Nephrol Dial Transplant (2003) 18: 694–702
DOI: 10.1093/ndt/gfg142
Original Article
The effect of growth hormone on the development of diabetic kidney
disease in rats
Daniel Landau1, Eytan Israel1, Inessa Rivkis2, Leonid Kachko3, Bieke F. Schrijvers4,
Allan Flyvbjerg4, Moshe Phillip5 and Yael Segev2
1
Department of Pediatrics, 2Department of Immunology and 3Department of Pathology, Soroka University Medical
Center, Ben Gurion University of the Negev, Beer Sheva, Israel, 4Medical Department M, Medical Research Laboratory M,
Institute of Experimental Clinical Research, Aarhus Kommunehospital, Aarhus C, Denmark and 5Felsenstein Medical
Research Center, Institute for Endocrinology and Diabetes, Schneider Children’s Medical Center of Israel, Petach Tikva,
Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
Abstract
Background. Nephropathy is the most severe complication of diabetes mellitus. We investigated the
effect of exogenous growth hormone (GH) administration on renal function and matrix deposition in the
streptozotocin (STZ) model of type I-diabetic rat.
Methods. Adult female STZ-diabetic rats (D), nondiabetic control rats injected with saline (C) and
control and diabetic rats injected with bovine GH for
3 months (CGH and DGH, respectively) were used.
Results. The usual renal hypertrophy seen in D
animals was more pronounced in the DGH group.
Creatinine clearance increased only in the D rats, but
not in the other groups, including DGH. Albuminuria
was observed in the D animals but was significantly
elevated in the DGH group. Glomeruli from DGH
animals showed more extensive matrix accumulation
(manifested as an increase in mesangialuglomerular
area ratio). Renal extractable insulin-like growth
factor (IGF-I) mRNA was decreased in the D and
DGH groups, but renal IGF-I protein was not
significantly increased. Renal IGF binding protein-1
was increased in the D groups and further increased in
the DGH group, at both the mRNA and protein levels.
Conclusions. GH-treated diabetic rats had less hyperfiltration and more albuminuria, concomitant with
more glomerular matrix deposition, when compared
with regular diabetic animals. This was associated with
a significant increase in renal IGFBP-1, and dissociated from IGF-I changes. Thus, in this model, GH
exacerbates the course of diabetic kidney disease.
Correspondence and offprint requests to: Daniel Landau, MD,
Pediatric Nephrology, Department of Pediatrics, Soroka Medical
Center, P.O. Box 151, Beer Sheva 84101, Israel.
Email: [email protected]
#
Keywords: diabetes insulin-dependent; insulin-like
growth factor; insulin-like growth factor binding
protein-1; somatotropin; steptozotocin
Introduction
Diabetic nephropathy is one of the main causes of
mortality in both insulin- and non-insulin-dependent
diabetes mellitus. It is currently the leading cause of
end-stage renal disease in the Western world, and the
only cause whose incidence is growing [1]. An increase
in kidney size is a very early change in diabetes, and it
is followed by obstruction of the glomerular capillary
lumen and loss of glomerular filtration and function.
Clinical findings include microalbuminuria followed
by proteinuria, hypertension and worsening renal
function [2].
Several studies have shown that the growth hormone
(GH)-insulin-like growth factor (IGF) system may
play a significant role in diabetic kidney disease and in
other nephropathies. Studies in animal models have
shown that the rapid increase in kidney size caused by
streptozotocin (STZ)-induced insulin-dependent diabetes is preceded by an increase in extractable renal
IGF-I [3]. Albuminuria usually occurs within 1 month
of the onset of diabetes. Both the renal hypertrophy
and the albuminuria can be prevented by administration of long acting somatostatin analogues [4]. Kidney
tissue expresses receptors not only for IGF-I but also
for GH [5]. Thus, even though most of the biologic
effects of GH are IGF-I mediated, GH may also act
independently of IGF-I. We have reported previously
an increase in serum GH levels in non-obese diabetic
(NOD) [6], as well as STZ-treated mice [7]. The
increase in circulating GH imitates the changes
2003 European Renal Association–European Dialysis and Transplant Association
GH effects in diabetic kidney disease
described in humans. Using the same models, we also
observed a blunting effect by GH receptor antagonist
on diabetic renal hypertrophy [8]. The molecular
mechanisms that mediate these effects are still
unknown.
The ability of the STZ-induced model of diabetic
kidney disease to imitate human diabetic nephropathy
is disturbed by the fact that even after a follow up of
6 months, there is no appearance of uraemia or
worsening of the proteinuria. However, in contrast
with human diabetes, GH secretion in STZ-diabetic
rats is inhibited [9]. Therefore, in this model, late
sclerotic changes in the rat glomerulus may fail to
develop because of a relative lack of elevation in serum
GH. This paucity of circulating GH may prevent the
activation of processes in the glomerular and tubular
cells that lead to glomerulosclerosis.
The purpose of the present study was to investigate
the role of GH in the induction of the advanced
sclerotic changes seen in diabetic nephropathy, using
the rat STZ-diabetic model. Given the major involvement of the renal IGF system in diabetic nephropathy
and the control of this system by GH, the influence of
such exogenous GH administration on kidney IGF
system gene expression was also examined.
Subjects and methods
Study design
Adult female Wistar rats with initial body weights of 200 g
were studied. Rats were housed three per cage in a room with
a 12:12 h artificial light cycle. Temperature and humidity
were kept in a controlled range. The animals had free access
to standard chow and tap water throughout the experiment.
Previous studies [11] have shown that this regimen can keep
rats with STZ-induced diabetes insulinopenic, hyperglycaemic
and alive for long periods of time without their becoming
ketotic. The animals were divided into four groups: controls
(C), injected with saline only; diabetic rats (D) given a single i.p.
injection of STZ in a dose of 55 mgukg; GH-treated
diabetic rats (DGH) given bovine GH (bGH) (Monsanto,
St Louis, MO), in a single daily s.c. dose of 10 mgukg body
weight; and GH-treated non-diabetic age-matched rats
(CGH), given the same daily bGH dose as the DGH group
(additional controls). Only animals with serum glucose levels
)18 mmolul and glucosuria without ketonuria were included
in the study. bGH was started as soon as the diabetic state
was determined ()24 h of glycosuria from the time of STZ
injection) and was given daily for 3 months. Body weights
were recorded monthly. Urine was tested for ketone bodies
weekly. Prior to death, the animals were placed in individual
metabolic cages for determination of 24-h urine volume,
glucose, creatinine and albumin. Creatinine clearance was
corrected for animal body weight. The animals were killed
by the end of 3 months. Blood was collected at that stage
for assessment of creatinine and IGF-I. Kidneys were
rapidly removed. The left kidney was snap-frozen in liquid
nitrogen and was used for protein and mRNA levels
determination. The right kidney was perfused and fixed
with alternating intra-aortic injections of PBS (0.02 M,
pH 7.4) and 10% neutral-buffered formalin until blanching
was achieved, for histopathologic assessment.
695
Histopathological assessment
A 2-mm thick, horizontally cut slice from the middle of the
left kidney (containing the papilla) was embedded in
Technovit. Sections of 4–5 mm thickness were cut on a
rotation microtome and stained with periodic acid-Schiff
(PAS) and haematoxylin–eosin. All glomeruli in the sections
were analysed by standard pathological criteria. All tissues
were coded and blindly evaluated for the following parameters. The index of mesangial expansion was determined by
a quantitative estimate of the width of the mesangial zones in
each glomerulus as a function of the total glomerular area.
Digital images were acquired through a light microscope
(Zeiss Axioplan 2, Germany) and a refrigerated camera
(Spot, Diagnostic Instruments, Inc.), using a TINA V2.10G
densitometry software (Raytest Isotopenmegerifte GmbH,
Germany). Light intensity was fixed and the same contrast
range was used for all measurements. Thirty glomeruli were
analysed per slide and four randomly selected animals were
chosen from each experimental group. Measurements were
performed by one investigator (I.R.) and repeated twice. The
intra-observer variability was 9%.
Immunohistochemistry
For immunohistochemistry studies, paraffin sections (4 mm)
were deparaffinized in xylene, hydrated in gradual ethanol
concentrations and reacted for 1 h at room temperature with
a monoclonal antimouse collagen type IV antibody (Zymed,
CA). This was followed by incubation with an appropriate
biotinylated second antibody for 30 min and with biotin–
avidin complex peroxidase for 30 min (Vectastain ABC kit,
Vector, CA). The reaction was developed with 3,39diaminobenzidine (DAB) as a substrate. The intensity of
the staining was evaluated under light microscopy in a
semiquantitative way (q1 to q3) for the different glomerular
areas.
Kidney IGF-I protein
Kidney protein extraction was performed as described
previously [7]. Briefly, 80–100 mg of tissue was homogenized
on ice in 1 M acetic acid (5 mlug tissue) with an Ultra Turrax
TD 25 and further disrupted with a Potter Elvehjelm
homogenizer. With this procedure, all IGF binding proteins
(IGFBPs) are removed from kidney tissue. After lyophilization, the samples were re-dissolved in phosphate buffer
(pH 8.0) and kept at 808C until the IGF-I assay was
performed in diluted extracts. Kidney IGF-I levels were
measured by radioimmunoassay (RIA) using a polyclonal
rabbit antibody (Nichols Institute Diagnostics, San Capistrano,
CA) and recombinant human IGF-I as standard
(Amersham International). The tissue IGF-I concentrations
were corrected for the contribution of entrapped
serumIGF-I. Mono-iodinated IGF-I {[125I-(Tyr31)]IGF-I}
was obtained from Novo-Nordisk AuS (Bag-Svaerd,
Denmark). Intra- and inter-assay coefficients of variation
were -5 and 10%, respectively, for both assays.
Western immunoblot analysis
Kidney tissue was homogenized on ice with a polytron
(Kinetica, Littau, Switzerland) in lysis buffer (50 mM Tris,
pH 7.4, 0.2% Triton X-100) containing 20 mM sodium
pyrophosphate, 100 mM NaF, 4 mM EGTA, 4 mM
Na3VO4, 2 mM PMSF, 0.25% aprotinin and 0.02 mguml
leupeptine. Extracts were centrifuged for 20 min at 17 000 g
696
at 48C and the supernatants collected and frozen. For the
detection of kidney IGFBP-1 homogenates were mixed with
5 3 sample buffer and boiled for 5 min, then 100 mg portions
of sample protein were loaded in each gel lane and subjected
to 10% SDS–polyacrylamide gel, and electroblotted into
nitrocellulose membranes. Blots were blocked for 1 h in TBS
buffer (10 mM Tris, pH 7.4, 138 mM NaCl) containing 5%
non-fat dehydrated milk, followed by overnight incubation
with polyclonal antibody against IGFBP-1 (Santa Cruz
Biotechnology, CA) diluted in TBS containing 5% dry milk.
After washing three times for 15 min in TBST (0.05% Tween20, the blots were incubated with secondary anti-mouse
antibody conjugated to horseradish peroxidase for 1 h at
room temperature and then washed again three times. The
antibody band was visualized by enhanced chemiluminiscense (ECL; Amersham, Life Sciences Inc.) and exposed to
Kodak-BioMax film (Eastman Kodak, Rochester, NY).
Protein expression was quantified densitometrically using
Fluorchem software (Alpha-Innotech, CA).
mRNA studies
Total RNA was prepared from frozen tissues by the
Tri-reagent
method
(Molecular
Research
Center,
Cincinnati, OH) and quantified by absorbency at 260 nm.
The integrity of the RNA was assessed by visual inspection
of the ethidium bromide-stained 28S and 18S RNA bands
after electrophoresis through 1.25%u2.2 M formaldehyde
gels. For northern blot analysis, 30 mg of total RNA were
electrophoresed on 1.3% agaroseu2.2 M formaldehyde gels
in 3-morpholinopropanesulfonic acid buffer. The RNA was
then transferred onto MagnaGraph (MSI, Westboro, MA)
nylon membranes and cross-linked to the membrane with a
UV cross-linker (Hoefer Scientific Instruments, San
Francisco, CA). The rat IGFBP-1 probe (a gift from
Dr L. Mathews, University of Oregon, USA) was radiolabelled with [32P]dCTP 3000 Ciummol (Amersham, UK)
by a random primed DNA labelling kit (Boehringer
Mannheim, Germany).
RNA hybridization was performed in a hybridization oven
(Micro-4, Hybaid Ltd, UK) at 658C for 20 h using a
hybridization solution [0.2 mM Na2HPO4 pH 7.2, 7% (vuv)
SDS, 1% (wuv) BSA and 1 mM EDTA]. The washings were
done in 0.4 3 SSC and 0.1% SDS at 658C. Gels were exposed
to Kodak X-Omat AR film (Eastern Kodak) at 708C with
two intensifying screens. The autoradiograms were quantified
with a PhosphorImager (Imagequant, Molecular Dynamics,
Sunnyvale, CA). Each experiment was repeated twice.
Evaluation of renal IGF-I mRNA was performed using
the RT–PCR method. The RNA samples were converted to
cDNA by adding to each sample of RNA (13 ml) 7 ml of
reverse transcriptase reaction mixture, containing: 1 ml of
Moloney murine leukemia virus-reverse transcriptase
(MMLV-RT; 200 Uuml, Sigma, Rechovot, Israel), 0.5 ml
DTT (0.1 M, Sigma), 0.5 ml RNase inhibitor (40 Uuml,
Sigma), 1 ml of oligo-d(T) 12–18 primer (0.5 mguml, Life
Technologies, BRL, Gaithesburg, MD) and 1 ml of dNTP
(2.5 nmoluml each nucleotide, Sigma). The reaction tube was
incubated for 1 h at 378C, then the volume of each sample
was adjusted to 60 ml and the enzyme inactivated by
incubation for 10 min at 658C.
IGF-I and b-actin cDNA were then amplified
by PCR using specific primers. IGF-I sense:
GGACCAGAGACCCTTTGCGGGG; IGF-I antisense:
GGCTGCTTTTGTAGGCTTCAGTGG; b-actin sense:
GACGAGGCCCAGAGCAAGAG;
b-actin
antisense:
GGGCCGGACTCATCGTACTC. Five microlitres of
D. Landau et al.
reverse transcription product was added to 45 ml of PCR
reaction mixture containing 32.75 ml of H2O, 2.5 ml of 59
primer (20 mM), 2.5 ml of 39 primer (20 mM), 2 ml of dNTP
(2.5 nmoluml each nucleotide, Sigma), 5 ml of 10 3 reaction
buffer and 0.25 ml Taq DNA polymerase (Sigma). A
negative control consisting of the reaction mixture without
the cDNA was included in each run. PCR was run for 20–
25 cycles with b-actin primers under the following
conditions: 90 s at 958C, then five to 10 cycles of 45 s
each at 958C, 90 s at 608C and 60 s at 728C. The last 15
cycles were run under the same conditions but at 728C.
Incubation was prolonged by 5 s in each cycle. PCR with
IGF-I primers was run with the same protocol, except that
the annealing temperature was 65 instead of 608C. Every
experiment was amplified with at least two different
number of cycles to ensure that amplification was at the
exponential phase of PCR. We found that 25–30 cycles for
IGF-I and 20–25 cycles for b-actin were in this range.
Under these conditions we also found a linear dose–
response of the PCR product to increasing doses of cDNA.
Fifteen microlitres of each sample containing amplified
cDNA were loaded onto an agarose gel (2%) containing
ethidium bromide (0.5 mguml). A DNA size marker was run
on the same gel (100 bp ladder, Life Technologies, BRL).
PCR products were quantified densitometrically using
Fluorchem software (Alpha-Innotech, CA). To correct for
differences in loading we corrected densitometric values of
IGF-I cDNA with corresponding values of b-actin cDNA
and the IGF-Iub-actin ratio was calculated.
Urinary albumin excretion
The urinary albumin concentration in urine samples from
24 h urine collections obtained prior to death was determined by RIA, using rat albumin antibody and standards.
The urine samples were stored at 208C until assayed.
Rabbit anti-rat antibody RARauAlb was purchased from
Nordic Pharmaceuticals and Diagnostics (Tilburg, The
Netherlands). For standard and iodination, globulin-free
rat albumin was obtained from Sigma Chemical Co.
(St Louis, MO). Urine creatinine values were assessed
simultaneously (using standard laboratory methods) to
calculate albuminucreatinine ratios (U [AlbuCreat]). Natural
logarithmic values for (U [AlbuCreat]) were calculated for
each animal, due to an abnormal distribution of albuminuria
data.
Statistics
One-way analysis of variance was used to evaluate differences between groups for multiple comparisons. The
Kruskal–Wallis modification for non-parametric data was
used as a first step, and the Mann–Whitney test for
differences between the groups was used subsequently. A P
value of -0.05 was considered as significant. Means are
given as "SEM.
Results
Body and kidney weight
The diabetic rats had a significant depression in weight
gain. By the end of the 3-month study period, body
weight had increased in the D group by only 9"4%,
and in the DGH group by 18"4% (P-0.05)
GH effects in diabetic kidney disease
(Figure 1A). In comparison, body weight in the C and
CGH groups increased by 40"5% and 53"6%,
respectively. The Mann–Whitney test identified
significant body weight difference (P-0.05) between:
C vs CGH; C vs D, C vs DGH, CGH vs DGH.
The ratio of kidney weight to body weight (KWuBW)
was significantly elevated in both the DGH and D
groups (160"6% and 200"19% of C; P-0.05 vs C by
Mann–Whitney test), but not in the CGH group
(93"2% of C; PsNS vs C but -0.05 vs DGH)
(Figure 1B).
Creatinine clearance and urine albumin excretion
Creatinine clearance was measured by 24 h urine
collection using metabolic cages prior to death.
Creatinine clearance increased non-significantly in the
D group (139"14% of C, P-0.05 by Mann–Whitney
test), and was unchanged in the other groups,
including the DGH group (101"28% of C)
(Figure 2A).
Albuminuria was measured from the same urine
collections. Twenty-four hour albumin excretion after
3 months of disease was 178"50, 4557"1198 and
Fig. 1. Body weight (BW) at death (A) (presented as the percentage
of BW at the beginning of the experiment) and kidney weight to body
weight (KWuBW) (B) (presented as the percentage of control
animals). C, control rats; CGH, C rats treated with bGH; D,
diabetic rats; DGH, D animals treated with bGH. Differences
between groups are significant by the Kruskal–Wallis test.
697
2497"602 mguday in the C, D and DGH groups,
respectively. The ratio of the natural logarithmic
values of urinary albumin to creatinine (mgumg) in
these samples increased in both diabetic groups
compared with C but was further increased in the D
and DGH groups (208"18% of C vs 184"14% of C;
P-0.05 by Mann–Whitney test) (Figure 2B).
Histopathological assessment
Diabetic non-treated animals showed global proliferation of mesangial cells, with expansion of mesangial
matrix, as well as thickening of capillary walls. The
glomeruli of the DGH animals showed more extensive
matrix expansion and thickening of capillary walls, in
comparison with both controls (C) and diabetic-nontreated animals (D). No significant changes were seen
in the CGH animals (Figure 3). These changes were
generalized all along the examined slides. No extensive
glomerulosclerosis or sclerotic nodule formation in
glomeruli was seen, and there was no tubulointerstitial
infiltration by inflammatory cells in any of the
experimental groups. Mesangial-to-glomerular area
ratio was significantly different between the groups
(124"4% and 158"3% of C in D and DGH groups,
Fig. 2. Creatinine clearance (expressed as percentage of C animals)
(A) and the ratio of the natural logarithmic values of urine albumin
to creatinine (mgumg) (B). C, control rats; CGH, C rats treated with
bGH; D, diabetic rats; DGH, D animals treated with bGH. The Ln
U (AlbuCreat) values between groups are signicantly different by the
Kruskal–Wallis test. ns6 animals per group.
698
D. Landau et al.
Fig. 3. (A) Glomerular representative changes in the experimental groups (PAS stain, 3 400 magnification). (A) Normal glomerulus in control
animals. (B) CGH group: glomerulus with mild focal thickening of peripheral capillary walls. (C) Untreated diabetic animals: proliferation of
mesangial cells, with expansion of mesangial matrix. Segmental thickening of capillary walls is also seen. (D) DGH group (diabetic animals
treated with bGH): proliferation of mesangial cells and expansion of mesangial matrix is more accentuated in comparison with D. The
capillary walls are also more prominently thickened. (B) Mesangial to glomerular area calculations (expressed as the percentage of nondiabetic control), based on the assessment of four animals in each experimental group and 30 glomerular tuftsuanimal. Differences between
groups are significant by the Kruskal–Wallis test.
respectively; P-0.05 when comparing D vs C, DGH vs
C and D vs DGH by Mann–Whitney test) (Figure 3B).
Immunohistochemistry for type IV collagen showed a
more increased type IV collagen intensity in both D
and DGH groups in comparison with non-diabetic
animals (Figure 4). Semiquantitative analysis of the
staining intensity showed no changes between the D
and DGH groups.
Renal IGF-I and IGFBP-1
Steady-state renal IGF-I mRNA levels were decreased
in both D and DGH groups (P-0.05 when comparing
C and CGH vs D, C and CGH vs DGH by the Mann–
Whitney test) (Figure 5). However, renal extractable
IGF-I protein was not significantly changed between
the experimental groups, after 3 months of diabetes
GH effects in diabetic kidney disease
699
Fig. 4. Immunohistochemistry for type IV collagen in controls (C), control animals treated with GH (CGH), diabetic non-treated animals (D)
and diabetic animals treated with GH (DGH). 3 400 magnification.
(131"13% and 127"8% of C; PsNS) (Figure 5).
Steady-state renal IGFBP1 mRNA levels were
increased in the D and DGH groups, more pronouncedly DGH (265"35% and 586"216% of C; P-0.05
when comparing C vs D and C vs DGH using the
Mann–Whitney test). Renal IGFBP1 mRNA levels
were not significantly elevated in the CGH group
(200"70% of C; P)0.1) (Figure 6). A similar pattern
was observed for IGFBP-1 protein content, as assessed
by western blot analysis (Figure 7), including a
significant difference between D and DGH animals
(151"6% and 214"14% of C, respectively; P-0.05 by
Mann–Whitney test).
Discussion
Previous investigations have suggested a role for GH in
the pathogenesis of experimental renal growth and
scarring. Mice transgenic for the bovine GH gene have
increased body weight and develop severe glomerulosclerosis, leading to uraemic death [11]. In this model,
there is a proportionate increase in kidney weight and
body size, but a disproportionate increase in glomerular volume. Hypophysectomy has been shown to
protect diabetic patients from developing nephropathy [12], and a GH-deficient strain of dwarf
rats is relatively protected from the tendency to
develop glomerulosclerosis after subtotal nephrectomy [13] or diabetic related renal hypertrophy [14].
The finding that kidney tissue expresses receptors
not only for IGF-I but also for GH [5] suggests
that even though many of the biologic effects of
GH are IGF-I mediated, there is a possibility of an
IGF-I-independent action of GH. In human insulindependent diabetes mellitus, serum GH levels are
increased, and inversely correlated with metabolic
control [15]. This is in contrast with the STZ rat
model, where serum GH levels are depressed [9]. In
our study, exogenous GH administration to STZ
rats worsened the renal changes, as measured by
different markers: albuminuria was higher, creatinine clearance was not elevated and there was more
matrix deposition (composed among others, by
type IV collagen) on light microscopy. On the other
hand, there was still no overt renal insufficiency in
this experimental design. Creatinine clearance was
increased in the diabetic animals (D) but not increased
in the diabetic animals treated with bGH (DGH). The
lack of hyperfiltration in the GH-treated diabetic
rats after 3 months of GH treatment could be due to
both a prevention of glomerular hyperfiltration by
GH or acceleration in the progression of nephropathy. However, albuminuria was markedly increased
in the DGH animals in comparison with D. The
glomeruli of the DGH animals also had more extensive
fibrotic elements. These changes at 3 months of
700
Fig. 5. (A) Renal IGF-I protein (shown as percentage of C animals).
C, control rats; CGH, C rats treated with bGH; D, diabetic rats;
DGH, D animals treated with bGH. ns5 animals per group.
Analysis of variance showed no significant differences between
groups. (B) Representative ethidium bromide stain of renal IGF-I
mRNA, using RT–PCR. Each two bands from the same experimental group are duplicates of the same animal. The lower panel
shows the amplification of the b-actin mRNA, using the same
amount of initial total RNA. (C) Densitometric analysis of five
separate RT–PCR reactions (representing five different animals per
group). The renal IGF-Iub-actin ratio is represented as the percentage
of non-diabetic control. Analysis of variance (Kruskal–
Wallis modification) showed a significant difference between the
experimental groups.
diabetes are remarkable, as the rat model is known to
be relatively resistant to the development of glomerulosclerosis [16].
Increased GH action on target tissues may be an
important risk factor (independent of IGF-I) for the
development of diabetic complications, such as
nephropathy and retinopathy. We have shown previously that liver GH receptor and GH-binding protein
(GHBP) mRNA levels, as well as liver membrane GH
binding assays were deeply decreased in NOD diabetic
D. Landau et al.
mice [17]. However, renal GH binding to kidney tissue
may be increased in diabetic rats [18]. In addition,
renal GHBP is up regulated in a model of long-term
diabetes [19], without a change in renal GHR mRNA
levels. Given the increased circulating GH levels
applied in this model as well as in human disease, an
increased biological effect of GH on the kidney tissue
(including sclerosis) is possible.
No significant increase in renal extractable IGF-I
protein was seen in this study. Previous reports on
short-term STZ diabetes models have described a
transient increase in extractable IGF-I protein in
association with a decrease in renal IGF-I mRNA
levels [3]. In contrast, there is a persistent increase in
IGFBP-1 mRNA and protein in both the STZ-injected
rat [10] and the NOD mouse model [8], suggesting
either an increased trapping of IGF-I from the
circulation or independent effects of IGFBP-1 on the
kidney. The lack of IGF-I protein accumulation in this
experiment favours the second possibility of independent IGFBP-1 action. For example, in the IGFBP-1
transgenic mouse model, glomerulosclerosis develops,
in spite of a dwarf phenotype, in association with low
IGF-I bioavailability and high circulating GH levels
[20].
GH therapy in this model did not improve somatic
growth. This could have been due to the catabolic
condition of these animals, given that they were not
injected with exogenous insulin. However, it is also
known that a GH resistance exists in the diabetic state,
perhaps caused by a decrease in the expression of
hepatic GH receptors [6], thereby decreasing circulating IGF-I, which is less available to the target tissues
involved in somatic growth. Alternatively, this state of
GH insensitivity could also exist in the peripheral
tissues.
Renal hypertrophy and glomerulomegaly are early
markers for the development of glomerulosclerosis. In
our study, renal hypertrophy was accentuated in the
DGH vs the D animals. The glomerular basement
membrane becomes thickened in insulin-dependent
diabetes, irrespective of disease duration [21].
Glomeruli show mesangial expansion, owing to the
increased deposition of extracellular matrix, which is
composed mainly of type IV collagen (mostly its a-2
fraction), laminin and fibronectin. The accumulation
of extracellular matrix within the mesangial areas is
the most common early finding in glomerular lesions
progressing to end-stage renal disease [22]. Progressive
tubulointerstitial fibrosis is also typical, particularly at
the time when reduction of the glomerular filtration
rate becomes apparent [23]. Histopathological examination of the kidney in the STZ model shows
mesangial cell proliferation but not advanced deposition of extracellular matrix or sclerosis [24], both
characteristic features of human diabetic nephropathy.
In our model, the increased matrix deposition in the
glomeruli in association with the relative decrease in
GFR and increase in albuminuria indicate that GH
accelerates the sclerotic process associated with diabetes. The exact pathways of these changes are not
GH effects in diabetic kidney disease
701
Fig. 6. (A) Northern blot analysis of kidney IGFBP-1 mRNA, using 30 mg total RNA, in the control (C), bGH treated C (CGH), untreated
diabetic (D) and bGH treated diabetic (DGH) rats. The 1.6 kb bands corresponding to the IGFBP-1 transcript are shown in the upper lane.
The ethidium bromide staining of the 28S ribosomal RNA of the samples appears in the lower lane. (B) A PhosphorImager-based
quantification of kidney IGFBP1 mRNA levels, from the autoradiogram shown in the upper panel is shown in the lower panel. The data are
expressed as the percentage of control. Analysis of variance (Kruskall–Wallis) showed a significant difference between the experimental
groups.
Fig. 7. (A) Western blot analysis of kidney IGFBP-1 content, using
100 mg tissue lysates, in the control (1), bGH treated C (2), untreated
diabetic (3) and bGH treated diabetic (4) rats. Total lysates were
separated by SDS–PAGE, followed by immunoblot analysis, using
anti-IGFBP-1 antibody. The 38 kDa band corresponds to IGFBP-1.
Shown is a blot, representative of five independent experiments. (B)
A densitometry analysis summarizing the five independent experiments. Data are expressed as the percentage of non-diabetic controls. Analysis of variance (Kruskall–Wallis) showed a significant
difference between the experimental groups.
clearly understood. Contradictory results using a
similar model were published previously by Nickels
et al. [25]. In that study there was no significant
difference in the severity of the histopathologic
changes between the GH-treated and non-treated
diabetic animals. Contrary to their findings, we did
find evidence for more severe renal damage in GHtreated diabetic rats. We have used a higher dose
(10 mgukg, equivalent to ;2 mguday for a 200 g
animal) and also opted to use bGH, which has a
better affinity to the GH receptor and does not bind to
the prolactin receptor as other GH formulations do
[26]. In comparison, the dose used by Nickels et al. was
only 0.2 mguday and ovine GH was used. This higher
dosage may cause significant metabolic and haemodynamic changes, such as blood pressure alterations,
which have not been measured in this study. This
concern should be addressed in future experiments.
However, even transgenic mice for GH do not develop
systemic hypertension [27].
In summary, diabetic rats treated with GH had less
hyperfiltration and more albuminuria, concomitant
with more glomerular sclerosis when compared with
placebo-treated diabetic controls. This was associated
with a significant increase in renal IGFBP-1, and was
702
dissociated from IGF-I changes. Thus, in this model,
GH exacerbates the course of diabetic renal changes.
Acknowledgements. We thank Dr M. Frieger, from the
Department of Epidemiology, Faculty of Health Sciences, Ben
Gurion University, for the statistical evaluation of data. We also
thank Chen Chayat, for the excellent technical assistance. This
study was partially funded by a grant from the American
Physicians Fellowship for Medicine in Israel.
References
1. US Renal Data System (USRDS) Annual Data Report. Atlas of
end stage renal disease in the United States. National Institutes
of Diabetes and Digestive and Kidney Diseases, Bethesda, MD,
2001
2. Viberti GC. Early function and morphological changes in
diabetic nephropathy. Clin Nephrol 1979; 12: 47–53
3. Flyvbjerg A. Putative pathophysiological role of growth factors
and cytokines in experimental diabetic kidney disease.
Diabetologia 2000; 43: 1205–1223
4. Landau D, Segev Y, Afargan M et al. A novel somatostatin
analogue prevents the development of renal diabetic complications in the non-obese diabetic (NOD) mouse. Kidney Int 2001;
60: 505–512
5. Chin E, Zhou J, Bondy C. Renal growth hormone gene
expression: relationship to renal insulin-like growth factor
system. Endocrinology 1992; 131: 3061–3066
6. Landau D, Segev Y, Eshet R, Flyvbjerg A, Phillip M. Changes in
the growth hormone-IGF-I axis in nonobese diabetic mice. Int J
Exp Diab Res 2000; 1: 9–18
7. Flyvbjerg A, Bennett WF, Rasch R, Kopchick JJ, Scarlett JA.
Effect of a long acting growth hormone receptor antagonist
(G120K-PEG) on renal enlargement, glomerular hypertrophy
and urinary albumin excretion in experimental diabetes in mice.
Diabetes 1999; 48: 377–382
8. Segev Y, Landau D, Rasch R, Rasch R, Flyvbjerg A, Phillip M.
Growth hormone receptor antagonism prevents early renal
changes in the non-obese diabetic (NOD) mouse. J Am Soc
Nephrol 1999; 10: 2374–2381
9. Tannenbaum GH. Growth hormone secretory dynamics in
streptozotocin diabetes: evidence of a role for endogenous
circulating somatostatin. Endocrinology 1981; 108: 76–82
10. Landau D, Chin E, Bondy C et al. Expression of insulin-like
growth factor binding proteins in the rat kidney: effects of longterm diabetes. Endocrinology 1995; 136: 1835–1842
11. Pesce CM, Striker LJ, Peten EP, Elliot S, Striker GE.
Glomerulosclerosis at both early and late stages is associated
with increased cell turnover in mice transgenic for growth
hormone. Lab Invest 1991; 65: 601–605
D. Landau et al.
12. Lundbaek K, Christensen NJ, Jensen VA et al. Diabetes,
diabetic angiopathy and growth hormone. Lancet 1970;
2: 131–133
13. El Nahas AM, Bassett AH, Cope GH. Role of growth hormone
in the development of experimental renal scarring. Kidney Int
1991; 40: 29–34
14. Muchaneta-Kubara EC, Sayed-Ahmend N, Besbas N et al.
Experimental diabetic renal growth: role of growth hormone and
insulin-like growth factor-I. Nephrol Dial Transplant 1994;
9: 1395–1401
15. Press M, Tamborlane WV, Sherwin RS. Importance of raised
growth hormone levels in mediating the metabolic derangements
of diabetes. N Engl J Med 1984; 310: 810–815
16. Feld LG, Cachero S, Ellis EN et al. Resistance to glomerular
injury in the diabetic biobreeding rat. Exp Physiol 1995;
80: 991–1000
17. Landau D, Segev Y, Eshet R, Fyvbjerg A, Phillip M. Changes in
the growth hormone-IGF-I axis in nonobese diabetic mice. Int J
Exp Diab Res 2000; 1: 9–18
18. Marshall SM, Flyvbjerg A, Frystyk J, Korsgaard L, Ørskov H.
Renal insulin-like growth factor I and GHR binding in
experimental diabetes and after unilateral nephrectomy in the
rat. Diabetologia 1991; 34: 632–639
19. Landau D, Domené H, Flyvbjerg A et al. Differential expression
of renal growth hormone receptor and its binding protein in
experimental diabetes mellitus. Growth Hormone IGF Res 1988;
8: 39–45
20. Doublier S, Seurin D, Fouqueray B et al. Glomerulosclerosis in
mice transgenic for human insulin-like growth factor-binding
protein-1. Kidney Int 2000; 57: 2299–2307
21. Hayashi H, Karasawa R, Inn H et al. An electron microscopy
study of glomeruli in Japanese patients with noninsulin
dependent diabetes mellitus. Kidney Int 1992; 41: 749–757
22. Laurie GW, Horikoshi S, Killen PD, Segui-Real B, Yamada Y.
In situ hybridization reveals temporal and spatial changes in
cellular expression of mRNA for a laminin receptor, laminin,
and basement membrane (type IV) collagen in the development
kidney. J Cell Biol 1989; 109: 1351–1362
23. Lane PH, Steffes MW, Fioretto P, Mauer SM. Renal interstitial
expansion in insulin dependent diabetes mellitus. Kidney Int
1993; 43: 661–667
24. Weil R. III, Nozawa M, Koss M, Weber C, Reemtsma K,
McIntosh R. The kidney in streptozotocin diabetic rats.
Morphologic, ultrastructural, and functional studies. Arch
Pathol Med 1976; 100: 37–49
25. Nickels DA, Sabnis SG, Antonovych TT, Poth M. Effect of
chronic growth hormone administration on diabetic nephropathy in the rat. Ann Clin Lab Sci 1993; 23: 462–468
26. Mode A, Tollet P, Wells T et al. The human growth hormone
(hGH) antagonist G120RhGH does not antagonize GH in the
rat, but has paradoxical agonist activity, probably via the
prolactin receptor. Endocrinology 1996; 137: 447–454
27. Peten EP, Striker LJ, Fogo A, Ichikawa I, Patel A, Striker GE.
The molecular basis of increased glomerulosclerosis after
blockade of the rennin angiotensin system in growth hormone
transgenic mice. Mol Med 1994; 1: 104–115
Received for publication: 3.12.01
Accepted in revised form: 21.10.02