|Home | About | Journals | Submit | Contact Us | Français|
The zebrafish (Danio rerio) is now the pre-eminent vertebrate model system for clarification of the roles of specific genes and signaling pathways in development. The zebrafish genome will be completely sequenced within the next 1–2 years. Together with the substantial historical database regarding basic developmental biology, toxicology, and gene transfer, the rich foundation of molecular genetic and genomic data makes zebrafish a powerful model system for clarifying mechanisms in toxicity. In contrast to the highly advanced knowledge base on molecular developmental genetics in zebrafish, our database regarding infectious and noninfectious diseases and pathologic lesions in zebrafish lags far behind the information available on most other domestic mammalian and avian species, particularly rodents. Currently, minimal data are available regarding spontaneous neoplasm rates or spontaneous aging lesions in any of the commonly used wild-type or mutant lines of zebrafish. Therefore, to fully utilize the potential of zebrafish as an animal model for understanding human development, disease, and toxicology we must greatly advance our knowledge on zebrafish diseases and pathology.
As the most numerous and phylogenetically diverse group of vertebrates, fish teach us important principles about fundamental processes in vertebrate evolution, development and disease processes. For over 100 years, fish from primitive hagfish to advanced reef fish have yielded unique insights into cell biology, physiology, development, and immunology (258, 418, 545). A high level of conservation of genetic programs controlling development and fundamental physiologic processes is present among all vertebrates, as well as between invertebrates and vertebrates. Endocrine systems are highly conserved between fish and other vertebrates. Fish possess most of the tissue types of mammals except breast, prostate and lung. Fish have served as useful sentinels to detect environmental hazards, and as efficient, cost-effective model systems for mechanistic toxicology and risk assessment for many decades (206, 208). In choosing a model system for conducting particular research, it is essential to realize that no single model is best for addressing all biomedical questions. Each model species has unique strengths and weaknesses (576).
Zebrafish and other aquarium fish species have distinct advantages as models for biomedical research including much lower husbandry costs than mammals. Oviparous species including the zebrafish have external fertilization and development, facilitating access for observation and manipulation of developing embryos. Oviparous fish species can be cloned quite easily, allowing genetic manipulation for study of haploids, triploids, or tetraploids with androgenesis or gynogenesis (107, 489). Zebrafish are easily housed in compact recirculating systems, breed continuously year-round, and have short generation times of approximately 3–5 months (125). The small size of adult zebafish allows efficient, low-cost evaluation of all major organs on a limited number of slides (168). The small size of embryos and fry minimizes the cost and waste volume for drug and toxicant studies. Thus minute amounts of expensive metabolites or new targeted drugs can be rapidly evaluated.
Among vertebrates, the zebrafish embryo has unrivaled optical clarity, allowing visual tracing of individual cell fates throughout organogenesis. Fluorescent dyes or other markers aid in visualization of cell lineages (104, 105, 273, 279, 474). If inhibitors of pigment formation are included in the rearing medium, cell lineages can be traced throughout the first week of zebrafish development, and immunohistochemical or immunofluorescence studies will reveal specific cell types in whole mount preparations (269). Histological sections of larvae over 1 week of age are required to localize specific cell types. Alternatively, confocal microscopy can optically section these thicker larvae (323). A wide array of histochemical markers for protein and gene expression allows identification of essentially all cell types, and often reveals functional capabilities such as synthesis of nitric oxide or specific neurotransmitters, during development of the major tissues (75, 77, 82, 103, 111, 171, 240, 261, 520, 533, 536, 579, 583).
Immunohistochemical studies have quantified cell proliferation and cell death in specific tissues during development (100, 284, 578). The optical clarity of the embryo coupled with detailed understanding of basic developmental processes and a well-established timetable for specific developmental milestones allows elegant embryonic manipulations to distinguish the relative influences of the genetic composition of a specific cell (cell autonomous effects) versus the influences of the genetic suite of its surrounding environment (non–cell-autonomous effects). For example, at a precise stage of development, specific neurons can be removed from the spinal cord using a micropipette, and can be replaced by those from fish of a different genotype. Then the impact on neuronal fate and innervation of skeletal muscle can be determined (142). Or during various time points in development, single cells, or cell clusters can be removed from specific anatomic fields in the embryo and relocated to other sites to clarify the processes controlling cell fate determination, and revealing when the fate of certain cell types is irreversibly specified (224, 356).
DNA or RNA constructs can be readily microinjected into embryos at the 1-cell or 2-cell stages to study effects of transient gene expression. More uniform tissue expression is achieved with RNA injection (228). With injection of RNA at the 2-cell stage, typically half of the embryo expresses the exogenous construct, with the other half acting as an elegant internal control. Using constructs with a promoter such as that from a heat shock gene, laser probes can elicit transient expression of injected constructs in precise cell types at exact stages of development (203, 238, 288).
Because of these advantages of zebrafish for basic developmental biology and molecular development studies, the zebrafish has emerged as the premier vertebrate model for clarification of the roles of specific genes and signaling pathways in development. The past decade has seen intense worldwide research into molecular genetic mechanisms in cell fate determination, pattern formation, morphogenesis and functional maturation of heart, blood vessels, brain, eye, ear, nose, neural crest, muscle, cartilage, bone, lymphomyeloid system, skin, kidney, and gonad.
Cell culture methods are established to create primary and immortal cell lines from adult tissues as well as from embryos (101, 182, 183, 210, 292). Also explants of embryos and adult tissues, such as whole brain, can be cultured to study cell-cell interactions and metabolism (509).
Several transinstitute funding initiatives by the National Institutes of Health (NIH) in the past 5 years have spawned a flurry of activity using the zebrafish as an animal model for understanding human development and disease. The NIH has funded research in zebrafish genomics for the past decade. Recently, Britain's Wellcome Trust has put the zebrafish on the fast track for complete genome sequencing, with a rough draft to be available by the summer of 2003. Among fish species, the most complete database on genomics, molecular genetics and embryology is available for the zebrafish. These data are accessible through the Zebrafish Information Network <http://zfin.org/ZFIN> (559) coordinated in conjunction with the NIH-funded Zebrafish International Resource Center (ZIRC) at the University of Oregon. Updated sequencing information is posted on the Sanger Institute/Wellcome Trust web site <http://trace.ensemble.org>.
Although a rough draft of the human genome is now available, the function of most mammalian genes remains unknown. Because of the rapid embryonic development, optical clarity of zebrafish embryos, and efficiency of mutant screens in zebrafish, we are likely to understand the role of every gene in development in the zebrafish before we have such knowledge for any other vertebrate. It will take much longer to clarify the role of these genes in diseases occurring later in life.
Zebrafish have 25 pairs of chromosomes compared to 23 pairs in humans. Evolutionary genetic data indicate that a whole-genome duplication event occurred early in the teleost lineage, after separation of the fish lineage from the tetrapod lineage. So for many genes of mammalian species, duplicate genes (paralogs) occur on separate chromosomes in bony fish. Not all of these gene duplicates have been preserved during evolution (40, 161, 242, 346, 413, 414). On average about 20% of the pairs of duplicated genes have been preserved due to neofunctionalization or subfunctionalization. In neofunctionalization, an entirely new function for a gene evolves. In subfunctionalization, for example, if the original gene of mammals was expressed in 4 tissues during development, each of the duplicate genes of a bony fish might be expressed in 2 of the 4 tissues (166). Although many scientists contend that fish are simpler model systems than mammals, in terms of the molecular genetics, fish are actually more complex. An additional whole-genome duplication likely occurred in trout and salmon following the initial teleost genome duplication, so the salmonids will often have 4 co-orthologs for each gene of mammals (68).
Mutations in one of a pair of duplicated genes in zebrafish may give rise to a simpler phenotype than mutation of the ortholog in mammals. Also, in cases in which mutation of the mammalian ortholog is embryonic lethal, mutation of one of the zebrafish paralogs may give rise to viable animals, allowing better definiton of gene function and clearer understanding of the roles of the gene in signaling pathways.
Investigators worldwide collaborated to create an oligonucleotide-normalized and expressed sequence tag-characterized zebrafish cDNA library which has been used for mapping zebrafish genes, for in situ hybridization studies of gene expression, and now is being used to develop microarray technology for comprehensive evaluation of stage-specific gene expression in zebrafish (97, 289). ZFIN contains a virtual web-based map of the zebrafish genome, integrating data from the various gene mapping projects. A physical map of the zebrafish genome is available, constructed using fluorescence in situ hybridization (FISH) (407). At least 28 groups of genes are syntenic (together on a single chromosome) in zebrafish and humans (28). This shared synteny has aided identification of and positional cloning of candidate disease genes (498, 499).
Over a decade ago, geneticist Christiane Nüsslein-Volhard chose zebrafish as the best vertebrate model system for clarifying the roles of specific genes in development. Patterning these studies in zebrafish upon her highly successful Nobel-Prize-winning saturation mutagenesis studies that elucidated the role of every gene and every signal pathway in fruit fly development, Nüsslein-Volhard collaborated with Wolfgang Driever, then at Massachusetts General Hospital, to conduct the first large-scale mutagenesis studies in zebrafish (72, 112, 199, 241, 303, 380). These studies were spectacularly successful, generating thousands of mutant lines of zebrafish with defects in essentially every major organ system, as well as mutant lines with defects in basic embryo patterning. A special issue of the journal Development in 1996 was devoted to describing this cornucopia of mutant lines. The success of these initial mutant screens has stimulated many research centers around the world to continue conducting mutant screens, working toward a complete understanding of molecular, cellular, and functional development of all zebrafish tissues and organ systems. Initial screens focused on several parameters including morphology of embryos using stereoscopic and Nomarsky optics, neuronal pathway arrangements, and markers for expression of panels of genes. Now the focus of mutant screens using zebrafish is broadening to consider functional parameters such as lipid digestion (64, 110, 153), behaviors such as mating, the startle response and feeding, and diseases such as neoplasia or skeletal malformations occurring later in life (5, 160, 397). The ingenuity of future mutant screens will be limited mainly by manpower and funding—these screens are very labor-intensive undertakings.
To date, totipotent embryonic stem cells are not available for any fish species. Therefore the homologous recombination methods used for creating knockout mice cannot be used in fish. Chemical mutagens including N-nitroso-N-ethylurea (ENU; 282), psoralens (11), radiation (543), insertional mutagenesis using pseudotyped retroviruses (173, 174, 186), or dominant negative transgenes can be used to create stable mutant lines of fish with “knocked down” gene function. For homozygous lethal or recessive mutations, screening of haploid embryos or early pressure-derived embryos (35) containing the genetic material of just one parent greatly speeds up identification of mutant phenotypes (87). Unlike mammalian species in which haploid embryos do not survive, haploid zebrafish embryos survive for 4 days. These haploid embryos exhibit a well-defined suite of brain abnormalities, however, most organs can be screened for abnormalities associated with mutant genes (543). If transient gene “knockdown” is desired, morpholino antisense oligonucleotides efficiently achieve null phenotypes for most genes evaluated early in development (106, 144, 209). An alternative strategy for gene “knockdown,” albeit a more technically challenging approach, is the use of ribozymes (544). Recently, the Cre-loxP system has been used in zebrafish embryos to achieve precise control of gene expression in a spatially and temporally restricted manner (587).
Transgenic zebrafish are useful for study of the phenotypes resulting from overexpression of selected genes globally, or in a tissue-targeted fashion (256, 313, 398). Transgenic technology also allows “knockdown” of gene function using dominant negative transgenes. Recently, growing numbers of tissue-specific promoters are becoming available to allow very precise targeting of transgene expression (188, 221, 222, 238, 255, 286). Transgenic zebrafish with the green fluorescent protein marker under control of a specific promoter can highlight stage-specific, tissue-specific and cell–type-specific expression of selected genes in developing and adult zebrafish (364, 402).
Transgenes are most commonly introduced by microinjection of constructs into newly fertilized eggs. Electroporation of eggs or sperm has also been used. The most efficient method for introducing transgenes into zebrafish eggs utilizes pseudotyped retroviruses (173, 174, 186); however, use of retrotransposons and integrase is continually being optimized to achieve higher efficiency of stable integration of microinjected transgenes in developing zebrafish (96, 262).
Mutant lines of zebrafish will help clarify the roles of specific genes and their associated signaling pathways and networks in the pathogenesis of toxicant-induced lesions. Double and triple mutants can clarify the interactions of suites of genes, and these multiple mutants can be produced more efficiently and cheaply in zebrafish than in rodents.
Numerous recent review articles have trumpeted praises of the zebrafish model for understanding human development and disease (5, 30, 64, 128, 131, 133, 162, 282, 304, 505, 530, 550, 552, 537, 539, 574, 592). Active research programs are now present worldwide focusing on normal and abnormal development of essentially all organ systems and tissues shared by zebrafish and man, as well as on a variety of pathologic lesions, physiologic processes, and disease conditions including aging, alcoholism, and drug addiction, for which zebrafish can provide mechanistic insight into human disease pathogenesis.
Several hemato-oncology research groups at Harvard Medical School including those of Leonard Zon and Thomas Look, Graham Lieschke's group from the Ludwig Institute for Cancer Research, Victoria, Australia, and Shu Lin's group at the University of Georgia, together with other collaborators have clarified the role of specific genes in normal and deranged hemopoietic development of zebrafish (6, 42, 53, 121, 310). Several mutant lines of zebrafish have clinical syndromes resembling diseases of humans, such as congenital sideroblastic anemia (67), X-linked sideroblastic anemia (581), hepatoerythropoietic porphyria (547), erythroid myeloproliferative disorders (319). A transgenic zebrafish expressing a RUNX1-CBF2T1 translocation shows a preleukemic syndrome (257). The genetic program for sequential specification of pluripotential precursors followed by divergence of distinct lineages of lymphoid, myeloid and erythroid cells is highly conserved from zebrafish to humans, indicating that zebrafish will continue to provide valuable insights into additional lymphomyeloid disorders of humans (6).
Although the immune system of zebrafish has been less well studied than that of several fish species important in commercial aquaculture such as rainbow trout and channel catfish, this field will grow in importance as we begin to understand more about infectious diseases of zebrafish. Like other bony fish, zebrafish have distinct lymphoid subsets, T and B cells. Willett et al (569-572) describe expression of the recombination activating genes, rag 1 and rag 2 in zebrafish thymus and anterior kidney, the major histolologic sites for production of T cells and B cells, respectively, during zebrafish development. The Zon laboratory at Children's Hospital in Boston, MA has developed mutant lines with defective T cell development. These mutant lines fall into 5 distinct complementation groups (505, 515, 516). Zebrafish possess several novel immune-type receptor genes in comparison to mammals. These genes may shed light on evolution of the innate and adaptive immune systems of vertebrates (314). Recent data indicate that zebrafish novel immune-type receptor genes may represent evolutionary intermediates in the establishment of the leukocyte receptor cluster for cytokine responses in mammals (584). Class I and II major histocompatability proteins are mapped and characterized in zebrafish (492, 493, 497). The promoter region of the Class II A genes is highly conserved between mammals and zebrafish, indicating conservation of gene regulatory mechanisms (492). Independent duplications of Bf and C3 components of the complement system occurred during evolution of the zebrafish (189). Haire et al (201) studied immunoglobulin isotypes and the T cell antigen receptor in zebrafish. Herbomel et al (216, 217) show that macrophages appear in embryos concurrent with embryonic erythrocytes, but at a distinct location compared to other hemopoietic cells. Macrophages develop initially in mesoderm anterior to heart, while primary hemopoiesis occurs ventral to the tail, near the yolk extension. These embryonic macrophages constitute a distinct lineage from the monocyte lineage which gives rise to macrophages in adult vertebrates including zebrafish (465). Embryonic macrophages require an intact M-CSF signaling system to disseminate properly to embryonic tissues, indicated by the failure of macrophage dispersal in the panther mutant lacking functional M-CSF receptors. Antigen receptors on nonspecific cytotoxic cells are similar in zebrafish and channel catfish (251).
Specialized methods are required to analyze coagulation parameters in zebrafish due to the small size of both early life stages and adults, however, clotting times can be evaluated in blood samples from larvae by 30 hours postfertilization (246, 249). Zebrafish possess both intrinsic and extrinsic clotting pathways containing the same clotting factors present in mammals. Anticoagulants such as warfarin and hirudin act similarly in zebrafish and mammals (244, 245, 247-250). Although nucleated, zebrafish thrombocytes are functionally similar to those of mammals and circulate in the blood by 36 hours postfertilization (247). Like mammals, thrombocyte activation in zebrafish is blocked by drugs inhibiting cyclooxygenase 1, such as aspirin, but not by inhibitors of cyclooxygenase 2 (195, 247). Screens to detect mutants with defective blood coagulation are underway and lines of zebrafish with specific abnormalities in the coagulation system are evident (244).
The clarity of zebrafish embryos and fry, and the ability of the fish to survive for several days without circulation due to the diffusion of gases through the skin make this an excellent model for study of cardiovascular development (86, 163-165, 309, 318, 321, 405, 456, 462, 479, 552, 553, 556, 582, 583). Mutant lines with defects in contractility (458, 580), rhythmicity (551), heart size, heart tube patterning, valve morphogenesis, and cardiac looping are available. The gridlock (gdl) mutant with abnormal circulation to the tail resembles coarctation of the aorta in man. Babin et al (25) and Durliat et al (138) cloned the zebrafish orthologs for the apolipoprotein E (apoE) and A-1 (apoA-1) genes of humans. These genes regulate lipid uptake and distribution in mammals and ApoE plays a role in Alzheimer's disease in man. ApoE and A-1 are highly expressed in the yolk syncytial layer in zebrafish embryos, a tissue which controls yolk assimilation, and apoE is expressed in brain and eye of developing zebrafish. The roles of these genes in cardiovascular and brain development and disease in zebrafish remain to be explored. Brant Weinstein of the National Institutes of Health coordinates a Web site featuring images of zebrafish cardiovascular development <http://mgchd1.nichd.nih.gov:8000/zfatlas> or <http://dir.nichd.nih.gov/lmg/uvo/WEINSLAB.html>. The site features current information on comparative vascular development of various organs.
Some of the most intense research in the zebafish model has focused on genetic mechanisms of cell specification and morphogenesis of the nervous system (52, 142, 357, 358). Mutant lines with extremely specific defects in most components of the nervous system are available. Also mutant lines with defects in neural connections are established (237). The ontogeny of specific behaviors is well defined (156, 172, 443), and neurologic functions such as sleep are being investigated (211). Calcium fluxes can be visualized in individual neurons in the central nervous system of live fry during behaviors such as the escape response (155, 496, 590). Neuronal metabolic profiles are described (534).
Migration pathways of neural crest cells and sequential specification of particular cell types are documented in zebrafish (143, 176, 275, 377, 513). Mutant lines with specific defects in certain aspects of neural crest development are available, with lesions including abnormal pigment patterns, abnormal jaw development, and abnormal enteric neural tissue (20, 140, 213, 265, 343, 394).
At least 50 mutant lines with lesions affecting specific cells or layers of the eye are available (46, 65, 327, 328, 375, 541). Also, more sophisticated tests of visual function and ocular motion are being developed (26, 146, 185, 306, 307). All zebrafish mutants with lamination defects in the retina have anomalies of the retinal pigment epithelium (RPE). This suggests that the RPE controls retinal lamination during morphogenesis of eye. The mosaic eyes (moe) mutant shows loss of normal retinal lamination, with loss of the normal localization of dividing retinal cells to the surface of the RPE. Cell–type-specific antibodies and riboprobes show that all retinal cell types differentiate normally, but are misplaced in the moe mutant (252). Normal and abnormal morpohologic and functional development of the olfactory system is understood in detail (287, 426, 561-563). Recent work by Jessen et al (253) shows that the recombination activation genes, rag1 and rag2, which function to generate diversity in the immune system, may perform a similar function in generating a large repertoire of olfactory receptors. Cellular and molecular mechanisms controlling development and function of ear (21, 44, 99, 361, 406, 557) and tastebuds (204) are less well understood than other sensory systems in zebrafish, but are currently being investigated.
Embryo and fry stages of zebrafish possess a pronephros with a single glomerulus, whereas juvenile and adult stages have a mesonephric kidney. Recent investigations have clarified molecular and cellular aspects of renal development in zebrafish (132, 134, 135, 324-326, 460, 462). Although many aspects of renal development are conserved from zebrafish to mammals, in contrast to mammals, glial-derived neurotrophic factor (GDNF) is not required for kidney morphogenesis in zebrafish (466). Several zebrafish mutants including double bubble resemble human autosomal dominant polycystic kidney disorders. Fleer and elipsa display renal-retinal dysplasia similar to that seen in the human Senior-Loken syndrome (135, 271). Zebrafish with inactivating mutations in the homeobox gene tcf2 (vhnf1) show kidney cysts, underdevelopment of pancreas and liver, and reduction in size of the otic vesicles (494). This zebrafish mutant line is a good model for study of the MODY5 syndrome (maturity-onset diabetes of the young, type V) in humans and familial GCKD (glomerulocystic kidney disease).
Much is understood regarding the timing and genetic mechanisms controlling development of the axial and appendicular skeleton of zebrafish (158, 362). Shannon Fisher of Johns Hopkins University School of Medicine is screening mutant lines using radiography to detect lesions in the skeletal system (160). The zebrafish chihuahua (chi) mutant shows skeletal dysplasia resembling osteogenesis imperfecta in man (159).
Molecular and cellular development of skeletal muscle, and neuromuscular function in normal and abnormal zebrafish are well studied, although not yet fully understood (39, 73, 74, 200, 224, 229, 236, 363, 376). Mutant screens have identified over 50 genes essential for normal somite development (136, 137, 486). Mutant lines with abnormalities similar to human Duchenne muscular dystrophy are available (54, 55, 83, 396). Mutant lines with altered motility due to abnormal neurotransmitter receptors on nerve and muscle are described (385, 459).
Studies of normal and abnormal development of the enteric tract and liver have continued since the initial mutant screens were conducted in 1996 (98, 193, 285, 390, 424, 425). Mutants with abnormally large or small livers, or with abnormal positioning of the gut and liver are documented (157). Guo et al (198) are using microarray analysis of stage-specific gene expression combined with forward and reverse genetic analyses to clarify the roles of specific genes in liver development. Cheng et al (88) describe dysplastic histologic features in gut of early life stages of certain mutant lines of zebrafish.
Growing international interest is focused on development and function of endocrine systems of vertebrates due to the concern regarding endocrine-disrupting agents in the environment. Whereas most endocrine organs are similar structurally and functionally in fish and mammals, fish have some unique features. Fish interrenal and chromaffin tissue, the homologs of adrenal medulla and cortex, respectively, are not encapsulated and are located surrounding veins in the anterior kidney. Zebrafish thyroid is also not encapsulated and is located surrounding the ventral aorta as it emerges from the bulbus arteriosus. Fish, like birds, have ultimobranchial bodies, homologs of mammalian medullary thyroid tissue, which are located in the transverse septum between heart and esophagus. Surprisingly, although the zebrafish does not have a parathyroid gland, it has functional parathyroid hormones and receptors (438, 439, 524).
Cellular and molecular mechanisms controlling pituitary development are active topics of current research (139, 293, 440, 448), however, knowledge about development of this organ is currently less than that regarding most organ systems of zebrafish. More information is available regarding development of endocrine pancreas (19, 45, 193, 349, 437), thyroid (66, 149, 150, 316, 317, 417, 425, 431, 558), and pineal gland (167, 192, 263, 589). Whitlock et al (564) show a dual origin of cells secreting gonadotropin-releasing hormone (GnRH), with these cells arising in both neural crest and in pituitary placode.
Confirmation of expression of the evolutionarily conserved vasa gene as a marker of germ cells in zebrafish has facilitated study of migration, cell proliferation and cell death patterns of this cell type (57). Weidinger et al (554, 555) and Stars-Gaiano and Lehmann (483) describe the normal migration patterns of zebrafish germ cells and document abnormal homing of germ cells to presumptive gonad in certain mutant lines. Uchida et al (523) document histologic changes occurring during early differentiation of gonad into patterns characteristic of male or female. Chiang et al (92, 93) reveal the role of sox9 genes in gonadal differentiation and in later gonadal function in zebrafish. Bauer and Goetz (32) report a variety of mutant lines of zebrafish with abnormal development of male and female gonads, including lines with defects in various stages of spermatogenesis or oocyte maturation. Some information is available on the structure of zebrafish GnRH (415, 517) and gonadotropin receptors (293). Wu et al (577) and Pang and Ge (392, 393) outline signaling pathways controlling ooctye maturation in zebrafish. Li et al (308) successfully achieved in vitro maturation of zebrafish oocytes, documenting a culture system that may prove useful in future toxicant studies. Likewise, an in vitro culture system is optimized to produce functional sperm from zebrafish germ cells (444). Zebrafish will provide a model system for study of gender-related differences in toxicant responses (113).
Much of the initial embryology studies as well as studies of mutant lines focused on cellular and molecular mechanisms for determination of basic body plans and cell fates (273, 276-279, 347, 351, 365, 366, 451, 452, 473, 474, 511). The initial Tuebingen mutant screen identified a wide variety of mutants with defective epiboly, axis pattterning, gastrulation, dorso-ventral patterning, or anterior-posterior patterning. An example of defective patterning clarifying mechanisms of head formation is the headless (hdl) mutant. This phenotype occurs due to defective tcf3 (T-cell factor 3). Appropriate tcf3 expression is necessary to repress specific wnt target genes, allowing head formation (270). The 1-eyed pinhead (oep) mutant shows defects in L-R asymmetry mimicking some situs inversus syndromes of man such as dextrocardia, cardiac septal defects, or gut malrotation associated with mutations in the CFC gene (47, 49, 194, 450).
Normal and abnormal craniofacial development in zebrafish has been studied intensely (350, 408, 409, 451). A variety of mutant lines show phenotypes resembling congenital malformations occurring in humans.
Karlovich of Stanford University and colleagues (259) cloned the Huntington's disease gene homolog from zebrafish, and will develop a zebrafish model for this progressive neurodegenerative disease of man. Leimer et al (302) cloned the presenilin gene of zebrafish to begin developing a model of Alzheimer's disease. Musa et al (367) studied early life stage expression of 2 zebrafish orthologs of human amyloid precursor protein, and found distinct expression patterns for each of these paralogs in zebrafish. Rubenstein et al (440) are developing a zebrafish model for the study of Parkinson's disease, and have demonstrated destruction of dopaminergic neurons in zebrafish embryos following treatment with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), an agent associated with drug-induced Parkinsonism in man.
The database on metabolism in zebrafish is much less complete than that available for certain other highly studied fish species such as rainbow trout. More data are available regarding Phase I than Phase II metabolism in zebrafish. Several cytochrome P450 enzymes from zebrafish have been mapped, but full cDNA sequences are in the public domain only for cyp19a, cyp19b, and cyp26. Cyp1a1 (P4501a1) activity is induced in adults (69, 518) as well as in early life stages of zebrafish (13), and in liver cell cultures (102, 214, 353) by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The aryl hydrocarbon receptor (AHR) signaling pathway has been studied extensively in zebrafish (1, 12, 416, 500, 501). Stage-specific expression of ahr, the ahr nuclear translocator (arnt), and cyp1a1 are documented in early life stages of zebrafish (13, 335, 336).
Paralogs of aromatase are present on two separate chromosomes in zebrafish, due to an ancient chromosomal duplication event. These enzymes play key roles in sex steroid metabolism, facilitating androgen conversion to estrogens. Zebrafish cyp19a is expressed in ovary, whereas cyp19b is expressed in brain (93). Appropriate tissue-specific expression of these enzymes is required for normal sexual development (76, 280, 510, 514). Steroid glucuronides, produced in zebrafish testis by metabolism of androgens, are potent pheromones stimulating mating behavior in females (529). Hu et al (235) and Lai et al (295) studied the regulation of steroidogenesis in zebrafish. Zebrafish cyp26 plays an essential role in retinoic acid catabolism in zebrafish and mammals, allowing precise local regulation of retinoid acitivity (320, 422).
Keizer et al (264) studied species differences in acetylcholine esterase inhibition by diazinon in fish. They found that zebrafish are relatively resistant to diazinon compared to other fish species, because their acetylcholine esterase is relatively resistant to inhibition by this pesticide. Donnarumma et al (130) investigated glutathione S-transferase activity in liver of several fish species. They found no activity of this enzyme toward 1,2-epoxy-3-(p-nitrophenoxy) propane in zebrafish liver.
Relatively scant data are available regarding pharmacokinetics of drugs or environmental agents in zebrafish. Hertl and Nagel (220) and Zok et al (591) studied bioconcentration and metabolism of various substituted anilines in zebrafish. Petersen and Kristensen (403) evaluated uptake and elimination of polycyclic aromatic hydrocarbons by early life stages of zebrafish. Gorge and Nagel (190) studied toxicokinetics and metabolism of lindane and atrazine in eggs, larvae and juvenile zebrafish. Andersson et al (9) compared PCB uptake in mature zebrafish from the diet, from intraperitoneal injection, and from implanted intraperitoneal silicone capsules. Neilson et al (374) investigated bioconcentration, metabolism, and toxicity of components of bleached pulp effluent to various life stages of zebrafish. Wiegand et al (568) evaluated uptake, metabolism and toxicity of the algal toxin, microcystin, following bath treatment of various early life stages. Labrot et al (294) studied toxicokinetics of lead and uranium in zebrafish. Wicklund et al (565) investigated effects of zinc on uptake, distribution and elimination of cadmium in zebrafish. Most published zebrafish studies report exposure concentrations, but not tissue concentrations of toxicants, so that quantitative comparison of sensitivity between zebrafish and other species is difficult.
For decades worldwide, the zebrafish has been used extensively for environmental toxicity testing, in acute and chronic bioassays to generate mandated water quality criteria. Because of the rapid development and optical clarity of early life stages, the bulk of the toxicology data for zebrafish regards early life stages. To date, the great power of our current knowledge in molecular genetics, genomics, and mechanistic developmental biology have not been fully utilized in toxicology studies. We now have the tools to answer very specific questions in many areas of toxicology and toxicologic pathology. Unfortunately, the strain and source of zebrafish used in many toxicology studies is not reported in the publications. Generalizations regarding the zebrafish based on data from a single genetic line will not likely predict responses for the many wild-type lines and thousands of mutant lines. Much of the basic toxicity data for zebrafish is in unpublished reports.
Ansari and Kumar (17) studied effects of the insecticide diazinon on liver protein and nucleic acid metabolism in adult zebrafish. Kumar and Ansari (290) investigated liver toxicity of malathion. Roex et al (429) assessed acute toxicity of parathion and chlorobenzene. Zok et al (591) studied acute toxicity of substituted anilines. Lanzky and Halling-Sorensen (297) found metronidazole to have low acute toxicity to zebrafish. Meinelt et al (341) investigated the influences of calcium concentration and humic acid on toxicity of acriflavine to juvenile zebrafish. Roche et al (428) studied acute and chronic toxicity of colchicine in zebrafish. Labrot et al (294) investigated acute toxicity of lead and uranium in zebrafish, finding lead relatively nontoxic and uranium highly toxic. Van den Belt et al (527) evaluated acute toxicity of cadmium-contaminated clay to zebrafish. Ultraviolet-B light induces oxidant stress in adult zebrafish (85).
Nearly every class of environmental contaminants has been evaluated for early life stage toxicity in zebrafish. However, only a few recent studies have investigated molecular mechanisms of toxicity. Early life stage toxicity of most metals and several organometallics have been investigated in zebrafish (91, 117-120, 184, 340, 378, 389, 391, 427, 436, 447, 470, 490, 495). Toxicity of a variety of pesticides, organochlorines, and halogenated aromatic hydrocarbons is reported (43, 62, 116, 145, 190, 191, 196, 197, 215, 354, 355, 383, 502, 566, 567). Willey and Krone (573) utilized the vasa gene as a marker of primordial germ cells to track alterations in their homing to gonad caused by endosulfan or nonylphenol. Dong et al (129) applied in situ terminal transferase-mediated nick-end-labeling staining (TUNEL) to demonstrate increased cell death in the dorsal midbrain of TCDD-treated embryos. These investigators could mimic the TCDD-induced cell death with the AHR agonist beta-naphthoflavone, and prevent the increased cell death with the AHR antagonist alpha-naphthoflavone. Another recent study which exploited current molecular genetic tools available for zebrafish demonstrates that TCDD does not inhibit primitive embryonic erythropoiesis, indicated by gata-1, gata-2 markers (41). However, TCDD prevents formation of later definitive erythropoiesis revealed by the scl marker. TCDD causes craniofacial malformations in fish and mammals. TCDD reduces normal expression of the signaling molecule sonic hedgehog (shh) in zebrafish jaw (503). Toxicity of various drugs, endogenous signaling molecules and hormones have been evaluated in zebrafish (33, 202, 218, 219, 224, 383, 384, 538, 588). Akimenko and Ekker (2) show that fin malformations induced by exogenous all-trans retinoic acid are associated with anterior duplication of expression domains of shh. Various industrial chemicals and waste show adverse effects in developing zebrafish (71, 79, 95, 147, 148, 177, 296, 368, 370, 373, 388, 531, 532). Physical stresses such as magnetic fields may perturb zebafish development (85, 179, 488). Skauli et al (469) show an additive interaction of magnetic field stress with the hormone progesterone in causing delayed hatching. The algal toxin microcystin suppresses growth following early life stage exposure (568).
Most structural classes of toxicants including metals, organochlorines, and pesticides (15, 18, 58, 148, 152, 429), halogenated aromatic hydrocarbons (387) substituted anilines (63), synthetic and natural estrogens (8, 526, 528), and other industrial chemicals (79, 230) have been evaluated for reproductive toxicity in zebrafish (61, 272, 293). Potential for disruption of endocrine systems in the zebrafish model has been assessed with a growing list of agents (301, 522).
Neurobehavioral effects of a variety of toxicants, drugs, and alcohol have been evaluated in developing or adult zebrafish (491). Samson et al (447) found that impairment of swimming and predator/prey behavior were much more sensitive indicators of toxicity in zebrafish exposed as early life stages to methylmercury than were mortality or morphologic lesions. Both early and recent studies have investigated pathologic lesions and functional impairment of developing zebrafish exposed to ethanol (51, 115, 291). Gerlai et al (181) are using zebrafish to assess genetic factors predisposing lines of fish to alcohol preference. Darland and Dowling (114) are screening mutant lines of zebrafish to identify those with increased preference for exposure to cocaine and for altered responses to cocaine. Thomas (506) studied effects of tetrahydrocannabinol (THC) from marijuana in developing zebrafish.
Assays for assessment of disease resistance and immune competence are developed to a much greater extent in the medaka (34, 78, 586) than in zebrafish. Treatment of zebrafish with zinc or copper did not consistently increase their susceptibility to intraperitoneally injected Listeria (435). Zinc exposure suppresses the humoral immune response against Proteus vulgaris, but not against infectious pancreatic necrosis virus in adult zebrafish (449).
Surrogate tests to predict chronic toxicity are gaining favor for environmental monitoring of human health and other sentinel species. Data regarding several biomarkers is available in zebrafish. Induction of the cytochrome P4501a1 (Cyp1a1) enzyme has been extensively used to indicate exposure of fish to halogenated hydrocarbons and polycyclic aromatic hydrocarbons. Activity of Cyp1a1 is induced in early life stages as well as adult zebrafish, and in zebrafish cell cultures (13, 69, 102, 214, 353, 518) following exposure to TCDD. A variety of assays of DNA damage may indicate exposure of fish to mutagens (126). Transgenic fish containing shuttle plasmid vectors serve as efficient indicators of exposure to environmental mutagens (3, 4). Schnurstein and Braunbeck (453) have developed an in vitro system using hepatocytes and gill cells to detect mutagenic agents. Troxel et al (519) measured DNA adducts in adult zebrafish injected with the carcinogen aflatoxin B1 (AFB1). Zebrafish are quite resistant to AFB1-induced neoplasia compared to the most sensitive vertebrate species, the rainbow trout. Surprisingly, adduct levels in zebrafish were just 4-fold less than those in the highly sensitive rainbow trout, and were comparable to adduct levels occurring in relatively sensitive mammalian species. Hsu and Deng (234) quantified adduct levels in various organs of zebrafish following bath exposure to benzo[a]pyrene. They found highest adduct levels in intestine and liver. Several investigators have measured plasma vitellogenin levels in zebrafish as an indicator of exposure to environmental estrogens (8, 522).
Carvan et al (80) developed transgenic lines of zebrafish with response elements to indicate exposure to specific toxicants. The constructs introduced into the zebrafish include aryl hydrocarbon, electrophile, metal, and estrogen response elements. Each reponse element is coupled to a reporter such as luciferase or green fluorescent protein. Zebrafish cell lines containing such reponse elements coupled to reporter genes are also available (81). Zebrafish embryos are sensitive indicators of toxic components in industrial and municipal waste and landfill leachates (169, 170, 243). Chronic bioassays using whole life-cycle tests with zebrafish can detect mortality, growth suppression, reproductive, and developmental toxicity, as well as behavioral toxicity (472).
Although the zebrafish was the first fish species in which experimental carcinogenesis was conducted (by Mearle Stanton of the National Cancer Institute in the 1960s; 481, 482), until recently, little additional neoplasia research has used zebrafish (269, 411, 412).
The most comprehensive carcinogenesis studies to date with zebrafish were conducted at Oregon State University (OSU) with funding from the US Army. This research evaluated carcinogenicity of a panel of structurally diverse carcinogens at three development stages in Florida wild-type zebrafish. Carcinogens evaluated in this work included N-nitrosodiethylamine (DEN), aflatoxin B1 (AFB1), methylazoxy-methanol acetate (MAMA), N-methyl-N′-nitro-N-nitrosoguanidine (MNNG), and 7,12-dimethyl-benz[a]anthracene (DMBA). We administered carcinogens by bath to late stage embryos or fry at 2–3 weeks postfertilization. In dietary carcinogenesis studies we fed 2-month-old juvenile fish a semipurified diet (300) containing carcinogen for 3–9 months. For each carcinogen and the 3 developmental stages, we treated fish with at least 3 graded doses of carcinogen up to the maximum tolerated dose.
Zebrafish exposed to most of the carcinogens as fry or embryos display a wide variety of neoplasm types derived from epithelial, mesenchymal, neural, and neural crest tissues. Liver is a primary target organ for most carcinogens, regardless of developmental stage of fish at exposure. We see the greatest diversity of histologic types of neoplasia following early life stage exposure to MAMA, MNNG, or DMBA (212, 476-478, 521). Table 1 indicates the patterns of target organs observed in our carcinogen studies with Florida wild-type zebrafish. Zebrafish of the Florida wt strain are remarkably resistant to carcinogenic effects of AFB1 at all life stages. In dietary studies, neoplasia was not observed following treatment with MNNG or DEN, and substantial incidences of neoplasia occurred with AFB1 only after dietary exposure for 9 months to 100 ppm. Typically low ppb concentrations of AFB1 are used for dietary carcinogenesis studies in rainbow trout to achieve high incidences of liver neoplasia (27). The relative resistance of juvenile zebrafish to most dietary carcinogens when compared to juvenile rainbow trout fed the same carcinogens may occur in part because the growth rate of zebrafish slows substantially as they approach maturity at 3–6 months of age, whereas trout continue to grow at a more rapid rate throughout life. The reasons for the relative resistance of Florida wild-type zebrafish to AFB1 at all life stages are not yet clear. Surprisingly, adduct levels following adult exposure were just 4-fold less than those observed in rainbow trout, and were comparable to adduct levels occurring in relatively sensitive mammalian species (518, 519). Although maximal percent incidences of total neoplasia in the most sensitive life stage were 67, 66, 62, 37, and 32 for DEN, DMBA, MAMA, AFB1, and MNNG, respectively, at most a 32% incidence of liver neoplasia or 15% incidence of gut neoplasia occurred in carcinogen treated fish.
Beckwith et al (36) of Pennsylvania State University report 100% incidence of cutaneous papillomas occurring in 18 zebrafish of the Florida wild-type line within 1 year following 3 adult bath exposures to 2.5–3 mM ENU. We hypothesized that early life stage exposure to ENU might speed up the time to tumor development, and also predicted that the Tuebingen long fin leopard (TL) mutant line with fin overgrowth would likely develop more papillomas at a faster rate on fins than wild-type lines. We administered single 1 hr bath exposures of the maximum tolerated dose of ENU (2.5 mM) to 3-week-old TL fry, and administered 3 doses of 2.5 mM ENU to Tuebingen wild-type (TU) fry at 3, 5, and 7 weeks of age. At 1 year posttreatment in the TL line, and at 1 and 2 years posttreatment in the TU line, we observed no papillomas. In carcinogen-treated fish of both lines, we documented a variety of neoplasm types including hemangiomas and hepatic and neural neoplasia not observed in vehicle or sham control fish of those lines (Spitsbergen and Kent, unpublished data). Several explanations are possible for the differences in our findings in the TL and TU lines at OSU compared to the Florida wild-type line at Pennsylvania State University. Perhaps only the Florida wild-type line has skin that responds to ENU. Perhaps ENU is carcinogenic in zebrafish only after adult exposure of specific genetic lines. Perhaps the Penn. State University zebrafish colony has a unique oncogenic virus and/or tumor promoter contributing to papilloma development. The zebrafish colony at the University of Oregon houses large numbers of zebrafish of the AB line treated as adults with a similar ENU protocol to that used at Pennsylvania State University. Yet cutaneous papillomas are seen at the University of Oregon only in eggbound female broodstock with prolapse of the distal intestine and subsequent chronic irritation of the vent, resulting in hyperplasia or papilloma of the skin of the vent.
In carcinogen studies at OSU, the spontaneous rate of neoplasia in the Florida wild-type line at 6–14 months of age is 1%, based on 3,000 untreated controls. The most common histologic types of spontaneous neoplasia in this line are seminoma of testis, hepatocellular adenoma, and adenoma of exocrine pancreas, with intestinal adenocarcinoma less common (476-478).
Michael Kent and Jan Spitsbergen of OSU and Monte Westerfield of Univeristy of Oregon are currently investigating spontaneous rates of neoplasia in wild-type and mutant lines of zebrafish. We are evaluating the relative roles of diet, husbandry systems, infectious agents, and genetic influences on rates of neoplasia in various zebrafish lines. As part of this research, we are comparing spontaneous tumor incidences at 2 years of age in replicate groups of AB wild-type zebrafish fed 2 different diets (commercial diet or semipurified diet) and raised at 3 different facitilites at Oregon State University and University of Oregon. We have begun to systematically examine large numbers of retired broodstock histologically over the past year. Since 1999, we have studied diagnostic cases from moribund fish, fish with gross lesions, or sentinel fish submitted to the diagnostic pathology service provided by the Zebrafish International Resource Center at the University of Oregon. We have examined approximately 500 fish in the diagnostic service. Among these diagnostic cases, the most common target tissues for neoplasia are ultimobranchial gland (26/500), testis (22/500), liver (15/500), gut (14/500), peripheral nerve (11/500), and thyroid (6/500). Less common target tissues in diagnostic cases are eye (5/500), nasal sensory neuroepithelium (3/500), blood vessel (2/500), fibroblast (1/500), brain (2/500), gill (1/500), lymphomyeloid system (1/500), pancreas (2/500), notochord (1/500), and pigment cell (1/500). In these studies of spontaneous neoplasia together with our ongoing and historical carcinogen studies, benign and malignant neoplasms occur in nearly every tissue of zebrafish. Tissues in which we have not yet observed neoplasia are pituitary, glial cells of brain, ovary, and the granulocytic lineage of blood (have seen myelodysplasia with granulocytes predominating). Sample sizes in our studies of retired broodstock are currently too small to draw clear conclusions. Fewer than 100 fish per line have been examined from most lines, often of a single cohort raised together in one facility. So far the spectrum of tumor types in most broodstock lines, wild-type or mutant, is similar to that in diagnostic cases, with tumors of testis, liver, and gut most common. AB wild-type zebrafish 1.5–2 years of age show incidences of total neoplasia of approximately 10–15%. Up to 50% of males of the AB line develop seminomas of testis by 1.5–2 years of age.
Medical doctor pathologists, veterinary pathologists and fish pathologists are collaborating to complete a review paper summarizing current information on neoplasia in zebrafish. This information was presented in poster form at scientific meetings (37, 38). This manuscript should be submitted within months, and information with images will be available on the World Wide Web through ZFIN.
Investigations of spontaneous and carcinogen-induced neoplasia in zebrafish reveal a variety of neoplasms that rarely occur even in carcinogen-treated vertebrates of other species (chordoma, pineocytoma, hepatoblastoma, ocular medulloepithelioma, olfactory esthesioneuroepithelioma). Development of zebrafish lines that show high incidences of specific neoplasms early in life will offer an efficient, cost-effective system for investigating molecular alterations occurring during various phases of the evolution of tumors and for developing novel anticancer therapy.
Jan Spitsbergen, Michael Kent, and Donald Buhler of Oregon State University are investigating carcinogen-induced neoplasia in various wild-type and mutant lines of zebrafish in order to identify lines that develop high incidences of specific histologic types of neoplasia early in life. We have identified several mutant lines that develop 70–90% incidences of liver neoplasia within 6–12 months posttreatment. Certain of these lines rapidly develop highly anaplastic neoplasia, including hepatoblastoma. Some of these mutant lines are susceptible to relatively high incidences of multiple tumor types including hemangioma/hemangiosarcoma, esthesioneuroepithelioma/esthesioneuroblastoma of nose, and myelodysplastic syndrome of hemopoietic tissue. Our finding of myelodysplasia at a relatively high incidence (50% incidence or greater) in mutant lines following carcinogen exposure is particularly exciting because, for unexplained reasons, myeloid hemopoietic neoplasia has not yet been reported in untreated or carcinogen-treated wild-type zebrafish under the conditions studied to date (475).
Leonard Zon and colleagues at Children's Hospital in Boston, Massachusetts are investigating carcinogen responsiveness of mutant lines of zebrafish selected for altered expression of cell cycle genes in embryos and young fry. Several mutant lines show increased cumulative incidences in neoplasia following fry bath treatment with MNNG or DMBA, compared to wild-type siblings when sampled at 3, 6, and 12 months posttreatment (464). Keith Cheng and colleagues at Pennsylvania State University have developed lines of zebrafish with genomic instability and are investigating neoplasm prevalences in these lines (359, 360). Cheng's group is also investigating lines of zebrafish showing dysplasia of various tissues at 7 days of development, predicting that such lesions will be associated with elevated neoplasm risk later in life (88). Thomas Look and colleagues at Harvard Medical School are developing a model for neuroblastoma by engineering overexpression of zebrafish mycn (420).
Functional motifs in oncogenes and tumor suppressor genes tend to be highly conserved among vertebrate species, so that study of oncogenes and tumor suppressor genes in zebrafish will likely provide valuable insights into molecular mechanisms of tumorigenesis and genetic susceptibility to tumors in humans (124, 525). Because interest in neoplasia and carcinogenesis in the zebrafish model has only recently begun to dramatically expand, unfortunately relatively few of the zebrafish orthologs of human tumor suppressor genes and oncogenes have been cloned, fully sequenced and/or mapped. Single orthologs of the human tumor suppressor genes WT1 (266, 471), P53 (90), MEN 1 (268), MADH2, MADH5 (127), PTC1, and PTC2 (305) are fully sequenced and mapped in zebrafish. Full cDNA sequence and mapping data are available for single orthologs of human oncogenes NRAS (89), RET (48), CMYC (455), CKIT (395, 421), PIM1 (239) and MDM2 (504). In contrast to findings in Xenopus, overexpression of mdm2 in developing zebrafish does not cause early increased tumor incidence (504).
There are scant baseline data on zebrafish pathologic lesions in infectious and noninfectious diseases. Shortly after the widespread use of mutant mice began, industry, academia, and government agencies realized that molecular biology and genetic approaches alone could not adequately predict and systematically characterize the complex phenotypes of mice with single or multiple gene mutations. For example, many lines of knockout mice for which the inactive genes are not related directly to the immune system still are immunodeficient and at high risk for opportunistic infection (60, 433, 575). Thus rigorously trained comparative pathologists are playing a growing role in research with mutant mice. To utilize the potential of zebrafish fully as an animal model for understanding human development, disease, and toxicology, we must greatly advance our knowledge of zebrafish diseases and pathology. Much of the information pertinent to zebrafish infectious diseases is in general reference texts regarding the topic of aquarium fish diseases (379, 487). Since 1999, Michael Kent of Oregon State University, Jennifer Matthews of the University of Oregon, and collaborators have investigated epidemiology, prevention and control of important infectious and noninfectious diseases affecting zebrafish colonies (50, 334).
Two infectious diseases that commonly occur in well-managed zebrafish colonies are microsporidiosis and mycobacteriosis. Pseudoloma neurophila, a microsporidian infects the central nervous system, cranial and spinal nerves, and skeletal muscle of zebrafish. Severely affected fish may be emaciated, ataxic, or have spinal malformations (123, 333). Michael Kent's research group has developed a polymerase chain reaction test to screen broodstock and derive specific pathogen-free lines. We are uncertain whether vertical transmission of microsporidia can occur in zebrafish. Unfortunately, these parasites are very difficult to inactivate and remove from recirculating husbandry systems. Zebrafish of various lines derived from eggs obtained in the main colony or new Zebrafish International Resource Center Colony at the University of Oregon do not develop microsporidiosis when reared in a flow-through husbandry system at Oregon State University, but do develop microsporidiosis when raised in the recirculating systems at University of Oregon. Thus horizontal transmission of this parasite seems to be important, however, the role of vertical vs horizontal transmission is still unclear. Renibacterium salmoninarum, an important bacterial disease of salmonids, is spread vertically within eggs, but only a few eggs are usually infected. Yet these few infected eggs serve as a source for horizontal transmission in early life stages held in incubation trays (151). Pseudoloma presents particular problems for neurobehavioral toxicity studies. Although Kent's group is investigating possible drug therapy, no recommended therapy methods are yet available.
Piscine mycobacteriosis remains a tenacious problem in aquarium fish colonies. It most often occurs as an opportunistic infection in fish over 1 year of age, especially those stressed by toxicant exposure or adverse environmental conditions. Therefore, mycobacteriosis is a problem in long-term tumor or aging studies. Highly pathogenic strains may cause devastating mortality in otherwise healthy colonies. The sources of infection and epizootiology are not well defined. Mycobacteria may enter fish from food or water, and may possibly be vertically transmitted. Currently no effective chemopreventative or therapeutic regimens are defined for fish (23). Because mycobacterial antigens are potent immune adjuvants, these agents can seriously confound research in disease resistance or immune responses.
Pseudocapillaria tomentosa, a nematode parasite, infects the gut of zebrafish. This parasite occurs commonly in colonies of Florida wild-type lines. Infection causes moderate to severe multifocal to diffuse hyperplasia and dysplasia of the intestine, as well as elevated incidences of intestinal neoplasia. Enteric adenomas and carcinomas occur in close proximity to profiles of nematodes in the gut. The parasite acts as a tumor promoter in carcinogen experiments. In dietary carcinogen studies, zebrafish fed DMBA developed more intestinal neoplasia if infected with the nematode (267).
To date, a pathogenic viral agent has not yet been isolated from or visualized in zebrafish tissues. We are looking for viral agents with ultrastructural studies of fish tissue, viral isolation procedures in fish cell lines, and disease transmission trials in which young zebrafish are injected with tumor tissue or tissue from diseased fish. All vertebrate species that have been studied intensely are host to multiple pathogenic viruses.
Zebrafish are certainly affected by pathogenic and oncogenic viruses that will eventually be identified. Salmonid viruses infectious hematopoietic necrosis virus and infectious pancreatic necrosis virus infect zebrafish and zebrafish cell lines (315), but do not cause clinical disease (298, 457).
A small amount of published information is available regarding zebrafish disease diagnostic procedures and infectious diseases. Astrofsy et al (22) outline diagnostic techniques for clinical investigation of laboratory zebrafish. Menudier et al (345) compared pathogenicity of various strains of Listeria in zebrafish and mice, showing that strains nonpathogenic in mice were sometimes highly pathogenic in zebrafish. Pullium et al (419) document epizootic motile aeromonad septicemia in a zebrafish colony associated with stress due to poor water quality. Edwardsiella ictaluri, a common pathogen of channel catfish, was isolated from moribund danio during a disease outbreak (546). Mills (352) reports on ecological factors influencing digenean trematode infections.
Zebrafish colonies fed commercial diets and maintained in standard recirculating systems have different patterns of neoplasia and spontaneous pathologic lesions than those observed in the Core Fish Facility of the Marine/Freshwater Center at Oregon State University. We have a flow-through system and feed fish a semipurified diet used for over 30 years in carcinogenesis studies in fish. We have not yet seen spontaneous mesenchymal neoplasia, spontaneous neoplasia of the central or peripheral nervous system, or spontaneous thyroid neoplasia in any lines of zebrafish in the Oregon State University center. In contrast, such neoplasms occur quite commonly in diagnostic cases from many labs around the world, even in fish just 6 months of age. These findings suggest that naturally occurring carcinogens may be present in the diet or environment in many zebrafish colonies. Our studies comparing neoplasm incidences in zebrafish of the AB line fed commercial or semipurified diets at 3 separate fish facilities will help shed light on the influences of diet and husbandry on neoplasm patterns.
In diagnostic cases from around the world and in about 50% of groups of retired broodstock from standard husbandry systems feeding commercial diets, we see mild to moderate hepatocyte megalocytosis with karyomegaly. From 10 to 100% of the fish in affected lots of broodstock show hepatic megalocytosis. Hepatocyte cytoplasmic volumes and nuclear volumes may be 5–50X normal. We have never seen this megalocytosis in untreated control fish of any line from the Core Fish Facility of the Marine/Freshwater Center at Oregon State University. Data from many vertebrates including zebrafish indicate that hepatocyte megalocytosis is caused by toxicant damage to DNA or the mitotic apparatus (207). The algal toxin microcystin causes hepatic megalocytosis in Atlantic salmon (10). We often see high incidences of hepatocyte megalocytosis in zebrafish following carcinogen exposure. In a given environment, some lines appear more prone to hepatic megalocytosis. In the flow-through quarantine facility at the University of Oregon, the TL line shows megalocytosis, but the TU, Cologne, another long fin, and knorrig lines do not. The toxicant sources causing megalocytosis are uncertain. Dietary components (commercial flake, paramecium cultures), algae or biofilm in hoses/lines, and metabolites of microbes in biofilters may contribute to the problem. In addition to increasing the spontaneous incidences of neoplasia, the toxicant(s) causing megalocytosis may also cause other health problems, reducing early life stage survival, longevity, reproductive potential, immune competence, and disease resistance.
In our studies of retired broodstock at OSU and U. of O. mild to severe multifocal to diffuse hyperplasia of bile ducts is present in essentially 100% of broodstock of the TL line by 1.5 year of age. The incidences and severity of these lesions are similar at a given age in fish raised at Oregon State University or University of Oregon. The lesions are not associated with significant inflammation, parasites or other infectious agents, however, the severity is greater in males than in females. Approximately 10% of TL line fish raised in flow-through conditions at the University of Oregon show biliary neoplasia spontaneously by 1.5 years of age. Bile duct hyperplasia is not linked to the long fin mutant gene, as broodstock are not homozygous for this dominant gene, and wild-type siblings have similar incidences and severity of bile duct hyperplasia. Although it is a double mutant, the TL line has been used as a background genetic line for perpetuating other mutant lines in many zebrafish colonies because it is hardy and fecund, so its background genetics are unique. Although the proliferated bile ducts may occupy 50% of the volume of liver of TL line fish by 1.5 year of age, except for elevated incidences of neoplasia, the fish appear clinically healthy and usually breed well. The bile duct hyperplasia seen in TL line zebrafish is not seen in TU wild-type line or in the another long fin (alf ) mutant line on the TU × AB background. The TL line will be a useful model for study of primary biliary cirrhosis and bile duct neoplasia, spontaneous and carcinogen-induced, but would be a poor line to use in toxicant studies in which subtle alterations in bile ducts must be discerned.
At many zebrafish facilities around the world, problems with skewed sex ratios in cohorts of zebrafish can interfere with natural breeding, and can complicate studies such as carcinogen or other toxicant bioassays where balanced sex ratios in control groups are desired. In contrast to mammalian species, sex determination in fish is more flexible, often reversible, and less dictated by genetic factors (29, 122, 442). The relative influences of and interactions between environment, genetics, and toxicants in sex determination in zebrafish are not yet clearly defined. However, genetic as well as environmental factors appear to influence sex determination. The TU line raised in any of the recirculating or flow-through systems at the University of Oregon tends to be predominately female, whereas the AB line reared in the same systems often has an excess of males. We do not see skewed sex ratios in control groups of Florida wt zebrafish in the Core Fish Facility of the Marine/Freshwater Center at Oregon State University, but we often see very skewed sex ratios with the German lines TU and TL, and the AB line. We do not see unbalanced sex ratios in rainbow trout in this facility.
The literature on nonneoplastic pathologic lesions in zebrafish is sparse compared to that for mammals. A few references on nutritional pathology are available (141, 342, 408). Vithelic and Hyde (540) document light-induced retinal damage in albino zebrafish. Ferretti and Geraudi (154) and Geraudi et al (180) show retinoic acid-induced cell death in healing wounds on regenerating fins. Few of the toxicant studies using zebrafish have investigated light microscopic or ultrastructural morphologic alterations. Hisaoka (225, 226) describe histochemical changes including depletion of hepatic glycogen, and light microscopic morphologic lesions including necrosis of neuroepithelium of the central nervous system occurring in zebrafish following embryo exposure to 2-acetylaminofluorene. Burkhardt-Holm et al (71) report ultrastructural alterations in liver, gill and erythrocytes following chronic exposure of embryonic or larval zebrafish to 4-chloroaniline. Oulmi and Braunbeck (388) describe ultrastructural alterations in liver and kidney in zebrafish following embryo microinjection of 4-chloroaniline.
Braunbeck et al (59) document ultrastructural alterations in liver of zebrafish including reduced numbers of peroxisomes following chronic exposure to 4-chloroaniline. Bresch et al (63) evaluated spinal integrity using radiography during a 3-generation life cycle study with 4-chloroaniline. Braunbeck et al (58) show hepatic steatosis occurring in zebrafish chronically exposed to lindane in a full life cycle test. Strmac and Braunbeck (490) indicate that the cardiovascular system is a primary target following sublethal exposure of early life stages of zebrafish to triphenyl tin acetate. They report ultrastructural alterations in liver including depleted glycogen, mitochondrial swelling, and cytoplasmic myelin figures. Dong et al (129) and Henry et al (215) describe light microscopic lesions in zebrafish during the early life stage toxicity syndrome caused by TCDD. As in early life stages of other species of fish exposed to TCDD, the cardiovascular system is the primary target organ system in zebrafish. Ansari et al (18) and Ansari and Kumar (15) show reduced numbers of mature ovarian follicles, hepatomegaly, and alterations in hepatic nucleic acid and protein (16, 17) in zebrafish following chronic sublethal exposure to malathion. Olsson et al (383) studied histologic lesions in offspring following maternal exposure to PCBs or 17-beta-estradiol. Maternal exposure to PCB-104 or 190 causes craniofacial malformations and scoliosis in offspring. Maternal exposure to PCB-104 or 17-beta-estradiol causes tubular nephrosis, and 17-beta-estradiol, PCB-104, or PCB-60 increases ooctye atresia in ovaries, and arrests spermatogenesis in testes in early life stages of progeny. Histologic alterations occur in gonads of both female and male zebrafish given acute bath exposures as adults to the synthetic estrogen 17-alpha-ethynylestradiol (526). Increased atretic follicles and arrested follicular development occur in the gonads of females. In males, spermatogenesis ceases at the spermatogonial stage.
The utility of early life stages of zebrafish in high throughput screening systems for drug development is already being exploited (369, 371, 404, 484). The small size of zebrafish embryos and fry, and their ability to be cultured during the first week of life in 96 well microtiter plates make this system ideal for drug discovery and safety testing. Most hydrophilic as well as lipophilic agents are readily absorbed from the culture medium of eggs or fry, facilitating efficient testing of new agents.
To advance the field of zebrafish toxicologic pathology toward the state of the art in mammalian toxicologic pathology, much more data regarding pathologic lesions following acute, subchronic and chronic toxicant exposure will be required. We need to develop a comprehensive database regarding spontaneous and toxicant-induced neoplasia and nonneoplastic lesions in the common wt and mutant lines of zebrafish. Spontaneous aging lesions in various strains of zebrafish need to be investigated. Comprehensive data on metabolism and pharmacokinetics of toxicants in various wt and mutant lines will be essential to support sophisticated toxicologic pathology research. Very little information is available on DNA repair enzymes in various lines of zebrafish. Once the full genomic sequence is available for zebrafish during the coming year, this information will speed the identification of and study of DNA repair enzymes. The cornucopia of markers for protein and gene expression generated for studies of zebrafish development need to be exploited in toxicant studies. To best utilize the zebrafish as a model system for understanding mechanisms in oncogenesis relevant to man, we must clarify the members of families of oncogenes and tumor suppressor genes to sort out their relative roles in tumorigenesis in zebrafish compared to mammals. Duplicate genes for some members of these families are likely to complicate the clarification of signaling pathways in the zebrafish. Microarray technology will likely help in unraveling the complex signaling networks in zebrafish.
In recent years, the Society of Toxicologic Pathology has coordinated working groups of pathologists focused on specific organ systems in rodents. The poster sessions at STP meetings and STP-sponsored monographs on proliferative and nonproliferative lesions for each organ system have established consensus in the field of toxicologic pathology regarding terminology for morphologic diagnosis of pathologic lesions in rodents. We need to establish similar pathology working groups to define diagnostic criteria for proliferative and nonproliferative lesions in the major organ systems of zebrafish. An inadequate number of scientists have sufficient training in anatomy, histology, and pathology of fish including zebrafish to support the growing need for pathology expertise in the study of zebrafish models for human disease. Some short courses in fish diseases and pathology such as the Aquavet program taught each spring at the Marine Biological Laboratory in Woods Hole, Massachusetts can provide an introduction to fish pathology for scientists who already have basic histology and pathology training.
However, more extended residency and/or graduate training for D.V.M. and M.D. pathologists as well as fish biologists will be required in order to establish adequate diagnostic skills to correctly interpret histologic lesions in fish tissues.
Research at Oregon State University and University of Oregon was funded by US Public Health Service grants ES011587-01 and P30ESO3850 from the National Institutes of Environmental Health Sciences, by grant 3P40RR12546 and its supplement 03S1 from the National Center for Research Resources, and by US Army contract DAMD 1791Z1043. We thank Tom Miller, Keri St. Clair, Janelle Bishop-Stewart, Sheila Cleveland, Chance MacDonald, Karen Larison, April Mazanac, Bill Trevarrow, and Dan Arbogast for technical support in our ongoing and past zebrafish studies.