near zero.
A teratogenic response from the DBP
exposures was evident in the rates of
vertebral fusion and neural and hemal
spine deformity A histogram of such
effects in offspring reared in uncontam-
mated water shows a clear dose-
response relationship (Figure 2). Evi-
dence for a teratogenic response was
also seen in offspring from TECP
exposed fish. There was no clear
evidence for teratogenic response in
offspring of PCP, TRCP, or bromoform
treated fish. However, offspring ex-
posed to bromoform developed ab-
normal dorsal fins.
The data demonstrate not only the
advantages of using rivulus in full life
cycle exposures, but also the sensitivity
of rivulus to chemically induced skeletal
malformations. This latter finding is
significant because to date there are no
aquatic vertebrate test animals routinely
used for determination of teratogenic
potential of various pollutants. Because
of the many attributes of rivulus
including genetic uniformity of clones, it
is anticipated that further development
of this aquatic vertebrate assay would
produce a powerful tool useful in water
quality management
Conclusions
Mutagenesis Bioassay:
Investigations with Rivulus
marmoratus and Other
Selected Fish Species
Objectives of this section include: 1)
to characterize the genetic system of
rivulus and to determine the long-term
effects of known mutagens, and 2) to
-------
80
70
60
£ 50
tl)
£ 40
30
20
W
DBP
m
Total Skeletal
Abnormalities
h m I c
Vertebral
Fusion
m
Deformed
Neural Spines
m
Abnormal
Pectoral Fins
Figure 2. Percentages of various skeletal abnormalities in offspring of DBP
exposed rivulus.
investigate possible short-term proce-
dures for assessing mutagenic effects
on the rivulus system. These objectives
were subsequently amended to include
other fish species when rivulus proved
unsuitable for the experiments we had
in mind.
The initial intent was to identify gene
markers in rivulus by starch gel elec-
trophoresis. Several fish from each
clone and 15 to 30 wild-caught fish
(caught from Naples, Florida) were
screened for 14 enzyme systems
representing an estimated 28 loci.
There were no electrophoretic dif-
ferences found between clones and
wild-caught fish. Thus, genetic distance
between clones, occurrence of self-
fertilization and segregation of alleles
could not be verified. The finding of no
electrophoretic difference supports the
view that the Florida populations of
rivulus were derived from a single small
founder population.
As part of the second objective we
attempted to establish a permanent line
of cultured cells from rivulus to provide
sufficient quantities of material to
investigate the metabolism of mutagens
and carcinogens and to evaluate sensi-
tivity to induction of mutation, DNA
repair and neoplastic transformation by
these chemicals. The attempts at tissue
culture of rivulus cells have not been
successful Alternatively, the culture of
toadfish (Opsanus tau) cells up to fourth
passage was successful.
The karyotype of rivulus was evaluated
for cytogenetic analysis, specifically
sister chromatid exchange (SCE) analy-
sis, and found unsuitable. Other marine
and freshwater species were surveyed
and the toadfish was found most
suitable, having relatively large chro-
mosomes. Attention was therefore
diverted from rivulus to the toadfish.
Culture of peripheral leukocytes, an
effective method for preparing meta-
phase chromosomes, was successful
with the toadfish. Leukocyte prepara-
tions were then used for sister chromatid
exchange assays. Rates of sister chro-
matid exchange in mammalian cells
show a dose-response relationship with
concentration of mutagens and carcin-
ogens. Techniques used to differentiate
sister chromatids in vitro, when applied
to toad fish leukocytes, were successful
and an increased rate of SCE was
obtained when cells were treated with
the mutagen ethyl methanesulfonate
No increase in rate of SCE was found
with bromoform exposure This line of
research is being pursued under another
grant.
Another set of experiments involved
characterization of the nature of the
toadfish cytochrome P450 system. A
series of comparative studies were done
to determine the effects of hepatic
microsomal enzyme preparations (S-9)
from rats and toadfish pretreated with
standard enzyme inducers (3-methyl-
cholanthrene jMCj and Arochlor |ACj)
and untreated on the metabolism of a
polycyclic aromatic hydrocarbon (benzo
(a) pyrene) and an aromatic amine (2-
aminoanthracene) to Salmonella muta-
gens.
Neither benzo (a) pyrene (BP) nor 2-
aminoanthracene (AA) were mutagenic
in the absence of S-9 protein. All fish S-
9 and S-9 from MC- and AC-pretreated
rats resulted in little activation The
extent of activation of BP and AA under
optimal conditions of S-9 type and
concentrations for fish and rat were
comparable. These results support the
growing evidence of similarities be-
tween fish and mammal enzyme systems
which metabolize pro-mutagenic xeno-
biotics.
Christopher C. Koenig, Daniel C. Abel, Courtney W. Klingensmith, and Michael
B. Maddock are with Grice Marine Biological Laboratory, College of Charleston.
Charleston, SC29412.
W. P. Davis is the EPA Project Officer (see below).
The complete report, entitled "Usefulness of the Self-Fertilizing Cyprinodontid
Fish, Rivulus marmoratus as an Experimental Animal in Studies Involving
Carcinogenesis, TeratogenesisandMutagenesis,"(OrderNo. PB82-249 194;
Cost: $13.5O. subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
Sabine Island
Gulf Breeze, FL 32561
U. S. GOVERNMENT PRINTING OFFICE: 1982/559-092/0531
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
0000329
------- |