A characteristic finding of childhood HIV-associated hemolytic uremic syndrome (HIV-HUS) is the presence of endothelial injury and microcystic tubular dilation, leading to a rapid progression of the renal disease. We have previously shown that a secreted fibroblast growth factor-binding protein (FGF-BP) is upregulated in kidneys from children affected with HIV-HUS and HIV nephropathy. Here, we sought to determine the potential role of FGF-BP in the pathogenesis of HIV-HUS. By immunohistochemical and in situ hybridization studies, we observed FGF-BP protein and mRNA upregulation in regenerating renal tubular epithelial cells from kidneys of HIV-Tg26 mice with late-stage renal disease, that is, associated with the development of microcystic tubular dilatation and accumulation of FGF-2. Moreover, FGF-BP increased the FGF-2-dependent growth and survival of cultured primary human renal glomerular endothelial cells and enhanced FGF-2-induced MAPK/ERK2 activation, as well as the proliferation of immortalized GM7373 endothelial cells. We propose that HIV-Tg26 mice are a clinically relevant model system to study the role of FGF-BP in the pathogenesis of HIV-associated renal diseases. Furthermore, the upregulation of FGF-BP by regenerating renal tubular epithelial cells may provide a mechanism by which the regenerative and angiogenic activity of FGF-2 in renal capillaries can be modulated in children with HIV-HUS and other renal disease.
- renal tubular epithelial and endothelial cells
- HIV-transgenic mice
the term hemolytic uremic syndrome (HUS) was first introduced in 1955 by Gasser et al. (17) to describe a heterogeneous group of diseases characterized by microangiopathic hemolytic anemia, thrombocytopenia, and acute renal failure. The classic form of HUS is a well-defined clinical entity in which there is a prodrome of diarrhea, usually bloody, as a result of Shiga-like, toxin-producing bacterial infection (34). In addition, other atypical and familiar forms of HUS, not associated with diarrhea, have been described in association with drugs, inborn errors of metabolism, genetic diseases, and infectious agents (34). HIV-1-infected children can develop an atypical form of HUS characterized by the presence of severe microangiopathic lesions in renal capillaries, in association with microcystic tubular dilation, tubulointerstitial changes, and rapid development of end-stage renal disease (37, 46).
Endothelial damage appears to be a primary pathogenic event in all of the classic and atypical forms of HUS (34). We have demonstrated that FGF-2, an endothelial cell growth factor that can be released by injured endothelial cells, is accumulated in the plasma and urine of children with classic and HIV-HUS, in correlation with the severity of the renal disease (31, 36, 46). In normal conditions, FGF-2 is stored as an inactive form in the extracellular matrix bound to heparan sulfate proteoglycans (HSPG) (18, 22, 40, 52). Interestingly, we have found a significant overexpression of FGF-2 and HSPG in the kidneys of both HIV-transgenic (Tg26) mice (33, 36, 37, 46) and HIV-infected children with renal disease (26). These findings suggest that FGF-2 released by injured endothelial cells can be sequestered from the circulation and stored in the kidneys bound to renal HSPG (26, 31, 33, 36). However, it is crucial to determine how FGF-2 is mobilized from the HSPG milieu to reach and activate its cognate receptor(s) on the target renal endothelial cells.
Fibroblast growth factor-binding protein (FGF-BP) was first characterized as a secreted protein that is able to bind to FGF-2 in a noncovalent, reversible manner and protect it from proteolytic degradation (51). Studies with FGF-BP-negative cell lines showed that expression of FGF-BP increased tumor growth and angiogenesis in a xenograft model (6–8, 50). These and other findings suggest that FGF-BP has a rate-limiting role modulating the angiogenic activity of FGF-2 during the process of tumor growth and embryo development (6, 27, 45). In previous studies, we have also found that FGF-BP is upregulated in regenerating renal tubular epithelial cells in children with HIV-HUS and other renal diseases (26). On the basis of these observations, it is tempting to speculate that the interaction of FGF-BP with FGF-2 may impact on FGF-2-mediated regeneration of renal capillaries adjacent to renal epithelial cells.
In previous studies, we have found that renal epithelial cells isolated from HIV-transgenic (Tg26) mice with renal disease synthesize and release significant levels of FGF-2 into the conditioned media (33). Here, we show for the first time that FGF-BP is upregulated in regenerating renal tubular epithelial cells in the same mouse model. Additionally, we demonstrate the ability of FGF-BP to increase FGF-2-induced survival activity of cultured primary renal glomerular endothelial cells, as well as to enhance FGF-2-dependent MAPK/ERK signaling and mitogenesis in bovine endothelial cells. Overall, when these findings are interpreted in the context of previous studies, they support the notion that FGF-BP, locally released by regenerating tubular epithelial cells, may enhance the angiogenic activity of FGF-2 in children with HIV-HUS and, in this manner, facilitate the healing of renal capillaries and tubules.
MATERIALS AND METHODS
HIV-1 Transgenic Mice
HIV-1 transgenic heterozygous mice (Tg26 line) have been previously described (11, 23). Plasma blood urea nitrogen (BUN) and creatinine levels, as well as urine protein/creatinine ratios in HIV-Tg26 and littermate nontransgenic control mice, were measured following standard techniques as previously described (23, 35).
This study was approved by the Institutional Review Board from the Children's National Medical Center. Renal sections obtained at autopsy from children with HIV-HUS (n = 2 ), HIV-nephropathy (n = 2), classic HUS (n = 1), and renal sections from HIV-infected children without renal disease (n = 2) were collected to validate the findings obtained in HIV-Tg26 mice. All patients were younger than 16 years of age. The diagnosis of HIV infection, HIV-HUS, HIV nephropathy, and classic HUS was done using standard clinical and pathologic criteria.
Renal tissues were fixed in 10% neutral buffered formalin and embedded in paraffin. Four-micrometer sections were deparaffinized by xylene and rehydrated with decreasing concentration of ethanol and finally H2O. Antigen retrieval was performed by heating the sections twice for 5 min in a microwave oven (2,450 mHz, 850 W) in a 0.01 M sodium citrate solution (pH 6.0). Endogenous peroxidase activity was blocked with 3% H2O2 in 100% methanol for 10 min. Tissues were incubated for 1 h at room temperature with affinity purified IgG fractions (2.5 μg/ml) from a rabbit polyclonal antibody directed against unique peptide sequence of FGF-2 (provided by Dr. Baird, PRIZM Pharmaceuticals, San Diego CA); anti-bovine FGF-2 mouse monoclonal antibody (Upstate Biotechnology, Lake Placid, NY); rabbit polyclonal FGF-BP antibody (1:50 dilution) (8); and biotinylated anti-proliferative cell nuclear antigen (anti-PCNA) monoclonal antibody (clone PC10) (Zymed Laboratories, South San Francisco, CA). Nonimmune mouse or rabbit IgGs were used as negative controls. The specificity of FGF-BP staining in control and diseased renal sections was confirmed by preabsorbing the primary FGF-BP antibody with a 20-fold molar excess of recombinant FGF-BP. Secondary staining was performed with a commercial streptavidin-biotin-peroxidase complex Histostain SP kit (Zymed) and aminoethyl carbazole, according to the manufacturer's instructions. Counterstaining of nuclei was performed by immersion in hematoxylin. Slides were dehydrated with increasing concentrations of ethanol and a final immersion in Xylene 100% and then mounted with Cytoseal 60 low-viscosity mounting medium (Richard-Allan Scientific, Kalamazoo, MI), and coverslips. Immunostaining for von Willebrand factor (vWF) (DAKO, Carpinteria, CA) was carried out as described previously (36, 38).
In Situ Hybridization for FGF-BP mRNA Expression in Tissues
Digoxigenin-labeled mouse and human FGF-BP probes.
A 750-bp mouse FGF-BP cDNA (GenBank #U49641) was subcloned into the EcoR I sites of the pRC/CMV vector between T7 and SP6 promoters. The vector was linearized with the restriction enzymes Hind III or Not I and utilized as a template for the synthesis of DIG-labeled antisense or sense riboprobes, respectively. In vitro transcription reactions were performed with SP6 or T7 RNA polymerase in the presence of DIG-UTP by using a DIG RNA labeling Kit (Roche, Indianapolis, IN). The reaction products were purified by phenol extraction and ethanol precipitation and stored at −70°C until use. Labeling efficiency of the riboprobe was estimated by comparison with 10-fold serial dilution of a digoxigenin-labeled control riboprobe and direct detection of the labeled riboprobe with anti-digoxigenin antibodies. Riboprobe concentrations were adjusted to be equivalent on the basis of the labeling efficiency before use in the in situ hybridization studies.
In situ hybridization.
Ten percent neutral buffered formalin-fixed and paraffin-embedded renal tissues were cut at 4 μm and floated onto 2% 3-aminopropyltriethoxysilane (Sigma Chemical, St. Louis, MO)-coated slides. Sections were heat-fixed for 30 min at 65°C and deparaffinized. After a 30-min incubation in 5 mmol/l levamisole, the sections were washed in PBS and in diethyl pyrocarbonate-treated water for 5 min, respectively, and then immersed in 0.2 mol/l HCl for 20 min. Then, sections were deproteinized by digestion with 20–40 μg/ml proteinase K (Sigma) for 10 min at 37°C, washed twice in 0.2% glycine for 5 min, once in PBS for 10 min, and postfixed with 1.5% paraformadehyde-1.5% glutaraldehyde for 1 min. After being rinsed twice in PBS, sections were dehydrated through a graded ethanol series and air-dried. The sections were hybridized with the mouse FGF-BP RNA sense or antisense probes (0.1–0.5 ng/μl) in 2× SSC, 10% dextran sulfate, 1× Denhardt's solution, 20 mM vanadyl ribonucleoside complex (Invitrogen, Carlsbad, CA), 0.1 M sodium phosphate under sealed coverslips and incubated overnight in a moist chamber at 42°C. Sections were washed in 0.2× SSC and then blocked with a blocking solution (50 mg/ml skimmed dried milk, 150 mM NaCl, 100 mM Tris·HCl, pH 7.8) at room temperature for 15 min followed by an incubation with alkaline phosphatase-conjugated antidigoxigenin antibody (Roche) (1:750 dilution) for 30 min at 37°C. Colorimetric detection with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Roche) was then performed, and the sections were mounted in 60% glycerol and examined by light microscopy. The negative controls included 1) hybridization with the sense probe, 2) RNase A (100 μg/ml in 10 mM Tris·HCl, pH 8.0, 1 mM EDTA) pretreatment before hybridization, and 3) omission of either the antisense RNA probe or the anti-digoxigenin antibody.
Total RNA was extracted from control and diseased tissue samples or cells using Trizol reagent (Invitrogen), according to the manufacturer's instructions. RT of 2 μg of total RNA was performed using the Reverse Transcription System (Promega, Madison, WI). The resulting first-strand cDNA was used as a template for a PCR reaction using the Advantage PCR Kit (Clontech, Palo Alto, CA) at an annealing temperature of 58°C for 30 cycles. The following primers were used for amplification: FGF-BP forward: 5′-GCAAGTTTG26TCACCAAAGACC-3′, FGF-BP reverse: 5′-CTG26CTTTCTCCTTCCTG26GGC-3′; β-actin forward: 5′-TG26CGTTACACCCTTTCTTG26AC-3′, β-actin reverse: 5′-ACTG26GTCTCAAGTCAGTG26TAC-3′.
PCR products were separated on 1.2% agarose gels and stained with ethidium bromide. Amplified cDNA bands were quantified by densitometry image scanning as described before (8). The product yield was expressed relative to β-actin.
Culture of endothelial cells.
Primary human glomerular endothelial cells were purified by elutriation from dispase-dissociated normal human cortical glomeruli (Cell Systems, Seattle, WA) and were a generous gift from Dr. Carl Soderland [Applied Cell Biology Research Institute (ACBRI), Kirkland, WA; Ref. (44)]. Cells were used between passages 2 and 4 and were seeded on Attachment Factor-coated Nunclon cell culture flasks (ACBRI). Primary glomerular endothelial cells were maintained in Cell System-Certified endothelial culture medium (ACBRI), containing the following endothelial cells growth supplement (ECGS): recombinant human epidermal growth factor (10 ng/ml), bovine brain extract, heparin (10 μg/ml), amphotericine B (50 ng/ml), gentamycin (50 mg/ml), and 5% (vol/vol) fetal bovine serum. Transformed fetal bovine aortic endothelial GM7373 cells, corresponding to the BFA-1c 1BPT clone (4, 19), were obtained from the National Institute of General Medical Sciences, Human Genetic Mutual Cell Repository, Coriell Institute for Medical Research (Camden, NJ) and were maintained in modified improved MEM (Invitrogen) supplemented with 10% (vol/vol) fetal bovine serum.
Survival assays of human glomerular endothelial cells.
Cells were plated in 6-well culture dishes, in the presence of ECGS. At 80% confluence, cells were starved for 12 h; then FGF-BP (30 ng/ml) and FGF-2 (5 and 30 ng/ml) (Invitrogen) were added, and the cells were cultured for 18 h. Cells were then incubated with 1 μCi/ml [3-H]-labeled thymidine for 6–12 h followed by the addition of 1 ml ice-cold 10% TCA for 20 min at 4°C. After the cells were washed, three times with ice-cold water, they were solubilized with 1 ml 0.5 M NaOH/16 mm dish at 37°C for 30 min; 0.5-ml aliquot samples were measured. The results were expressed in counts per minute.
Proliferation assay of immortalized endothelial cells.
GM7373 cells (104) were seeded in triplicate in 96-well plates for 8 h in the presence or absence of sodium chlorate (30 mM; Sigma) and serum-starved overnight. Cells were then stimulated with 5 and 10 ng/ml of human recombinant FGF-2 (Invitrogen) in the presence or absence of recombinant human FGF-BP (6 ng/ml) for 48 h. Cell proliferation was assessed using the tetrazolium-based WST-1 reagent (Roche), according to the manufacturer's instructions.
Forty percent confluent GM7373 cells were plated in complete medium in the presence or absence of 50 mM sodium chlorate for 36 h. Cells were then serum-starved overnight in the presence or absence of sodium chlorate and treated for 5 min with 0.5, 1, and 2 ng/ml of human recombinant FGF-2 (Invitrogen) in the presence or absence of FGF-BP (5 ng/ml) (45). Cells were then washed with cold PBS, pH 7.4, and subsequently lysed at 4°C in a buffer containing 50 mM Tris·HCl, pH 8, 150 mM NaCl, 40 mM b-glycerophosphate, 1 mM EGTA, 0.25% sodium deoxycholate, 1% Nonidet P-40, 50 mM sodium fluoride, 20 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 2 mg/ml leupeptin, 2 mg/ml aprotinin, 1 mg/ml pepstatin, and 100 mg/ml pefabloc. Detection of phosphorylated proteins and phosphorylated ERK1/2 from total cell lysates was carried out using an anti-phosphotyrosine monoclonal antibody (4G10; Upstate Biotechnology, Lake Placid, NY) and an anti phospho-p44/42 MAP Kinase (Thr202/Tyr204) rabbit polyclonal antibody (Cell Signaling, Beverly, MA), respectively, as described previously (45).
Isolation of recombinant FGF-BP protein was done as described previously (45).
GraphPad Prism Software was used for statistical analysis. Data are expressed as means ± SE. Differences between two groups were compared by Student's t-test. When more than two means were compared, data were evaluated for statistically significant differences by one-way ANOVA followed by multiple comparisons using the Student-Neuman-Keuls test. The nonparametric Kruskal-Wallis test was used whenever the distribution of the data did not follow a Gaussian distribution. GM7373 proliferation was analyzed by two-way ANOVA with Bonferroni's post hoc test. Statistical significance was defined as P < 0.05.
Indicators of Late-Stage Renal Disease in HIV-Tg26 Mice
The natural history of renal disease in HIV-Tg26 transgenic mouse model has been characterized (23, 33). Interestingly, although born with normal kidneys, HIV-Tg26 mice, at ∼40–50 days of life, develop renal disease with high incidence and with many of the pathologic features typical of human immunodeficiency virus associated nephropathy (HIVAN). These renal lesions include hypertrophy and hyperplasia of glomerular epithelial cells, focal and global glomerulosclerosis, varying degrees of tubular distortion and dilation, as well as kidney enlargement (23). These pathologic features are also associated with the development of severe proteinuria and azotemia, as indicated by the following parameters from control vs. HIV-Tg26 mice: urinary protein/creatinine ratios 8.5 ± 0.8 vs. 51 ± 6.6; plasma BUN 23 ± 1.5 vs. 43 ± 4.8 mg/dl; plasma creatinine 0.45 ± 0.0016 vs. 0.6 ± 0.009 mg/dl (mean ± SE, n = 5; P < 0.05 for all parameters compared with controls, respectively).
FGF-BP Protein Upregulation During Late-Stage Renal Disease in HIV-Tg26 Mice
To determine localization and pattern of expression of FGF-BP during disease progression in HIV-Tg26 mouse kidneys, we performed immunohistochemical studies on renal sections from transgenic and wild-type mice. At 20 days of age, HIV-Tg26 mice with no histological evidence of renal disease showed minimal or no expression of FGF-BP in renal tubules (Fig. 1, Bj), with levels comparable to those obtained from 20-day-old control nontransgenic littermates (Fig. 1, Bi). However, 40-day-old HIV-Tg26 mice, showing microcystic dilation, as well as regenerative changes of renal tubules, displayed a significant upregulation of FGF-BP (Fig. 1, Bf). It is noteworthy that no FGF-BP staining was detected in kidney specimens from nontransgenic littermates or from 40-day-old HIV-Tg26 mice stained with nonimmune antiserum (Fig. 1, B, e and g). The expression of FGF-BP was predominantly upregulated in renal tubular epithelial cells, which also stained positive for the proliferating nuclear antigen (PCNA) (Fig. 1, Bh). We detected a significant upregulation (∼ 8-fold increase) in the number of proliferating renal tubular cells compared with that of HIV-Tg26 mice without renal disease or control littermates (not shown). These findings correlate with FGF-2 accumulation in the renal interstitium surrounding renal tubular epithelial cells undergoing proliferative changes in HIV-Tg26 mice (33). Moreover, FGF-BP expression in HIV-Tg26 renal tubules coincided with that of PCNA (Fig. 1, Bh), suggesting that FGF-BP synthesis is predominantly restricted to regenerating tubular epithelial cells.
FGF-BP Protein Upregulation in the Kidneys of HIV-Tg26 Mice Mimics the Renal Changes Seen in Children with HIVAN
In a previous study, we had shown upregulation of FGF-BP in the kidneys of children with HUS and other renal disease (26). Renal tubules of children with HIVAN showed a significant upregulation of FGF-BP protein in contrast to renal sections from HIV-infected children without renal disease (Fig. 1, A, a and b). Specificity of the FGF-BP staining was assured by preabsorption of the antibody with recombinant human FGF-BP protein (Fig. 1, Ac). As in the HIV-Tg26 mice, the accumulation of FGF-BP in human HIVAN was predominantly observed in renal tubules undergoing regenerative/proliferative changes, as evidenced by double-staining for PCNA and FGF-BP (Fig. 1, Ad). Thus the transgenic model appears to mimic the human disease.
FGF-BP mRNA Upregulation During the Late-Stage Renal Disease in HIV-Tg26 Mice
The upregulation of FGF-BP protein in the transgenic model was paralleled at the mRNA level. In situ hybridization for FGF-BP mRNA in sections from diseased kidneys from HIV-Tg26 mice showed the same pattern of distribution of FGF-BP mRNA and protein. A significant upregulation of FGF-BP mRNA was detected in regenerating and/or injured renal tubules from 40-day-old HIV-Tg26 mice with renal disease (Fig. 2, Ab). Signal specificity was confirmed by the absence of staining in sections incubated with a sense FGF-BP probe (Fig. 2, Aa). Furthermore, no staining was detected in renal sections from young (20-day-old) or adult (40-day old) HIV-Tg26 without renal disease or control littermates (data not shown). Similarly, RT-PCR of diseased kidneys from 40-day-old HIV-Tg26 mice showed a 2.2-fold increase of FGF-BP mRNA, compared with the levels obtained from control littermate kidneys. Although the differential FGF-BP mRNA expression detected from the RT-PCR analysis appears smaller than that observed from the in situ hybridization, it is likely to be due to the dilution effect of mRNA from other cell types in the tissue. It is noteworthy that no significant differences of FGF-BP mRNA expression were detected between control and HIV-Tg26 mice during the early stages of renal disease (data not shown). We conclude from these data that the renal tubular epithelial cells from HIV-Tg26 mice induce expression of FGF-BP mRNA and do not accumulate the protein from the renal interstitium.
Effect of FGF-BP on the Growth and Survival of Primary Glomerular Endothelial Cells
Primary human endothelial cells isolated from renal glomeruli were used to assess the effect of FGF-BP on cell growth. Cytoplasmic expression of von Willebrand factor was used as histological marker to confirm the endothelial characteristics of these cells (Fig. 3A) (36, 38). In the absence of added endothelial cell growth factors, primary glomerular endothelial cells detached from the culture dishes and underwent cell death but were rescued by the addition of FGF-2 (Fig. 3B). The viability of the cultured primary glomerular endothelial cells in the presence of FGF-2 was confirmed by [3H]-labeled thymidine incorporation studies (Fig. 3C). The addition of FGF-BP alone did not affect the growth rate or viability of the cells. However, inclusion of FGF-BP with FGF-2 significantly increased FGF-2-mediated cell survival (P < 0.01). These data, altogether, indicate that the interaction of FGF-BP with FGF-2 is associated with a significant enhancement of FGF-2-induced effects on primary human renal endothelial cells. Therefore, the upregulation of both FGF-2 and FGF-BP detected in HIVAN may thus have a functional impact also on the epithelial/endothelial interactions.
Interaction of FGF-BP, FGF-2 and Heparan Sulfates in Endothelial Cells
FGF-2, like most other FGFs, binds to extracellular HSPGs that serve to store these growth factors and modulate FGF receptor signaling. We had reported earlier that different glycosaminoglycans inhibit FGF-1 and FGF-2 binding to FGF-BP (45), and thus we hypothesized that heparan sulfates would also impact on the FGF-BP interaction with FGF-2 in intact cells. To investigate these interactions with respect to endothelial signaling, we selected an established, immortalized endothelial cell line, GM7373, and used a standard protocol to modulate the proteoglycan sulfate content of the cells by growing the cells in the absence or presence of sodium chlorate. After sodium chlorate pretreatment with 10 to 50 mM, a >70% reduction in proteoglycan sulfation of GM7373 and bovine aortic endothelial cells was reported by different laboratories (3, 30) and this protocol was used by others to evaluate the impact on FGF signaling in different cell types and via different FGF receptor subtypes (9, 10, 20). Here, we stimulated the immortalized endothelial cells with and without sodium chlorate pretreatment and different concentrations of FGF-2 ± FGF-BP and analyzed cells for the induction of protein phosphorylation or followed the growth response of the cells. Anti-phosphotyrosine blots of cells treated with FGF-2 for 5 min showed a distinct induction of p42/p44 at 2 ng/ml of FGF-2. Addition of FGF-BP reduced this threshold concentration of FGF-2 needed to initiate signal transduction to 1 ng/ml (Fig . 4A, top). After growth of the cells in the presence of sodium chlorate, the stimulation by FGF-2 is seen at 1 ng/ml, and the addition of FGF-BP reduced this threshold concentration of FGF-2 to 0.5 ng/ml (Fig. 4A, bottom). The nature of the p42/p44 bands was confirmed by immunoblotting the same whole cell lysates with an anti phospho ERK antibody (Fig. 4B). We conclude from this that less FGF-2 is required for signaling when FGF-BP is present.
Induction of cell growth was monitored 48 h after growth factor stimulation and required higher concentrations of FGF-2 than the threshold concentrations needed for signaling. Cell proliferation was induced by 2.5-fold with 10 ng/ml of FGF-2. The inclusion of FGF-BP enhanced this FGF-2-induced cell proliferation to 14.7-fold (P < 0.01) (Fig. 4C, left). After chlorate pretreatment of the cells, FGF-2 at 5 and at 10 ng/ml did not stimulate proliferation of the cells. Inclusion of FGF-BP, however, led to a maximum stimulation at both concentrations of FGF-2 (Fig. 4C, right). These data suggest to us that heparan sulfate proteoglycans can modulate the FGF-BP interaction with FGF-2, as well as the resulting cell signaling in endothelial cells. Moreover, FGF-BP appears to be able to supplement heparan sulfate function required for FGF growth stimulation in these cells. A model to explain these interactions is shown in Fig. 5 and discussed below.
In recent years, several studies have demonstrated the important role of FGF-2 in the pathogenesis of renal diseases (12, 14, 24, 28, 35, 53, 54). However, to date, it is not clearly understood how FGF-2 trapped in the renal extracellular matrix can induce angiogenic effects in renal capillaries (49). In this study, we describe a potential mechanism by which the renal angiogenic activity of FGF-2 in children with HIV-HUS could be controlled. More specifically, we have found that a secreted FGF-binding protein produced in renal tubular epithelial cells of both HIV-Tg26 mice and HIV-infected children with renal disease (26) enhances the mitogenic activity of FGF-2 and the survival of cultured human renal glomerular endothelial cells. These findings, when interpreted in the context of our previous studies, may be relevant for the pathogenesis of HIV-HUS, HIV-associated nephropathy, and classic HUS (26, 31, 32).
FGF-1 and FGF-2 lack a classic signal peptide for secretion and are not released from cells following a classic secretory pathway (25, 49). Both growth factors are deposited into the extracellular matrix of the cells and tightly bound to HSPGs, and the biological activity of these FGFs is quenched by this binding (2, 22, 29, 40, 43, 47, 49, 52). The mechanisms by which FGF-2 is released and activated in the context of HUS are not clearly understood. We have found high levels of FGF-2 in the plasma and urine of HIV-infected children with classic and HIV-HUS. These patients also show an increased expression of renal heparan sulfate proteoglycans that may facilitate the renal accumulation of FGF-2 (31, 36). In addition, heparanases release by activated peripheral blood mononuclear cells (16, 48), and/or proteolytic enzymes, known to be activated in HUS (i.e, plasmin, thrombin), can digest the protein backbone of the HSPG and release biologically active FGFs from the renal extracellular matrix (1, 13, 22, 47). Here, we report that FGF-BP may represent an alternative mechanism by which the angiogenic activity of FGF-2 can be regulated in the kidney of HIV-infected children with renal disease (26).
Although FGF-PB induced only a modest increase in the FGF-2-induced survival of primary human renal glomerular endothelial cells (see Fig. 3), these results may be biologically relevant in vivo. It should be considered that lethally injured glomerular endothelial cells release FGF-2, and therefore, the ability of FGF-BP to modulate the angiogenic activity of FGF-2 decreases progressively in correlation with the accumulation of FGF-2 in the culture media, which prompted us to also use immortalized endothelial cells for more extensive parallel studies (see Fig. 4). FGF-BP may play a critical role, in particular, during the early stages of renal tubular injury, in which relatively low concentrations of FGF-2 are initially released into the renal interstitum and FGF-2 needs to reach its specific target cells in a biologically active form. The reduction of extracellular heparan sulfates in the immortalized endothelial cells reveals the effect of FGF-BP on FGF-2 growth signaling and supports the notion of a positive modulatory role for FGF-BP (Fig. 4C, right). In addition, in vivo, FGF-BP may further increase the angiogenic activity of FGF-2 by preventing its proteolytic degradation (2, 29, 43). FGF-2 also binds specifically and with high affinity to fibrin and fibrinogen (41, 42), which can potentiate its mitogenic activity. In this manner, fibrin can provide a matrix to support the angiogenic activity of the FGF-2/FGF-BP complex and regulate the process of fibrinolysis and thrombus reorganization, which are essential for the recovery of renal function (34). Alternatively, the FGF-2/FGF-BP complex may also modulate the outcome of HUS by inducing hemodynamic changes through changes in vascular tone (54) contractility (32) and blood pressure (5). Taken together, our data suggest that FGF-BP, acting in synergy with FGF-2, may enhance the regeneration of renal capillaries in children with HIV-HUS by different mechanisms. It is tempting to speculate that, in this manner, FGF-BP may participate in a feedback mechanism to facilitate the regeneration of renal tubules by promoting the survival, regeneration, and proliferation of peritubular endothelial capillaries located in close proximity to the injured tubules (21).
The accumulation of FGF-2 (33) and upregulation of FGF-BP in the kidney of HIV-Tg2626 mice and HIV-infected children with renal disease strongly suggest that both factors may play a relevant role in the pathogenesis of the renal disease. Interestingly, we found that FGF-2 (33) and FGF-BP are upregulated only during the late stages of the renal disease and in association with the development of the tubulointerstitial proliferative lesions in these mice. Since the expression of HIV-1 genes in the kidney is significantly reduced during the late stage of the renal disease (33, 39), our findings suggest that FGF-2 and FGF-BP may play an active role in the pathogenesis of these renal proliferative changes. In support of this notion, we have previously reported that renal tubular epithelial cells harvested from the urine of children with HIV-HUS express high levels of FGF-BP and FGF-2 and that FGF-BP enhances the proliferation of these cells in the presence of FGF-2 (26, 36, 37, 46). Our findings in the HIV-Tg26 model are also clinically relevant, because they mimic the upregulation of FGF-BP seen in renal sections from children with HIV-HUS and HIVAN (26). Thus these mice should provide an excellent small animal model system to determine the role of FGF-BP in the pathogenesis of HIV-associated renal diseases in vivo.
In summary, to the best of our knowledge, we have demonstrated for the first time the presence and upregulation of the secreted FGF-BP in renal tubular epithelial cells of HIV Tg26 mice with renal disease, which mimics the clinical situation seen in HIV-infected children. In addition, we have described a novel mechanism by which the renal angiogenic activity of FGF-2 can be modulated in the context of childhood HIV-HUS. These findings may not be limited to the pathogenesis of HIV-HUS and could be relevant for children with HIV-associated nephropathy (26), classic HUS (31), and renal development (26). Obviously, FGF-BP provides another mechanism of fine-tuning for FGF receptor signaling that is also impacted by extracellular matrix heparan sulfate proteoglycans (15), and we provide a model that would encompass these different players (Fig. 5).
This study was supported by National Institutes of Health Grants R-O1-DK-49419 and HL-55605 (P. E. Ray); R-O1-CA71508 (A. Wellstein) and Fundación Argentina para el Desarrollo Infantil (P. E. Ray).
We thank Dr. Carl Soderland from the Applied Cell Biology Research Institute, Seattle, WA, for providing primary human renal glomerular endothelial cells.
↵* These authors contributed equally to the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society