inbred fish strains in experimental research

guus bongers bongers at
Mon Oct 23 09:20:28 EST 2000

The following paper is a more or less elaborate abstract from the article
"Development and use of genetically uniform strains of common carp in
experimental animal research", published in Laboratory Animals (1998,
32:349-363) by ABJ Bongers, M Sukkel, G Gort, J Komen and CJJ Richter.

I hope this paper will add substantial information on the use of genetically
uniform fish strains. To improve readability, I have minimized references to
literature. Please send comments or questions to bongers at

The advantages of fish as experimental animal are numerous. Great diversity
between species exists, they are highly fecund and, in most cases, they have
large eggs with external fertilization. They are used mainly in routine
toxicity testing (60 %) of which the majority is required by law. 
While in mammalian research, inbred strains of in particular mice and rats
made substantial contributions to many areas of biomedical research, very
little demands are made thusfar on fish. The only requirements concern in
most cases health status and size homogeneity. Frequently, experimental fish
with unknown genetic background and raising history are used. Some
genetically uniform strains of fish have been produced through conventional
inbreeding and have shown to be suitable for immunological, radiation and
genetic research  (medaka, Oryzias latipes: Hyodo-Taguchi and Egami, 1985;
platyfish, Xiphophorus maculatus: Kallman, 1984).
In conventional inbreeding, approx. 20 generations of full sib mating are
needed to obtain (near) homozygous animals. However, because most fish have
external fertilization, manipulation with the sexual products is possible
and fully homozygous (inbred) fish can be produced in only one generation.
The principle is to (1) eliminate the genetic contribution of one of the
parents, followed by (2) artificially doubling the haploid genome : 
(1) Elimination of genetic material of one of the parents is done by
treating gametes before fertilization with ionizing radiation like
gamma-rays or ultraviolet irradiation. Ionizing radiation causes breakdown
of chromosomes into small fragments. UV-irradiation initiates the formation
of thymidine-dimers in adjacent base-pairs, rendering the DNA inactive.
After fusion of gametes of which one parental genome is inactivated, haploid
zygotes are produced. Without any further treatment development proceeds,
but haploids die around the moment of hatching. However, (2) the haploid
state of the zygote can be changed into a diploid state by suppressing the
first cleavage using physical shocks (temperature, pressure or a combination
of the two), applied at the metaphase of the first mitosis. After this
treatment, a new cell cycle is initiated, starting with DNA-replication.
Because an exact copy of the DNA is made, all homologues are fully
identical, thus a 100 % homozygous individual will be generated. Gynogenesis
(all female inheritance) involves the irradiation of the paternal genome;
androgenesis (all male inheritance) is achieved after irradiating the
maternal genome. Thusfar, androgenesis has been applied with much less
success than gynogenesis. Irradiating eggs is more complicated than
irradiating a sperm suspension, due to the relatively large size and
adhesive chorion. Androgenesis is valuable since phenotypic effects of
(maternal) cytoplasmic constituents can be studied, genotypes can be
recovered from cryopreserved sperm and the generation interval can be
decreased since in general, male fish sexually mature earlier than their
female conspecifics.
To proof the absence of genetic contribution of one of the parents, dominant
morphological traits are most frequently used as "genetic markers". Proof of
homozygosity can be achieved by Mendelian segregation of (heterozygous)
alleles in the parent, isozymes, skin transplantations and by microsatellite
F1 hybrids (heterozygous clones) are produced by crossing two homozygous
(not related) individuals. From F1 hybrids, recombinant inbred strains can
be produced by gynogenesis or androgenesis, and repeated backcrossing of a
F1 hybrid to an inbred strain yields a congenic strain.

Outbred strains are defined as a closed colony (> 4 generations) of animals
with a limited increase of the coefficient of inbreeding (< 1 % per
generation, van Zutphen, 1993) and are assumed to be genetically variable
within the colony. There are two main objections to the use of outbred
strains. First, especially in small populations maintained as closed
colonies for long periods, high levels of inbreeding may accumulate in
outbred stocks even when the mating of close relatives is avoided (Festing,
1993). When the level of inbreeding is kept within acceptable limits,
genetic drift (reduction in gene frequencies, due to for example
unintentional selection) can still result in reduced genetic variation.
Secondly, if genetic variability in outbred strains is high, the increased
experimental "noise" could obscure true treatment effects or is mistaken for
a treatment effect. High quality animal experiments have 1) a high
replicability (= low variation between replicates during a single
measurement), 2) a high repeatability (= low variation between tests within
the same laboratory) and 3) a high reproducability (= low variation between
tests from different laboratories) (Dave, 1993). It is obvious that using
outbred strains will especially decrease reproducability. 
When testing a panel of inbred strains in a factorial design, differences in
main effects of the strains reflect the amount of (genetic) variation
present within the experiment. In the ideal situation, this amount of
genetic variation should reflect genetic variation in the outbred
population, so generalization of test results would become possible. Optimal
allocation of a fixed number of experimental units within and between inbred
strains will depend on the quantity to be estimated and possible
constraints, for example with respect to costs. If the main interest lies in
the strain-mean, the number of strains should be as large as possible.
However, more units per strain should be taken if the within strain variance
component gets larger compared to the between strain component. Also, if the
cost per strain gets higher compared to cost within strain, more units per
strain should be taken. 
Genetic variance in homozygous populations only consists of additive genetic
variance (VA). Dominance-effects (VD) are absent due to the absence of
heterozygotes. Interaction-effects (VI) are thought to be negligible. When a
panel of gynogenetic or androgenetic inbred strains is to be used for
experimental purposes, inbred strains are to be selected from different
families. Selecting strains from the same family is equal to testing within
a genetic range of VA, while selecting strains from different families
equals testing within 2VA. When outbred strains are used, results obtained
are representative for a range of genotypes with variance VA + VD + VI. When
F1 hybrids are produced from not related inbred strains (outbreeding, F =
0), results are also representative for VA + VD + VI. Whether the absence of
dominance or interaction effects in homozygous inbred strains will affect
the outcome will depend on the trait under investigation.
Phenotypic uniformity in inbred strains is depending on how much of the
variance in outbreds is of genetic origin. For characteristics with high
heritability (i.e., a large part of the phenotypic variance is additive
genetic variance), inbred strains are likely to be more uniform than outbred
strains. For characteristics with low heritability, this may not be the
case. However, the increased susceptibility of homozygous inbred strains for
environmental sources of variation can offset the reduced genetic variance,
with an increase of the total phenotypic variance as a result. So, when
inbred strains of fish are to be used, they should be derived from crossing
homozygous clonal conspecifics since F1 hybrids combine both genetic and
phenotypic uniformity (Festing, 1979; Falconer and Mackay, 1996).
Coefficients of variation in F1 hybrids of fish are still relatively large
when compared to inbred strains of laboratory rodents and vary greatly
between strains. According to Festing (1979), residual variation in mammals
can be attributed to competition between animals, chance variation in utero,
chance contamination by micro-organisms etc. The fact that fish are
poikilothermic make them more susceptible to similar sources of variation,
resulting in a large phenotypic variance, compared to homeothermic animals
(Allendorf et al., 1987).


Allendorf FW, Ryman N and Utter FM (1987). Genetics and fishery management.
Past present and future. In: Population genetics and fisheries management
(Ryman, N and Utter, FM, eds). University of Washington Press, Seattle,
London, 420 pp.

Dave G (1993). Replicability, Repeatability and Reproducability of
embryo-larval toxicity-tests with fish. In: Progress in standardization of
aquatic toxicity tests with fish (Soares, A.M.V.M. and Calow, P., eds).
Lewis Publishers, London

Falconer DS and Mackay TFC (1996) Introduction to quantitative genetics (4th
edition). Longman Group Ltd., England, 464 pp.

Festing MFW (1979) Inbred strains in biomedical research. MacMillan Press
Ltd., London, 483 pp.

Festing MFW (1993) Genetic variation in outbred rats and mice and its
implications for toxicological screening. J. Exp. Anim. Sci. 35, 210-220.

Hyodo-Taguchi Y and Egami N (1985) Establishment of inbred strains of the
medaka Oryzias latipes and the usefulness of the strains for biomedical
research. Zool. Science 2, 305-316.

Kallman KD (1984) A new look at sex determination in poeciliid fishes. In:
Evolutionary genetics of fishes (Turner BJ ed.), Plenum press, New York and
London, pp 95-171.

Zutphen LFM van (1993) Toxicity testing and genetic quality control. J. Exp.
An. Sci. 35, 202-209.


More information about the Toxicol mailing list