AJP-Renal Articles in PresS. Published on August 13, 2002 as DOI 10.1152/ajprenal.00026.2002
Urinary Tumor Necrosis Factor Contributes to Sodium Retention
and Renal Hypertrophy During Diabetes
1
1,2
Keith DiPetrillo1, Bonita Coutermarsh
, and Frank Gesek
.
1
Department of Pharmacology and Toxicology, Dartmouth Medical School, Hanover,
NH 03755
2
Address correspondence to: Frank A. Gesek, Ph.D., Dartmouth Medical School, 7650
Remsen, Hanover, NH 03755 Phone: 603-650-1313, Fax: 603-650-1129, email:
[email protected]
Running Title: TNF Contributes to Early Diabetic Nephropathy.
Copyright 2002 by the American Physiological Society.
2
ABSTRACT:
Nephropathy is a major contributor to overall morbidity and mortality in diabetic patients.
Early renal changes during diabetes include Na retention and renal hypertrophy. Tumor
Necrosis Factor (TNF) is elevated during diabetes and is implicated in the pathogenesis of
diabetic nephropathy. We tested the hypothesis that TNF contributes to Na retention and
renal hypertrophy during diabetes. Rats with streptozotocin-induced diabetes exhibit
increased urinary TNF excretion, Na retention, and renal hypertrophy through the first 20
days of diabetes. Administration of a soluble TNF antagonist (TNFR:Fc) to diabetic rats
reduces urinary TNF excretion and prevents both Na retention and renal hypertrophy. TNF
stimulates Na uptake in distal tubule cells isolated from diabetic rats, providing a possible
mechanism for TNF-induced Na retention. We conclude that urinary TNF contributes to
early diabetic nephropathy, and may serve as a valuable diagnostic marker. Furthermore,
inhibition of TNF during diabetes may attenuate early pathological changes in diabetic
nephropathy.
MeSH Keywords: Distal Tubule, Albuminuria, Sodium Retention, Hypertrophy, TNFR:Fc
3
INTRODUCTION:
Diabetes is the leading cause of end-stage renal disease in the United States and
Europe (24). Diabetic nephropathy contributes substantially to overall morbidity and
mortality in diabetic patients, either directly or indirectly as a risk factor for cardiovascular
disease. Microalbuminuria (defined as urinary albumin/creatinine ratio of 300mg/g) is the
earliest clinical sign of diabetic nephropathy (24) and is associated with an accelerated
decline in glomerular filtration rate (GFR) in diabetic patients, compared to
normoalbuminuric patients (24, 33). Microalbuminuria is an independent risk factor for the
development of ischemic heart disease (5) and cardiovascular events (myocardial infarction,
stroke, hospitalization for congestive heart failure, or death) (11). Nephropathy is associated
with left ventricular hypertrophy and impaired diastolic relaxation in type I diabetic patients
(39). Nephropathy also underlies the development of hypertension in type I diabetic patients
(24). Improved understanding of the early pathological changes in diabetic nephropathy may
identify novel clinical markers for detecting early renal disease, as well as elucidate new
therapeutic strategies to diminish these changes and reduce the progression to end-stage renal
disease.
The principal renal alterations that occur during the initial stage of diabetic
nephropathy are hypertrophy and hyperfunction (see Mogensen et al. for review (23)). Renal
enlargement is common in diabetic patients (4, 21) and may predict progression to overt
diabetic nephropathy (4). Na retention is a manifestation of hyperfunction and is observed in
diabetic patients prior to the onset of albuminuria (29, 41, 45). Na retention likely contributes
to the onset of hypertension (41, 44, 46), which develops from underlying renal dysfunction
in type I diabetic patients (24). Experimental evidence suggests that Na retention may also
contribute to organ hypertrophy. Elevated dietary Na can lead to renal hypertrophy and
induce TGF-β expression in non-diabetic rats (48). TGF-β has been identified as a
4
hypertrophic factor during diabetic nephropathy (see Reeves and Andreoli for review (36)).
Moreover, Na restriction reduces renal hypertrophy in diabetic rats (2). Since hypertension
(24) and renal hypertrophy (4, 21) are frequently observed in diabetic patients, Na retention
may represent a significant manifestation of altered renal function that occurs during
diabetes.
Renal enlargement during early diabetic nephropathy results primarily from tubular
hypertrophy (34). Both proximal and distal tubules are enlarged due to hypertrophy and
hyperplasia (34). Early degenerative changes in diabetes occur in the distal tubule (DT) and
precede changes in proximal tubules (PT) and glomeruli (47). Tubular function is also
altered early in the course of diabetes. Urinary excretion of alpha-1-microglobulin and
Tamm-Horsfall protein (THP), markers of PT and DT dysfunction, respectively, are
increased in diabetic patients prior to the development of albuminuria (35). Altered tubular
function during diabetes likely contributes to Na retention.
The pathological events underlying the initial changes of diabetic nephropathy are not
well defined. Inhibitory monoclonal antibodies against TGF-β attenuate kidney enlargement
in diabetic mice, illustrating that TGF-β participates in renal hypertrophy during diabetes
(50). Thomson et al. propose that hypertrophy of the proximal tubule early in diabetes
increases reabsorption and, subsequently, elevates GFR (43). They demonstrate that
enhanced ornithine decarboxylase (ODC) activity contributes to renal hypertrophy and
hyperfunction. However, inhibition of either pathway does not fully reduce renal
hypertrophy, suggesting the existence of additional pathogenic mechanisms.
Tumor Necrosis Factor (TNF) is elevated during diabetes (3, 9, 22, 25) and TNF is
implicated in the development of diabetic nephropathy (8, 18, 27, 42). In the current study,
we tested the hypothesis that TNF underlies Na retention and renal hypertrophy during
diabetes. To test this hypothesis, we administered a specific inhibitor of TNF to diabetic rats
5
and measured Na balance and wet kidney weight. We also examined the effects of TNF
inhibition on the urinary excretion of TNF, and investigated the ability of TNF to directly
stimulate Na uptake in isolated DT cells.
6
METHODS:
Animals. Twelve-week-old male Sprague-Dawley rats were purchased from Charles River
Laboratories (Wilmington, Massachusetts) and diabetes was induced by tail vein injection of
50mg/Kg streptozotocin (STZ; Sigma, St. Louis, Missouri) dissolved in 0.9% sodium
chloride (NS). Blood glucose levels were measured before and three to five days after STZ
injection to ensure the onset of hyperglycemia. Hyperglycemia was also confirmed by
measuring urine glucose using Ketodiastix (Bayer Corporation, Elkhart, Indiana) for rats
housed in metabolic cages. Control and diabetic rats were treated with the anti-TNF agent
TNFR:Fc, a soluble TNF receptor fusion protein. Subcutaneous TNFR:Fc (2mg/Kg;
Immunex Corporation, Seattle, Washington) injections started one day prior to STZ injection
and continued twice weekly for 20 days.
Metabolic Cage Studies. Animals were housed in metabolic cages to determine whole animal
Na balance. Food intake was determined by carefully weighing the amount of food provided
to each rat housed in a metabolic cage. After 24 hours, the remaining food was collected.
Small particles of chow that fell through the floor of the metabolic cage were collected with
feces by design of the metabolic cage. These small pieces were separated from feces,
weighed with the remaining food, and subtracted from the starting weight to determine total
food intake per 24 hours. Na intake was subsequently calculated as a percentage of food
intake (49) based upon the Na content of the chow provided by the manufacturer (HarlanTeklad, Madison, Wisconsin). All animals received the same lot of rat chow, ensuring
consistent Na content in the food.
Water intake and urine output were also measured in animals housed in metabolic
cages. Urine glucose and ketone concentrations were determined using Ketodiastix (Bayer
Corporation, Elkhart, Indiana). Urine Na, creatinine and albumin concentrations were
quantified at the clinical laboratory at Dartmouth-Hitchcock Medical Center (Lebanon, New
7
Hampshire). Na balance was calculated as the difference between Na intake (determined as a
percentage of food intake) and Na output (urine [Na] x urine volume). This method has been
used successfully to determine Na balance in rats (49). For albumin measurements, a
standard curve (r2 = 0.99) was generated using rat albumin (Sigma, St. Louis, Missouri) with
a lower limit of 5µg/mL. At the conclusion of metabolic cage experiments (day 20), kidneys
were excised and weighed to determine wet kidney weight, an established index of renal
hypertrophy during diabetes (43, 50).
Distal Tubule Cell Isolation. An in situ enzyme digestion procedure previously described
was used to isolate PT and DT cells from diabetic and control rat kidneys (16). Briefly,
kidneys were perfused with a collagenase/ hyaluronidase (1/3 mg/mL) mix and subsequently
excised. The kidneys were sliced longitudinally and the renal medulla was removed. The
remaining renal cortex contained primarily proximal and distal tubule segments. The cortex
was sliced and incubated with collagenase (1mg/mL) for 10 minutes at 37°C. The mixture
was filtered through wire mesh to remove tissue remnants and the cells were incubated with a
magnetic particle coupled to a primary antibody recognizing Tamm-Horsfall protein (THP)
(Polysciences Inc, Warrington, Pennsylvania)(32). The flasks were affixed to magnet to
selectively remove THP-containing DT cells, comprising cells from both the thick ascending
limb (TAL) and distal convoluted tubule (DCT) (30, 40). The remaining cells represent
primarily PT cells. DT cells isolated by this procedure exhibited characteristic DT Na and Ca
transport pathways, including calcitonin and parathyroid hormone-stimulated Ca transport
(10, 13), amiloride and chlorothiazide inhibitable Na uptake (14), and phenylephrinestimulated Na transport (12). The specificity of PT cells isolated by this method was
confirmed by the presence of ethylisopropylamiloride (EIPA)-inhibitable and phloridzinsensitive Na uptake (15).
8
22
+
Sodium Uptake Assay. To quantify Na entry, uptake of Na into isolated DT cells was
measured using a rapid filtration technique described in previous reports (15). DT cells were
incubated with recombinant rat TNF (Research Diagnostics, Inc., Flanders, New Jersey) at
37°C prior to adding 22Na. Entry of 22Na into cells was terminated after 1 min by rapid
addition of ice-cold isoosmotic buffer and cells were filtered onto Whatman GF/C filters
using a Millipore 12-port manifold. Nonspecific binding of 22Na to filters and cells was
determined and subtracted to calculate uptake. Uptake measurements were normalized to
total cellular protein quantified by the Pierce BCA protein assay (Pierce, Rockford, Illinois),
using BSA as a standard.
Detection of TNF and THP. Urine TNF levels were measured by ELISA using the OptEIA™
Rat TNF Set (Pharmingen, San Diego, California) according to the accompanying protocol,
using recombinant rat TNF as a standard. Urinary THP was also measured by ELISA.
Briefly, Nunc Maxisorp 96-well ELISA plates were coated with goat anti-human THP
antibody (ICN, Irvine, California; 1:1000) diluted in coating buffer (0.1M Carbonate, pH
9.5). Urine samples were incubated at room temperature for 3 hours, each well was washed,
and HRP-conjugated sheep anti-human THP antibody (Biogenesis, Kingston, New
Hampshire) was added for 1 hour. TMB substrate set (Pharmingen, San Diego, California)
was used for color development. A standard curve using rat THP (Biogenesis, Kingston,
New Hampshire) was generated for each experiment.
Immunofluorescence and Confocal Microscopy. DT cells were grown on transwell filters to
maintain membrane polarization (Costar, Cambridge, Massachusetts) for confocal
microscopy or on glass coverslips for fluorescence microscopy. The cells were fixed with
methanol-free paraformaldehyde in phosphate-buffered saline (PBS) solution. DT cells were
incubated overnight at 4°C with primary antibodies to TNFR1 or TNFR2 (Santa Cruz, Santa
Cruz, California). DT cells were incubated with secondary Ab coupled to Alexa 568
9
(Molecular Probes, Eugene, Oregon) and treated with ProLong Antifade reagent (Molecular
™
Probes, Eugene, Oregon). Confocal images were captured using a Zeiss microscope attached
to a BioRad Model MRC-1024 scanning confocal system, followed by three-dimensional
reconstruction using BioRad imaging software. Immunoflouresence images were visualized
using a fluorescence microscope (Olympus America Inc., Melville, New York).
Statistical Analysis. Metabolic cage data for each treatment group are presented as mean ±
SEM for Cont (n=3), Diab (n=8), Diab+TNFR:Fc (n=9), and Cont+TNFR:Fc (n=3).
Statistical significance was determined using ANOVA followed by Student Newman-Keuls
multiple comparisons test. P<0.05 was considered significant.
10
RESULTS:
Urinary TNF Excretion is Increased during Diabetes. Renal TNF expression is increased in
animal models of diabetes (26, 42). To determine if renal TNF protein is increased and
released during diabetes, urinary TNF excretion was measured. A significant increase in
urinary TNF excretion was observed within the first 3 days following induction of diabetes
(Fig. 1). Elevated urinary TNF excretion continued through the first 20 days of diabetes.
Urinary TNF excretion was also measured in Diab+TNFR:Fc rats. TNFR:Fc is a soluble
TNF antagonist consisting of two TNFR2 extracellular domains fused to the Fc portion of
human IgG1 (6). TNFR:Fc binds to soluble TNF and prevents TNF from interacting with its
cognate cell surface receptors. In Diab+TNFR:Fc rats, TNFR:Fc administration significantly
reduced urinary TNF excretion, nearly to control levels, throughout the 20 study period (Fig.
1).
TNFR:Fc does not Alter the Metabolic Profile of Diabetic Rats. To accurately determine the
role of TNF in the development of diabetic nephropathy, we first determined whether TNF
inhibition alleviated the metabolic abnormalities of diabetes, since improved metabolic
control could prevent the development of diabetic nephropathy (1, 24, 37). Urine glucose
was significantly enhanced in Diab rats from day 3 through day 20 (Fig. 2A). TNFR:Fc
therapy did not affect urinary glucose at any day of the study period. The urinary ketone
concentration in the Diab group was not significantly increased above control until days 17
and 20 (Fig. 2B). TNFR:Fc administration did not alter ketonuria in Diab rats throughout the
first 17 days of diabetes, but did significantly reduce ketonuria at day 20 (Fig. 2B). Water
and food intake, as well as urine output, were increased equally in both the Diab and
Diab+TNFR:Fc groups (Figs. 2C, 2D, and 2E). Although food intake increased, Diab rats
lost weight during the initial week of diabetes, and subsequently maintained stable body
weight throughout the rest of the study period (Fig. 2F). TNFR:Fc treatment did not alter
11
body weight changes during diabetes. Urine glucose and ketones were virtually undetectable
in Cont rats. Water intake, urine output, and food intake remained stable in Cont rats as they
progressively gained body weight throughout the 20-day study period. TNFR:Fc
administration did not affect these metabolic indices in Cont+TNFR:Fc rats.
TNFR:Fc Reduces Na Retention in Diabetic Rats. To investigate the role of TNF in the
maintenance of Na homeostasis during diabetes, we measured dietary Na intake and urinary
Na output in Diab and Diab+TNFR:Fc rats. Dietary Na intake remained stable in Cont and
Cont+TNFR:Fc rats throughout the 20-day study period (Fig. 3A). In contrast, dietary Na
intake was increased above control equally in Diab and Diab+TNFR:Fc rats starting at day 7
and continuing through day 20 (Fig. 3A).
Urinary Na output was constant throughout the study period in Cont rats (Fig. 3B).
TNFR:Fc therapy did not alter urinary Na output in Cont+TNFR:Fc rats. Na output was
elevated in Diab rats compared to Cont, probably due to markedly increased dietary Na
intake in Diab rats (Fig. 3B). TNFR:Fc administration significantly enhanced urinary Na
output in Diab+TNFR:Fc rats throughout the first 17 days of the study period (Fig. 3B). The
ability of TNFR:Fc to augment urinary Na excretion during diabetes was independent of
therapeutic effects on the metabolic state of the animals since TNFR:Fc did not alter
metabolic indices of diabetes through day 17 (Fig. 2).
After quantifying dietary Na intake and urinary Na output, we calculated the
difference between Na intake and output to determine Na balance. Na balance was stable in
Cont and Cont+TNFR:Fc rats during the study period (Fig. 3C). Diab rats exhibited
significant Na retention compared to Cont, as early as day 7 of diabeted, and continued to
retain Na through day 20 of the study. TNFR:Fc treatment significantly reduced Na retention
in Diab+TNFR:Fc rats (Fig. 3C).
12
TNFR:Fc Reduces Renal Hypertrophy during Diabetes. Experimental studies demonstrate
that excess Na contributes to renal hypertrophy (2, 48). Increased kidney size is common in
diabetic patients (4, 21). Based upon our findings that TNF inhibition reduced Na retention
during diabetes, we tested the hypothesis that TNF inhibition would reduce renal hypertrophy
during diabetes. As an index of renal hypertrophy, we measured average wet kidney weight
from TNFR:Fc treated and untreated control and diabetic rats. Wet kidney weight is an
accurate measure of renal hypertrophy during diabetes (43, 50). Kidney weight was
significantly increased in Diab rats (day 20) compared to Cont (Fig. 4). TNFR:Fc did not
affect wet kidney weight in Cont+TNFR:Fc rats. In contrast, TNFR:Fc therapy significantly
reduced renal hypertrophy in Diab+TNFR:Fc rats (Fig. 4).
Na Retention and Renal Hypertrophy Precede Albuminuria. The development of
microalbuminuria, defined as urinary albumin/creatinine ratio of 30-300mg/g (24), is used as
a marker of glomerular dysfunction during diabetes (24). Albuminuria is the earliest clinical
marker of diabetic nephropathy and albuminuria correlates with progression to end-stage
renal disease (24, 33) and cardiovascular morbidity and mortality (5, 11, 39). We measured
urinary albumin/creatinine ratios at 10 and 20 days after the onset of diabetes to determine if
Na retention and renal hypertrophy precede albuminuria during diabetes. Control rats did not
exhibit measurable urinary albumin excretion at 10 or 20 days. None of the diabetic rats
displayed microalbuminuria on day 10, at which point Na retention was established. On day
20, when diabetic rats exhibited significant renal hypertrophy, only two of eight diabetic rats
exhibited microalbuminuria (albumin/creatinine ratios of 191.3 and 146.0 mg/g), whereas the
remaining diabetic rats were normoalbuminuric.
TNF Stimulates Na Uptake in Distal Tubule Cells Isolated from Diabetic Rats. Tubular
structure and function are altered in the early stages of diabetes (31). To determine if TNF
induces Na retention by stimulating tubular Na uptake, we measured Na uptake in PT and DT
13
cells freshly isolated from Cont and Diab rats. Basal Na uptake in PT cells was unchanged
following 10 days of diabetes (Fig. 5A). Acute exposure to either 10 or 100ng/mL TNF did
not stimulate Na uptake in PT cells isolated from Cont or Diab rats (Fig. 5A).
DT function is important for the regulation of Na homeostasis and is altered during
diabetes (31, 35). We confirmed DT tubule dysfunction in our animal model of diabetes by
measuring urinary THP excretion by ELISA. Diab rats exhibited a significant increase in
THP excretion from day 3 through day 20 of the study period (data not shown). Since DT
function was altered early in diabetes, we hypothesized that TNF induced Na retention by
stimulating Na transport in DT cells. To test this hypothesis, we isolated DT cells from
control rats or after 10 or 20 days of diabetes and quantified TNF-stimulated Na uptake.
Basal rates of Na transport were unaltered in DT cells following 10 or 20 days of diabetes
(Fig. 5A; Cont basal rate, 256.6 ± 5.7 nmols min-1 mg protein-1). TNF had no effect on Na
transport in DT cells from control rats. In contrast, TNF stimulated Na uptake in DT cells
isolated from diabetic rats at day 10 and 20. This differential activation of Na transport by
TNF occurred over a wide concentration range. 10ng/mL TNF stimulated Na uptake in Diab
rat DT cells, whereas DT cells from control rats responded minimally to TNF concentrations
up to 1µg/mL (Fig. 5B).
Expression and Cellular Localization of TNF receptors. TNF modulates cellular function
through distinct membrane receptors, type I (TNFR1) and type II (TNFR2) TNF receptors
(see Rothe et al. for review (38)). To examine the expression and cellular localization of
TNF receptors in DT cells, immunofluorescence and confocal studies were conducted with
antibodies specific for TNFR1 and TNFR2 (7). In immunofluorescence studies using nonpolarized DT cells grown on glass coverslips, both TNFR1 and TNFR2 were observed in DT
cells isolated from control rats (Fig. 6). TNFR1 and TNFR2 were also expressed in DT cells
isolated from diabetic rats (data not shown). TNF receptor localization was examined by
14
confocal microscopy using DT cells grown on filters to maintain polarity of apical and basal
membranes. DT cells are polarized in vivo, such that the apical membrane contacts the urine
and the basal membrane faces the blood. In confocal images, TNFR1 was clearly localized to
the apical portion of DT cells, whereas TNFR2 was expressed diffusely throughout the cell
(Fig. 6). Membrane localization of either TNF receptor subtype was not altered during
diabetes (data not shown).
15
DISCUSSION:
The focus of this study was to elucidate the role of TNF in the development of early
diabetic nephropathy. Our results implicate TNF as an important factor underlying diabetic
nephropathy. Diabetic rats exhibited enhanced urinary TNF excretion, Na retention, and
renal hypertrophy. Administration of the TNF antagonist TNFR:Fc significantly decreased
urinary TNF and prevented Na retention and renal hypertrophy. These observations imply
that enhanced TNF production and release is necessary to induce Na retention and renal
hypertrophy in diabetes.
The mechanism of TNF-induced Na retention during diabetes is presently unclear,
although several findings suggest that augmented tubular Na transport underlies Na retention
in response to elevated TNF. First, TNF directly stimulated Na uptake in DT cells isolated
from diabetic rats. Second, TNFR:Fc treatment augmented urinary Na output during diabetes
compared to Diab rats, despite equivalent Na and water intake and urine output, implying that
TNFR:Fc inhibited tubular Na transport. Finally, TNF was undetectable in serum samples
from any of the groups, whereas urinary TNF was markedly increased in Diab rats,
suggesting that TNF actions were mediated from the lumen of the nephron. Taken together,
these findings implicate enhanced tubular Na transport as the mechanism underlying TNFinduced Na retention.
The role of DT cells underlying Na retention during diabetes is intriguing. The data
we provide shows that DT cells are sensitized to the acute effects of TNF on Na transport
during diabetes. TNF does not activate Na uptake in DT cells isolated from control rats, but
TNF significantly increases Na transport in DT cells from diabetic rats. The factor(s) leading
to DT sensitization during diabetes are presently unclear, but may include increased urinary
TNF or glucose excretion. DT sensitization is explored in detail in the accompanying
manuscript. Second, DT cells normally reabsorb only a small percentage of the filtered Na
16
load, but are critical for the fine regulation of Na homeostasis. Our findings show that
perturbations in DT Na transport may induce profound changes in whole animal Na retention.
However, TNFR:Fc treatment does not fully reverse Na retention in diabetic rats, which
suggests that additional mechanisms may in part contribute to Na retention. The ornithine
decarboxylase mediated increase in PT Na reabsorption during diabetes may contribute to net
Na retention (43). Alternatively, elevated urinary glucose concentrations may enhance
proximal tubule Na reabsorption via constitutively expressed Na-glucose transporters.
TNF contributes to both Na retention and renal hypertrophy in the early stages of
diabetes; however, the mechanisms for TNF actions have not been determined. Our
experiments do not fully differentiate between renal hypertrophy induced by direct TNF
actions and effects secondary to Na retention. High Na feeding can induce renal TGF-β
expression and hypertrophy in rats (48). In theory, the hypertrophic effects of TNF during
diabetes could be mediated by Na retention, with subsequent TGF-β effects. Experiments to
inhibit Na retention without affecting urinary TNF excretion in diabetes are needed to
distinguish between these possibilities. At the cellular level, TNF-stimulated Na uptake
could be mediated by bumetanide-sensitive Na-K-2Cl cotransporters in TAL cells, or by
amiloride-sensitive epithelial Na channels (ENaC) or thiazide-inhibitable Na-Cl
cotransporters in DCT cells. In other studies, we show that TNF-stimulated Na uptake was
inhibited by amiloride, implicating ENaC as the Na transport mechanism underlying TNFstimulated Na uptake in DT cells (companion manuscript).
Elevated urinary TNF correlated with Na retention and renal hypertrophy, and
reduction of urinary TNF with TNFR:Fc therapy ameliorated both Na retention and
hypertrophy. Serum TNF was undetectable in all of the treatment groups, implicating urinary
TNF in these early renal changes during diabetes. The finding that TNF receptors localize to
the apical portion of DT cells further supports the theory that urinary TNF regulates renal
17
function during diabetes. Although urinary TNF appears to be a critical mediator of diabetic
nephropathy, the source of urinary TNF remains to be elucidated. Nakamura et al.
demonstrate increased TNF mRNA in glomeruli during diabetes (26). Proximal tubule (PT)
cells are also capable of producing TNF (19, 20). TNF secretion from PT segments could
subsequently affect DT function in a paracrine manner. TNF could also be produced in DT
cells and act upon nearby DT cells in an autocrine fashion. PT cells exhibited enhanced TNF
mRNA and protein expression following 10 days of diabetes in preliminary experiments in
our laboratory, whereas DT TNF mRNA and protein expression was unchanged (unpublished
observations). These findings implicate the PT as an important source of urinary TNF and a
paracrine regulator of DT function during diabetes. We propose that therapies targeted at
reducing PT TNF production should decrease urinary TNF excretion and replicate the
therapeutic effects of TNFR:Fc. Pentoxifylline, an agent that inhibits TNF synthesis, may be
effective in this regard.
Our results obtained using an animal model of type I diabetes correlate well with
clinical observations of Na retention (29, 45) and renal hypertrophy (4, 21) in diabetic
patients preceding the onset of microalbuminuria. Navarro et al. implicate TNF in the
development of nephropathy in diabetic patients by demonstrating that the TNF inhibitor
pentoxifylline reduces albuminuria (28). Pentoxifylline also decreases microalbuminuria and
overt proteinuria in type II diabetic patients (17). The presence of urinary albumin is
currently the earliest clinical marker of renal dysfunction in diabetic patients (24). Our
finding that urinary TNF excretion was significantly increased early in diabetes, prior to the
development of microalbuminuria, suggests that TNF may be a useful marker of renal
dysfunction during diabetes. Elevated urinary TNF may be a clinically relevant marker for
early detection of nephropathy, based upon evidence we provide that TNF participates in
18
renal hypertrophy, which is a predictor of the severity of renal dysfunction in diabetic
patients (4).
In summary, these findings provide novel information regarding the role of TNF in
early diabetic nephropathy by examining the effects of TNF on isolated kidney cells and
utilizing a specific TNF inhibitor (TNFR:Fc) in vivo. We demonstrate that TNF is a critical
factor underlying the early pathological changes during diabetic nephropathy, including Na
retention and renal hypertrophy. Urinary TNF excretion is increased during diabetes, and
reduction of urinary TNF by administration of TNFR:Fc prevents Na retention and
hypertrophy. Based on these findings, urinary TNF excretion is likely a valuable diagnostic
marker for renal dysfunction during the early stages of diabetes. Our finding that increased
urine TNF excretion precedes proteinuria is particularly important since proteinuria is
currently the earliest clinical marker of renal dysfunction during diabetes. Furthermore,
inhibition of TNF may represent a novel therapeutic target for preventing the progression of
diabetic nephropathy.
19
ACKNOWLEDGEMENTS:
This work was supported by NIH Grants DK07301 and DK57673, and the Joseph Shankman
Award from the National Kidney Foundation of Massachusetts, Rhode Island, New
Hampshire and Vermont. Additional support was provided through the Hitchcock
Foundation and by the Albert J. Ryan Foundation. We thank Dr. Bruce Stanton, director of
the Dartmouth Cystic Fibrosis Cell Biology/Cell Culture Core (STANTO97RO), for
scientific and technical support. We also thank Kathleen Picha (Immunex Corporation,
Seattle, WA) for helpful discussions and for providing TNFR:Fc.
20
REFERENCES:
1. The effect of intensive treatment of diabetes on the development and progression of longterm complications in insulin-dependent diabetes mellitus. The Diabetes Control and
Complications Trial Research Group. N Engl J Med 329: 977-86., 1993.
2. Allen, T. J., M. J. Waldron, D. Casley, G. Jerums, and M. E. Cooper. Salt restriction
reduces hyperfiltration, renal enlargement, and albuminuria in experimental diabetes.
Diabetes 46: 19-24., 1997.
3. Argiles, J. M., J. Lopez-Soriano, and F. J. Lopez-Soriano. Cytokines and diabetes: the
final step? Involvement of TNF-alpha in both type I and II diabetes mellitus. Horm Metab
Res 26: 447-9, 1994.
4. Baumgartl, H. J., G. Sigl, P. Banholzer, M. Haslbeck, and E. Standl. On the prognosis of
IDDM patients with large kidneys. Nephrol Dial Transplant 13: 630-4., 1998.
5. Borch-Johnsen, K., B. Feldt-Rasmussen, S. Strandgaard, M. Schroll, and J. S. Jensen.
Urinary albumin excretion. An independent predictor of ischemic heart disease.
Arterioscler Thromb Vasc Biol 19: 1992-7., 1999.
6. Choy, E. H., and G. S. Panayi. Cytokine pathways and joint inflammation in rheumatoid
arthritis. N Engl J Med 344: 907-16., 2001.
7. Cottin, V., A. A. Van Linden, and D. W. Riches. Phosphorylation of the tumor necrosis
factor receptor CD120a (p55) recruits Bcl-2 and protects against apoptosis. J Biol Chem
276: 17252-60., 2001.
8. Cottone, S., A. Vadala, M. C. Vella, G. Mule, A. Contorno, and G. Cerasola. Comparison
of tumour necrosis factor and endothelin-1 between essential and renal hypertensive
patients. J Hum Hypertens 12: 351-4, 1998.
9. Foss, M. C., N. T. Foss, G. M. Paccola, and C. L. Silva. Serum levels of tumor necrosis
factor in insulin-dependent diabetic patients. Braz J Med Biol Res 25: 239-42, 1992.
10. Friedman, P. A., and F. A. Gesek. Cellular calcium transport in renal epithelia:
measurement, mechanisms, and regulation. Physiol Rev 75: 429-71, 1995.
11. Gerstein, H. C., J. F. Mann, Q. Yi, B. Zinman, S. F. Dinneen, B. Hoogwerf, J. P. Halle, J.
Young, A. Rashkow, C. Joyce, S. Nawaz, and S. Yusuf. Albuminuria and risk of
cardiovascular events, death, and heart failure in diabetic and nondiabetic individuals.
Jama 286: 421-6., 2001.
12. Gesek, F. A. Alpha1- and alpha2-adrenoceptor control of sodium transport reverses in
developing hypertension. Hypertension 33: 524-9., 1999.
13. Gesek, F. A., and P. A. Friedman. On the mechanism of parathyroid hormone stimulation
of calcium uptake by mouse distal convoluted tubule cells. J Clin Invest 90: 749-58,
1992.
14. Gesek, F. A., and P. A. Friedman. Sodium entry mechanisms in distal convoluted tubule
cells. Am J Physiol 268: F89-98., 1995.
15. Gesek, F. A., and A. C. Schoolwerth. Hormonal interactions with the proximal Na(+)-H+
exchanger. Am J Physiol 258: F514-21., 1990.
16. Gesek, F. A., D. W. Wolff, and J. W. Strandhoy. Improved separation method for rat
proximal and distal renal tubules. Am J Physiol 253: F358-65, 1987.
17. Guerrero-Romero, F., M. Rodriguez-Moran, J. R. Paniagua-Sierra, G. Garcia-Bulnes, M.
Salas-Ramirez, and D. Amato. Pentoxifylline reduces proteinuria in insulin-dependent
and non insulin- dependent diabetic patients. Clin Nephrol 43: 116-21., 1995.
18. Hasegawa, G., K. Nakano, M. Sawada, K. Uno, Y. Shibayama, K. Ienaga, and M. Kondo.
Possible role of tumor necrosis factor and interleukin-1 in the development of diabetic
nephropathy. Kidney Int 40: 1007-12, 1991.
21
19. Jevnikar, A. M., D. C. Brennan, G. G. Singer, J. E. Heng, W. Maslinski, R. P. Wuthrich,
L. H. Glimcher, and V. E. Kelley. Stimulated kidney tubular epithelial cells express
membrane associated and secreted TNF alpha. Kidney Int 40: 203-11, 1991.
20. Jevnikar, A. M., R. P. Wuthrich, D. C. Brennan, W. Maslinski, L. H. Glimcher, and V. E.
Rubin-Kelley. TNF-alpha is expressed on the surface of kidney proximal tubular cells.
Transplant Proc 23: 231-2, 1991.
21. Kleinman, K. S., and L. G. Fine. Prognostic implications of renal hypertrophy in diabetes
mellitus. Diabetes Metab Rev 4: 179-89., 1988.
22. Lechleitner, M., T. Koch, M. Herold, and F. Hoppichler. Relationship of tumor necrosis
factor-alpha plasma levels to metabolic control in type 1 diabetes [letter]. Diabetes Care
22: 1749, 1999.
23. Mogensen, C. E., C. K. Christensen, and E. Vittinghus. The stages in diabetic renal
disease. With emphasis on the stage of incipient diabetic nephropathy. Diabetes 32 Suppl
2: 64-78., 1983.
24. Molitch, M. E., R. A. DeFronzo, M. A. Franz, W. F. Keane, C. E. Mogensen, H. Parving,
and M. W. Steffes. Diabetic Nephropathy. Diabetes Care 23: S69-S73, 2000.
25. Muc-Wierzgon, M., M. Baranowski, and K. Madej. [Monitoring of changes in levels of
endogenous TNF alpha in patients with uncontrolled insulin dependent diabetes mellitus].
Pol Arch Med Wewn 97: 426-32, 1997.
26. Nakamura, T., M. Fukui, I. Ebihara, S. Osada, I. Nagaoka, Y. Tomino, and H. Koide.
mRNA expression of growth factors in glomeruli from diabetic rats. Diabetes 42: 450-6,
1993.
27. Navarro, J. F., and C. Mora. Antiproteinuric effect of pentoxifylline in patients with
diabetic nephropathy. Diabetes Care 22: 1006-8., 1999.
28. Navarro, J. F., C. Mora, A. Rivero, E. Gallego, J. Chahin, M. Macia, M. L. Mendez, and
J. Garcia. Urinary protein excretion and serum tumor necrosis factor in diabetic patients
with advanced renal failure: effects of pentoxifylline administration. Am J Kidney Dis 33:
458-63, 1999.
29. O'Hare, J. A., J. B. Ferriss, D. Brady, B. Twomey, and D. J. O'Sullivan. Exchangeable
sodium and renin in hypertensive diabetic patients with and without nephropathy.
Hypertension 7: II43-8, 1985.
30. Peach, R. J., W. A. Day, P. J. Ellingsen, and A. R. McGiven. Ultrastructural localization
of Tamm-Horsfall protein in human kidney using immunogold electron microscopy.
Histochem J 20: 156-64., 1988.
31. Pfleiderer, S., L. B. Zimmerhackl, R. Kinne, F. Manz, G. Schuler, and M. Brandis. Renal
proximal and distal tubular function is attenuated in diabetes mellitus type 1 as
determined by the renal excretion of alpha 1- microglobulin and Tamm-Horsfall protein.
Clin Investig 71: 972-7, 1993.
32. Pizzonia, J. H., F. A. Gesek, S. M. Kennedy, B. A. Coutermarsh, B. J. Bacskai, and P. A.
Friedman. Immunomagnetic separation, primary culture, and characterization of cortical
thick ascending limb plus distal convoluted tubule cells from mouse kidney. In Vitro Cell
Dev Biol 27A: 409-16, 1991.
33. Rachmani, R., Z. Levi, M. Lidar, I. Slavachevski, E. Half-Onn, and M. Ravid.
Considerations about the threshold value of microalbuminuria in patients with diabetes
mellitus: lessons from an 8-year follow-up study of 599 patients. Diabetes Res Clin Pract
49: 187-94., 2000.
34. Rasch, R., and J. Dorup. Quantitative morphology of the rat kidney during diabetes
mellitus and insulin treatment. Diabetologia 40: 802-9, 1997.
22
35. Rasch, R., O. Torffvit, S. Bachmann, P. K. Jensen, and N. O. Jacobsen. Tamm-Horsfall
glycoprotein in streptozotocin diabetic rats: a study of kidney in situ hybridization,
immunohistochemistry, and urinary excretion. Diabetologia 38: 525-35, 1995.
36. Reeves, W. B., and T. E. Andreoli. Transforming growth factor beta contributes to
progressive diabetic nephropathy. Proc Natl Acad Sci U S A 97: 7667-9., 2000.
37. Reichard, P., B. Y. Nilsson, and U. Rosenqvist. The effect of long-term intensified insulin
treatment on the development of microvascular complications of diabetes mellitus. N
Engl J Med 329: 304-9., 1993.
38. Rothe, J., G. Gehr, H. Loetscher, and W. Lesslauer. Tumor necrosis factor receptorsstructure and function. Immunol Res 11: 81-90, 1992.
39. Sato, A., L. Tarnow, and H. H. Parving. Prevalence of left ventricular hypertrophy in
Type I diabetic patients with diabetic nephropathy. Diabetologia 42: 76-80., 1999.
40. Sikri, K. L., D. P. Alexander, and C. L. Foster. Localization of Tamm-Horsfall
glycoprotein in the normal rat kidney and the effect of adrenalectomy on its localization
in the hamster and rat kidney. J Anat 135: 29-45., 1982.
41. Strojek, K., W. Grzeszczak, B. Lacka, J. Gorska, C. K. Keller, and E. Ritz. Increased
prevalence of salt sensitivity of blood pressure in IDDM with and without
microalbuminuria. Diabetologia 38: 1443-8., 1995.
42. Sugimoto, H., K. Shikata, J. Wada, S. Horiuchi, and H. Makino. Advanced glycation end
products-cytokine-nitric oxide sequence pathway in the development of diabetic
nephropathy: aminoguanidine ameliorates the overexpression of tumour necrosis factoralpha and inducible nitric oxide synthase in diabetic rat glomeruli. Diabetologia 42: 87886, 1999.
43. Thomson, S. C., A. Deng, D. Bao, J. Satriano, R. C. Blantz, and V. Vallon. Ornithine
decarboxylase, kidney size, and the tubular hypothesis of glomerular hyperfiltration in
experimental diabetes. J Clin Invest 107: 217-24., 2001.
44. Weder, A. B. Sodium metabolism, hypertension, and diabetes. Am J Med Sci 307 Suppl 1:
S53-9., 1994.
45. Weidmann, P., C. Beretta-Piccoli, and B. N. Trost. Pressor factors and responsiveness in
hypertension accompanying diabetes mellitus. Hypertension 7: II33-42, 1985.
46. Weidmann, P., L. M. Boehlen, and M. de Courten. Pathogenesis and treatment of
hypertension associated with diabetes mellitus. Am Heart J 125: 1498-513., 1993.
47. Yong, L. C., and A. F. Bleasel. Pathological changes in streptozotocin induced diabetes
mellitus in the rat. Exp Pathol 30: 97-107, 1986.
48. Yu, H. C., L. M. Burrell, M. J. Black, L. L. Wu, R. J. Dilley, M. E. Cooper, and C. I.
Johnston. Salt induces myocardial and renal fibrosis in normotensive and hypertensive
rats. Circulation 98: 2621-8., 1998.
49. Zhou, M. S., Y. Nishida, H. Yoneyama, Q. H. Chen, and H. Kosaka. Potassium
supplementation increases sodium excretion and nitric oxide production in hypertensive
Dahl rats. Clin Exp Hypertens 21: 1397-411., 1999.
50. Ziyadeh, F. N., B. B. Hoffman, D. C. Han, M. C. Iglesias-De La Cruz, S. W. Hong, M.
Isono, S. Chen, T. A. McGowan, and K. Sharma. Long-term prevention of renal
insufficiency, excess matrix gene expression, and glomerular mesangial matrix expansion
by treatment with monoclonal antitransforming growth factor-beta antibody in db/db
diabetic mice. Proc Natl Acad Sci U S A 97: 8015-20., 2000.
23
FIGURE LEGENDS:
Figure 1. TNFR:Fc reduces urinary TNF excretion during diabetes. Urinary TNF was
detected by ELISA and expressed as total TNF per 24 hours to account for marked
differences in urine volume between control and diabetic rats. Symbols denote mean ± SEM
of n=3 (Cont and Cont+TNFR:Fc), n=8 (Diab), or n=9 (Diab+TNFR:Fc) rats.
*, P<0.05 versus Diab.
Figure 2. TNFR:Fc does not affect metabolic parameters during diabetes. Effect of
TNFR:Fc treatment on urine glucose concentration (A), ketonuria (B), water intake (C), urine
output (D), food intake (E), and body weight (F). Body weight is expressed as a percentage
of body weight on day 0 (baseline). Symbols denote mean ± SEM of n=3 (Cont and
Cont+TNFR:Fc), n=8 (Diab), or n=9 (Diab+TNFR:Fc) rats.
*, P<0.05 versus Diab.
Figure 3. TNFR:Fc augments urinary Na output and reduces Na retention during
diabetes. Effect of TNFR:Fc treatment on dietary Na intake (A) and urinary Na output (B) in
control and diabetic rats. C. Effect of TNFR:Fc therapy on Na balance in control and
diabetic rats. Na balance was calculated as the difference between dietary Na intake and
urinary Na output for individual rats on each day. Symbols denote mean ± SEM of n=3
(Cont and Cont+TNFR:Fc), n=8 (Diab), or n=9 (Diab+TNFR:Fc) rats.
*, P<0.05 versus Diab.
Figure 4. TNFR:Fc prevents renal hypertrophy during diabetes. Effect of TNFR:Fc
treatment on wet kidney weight in control and diabetic (day 20) rats. Bars denote mean ±
SEM of n=3 (Cont and Cont+TNFR:Fc), n=8 (Diab), or n=9 (Diab+TNFR:Fc) rats.
*, P<0.05 versus Cont.
Figure 5. TNF stimulates Na uptake selectively in DT cells during diabetes. A. Active
Na transport in PT cells isolated from control or diabetic (day 10) rats. Cells were exposed to
TNF for 15 minutes. Bars denote mean ± SEM of triplicate determinations of each
experimental condition for n=3 rats. B. Basal and TNF-stimulated (10ng/mL; 15 min.) Na
uptake in DT cells isolated from control or diabetic rats. C. TNF-stimulated Na uptake is
selective in DT cells from diabetic rats over a wide dose range. Bars (B) and symbols (C)
denote mean ± SEM for triplicate determinations of each experimental condition of n=5
(Diab 20d) or n=6 (Cont and Diab 10d) rats.
#
, P<0.05 versus basal. *, P<0.05 versus Cont.
Figure 6. DT cells express TNF receptors. TNF receptors visualized by
immunofluorescence (top panels) and confocal microscopy (bottom panels). Non-polarized
cells were grown on glass coverslips for immunofluorescence images. Confocal images
shown represent Z-sections through polarized cells grown on filters.
Figure 1
Cont
Diab
50000
Diab+TNFR:Fc
40000
30000
20000
10000
*
*
*
*
*
*
13
17
20
0
0
3
7
10
Days
Figure 2
A
Cont
Diab + TNFR:Fc
Diab
Cont + TNFR:Fc
B
2000
1600
160
120
1200
80
800
400
0
0
0
C
3
7
10 13
Days
17
20
D
300
250
200
200
150
150
100
100
50
50
0
0
3
7
10 13
Days
17
20
F
60
0
3
7
10 13
Days
17
20
0
3
7
10 13
Days
17
20
0
3
7
10 13
Days
17
20
300
250
0
E
*
40
130
120
50
110
40
100
30
90
20
80
0
3
7
10 13
Days
17
20
Figure 3
A
Cont
Diab + TNFR:Fc
Diab
Cont + TNFR:Fc
8
7
6
5
4
3
2
1
B
0
3
7
8
10
Days
13
17
20
7
*
*
*
*
6
5
4
3
* *
2
1
C
0
3
7
5
10
Days
13
17
20
4
3
*
2
*
*
1
0
3
7
10
Days
13
17
20
Figure 4
3.0
*
2.5
2.0
1.5
0.0
Cont
Diab
Diab + TNFR:Fc Cont + TNFR:Fc
Figure 5
A
12
Cont
Diab 10d
10
8
0
basal
TNF (10ng/mL)
TNF (100ng/mL)
B
450
Cont
#
Diab 10d
400
#
Diab 20d
350
300
250
0
basal
C
450
TNF
Cont
400
Diab 10d
*
*
*
100
1000
350
300
250
200
0
0.01
0.1
1
10
[TNF] (ng/mL)
Figure 6