Diabetes-Induced Disruption of Gap Junction Pathways within the Retinal Microvasculature

Abstract

purpose. Microvascular damage caused by diabetes is a leading cause of visual loss. Identifying events early in the course of diabetic retinopathy may help in understanding and, perhaps, preventing this disorder. The hypothesis that cell-to-cell communication within the retinal microvasculature may be affected soon after the onset of diabetes was tested.

methods. Streptozotocin was used to induce diabetes in rats. To assess cell-to-cell coupling the gap junction-permeant tracer, Neurobiotin, was delivered via patch pipettes into pericytes located on microvessels freshly isolated from the retinas of diabetic and control animals. Subsequently, immunohistochemical methods revealed the extent of the intercellular spread of the tracer. Electrophysiological methods were also used to detect intercellular communication.

results. In retinal microvessels of control rats, Neurobiotin spread hundreds of micrometers from the tracer-loaded pericytes. However, within days after the onset of diabetes, this cell-to-cell coupling was dramatically reduced. In contrast, microvessels of insulin-treated diabetic rats showed no significant loss of intercellular communication. Consistent with protein kinase C (PKC) playing a role in the diabetes-induced inhibition of gap junction pathways, exposure of microvessels to a PKC activator (phorbol myristate acetate) markedly reduced tracer coupling.

conclusions. Within retinal microvessels there is extensive cell-to-cell coupling, which is markedly reduced soon after the onset of streptozotocin-induced diabetes. The closure of gap junction pathways disrupts the multicellular organization of retinal microvessels and may contribute to vascular dysfunction.

Microvessels of the retina are particularly vulnerable to damage induced by diabetes. An event early in the course of diabetic retinopathy is the loss of pericytes, ¹ which are cells positioned on the ablumenal surface of a capillary’s endothelial-lined lumen. However, well before the onset of pericyte death, diabetes causes functional changes in the retinal vasculature. These include alterations in the rate of blood flow, ^{2 3} breakdown in the blood–retinal barrier ^{4 5} and loss of autoregulatory mechanisms for adjusting capillary perfusion to meet local metabolic demand. ⁶ Because these functions almost certainly depend on interactions among vascular cells, we hypothesized that diabetes may interfere with cell-to-cell communication within microvessels of the retina. 

One mechanism by which cells may interact in the retinal microvasculature is via gap junctions. ⁷ Although definitive evidence of cell-to-cell coupling within retinal microvessels is scant, it seems likely that intercellular pathways exist because ultrastructural studies have demonstrated gap junctions between pericytes and endothelial cells in other parts of the brain. ^{8 9} To establish that retinal microvessels do have intercellular pathways, we loaded pericytes with Neurobiotin (MW, 286 Da), which is a tracer that diffuses through gap junctions. Subsequently, we looked for the intercellular spread of this tracer. In other experiments, analysis of membrane capacitive currents helped to confirm that pericytes are normally coupled to other cells. Because these tracer and electrophysiological assays are not currently feasible in vivo, we used microvessels that were freshly isolated from the retina. ¹⁰  

To assess the effect of diabetes on gap junction pathways within retinal microvessels, we injected rats with streptozotocin (STZ), which destroys insulin-secreting cells of the pancreas and, thereby, creates an experimental model of type 1 diabetes. Here we report that normally there is extensive cell-to-cell coupling within retinal microvessels. This intercellular communication is dramatically reduced soon after the onset of diabetes. However, when diabetic animals are treated with insulin, the gap junction pathways persist, and the multicellular organization of the retinal microvasculature is unaffected. 

Methods

Rat Model of Diabetes

After an overnight fast, 25 8.9 ± 0.1-week-old Long-Evans rats (Harlan Sprague–Dawley, Inc., Indianapolis, IN) received an intraperitoneal injection of STZ (75 mg/kg) diluted in 0.8 ml citrate buffer. This study conformed to the guidelines of the Association for Research in Vision and Ophthalmology and the University of Michigan University Committee on the Use and Care of Animals. When blood glucose levels were measured (One Touch Basic; Lifescan, Milpitas, CA) 2 days later, three rats had glucose levels of <200 mg/dl and received a second dose of STZ. Another group of rats was injected with the same amount of buffer solution. Animals were maintained on a 12-hour alternating light/dark cycle and received food and water ad libitum. Immediately before the harvesting of the retinal microvessels, the blood glucose level was 339 ± 14 mg/dl in the diabetic rats that were not treated with insulin. A subset of diabetic rats were treated three times per day for 6 to 8 days with SC administered insulin (mean dose, 22.8 U bovine insulin/kg per day); for this group the blood glucose was 103 ± 26 mg/dl at the time of death. The number of rats used in this study was 19, 22, and 3 for the control, untreated diabetic, and insulin-treated diabetic groups, respectively. Another 15 normal rats were used to test the effects doxyl-stearic acid, CO₂, metabolic inhibitors, and phorbol myristate acetate on microvessels isolated from the retina. 

Microvessel Isolation

Freshly isolated retinal microvessels were prepared using a modified “tissue-print” method. ¹⁰ For each experiment, retinas from a euthanatized rat were rapidly removed and incubated in 2.5 ml Earle’s balanced salt solution (Life Technologies, Grand Island, NY) supplemented with (mM): 0.5 EDTA, 1.5 CaCl₂, 1 MgSO₄, 20 glucose, 26 sodium bicarbonate, 15 U papain (Worthington Biochemicals, Freehold, NJ), 0.04% DNase, and 2 mM cysteine for 30 minutes at 30°C, while 95% O²–5% CO₂ was bubbled through to maintain pH and oxygenation. After transfer to solution A (140 mM NaCl, 3 mM KCl, 1.8 mM CaCl₂, 0.8 mM MgCl₂, 10 mM Na-Hepes, 15 mM mannitol, and 5 mM glucose at pH 7.4 with osmolarity adjusted to 310 mosmol l⁻¹), each retina was briefly sandwiched gently between two glass coverslips (diameter, 15 mm; Warner Instrument Corp., Hamden, CT). Vessels adhered to the coverslip that contacted the vitreal surface of the retina. We do not know whether the adhering microvessels are from the inner and/or deeper vascular layers of the retina. The coverslip containing microvessels was placed in a recording chamber containing a solution A. 

Tracer Labeling

Experiments were performed at room temperature using microvessels that had been isolated from the retina within 2 hours. Vessels were examined at ×400 magnification with an inverted microscope equipped with phase-contrast optics. Pericytes were identified by their characteristic location on the ablumenal wall of microvessels that had outer diameters of <7 μm. ¹¹ A patch pipette containing 0.5% Neurobiotin (Vector Laboratories, Burlingame, CA), 25 mM KCl, 105 mM potassium aspartate, 1.5 mM CaCl₂, 2 mM MgCl₂, 3 mM K-EGTA, and 10 mM K-Hepes (pH 7.4, osmolarity ∼280 mosmol/l) was mounted in the holder of a Dagan 3900 patch-clamp amplifier (Dagan Corp., Minneapolis, MN) and positioned onto the soma of a retinal pericyte located on a freshly isolated retinal microvessel. Application of gentle suction to the back end of the pipette created a gigaohm seal. After the patch of cell membrane at the tip of the pipette was broken by applying progressively increasing amounts of suction via a pneumatic transducer (Bio-Tek, Winoosik, VT), a voltage of +50 mV was applied to the pipette for 5 minutes to enhance movement of Neurobiotin from the pipette into the sampled pericyte. The Neurobiotin-containing pipette was then removed, and the microvessel was left undisturbed for 40 minutes; the location of the sampled pericyte was documented in a sketch. In some experiments (see Fig. 4B ), Neurobiotin was loaded using a similar protocol into vascular smooth muscle cells encircling retinal vessels, which had outer diameters of 15 to 40 μm. After approximately 18 hours of fixation at 4°C in a phosphate buffer (PBS) containing 4% paraformaldehyde, the microvessel-containing coverslip was washed in PBS, and the endogenous peroxidase activity of the vessels was blocked by exposure to 0.03% hydrogen peroxide in PBS for 30 minutes. After a 60-minute incubation in 0.5% Triton X-100, the coverslip was exposed for 40 to 65 hours at 4°C to a horseradish peroxidase–streptavidin solution (1 μg/ml, RTU; Vector Laboratories) supplemented with 0.5% Triton X-100. Subsequently, microvessels were developed for 5 minutes in diaminobenzidine plus nickel (DAB kit; Vector Laboratories). After the vessels were counterstained with methylgreen, they were viewed at ×100 magnification with a microscope equipped with bright field optics, and the extent of Neurobiotin staining was measured. 

In experiments to help confirm that a large molecular weight tracer remains within a sampled pericyte, fluorescein isothiocynate (FITC)-dextran (MW, 2,000,000) was loaded into pericytes via a patch pipette containing 50 mg/ml of this tracer. Forty minutes after removing the pipette, the coverslip containing the microvessel was mounted with Gel/mount (Biomeda, Foster City, CA) on a microscope slide and viewed at ×500 with an Olympus (Lake Success, NY) Vanox microscope equipped with fluorescent illumination. 

Electrophysiology

We used the perforated-patch configuration of the patch-clamp technique to record the capacitive currents of pericytes located on microvessels, which had been isolated from the retina within 2 hours. Analysis of current relaxations evoked by a 10-mV voltage step permitted the detection of cell-to-cell coupling. ^{12 13 14} In these experiments, the pipette solution consisted of 50 mM KCl, 65 mM K₂SO₄, 6 mM MgCl₂, 10 mM K-Hepes, 240 μg ml⁻¹ amphotericin B, and 240 μg ml⁻¹ nystatin at pH 7.4, with the osmolarity adjusted to 280 mosmol l⁻¹. The pipettes, which had resistances of approximately 5 MΩ, were mounted in the holder of a patch-clamp amplifier (Dagan Corp.) and sealed to the cell bodies of pericytes. Compensation of the pipette capacitive current was applied via circuits within the amplifier. As amphotericin/nystatin perforated the patch, the access resistance decreased to <20 MΩ for the pericytes studied. Currents were evoked by a voltage step protocol controlled by pClamp 8 software (Axon Instruments, Inc., Foster, CA), filtered at 10 kHz with a four-pole Bessel filter, digitally sampled at 50-μsec intervals using a Digidata 1200B acquisition system (Axon Instruments), and stored by a Pentium class computer. Curve-fitting software (Origin 6; Microcal, Northampton, MA) was used to calculate exponential functions describing the decay of the membrane capacitive currents. 

Chemicals

Chemicals were from Sigma/RBI (St. Louis, MO) unless noted otherwise. 

Statistics

Data are given as means ± SEM. Unless otherwise noted, probability was evaluated by the Student’s t-test. 

Results

To assess cell-to-cell coupling within the microvasclature, we used patch pipettes to load Neurobiotin into pericytes that were located on microvessels freshly isolated from the rat retina. After allowing 5 minutes for the tracer to enter the pericyte, the pipette was removed, and 40 minutes was allotted for intercellular diffusion. Subsequently, we found that Neurobiotin had spread hundreds of micrometers from the sampled pericyte (Figs. 1A 2 ; see also Fig. 4A ). In a series of 28 microvessels from 13 rats, the length of microvessel stained with this tracer was 452 ± 31 μm. This is nearly 14-fold longer than the length (33 ± 7 μm, n = 20) of staining observed under experimental conditions in which Neurobiotin was restricted to the tracer-loaded perictye (details of experiments in which single pericytes were stained are presented below). 

Consistent with Neurobiotin passing through gap junctions within the retinal microvasculature, the spread of this tracer was dramatically reduced by DSA (30 μM, 16-doxyl-stearic acid), which is a gap junction uncoupler that we tested on seven microvessels from four rats (Figs. 1B 2) . We also tested the effect of ischemia, which is known to close gap junctions in other cellular systems. ¹² Exposure of six microvessels from 3 rats to a bathing solution containing inhibitors of ATP synthesis markedly (P < 0.001) diminished the spread of the tracer from sampled pericytes (Figs. 1C 2) . Similarly, saturating the bathing solution with CO₂ significantly (P < 0.001) reduced tracer spread in five microvessels from three rats (Fig. 2) . In other experiments, we loaded pericytes with FITC-dextran (MW, 2,000,000), which is too large to pass through gap junctions. In each of three pericytes loaded with this tracer, fluorescence was limited to the sampled cell. The spread of fluorescence was 23 ± 8 μm, which was not significantly (P = 0.25) different from the length (33 ± 7 μm) of Neurobiotin staining under conditions in which only a single pericyte was labeled. From these various experiments, we conclude that Neurobiotin spreads via gap junctions from a pericyte to other microvascular cells. 

Analyzing the membrane capacitive currents of pericytes provided further evidence for cell-to-cell coupling in retinal microvessels. Under control conditions, the rate of decay of the transient current evoked by a voltage step was poorly fit by a single exponential function (Fig. 3 , left). Rather, this current decayed as the sum of multiple exponential functions of time. A complex rate of decay, such as that observed under control conditions, is predicted for cells that are interconnected in series. ^{13 14 15} However, with exposure of a retinal microvessel to a gap junction uncoupler, the decay of a pericyte’s capacitive current could then be fit by a single exponential function (Fig. 3 , right) as predicted for a single cell. ¹⁶ Electrophysiological findings similar to those shown in Figure 3 were made in five pericytes (from 3 rats) that were located on freshly isolated retinal microvessels. Taken together, our tracer studies and electrophysiological experiments suggest that pericyte-containing microvessels of the retina are organized via gap junctions into extensive multicellular networks. 

Although Neurobiotin extended for long distances from the sampled pericytes, we did not detect diffusion of this tracer into larger adjoining vessels that were enveloped by vascular smooth muscle cells (Fig. 4A) . A similar compartmentalization of tracer coupling was observed in each of five sampled microvessels (from 3 rats) that branched from a larger, myocyte-encircled vessel. Conversely, Neurobiotin loaded into vascular smooth muscle cells did not spread (n = 8 vessels from 7 rats) into branching microvessels that were covered by pericytes (Fig. 4B) . Thus, these tracer experiments indicate that pericyte-containing microvessels possess an extensive intercellular network that appears to be separate from the vascular smooth muscle system. 

The question arose as to whether Neurobiotin spreads from a sampled pericyte exclusively to other pericytes or also to endothelial cells. Although a definitive answer may require electron microscopic evaluation, our findings suggest that this tracer does pass from pericytes to the vascular endothelium. This is based on our observation that single pericytes stained with Neurobiotin typically did not completely cover the underlying endothelium (Figs. 1B 1C) . However, when this tracer spread from a sampled pericyte, the entire width of the retinal microvessel was stained (Figs. 1A 4A) . Thus, it appears that Neurobiotin can spread from pericytes to endothelial cells in the retinal microvasculature. 

Because microvascular dysfunction occurs early in the course of diabetes, ^{2 3 4 5 6} we hypothesized that the multicellular organization of these vessels may be disrupted in diabetic animals. To test this possibility, we assessed cell-to-cell communication in microvessels freshly isolated from rats made diabetic by STZ. In this experimental model, the spread of Neurobiotin from labeled pericytes was markedly reduced soon after the onset of hyperglycemia (Figs. 5 6A) . In contrast, there was no decrease in the spread of Neurobiotin in animals that received only the buffer that was used to dissolve STZ. Specifically, within 6 days after STZ administration, the distance that the tracer spread decreased by 38% (P = 0.008) to 278 ± 68 μm (n = 8 microvessels from 3 rats). During the second week of diabetes, labeling diminished to 215 ± 31 μm (n = 28 microvessels from 16 rats). This reduction in cell coupling persisted for at least 2 months, at which time Neurobiotin labeled only 132 ± 67 μm (n = 6 microvessels from 3 rats) of the sampled microvessels. 

Quantification of the incidence of cases in which Neurobiotin appeared to remain exclusively in the tracer-loaded pericyte (e.g., Fig. 6A ) provided further evidence that diabetes disrupted intercellular communication. A lack of evidence for cell-to-cell coupling was observed in 2.5% (1 of 40) of the microvessels studied in controls. In contrast, during the second week of diabetes, we did not detect a diffusion of Neurobiotin from 15.8% (6 of 38) of the sampled pericytes. This was a significant (P = 0.03, Fisher’s exact test) change, which persisted for at least 2 months of diabetes, at which time two of six diabetic microvessels had tracer limited to the Neurobiotin-loaded pericytes. Taken together, our experiments with diabetic rats indicate that STZ-induced diabetes causes a loss of cell-to-cell coupling in microvessels of the retina. 

Finding that diabetes reduced gap junction communication led us to test the possibility that this disruption of microvascular organization could be ameliorated with insulin therapy. We found that the microvessels of STZ-injected rats, which were hyperglycemic for 3 days and then treated for 6 to 8 days with insulin three times per day, showed significantly (P = 0.004) more spread of Neurobiotin than was observed in the microvessels of untreated diabetic animals (Figs. 6 and 7) . With this treatment protocol, tracer coupling within the retinal microvessels was not significantly (P = 0.4) different from in the nondiabetic controls. These results indicate that it is very unlikely that the loss of cell-to-cell coupling in STZ-induced diabetes was due to a direct toxic effect of STZ on the microvessels. Rather, hyperglycemia and/or insulin depletion appear to be the causative factors for a rapid loss of intercellular coupling within the retinal microvasculature. 

One mechanism by which diabetes may damage blood vessels is secondary to a glucose-induced activation of protein kinase C (PKC). ¹⁷ To begin to examine this possibility, we assessed the effect of a PKC activator, phorbol myristate acetate (PMA). In a series of experiments, we observed that exposure of freshly isolated retinal microvessels to 100 nM PMA for 66 ± 10 minutes (n = 10 microvessels from 5 rats) was associated with a significant (P < 0.001) reduction in the intercellular spread of Neurobiotin from tracer-loaded pericytes. Namely, this phorbol ester reduced the length of Neurobiotin staining to 56 ± 16 μm, which was 87% less than in the controls. Also, in the PKC group, Neurobiotin staining was limited exclusively to the tracer-loaded pericyte in 5 of 10 sampled microvessels. This contrasts (P = 0.003, Fisher’s exact test) with only 1 of 28 sampled pericytes lacking cell-to-cell coupling in the control group. These experiments with PMA are consistent with the hypothesis that an elevation of PKC isoforms may play a role in the closure of gap junction pathways within the retinal microvasculature. 

Discussion

The results show that there is extensive cell-to-cell coupling within retinal microvessels freshly isolated from the adult rat. Soon after the onset of experimental diabetes, this intercellular network is compromised. However, loss of cell coupling did not occur when diabetic rats were treated with insulin. Consistent with a putative role for protein kinase C in mediating diabetic complications, ¹⁷ exposure to a phorbol ester markedly reduced the spread of Neurobiotin within isolated retinal microvessels. 

This is the first demonstration of tracer coupling within a pericyte-containing microvessel. Although not technically feasible at present, it will be important to obtain in vivo confirmation of our tracer and electrophysiological evidence for cell-to-cell coupling within retinal microvessels. Our observation that Neurobiotin spreads from pericytes to other microvascular cells concurs with ultrastructural evidence of gap junctions located between pericytes and endothelial cells in the central nervous system. ^{8 9} Our findings also extend those of Larson and colleagues, ^{18 19} who detected the expression of connexin 43 by cultured pericytes and the junctional transfer of small molecules between brain pericytes and endothelial cells in culture. 

Our conclusion that gap junctions extensively couple cells of the retinal microvasculature is similar to inferences concerning the functional linkage of cells within capillaries of skeletal muscle. ^{20 21 22} However, unlike the situation in skeletal muscle, ^{20 21 22} we did not detect intercellular communication between pericyte-containing retinal vessels and those encircled by vascular myocytes. This difference may reflect different functional roles for the microvessels in the retina compared with those in muscle. For example, capillaries are thought to have a passive role in peripheral tissues where precapillary sphincters regulate local perfusion. ^{20 21 22 23} In contrast, because the retinal vasculature lacks precapillary sphincters, ²⁴ receives no extrinsic innervation, ²⁵ and contains the highest density of contractile pericytes, ²⁶ blood flow in the retina may be regulated, at least in part, at the capillary level. ^{27 28 29} Thus, the compartmentalization of gap junction pathways within the retinal vasculature is consistent with the idea that the pericyte-containing microvessels constitute functional units that may actively regulate retinal blood flow. 

Our finding that cell-to-cell coupling within the retinal microvasculature is reduced within 6 days after the administration of streptozotocin is one of the earliest functional changes observed in an experimental model of diabetes. This decrease in intercellular communication is unlikely to be due to a direct toxic effect of STZ because cell coupling was normal in STZ-injected rats that were subsequently treated with insulin. Also, the progressive loss of cell coupling during 2 months of diabetes suggests that hyperglycemia and/or insulin deficiency are the causative factors for a breakdown in the multicellular organization of the retinal microvasculature. Clearly, future studies are needed to elucidate the mechanism(s) by which this gap junction communication is compromised early in the course of diabetes. 

In addition to the closure of gap junction pathways, a variety of other retinal vascular changes are observed in rats with STZ-induced diabetes. Changes detected by the beginning of the second week include an alteration in the rate of retinal blood flow ² and a breakdown of the blood–retinal barrier. ⁵ Although the interrelationships of these vascular abnormalities are unclear, it may be that the rapid loss of gap junction communication disrupts intercellular pathways that are essential for the maintenance of normal blood flow parameters and barrier properties. In addition, over a longer term, the loss of cell-to-cell communication may metabolically isolate pericytes and, thereby, contribute to their demise early in the course of diabetic retinopathy. 

²  TK and HO contributed equally to the work.

 Supported by National Institutes of Health Grants EY12505 and EY07003 and the Japanese Ministry of Education. DGP received a Research to Prevent Blindness Senior Scientist Award.

 Submitted for publication December 15, 2000; revised February 20, 2001; accepted March 27, 2001.

 Commercial relationships policy: N.

 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.

 

Neurobiotin staining of microvessels freshly isolated from the rat retina. The gap junction-permeant tracer, Neurobiotin, was loaded into pericytes under control conditions (A), in the presence of 30 μM DSA for 4 minutes (B) and under conditions of chemical ischemia induced for 23 minutes by 0.5 mM iodoacetate and 0.5μ M antimycin A (C). Arrows show the pericyte that was loaded with Neurobiotin via a patch pipette. Consistent with intercellular communication via gap junctions, DSA and ischemia profoundly reduced the spread of Neurobiotin within retinal microvessels.

Figure 2.

 

Effects of gap junction uncouplers. The bathing solution during the loading of Neurobiotin into pericytes consisted of solution A with no additives (control), with 30 μM DSA for 4.1 ± 0.3 minutes, with 0.5 mM iodoacetate (IAA) and 5 μM antimycin A (AA) for 21 ± 2 minutes, or saturated with CO₂ for 22 ± 8 minutes. A total of 46 microvessels from 26 rats were sampled (12 ± 5 vessels per group). In retinal microvessels, conditions that close gap junctions limit the spread of Neurobiotin from sampled pericytes.

Figure 3.

 

Effect of octanol, a gap junction uncoupler, on the membrane capacitive current of a pericyte located on a freshly isolated microvessel. Each panel shows the voltage step applied, the evoked current and the calculated first order exponential fit (dotted lines) for the decay of this current. Initially (left), the recording chamber was perfused with solution A. Subsequently (right), the perfusate was supplemented with 0.5 mM octanol. The sensitivity of the capacitive current to octanol is consistent with gap junctions interconnecting cells within retinal microvessels.

Figure 4.

 

Neurobiotin staining in retinal vascular complexes. In both (A) and (B), one or more pericyte-containing microvessels branched from a larger vessel, which was encircled by vascular smooth muscle. Black arrows, site of tracer loading; white arrows, vessel that lacked tracer staining. In the retinal vasculature, the multicellular network of pericyte-containing vessels and of myocyte-enveloped vessels does not interact via Neurobiotin-permeable pathways.

Figure 5.

 

Effect of STZ-induced diabetes on the extent of Neurobiotin staining in retinal microvessels. At various times after rats received STZ or vehicle only, their retinal microvessels were isolated, and one pericyte per sampled microvessel was loaded with Neurobiotin. Totals of sampled microvessels were 30 (10 ± 4 per point) from 13 rats and 49 (12 ± 1 per point) from 22 rats for the controls and the diabetic group, respectively. Diabetes, which is induced by STZ, causes a rapid reduction in cell-to-cell coupling within retinal microvessels.

Figure 6.

 

Neurobiotin staining of retinal microvessels from untreated and insulin-treated diabetic rats. (A) A retinal microvessel isolated from a rat made diabetic by STZ injection 9 days earlier. The blood glucose level before death was 423 mg/dl. (B) An isolated retinal microvessel from a diabetic rat treated with insulin for 7 days before death; immediately before killing, the blood glucose level was 56 mg/dl. Arrows, the pericyte that had been loaded with the tracer. The intercellular spread of Neurobiotin decreases in retinal microvessels of diabetic rats.

Figure 7.

 

Effect of insulin treatment on the extent of Neurobiotin staining in retinal microvessels of diabetic rats. Control animals (n = 3) received vehicle only; both diabetic groups were injected with STZ (n = 10); a subset of diabetics (n = 3) received insulin three times per day for 6 to 8 days before death. Microvessels were isolated 10.2 ± 0.3 days after the onset of experimental manipulations (i.e., injection of STZ or vehicle). The number of microvessels studied was 6, 11, and 19 in the control, insulin-treated diabetic, and untreated diabetic groups, respectively. Treatment with insulin prevents the diabetes-induced reduction in cell-to-cell coupling within retinal microvessels.

The authors thank Qing Li for providing instruction on the creation of STZ-induced diabetic rats. They also thank Qing Li, Bret Hughes, Peter Hitchcock and David Wu for helpful discussions.