Tissue factor and its inhibitor in human non-crescentic glomerulonephritis—immunostaining vs plasma and urinary levels

Nephrology Dialysis Transplantation, Dec 2006

Background. Tissue factor (TF)—the most potent trigger of coagulation and emerging antiapoptotic, proliferative and angiogenic factor, along with its principal inhibitor (tissue factor pathway inhibitor, TFPI) are known to be involved in crescentic glomerulonephritis (GN). We studied the relationship between plasma and urinary levels as well as renal biopsy immunostaining of TF and TFPI antigens with reference to some clinical parameters in human chronic non-crescentic GN. Methods. We examined plasma and urinary levels of TF and total TFPI (pre-biopsy, ELISA) and the intensity of TF, TFPI 1 and TFPI 2 staining (immunoperoxidase histochemistry) in kidney biopsy specimens from 30 chronic GN patients. Results. Plasma and urinary TF (uTF) were higher in patients than in 18 healthy individuals. In normal kidneys, TF and TFPI 1/2 antigens were undetectable in glomeruli while a distinct staining of both TFPI variants was observed in tubules and interstitial microvessels. In diseased kidneys, TF was strongly expressed in glomeruli but was undetectable in tubules. In contrast, staining for TFPI 1/2 was observed in glomeruli and tubules. Neither plasma nor urinary levels of the markers correlated with the intensity of TF and TFPI 1/2 staining in biopsy specimens. uTF was significantly associated with creatinine clearance (R = 0.489, P = 0.006) and urinary TFPI (R = 0.554, P = 0.014), and tended to be lower in proliferative vs non-proliferative GN [83 (0–617) vs 281 (10–805) pg/ml; P = 0.06]. Conclusion. The intrarenal TF/TFPI system is profoundly disturbed in chronic GN. Plasma and urinary concentrations of TF and TFPI probably do not reflect genuine activity of the disease, likely due to a confounding effect of kidney insufficiency. uTF measurement seems to be helpful in initial identification of proliferative GN, yet further studies are required to validate its use as a marker of glomerular injury in chronic GN.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://ndt.oxfordjournals.org/content/21/12/3450.full.pdf

Tissue factor and its inhibitor in human non-crescentic glomerulonephritis—immunostaining vs plasma and urinary levels

Beata Naumnik 1 2 Jacek Borawski 1 2 Lech Chyczewski 0 1 Krystyna Pawlak 1 2 Michal Mysliwiec 1 2 0 Department of Clinical Molecular Biology, Medical University , Bialystok, Poland 1 Department of Nephrology and Transplantology with Dialysis Unit, Medical University , 14 Zurawia St, 15-549 Biatystok, Poland 2 Department of Nephrology and Transplantology with Dialysis Unit Background. Tissue factor (TF)the most potent trigger of coagulation and emerging antiapoptotic, proliferative and angiogenic factor, along with its principal inhibitor (tissue factor pathway inhibitor, TFPI) are known to be involved in crescentic glomerulonephritis (GN). We studied the relationship between plasma and urinary levels as well as renal biopsy immunostaining of TF and TFPI antigens with reference to some clinical parameters in human chronic non-crescentic GN. Methods. We examined plasma and urinary levels of TF and total TFPI (pre-biopsy, ELISA) and the intensity of TF, TFPI 1 and TFPI 2 staining (immunoperoxidase histochemistry) in kidney biopsy specimens from 30 chronic GN patients. Results. Plasma and urinary TF (uTF) were higher in patients than in 18 healthy individuals. In normal kidneys, TF and TFPI 1/2 antigens were undetectable in glomeruli while a distinct staining of both TFPI variants was observed in tubules and interstitial microvessels. In diseased kidneys, TF was strongly expressed in glomeruli but was undetectable in tubules. In contrast, staining for TFPI 1/2 was observed in glomeruli and tubules. Neither plasma nor urinary levels of the markers correlated with the intensity of TF and TFPI 1/2 staining in biopsy specimens. uTF was significantly associated with creatinine clearance (R 0.489, P 0.006) and urinary TFPI (R 0.554, P 0.014), and tended to be lower in proliferative vs non-proliferative GN [83 (0-617) vs 281 (10-805) pg/ml; P 0.06]. Conclusion. The intrarenal TF/TFPI system is profoundly disturbed in chronic GN. Plasma and urinary concentrations of TF and TFPI probably do not reflect genuine activity of the disease, likely due to a confounding effect of kidney insufficiency. Introduction Tissue factor (TF) is the principal initiator of the extrinsic coagulation pathway, the major pathway for inducing fibrin deposition in vivo. TF is frequently encrypted in the plasma membrane of cells in contact with blood, and is exposed only after stimulation by certain agonists including endotoxin, inflammatory cytokines, activated lymphocytes, growth factors and oxidized low-density lipoproteins [1]. Tissue factor pathway inhibitor (TFPI), the leading inhibitor of TF-initiated coagulation, is a circulating and endothelial-bound serine protease that inhibits factor VIIa/TF complexes in a factor Xa-dependent manner. Its two isoforms (TFPI 1 and TFPI 2) revealed similarities in the overall domain organization as well as considerable amino acid sequence homology; nevertheless, TFPI 2 has been reported to exhibit significant factor Xa-independent inhibitory activity against the VIIa/TF complex [2]. Recently, new TFPI functions have been reportedantiproliferative (mainly for TFPI 1) and mitogenic (for TFPI 2) [3]. Several human and experimental models of glomerulonephritis (GN), especially those with crescentic formation, show alterations in kidney tissue TF expression [4]. Studies using fibrinolytic agents, antibodies to TF or its biological inhibitor TFPI showed significant attenuation of crescentic GN development [57]. In addition, TF inhibition was associated with decreased glomerular inflammation and macrophage influx, and contributed to renal function preservation and reduction of proteinuria [7,8]. These data suggest that TF/TFPI imbalance in humans is important in the initiation of glomerular fibrin deposition and exacerbation of glomerular inflammation. The source of glomerular procoagulant activity may be either bloodborne or generated by resident glomerular cells [9]. TF expression can be induced in glomerular resident cells by monocyte-derived cytokines, mainly interleukin-1 [4]. Simultaneously, it is suggested that early regulation of TF activity is largely a result of functional up-regulation of constitutive TF in intrinsic glomerular cells, whereas in an advanced stage of the disease infiltrating macrophages are the major sites of TF synthesis [10]. Increased amounts of TF in glomerular diseases are found not only in the renal tissue, but also in the urine [11,12]. Two possible sources of urinary TF (uTF) are thus implicated: it could be a tissue (excretory) product which passes through the glomerular filter or a secretory product which originates in renal parenchyma [1]. Up-to-date studies are ambiguous and were performed exclusively in rapidly progressive crescentic GN [4,5,7,8,10,14]. The intrarenal TF/TFPI system in chronic GN has not been analysed yet. Therefore, we aimed to study the relationships between TF and TFPI levels in plasma and urine and their immunostainings in renal biopsy specimens from patients with chronic non-crescentic GN. The results also were analysed in relation to clinical markers of renal function and glomerular injury. Patients Urine and blood were collected from 48 subjects: 30 clinically stable patients with chronic GN who were to undergo kidney biopsy, and from 18 healthy, nonsmoking volunteers (controls). Table 1 summarizes the clinical data of the patients and controls. Kidney biopsy procedures were uneventful. The morphological reports were then retrieved and classification of GN was made according to the criteria of the Collaborating Centre for the Histological Classification of Renal Diseases of the WHO. The patients were divided into those with proliferative GN [type I membrano-proliferative (n 7) and mesangioproliferative (n 8)] and those with non-proliferative GN [membranous (n 4), focal segmental glomerulosclerosis (n 10) and fibrillar GN (n 1)]. Each kidney biopsy specimen was studied with light microscopy, electron microscopy and immunofluorescence for C3 of complement, IgG, IgA, IgM and fibrin. According to the glomerular finding of C3 and at least two other deposits, two subgroups of patients were distinguished: with immune complex GN (ICGN) [type I membrano-proliferative (n 7), mesangioproliferative (n 8) and membranous (n 4)] and with non-ICGN [focal segmental glomerulosclerosis (n 10) and fibrillar GN (n 1)]. Two sections of normal human kidney were obtained from two different cadaveric renal transplant donors whose kidneys were technically unsuitable for transplantation (aberrant renal arteries); they served as controls for immunohistochemical analysis. Before the biopsy, none of the patients received glucocorticosteroids or other immunosuppressive agents, non-steroidal anti-inflammatory drugs, including aspirin, heparin, oral anticoagulants, antiplatelet drugs, or lipid-lowering agents. No patient was seropositive for HIV, hepatitis C virus antibodies or hepatitis B virus antigen; none had biochemical evidence of liver injury and their routine haemostatic tests were normal. Nineteen patients (63.3%) had hypertension and all were treated with angiotensin I converting enzyme (ACE) inhibitors. Twelve patients were receiving diuretics, six were recieving b-blockers and three had compensated chronic heart failure. None of the patients and controls had suffered from any acute inflammatory or infectious diseases in a month preceding the renal biopsy. Study protocol was approved by our institutional ethical committee, and informed consent was obtained from each of the participants. The study was performed in conformity with the Helsinki Declaration. Methods Immunohistochemical analyses For TFPI-1 and TFPI-2 detection, purified goat polyclonal antibody (Santa Cruz Biotechnology Inc., CA, USA: TFPI-1 C-20 sc-18713; TFPI-2 C-20 sc-18719) and goat ABC staining system (Santa Cruz Biotechnology Inc., CA, USA: sc-2023) Age (years) Male/female (n) Serum creatinine (mg/dl) Creatinine clearance (ml/min)* Serum albumin (g/dl) Urinary protein lost (g/1g Cr/24 h) Urinary protein lost (g/24 h) Controls (n 18) Patients (n 30) 45.5 (18.072.0) 11/19 0.90 (0.603.50) 49.3 (14.7129) 2.81 (0.974.05) 0.67 (0.161.16) 1.03 (0.014.5) Data are shown as median (full range) or number. P by MannWhitney U or chi-square tests when appropriate. NS, not significant. *Calculated with Cockcroft and Gault formula. were used. For TF detection, anti-human TF, type 1 mouse antibody (Calbiochem, Darmstadt, Germany: cat. No 612161) and En Vision Dual Link (Dako, Glostrup, Denmark) reagent were used. Diaminobenzidine (DAB) (Dako, Glostrup, Denmark) as chromogen was used for visualization of the results. Biopsy specimens embedded in Jung tissue freezing medium were cut on 4 mm cryostat sections and applied to pre-treated slides. Fixed sections (in cold acetone for 10 min) were washed in phosphate-buffered saline (PBS) (intended for TFPI-1 and TFPI-2 detection) or in Tris (intended for TF detection). The next steps were carried out in a humidified chamber at room temperature. To quench endogenous peroxidase activity, slides were incubated for 10 min in 1% hydrogen peroxide diluted in PBS and for 1 h in 1.5% blocking serum for TFPI-1 and TFPI-2 detection or 3% hydrogen peroxide diluted in Tris for TF detection. Then, the sections were incubated for 30 min with primary antibody diluted 1 : 50. For TFPI-1 and TFPI-2 detection, after this, step sections were incubated with biotinylated secondary antibody for 30 min and with AB enzyme reagent for 30 min, and after the wash with PBS they were incubated with DAB chromogen. For TF detection, slides were incubated with peroxidase-conjugated En Vision Dual Link reagent, and after the wash with Tris they were incubated with DAB. Sections were counterstained with haematoxylin for 5 s. In the negative control the primary antibodies were omitted. The intensity of TF and TFPI 1/2 staining in 30 percutaneous renal biopsy specimens was examined prospectively by two independent histopathologists and scored on a scale from 0 to 3. Glomeruli in which there was no expression of detected antigens were scored 0 and those with weak staining were scored 1, while those with medium staining were 2 and with intense staining were 3 (Figures 1AD and 2AD). Quantitative analysis of TF and TFPI 1/2 staining intensity for each patient was calculated as mean of the total glomeruli score in biopsy specimen (Table 2). Plasma TF and total TFPI antigen In healthy controls and patients, fasting blood samples were obtained without stasis from the antecubital vein punctured with 19-gauge needle. Cell-free plasma was prepared by centrifugation at 3000 g for 20 min at room temperature, aliquoted and stored in plastic tubes at 408C until further processing. Plasma antigens of TF and total TFPI were measured in duplicate using commercial two-antibody sandwich assay kits (Imubind, American Diagnostica Inc., Greenwich, CT, USA). The monoclonal antibodies used in TF assay recognize both free and VIIa-complexed TF particles. The total TFPI assay detects both intact and truncated forms of the inhibitor as well as its complexes with TF/VIIa. The detection limits of the TF and total TFPI assays are 10 pg/ml and 0.36 ng/ml, respectively. The measurements were performed on 400 SFC photometer (SLT-Labinstruments, Gro ding/Salzburg, Austria) calibrated using the supplied Table 2. TF and TFPI 1/2 glomerular staining (immunohistochemistry) in 30 percutaneous renal biopsy specimens scored on a scale from 0 to 3 Score TF (n;%) TFPI 1/2 (n;%) 0 TF, tissue factor; TFPI, tissue factor pathway inhibitor; n;%, number and percentage of patients with particular scores. Quantitative analysis of TF and TFPI 1/2 staining for each patient was performed as described in Methods. by vortex mixing. The resulting sample was then placed on a slow mixer (3000 g) for 30 min at room temperature to ensure complete solubilization, aliquoted and stored in plastic tubes at 408C until further processing. The assay for both TF and total TFPI antigens was conducted in duplicate using an enzyme-linked immunosorbent assay (ELISA), according to the manufacturers instructions. The results were corrected for the dilution of the urine using the creatinine (Cr) concentration of the sample. Final results were expressed as uTF pg/ml per 1 g Cr: uTF (pg/ml)/Cr (g/24 h) and uTFPI ng/ml per 1g Cr: uTFPI (ng/ml)/Cr (g/24 h). reference samples and standards. For calculations of the results, a computer and curve-fitting software were used. The inter- and intra-assay coefficients of variations were <10%. Urinary TF and total TFPI antigen Urine samples were collected from each patient and controls into sterile universal containers without preservative. One milligram of urine was placed in a 1.5 ml microcentrifuge tube and centrifuged at 51 000 g for 90 min at 48C. The supernatant was then discarded and the pellet obtained from the centrifugation process was solubilized in 1 ml of 15 mM b-octyl-glucopyranoside (BOG, Sigma, Poole, Dorset, UK) Statistical analysis The results were presented as mean SD or median (full range) depending on their normal or skewed distribution provided by ShapiroWilks W test. For intra- and intergroup comparison, the non-parametric Friedmans one-way analysis of variance (ANOVA), KruskalWallis ANOVA, MannWhitneys U, chi-square or unpaired Students t-tests were used when appropriate. Bivariate correlations were assessed using Spearmans regression analysis. Each P was two-tailed, and P values <0.05 were considered significant. Computations were performed using Statistica 6.0 PL (StatSoft, Tulsa, OK, USA). TF, tissue factor; TFPI, tissue factor pathway inhibitor. Data are shown as mean 1 SD or median (full range). P by MannWhitney U test. NS, not significant. Controls (n 18) Patients (n 30) 283 145 127 (0805) 132 (01512) 82.2 27.2 0.47 (0.460.48) 0.50 (0.111.37) Plasma TF (pg/ml) Urinary TF (pg/ml) Urinary TF (pg/ml per 1g Cr) Plasma total TFPI (ng/ml) Urinary total TFPI (ng/ml) Urinary total TFPI (ng/ml per 1g Cr) TF and TFPI staining in renal biopsy specimens In normal kidney specimens, TF, TFPI 1 and TFPI 2 antigens were undetectable in glomeruli whereas a distinct staining of both TFPI variants was observed in tubules and interstitial microvessels (Figures 1A and 2A). The pattern of both TFPI 1 and 2 staining overlapped in immunohistochemical slides, therefore they are further discussed as one TFPI 1/2. The overlapping of TFPI 1 and 2 variant patterns in our immunohistochemical analysis probably resulted from using antibodies which were made towards a peptide mapping within the integral region homology in TFPI 1 and TFPI 2. In patients with chronic GN, a total of 311 glomeruli were examined (a mean of 10.2 glomeruli per biopsy, range 325). All biopsies showed glomeruli at various stages of damage. A total of 108 glomeruli (34.7%) from nine of all the patients studied (30%) showed a minor injury. Proliferative lesions were observed in a total of 154 glomeruli (49.5%) from 15 patients (50%) and all of them were fibrin-positive. In diseased kidneys, TF was strongly expressed in glomeruli (both in minimal injury regions as well as in proliferative and necrotizing lesion regions) but was undetectable in tubules (Figure 1BD, Table 2). In contrast, staining for both TFPI variants was observed in glomeruli and tubules (Figure 2BD, Table 2); they were undetectable in glomeruli with minimal damage. Interestingly, the intensity of both TFPI variants expression in tubulointerstitium was similar in control and diseased kidneys. In glomeruli with proliferative lesions, TFPI was not observed on capillary endothelia, but stained weakly in the mesangia. We did not observe a direct correlation between TF and TFPI 1/2 expression in biopsy slides, either in glomeruli or in tubules. Plasma and urinary TF and total TFPI in patients and controls Plasma and uTF levels were markedly higher in patients than in healthy subjects (Table 3). Interestingly, TF was detected in urine of only three out of 18 controls (n1 23.4, n2 98.6 and n3 19.4 pg/ml). Neither plasma nor urinary levels of the markers correlated with the intensity of TF and TFPI 1/2 staining in biopsy specimens. uTF was directly associated with Cr clearance (R 0.489, P 0.006, Figure 3A) and negatively with serum Cr (R 0.416, P 0.022) and urea levels (R 0.594, P 0.003). There were no significant correlations between uTF and the loss of protein in urine (24 h proteinuria, urinary protein per 1 g of excreted Cr) or between uTF and serum albumin, either. uTF also was positively associated with urinary TFPI (R 0.554, P 0.014, Figure 3B). All urinary TFPI data that occurred below the assay detection limit were omitted. uTF was numerically lower in proliferative (n 15) vs non-proliferative GN (n 15) [83 (0617) vs 281 (10805) pg/ml]; the 3.5-fold difference was, however, on the verge of significance (P 0.06; Figure 4A). No differences in uTF [82 (0805) vs 280 (11458) pg/ml; P 0.116] and uTFPI [0.38 (0.250.91) vs 0.47 (0.210.9) ng/ml; P 0.576] were found between the ICGN (n 19) and non-ICGN (n 11) subgroups (Figure 4CD). Discussion Tissue factor, which can be found in the urine, has a molecular weight of 43 kDa and occurs in a lipid-bound form [12]. Several histochemical and immunolocalization studies have addressed the issue of uTF origin to the glomerulus, while others reported signal within renal tubules [13]. However, the low concentrations of uTF observed in patients with tubular pathology and the lack of associations between uTF and both retinol-binding protein and b-N-acetyl-D-glucosaminidase suggest that uTF level is independent of tubular impairment [11]. Some reports showed that uTF may be a valuable marker of some malignant and inflammatory diseases [1,11,12]. Our study indicates that uTF excretion in patients with chronic GN is significantly higher than in healthy controls, and that uTF may serve as a marker of glomerular injury. It has been postulated that uTF may reflect the aetiopathogenesis of GN, especially the form with R = 0.489, P= 0.006 R = 0.554, P = 0.014 0 0 250 500 750 1000 Urinary TF (pg/ml) Urinary TF (pg/ml) Fig. 3. Spearmans regression between uTF antigen and Cr clearance (A) and total urinary TFPI (B) in chronic GN patients. TF, tissue factor; TFPI, tissue factor pathway inhibitor. 600 400 200 0 1.0 P = 0.06 P = 0.116 Proliferative Non-proliferative ICGN Non-ICGN Proliferative Non-proliferative ICGN Non-ICGN P = 0.595 P = 0.576 immune complexes and crescent formations [7,9,14]. High levels of uTF observed in the ICGN are probably related to immunoglobulin and complement deposition, mononuclear cell infiltration and fibrin formation [12]. This may cause direct or indirect activation of the resident glomerular cells or infiltrating inflammatory cells to express increased amounts of TF. The presence of fibrin or fibrin-related products acts as a macrophage-aggregating agent and stimulates mesangial cells directly through cytotoxic effects. In addition, changes in membrane structures cause TF de-encryption that would result in enhanced TF procoagulant activity and cell apoptosis [12]. Therefore, the escalation of glomerular damage might be reflected by the amount of uTF produced and excreted in urine. In this study, we found lower uTF levels in patients with proliferative GN compared with the non-proliferative GN. However, we did not observe any differences between the ICGN and nonICGN patients. The fact that uTF occurred to be lower in proliferative non-crescentic GN could be against the hypothesis that more active inflammation generates and/or releases higher amounts of TF that can be detected in the urine. The issue requires further investigations. The novel finding of our study is also a direct relation between uTF and TFPI. It is plausible that TF is excreted as a complex formed with TFPI. In experimental crescentic GN, TFPI bound to TF (as it is shed) may contribute to its accumulation in acellular regions of the crescents [8]. In human atheromatous plaque, TF and TFPI colocalize in acellular regions of intimal lesions; however, there are no reports of TFPI production by human cells of epithelial origin. In normal human kidney, TFPI is found only in association with endothelium of small interstitial vessels and is absent in glomerulus, which is in contrast to the constitutive expression of TF by epithelial and mesangial cells [15]. Our study confirmed the TFPI localization mainly to interstitial microvessels in normal kidney, but simultaneously we observed the intense staining of both TFPI variants in the tubules and the lack of TF expression in healthy glomeruli. The intriguing TFPI localization in the tubules is reported for the first time, and the reason for this is not clear; it may result from in situ synthesis, reabsorption or both. In diseased kidneys, a strong TF expression in glomeruli, both in minimal injury regions as well as in proliferative and necrotizing lesions, was observed. In contrast, staining for both the TFPI variants was revealed in the glomeruli and tubules; they were, however, undetectable in glomeruli with minimal damage. The reason for this phenomenon is currently unclear. In experimental crescentic GN in rabbits, glomerular TFPI was constitutively expressed and then down-regulated early in the course of the disease, simultaneously with the maximal up-regulation of glomerular TF activity [10,14]. This may contribute to the greater increase of TF activity compared with TF antigen deposition at the early stage of the disease. In human crescentic GN, TFPI is strongly expressed in fibrotic crescents, and its amount is inversely correlated with the presence of fibrin-related antigen, being a sensitive marker of early and active disease [10]. According to Cunningham et al. [8], this late overexpression of TFPI may inhibit TF activity and thus limit fibrin deposition in advanced crescentic GN. No studies, however, addressed this interesting issue in non-crescentic human GN. Anyway, we did not observe the correlation between TF and TFPI 1/2 expression in our biopsies, either in the glomeruli or in the tubules. Also, we found no associations between the intensity of TF and TFPI 1/2 staining in the specimens and their plasmatic and urinary concentrations as well as with the clinical measures of renal function. These are related to the material rather than the method as only one section was examined, which is a common practice in morphometric studies of percutaneous renal biopsy specimens. The small size of the sample implies that it might not be fully representative of the severity of kidney disease, due to possible asymmetrical distribution of the lesions. This was exemplified by the lack of correlation between the high index of chronic damage in biopsy specimen and the clinical advancement of renal failure in a number of cases [16]. Also, Lwaleed et al. [12] found no associations between uTF activity and either tuft morphological changes or interstitial cell infiltration in patients with diffuse GN. On the other hand, it is well-known that the Cockcroft and Gault formula used for Cr clearance calculation tends to overestimate the glomerular filtration rate in advanced renal failure and to underestimate it at early stages of the disease [17]. This limitation may have also contributed to the weak relations between immunohistochemical and biochemical findings in our chronic GN patients. Interestingly, we found that the uTF level remarkably and positively correlated with renal excretory function, which is in contrast to the findings of Lwaleed et al. [11]. This is of importance if the assay is to be used in clinical practice. So far only a weak correlation between uTF and urinary protein excretion (expressed as protein/Cr index) has been reported [11]. However, no association between uTF and proteinuria or serum albumin was observed in our patients; this could hypothetically be the consequence of the almost universal use of ACE inhibitorspotent antiproteinuric, notably antithrombotic, actually pleiotropic and thus likely confounding drugs [1820]. In conclusion, intrarenal TF/TFPI system is profoundly disturbed in chronic non-crescentic GN and may be involved in the development and/or progression of the disease. Plasma and urinary concentrations of TF and TFPI do not reflect genuine activity of GN, likely due to the confounding effects of renal failure and pharmacotherapy. Further studies are required to evaluate the utility of uTF as a marker of glomerular injury. Acknowledgements. The authors thank Dr Oksana Kovalchuk for her technical assistance. This work was supported by grant No 3-54 958 from the Medical University, Bialystok, Poland. Conflict of interest statement. None declared.


This is a preview of a remote PDF: http://ndt.oxfordjournals.org/content/21/12/3450.full.pdf

Beata Naumnik, Jacek Borawski, Lech Chyczewski, Krystyna Pawlak, Michal Mysliwiec. Tissue factor and its inhibitor in human non-crescentic glomerulonephritis—immunostaining vs plasma and urinary levels, Nephrology Dialysis Transplantation, 2006, 3450-3457, DOI: 10.1093/ndt/gfl365