Increased ocular levels of IGF-1 in transgenic mice lead to diabetes-like eye disease (original) (raw)
IGF-1 is overexpressed in eyes of transgenic mice. Transgenic mice overexpressing IGF-1 under the control of the RIP-I show increased IGF-1 in β cells (31). These mice are normoglycemic and normoinsulinemic and have body weight similar to that of control mice (31). Surprisingly, ocular alterations such as cataracts and megaloglobus or buphthalmos appeared in old (more than 6 months) transgenic mice. RT-PCR analysis demonstrated the expression of the transgene (Figure 1A), indicating that transcription from the RIP-I promoter was active in the eye of transgenic mice. This finding is consistent with the reported expression of preproinsulin in the eye (36). Northern blot analysis showed high IGF-1 expression in the retina (Figure 1B). In addition, by whole-mount in situ hybridization, IGF-1 expression was detected in the outer plexiform layer and the inner segment of photoreceptors of the transgenic retina (Figure 1, C and D). Thus, IGF-1 was overexpressed in photoreceptors of the transgenic eyes. This finding was parallel to high levels of IGF-1 in the aqueous humor of both 2- and 14-month-old transgenic eyes (Figure 2). Serum IGF-1 levels were similar in both control and transgenic mice (control, 300 ± 20 ng/ml, n = 8, versus transgenic, 320 ± 35 ng/ml, n = 8), however. Furthermore, glucose concentration in the aqueous humor of both 2- and 12-month-old transgenic eyes was similar to that of controls (at 12 months: control, 126.1 ± 18.1 mg/dl, n = 5, versus transgenic, 130.7 ± 32.2 mg/dl, n = 5) and was consistent with the normoglycemia of these mice (at 12 months: control, 138 ± 8 mg/dl, n = 7, versus transgenic, 142 ± 6 mg/dl, n = 7). This indicated that local ocular hyperglycemia was not responsible for the alterations in transgenic eyes.
IGF-1 is overexpressed in eyes of transgenic mice. (A) RT-PCR confirmed that IGF-1 in the eye was a result of transgene expression. Pancreas mRNA was used as a positive control and liver mRNA as a negative control. The amplified band hybridized with an IGF-1 cDNA probe confirms the specificity of the RT-PCR. (B) Northern blot analysis was performed using samples from retina, retinal pigmented epithelium (RPE), choroid (Cho), sclera (Scl), and from the whole eye. Endogenous IGF-1 was detected only after long exposure (data not shown). (C) IGF-1 expression in transgenic retina by in situ hybridization. Sense probe did not give any signal. (D) In the transgenic retina, purple staining indicated expression of IGF-1 in the outer plexiform layer and in the inner segment of photoreceptors. No IGF-1 signal was observed in the retina of control mice. Con, control; GCL, ganglion cell layer; INL, inner nuclear layer; IS, inner segment of photoreceptors; L, lens; ONL, outer nuclear layer; OPL, outer plexiform layer; R, retina; Tg, transgenic. Scale bars: 72 μm.
IGF-1 levels in aqueous humor. IGF-1 was detected by ELISA in the aqueous humor of 2- and 14 month-old transgenic mice, but not in controls (ND, not detected). Results are mean ± SEM of eight mice in each group. **P < 0.01.
Transgenic mice overexpressing IGF-1 develop nonproliferative retinopathy. Increased production of IGF-1 in the retina of transgenic mice led to the progressive development of all stages of diabetic-like retinopathy in both eyes. One-day-old transgenic mice did not present proliferation of hyaloid vessels in the eye (data not shown). In contrast, features of nonproliferative diabetic retinopathy were observed in 2-month-old mice. Thus, transgenic eyes showed loss of pericytes (Figure 3). The quantitative study of retinal digest preparations showed a decrease of about 36% in the number of pericytes per square millimeter of capillary area (Figure 3, B and C). In addition, although we observed a slight decrease in the total endothelial cell number per square millimeter of capillary area in retinal digest preparations of transgenic eyes, the difference was not significant (Figure 3C). Nevertheless, after laser confocal microscopy analysis, several acellular capillaries were observed in the transgenic retina (Figure 3A). Furthermore, thickening of the capillary basement membrane was also observed, as detected by increased collagen IV immunostaining (Figure 4A) and by transmission electron microscopy (Figure 4B). Collagen IV is also highly increased in human diabetic retinal blood vessels (37). Measurement of the capillary basal membrane thickness in transmission electron microscopy images of transgenic retinas showed an increase of about 1.8-fold (Figure 4C). Similarly, direct injection of IGF-1 into the vitreous cavity of pigs lead to significant capillary basement membrane thickening (38).
Pericyte loss in retinal capillaries from transgenic mice overexpressing IGF-1. (A) Confocal analysis of retinal capillaries in flat-mounted retinas immunohistochemically marked with anti-collagen IV (green). Nuclei were counterstained with propidium iodide (red). Acellular capillaries, lacking pericytes and endothelial cells, were observed (arrowheads) in transgenic retina. The images are single confocal sections. (B) Retinal digest preparations were obtained as indicated in Methods. A representative area is shown (left panel). Endothelial cell nuclei are placed within the vessel wall, while pericyte nuclei are placed more laterally on the vessel wall (right panel). (C) The number of pericytes (left) and endothelial cells (right) per square millimeter of capillary area were determined as indicated in Methods. Results are mean ± SEM of five mice in each group. **P < 0.01. E, endothelial cell; P, pericyte. Scale bars: 35 μm (A), 43 μm (B, left panel), and 11 μm (B, right panel).
Thickening of capillary basement membrane was observed in transgenic retinas. (A) Confocal analysis of retinal capillaries in paraffin sections of retinas immunohistochemically marked with anti-collagen IV (green). Nuclei were counterstained with propidium iodide (red). (B) Transmission electron microscopic analysis of retinal capillaries (upper panels). Thickening of basement membrane (arrows) of transgenic retinas is shown (lower panels, magnification of insets in corresponding upper panels). (C) The basal membrane thickness was measured as indicated in Methods. Results are mean ± SEM of five mice in each group. **P < 0.01. Er, erythrocyte. Scale bars: 5.6 μm (A), 2.7 μm (B, upper panel), and 487 nm (B, lower panel).
Transgenic mice overexpressing IGF-1 develop proliferative retinopathy. Transgenic mice aged 6 months and older presented altered retinal vascularization (Figure 5A) and showed most of the features found in human diabetic retinopathy, such as intraretinal microvascular abnormalities (IRMAs) and neovascularization (3). Venule dilatation (about 60%) was noted in 6-month-old transgenic mice (control 30.6 ± 0.8 μm, n = 8, versus transgenic, 53.1 ± 3.7 μm, n = 8; P < 0.01). Furthermore, these mice showed IRMAs, such as intraretinal capillary vessels branching with an anomalous frequency and angulation (Figure 5B). Neovascularization was also observed in transgenic eyes since new vessels formed inside the retina (Figure 5, A, C, and D) and in the vitreous cavity (Figure 6). Preretinal new vessels appeared as looplike vessels that originated from veins and invaded areas of nonperfusion (Figure 5, A and C). Moreover, transgenic eyes showed dense areas of intraretinal neocapillaries (Figure 5D). Blood vessels were detected in the vitreous cavity of 6- and 12-month-old transgenic mice, as confirmed by the presence of vWF, a marker of endothelial cells (Figure 6A). Injection of Mercox resin in the thoracic aorta demonstrated that intravitreous vessels were functional and connected to the general circulation. Consistent with the morphology of angiogenic vessels (39), casts of intravitreous vessels showed more prominent and numerous imprints of endothelial nuclei than intraretinal vessels (Figure 6B).
IRMAs and neovessels in transgenic retina. (A) Flat-mounted, FITC-dextran–perfused retinas from transgenic eyes showed extensive areas of non-perfusion (asterisks) and neovessels (arrowheads and inset). (B) Capillary IRMA inside the retina is shown. (C) Neovessel originating from a venule (magnification of inset in A, right panel). This vessel had more endothelial cell nuclei after staining with propidium iodide and analysis with laser confocal microscopy (data not show). (D) Intraretinal dense area of neocapillaries in transgenic retina originating from a central feeding vessel (arrowhead). Scale bars: 630 μm (A), 188 μm (B), 87 μm (C), and 105 μm (D).
Neovessels in the vitreous cavity of transgenic eyes. (A) Blood vessels inside the vitreous (left panel) were immunodetected (right panel, magnification of inset in left panel) with anti-vWF (green). Nuclei were counterstained with propidium iodide. (B) Scanning electron microscopy analysis of Mercox vascular casts from an intravitreous neovessel (left panel) and intraretinal normal vessel (right panel). The intravitreous neovessels showed imprints of endothelial nuclei (arrowheads), which appeared to be more prominent and numerous than intraretinal vessels. Scale bars: 36 μm (A) and 11 μm (B).
Transgenic retinas show increased GFAP expression in glial cells. Similarly to human diabetic retinopathy, in which GFAP is upregulated in Müller cells (40), transgenic retinas overexpressed GFAP. This was detected by immunohistochemistry (Figure 7A) and Western blot analysis (Figure 7B). In control mice, GFAP staining was mainly confined to the ganglion cell layer around the preretinal vessels, corresponding to astrocytes and Müller cells. In contrast, transgenic mice showed highly increased GFAP in astrocytes and Müller cells, but Müller cells also expressed GFAP in the radial cytoplasmic processes extending throughout the retina (Figure 7A).
Structure of retina is altered in transgenic eyes. (A) Immunohistochemical analysis of GFAP expression in retina. Glial cells from transgenic retinas overexpressed GFAP. (B) Western blot analysis of GFAP protein in whole eye. Both 3- and 15-month-old transgenic mice showed high increase in GFAP protein. Scale bars: 81 μm (A).
Old transgenic mice show retinal detachment. In advanced diabetic eye disease, neovascularization of the vitreous may form strong fibrous adhesions between it and the retina, which may contract and thus separate the neuroretina from the retinal pigment epithelium and lead to retinal detachment (3, 41). Transgenic retina overexpressing IGF-1 folded and detached in about 75% of eyes in mice more than 6 months old (n = 21) (Figure 8, left panel). Intravitreous vessels were observed between folds of transgenic detached retinas (Figure 8, inset). Macrophages in the subretinal space, a common feature in long-standing retinal detachment (42), were also observed in the transgenic eyes (Figure 8, right panel).
The retina was folded and detached in the transgenic eye (left panel). Note the presence of vessels between retinal folds (inset magnified in upper right panel). Arrows point to endothelial cell nuclei. Stimulated macrophages, labeled with Griffonia simplicifolia lectin (green), were observed in the outer surface of detached flat-mounted transgenic retinas (lower right panel). Nuclei were counterstained with propidium iodide (red). The image is a single confocal section. CH, choroid; PE, retinal pigment epithelium; S, sclera; SE, subretinal space formed by the separation between neuroretina and pigment epithelium. Scale bars: 354 0m (left panel) and 16 μm (right panel).
IGF-1 increases VEGF expression in the eye of transgenic mice. VEGF is a potent angiogenic factor, and IGF-1 may induce its expression in vitro (29). To examine whether the vascular alterations observed in transgenic eyes also correlated with an increase in VEGF, we measured VEGF by Northern blot and Western blot analysis. About a twofold increase in VEGF transcripts in transgenic eyes was detected (Figure 9A). Furthermore, a high increase in VEGF protein was also found in both 3- and 15-month-old transgenic eye extracts compared with controls (Figure 9B). These findings correlated with increased numbers of VEGF-expressing retinal cells after immunohistochemical analysis of flat-mounted retinas (Figure 10A). In humans proliferative diabetic retinopathy glial cells produce VEGF (43). Similarly, glial cells were in the transgenic retina cells overexpressing VEGF because they also coexpressed GFAP (Figure 10B).
VEGF expression in the transgenic eye. (A) Total RNA was obtained from whole eyes and analyzed by Northern blot as indicated in Methods. Expression of VEGF transcripts was increased in 15-month-old transgenic mice. (B) Western blot analysis of VEGF protein in whole eye. Both 3- and 15-month-old transgenic mice showed a significant increase in VEGF protein. Actin documents equal loading.
Immunohistochemical analysis of VEGF. (A) Only the flat-mounted 6-month-old transgenic retinas showed extensive groups of cells producing VEGF (red). Inset shows the location in the retina. (B) VEGF (red) was detected in GFAP-expressing glial cells (green). Scale bars: 44 μm (A) and 23 μm (B).
Transgenic mice develop rubeosis iridis and neovascular glaucoma. The increase of IGF-1 in the eyes of transgenic mice was parallel to neovascularization of the iris (rubeosis iridis) (Figure 11). New vessels from the anterior face of the iris invaded corneal stroma, producing anterior synechias, the adhesions between the iris and the cornea (Figure 11A). Anterior synechias hampered drainage of aqueous humor in the iridocorneal angle and led to buphthalmos (Figure 11B, left panel). Eyes from 6-month-old mice showed a high increase in the size of the anterior chamber of the eyeball (Figure 11B, right panels) and in aqueous humor weight (about 2.6-fold) (control 2.0 ± 0.3 mg, n = 8, versus transgenic, 5.2 ± 0.5 mg, n = 14; P < 0.01). Transgenic mice also showed an increase in eye weight (control 22.7 ± 0.4 mg, n = 8, versus transgenic, 34.0 ± 0.8 mg, n = 14; P < 0.0001). Nevertheless, the aqueous humor/eye-weight ratio was higher (about 74%) in transgenic mice. When IOP was determined in 6-month-old mice, a decrease (about 37%) was observed in transgenic eyes (control, 6 ± 0.6 mmHg, n = 8, versus transgenic, 3.8 ± 0.7 mmHg (n = 5); P < 0.05). This agreed with the fact that a period of prolonged rise of IOP may be followed by a return to normotension or even hypotension (44). All these findings suggested that transgenic mice probably developed glaucoma. In addition, scanning electron microscopic analysis of critical-point dried transgenic eyes showed globular cells blocking the iridocorneal angle, while control eyes presented a trabecular meshwork free of these cells (Figure 11C). This may have reduced the outflow of aqueous humor. Furthermore, separation of ganglion cell nuclei greater than 20–30 μm indicates ganglion cell loss, a feature of glaucoma (45, 46). In contrast to control retinas, most transgenic retinas of 12-month-old mice presented highly separated (greater than 30 μm) ganglion cell nuclei (Figure 12), which suggested that these mice may have developed glaucoma.
Transgenic mice develop rubeosis iridis and neovascular glaucoma. (A) The transgenic eye developed synechias (two connected arrows) that hampered the drainage of aqueous humor (left panel). Neovessels from the iris, marked with anti-vWF (green), invade the cornea (right panel, magnification of inset in left panel). (B) Transgenic mice developed buphthalmos. The length of transgenic anterior chamber was higher than that in controls. (C) Scanning electron microscopy analysis of iridocorneal angle. (Insets in left two panels are magnified in the panels to their right.) Transgenic eyes showed an accumulation of cells (arrowheads) occluding the trabecular meshwork. In contrast, control mice presented an unobstructed trabecular meshwork (arrows). Ach, anterior chamber; EM, extraocular muscles; I, iris; ICA, iridocorneal angle; ONe, optic nerve; PB, pupillary border; Pu, pupil opening. Scale bars: 460 μm (A), 2,100 μm (B), and 900 μm (C).
Ganglion cell loss in transgenic retina. Histological sections of retina from 12-month-old transgenic mice showed decreased number of retinal ganglion cells (arrows). Scale bars: 45 μm.
Transgenic mice develop cataracts. Cataracts, opacities in any part of the lens, are frequent in diabetic patients (1, 47) and were observed in about 85% (n = 16) of 6-month-old transgenic mice after lens extraction (Figure 13A, left panel). Moreover, at 12 months all transgenic mice showed obvious cataracts by external observation (Figure 13A, right panel). Histological analysis of lenses with cataracts showed that the opacity resulted from proliferation and migration of epithelial cells to the subcortical area (Figure 13B).
Transgenic mice develop cataracts. (A) Lens from transgenic mice developed opacities (left panel), which were seen macroscopically in all 12-month-old mice (right panel). (B) Histopathological analysis of transgenic lens showed migration of epithelial cells toward the lens subcortical area (arrow). (Insets are magnified in the corresponding rectangles.) C, cornea. Scale bar: 670 μm.