Rod Photoreceptor Dysfunction in Diabetes: Activation, Deactivation, and Dark Adaptation

Abstract

purpose. To examine photoreceptor function in diabetes in detail by evaluating photoreceptor light activation, deactivation of the photoresponse, and recovery of the photoreceptor after bleaching (dark adaptation) in rats made diabetic with streptozotocin (STZ).

methods. Animals were assigned to treated and control groups. Light activation in rod photoreceptors was established using a paired-flash electroretinogram (ERG) protocol, and the leading edge of the a-wave was modeled with the mechanisms mediating phototransduction. Deactivation of the photoreceptor response was evaluated at three luminous exposures (1.4–2.2 log cd · m/s^–2) using a variable interstimulus interval (ISI) paradigm. Dark adaptation was evaluated at 90-second intervals for 30 minutes after approximately 20% pigment bleach. At each time point, a paired-flash signal (1.4 log cd · s/m^–2) was used to extract rod responses.

results. Diabetic animals showed decreased amplitudes of the photoreceptor response 12 weeks after diabetes induction. No difference was found in the rate of deactivation of the photoresponse in diabetic rats. Normalized amplitudes showed that diabetic animals had significantly faster dark adaptation (P < 0.01) than did controls.

conclusions. Although photoreceptor activation was abnormal, deactivation was unaltered after 12 weeks of diabetes. The faster relative recovery found in diabetes after bleach, in the presence of normal pigment dynamics, may reflect a decrease in outer segment lengths.

Diabetes is a leading cause of blindness in the western world. Its pathophysiology involves pericyte loss, ¹ vascular leakage, and angiogenesis in the retinal vasculature, changes to inner and outer retinal neurons, ² and alterations to glial cell structure and function. ^{3 4} This has led to the observation, by our group ^{5 6} and others, ^{2 3 7} that early neuronal deficits occur in diabetes. In particular, the photoreceptors—the most metabolically demanding neuron in the central nervous system ⁸ —show dysfunction as early as 2 days after diabetogenesis. ⁵ This is in addition to the well-described losses of the ERG oscillatory potentials in human and animal models. ^{9 10 11 12}  

Photoreceptor function can be evaluated with the ERG, from which activation of the photoresponse can be modeled from the a-wave. ^{13 14 15} Light activation produces a cascade that depresses the dark current of the photoreceptor, so-called because it is maximal in the dark, by closing ionic channels in the rod outer segments. Deactivation of the photoresponse begins with the phosphorylation of activated rhodopsin by rhodopsin kinase (RK), followed by its binding by arrestin. ^{16 17} This can also be monitored with the ERG by using a paired flash with varying interstimulus intervals (ISIs). ^{18 19} In the presence of flashes that give little bleach, this paradigm gives rise to an exponential recovery after an initial delay. ¹⁸ Finally, the dark adaptation or recovery of rhodopsin after pigment bleach can be assessed by measurement of growth in the dark current as a function of time. ²⁰  

Although the literature shows that photoreceptor activation is abnormal in diabetes, the nature of the loss is uncertain. Abnormal response amplitudes have been noted in diabetic rats in the presence of normal sensitivity. ^{5 6} However, sensitivity has been found to be abnormal in humans with diabetes. ²¹ Reduced levels of the α-subunit of G-protein in diabetic rats ^{22 23} implies an abnormality in sensitivity, but this has not been found in rats. Possible mechanisms for amplitude loss include reduction in the number and length of rod photoreceptors ²⁴ and reduction in the density of the gated channels supporting the dark current. Apart from these outer segment factors, diabetes-induced changes in the energy-dependent Na⁺/K⁺-ATPase ^{25 26} can also affect the a-wave. Na⁺/K⁺-ATPase has its highest density in photoreceptor inner segments ²⁷ and is responsible for sustaining the dark current. Hence, the ERG affords insight into energy- and adenosine triphosphate (ATP)–derived processes in normal outer segments because a-wave activation can be specified in terms of amplitude (ATP dependent) and sensitivity (ATP independent) of the phototransduction cascade. Moreover, the prospect for sensitivity changes must be revisited with large sample sizes to enhance experimental power given the recent reports for altered transducin activity and the reduced optical density of photopigment ²⁸ in diabetic rodents, each of which should selectively alter sensitivity. 

However, a full understanding of the diabetic status requires examination of other photoreceptor functions, such as deactivation of the photoresponse and dark adaptation (involving the photoreceptor/retinal pigment epithelium [RPE] complex), to define the nature of the loss. This is especially relevant given that electrooculogram amplitudes, thought to reflect ionic fluxes across the RPE, have been shown to fluctuate with elevation of blood glucose in healthy human subjects. ²⁹ Our approach has been to consider not only the activation phase of the light response but the deactivation of the photoresponse and dark adaptation. 

Activation of rhodopsin by light yields its activated intermediary, metarhodopsin II, whose inactivation begins with the phosphorylation of the activated state by RK, ^{16 30} a process that also requires ATP. This is followed by the binding of arrestin to the phosphorylated molecule. ¹⁷ Recent studies have found RK is increased in diabetic rats. ²² The significance of this finding is difficult to interpret, but the increased RK might produce faster recovery because animals lacking RK show delays in recovery kinetics. ^{31 32} To our knowledge, deactivation of the photoresponse has not been studied in diabetes. However, given the ATP dependency and the reports for abnormal dark adaptation in humans, ^{33 34 35 36} deactivation seems a useful assay for understanding the diabetic effects on neural function. 

Abnormalities in dark adaptation and absolute threshold have also been reported in human subjects with diabetes. ^{33 34 35 36} The higher absolute threshold is consistent with in vitro studies that note reduced rhodopsin ²⁸ and vitamin A ³⁷ (all-trans retinal) levels in diabetic animals, with the latter also expected to produce a slowing in the recovery process. ³⁸ However, the nature of the time course for dark adaptation remains controversial from human studies. Amemiya et al. ³⁴ and Henson and North ³⁵ report slower rates of recovery and elevated absolute thresholds, whereas Arden et al. ³³ found normal or faster recovery times, especially in the early phase of one patient with diabetes. As noted previously, this is consistent with the reported increase in RK ²² given that the early phase of dark adaptation is dependent on the deactivation process. ³⁴ We suggest that measuring both deactivation and dark adaptation provide independent evaluation of the abnormality of these processes. 

In this study, the nature of the photoreceptor response was examined in diabetes in a large sample of rats. We considered the activation, deactivation, and recovery of visual sensitivity after bleach (dark adaptation) in animals made diabetic with streptozotocin (STZ). Because each of these functions relies on different anatomic and biochemical mechanisms, the nature of the change of these functions in diabetes provides us with more information regarding the underlying retinal disorder. 

Materials and Methods

Animals

Eighty-one Sprague-Dawley male rats (aged 20 weeks) were tested in two cohorts to reduce test duration and time under anesthesia. Twenty-three control and 22 diabetic animals were used to measure activation of the photoresponse. Fifteen of each of these groups (total, 30) were randomly allocated to the deactivation experiment. The second cohort comprised a different group of 17 control and 19 diabetic animals for the measurement of dark adaptation. Animals were randomly assigned to treatment or control, with the treated group receiving STZ injection into the tail vein (50 mg/kg, in trisodium citrate buffer [pH 4.5]) at 8 weeks of age. Control animals received a sham injection of an equivalent volume (1 mL/kg) of 0.01 M trisodium citrate buffer (pH 4.5). Diabetes was confirmed by weekly assays of blood glucose (>15 mmol/L) and HbA_1C (>7.0%) measurement 12 weeks after diabetogenesis. All diabetic animals received 2 U/d insulin (human protophane; Novartis Pharmaceuticals Australia Pty. Ltd., North Ryde, NSW, Australia). All experiments adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by our institutional ethics committee. 

Electroretinography

Retinal function was measured in each animal 12 weeks after STZ treatment because diabetic changes have been found at this age. ^{5 6} Data collection occurred after the administration of general anesthesia with a mixture of ketamine and xylazine (60:5 mg/kg; Therapon Pty. Ltd., Burwood, VIC, Australia) and corneal anesthesia with proxymetacaine (Ophthetic 0.5%; Allergan, Frenchs Forest, NSW, Australia), under dim red illumination (λ_max = 650 nm) and after overnight dark adaptation (>12 hours), to maximize retinal sensitivity. ³⁹ Electroretinograms were recorded using a mecablitz flash unit (60CT4; CCT 5500°K; Metz, Zirndorf, Germany) presented by way of a Ganzfeld sphere. Light sources were attenuated with calibrated neutral-density filters (Kodak Wratten; Eastman Kodak, Rochester, NY) or by varying flash exposure. The unattenuated flash (¼ aperture) was measured as having a duration of 3.81 ms (8% amplitude) and a luminous energy of 3.41 log cd · s/m^–2 (IL1700; International Illumination) or 3.84 log scot cd · s/m^–2 and was calculated to produce 4.16 × 10⁶ photoisomerizations · rod^–1 in the rat eye. ⁴⁰ We have chosen to report photopic luminous energy units throughout this article. ERG waveforms were recorded after pupillary dilation with tropicamide (Mydriacyl 0.5%; Allergan) using custom-made silver–silver chloride electrodes referenced to a stainless steel ground inserted in the tail. Responses were amplified (gain ×5000; −3 dB at 1 and 1000 Hz; Powerlab Dual Bioamplifier, ADInstruments, Castle Hill, NSW, Australia) and digitized at 10 kHz over a 250-ms epoch. 

Activation of the Photoreceptor Response

Rod activation was established in control (n = 23) and STZ (n = 22) animals over an ensemble of luminous energies (1.4, 1.7, and 1.9 log cd · s/m^–2) with a twin-flash paradigm, ⁴¹ as described previously. ⁵ These energies were chosen to saturate the rod response with minimal pigment bleach (<0.5%) and to yield cone-isolated waveforms with a probe flash at an ISI of 0.8 seconds. Rod responses were obtained by digital subtraction of the cone response from the mixed rod/cone waveform, and the leading edge of the saturating rod a-wave (rod PIII) was modeled on the biochemical mechanisms mediating phototransduction, as described by equation 1 , ¹⁴  

\[\mathrm{PII}I(i,\ t)\ {=}\ {[}1\ {-}\ \mathrm{exp}{\{}{-}i\ {\cdot}\ S\ {\cdot}\ (t{-}t_{\mathrm{d}})^{2}{\}}{]}\ {\cdot}\ R_{\mathrm{max}},\]

where the amplitude of the PIII (μV) represents the sum of individual rod responses throughout the eye and can be described as a function of flash energy (i, log cd · s/m^–2) and time (t, seconds) after the onset of flash. The sensitivity parameter (S, cd^–1 · s^–3/m²) scale, i, R _max (Rm _PIII) is the saturated amplitude of the PIII response (μV), and t _d is a fixed delay (seconds) that accounts for delays in the phototransduction cascade and the recording system. ¹⁴ Parameter optimization for R _max and S was achieved with the Solver module of a spreadsheet (Excel; Microsoft, Redmond, WA) that minimized the sum-of-squares (SS) error term over the ensemble of light intensities. For this fitting, the delay (t _d), was fixed to the average value found in control animals (3.62 ms). 

Deactivation of the Photoresponse

Deactivation of the photoresponse was measured at three luminous exposures (2.2, 1.9, 1.4 log cd · s/m^–2) with 15 animals in each group. After manipulation of electrodes, animals were dark adapted for another 10 minutes before signal collection. A baseline paired-flash signal was collected at the luminous energy to be used for that paradigm (ISI, 0.8 seconds), after which a binary sequence of ISIs (0.5–128 seconds) was used to measure deactivation similar to the method described by Birch et al. ¹⁹ We calculated that the amount of pigment bleach produced by our brightest flash was less than 0.5%. 

The delay and recovery time course of the photoresponse can be modeled at a given intensity by the delayed exponential specified by Pepperberg et al. ¹⁸ (equation 2) .  

\[\mathit{A/A}_{\mathrm{max}}\ {=}\ 1\ {-}\ \mathrm{exp}{[}{-}(t\ {-}\ T_{c})/{\tau}_{r}{]}.\]

For this purpose, the amplitude of the photoresponse was taken at the criterion time of 6.5 ms (A _T6.5), which returned approximately 80% of the maximum a-wave response. This time point was chosen to give an estimate of the photoreceptor response before postreceptor intrusion. ^{20 42} Fitting was achieved by floating all variables and minimizing the SS error term in a spreadsheet (Excel) with the Solver module. 

Dark-Adaptation Protocol

The recovery of rhodopsin, or dark adaptation, was measured in 17 control and 19 STZ-treated animals by observing growth in the dark current after bleach. After electrode placement (in dim red light), animals were dark adapted for another 10 minutes before a baseline paired-flash signal was collected using 1.4 log cd · s/m^–2 (ISI, 0.8 seconds). Bleaching was achieved by six consecutive flashes (1-second ISI) at a luminous exposure of 3.1 log cd · s/m^–2 delivered through the Ganzfeld sphere. This procedure was calculated to produce 18% to 20% pigment bleach. 

After bleach, paired-flash signals (ISI, 0.8 seconds) were collected at a luminous exposure of 1.4 log cd · s/m^–2 at 90-second intervals, from 1 minute to 29.5 minutes. Signals were passed through an amplifier, as detailed previously, and were stored on a computer for post hoc analysis. 

Rod responses were extracted using the paired-flash protocol, as detailed previously, ^{5 41} and a-wave amplitudes were taken at a 6.5-ms criterion (A _T6.5). A theoretical model for analyzing the time course of dark adaptation in terms of the underlying mechanisms was used, as suggested by Thomas and Lamb ²⁰ and as adapted by Kennedy et al. ¹⁶ This model describes the recovery of the photoreceptor response after bleach in terms of two phases (equation 3) ,  

\[A_{T6.5}{=}A_{T6.5}({\infty})/{[}1{+}c_{1}\ \mathrm{exp}({-}k_{1}t)\ {+}\ c_{2}\ \mathrm{exp}({-}k_{2}t){]},\]

where A _T6.5(∞) is the baseline amplitude, c ₁ and c ₂ are desensitizing constants for the first and second phases of recovery respectively, and k ₁ and k ₂ are the respective recovery time constants. Fitting was achieved by floating all variables and minimizing the SS error term in a spreadsheet (Excel) with the Solver module. 

Statistical Analysis

Data normality was tested using a Komologrov-Smirnov test (StatView, version 5.0.1; SAS Institute, Cary, NC), and homogeneity was evaluated with a variance ratio (σ² maximum/σ² minimum). Dark-adaptation parameters (c ₁, c ₂, k ₁, k ₂) were analyzed with one-way ANOVA, with group as the dependent variable. Deactivation of the photoresponse (T _c and τ _r ) and dark adaptation were analyzed with repeated-measures ANOVA with time nested within group the dependent variable. In these cases, a Geiser-Greenhouse corrected F-ratio was used to safeguard against type 1 errors of nested designs. ⁴³ In the presence of significant interaction effects, analysis of simple comparisons was performed with one-way ANOVA. In cases of non-Gaussian data, log transformation ⁴³ of the raw data yielded normality (e.g., sensitivity) in all cases, whereas for heterogenous data, a more stringent alpha level (0.025) was applied to protect against type 2 errors. ⁴³  

Based on data from pilot trials, we calculated that a sample size of 15 should detect a 50% (0.3 log unit) change in parameter, with a power of 0.8 and an alpha of 0.05. When we found nonsignificant outcomes, we reported the experimental power for detection of a 50% (0.3 log unit) change for our sample size given the assayed experimental variability (example, [0.80]). 

Results

Average blood glucose levels, weights, and HbA_1C values at 12 weeks after treatment for the diabetic and control animals are shown in Table 1 . All diabetic animals had blood glucose levels greater than 15 mmol/L and HbA_1C levels greater than 7.0%, and all displayed characteristic weight loss (36%) at 12 weeks. 

Photoreceptor Activation Is Decreased at 12 Weeks of Diabetes

The photoreceptor response to light was decreased in amplitude 12 weeks after the induction of diabetes, as reported previously for different groups of animals. ^{5 6} Representative waveforms of a control and a diabetic animal over several luminous energies are shown in Figure 1A , with the group data summarized in Figures 1B and 1C . Although the saturated amplitude was significantly smaller (−16%) with diabetes (control versus STZ: Rm _PIII −434.5 ± 19.6 μV versus −364.5 ± 22.2 μV; F _1,26 = 10.2; P < 0.025), the sensitivity of the phototransduction cascade (3.66 ± 0.04 versus 3.60 ± 0.04 log cd^–1 · s^–3/m²; F _1,26 = 1.18; P > 0.025) was not affected by diabetes given that our sample size was adequate to detect a 50% change in sensitivity [0.98]. 

Photoresponse Deactivation Remains Unaltered in Diabetes

Representative waveforms for control and diabetic animals showing the time course for deactivation in the photoresponse are plotted in Figure 2 . The reduced saturated amplitude of the diabetic animal means that we considered deactivation normalized to the maximum potential for each group. With amplitudes normalized to A _max, repeated-measures ANOVA found no significant interaction effect (F_1,3 = 0.092; P > 0.025) and no significant group effect (F _1,16 = 0.001; P > 0.025). This indicates common deactivation kinetics in control and diabetic animals, as evident in Figure 3 , where average group (± SEM) recovery of the normalized a-wave amplitude is shown for each luminous energy. The critical delay (T _c) was not significantly different in the diabetic animals at any luminous exposure than it was in controls (e.g., 1.4 log cd · s/m^–2 2.13 ± 0.23 versus 2.13 ± 0.42 second; RmANOVA F_1,3 = 0.092; P > 0.05 [0.81]). Similarly, the time constant for deactivation (τ _r ) was unaffected by diabetes (e.g., 1.4 log cd · s/m^–2 3.14 ± 0.24 versus 4.04 ± 0.54 second: RmANOVA F_1,3 = 0.019; P > 0.05 [0.68]). Because there were no significant differences between groups, the delayed exponential increase (equation 2)was optimized as an ensemble across control and diabetic data sets to show trends (Figure 3) . 

Dark Adaptation Is Faster in Diabetic Animals

Dark adaptation is shown at various time points after 20% pigment bleach by representative waveforms in Figure 4 . The decreased a-wave amplitude noted earlier was confirmed in a different group of diabetic animals (−16% versus −24%) before their bleaching (control versus STZ: −410.9 ± 12.4 versus −311.5 ± 27.3; F_1,34 = 11.88; P < 0.025). 

To examine recovery dynamics in the presence of these absolute amplitude differences, recovery was expressed as a normalized amplitude relative to each animal’s own baseline. Mean (± SEM) normalized rod responses for control and STZ animals at each time point after bleach are shown on a semi-log plot in Figure 5 . This figure indicates that diabetic animals experience faster recovery after bleach than do control animals, which was confirmed by the significant interaction effect of the repeated-measures ANOVA (F_1,19 = 3.1; P < 0.025) performed on these data. 

Interestingly, no significant difference was found between the parameters that describe the time course of the two phases of rod recovery (c ₁, c ₂, k ₁, k ₂), as can be seen in Figure 6 . The fact that the early phase is not affected supports the nonsignificant finding from the deactivation experiment because these two processes should be closely related. However, the lack of significant finding in the late phase is perplexing given our significant ANOVA outcome and will be addressed in the Discussion. The trend in the k ₂ value is in the right direction because it indicates faster kinetics in the diabetic group though it fails to reach statistical significance [0.36]. 

Discussion

We have found photoreceptor function to be altered in several ways by diabetes. The magnitude of the dark current generated by the photoreceptors, as measured by the leading edge of the a-wave of the ERG, was decreased 16% to 24% at 12 weeks after STZ injection. Deactivation of the photoreceptor response was unaffected by diabetes, whereas dark adaptation occurred more quickly in diabetic animals than in control animals. 

The observation of reduced saturated response in the presence of normal sensitivity (Fig. 1)supports our previous data ^{5 6} and is consistently made in STZ-treated diabetic rats, ⁶ particularly in animals treated with insulin, as in our study. Interestingly, in human patients with diabetes, changes in the sensitivity of the photoreceptor response have been noted in the presence of normal amplitudes. ²¹  

The amplitude loss noted in our diabetic animals could have occurred for a number of reasons, among them reduction in length or number of photoreceptors ²⁴ or change in channel density within the photoreceptor outer segments. Changes to the Na⁺/K⁺-ATPase activity in the inner segment or reduced efficiency in the potassium channels can also reduce the dark current and result in lower amplitude. ⁴⁴ A decrease in Na⁺/K⁺-ATPase activity has been reported in diabetic animals ^{25 26} and most likely contributes to the reduced a-wave amplitude found in our study (Fig. 1) . That sensitivity remains unaffected in the presence of the reported changes in humans and the reduced transducin levels in diabetic rats ^{22 23} is perplexing. However, we believe this is a reliable outcome given that the power of our experiment to detect a 50% change in sensitivity was 0.98 and given the replication of this finding in two other groups of animals. ^{5 6} In this study and in other studies in which sensitivity was normal, rats were treated with daily injections of insulin. Kowluru et al. ²³ show that daily insulin (2 IU/d) can reverse the G-protein deficiencies induced by diabetes. This may explain our findings of normal sensitivity in rats given insulin, implying that human patients with abnormal sensitivity have hyperglycemia or poor control of glycemia or experience sensitivity changes because of low insulin levels. The exact cause for this loss is unclear and warrants further investigation. 

Deactivation of the photoresponse was unchanged at 12 weeks of diabetes. Light capture causes conversion of rhodopsin to its active state (metarhodopsin II), which activates the G-protein cascade and gives rise to phototransduction (for a review, see Burns and Baylor ⁴⁵ ). Inactivation of metarhodopsin II requires the phosphorylation of activated rhodopsin by RK, ^{16 30} followed by the binding of arrestin to the phosphorylated molecule. ¹⁷ Given that we found no significant difference in the deactivation kinetics of control and diabetic animals, we conclude that the bioavailability of ATP is not altered by diabetes, in agreement with previous studies that report no difference in retinal ATP content in diabetic and galactosemic animals. ⁴⁶ These observations indicate that a reduction in the availability or production of ATP is not contributory to the losses in photoreceptor activation found in this study (Fig. 1)and reported previously. ^{5 6}  

Unexpectedly, the rate of dark adaptation in diabetic animals was faster than in controls. This finding is particularly interesting in that recovery dynamics, when modeled using the equations of Kennedy et al., ¹⁶ were unchanged. However, the lack of a positive finding in these parameters most likely results from our undersampling of the late phase of adaptation, which can take more than 3 hours for complete recovery in Sprague-Dawley rats. ⁴⁷ Studies in rats have shown that recoverin is reduced by diabetes ²² ; this would yield a delay in the onset of the late phase of dark adaptation, potentiating our undersampling. Not surprisingly, the experimental power for these parameters is low [0.36]. We think ANOVA provides a truer reflection of the differences in the data given the undersampling. The fact that the early phase of this response is not altered is not surprising because it most likely is mediated by mechanisms common to those measured by photoresponse deactivation and found to be normal. However, the second, or slow, phase of dark adaptation involves restoration of the photopigment molecule to the 11-cis isomer through a series of reactions known as the retinoid cycle. ⁴⁸ The retinoid cycle begins after the capping of the phosphorylated rhodopsin by arrestin and leads to all-trans retinoid release from the opsin molecule and its transfer to the RPE, ⁴⁹ where it is reconverted to its 11-cis isomer and is transported back to the photoreceptor (for a review, see Lamb and Pugh ⁴⁸ ). Faster kinetics of dark adaptation implies that the recycling of photopigment across the RPE, or the retinoid cycle, could be normal or faster. One plausible reason for faster recovery would be that there is less pigment in the diabetic retina; therefore, less would be bleached by the light source, making recovery relatively faster in the presence of a normal retinoid cycle. Reduced rhodospin content has been reported in diabetes ²⁸ and may explain the reduced amount of arrestin present, consequent to the lower levels of rhodopsin. This decrease in rhodopsin would result in reduced optical density in normal rods and would yield lower sensitivity, which is in contrast to our data. However, optical density could remain normal if it were associated with a proportional reduction in the length of the rod outer segments, another finding reported in early diabetes. ² These changes would provide an explanation for our findings of reduced saturated amplitude with no change in sensitivity in diabetic animals. They would also explain the faster adaptation after bleach. 

In conclusion, we found photoreceptor function to be altered at 12 weeks of diabetes, likely reflecting a decrease in the amount of rhodopsin present in the rod outer segments associated with a proportional decrease in outer segment lengths. These changes would explain our normal sensitivity, faster than expected dark adaptation, and smaller than expected saturated photoreceptoral response. Given past reports of dysfunction in Na⁺/K⁺-ATPase activity in diabetes, ^{25 26} it is also possible that this mechanism contributes to reduced photoreceptor amplitude, magnifying the loss. Moreover, our work shows that the latter deficit does not arise from a decrease in the availability of ATP because deactivation of the photoresponse, which also requires phosphorylation, is normal in diabetes. Contrasting our findings in rats with those of humans suggests that the sensitivity parameter might provide a useful index of diabetes control, but this needs further clarification. 

 Supported by Australian Research Council Grant ARC-LP0211474 and National Health and Medical Research Council Grant 208950.

 Submitted for publication November 22, 2005; revised February 14, 2006; accepted May 1, 2006.

 Disclosure: J.A. Phipps, None; P. Yee, None; E.L. Fletcher, None; A.J. Vingrys, None

 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.

 

Table 1.

 

Weight, Blood Glucose and HbA_1C Values for Diabetic and Control Animals at 12 Weeks

Treatment Group	Weight (g)	Blood Glucose (mmol/L)	HbA1C (%)
Control (n = 40)	402.6 ± 6.9	4.9 ± 0.1	3.18 ± 0.1
Diabetic (n = 41)	256.3 ± 7.3*	20.7 ± 0.7*	10.6 ± 0.5*

 All values are mean ± SEM.

*  Significantly different from control values (P < 0.025).

Figure 1.

 

(A) Representative control (filled circles, solid lines) and diabetic (unfilled circles, dashed lines) a-waves with the associated modeled PIII waveform showing that diabetic animals have reduced maximum amplitude (Rm _PIII) with normal sensitivity. (B) Group results for Rm _PIII magnitude. Squares represent group average data (± SEM; n = 23 control; 22 STZ). (C) Group results for the sensitivity parameter showing no significant difference between control and STZ animals.

Figure 2.

 

Representative waveforms for a control (filled circles) and a diabetic (unfilled circles) animal at the luminous exposure of 2.2 log cd · s/m^–2 for various ISIs (seconds; numerals to the right).

Figure 3.

 

Photoresponse deactivation plotted as the normalized amplitude (A/A _max) of a paired flash stimulus as a function of interstimulus interval for three luminous exposures: (A) 2.2 log cd · s/m^–2, (B) 1.9 log cd · s/m^–2, (C) 1.4 log cd · s/m^–2. Control (filled symbols; n = 15) and diabetic (unfilled symbols; n = 15) animals. Solid lines show the exponential rise fitted as an ensemble over control and diabetic groups. Error bars represent ± SEM.

Figure 4.

 

Dark adaptation of representative waveforms for a control (A) and a diabetic (B) animal measured at various times after bleach. Dashed line: Criterion time for a-wave amplitude estimation (6.5 ms). Numbers superimposed over the waveforms represent the times after bleach (minutes).

Figure 5.

 

Dark adaptation of rod a-waves normalized to the average baseline value for contol (filled circles, solid line; n = 17) and diabetic (open circles, dotted line; n = 19) animals. Lines represent the best-fitting, two-phase recovery process returned from the model of Kennedy et al. ¹⁶

Figure 6.

 

Mean (± SEM) recovery parameters for control (filled bars, n = 17) and diabetic (open bars, n = 19) animals for (A) the rate constant of recovery and (B) the degree of bleach-induced desensitization for the two components of recovery.