The Cytoplasmic Domain of Tissue Factor Restricts Physiological Albuminuria and Pathological Proteinuria Associated with Glomerulonephritis in MiceApostolopoulos J. · Moussa L. · Tipping P.G.
Centre for Inflammatory Diseases, Department of Medicine, Monash University, Clayton, Vic., Australia Corresponding Author
Background/Aims: Tissue factor (TF) is a transmembrane protein that is essential for coagulation. TF is expressed on podocytes and its cytoplasmic domain has cell signalling functions in epithelial cells. Methods: Mice lacking the cytoplasmic domain of TF (TFCT–/– mice) were used to study its role in physiological albuminuria and pathological proteinuria following induction of glomerulonephritis (GN). Results: Absence of the cytoplasmic domain of TF was associated with increased albuminuria, podocyte effacement, reduced podocyte numbers and increased spontaneous glomerular tumour necrosis factor α(TNFα) production under physiological conditions. In mice developing GN, absence of the cytoplasmic domain of TF resulted in increased proteinuria and enhanced renal TNFα production without altering other parameters of renal inflammation and injury. Studies in TFCT–/– chimeric mice (created by bone marrow transplantation) showed increased proteinuria and renal TNFα mRNA in GN was associated with absence of the cytoplasmic domain of TF in the kidney and was independent of the leucocyte phenotype. Conclusion: These studies demonstrate that the cytoplasmic domain of TF contributes to renal albumin retention and its renal expression protects against proteinuria in leucocyte-mediated renal inflammation. Increased glomerular production of TNFα in the absence of cytoplasmic domain of TF may contribute to podocyte injury resulting in albuminuria and proteinuria.
Copyright © 2010 S. Karger AG, Basel
Tissue factor (TF) is a single chain type I transmembrane glycoprotein with a 219-amino acid N-terminal extracellular domain, a 23-amino acid transmembrane domain and a cytoplasmic domain that contains 21 amino acids in humans (20 amino acids in mice). TF has significant structural homology with type II cytokine receptors . It does not have any intrinsic enzymatic activity, but through its ability to bind coagulation factor VII (FVII) and accelerate its auto-activation, TF provides the primary trigger for coagulation in vivo. The coagulant function of TF is dependent on FVII binding capacity of the extracellular domain and tethering in the cell membrane via the trans-membrane domain  but is completely independent of the cytoplasmic domain .
In addition to its essential role in haemostasis, TF has important coagulation independent roles in embryogenesis , angiogenesis , tumour growth  and metastasis  and cell migration [7,8]. The cytoplasmic domain of TF has two serine residues that can be phosphorylated by protein kinase C-dependent mechanisms [9,10]. Cell responses shown to require the cytoplasmic domain of TF include induction of Ca2+ fluxes in myeloid cell lines , adhesion, migration and spreading of fibroblasts , smooth muscle migration associated with vascular remodelling following injury , metastatic tumour growth and protease-activated receptor-2 (PAR-2)-dependent angiogenesis . In bladder carcinoma cell lines, the TF cytoplasmic domain has been shown to activate Rac1 and p38 mitogen-activated protein kinase (MAPK) and promote adhesion, spreading and migration .
TF is expressed widely throughout the body, particularly in the brain, heart, lungs and kidneys . Expression of TF in the adventitial layer of blood vessels forms a ‘haemostatic envelope’ that results in rapid activation of coagulation following vascular injury. In the kidney, TF expression is observed predominantly in the glomerulus on podocyte and parietal epithelial cells [15,16,17]. Foot processes of podocytes are essential to the size-selective nature of the glomerular filtration barrier. Both the junctions between adjacent foot processes that form the slit diaphragm complex and interactions between the sole of the foot process and the glomerular basement membrane (GBM) involving α3β1-integrin and dystroglycan are critical to filtration. In vitro, TF inhibits α3β1-integrin-dependent migration of keratinocytes when the cytoplasmic domain is not phosphorylated . Interactions of slit-diaphragm proteins, including nephrin and podocin with the cytoskeleton are also critical in maintenance of podocyte morphology and function. The cytoplasmic domain of TF has been shown to interact with cytoskeletal structures through actin-binding protein-280 (ABP-280) that cross-links actin filaments, promotes orthogonal branching and contributes to maintenance of epithelial cell structure, adhesion and migration [12,14]. The ability of TF to modulate integrin adhesion and APB-280 cytoskeletal interactions in epithelial cells in vitro raises the possibility that via similar interactions in podocytes TF may contribute maintenance of the glomerular filtration barrier.
We have previously demonstrated that mice lacking the cytoplasmic domain of TF (TFCT–/– mice) are protected from systemic inflammatory injury induced by lipopolysaccharide  and from arthritis . Unexpectedly, these did not show significant protection from leucocyte-mediated glomerular injury and developed enhanced proteinuria compared to wild-type (WT) mice. To further explore the role of the cytoplasmic domain of TF in proteinuria, we compared urinary albumin excretion in TFCT–/– and WT mice under normal physiological conditions and used bone marrow chimeric mice to determine whether the expression of the cytoplasmic domain of TF from leucocytes or from intrinsic renal cells contributes to the increased proteinuria observed during development of glomerulonephritis (GN).
TFCT–/– mice, backcrossed 9 generations onto a C57BL/6 background, were housed under specific pathogen-free conditions and maintained by usual husbandry methods. C57BL/6 mice were used as WT controls. Male mice were allocated for study at 4, 6, 8 and 14 weeks of age. Experimental protocols were approved by the Monash University Animal Ethics Committee.
Mice were housed for 24 h in metabolic cages with free access to water to collect urine. Urine collections were centrifuged at 13,000 rpm for 5 min and stored at –20°C prior to assay. For PAGE analysis, 10 µl of urine was added to 10 µl of 2× gel loading buffer was loaded onto a 10% polyacrylamide-SDS gel, run at 100 V in an SDS-PAGE buffer system for 1 h and then stained in Coomassie blue for 30 min. Gels were de-stained overnight and imaged using a Canon D1250U2F scanner. Urinary albumin concentrations were measured by ELISA (Bethyl Laboratories Inc., Montgomery, Tex., USA), according to the manufacturer’s instructions and results expressed as µg/24 h. The presence of vitamin D-binding protein (VDBP) and transferrin in urine was examined by Western blotting using polyclonal rabbit anti-Gc globulin at 1:200 (DakoCytomation, catalogue No.: A0021) and polyclonal rabbit anti-transferrin at 1:1,000 (DakoCytomation, catalogue No.: A0061). The secondary antibody used was swine anti-rabbit horseradish peroxidase 1:5,000 (DakoCytomation, catalogue No.: P0399). Urine from mice with tubular proteinuria served as a positive control.
Renal tissue was collected into Bouin’s fixative and 5-µm paraffin sections were stained with periodic acid-Schiff for routine light microscopy. Tissue for electron microscopy was fixed in 1.25% glutaraldehyde in 0.2 M sodium cacodylate overnight. The tissue was then post-fixed in 1% osmium tetroxide for 2 h and stained with 2% uranyl-acetate en bloc for 1.5 h. It was then dehydrated in 50% alcohol for 10 min and 75% alcohol for 15 min, infiltrated with LR White™ resin for 1 h and then in resin overnight. The tissue was embedded in LR White resin and polymerized at 58°C for 2 h. Sections were cut at 0.5 µm and stained with methylene blue. Thin sections were then cut using a Reichert OM U3 ultramicrotome, stained with lead citrate, and examined using a Phillips 301 electron microscope.
The histological pattern of glomerular podocin and nephrin expression was assessed on 6-µm frozen tissue sections by indirect immunofluorescence, using rabbit anti-human podocin (Alpha Diagnostic International, Tex., USA) at a dilution of 1:1,000 and rabbit anti-rat nephrin (Alpha Diagnostic) at a dilution of 1:1,000. In both cases, the secondary antibody was biotinylated swine anti-rabbit IgG (DAKO, Denmark) at a dilution of 1:100 that was followed by steptavidin-FITC (DAKO) at a dilution of 1:100.
mRNA was quantified byreal-time PCR. Total RNA was prepared from fresh kidney, from glomeruli isolated following renal perfusion with Dynal beads and from cells collected from urine by centrifugation at 13,000 g for 5 min using Trizol reagent (Invitrogen, catalogue No. 315596-018), according to the manufacturer’s instructions. cDNA synthesis was performed using 10 µl of RNA with a High Capacity cDNA Reverse Transcription cDNA Kit (Applied Biosystems). The expression of selected genes was performed in individual real-time PCR reactions in a total of 20 µl using 10 µl Taqman Universal PCR Master Mix (2×), 1 µl 20× Taqman Gene Expression Assay Mix (containing specific Taqman FAM dye-labeled minor groove binder probes for each gene) and 9 µl of cDNA (diluted in RNase/DNase-free water) with the following cycler conditions: initial setup, 2 min at 50°C, then a cycle of 10 min at 95°C; denature 15 s at 95°C and anneal/extend, 1 min at 60°C. The ΔΔCt value for each reaction was used to calculate a ratio for gene expression relative to 18S.
Podocytes were identified by WT-1-positive nuclear staining in 5-µm formalin-fixed paraffin kidney sections. WT-1 staining was performed using rabbit anti-mouse WT-1 polyclonal antibody at 1:100 in 1% BSA/PBS (Santa Cruz sc-192) and biotinylated swine anti-rabbit IgG (1:300) in 1% BSA/PBS/10% mouse serum, followed by diaminobenzadine black solution (Dako) and counter staining in nuclear fast red. Podocytes were counted in 25 equatorially sectioned glomeruli per animal and expressed as podocytes per glomerular cross section.
For kidney tumour necrosis factor α(TNFα) measurements, kidney tissue lysates were prepared from snap frozen kidney by homogenization in 1 ml of lysis buffer (0.1 M Tris HCl pH 7.5, 0.15 M NaCl, 1% Triton X-100) followed by centrifugation at 13,000 g for 10 min. To determine, glomerular TNFα production, 1,000 freshly isolated glomeruli were cultured in 2 ml of DMEM for 48 h. Supernatants were collected and stored at –20°C. Urine and serum were collected and stored at –20°C. Microtitre plates were coated with 2 µg/ml anti-mouse TNFα monoclonal antibody (clone AF-410NA, R&D Systems) in carbonate bicarbonate buffer and incubated at 4°C overnight. The plates were washed 3 times with 0.05% Tween/PBS and blocked by 200 µl of 1% BSA/PBS for 2 h with gentle agitation. The samples and recombinant standards were diluted 1:10 in blocking buffer, added to the plates and incubated at 4°C overnight. The plates were washed and then incubated with 250 ng/ml biotinylated anti-mouse TNFα monoclonal antibody (clone BAF40, R&D Systems), then amplified further by incubation with 1 µg/ml Extravidin® solution (Sigma) at 37°C for 30 min followed by anti-avidin-biotin solution (0.25 µg/ml) for 1 h at room temperature. Streptavidin/horseradish peroxidase (1/2,000) was added for 30 min, then developed in tetramethylbenzadine; the development was stopped in 0.5 M H2SO4 and plates were read on the microplate reader at 450 nm. Protein levels were calculated from values derived from the respective standard curve.
Mice between 8 and 9 weeks of age were injected intravenously with 18 mg of sheep anti-mouse anti-GBM antibody as previously described . Twenty-four-hour urine collections were performed immediately prior to initiation of GN and on days 7, 14 and 21 after administration of anti-GBM antibody. At the end of day 21, renal tissue and blood was collected for analysis.
Urinary protein levels were measured on aliquots from 24-hour collections using a modified Bradford assay by reference to BSA standards. Blood samples were allowed to clot overnight at 4°C, and spun at 1,200 g for 10 min to collect serum. Creatinine levels were measured using an enzymatic creatininase assay on a Roche Cobas Bioanalyser (CREA Plus, Boehringer Mannheim, catalogue No. 1775685).
Renal tissue was collected into Bouin’s fixative, and 5-µm paraffin-embedded sections were stained with periodic acid-Schiff reagent to assess renal injury by light microscopy.Glomerular crescent formation was assessed in 50 glomeruli per animal. Glomeruli were considered to exhibit crescent formation when two or more layers of cells were observed in Bowman’s space. Fibrinogen staining was performed using a rabbit anti-mouse fibrinogen antibody (a gift from Dr. J. Degen, Children’s Research Foundation, Cincinnati, Ohio, USA), a biotinylated secondary antibody, and incubation with an avidin-biotinylated enzyme complex solution. Fibrinogen staining was scored on a 0–3 scale using a blinded protocol on a minimum of 50 glomeruli. A score of 0 was given to glomeruli with little or no staining, a score of 1 was given if fibrin was deposited in at least 25% of the glomerulus, a score of 2 was given if fibrin was deposited in at least 50% of the glomerulus, and a score of 3 was given when more than 75% of the glomerulus was positive for fibrin.
Frozen tissue sections from periodate-lysine paraformaldehyde-fixed renal tissue were stained with an anti-CD68 monoclonal antibody (FA/11) using a 3-layer immunoperoxidase technique as previously described  to quantify glomerular macrophage accumulation. The number of positive cells/glomerular cross section was counted using a blinded protocol on a minimum of 20 equatorial or near equatorial glomerular cross sections to determine the mean number of glomerular macrophages for each animal.
Circulating antibodies (mouse anti-sheep globulin IgG) to the nephritogenic antigen were measured by ELISA. Microtitre plates (Greiner, Longwood, Fla., USA) were coated overnight with 10 µg/ml normal sheep globulin and mouse serum was added in serial dilutions of 1:50 to 1:1,600. After incubation at 37°C for 1 h, plates were incubated with peroxidase-conjugated sheep anti-mouse IgG (1 in 2,000), followed by tetramethylbenzadine solution (Sigma, St. Louis, Mo., USA) as a chromogenic substrate. The reaction was stopped using 0.1 M H2SO4 and absorbance was measured at 450 nm on a microplate reader (Biorad, Hercules, Calif., USA).
Bone marrow chimeric mice were generated as previously described . TFCT–/– mice recipients were transplanted with WT bone marrow (WT >> TFCT–/–) and vice versa (TFCT–/– >> WT). WT mice transplanted with WT bone marrow (WT >> WT) served as transplant controls (sham chimeras). Recipient mice at 6 weeks of age were irradiated with 2 doses of 550 rad gamma irradiation and injected with bone marrow from donor mice within 24 h. A period of 8 weeks was allowed for bone marrow reconstitution before the initiation of GN. Previous bone marrow transplant studies have shown that approximately 95% of leucocytes in recipient mice are derived from donor bone marrow after transplantation . GN was initiated by injection of anti-GBM antibodies when mice were 14 weeks of age and urine and blood was collected as described above.
Results are expressed as the mean ± standard error of the mean (SEM). Statistical comparisons involving two groups were performed by the Student t test and comparisons involving more than 2 groups were performed by a one-way ANOVA, followed by Dunnett’s multiple comparison test (GraphPad Prism, GraphPad Software Inc., San Diego, Calif., USA).
Deletion of the Cytoplasmic Domain of TF Results in Albuminuria and Podocyte Effacement
SDS-PAGE analysis of urine from 8-week-old mice demonstrated prominent protein bands with a molecular weight consistent with albumin in TFCT–/– mice (fig. 1a, lanes 6–10) compared with WT mice (lanes 1–5). Renal excretion of mouse major urinary proteins  was similar in both strains indicating similarities in the handling of low-molecular-weight proteins in the 18- to 19-kDa range. Quantification by ELISA showed similar 24-hour urinary albumin excretion in TFCT–/– and WT mice at 4 weeks of age, but significant increases (between 2- and 3-fold) in TFCT–/– mice at 6, 8 and 14 weeks of age (fig. 1b). TFCT–/– mice appeared otherwise phenotypically normal. There were no differences between their body weight, serum creatinine or urine volumes between TFCT–/– and WT mice (data not shown). The renal histology by light microscopy did not reveal any morphological abnormalities in glomeruli or tubulo-interstitial areas in TFCT–/– mice (fig. 2b) compared with WT (fig. 2a). However, on electron microscopy, effacement of podocyte foot processes was clearly demonstrated in TFCT–/– mice (fig. 2d and inset) but podocyte morphology was intact in WT mice (fig. 2c). The absence of VDBP or transferrin in the urine of TFCT–/– mice indicated that proximal tubular protein reabsorption was functionally intact, consistent with the result in WT mice (data not shown).
|Fig. 1. TFCT–/– mice have increased urinary albumin excretion. a Urine samples from 8-week-old WT mice (lanes 1–5) and TFCT–/– mice (lanes 6–10) were loaded on SDS-PAGE gels and stained with Coomassie blue. TFCT–/– mice had prominent protein bands corresponding to albumin (67 kDa) compared with WT mice in their urine. MUPs represent the ‘major urinary proteins’ [45, 46] or ‘low-molecular-weight proteins’ routinely seen in normal male mouse urine . b Quantitative albuminuria (measured by ELISA) on 24-hour urine collections was significantly higher in TFCT–/– mice compared with WT mice at 6, 8 and 14 weeks of age (* p = 0.01, ** p = 0.02, *** p = 0.04, n = 15 per group).|
|Fig. 2. TFCT–/– mice have normal renal morphology by light microscopy at 8 weeks of age: WT (a) and TFCT–/–(b). ×400. However, electron microscopy revealed effacement of podocyte foot process in TFCT–/– mice (indicated by arrows in d and inset) compared to normal podocyte morphology in WT mice (c). The bars represent 5 µm.|
Podocin and Nephrin Distribution Is Altered in the Absence of the Cytoplasmic Domain of TF, without Quantitative Changes in mRNA Expression
Glomerular expression of podocin and nephrin showed a uniform distribution with more prominent linear staining along the GBM in WT mice (fig. 3). In TFCT–/– mice, the expression of podocin and nephrin showed a less uniform, segmented distribution pattern with variable staining intensity throughout the glomerulus. Quantitative RT-PCR showed no significant differences in kidney podocin or nephrin mRNA between TFCT–/– and WT mice. TF mRNA expression in whole kidney and isolated glomeruli (data not shown) was also similar.
|Fig. 3. The pattern of podocin (a) and nephrin (b) expression in the glomeruli of TFCT–/– mice is abnormal. Podocin and nephrin staining in WT mice was uniform along capillary membranes throughout the entire glomerulus. However, in TFCT–/– mice, the pattern of podocin and nephrin staining was different from WT, with some segmental areas showing no expression. Renal podocin and nephrin mRNA levels, however, were similar in both groups of mice.|
Absence of the Cytoplasmic Domain of TF Is Associated with Reduced Podocyte Numbers and Increased Nephrin mRNA in Urine
Loss of podocytes can play an important role in the progression of renal disease  and podocyte numbers have been shown to be a strong predictor of long-term urinary albumin excretion and rapid progression of glomerulosclerosis in type II diabetes . We therefore examined whether development of albuminuria in TFCT–/– mice was associated with reduced podocyte numbers and their loss into the urine. Podocytes were identified by WT-1 staining (fig. 4a) and counted in at least 20 equatorial glomerular cross sections per mouse. Podocyte numbers were reduced by 38% in 8-week-old TFCT–/– mice (fig. 4b). Phase-contrast microscopy revealed significantly greater numbers of cells in the urine of TFCT–/– mice (data not shown) and real-time PCR analysis demonstrated increased levels of nephrin mRNA in pelleted urinary cells from TFCT–/– mice compared to WT mice (fig. 4c), suggesting the presence of significantly greater numbers of podocyte in urine from TFCT–/– mice.
|Fig. 4. Podocyte numbers, counted as WT-1 positive cells (indicated by arrows) per equatorial glomerular cross section (gcs) (a) are significantly reduced in TFCT–/– mice compared to WT mice at 8 weeks of age (p < 0.0001, n = 8 per group, b). The urine levels of nephrin mRNA were 5-fold higher in 8-week-old TFCT–/– mice compared to WT mice (* p < 0.05, n = 8 per group, c), suggesting shedding of podocytes in the urine.|
TNFα Is Constitutively Expressed in the Kidney when the Cytoplasmic Domain of TF Is Absent
Dysregulated expression of proinflammatory cytokines (IL-1, IL-6 and TNFα) has previously been demonstrated in TFCT–/– mice with systemic inflammatory responses . However, the ability of the cytoplasmic domain of TF to affect cytokine production in the kidney is unknown. Quantitative RT-PCR showed renal TNFα mRNA levels were significantly higher in 8-week-old TFCT–/– mice (fig. 5) compared with WT mice, whereas IL-1β and IL-6 mRNA was unaltered. Measurement of TNFα protein by ELISA showed significant amounts of TNFα in kidneys (14 ± 1.2 pg/mg kidney) and urine (91 ± 20 pg/ml) of TFCT–/– mice. In contrast, TNFα was undetectable in the kidney and urine of WT mice. Isolated glomeruli from TFCT–/– mice spontaneously produced TNFα in vitro (183 ± 8 pg/103 glomeruli/48 h), whereas no TNFα production was detected in WT glomeruli (fig. 5b). TNFα was undetectable in the serum of both TFCT–/– mice and WT mice.
|Fig. 5. TFCT–/– mice under normal physiological conditions show increased renal TNFα mRNA expression compared to WT mice (* p < 0.001, n = 8 per group), whilst IL-1 and IL-6 levels remained unaltered (a). Cultured glomeruli from TFCT–/–mice (b) spontaneously produce significant amounts of TNFα in culture (* p < 0.0001, n = 5 per group), whereas TNFα production by glomeruli from WT mice is not detectable (ND).|
Absence of the Cytoplasmic Domain of TF Increases Proteinuria in Inflammatory Renal Injury in GN
Although albuminuria is significantly greater in TFCT–/– mice compared to WT in the non-diseased state (fig. 1b), ‘baseline’ proteinuria was not significantly different (WT, 1.34 ± 0.20 mg/24 h; TFCT–/–, 1.72 ± 0.37 mg/24 h, p = 0.31, n = 22 per group). However, in response to inflammatory glomerular injury, TFCT–/– mice developed significantly greater proteinuria than WT mice at both day 14 and day 21 after the initiation of GN (fig. 6a). Albuminuria was also significantly greater at day 21 in TFCT–/– mice (fig. 6c).
|Fig. 6. The response of TFCT–/– mice to inflammatory glomerular injury. TFCT–/– mice had significantly more proteinuria than WT mice (a) at day 14 and 21 after initiation of α-GBM GN (* p < 0.05, n = 6 per group). However, other indicators of inflammatory injury, including serum creatinine (b) and glomerular macrophage recruitment (d) were not significantly different at day 21. Albuminuria at day 21 (c) (* p < 0.05, n = 6 per group). gcs = Glomerular cross section.|
Other indices of renal inflammation and injury were not significantly different after initiation of GN. Serum creatinine levels were similar (fig. 6b), as was the histological severity of injury as indicated by similar glomerular crescent formation at day 21 (WT, 29.2 ± 4.6% glomeruli affected; TFCT–/–, 30.4 ± 5.1% glomeruli affected, n = 6 per group). Glomerular fibrin deposition was not significantly different at day 21 (WT, 0.2 ± 0.1; TFCT–/–, 0.3 ± 0.1, n = 6 per group). Renal TF mRNA expression was similar in both groups of mice (WT, 0.1 ± 0.03; TFCT–/–, 0.1 ± 0.01, n = 6 per group). These results indicate that increased proteinuria was not associated with evidence of enhanced histological and functional injury in TFCT–/– mice with GN.
The Cytoplasmic Domain of TF Does Not Affect Nephritogenic Immune Responses or Glomerular Macrophage Recruitment
The humoral immune response to the nephritogenic antigen, indicated by circulating levels of autologous mouse anti-sheep IgG, was similar in WT and TFCT–/– mice (data not shown) and glomerular recruitment of cellular effectors (i.e. macrophages) of injury was similar in both groups (fig. 6d).
Increased Proteinuria in GN Depends on the Absence of the Cytoplasmic Domain of TF in the Kidney
Bone marrow transplantation was used to generate mice with absence of the cytoplasmic domain of TF restricted to either bone marrow or non-bone marrow derived cells. Bone marrow transplantation did not affect baseline proteinuria prior to induction of GN (fig. 7a). After induction of GN, the ‘sham chimeras’, i.e. WT mice transplanted with WT bone marrow (WT >> WT), showed proteinuria, serum creatinine and glomerular macrophage accumulation similar to non-transplanted WT mice with GN (fig. 7), indicating that the bone marrow transplantation protocol does not alter development of the disease. The mice lacking the cytoplasmic domain of TF in bone marrow derived cells but with WT kidneys (TFCT–/– >> WT) developed similar proteinuria to the sham chimeras (WT >> WT). However, in mice with WT bone marrow, lacking the cytoplasmic domain of TF in their kidneys (WT >> TFCT–/–), proteinuria was significantly increased compared to sham chimeras (p < 0.05) and equivalent to the proteinuria seen in the non-transplanted TFCT–/– mice with GN, indicating that the renal phenotype determines the propensity to increase proteinuria in GN. Serum creatinine and glomerular macrophage accumulation were not different between WT >> WT, WT >> TFCT–/– or TFCT–/– >> WT chimeras with GN.
|Fig. 7. Increased proteinuria in TFCT–/– mice with anti-GBM GN is associated with absence of the cytoplasmic domain of TF in the kidney. a WT >> TFCT–/– chimeras had significantly elevated proteinuria at day 21 compared to WT >> WT (sham) chimeras (* p < 0.05, n = 6 per group) whilst proteinuria levels in TFCT–/– >> WT chimeras were similar to sham chimeras. Other indices of glomerular injury and inflammation, including serum creatinine (b) and glomerular macrophage recruitment (c), were not significantly different between chimera groups. gcs = Glomerular cross section.|
Renal Absence of the Cytoplasmic Domain of TF Increases TNFα Expression in GN
TNFα protein (fig. 8a) and TNFα mRNA (fig. 8b) were significantly increased during GN in TFCT–/– mice compared to WT mice. In chimeric mice with GN (fig. 8c), mice lacking the cytoplasmic domain of TF in the kidney (WT >> TFCT–/–) had significantly higher renal TNFα mRNA during development of GN than sham chimeras (WT >> WT), Renal TNFα mRNA was similar in TFCT–/– >> WT and WT >> WT chimeras. Renal IL-1 and IL-6 mRNA were also slightly elevated in TFCT–/– >> WT compared with WT >> WT and WT >> TFCT–/– chimeras. TNFα protein (fig. 8d) was significantly increased in WT >> TFCT–/– chimeras compared to sham chimeras (WT >> WT), confirming that absence of the cytoplasmic domain of TF in the kidney is associated with increases in both TNFα mRNA and protein in mice with GN.
|Fig. 8. Increased renal TNFα expression in TFCT–/– mice with anti-GBM GN is associated with absence of the cytoplasmic domain of TF in the kidney. Renal TNFα protein expression (a) and TNFα mRNA levels (b) were significantly greater in TFCT–/– mice compared with WT mice at day 21 of GN (* p < 0.05, n = 6 per group). In chimeric mice with GN, renal TNFα mRNA levels (c) were increased significantly in WT >> TFCT–/– chimeras compared to WT >> WT (sham) chimeras and TFCT–/– >> WT chimeras (** p < 0.0001, n = 6 per group). There was a small but significant increase in IL-1 and IL-6 mRNA levels in the TFCT–/– >> WT chimeras compared to WT >> WT (sham) chimeras and WT >> TFCT–/– chimeras (* p < 0.05, n = 6 per group). Renal TNFα protein levels (d) were also increased significantly in WT >> TFCT–/– chimeras compared to WT >> WT (sham) chimeras (* p < 0.05, n = 6 per group).|
TF is strongly expressed in epithelial cells [15,26] including glomerular epithelial cells [15,16,17] but its function at epithelial interfaces is unclear. Podocytes are specialized epithelial cells that express TF  and play a pivotal role in the structure and function of the glomerular filtration barrier. The current studies show that absence of the cytoplasmic domain of TF in mice results in spontaneous albuminuria in association with podocyte foot process effacement, consistent with a role for TF in maintenance of normal podocyte structure and function.
TFCT–/– mice did not show any histological abnormalities in glomeruli or tubulo-interstitial areas by routine light microscopy. In addition, absence of VDBP and transferrin in the urine indicates intact proximal tubular function in these mice. Reabsorption of these proteins is mediated by the receptors megalin and cubulin  and their presence in the urine can be a sensitive index of proximal tubular injury . However, electron microscopy revealed podocyte effacement, similar to that seen in human minimal change disease, indicating a role for the cytoplasmic domain of TF in maintenance of normal podocyte structure in these mice. The glomerular distribution of podocin and nephrin was disrupted in a pattern similar to changes observed at the early stages of development of spontaneous albuminuria and podocyte injury in Munich Wistar Fromter rats  and in homocysteine-induced injury in Sprague-Dawley rats . The reduced podocyte numbers in glomeruli and the increased presence of cells in the urine expressing nephrin mRNA suggest that shedding of podocytes may be a consequence of effacement and detachment that results in loss of podocytes from glomeruli. Podocyte detachment and shedding have been observed in association with microalbuminuria and the early stages of several progressive experimental and human renal diseases , including puromycin aminonucleoside nephrosis  and diabetes .
There are no previous studies that provide evidence for a direct role of TF in podocyte structure or function. However, studies in cultured epithelial cells [12,14,34] have shown interactions between the cytoplasmic domain of TF and elements of the cytoskeleton and intergins that are important in cell structure, attachment and migration. TF co-localizes with ABP-280 on microspikes and lamellipodes at the leading edge of spreading epithelial cells in close proximity with actin and α-actinin, at branching points and at contact points with adjacent cells  and interactions between the cytoplasmic domain of TF and ABP-280 induce cytoskeletal rearrangement  that facilitates cell adhesion and migration. Activation of p38 MAPK is associated with cytoskeletal rearrangement in podocytes that can result in proteinuria  and FVIIa binding to TF activates p38 MAPK and GTPase Rac1 (Rac1) signalling pathways and promotes epithelial cell migration . It is possible that similar interactions between the cytoplasmic domain of TF and ABP-280 may contribute to the maintenance of podocyte structure and function.
TF has also been demonstrated to interact with α3β1-integrin, which is involved in adhesion of the sole of podocyte foot processes to laminin-5 in the GBM . Adhesion to laminin-5 via α3β1-integrin activates focal adhesion kinase (FAK) and extracellular signal-regulated kinase (ERK) and promotes survival of epithelial cells, whilst inhibition of either FAK or ERK signalling leads to apoptosis [36,37]. Phosphorylation of the cytoplasmic domain is required for TF to inhibit α3β1-integrin-dependent migration of epithelial cell lines on laminin-5 .
Interestingly, TFCT–/– mice showed increased levels of TNFα (mRNA and protein) in their kidneys and urine under normal physiological conditions, and glomeruli isolated from TFCT–/– mice spontaneously produced TNFα in vitro. Serum TNFα was undetectable, indicating that systemic TNFα production is not dysregulated. To our knowledge, spontaneous production of TNFα in the kidney under physiological conditions has not previously been reported. We have previously observed that the kidney is the major source of TNFα production in crescentic GN . Podocytes produce TNFα in response to injurious stimuli, such as lipopolysaccharide in vitro  and expression of TNFα by podocytes and increased urine TNFα has been reported in human membranous nephropathy . Thus, increased glomerular TNFα production may represent a ‘response to injury’ of podocytes in TFCT–/– mice.
Alternatively, it is possible that the spontaneous glomerular TNFα production may be the primary abnormality that leads to podocyte injury and detachment in TFCT–/– mice. Infusion of TNFα has been shown to induce proteinuria in rabbits ; TNFα induces actin reorganization in glomerular epithelial cells  and is associated with disrupted attachment of podocytes to the GBM. In addition, TNFα increases albumin permeability in isolated glomeruli and can lead to proteinuria . Increased TNFα production in TFCT–/– mice may directly contribute to their podocyte injury. The mechanism by which the cytoplasmic domain of TF could regulate TNFα production in podocytes is unknown. PAR-2 activation can induce TNFα expression  and the cytoplasmic domain of TF has been shown to suppress PAR-2 mediated cellular responses , thus the potential for TF to tonically suppress TNFα production via PAR-2-dependent mechanisms is worthy of exploration.
To investigate whether the podocyte abnormalities in TFCT–/– mice have a significant effect on renal response to inflammatory injury, we used a mouse model of anti-GBM GN. Following induction of anti-GBM GN, TFCT–/– mice developed twice the proteinuria of WT mice, without significant differences in their nephritogenic antibody levels, glomerular macrophage recruitment, crescent formation or serum creatinine. Renal expression of TF mRNA and glomerular fibrin deposition were also unaltered consistent with previous observations in these mice that deletion of the cytoplasmic domain of TF does not affect the coagulant function of TF . Studies with chimeric mice confirmed that the lack of expression of the cytoplasmic domain of TF in the kidney rather than in leucocytes determines this increased susceptibility to proteinuria in GN. This supports the conclusion that the increased susceptibility to proteinuria is due to the glomerular changes associated with the absence of the cytoplasmic domain of TF, rather than that alterations of inflammatory responses of leucocytes as observed with these mice represent systemic inflammation due to endotoxin challenge . Increased renal TNFα production was also observed in TFCT–/– mice with GN in association with increased proteinuria. Studies in chimeric mice showed that increased renal TNFα was associated with absence of the cytoplasmic domain of TF in the kidney. Serum levels of TNFα were similar in WT and TFCT–/– mice with GN confirming the absence of systemic dysregulation of TNFα production.
In summary, this study shows that the cytoplasmic domain of TF contributes to the structural integrity of podocytes, the renal retention of albumin and suppression of glomerular production of TNFα production under normal physiological conditions. Absence of the cytoplasmic domain of TF is associated with altered glomerular distribution of podocin and nephrin, and podocyte shedding into the urine. The mechanistic relationship between glomerular TNFα production and podocyte abnormalities is unknown. These subtle abnormalities in TFCT–/– mice lead to a more profound susceptibility to proteinuria following inflammatory injury in GN. The susceptibility to inflammatory glomerular injury is dependent on the absence of the cytoplasmic domain of TF in the kidney, independent of the leucocyte phenotype or the systemic or local glomerular inflammation response, but is associated with increased renal TNFα production.
General technical assistance was ably provided by Mrs. Rosemary Genovese and Ms. Emily Wilson. For electron microscopy studies, the technical assistance of Mr Paul Crammer and the advice of Prof. Terry Cook are gratefully acknowledged. We would also like to thank Dr. David Power for performing the transferrin and VDBP Western blots.
Jim Apostolopoulos, Department of Medicine, Monash University
Level 5 Block E, Monash Medical Centre
246 Clayton Road, Clayton, Vic. 3168 (Australia)
Tel. +61 3 9594 5534, Fax +61 3 9594 6495
J.A. and L.M. contributed equally to this work.
Received: December 22, 2009
Accepted: April 16, 2010
Published online: July 28, 2010
Number of Print Pages : 12
Number of Figures : 8, Number of Tables : 0, Number of References : 47
Nephron Experimental Nephrology
Vol. 116, No. 4, Year 2010 (Cover Date: November 2010)
Journal Editor: Hughes J. (Edinburgh)
ISSN: 1660-2129 (Print), eISSN: 1660-2129 (Online)
For additional information: http://www.karger.com/NEE