- Zebrafish (Danio rerio) have become an increasingly important model organism for cancer research. There are several cancer models in which neoplasms are induced chemically or genetically. However, spontaneous (unknown etiology) neoplasms are rather common in zebrafish, particularly in older fish. For over a decade, spontaneous intestinal neoplasia and preneoplastic intestinal changes including hyperplasia, dysplasia and enteritis have been observed in zebrafish from several laboratories submitted to the ZIRC (Zebrafish International Resource Center) diagnostic service. A retrospective analysis showed that 2% of the total fish submitted to the service demonstrate these lesions. These affected fish were submitted from 18 facilities from laboratories in the United States and other countries. One facility (referred to as the primary facility) was particularly affected, representing approximately 74% of the fish with lesions. Tumor prevalence appeared similarly distributed between sexes and generally occurred in zebrafish greater than one year of age, although neoplastic changes were observed in fish as young as 6 months. Eleven lines displayed these preneoplastic and neoplastic changes, including wild-types and mutants. Zebrafish submitted as normal sentinel fish affected with these intestinal lesions demonstrated that these lesions are most often subclinical and emphasize that as an extra-experimental variable these lesions could have underappreciated effect in research.
Given the distribution of the affected fish defined from the retrospective study, possible etiologic agents include a variety of environmental elements or infectious agents. To further investigate potential etiologic agents versus a carcinogen in the diet, a diet challenge was conducted previously by feeding the same diet fed at the primary facility to fish held at a different location. No appreciable preneoplastic or neoplastic lesions developed in the experimental fish from this study, indicating that diet is not the cause. Furthermore, lesion prevalence between different populations of age and strain matched fish at the primary facility were dramatically different. This indicates that a water-borne carcinogen in the juvenile and adult rearing laboratories not the cause, as all of the affected fish were held in the same water supplies in recirculating systems at each facility.
An infectious agent, therefore, could be the cause, as both bacteria and viruses are recognized as causes of cancers in fishes. I conducted a transmission study to investigate this etiology. The experiment included subjecting naïve fish to cohabitation with affected fish, feeding of a tissue homogenate from a pool of affected fish, and intraperitoneal (IP) injection of filtered and non-filtered affected tissue homogenate. Two fish from the cohabitation study sampled at 8 mo. and two fish from the injected filtrate study sampled at 7 mo. exhibited early hyperplastic changes. However, comprehensive histological evaluation at the termination (10 mo. post-exposure [11 mo. post-exposure for cohabitation fish]) of the study did not recognize any preneoplastic or neoplastic intestinal lesions. However, the study was terminated earlier than the original protocol due to high mortality attributed to Piscinoodinium pillulare infections amongst the fish in the study.
I conducted a comprehensive histological review of these neoplasms through a retrospective examination of 9,508 zebrafish provided by the ZIRC diagnostic service. The neoplasms were classified either as adenocarcinoma or
small cell carcinoma, with a few exceptions (carcinoma not otherwise specified,
tubular adenoma, and tubulovillous adenoma) based upon histomorphologic presentation. In mammals, these neoplasms usually arise from the gastrointestinal epithelium or neural-endocrine cells (e.g., carcinoids). Hence we proposed that cells of origin for these neoplasms in zebrafish were either intestinal epithelial or gut-derived neural endocrine origin. We subjected tissue sections of several of these neoplasms to a panel of mammalian antibodies directed toward epithelial (Cytokeratin Wide Spectrum Screening [WSS], AE1/AE3) or neural (S100, and chromogranin A) tissues. We also investigated the specificity of these antibodies using Western blot analysis, comparing human and zebrafish profiles. WSS and AE1/AE3 were relatively reactive with approximately half of the neoplasms analyzed (staining positive with these markers). S100 and chromogranin A (neural markers) did not specifically stain the cells within the neoplasms. The positive cytokeratin association (WSS and/or AE1/AE3) for most of the neoplasms, while negative for neural and neuroendocrine markers (S100 and chromogranin A respectively), indicates that these intestinal neoplasms are of common epithelial origin. Perhaps those that were negative for cytokeratin may have further progressed to a state of dedifferentiation as most of the neoplasms that were negative were classified as small cell carcinomas and the cell type of these neoplasms are characterized as cells with a small amount of cytoplasm and lacking features of typical epithelial cells.
In conclusion, based on data to date, the neoplasms are epithelial in origin, and three plausible causes for the lesions should be considered; 1) a microorganism in adult facilities, 2) a microorganism in larval nurseries, and 3) a chemical carcinogen in individual nursery tanks.