|Home | About | Journals | Submit | Contact Us | Français|
Congenital and acquired deficiencies of complement regulatory proteins are associated with pathologic complement activation in several renal diseases. To elucidate the mechanisms by which renal tubular epithelial cells (TECs) control the complement system we have examined the expression of complement regulatory proteins by the cells. We found that Crry is the only membrane bound complement regulator expressed by murine TECs, and its expression is concentrated on the basolateral surface. Consistent with the polarized localization of Crry, less complement activation was observed when the basolateral surface of TECs was exposed to serum than when the apical surface was exposed. Furthermore, greater complement activation occurred when the basolateral surface of TECs from Crry−/−fB−/− mice was exposed to normal serum compared with TECs from wild-type mice. Complement activation on the apical and basolateral surfaces was also greater when factor H, an alternative pathway regulatory protein found in serum, was blocked from interacting with the cells. Finally, we injected Crry−/−fB−/− mice and Crry+/+fB−/− mice with purified factor B (an essential protein of the alternative pathway). Spontaneous complement activation was seen on the tubules of Crry−/−fB−/− mice after injection with factor B, and the mice developed acute tubular injury. These studies indicate that factor H and Crry both regulate complement activation on the basolateral surface of TECs and that factor H regulates complement activation on the apical surface. Congenital deficiency of Crry or reduced expression of the protein on the basolateral surface of injured cells, however, permits spontaneous complement activation and tubular injury.
Uncontrolled complement activation causes renal injury in a number of diseases including immune-complex glomerulonephritis (1–3), ischemic acute kidney injury (AKI) (4, 5), and acute renal allograft rejection (6). In addition, complement activation contributes to the progression of proteinuric renal diseases (7, 8). Recent reports have strongly linked functional deficiencies of the complement regulatory proteins membrane cofactor protein (MCP; CD46) and factor H with the development of atypical hemolytic uremic syndrome (aHUS) (9, 10). Furthermore, inadequate regulation of fluid phase alternative pathway activation is a primary cause of dense deposit disease (DDD) (11, 12).
All cells in the body express complement regulatory proteins, but the kidney appears to be particularly susceptible to injury in patients with DDD and aHUS caused by defective complement regulation, despite the potential of these systemic defects to damage any organ system (10, 13). The predilection for the kidney as the site of injury suggests that control of the complement system within the kidney by endogenous complement proteins is easily overwhelmed or disrupted, and that inadequate control of the complement system predisposes individuals to injury of the kidney.
The alternative pathway of complement appears to mediate renal injury in most of the diseases associated with defective complement regulation (5, 11, 14, 15). The alternative pathway is continually activated in the fluid phase through a process called “tickover”. Tickover generates C3b which can bind to amino and hydroxyl groups on the surface of cells. Bound C3b then catalyzes further alternative pathway activation unless actively inhibited by proteins such as MCP or factor H. Because invasive pathogens typically lack these regulatory proteins, the C3b generated by tickover can bind to the pathogen surface and trigger further complement activation, helping to eliminate the invading organism. Regulation of the alternative pathway, therefore, is an important mechanism by which the innate immune system discriminates between host and pathogen.
Because there is continuous auto-activation of the alternative pathway, adequate expression of complement regulatory proteins by host cells is critical for preventing autologous injury by the complement system. The complement regulatory proteins expressed on human cells are MCP, decay accelerating factor (DAF; CD55) and CD59 (16). Factor H is a circulating protein that regulates the alternative pathway in the fluid phase and also controls activation on cell surfaces (17, 18). All three membrane-bound complement regulatory proteins are expressed within the kidney, but MCP is the only inhibitor found in abundance on human TECs (19). MCP is a transmembrane protein that regulates both the classical and the alternative pathway of complement by acting as a cofactor for factor I mediated cleavage of C4 and C3 activation fragments (20). Crry is the murine homologue of MCP, and we have previously found that Crry is expressed on the basolateral surface of mouse tubular epithelial cells in vivo. Similar to what has been found in human kidneys, DAF and CD59 were not detected by immunofluorescence microscopy of mouse tissue (21).
Expression of Crry at the surface of the renal tubules is disrupted by ischemia, and complement activation on the basolateral surface of the tubules correlates with the loss of Crry at this location (21). Furthermore, mice partially deficient in Crry display greater susceptibility to ischemic injury than wild-type controls (21). In addition, Bao et al performed cross-transplantation of kidneys from Crry−/−C3−/− mice into wild-type hosts and demonstrated that kidneys lacking Crry are destroyed by uncontrolled complement activation (22). These studies highlight the important role that Crry plays in protecting TECs from complement-mediated injury after warm ischemia or transplantation.
In the current study, we have used in vitro and in vivo experiments to further explore the role of Crry in regulating complement activation in the surface of TECs. We hypothesized that control of the complement system on the surface of TECs is critically dependent upon Crry, and that loss of complement regulation by Crry is sufficient to permit alternative pathway activation on the cell surface even in the absence of cellular injury or hypoxia. We also hypothesized that circulating factor H cannot fully compensate for reduced expression of Crry on the TEC surface. To test these hypotheses we have examined which of the regulatory proteins are expressed on TECs, whether expression of Crry is altered by cellular hypoxia, and whether cells that do not express Crry are susceptible to complement mediated injury. By examining the mechanisms by which the complement system is regulated on the renal tubules, we hope to gain insight into the pathogenesis of complement-mediated diseases of the kidney.
Primary antibodies to Crry (BD Biosciences), DAF (Cedarlane), CD59 (Abcam), pan-cytokeratin (Sigma), Megalin (Santa Cruz Biotechnology), collagen type IV (Millipore), and Na+K+ ATPase (Upstate Biotechnology) were used in these studies, and species appropriate secondary antibodies were obtained from Jackson ImmunoResearch. To detect C3 activation fragments, an HRP-conjugated goat polyclonal antibody to mouse C3 (MP Biomedicals) was used for Western blot analysis and a FITC-conjugated form of the same antibody was used for FACS analysis. CD11b was detected by staining sections with an Alexa-Flour 488 conjugated rat anti mouse CD11b antibody (Invitrogen). Complement activation experiments were performed by adding 10% normal mouse serum to the cell media. To block factor H present within the serum from inhibiting the alternative pathway on the surface of cells, a recombinant murine protein referred to as recombinant factor H domains 19-20 [rH 19-20, similar to a human recombinant protein previously described (23)] was added to the reaction at a final concentration of 50 μg/ml. This protein comprises the 19th and 20th short consensus repeats (SCRs) of factor H and competes with the C-terminal polyanion and C3b binding region of the full-length factor H protein, but does not interfere with the N-terminal complement regulatory region (23).
Mice with targeted deletion of the gene for Crry (24) and for factor B (25) were generated as previously described. To generate mice with homozygous deficiency of Crry, Crry+/− mice and fB−/− mice were intercrossed to produce Crry−/−fB−/− mice. Animal care before and during the experimental procedures was conducted in accordance with the policies of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All animal procedures were approved by the University of Colorado Denver – Health Sciences Programs Animal Care and Use Committee.
Primary culture of renal tubular epithelial cells was performed by isolating the kidneys from female C57Bl/6 or Crry−/−fB−/− mice at less than 12 weeks of age. After removing the capsule, cortical tissue was minced with sterile blades and added to 6 mL of 1 mg/mL Type 1A Collagenase (Sigma) dissolved in PBS. The samples were incubated at 37° C for 30 minutes with 1 minute of vortexing every 10 minutes. The cell suspension was then passed sequentially over 100 μm, 70 μm, and 40 μm filters. The filtered cells were centrifuged at 1000 rpm for 5 minutes and were then resuspended in epithelial growth medium [DMEM/F12 (Invitrogen), 2% heat-inactivated fetal calf serum (Hyclone), Insulin 5 mg/L (Gibco), Transferrin 5 mg/L (Invitrogen), 10−12 M T3 (Sigma), 1% Penicillin-Streptomycin (Invitrogen), and Hydrocortisone 40 ng/L (Sigma)]. Cells were used within four weeks and testing for mycoplasma was not performed. In some experiments the cells were grown on Transwell filters (Costar). Trans-epithelial resistance (TER) of the cell monolayer on the Transwell filters was determined by measuring the resistance across the monolayer with an EVOM volt-ohmmeter (World Precision Instruments). The value for cell monolayers was determined by subtracting the TER for filters without cells and then multiplying by the surface area of the filters.
To isolate protein from the apical or basal surfaces the cells were incubated with 2 mM Sulfo-NHS-LC-Biotin (Pierce) added to the apical compartment for 30 minutes. The cells were lifted using a cell scraper, sonicated, and centrifuged for 15 minutes at 2000 RPM. The supernatant was then centrifuged at 17,000 g in an ultracentrifuge for 20 minutes at 4° C. The pellet was re-suspended by sonication in RIPA buffer. The protein was incubated with streptavidin-coupled dynabeads (Invitrogen) while rotating for one hour at 4° C. The mixture was then applied to a magnetic column and the unbound supernatant was saved as the basolateral fraction. The bound beads were then incubated with an equal volume of RIPA buffer and heated to 95° C for 5 minutes. The apical and basolateral proteins were then separated by SDS-PAGE and probed for Crry.
To induce chemical hypoxia in the cells, the media was removed and replaced with 1 μM Antimycin A (Sigma) in DMEM without glucose. ATP measurements of cell extracts were performed by washing cell monolayers in PBS, harvesting the cells into buffer containing 0.220 M mannitol, 0.070 M sucrose, 0.5 mM EGTA, 2 mM HEPES, and 0.1% fatty acid free BSA. The extracts were incubated on ice for 30 minutes and then centrifuged at 750 g for 5 minutes. The ATP in the supernatant was then determined using a bioluminescence assay (Molecular Probes). ATP values were expressed as a percentage (%) of control samples. Complement activation experiments were performed by incubating the cells at 37° C for one hour with 10% normal mouse serum in the media. In some experiments the cells were grown to confluence on Transwell filters and serum was added to either the apical or basal chamber.
Surface expression of Crry, DAF, and CD59 was examined by flow cytometry, and deposition of C3 fragments onto the cell membrane was examined in a similar fashion. TECs were grown to confluence. Some cells were treated with Antimycin A as described in specific experiments. The cells were released from the plates by treatment with Accutase (Innovative Cell Technologies, Inc.) and washed in PBS. For complement activation experiments, the cells were then incubated in 10% mouse serum at 37°C for 1 hr. In some experiments rH 19-20 was added to the reaction as described. Staining of surface proteins was performed by incubating the cells with primary antibody at 4°C for 1 hr, followed by washing the cells in PBS, and incubating them with appropriate secondary antibodies when necessary. Cells were then washed and resuspended in 500 μL of PBS, run on a FACSCalibur machine (BD Biosciences), and the results were analyzed with CellQuest Pro software (BD Biosciences).
Mouse complement factor B was purified from normal mouse plasma by affinity purification and size exclusion chromatography. The affinity column was created by coupling a monoclonal antibody to mouse factor B [mAb 1379, (26)] to CNBr-Activated Sepharose (GE Healthcare) according to the manufacturer’s instructions. C57BL/6J mice were bled by cardiac puncture, and the blood was collected into syringes containing 50 μl of 500 mM EDTA in order to prevent alternative pathway activation. The blood was centrifuged at 10,000 rpm for 15 minutes and the plasma was collected. The plasma was then diluted 1:1 with buffer (EACA 50 mM, EDTA 10 mM, benzamidine 2 mM in PBS, pH 7.4), passed through a 0.22 μm filter (Corning), and added to the affinity column. The column was washed with 10 column volumes of buffer, and the factor B was eluted using 5 M LiCl2. The buffer in the eluent was exchanged with PBS and the eluted proteins were then separated on a HiLoad 26/60 Superdex column (GE Healthcare) using an AKTA FPLC (GE Healthcare). The purified factor B was subjected to electrophoresis on a 10% Bis-Tris gel (Invitrogen) and stained with Coomassie. The protein appeared greater than 95% pure by visual inspection, and we have previously found that protein purified by this method restores alternative pathway activity when added to serum from fB−/− mice.
For immunofluorescence microscopy, sagittal sections of the kidneys were snap frozen in OCT compound (Sakura Finetek, U.S.A., Inc.). Five μm sections were cut with a cryostat and stored at −80°C. The slides were later fixed with acetone and stained with antibody to mouse C3, CD59, collagen type IV, or to mouse CD11b. The slides were then counterstained with hematoxylin (Vector Laboratories, Inc.). To assess the extent of C3 deposition, ten high-powered fields (400X) in the outer medulla of each section were examined. The number of fields with positive C3 staining was averaged for each experimental group. To assess neutrophil infiltration, ten high-powered fields in the outer medulla of each section were examined. The number of CD11b positive cells was expressed as an average per high power field. To confirm that isolated TECs retained an epithelial cell phenotype, the cells were stained with anti-pan cytokeratin or anti-Megalin, and nuclei were visualized with Dapi. Slides were imaged using a Nikon T-2000 inverted microscope and analyzed with Slidebook 4.2 software (Intelligent Imaging Innovations). Cells on coverslips were fixed and permeabilized in cold methanol:acetone (1:1) for 5 minutes and then stained for Crry. Images of the cells were acquired on a Zeiss LSM 50 META confocal microscope.
C3a in cell supernatants was measured by ELISA using monoclonal antibodies and standards from BD Biosciences according to the manufacturer’s instructions. To ensure that variations among serum batches did not affect C3a generation, values are only compared for cells treated simultaneously with the same batch of pooled serum. To determine the amount of C3a generated by contact with the TEC surface, the values were corrected for the amount of C3a generated in serum incubated with media at 37° without cells present. To determine whether rH 19-20 affects complement activation in the fluid phase, 10% serum was incubated with or without rH 19-20. The reaction was sampled at 10, 20, and 30 minutes, and C3a levels were measured.
Serum urea nitrogen levels were measured using a QuantiChrom Urea Assay Kit (BioAssay Systems) according to the manufacturer’s instructions.
Statistical comparisons were performed using Graphpad Prism 5.0 software. A T-test was used for comparison of two groups, and multiple groups were compared using a one-way ANOVA with a Tukey post-test. A P-value of less than 0.05 was considered statistically significant. Results are reported as mean ± SEM.
Immunofluorescence microscopy of mouse (21) and human (19) tissue suggests that Crry (MCP in humans) is expressed on the tubular epithelium, and that there is no detectable DAF or CD59 expression in vivo (19). To determine whether in vitro growth alters the expression of the complement regulatory proteins, TECs were grown in primary culture. Freshly isolated TECs were grown in medium selective for epithelial cells for at least two weeks. The cells were stained for cytokeratin, confirming that > 95% of the cells were epithelial (Figure 1A), and staining for Megalin indicated that they were primarily proximal tubular epithelial phenotype (Figure 1B). FACS analysis was performed using monoclonal antibodies specific for Crry, DAF, and CD59 (Figure 1C-E). Consistent with staining of the tissue sections, the Crry was abundantly expressed on the surface of the cells, but DAF and CD59 were not detected. Therefore, cultured TECs retain their original phenotype with regard to expression of complement regulatory proteins.
It has previously been reported that the alternative pathway of complement is activated on TECs exposed to fresh serum (27, 28). We grew TECs on Transwell filters until a stable TER was obtained. We biotinylated proteins on the apical or basal surface, isolated the biotinylated proteins using streptavidine beads and then performed a Western blot for Crry (Figure 2A). More Crry was isolated from the basolateral surface than the apical surface, supporting the polarized expression of this protein. Cells on Transwell filters were then exposed to 10% complement sufficient mouse serum on either the apical or the basal surface for one hour. Cell lysates were then blotted for C3 activation fragments (Figure 2B). Abundant fixation of C3 was seen in samples exposed to serum on their apical surface, whereas much less C3 was seen in cells exposed on the basal surface.
To determine whether complement activation on the TEC cell surface is controlled by in part by factor H in the serum, the interaction of factor H with the cell surface was blocked using a recombinant dominant negative inhibitor containing the carboxy-terminus of factor H (23). C3a was measured in the supernatants of the cell cultures (Figure 2C), confirming that more C3a is generated when the serum is exposed to the apical surface than the basolateral. The difference in C3a generation on the apical and basal surfaces was not as great as the difference in C3 fixation to these surfaces (Figure 2B). This may be due to fluid phase C3a generation in both the apical and basal compartment which obscures the difference in C3a generation on the cell surfaces. When cells were grown on Transwell filters and exposed to 10% mouse serum in either the apical or basal compartments, the addition of rH 19-20 (50 μg/mL) caused greater generation of C3a (Figure 2C), and that the amount of C3a generated on either surface is greater with the addition of rH 19-20. When the same concentration of rH 19-20 was incubated with serum in the absence of cells there was a slight increase in the amount of C3a generated (Figure 2D), suggesting that the rH 19-20 may have a modest effect on fluid phase activation. The quantity of C3a generated was not sufficient to account for the increase in C3a seen when the rH 19-20 was applied to cells, however. These results indicate that complement activation is regulated more effectively on the basolateral surface of TECs than on the apical surface, and that factor H present in the serum contributes to the regulation of complement activation on both surfaces. Factor H is found in the urine of some patients with renal disease, indicating that this interaction may be functionally important in proteinuric patients (29, 30).
We used Antimycin A to induce chemical hypoxia in the TECs. After treatment of the cells with 1 μM Antimycin A for one hour, levels of ATP in the cells were reduced by 60±15%, P<0.001 (Figure 3A). The integrity of the monolayer, as assessed by the TER, was also disrupted by this treatment protocol (Figure 3B). Flow cytometry of surface Crry demonstrated that the surface levels of this protein were decreased on hypoxic cells relative to unmanipulated controls (Figure 3C). Flow cytometry was performed to measure the deposition of C3 on the surface of the cells after exposure to 10% mouse serum. When gated on live cells, the amount of C3 deposited on hypoxic cells was greater than that deposited on unmanipulated cells (Figure 3D), although the effect was modest. The overall level of Crry in the cells was not changed (Figure 3E), indicating that the protein may be internalized in response to hypoxia. When 10% serum was applied to the cells in the absence of Antimycin no change in the surface Crry levels was seen (data not shown).
Cells were grown on coverslips until confluent and were then subjected to the same protocol of treatment with Antimycin A. At baseline, confocal microscopy of the cells for the Na+K+ ATPase and for Crry demonstrates that both proteins are expressed on the basolateral aspect of the cells (Figure 3F). After treatment of the cells with Antimycin A the basolateral concentration of both proteins was reduced, and accumulated regions of Crry were redistributed towards the apical portion of the cells (Figure 3G).
DAF and CD59 expression can be induced on some renal cell types in response to complement activation (31, 32). We sought to determine whether chemical hypoxia or complement activation on the TECs induces DAF or CD59 expression. TECs were treated with 1 μM Antimycin A for one hour and flow cytometry was performed. Neither DAF nor CD59 could be detected on the surface of the TECs after treatment with Antimycin A (data not shown). Confluent cells were also exposed to 10% mouse serum to activate complement on the cell surface, and were then examined by flow cytometry. Surface DAF and CD59 were not detected on TECs after they were exposed to mouse serum (data not shown).
To determine whether the absence of Crry renders TECs susceptible to complement activation, even without concomitant hypoxia, we generated mice with homozygous deficiency of Crry. Homozygous deficiency of Crry is lethal in utero due to uncontrolled complement activation in the placenta (24), but deficiency of complement factor B (a protein necessary for alternative pathway activation) rescues the phenotype and allows the appropriate number of Crry−/−fB−/− births (33). Therefore, we bred mice doubly deficient in Crry and in complement factor B (Crry−/−fB−/− mice) and confirmed the genotype by PCR (Figure 4A). The absence of Crry on TECs cultured from these mice was confirmed with flow cytometry (Figure 4B). When TECs from Crry−/−fB−/− or from wild-type control mice were exposed to normal mouse serum, more C3 deposition was seen on Crry deficient cells (Figure 4C). Cells were grown on Transwell filters and exposed to 10% mouse serum on their apical or basal surface (Figure 4D). The amount of C3a generated after Crry−/−fB−/− TECs were exposed to serum on their basal surface was greater than that generated when TECs from wild-type mice received the same treatment (Figure 4D), and the levels were similar to those produced when rH 19-20 was added to the reaction (Figure 2C). Furthermore, the difference in C3a generated on the apical and basal surfaces of TECs from wild-type mice was no longer seen when Crry−/−fB−/− TECs were used, indicating that the basolateral expression of Crry accounts for the greater control of complement activation on the basolateral surface. More C3a was generated when the apical surface of the Crry−/−fB−/− TECs was exposed to serum in the presence of rH 19-20 than when it was exposed to serum alone (Figure 4E), indicating that factor H still regulates complement on this surface in the absence of Crry.
We purified factor B from mouse plasma by affinity chromatography followed by size exclusion chromatography (Figure 5A). We injected Crry−/−fB−/− mice and Crry+/+fB−/− control mice with 50 μg of the purified factor B via the tail vein in order to restore alternative pathway activity. Immunofluorescence microscopy demonstrated all of the Crry−/−fB−/− had C3 detectable C3 deposits after reconstitution with factor B. Only one of the Crry+/+fB−/− control mice had detectable C3 after injection with factor B. When the percent of high-powered fields containing C3 deposits was assessed, a greater degree of C3 deposition was seen in the kidneys of Crry−/−fB−/− mice than in Crry+/+fB−/− mice (Figure 5B). C3 deposition was seen along the proximal tubule at the tubular pole of the glomerulus and faintly within the mesangium (Figure 5C). Patches of tubular staining were also seen in the outer medulla (Figure 5D). It is not clear why the deposition was patchy, but the tubulointerstitial deposition that occurs in wild-type mice is also similarly patchy suggesting regional heterogeneities that may foster complement activation.
C3 deposition on the tubules of wild-type mice is typically seen along the tubular basement membrane. Reconstitution of the Crry+/+fB−/− mice did not restore the wild-type pattern of patchy tubular C3 deposition. It is possible that the quantity of factor B that we used was too low or too short-lived to support complement activation in the tubulointerstitium of Crry+/+fB−/− mice. In the reconstituted Crry−/−fB−/− mice some of the C3 staining in the proximal tubules appeared to be cytosolic (unlike what is seen in wild-type mice), and was located within the tubular basement membrane (Figure 5G).
SUN values in unmanipulated Crry+/+fB−/− and Crry−/−fB−/− mice were 15 ± 2 and 10 ± 1, respectively. By 24 hours after reconstitution with factor B, SUNs in Crry−/−fB−/− mice were significantly elevated (85 ± 24 mg/dL in Crry−/−fB−/− mice versus 16 ± 7 in Crry+/+fB−/− mice, P < 0.01; Figure 6A). Immunofluorescence staining demonstrated infiltration of neutrophils in the kidneys of Crry−/−fB−/− mice reconstituted with factor B (Figures 6B and 6C). No evidence of neutrophil infiltration was seen in the kidneys of Crry+/+fB−/− mice reconstituted with factor B.
Finally, we sought to determine whether the tubulointerstitial complement activation would induce expression of other complement regulatory proteins. We examined tissue sections from Crry−/−fB−/− mice for expression of CD59. CD59 was detected in the mesangium of unreconstituted mice (Figure 7A) and mice reconstituted with factor B (Figure 7B). Tubulointerstitial expression was not detected in either group of mice.
The kidney is particularly susceptible to complement mediated injury in a number of clinical settings, and congenital deficiency or defects in the complement regulatory proteins MCP and factor H are strongly associated with the development of renal disease. In the current study we demonstrated that Crry (the murine homologue of MCP in the kidney) is the only membrane bound regulatory of complement expressed by murine TECs. Crry is expressed on the cell membrane, and its expression is concentrated in the basolateral portion of the cell. Polarized TECs regulate complement more efficiently on the basolateral surface of the cells than on the apical surface, in part due to Crry expression at this site. As with renal ischemia/reperfusion (I/R) (21), chemical hypoxia of the TECs causes a reduction in surface Crry levels, and the distribution within the cell is also altered.
Spontaneous complement activation on the surface of TECs is also controlled by endogenous factor H. When rH 19-20 was added to TECs it permitted increased spontaneous deposition of C3 on both the apical and the basal surfaces of the TECs. Thus, both factor H and Crry are important for regulating the complement system on the surface of TECs and preventing autologous injury. Polymorphisms in both factor H and MCP (9, 10) are associated with the development of aHUS. Although endothelial injury is regarded as the trigger of aHUS, tubular injury and cortical necrosis are also well described findings (34), reflecting the vulnerability of the TEC to complement mediated injury. Our findings demonstrate that control of the complement system on the TEC requires proper functioning of both of these proteins and proper localization of the membrane inhibitor on the basolateral surface when it is exposed to serum complement proteins. Hypoxia of the cells disrupts the organization of the cell membrane and renders the cell susceptible to complement activation. This may be due to the overall reduction in surface Crry. It may also be due to increased access of serum complement proteins to the apical surface as the integrity of the monolayer is lost.
Tickover of the alternative pathway in the fluid phase deposits C3b on nearby surfaces, and the deposited C3b causes auto-activation unless effectively inhibited by complement regulatory proteins (35). Several studies have indicated that the absence or functional loss of Crry renders the kidney susceptible to tubular injury (21, 22, 36). The studies presented herein demonstrate that the loss of surface regulation by Crry is sufficient to permit alternative pathway activation on TECs, even without cellular ischemia or another cellular insult. The presence of sporadic C3 in the tubulointerstitium of wild-type mice demonstrates that this is a location of basal complement activation. This constitutive tubulointerstitial activation may help explain why the kidney is susceptible to spontaneous injury in the setting of inadequate complement regulation.
The susceptibility of the kidney to complement mediated injury is likely influenced by fact that expression of the membrane bound inhibitor by the TECs is restricted to the basal surface of the cells, and levels are decreased in response to cellular stress and injury. It has also recently been demonstrated that properdin binds to the apical surface of TECs and may catalyze alternative pathway activation on this surface (28). Factor H regulates the alternative pathway on both the apical and the basal surface of TECs, but regulation of complement by factor H is inadequate to prevent spontaneous complement activation on the apical surface of the cells, and it cannot compensate for the loss of regulation by Crry on the basolateral surface of the cells. Likewise, expression of Crry on the basolateral surface of TECs does not fully prevent complement activation when the C-terminal membrane-interacting portion of factor H was blocked by rH 19-20. Although DAF and CD59 are not expressed in the interstitium of normal kidneys, they can be detected in the tubulointerstitium of some diseased kidneys (31, 32). Therefore, the in vivo dependence of TECs on protection by MCP may not be as stringent as our current results would suggest.
The reduction of surface Crry on hypoxic cells may be an active response to stress. Cervical epithelial cells have been demonstrated to down-regulate surface MCP after infection with piliated Neisseria gonorrhea (37), and down-regulation of surface regulatory proteins may be a mechanism by which epithelial cells foster the immune response to invasive pathogens. Renal TECs form a barrier epithelium, and they may play an important role in the anti-microbial response to pathogens (38). In aseptic diseases such as I/R, however, such a response to stress could be maladaptive. Our findings help explain why the TECs are a target for uncontrolled complement activation in several clinical settings, including proteinuric states and ischemic injury. The mechanisms we have described may also apply to other diseases in which the TECs sustain cellular stress or injury.
Recent studies have examined complement regulation in Crry−/− mice (39, 40). Although maternal complement proteins injure the placentae of Crry−/− embryos (24), if the mother is deficient in alternative pathway proteins (either C3 or factor B) then Crry−/− pups can be generated. Two groups have bred Crry−/− mice that do not have genetic deletion of the C3 or factor B genes. The Crry−/− mice displayed increased consumption of factor B and C3, and levels of these two proteins were decreased in plasma (39, 40). Although the alternative pathway activation was presumably occurring on tissue surfaces, no injury phenotype was described and pathologic complement activation in the glomeruli was not seen. Renal function was not reported in the studies, but the authors stated that the mice were followed for over a year without evidence of renal damage (39). Just as the chronic, diffuse consumption of the complement proteins allows successful parturition in these mice, the extrarenal consumption of factor B and C3 may prevent complement activation on the tubules from reaching a level sufficient to cause renal injury.
Several previous studies have examined the role of Crry in protecting the kidney from complement-mediated injury of the tubules (21, 22, 36). One of these studies demonstrated that basolateral expression of Crry in mice is disrupted by renal ischemia (21). Another study demonstrated progressive complement mediated injury of Crry−/− kidneys transplanted into wild-type hosts (22). In the current study, the acute infusion of factor B restored sufficient activity to cause foci of complement deposition and tissue injury within the kidney. Thus, in a setting that is not potentially complicated by effects of warm ischemia or transplantation-related I/R, our results clearly show that the absence of Crry renders TECs susceptible to spontaneous complement mediated injury.
Although several different renal diseases are closely linked to uncontrolled activation of the alternative pathway, the molecular mechanisms appear to be distinct in several of the models. In DDD, for example, activation in the absence of adequate factor H function appears to occur primarily in the fluid phase (11, 41). The tropism of systemically generated C3 fragments for the glomerular basement membrane is not, as yet, explained. Atypical HUS involves defects in alternative pathway regulation that could potentially cause injury on many host cells but which usually manifests as renal injury (10). The mechanisms of complement regulation on the TECs are distinct from those that protect the glomerular endothelial cells and the GBM, and yet it is striking that the kidney is such a common target of injury in patients with defects in complement regulatory proteins. Other factors, such as the large percentage of cardiac output that the kidneys receive, could possibly contribute to the development of renal injury as a consequence of global defects in complement regulation.
In summary, we have found that Crry is the only cell surface complement regulatory protein expressed by murine TECs, but factor H in the serum also contributes to complement regulation on the TEC surface. At baseline, Crry is concentrated on the basolateral surface, and complement regulation is more efficient on this surface than on the apical surface. Cellular hypoxia disrupts the TEC tight junctions, reduces surface levels of Crry, and potentiates complement activation on the TEC membrane. Finally, we have performed studies demonstrating that complement is spontaneously activated on TECs with genetic deletion of Crry when they are exposed to an intact alternative pathway, and reconstitution of Crry−/−fB−/− mice with purified factor B protein induced acute kidney injury. These results underscore the importance of Crry for protecting TECs from autologous injury. These findings also help to explain the particular vulnerability of TECs to complement mediated injury in several clinical settings, and indicate that therapies capable of restoring control of the complement system at the TEC surface may be beneficial in renal tubulo-interstitial disease.
This work was supported in part by National Institutes of Health Grants DK076690 and DK77661 (JMT), AI311052 (VMH), DK035081 (MKP), and American Heart Association grant 0735101N (VPF).