Des(1–3)IGF-1 Treatment Normalizes Type 1 IGF Receptor and Phospho-Akt (Thr 308) Immunoreactivity in Predegenerative Retina of Diabetic Rats

Little is known about interventions that may prevent predegenerative changes in the diabetic retina. This study tested the hypothesis that immediate, systemic treatment with an insulin-like growth factor (IGF)-1 analog can prevent abnormal accumulations of type 1 IGF receptor, and phospho-Akt (Thr 308) immunoreactivity in predegenerative retinas of streptozotocin (STZ) diabetic rats. Type 1 IGF receptor immunoreactivity increased approximately 3-fold in both inner nuclear layer (INL) and ganglion cell layer (GCL) in retinas from STZ rats versus nondiabetic controls. Phospho-Akt (Thr 308) immunoreactivity increased 5-fold in GCL and 8-fold in INL of STZ rat retinas. In all cases, immunoreactive cells were significantly reduced in STZ des(1–3)IGF-1–treated versus STZ rats. Preliminary results suggested that vascular endothelial growth factor (VEGF) levels may also be reduced. Hyperglycemia/ failure of weight gain in diabetic rats continued despite systemic des(1–3)IGF-1. These data show that an IGF-1 analog can prevent early retinal biochemical abnormalities implicated in the progression of diabetic retinopathy, despite ongoing hyperglycemia.

The pathogenesis of diabetic retinopathy is a complex process involving ischemic/hyperglycemic and growth factor retinal insults that can result in neovascularization and vision loss. The incidence of diabetic retinopathy can be reduced somewhat when blood glucose is well-controlled [1]. Although early glucose control may be important in delaying the onset of diabetic retinopathy, glucose control alone, unfortunately, cannot halt the eventual progression of retinopathy [2]. There is a pressing need for novel interventions to supplement glycemic control.
Insulin-like growth factor-1 (IGF-1) is among several factors that have been suggested to regulate predegenerative abnormalities, including early elevation of vascular endothelial growth factor (VEGF) levels in the retina [3,4]. VEGF has been identified as a causative factor in retinal neovascularization as well as vascular permeability [5,6] associated with proliferative diabetic retinopathy. There is controversy as to whether serum or vitreous IGF-1 levels correlate with the progression of retinal neovascularization in clinical diabetes. Some studies report no correlation [7], whereas others report correlations with either elevated or decreased levels of IGF-1 in the vitreous or serum of diabetic patients with retinopathy [8][9][10]. Disparities in these reports may be due to the methods and biological samples used for analysis (mRNA versus protein, vitreous versus serum), but also to differences in the extent of blood-retinal barrier (BRB) breakdown at the time of sample collection. Recent studies point to increased STZ and Glucose Diagnostic Kit 510A were purchased from Sigma Chemical (St. Louis, MO). Des(1-3)IGF-1 was from GroPep (Adelaide, Australia). Primary rabbit polyclonal antibody against Phospho-Akt (Thr 308) (Cell Signaling Technologies, Beverly, MA) as well as mouse monoclonal antibodies against type 1 IGF receptor (Calbiochem, San Diego, CA) and VEGF (Calbiochem) were obtained. Alzet osmotic minipumps (0.5 µL/h; 2-week duration) were from Durect (Cupertino, CA).

Animals
Animal experiments were performed in accordance with National Institutes of Health (NIH) guidelines (DHEW publication NIH80-23). Sprague-Dawley (Harlan Sprague-Dawley, Indianapolis, IN) male rats were maintained on 20 g per day of rat chow until the study, and chow and water were provided ad libitum thereafter. Rats (12 weeks old) were randomly assigned to treatment groups (5 rats per group). All solutions to be administered to rats were sterilized by passage through 0.2-µm Acrodisk filters (Pall Corp., Ann Arbor, MI). Diabetes was induced by intraperitoneal [IP] administration of 50 mg/kg STZ, whereas nondiabetic rats were administered solvent (10 mM sodium citrate in 0.9% NaCl, pH 4.5). The treatment groups were as follows: ND, non-diabetic; STZ-veh, diabetic with subcutaneous osmotic minipumps releasing vehicle (1 mM acetate, pH 6) for 2 weeks; or STZ-des, diabetic rats with pumps releasing des(1-3)IGF-1 (5 µg/rat/day) for 2 weeks. Two weeks later, the animals were euthanized, and the eyes were placed in 4% paraformaldehyde in phosphatebuffered saline (PBS). The fixed eyes were embedded in paraffin and cut into 4-µm-thin sections. Tail blood was withdrawn for glucose assays 1 day after STZ or vehicle treatment as well as at 2 weeks.

Immunohistochemistry
Paraffin-embedded retinal tissue sections were rehydrated through xylene and a series of graded alcohol concentrations. Tissue sections were incubated in 0.25% Triton X-100 for 5 minutes. After a rinse in PBS, sections were incubated for 1 hour with primary antibody. After rinsing 3 × 5 minutes in PBS, sections were incubated with a 1:1500 dilution of biotinylated goat anti-rabbit or anti-mouse immunoglobulin (Vector Laboratories, Burlingame, CA) for 60 minutes. Tissue sections were incubated for 20 minutes with horseradish peroxidase-conjugated avidin (Elite kit, Vector Laboratories). The sections were rinsed in 0.05 M Tris and antigens were detected with a diaminobenzidine (DAB) kit (Pierce); the brown/black reaction product was visualized by light microscopy. Negative controls consisted of incubation with 5% goat serum without primary antibody, and did not generate any detectable reaction product. After staining, immunoreactive cells in the ganglion cell layer (GCL) and the inner nuclear layer (INL) were counted in 3 random 500-µm-long IGF-1 ANALOG MODULATES RETINAL TYPE 1 IGF RECEPTOR AND PHOSPHO-AKT 47 segments of the 4-µm-thick retinal cross-sections taken from each eye of the rat. There were approximately 100 cells along the length of each 500-µm segment. Sample labels were not visible to observers at the time of cell counting.

Statistical Analysis
Results are expressed as means ± SD for numbers of immunoreactive cells per 500-µm segment for each treatment group. Cell counts were analyzed with Fisher's post hoc least significant differences test. Differences between group means were accepted as significant at P < .05.

RESULTS
A 2-week duration of STZ diabetes was selected for this study to examine early pathological changes that may precede degenerative events. An immunohistological approach was chosen in order to identify the specific retinal cell layers affected.

Des(1-3)IGF-1 Treatment Did Not Prevent Hyperglycemia Nor Weight Loss in Diabetic Rats
Excessively high levels of IGF-1 or IGF analogues may cross-occupy the insulin receptor and ameliorate weight loss and hyperglycemia. The low dose of des(1-3)IGF-1 used in this experiment was not expected to alter these parameters; nevertheless, measurements were taken and the results are shown in Figure 1.
As seen in Figure 1 (top panel), significant differences in weight loss were not observed in des(1-3)IGF-1 versus vehicle-treated diabetic groups. Nondiabetic rats gained approximately 51 g, whereas diabetic rats weighed significantly less. No difference in weight was observed between STZ-veh and STZ-des groups. At termination of the experiment, serum glucose concentrations were measured as well (Figure 1, bottom panel). The diabetic rats were clearly hyperglycemic. Des(1-3)IGF-1 treatment did not reduce hyperglycemia in diabetic rats.

Effect of Des(1-3)IGF-1 Treatment on Type 1 IGF Receptor Immunoreactivity
There was a low level of IGF-1 receptor immunoreactivity in the GCL, INL, and BRB in the nondiabetic retina ( Figure  2A). Immunoreactivity in all of these areas appeared to be increased in retinas from STZ-veh rats ( Figure 2B). On the other hand, such immunoreactivity appeared to be reduced in STZ-des versus STZ-veh retinas, and was similar to that of the nondiabetic group ( Figure 2C).
To determine whether these differences were significant, type 1 IGF receptor-immunoreactive cells were counted in the GCL and INL in retinas from all rats. Type 1 IGF receptor immunoreactivity was significantly increased (P < .0001) in both the GCL ( Figure 3A) and INL ( Figure 3B) in STZ-veh versus the nondiabetic group. With des(1-3)IGF-1 treatment, type 1 IGF receptor immunoreactivity returned nearly to control levels. Immunoreactivity was significantly reduced (P < .0001) in STZ-des versus STZ-veh groups ( Figure 3A, B).

Preliminary Studies on Effect of Des(1-3)IGF-1 Treatment on VEGF Immunoreactivity
In anticipation of future studies, an initial examination of VEGF immunoreactivity was performed to determine whether des(1-3)IGF-1 treatment might prevent an increase in VEGF immunoreactivity. Adjacent sections of retinal tissue from the foregoing experiments were examined. The nondiabetic control group showed a basal level of VEGF immunoreactivity that was mainly associated with retinal endothelial cells ( Figure 4A). A qualitative change was observed in STZ-veh rats, and VEGF immunoreactivity appeared in retinal pigmented epithelial cells (RPEs) ( Figure 4B). This increase in RPE-associated VEGF immunoreactivity was prevented by treatment of diabetic rats with des(1-3)IGF-1 ( Figure 4C). Occasional cells of the inner retina, however, stained positively for VEGF in STZ-des as well as ND control tissues.

Effect of Des(1-3)IGF-1 Treatment on Phospho-Akt (Thr 308) Immunoreactivity
There was a basal level of the apoptotic-stress response protein phospho-Akt (Thr 308) immunoreactivity in the GCL and INL in the nondiabetic retina ( Figure 5A). Immunoreactivity in both of these areas appeared to be increased in the retina from STZ-veh rats ( Figure 5B). On the other hand, such immunoreactivity appeared to be reduced in STZ-des versus STZ-veh retinas, and was similar to the ND group ( Figure 5C).

DISCUSSION
Type 1 IGF receptor and phospho-Akt (Thr 308) immunoreactivity were increased in the GCL and INL of the rat retina 2 weeks after induction of diabetes with STZ. These changes, seen at 2 weeks, are among the earliest biochemical abnormalities that have been detected in the eye in diabetes, which coincide with VEGF up-regulation and BRB breakdown [3,4]. These predegenerative biochemical abnormalities were prevented by the subcutaneous administration of des(1-3)IGF-1 at time of onset of diabetes. Interestingly, treatment with des(1-3)IGF-1 was effective independently of poor metabolic control. These data suggest that treatment with IGF-1 or its analogues may be protective if administered early in the course of diabetes. There is evidence for synergistic effects between IGF-1 and VEGF on retinal endothelial cell proliferation/survival [21], and concern remains regarding the potential neovascularizing effects of IGF-1 in the retina. The present acute study brings new data suggesting that the role of IGF may be more complex than previouslyappreciated.

The Pathophysiology of Diabetic Neurological Disturbances is Mimicked by a Reduction of IGF Activity in Nondiabetic Conditions and IGF-1 Administration May Protect Against Deleterious Effects of IGF Depletion in Diabetes
Reduced axonal diameters, diminished conduction velocity, impaired nerve regeneration, and neuronal death are major pathological features of clinical diabetic neurological disturbances. A reduction of IGF activity in nondiabetic animals can mimic these effects. For example, anti-IGF antibodies impair nerve regeneration [22,23], and administration of anti-IGF antibodies or IGFBPs can cause neuronal death [24]. IGF-1null mice have reduced axon diameters and nerve conduction velocity [25] as well as neuron loss [26]. IGF-1 and IGF-2 mRNA levels are reduced in various tissues, including nerves, brain, and spinal cord in diabetic rodents [16,27,28]. IGF-1 gene expression is reduced in retina from diabetic patients and rodents [29]. IGF-1 gene expression is reduced in liver, the primary source of circulating IGF-1 [30,31], and circulating IGF-1 levels are reduced in diabetic rats [32,33], as well as in type 1 and type 2 diabetic patients [34,35]. Thus, there is a profound loss of IGF-1 support for various tissues early in diabetes. In the present study, immediate des(1-3)IGF-I treatment protected against early predegenerative changes in retina.

Des(1-3)IGF-1 Treatment is Effective Despite Poor Metabolic Control
The earliest detection of retinal neural degeneration in STZ diabetic rats is 4 weeks [13]. Consequently we examined for predegenerative changes at 2 weeks after the induction of diabetes. Des(1-3)IGF-1 protected against predegenerative retinal abnormalities independently of poor metabolic control evidenced by continued hyperglycemia and failure of weight gain. This suggests that the early predegenerative biochemical changes that were observed were possibly not a consequence of acute hyperglycemia per se. Alternatively, these µm.

FIGURE 6
Phospho-Akt (Thr 308) immunoreactivity was increased in STZ rat retina, and such increase was prevented by des(1-3)IGF-1 treatment. The mean numbers of type 1 IGF receptor immunoreactive cells per 500-µm-length retinal sections were calculated: (A) Ganglion cell layer (GCL) and (B) inner nuclear layer (INL). The mean immunoreactive cell count was significantly increased in STZ-veh versus nondiabetic control retina (P < .0001). This count was significantly reduced in STZ-des versus STZ-veh retina (P < .0001). Error bars indicate ± SD.
abnormalities were a consequence of the loss of IGF-1 activity in diabetes. This is consistent with the observation that low doses of IGFs prevent diabetic neuropathy in type 1 [9,15] diabetic rats despite hyperglycemia and weight loss, and in type 2 [16] diabetic rats despite hyperglycemia and weight gain. The present studies show that des(1-3)IGF-1 treatment was effective independently of continued weight loss. It might be considered that large pharmacologic doses of IGF-1 can cross-occupy the insulin receptor and reduce hyperglycemia. This occurs at doses that exceed by severalfold the 31-µg/rat/day IGF-1 production in liver. By contrast, the des(1-3)IGF-1 dose used in this study (5 µg/rat/day), and the 4.8 µg/rat/day IGF-1 used elsewhere [9,15], were too low to ameliorate hyperglycemia in diabetic rats.

IGFBP May Not be Essential for Protection
Des(1-3)IGF-1 is a naturally occurring truncated form of IGF-1 that is missing the N-terminal tripeptide important for binding to IGFBP [36,37]. Consequently, it is more potent than IGF-1 in vitro [38] and in vivo [33] due to reduced sequestration to IGFBP. It binds with 25-fold lower affinity to IGFBP-3 [39], has markedly reduced affinity for IGFBP-1, and 40-fold lower affinity to IGFBP-4 and -5 but retains similar affinity for the type 1 IGF receptor [39][40][41][42]. The data in the present study suggest that IGF-1 binding to IGFBP is not essential for protection against early predegenerative changes in type 1 IGF receptor and phospho-Akt (Thr 308) levels in retina.

IGF-1 Presence in Diabetic Retina
Elevated IGF-1 levels in the retina do not seem to originate from the retina itself, because IGF-1 mRNA levels are actually reduced in retinas from patients with 7-year duration diabetes as well as rats with 3-to 7-week duration STZ diabetes [29,43]. The predominant source of the elevated vitreous IGF-1 levels is a breakdown of the BRB because various serum proteins are increased in the vitreous together with IGF-1, although at least some of the IGF-1 may be of intraocular origin [10,44,45].
Circulating IGF-I levels are reduced 50% in type 1 and type 2 diabetic patients [34]. Despite this decrease, serum IGF-1 levels remain at least 20-to 50-fold higher than vitreous IGF-1 levels; hence, the increased permeability of retinal capillaries in diabetes may contribute to the elevated total IGF-1 levels in the eye. Therefore, vitreous IGF-1 levels may initially be reduced in early stages of diabetes as a consequence of reduced retinal IGF-1 mRNA levels in patients and rats. With chronic diabetes, increased permeability may result in elevated vitreous IGF-1 levels. Factors that influence the rate at which the BRB breaks down may explain at least in part the variability in vitreous IGF-1 levels reported in various clinical studies [7][8][9][10].

Type 1 IGF Receptor in Diabetes
The IGF-1 receptor (IGF-1R) appears to be under complex regulation in diabetes. In diabetes, IGF-1R mRNA levels are reduced in rat superior cervical ganglia [46], heart [47], and muscle [48], whereas IGF-1R protein levels are decreased in rat hippocampus [49]. Yet, retina [50] and endothelial cells 54 A. KUMMER ET AL. cultured from human retina [50] have elevated levels of IGF-1R immunoreactivity. Consistent with this observation, the present results show that IGF-1R immunoreactivity was significantly increased in vivo in retina from STZ-veh versus nondiabetic rats (Figures 4, 5), and the diabetic rat may provide a model for studying this biochemical abnormality. Immediate des(1-3)IGF-1 administration prevented this increase in IGF-1R immunoreactivity in diabetic rats (Figures 4, 5); but it is unclear whether this effect is at the transcriptional or translational level. Increased receptor immunoreactivity in the early stages of experimental diabetes does not appear to result from hyperglycemia because it is prevented by des(1-3)IGF-1 irrespective of ongoing hyperglycemia.
These results are discordant with the observation that IGF-1R immunoreactivity is not increased in retina from 8-week diabetic rats [43]. This difference is possibly due to 8-versus 2-week duration of STZ diabetes. Permeability of the BRB is increased after 3-week STZ diabetes; perhaps the associated increase in vitreous IGF-1 levels [45] may lead to downregulation of IGF-1R immunoreactivity in chronic disease.

Putative Effect of Des(1-3)IGF-1 Treatment on VEGF
VEGF is among the leading candidates as the primary mediator of proliferative retinopathy. It can induce vascular endothelial cell proliferation, migration, and vasopermeability. Inhibitors of phosphorylation mediated by the VEGF receptor can completely block retinal neovascularization [51].
VEGF may accumulate in the retina from retinal and vascular sources. VEGF accumulates in the vitreous humor of patients with proliferative diabetic retinopathy [52]. Patients with proliferative diabetic retinopathy have increased VEGF mRNA content in the GCL, INL, and outer nuclear layer, and this seems to be associated with ischemic regions of retina [53]. VEGF immunoreactivity may occur early, prior to evidence of retinal ischemia [54]. An increase in VEGF mRNA is also observed in the GCL and INL in the retina of STZ diabetic rats [32,55]. The increase in immunoreactive VEGF labeling is associated with increasing breakdown of the BRB, and is most prominent in the nerve fiber layer near the optic disk and in perivascular areas in diabetic rats [56]. These are the sites of BRB breakdown and neovascularization observed clinically. Early up-regulation of VEGF in diabetic retina is also associated with antioxidative defense mechanisms [57] and the formation of advanced glycation end products [58]. Hypoxia/ischemia, characteristic of diabetic retinal tissues, is a strong inducer of VEGF and may contribute to the activation of oxidative stress mechanisms in the diabetic retina (for review, see [59]). Our own previ-ous studies have shown that neuroretinal cell death under hypoxic conditions can be inhibited by IGF-1 and its analogues in vitro [60].
Our VEGF immunostaining ( Figure 4) showed a clustered pattern, which unfortunately did not lend itself to quantitative counts of random retinal fields. Yet, treatment with des(1-3)IGF-1 appeared consistently to reduce RPEassociated VEGF immunoreactivity. Consequently, these morphological data should be viewed with caution. Our preliminary results seem to indicate that the increase in VEGF immunoreactivity in the perivascular regions of retinas of diabetic rats is reduced by the administration of des(1-3)IGF-1 ( Figure 4). This is potentially due to reduced VEGF permeability, or other causes. Further studies are underway.

Phospho-Akt (Thr 308) and the Diabetic Stress Response
The serine/threonine protein kinase Akt (also known as PKB and Rac) plays a critical role in regulating the balance between survival and apoptosis in a variety of systems [61,62]. In the context of the present study, it is also noteworthy that Akt is proposed to be an important downstream target of phosphoinositol (PI) 3-kinase in insulin-mediated processes. There are high levels of PKB-β expression in insulin-responsive adipose tissue [63], whereas PKB-β-deficient mice exhibit manifestations of type II diabetes, including hyperglycemia and insulin resistance [64]. Mechanical stretch of retinal pericytes, proposed to exacerbate diabetic retinopathy, is also associated with increases in expression of both VEGF and activated phospho-Akt [65].
Akt phosphorylation was somewhat enhanced (123%) in STZ diabetic and galactosemic rats versus control, as measured by Western immunoblot of whole retinal cell lysates [43]. In our study, the differences are much more striking between control and diabetic retinal tissues, due to the specificity of our analysis of GCL and INL regions of the retina. Our results support the Gerhardinger suggestion that increased activation/phosphorylation of Akt may reflect a stress response in the retina, possibly through a p38/HOG1 kinase cascade of events [32,50,66]. Des(1-3)IGF-1 can cross the blood-central nervous system barrier [67], and might also cross the BRB. Consequently, one attractive interpretation of these data is that des(1-3)IGF-1 may have entered the eye and affected the phosphorylation and activation state of Akt. In our own previous studies [68], we have shown that the ability of insulin to rescue retinal cell cultures from cell death is mediated through the PI 3-kinase/Akt pathway, by the inhibition of caspase-3 activation. Therefore, the present observation that increased phospho-Akt (Thr 308) immunoreactivity is an early event in the course of STZ-induced diabetes appears to IGF-1 ANALOG MODULATES RETINAL TYPE 1 IGF RECEPTOR AND PHOSPHO-AKT 55 be important to our understanding of cell signaling and cell death mechanisms in diabetes-associated retinal degeneration. In fact, preliminary data from a separate, longer-term experiment show that apoptotic cell death is elevated and IGF-1 administration prevents such elevation in retina from diabetic rats (Seigel et al., unpublished data). This implies strongly that preventing these early predegenerative changes in diabetic retina may prevent the loss of retinal cells. The identification of predegenerative diabetic changes offer potential targets for future interventions, including therapy possibly with IGF-1 and its analogs.