Altered Expression of Retinal Occludin and Glial Fibrillary Acidic Protein in Experimental Diabetes

Alistair J. Barber; David A. Antonetti; Thomas W. Gardner; The Penn State Retina Research Group

Abstract

purpose. To investigate how diabetes alters vascular endothelial cell tight junction protein and glial cell morphology at the blood–retinal barrier (BRB).

methods. The distribution of the glial marker, glial fibrillary acidic protein (GFAP), and the endothelial cell tight junction protein occludin were explored by immunofluorescence histochemistry in flatmounted retinas of streptozotocin (STZ)-diabetic and age-matched control rats, and in BB/Wor diabetes-prone and age-matched diabetes-resistant rats.

results. GFAP immunoreactivity was limited to astrocytes in control retinas. Two months of STZ-diabetes reduced GFAP immunoreactivity in astrocytes and increased GFAP immunoreactivity in small groups of Müller cells. After 4 months of STZ-induced diabetes, all Müller cells had intense GFAP immunoreactivity, whereas there was virtually none in the astrocytes. BB/Wor diabetic rats had similar changes in GFAP immunoreactivity. Occludin immunoreactivity in normal rats was greatest in the capillary bed of the outer plexiform layer and arterioles of the inner retina but much less intense in the postcapillary venules. Diabetes reduced occludin immunoreactivity in the capillaries and induced redistribution from continuous cell border to interrupted, punctate immunoreactivity in the arterioles. Forty-eight hours of insulin treatment reversed the pattern of GFAP and occludin immunoreactivity in the STZ-diabetic rats.

conclusions. Diabetes alters GFAP expression in retinal glial cells, accompanied by reduction and redistribution of occludin in endothelial cells. These changes are consistent with the concept that altered glial–endothelial cell interactions at the BRB contribute to diabetic retinopathy.

Diabetic retinopathy is an ocular complication of diabetes with characteristic vascular and neurodegenerative components.¹ ² Some of the changes that have been reported to occur soon after the onset of diabetes are reminiscent of changes that occur in many types of central nervous system injury. Increased vascular permeability to sodium fluorescein,³ increased leukocyte adhesion,⁴ and elevated apoptosis of both neural and vascular cells⁵ ⁶ are integral components that precede the proliferative phase of the disease. Furthermore, the upregulation of GFAP that is typical after ischemia-reperfusion and other models of central nervous system injury⁷ ⁸ is induced in both human and rat retina.⁹ ¹⁰

The retina contains two types of macroglial cells. The most abundant are the Müller cells, which project from the retinal ganglion cell layer to the photoreceptors, whereas the astrocytes, which originate in the optic nerve and migrate into the retina during development,¹¹ ¹² reside as a single cell layer adjacent to the inner limiting membrane. The function of astrocytes is closely tied to that of the vasculature, because they are mostly present in vascularized retinas¹³ ¹⁴ and may guide the lateral growth of blood vessels into the retina during development.¹⁵ Müller cells are also closely associated with astrocytes, endothelial cells, and neurons and play a part in regulating the blood–retinal barrier (BRB).¹⁶ Normally, the astrocytes of the retina express GFAP, whereas Müller cells do not.¹⁷

The BRB normally regulates passive diffusion of solutes from the blood to the retinal parenchyma. The tight junctions that form the continuous seal between the vascular endothelial cells are a primary feature of the BRB.¹⁸ Occludin is an important transmembrane protein of the tight junction, responsible for forming the permeability barrier.¹⁹ ²⁰ ²¹ Tight junction protein content is reduced during the first few weeks of streptozotocin-induced diabetes (STZ-diabetes) in rats, correlating with increased permeability to serum albumin.²² There is an inverse relationship between tight junction protein content and endothelial cell permeability.²³ ²⁴ Therefore, in vivo reduction of tight junction protein content correlates with increased vascular permeability.

The glial cells and blood vessels of the retina are in close apposition and are likely to communicate directly with each other.²⁵ It has been suggested that the increase in vascular permeability in diabetic retinopathy may be due to effects of diabetes that alter the neural components of the retina, giving rise to a breakdown in the interactions between neurons, glia, and endothelial cells.²⁶ ²⁷ Breakdown of the BRB in diabetes may be due to elevated production of vascular endothelial growth factor by glial cells and neurons.²⁵ ²⁸ ²⁹ ³⁰ Conversely, tight junction protein expression is increased by factors secreted from astrocytes,³¹ which is likely to be a mechanism by which the BRB phenotype is maintained in vivo.³² Therefore, the integrity of the BRB, and thus the degree of vascular permeability, depends on factors released from the glial cells of the retina.

In this study, we examined changes in the relationship between vascular endothelial cells and glia in the retinas of STZ-diabetic and BB/Wor rats, an inbred genetic model of spontaneous-onset diabetes.³³ ³⁴ Retinal changes develop in BB rats that are similar to those in both STZ-treated rats and humans with diabetes.³⁵ The distribution of both GFAP and occludin was examined by flatmounted retina immunofluorescence histochemistry at various times after the onset of diabetes. The data presented describe an unexpected change in the expression of these proteins in diabetic rat retinas that extends previous observations of cryostat sections.¹⁰ Furthermore, the change in GFAP and occludin is reversed by a 2-day period of systemic insulin administration.

Methods

Materials and Reagents

STZ, paraformaldehyde, phosphate-buffered saline (PBS), donkey serum, and Triton X-100 were all purchased from Sigma (St. Louis, MO). Insulin administered to STZ-diabetic rats was a 1:1 mixture of regular Lente and Ultralente recombinant human insulin (Humulin; Eli Lilly, Indianapolis, IN).

Animals

For STZ-diabetic and age-matched control rats, male Sprague–Dawley rats weighing 150 to 175 g (Charles River, Wilmington, MA) were housed in the Penn State University College of Medicine animal facility in accordance with the Institutional Animal Care and Use Committee guidelines. All animal experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. All rats were group-housed in suspended wire-bottomed cages with food and water administered ad libitum, under a 12-hour light–dark schedule. Diabetes was induced by injection of STZ into a tail vein (65 mg/kg freshly dissolved in citrate buffer, pH 4.5) and confirmed 3 days later by a blood glucose level higher than 250 mg/dl (Lifescan; Johnson & Johnson, Milpitas, CA). No STZ-diabetic rats were routinely given insulin at any time during housing. A group of five 4-month STZ-diabetic rats were injected subcutaneously with 10 U of insulin twice daily for 2 days. The final injection was 4 to 5 hours before death. At death each rat was weighed, and blood glucose measured again (see Table 1 for data). Rats were killed under deep ether anesthesia followed by decapitation, and both eyes were enucleated immediately. The retinas were dissected in ice-cold PBS and fixed in fresh 2% paraformaldehyde for 10 minutes at room temperature before processing for immunohistochemistry.

BB/Wor rats were shared with Kevin McVary (Urology Department, Northwestern University Medical School, Chicago, IL). Seven diabetes-prone and six diabetes-resistant rats were used. The rats were raised in isolation by Dennis L. Gubersky (Biomedical Research Models, Worcester, MA). See Table 1 for the average blood glucose measurements during the entire period of diabetes. The mean age of the diabetes-prone group was 194 ± 1.8 days, whereas that of the diabetes resistant group was 125 ± 2.7 days. The mean duration of diabetes was 123.2 ± 2.8 days. The mean dose of insulin given to the diabetes-prone rats the day before death was 3.3 ± 0.4 U of rat insulin, which was a typical average daily dose. The rats were killed at Northwestern by isoflurane inhalation followed by decapitation. The heads were immersed in 10% normal buffered formalin for 24 hours, and then the eyes were dissected by one of the authors (AJB) at Penn State University College of Medicine. The retinas were rinsed in PBS and immediately processed for immunohistochemistry.

Immunohistochemistry and Microscopy

Whole retinas were blocked and permeabilized in 10% donkey serum with 0.3% Triton in PBS, for 1 to 2 hours. The retinas were transferred to primary antibodies diluted in block solution and incubated for three days at 4°C. The primary antibodies were mouse anti-GFAP (1:50; Roche Diagnostics, Indianapolis, IN) and rabbit anti-occludin (1:2000; Zymed, San Francisco, CA). The retinas were transferred to the secondary antibody for 24 hours at 4°C after extensive washing in PBS with 0.3% Triton. Secondary antibodies (Jackson ImmunoResearch, West Grove, PA) were CY2-conjugated donkey anti-mouse F(ab′)₂ (1:1000) and rhodamine red X–conjugated donkey anti-rabbit F(ab′)₂ (1:2000). Specimens were mounted (Aquamount; Polysciences, Warrington, PA) and viewed with a fluorescence microscope (BH-2; Olympus, Lake Success, NY) mounted with a video camera (3CCD; Sony, Tokyo, Japan) attached to a computer running image analysis software (Optimus; Media Cybernetics, Silver Spring, MD). All digital images were prepared from 640 × 480 pixel originals, with an original resolution of 213 pixels/in. Comparative digital images from diabetic and control samples were grabbed using identical brightness and contrast settings.

Optical sectioning for occludin distribution in arterioles was performed with a confocal microscope (Carl Zeiss, Thornwood, NY) with a× 40 objective and ×75 digital zoom. All confocal pictures were taken with equivalent brightness and contrast settings.

Statistical Analysis

Statistical comparisons were made on computer by unpaired t-test (Excel 98; Microsoft, Redmond, WA) or by one-way analysis of variance with post hoc Student–Newman–Keuls multiple comparisons test (Instat 2.0, GraphPad Software, San Diego, CA) for data sets containing more than two groups. Significance tests were performed with α = 0.05.

Results

Animals

The average weight and blood glucose values for the rats at the time of death are given in Table 1 . There was a significant weight loss in STZ-diabetic rats, compared with age-matched controls, both 2 and 4 months after induction of diabetes (P < 0.001). There was also an elevation in blood glucose at the time of death in these rats (P < 0.001). Four-month STZ-diabetic rats treated with insulin for 48 hours had reduced blood glucose at death that was significantly different from the untreated diabetic rats (P < 0.001) and the age-matched control group (P < 0.05). The blood glucose of BB/Wor rats was elevated in the diabetes-prone group compared with the diabetes-resistant group despite continuous insulin administration throughout the entire period of diabetes (P < 0.001).

GFAP Immunoreactivity

Previous immunoblot analysis data have shown that diabetes increases the total content of GFAP in retinas.⁹ ¹⁰ To further examine the effects of diabetes, GFAP distribution was assessed in the whole retina. GFAP immunofluorescence histochemistry was performed on flatmounted preparations of retinas from STZ-diabetic and age-matched control rats after 2 or 4 months of diabetes. Similar immunohistochemistry was also performed on retinas from spontaneously diabetic BB rat retinas. In control rat retinas, GFAP immunoreactivity was limited to the astrocytes (Fig. 1A ), as expected from results of other studies.⁹ ¹⁰ ¹⁵ In the 2-month STZ-diabetic retinas the intensity of GFAP immunoreactivity was reduced in the astrocytes (Fig. 1B) . The astrocytes were also hypertrophied, with enlarged cell bodies and multiple processes. There were also scattered regions of Müller cells with GFAP immunoreactivity determined by focusing deeper into the flatmounted retinas (Figs. 1C 1D) . Some of these regions contained autofluorescent particles that were visible in both the rhodamine and the fluorescein channels of the microscope (Fig. 1C) .

The retinas of the 4-month STZ-diabetic rats had virtually no GFAP immunoreactivity in astrocytes compared with their age-matched control subjects (Fig. 2A 2B ). The rats treated with insulin for 48 hours had increased astrocytic GFAP immunoreactivity compared with the diabetic rats (Fig. 2C) . The Müller cells in these retinas were observed by focusing through to the outer plexiform layer (identified by capillary occludin immunoreactivity in the rhodamine channel of the microscope). The Müller cells were immunoreactive for GFAP (Figs. 2D 2E) and appeared as many points of immunofluorescence, because these spindle-shaped cells are observed “end-on” in the flatmounted retina preparation. After 48 hours of insulin treatment, GFAP immunoreactivity remained positive in the Müller cells but was slightly reduced in intensity compared with the untreated STZ-diabetic rats (Fig. 2F) .

The diabetes-prone BB rats also had reduced GFAP immunoreactivity compared with the age-matched diabetes-resistant control animals (Figs. 3A 3B ). The astrocytes were hypertrophied, and GFAP immunoreactivity was associated with many long projections from the ganglion cell layer into the inner plexiform layer (arrows, Figs. 3B 3C ). Some of the retinas also contained positive GFAP immunoreactivity in the Müller cells (Fig. 3B) . These data strongly suggest that diabetes induces a differential GFAP expression pattern in the two types of macroglial cell in the retina.

Occludin Immunoreactivity

Previous immunoblot analysis studies show that diabetes reduces the total occludin content in rat retina.²² To further examine the effect of diabetes on the local distribution and expression of tight junction proteins, flatmount immunohistochemistry for occludin was also performed in the retinas described earlier. In normal rat retinas, there was occludin immunoreactivity at the cell borders in the arterioles of the inner retina (Figs. 4A 4B ). This distribution was accompanied by a small amount of punctate immunoreactivity that appeared to be located within the endothelial cell cytoplasm of these vessels. The capillaries of this layer had reduced immunoreactivity (Fig. 4C) compared with the capillaries of the outer plexiform layer (Fig. 4D) . The postcapillary venules of the inner retinal vascular bed had less occludin immunoreactivity, and the immunofluorescence intensity diminished to an almost undetectable level as the venules approached the optic disc (Figs. 4E 4F) . Thus, there was a progressive decline in occludin immunoreactivity from the arterial to the venular sides of the normal vascular tree.

After 4 months of STZ-diabetes the intensity of occludin immunoreactivity was reduced in the outer plexiform capillary bed compared with control animals (Figs. 5A 5B ). The immunofluorescence in retinas of rats treated with insulin for 48 hours approached the intensity of the control retinas (Fig. 5C) . The immunofluorescence intensity of the vessels of the inner retina was also reduced by diabetes. The reduction was most apparent in major arterioles after both 2 and 4 months of STZ-diabetes, which contained regions of heavily punctate immunoreactivity and an absence of distinct junctional distribution that was not observed in controls (Fig. 6A 6B ). The distribution and intensity of occludin immunoreactivity in the retinas from rats treated with insulin were similar to that in control rats (Fig. 6C) .

Optical sectioning of the major arterioles of the inner retina by confocal microscopy revealed that limited punctate occludin immunoreactivity was found within the cells of the vessel walls in control retinas. This punctate immunoreactivity was often associated with continuous cell border immunoreactivity (arrow, Fig. 7A ). The occurrence of punctate occludin immunoreactivity in major arterioles increased dramatically in the retinas from diabetic rats and appeared in regions where cells contained reduced or interrupted junctional immunoreactivity (Fig. 7B 7C) . These regions of highly punctate occludin immunoreactivity were always in major arterioles a short distance from the optic disc. The punctate pattern of occludin immunoreactivity was not found in any of the retinas from rats treated with insulin (Fig. 7D) . Occludin immunoreactivity was not present in the BB rat retinas due to the more extensive formalin fixation of this tissue. These findings indicate that diabetes induces regional variation in occludin expression.

Discussion

Previous studies have demonstrated that retinal glial cell metabolism is altered in diabetes.¹⁰ ³⁶ These changes include an increase in GFAP in the Müller cells of the retina⁹ and a decrease in the activity of glutamine synthetase.³⁷ Immunohistochemistry data from cryostat sections have been interpreted as indicating that STZ-diabetes increases GFAP expression in both Müller cells and astrocytes.¹⁰ The present study used immunohistochemistry applied to the flatmounted retinas and demonstrated that the increased GFAP immunoreactivity in Müller cells was accompanied by a dramatic reduction in the astrocytes. This finding is unexpected and emphasizes the value in using the flatmounted retina preparation to investigate the cell biology of the retina. The flatmount approach permits optical sectioning to accurately examine the juxtaposition of different cell types, whereas cryostat sectioning can compress and distort the tissue.

Diabetes increased GFAP in Müller cells while decreasing its expression in astrocytes. Furthermore, the morphologic changes in astrocytes are typical of reactive gliosis. The meaning of this surprising discovery is unclear, but the implication is that the astrocytes and Müller cells of the retina are affected by diabetes in different ways. Although studies of GFAP knockout mice suggest that this protein is not necessary for survival or blood–brain barrier integrity in vivo,³⁸ the dynamics of synaptic transmission are altered in the hippocampus of these mice.³⁹ Furthermore, studies on cortical astrocytes in culture indicate that GFAP is involved in regulating glutamine levels in astrocytes.⁴⁰ It is also required for induction of blood–brain barrier characteristics in aortic endothelial cells and for reactivity to β-amyloid protein.⁴¹ ⁴² Therefore, a reduction in GFAP expression in retinal astrocytes during diabetes may be linked with altered metabolic capacity and a reduced ability to induce and maintain BRB characteristics in endothelial cells.

The striking effect of diabetes on GFAP distribution was partially reversed by a short period of insulin administration, in which relatively large doses of insulin were administered to some of the STZ-diabetic rats. It should be noted that this did not establish normoglycemia, although blood glucose levels were reduced compared with the untreated STZ-diabetic rats. In these rats the astrocytic expression of GFAP was elevated toward normal, whereas the GFAP content in the Müller cells was reduced. A similar short-term administration of insulin also restored glutamine synthetase activity in STZ-diabetic rats.³⁷ These data suggest that systemic insulin may acutely regulate several aspects of glial cell metabolism in the retina. We hypothesize that prolonged insulin treatment would completely normalize the aberrant GFAP distribution in the retina of STZ-diabetic rats. While the blood glucose levels remained high in the BB/Wor rats, they were repeatedly injected with insulin on a daily basis. The less intense Müller cell GFAP immunoreactivity in the BB rat retinas suggests that repeated exposure to insulin is required to maintain the normal pattern of GFAP distribution.

Other evidence supports the possibility that insulin regulates GFAP expression in astrocytes. Insulin altered the morphology of astrocytes cultured from mouse brain and increased the expression of both GFAP mRNA and peptide.⁴³ Insulin also increased GFAP expression threefold and increased the expression and activity of glutamine synthetase⁴⁴ in rat brain astrocytes. Furthermore, a constant presence of insulin was necessary to maintain these effects, suggesting that it may be important for normal astrocytic function.

Occludin expression was also examined in normal and STZ-diabetic retinas. In the control retinas, the occludin content of the arterioles and venules of the inner retina had a strikingly differential distribution. The arterioles were strongly immunoreactive for occludin, whereas the venules had less intense immunoreactivity that diminished further as they approached the optic disc. The capillaries of the outer plexiform layer had the most intense immunoreactivity. This differential pattern of occludin expression in major arterioles and venules has been described in other tissues.⁴⁵ It is not yet clear how this relates to the vascular permeability properties of major arterioles and venules in the retina.

The reduction in occludin content, previously reported in 3-month STZ-diabetic rats,²² was confirmed by reduced immunoreactivity, particularly in the outer plexiform layer capillary bed. This reduction was accompanied by a marked redistribution of occludin immunoreactivity in the large arterioles of the inner retina. Some regions of these vessels, often close to the optic disc, had reduced tight junction immunoreactivity. These regions also had punctate immunoreactivity, giving the impression of a redistribution from cell borders to the endothelial cytoplasm. Prior immunohistochemistry in cryostat sections did not reveal the details reported here, which again emphasizes the value of the flatmounted retina preparation.

The distribution of occludin immunoreactivity is similar to that seen in vitro in bovine retinal endothelial cells and Madin–Darby canine kidney cells during growth factor–induced permeability. Therefore, the punctate occludin immunoreactivity we have described may represent the“ junction-containing vesicles” proposed recently by Antonetti et al.²⁷ as a mechanism for tight junction–regulated endothelial cell permeability.

These data strongly suggest that a primary source of vascular permeability in the STZ-diabetic rat retina is at the arterioles of the inner retina. This agrees with previous studies on vascular permeability in which extravasation of endogenous serum albumin was detected by immunohistochemistry. Albumin leakage was mostly associated with the vessels of the inner retina in human tissue and STZ-diabetic rats detected by both light and electron microscopy.⁴⁶ ⁴⁷ ⁴⁸ It is notable that the arterioles of control retinas also had a small amount of punctate immunoreactivity that was always accompanied by strong cell border immunoreactivity. This may imply that some movement of occludin from the cell border to the cytoplasm occurs during normal BRB function.

The redistribution of GFAP in the astrocytes and Müller cells of diabetic rats was reflected in changes in occludin distribution in vascular endothelial cells. Diabetes reduced the content of occludin in the capillaries of the outer plexiform layer, whereas the arterioles of the inner retina contained regions of occludin with a punctate distribution. These findings imply that astrocytes and Müller cells may respond differently to diabetes and that the subtypes of blood vessels may regulate their tight junctions by different mechanisms.

In summary, both STZ-diabetes and spontaneous BB/Wor diabetes induced changes in retinal astrocyte morphology and GFAP expression that were indicative of hypertrophy. These changes were accompanied by a reduction and redistribution of occludin. Insulin reversed the effects of diabetes on both occludin and GFAP distribution. Taken together, these data suggest that impaired vascular–glial cell interactions play an important role in the development of diabetic retinopathy.

Supported by Fight For Sight, Research Division of Prevent Blindness America (AJB), The Pennsylvania Lions Sight Conservation and Eye Research Foundation (AJB, DAA), The Juvenile Diabetes Foundation (DAA, TWG), American Diabetes Association (TWG), National Institutes of Health Grant EY 12021 (TWG), and Jack and Nancy Turner, Athens, Georgia (PSRRG).

Submitted for publication February 17, 2000; revised May 19, 2000; accepted May 31, 2000.

Commercial relationships policy: N.

Corresponding author: Thomas W. Gardner, Department of Ophthalmology, H166, Penn State University College of Medicine, Milton S. Hershey Medical Center, 500 University Drive, Hershey, PA 17033. [email protected]

Table 1.

View Table

Weight and Blood Glucose of Diabetic and Age-Matched Control Rats at Death

Table 1.

Weight and Blood Glucose of Diabetic and Age-Matched Control Rats at Death

Group	n	Weight (g)	Blood Glucose (mg/dl)
Two-month control	4	456.3 ± 22.25	80.5 ± 8.96
Two-month STZ	5	291.8 ± 9.00^*	381.4 ± 21.28^*
Four-month control	5	632.8 ± 31.14	87.0 ± 2.17
Four-month STZ	4	371.0 ± 17.97^{, †}	426.8 ± 43.33^{, †}
Four-month STZ+insulin	5	358.4 ± 15.33^{, †}	226.8 ± 65.95^{, ‡}
BB/Wor-DR	6	NA	125.6 ± 4.4
BB/Wor-DP	7	424.6 ± 18.54	314.0 ± 62.3^{, §}

Data are means ± SEM. DP, diabetes prone; DR, diabetes resistant; NA, data not available.

* P < 0.001, t-test with corresponding 2-month age-matched control group.

† P < 0.001, Student–Newman–Keuls, compared with 4-month age-matched control group.

‡ P < 0.001, Student–Newman–Keuls, compared with 4-month age-matched control group and P < 0.05 compared with untreated 4-month STZ-diabetic group.

§ P < 0.001 compared with diabetes-resistant group.

Figure 1.

View Original Download Slide

STZ-diabetes altered GFAP expression in astrocytes and Müller cells. Immunofluorescence histochemistry was performed to detect the astrocyte-specific intermediate filament GFAP in flatmounted retinas from age-matched control and STZ-diabetic rats 2 months after induction of diabetes. (A) Astrocytes were intensely immunoreactive for GFAP in age-matched control retinas. (B) Intensity of GFAP immunoreactivity was reduced in retinas of STZ-diabetic rats. (C) Some regions of astrocytes in the retinas from diabetic rats contain autofluorescent particles that can be seen in both the fluorescein and rhodamine channels of the microscope (small arrow), close to a blood vessel (large arrowheads). (D) The same field as (C) focused at a point below the astrocytes. Large arrowheads: blood vessel. Some of the Müller cells had GFAP immunoreactivity (wide arrow) and some autofluorescent particles could also be seen (narrow arrow). Scale bar, 50μ m.

Figure 2.

View Original Download Slide

Insulin partially reversed the differential GFAP expression that was induced in astrocytes and Müller cells by diabetes. Immunofluorescence histochemistry was performed to detect GFAP in flatmounted retinas from age-matched control and STZ-diabetic rats 4 months after induction of diabetes, with or without 48 hours of insulin treatment. (A) Astrocytes in age-matched control retinas were intensely immunoreactive for GFAP. (B) GFAP immunoreactivity in the astrocytes of STZ-diabetic rats was almost undetectable. (C) GFAP immunoreactivity in the astrocytes of STZ-diabetic rats was elevated after 48 hours of insulin treatment. (D) Focusing on the outer plexiform layer reveals that the Müller cells of the age-matched control rats had no GFAP immunoreactivity. (E) The Müller cells of the STZ-diabetic rats were intensely immunoreactive for GFAP. (F) After insulin treatment the STZ rats had reduced GFAP in the Müller cells, compared with that in untreated diabetic rats. Scale bar, 50 μm.

Figure 3.

View Original Download Slide

GFAP immunoreactivity and distribution were altered in a genetic model of spontaneous-onset diabetes. Immunofluorescence histochemistry was performed to detect GFAP in flatmounted retinas of age-matched BB/Wor diabetes-resistant and diabetes-prone rats, 4 months after the spontaneous onset of diabetes. (A) Diabetes-resistant rat retinas were intensely immunoreactive for GFAP. (B) The retinas of diabetes-prone rats had reduced GFAP immunoreactivity in the astrocytes, elevated immunoreactivity in Müller cells, and long immunoreactive processes from the astrocytes into the inner plexiform layer (arrows). (C) Other diabetes-prone rat retinas had less GFAP immunoreactivity in Müller cells but also had immunoreactive processes associated with astrocytes (arrows). Scale bar, 100 μm.

Figure 4.

View Original Download Slide

Occludin was differentially distributed in the blood vessels of the normal rat retina. Immunofluorescence histochemistry was performed to detect the tight junction protein occludin in flatmounted retinas of normal rats. (A) Occludin immunoreactivity was intense in the cell borders of main arterioles, and also could be detected as punctate immunoreactivity within cells (arrow). (B) The cell borders of smaller arterioles were also immunoreactive for occludin. (C) Occludin immunoreactivity in the capillaries of the inner retina (arrowheads) was less than that of the arterioles. (D) Occludin immunoreactivity of the capillaries of the outer plexiform layer was as intense as that of the arterioles. (E) Occludin immunoreactivity of the postcapillary venules (arrowheads) of the inner retina was diminished. (F) Immunoreactivity of the main venules (arrowheads) was further reduced as they approach the optic disc (right). Scale bar, 20 μm.

Figure 5.

View Original Download Slide

Insulin reversed the reduction in occludin expression that was induced by diabetes. Immunofluorescence histochemistry was performed to detect occludin in flatmounted retinas from age-matched control and STZ-diabetic rats 4 months after the induction of diabetes, with or without 48 hours of insulin treatment. (A) The cell borders of the capillaries of the outer plexiform layer of age-matched control rats were intensely immunoreactive for occludin. (B) Occludin immunoreactivity was reduced in similar capillaries of STZ-diabetic rats. (C) Occludin immunoreactivity was elevated in the capillaries of retinas from STZ-diabetic rats treated with insulin for 48 hours, when compared with the untreated diabetic rats. Scale bar, 50 μm.

Figure 6.

View Original Download Slide

Diabetes induced a punctate redistribution of occludin in main arterioles that was reversed by insulin. Immunofluorescence histochemistry was performed to detect occludin in flatmounted retinas from age-matched control and STZ-diabetic rats 4 months after the induction of diabetes, with or without 48 hours of insulin treatment. (A) The cell borders of major arterioles had occludin immunoreactivity in a control retina. (B) Occludin immunoreactivity had a punctate distribution in some regions of major arterioles of STZ-diabetic rats. (C) Insulin treatment of STZ-diabetic rats abolished the punctate occludin immunoreactivity in major arterioles. Scale bar, 50 μm.

Figure 7.

View Original Download Slide

Diabetes induced tight junction disorganization in major arterioles. Immunofluorescence histochemistry was performed to detect occludin and was viewed by confocal microscopy in flatmounted retinas from age-matched control and STZ-diabetic rats 4 months after the induction of diabetes with or without 48 hours of insulin treatment. (A) Occludin immunoreactivity in control retinas had both a continuous cell border and punctate cellular distribution. The punctate immunoreactivity was often associated with the cell border (arrow). (B, C) The punctate occludin immunoreactivity (arrows) was more abundant in major arterioles of STZ-diabetic rats, whereas the occludin immunoreactivity at the cell borders was often interrupted. (D) The frequency of punctate immunoreactivity for occludin was markedly reduced in the major arterioles of rats treated with insulin for 48 hours.

The authors thank Teresa A. Barber and Ellen B. Wolpert for their help with editing this manuscript.

Engerman RL, Kern TS. Retinopathy in animal models of diabetes. Diabetes Metab Rev. 1995;11:109–120. [CrossRef] [PubMed]

Wolter JR. Diabetic retinopathy. Am J Ophthalmol. 1961;51:1123–1139. [PubMed]

Do Carmo A, Ramos P, Reis A, Proenca R, Cunha–Vaz JG. Breakdown of the inner and outer blood retinal barrier in streptozotocin-induced diabetes. Exp Eye Res. 1998;67:569–575. [CrossRef] [PubMed]

Miyamoto K, Khosrof S, Bursell SE, et al. Prevention of leukostasis and vascular leakage in streptozotocin-induced diabetic retinopathy via intercellular adhesion molecule-1 inhibition. Proc Natl Acad Sci USA. 1999;96:10836–10841. [CrossRef] [PubMed]

Barber AJ, Lieth E, Khin SA, Antonetti DA, Buchanan AG, Gardner TW. Neural apoptosis in the retina during experimental and human diabetes: early onset and effect of insulin. J Clin Invest. 1998;102:783–791. [CrossRef] [PubMed]

Mizutani M, Kern TS, Lorenzi M. Accelerated death of retinal microvascular cells in human and experimental diabetic retinopathy. J Clin Invest. 1996;97:2883–2890. [CrossRef] [PubMed]

Osborne NN, Larsen AK. Antigens associated with specific retinal cells are affected by ischaemia caused by raised intraocular pressure: effect of glutamate antagonists. Neurochem Int. 1996;29:263–270. [CrossRef] [PubMed]

Huxlin KR, Dreher Z, Schulz M, Dreher B. Glial reactivity in the retina of adult rats. Glia. 1995;15:105–118. [CrossRef] [PubMed]

Mizutani M, Gerhardinger C, Lorenzi M. Muller cell changes in human diabetic retinopathy. Diabetes. 1998;47:445–449. [CrossRef] [PubMed]

Lieth E, Barber AJ, Xu B, et al. Glial reactivity and impaired glutamate metabolism in short-term experimental diabetic retinopathy. Penn State Retina Research Group. Diabetes. 1998;47:815–820. [CrossRef] [PubMed]

Watanabe T, Raff MC. Retinal astrocytes are immigrants from the optic nerve. Nature. 1988;332:834–837. [CrossRef] [PubMed]

Ling TL, Mitrofanis J, Stone J. Origin of retinal astrocytes in the rat: evidence of migration from the optic nerve. J Comp Neurol. 1989;286:345–352. [CrossRef] [PubMed]

Schnitzer J. Retinal astrocytes: their restriction to vascularized parts of the mammalian retina. Neurosci Lett. 1987;78:29–34. [CrossRef] [PubMed]

Schnitzer J. Astrocytes in the guinea pig, horse, and monkey retina: their occurrence coincides with the presence of blood vessels. Glia. 1988;1:74–89. [CrossRef] [PubMed]

Zhang Y, Stone J. Role of astrocytes in the control of developing retinal vessels. Invest Ophthalmol Vis Sci. 1997;38:1653–1666. [PubMed]

Hollander H, Makarov F, Dreher Z, van Driel D, Chan-Ling TL, Stone J. Structure of the macroglia of the retina: sharing and division of labour between astrocytes and Muller cells. J Comp Neurol. 1991;313:587–603. [CrossRef] [PubMed]

Bignami A, Dahl D. The radial glia of Muller in the rat retina and their response to injury: an immunofluorescence study with antibodies to the glial fibrillary acidic (GFA) protein. Exp Eye Res. 1979;28:63–69. [CrossRef] [PubMed]

Sagaties MJ, Raviola G, Schaeffer S, Miller C. The structural basis of the inner blood–retina barrier in the eye of Macaca mulatta. Invest Ophthalmol Vis Sci. 1987;28:2000–2014. [PubMed]

Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, Tsukita S. Occludin: a novel integral membrane protein localizing at tight junctions. J Cell Biol. 1993;123:1777–1788. [CrossRef] [PubMed]

Hirase T, Staddon JM, Saitou M, et al. Occludin as a possible determinant of tight junction permeability in endothelial cells. J Cell Sci. 1997;110:1603–1613. [PubMed]

McCarthy KM, Skare IB, Stankewich MC, et al. Occludin is a functional component of the tight junction. J Cell Sci. 1996;109:2287–2298. [PubMed]

Antonetti DA, Barber AJ, Khin S, et al. Vascular permeability in experimental diabetes is associated with reduced endothelial occludin content: vascular endothelial growth factor decreases occludin in retinal endothelial cells. Diabetes. 1998;47:1953–1959. [CrossRef] [PubMed]

Gardner TW, Lesher T, Khin S, Vu C, Barber AJ, Brennan WA, Jr. Histamine reduces ZO-1 tight-junction protein expression in cultured retinal microvascular endothelial cells. Biochem J. 1996;320:717–721. [PubMed]

Gardner TW. Histamine, ZO-1 and increased blood–retinal barrier permeability in diabetic retinopathy. Trans Am Ophthalmol Soc. 1995;93:583–621. [PubMed]

Stone J, Dreher Z. Relationship between astrocytes, ganglion cells and vasculature of the retina. J Comp Neurol. 1987;255:35–49. [CrossRef] [PubMed]

Gardner TW, Lieth E, Antonetti DA, Barber AJ. A new hypothesis on mechanisms of retinal vascular permeability in diabetes. Friedman EA L’Esperance FA. eds. Diabetic Renal–Retinal Syndrome. 1998;169–179. Kluwer Academic Publishers Boston.

Antonetti DA, Lieth E, Barber AJ, Gardner TW. Molecular mechanisms of vascular permeability in diabetic retinopathy. Semin Ophthalmol. 1999;14:240–248. [CrossRef] [PubMed]

Adamis AP, Miller JW, Bernal MT, et al. Increased vascular endothelial growth factor levels in the vitreous of eyes with proliferative diabetic retinopathy. Am J Ophthalmol. 1994;118:445–450. [CrossRef] [PubMed]

Stone J, Itin A, Alon T, et al. Development of retinal vasculature is mediated by hypoxia-induced vascular endothelial growth factor (VEGF) expression by neuroglia. J Neurosci. 1995;15:4738–4747. [PubMed]

Amin RH, Frank RN, Kennedy A, Eliott D, Puklin JE, Abrams GW. Vascular endothelial growth factor is present in glial cells of the retina and optic nerve of human subjects with nonproliferative diabetic retinopathy. Invest Ophthalmol Vis Sci. 1997;38:36–47. [PubMed]

Gardner TW, Lieth E, Khin SA, et al. Astrocytes increase barrier properties and ZO-1 expression in retinal vascular endothelial cells. Invest Ophthalmol Vis Sci. 1997;38:2423–2427. [PubMed]

Janzer RC, Raff MC. Astrocytes induce blood-brain barrier properties in endothelial cells. Nature. 1987;325:253–257. [CrossRef] [PubMed]

Chappel CI, Chappel WR. The discovery and development of the BB rat colony: an animal model of spontaneous diabetes mellitus. Metab Clin Exp. 1983;32:8–10. [CrossRef] [PubMed]

Mordes JP, Desemone J, Rossini AA. The BB rat. Diabetes Metab Rev. 1987;3:725–750. [CrossRef] [PubMed]

Sima AA, Chakrabarti S, Garcia–Salinas R, Basu PK. The BB-rat: an authentic model of human diabetic retinopathy. Curr Eye Res. 1985;4:1087–1092. [CrossRef] [PubMed]

Lieth E, Ratz MJ, LaNoue KF, et al. Elevated glutamate in retinas of short-term diabetic rats [ARVO Abstract]. Invest Ophthalmol Vis Sci. 1998;39(4)S826.Abstract nr 3830

Lieth E, LaNoue KF, Antonetti DA, Ratz M, Penn State Retina Research Group. Diabetes reduces glutamate oxidation and glutamine synthesis in the retina. Exp Eye Res. 2000;70:723–730. [CrossRef] [PubMed]

Pekny M, Leveen P, Pekna M, et al. Mice lacking glial fibrillary acidic protein display astrocytes devoid of intermediate filaments but develop and reproduce normally. EMBO J. 1995;14:1590–1598. [PubMed]

McCall MA, Gregg RG, Behringer RR, et al. Targeted deletion in astrocyte intermediate filament (GFAP) alters neuronal physiology. Proc Natl Acad Sci USA.. 1996;93:6361–6366. [CrossRef]

Pekny M, Eliasson C, Siushansian R, et al. The impact of genetic removal of GFAP and/or vimentin on glutamine levels and transport of glucose and ascorbate in astrocytes. Neurochem Res. 1999;24:1357–1362. [CrossRef] [PubMed]

Pekny M, Stanness KA, Eliasson C, Betsholtz C, Janigro D. Impaired induction of blood-brain barrier properties in aortic endothelial cells by astrocytes from GFAP-deficient mice. Glia. 1998;22:390–400. [CrossRef] [PubMed]

Xu K, Malouf AT, Messing A, Silver J. Glial fibrillary acidic protein is necessary for mature astrocytes to react to beta-amyloid. Glia. 1999;25:390–403. [CrossRef] [PubMed]

Toran–Allerand CD, Bentham W, Miranda RC, Anderson JP. Insulin influences astroglial morphology and glial fibrillary acidic protein (GFAP) expression in organotypic cultures. Brain Res. 1991;558:296–304. [CrossRef] [PubMed]

Aizenman Y, de Vellis J. Synergistic action of thyroid hormone, insulin and hydrocortisone on astrocyte differentiation. Brain Res. 1987;414:301–308. [CrossRef] [PubMed]

Kevil CG, Okayama N, Trocha SD, et al. Expression of zonula occludens and adherens junctional proteins in human venous and arterial endothelial cells: role of occludin in endothelial solute barriers. Microcirculation. 1998;5:197–210. [PubMed]

Vinores SA, Gadegbeku C, Campochiaro PA, Green WR. Immunohistochemical localization of blood–retinal barrier breakdown in human diabetics. Am J Pathol. 1989;134:231–235. [PubMed]

Vinores SA, Van Niel E, Swerdloff JL, Campochiaro PA. Electron microscopic immunocytochemical demonstration of blood–retinal barrier breakdown in human diabetics and its association with aldose reductase in retinal vascular endothelium and retinal pigment epithelium. Histochem J. 1993;25:648–663. [CrossRef] [PubMed]

Vinores SA, McGehee R, Lee A, Gadegbeku C, Campochiaro PA. Ultrastructural localization of blood–retinal barrier breakdown in diabetic and galactosemic rats. J Histochem Cytochem. 1990;38:1341–1352. [CrossRef] [PubMed]

Figure 1.

View Original Download Slide