Abstract
purpose. For many vascular diseases, aging appears to be an independent risk factor for severity of vascular complications, and blood vessels of aged individuals often demonstrate exaggerated repair responses to injury. This study was undertaken to determine the influence of aging on the severity of neovascularization in a mouse model of laser-induced choroidal neovascularization (CNV).
methods. CNV was induced in young (2-month-old) and aged (16-month-old) C57BL/6 mice by making four separate choroidal burns in each eye with a diode red laser (650 nm). At 1, 2, and 4 weeks, the left eyes were removed for histopathology, and the right eyes were removed for flatmount analysis of CNV surface area, vascularity, and cellularity.
results. Aged mice demonstrated a much larger area of CNV than did young mice (3.81 ± 1.28 vs. 1.36 ± 0.99 disc areas, P < 0.001) at 2 weeks, when the lesions showed maximum growth. Aged mice also demonstrated higher ratios for vascularity and cellularity of the CNV (1.34 ± 0.06 vs. 1.03 ± 0.11, P < 0.0001 and 4.06 ± 1.19 vs. 1.91 ± 0.81, P < 0.002 at 2 weeks, respectively). Histopathology revealed that CNV in older eyes was larger, thicker, and more cellular than in young eyes.
conclusions. In mice, age is associated with more severe CNV, defined as larger surface area, greater vascularity, and greater cellularity. Age–related systemic susceptibility factors, independent of local changes in the retina, may contribute to the greater severity of CNV in older than in younger individuals.
Choroidal neovascularization (CNV) is the major vision-threatening complication associated with several common retinal degenerative or inflammatory diseases, especially age-related macular degeneration (AMD),
1 2 3 4 5 pathologic myopia, angioid streaks, and ocular histoplasmosis.
6 7 8 The severity of CNV (defined as size of affected area, amount of hemodynamic instability, and tendency to form large disciform scars) appears to be strongly associated with both the underlying degenerative disorder and the age of the affected individual. CNV in ocular histoplasmosis or myopia, which occurs in young individuals, tends to present as small classic CNV, shows development of minor exudation, and often evolves into relatively small fibrotic scars.
9 10 11 In AMD, CNV typically grows large, is hemodynamically unstable (i.e., exudes plasma and tends to hemorrhage) and often progresses into large fibrovascular scars.
The specific biological factors that predispose to more severe CNV are unknown. Disease-specific differences in the severity of subretinal degenerative changes are often proposed as a significant factor in the severity of CNV. For instance, in ocular histoplasmosis, RPE degeneration is usually localized to small inflammatory foci, which are the site of CNV’s formation.
12 13 14 15 In AMD, however, widespread RPE degeneration, deposits in Bruch membrane, and choriocapillaris damage are present. These degenerative changes may contribute significantly to the nature and severity of CNV.
16 17 18
However, another possible mechanism contributing to CNV’s severity is the presence of age-related changes in systemic biology,
19 20 21 22 independent of subretinal degeneration. In particular, aging individuals often demonstrate dysfunctional blood vessel repair after vascular injury, leading to increased endothelial and smooth muscle proliferation, abnormal repair of the extracellular matrix, excessive fibrosis, and even angiogenesis.
23 This concept of age-related “dysfunctional response to vascular injury” is relevant in several age-related vascular diseases, such as renal glomerulosclerosis
24 25 26 27 28 and coronary neointimal proliferation after angioplasty.
29 30 31 32 33 34 35 The specific role of aging as a pathogenic mechanism of CNV’s formation has not been evaluated.
The purpose of this study was to examine the role of age as an independent risk factor for the severity of CNV in a mouse model of diode laser injury to the choroid. We observed that older mice demonstrated much more severe CNV, defined as larger surface area, greater vascularity, and greater cellular density. However, collagen content and morphologic features were similar.
Mice used in this study were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. A total of 43 female, normal C57BL/6 mice age 2 months (n = 21; weight, 18–20 g) and 16 months (n = 22; weight, 25–30 g), at the onset of the study, were purchased from the National Institute on Aging (Bethesda, MD).
Mice were anesthetized with an intramuscular administration of ketamine hydrochloride (42.8 mg/kg; Fort Dodge Animal Health, Fort Dodge, IA), xylazine (8.5 mg/kg; Phoenix Scientific, Inc., St. Joseph, MO), and acepromazine (1.4 mg/kg; Phoenix Scientific, Inc.). Pupillary dilation was achieved by using 1.0% tropicamide (Alcon Laboratories, Fort Worth, TX) and 2.5% phenylephrine hydrochloride (Akorn, Buffalo Grove, IL). A diode red laser (650 nm; OcuLight SLx; Iris Medical Instruments, Inc., Mountain View, CA) was delivered to the retina through a slit lamp biomicroscope (Haag-Streit; Mason, OH) using a 22 × 22-mm coverglass (Baxter Health Care Corp., McGaw Park, IL) as a contact lens. Treatment parameters were chosen to produce a cavitation bubble in the choroid without hemorrhage. This was achieved in both young and old mice with 75-μm spot size, 150-mW intensity, and 100-ms duration. Mice with obvious media opacities (i.e., cornea and/or lens) were excluded. Four laser burns were performed in the 12, 3, 6, and 9 o’clock positions of the posterior pole around the optic nerve of both eyes of the animals. One, 2, and 4 weeks later, eyes were examined for CNV by the various methods described in the following sections.
Sodium fluorescein (0.03 mL of 10 mg/mL; Akorn) was injected into the heart of the anesthetized mice. The angiograms were recorded using a camera (MD-R; Nikon, Garden City, NY) attached to an operating microscope (Op-Mi6; Carl Zeiss, Thornwood, NY) and a custom-made angiography machine with a power supply flash of 50 to 400 joules (Carl Zeiss). Photographs were taken at approximately 1 and 5 minutes after dye injection. Mice were then perfused through the heart with a mixture of high-molecular-weight fluorescein-isothiocyanate (FITC)-dextran (Sigma, St. Louis, MO): 2 × 106 molecular weight + 4 × 104 molecular weight in a proportion of 2:1, in lactated Ringer’s solution at a concentration of 10 mg/mL (0.6 mL of this solution was injected into each animal). The right eye was enucleated for flatmount preparation. Each animal was then perfused and fixed through the heart with formaldehyde, buffered 10% (Ricca Chemical Company, Arlington, TX). Left eyes were enucleated and stored in fixative for 24 hours.
Fixed globes from young and old mice at 1, 2, or 4 weeks after laser treatment were embedded in paraffin and processed for standard hematoxylin and eosin (H&E) staining, to assess standard morphology, or Masson’s trichrome staining, to assess collagen and matrix formation. Serial thick sections were examined from the entire retina, and a single specimen representing the largest and thickest specimen within the examined specimens was evaluated for each eye. H&E- and Masson’s trichrome–stained sections were digitized using a light microscope (Labophot; Nikon) connected to a color video camera and a frame grabber (DEI-750; Optronics, Goletta, CA). Using image analysis software (Photoshop 5.5; Adobe, San Diego, CA), the diameter and maximum thickness of CNV were calculated from H&E specimens in pixel dimensions. The relative collagen content was approximated from the Masson’s trichrome–stained specimens by determination of the ratio of the average luminosity of the blue channel from the extracellular matrix of the lesion divided by the average density of the blue channel taken from representative areas within the sclera.
A technique to visualize FITC-dextran–perfused vessels within CNV was modified to allow simultaneous visualization of cell nuclei within the lesion. The anterior segment and the neurosensory retina were removed. Four radial relaxing incisions were made in the remaining sclera–choroid–RPE complex. RPE was removed using microsponges (Alcon Laboratories) after placing the tissue on a glass slide with the laser spots facing upward under an operating microscope (Op-Mi6; Carl Zeiss). Cellular nuclei were stained with a mixture of 1 mg/mL digitonin (Sigma) and 0.5 mg/mL propidium iodide (PI; Sigma). Green FITC and red propidium fluorescence within the neovascular complexes was visualized under an immunofluorescence microscope (Axiophot Photomicroscope; Carl Zeiss), in which an objective of ×4 and FITC filter (exciter filter: 450–490 nm; dichroic mirror: 510 nm; LP Filter: 520 nm) and PI filters were used, respectively (exciter filter: 510–561 nm; dichroic mirror: 580 nm; LP filter: 590 nm). A color video camera and a frame grabber (DEI-750; Optronics) were used to digitize the images.
Surface area of CNV was determined by using either FITC-dextran fluorescence or PI fluorescence, and outlining the margins of the lesion with the image-analysis software (Photoshop; Adobe). The area in pixels was normalized by dividing by the average area of the disc measured in 10 independent eyes. Five eyes were examined 4 hours after laser treatment, to determine the average laser spot size (0.48 disc areas). A CNV was defined as present if the surface area of an individual lesion was more than 0.50 disc areas.
An index of vascularity was approximated by calculating the ratio of the average luminosity of the green channel of FITC images from five samples within the CNV compared with five independent samples from the normal choroid away from the CNV (i.e., background luminosity). An index of cellular density was approximated by calculating the average luminosity of the red channel within the lesion divided by the average density from the normal choroid away from the CNV (i.e., background choroidal cellular density).
Morphometric data for different lesions in each eye were averaged to provide one value per eye. The mean and SD for these measures for each group was calculated and probabilities (t-test) were determined by computer (Prism 3.0; GraphPad Software Inc., San Diego, CA). Results at P < 0.05 were considered statistically significant for all forms of statistical analysis used.
Comparison of fluorescein angiography between young and old mice revealed a similar frequency of fluorescein leakage in approximately 75% of the cases. Young mice revealed the typical pattern of early focal hyperfluorescence that increased in size over time. However, older mice demonstrated larger areas of early hyperfluorescence and much greater areas and intensity of late fluorescein leakage, suggesting pooling of subretinal fluid
(Fig. 1) .
Histopathology confirmed that CNV in older mice had much larger diameter and thicker center
(Fig. 2) , which was confirmed with quantitative analysis
(Fig. 3) . Morphologic features of CNV between young and old were similar, except for the presence of prominent RPE cystlike structures underlying the CNV in the larger lesions of older mice. These pseudocystic cavities did not fill with fluorescein or FITC-dextran and were similar to those observed in the primate model.
36
Flatmount analysis better demonstrated the increased size, vascularity, and cellularity in CNV from old mice
(Fig. 4) . CNV (defined as surface area >0.5 disc areas) was observed in 68 (94%) of 72 lesions in 2-month-old mice and in all 76 (100%) the lesions induced in 16-month-old animals. CNV in young mice reached maximum size within 14 days and showed four small, discrete circular lesions with poorly defined margins by FITC. In contrast, CNV in older mice were much larger, with irregular borders and extensions. CNV from different laser spots often became confluent. The margins of the lesion outlined by PI fluorescence (i.e., cell nuclei) was slightly larger than the margins by FITC, suggesting the presence of more avascular cellular growth at the margins of the CNV in older mice. Quantitative analysis of the surface area of CNV over time confirmed that older mice had significantly larger areas of CNV
(Figs. 5A 5B) .
The relative vascularity of CNV in young and old mice was calculated by measuring the ratio of FITC fluorescence in the lesion compared with that in the background choroid. The data showed a significantly greater luminosity index in the older mice
(Fig. 5C) .
We also evaluated the cell density in CNV by measuring the ratio of fluorescent luminosity from the PI nuclear staining in the lesion compared with that in the background choroid. This cellularity index also was significantly greater in the older mice at all time points
(Fig. 5D) .
Finally, we measured the amount of extracellular matrixdeposition, by using Masson’s trichrome stain. Qualitative assessment indicated that, as expected, a greater matrix area was present in the larger CNV of older mice
(Fig. 6) . However, no statistically significant difference was found in the relative density of matrix deposition evaluated by relative Masson’s trichrome uptake within the CNV
(Fig. 7) , suggesting that there is no major difference in the relative collagen concentration per unit area of extracellular matrix.
We evaluated the severity of CNV induced by laser injury to the choroid in young and old mice. We found that CNV in older mice were larger, thicker, more cellular, and probably more vascularized than CNV in younger mice. However, we did not observe a difference in frequency of fluorescein leakage or in relative extracellular matrix concentration. Nevertheless, the amount of leakage and amount of matrix was probably greater in older mice, because of the increased size of the CNV.
The effect of aging on the formation of pathologic or physiologic neovascularization in nonocular tissues is controversial, and only a few studies have addressed the question. Some studies have suggested the existence of impaired angiogenesis in aging animals, especially in models based on postischemic neovascularization. For example, studies investigating neovascularization after hindlimb ischemia have concluded that aging impairs physiologic neovascularization.
37 Also, partial impairment in border zone angiogenesis was observed more often in older than in younger patients with stroke.
38 Proposed mechanisms have included the age-related loss of expression of angiogenic growth factors, loss of the proliferative capacity of aging endothelial cells, or loss of the expression of endothelial enzymes responsible for digesting the extracellular matrix.
39 40 41 42 43 44 45
Other studies, however, suggest that increased neovascular responses occur in older animals, especially those using models of vascular injury. In fact, abnormal vascular repair after blood vessel injury in aged individuals has been demonstrated in several nonocular vascular beds. Aortic injury in rats resulted in increased neointimal proliferation of vascular smooth muscle cells in aged rats when compared with younger animals.
46 47 Aging also increases susceptibility to injury-induced nephropathy in animals, resulting in progressive glomerular basement membrane thickening and blood vessel proliferation.
48 49 In addition, aging appears to be associated with increased vasoproliferation in a mouse model for neurodegeneration.
50 51 52 Proposed age-related biological changes that might influence excessive vascular repair in aging include loss of soluble growth inhibitors from the blood (i.e., sex hormones),
53 54 55 56 increase in soluble growth stimulants (i.e., cholesterol and lipids),
57 58 59 60 changes in immune function (i.e., low-grade inflammation)
19 20 21 22 or changes in the circulating vascular progenitors cells responsible for regenerating injured vessels (i.e., increase in the number of circulating vascular progenitors).
61 62 63
Our experiments were designed to evaluate the impact of aging on the severity of CNV. Although we cannot rule out the contribution of age-related changes in the biology of the choriocapillaris or Bruch membrane in older mice, we believe that the best explanation for our data is the presence of age-related dysregulation in choroidal vascular repair after laser injury, resulting in the abnormal regulation of CNV growth and development in older mice. The present study does not address the mechanisms regulating the severity of CNV, although we are currently evaluating the contribution of inflammatory stimuli. Furthermore, the effect of age must be confirmed in another CNV model based on angiogenic factor overexpression,
64 rather than laser injury.
These results also indicate that it is feasible to measure different parameters of CNV to assess severity. Although plasma exudation and hemorrhage are considered to be complications that most acutely cause loss of vision in humans, the presence of CNV in the absence of leakage (i.e., after successful photodynamic therapy) is still associated with chronic progressive loss of vision.
65 66 67 68 For example, even in the absence of active leakage, CNV may be associated with abnormal overlying RPE,
69 70 71 cells within CNV may produce cytotoxins injurious to photoreceptors,
72 73 74 75 76 or CNV may disrupt transport between retina and choroid. Thus, it is possible that measurement of cell density, matrix deposition, and other parameters may be useful markers to evaluate long-term complications of CNV.