Retinopathy Validation

VALIDATION OF MOUSE MODELS OF DIABETIC RETINOPATHY

  1. Definition of human diabetic retinopathy


    2. Diabetes is the leading cause of blindness in the United States for patients age 20 to 74. This vision loss or impairment usually occurs as a result of proliferative diabetic retinopathy or macular edema. Proliferative diabetic retinopathy causes vision loss as a result of unregulated growth of abnormal blood vessels out of the retina, resulting in excessive permeability, hemorrhages, and formation of fibrovascular scars. Macular edema does not cause total vision loss, but can cause substantial loss of visual acuity and usable vision.

    Diabetes results in fewer perfused capillaries in the retina and in excessive vascular leakage. Available evidence has suggested that the vascular changes contribute to decreased availability of oxygen, stimulating the production and release of permeability factors and vasoproliferative factors, and culminating with retinal neovascularization in some patients. Nevertheless, it has not been directly demonstrated that the retina becomes ischemic in diabetes. Diabetic retinopathy is a progressive disorder, and several lesions of diabetic retinopathy have been identified which predict the likelihood of vision loss in diabetic patients. These include the extent of capillary nonperfusion, the number of microaneurysms, fluid and lipid accumulation in the macula, and the extent and location of intra-retinal neovascularization.

    Diabetes-induced apoptosis or degeneration of retinal ganglion cells also has been observed in diabetic patients, but the contribution of this change to vision loss in diabetes is unknown at present.

  2. Areas where current rodent models fall short

    3. Rodent models of diabetic retinopathy do develop thickening of basement membrane surrounding retinal capillaries, altered permeability of the vasculature, accelerated death of capillary cells and neuronal cells. However, they have not been found to reproducibly develop retinal microaneurysms in diabetes. Moreover, nonprimate animal models having diabetes have not been documented to develop four key features that are integral to the human disease:


    1. Pre-retinal neovascularization
    2. Retinal or macular edema (only primates have a macula, so macular edema cannot be studied in nonprimates)
    3. Large patches of capillary nonperfusion or retinal ischemia
    4. Evidence of visual loss or impairment (other than due to cataract)

    Mice have only been used rarely in studies of diabetic retinopathy to date, and until now there has been no systematic characterization of any mouse models of diabetic retinopathy.


  3. Retinal phenotyping

    4. Electrophysiologic tests of retinal function currently are the only way available to gauge visual function in mice or other species, and this can be measured at multiple timepoints while animals are alive. Rapidly developing cataract in diabetic rats limits this technique for long-term studies, but most other laboratory animals (such as mice, cats and dogs) develop cataract considerably slower. Retinal thickness might be measured noninvasively as a surrogate for retinal edema, as is being done currently in clinical studies. High-resolution optical coherence tomography likely could measure the retinal thickness, but this has not yet been reported using mice. Alternate methods to assess retinal edema are terminal procedures, and include lyophilization of the retina to determine water content, or quantitating vascular permeability.

    Neovascularization is very difficult to assess noninvasively in the mouse. Fluorescein angiography is used to document neovascularization in patients, but the small size and dramatic curvature of the eye make this very difficult to do reliably and reproducibly in the mouse. Histologic assessment of pre-retinal neovascularization is a quantitative and relatively easy procedure, and gives positive identification of the location of the new vessels. Additionally, measurement of the number of acellular (nonperfused) retinal capillaries gives a quantitative surrogate parameter that should predict the progression to the advanced, neovascular stages of the retinopathy. Mechanisms for the diabetes-induced increase in vasoproliferative factors (or neovascularization) in diabetes, or whether ischemia develops in diabetes has not been explored in mice.

    Thus, evaluation of electrophysiology and retinal thickness might be made at multiple time points while animals are alive, but assessment of changes in the retinal vasculature should be done when animals are killed after long durations of diabetes.

  4. Proposed target for animal models

    1. Retinal cell apoptosis (capillary and neuronal)
    2. Capillary obliteration and dropout
    3. Retinal edema
    4. Pre-retinal neovascularization

  5. Recommended screening strategy

    6. Screening should be performed by light microscopy. Both eyes should be fixed in 10% buffered formalin (pH 7.0). One eye should be embedded in paraffin, sections cut (2-4 um; cornea to optic nerve axis) and stained with H&PAS and H&E. These are for evaluation of pre-retinal neovasculariation and retinal neurodegeneration. The retinal vasculature is prepared from the other formalin-fixed retina for quantitation of capillary cell apoptosis, pericyte loss, and acellular capillaries. Lesions should be assessed at 6 months of diabetes (an early timepoint which should detect accelerated development of retinopathy if present), and at 9-12 months of diabetes (to show progression of pathology with duration of disease). Inasmuch as severity of hyperglycemia is a critical determinant of the development and severity of retinopathy, measurement of glycated hemoglobin or HbA1c while animals are alive are valuable pieces of information.


  6. Acceptable lesions of advanced diabetic retinopathy (increasing order from earlier changes to advanced changes)

    1. Greater than 3-fold increase in number of acellular capillaries compared to age matched nondiabetic controls. (Diabetic C57Bl/6J mice can double the number of acellular capillaries within one year of diabetes, but this is not sufficient to cause advanced or proliferative changes in the retina.)
    2. Statistically significant increase in fluid accumulation in retina or retinal thickening compared to age-matched controls.
    3. Intra-retinal neovascularization.
    4. Pre-retinal neovasculariation.

    Demonstration of accelerated neurodegeneration or loss of retinal ganglion cells compared to age matched controls also will be an important lesion to document, but interpretation of the significance of that change will need to be clarified in the future. Several mouse strains are known to undergo spontaneous retinal degeneration (including DBA/2J, Tubtub/Tubtub, C57BL/6J-Tyrc-2J/J, C3Fe.CGr(Cg)-nr/J, BALBCrds, and retinal degeneration (rd)), so it is critical to make sure that neurodegeneration in models studied is due to diabetes.


  7. Methods

    1. Sample preparation.
      To evaluate retinopathy in animal models that other investigators develop, we would need to receive both eyes from at least 5 and preferably 10 experimental animals, and an equal number of age-matched normal controls. The controls are essential to include because many of the lesions characteristic of the retinopathy are found also in nondiabetic animals and humans (although quantitatively fewer). I would suggest that investigators maintain the animals for as long as possible (to increase the likelihood that pathology would develop), and then carefully remove the eyes and fix in 10% buffered formalin (pH 7.0). Both eyes should be received at Case Western Reserve University within about 4 days of death and fixation if apoptosis is to be assessed (if apoptosis is not to be assessed, the duration in formalin is not important).
    2. Trypsin digest procedure
      After slicing the globe along the equator, the anterior eye (cornea, lens) are discarded. The retina is freed carefully, taking care to leave the optic nerve attached to the retina during the isolation. Each retina is put by itself into a tissue sample holder, and washed in slowly flowing water overnight. Incubate the washed retina in 3% trude trypsin (Difco) containing 0.2 M sodium fluoride in 0.1 M Tris buffer (pH 7.8). The source of this crude trypsin is critical; purified trypsin does not work! After shaking 30-45 min at 37 centigrade, strip the vitreous away and discard. Gently brush retina with a single cat whisker or strand from a sable hair paintbrush to dislodge neural elements. Sometimes more vigorous brushing is required, or longer digestion time. Use one “brush” with the single hair to push retina down and hold it in place, and use another “brush” to stroke the surface of the retina to dislodge neural cells. After about 10 minutes, vessel network should be visible and free from debris. Place clean mounting dish under dissecting scope, fill dish with double distilled water, and place a cleaned labeled microscope slide in dish. Transfer cleaned retinal vessels to slide, and “tack" vessels onto very clean microscope slide with the hair “brush”. Let water drain out from UNDERNEATH the microscope slide containing isolated vessels; that way, gravity and fluid motion help keep vessels in one place on the slide. Air dry onto slide. Stain vessels with hematoxylin and PAS, dehydrate and coverslip
    3. Quantitation of cell death/DNA damage in isolated retinal vasculature using TUNEL technique.
      Acceleration of retinal capillary cell death in vivo and in vitro will be demonstrated by a technique reported previously by us (1; 2) using a commercially available kit to demonstrate TUNEL-positive cells (In Situ Cell Death Kit, Roche Molecular Biochemicals, Indianapolis, IN). The isolated vasculature is dried onto a microscope slide, and the vessels or the cultured cells are then incubated with terminal deoxynucleotidyl transferase to add deoxynucleotide to the free 3'-OH end of DNA breaks. To be identified as apoptotic, cells must be TUNEL-positive and have evidence of nuclear fragmentation or shrinkage. TUNEL-positive cells are identified, counted and photographed under immunofluorescence, and then the vessels re-stained with PAS and hematoxylin to positively identify which cell types are TUNEL-positive (using high power light microscopy). These samples then are restained with PAS and hematoxylin for quantitation of vascular pathology (below).
    4. Quantitation of capillary histopathology.
      From trypsin digest preparations of the retinal vasculature, we plan to obtain such data as the frequency of pericyte ghosts, acellular capillaries, varicose capillaries, and ratio of endothelial cells to pericytes. Endothelial cell and pericyte counts are determined in the mid-retina on approximately 1300 capillary cells per retina, and the number of acellular capillaries is counted in multiple fields of mid-retina (one field adjacent to each of the 5 to 7 retinal arterioles radiating out from the optic disc) and expressed per mm2 of retinal area examined. Pericyte "ghosts", indicating where pericytes had been lost along the capillary, are counted only on capillaries that possess one or more endothelial cells. All of these methods have been successfully used in our laboratory for many years (2-12).
    5. Neurodegeneration.
      From tissue sections that pass through the optic nerve (to allow sampling of reproducible areas among animals), we will determine the number of ganglion cells per unit length of retinal surface and thickness of retinal layers. Careful attention will be paid to making sure that measurements are done reproducibly in the same areas of retina between animals, since number of ganglion cells and retinal thickness vary across the retina. These measurements will be made in all animals adjacent to the optic nerve and at a reproducible point midway between the optic nerve and pars plana.
    6. Vascular permeability assessment.
      Not to be performed at present
    7. Electroretinogram
      Not to be performed at present

  8. Sample preparation for labs sending eyes for evaluation

    1. What we want:
      Both eyes in formalin from at least 5 animals per experimental group. The longer the duration of diabetes, the better (C57bl/6 mice require at least 6 mos diabetes before we can detect retinal microvascular lesions). Include also age-matched nondiabetic animals and diabetic controls for comparison.
    2. Eye collection:
      All tissue needs to be collected fresh (within 10-15 min of death) to get best results. Tissue post-mortem more than 1 hour does not work!
      • Anesthetize animal (we have not yet found that any particular anesthetic is better or worse than others).
      • Place thumb and pointer finger on either side of eye. By spreading thumb and pointer finger, stretch the skin around the eye.
      • Using fine, sharp scissors, make 3-5 deep (5-10mm) cuts while forcing scissors into the eye socket. The eye now should be attached only by the optic nerve leading into the brain. Push the scissor back into the eye socket again to cut the nerve, so that there is some optic nerve still attached to the eye. Do not push on the skull to protrude the eye and then cut horizontal across the skull; this can damage the eye.
      • Drop both eyes into 10% buffered formalin (pH 7.4) in a microfuge tube. Variances in pH seem to influence how well the tissue can be stained later, so check the pH (especially if the fixative is prepared in your lab). Label each tube with animal number. Also send a summary sheet listing all animals and experimental description with date of death; please put your name and email address on the sheet as well. Group identifications need not be on this sheet until analysis is completed. Parafilm each tube to minimize the likelihood of leaks, and put tubes in a zip-lock bag.
      • We will need to receive both eyes within about 4 days of death and fixation if apoptosis is to be assessed (if apoptosis is not to be assessed, the duration in formalin is not important). Please give me a few days warning that tissue about when tissue will be shipped so that we can watch for it. Email tracking number once it is sent.


    1. Mizutani M, Kern TS, Lorenzi M: Accelerated death of retinal microvascular cells in human and experimental diabetic retinopathy. J. Clin. Invest. 97:2883-2890, 1996
    2. Kern TS, Tang J, Mizutani M, Kowluru RA, Nagaraj RH, Romeo G, Podesta F, Lorenzi M: Response of capillary cell death to aminoguanidine predicts the development of retinopathy: comparison of diabetes and galactosemia. Invest Ophthalmol Vis Sci 41:3972-3978, 2001
    3. Engerman RL, Bloodworth JMB, Jr, Nelson S: Relationship of microvascular disease in diabetes to metabolic control. Diabetes 26:760-769, 1977
    4. Engerman RL, Kern TS: Experimental galactosemia produces diabetic-like retinopathy. Diabetes 33:97-100, 1984
    5. Engerman RL, Kern TS: Progression of incipient diabetic retinopathy during good glycemic control. Diabetes 36:808-812, 1987
    6. Engerman RL, Kern TS: Aldose reductase inhibition fails to prevent retinopathy in diabetic and galactosemic dogs. Diabetes 42:820-825, 1993
    7. Engerman RL, Kern TS: Retinopathy in galactosemic dogs continues to progress after cessation of galactosemia. Arch Ophthalmol 113:355-358, 1995
    8. Kern TS, Engerman RL: Galactose-induced retinal microangiopathy in rats. Invest Ophthalmol Vis Sci 36:490-496, 1995
    9. Kern TS, Engerman RL: Comparison of retinal lesions in alloxan-diabetic rats and galactose-fed rats. Curr Eye Res 13:863-867, 1994
    10. Kern TS, Engerman RL: A mouse model of diabetic retinopathy. Arch. Ophthalmol. 114:986-990, 1996
    11. Kern TS, Engerman RL: Pharmacologic inhibition of diabetic retinopathy: Aminoguanidine and aspirin [ARVO abstract]. Invest. Ophthalmol. Vis. Sci., 2000
    12. Kowluru R, Tang J, Kern TS: Abnormalities of retinal metabolism in diabetes and galactosemia. VII. Effects of long-term administration of antioxidants on retinal oxidative stress and the development of retinopathy. Diabetes 50:1938-1942, 2001