Development of a Hypoxia-Induced Zebrafish Choroidal Neovascularization Model - Zebrafish: Methods for Assessing Drug Safety and Toxicity - page 208

accompanied by edema in the surrounding tissue, a phenotype present in AMD in
humans. In normal zebrafish, there are no vessels present in the retina. In the CNV
model, immunohistochemistry demonstrated that angiogenic vessels were present in
the retina, indicating that neovascularization is a pathological phenotype. Further-
more, the ISV pattern was normal, indicating specificity of CoCl 2 -induced neovas-
cularization in CVP and absence of toxicity.

16.4.2 Isolation of Eyes from Whole Zebrafish Using
Collagenase Treatment Facilitates CNV Assessment
and Quantitation of Drug Effects

Since the CVP is located in the back of the eyes, examination of abnormal
angiogenesis is difficult in intact, whole mount Phy-V-stained zebrafish. In order to
optimize assessment of abnormal eye angiogenesis, we dissociated intact eyes from
whole zebrafish. Initially, we used forceps to extract eyes from whole animal;
however, this procedure was tedious and often caused eye damage. To loosen eyes,
we then tried collagenase treatment on whole mount immunostained zebrafish
followed by gentle shaking. This method permitted examination of abnormal
neovascularization specifically in the CVP region in the back of the eye with no
background signal from other sites in whole zebrafish.

16.4.3 Hypoxia Is a Major Mechanism of Abnormal
Angiogenesis in the Zebrafish CNV Model

We hypothesized that hypoxia is the major mechanism of CoCl 2 -induced CNV in
zebrafish. To support this hypothesis, using RT-PCR and in situ hybridization, we
demonstrated that HIF-1 and VEGF expression was upregulated in CoCl 2 -treated

a

zebrafish eyes.

16.4.4 Inflammation Plays an Important Role in the
Zebrafish CNV Model

We tested three antiangiogenic compounds, Celebrex, genistein, and SU5416, in the
CoCl 2 -induced CNV model. Not only did Celebrex and genistein inhibit abnormal
angiogenesis in CVP, but they also inhibited the edematous inflammatory response.
SU5416, a VEGF receptor inhibitor, is antiangiogenic in mice (Shaheen et al., 1999).
However, in clinical trials, this compound was shown to induce serious toxic side
effects (Kindler et al., 2001). As expected, co-treatment with CoCl 2 and either
Celebrex or genistein prevented CNV formation. In contrast, SU5416 did not. These
results imply that, in addition to angiogenesis, inflammation plays a crucial role in
CNV formation. Assessment of additional compounds involved in both angiogenesis
and inflammation will further support this hypothesis andmay lead to identification of
macular degeneration drug candidates.