Abstract
Background. Under proteinuric conditions, ultrafiltrated albumin can induce an inflammatory and fibrotic response in proximal tubular cells. It is unclear whether albumin per se or compounds bound to albumin are nephrotoxic. Some studies have supported the toxicity of albumin-bound fatty acids; however, these compared untreated, fatty acid containing, albumin and delipidated albumin. To prevent confounding by the delipidation procedure, we compared delipidated albumin and oleic acid (OA)-loaded delipidated albumin in two models: the classical rat protein overload and the Axolotl. The latter had an amphibian kidney with a subset of nephrons that drained the peritoneal cavity, so that i.p. injection of albumin selectively targeted open but not closed nephrons and was used to prevent removal of fatty acids from albumin in the circulation.
Methods. Protein overload was induced in Wistar rats (groups n = 8, for 12 weeks) and Axolotl (groups n = 10, for 10 days) by daily i.p. injections of 1 g (rat) or 50 mg (Axolotl) of fatty acid-free bovine serum albumin (BSA), fatty acid-free BSA with addition of six molecules of oleic acid (OA-BSA) or saline (SAL).
Results. After 12 weeks, proteinuria and SBP were increased in BSA and OA-BSA rats compared to saline-injected controls (P < 0.05), but OA loading had no additional effect compared to albumin alone. This was also true for glomerular and interstitial inflammation, fibrotic changes and focal glomerulosclerosis (OA-BSA versus BSA, all P = ns). Axolotls injected with OA-loaded albumin showed comparable protein storage in tubular epithelial cells, tubular dilatation and peritubular fibrosis around tubules draining the peritoneal cavity compared with Axolotls injected with albumin alone. This was also true for TGF-β (inflammation marker) and collagen I (fibrosis marker) (OA-BSA versus BSA, all P = ns).
Conclusions. In the Axolotl and chronic rat model, OA loading of albumin did not aggravate renal damage compared to albumin alone. Although in vitro studies clearly show induction of changes in cultured tubular epithelial cells exposed to albumin-bound fatty acids that are consistent with a role in induction of tubulointerstitial disease, our experiments suggest that currently available models for demonstrating such an effect in vivo are insufficient.
Introduction
Under proteinuric conditions, ultrafiltrated proteins reflect the degree of glomerular damage and play an active role in the pathogenesis of chronic tubulointerstitial damage [1,2]. Albumin is the major protein in the ultrafiltrate and nephrotic urine. High in vivo and in vitro exposure of proximal tubular cells (PTC) to albumin has been shown to induce a pro-inflammatory and pro-fibrotic response [3–5]. Without specific treatment, albumin isolated from plasma always carries other molecules, including fatty acids (FA) [6]. It has been argued that not albumin per se, but rather the FA that are bound to albumin, are toxic to PTC. Macrophage chemotactic activity and apoptosis were increased in cultured PTC in response to exposure to untreated albumin that contains FA, compared to delipidated albumin that does not contain FA [7,8]. Untreated FA-containing albumin induced more tubulointerstitial inflammation and glomerular injury in animal protein-overload models than delipidated FA-free albumin [7,9,10]. However, the delipidation procedure may remove other substances besides FA, or modify chemical reactivity or structure of albumin. It remains therefore unclear whether the observed effects are due to FA or the procedure of delipidation itself. Moreover, in rodent protein-overload studies, the albumin-bound FA may never appear in the glomerular ultrafiltrate, because in the circulation FA are loosely bound to albumin, with a t1/2 in the order of milliseconds [11]. Once FA are not bound to albumin, they tend to be taken up and oxidized in tissues, rather than to stay in the circulation [6,12,13]. Injection doses of palmitate have been shown to almost completely disappear from the circulation within minutes after intravenous injection [6].
We aimed to prevent any potential confounding by the delipidation procedure, by comparing delipidated albumin with delipidated albumin that was subsequently loaded with FA, rather than regular albumin with delipidated albumin. We selectively loaded delipidated albumin with oleic acid (OA) because this is the most abundant FA present in human serum [11]. Moreover, OA has been shown to induce fibronectin [14] and ROS production in vitro [15]. To this end, we studied the chronic effects of FA-loaded delipidated albumin versus delipidated albumin on renal damage in the classical rat protein-overload model. We also aimed to prevent the occurrence of removal of FA during the passage of the complex through the circulation, by the use of the protein-overload model with FA-loaded delipidated albumin versus delipidated albumin in Axolotls (Ambystoma Mexicanum) rather than rats. This amphibian kidney has a unique anatomy with closed and open nephrons; the latter drain the peritoneal cavity via a nephrostome. Injection of proteins into the peritoneal cavity will cause exposure of PTC of nephrons with nephrostomes without the passage of albumin through the circulation [16].
Materials and methods
Albumin solutions for protein overload (rat and Axolotl)
Fatty acid-free bovine serum albumin (BSA) (Sigma-Aldrich, Zwijndrecht, The Netherlands, catalog no. A-3803) was dissolved in phosphate buffered saline (PBS, pH 7.2) to a solution of 30%. BSA was complexed with OA (Sigma-Aldrich, catalog no. O-1008) as previously described, resulting in OA-loaded albumin (OA-BSA) with an OA:BSA molar ratio of 5.2:1 [17]. Low endotoxin levels of the solutions were previously confirmed in our laboratory [17].
Protein overload in rat
Protein-overload nephropathy was induced in adult male Wistar rats (Harlan, Horst, The Netherlands), weighing 220.7 ± 8.6 g (mean ± SD). Rats were housed in a light- and temperature-controlled environment, and had free access to water and standard rat chow. All rats received intraperitoneal injections six times a week for 12 consecutive weeks under isoflurane anaesthesia. Rats were injected with either 3.4 mL of saline (SAL, n = 8), 1 g of fatty acid-free BSA (BSA, n = 8) or OA-loaded BSA (OA-BSA, n = 8). All procedures were approved by the Committee for Animals Experiments of the University of Groningen, and the Principles of Laboratory Animal Care (NIH publication no. 85-23) were followed.
Bodyweight was measured weekly. Urine was collected every 4 weeks during a 24-h stay in metabolic cages with access to drinking water only. Tail vein blood samples were taken every 4 weeks under anaesthesia. Systolic blood pressure (SBP) was measured in conscious rats by the tail-cuff method. After 12 weeks, the rats were anaesthetized, a blood sample was taken by cannulation of the aorta and the kidneys were perfused with saline. A coronal tissue slice was snap-frozen in isopentane and stored at −80°C. Another coronal tissue slice was fixed in 4% paraformaldehyde and processed for paraffin embedding.
Measurements in blood and urine
Proteinuria was measured colorimetrically with the pyrogallol red–molybdate method. Plasma and urine creatinine levels were determined by the Jaffé method (Merck Mega, Darmstadt, Germany).
Immunohistochemistry
Paraffin sections (4 μm) were stained with periodic acid-Schiff (PAS) to evaluate focal glomerulosclerosis (FGS). Immunostaining was performed on paraffin sections for α-smooth muscle actin (α-SMA, clone 1A4, dilution 1:15 000, Sigma, St Louis, MO, USA) and macrophages (ED1, dilution 1:1000, Serotec, Oxford, UK). Deparaffinized sections were subjected to heat-induced antigen retrieval by overnight incubation in 0.1 M Tris/HCl buffer (pH = 9.0) at 80°C. Endogenous peroxidase was blocked with 0.075% H2O2 in PBS (30 min). Primary antibodies diluted in 1% BSA/PBS were incubated for 60 min at room temperature. Binding was detected using sequential incubations (30 min) with appropriate peroxidase (PO)-labelled secondary antibodies (DakoCytomation), diluted in PBS with 1% BSA and 1% human serum. Peroxidase activity was developed using 3,3′-diaminobenzidine tetrachloride (DAB, 10 min). Sections were counterstained with haematoxylin. Appropriate isotype and PBS controls were consistently negative.
Measurement of renal damage
FGS was semi-quantitatively scored (scale 0–4) in PAS-stained sections and expressed as the mean score of 50 glomeruli per kidney. FGS was scored positive when mesangial matrix expansion and adhesion of the visceral epithelium to Bowman's capsule were simultaneously present. A score of 1 was given when 25% of the glomerulus was involved, 2 for 50%, 3 for 75% and 4 for 100%. Computerized morphometry was used to measure glomerular α-SMA (50 glomeruli) and interstitial α-SMA (30 randomly selected cortical fields; vessels and glomeruli were excluded). The total staining area was divided by the total surface area and expressed as percentage. For glomerular macrophages (Mϕ), the number of positive cells per glomerulus (n = 50) was counted. The number of interstitial macrophages was counted per interstitial field (average of 30 fields per kidney) using a 10 × 10 grid at a magnification of 200× (with a total area of 0.25 mm2 per interstitial field); vessels and glomeruli were excluded from measurements. Interstitial macrophages are presented as number per interstitial field.
Protein overload in Axolotl
Eighteen-month-old neotenic Axolotls of both sexes, weighing between 80 and 120 g, were derived from the Axolotl Colony of the University of Indiana. Animals were held in tanks of aerated tap water at a constant temperature of 18°C with a 12-h light (06.00–18.00 hours) and 12-h dark cycle (18:00–06:00 hours). They were fed with pellets of fish food. One week prior to the study, animals were randomly allotted to three groups of 10 animals each. The daily BSA dose was chosen based on work of Gross et al. [16] who did not find tubulointerstitial disease after intraperitoneal injection of glycated human albumin at a concentration of 4.5 g/dL for 6 days (the daily injected amount of albumin was ∼22.5 mg) and work of Hein et al. [18] who found significant induction of tubulointerstitial disease after intraperitoneal injection of 0.5 mL of human albumin (probably corresponding to ∼20 mg of albumin) for 6 consecutive days. To prevent potential underdosing, the animals received daily intraperitoneal injections for 10 consecutive days of either saline (SAL), 50 mg fatty acid-free BSA (BSA, n = 10) or OA-loaded BSA (OA-BSA, n = 10). At the end of the observation period, a blood sample was obtained under general anaesthesia (3-aminobenzotic acid ethyl ester, A-5040; 10 g/L water in the tank; Sigma) and retrograde perfusion was performed via the main heart ventricle. For immunohistochemistry, animals were perfused with ice-cold isotonic saline. The kidneys were then excised. One part was snap-frozen and the other part was fixed with 4% formalin.
Light microscopy
Paraffin sections (4 μm) were stained with a connective tissue Ladewig stain [19] and examined using light microscopy at a magnification of 100×. The changes were quantified by a ‘blinded’ examiner (MvT), who was unaware of the assignment to the treatment groups, using a score system by comparison with a set of photos with standardized lesions, i.e. tubular dilatation, protein droplet content of tubular epithelial cells and peritubular interstitial fibrosis around protein storing tubules. For protein droplet content, scores were assigned for the quantity of the tubules containing protein droplets (0–4: 0 is 0; 1 is <25%; 2 is 25–50%; 3 is 50–75%; 4 is >75% of the tubules containing protein droplets) and for the quality of the droplets according to the size of the droplets. The size of the droplets was scored semi-quantitatively on a scale of 1–3: small droplets (∼0.5 μm) is 1; moderate droplets (∼2 μm) is 2 and large droplets (∼5 μm) is 3. The ultimate score for protein droplet content was given by multiplying the quantity and quality score (range 0–12). For peritubular interstitial fibrosis, the following scores were assigned: 0 is no change; 1 is minimal change; 2 is moderate change; 3 is marked change and 4 is very pronounced change. For tubular dilatation a score of 0 corresponded to an average diameter of 50 μm, 1 to 100 μm, 2 to 150 μm and 4 to 250 μm or more.
Immunohistochemistry
Immunostaining was performed as described above on paraffin sections for TGF-β (TGF-β1, SC-146, dilution 1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA) and anti-collagen I (rabbit anti rat collagen I, AB 755, Lot 131 DDM, dilution 1:200, Chemicon International Inc., Temecula, CA, USA).
Glomerular, tubular and interstitial structures were assessed using a score system. Tubular epithelial cells and interstitial cells were separately quantified, evaluating the area and the intensity of staining. The scores were defined as 0 is no staining, 1 is minimal staining, 2 is moderate staining, 3 is marked staining and 4 is very pronounced staining.
Oil Red O staining and quantification in rat and Axolotl kidneys
Frozen sections (4 μm) were fixed with 8% formalin for 10 min. After washing with demi-water, sections were briefly kept in 60% isopropanol, and then stained for 10 min in a freshly diluted Oil Red O solution (6 parts Oil Red O stock solution and 4 parts water; Oil Red O stock solution is 0.5% Oil Red O in isopropanol). The stain was then removed, and the sections were briefly kept in 60% isopropanol (to remove the excess staining), after which the sections were washed with water and counterstained with haematoxylin. Tubular epithelial cells in the cortex area were semi-quantitatively scored for their area covered by fat droplets (scale 0–4): 0 accounts for no fat droplets; 1/2 for <10% of the cortex area containing droplets; 1 for 10–25%; 2 for 25–50%; 3 for 50–75% and 4 for >75% of the cortex area containing fat droplets.
Fatty acid spectra of rat and Axolotl kidneys
Pieces (∼0.5 g) of renal cortex of rats and Axolotls were homogenized. Spectra of FA of these homogenates were determined by gas chromatography as previously described [20].
Statistical analysis
Data are presented as median and range. Statistical analyses were performed with SPSS Version 12.0 (SPSS Inc., Chicago, IL, USA). Differences between groups were detected with the non-parametric Kruskall–Wallis test. A two-sided P-value <0.05 indicated statistical significance.
Results
Rat: clinicopathological parameters and renal damage
Clinicopathological parameters are shown in Table 1. At Week 12, bodyweight was the same in all treatment groups. The BSA- and OA-BSA-overloaded rats did not differ in urinary protein excretion, creatinine clearance and SBP, but these parameters were in both groups—as expected—significantly increased when compared to saline-injected controls (P < 0.001 for urinary protein excretion and creatinine clearance; P < 0.05 for SBP). The course of development of proteinuria in the different groups is shown in Figure 1. At all time points there was an insignificant trend towards less proteinuria in the OA-BSA-overloaded rats compared with the BSA-overloaded rats.
Clinicopathological parameters rat model at Week 12
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 440.8 ± 31.9 | 438.3 ± 47.3 | 426.7 ± 37.4 |
| Proteinuria (mg/day) | 15.7 ± 4.3 | 139.0 ± 55.6# | 103.3 ± 59.6# |
| SBP (mmHg) | 126.6 ± 5.9 | 137.1 ± 6.3* | 141.1 ± 7.3* |
| CrCl (ml/min/100 g bodyweight) | 0.751 ± 0.087 | 1.098 ± 0.194# | 1.103 ± 0.187# |
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 440.8 ± 31.9 | 438.3 ± 47.3 | 426.7 ± 37.4 |
| Proteinuria (mg/day) | 15.7 ± 4.3 | 139.0 ± 55.6# | 103.3 ± 59.6# |
| SBP (mmHg) | 126.6 ± 5.9 | 137.1 ± 6.3* | 141.1 ± 7.3* |
| CrCl (ml/min/100 g bodyweight) | 0.751 ± 0.087 | 1.098 ± 0.194# | 1.103 ± 0.187# |
SAL, saline; BSA, BSA-injected rats; OA-BSA, oleic acid-loaded BSA-injected rats; SBP, systolic blood pressure; CrCl, creatinine clearance.
*P < 0.05 versus saline; #P < 0.001 versus saline.
Clinicopathological parameters rat model at Week 12
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 440.8 ± 31.9 | 438.3 ± 47.3 | 426.7 ± 37.4 |
| Proteinuria (mg/day) | 15.7 ± 4.3 | 139.0 ± 55.6# | 103.3 ± 59.6# |
| SBP (mmHg) | 126.6 ± 5.9 | 137.1 ± 6.3* | 141.1 ± 7.3* |
| CrCl (ml/min/100 g bodyweight) | 0.751 ± 0.087 | 1.098 ± 0.194# | 1.103 ± 0.187# |
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 440.8 ± 31.9 | 438.3 ± 47.3 | 426.7 ± 37.4 |
| Proteinuria (mg/day) | 15.7 ± 4.3 | 139.0 ± 55.6# | 103.3 ± 59.6# |
| SBP (mmHg) | 126.6 ± 5.9 | 137.1 ± 6.3* | 141.1 ± 7.3* |
| CrCl (ml/min/100 g bodyweight) | 0.751 ± 0.087 | 1.098 ± 0.194# | 1.103 ± 0.187# |
SAL, saline; BSA, BSA-injected rats; OA-BSA, oleic acid-loaded BSA-injected rats; SBP, systolic blood pressure; CrCl, creatinine clearance.
*P < 0.05 versus saline; #P < 0.001 versus saline.
Proteinuria in the three different treatment groups: saline-injected control rats (SAL), delipidated albumin (BSA) and oleic acid-loaded delipidated albumin (OA-BSA).
The scores of histological parameters are given in Figure 2. Representative photographs of histological sections are shown in Figure 3. Between BSA and OA-BSA groups, there were no significant differences in glomerular (mean ± SD 3.9 ± 1.1 versus 4.4 ± 1.3 macrophages per glomerulus, respectively, P = 0.46) and interstitial macrophages (87.4 ± 28.6 versus 84.7 ± 25.8 macrophages per interstitial field, respectively, P = 0.60). Also, glomerular (1.2 ± 0.6 versus 1.2 ± 0.8% per glomerulus, respectively, P = 0.92) and interstitial (2.0 ± 1.0 and 1.7 ± 0.4% per interstitial field, respectively, P = 0.83) α-SMA expression did not significantly differ between BSA- and OA-BSA-overloaded rats. The same was true for FGS (0.10 ± 0.08 versus 0.07 ± 0.06 arbitrary units, respectively, P = 0.26). Compared to the saline-injected control group, the extent of glomerular macrophages, interstitial α-SMA expression and FGS was significantly increased in both BSA- and OA-BSA-overloaded rats (respectively, values were 2.1 ± 0.6, 0.6 ± 0.2 and 0.0 ± 0.0 in control rats, P < 0.01 for all six comparisons). Scores for interstitial macrophages and glomerular α-SMA expression in BSA and OA-BSA were not significantly different from saline-injected controls (respectively, values were 64.5 ± 32.3 and 0.9 ± 0.6 in control rats, P-values ranged between 0.15 and 0.25 for the four comparisons).
Scores of renal damage parameters in rats of the three different treatment groups: saline-injected control animals (SAL), animals injected with delipidated albumin (BSA) and oleic acid-loaded delipidated albumin (OA-BSA).
Protein overloading in rat. (A) In saline-injected control kidneys, interstitial α-SMA expression is not very abundant, whereas glomerular α-SMA expression is more pronounced. (B) In protein-overloaded kidneys (both BSA and OA-BSA), increased interstitial α-SMA expression was found, while glomerular α-SMA expression was not changed. (C) In saline-injected control kidneys, little interstitial and glomerular macrophages were found. (D) In protein-overloaded kidneys (both BSA and OA-BSA), the number of interstitial macrophages was not changed, while an increased number of glomerular macrophages were found. An immunohistochemical stain is shown. Magnifications are 200×.
Axolotl: clinicopathological parameters and renal damage
Clinicopathological parameters are shown in Table 2. At Day 10 bodyweight, kidney weight and heart weight were the same in all treatment groups. The scores of histological parameters are given in Figure 4. Representative photographs of histological sections are shown in Figure 5. In the BSA- and OA-BSA-injected animals, tubular protein storage was only noted in tubules that drained the peritoneal cavity via a nephrostome (Figure 5A), confirming the selective uptake of proteins in PTC in open, but not in closed, nephrons [16]. Furthermore, tubular dilatation and peritubular fibrosis were found in areas in close vicinity to protein-loaded tubules (Figure 5B). Between BSA and OA-BSA groups, there were no significant differences in tubular protein storage (4.6 ± 3.9 versus 5.9 ± 3.4 arbitrary units, respectively, P = 0.27), tubular dilatation (2.3 ± 1.0 versus 2.1 ± 1.1 arbitrary units, respectively, P = 0.63) and fibrosis (1.6 ± 0.4 versus 1.3 ± 0.7 arbitrary units, respectively, P = 0.28) (determined on a connective tissue Ladewig stain). Compared to the saline-injected control group, the extent of tubular protein storage and tubular dilatation was significantly increased in both BSA and OA-BSA-overloaded Axolotls (respectively, values were 1.5 ± 2.4 and 0.7 ± 0.9 in controls, both P < 0.05). Fibrosis was significantly increased in BSA (P < 0.001) and non-significantly in OA-BSA (P = 0.07) overloaded Axolotls (value was 0.8 ± 0.6 in controls). Immunohistochemical stainings confirmed these results. TGF-β staining, used as inflammatory marker, was found in epithelial cells (Figure 5C and D). TGF-β staining in Axolotl renal cortical tissue in our experiments seems to differ from the same staining in a previous paper on Axolotl renal cortical tissue [16]. It should be realized that in the latter paper, TGF-β staining was performed after repeated intraperitoneal injection of bovine serum, which contains more proteins than albumin alone, whereas we injected only albumin (or albumin with OA). It seems that bovine serum gives a more pronounced TFG-β induction than albumin alone. However, localization seems to be most pronounced at the apical tubular membrane in both studies. Collagen I staining, used as a marker of fibrosis, was found in epithelial and interstitial cells (Figure 5E and F). We applied no specific markers for differentiation between proximal and distal tubular epithelial cells. So, it cannot be judged whether collagen I expression in tubules, like presented in Figure 5F, is localized in proximal or distal tubules. Between BSA and OA-BSA groups, there were no significant differences in TGF-β (3.2 ± 0.6 versus 2.8 ± 0.8 arbitrary units, respectively, P = 0.49) and collagen I (7.9 ± 7.3 versus 4.6 ± 3.3 arbitrary units, respectively, P = 0.39) staining. TGF-β and collagen I staining in both BSA and OA-BSA groups were significantly increased when compared to saline-injected controls (0.9 ± 0.5 and 1.6 ± 2.4 in controls, respectively, P < 0.05 for all four comparisons).
Clinicopathological parameters Axolotl model at Day 10
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 113 ± 15 | 107 ± 13 | 110 ± 12 |
| Kidney weight (mg) | 103.6 ± 48.2 | 87.1 ± 18.9 | 108.3 ± 47.4 |
| Heart weight (mg) | 50.2 ± 21.6 | 39.0 ± 9.0 | 56.8 ± 21.2 |
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 113 ± 15 | 107 ± 13 | 110 ± 12 |
| Kidney weight (mg) | 103.6 ± 48.2 | 87.1 ± 18.9 | 108.3 ± 47.4 |
| Heart weight (mg) | 50.2 ± 21.6 | 39.0 ± 9.0 | 56.8 ± 21.2 |
SAL, saline; BSA, BSA-injected Axolotls; OA-BSA, oleic acid-loaded BSA-injected Axolotls.
Clinicopathological parameters Axolotl model at Day 10
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 113 ± 15 | 107 ± 13 | 110 ± 12 |
| Kidney weight (mg) | 103.6 ± 48.2 | 87.1 ± 18.9 | 108.3 ± 47.4 |
| Heart weight (mg) | 50.2 ± 21.6 | 39.0 ± 9.0 | 56.8 ± 21.2 |
| SAL | BSA | OA-BSA | |
|---|---|---|---|
| Bodyweight (g) | 113 ± 15 | 107 ± 13 | 110 ± 12 |
| Kidney weight (mg) | 103.6 ± 48.2 | 87.1 ± 18.9 | 108.3 ± 47.4 |
| Heart weight (mg) | 50.2 ± 21.6 | 39.0 ± 9.0 | 56.8 ± 21.2 |
SAL, saline; BSA, BSA-injected Axolotls; OA-BSA, oleic acid-loaded BSA-injected Axolotls.
Scores of renal damage parameters in Axolotls of the three different treatment groups: saline-injected control animals (SAL), animals injected with delipidated albumin (BSA) and with oleic acid-loaded delipidated albumin (OA-BSA).
Protein overloading in Axolotl. (A) Ciliated funnel on the peritoneal surface of the Axolotl kidney. Cilia are present on the peritoneal surface and in the funnel (arrow). After protein loading, protein droplets are found in tubules that drain the peritoneal cavity, and not in other tubules. (B) After protein loading, in tubules that drain the peritoneal cavity, the tubular epithelial cells were swollen and filled with protein droplets (arrows). Sometimes tubular dilatation was observed in tubules with protein droplets (stars). Accumulation of interstitial tissue around the tubules with protein droplets was also seen (arrowheads). (C) In saline-injected control kidneys, virtually no TGF-β expression was found (arrows). (D) In protein-overloaded kidneys, increased tubular TGF-β expression was found (arrows). (E) In saline-injected control kidneys, only little tubular collagen I expression was found. (F) In protein-overloaded kidneys, increased collagen I expression was found. In (A) and (B), a Ladewig stain is shown, while (C–F) show an immunohistochemical stain. Magnifications are 100× (A–B) and 200× (C–F).
Oil Red O staining for quantification of lipid droplets
In rats, the area of cortical tubular epithelial cells that contained lipid droplets did not differ significantly between BSA and OA-BSA treatment [median (interquartile range) 0.3 (0.0–0.5) versus 0.5 (0.1–0.5) respectively, P = 0.69]. In Axolotls, there was a non-significant trend towards higher Oil Red O staining in tubular epithelial cells in the OA-BSA-treated group than in the BSA-treated group [3.0 (1.3–3.5) versus 1.5 (0.0–2.8) respectively, P = 0.17].
Fatty acid spectra of kidney cortex homogenates
In rats, no differences, and even no trends for differences, were present between kidney cortex of rats treated with OA-BSA and BSA, neither if contents of FA—and more specifically of OA—are presented per gram tissue nor if contents of individual FA—including OA—are presented as percentage of total amount of FA (Table 3). Results are clearly different for Axolotls. If amounts of FA are presented as total amounts per gram tissue, a trend towards higher amount of OA in the kidney cortex of OA-BSA-treated Axolotls versus BSA-treated Axolotls appears to be present, the results not reaching significance mainly because of large variation in contents in the OA-BSA-treated group (Table 4). However, if amounts of FA are presented as percentage of the total amount of FA, variation is much lower, and clear differences in fatty acid content between OA-BSA- and BSA-treated Axolotls become apparent. In OA-BSA-treated Axolotls, OA comprised 30.8% of the total FA present compared to 17.0% in the BSA-treated Axolotls (P = 0.003) (Table 4). Relative amounts of palmitic acid, stearic acid, vaccenic acid, linoleic acid, docosahexenoic acid and nervonic acid decreased in response to treatment with OA-BSA compared to BSA.
Quantification of fatty acid contents of rat kidney cortex
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 17.5 ± 1.6 | 15.8 ± 3.0 | 0.19 | 24.7 ± 0.7 | 24.1 ± 0.6 | 0.12 |
| Stearic acid | C18:0 | 13.3 ± 0.6 | 12.4 ± 2.5 | 0.38 | 18.9 ± 1.8 | 18.9 ± 0.4 | 0.96 |
| Vaccenic acid | C18:1ω7 | 1.8 ± 0.3 | 1.6 ± 0.3 | 0.16 | 2.6 ± 0.2 | 2.4 ± 0.1 | 0.18 |
| Oleic acid | C18:1ω9 | 5.3 ± 2.5 | 4.3 ± 0.6 | 0.31 | 7.3 ± 2.4 | 6.7 ± 0.7 | 0.50 |
| Linoleic acid | C18:2ω6 | 8.2 ± 2.5 | 7.0 ± 1.2 | 0.29 | 11.4 ± 2.0 | 10.8 ± 0.6 | 0.50 |
| Arachidonic acid | C20:4ω6 | 18.2 ± 1.7 | 18.1 ± 4.0 | 0.95 | 25.8 ± 2.7 | 27.5 ± 1.3 | 0.17 |
| Eicosapentaenoic acid | C20:5ω3 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.45 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.73 |
| Docosahexenoic acid | C22:6ω3 | 1.1 ± 0.2 | 1.1 ± 0.2 | 0.98 | 1.5 ± 0.3 | 1.6 ± 0.1 | 0.36 |
| Nervonic acid | C24:1ω9 | 0.4 ± 0.1 | 0.5 ± 0.1 | 0.56 | 0.6 ± 0.1 | 0.7 ± 0.0 | 0.04 |
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 17.5 ± 1.6 | 15.8 ± 3.0 | 0.19 | 24.7 ± 0.7 | 24.1 ± 0.6 | 0.12 |
| Stearic acid | C18:0 | 13.3 ± 0.6 | 12.4 ± 2.5 | 0.38 | 18.9 ± 1.8 | 18.9 ± 0.4 | 0.96 |
| Vaccenic acid | C18:1ω7 | 1.8 ± 0.3 | 1.6 ± 0.3 | 0.16 | 2.6 ± 0.2 | 2.4 ± 0.1 | 0.18 |
| Oleic acid | C18:1ω9 | 5.3 ± 2.5 | 4.3 ± 0.6 | 0.31 | 7.3 ± 2.4 | 6.7 ± 0.7 | 0.50 |
| Linoleic acid | C18:2ω6 | 8.2 ± 2.5 | 7.0 ± 1.2 | 0.29 | 11.4 ± 2.0 | 10.8 ± 0.6 | 0.50 |
| Arachidonic acid | C20:4ω6 | 18.2 ± 1.7 | 18.1 ± 4.0 | 0.95 | 25.8 ± 2.7 | 27.5 ± 1.3 | 0.17 |
| Eicosapentaenoic acid | C20:5ω3 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.45 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.73 |
| Docosahexenoic acid | C22:6ω3 | 1.1 ± 0.2 | 1.1 ± 0.2 | 0.98 | 1.5 ± 0.3 | 1.6 ± 0.1 | 0.36 |
| Nervonic acid | C24:1ω9 | 0.4 ± 0.1 | 0.5 ± 0.1 | 0.56 | 0.6 ± 0.1 | 0.7 ± 0.0 | 0.04 |
BSA, BSA-injected rats; OA-BSA, oleic acid-loaded BSA-injected rats.
Data expressed as mean ± SD. Only fatty acids with a relative presence of ≥1% are presented in the table. Eicosapentaenoic acid and nervonic acid are included in the table to allow comparison with fatty acid contents of Axolotl kidney cortex.
Quantification of fatty acid contents of rat kidney cortex
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 17.5 ± 1.6 | 15.8 ± 3.0 | 0.19 | 24.7 ± 0.7 | 24.1 ± 0.6 | 0.12 |
| Stearic acid | C18:0 | 13.3 ± 0.6 | 12.4 ± 2.5 | 0.38 | 18.9 ± 1.8 | 18.9 ± 0.4 | 0.96 |
| Vaccenic acid | C18:1ω7 | 1.8 ± 0.3 | 1.6 ± 0.3 | 0.16 | 2.6 ± 0.2 | 2.4 ± 0.1 | 0.18 |
| Oleic acid | C18:1ω9 | 5.3 ± 2.5 | 4.3 ± 0.6 | 0.31 | 7.3 ± 2.4 | 6.7 ± 0.7 | 0.50 |
| Linoleic acid | C18:2ω6 | 8.2 ± 2.5 | 7.0 ± 1.2 | 0.29 | 11.4 ± 2.0 | 10.8 ± 0.6 | 0.50 |
| Arachidonic acid | C20:4ω6 | 18.2 ± 1.7 | 18.1 ± 4.0 | 0.95 | 25.8 ± 2.7 | 27.5 ± 1.3 | 0.17 |
| Eicosapentaenoic acid | C20:5ω3 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.45 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.73 |
| Docosahexenoic acid | C22:6ω3 | 1.1 ± 0.2 | 1.1 ± 0.2 | 0.98 | 1.5 ± 0.3 | 1.6 ± 0.1 | 0.36 |
| Nervonic acid | C24:1ω9 | 0.4 ± 0.1 | 0.5 ± 0.1 | 0.56 | 0.6 ± 0.1 | 0.7 ± 0.0 | 0.04 |
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 17.5 ± 1.6 | 15.8 ± 3.0 | 0.19 | 24.7 ± 0.7 | 24.1 ± 0.6 | 0.12 |
| Stearic acid | C18:0 | 13.3 ± 0.6 | 12.4 ± 2.5 | 0.38 | 18.9 ± 1.8 | 18.9 ± 0.4 | 0.96 |
| Vaccenic acid | C18:1ω7 | 1.8 ± 0.3 | 1.6 ± 0.3 | 0.16 | 2.6 ± 0.2 | 2.4 ± 0.1 | 0.18 |
| Oleic acid | C18:1ω9 | 5.3 ± 2.5 | 4.3 ± 0.6 | 0.31 | 7.3 ± 2.4 | 6.7 ± 0.7 | 0.50 |
| Linoleic acid | C18:2ω6 | 8.2 ± 2.5 | 7.0 ± 1.2 | 0.29 | 11.4 ± 2.0 | 10.8 ± 0.6 | 0.50 |
| Arachidonic acid | C20:4ω6 | 18.2 ± 1.7 | 18.1 ± 4.0 | 0.95 | 25.8 ± 2.7 | 27.5 ± 1.3 | 0.17 |
| Eicosapentaenoic acid | C20:5ω3 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.45 | 0.1 ± 0.0 | 0.1 ± 0.0 | 0.73 |
| Docosahexenoic acid | C22:6ω3 | 1.1 ± 0.2 | 1.1 ± 0.2 | 0.98 | 1.5 ± 0.3 | 1.6 ± 0.1 | 0.36 |
| Nervonic acid | C24:1ω9 | 0.4 ± 0.1 | 0.5 ± 0.1 | 0.56 | 0.6 ± 0.1 | 0.7 ± 0.0 | 0.04 |
BSA, BSA-injected rats; OA-BSA, oleic acid-loaded BSA-injected rats.
Data expressed as mean ± SD. Only fatty acids with a relative presence of ≥1% are presented in the table. Eicosapentaenoic acid and nervonic acid are included in the table to allow comparison with fatty acid contents of Axolotl kidney cortex.
Quantification of fatty acid contents of Axolotl kidney cortex
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 23.5 ± 15.9 | 18.1 ± 16.8 | 0.51 | 20.3 ± 0.6 | 17.1 ± 2.3 | 0.003 |
| Stearic acid | C18:0 | 9.2 ± 6.3 | 6.8 ± 6.2 | 0.44 | 8.1 ± 1.0 | 6.7 ± 1.5 | 0.05 |
| Vaccenic acid | C18:1ω7 | 5.1 ± 3.5 | 4.1 ± 4.1 | 0.59 | 4.4 ± 0.3 | 3.6 ± 0.3 | <0.001 |
| Oleic acid | C18:1ω9 | 19.6 ± 13.0 | 36.0 ± 38.6 | 0.27 | 17.0 ± 1.9 | 30.8 ± 9.6 | 0.003 |
| Linoleic acid | C18:2ω6 | 28.1 ± 19.2 | 22.9 ± 23.9 | 0.63 | 23.8 ± 1.6 | 20.0 ± 1.8 | <0.001 |
| Arachidonic acid | C20:4ω6 | 6.7 ± 4.6 | 4.5 ± 3.8 | 0.30 | 5.9 ± 1.2 | 4.8 ± 1.9 | 0.18 |
| Eicosapentaenoic acid | C20:5ω3 | 6.0 ± 4.1 | 4.4 ± 3.9 | 0.40 | 5.2 ± 0.6 | 4.5 ± 1.8 | 0.25 |
| Docosahexenoic acid | C22:6ω3 | 6.4 ± 4.3 | 4.6 ± 4.5 | 0.42 | 5.5 ± 0.3 | 4.3 ± 1.2 | 0.02 |
| Nervonic acid | C24:1ω9 | 2.8 ± 2.0 | 1.9 ± 1.7 | 0.35 | 2.4 ± 0.4 | 1.9 ± 0.4 | 0.02 |
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 23.5 ± 15.9 | 18.1 ± 16.8 | 0.51 | 20.3 ± 0.6 | 17.1 ± 2.3 | 0.003 |
| Stearic acid | C18:0 | 9.2 ± 6.3 | 6.8 ± 6.2 | 0.44 | 8.1 ± 1.0 | 6.7 ± 1.5 | 0.05 |
| Vaccenic acid | C18:1ω7 | 5.1 ± 3.5 | 4.1 ± 4.1 | 0.59 | 4.4 ± 0.3 | 3.6 ± 0.3 | <0.001 |
| Oleic acid | C18:1ω9 | 19.6 ± 13.0 | 36.0 ± 38.6 | 0.27 | 17.0 ± 1.9 | 30.8 ± 9.6 | 0.003 |
| Linoleic acid | C18:2ω6 | 28.1 ± 19.2 | 22.9 ± 23.9 | 0.63 | 23.8 ± 1.6 | 20.0 ± 1.8 | <0.001 |
| Arachidonic acid | C20:4ω6 | 6.7 ± 4.6 | 4.5 ± 3.8 | 0.30 | 5.9 ± 1.2 | 4.8 ± 1.9 | 0.18 |
| Eicosapentaenoic acid | C20:5ω3 | 6.0 ± 4.1 | 4.4 ± 3.9 | 0.40 | 5.2 ± 0.6 | 4.5 ± 1.8 | 0.25 |
| Docosahexenoic acid | C22:6ω3 | 6.4 ± 4.3 | 4.6 ± 4.5 | 0.42 | 5.5 ± 0.3 | 4.3 ± 1.2 | 0.02 |
| Nervonic acid | C24:1ω9 | 2.8 ± 2.0 | 1.9 ± 1.7 | 0.35 | 2.4 ± 0.4 | 1.9 ± 0.4 | 0.02 |
BSA, BSA-injected Axolotls; OA-BSA, oleic acid-loaded BSA-injected Axolotls.
Data expressed as mean ± SD. Only fatty acids with a relative presence of ≥1% are included in the table.
Quantification of fatty acid contents of Axolotl kidney cortex
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 23.5 ± 15.9 | 18.1 ± 16.8 | 0.51 | 20.3 ± 0.6 | 17.1 ± 2.3 | 0.003 |
| Stearic acid | C18:0 | 9.2 ± 6.3 | 6.8 ± 6.2 | 0.44 | 8.1 ± 1.0 | 6.7 ± 1.5 | 0.05 |
| Vaccenic acid | C18:1ω7 | 5.1 ± 3.5 | 4.1 ± 4.1 | 0.59 | 4.4 ± 0.3 | 3.6 ± 0.3 | <0.001 |
| Oleic acid | C18:1ω9 | 19.6 ± 13.0 | 36.0 ± 38.6 | 0.27 | 17.0 ± 1.9 | 30.8 ± 9.6 | 0.003 |
| Linoleic acid | C18:2ω6 | 28.1 ± 19.2 | 22.9 ± 23.9 | 0.63 | 23.8 ± 1.6 | 20.0 ± 1.8 | <0.001 |
| Arachidonic acid | C20:4ω6 | 6.7 ± 4.6 | 4.5 ± 3.8 | 0.30 | 5.9 ± 1.2 | 4.8 ± 1.9 | 0.18 |
| Eicosapentaenoic acid | C20:5ω3 | 6.0 ± 4.1 | 4.4 ± 3.9 | 0.40 | 5.2 ± 0.6 | 4.5 ± 1.8 | 0.25 |
| Docosahexenoic acid | C22:6ω3 | 6.4 ± 4.3 | 4.6 ± 4.5 | 0.42 | 5.5 ± 0.3 | 4.3 ± 1.2 | 0.02 |
| Nervonic acid | C24:1ω9 | 2.8 ± 2.0 | 1.9 ± 1.7 | 0.35 | 2.4 ± 0.4 | 1.9 ± 0.4 | 0.02 |
| Name | Number of C atoms and double bounds | Absolute fatty acid content (μmol/g) | Relative fatty acid content (%) | ||||
|---|---|---|---|---|---|---|---|
| BSA | OA-BSA | P-value | BSA | OA-BSA | P-value | ||
| Palmitic acid | C16:0 | 23.5 ± 15.9 | 18.1 ± 16.8 | 0.51 | 20.3 ± 0.6 | 17.1 ± 2.3 | 0.003 |
| Stearic acid | C18:0 | 9.2 ± 6.3 | 6.8 ± 6.2 | 0.44 | 8.1 ± 1.0 | 6.7 ± 1.5 | 0.05 |
| Vaccenic acid | C18:1ω7 | 5.1 ± 3.5 | 4.1 ± 4.1 | 0.59 | 4.4 ± 0.3 | 3.6 ± 0.3 | <0.001 |
| Oleic acid | C18:1ω9 | 19.6 ± 13.0 | 36.0 ± 38.6 | 0.27 | 17.0 ± 1.9 | 30.8 ± 9.6 | 0.003 |
| Linoleic acid | C18:2ω6 | 28.1 ± 19.2 | 22.9 ± 23.9 | 0.63 | 23.8 ± 1.6 | 20.0 ± 1.8 | <0.001 |
| Arachidonic acid | C20:4ω6 | 6.7 ± 4.6 | 4.5 ± 3.8 | 0.30 | 5.9 ± 1.2 | 4.8 ± 1.9 | 0.18 |
| Eicosapentaenoic acid | C20:5ω3 | 6.0 ± 4.1 | 4.4 ± 3.9 | 0.40 | 5.2 ± 0.6 | 4.5 ± 1.8 | 0.25 |
| Docosahexenoic acid | C22:6ω3 | 6.4 ± 4.3 | 4.6 ± 4.5 | 0.42 | 5.5 ± 0.3 | 4.3 ± 1.2 | 0.02 |
| Nervonic acid | C24:1ω9 | 2.8 ± 2.0 | 1.9 ± 1.7 | 0.35 | 2.4 ± 0.4 | 1.9 ± 0.4 | 0.02 |
BSA, BSA-injected Axolotls; OA-BSA, oleic acid-loaded BSA-injected Axolotls.
Data expressed as mean ± SD. Only fatty acids with a relative presence of ≥1% are included in the table.
Discussion
In the current study, we show that OA loading of albumin has no additional effect on renal damage when compared to albumin alone. In both the Axolotl and the chronic rat model, proteinuria, renal inflammation and fibrotic changes were comparable in the groups injected with delipidated OA-loaded albumin and delipidated albumin. These results suggest that the delipidation procedure itself may have been responsible for observations of less renal injury after injection of delipidated albumin versus regular albumin in previous studies.
Other rodent protein-overload studies have provided evidence for the detrimental effects of albumin-bound FA on renal structure [9,10]. The experimental set-up in these studies was different from our set-up, for example, Thomas et al. [10] did the experiments in Lewis rats and gave twice as much albumin for 7 days. This makes a direct comparison with our study difficult. The most prominent difference with our study was, however, that in the above-described experiments, a comparison was made between untreated albumin—which always carries other molecules, including FA—and delipidated albumin. In our study, we compared delipidated albumin with delipidated albumin that was selectively loaded with OA. Delipidated OA-loaded albumin has been shown to induce ROS production in vitro [15]. In addition, in a short-term, i.e. 3 weeks, rat protein-overload model, we have previously shown that OA-loaded albumin induced more renal inflammation and prefibrotic changes than delipidated albumin, although no effect was seen on the level of proteinuria [17]. Surprisingly, in the current long-term, i.e. 12 weeks, protein-overload model, renal inflammation and fibrosis, as well as proteinuria, were similar in BSA- and OA-BSA-overloaded rats. The discrepancy between short- and long-term effects of OA-BSA may have different explanations. The results of the short-term study can be a chance hit. The same may of course apply to the results of this study; however, the results of our current long-term protein-overload study did not approach statistical significance, and if trends existed, they were more in the direction of a protective effect of OA than a harmful effect. It is also possible that in the current long-term protein-overload model, the additional effects of OA on renal damage that were seen in the short-term model were overruled by the deleterious effects of albumin and the vicious circle of tubulointerstitial damage and inflammation that was consequently induced. If this is true, the contribution of FA in the long term, i.e. the clinically more relevant time period, seems neglectable. Obviously, before a definite conclusion about the absence of an effect of albumin loaded with OA on induction of tubulointerstitial damage can be drawn, our findings require independent confirmation by other investigators.
Another point that needs to be addressed in previous rodent protein-overload studies is the possibility that the albumin-bound FA may never appear in the glomerular ultrafiltrate. Although albumin is the major carrier of FA in the circulation, FA are only loosely bound to albumin [11]. FA that are not bound to albumin are rapidly taken up by the tissues and consequently disappear from the circulation [6]. Consistent with this notion, we found no difference in Oil Red O staining in tubular epithelial cells and no difference in fatty acid spectra of homogenized renal cortex between BSA- and OA-BSA-overloaded rats. To overcome the potential removal of FA from albumin during passage through the circulation, we therefore evaluated the direct effects of albumin-bound FA on PTC in the Axolotl. This amphibian kidney has a unique anatomy in which ciliated peritoneal funnels, nephrostomes, connect the proximal tubule to the peritoneal cavity. Due to these nephrostomes, the Axolotl kidney possesses two different sets of nephrons: ‘normal’ closed nephrons, in which glomerulus and tubule form a closed unit, and open nephrons that connect via a nephrostome to the peritoneal cavity. Hein et al. [18,21] and Gross et al. [16] have previously shown that the Axolotl kidney can serve as a model to study tubulointerstitial activation and induction of interstitial fibrosis by protein loading. Injection of protein into the peritoneal cavity fails to expose closed nephrons to a protein load, but causes selective uptake and storage of proteins in tubular epithelial cells of nephrons with nephrostomes [16]. We confirmed this selective uptake and storage of proteins in PTC belonging to open nephrons, but we furthermore showed that OA loading of albumin did not aggravate renal damage compared to albumin alone. Although Gross et al. [16] could not demonstrate an effect of glycated human albumin (∼22.5 mg injected daily for 6 days) compared to saline-injected controls, our results confirm preliminary results of another study by Gross et al. [22], in which it was shown that Axolotls overloaded with lipid poor and lipid rich albumin—for 10 days with 25 mg of albumin—had equal amounts of tubular protein storage and peritubular fibrosis. However, like in rats in the circulation, it cannot be excluded that the FA dissociate from albumin in the peritoneal cavity in the Axolotl model. We found, however, a significantly higher percentage of OA in renal cortex of OA-BSA-overloaded Axolotls than that of BSA overloaded, which strongly supports the notion that albumin-bound OA reached the tubular lumen in Axolotls.
In cultured PTC, the effects of delipidated albumin in comparison with untreated, lipidated albumin are not conclusive. Although several studies demonstrated that FA that are bound to albumin aggravated the deleterious effects on PTC compared to delipidated albumin [7,8], other studies have failed to show this. PTC stimulated with untreated lipidated albumin and delipidated albumin showed equal levels of proliferation and hypertrophy and synthesized equal levels of endothelin-1 and MCP-1 [23–25]. The initiation of inflammatory and fibrotic cascades is likely to be dependent on endocytosis of albumin. It appears that renal tubular cell lines with low rates of endocytic uptake are resistant to the pro-fibrotic effects of albumin [26].
Which proteins play the most predominant role in the activation of PTC towards a more pro-inflammatory and pro-fibrotic state is still unanswered. In addition to albumin, and albumin-bound compounds, several other filtered proteins have been suggested to induce changes in PTC, including proteins of the complement system, immunoglobulins and growth factors. To what extent these proteins contribute to inflammatory and fibrotic effects is unclear, although some studies indicate differential and protein-specific effects [23,27,28].
How can proteins, or more specific albumin, or albumin-bound FA exert an effect on PTC? Although the exact pathway has not been established, evidence suggests that the process involves the initial endocytic uptake of albumin. In humans and rodents, albumin in the tubular lumen binds to the receptors megalin and cubulin in the luminal surface of PTC [1,29]. After binding, it is then internalized. Additional uptake mechanisms of albumin have been suggested, including endocytosis via the epidermal growth factor (EGF) receptor [30] and reabsorption via a high-capacity transcellular pathway that transfers intact albumin into the peritubular blood supply [31]. However, to what extent these different pathways contribute is unclear. A key consequence of albumin overload is the production of the pro-inflammatory cytokine MCP-1 [25], the pro-fibrotic cytokine TGF-β [32] and collagen [26], and these are dependent on the endocytosis of albumin. However, a recent study has shown that induction of TGF-β by exposure to albumin does not depend on the endocytosis of albumin [33]. Whether the above-mentioned pathways are also present in the Axolotl kidney is unclear, but as it is shown that proteins accumulate in PTC, it seems very likely that at least some of them are present. Nevertheless, we did not observe differences in response to tubular loading with delipidated albumin and OA-loaded albumin.
In conclusion, we studied the effects of OA-complexed albumin on the induction of renal damage. To overcome potential confounding by the delipidation procedure, we compared delipidated albumin with OA-loaded albumin. We compared the effects in the classical rat protein-overload model, in which albumin reaches the kidney via the circulation, and in the Axolotl in which albumin is directly delivered to the tubules and thereby omits the possible disturbing effects of the circulation on albumin. In both models, we demonstrated that OA loading of albumin has no additional effect on renal damage when compared to albumin alone. Although in vitro studies clearly show induction of changes in cultured tubular epithelial cells exposed to albumin-bound FA that are consistent with a role in induction of tubulointerstitial disease, our experiments suggest that currently available rat and mice models for demonstrating such an effect in vivo are insufficient. One reason may be that FA attached to albumin never reach the tubular lumen. In this respect, the Axolotl model, in which there are tubules that directly drain the peritoneal cavity, seems more appropriate.
The authors wish to thank Marian Bulthuis, Heike Ziebart and Ingrid Martini for skilled technical assistance. Graduate School Groningen University Institute for Drug Exploration (GUIDE); J.K. de Cock Foundation (Groningen, The Netherlands) provided funding to this study.
Conflict of interest statement. None declared.
Comments