Jason N. Berman

Zebrafish xenografts as a tool for in vivo studies on human cancer

The zebrafish has become a powerful vertebrate model for genetic studies of embryonic development and organogenesis and increasingly for studies in cancer biology. Zebrafish facilitate the performance of reverse and forward genetic approaches, including mutagenesis and small molecule screens. Moreover, several studies report the feasibility of xenotransplanting human cells into zebrafish embryos and adult fish. This model provides a unique opportunity to monitor tumor‐induced angiogenesis, invasiveness, and response to a range of treatments in vivo and in real time. Despite the high conservation of gene function between fish and humans, concern remains that potential differences in zebrafish tissue niches and/or missing microenvironmental cues could limit the relevance and translational utility of data obtained from zebrafish human cancer cell xenograft models. Here, we summarize current data on xenotransplantation of human cells into zebrafish, highlighting the advantages and limitations of this model in comparison to classical murine models of xenotransplantation.

Leukaemia xenotransplantation in zebrafish – chemotherapy response assay in vivo

As the molecular understanding of cancer pathogenesis has evolved, so has the need for innovative assays to identify novel targeted therapeutic agents. Characterization of specific tumour-drug interactions would enable the design of more personalized treatment with improved outcomes and reduced toxicity. The zebrafish has emerged as a robust animal model of human malignancy due to conserved genetics and cell biology (Amatruda & Patton, 2008; Payne & Look, 2009). Xenotransplantation of human cancer cells into zebrafish was initiated in 2005 using a melanoma xenograft (Lee et al, 2005). Subsequent studies determined that incubation at 35 C enabled normal growth of injected human cell lines without compromising zebrafish embryogenesis and established the yolk sac and 48 h post-fertilization (hpf) as the ideal anatomic location and developmental stage for injection (Haldi et al, 2006). At this stage, adaptive immune responses are not yet established in the embryo, permitting injection of human cells without requiring immunosuppression (Lam et al, 2004; Haldi et al, 2006; Nicoli et al, 2007). Proliferation of xenografted melanoma cells was documented using a haemocytometer (Haldi et al, 2006); while proliferation of xenografted leukaemia cells was determined by fluorescence and embryo demise (Pruvot et al, 2011). These reports provided a critical foundation, however further refinement and proof-of-principle studies are necessary for this approach to be practically applied to cancer drug discovery and clinically employed for therapeutic drug responses. To this end, we transplanted K562 and NB-4 human leukaemia cell lines into casper embryos, a zebrafish combinatorial pigment mutant (White et al, 2008), and developed a robust cell proliferation assay demonstrating in vivo targeted therapeutic inhibition of these leukaemias. Five million K562 or NB-4 cells were pelleted by centrifugation (5 min at 100 g), resuspended in phosphatebuffered saline (PBS) containing CM-DiI (Invitrogen, Burlington, ON, Canada) (5 lg/ml) and incubated for 5 min at 37 C and 20 min at 4 C. Cells were pelleted, washed twice with PBS, and re-suspended in 500 ll RPMI medium. Dechorinated 48-h casper embryos (adults maintained as described by Westerfield, 1995) were anaesthetized with tricaine and 25–50 labelled cells injected into the yolk sac (adapted from (Haldi et al, 2006)). Following 1 h recovery at 28 C, embryos were maintained at 35 C, and screened for fluorescence at the injection site. At 24 h postinjection (hpi), imatinib mesylate (IM), all-trans retinoic acid (ATRA), or vehicle (dimethyl sulfoxide, DMSO) were added to the water for 48 h. Proliferation of leukaemia cells was monitored by live-cell microscopy. Embryos (15–20) were dissociated at 24 and 72 hpi with or without drug, incubated in protease solution (0Æ25% trypsin, 1 mmol/l EDTA, 2Æ25 mg/ml collagenase (Sigma-Aldrich, Oakville, ON, Canada) in 1Æ2 ml sterile PBS) for 45 min at 35 C and homogenized. Cells were pelleted by centrifugation (400 g · 5 min) and resuspended in 0Æ9· PBS with 5% fetal bovine serum at 10 ll per embryo (Fig. 1). Droplets were imaged as a 5 · 4 mosaic (Axio Observer Z, Colibri LED light source, Axiocam Rev 3.0 CCD camera, Carl Zeiss, Wetzlar, Germany) (Fig. S1). Leukaemia cells were counted in four internal fields using a semi-automated macro (image j; NIH, Bethesda, MD, USA). Co-staining with 10 nmol/l DRAQ5 viable nuclear stain confirmed enumeration of intact leukaemia cells (Biostatus Ltd. Leicester, UK) (Fig. S2). Using four micrographs from five boluses repeated in triplicate, average cell counts per embryo were determined. The use of zebrafish in these experiments was approved by the Dalhousie University Committee on Laboratory Animals. Injected embryos tolerated the presence of human leukaemia cells for up to 7 days, during which time engrafted leukaemia cells proliferated and circulated within the embryonic vasculature (Fig. 2A; Movies S1 and S2). K562 cell numbers increased from 90 to 230 cells per embryo (2Æ6fold) and NB-4 cells from 50 to 170 per embryo (3Æ4-fold) after 48 h (Fig. 2B). To evaluate this xenotransplantation assay as a means of determining specific tumour-drug interactions in vivo, we exploited the known responses to targeted therapies of the molecular alterations in each leukaemia cell line. Specifically, we used IM to target BCR-ABL1 in engrafted K562 cells and ATRA to target PML-RARA in engrafted NB-4 cells. Toxicity curves were generated for both 48-h K562 xenotransplanted zebrafish embryos exposed to IM and NB-4 xenotransplanted embryos treated with ATRA (Fig. S3). We subsequently used 50% of the maximum tolerated dose (MTD) (20 lmol/l IM; 0Æ2 lmol/l ATRA). K562 xenotransplanted embryos exposed to IM or NB-4 xenotransplanted embryos exposed to ATRA for 48 h, displayed a significant decrease in leukaemia cells compared to vehicle-treated control embryos (P < 0Æ001) (Fig. 2C). Reciprocal experiments failed to show a significant inhibition in proliferation (Fig. 2C), confirming that the correspondence

Translational Activation of HIF1α by YB-1 Promotes Sarcoma Metastasis

Metastatic dissemination is the leading cause of death in cancer patients, which is particularly evident for high-risk sarcomas such as Ewing sarcoma, osteosarcoma, and rhabdomyosarcoma. Previous research identified a crucial role for YB-1 in the epithelial-to-mesenchymal transition (EMT) and metastasis of epithelial malignancies. Based on clinical data and two distinct animal models, we now report that YB-1 is also a major metastatic driver in high-risk sarcomas. Our data establish YB-1 as a critical regulator of hypoxia-inducible factor 1α (HIF1α) expression in sarcoma cells. YB-1 enhances HIF1α protein expression by directly binding to and activating translation of HIF1A messages. This leads to HIF1α-mediated sarcoma cell invasion and enhanced metastatic capacity in vivo, highlighting a translationally regulated YB-1-HIF1α axis in sarcoma metastasis.

CRISPR MultiTargeter: A Web Tool to Find Common and Unique CRISPR Single Guide RNA Targets in a Set of Similar Sequences

Genome engineering has been revolutionized by the discovery of clustered regularly interspaced palindromic repeats (CRISPR) and CRISPR-associated system genes (Cas) in bacteria. The type IIB Streptococcus pyogenes CRISPR/Cas9 system functions in many species and additional types of CRISPR/Cas systems are under development. In the type II system, expression of CRISPR single guide RNA (sgRNA) targeting a defined sequence and Cas9 generates a sequence-specific nuclease inducing small deletions or insertions. Moreover, knock-in of large DNA inserts has been shown at the sites targeted by sgRNAs and Cas9. Several tools are available for designing sgRNAs that target unique locations in the genome. However, the ability to find sgRNA targets common to several similar sequences or, by contrast, unique to each of these sequences, would also be advantageous. To provide such a tool for several types of CRISPR/Cas system and many species, we developed the CRISPR MultiTargeter software. Similar DNA sequences in question are duplicated genes and sets of exons of different transcripts of a gene. Thus, we implemented a basic sgRNA target search of input sequences for single-sgRNA and two-sgRNA/Cas9 nickase targeting, as well as common and unique sgRNA target searches in 1) a set of input sequences; 2) a set of similar genes or transcripts; or 3) transcripts a single gene. We demonstrate potential uses of the program by identifying unique isoform-specific sgRNA sites in 71% of zebrafish alternative transcripts and common sgRNA target sites in approximately 40% of zebrafish duplicated gene pairs. The design of unique targets in alternative exons is helpful because it will facilitate functional genomic studies of transcript isoforms. Similarly, its application to duplicated genes may simplify multi-gene mutational targeting experiments. Overall, this program provides a unique interface that will enhance use of CRISPR/Cas technology.

Optimized cell transplantation using adult rag2 mutant zebrafish

Cell transplantation into adult zebrafish has lagged behind mouse models owing to the lack of immunocompromised strains. Here we have created rag2E450fs mutant zebrafish that have reduced numbers of functional T and B cells but are viable and fecund. Mutant fish engraft muscle, blood stem cells and various cancers. rag2E450fs mutant zebrafish are the first immunocompromised zebrafish model that permits robust, long-term engraftment of multiple tissues and cancer.

The Hox cofactors Meis1 and Pbx act upstream of gata1 to regulate primitive hematopoiesis

During vertebrate development, the initial wave of hematopoiesis produces cells that help to shape the developing circulatory system and oxygenate the early embryo. The differentiation of primitive erythroid and myeloid cells occurs within a short transitory period, and is subject to precise molecular regulation by a hierarchical cascade of transcription factors. The TALE-class homeodomain transcription factors Meis and Pbx function to regulate embryonic hematopoiesis, but it is not known where Meis and Pbx proteins participate in the hematopoietic transcription factor cascade. To address these questions, we have ablated Meis1 and Pbx proteins in zebrafish, and characterized their molecular effects on known markers of primitive hematopoiesis. Embryos lacking Meis1 and Pbx exhibit a severe reduction in the expression of gata1, the earliest marker of erythroid cell fate, and fail to produce visible circulating blood cells. Concomitant with a loss of gata1, Meis1- and Pbx-depleted embryos exhibit downregulated embryonic hemoglobin (hbae3) expression, and possess increased numbers of pu.1-positive myeloid cells. gata1-overexpression rescues hbae3 expression in Pbx-depleted; meis1-morphant embryos, placing Pbx and Meis1 upstream of gata1 in the erythropoietic transcription factor hierarchy. Our study conclusively demonstrates that Meis1 and Pbx act to specify the erythropoietic cell lineage and inhibit myelopoiesis.

NUP98-HOXA9-transgenic zebrafish develop a myeloproliferative neoplasm and provide new insight into mechanisms of myeloid leukaemogenesis.

NUP98‐HOXA9 [t(7;11) (p15;p15)] is associated with inferior prognosis in de novo and treatment‐related acute myeloid leukaemia (AML) and contributes to blast crisis in chronic myeloid leukaemia (CML). We have engineered an inducible transgenic zebrafish harbouring human NUP98‐HOXA9 under the zebrafish spi1(pu.1) promoter. NUP98‐HOXA9 perturbed zebrafish embryonic haematopoiesis, with upregulated spi1expression at the expense of gata1a. Markers associated with more differentiated myeloid cells, lcp1, lyz, and mpx were also elevated, but to a lesser extent than spi1, suggesting differentiation of early myeloid progenitors may be impaired by NUP98‐HOXA9. Following irradiation, NUP98‐HOXA9‐expressing embryos showed increased numbers of cells in G2‐M transition compared to controls and absence of a normal apoptotic response, which may result from an upregulation of bcl2. These data suggest NUP98‐HOXA9‐induced oncogenesis may result from a combination of defects in haematopoiesis and an aberrant response to DNA damage. Importantly, 23% of adult NUP98‐HOXA9‐transgenic fish developed a myeloproliferative neoplasm (MPN) at 19–23 months of age. In summary, we have identified an embryonic haematopoietic phenotype in a transgenic zebrafish line that subsequently develops MPN. This tool provides a unique opportunity for high‐throughput in vivo chemical modifier screens to identify novel therapeutic agents in high risk AML.

Hooking the big one: the potential of zebrafish xenotransplantation to reform cancer drug screening in the genomic era

The current preclinical pipeline for drug discovery can be cumbersome and costly, which limits the number of compounds that can effectively be transitioned to use as therapies. Chemical screens in zebrafish have uncovered new uses for existing drugs and identified promising new compounds from large libraries. Xenotransplantation of human cancer cells into zebrafish embryos builds on this work and enables direct evaluation of patient-derived tumor specimens in vivo in a rapid and cost-effective manner. The short time frame needed for xenotransplantation studies means that the zebrafish can serve as an early preclinical drug screening tool and can also help personalize cancer therapy by providing real-time data on the response of the human cells to treatment. In this Review, we summarize the use of zebrafish embryos in drug screening and highlight the potential for xenotransplantation approaches to be adopted as a preclinical tool to identify and prioritize therapies for further clinical evaluation. We also discuss some of the limitations of using zebrafish xenografts and the benefits of using them in concert with murine xenografts in drug optimization.

Zebrafish as a model organism for blood diseases

The zebrafish (Danio rerio) animal model offers a unique opportunity to discover and study novel genes required for the control of normal vertebrate haematopoiesis. It is well suited for both developmental and genetic analysis: as an example, an anaemic mutant zebrafish line isolated in a genome-wide chemical mutagenesis screen has led to the identification of ferroportin 1, the orthologue of which was shown to be mutated in human type IV haemochromatosis. In addition, a zebrafish model of T-cell leukaemia has been established, opening the way for the use of this organism to dissect the aberrant genetic programmes that lead to haematologic malignancies. It should be possible to more precisely delineate molecular pathways responsible for human blood diseases, and thereby validate new target proteins for therapeutic intervention, by using a combination of forward genetic mutagenesis screens and technologies for transgenic and antisense ‘knockdown’ experiments. From its roots as a means to study genetic programmes controlling vertebrate development, the zebrafish is rapidly becoming a valuable animal disease model. The zebrafish complements mouse models because of its inherent forward genetic capacity, which allows the unbiased detection of mutations based on phenotype. This review article will briefly describe the history and features of the zebrafish as a model organism, with a particular focus on advances and applications related to haematopoiesis and haematopoietic diseases.

Hace1 controls ROS generation of vertebrate Rac1-dependent NADPH oxidase complexes

The Hace1-HECT E3 ligase is a tumor suppressor that ubiquitylates the activated GTP-bound form of the Rho family GTPase Rac1, leading to Rac1 proteasomal degradation. Here we show that, in vertebrates, Hace1 targets Rac1 for degradation when Rac1 is localized to the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase holoenzyme. This event blocks de novo reactive oxygen species generation by Rac1-dependent NADPH oxidases, and thereby confers cellular protection from reactive oxygen species-induced DNA damage and cyclin D1-driven hyper-proliferation. Genetic inactivation of Hace1 in mice or zebrafish, as well as Hace1 loss in human tumor cell lines or primary murine or human tumors, leads to chronic NADPH oxidase-dependent reactive oxygen species elevation, DNA damage responses and enhanced cyclin D1 expression. Our data reveal a conserved ubiquitin-dependent molecular mechanism that controls the activity of Rac1-dependent NADPH oxidase complexes, and thus constitutes the first known example of a tumor suppressor protein that directly regulates reactive oxygen species production in vertebrates.