The Squirrel Monkey: Characterization of a New-World Primate Model of Experimental Choroidal Neovascularization and Comparison with the Macaque

Abstract

purpose. To evaluate and characterize the New-World squirrel monkey as a primate model for experimental choroidal neovascularization (CNV) studies and to compare it with the current Old-World macaque monkey model.

methods. Fibrovascular tissues (FVT) were elicited in 12 maculae of seven squirrel monkeys by laser photocoagulation using optimized laser parameters (532 nm, 0.05 second, 75 μm, 650 mW). Follow-up fundus and fluorescein angiography (FA) examinations were conducted on postlaser days 30 and 35, followed by euthanasia and histologic analysis of tissues. For comparative evaluations, FVT development also was induced and analyzed in eight maculae of four macaque monkeys with laser parameters previously used in this species (514 nm, 0.1 second, 50 μm, 390 and 455 mW).

results. FVT developed in both primate species, consisting of fibrous tissue that contained vessels that ranged from sparse but identifiable capillaries to well-established neovascular networks. Overall, 65% of the photocoagulation sites in the squirrel monkey and 37% of sites in macaque monkey elicited development of FVT. Localized FVT ranged from modest to extensive thickenings of the choriocapillaris layer. Unexpectedly, 76% of the FVT sites in squirrel monkey eyes and 27% of the sites in macaque eyes showed diffuse FVT that expanded beyond the original photocoagulation sites, accompanied by neovascular infiltration of the retina.

conclusions. Like the macaque, the squirrel monkey can be considered a useful primate model for experimental CNV investigations, while additionally offering certain species-specific advantages. Diffuse FVT permit studies of antiangiogenic therapies in areas distant from laser photocoagulative trauma sites.

Although the etiological factors that underlie age-related macular degeneration (AMD) are yet to be resolved, an important pathologic component of exudative (i.e., wet) AMD is the development of choroidal neovascularization (CNV). The potential of novel AMD treatments is ideally evaluated by inducing experimental subretinal CNV in animal models. Primates constitute the preferred vascular animal model, because their eyes are morphologically and functionally most analogous to humans. ^{1 2 3} Using mechanically, chemically, and laser-induced trauma to establish focal breaks in Bruch’s membrane, Ryan ⁴ first described the subretinal neovascularization (SRN) primate model in 1979, in Old-World (Cercopithecoidea) stump-tailed (Macaca speciosa), and rhesus (Macaca mulatta) macaque monkeys. Subsequently, Ryan and associates ^{5 6 7 8 9 10 11 12 13 14 15 16 17} as well as Miller and associates ^{18 19 20 21 22} have further refined and used the primate model, while extensively using another Old-World species, the long-tailed or cynomolgus (Macaca fascicularis) macaque monkey. ^{5 7 8 9 10 11 12 13 14 17 18 19 20 21 22}  

Currently, a variety of factors limit the use of macaque monkeys in research, including limited availability from domestic and overseas suppliers, import and quarantine restrictions, more stringent U.S. Department of Agriculture housing and care regulations, and increased procurement and per diem costs. Moreover, Old-World macaques, particularly those caught in the wild in Africa, Asia, and Pacific Island locations, pose potential health risks to humans because of the possibility of disease transmission. Specific health risks include, but certainly are not limited to, human and simian AIDS (e.g., HIV-2, SIV_mac, SIV_stm, SIV_cyn, and SIV_HU), B virus (Herpesvirus simiae), monkeypox (orthopoxvirus and tanapox), and hepatitis A and B. ²³  

In an attempt to deal with these issues, New-World (platyrrhini) squirrel monkeys were evaluated to determine their feasibility as an alternative primate model. Rather than CNV, SRN, or choroidal neovascular membrane (CNVM), this investigation has designated the use of the term “fibrovascular tissues” (FVT) to describe more accurately the combined histopathologic presence of both choroidal neovascularization and fibroplasia, dual components originating from the choriocapillaris at or around the lesion sites. In addition, FVT may remain subretinal in their development, or they may continue with retinal infiltration. Stimulus parameters were optimized so that focal laser-induced photocoagulation sites elicited maximum FVT development. Finally, FVT formations in squirrel monkey eyes were directly compared with those elicited in eyes of cynomolgus macaque monkeys to determine similarities and differences. 

Methods

Animals

Twelve young adult (ages 2–4 years) female squirrel monkeys (Saimiri sciureus; Charles River Laboratories, Houston, TX) and four young adult (two male and two female) cynomolgus (Macaca fascicularis) monkeys (Covance Research, Alice, TX) were used to evaluate formation of FVT. Experiments involving squirrel monkeys were conducted at the Indiana University Laboratory Animal Resources facility (Bloomington, IN), and macaque research was conducted at Alcon Research Laboratories (Fort Worth, TX). Tissue processing and analyses for both species were performed at the Indiana University School of Medicine (Indianapolis, IN). All procedures were performed with strict adherence to the ARVO Statement for the Use of Animals and Ophthalmic and Vision Research and the guidelines for animal care and experimentation prepared by the Indiana University Institutional Animal Care and Use Committee and by the Alcon Institutional Animal Care and Use Committee. 

For experimental procedures, including laser photocoagulation, ocular examination, and photography, squirrel monkeys (weight range, 600–950 g) received intramuscular (IM) ketamine at 40 mg/kg and acepromazine at 2 mg/kg for anesthesia, along with IM atropine at 0.05 mg/kg, to minimize bronchial secretions. Maintenance amounts of this mixture (10%–15% of the original dose) were administered at 45-minute intervals, when necessary. For all ophthalmic procedures, topical 0.8% tropicamide and 2.5% phenylephrine hydrochloride were administered for pupillary dilation and cycloplegia. 

Cynomolgus monkeys (weight range, 2.3–2.5 kg) received IM ketamine at 10 mg/kg, with intravenous pentothal at 15 to 20 mg/kg to supplement anesthesia. For ophthalmic procedures, topical 1.0% tropicamide and 2.5% phenylephrine hydrochloride were administered to achieve maximum pupillary dilation and cycloplegia. 

Laser Photocoagulation

After undergoing anesthesia and pupillary dilation, animals were positioned before a slit lamp (Carl Zeiss Meditec, Jena, Germany) laser-delivery system. The fundus was visualized using a Goldmann-type plano fundus contact lens (Model OGFA; Ocular Instruments, Inc., Bellevue, WA) with 2.5% hydroxypropyl methylcellulose solution as the cushioning agent. 

During initial squirrel monkey experiments, laser parameters for eliciting optimal FVT photocoagulation sites were established with a green (532-nm wavelength) solid-state diode laser (OcuLight GL; Iris Medical Instruments, Inc., Mountain View, CA) and with a red (664-nm wavelength) diode laser (model DD4; Miravant Systems, Inc., Santa Barbara, CA), which was modified to provide subsecond pulse duration and reduced spot size (with a specially modified slit lamp adapter model DD3-SLA, Iris Medical Instruments, Inc.). Laser power was verified with a power meter (Radiometer model IL1400A, with Detector Head model SPL024F; International Light, Inc., Newburyport, MA). In each squirrel monkey eye, a series of 9 or 12 photocoagulation sites (Fig. 1) were placed within the macula, arranged in a grid-like pattern (consisting of three horizontal rows by three to four vertical columns). ^{5 6 19 20}  

For induction of photocoagulation sites in the cynomolgus monkeys, a green (514-nm wavelength) argon laser (Lambda Plus PDL II photodynamic laser; Coherent Inc., Santa Clara, CA) was used. Laser power was verified with a power meter (FieldMaster, Coherent Inc.). In the slightly larger cynomolgus eye, 12, 16, or 20 photocoagulation sites were similarly arranged in a grid-like pattern of four rows by four to five columns within the macular region. 

For both species, deliberate care was taken to avoid placing photocoagulation sites within the immediate vicinity of the fovea (within approximately 150 μm). Species-specific laser power settings were established and used that most reliably produced acute vapor bubbles, suggestive of the rupture of Bruch’s membrane. These experimental conditions also most reliably produced fibrovascular proliferations that emanate from disrupted Bruch’s membrane and which were capable of retinal infiltration. Within each eye, laser power settings remained constant from site to site (with no minor titrations in power), and individual photocoagulation sites were laser treated only once, regardless of whether vapor bubble formation occurred. This procedure later permitted a better determination and comparison of overall FVT development at photocoagulation sites within each eye. Nevertheless, other factors, related to the eccentricity of photocoagulative site locations (i.e., variations in the incident beam angle with respect to different macular sites), accuracy in laser focusing, and variations in macular pigmentation density, could represent potential sources of variability that might affect development of FVT. ^{24 25}  

Ophthalmologic and Histologic Assessment

Each squirrel monkey underwent a baseline ophthalmologic examination, 7 to 14 days before induction of laser photocoagulation sites, and subsequently, a follow-up examination on post–laser-treatment day 30. This assessment included fundus photography and fluorescein angiography (FA) with a fundus camera (FK-30; Carl Zeiss Meditec) modified (with the inclusion of a +9-D, antireflective-coated spherical lens) to image the fundus in the small-diameter eye (∼14.5 mm, axial length). For early- and late-phase FA photography, 25% sodium fluorescein (0.1 mL/kg) was administered intravenously (through the saphenous leg vein). On postlaser day 35, squirrel monkeys received a final ophthalmic examination, fundus photography, and fluorescein angiography, immediately before euthanasia. 

Cynomolgus monkeys received FA assessment on postlaser day 21 immediately before euthanasia. Animals received 10% sodium fluorescein (0.14 mL/kg) administered intravenously (through the saphenous leg vein). Photocoagulation sites were viewed and photographed with a TRC-501A Retinal Camera and ImageNet 2000 system (Topcon Medical Systems, Inc., Paramus, NJ). 

Eyes from both species were enucleated immediately after euthanasia and eyecup preparations fixed in 4% phosphate-buffered paraformaldehyde solution (overnight at room temperature). For each eye, a single square-shaped tissue block (approximately 8 to 10 mm/side), containing the photocoagulation sites, optic disc and fovea, was hand sectioned from the eyecup preparation. Tissue sections were dehydrated, embedded in paraffin, serially sectioned (6 μm thickness), and hematoxylin and eosin (H&E) stained for light microscopy using a regressive hematoxylin procedure (to label cell nuclei), followed by eosin counterstaining (as a cytoplasm stain and to highlight macrophages). Each laser lesion site recovered was individually evaluated and photographed. 

Histologic specimens were qualitatively assessed by one reader (MHC) and results verified by a second (TAC) to analyze each laser lesion for the presence or absence of experimentally induced FVT. Within each laser photocoagulation site, consecutive 6-μm radially oriented, H&E-stained tissue sections were visually inspected to determine the presence or absence of red blood cell (RBC)–containing luminal structures, in addition to the collagenous tissue containing fibroblasts. Each recovered lesion site was evaluated in its entirety to quantitate the fibrovascular response. Maximum FVT thickness was measured from its distal origins in the choriocapillaris, which, at some sites had expanded distally into the deeper choroidal and inner scleral layers, to its proximal limits, which remained subretinal at some sites, whereas in other sites, it continued proximally with distal and even proximal retinal infiltration. Occasionally, individual neovessels, originating from the central FVT mass, infiltrated into the retinal layers; however, sparse intermittent vessels of this type were not included in the FVT thickness measurements. 

Results

Determination of Laser Stimulus Parameters for Neovascular Development

Initial attempts to elicit FVT were conducted in three squirrel monkeys using a red (664 nm) diode laser (Table 1 , animals 1–3). For these experiments, the laser spot diameter was 70 μm. Unfortunately, the available laser produced a maximum power level that measured only 127 mW, which was found to be insufficient for the task of evoking FVT at the photocoagulation sites. In an attempt to compensate for the lower-than-desired power component, light beam duration times were titrated upward, ranging from 0.1 to 1.0 seconds. In only a few instances were acute vapor bubbles observed. Histologic examination of the tissues indicated that although longer duration times increased the overall extent of focal damage to both the choroid and retina, only fibroplasia had actually appeared at the sites, along with thermal scar tissue produced by the burn. There was no indication of neovascularization. In this type of laser lesion model, focal trauma is a necessary component, but may not be adequate in itself to elicit an effective neovascular response. 

Subsequent experiments were performed in nine squirrel monkeys (Table 1 , animals 4–12) using the green (532 nm) diode laser described in the Methods section (Iris Medical Instruments, Inc.). The spot diameter for this laser was 75 μm. For early tests in four animals (eight eyes; Table 1 , animals 4–7), photocoagulation sites were evaluated and compared using different power levels and two different beam durations (0.05 second and 0.1 second). Laser light powers, ranging from 350 to 650 mW (verified by power meter) were evaluated to achieve reliable vapor bubble formation. Vapor bubble appearance denotes laser-induced acute rupture of Bruch’s membrane, is thought to represent a consistent criterion for later development of CNV (particularly in conjunction with retinal infiltration), and is often designated as a commonly desired end point ²⁶ (Das A, et al. IOVS 2003;44:ARVO E-Abstract 3941) in laser-induced CNV animal models. ^{24 27 28} At the higher power levels and at both duration times, well-developed fibroplasia was typically evident at most of the photocoagulation sites, whereas at some sites, neovascularization was also apparent, thus constituting the formation of FVT. Neovascularization was most apparent at sites where the higher (650 mW) power level was used in conjunction with the shorter (0.05 second) duration. Because of the limited number of animals available, these parameters consequently were used to elicit development of FVT at photocoagulation sites placed within the remaining five animals (Table 1 , animals 8–12). 

Because only four cynomolgus monkeys were available for this investigation, laser stimulus parameters for eliciting FVT with the argon (514 nm) laser were based on previously published criteria. ^{18 20} Laser spot diameter was designated at 50 μm and duration at 0.1 second (Table 1) . Photocoagulation sites were induced using two power levels, 390 and 455 mW (verified by power meter). One male and one female were evaluated at each power level. Although both power levels were capable of inducing FVT, the 455-mW power level evoked the more pronounced neovascular changes, as further characterized in a section that follows. 

Characterization of FVT Development in the Squirrel Monkey

The overall incidence of FVT at photocoagulation sites in the squirrel monkey was 65% (55/84 recovered sites). Photocoagulation sites eliciting localized FVT in 8 of the 12 eyes (13/39 sites; 39%) produced modest to moderate subretinal thickening (∼5–40 μm) of the choriocapillaris layer that extended laterally from approximately 10 to 80 μm beyond the 75-μm diameter edge of the laser photocoagulation site, as evidenced by fundus and FA photography (Fig. 2A) . Light microscopic examination revealed that these sites of localized FVT development typically consisted of fibrous tissue containing neovessels that ranged from sparse but identifiable perfused capillaries to well-developed neovascular networks (Figs. 3B 3C 3D) . 

Photocoagulation sites in the remaining four eyes (42/45 sites; 93%, which represents 76% of the total FVT sites) elicited unanticipated fibrous and choroidal neovascular propagations that extended well beyond the immediate boundaries of the associated laser photocoagulation sites and that typically merged into a common latticework that interconnected the neighboring FVT sites (Fig. 1A ; large arrows). The tangential spread of these diffuse FVTs over hundreds of micrometers were clearly observable in fundus and FA photography (Figs. 2A 2B 2C) and rendered affected tissue regions a translucent, cloudy gray (Fig. 2D) . Macular photocoagulation sites that were located in the immediate vicinity of the fovea were most particularly affected, whereas sites located more peripherally (temporally) commonly demonstrated less (Fig. 2A ; left column of photocoagulation sites) or no neovascular development. In Figures 2C and 2D a fourth indiscernible column of three laser photocoagulation sites situated in the temporal retina (peripheral to the first three) failed to elicit formation of FVT. 

Histologic examinations of photocoagulation sites in the central macula (Figs. 4 5) revealed aggressive development of FVT, originating from the RPE–choriocapillaris border at the original photocoagulation sites. Moving tangentially, FVT growth fronts (Figs. 4A 4B 4D 5) expanded from 600 to 800 μm into surrounding (nonlasered) retinal tissues, initially detaching the retina and RPE from the choroid. Just behind the leading edge of the neogrowth front, the FVT mass also expanded in a proximal (vitreous) direction (Figs. 4 5) , resulting in the destruction of the photoreceptor layer and subsequently damage to the deeper retinal layers. Extensive vascular development from this principal mass of the diffuse FVT often continued to infiltrate well into the proximal retinal layers. The radial thickness of the principal FVT mass through the distal retinal–choroidal junction space typically ranged from approximately 75 to more than 150 μm (not including neovascular infiltration of the retina). 

Lateral expansion of a diffuse FVT growth front progressed tangentially from the central photocoagulation site in a roughly concentric fashion, often similar to that of other localized FVT sites (Fig. 2A , small arrows). In addition, portions of the neogrowth front expanded asymmetrically (Fig 2A , large arrows), with preferential expansion toward similar neogrowth fronts that emerged from neighboring photocoagulation sites. These neighboring FVT growth fronts often merged into one another, leaving little or no histologic evidence of their individual borders (Figs. 2C 2D) . Although originating far away, the mergences of expanding FVT fronts from different photocoagulation sites were observed to reach and envelop the fovea, as shown in fundus and FA photographs (Fig. 2C) and in flatmount and radial histologic sections (Figs. 2D 5D) . 

The presence of what appeared and what was considered to be tufted, brown-pigmented macrophages in retinal tissues located directly at, or in the immediate vicinity of, photocoagulation sites was typically apparent with light microscopy, particularly after enhancement by eosin counterstaining (Fig. 5) . At sites of localized FVT activity, scattered, individual macrophages were evident (Figs. 3B 3D) , whereas from sites where diffuse FVT had developed, macrophage concentrations were particularly high along the leading tangential edge of the expanding FVT growth front (Figs. 4B 4D 5) in conjunction with the expansion of the retinal pigment epithelium layer in this region. Immediately at, and just behind, the neogrowth front (Fig. 5) , large populations of these supposed macrophages appeared to be engaged in phagocytosis of cells within the distal retina, and in particular, in the elimination of photoreceptor outer segments (Figs. 4 5) . 

Characterization of FVT Development in the Cynomolgus Monkey

The general incidence of FVT development at laser photocoagulation sites in this study, determined by FA (Fig. 6) and by histologic evaluations of sites (Fig. 7 8) , was 37% (44/119 photocoagulation sites recovered). This incidence of 37% falls within the range of CNV previously reported in the macaque model (32%–80%, mean: 42%). ^{5 7 8 10 12 16} Although the remainder of the photocoagulation sites did not exhibit the presence of neovessels, extensive fibroplastic development often was evident at these sites. Localized FVT sites (Fig. 7) , as well as nonvascular fibroplasia sites demonstrated similar radial thicknesses (5–60 μm). Unexpectedly, FVT in cynomolgus monkeys (Fig. 6) , like those found in the squirrel monkey (Fig. 2) , were capable of developing as either localized FVT (27%, which represents 73% of the total FVT sites) in the immediate vicinity of the photocoagulation site, or as diffuse FVT (10%, which represents 27% of the total FVT sites) that spread tangentially from the initial laser photocoagulation site and into surrounding normal tissue. As in the squirrel monkey (Fig. 5D) , spreading diffuse FVT encroached directly on the central fovea. Localized and diffuse FVT (accompanied by neovascular infiltration of the retina) were optimally evoked from photocoagulation sites in the cynomolgus eye the argon laser set at 455 mW. 

Some relative differences in neovascularization seemed apparent between the two species, based on both FA and histologic evidence. In cynomolgus eyes (Figs. 6A 6B 7B 7C 7D) , the tangential spread of localized FVT from the original photocoagulation sites typically resulted in marginally greater lesions (30–125 μm beyond the edge of the 50-μm photocoagulation sites) compared with those (10–80 μm beyond the edge of the 75-μm photocoagulation site) measured in squirrel monkeys (Figs. 2 3) . The overall extent of neovascular development within sites from both species varied from modest subretinal thickening to extensive incursion with accompanying vascular infiltration of the retinal layers. 

Although instances (Figs. 6C 6D) of diffuse FVT development occurred in the cynomolgus monkey, consisting of both profuse fibroplastic and neovascular components (Fig. 8) , only eight photocoagulation sites from five of eight eyes were involved, and the tangential spread (ranging from ∼200–400 μm beyond the photocoagulation site) was not as extensive as that observed in the squirrel monkey (Fig. 2) . Similarly, the mergence of adjacent FVT sites was evident in both species, but in cynomolgus monkeys (Fig. 6D) this mergence occurred only in two instances between two neighboring photocoagulation sites and was similar to previously reported FA findings. ⁵ Unlike the squirrel monkey, in the macaque the radial thickness of diffuse FVT development resulted in a wide variability of values, ranging from approximately 15 to 60 μm around some sites (Fig. 8A) to approximately 150 to 200 μm around other sites (Figs. 8B 8C 8D) . 

Discussion

The objectives of this study were to evaluate and characterize the squirrel monkey as an experimental CNV animal model and further to compare its performance with that of the macaque CNV model. Both species exhibit similar, although not necessarily identical, characteristics for FVT development. Based on the current results, the squirrel monkey can be considered a useful “New-World” alternative to macaques as a primate CNV model. Each species provides useful traits for studying aspects of angiogenesis, in terms of proliferation, neovascularization, and migration. Unfortunately, the ability to make absolute comparisons and to quantify differences between the two species was confounded by the following factors: (1) the limited number of primate animals available, (2) partial differences in available technical resources at the two project facilities, (3) possible differences in FVT development resulting from differences in the lasers used and the optimized laser stimulus parameters for inducing trauma in each species, (4) possible differences in the robustness of inflammatory and/or wound-healing processes, and (5) possible variations in the relative distances between consecutive photocoagulation sites. Nevertheless, there is no question that both species demonstrated similar basic characteristics for both localized and diffuse FVT development. 

Laser photocoagulation produces a site of tissue injury that evokes a wound-healing response that includes early blood cell infiltration to the lesion site, secondary blood vessel formation, and ultimately the formation of fibrous scar tissue. The formation of fibrovascular proliferations that originate from the choroid is the basis for CNV modeling. In this investigation, each laser light parameter (wavelength, power, duration, and spot beam diameter) was considered an independent variable that could affect the subsequent extent of neovascular development in the model by influencing the extent of acute traumatic injury to the choroid-Bruch’s membrane-RPE complex. In particular, acute rupture of Bruch’s membrane, denoted by a vapor bubble at the site of laser injury, ^{24 27 28} historically has been sought as an acute end point ²⁶ (Das A, et al. IOVS 2003;44:ARVO E-Abstract 3941) to achieving CNV development. Bruch’s membrane-RPE may normally function as an antiangiogenic barrier in the normal eye, perhaps, in part, by the elaboration of pigment epithelium-derived factor (PEDF), a potent antiangiogenic protein, ^{29 30 31 32} and possibly by endogenous inhibitors of matrix metalloproteinases that are involved in the initial steps in neovascularization. ^{33 34 35 36 37 38 39}  

Small-diameter photocoagulation sites, brief laser duration times, and appropriately sufficient laser power levels seem generally to promote an acute rupture of Bruch’s membrane. Inappropriately high laser power levels can lead to massive acute traumatic injury with hemorrhage, which can be observed during histologic analysis. Insufficient laser power levels and longer laser treatment duration do not rupture Bruch’s membrane and result in lower CNV yield, although CNV may still sometimes develop posterior to the intact Bruch’s membrane, as evidenced on thin serial section histologic analysis (Criswell M, et al. IOVS 2000;40:ARVO Abstract 1222 and unpublished results using rat CNV laser trauma model, 1999–2000). 

With respect to the laser light wavelengths that were used to create photocoagulation sites in the two primate species, the resultant comparative data of this investigation were based on a wavelength difference of only 18 nm between the two lasers (squirrel monkey: 532 nm, diode laser; cynomolgus monkey: 514 nm, argon gas laser). Unlike previous studies in which effects of argon and krypton lasers have been compared, ^{40 41 42} in this investigation there were no apparent differences in light absorption or in subsequent tissue damage at the trauma site that were wavelength related. 

Based on the criterion of late-stage fluorescein leakage around photocoagulation sites, some investigators studying macaques have reported that maximum development of CNV occurred approximately 2 to 4 weeks after photocoagulation and that spontaneous CNV regression (i.e., involution), as marked by decreased FA leakage, commenced at approximately 3 to 7 weeks and then gradually progressed (over a period from approximately 2 to 13 months) until leakage was no longer apparent at the site. ^{5 6 7 8 9 12 16 22} Nevertheless, investigators also have proposed that changes in FA staining patterns may not necessarily be correlated with histologic evidence of the presence of CNV. ^{6 8 9 10} In squirrel monkeys, regression of late-stage fluorescence leakage was not apparent at 35 days; meanwhile, histologic assessment demonstrated that both neovessels within the FVT and those vessels infiltrating the retina appeared viable at 35 days. Longitudinal study of this characteristic was not continued beyond this time point. 

Determinations of initial neovascular formation, the extent of neovascular development, and the determination of vascular regression typically rely on fundus and fluorescein photographic data (particularly late-phase FA leakage). Although these techniques are valuable evaluative tools (and necessary in human clinical studies), fundus photography unfortunately provides only a topographical view of superficial changes in the macula, and FA (as well as other dye-labeling methods) presents possible interpretative limitations because of leakage and obscured vessel imagery with increased tissue depth. ^{6 8 9 22 23 43 44 45} Rather than relying solely on typical FA (which is inherently capable of providing unreliable false-positive or false-negative results) to determine the presence and extent of neovascular development, qualitative and quantitative aspects of CNV development in both the squirrel and cynomolgus monkey CNV models were characterized in this animal investigation by using direct histopathologic evidence as the primary gold standard ⁴⁶ method for evaluation, in conjunction with fundus and FA evidence. Some primate studies have presented high-magnification light and electron micrographs to illustrate neovascular changes within the choriocapillaris (usually in the immediate vicinity of the photocoagulation site). ^{6 8 9} However, remarkably few reports have demonstrated localized FVT formation with histologic evidence ^{8 10 14 46 47} or have directly correlated development of CNV-FVT with in vivo tissue changes depicted by angiographic evidence. ^{8 10 46 47 48} The distinct neovascular changes reported herein have not been mentioned in connection with the macaque primate model, although a recent report by Obana et al. ⁴³ included a figure of neovascular formation (termed “subretinal proliferative tissue”) in cynomolgus monkeys which, if placed in context with this present study, would be described as demonstrating development of diffuse FVT. 

As mentioned earlier, previous CNV investigations using the macaque model have reported the incidence of CNV as ranging from 32% to 80% (principally based on FA data). Considering only those localized photocoagulation sites (initially identified by FA and subsequently confirmed by histologic examination) where modest to moderate levels of FVT have occurred, the 27% incidence of macaque FVT and the 39% incidence of squirrel monkey FVT in this investigation were comparable to localized findings in previous studies. 

Qualitative histologic analyses of squirrel and cynomolgus monkeys in this study revealed that the extent of neovascularization at identified FVT sites ranged from modest to extensive. Compared with primates, in the rat laser trauma model, aggressive FVT development, containing rich neovascularization, can be documented in conjunction with yields approaching 100% of the recovered photocoagulation sites. ^{49 50} In the present study, a lower but significant proportion of photocoagulation sites in squirrel monkey eyes elicited FVT that were localized to the immediate vicinity of the photocoagulation site and that consisted of fibroplastic and limited neovascular components (similar to previous reports in the macaque literature). 

Of great interest was the discovery that in eyes of some squirrel monkeys (using seemingly identical laser stimulus conditions), most of the photocoagulation sites exhibited diffuse development of FVT that, once initiated, infiltrated deep into the retina and that spread beyond the diameter of the original trauma site for hundreds of micrometers, tangentially through normal (nonlasered) retinal and choroidal tissues. In those cases, neovascularization was a prominent feature within the principal FVT mass. Even more unanticipated, however, was the discovery that this diffuse type of FVT formation also occurred in cynomolgus monkeys. 

In squirrel monkeys, the prominent development of diffuse FVT, involving multiple photocoagulation sites, seemed to occur selectively and bilaterally in only certain animals. In macaques, diffuse FVT occurred in a more generalized fashion across the population, either unilaterally or bilaterally, whereas diffuse FVT in this species were not as extensive in their development compared with FVT in the squirrel monkey. Reasons for possible interspecies differences, and more important, possible intraspecies variations within squirrel monkeys require further study. Whether the occurrence of localized versus diffuse FVT results from differences in tissue or vascular conditions that may exist across the macular region or from differences in the laser light delivery at the central versus peripheral photocoagulation sites is not known. 

The presence of what are proposed to be highly pigmented and tufted macrophages may play an important role in the ability of FVT to expand far beyond the original boundaries of the photocoagulation site. Other investigators also have observed these cells, although their origin (perhaps sequestered macrophages or conceivably modified RPE cells) remains unresolved. ^{10 13 15} As indicated earlier and as reported by others, ^{8 9 10 12 13 14 15 44 51 52} macrophages are prominent along the expanding surface of the tangential FVT growth front and conceivably are involved in the initial detachment and elimination of the photoreceptor outer segments. To a somewhat lesser extent, these cells may also assist vascular and nonvascular FVT components behind the neogrowth front to infiltrate the more proximal layers of the retina. 

Similar tangential expansion and mergence of FVT also can be observed between successive photocoagulation sites in the rat eye, suggesting that preferential tangential growth may be related to the expression of angiogenic promoters (Criswell M et al., unpublished results, 2002–2003). Based on the current data available, this augmented neovascularization seems to be somewhat more common in occurrence and possibly more pronounced in effect in the squirrel monkey than in cynomolgus. The underlying reason(s) why such differences in neovascular development may result between some photocoagulation sites in a small population of eyes from both species (possibly involving a selective predisposition, a morphologic variation, or perhaps merely a protocol difference) remains unclear. 

Beyond their established usefulness as a primate CNV model, squirrel monkeys demonstrate additional useful characteristics for primate research. Squirrel monkeys are less than half the size and weight of macaques. Females (with shorter incisors) are relatively easy to handle and sedate. Older-style macaque cages, that are no longer appropriate for long-term use, provide more than ample space for squirrel monkeys, the only caveat being that smaller females occasionally can manage to squeeze between the front bars. As primates, squirrel monkeys can still act as vectors for human diseases, in that they possess species-specific pathogens, and feral animals may harbor unusual parasitic agents (particularly Trypanosoma cruzi) ⁵³ ; however, squirrel monkeys generally tend to possess fewer indigenous pathogens than do Old-World macaques. ²⁰ During quarantine, special consideration should be given to tuberculosis testing. 

The squirrel monkey CNV model, like the macaque model, is a useful experimental means for modeling choroidal neovascularization in an animal species, especially because both species and humans share many common ocular and visual traits. ^{54 55 56 57 58 59 60 61 62} These animal models are useful in mimicking pathologic conditions of CNV development that occur in human AMD patients. However, it should be emphasized that laser-trauma–induced FVT in animal models are not identical with human exudative AMD. In addition, Shen et al. ⁴⁷ have indicated that the relatively low CNV yield rates in primates (compared with the higher yields obtained with the rat model) may necessitate the use of a greater number of primates for meaningful statistical analyses of drug evaluation data; alternatively, further refinement of the primate model may be necessary to increase the percentage of photocoagulative sites that elicit neovascularization. 

Nevertheless, primate models permit useful testing of potential AMD treatments and therapies and investigations of the underlying etiology for AMD. The macaque model already has been used to evaluate novel photodynamic therapy (PDT) agents, including verteporfin (Visudyne; Novartis AG, Basel, Switzerland.), ^{18 19 20 21 63} mono-l-aspartyl chlorin e6 (NPe6; Meiji Seika Kaisha, Ltd., Tokyo, Japan), ^{64 65 66} ATX-S10(Na) (Photochemical Inc., Okayama, Japan), ^{43 67} and motexafin lutetium (Optrin; Pharmacyclics Inc., Sunnyvale, CA), whereas the squirrel monkey model has been used in the testing of the MV6401 agent (PhotoPoint; Miravant Pharmaceuticals Inc., Santa Barbara, CA) (Pratt LM, et al. IOVS 2001;42:ARVO Abstract 2353; Ciulla TA, et al. IOVS 2002:43:ARVO E-Abstract 614). ⁶⁸ Diffuse FVT offer special opportunities to evaluate neovascularization and potential methods for treatment in areas of the choroid and retina that were initially unaffected by laser photocoagulation. There also is evidence to suggest that some squirrel monkeys may have retinal degenerations, with characteristics similar to those associated with retinitis pigmentosa, ⁶⁹ whereas some very old animals (older than 10 years) exhibit vascular changes that are characteristic of macular degeneration (Criswell M, et al., unpublished results, 2001). Finally, a current investigation that compares CNV growth factor expression with molecular and immunocytochemical techniques in the rat CNV laser trauma model (Hu W, et al. IOVS 2003;44:ARVO E-Abstract 3937; Criswell M, et al. unpublished results, 2003) also will include primate tissues as they become available for such analyses. 

 Supported in part by National Eye Institute Grant R44 EY11191, Business Innovation Research Program through Miravant Medical Technologies; by Miravant Medical Technologies Inc., Santa Barbara, California; by Alcon Research, Ltd., Fort Worth, Texas; and by an unrestricted grant from Research to Prevent Blindness, Inc. to the Department of Ophthalmology, Indiana University.

 Submitted for publication July 10, 2003; revised October 7, 2003; accepted October 26, 2003.

 Disclosure: M.H. Criswell, Miravant Medical Technologies, Inc. (F) and Alcon Research, Ltd. (F); T.A. Ciulla, Miravant Medical Technologies, Inc. (F) and Alcon Research, Ltd. (F); T.E. Hill, None; W. Small, Miravant Medical Technologies, Inc. (E); R.P. Danis, Miravant Medical Technologies, Inc. (F); W.J. Snyder, Miravant Medical Technologies, Inc. (E); L.A. Lowseth, Alcon Research, Ltd. (E); D.L. Carson, Alcon Research, Ltd. (E)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Schematic diagram depicting the relative placement and organization (three rows by four columns) of laser photocoagulation sites (dots) within the macula of the squirrel monkey. White: areas of possible FVT that can develop at and between the photocoagulation sites. Connections between sites denote routes of possible diffuse FVT development within the central macula. Neither localized, nor diffuse FVT development was evident as sites within the fourth, temporal-most column (far left), whereas localized and diffuse FVT developed within this same region in cynomolgus eyes. F, location of the fovea; OD, indicates the location of the optic disc. Reprinted with permission from Indiana University School of Medicine Visual Media Services.

Table 1.

 

Specific Laser Light Parameters Used in Individual Squirrel and Cynomolgus Macaque Monkey Maculae To Determine Optimal Conditions for Eliciting FVT Development at Photocoagulation Sites

	Eye	Laser (nm)	Laser Power (mW)	Beam Duration (sec)	Beam Diameter (μm)	Vapor Bubbles per Total Sites	FVT Sites Per Evaluated Sites
Squirrel monkey
1 Female	OD	664	104	0.1, .25, 0.5, 1.0	70	0/50	0/50
	OS	664	127	0.1, .25, 0.5, 1.0	70	3/50	0/50
2 Female	OD	664	127	0.1, .25, 0.5, 1.0	70	2/25	0/25
	OS	664	127	0.1, .25, 0.5, 1.0	70	0/25	0/25
3 Female	OD	664	127	0.1, .25, 0.5, 1.0	70	4/25	0/25
	OS	664	127	0.1, .25, 0.5, 1.0	70	0/25	0/25
4 Female	OD	532	350	0.05	75	2/9	0/9
	OS	532	450	0.05	75	7/9	1/9
5 Female	OD	532	400	0.05	75	4/9	0/9
	OS	532	500	0.05	75	6/9	1/9
6 Female	OD	532	450	0.01	75	9/9	2/9
	OS	532	650*	0.01	75	9/9	3/9
7 Female	OD	532	450	0.01	75	9/9	1/9
	OS	532	650*	0.01	75	9/9	3/9
8 Female	OD	532	650*	0.05	75	9/9	1/3, †
	OS	532	650*	0.05	75	9/9	2/3, †
9 Female	OD	532	650*	0.05	75	9/9	0/3, †
	OS	532	650*	0.05	75	9/9	1/3, †
10 Female	OD	532	650*	0.05	75	9/9	0/3, †
	OS	532	650*	0.05	75	9/9	3/3, †
11 Female	OD	532	650*	0.05	75	9/9	9/9, ‡
	OS	532	650*	0.05	75	12/12	12/12, ‡
12 Female	OD	532	650*	0.05	75	12/12	11/12, ‡
	OS	532	650*	0.05	75	12/12	10/12, ‡
Macaque monkey
1 Female	OD	514	390	0.1	50	12/12	4/12
	OS	514	390	0.1	50	12/12	6/12
2 Male	OD	514	390	0.1	50	19/20	4/20
	OS	514	390	0.1	50	15/16	3/15
3 Female	OD	514	455	0.1	50	12/12	5/12
	OS	514	455	0.1	50	16/16	6/16, ‡
4 Male	OD	514	455	0.1	50	16/16	9/16, ‡
	OS	514	455	0.1	50	16/16	7/16

 In squirrel monkey, diode laser wavelengths of 664 and 532 nm were tested, and in macaques, an argon gas laser emitted light at wavelength 514 nm.

*  Photocoagulation sites produced at 650-mW power were included in subsequent FVT analyses.

†  Additional photocoagulation sites from animals 8 through 10 were included in a different investigation, and these data sites were excluded from FVT analyses in this study.

‡  Includes diffuse and localized FVT sites.

Figure 2.

 

FVT development around photocoagulation sites in the squirrel monkey. (A–C) FA at 30 days after induction of photocoagulation sites. Each photograph (A–C) illustrates progressively greater levels of FVT development at and between the photocoagulation sites. (A, small arrows) Concentric development of FVT; (A, large arrows) Diffuse FVT between neighboring sites. The extensive FVT proliferation between the nine sites in (C) continued to be evident in flat-mount tissue from the same region (D).

Figure 3.

 

Radial histologic views (6 μm thickness, H&E stained) of the normal retina (A) and of localized FVT development at and expanding slightly beyond photocoagulation sites (B–D) in the squirrel monkey. Retinal infiltration by neovessels containing erythrocytes was evident at each of these sites. Bar, 75 μm.

Figure 4.

 

Radial histologic views of diffuse FVT development in the squirrel monkey. (A, B, D) Micrographs demonstrating FVT development at the expanding neogrowth fronts where, in each instance, they caused detachment of the retina from the choroid. (A) Principal FVT mass is located to the right and the growth front expanded tangentially to the left (in a nasal direction). (B) The growth front expanded tangentially to the right (in a temporal direction, away from the optic disc). (C) Region in the middle of diffuse FVT mass that contained numerous small- to medium-sized vessels. (D) Diffuse neogrowth front similar to that shown in (B), except that the plane of histologic sectioning has been rotated by 90° (orthogonal view). Note the presence of large pigmented cells (presumably macrophages) at the RPE–choriocapillaris junction and individually and in groups within the FVT mass. Bar, 75 μm.

Figure 5.

 

(A–C) An advancing diffuse FVT growth front in the squirrel monkey that produced retinal detachment from the choroid. Note the presence of aggregated pigmented cells (presumed macrophages) along the FVT front and their apparent phagocytic activity on the retinal photoreceptors. (D) Diffuse FVT growth front is shown as it encroached on the fovea. Bars: (A, B) 75 μm; (C, D) 150 μm.

Figure 6.

 

FA demonstrating partial localized (A, B) and mixed localized and diffuse (C, D) FVT development in the maculae of the cynomolgus macaque monkey at 21 days after induction of laser photocoagulation sites.

Figure 7.

 

Radial histologic views of normal retina (A) and localized development of FVT at photocoagulative sites (B–D) in a cynomolgus monkey. Note the similarities in fibroplastic development, vascular infiltration of the retina and the presence of pigment cells (macrophages) between the macaque and the squirrel monkey (Fig. 3) . Bar, 75 μm.

Figure 8.

 

Radial histologic views of diffuse FVT in the cynomolgus monkey. (A) Well-developed, diffuse FVT with a small radial thickness (20–55 μm) is shown. (B) Diffuse FVT developed from the photocoagulation site. Note the direct infiltration of the retina located at the site. (C) Region in the middle of a diffuse FVT mass containing numerous viable vessels, similar to that in the squirrel monkey (Fig. 4C) . (D) Continuation of the same FVT mass shown in (C), illustrating the FVT neogrowth front. Note the pigmented (macrophage) cell activity along and within the leading edge of the front as it expanded and separated the retina from the choroid. Bar, 75 μm.

The authors thank Russell L. Schmidt, Director of Indiana University Laboratory Animal Resources (Bloomington, IN), for extensive contributions and assistance and Lisa L. Bird-Turner and Robert S. Flack for lending technical expertise and assistance.