Significant glomerular basement membrane thickening in hyperglycemic and normoglycemic diabetic-prone BB Wistar rats


  • Edward C. Carlson,

    Corresponding author
    1. Department of Anatomy and Cell Biology, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, North Dakota
    • Department of Anatomy and Cell Biology, Box 9037, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, ND 58202
    Search for more papers by this author
    • Fax: 701-777-2477

  • Richard C. Vari,

    1. Department of Pharmacology, Physiology, and Therapeutics, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, North Dakota
    Search for more papers by this author
  • Janice L. Audette,

    1. Department of Anatomy and Cell Biology, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, North Dakota
    Search for more papers by this author
  • Michelle A. Finke,

    1. Department of Anatomy and Cell Biology, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, North Dakota
    Search for more papers by this author
  • Michael J. Ressler

    1. Department of Anatomy and Cell Biology, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, North Dakota
    Search for more papers by this author


The diabetic-prone BioBreeding Wistar rat (BB/DP) is an autoimmune model of insulin-dependent diabetes mellitus. Approximately 80–90% of the animals are hyperglycemic (BB/DPh) by 90–120 days of age while those that do not become diabetic in adolescence (BB/DPn) remain normoglycemic for life. Likewise, rats in the diabetes-resistant (BB/DR) strain are normoglycemic. Although renal morphological studies have been carried out in this model, ultrastructural observations of age- and diabetes-related extracellular matrix (ECM) changes, including glomerular basement membrane (GBM) morphometry, are not available. Moreover, possible renal changes in the relatively uncommon BB/DPn control animals have not been reported. The current electron microscopic study was carried out to investigate temporal changes in detergent-treated acellular ECM in BB/DPh rats at 2 weeks, 3 months, 6 months, and 1 year postonset of moderate hyperglycemia. Age-matched BB/DR and BB/DPn control animals were also examined. Our data demonstrate age- and diabetes-related alterations in mesangial matrix distributions and GBM widths and show for the first time significant increases in GBM thickening in both hyperglycemic (BB/DPh) and normoglycemic (BB/DPn) rats when compared to age-matched BB/DR controls. Surprisingly, the rate of increase is greatest in BB/DPn animals. Although the pathogenesis of diabetic basement membrane disease is not completely understood, GBM thickening is widely regarded as a morphological consequence of hyperglycemia. However, data in the current investigation show that ECM alterations, including significantly increased GBM thickness, may occur in genetically diabetic animals in the absence of hyperglycemia. © 2004 Wiley-Liss, Inc.

The BioBreeding Wistar rat (BB rat) is an autoimmune model of insulin-dependent diabetes mellitus (IDDM) discovered in 1974 in a colony of pathogen-free Wistar rats (Chappel and Chappel, 1983). Like human IDDM, the onset of clinical symptoms in diabetes-prone (BB/DP) BB rats appear during adolescence, generally between 60 and 120 days of age (Butler et al., 1983; Chappel and Chappel, 1983). These are characterized by rapid weight loss, polyuria, polydipsia, ketonuria, and finally death unless exogenous insulin is administered (Chappel and Chappel, 1983; Like and Rossini, 1984). The incidence of diabetes through 120 days of age is 60–90% in most colonies and averaged 86% in a viral antibody-free NIH colony in Worcester, Massachusetts (Crisa et al., 1992). The diabetes exhibited by BB rats is not attributed to a single dominant or recessive gene, but the etiology seems to include a genetic component (Chappel and Chappel, 1983). BB/DP rats that do not become diabetic during adolescence (BB/DPn) generally remain normoglycemic for life. Likewise, diabetes-resistant (BB/DR) rats are normoglycemic. Neither BB/DR nor BB/DPn rats show clinical symptoms of diabetes.

During the past 20 years, characterization of and experimentation with the diabetic BB rat increased dramatically. Studies of endogenous hormonal environment have been targeted in a number of investigations that have identified alterations in plasma insulin, glucagon, growth hormone, gonadal hormones, pituitary hormones, and thyroid hormones (Nakhooda et al., 1976; Rossini et al., 1983; Mordes et al., 1987; Cameron et al., 1990; Verhaeghe et al., 1990; Crisa et al., 1992). Significantly, pancreata from these animals showed islet fibrosis and an almost total absence of β-cells (Chappel and Chappel, 1983).

Owing to the well-known relationship of IDDM and renal disease, numerous studies of experimental models of diabetes have centered on the kidney. For these investigations, the BB rat model may be preferable to those induced experimentally by cytotoxins such as streptozotocin, which is a reliable inducer of diabetes, but may be somewhat nephrotoxic (Arison et al., 1967).

Previous morphological analyses in the BB rat show significant renal alterations in hyperglycemic (BB/DPh) rats over age-matched BB/DR animals, including increased thickness in glomerular basement membrane (GBM) (Cohen et al., 1987; Chakrabarti and Sima, 1991; Feld et al., 1995), increases in mesangial volume (Cohen et al., 1987; Chakrabarti and Sima, 1991), and a loss of GBM anionic sites (AS) (Chakrabarti and Sima, 1991). Nevertheless, important information is not available regarding the correlation of nephropathic changes in the extracellular matrix (ECM) associated with the hyperglycemia of diabetes. Most particularly, any possible changes in the relatively uncommon (∼ 10%) but important BB/DPn control animals have not been reported. Accordingly, the current electron microscopic investigation was carried out to investigate carefully the temporal changes in GBM and other ECM components in BB/DR and BB/DPh rats at 2 weeks, 3 months, 6 months, and 1 year postonset of moderate hyperglycemia. In addition, BB/DPn animals were studied at 3 months, 6 months, and 1 year beyond the age (∼ 90 days) at which they would normally be expected to become hyperglycemic. At each time period, AS integrity, acellular mesangial matrix (MM) morphology and GBM thickness were evaluated. Our results show for the first time that, although GBM thickening in diabetes is widely regarded as a morphological consequence of hyperglycemia, both hyperglycemic (BB/DPh) and normoglycemic (BB/DPn) rats show significant GBM thickening when compared to age-matched BB/DR controls.


Mixed-gender BB/DP and age-matched BB/DR rats were received from the laboratory of Dr. Arthur Like, Department of Pathology, University of Massachusetts Medical Center (Worcester, MA). The animals generally were received within 30 days of birth and prior to the onset of diabetes. Onset was defined by a positive (≥ 250 mg/dl) reading of glycosuria (tested daily) as determined by Chemstrip uG urine glucose reagent strips (Boehringer Mannheim). Blood glucose concentrations were determined by Chemstrip bG reagent strips, which were read and recorded three times per week with an Accu-Chek III blood glucose monitoring system (Boehringer Mannheim). Diabetic animals were maintained at 300 ± 50 mg/dl by the Limplant sustained-insulin-release system (Linshin, Canada). By 90 days of age, approximately 90% of the BB/DP animals were hyperglycemic and were designated BB/DPh. The remaining ∼ 10% remained normoglycemic and were designated BB/DPn. BB/DPh animals were divided into groups that were designated 2 weeks, 3 months, 6 months, and 1 year postonset of diabetes. Similar age-matched BB/DR groups were established. BB/DPn animals included 3-month, 6-month, and 1-year groups only. Each group, regardless of type, had at least three animals. All animals were given Purina Rat Chow and water ad libitum in air-filtered metabolic cages with 12-hr alternating light-dark cycles. They were maintained in accordance with the regulations of the NIH Guide for the Care and Use of Laboratory Animals.


At 2 weeks, 3 months, 6 months, or 1 year postonset of diabetes, BB/BPh animals were killed by lethal dose of sodium pentobarbital. Age-matched BB/DR and BB/DPn animals were killed at similar times. Kidneys from all animals were removed, weighed, decapsulated, divided into two aliquots, and placed in cold 0.9% saline.

Tissues from aliquot 1 were prepared for conventional transmission electron microscopy (TEM) with or without prior en bloc polyethyleneimine (PEI) staining (Rada and Carlson, 1991). PEI is a cationic molecule that can be made electron-dense by the addition of 2% phosphotungstic acid; it binds to AS and is readily imaged by TEM. Aliquot 2 samples were subjected to sequential detergent extraction (Carlson et al., 1978) and subsequently were fixed and postfixed for TEM or scanning electron microscopy (SEM) as described previously (Carlson et al., 1997). Thick sections were stained with toluidine blue (1% in 1% sodium borate) and observed by bright-field microscopy. All TEM specimens were observed and photographed in a Hitachi 7500 TEM.

Fixed tissues to be examined by SEM were immersed in liquid N2-cooled Freon 22, cleaved with a liquid N2-cooled razor blade, and prepared as previously described (Carlson et al., 1997). Cleaved tissue samples were mounted on aluminum specimen stubs with colloidal silver and coated with a layer of gold-palladium in a Denton Desk II sputter coater. Fractured glomeruli were observed and photographed in a Hitachi 4700 field emission SEM.


Morphometry of peripheral GBM was carried out as previously described (Carlson et al., 2003). The method was a modification (Dische, 1992) of the orthogonal intercept procedure of Jensen et al. (1979). GBM measurements were carried out on acellular glomeruli from 11 groups of animals, including 4 BB/DR groups, 4 BB/DPh groups, and 3 BB/DPn groups. Each group comprised three or four viable animals at the time of sacrifice, though several died between 6 and 12 months after onset of hyperglycemia.

Two to four sets of 20 identical magnification (37,500 diameters) transmission electron micrographs of GBMs from at least three acellular glomeruli from each animal in each group were used in this study. Randomized measurements of peripheral GBMs (mesangial and stalk GBM regions were excluded) were made with a measuring ruler with a logarithmic scale utilizing a digitizing tablet and appropriate software (Bioquant, R&M Biometrics). “True” GBM thickness in each micrograph set was generated from the measurements according to the orthogonal intercept technique. Approximately 350 measurements were made from each set of micrographs and an average of 3,500 measurements were made for each data point (Fig. 7).


Distributions of true GBM thickness values, derived from orthogonal intercepts on sets of 20 electron micrographs for each animal age and type, were statistically compared between BB/DPh animals and BB/DR or BB/DPn controls using the nonparametric Wilcoxon-Mann-Whitney rank-sum test. Significance value for all tests was set at P < 0.05.


Upon arrival by Federal Express, all BB/DP and BB/DR rats appeared active and healthy. This demeanor continued in both groups until the animals were about 60–90 days of age when approximately 80–90% of the BB/DP rats developed IDDM and were designated BB/DPh. BB/DP animals that did not become diabetic were separated from the group and designated BB/DPn. The first indication of diabetes in BB/DPh animals was a dramatic increase in urine excretion followed by the development of lethargy, polydipsia, and coat yellowing. As the disease progressed, muscle wasting occurred in the extremities and abdomens became distended. In contrast, normoglycemic (BB/DR and BB/DPn) controls remained active and alert and showed no evidence of increased urine production.

Body Weights

At the onset of diabetes, BB/DP and age-matched BB/DR rats weighed approximately the same (∼ 260 g; Fig. 1A). As the diabetes progressed, body weights of the BB/DPh rats increased very slowly and leveled at about 380 g, while age-matched normoglycemics steadily increased in weight and averaged about 455 g at 1 year.

Figure 1.

A: Weekly body weight averages (mean ± SEM) in grams. BB/DR and BB/DPn rats are indicated by open circles and BB/DPh rats by open squares. B: Decapsulated kidney weight averages (mean ± SEM) in grams. BB/DR and BB/DPn rats are indicated by open squares and BB/DPh rats by open circles.

Kidney Weights

BB/DPh decapsulated kidney weights averaged 1.21 ± 0.09 g at 2 weeks postonset of hyperglycemia and increased to 1.64 ± 0.18 g at 3 months, while weights from normoglycemic animals averaged 1.06 ± 0.01 and 1.14 ± 0.67 g, respectively (Fig. 1B). At 2 weeks and 3 months, respectively, BB/DPh kidney weights were 145% and 44% greater than age-matched controls. After 6 months of hyperglycemia, BB/DPh mean decapsulated kidney weights increased to 1.78 ± 0.08 g and by 1 year reached 1.81 ± 0.14 g. In contrast, mean kidney weights in age-matched normoglycemic rats increased to only 1.34 ± 0.08 g at 6 months and to 1.44 ± 0.03 g at 1 year. BB/DPh kidney weights were 33% and 27% larger than age-matched normoglycemic controls at 6 months and 1 year, respectively. Overall, kidneys from BB/DPh animals increased approximately 50% at 1 year postonset of hyperglycemia, while age-matched normoglycemic rats increased approximately 35%.

Blood Glucose Levels

In BB/DPh rats, blood glucose was maintained at 300 ± 50 mg/dl (Fig. 2) by the Limplant sustained-release-insulin implant system. Levels initially dropped sharply as the result of the insulin implants, but averaged 301 mg/dl at all times postonset of diabetes. In BB/DR and BB/DPn animals, blood glucose levels remained normal (125 ± 20 mg/dl) throughout the course of study.

Figure 2.

Weekly blood glucose averages (mean ± SEM) of all BB/DPh rats in weeks 1–52 postonset of hyperglycemia.

Light Microscopic (LM) Studies

LM examination of renal tissues subjected to sequential detergent extractions and endonuclease treatment showed typical cortical histoarchitectural patterns despite the absence of cells. Structural ECM components and all basement membranes were intact. GBMs were particularly well preserved with clearly recognizable peripheral loops, stalk regions, and mesangial areas. LM preparations were used primarily for determination of glomerular position and orientation. However, in these sections, acellular glomeruli from 1-year animals in all groups (BB/DR, BB/DPh, BB/DPn) were compared (Fig. 3) and no significant structural differences were identified. Glomerular diameters (110–138 μm) were similar in all groups and averaged 126.8 ± 6.2 μm (n ≥ 25). Mesangial regions were similar in all groups with only slight increases in MM in BB/DPh animals.

Figure 3.

Representative light micrographs of acellular renal glomeruli in tissues from 1 year: (A) BB/DR, (B) BB/DPh, and (C) BB/DPn rats showing similar glomerular size and distribution of glomerular basement membranes. Glomeruli from BB/DPh rats show slightly increased mesangial matrix compared to those from BB/DPh and BB/DPn rats. Magnifications, 385×.

Electron Microscopic Studies

Anionic sites.

PEI-positive electron-dense particles appeared within all basement membranes and other components of the renal ECM and provided an estimation of the size, location, and density of AS. They were particularly prominent in GBMs where they were larger and more evenly distributed in the lamina rara externa (LRE) than in the lamina rara interna (LRI) or lamina densa (LD).

AS staining patterns were similar in GBMs of all animal groups at all ages. In 1-year animals (Fig. 4), linear intersite distances (at LD/LRE interfaces) averaged 63.7 ± 0.915 nm in normoglycemic BB/DR and BB/DPn rats, while in hyperglycemic BB/DPh rats they averaged 62.0 ± 0.938 nm. Linear AS distributions (at LD/LRE interfaces) averaged 16.42 ± 2.3/nm in BB/DR and BB/DPn rats, and 15.61 ± 2.9/nm in age-matched BB/DPh animals. Distributions of smaller AS within the LRE and LD of peripheral GBMs were variable but similar in all animal groups.

Figure 4.

Representative transmission electron micrographs of renal glomerular basement membranes in renal tissues from 1 year: (A) BB/DR, (B) BB/DPh, and (C) BB/DPn rats showing PEI-positive anionic sites similarly distributed in all groups. Magnifications, 34,000×.

Anionic sites were characterized in renal tissues of 1-year animals by preincubation with heparitinase specifically to degrade heparan sulfate proteoglycan (Linker and Hovingh, 1972). In these preparations, PEI-positive sites were virtually eliminated, and only a few remained within the GBM or weakly associated with glomerular epithelial cells (data not shown).

Acellular renal tissues.

TEM and SEM analyses of acellular renal cortices showed that detergent solubilization rendered the tissues completely acellular while structural ECM components and all basement membranes remained intact (Figs. 5 and 6). By TEM, acellular glomeruli in rats of all ages showed typical histoarchitectural arrangements. Smooth and thin peripheral GBMs radiated from axial mesangial regions forming capillary loops.

Figure 5.

Representative transmission (AC) and scanning (DF) electron micrographs of acellular glomeruli from age-matched BB/DR (A and D), BB/DPh (B and E), and BB/DPn (C and F) rats at 3 months postonset of hyperglycemia. Glomerular basement membranes are indicated by opposing arrows. MM, mesangial matrix. Magnifications, 6,000× (A–C); 6,200× (D–F).

Figure 6.

Representative transmission (AC) and scanning (DF) electron micrographs of acellular glomeruli from age-matched BB/DR (A and D), BB/DPh (B and E), and BB/DPn (C and F) rats at 1 year postonset of hyperglycemia. Glomerular basement membranes are indicated by opposing arrows. MM, mesangial matrix. Magnifications, 6,000× (A–C); 6,200× (D–F).

When acellular glomeruli from 3-month-old BB/DR and age-matched BB/DPh and BB/DPn rats were examined by TEM, only slight increases in GBM widths were noted in hyperglycemic animals (compare Fig. 5A and B). Interestingly, however, slight increases in GBM thicknesses were also recognized in age-matched normoglycemic BB/DPn animals (compare Fig. 5A and C). In all animal groups, MM material was confined to centrolobular regions and did not extend onto peripheral GBMs.

SEM observations of acellular GBMs in 3-month-old BB/DPh (Fig. 5E) or BB/DPn (Fig. 5F) animals showed no increase in MM over BB/DR rats (Fig. 5D), and the three-dimensional morphology of MM material showed normal fenestrated plates in the axial regions of glomeruli from all animal groups.

At 1 year, TEM examination of acellular glomeruli in BB/DR rats showed peripheral GBMs with smooth regular features but increased in width compared to those seen at 3 months (compare Figs. 6A and 5A). MM material partially covered internal surfaces of GBM axial regions and was deposited somewhat more peripherally (on the endothelial side of the GBM) than in younger animals. However, BB/DR rats showed no significant age-related mesangial expansion. GBMs in 1-year BB/DPh rats were more irregular and significantly thicker than age-matched controls (compare Fig. 6A and B). GBM convolutions associated with axial regions were more irregular and the enclosed mesangial spaces were occupied by substantial MM material, and the matrix was deposited farther onto the endothelial sides of peripheral GBM than those in age-matched BB/DR rats. Accordingly, capillary lumenal spaces appeared somewhat reduced. MM in these animals contained small aggregates of striated collagen fibrils, though mesangial regions were not substantially expanded and glomerulosclerotic nodules were not seen.

TEM and SEM studies of 1-year BB/DPn animals demonstrated an expected age-related increase over younger animals of the same type (compare Figs. 6C and 5C, 6D and 5D). However, an unexpected finding was a substantial increase in GBM width over age-matched BB/DR animals (compare Fig. 6C and A). These increases were also evident by SEM (compare Fig. 6F and D). SEM also showed interwoven fenestrated MM plates in all groups of 1-year animals, but they extended farther onto peripheral GBM capillary loops in BB/DPh and BB/DPn rats than in their younger counterparts (compare Figs. 6E and 5E, 6F and 5F) and were much less evident in BB/DR controls (Fig. 6D). In addition, individual fibers, which make up the MM meshwork in BB/DPh and BB/DPn rats, were more substantial than those seen in BB/DR controls.

GBM morphometry.

During the course of TEM observations, it became evident that GBMs increased in width, not only with normal aging but also as a result of hyperglycemia. Much less expected, however, was an apparent increase in GBM width in normoglycemic BB/DPn animals relative to age-matched normoglycemic BB/DR controls.

In an effort to confirm these early impressions, acellular GBMs from all animal groups were subjected to morphometric analyses (Table 1, Fig. 7). These studies confirmed that age-related thickening occurred in normoglycemic BB/DR rats and mean GBM thickness in these animals increased from ∼ 85 nm at 2 weeks to ∼ 205 nm at 1 year of age. The rate of increase in thickness was much greater, however, in BB/DPh rats, where GBM widths increased from a mean of ∼ 97 nm at 2 weeks postonset of diabetes to ∼ 284 nm at 1 year. GBM thickness in the BB/DPh rats was ∼ 14% greater than their age-matched BB/DR counterparts at 2 weeks postonset of hyperglycemia, and this progressed to +55% at 1 year.

Table 1. Glomerular Basement Membrane Thickness in Age-Matched BB/DR, BB/DPh, and BB/DPn Rats
AgeBB/DRaBB/DPhaBB/DPna% changecPd
RangeMean (n)bRangeMean (n)bRangeMean (n)b
  • a

    Values are mean true basement membrane thickness derived from orthogonal intercepts as described in text.

  • b

    Each (n) represents one set of 20 electron micrographs (approximately 350 measurements per set).

  • c

    Relative to mean of BB/DR animals.

  • d

    P value determined by Mann-Whitney rank-sum test.

2 weeks68.4–110.384.86 ± 11.86 (9)      
   85.9–107.196.68 ± 7.73 (9)  +13.920.018
3 months91.4–138.4104.42 ± 14.14 (10)      
   98.0–127.7115.29 ± 10.74 (9)  +10.410.045
     132.1–181.5156.81 ± 18.0 (6)+50.170.002
6 months126.1–162.7144.39 ± 13.44 (9)      
   154.9–183.6172.14 ± 9.42 (9)  +19.210.0004
     153.1–220.8175.6 ± 21.3 (12)+21.640.0004
1 year174.0–262.5205.06 ± 27.36 (7)      
   244.2–339.4284.34 ± 31.5 (10)  +38.660.0007
     258.6–368.9297.8 ± 40.9 (9)+45.230.005
Figure 7.

Glomerular basement membrane thickness in nanometers expressed as mean ± SD in age-matched BB/DR, BB/DPh, and BB/DPn rats at 3 months, 6 months, and 1 year postonset of diabetes. Circles indicate diabetic-resistant controls (BB/DR), squares indicate diabetic-prone hyperglycemics (BB/DPh), and triangles indicate diabetic-prone normoglycemics (BB/DPn).

The unexpected apparent increase in GBM widths in BB/DPn animals was also confirmed (Table 1, Fig. 7). Surprisingly, in these normoglycemic animals, both mean GBM widths and rates of thickening not only significantly exceeded those of age-matched BB/DR rats, but were substantially greater than those seen in age-matched hyperglycemic BB/DPh animals.


A number of investigators have focused on renal morphological changes in the spontaneously diabetic BB rat model of IDDM. However, data regarding the development of ECM changes remain incomplete, especially in animals expressing moderate (∼ 300 mg/dl) hyperglycemia. Accordingly, the current investigation focused on progressive ultrastructural alterations in the glomerular ECM.

Body Weights

BB/DPh rats in the current study showed lower steady-state body weight averages than controls. This was not unexpected, as data from other studies (Cohen et al., 1987; Feld et al., 1995) indicated that moderately hyperglycemic BB rats achieved lower body weights than normoglycemic controls. Although the precise reasons for body weight reduction in diabetics are not fully understood, overall activity levels of BB/DPh rats in the current study decreased, and upper and lower limb musculatures atrophied at the time they developed classic symptoms of diabetes (i.e., polydipsia, polyuria, and hyperglycemia). Other reports likewise indicated weight decline correlated with onset of polyuria and glycosuria in BB rats (Nakhooda et. al., 1976; Chappel and Chappel, 1983), and it has been suggested that protein may be mobilized from muscle tissue as an alternative energy source in these animals.

Kidney Weights

BB/DPh rats in the current study showed a 44% average increase in decapsulated kidney weight at 3 months postonset of hyperglycemia when compared to age-matched controls. This trend continued and at 1 year diabetic kidney weight averages exceeded controls by 27%. Similarly, Abrass et al. (1988) and Reddi and Camerini-Davalos (1990) demonstrated that whole kidneys in streptozotocin-induced diabetic rats enlarge shortly after onset of hyperglycemia. Although clear explanations for diabetic renal hypertrophy remain elusive, it has been suggested that at least in streptozotocin-induced diabetes, depressed proteolysis contributed to increase kidney protein content (Shechter et al., 1994). However, it is not possible to invoke a similar mechanism for BB/DPh animals as their hyperglycemia develops as a result of autoimmune β-cell destruction, while streptozotocin-induced diabetic rats are hyperglycemic as a result of cytotoxicity of the same pancreatic islet cells. One hypothesis that may be common to all models, however, is that renal RNA/DNA ratios are increased in hyperglycemic environments (Gotzsche et al., 1981) and may lead to increased renal protein levels (Seyer-Hansen, 1976).

Blood Glucose Levels

In the current study, blood glucose concentrations in control rats averaged ∼ 125 mg/dl both at 90 days of age and immediately before death and were similar to those reported by Brown et al. (1983). For diabetics, sustained-insulin-release implants maintained blood glucose levels at 300 ± 50 mg/dl. Cycling increases in blood glucose concentrations were observed, however, apparently due to gradual erosion and subsequent depletion of insulin implants. Occasional precipitous decreases in blood glucose concentrations were also noted that generally could be attributed to new Limplant implantation.

Light Microscopic Studies

We have previously shown in several species that at the level of LM, detergent treatment completely removed cellular components while glomerular ECM components, including various renal basement membranes, maintain their normal histoarchitectures (Carlson and Hinds, 1981, 1983; Carlson and Kenney, 1982; Carlson and Chatterjee, 1983; Carlson and Surerus, 1986; Marion and Carlson, 1989; Carlson et al., 1997). Likewise in the current study, whole acellular glomeruli from BB/DR, BB/DPh, and BB/DPn animals at 1 year postonset of diabetes showed normal morphological features and were virtually indistinguishable with only minor increases in MM in BB/DPh kidneys. No glomerulosclerotic nodules similar to those described by Kimmelsteil and Wilson (1936) were seen in any of the BB rat types. This was somewhat surprising, as increased MM is considered a hallmark for diabetes in human and some animal models (Velasquez et al., 1990; Viberti et al., 1994). However, mesangial expansion in moderately hyperglycemic BB rats generally is reportedly unremarkable or absent (Brown et al., 1983; Cohen et al., 1987).

Electron Microscopic Studies

Anionic sites.

In the current study, GBM anionic sites were characterized as heparan sulfate proteoglycan and their densities were similar in BB/DPh and age-matched controls. These results varied from those of several investigators, including Chakrabarti and Sima (1989), who demonstrated significantly reduced site density using the cationic dye cuprolinic blue in 6-month diabetic BB rats. Likewise, significant reductions in PEI or cuprolinic blue labeling have been demonstrated in streptozotocin-induced diabetic rats at up to 8 months of diabetes (Moriya et al., 1993; van den Born et al., 1995a).

Reductions in GBM AS have been tightly linked with increased filtration unit permeability and significant proteinuria (Kanwar, 1984). In this regard, reduced GBM-heparan sulfate proteoglycan is associated with increased capillary permeability and passage of proteins into the ultrafiltrate (Kanwar et al., 1980; Rosenzweig and Kanwar, 1983). Also, Goode et al. (1995) concluded that a reduction in the density of cationic gold-labeled GBM-associated AS in patients with chronic diabetes (greater than 15 years) correlated closely with increased urine protein excretion. Moreover, van den Born et al. (1995b) showed that at 8 months postinduction of diabetes, streptozotocin diabetic rats demonstrated increased capillary permeability to albumin and IgG that correlated with relative decreases in GBM-heparan sulfate proteoglycan. Accordingly, reduction of AS density in GBMs is widely regarded as contributory to diabetic proteinuria. However, since only modest levels of urine protein have been reported for moderately hyperglycemic BB rats (Cohen et al., 1987; Velasquez et al., 1990; Zamlauski-Tucker et al., 1992; Feld et al., 1995), it seems possible that maintenance of normal complements of AS as shown in the current investigation may be related functionally to known relatively low levels of proteinuria in diabetic BB rats.

It is difficult to speculate on the reasons why AS density in the current study conflicts with those of other investigators. However, our PEI staining procedures differ from those utilizing cuprolinic blue (Chakrabarti and Sima, 1989; van den Born et al., 1995a) and the cationicity of these two probes may not be identical. Moreover, the studies were carried out at disparate times postonset of hyperglycemia. Future combined AS/functional studies including creatinine clearance and assessment of proteinuria in BB/DR, BB/DPh, and BB/DPn rats may shed some light on this apparent disparity.

Acellular renal tissues.

Although our LM observations of acellular glomeruli from BB/DR, BB/DPh, and BB/DPn rats at 1 year postonset of diabetes showed few structural differences, increased resolution at the level of SEM and TEM showed several important distinctions. Simple TEM inspection showed that at 3 months, MM was only slightly increased in BB/DPh kidneys over BB/DR controls. However, GBMs in the diabetics appeared markedly thickened, and surprisingly GBMs in BB/DPn animals also appeared wider than those in BB/DR controls. Similar though exaggerated differences were observed in 1-year animals, where modestly increased MM was more extensively distributed in BB/DPh animals and GBMs were substantially widened.

However, by far the most unexpected result in the current investigation related to substantially increased GBM widths in 1-year BB/DPn rats relative to age-matched BB/DR controls. Although BB/DPn rats never became diabetic and remained normoglycemic until death, their GBM thickness appeared similar or even greater than those seen in age-matched hyperglycemics.

GBM morphometry.

Morphometric analyses of all animal types showed substantial increases in GBM thickness with age, and these data were consistent with several studies that indicated this parameter was a reliable biomarker of aging in rats (Yagahashi et al., 1978; Cohen et al., 1987; Feld et al., 1995). Our morphometric data also showed that GBMs in BB/DPh animals were significantly thicker than age-matched BB/DR controls at 3 months, 6 months, and 1 year postonset of hyperglycemia. This was also not unexpected as other investigators showed that as early as 6 months postonset of hyperglycemia, GBMs in BB/DP rats were thicker than age-matched BB/DR rats (Brown et al., 1983; Cohen et al., 1987; Chakrabarti and Sima, 1989; Feld et al., 1995). Also, Feld et al. (1995) showed that at 1 year, GBMs were 25% wider in BB/DP rats than in age-matched BB/DR controls.

Age-related GBM thickness data in normoglycemic BB/DPn animals have not been reported previously. This may be related to several factors, including the relative scarcity of the animals (∼ 10% of BB/DP rats), and a belief that diabetic-resistant (BB/DR) rats served as the most appropriate normoglycemic control for diabetics. As expected, morphometric analyses confirmed our early impressions that GBMs in BB/DPn rats were wider than predicted for a normoglycemic animal group. Remarkably, they not only significantly exceeded those of age-matched normoglycemic BB/DR rats, but were substantially thicker than diabetic BB/DPh rats at 3 months, 6 months, or 1 year.

The precise mechanisms by which basement membranes increase in width are not clear, but most investigators agree that GBM widths increase with age and in several diseases, most notably diabetes mellitus. A number of pathogenetic mechanisms for diabetic basement membrane disease were reviewed recently by Tsilibary (2003). They include hyperglycemia-induced increased type IV collagen synthesis, decreased expression of matrix metalloproteinases (MMP-2 and -3), and increased tissue inhibitors of metalloproteinase (TIMP). Vascular endothelial growth factor (VEGF) may also be involved, as anti-VEGF antibody treatment reduces GBM thickening. Moreover, oxygen radicals/oxidative stress and advanced glycation end products (AGEs) may play a role since AGE inhibitors (e.g., aminoguanidine, which also has antioxidant properties) attenuate diabetic nephropathy. In addition, the polyol enzymatic pathway is stimulated by hyperglycemia and has been implicated in the pathogenesis of chronic diabetic complications. In this regard, diabetic hyperpermeability, an early feature of diabetic microangiopathy, is reduced by aldose reductase inhibitors.

A unifying hypothesis for these mechanisms has been proposed (Nishikawa et al., 2000) and suggests that since increased blood glucose apparently increases the generation of reactive oxygen species (ROS), activates aldose reductase, and induces AGE formation, hyperglycemia-driven ROS production may represent a final common pathway leading to diabetic microvascular damage, including increased basement membrane thickness.

It is clear from a plethora of animal and clinical studies (Skyler, 1996; Feener and King, 2001; Tsilibary, 2003) that most hypotheses on the etiology of diabetic GBM thickening begin with hyperglycemic microvascular damage. A direct causal relationship between chronic hyperglycemia and microvascular complications of diabetes has been established by data from the Diabetes Control and Complications Trial (1993), a randomized multicenter prospective control study. However, data in the current study strongly suggest that microvascular alterations may also occur in genetically diabetic animals in the absence of hyperglycemia.

It should be pointed out that although a large number of GBM measurements were made, the current study focused on relatively few animals, and our data should be confirmed in future studies using larger populations. Furthermore, recent studies have shown that in BB/Wistar rats, females are more resistant than males to activation of protein kinase C and decreased Na+-K+ adenosine triphosphatase activity, which generally are associated with diabetic complications (Sieber et al., 2001). Since the animals used in the current study were of mixed gender, it may be possible that gender-specific differences in GBM thickness were masked in our data set. This also must be considered in future experiments.

In summary, it is difficult to reconcile our GBM thickness data with numerous morphometric studies, including our own (Carlson et al., 1997, 2003), that indicate increased microvascular basement membrane thickness is a concomitant of hyperglycemia. It is also difficult to escape the conclusion that high blood glucose may be sufficient but not required for excessive GBM thickening to occur. Although this clearly is contrary to widely held theories on the mechanisms by which microvessel basement membranes increase in width, it must be pointed out that such mechanisms are not completely understood and likely are multifactorial. Accordingly, resolution of conflicting correlative data on GBM thickness and hyperglycemia requires further study, including careful analyses of prediabetic basement membrane thickening in other spontaneous models of the disease (e.g., NOD mice). Additionally, a reconsideration of early TEM investigations that demonstrate increased capillary basement membrane thickness in genetically diabetic but normoglycemic patients may be appropriate (Barbosa and Saner, 1984).


The authors gratefully acknowledge the technical assistance of Minto Porter. Supported by the North Dakota Lions Foundation (to E.C.C.)