FIGURE SUMMARY
Title

Evaluating human cancer cell metastasis in zebrafish

Authors
Teng, Y., Xie, X., Walker, S., White, D.T., Mumm, J.S., and Cowell, J.K.
Source
Full text @ BMC Cancer

Zebrafish faithfully report metastasis in human cancer cells. Confocal imaging (shown as z-stack images using 40 × magnification) shows that (a) metastatic MDA-MB-231 (MDA231) breast cancer cells (above) are dispersed throughout the fish body (arrows), whereas non-invasive T47D cells (below) remain in the confines of the yolk sac. These results are mirrored by the Transwell invasion assays, shown in each case to the right. Similarly (b) invasive DU145 prostate cancer cells show extensive spread throughout the fish compared with non-invasive LNCaP cells which do not. In colon cancer cells (c), highly invasive SW620 cells metastasize extensively compared with non-invasive HT29 cells. Relative invasion potential of pancreatic cancer cells (d) shows that more cells are distributed throughout the fish body for the relatively more invasive AsPC-1 cells compared with less invasive BxPC3 cells. Assessment of fish showing metastasis (see text) in each experiment compared with the invasion potential is shown in (e) showing the close correlation between relative invasion in vitro using the transwell assay and in vivo metastasis in fish. ** P > 0.001; * P >0.05. Using the Tg(kdrl:EGFP) transgenic zebrafish (f), cancer cells (red) that have spread throughout the vasculature (green) can be seen for MDA231 as show by successive imaging of 4 hours (200 × magnification). The presence of MDA231 cells that have extravasated into the body of the fish can clearly be seen (the white arrow tracks movement of a single cell and yellow arrow indicates movement of another).

Metastasis of primary human lung tumor cells in zebrafish. Cells derived from short term culture of a primary human lung tumor with an undifferentiated phenotype (#8) show extensive spread throughout the body of the fish (a, left) and invasion in vitro (a, right). In contrast, cells from a well-differentiated adenocarcinoma of the lung (#9) show only minimal dissemination in the fish and limited invasion in vitro. From cohort studies (b), the number of fish showing metastasis is significantly greater in #8 (shown as z-stack images using 40 × magnification). This relationship is maintained in in vitro invasion assays (b, right).

Comparable WASF3-induced metastasis of DU145 cells in mice and fish. Knockdown of WASF3 in DU145 cells using two different shRNA constructs (shW3-1 and shW3-2) shows significant reduction in protein levels (a). DU145 cells transfected with a control shRNA (shGFP) show high level invasion in vitro(b) but knockdown of WASF3 in these cells using two different shRNAs (shW3-1 and shW3-2) leads to significantly reduced invasion. In fish (c), no, or reduced numbers, of WASF3 knockdown cells (shW3-1 and shW3-2) spread throughout the body of the fish, whereas the control cells (shGFP) spread extensively (arrows) (shown as z-stack images using 40 × magnification). Quantitation of invasion potential shows a significant reduction (** p >0.001) in the WASF3 knockdown cells compared with control shGFP cells (d, above). This reduced invasion is mirrored by the metastasis efficiency in cohort studies of fish (d, below). In mice (e), tumor nodules (arrows) on the surface of the lungs can be seen in the control shRNA treated cells compared with cells expressing shW3-1 (left) 3 months after tail veil injection. Histological examination of the lungs of these mice shows large tumor foci from the control cells (arrows) compared with small tumor foci for the WASF3 knockdown cells.

Quantitation of metastatic cells in zebrafish using Fiji. Fluorescent (a, left) and bright field (a, center) images of Dil-stained (red) human MDA231 cancer cells are captured by confocal microscopy. Individual red cells are counted relative to their location throughout the fish (a, center) using Fiji. Cells within the yolk sac (boxed) can be eliminated from the cell count (a, right). Cells were counted using this approach (b) either throughout the body of the fish (left) or confined to the tail region (right) in the same fish over a 48–120 hours period after injection. Cells visualized by the fluorescent dye (arrows) and following Fiji analysis (right) were plotted from the same 10 different fish at the three time points showing (c) no significant difference in mean number of cells counted in the whole fish body (left) up to 80 hours. Although the number of metastatic cells was reduced when cells restricted to the tail region were counted (right), there was again no significant difference in metastatic cells up to 80 hours after injection. At later stages (120 hpi), an ~25% reduction in metastatic cell numbers was observed in both the fish body and the tail alone. Data are presented as the mean of three independent experiments (n = 3) ± SEM, P < 0.05.

Relative metastatic potential in human cancer cells during progression from low to high grade tumors. When immortal, non-transformed MCF10A cells (M-I) were introduced into the zebrafish metastasis model (a), no spread was seen in any of the fish analyzed at 60 hpi. The cells (M-II) recovered from low grade hyperplastic lesions in mice from MCF10AT cells (MCF10A cells transformed with activated HRAS) show increased metastatic potential (arrows). M-III cells, which were derived from well differentiated carcinomas derived from M-II cells, show reduced metastatic potential, whereas M-IV cells, recovered from aggressive adenocarcinomas derived from M-II cells, show enhanced metastatic potential. The same relationship was seen at 120 hpi (b) (shown as z-stack images using 40× magnification). Quantitation of metastatic cells throughout the fish body using Fiji (c) confirms the relationship between tumor grade and metastatic potential in fish (left). At 120 hpi, although the relative proportion of metastatic cells is seen in the M-II/M-III/M-IV comparison, the overall number of cells in each tumor grade was reduced. When the same comparison was performed using the in vitro invasion assay (d), the relative proportion of invading cells in the different sub groups was again maintained. Data are presented as the mean of three independent experiments (n = 3) ± SEM, **P < 0.01.

Suppression of invasion in vitro leads to suppression of metastasis in zebrafish following loss of JAK1 or JAK2 expression. Western blot analysis confirmed JAK-deficiencies in U4C and γ 2A cells (a). When human 2C4 fibrosarcoma cells are injected into zebrafish (b, left) (shown as z-stack images using 40× magnification), extensive spread throughout the body of the fish can be seen over 120 hpi. This metastasis potential correlates with in vitro invasion (b, right). In both JAK1-deficient U4C cells and JAK2-deficient γ2A cells, reduced levels of metastatic spread are seen (c, left), which correlates with reduced invasion in vitro (c, right).

Acknowledgments
This image is the copyrighted work of the attributed author or publisher, and ZFIN has permission only to display this image to its users. Additional permissions should be obtained from the applicable author or publisher of the image. Full text @ BMC Cancer