Diabetes-induced changes of the rat ERG in relation to hyperglycemia and acidosis

Abstract

Purpose:

To understand the mechanism of changes in the c-wave of the electroretinogram (ERG) in diabetic rats, and to explore how glucose manipulations affect the c-wave.

Methods:

Vitreal ERGs were recorded in control and diabetic Long-Evans rats, three to 60 weeks after IP vehicle or streptozotocin. A few experiments were performed on Brown Norway rats. Voltage responses to current pulses were used to measure the transepithelial resistance of the retinal pigment epithelium (RPE).

Results:

During development of diabetes the b-wave amplitude progressively decreased to about half of the initial amplitude after a year. In contrast, the c-wave was strongly affected from the very beginning (3 weeks) of diabetes. In control rats the c-wave was cornea-positive at lower illuminations but was cornea-negative at higher (photopic) illumination. In diabetics, the whole amplitude-intensity curve was shifted toward negativity. The magnitude of this shift was markedly affected by acute glucose manipulations in diabetics but not in controls. Increased blood glucose made the c-wave more negative, and decreased blood glucose with insulin had the opposite effect. Experimentally induced acidification of the retina had a small effect that was different from diabetes, shifting the c-wave toward positivity, slightly in controls and more noticeably in diabetics. One reason for the significant negativity of the diabetic ERG was a decrease of the cornea-positive response of the RPE due to a decrease of the transepithelial resistance.

Conclusions:

The ERG c-wave is more negative in diabetics than in control animals, and is far more sensitive to changes in blood glucose. The increased negativity is largely if not entirely due to changes in the transepithelial resistance of the RPE, an electrical analog of the breakdown of the blood-retinal barrier observed in other studies. The sensitivity of the c-wave to glucose in diabetics may also be due to changes in transepithelial resistance.

Introduction

Diabetic retinopathy (DR) is a leading cause of visual impairment.^1,2 While the most severe clinical effects in humans are the ones that affect the retinal vasculature (capillary loss, microaneurysms, leakage, neovascularization), it is clear that there are neural effects in humans and in animal models of diabetes that are likely to be independent of the vascular ones.^3–6 These neural effects range from multiple effects on photoreceptors,^5,7 including decreases in response amplitude and sensitivity.^8–10 to loss of ganglion cells,¹¹ to changes in the timing and amplitude of components of the ERG and EOG.^12–15 The earliest ERG changes are those in the oscillatory potentials^16–19 and in the c-wave of the ERG.^20,21 The oscillatory potentials can be recorded more easily in humans, and have received much more attention, but the c-wave changes are of interest in understanding the role of the retinal pigment epithelium (RPE) in the pathology of DR.

A decrease in c-wave amplitude, or even an inversion from positive to negative, was first observed by Pautler and Ennis²¹ in pigmented Long-Evans rats. They noted a decrease in c-wave amplitude at two weeks, with respect to non-diabetic controls, and it became negative at 4 weeks and even more negative at 19 weeks, whereas there was no effect on the b-wave until 19 weeks. MacGregor and Matschinsky²⁰ obtained similar results on Long-Evans rats but only studied them for 6 weeks. Samuels et al.¹³ found a decrease in the c-wave at two weeks in C57Bl/6J mice, but this was accompanied by decreases in the a- and b-waves.

The obvious starting point of the events leading to diabetic retinopathy is elevation of the glucose level in the blood. This is expected to be associated with the increase of glucose concentration in extracellular space of every tissue of the body, including the retina. One of the early consequences of the increased glucose level is accumulation of acidic wastes that lead to acidosis. An abundance of glucose could stimulate the cell’s energy metabolism to shift toward intensification of glycolysis, producing more acidic waste. Indeed, retinal acidosis develops at one stage of diabetes in rats.²²

In this paper we recorded the ERG of diabetic and control rats over a year, and investigated the consequences of acute manipulations of blood glucose and acidosis to better understand the damage initiated by diabetes in the ERG, particularly the c-wave.

Methods

Animal experiments were performed in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by Northwestern University’s Institutional Animal Care and Use Committee. The procedures to obtain electrophysiological recordings have been described in detail previously.^22,23 Briefly, rats were surgically prepared under isoflurane anesthesia. They were anesthetized during recordings with intravenous urethane supplemented by 0.5% inspired isoflurane, and artificially respirated after paralysis with pancuronium bromide. Animals were maintained for several hours and their body temperature, blood pressure, arterial PO₂, PCO₂, pH and glucose were monitored and controlled. Vitreal ERGs from adult Long Evans rats (32 controls and 45 diabetics) of different ages (up to 1 year), were recorded with microelectrodes positioned in the vitreous with a reference outside the eye under the skin on the back. Brown Norway rats (5 controls and 5 diabetics) were also tested for comparison. In these experiments the intraretinal ERGs (local ERGs) were collected routinely, but only a few of them will be shown in this work since the focus here is on the vitreal ERG as it reflects activity of the whole retina. Additionally, recordings with double-barreled H⁺-selective microelectrodes were conducted (for fabrication of ion-selective microelectrodes (see Reference ²⁴). The reference barrel of the H⁺-selective microelectrode was used for ERG recordings. Some ERGs were obtained during measurements of PO₂ in the retina, but oxygen data will be described separately. Animals were euthanized at the end of recordings without recovering consciousness.

Diabetes was induced with a single intraperitoneal injection of streptozotocin (Axxora LLC, Farmingdale, NY; 65 mg/kg rat) in 0.05 mol/l sodium citrate buffer (pH 5) in a volume of 0.01 ml/g rat. Age-matched controls received a single intraperitoneal injection of 0.05 mol/l sodium citrate buffer only (0.01 ml/g rat). Rats were weighed weekly and nonfasting blood glucose levels were measured from the tail vein using a Bayer CONTOUR Meter (Bayer HealthCare LLC, Mishawaka, IN)

During recordings intravenous glucose injections were used to increase blood glucose in both control and diabetic rats; in the case of diabetics, those animals with relatively low glucose (300 – 400 mg/dl) were chosen for the procedure. After such injections the blood glucose significantly increased, sometimes above the maximum reading of the meter (>600 mg/dl). Blood glucose increased rapidly, and intensity series were recorded an average of 71 min (26 to 313 min) after starting the glucose infusion. In other experiments the high blood glucose level in diabetics was reduced by intravenous insulin. For acute acidification of the retina 15 mg/kg of the carbonic anhydrase blocker dorzolamide was given intravenously in a single injection during experiments.²⁵ Blood pH rapidly became acidotic, and intensity series were recorded an average of 77 min (22 to 169 min) after the injection. Finally, to induce systemic metabolic acidosis, control rats were given 500mM NH₄Cl and 2% sucrose in their drinking water for several days prior to recordings, as described earlier.²⁶

To evaluate the electrical resistance across the retinal pigment epithelium, constant current electrical pulses of 0.5 second in duration were sent from a macroelectrode on the cornea to a reference outside the eye (on the scleral side of the eye in electrical terms). The resulting voltages were recorded by the microelectrode inside the eye, first on the basal side of the retinal pigment epithelium (RPE) and then, during electrode withdrawal, in the subretinal space outside the apical RPE. The ratio of the amplitude of the basal to apical voltages served as a measure of the transepithelial resistance across the RPE.²⁷

Individual flashes of diffuse white light were used to evoke ERGs. The light was attenuated by neutral density filters and light intensities are presented here as log relative illumination compared to the maximal illumination produced by our stimulator (750 lux; 0 log units relative illumination). For reference, rod saturation occurred at about −1.5 log relative illumination, so lower values are in the scotopic range, and higher illuminations are in the low photopic range. Even the strongest stimuli did not evoke a visible a-wave. The duration of the light flashes was 2.5 seconds, and they were given infrequently, so they did not disturb the prevailing state of dark adaptation even at the highest intensities. On the other hand, this duration was long enough to allow development of the c-wave of the ERG, which was the center of attention in this paper. All individual responses shown are from single flashes, not averages of multiple flash presentations. The averaged intensity series in the figures generally used one series per animal per condition, but in a few cases two series from an animal were included in the averages.

Results

I. The diabetes-induced changes of the rat ERG

The main features of the ERG of a normal rat are shown in Fig. 1, left column, where ERGs evoked by 2.5 sec light flashes of different illumination (intensity series) are presented. The b-wave amplitude increased with illumination, but the slower potential, the c-wave, did not. The c-wave of the rat ERG does not look like the classical exponentially rising cornea-positive potential, however, it is still mostly the result of two simultaneous events of opposite polarity.^28,29 In most vertebrates the cornea-positive transepithelial potential (TEP) generated across the retinal pigment epithelium is larger than the cornea-negative transretinal potential generated by glial Müller cells (slow PIII or sPIII). In rats these potentials have about the same amplitude and practically cancel each other. Additionally, they grow differently as light intensity increases. In the scotopic range the transepithelial potential of the RPE exceeds sPIII; as a result, the c-wave at the end of a 2.5 second stimulation has slightly cornea-positive polarity. But with an increase of illumination sPIII was larger than the TEP so the c-wave recorded in the vitreous had cornea-negative polarity.

Fig. 1:

Samples of ERG intensity series from a control rat and three diabetic animals at different times after STZ injection or vehicle. The gray bar shows the stimulus period of 2.5 sec. The voltage scaling is the same in all traces. The vertical distance between traces is 1 mV. The stimulus strength is shown as log relative illumination where zero is 750 lux.

Figure 1 also shows examples of ERG intensity-response series for three diabetics, and Figure 2 shows the b- and c-wave at −0.5 log relative illumination for all animals, where each point represents one animal. The b-wave recorded in the vitreal ERG progressively decreased during the development of diabetes. It was not different in diabetics and controls in the period from one to 10 weeks (t-test; p = 0.12), but then decreased so that after a year the b-wave amplitude diminished by half. In contrast, changes in the c-wave were already apparent 3 weeks after STZ injection and the c-wave was significantly different in diabetics and controls in the one to 10 week period (t-test; p = 0.011). ERGs of diabetics were distinguishable by a large negative c-wave under high scotopic and photopic illumination (Fig. 1 and Fig. 2) and the whole amplitude-intensity curve was shifted toward negativity (Fig. 3). Furthermore, the early change in the c-wave in diabetics did not progress significantly with the duration of diabetes at −0.5 log illumination. The slope of the regression line for the diabetic c-wave in Fig. 2 was not significantly different from zero, whereas the b-wave continued to change with time. We tested whether the relatively large variability of the b- and c-waves of the diabetic animals was due to variation of blood glucose or variation of animal weight, but neither contributed substantially to variability of either the b- or c-wave (r² < 0.04 for all relationships between b- or c-wave and glucose or weight).

Fig. 2:

Each point represents the amplitude of the ERG b- or c-wave of control (circles) or diabetic (triangles) rats from an intensity-response series at −0.5 log relative illumination vs. time after STZ or vehicle. (n = 32 control rats; 38 diabetics). The slopes of the regression lines were not different from zero for the b- or c-wave in control animals (p = 0.83 and 0.46), or for the c-wave in diabetic animals (p = 0.09). The slope for the b-wave of diabetics was significantly different from zero (p = 0.0004). We do not have a good explanation for the smaller b-wave in the three oldest control rats, but note that the c-wave in these animals was in the normal range.

Fig. 3:

Average (± SEM) amplitudes of the b-wave (part A) and c-wave (part B) in controls (solid line; filled symbols; n=40 intensity-response series) and diabetics (long dashed lines 0–20 weeks, n=25; short dashed lines 21–40 weeks, n=19; dotted lines 41–60 weeks, n=10). Naka-Rushton fits to the b-wave showed that the half-saturating illumination did not change (even in the presence of diabetic cataracts).

The saturation of the b-wave happened at the same illumination, near −1.0 log unit, in both control and diabetic animals (Fig. 3), and there was no change in the half-saturation value obtained from fits to the Naka-Rushton equation, despite the cataract that always accompanied development of diabetes. Therefore, retinal sensitivity was not affected up to one year of diabetes.

The ERGs of Brown Norway rats were recorded at 39.2 ± 3.0 (SD) weeks after STZ or vehicle. The b-wave in control Brown Norway rats was about half the amplitude of the b-wave in control Long Evans rats and the c-wave was also smaller. Both components in Brown Norways were affected by diabetes in a similar way as in Long Evans animals (Supplement - Fig. S1), with the b-wave becoming smaller and the c-wave becoming more negative.

II. Glucose affects the c-wave of diabetics ERG, but not of controls

Since the diabetes-induced negativity of the c-wave was already apparent just 3 weeks after STZ injection (the earliest time the recordings were made), it was possible that those c-wave changes were caused by hyperglycemia rather than changes due to diabetes. However, measurements on control rats showed that acutely produced hyperglycemia affected the ERG very little. If anything, there was an increase in both the b-wave and the c-wave when control animals had higher glucose (Fig. 4), but neither change was significant. However, acute manipulation of blood glucose level had a strong effect on the c-wave of diabetic animals, and this effect was in the same direction as that produced by diabetes. Intravenous glucose led to a significant negative shift of the c-wave amplitude-intensity curve (Fig. 5 top panels) (two factor ANOVA with pre- or post-glucose as one category and the three highest illuminations as another category; p = 0.015 for the effect of glucose). Lowering glucose in diabetics with insulin produced a significant c-wave shift in the positive direction (Fig. 5 lower panels) (ANOVA as for the effect of high glucose; p < 0.001).

Fig. 4:

Effects of acute blood glucose elevation on the intensity response behavior of control rats. A: Records from the same animal at the control glucose level of 164 mg/dl and after elevation to 484 mg/dl. B: Average effect on the intensity-response series, before (solid line and filled symbols; n=10) and after (dashed line and open symbols; n=10) glucose elevation. Average glucose at the time of recordings was 127 ± 43 mg/dl (mean ± SD) before elevation and 481 ± 78 mg/dl after. Amplitudes here and in other figures show mean ± SEM.

Fig. 5:

Top panels: Effects of acute blood glucose elevation on the intensity response behavior of diabetic rats. A: Records from the same animal at the initial glucose level of 381 mg/dl and after elevation to 578 mg/dl. B: Average effect on the intensity-response series. Solid lines and filled symbols are before glucose elevation (n=6), dashed line and open symbols are after glucose elevation (n = 6). Average glucose at the time of recordings was 377 ± 72 mg/dl before elevation and 558 ± 48 mg/dl after. Bottom panels: Effects of acute insulin-mediated decrease of blood glucose on the ERG of diabetic rats. C: Records from the same animal before (at 540 mg/dl glucose) and after (227 mg/dl glucose) insulin. D: Average effect on the intensity-response series. Solid lines and filled symbols - before insulin (n=4); dashed lines and filled symbols – after insulin (n = 4). Average glucose at the time of recordings was 543 ± 48 mg/dl before insulin and 240 ± 49 mg/dl after.

III. Acidosis has no apparent contribution to diabetes-induced changes of the c-wave.

As was previously shown,²² diabetes in the STZ rat is accompanied by acidosis. But it seems that acidosis does not play an important role in the c-wave changes related to diabetes. The negativity of the c-wave is prominent at the very first month of diabetes, when acidosis has not yet developed, and it persists at later stages (up to a year), when retinal acidosis decreases toward normal. In two cases when acute acidification of the retina of control rats was produced by intravitreal dorzolamide, there was a small increase in the c-wave (Fig. 6, top panels). Producing acute acidosis of the diabetic rat retina with dorzolamide led to suppression of the b-wave, and, as in controls, an increase of the c-wave with mid-range light intensity. The whole amplitude-intensity curve of the c-wave was shifted in the positive direction (Fig. 6, bottom panels). This was opposite to the effect of diabetes alone and was more pronounced than the effect of acidosis in control rats. These effects of acute acidosis in control and diabetic rats are in the same direction as those produced by acidosis (hypercapnia) in cat retina, that is, a decrease in the b-wave and an increase in the c-wave.^30,31 Artificial chronic retinal acidification produced in control rats by adding NH₄Cl to the drinking water also increased the c-wave, but this had little effect on the b-wave (Supplement - Fig. S2).

Fig. 6:

Top panels: Effects of dorzolamide-induced (acute) acidosis on ERG of control rats. A: Records from the same animal before and after injection of dorzolamide. Values under the intensity series are [H⁺] in the blood and maximum [H⁺] in the retina at the time of recording. B: Average effect on the intensity-response series. Solid lines - before dorzolamide (n=5); dotted lines - after dorzolamide (n=2). The average maximum [H⁺] in the retina was 74.8 ± 7.7 nM before acidosis and 128.0 ± 35.4 nM after acidosis. Bottom panels: Effects of dorzolamide-induced (acute) acidosis on the ERG of diabetic rats. C: Records from the same animal before and during acidosis. D: Intensity-response series. Solid lines – before dorzolamide (n=6), dotted lines - after dorzolamide (n=6). The average maximum [H⁺] in the retina was 77.2 ± 7.3 nM before acidosis and 198.8 ± 81.0 nM after acidosis.

IV. Intraretinal ERG, c-wave, responses of RPE and MC and electrical pulses.

The c-wave is mainly a result of the subtraction of the large cornea-negative slow PIII generated by Müller cells from the cornea-positive transepithelial potential from the RPE (Fig. 7), which is larger in most vertebrates, but not at all illuminations in rats. As a result, in normal rats the resulting c-wave is slightly positive or slightly negative (Fig. 3, Fig. 7) depending on stimulus intensity.

Fig 7:

Comparison of the vitreal ERG (marked as ERG), the transepithelial potential across the RPE and the transretinal potential across the neural retina (marked as Retina). Light flash of 2.5 s in duration had an illumination of −2.5 log units (part A) and −0.5 log units (part B). The vitreal ERG was recorded with a microelectrode tip in the vitreous. This is the same electrical configuration as in traditional electroretinography with an active electrode positioned on the cornea. The transepithelial potential was recorded with the electrode tip in the retina in the subretinal space close to the RPE. In both A and B the reference point was a reference electrode outside the eye. The transretinal potential was calculated by subtracting the transepithelial potential from the vitreal ERG.

One possible explanation of the negative shift of the c-wave in diabetic rats is a decrease of the transepithelial potential due to a decrease of the transepithelial resistance (eq. 15 in reference ²⁸). To test this hypothesis, the voltage amplitudes of constant current electric pulses sent across the eye were measured on both sides of the RPE in animals that had been diabetic for 5 to 11 weeks. Several measurements were made in each animal, and then averaged to give a single value for each animal that was used in statistical tests. The ratio of the amplitudes on the apical and basal sides of the RPE gives a measure of resistance, with higher values meaning higher transepithelial resistance. Diabetics had significantly reduced resistance of the RPE (Fig. 8) (p = 0.028; n=8 diabetic and n=8 control rats; t-test). There was not enough data for statistical tests of other conditions, but neither elevation of blood glucose in control animals nor acidosis appeared to affect the transepithelial resistance (ratio = 1.42 ± 0.07 (SEM) in high glucose, n= 4 rats; ratio = 1.28 in acidosis, n=1 rat). In addition, dorzolamide did not change the resistance in diabetic animals (without dorzolamide, 1.14 ± 0.06, n=8 animals; after dorzolamide, 1.17 ± 0.07; n=6 animals). Note that dorzolamide strongly influenced both the b- and c-waves (Fig. 6), so that the additional effect of dorzolamide on the diabetic ERG was probably due to a change in b- and c-wave generation rather than a resistance effect.

Fig 8.

Ratio of the amplitude of the electrical pulses measured in the subretinal space (on the apical side of the RPE) to those in the choroid (on the basal side of the RPE) in control animals (ctr), and in diabetics (db). The colored boxes show the 25–75% quartile range; the dot is the mean; the horizontal line is the median; the whiskers are at 1.5 times the interquartile range. The star is an outlier. Eight animals in each group.

Discussion

Changes in the ERG of diabetic rats have been known for decades, including the development of c-wave negativity at the very beginning of diabetes.^21,32,33 These groups studied Long-Evans rats, and we have extended this to Brown Norway rats as well. Knockdown of the insulin receptor produces diabetic “TetO” rats, which also have a reduced c-wave ³⁴. Similar changes in the c-wave occur in mice.^13,14,35 The fast oscillation, which originates in the RPE ^36,37 is also reduced in diabetic humans ¹⁵ and mice,^13,14,35 and the response of the RPE to intravenous bicarbonate is also reduced in humans.¹⁷ Perhaps surprisingly, the other slow light-evoked response of the RPE, the light peak, appears to be more resistant to diabetes.^13,14,17

The four major new things shown in this paper are: 1) after the initial decrease, the c-wave does not decrease further, 2) the c-wave diabetic rats becomes sensitive to glucose concentration, but it is not in control animals, 3) the effect of diabetes is not related to diabetes-induced acidosis or to acidification in general, and 4) changes in the c-wave are at least partly attributable to a decrease in transepithelial resistance. Some of these points require further discussion.

The lack of significant change in the c-wave over the course of a year (Fig. 3) is a new finding in type 1 diabetic animals, although it was also found in type 2 Lepr^db/db mice over 24 weeks.¹³ Previously the longest period of diabetes studied in type 1 animals was six months,³⁵ and because most studies have had only time point, it was not possible to tell whether there was a progressive reduction in the c-wave. Considering that the b-wave (Fig. 2, Fig. 3)³⁸ and photoreceptor electrical and structural changes progress in at least some rodents,⁷ it may be surprising that the c-wave remains stable after the initial decrease. It is possible that the potassium change underlying both the RPE c-wave and slow PIII continues to decline, making both components smaller. (Tarchick et al.¹⁴ reported that slow PIII was decreased in diabetic Nxy^nob mice (which lack a b-wave), but this was assessed at about 200 ms after stimulus onset, and at this time fast PIII would also be present.) If both components of the c-wave change, their difference, the vitreal c-wave, can be affected very little.

Many studies have indicated that the blood-retinal barrier at the RPE can be compromised by diabetes.³⁹ Most dye permeation studies^40–43 have shown leakage across the RPE in diabetics, although this is not universal.^43,44 Additional evidence of an RPE defect comes from the decrease in the ability to remove fluid from the subretinal space,⁴⁵ Thus, it is not surprising that the electrical resistance of the RPE would be reduced, but this has not previously been measured in diabetes, and it appears to change faster than transepithelial transport.

Transepithelial resistance is a composite of the paracellular resistance across the tight junctions and the transcellular resistance across the apical and basal membranes. For a given size of the c-wave voltage generated at the apical membrane, apical or basal membrane resistance would have to increase in order to decrease the transepithelial c-wave (eq. 15 in reference ²⁸), which seems unlikely. Thus, especially considering the evidence about transepithelial leakage, it is probable that the resistance of the tight junctions decreases in diabetes, and that this is the reason for the negativity in the c-wave. Transepithelial resistance changes also account for other situations in which the c-wave amplitude changes.^27,46,47

The effect of glucose on the c-wave of diabetics may also be related to resistance, because the changes in the c-wave are in the right direction for further reductions in resistance at high glucose to be the explanation, and because acute glucose manipulations did not change the b-wave. During the terminal experiments, baseline blood glucose values in diabetics were generally more than 100 mg/dl lower than when glucose was measured at regular intervals to check the progress of diabetes. This means that what we designate here as acutely elevated glucose may be more representative of the situation when the animals were in their cages, and the more negative c-wave at high glucose may be due to the even lower resistance of the RPE in awake diabetic animals. Manipulating glucose during the acute experiments may have allowed adjustments in the transepithelial resistance over a continuum, resulting in corresponding changes in the c-wave: more positive when glucose was low and resistance was high, and more negative when glucose was high and resistance was low. Unfortunately, this has to remain a suggestion, because we did not measure resistance under the high and low glucose conditions. The mechanism by which glucose could have influenced resistance in the first place is uncertain, but as noted above, the RPE barrier is known to be affected in diabetes. One possibility is that oxidative stress, which occurs in diabetes, can increase the permeability of the RPE barrier.⁷ How acute lowering of glucose would repair the barrier in diabetics is unknown, but if the c-wave changes are all due to resistance effects, it would suggest that barrier function is more labile in diabetics. The lack of a c-wave change with glucose in control animals would be explained by the barrier being tighter, not damaged, and not labile.

Glucose did not influence the c-wave in control rats, but in normal humans acutely elevated glucose increased the fast oscillation.⁴⁸ It is not known whether this is a resistance change or a change in the generation of the fast oscillation. In contrast to the present findings, glucose increased the c-wave and standing potential of the isolated perfused (normal) cat eye, and decreased glucose had the opposite effect.⁴⁹ This result in cats could be due to the c-wave generator being larger at high glucose, because there was also an increase in the b-wave at high glucose. However, Macaluso et al.⁴⁹ found that slow PIII was not affected by glucose, which suggests that part of their result may have been due to a resistance change. High glucose is accompanied by hyperosmolarity, which can also influence the RPE. Madachi-Yamamoto et al.⁵⁰ found that IV administration of Fructmanit (fructose and mannitol) decreased the standing potential, so the results of Macaluso et al. were not due to hyperosmolarity. At present we cannot fully explain why the the changes caused by acute glucose manipulations are different in cat and diabetic rat.

In summary, consistent with previous work, we found that the c-wave of the ERG was an early indicator of diabetic changes, and we can now attribute this to changes in transepithelial resistance of the RPE. In diabetics we suggest that the resistance can fluctuate with glucose levels. While the c-wave itself cannot easily be recorded in humans, changes in the RPE barrier could be partly responsible for diabetic changes in the outer retina of humans.

Data availability statement

The data that support the findings of this study are available from the corresponding author, RAL, upon reasonable request.