We, the authors declare that no competing interests exist.
Diabetic retinopathy is a vascular disease of the retina characterised by hyperglycaemic and inflammatory processes. Most animal models of diabetic retinopathy are hyperglycaemia-only models that do not account for the significant role that inflammation plays in the development of the disease. In the present study, we present data on the establishment of a new animal model of diabetic retinopathy that incorporates both hyperglycaemia and inflammation. We hypothesized that inflammation may trigger and worsen the development of diabetic retinopathy in a hyperglycaemic environment. Pro-inflammatory cytokines, IL-1β and TNF-α, were therefore injected into the vitreous of non-obese diabetic (NOD) mice. CD1 mice were used as same genetic background controls. Fundus and optical coherence tomography images were obtained before (day 0) as well as on days 2 and 7 after intravitreal cytokine injection to assess vessel dilation and beading, retinal and vitreous hyper-reflective foci and retinal thickness. Astrogliosis and microgliosis were assessed using immunohistochemistry. Results showed that intravitreal cytokines induced vessel dilation, beading, severe vitreous hyper-reflective foci, retinal oedema, increased astrogliosis and microglia upregulation in diabetic NOD mice. Intravitreal injection of inflammatory cytokines into the eyes of diabetic mice therefore appears to provide a new model of diabetic retinopathy that could be used for the study of disease progression and treatment strategies.
Diabetic retinopathy (DR) is the most common microvascular complication of diabetes leading to vascular breakdown in the retina. In 2014, approximately 370 million people suffered from diabetes worldwide [
A key limitation to research into DR intervention strategies is the lack of comprehensive models that mimic the various stages of the disease. One possible reason is that most animal models used are diabetes, i.e., hyperglycaemia-only, models that do not account for the role that inflammation may play in the development of diabetic retinopathy in the eye. It is currently believed that hyperglycaemia causes inflammation; therefore, most of the literature assumes that DR is a result of prolonged hyperglycaemia [
We therefore investigated whether a model of DR can be established by introducing pro-inflammatory cytokines into the vitreous of hyperglycaemic diabetic mice based upon the hypothesis that intravitreal injection of pro-inflammatory cytokines into the eyes of diabetic mice could induce DR. As a result, we investigated the effect of the inflammatory cytokines, IL-1β and TNF-α, in non-obese diabetic (NOD) mice. NOD mice spontaneously develop type I diabetes due to T-cell mediated autoimmune destruction of their β-islet cells [
Female, 15-week old NOD mice (NOD/ShiLtJ; Stock No: 001976) obtained from Jackson Laboratory (Bar Harbor, ME, USA) and CD1 mice (Crl:CD1(ICR); Strain code: 022) obtained from Charles River Laboratories (Wilmington, MA, USA), were used in this study. A group of Swiss mice derived from a non-inbred stock in the laboratory of Dr. de Coulon, Centre Anticancereux Romand, Lausanne, Switzerland served as progenitors for both NOD and CD1 mouse strains [
Mice were weighed before and 1, 2, and 7 days after intravitreal injection. Weight measurements were performed consistently at the same time of the day. The mouse length was measured from the tip of the snout to the base of the tail using a ruler. The body mass index (BMI) was calculated by dividing the mass of the mouse (g) by the square of the animal length (mm2). In comparing the blood glucose levels between CD1 and NOD mice, the NOD mouse glucose levels were not normally distributed as confirmed by a Shapiro-Wilk normality test. As a result, nonparametric Mann-Whitney test was used to analyse statistical differences in blood glucose between the two mouse strains.
Non-fasting blood glucose levels were determined immediately after anaesthesia. A 25G needle (BD Bioscience, CA, USA) was used to prick the lateral tail vein. Blood glucose measurements were carried out using a glucose meter (Freestyle Optium H Glucometer, UK) and test strips (FreeStyle Optium, UK). Readings above the upper limit of the equipment (27.8 mmol/L) were treated as 27.8 mmol/L during data analysis. After collection of the blood sample, pressure was applied to the tail for a few seconds until bleeding stopped.
Mouse recombinant IL-1β (#RMIL1BI, Thermo Fisher Scientific, MA, USA) and mouse recombinant TNF-α (#RMTNFAI, Thermo Fisher Scientific) were intravitreally injected into both eyes of CD1 (non-diabetic) and NOD (diabetic) mice. A pilot study with three different cytokine concentrations (50, 100 and 500 ng/mL per cytokine) was carried out prior to this study to determine the optimum cytokine dose required to elicit moderate ocular effects. Control animals received an intravitreal injection of PBS. A volume of 1 μL of cytokines at a concentration of 500 ng/mL each was injected into the vitreous using a 10 μL Hamilton syringe attached to a 30G ½″ needle (Terumo Medical Corporation, NJ, USA). Injections were performed in anaesthetized mice using a dissection microscope to visualise the needle and avoid damage to the lens. The injection was consistently performed on the temporal side immediately posterior to the limbus at the corneal-scleral intersection.
Mice were anaesthetised and pupils were dilated with 1% tropicamide (Minims, UK). The cornea was kept moist using a lubricating eye gel (GenTeal® Gel; Alcon, Switzerland) and the Micron IV imaging system (Phoenix Research Labs; CA, USA) was used to obtain both fundus and image-guided OCT images before (day 0) as well as on days 2 and 7 after intravitreal cytokine injection. Vessel dilation and beading were assessed from fundus images. Vessel dilation was qualitatively defined as an increase in vessel diameter compared to day 0, before treatment. Vessel tortuosity was qualitatively assessed as a loss of vascular rigidity and increase in vessel ‘waviness’. Vessel beading was qualitatively defined as the presence of inconsistent vascular tone within a blood vessel compared to day 0. The researcher was masked to the strain and treatment received in order to reduce any bias. OCT images were also taken from the periphery of the eye cup, where lesions would mostly occur in the human condition [
Hyper-reflective foci (HRF) observed in OCT scans of retina and vitreous in human patients represent migratory inflammatory infiltrating cells [
The thickness of retinal layers was measured from OCT images between the nerve fibre layer (NFL) and the choroid using the ‘draw line’ and ‘measure’ functions in ImageJ software version 1.46r (National Institutes of Health, MD, USA). Due to the nature of OCT imaging in mice, it was difficult to differentiate between NFL, ganglion cell layer (GCL) and inner plexiform layer (IPL) and all three layers were thus quantified together. Layer thicknesses were acquired from OCT images taken consistently approximately 0.25 mm superiorly to the ONH (red line;
(A) Fundus image showing the position (red line) at which OCT scans were taken to quantify retinal layer thicknesses. (B) OCT image obtained from the position indicated by the red line in (A) showing the different retinal layers quantified. (C) Pseudo-colour OCT image showing the log of the backscattered light intensity. Fluid-filled areas and blood vessel shadows (white arrows) appear hypo-reflective and therefore black in colour OCT images while cell-dense and hyper-reflective areas range in colour from blue to red (red arrows). NFL = Nerve fibre layer; GCL = Ganglion cell layer; IPL = Inner plexiform layer; INL = Inner nuclear layer; OPL = Outer plexiform layer; ONL = Outer nuclear layer; IS/OS = Inner segment/outer segment.
Tissues were collected following CO2 asphyxiation of the mice. Immediately after euthanasia, eye globes and attached optic nerves were removed and fixed in 4% paraformaldehyde in 0.1 M PBS, pH 7.4 for 1 h. Eyes and their optic nerves were then washed in PBS and passed through 10, 20 and 30% sucrose solutions before embedding the eyes in optimal cutting temperature medium (Sakura, Netherlands). Eye globes were sectioned at 12 μm thickness. Tissue sections were rinsed in PBS before blocking with 10% normal goat serum and 0.1% TritonX-100 in PBS for 1 h at room temperature. Primary and secondary antibodies were diluted in the blocking solution. Sections were subsequently incubated at 4 °C overnight with anti- ionized calcium-binding adapter molecule 1 (Iba1) (goat polyclonal Iba1; #ab178846; 1:100; Abcam, UK) and anti-glial fibrillary acidic protein (GFAP)-Cy3 used as a marker for astrocytes and hyper-reactive Müller cells (mouse monoclonal GFAP; 1:1000; C9205; Sigma-Aldrich, MO, USA). Sections were then washed in PBS and incubated at room temperature for 2 h in donkey anti-goat Alexa-488 (1:500; Jackson Immuno Research, PA, USA). Nuclei were stained using DAPI (1 μg/mL; D9542; Sigma-Aldrich, MO, USA). Sections were washed and mounted using anti-fade reagent (Citifluor™; Electron Microscopy Sciences, PA, USA) and coverslips were sealed with nail polish.
All images were taken with a CCD camera mounted on an Olympus FV1000 confocal laser scanning microscope (Olympus, Japan) and processed using FV-10 ASW 3.0 Viewer and ImageJ software. Three images were taken per eye. Using ImageJ, each image was split into its RGB channels with GFAP in the red, Iba1 in the green, and DAPI in the blue channel. For GFAP immunohistochemical analysis, ImageJ was used to quantify the integrated density (area covered by GFAP labelling × mean grey value) within the optic nerve image. The researcher was masked to the mouse strain and treatment used during GFAP and Iba1 immunofluorescence quantification in order to reduce any bias.
All data are given as arithmetic mean + SEM. Data was first tested for normality using the Shapiro-Wilk test. Since blood glucose data was found to be not normally distributed, it was tested using the nonparametric Mann-Whitney test. All other data was normally distributed, and was tested using Student’s t-test, one-way or two-way ANOVA with
At 15 weeks of age, NOD mice were smaller than CD1 mice (
(A) NOD and CD1 mice showing that NOD mice were generally smaller than CD1 mice. (B) A comparison of mass, length, BMI and blood glucose showed that NOD mice were smaller in mass and length (p < 0.0001 for both) and had a significantly lower BMI (p = 0.0279) compared to CD1 mice. However, NOD mice had higher blood glucose levels than CD1 mice (p = 0.0120). Results are expressed as mean ± SEM. Statistical comparisons between NOD and CD1 were carried out using Mann-Whitney test *p ≤ 0.05; ****p < 0.0001; n = 12 eyes per strain. (C) Fundus (top) and OCT (bottom) images of CD1 and NOD mouse retinas showed no differences in retinal vasculature and overall integrity between the two mouse strains. (D) Retinal layer thickness measurements were obtained from OCT images and showed no differences between NOD and CD1 mice at baseline. Results are expressed as mean + SEM; Statistical analysis was carried out using a two-way ANOVA with Tukey’s test for multiple comparisons; n = 12 eyes per strain.
Baseline fundus and OCT images from both mouse strains were obtained with representative images shown in
Pro-inflammatory cytokines, IL-1β and TNF-α, or PBS used as control injection were introduced into the vitreous of both mouse strains and ocular assessments were carried out before injection (day 0) as well as 2 and 7 days after injection. Fundus images revealed no change in vascular morphology in PBS injected CD1 and NOD mice. However, there was an increase in vessel dilation on day 2 compared with baseline in both cytokine treated mouse strains (yellow arrows;
Fundus images showing vascular changes in saline-injected and pro-inflammatory cytokine-treated CD1 and NOD mice. On day 2, vessel dilation (see zoomed images and compare yellow arrows on day 2 with equivalent vessels of the same retina on day 0) increased in both cytokine-treated CD1 and NOD mice while vessel tortuosity (compare white arrow on day 2 with equivalent vessels of the same retina on day 0) was only observed in cytokine-treated NOD mice. Moreover, vessel beading was only seen in cytokine-treated NOD mice and only on day 7 (compare red arrows on day 7 with equivalent vessels of the same retina on day 0).
Vitreous HRF were not observed in PBS injected CD1 or NOD mice (data not included in graph). However, fundus and OCT images revealed the presence of vitreous HRF in both pro-inflammatory cytokine-injected mouse strains. In order to assess the severity of vitreous HRF, a grading scale was developed (
Grade 0 was characterized by no or very few HRF. Grade 1 showed significantly more HRF in the vitreous (red arrows) seen in OCT images only. Grade 2 and 3 were characterized by clumping of vitreous HRF (red circles) creating a ‘window effect’ that obstructed the view of the underlying retina in both fundus and OCT images. Vitreous HRF were classified as grade 2 if HRF formed clumps obstructing up to 25% of the fundus and OCT image area with grade 3 referring to HRF clumps obstructing more than 25% of the fundus and OCT images.
There was an increase in vitreous HRF severity in cytokine-treated NOD mice at day 2 (p < 0.0001) and day 7 (p = 0.0001) compared to day 0 (
(A) Fundus and OCT images showed that vitreous HRF were present in both pro-inflammatory cytokine-treated mouse strains, but were more pronounced in NOD than in CD1 mice. (B) 3D volume OCT images depict the extent of vitreous debris severity in NOD mice on days 2 and 7. (C) The vitreous HRF grading scale was used to compare the vitreous HRF severity between both mouse strains. There was no statistically significant change in severity over time in CD1 mice. However, there was a statistically significant increase in HRF severity in NOD mice at day 2 (p < 0.0001) and day 7 (p = 0.0001) compared to day 0. More importantly, the vitreous HRF severity in cytokine-treated NOD on day 2 (p = 0.0012) and day 7 (p = 0.0428) was significantly higher than in cytokine-treated CD1 eyes. Results are expressed as mean ± SEM; Statistical comparisons were carried out using two-way ANOVA with Tukey’s multiple comparisons test. #denotes statistically significant differences compared to day 0. ###p ≤ 0.001; ####p < 0.0001. *denotes significant differences between cytokine-injected CD1 and NOD mice. *p ≤ 0.05; **p ≤ 0.01; n = 12 eyes per strain.
OCT images showed the presence of intra-retinal abnormalities in cytokine injected NOD mice only with an increase in retinal layer disruption, particularly in the inner segment/outer segment of photoreceptors (IS/OS) on day 2 (white arrow;
(A) OCT images revealed that saline injection did not affect retinal layer integrity in either mouse strain. Pro-inflammatory cytokine injection did not affect retinal layer integrity in CD1 mice. However, cytokine-injection in NOD mice resulted in severe disruption of the IS/OS on day 2 (white arrow) and retinal oedema on day 7 (red arrow). (B) Retinal layer thickness was quantified from OCT images and expressed as a percentage of the baseline thickness. The thickness of the NFL-GCL-IPL increased over time in cytokine-treated NOD mice and this increase was significant on day 2 compared to NOD baseline values (p = 0.0025) and CD1 mice at day 2 (p = 0.0188). At day 7, cytokine-treated NOD mice again showed thicker NFL-GCL-IPL compared to cytokine-treated CD1 mice (p = 0.0410). At both day 2 and day 7, no significant increase was found in the thickness of the INL, OPL, ONL, IS/OS, or choroid in NOD or CD1 compared to their baseline Results are expressed as mean ± SEM; Statistical comparisons were carried out using two-way ANOVA with Tukey’s multiple comparisons test. #denotes statistically significant differences compared to day 0. ##p ≤ 0. *denotes significant differences between cytokine-injected CD1 and NOD mice. *p ≤ 0.05. n = 12 eyes per strain.
Changes in retinal thickness were then quantitatively assessed from OCT images (
OCT images revealed that retinal HRF were present in cytokine treated NOD mice only (
(A) Different classes of intra-retinal HRF (red arrows) were observed within the ONL in pro-inflammatory cytokine-injected NOD mouse retinas. The ONH is indicated by asterisks. (B) A pseudo-colour OCT image of class II graded retinal HRF seen in (A) showing the hypo-reflective regions of sub-retinal fluid deposition between the IS/OS layers and the RPE (white arrows).
Classes | Description | |
---|---|---|
Day 2 | Day 7 | |
HRF seen in the IS/OS | HRF disappeared | |
HRF seen within the IS/OS and ONL and distorted OPL | HRF disappeared, appearance of sub-retinal fluid | |
Columnar HRF seen within the ONL and distorted OPL | Columnar HRF remained and extended into the inner retinal layers |
GFAP was used as a marker of glial reactivity following intravitreal administration of pro-inflammatory cytokines (
(A) Immunohistochemical images showing GFAP expression in PBS-injected or cytokine-treated CD1 and NOD mice. GFAP labelling was evident only within the GCL in PBS-injected CD1 and NOD mouse retinas. In cytokine-treated CD1 retinas, GFAP expression was seen in the GCL but also extended to the IPL. Cytokine-treated NOD retinas, on the other hand, showed GFAP expression extending from the GCL to the ONL. As with OCT analysis, retinal layers of cytokine-treated NOD mice appeared much thicker than those of other treatment groups. (B) Immunohistochemical images showing the presence of Iba1-positive cells within the GCL, IPL, and INL of cytokine-treated CD1 as well as PBS- and cytokine-injected NOD mice (white arrows). Moreover, cytokine-treated NOD retinas revealed Iba1-positive cells (red arrows) within the OPL which were not observed in other groups. Iba1-positive cells were absent in the choroid of PBS-injected CD1 and NOD mice. However, Iba1-positive cells were observed in the choroid of cytokine-treated CD1 and NOD mice as indicated by white arrows. GCL = Ganglion cell layer; IPL = Inner plexiform layer; INL = Inner nuclear layer; OPL = Outer plexiform layer; ONL = Outer nuclear layer; IS/OS = Inner segment/outer segment; RPE = retinal pigment epithelium. Scale bar: 25 μm.
Iba1-positive cells were present in the GCL, IPL and OPL of NOD mice injected with cytokines, NOD mice injected with PBS, and CD1 mice injected with cytokines (
Cytokine-treated NOD mice showed significantly higher GFAP expression in the optic nerve relative to PBS treated NOD mice (p < 0.0001) or even cytokine treated CD1 mice (p = 0.0003) (
(A) Immunohistochemical images showing GFAP expression in the optic nerve of PBS-injected and cytokine-treated CD1 and NOD mice. GFAP labelling was evident in all treatment groups but appeared highest in cytokine-treated NOD mice. (B) Quantification of the integrated density of GFAP in the optic nerve of PBS-injected and cytokine-treated CD1 and NOD mice revealed that GFAP expression was higher in cytokine-treated compared to PBS-injected NOD (p < 0.0001) and cytokine-treated CD1 mice (p = 0.0003). (C) Immunohistochemical images showing the presence of Iba1-positive cells within the optic nerve in all treatment groups. Compared to PBS-injected groups, Iba1-positive cells in cytokine-treated eyes displayed long, extended cellular processes (white arrows). White boxes in the top row have been zoomed in on the bottom to highlight the differences in Iba1-positive cells. Statistical comparisons were carried out using two-way ANOVA with Tukey’s multiple comparisons test. Scale bar: 25 μm; Data presented as mean + SEM. ***p ≤ 0.001; ****p ≤ 0.0001; n = 6 eyes for PBS injected groups; n = 12 eyes for cytokine-treated groups.
Clinically, DR is associated with endothelial cell death, pericyte loss, BRB breakdown, intra-retinal microvascular abnormalities, retinal and optic nerve neovascularisation and macular oedema [
We tested the hypothesis that introducing pro-inflammatory cytokines into the eye of NOD mice may be able to induce ocular lesions resembling signs of DR. Interestingly, despite receiving the same dose of cytokines, CD1 mice did not show as many of the signs seen in NOD mice. The ocular signs observed in CD1 mice included vessel dilation, vitreous HRF (average integrated vitreous HRF severity of 1 on day 7), retinal and optic nerve astrogliosis and microgliosis (
Pathologies | Summary of results from present study | Supporting evidence from clinical studies | |
---|---|---|---|
CD1 | NOD | ||
[ |
|||
[ |
|||
[ |
|||
[ |
|||
[ |
|||
[ |
|||
[ |
|||
[ |
- indicates absence of pathology; + indicates presence of pathology; ++ indicates higher levels of pathology
The pro-inflammatory cytokines had a more severe effect in diabetic NOD mice. The additional signs observed in NOD mice included vessel beading, more severe vitreous HRF (average integrated vitreous HRF severity of 9 on day 7), retinal HRF, sub-retinal fluid accumulation, increased NFL-GCL-IPL thickness, and more severe retinal and optic nerve astrogliosis and microgliosis (
Unlike vitreous HRF, retinal HRF were only observed in cytokine injected NOD mice. These were divided into three classes based on their morphology and the resulting pathology on day 7. Interestingly, the most common class of retinal HRF, Class II, resulted in deposition of fluid between the IS/OS and the RPE by day 7. This is significant as retinal oedema is a classic sign of more advanced DR [
Increased GFAP expression has long been used as a marker of retinal astrogliosis in models of DR [
This investigation has shown that DR models incorporating hyperglycaemia and inflammation seem to result in a more reliable model of DR, where retinal damage is pronounced or accelerated in the presence of inflammation [
A possible limitation to the use of NOD mice to explain mechanisms of DR could be their proneness to autoimmunity due to a genetic defect that causes cytokine dysregulation and a reduced IL-1 response [
In conclusion, intravitreal inflammatory cytokine injection in NOD mice appears to provide a novel model for the study of signs consistent with the development of DR in humans. The model may provide an opportunity to further study the molecular basis of the disease in order to understand the disease pathology. While the present study focused on the eye, it is plausible that inflammation plays a similar role in exacerbating hyperglycaemia-mediated injury in other microvascular complications of diabetes such as nephropathy and neuropathy, thus the concept could be expanded for these indications.
Retinal sections stained with H&E showed an increase in IPL thickness in both CD1 and NOD mice following intravitreal pro-inflammatory cytokine administration. However, the increase in IPL thickness was more pronounced in NOD + cytokines compared to CD1 + cytokines.
(TIF)
O.O.M. received a doctoral scholarship from the Buchanan Ocular Therapeutics Unit, Faculty of Medical and Health Sciences, University of Auckland, New Zealand. C.R.G. holds the W&B Hadden Chair in Ophthalmology.