Relationship of Microvascular Disease in Diabetes to Metabolic Control Ronald Engerman, Ph.D., J. M. B. Bloodworth, Jr., M.D., and Susan Nelson, B.S., Madison, Wisconsin
SUMMARY
Dogs were made alloxan-diabetic and randomly distributed into either of two prospective treatment groups. In one group it was intended that the metabolic signs ofdiabetes be controlled poorly, and commercial insulin was administered in doses inadequate to prevent chronic, severe hyperglycemia and glucosuria. In the other group it was intended that the metabolic disorder be well controlled, and the animals received food and commercial insulin twice daily such that the hyperglycemia and glucosuria became mild or infrequent. Experimental improvement of the carbohydrate disorder was accompanied by amelioration of hyperUpemia and other clinical signs of
deficient insulin activity. By 60 months of diabetes, retinal capillary aneurysms, pericyte ghosts, obliterated vessels, and other microvascular abnormalities typical of diabetes were apparent in each animal of the poor-control group. Better control was found to reduce significantly the incidence and severity of microvascular lesions. The data suggest that the mechanism responsible for diabetic retinopathy is initiated as a result of deficient insulin activity and that the development of the microvascular complications of diabetes are preventable and may be inhibited by careful control of the metabolic disorder. DIABETES 25:760-69, August, 1977.
The relationship between microvascular disease and the control of diabetes has been the subject of many investigations and much controversy.1"12 Much of the clinical literature allows that the frequency and/or severity of microvascular disease, especially that seen in the retina, tends to be least among those diabetics in whom hyperglycemia and glucosuria remain infrequent and mild. The data fail to rule out a possibility that the retinal and other lesions result not from poor control but instead from some disorder that coincidentally prevents good control. Retinal and other microvascular lesions typical of
diabetes mellitus in man have been found by us to develop in dogs that have been made diabetic experimentally and allowed to remain hyperglycemic and glucosuric for several years. 13 ' 15 The evidence that the retinal and other vascular lesions of diabetes mellitus in man can be elicited in dogs by alloxan diabetes, 13 " 15 by metasomatotropin diabetes, 13 " 16 and by spontaneous diabetes 17 " 19 lends support to a working hypothesis that the vascular abnormalities are initiated as a result of deficient insulin activity rather than as a result of postulated hereditary defects associated with diabetes. The present investigation was undertaken to determine whether or not the retinal and other microvascular lesions of diabetes are preventable and may be inhibited particularly by careful treatment of the metabolic disorder with commercial insulin.
Presented in part at the Thirty-fifth Annual Meeting of the American Diabetes Association in New York City, June, 1975. From the Departments of Ophthalmology and Pathology, University of Wisconsin, Madison. Address reprint requests to Dr. Ronald Engerman, Department of Ophthalmology, University of Wisconsin, 1300 University Avenue, Madison, Wisconsin 53706. Accepted for publication February 14, 1977. 760
METHODS Animals. Healthy adult dogs of mixed breeds and both sexes were examined; those having normal eyes DIABETES, VOL. 2 6 , NO. 8
RONALD ENGERMAN, PH.D., AND ASSOCIATES
and no hyperglycemia, glucosuria, or other signs of diabetes were selected for the experiment. All animals were housed in metabolism cages and were offered a dry ration (Fromm's Dog Meal, Thiensville, Wisconsin). Soon after purchase their ages were estimated with the assistance of a veterinarian. For each animal we recorded the least age and greatest age readily compatible with the animal's dentition. The mean age reported for each experimental group consists therefore of two values: a mean of the least ages and a mean of the greatest ages. Age is reported in the t a b l e s ^ of the date tissue was obtained for microscopy. Treatment groups. The experiment was conducted as outlined schematically in figure 1. Diabetes was induced by intravenous injection of alloxan monohy-
food were provided twice daily, in the morning and evening. In the good-control group, efforts to reduce hyperglycemia and glucosuria were begun three to 10 weeks after administration of the alloxan, and doses of insulin thereafter were selected twice daily (morning and evening) in the light of the results of equally frequent urinalyses, occasional blood glucose assays, and the animal's past history. The amount of food given in the morning amounted usually to about twice that given in the evening, and the total amount offered daily was restricted to that which a given animal could be expected to consume consistently without showing signs of obesity. Asymptomatic hypoglycemia seemed not uncommon in this group, and
-UNTREATED NONDIABETIC (4) -INSULIN BUT NONDIABETIC (3)
NONDIABETIC (10)
i-ALLOXAN BUT NONDIABETIC (3)
Healthy adult dogs Housed in metabolism— cages Eyes normal
FIGURE 1
ALLOXAN —
—PROSPECTIVE POOR CONTROL (10.) Food ad lib Insulin once/day 4-7 days/week Urinalysis daily Glucosuria severe Body weight subnormal
Design of the experiment (number of dogs).
L-DIABETIC— 60 months 1—PROSPECTIVE GOOD CONTROL (.10) Pood 2x daily Insulin 2x daily Urinalysis 2x daily Glucosuria minimal Body weight normal
drate (40 to 100 mg./kg.) after a 24-hour fast. Animals remaining diabetic after three to 10 weeks were then assigned at random* to either of two treatment groups, both of which were given insulin. In one group (prospective poor control), the diabetes was controlled poorly by intention, and hyperglycemia and glucosuria were allowed to remain chronic and severe; insulin was injected daily in six animals and on four or five days each week in the remaining four of these animals (D157, D182, D185, D202). This periodic deprivation of insulin in the four animals usually elicited ketonuria, whereas spontaneous ketonuria otherwise occurred rarely in the remainder of the group. In the other group (prospective good control) the control of diabetes was intended to be good, and, in an effort to minimize hyperglycemia and glucosuria, insulin and a measured quantity of •Following a trial period of three to 10 weeks of poorly controlled diabetes, animals were selected for the good-control group by tossing a coin or drawing lots. Two animals that had been made diabetic before studies of that group could be begun are included in the poor-control group. AUGUST, 1977
symptoms of hypoglycemia such as tremors, incoordination, or convulsions were treated with additional food and/or parenteral glucose. Procedures necessary to maintain good control were adopted gradually during the course of the experiment, and hyperglycemia and glucosuria were allowed more often in the first of the animals assigned to the good-control group than in animals entering the experiment after greater experience had been gained. Excluded from this report are two dogs that died in hypoglycemia at eight and 16 months of diabetes during an exploratory unsuccessful attempt to maintain aglucosuria while providing insulin once daily and food ad libitum. The two animals were seldom aglucosuric, and no retinopathy developed. Insulin for injection consisted of isophane insulin suspension (NPH, Lilly) supplemented on some occasions with soluble regular insulin (Iletin, Lilly). These commercial preparations consist of mixtures of beef and pork pancreatic extracts. Insulin was injected subcutaneously each morning (about 9 a.m.), and the good-control group received an additional injection each evening (about 7 p.m.) that usually amounted to 761
MICROVASCULAR DISEASE A N D METABOLIC CONTROL
one-third to one-fourth of the total daily dose. On termination of the study, a minimum dose and a maximum dose encompassing 95 per cent of the doses were determined for each animal, and these values were used to calculate the mean minimum and mean maximum dose reported in table 2. Ten nondiabetic dogs were housed and fed among the experimental dogs for comparison and consisted of the following three subgroups: (1) three dogs that had received alloxan but had failed to become diabetic and were killed three to six years afterwards; (2) three seemingly normal dogs that, although not diabetic, were injected with isophane insulin suspension (NPH, Lilly) daily for five to six years in gradually increasing doses up to 6 to 10 U. per day (about 1 U./kg. body weight); (3) four seemingly normal animals that received no other treatment and were killed after three to six years of study. The results show no distinct difference among the three groups and therefore have been pooled in the tables. Blood and urine chemistry. Blood was sampled in the morning before injection of insulin and was assayed usually two or three times per year for serum triglyceride, 19 nonesterified fatty acid (NEFA), 20 and total cholesterol (modification of Lieberman-Burchard reaction). Serum triglyceride levels were measured after an overnight fast. The values reported for serum cholesterol and NEFA were obtained without forcibly fasting the animal; fasting levels were at times compared and found to yield essentially identical results. The lipid values reported are those obtained during the 24-month period immediately preceding histologic examination of tissue. Plasma glucose was measured on randomly selected days in nonfasted animals by a glucose oxidase method (Glucostat, Worthington Biochemicals, Freehold, N. ].). The mean (±S.E.) for each group reported in table 2 was calculated from 10 values, each representing a fiveyear mean for one of the animals. Urine was collected continuously and was tested daily or, in the case of the good-control group, twice daily, for glucose (Tes-Tape, Lilly) and ketones (Acetest, Ames). The quantity and concentration of urinary reducing sugar was measured by Benedict's quantitative method twice a week in the case of the poor-control group, and on any occasion that a 3+ or 4 + Tes-Tape reaction was found in animals of the good-control group. Proteinuria was tested at weekly intervals qualitatively by Albustix (Ames) and quantitatively by the sulfosalicylic acid method. Urinary protein of up to 30 mg./dl. is normal for dogs in this 762
laboratory. Proteinuria greater than normal developed in three of the poor-control group and two of the good-control group and showed no particular association with retinal lesions. Assay of lesions. Fundi were examined with a binocular indirect ophthalmoscope and were photographed with a Zeiss fundus camera in color and again after injection of a 10 per cent aqueous solution of sodium fluorescein (0.1 cc. per kilogram body weight) for angiography. Animals were killed with an overdose of pentobarbital. Eyes and other tissues were excised for histologic study at necropsy or surgically under pentobarbital anesthesia. A small sample of posterior retina from one eye of each animal was fixed in osmium tetroxide or glutaraldehyde solutions for electron microscopy. The remaining retina (about 90 per cent) was fixed for light microscopy usually in 10 per cent formalin, after which it was incubated with trypsin in order to isolate and stain the vasculature. 21 Retinas from a few eyes were fixed in Heidenhain's Susa fluid and were stained with a periodic acid-Schiff (PAS) reaction and mounted whole after incubation one hour with alpha amylase (0.1 per cent in 0.1 M phosphate buffer pH 7) to inhibit background staining and enhance the visibility of blood vessels.22 Conventional paraffin sections were prepared from a variety of other tissues, and samples of kidney, skeletal muscle, and retina were prepared for ultrastructural examination of capillary basement membrane thickness. These will be reported elsewhere. The severity of retinopathy as reported herein was estimated from histologic specimens. Whole mounts of the retina or trypsin-digest preparations of the retinal vessels were examined thoroughly, and the number of saccular capillary aneurysms present in each retina, excluding the anterior quarter of retina, was recorded. The capillary aneurysms recorded were clearly recognizable, but there also occurred questionable aneurysmal structures that could not be clearly identified and presumably included some debris, tangled vessels, and/or exudate as well as possibly some aneurysms. Questionable lesions are not discussed further herein but were more numerous in the poor-control group than in the two other groups. The severity of capillary obliteration was assessed by scanning 20 microscope fields of trypsinized vessels from a given animal and enumerating the acellular capillaries observed. Each field covered an area of 0.5 sq. mm. Vessels of the immediate peripapillary retina and anterior quarter of the retina were excluded from DIABETES, VOL. 2 6 , NO. 8
RONALD ENGERMAN, PH.D., AND ASSOCIATES
the count. A vessel was recorded as an acellular capillary if it was at least 50 micra long and about 2 to 10 micra wide and no endothelial cell or pericyte nuclei were present. Any retina in which acellular capillaries exceeded 200 per 10 sq. mm. was considered to be severely affected and for statistical analysis was recorded as 200 per 10 sq. mm. The severity of pericyte loss was estimated by scanning capillaries until 500 to 1,000 capillary cells (total of endothelial cells and intramural pericytes) were counted and recording the number of ghosts of intramural pericytes encountered coincidentally on these same capillaries. From these data the ratio of ghosts to capillary cells was calculated. RESULTS Prevalence of retinopathy in poorly controlled diabetes.
Prior to undertaking a study of control, before it was known whether or not significant angiopathy would develop in alloxan diabetes, all diabetic animals were controlled intentionally poorly. A total of 26 alloxan-diabetic dogs have been studied histologically in this laboratory after one to six years of intentionally poor metabolic control (table 1). The prevalence and severity of retinopathy in such animals are found to be negligible until many months after the onset of diabetes, but after about three years tend to increase with
increasing durations of diabetes. As shown in table 1, one eye of some animals was removed surgically two or three years before the animal was killed, and it invariably was found to have less extensive or severe abnormalities than the fellow eye, examined at the time of death. The 10 animals studied at 60 months of poorly controlled diabetes were those intended for comparison with the 10 dogs randomly assigned to the goodcontrol group. The present experimental observations, as reported below, pertain to the 20 diabetic dogs studied for 60 months. The metabolic disorder. Animals of the poor-control group received amounts of insulin insufficient to maintain a normal body weight and to prevent chronically severe glucosuria. Hyperphagia, polyuria, and hyperglycemia were consistently evident, and the excretion of urinary reducing sugar during the 60 months of diabetes amounted to a mean of 124 gm. per day at a mean concentration of 6.5 per cent (table 2). Hyperlipemia invariably appeared, and, when measured during the terminal two years preceding autopsy, the serum nonesterified fatty acid (NEFA), triglyceride, and cholesterol were found to be significantly greater (P < 0.05) than in nondiabetic animals (table 2). Animals that had been assigned to the good-control group received, on the average, as much or more insu-
TABLE 1 Prevalence of retinopathy in 26 dogs that were alloxan-diabetic and chronically glucosuric for one to six years. The number of capillary aneurysms occurring in one eye of the animal is reported, and the number in the fellow eye is recorded in parentheses if studied at the same duration of diabetes. Dog
no. D61 D37 D44 D138 D118* D64* D91* D99* D143* D157* D182* D185* D202* D211 D225 D226 D29 D48
Diabetes duration (months) 14 15 15 20 24 36 36 36 36 36 36 36 36 36 36 36 37 38
Aneurysms per retina 0(0) 0(0) 0(0) 0(0) 1 14 0 0 0 0 0 0 2 1 1
Age (yr.)
Dog no.
2
6-8 3-4 3-4 6-8 6-8 8-10 6-8 7-9 8-10 3-5 6-8 7-9 4-6 6-8 5 5
0 12
11-13
D64 S55* D59* D28 D186 D38 D40 D91 D99 D118 D143 S157 D182 D185 D202 D180 D55 D59 D140
6-8
Diabetes duration (months) 46 51 51 53 55 60 60 60 60 60 60 60 60 60 60 65 72 72 72
Aneurysms per retina 86 3 0 20 (17) 4 (4) 52(11) 14 (9) 12 7 20 66 11 27 5 32 75 (50)
60 15 17 (12)
Age (yr.) 9-11 8-10 7-9 6-8 8-10 6-8 6-8 8-10 9-11 9-11 10-12 6-7 8-10 9-11 6-8 7-9 10-12 9-11 8-10
•Surgically obtained tissue.
AUGUST, 1977
763
MICROVASCULAR DISEASE AND METABOLIC CONTROL
TABLE 2 Inhibition of metabolic symptoms of diabetes in the prospective good-control group
n Nondiabetic
10
Poor control
10
Good control
10
Insulin U./day*
4-12
14-21
Urine sugar* Sugargm. free gm. /day /dl. days 0 0 100% 124 ±17 4
±.4
6.5 ±.2
SUMMARY
Dogs were made alloxan-diabetic and randomly distributed into either of two prospective treatment groups. In one group it was intended that the metabolic signs ofdiabetes be controlled poorly, and commercial insulin was administered in doses inadequate to prevent chronic, severe hyperglycemia and glucosuria. In the other group it was intended that the metabolic disorder be well controlled, and the animals received food and commercial insulin twice daily such that the hyperglycemia and glucosuria became mild or infrequent. Experimental improvement of the carbohydrate disorder was accompanied by amelioration of hyperUpemia and other clinical signs of
deficient insulin activity. By 60 months of diabetes, retinal capillary aneurysms, pericyte ghosts, obliterated vessels, and other microvascular abnormalities typical of diabetes were apparent in each animal of the poor-control group. Better control was found to reduce significantly the incidence and severity of microvascular lesions. The data suggest that the mechanism responsible for diabetic retinopathy is initiated as a result of deficient insulin activity and that the development of the microvascular complications of diabetes are preventable and may be inhibited by careful control of the metabolic disorder. DIABETES 25:760-69, August, 1977.
The relationship between microvascular disease and the control of diabetes has been the subject of many investigations and much controversy.1"12 Much of the clinical literature allows that the frequency and/or severity of microvascular disease, especially that seen in the retina, tends to be least among those diabetics in whom hyperglycemia and glucosuria remain infrequent and mild. The data fail to rule out a possibility that the retinal and other lesions result not from poor control but instead from some disorder that coincidentally prevents good control. Retinal and other microvascular lesions typical of
diabetes mellitus in man have been found by us to develop in dogs that have been made diabetic experimentally and allowed to remain hyperglycemic and glucosuric for several years. 13 ' 15 The evidence that the retinal and other vascular lesions of diabetes mellitus in man can be elicited in dogs by alloxan diabetes, 13 " 15 by metasomatotropin diabetes, 13 " 16 and by spontaneous diabetes 17 " 19 lends support to a working hypothesis that the vascular abnormalities are initiated as a result of deficient insulin activity rather than as a result of postulated hereditary defects associated with diabetes. The present investigation was undertaken to determine whether or not the retinal and other microvascular lesions of diabetes are preventable and may be inhibited particularly by careful treatment of the metabolic disorder with commercial insulin.
Presented in part at the Thirty-fifth Annual Meeting of the American Diabetes Association in New York City, June, 1975. From the Departments of Ophthalmology and Pathology, University of Wisconsin, Madison. Address reprint requests to Dr. Ronald Engerman, Department of Ophthalmology, University of Wisconsin, 1300 University Avenue, Madison, Wisconsin 53706. Accepted for publication February 14, 1977. 760
METHODS Animals. Healthy adult dogs of mixed breeds and both sexes were examined; those having normal eyes DIABETES, VOL. 2 6 , NO. 8
RONALD ENGERMAN, PH.D., AND ASSOCIATES
and no hyperglycemia, glucosuria, or other signs of diabetes were selected for the experiment. All animals were housed in metabolism cages and were offered a dry ration (Fromm's Dog Meal, Thiensville, Wisconsin). Soon after purchase their ages were estimated with the assistance of a veterinarian. For each animal we recorded the least age and greatest age readily compatible with the animal's dentition. The mean age reported for each experimental group consists therefore of two values: a mean of the least ages and a mean of the greatest ages. Age is reported in the t a b l e s ^ of the date tissue was obtained for microscopy. Treatment groups. The experiment was conducted as outlined schematically in figure 1. Diabetes was induced by intravenous injection of alloxan monohy-
food were provided twice daily, in the morning and evening. In the good-control group, efforts to reduce hyperglycemia and glucosuria were begun three to 10 weeks after administration of the alloxan, and doses of insulin thereafter were selected twice daily (morning and evening) in the light of the results of equally frequent urinalyses, occasional blood glucose assays, and the animal's past history. The amount of food given in the morning amounted usually to about twice that given in the evening, and the total amount offered daily was restricted to that which a given animal could be expected to consume consistently without showing signs of obesity. Asymptomatic hypoglycemia seemed not uncommon in this group, and
-UNTREATED NONDIABETIC (4) -INSULIN BUT NONDIABETIC (3)
NONDIABETIC (10)
i-ALLOXAN BUT NONDIABETIC (3)
Healthy adult dogs Housed in metabolism— cages Eyes normal
FIGURE 1
ALLOXAN —
—PROSPECTIVE POOR CONTROL (10.) Food ad lib Insulin once/day 4-7 days/week Urinalysis daily Glucosuria severe Body weight subnormal
Design of the experiment (number of dogs).
L-DIABETIC— 60 months 1—PROSPECTIVE GOOD CONTROL (.10) Pood 2x daily Insulin 2x daily Urinalysis 2x daily Glucosuria minimal Body weight normal
drate (40 to 100 mg./kg.) after a 24-hour fast. Animals remaining diabetic after three to 10 weeks were then assigned at random* to either of two treatment groups, both of which were given insulin. In one group (prospective poor control), the diabetes was controlled poorly by intention, and hyperglycemia and glucosuria were allowed to remain chronic and severe; insulin was injected daily in six animals and on four or five days each week in the remaining four of these animals (D157, D182, D185, D202). This periodic deprivation of insulin in the four animals usually elicited ketonuria, whereas spontaneous ketonuria otherwise occurred rarely in the remainder of the group. In the other group (prospective good control) the control of diabetes was intended to be good, and, in an effort to minimize hyperglycemia and glucosuria, insulin and a measured quantity of •Following a trial period of three to 10 weeks of poorly controlled diabetes, animals were selected for the good-control group by tossing a coin or drawing lots. Two animals that had been made diabetic before studies of that group could be begun are included in the poor-control group. AUGUST, 1977
symptoms of hypoglycemia such as tremors, incoordination, or convulsions were treated with additional food and/or parenteral glucose. Procedures necessary to maintain good control were adopted gradually during the course of the experiment, and hyperglycemia and glucosuria were allowed more often in the first of the animals assigned to the good-control group than in animals entering the experiment after greater experience had been gained. Excluded from this report are two dogs that died in hypoglycemia at eight and 16 months of diabetes during an exploratory unsuccessful attempt to maintain aglucosuria while providing insulin once daily and food ad libitum. The two animals were seldom aglucosuric, and no retinopathy developed. Insulin for injection consisted of isophane insulin suspension (NPH, Lilly) supplemented on some occasions with soluble regular insulin (Iletin, Lilly). These commercial preparations consist of mixtures of beef and pork pancreatic extracts. Insulin was injected subcutaneously each morning (about 9 a.m.), and the good-control group received an additional injection each evening (about 7 p.m.) that usually amounted to 761
MICROVASCULAR DISEASE A N D METABOLIC CONTROL
one-third to one-fourth of the total daily dose. On termination of the study, a minimum dose and a maximum dose encompassing 95 per cent of the doses were determined for each animal, and these values were used to calculate the mean minimum and mean maximum dose reported in table 2. Ten nondiabetic dogs were housed and fed among the experimental dogs for comparison and consisted of the following three subgroups: (1) three dogs that had received alloxan but had failed to become diabetic and were killed three to six years afterwards; (2) three seemingly normal dogs that, although not diabetic, were injected with isophane insulin suspension (NPH, Lilly) daily for five to six years in gradually increasing doses up to 6 to 10 U. per day (about 1 U./kg. body weight); (3) four seemingly normal animals that received no other treatment and were killed after three to six years of study. The results show no distinct difference among the three groups and therefore have been pooled in the tables. Blood and urine chemistry. Blood was sampled in the morning before injection of insulin and was assayed usually two or three times per year for serum triglyceride, 19 nonesterified fatty acid (NEFA), 20 and total cholesterol (modification of Lieberman-Burchard reaction). Serum triglyceride levels were measured after an overnight fast. The values reported for serum cholesterol and NEFA were obtained without forcibly fasting the animal; fasting levels were at times compared and found to yield essentially identical results. The lipid values reported are those obtained during the 24-month period immediately preceding histologic examination of tissue. Plasma glucose was measured on randomly selected days in nonfasted animals by a glucose oxidase method (Glucostat, Worthington Biochemicals, Freehold, N. ].). The mean (±S.E.) for each group reported in table 2 was calculated from 10 values, each representing a fiveyear mean for one of the animals. Urine was collected continuously and was tested daily or, in the case of the good-control group, twice daily, for glucose (Tes-Tape, Lilly) and ketones (Acetest, Ames). The quantity and concentration of urinary reducing sugar was measured by Benedict's quantitative method twice a week in the case of the poor-control group, and on any occasion that a 3+ or 4 + Tes-Tape reaction was found in animals of the good-control group. Proteinuria was tested at weekly intervals qualitatively by Albustix (Ames) and quantitatively by the sulfosalicylic acid method. Urinary protein of up to 30 mg./dl. is normal for dogs in this 762
laboratory. Proteinuria greater than normal developed in three of the poor-control group and two of the good-control group and showed no particular association with retinal lesions. Assay of lesions. Fundi were examined with a binocular indirect ophthalmoscope and were photographed with a Zeiss fundus camera in color and again after injection of a 10 per cent aqueous solution of sodium fluorescein (0.1 cc. per kilogram body weight) for angiography. Animals were killed with an overdose of pentobarbital. Eyes and other tissues were excised for histologic study at necropsy or surgically under pentobarbital anesthesia. A small sample of posterior retina from one eye of each animal was fixed in osmium tetroxide or glutaraldehyde solutions for electron microscopy. The remaining retina (about 90 per cent) was fixed for light microscopy usually in 10 per cent formalin, after which it was incubated with trypsin in order to isolate and stain the vasculature. 21 Retinas from a few eyes were fixed in Heidenhain's Susa fluid and were stained with a periodic acid-Schiff (PAS) reaction and mounted whole after incubation one hour with alpha amylase (0.1 per cent in 0.1 M phosphate buffer pH 7) to inhibit background staining and enhance the visibility of blood vessels.22 Conventional paraffin sections were prepared from a variety of other tissues, and samples of kidney, skeletal muscle, and retina were prepared for ultrastructural examination of capillary basement membrane thickness. These will be reported elsewhere. The severity of retinopathy as reported herein was estimated from histologic specimens. Whole mounts of the retina or trypsin-digest preparations of the retinal vessels were examined thoroughly, and the number of saccular capillary aneurysms present in each retina, excluding the anterior quarter of retina, was recorded. The capillary aneurysms recorded were clearly recognizable, but there also occurred questionable aneurysmal structures that could not be clearly identified and presumably included some debris, tangled vessels, and/or exudate as well as possibly some aneurysms. Questionable lesions are not discussed further herein but were more numerous in the poor-control group than in the two other groups. The severity of capillary obliteration was assessed by scanning 20 microscope fields of trypsinized vessels from a given animal and enumerating the acellular capillaries observed. Each field covered an area of 0.5 sq. mm. Vessels of the immediate peripapillary retina and anterior quarter of the retina were excluded from DIABETES, VOL. 2 6 , NO. 8
RONALD ENGERMAN, PH.D., AND ASSOCIATES
the count. A vessel was recorded as an acellular capillary if it was at least 50 micra long and about 2 to 10 micra wide and no endothelial cell or pericyte nuclei were present. Any retina in which acellular capillaries exceeded 200 per 10 sq. mm. was considered to be severely affected and for statistical analysis was recorded as 200 per 10 sq. mm. The severity of pericyte loss was estimated by scanning capillaries until 500 to 1,000 capillary cells (total of endothelial cells and intramural pericytes) were counted and recording the number of ghosts of intramural pericytes encountered coincidentally on these same capillaries. From these data the ratio of ghosts to capillary cells was calculated. RESULTS Prevalence of retinopathy in poorly controlled diabetes.
Prior to undertaking a study of control, before it was known whether or not significant angiopathy would develop in alloxan diabetes, all diabetic animals were controlled intentionally poorly. A total of 26 alloxan-diabetic dogs have been studied histologically in this laboratory after one to six years of intentionally poor metabolic control (table 1). The prevalence and severity of retinopathy in such animals are found to be negligible until many months after the onset of diabetes, but after about three years tend to increase with
increasing durations of diabetes. As shown in table 1, one eye of some animals was removed surgically two or three years before the animal was killed, and it invariably was found to have less extensive or severe abnormalities than the fellow eye, examined at the time of death. The 10 animals studied at 60 months of poorly controlled diabetes were those intended for comparison with the 10 dogs randomly assigned to the goodcontrol group. The present experimental observations, as reported below, pertain to the 20 diabetic dogs studied for 60 months. The metabolic disorder. Animals of the poor-control group received amounts of insulin insufficient to maintain a normal body weight and to prevent chronically severe glucosuria. Hyperphagia, polyuria, and hyperglycemia were consistently evident, and the excretion of urinary reducing sugar during the 60 months of diabetes amounted to a mean of 124 gm. per day at a mean concentration of 6.5 per cent (table 2). Hyperlipemia invariably appeared, and, when measured during the terminal two years preceding autopsy, the serum nonesterified fatty acid (NEFA), triglyceride, and cholesterol were found to be significantly greater (P < 0.05) than in nondiabetic animals (table 2). Animals that had been assigned to the good-control group received, on the average, as much or more insu-
TABLE 1 Prevalence of retinopathy in 26 dogs that were alloxan-diabetic and chronically glucosuric for one to six years. The number of capillary aneurysms occurring in one eye of the animal is reported, and the number in the fellow eye is recorded in parentheses if studied at the same duration of diabetes. Dog
no. D61 D37 D44 D138 D118* D64* D91* D99* D143* D157* D182* D185* D202* D211 D225 D226 D29 D48
Diabetes duration (months) 14 15 15 20 24 36 36 36 36 36 36 36 36 36 36 36 37 38
Aneurysms per retina 0(0) 0(0) 0(0) 0(0) 1 14 0 0 0 0 0 0 2 1 1
Age (yr.)
Dog no.
2
6-8 3-4 3-4 6-8 6-8 8-10 6-8 7-9 8-10 3-5 6-8 7-9 4-6 6-8 5 5
0 12
11-13
D64 S55* D59* D28 D186 D38 D40 D91 D99 D118 D143 S157 D182 D185 D202 D180 D55 D59 D140
6-8
Diabetes duration (months) 46 51 51 53 55 60 60 60 60 60 60 60 60 60 60 65 72 72 72
Aneurysms per retina 86 3 0 20 (17) 4 (4) 52(11) 14 (9) 12 7 20 66 11 27 5 32 75 (50)
60 15 17 (12)
Age (yr.) 9-11 8-10 7-9 6-8 8-10 6-8 6-8 8-10 9-11 9-11 10-12 6-7 8-10 9-11 6-8 7-9 10-12 9-11 8-10
•Surgically obtained tissue.
AUGUST, 1977
763
MICROVASCULAR DISEASE AND METABOLIC CONTROL
TABLE 2 Inhibition of metabolic symptoms of diabetes in the prospective good-control group
n Nondiabetic
10
Poor control
10
Good control
10
Insulin U./day*
4-12
14-21
Urine sugar* Sugargm. free gm. /day /dl. days 0 0 100% 124 ±17 4
±.4
6.5 ±.2
