Laser-induced choroidal neovascularization model in mice

Loading...

Taking too long?

Reload document

| Open in new tab

Laser-induced choroidal neovascularization model in mice

Vincent Lambert1, 2, Julie Lecomte2, Sylvain Hansen2, Silvia Blacher2, Maria-Luz Alvarez

Gonzalez2, Ingrid Struman3, Nor Eddine Sounni2, Eric Rozet4, Pascal de Tullio5, Jean Michel

Foidart2, Jean-Marie Rakic1, Agnès Noel2

1Department of Ophthalmology, University Hospital (CHU), Liège, Belgium. 2Laboratory of

Tumor and Development Biology, GIGA-Cancer, University of Liège, Liège, Belgium. 3Unit

of Molecular Biology and Genetic Engineering, GIGA-Cancer, University of Liège, Liège,

Belgium. 4Analytical Chemistry Laboratory (CIRM), University of Liège, Liège, Belgium.

5Drug Research Center (CIRM), University of Liège, Liège, Belgium.

Correspondence should be addressed to A.N. (agnes.noel@ulg.ac.be; phone: +32-43662569;

fax: +32-43662936)

ABSTRACT

The mouse model of laser-induced choroidal neovascularization (CNV) has been used

extensively in studies of the exudative form of age-related macular degeneration (AMD). This

experimental in vivo model relies on laser injury to perforate Bruch’s membrane, resulting in

sub-retinal blood vessel recruitment from the choroid. By recapitulating the main features of

the exudative form of human AMD, this assay has served as the backbone for testing anti-

angiogenic therapies. This standardized protocol can be applied to transgenic mice and can

include treatments with drugs, recombinant proteins, antibodies, adenovirus and pre-miR to

aid in the search for new molecular regulators and the identification of novel targets for

innovative treatments. This robust assay requires 7-14 days to complete, depending on the

treatment applied and whether immunostaining is performed. This protocol includes details of

1

how to induce CNV, including laser induction, lesion excision, processing plus different

approaches to quantify neoformed vasculature.

INTRODUCTION

Age-related macular degeneration (AMD) is the leading cause of vision loss in Europe, the

USA and Australia 1. Almost two-thirds of the population over 80 years old will have signs of

AMD 2-4 resulting from the wet or exudative form, which is characterized by the presence of

drusen and choroidal neovascularization (CNV). The first studies performed on CNV

associated with AMD aimed to compare mRNA and protein levels in human neovascular

membranes excised surgically versus intact choroids. Experimental animal models rapidly

became essential for elucidating the cellular and molecular mechanisms involved in CNV

pathogenesis and for the screening of new drugs. Indeed, no in vitro model developed to date

recapitulates the complex CNV-associated processes that involve, at a minimum, several cell

types, such as inflammatory cells, endothelial cells, pericytes, bone marrow (BM)-derived

cells, myofibroblasts and glial cells 5, 6. The in vivo exploration of CNV currently involves

several murine models exhibiting the spontaneous development of CNV resulting from a

defective complement-activating pathway, deletion in a chemokine/chemokine receptor,

oxidative damage or aging

7. Currently,

the

laser-induced Bruch’s membrane

photocoagulation model is the most widely accepted and most frequently utilized

experimental murine CNV model. The model described here consists of the laser impact

rupturing of Bruch’s membrane, which leads to the growth of new blood vessels from the

choroid into the sub-retinal space, mimicking the main characteristics of the exudative form

of human AMD 8 and offering the opportunity to explore the molecular mechanism of CNV

through the use of a large panel of transgenic mice. The model has proven to be suitable for

testing the efficacy of new drugs through systemic or local (intraocular) administration and

2

has shown predictive value for drug effects in patients with AMD, for example with VEGFR-

trap 9, 10 or anecortave acetate 11, 12. This model is also appropriate for identifying new

potential targets using siRNA/miRNA technology 13.

Development of the CNV assay

The Campochiaro group was the first to induce CNV through laser injury of Bruch’s

membrane in mice 8. In this model, inflammatory cells are thought to be potent initiators of

the angiogenic process partly through their capacity to release a series of pro-angiogenic

factors 6. Both neutrophil and macrophage depletion reduce CNV formation 14-16. The

recruitment of BM-derived cells (endothelial progenitor cells, pericytes, mesenchymal stem

cells) in CNV lesions has been reported 5, 17-19 and mimics the vasculogenic process observed

in human patients 17. Vascular endothelial growth factor (VEGF-A) rapidly emerged as a

potent angiogenic factor involved in CNV formation, diverting much of the focus of research.

This intensive research led to strategies aimed towards blocking VEGF or its receptors that

have been approved by the US Food and Drug Administration (FDA) for AMD treatment 20-22.

This achievement, based on the use of the laser-induced CNV assay, has validated

angiogenesis as an important target for AMD treatment 6. Transgenic mouse technologies

have since allowed rapid expansion in the exploration of key regulators of CNV, leading to

the identification of other important VEGF family members, such as placental growth factor

(PlGF) 23, 24 and proteases 19, 25-28. More recently, the interplay between complement factor H,

the complement membrane attack complex, a chemokine (CCL2) and VEGF during CNV

development has been established 29, 30.

Applications of the CNV assay

3

The CNV assay is a robust, relevant model that mimics AMD disease, with clear advantages

over other recently reviewed in vivo models 7. The assay recapitulates the complex biological

processes involved in the exudative form of AMD disease (inflammation and angiogenesis)

and is relatively rapid to develop on an easily accessible biological tissue that can be flat-

mounted or prepared for histological sectioning or on protein or nucleic acid extractions.

Furthermore, in contrast to the transgenic models that overexpress apolipoprotein E 31 or with

a superoxide dismutase 1 (SOD1) deletion 32, which are long-term assays requiring senescent

animals, the laser-induced CNV assay can be applied to a panel of young wild-type or

transgenic mice. The model is applicable to transgenic (knock-out or knock-in) mice. In

addition, viruses, cells or compounds, including neutralizing antibodies, si/shRNA, pre-miR,

recombinant proteins, nanoparticles and drugs, can be administered via different pathways

(intraperitoneal injection, intravenous injection, per os, drinking water, intravitreous injection,

subretinal injection, BM engraftment and others) and combined or not with genetic

manipulation 9, 33-37.

Limitations of the CNV assay

In addition to the absence of a defined macula in mice, the rodent laser trauma model is

obviously unable to mimic the complexity of human pathology 4. This model is generated

with a wound-healing reaction that follows an insult at the level of Bruch’s membrane and

relies heavily on inflammation 38, 39; in contrast, in AMD, genetic susceptibility plays a major

role 40. Factors important for the generation of the model may or may not occur in patients

with exudative AMD. The involvement of VEGF 41 and placental growth factor 24 in the

progression of experimental CNV has led to the clinical development of specific antagonists,

such as ranibizumab 21 and aflibercept (VEGF trap-eye) 10. Key features of AMD, such as

sub-retinal pigmented epithelium (RPE) deposits (drusen) and the influence of age, are either

4

absent or not crucial in this experimental model. Limitations of the assay also include the

requirement of pigmented mice (BL6 mice are preferred) for laser reaction, the necessity of a

trauma to induce CNV (no spontaneous development) and spontaneous regression (after 14-

21 days). This model is suitable to study exudative AMD but not the atrophic form of this

ocular disorder. An important drawback of the model is that it cannot be used to test primate-

specific reagents. To overcome this limitation, it is sometimes possible to use mice expressing

the human homolog of the protein of interest. For example, to test human-specific VEGF

antagonists, a transgenic mouse strain that expresses human VEGF in photoreceptors has been

used as an alternative 42.

Overview of the procedure

The general protocol for CNV induction involves laser burning of Bruch’s membrane

following mouse anesthesia and pupil dilatation, eye resection at defined time point(s),

imaging and quantification of angiogenesis and/or inflammation (Fig. 1). After anesthesia and

pupil dilatation, a laser burn is first induced using a green Argon laser focused on the RPE.

The presence of a bubble confirms the success of the laser impact. Mice are housed in the

animal facility during the indicated time periods. During this period, treatments can be

applied to the mice. After sacrifice, the eyes are resected, and the choroid is either flat-

mounted for immunohistological staining, embedded in paraffin or frozen or used for protein

or nucleic acid extraction. Blood sample analysis can also be performed, such as for

metabolomics studies (Fig. 1).

Experimental design

Experimental animals. This protocol is applicable to mice older than 2 weeks when the eyes

are open and can be used to evaluate the effects of exogenous agents and/or genetic knock-in

5

or knock-out. Adult C57BL/6J mice 6-8 weeks old weighing 18-20 g are the most appropriate

for the assay, although the model can be applied to other mouse backgrounds provided they

have pigmented eyes. It is worth noting that aged mice exhibit more severe CNV 43. The sex

of the mice does not appear to influence the assay substantially, except for older female mice,

which develop more severe CNV than do males 44. Laser burns are induced typically at the 3,

6, 9 and 12 o’clock positions around the optic disc in compliance with national and local

ethical committees (Supplemental video 1).

All mice must be treated in accordance with the ARVO Statement for the Use of Animals in

Ophthalmic and Vision Research (www.arvo.org). Anesthesia and euthanasia must be

performed in accordance with the local animal use committee. The use of ketamine and

xylazine mixtures is not recommended because it can lead to serious side effects. At high

anesthetic doses, an acute and reversible cataract can occur, leading to lens clouding, which

can reduce the visibility of the eye fundus and disturb laser focusing 45.

Because of the inherent variability in animal experiments and because some laser impacts do

not result in CNV (approximately 70% of impacts are successful), it has proven necessary to

use at least 6 animals, with 4 impacts per eye per experimental group, and to repeat

experiments at least twice. Two types of control groups are recommended. The first control

group is not subjected to laser induction when comparing transgenic mice and/or evaluating

proteins or RNAs or performing metabolomics studies. The second control group consists of

mice subjected to laser burns and injected with the vehicle buffer alone for drug efficiency

testing. When evaluating pre-miR or siRNA, appropriate scramble sequences are

recommended as controls. In cases of BM engraftment, an experimental group consisting of

irradiated and BM-engrafted animals is mandatory because we have observed that BM

engraftment can impact CNV formation 5, 19. In general, the use of one eye for drug testing

and the second eye as a control is useful because it controls for variation between mice.

6

However, systemic dissemination of the drug is likely to occur. For example, intraocular

injections of bevacizumab suppress subretinal neovascularization in the injected eye and also

cause significant suppression in the fellow eye, indicating a systemic effect. In contrast, this is

not the case for intraocular injections of ranibizumab 42.(cid:3)Therefore, investigators should rule

out systemic effects in the fellow eye with the drug being tested.

Statistical analysis. To compare the effects of different treatments or compounds (I),

including the control groups, on the chosen response, a linear mixed model has to be

constructed that corresponds to the experimental setup. The following model can be used:

!!”#$ ! ! ! !! ! !! ! !!!!! ! !!”#$

where !!”#$ is the response measured (CNV surface); !! denotes the fixed effect of the

treatments i=1,..,I (staining background, drugs); !! is the random effect due to animal

variability (j=1,…,6) (individual mouse variability); !!!!! is the random effect of the eyes

(k=1,2) nested into the animal factor (right or left eye) and !!”#$ represents the residual

random error (l=1,…,4 impacts). The statistical significance of the treatment factor !! of the

model is assessed by comparing its p-value to the significance level e.g. ! ! !!!”. If found

significant (p-value