| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Nephrology, Clinical Sciences, Lund University, Lund, Sweden
Submitted 12 June 2007 ; accepted in final form 14 August 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
20 mM for 3 or 9 wk, were compared with age-matched controls. Glomerular clearances of native albumin (Cl-HSA) and neutralized albumin (Cl-nHSA) were assessed using a renal uptake technique. Glomerular filtration rate and renal plasma flow were assessed using 51Cr-EDTA and [125I]iodohippurate, respectively. In a separate set of animals, diabetic for 9 wk, and in controls, glomerular sieving coefficients (
) for neutral FITC-Ficoll (molecular radius: 15–90 Å) were assessed using size exclusion chromatography. At 3 wk of diabetes, Cl-HSA and Cl-nHSA remained unchanged, indicating no alteration in either size or charge selectivity. By contrast, at 9 wk of diabetes, there was a twofold increase of Cl-HSA, whereas Cl-nHSA remained largely unchanged, at first suggesting a glomerular charge defect. However, according to a two-pore model, the number of large pores, assessed from both Ficoll and Cl-HSA, increased twofold. In addition, a small reduction in proximal tubular reabsorption was observed at 3 wk, which was further reduced at 9 wk. In conclusion, no functional changes were observed in the glomerular filtration barrier at 3 wk of STZ-induced diabetes, whereas at 9 wk there was a decrease in size selectivity due to an increased number of large glomerular pores. sieving coefficient; proteinuria; capillary permeability; fractional clearance; macromolecules; diabetic nephropathy
By contrast, late DNP is characterized by a prominent, nonselective proteinuria that develops as a result of gradual deterioration of the glomerular barrier. The structural changes occurring during the development of early to late DNP involve thickening of the glomerular and tubular basement membranes and, later, a decreased filtration surface area combined with loss of podocytes, which is furthermore accompanied by a clearly reduced size selectivity of the glomerular filter (15, 16, 38). To date, the permselective properties of glomerular barrier in diabetic nephropathy have mainly been assessed by using dextran as a probe for glomerular permeability. It is now well established, however, that dextran is hyperpermeable across the glomerular filter. Therefore, it is insensitive to small changes in the glomerular size selectivity. Only in advanced DNP, i.e., during the macroalbuminuric phase, have increases in the large-pore/shunt pathway been detected using dextran (7, 14, 24, 31). Recently, Ficoll, a polysaccharide that is less hyperpermeable across the glomerular filter than dextran, was employed as a marker for glomerular permeability in patients with early DNP and microalbuminuria (1). With the use of this less hyperpermeable macromolecular probe, an increase in the large-pore pathway was indeed detected (1), implying that a decreased size selectivity is responsible for the increased permeability to albumin in early DNP. Furthermore, in a recent careful micropuncture study, albuminuria in early streptozotocin (STZ)-induced diabetes was found to be due to a reduced proximal tubular reabsorption of albumin (29).
The present study was performed to investigate the functional alterations of the glomerular barrier occurring in early DNP. The aim was to clarify whether the major early injury could be ascribed to size- or to charge-selective changes in the filter, or perhaps to both. This was accomplished by studying rats exposed to hyperglycemia in poorly treated STZ-induced diabetes for either 3 or 9 wk. Glomerular size and charge selectivity were assessed in vivo using two different approaches. The glomerular clearances of native (human serum albumin, HSA) and neutralized albumin (nHSA) were measured using a renal tissue uptake technique. Furthermore, after 9 wk of diabetes, when a perturbation in the albumin transport was detected, glomerular sieving coefficients () for Ficoll molecules with a broad size distribution (radius: 15–90 Å) were measured. This approach enabled us to provide, for the first time, a precise evaluation of both charge and size selectivity in the STZ-diabetic model in vivo.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Diabetic animals.
Forty-nine animals, diabetic for either 3 or 9 wk, were purchased from Møllegaard, arriving a few days before the experiments. Briefly, rats weighing between 140 and 160 g were made diabetic by a single intravenous injection of STZ [N-(methylnitroscocarbamolyl)-
-D-glucosamine, 90 mg/kg; Biochimika, Sigma-Aldrich Denmark]. Blood glucose levels were measured every morning. When glucose was detected in the urine and the plasma concentration of glucose was >25 mM, the animals were considered diabetic. The rats received daily subcutaneous injections using 0.5 IU insulin (porcine insulin, 40 UI/ml; Caninsulin, Intervet, Skovlunde, Denmark). Rats with a plasma glucose concentration higher than 17 mM were given 0.2 IU more insulin than the day before, whereas rats with a glucose concentration lower than 13 mM were given 0.2 IU less than the day before. The aim was to keep the animals at a blood glucose level between 18 and 25 mM. After either 3 or 9 wk of diabetes, glomerular filtration rate (GFR), renal plasma flow (RPF), and glomerular permselectivity were determined.
Surgery. The rats were anesthetized intraperitoneally using 60 mg/kg pentobarbital sodium and placed on a heating pad to maintain body temperature at 37°C. The tail artery was cannulated (PE-50 cannula) for arterial pressure recordings on a polygraph (model 7B; Grass Instruments, Quincy, MA) and for the administration of drugs. A tracheotomy was performed using a PE-240 tube. The left carotid artery and left jugular vein were cannulated (PE-50) for blood sampling and infusions, respectively. After surgery, the animal was allowed to recover for at least 30 min.
Glomerular filtration rate and renal plasma flow. GFR and RPF were determined at 3 (n = 10) and 9 wk (n = 8) of diabetes (D-3w and D-9w, respectively) and in their respective control groups [C-3w (n = 4) and C-9w (n = 6)]. A catheter was placed in the urinary bladder via an abdominal incision for urine collection. 51Cr-EDTA (0.37 MBq; Amersham Biosciences, Little Chalfont, UK) and [125I]iodohippurate (0.08 MBq; Amersham Biosciences) were given together as a bolus dose, immediately followed by a constant infusion (3 ml/h) of the respective tracer (0.37 MBq/ml 51Cr-EDTA and 0.08 MBq/ml [125I]iodohippurate in 0.9% NaCl). Five blood samples were collected from the carotid artery during a 20-min period. During the same period, urine was collected. After the rats were volume loaded with 2 ml of horse serum (SVA, Uppsala, Sweden), GFR and RPF were assessed for another 20-min period. At the end of the experiment a sample was collected from the renal vein, and the extraction fraction of [125I]iodohippurate was calculated and used for determination of its clearance.
Clearance of neutralized albumin and native albumin.
Clearance of nHSA [Stokes-Einstein (SE) radius: 35.0 Å] and native (negatively charged) HSA (SE radius: 35.5 Å) was measured at 3 (n = 8) and 9 wk of diabetes (n = 10) and in the respective control groups (n = 6 and n = 8) by using a tissue uptake technique as described in detail elsewhere (13). Briefly, 125I-HSA (0.2 MBq; Institute for Energy Technique, Kjeller, Norway) was administered as a bolus dose together with 131I-nHSA (0.15 MBq) in the tail artery. nHSA was prepared by Dr. Olav Tenstad (University of Bergen, Bergen, Norway) by a graded modification of the COOH groups and was labeled with 131I by using 1,3,4,6-tetrachloro-3,6-diphenylglycouril (Iodo-Gen) as described at some length previously (13). Six blood samples (25 µl) and one urine sample were collected during an 8-min period. To eliminate the tracer from the renal vasculature, a whole body washout (using a 1:2 mixture of 0.9% saline and heparinized horse serum) was performed via the carotid artery (20 ml/min) for 8 min, after the inferior vena cava had been freed and cut open for collection of the rinse fluid. The kidneys were dissected free, and the cortex and the urine sample were assessed with respect to radioactivity. Urine samples were precipitated using trichloroacetic acid (TCA), and the amount of free iodine was calculated. Clearance of albumin (nHSA or HSA) was calculated as the cortical tracer mass plus the precipitable urine mass of tracer divided by the average plasma tracer concentration and time. The fractional tubular albumin excretion was obtained from the precipitable urine mass divided by the total mass of albumin recovered in the urine and the renal cortex, and the fractional (proximal) tubular albumin reabsorption was obtained from 1 minus this entity.
Sieving of FITC-Ficoll.
A separate set of diabetic rats (n = 13) and controls (n = 8) was used for assessing for Ficoll at 9 wk. A mixture of fluorescein isothiocyanate (FITC)-labeled Ficoll-400 (1 mg) and Ficoll-70 (42 µg) (TdB Consultancy, Uppsala, Sweden) was administered as a bolus dose together with FITC-inulin (0.5 µg) and 51Cr-EDTA (0.37 MBq). The bolus was followed by a constant infusion of 3 ml/h (Ficoll-70, 94.5 µg/min; Ficoll-400, 3 mg/min; inulin, 1.5 µg/min; 51Cr-EDTA, 0.019 MBq/min). Twenty minutes after Ficoll administration, a laparotomy was performed and catheters (PE-10 coupled to PE-50) were placed in the left and the right ureter and used for urine collection. Urine was collected during a 5-min period during which one midpoint (2.5 min) blood sample was taken. Plasma and urine were assessed on a size exclusion high-performance chromatography system (Waters) with an Ultrahydrogel 500 column (Waters) and calibrated as described in detail previously (2).
Urinary albumin/creatinine concentration ratio. The albumin/creatinine concentration ratios (ACR) were obtained by sampling urine in metabolic cages for 4 h two days before the experiments. The albumin concentration in urine was assessed using a simplified (one-step incubation) enzyme-linked immunosorbent assay (ELISA), as described at some length previously (19, 33). Briefly, the plates were coated with rat albumin (Sigma) and then incubated with the urine sample, a clonal rabbit anti-rat antiserum (diluted 1:2,000; Nordic Immunological Laboratories, Tillberg, The Netherlands), and an anti-rabbit IgG conjugated with alkaline phosphatase. The detection limit was 16 µg/l, and the intra- versus interassay variations were 11.8 and 12.8%, respectively. Urinary creatinine was analyzed using the Jaffé reaction.
Statistics. Statistical significance among groups was tested using a one-way ANOVA with the significance level set to P < 0.05. Comparison among means was performed using Bonferroni's correction.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
|
for HSA and nHSA.
The clearance of native albumin (HSA) and neutralized albumin (nHSA) did not increase at all after 3 wk of diabetes compared with the control group. However, after 9 wk of diabetes duration the clearance of HSA was significantly increased (1.8-fold) compared with control (7.21 ± 0.315 x10–4 vs. 4.06 ± 0.714 x10–4 ml/min; P < 0.01), indicating an alteration in glomerular barrier characteristics. By contrast, the clearance of nHSA was not significantly different among any of the groups (Table 1). Thus the clearance ratio nHSA/HSA showed a significant decrease at 9 wk of diabetes (P < 0.01) (Fig. 1). Calculating the tubular reabsorption of HSA after 3 wk of diabetes indicates a significant (6.7%) decrease from 95.2 ± 1.1% in the control to 88.8 ± 2.3% (P < 0.05 ) and a 14% reduction at 9 wk of diabetes (to 82 ± 2.2%; P < 0.05). The tubular reabsorption of nHSA was 99.2 ± 0.1% in the control groups and decreased in the diabetic situation to 97.7 ± 0.4% (P < 0.05) and 95.5 ± 0.5% (P < 0.05) after 3 and 9 wk, respectively. for HSA was calculated by dividing the HSA clearance by the corresponding average GFR assessed in the parallel rats used for this purpose (see data above). for HSA was significantly increased only after 9 wk of diabetes (Table 1), whereas for nHSA was slightly, but significantly, reduced at both 3 and 9 wk of diabetes.
for FITC-Ficoll.
To evaluate whether the increase in clearance of native albumin after 9 wk of diabetes was due to a reduced charge selectivity or to an altered number of large pores, we investigated the sieving coefficients for FITC-Ficoll in the molecular range of 15–90Å. Our data show that after 9 wk of diabetes there was indeed a significant increase in for molecules larger than 55Å in radius (Fig. 2).
|
L) was calculated for HSA by using the two-pore model, assuming a filtration pressure of 9 mmHg (to assess the LpS), an osmotic pressure gradient (
) of 28 mmHg, and a large pore of 112 Å (120 Å minus a Debye length of 8 Å for negatively charged HSA). The "apparent" L, calculated from the for HSA, was significantly increased only after 9 wk of diabetes (Fig. 3). The parameters generated from the best fit of the measured Ficoll sieving curves to the two-pore model are shown in Table 2. After 9 wk of STZ-induced diabetes, L had increased twofold compared with that in the control group (Fig. 3), indicating an increased fraction of large pores in the glomerular filter. Furthermore, in the diabetic animals the capillary pore area (A0/x) remained unchanged, whereas the small pore radius (rs) decreased slightly (but significantly) from 46.1 ± 0.06 to 45.3 ± 0.14 Å compared with control (P < 0.05).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
20 mM), a decreased clearance ratio of nHSA to HSA was noticed, at first suggesting a reduction in the amount of negative charges in the glomerular barrier. Furthermore, concomitantly with the decreased nHSA/HSA ratio, we observed an increased clearance of large Ficoll molecules (>55 Å), exactly matching the increased excretion of HSA, suggesting that the main alteration in the glomerular barrier was an increase in the large pore number.
In humans, overt proteinuria due to DNP seldom appears until a decade or so has elapsed from the onset of type 1 diabetes mellitus. The first sign of glomerular dysfunction is microalbuminuria. This early excretion of (negatively charged) albumin has been ascribed mainly to loss of anionic sites of the glomerular filtration barrier, according to the so-called Steno hypothesis. Indeed, a decreased selectivity index, i.e., a reduced ratio between (neutral) IgG or (neutral) 2-macroglobulin and albumin (negatively charged), has been observed in several studies in microalbuminuric patients (4, 12, 32) and seems to support the concept of an alteration in the charge barrier. Furthermore, favoring the Steno hypothesis is the finding of a decreased amount of heparan sulfate (HS) side chains of the glomerular basement membrane (GBM) proteoglycans in patients with DNP (36), a decrease that is correlated with the degree of albuminuria (27). It should be noted, however, that the selectivity index is assessed in urine modified by tubular reabsorption and does not represent the ratio of protein concentrations in the primary urine. Furthermore, although the enzyme heparanase was found to be increased in STZ-induced diabetic rats, this was actually not found to imply a significant reduction in the GBM HS content (37). Furthermore, no change in HS sulfation or charge density has been detected in early STZ-induced diabetes (35). Hence, the contribution of a decreased charge selectivity in the urinary hyperexcretion of albumin in early DNP is controversial.
The early changes occurring in the kidney after the onset of diabetes are nephron hypertrophy and an increase in GFR and RPF. In accordance with this, we found that GFR was increased compared with control in the very early development of STZ-induced DNP (3 wk), both in the basal situation and after volume expansion. At 9 wk of diabetes duration, however, the increment in GFR, both before and after volume expansion, was markedly blunted, in agreement with previous data (17). At 3 wk of diabetes we did not find any increase whatsoever in the glomerular clearance of either native or neutralized albumin compared with control. This indicates that no deterioration of the glomerular barrier function had occurred at this point. However, we found a reduction in tubular reabsorption of both native and neutralized albumin (P < 0.05) after 3 wk of diabetes that was further enhanced at 9 wk of diabetes. Note that a reduction in fractional tubular albumin reabsorption from 95–96 to 89%, as found in the present study, at an unchanged GFR would imply a near threefold increment in albumin excretion, adding to the increases in albumin excretion that would follow upon increments in GFR. Actually, we found a sevenfold increment in rat albuminuria at 3 wk of diabetes duration that, at least partly, may be explained by a combination of reduced proximal tubular albumin reabsorption and an increased GFR. These results are in essential agreement with the findings of Tojo et al. (29), who showed, using a careful micropuncture technique, that the reabsorption of albumin in the proximal tubule was reduced, suggesting a dysfunction in the proximal tubular endocytotic process at early stages of diabetes. After 3 wk of diabetes, the reduction in proximal tubular reabsorption of HSA compared with nHSA was 6.7 vs. 1.6%, respectively, and after 9 wk, 13.9 vs. 3.7%. Hence, we found that the reabsorption of HSA, compared with its neutral counterpart, was reduced. Thus microalbuminuria may be chiefly secondary to alterations in the proximal tubular protein reabsorption, affecting negatively charged albumin more (18) than neutral and larger macromolecules (IgG or 2-macroglobulin). In fact, the renoprotective effects of angiotensin converting enzyme inhibitors and angiotensin II receptor blockers seem not only confined to reductions in glomerular hyperfiltration but also have recently been shown to restore albumin reabsorption in the proximal tubules (30).
The glomerular barrier size selectivity is bimodal in that the glomerular filter can be described as a membrane having two different size-selective pathways (21), "small pores" (rs 37.5 Å) and a low number of "large pores" (rL 120Å, comprising one part per million of the small pores) (13, 28). According to the two-pore concept, native albumin, because of its net negative charge, is excluded from the small pores and normally passes to the urine exclusively through the large pore pathway. If the negative charges in the small pore pathway were to decrease, this would lead to an increased glomerular passage of native albumin. nHSA, on the other hand, because of its lack of net negative charge, passes mainly through the small pores and should be only marginally affected by a change in pore charge. This is because pore charge critically influences the effective pore radius to negatively charged species in the small pores but not to a significant extent in the large pores, according to the Debye-Hückel theory of ion-ion interaction (21). In the present study, the sieving coefficient of nHSA was 10-fold higher than that of HSA. After 9 wk of diabetes duration, the clearance of HSA increased twofold, whereas that of nHSA was unchanged, yielding a reduced nHSA/HSA ratio. This could, in theory, be interpreted as a decrease in the negative charge of the small pores. However, the sieving coefficients for neutral FITC-Ficoll molecules in the SE radius range of >55 Å were significantly increased (Fig. 2) at this point, suggesting an increased macromolecular flux through the large pore pathway, whereas the rL was not significantly different from control. In line with this, the fractional ultrafiltration coefficient LpS accounted for by the large pores (L) was twofold higher after 9 wk of DNP compared with control rats. Calculating an "apparent" L from the HSA clearance data yielded an increase in L almost identical to that obtained with Ficoll (Fig. 3), implying that the main structural change of the filtration barrier in early STZ-induced diabetes must have been an increase in the number of large pores and not primarily a decrease in the negative charge of the glomerular filter.
Among diabetic patients with overt proteinuria, a reduced size selectivity of the glomerular barrier has been commonly noted by using dextran as a macromolecular marker (7, 14, 24, 31). However, in patients with microalbuminuria, clear evidence for a decreased size selectivity comes from one previous study only, where the more sensitive glomerular sieving probe, Ficoll, was employed (1). Dextran has been shown to be hyperpermeable through the glomerular filter compared with either proteins or Ficoll (5, 20). Furthermore, we have shown that Ficoll molecules also appear hyperpermeable through the (small) pores of the glomerular filter when they approach the pore radius in size (SE radius >60% of rs). However, we also found, somewhat surprisingly, that the permeability of Ficoll in the size range of 55–75 Å (SE radius <60% of rL) reflects the sieving of large proteins (of equivalent size) across the large pores (23). Actually, to be able to accurately measure the glomerular passage of very large Ficoll molecules, we introduced a mixture containing high concentrations of high molecular weight Ficoll (Ficoll-400 kDa) and a lower amount of Ficoll-70 kDa molecules. Thus, by enhancing the concentrations of very large Ficoll molecules in both plasma and urine, we have produced a very sensitive method for assessing glomerular sieving of Ficoll molecules >55 Å in radius.
The results from the present study are partly in contrast to recent long-term experiments in which the insulin-dependent diabetic NOD mice were investigated with regard to glomerular permselectivity, also utilizing Ficoll as a glomerular size probe. In these mice, the sieving coefficients for Ficoll (12–70 Å), measured in the cooled isolated perfused kidney (cIPK), did not increase in the diabetic mice after either 10 or 40 wk of diabetes duration (9). At the same time, the clearance of albumin was increased threefold at 40 wk of diabetes, indicating a defect in the charge barrier. However, in the cIPK, the number of large pores is much higher than measured in vivo (10), making this model less sensitive to changes in the large pore number compared with the model used in the present study. As a result, changes in size selectivity to very large macromolecules may have been overlooked in the cIPK.
Podocyte integrity is very important for the glomerular barrier function, because podocytes "embrace" the capillaries and stabilize the whole glomerular barrier. Since the discovery of nephrin in 1998 and the association of the loss of barrier function with the mutation in the nephrin gene (11), the podocyte slit diaphragm (PSD) has been considered to play a major role in the size-selective properties of the glomerular barrier. However, a recent study indicated that the major hindrance to macromolecules is most probably located more proximally to the PSD, i.e., in the endothelium or the GBM (8, 22). Although injury to any of the three layers of the glomerular filter (the endothelial cells, the GBM, or the podocytes) results in loss of glomerular size selectivity, it is important to stress the integrity of all layers acting in concert for the barrier to function properly. In a recent study Siu et al. (26) showed that STZ-diabetic rats with no insulin substitution had a decreased podocyte number already after 2 wk. Treating the rats with insulin for 6 wk resulted in a much milder phenotype, where the podocyte number was not significantly reduced compared with control. It is reasonable to speculate that after 9 wk of poorly controlled diabetes, the podocyte density would be somewhat reduced, which in part could be responsible for a disintegration of the glomerular barrier, reducing its size selectivity.
In conclusion, the present study indicates that in the very early stage of STZ-diabetes, albuminuria is due to a reduced proximal tubule reabsorption and not to functional alterations in the glomerular filtration barrier. However, this stage is followed by an increase in glomerular albumin clearance, primarily resulting from a reduced size selectivity of the glomerular filter.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
) in rats. Kidney Int 69: 1326–1332, 2006.[Web of Science][Medline]This article has been cited by other articles:
|
|
 |
|
  E. Gagliardini, D. Corna, C. Zoja, F. Sangalli, F. Carrara, M. Rossi, S. Conti, D. Rottoli, L. Longaretti, A. Remuzzi, et al. Unlike each drug alone, lisinopril if combined with avosentan promotes regression of renal lesions in experimental diabetes Am J Physiol Renal Physiol, November 1, 2009; 297(5): F1448 - F1456. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Axelsson, I. Mahmutovic, A. Rippe, and B. Rippe Loss of size selectivity of the glomerular filtration barrier in rats following laparotomy and muscle trauma Am J Physiol Renal Physiol, September 1, 2009; 297(3): F577 - F582. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Grande, C. Rippe, A. Rippe, A. Rahman, K. Sward, and B. Rippe Unaltered size selectivity of the glomerular filtration barrier in caveolin-1 knockout mice Am J Physiol Renal Physiol, August 1, 2009; 297(2): F257 - F262. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. G. Navar Glomerular permeability: a never-ending saga Am J Physiol Renal Physiol, June 1, 2009; 296(6): F1266 - F1268. [Full Text] [PDF]
|
|
|
|
|
|
J. Axelsson, A. Rippe, D. Venturoli, P. Sward, and B. Rippe Effects of early endotoxemia and dextran-induced anaphylaxis on the size selectivity of the glomerular filtration barrier in rats Am J Physiol Renal Physiol, February 1, 2009; 296(2): F242 - F248. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Rippe Free water transport, small pore transport and the osmotic pressure gradient three-pore model of peritoneal transport Nephrol. Dial. Transplant., July 1, 2008; 23(7): 2147 - 2153. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
, control; 
, diabetes) after 9 wk of poorly controlled diabetes compared with age-matched controls. Ficoll 