| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Division of Medicine and
Department of Histopathology, Imperial College School of Science, Technology, and Medicine, Hammersmith Hospital, London, United Kingdom; and
Centre Hospitalier Universitaire Vaudoise, Laboratoire de Rhumatologie, Lausanne, Switzerland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Accelerated nephrotoxic nephritis is a model of immune complex-mediated glomerular inflammation. The role of complement has been explored in this model by depleting mice of C3 using cobra venom factor (CVF). In BALB/c mice, which showed a Th2-biased immune response and a thrombotic pattern of injury, disease was complement dependent. However, the severity of disease in C57BL/6 mice, which showed a Th1-biased response and developed a somewhat different, crescentic pattern of nephritis, was not influenced by experimental complement depletion (17). We developed C1q-deficient mice to explore the mechanism by which C1q deficiency may predispose to SLE and previously reported that a proportion of these mice develop spontaneous glomerulonephritis (15). In this study, we used the model of accelerated nephrotoxic nephritis in mice to investigate further the role of complement, and C1q in particular, in immune-mediated renal disease. We studied three groups of animals: C1q-deficient (C1qa-/-) mice, mice deficient in both factor B and C2 (H2-Bf/C2-/-), and animals triply deficient in C1q, factor B, and C2, (C1qa/H2-Bf/C2-/-), to dissect the relative contributions of the classical and alternative pathways to pathogenesis.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
C1qa-/-, H2-Bf/C2-/-, and C1qa/H2-Bf/C2-/- mice were developed as previously described (15, 18, 19). Age-, strain-, and sex-matched mice were used in all experiments. All complement-deficient and wild-type animals used in these studies were of mixed genetic background 129/Sv x C57BL/6 and aged 6–12 wk. C1qa-/- mice on this mixed genetic background have been shown to develop spontaneous glomerulonephritis and autoantibodies at 8 mo of age (15). However, at 3 mo only 4 of 120 C1qa-/- mice on this background had a weakly positive anti-nuclear Ab (1/80) compared with 1 of 80 wild-type mice. Renal histology of untreated 129/Sv x C57BL/6 wild-type and C1qa-/- mice, aged 2–3 mo, was examined histologically (n = 20 and n = 16, respectively), and none was abnormal. Mice were kept in a standard animal house environment, and experiments were performed according to institutional guidelines.
Preparation of rabbit nephrotoxic globulin
Whole mouse kidneys were passed through two sieves (250 and 106 µm pore size) and collected on a 63-µm pore size sieve. The material collected was washed twice in cold PBS (pH 7.2) with centrifugation at 700 x g for 5 min between each wash, resuspended, and sonicated at 8-µm amplitude in 30-s bursts on ice. Disruption of glomeruli was verified by microscopy. The material was washed in cold PBS twice and in distilled water once with centrifugation at 200 x g for 15 min between each wash. Finally, the pellet was resuspended in distilled water, washed twice, with centrifugation at 700 x g for 5 min between washes, and frozen onto the surface of a centrifuge tube in liquid nitrogen. It was lyophilized overnight at -70°C and stored at -20°C. Male and female New Zealand White rabbits were immunized s.c. with 1 mg of this preparation of lyophilized glomeruli in CFA (Sigma, Poole, U.K.). Two weeks later they were immunized again with 1 mg in IFA (Sigma), and after an additional 2 wk they were exsanguinated by cardiac puncture. The serum was heat inactivated at 55°C for 30 min, and a globulin fraction was obtained by precipitation in 50% ammonium sulfate. The precipitate was resuspended in PBS and extensively dialyzed against PBS. It was sterilized using a 0.2-µm pore size filter and stored at -70°C. The concentration of rabbit IgG in the preparation was measured by radial immunodiffusion using a sheep anti-rabbit IgG Ab (Serotec, Oxford, U.K.) and rabbit IgG standards (Sigma).
Induction of glomerulonephritis in mice
Mice were initially immunized with 200 µg of rabbit IgG (Sigma) in CFA (Sigma) administered i.p. Seven days later the animals were injected with 1 mg of rabbit nephrotoxic globulin (prepared as described above) via the tail vein. At selected times up to 4 days after injection of nephrotoxic Ab, mice were anesthetized by i.p. injection of midazolam and fentanyl and exsanguinated before harvesting the kidneys. Whenever possible, mice were sacrificed in groups to allow comparison at identical time points. However, some C1qa-/- mice died prematurely.
Histological studies and quantitative immunofluorescence
Kidneys were fixed for 4 h in Bouin’s solution, transferred to 70% ethanol, and embedded in paraffin or were snap-frozen in isopentane and stored at -70°C. Sections embedded in paraffin were stained with periodic acid-Schiff (PAS) reagent. Frozen sections were cut at a thickness of 5 µm. All light microscopic analysis and quantitative immunofluorescence were performed by an observer without knowledge of the sample identity. Glomerular thrombosis was assessed by grading the degree of PAS-positive material as follows: grade 0: no glomeruli with PAS-positive material; grade 1, 0–25%; grade 2, 25–50%; grade 3, 50–75%; and grade 4, 75–100%. Immunohistochemistry, performed as described previously (20), confirmed that PAS-positive material contained large amounts of fibrin (data not shown). Neutrophils and apoptotic cells were identified by their characteristic morphology on PAS-stained sections, and 50 glomeruli per section were counted at a magnification of x1000 under oil immersion.
Mouse anti-rabbit IgG (Sigma) and rabbit anti-mouse IgG (Southern Biotechnology Associates, Birmingham, AL) mAb (FITC-conjugated) were used for direct immunofluorescence studies. Indirect immunofluorescence was performed using rat anti-mouse C1q, C3, or C4 mAb (Cedarlane Laboratories, Ontario, Canada). The secondary Ab was FITC-conjugated monoclonal mouse anti-rat light chain (Sigma). All incubations were performed for 1 h at room temperature, and all Abs were used at a 1/50 dilution in PBS (pH 7.2) except for mouse anti-rabbit IgG (1/200). In quantitative immunofluorescence studies, to exclude artifacts due to variable decay of the fluorochrome all sections from one experiment were stained and analyzed at the same time. Sections were examined at x400 magnification using an Olympus BX4 fluorescence microscope (Olympus Optical, London, U.K.). A Photonic Science Color Coolview digital camera (Photonic Science, East Sussex, U.K.) was attached to the microscope and, using Image-Pro Plus software (Media Cybernetics, Silver Spring, MD), images were captured for analysis. Glomeruli were selected manually on the computer screen, and the fluorescence intensity over each glomerulus was measured. For each section, 20 glomeruli were examined, and the mean fluorescence intensity was recorded, with results expressed in arbitrary fluorescence units (AFU).
Peripheral blood leukocyte counts
Blood was collected into potassium EDTA as described above, followed by 1/5 dilution in 2% acetic acid and counting the total white cells using a hemocytometer. Neutrophil numbers as a percentage of the total leukocyte count were assessed by visualization of cells under phase contrast microscopy.
Serum creatinine and albumin
Serum creatinine and albumin were measured at Department of Chemical Pathology, Hammersmith Hospital, using an Olympus AU600 analyzer (Olympus Diagnostics, London, U.K.), with albumin analyzed by the bromocresol green method. Baseline measurements in untreated mice showed that the normal ranges of serum creatinine and albumin did not differ between the wild-type and C1qa-/- mice used in the study. Serum albumin levels in 15 wild-type and 11 C1qa-/- age- and sex-matched mice were 25 (range, 9–28), and 24 (range, 19–28) g/L, respectively. Baseline serum creatinine levels in these mice were 43 (range, 36–52) and 44 (range, 39–54) µmol/ml, respectively.
Albumin clearance
Mice were housed in metabolic cages for 24 h for urine collection. The albumin concentration was measured by radial immunodiffusion. Samples and standards (mouse albumin; Sigma) were placed in wells (6 µl/well) in 1.2% agarose in PBS containing rabbit anti-mouse albumin (Biogenesis, Poole, U.K.). Gels were dried and stained with Coomassie Blue. Concentrations in samples were calculated with reference to a standard curve. Albumin clearance was calculated from the formula: (urinary albumin concentration x volume)/plasma albumin concentration. Baseline urinary albumin measurements in wild-type, C1qa-/-, and C1qa/H2-Bf/C2-/- mice showed concentrations of <50 µg/ml in all cases (n = 10, 10, and 5, respectively).
Measurement of the murine anti-rabbit IgG immune response
ELISA plates (Nunc Maxisorb, Life Technologies, U.K.) were coated overnight at 4°C with 10 µg/ml of rabbit IgG (Sigma) or 5 µg/ml goat anti-mouse Ig(H+L) (Southern Biotechnology Associates) in PBS (pH 7.2). Plates were blocked with 1% BSA in PBS. Blocking and all subsequent incubations were performed for 1 h at 37°C. Plates were washed twice in PBS-0.075% Tween and twice in PBS (pH 7.2). Wells coated with goat anti-mouse Ig were incubated with 50 µl of serial dilutions of mouse IgG1, IgG2a, IgG2b, or IgG3 (Sigma) to generate a standard curve. Wells coated with rabbit IgG were incubated with 50 µl of sera, diluted as required in 1% BSA-PBS-0.05% Tween. Plates were washed as before, incubated with alkaline phosphatase-conjugated goat anti-mouse IgG1, IgG2a, IgG2b, or IgG3 diluted 1/2000 in 1% BSA-PBS-0.05% Tween (Southern Biotechnology Associates), followed by further washing, and developed using the substrate p-nitrophenyl phosphate (Sigma). The OD of the reaction mixture at 405 nm wavelength was measured using an ELISA reader (Titer-Tek, ICN Flow, Costa Mesa, CA). The concentration of mouse IgG subclasses in the samples was calculated with reference to the standard curve.
RNase protection assay (RPA)
The genes analyzed were protease-activated receptor-1 (PAR-1),
tissue factor pathway inhibitor (TFPI), tissue factor (TF), early
growth response-1 (Egr-1), vascular endothelial growth factor (VEGF),
urokinase-plasminogen activator (uPA), and plasminogen activator
inhibitor-1 (PAI-1). cDNA fragments of different size for each of the
chosen genes were recovered by RT-PCR and subsequently subcloned into
the pGEM-T plasmid (Promega, Madison, WI). Positive clones with
antisense orientation with respect to the T7 promoter were identified
by colony PCR, and their complete sequences were verified by DNA
sequencing. Radioactive multiriboprobe preparation and the RPAs were
performed according to standard protocols (21). Briefly,
antisense [32P]UTP-labeled riboprobes were
synthesized by in vitro transcription of the DNA template set, which
contained all the linearized plasmid cDNAs pooled in equimolar amounts.
DNase I treatment was performed to remove the DNA templates, and
riboprobes were purified by phenol/chloroform extraction, followed by
ethanol precipitation using glycogen as carrier. For each sample 
5
µg of total RNA was hybridized overnight at 52°C with 3 x
105 cpm of the labeled multiprobe. Samples were
first treated with an RNase cocktail (Ambion, Austin, TX). RNase was
then removed by proteinase K treatment, and samples were purified by
phenol/chloroform extraction, followed by ethanol precipitation using
glycogen as carrier. Protected fragments were resolved through a 5%
sequencing gel. Quantification of all the experimental
samples for all time points studied was performed by analyzing the gels
with an InstantImager apparatus and software (Packard, Meriden, CT).
For each protected band analyzed, a counts per minute/square millimeter
value was obtained. This value was corrected for sample loading errors
by normalizing with the respective counts per minute/square millimeter
value calculated for the constitutively expressed GAPDH gene that was
also included in the template set.
Statistics
Results are expressed as the median (range). Nonparametric tests were applied throughout. The Mann-Whitney U test was used to analyze the differences between two groups of mice. Where three groups of mice were compared, the Kruskal-Wallis test was used, with Dunnett’s multiple comparisons test (p values given are from Dunnett’s test). GraphPad Prism (GraphPad Software, San Diego, CA) was used to analyze the data. Differences were considered significant when p < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The patterns of injury that developed following the induction of
accelerated nephrotoxic nephritis were compared in wild-type,
C1qa-/-, and
C1qa/H2-Bf/C2-/- mice. There were premature
deaths in C1q-deficient mice. In the experiment shown in Fig. 1
, a–c, six
C1qa-/- mice died on day 2, and one died on day
3. The remaining three C1qa-/- mice and all
wild-type and C1qa/H2-Bf/C2-/- mice were
sacrificed on day 4. Glomerular histology in
C1qa-/- and
C1qa/H2-Bf/C2-/- mice showed extensive
PAS-positive material representing glomerular thrombosis, whereas
wild-type mice typically showed little thrombosis and only mild
hypercellularity. Representative glomeruli from each group of mice are
shown in Fig. 2. Significantly more
thrombosis was seen in C1qa-/- and
C1qa/H2-Bf/C2-/- mice than in wild-type mice,
as shown in Fig. 1a (p < 0.001 and
p < 0.01 for C1qa-/- and
C1qa/H2-Bf/C2-/-, respectively). More severe
renal injury in C1qa-/- and
C1qa/H2-Bf/C2-/- mice was also evident from
analysis of serum albumin and creatinine levels, with serum albumin
used as an indirect index of proteinuria. Four days after injection of
nephrotoxic Ab, serum albumin levels were significantly lower in both
C1qa-/- and
C1qa/H2-Bf/C2-/- mice than in wild-type animals
(Fig. 1b). Results (expressed as the median in grams per
liter) were: wild-type, 21.5 (range, 14–31);
C1qa-/-, 11.5 (range, 10–14);
C1qa/H2-Bf/C2-/-, and 11 (range, 8–18;
p < 0.01 for each compared with wild-type mice). Serum
creatinine was higher in C1qa-/- than in
wild-type mice (Fig. 1c). Results were: wild-type, 41.5
µmol/l (range, 33–54); C1qa-/-, 81 µmol/l
(range, 61–179); C1qa/H2-Bf/C2-/-, 42 (range,
28–45; p < 0.001 for C1qa-/-
compared with wild-type mice).
|
|
Increased glomerular injury was not seen in mice deficient in both factor B and C2
Wild-type and C1qa-/- mice were compared
with H2-Bf/C2-/- mice after induction of
nephritis to assess the effect of disrupting the classical pathway
distally to C1q, at the level of C2. All mice were sacrificed on day 4.
The degree of glomerular injury in H2-Bf/C2-/-
mice was not significantly different from that in wild-type mice, but
was significantly less severe than that in
C1qa-/- mice (p <
0.01), which, as described above, developed severe glomerular
inflammation and thrombosis. The histologic appearance of
representative glomeruli is illustrated in Fig. 2
, and quantitative
data comparing thrombosis are shown in Fig. 1d. Serum
albumin was significantly higher in H2-Bf/C2-/-
animals than in C1qa-/- mice and was similar to
that in wild-type mice as shown in Fig. 1e. Results
(expressed as the median) were: wild-type mice, 23 g/L (range, 16–24);
C1qa-/-: 12 g/L (range, 8–15);
H2-Bf/C2-/-: 22 g/L (range, 10–25;
p < 0.05 for H2-Bf/C2-/-
compared with C1qa-/-). Serum creatinine was
significantly lower in H2-Bf/C2 -/- mice than
in C1qa-/- mice (p <
0.05) and was not significantly different from that in wild-type mice,
as shown in Fig. 1f. Serum creatinine levels were: wild-type
mice, 41 µmol/L (range, 37–58); C1qa-/-, 58
(range, 37–88); H2-Bf/C2-/-, 37 (range,
35–44; p < 0.05 for
H2-Bf/C2-/- compared with
C1qa-/-).
These data showed that the increased glomerular injury seen in the C1qa-/- mice did not occur in H2-Bf/C2-/- mice, suggesting that the complement proteins that have a critical role in protecting from the development of renal inflammation in this model lie proximal to C2 in the classical pathway of complement activation.
Quantitation of glomerular complement proteins and glomerular mouse and rabbit IgG by immunofluorescence
Interpretation of immunofluorescence studies performed on histological material obtained on day 4 after the induction of nephritis, when severe glomerular thrombosis was present in C1qa-/- and C1qa/H2-Bf/C2-/- mice, proved unreliable. Therefore, we used material taken at baseline (before the induction of nephritis) and 1 day after induction of disease for all quantitative immunofluorescence studies. Following immunostaining of kidney sections, the site of protein deposition was determined by direct microscopy, and quantification was performed by computer-aided image analysis as described in Materials and Methods. We studied the deposition of complement proteins C1q, C4, and C3 and both rabbit and mouse IgG.
Complement proteins
Frozen sections of kidney from wild-type and
C1qa-/- mice (n = 6/group) were
stained for C1q, C4, and C3. Representative sections are shown in Fig. 3.
|
, top panel).
C4.
Mesangial C4 deposits were seen in both wild-type and
C1qa-/- mice at baseline, with significantly more present
in wild-type mice (Fig. 3). One day after induction of disease,
wild-type mice showed the presence of additional linear glomerular
capillary wall C4, which was not seen at baseline, although there was
no increase in the amount of C4 deposited in wild-type mice on day 1
compared with baseline. The large and variable amount of C4 present at
baseline may have prevented detection of the small increase in the
total amount of C4 resulting from capillary wall deposition at day 1.
There was no change in the very small amount of C4 present in the
glomeruli of C1qa-/- mice following
induction of disease compared with the baseline, compatible with
failure to activate the classical complement pathway in these animals.
On day 1, as at baseline, there was significantly more C4 in the
glomeruli of wild-type mice compared with C1q-deficient animals.
Results (expressed as the median in AFU) at baseline were: wild-type,
43.1 (range, 33.1–65); and C1qa-/-, 9.05 (range,
6.6–12; p < 0.005). Day 1 results were:
wild-type, 45.35 (range, 23.8–56.2); and C1qa-/-, 11.55
(range, 11.3–45.9; p < 0.05).
C3.
Both wild-type and C1qa-/- mice had sparse
mesangial C3 deposits at baseline. In both wild-type and
C1qa-/- mice on day 1 after disease induction,
there was also linear capillary wall C3 (Fig. 3). The amount of C3 was
quantified, and there was a significant increase in the amount of
glomerular C3 in both wild-type and C1qa-/-
mice on day 1 compared with baseline levels. No significant differences
were detected between wild-type and C1qa-/-
animals at either time point. Results (expressed as the median) at
baseline were: wild-type, 19.9 AFU (range 15–32.4); and
C1qa-/-, 17 AFU (range, 12.4–28). Day 1
results were: wild-type, 45.25 AFU (range, 32.9–51.3); and
C1qa-/-, 36.9 AFU (range, 14.9–58.8; on day 1
compared with baseline: p < 0.005 for wild-type and
p < 0.05 for C1qa-/-). It
seemed likely that the glomerular deposition of C3 following the
induction of inflammation reflected alternative pathway activation, as
we documented significantly increased amounts of the protein in
C1qa-/- mice. This was confirmed in studies
performed in C1qa/H2-Bf/C2-/- mice, unable to
activate either the alternative or classical complement pathway. In
sections from these mice stained for C3 at day 1 after injection of
nephrotoxic Ab (n = 4), we were able to demonstrate
only sparse mesangial C3. This was of a similar magnitude to that seen
in wild-type, C1qa-/-, and
C1qa/H2-Bf/C2-/- mice at baseline.
Immunoglobulin
Quantitation of rabbit IgG in the kidney was performed by immunofluorescence to examine the possibility that the increased disease in C1q deficient mice might be due to increased binding of nephrotoxic Ab. One day after induction of nephritis by the injection of rabbit nephrotoxic globulin no significant differences were observed between wild-type and C1qa-/- mice (n = 5–8/group). Levels (expressed as the median) were: wild-type, 21.25 AFU (range, 13.8–29.55); and C1qa-/-, 25.75 AFU (range, 18.05–32.2).
We also measured the deposition of mouse IgG using the same technique
to explore the possibility that the increased disease in
C1qa-/- mice was due to the presence of more
autologous Ab. Data on day 1 following the injection of nephrotoxic
globulin (two separate experiments) are shown in Fig. 4. In an experiment comparing wild-type
and C1qa-/- mice (n =
5–7/group), there was significantly more IgG present in the glomeruli
of C1qa-/- mice. Results (median) were:
wild-type, 19.95 AFU (range, 13.55–25.6); and
C1qa-/-, 33.43 AFU (range, 17.65–72.85;
p < 0.05). Data are illustrated in Fig. 4a.
In a second experiment comparing wild-type,
C1qa-/-, and
C1qa/H2-Bf/C2-/- mice (n =
6–11/group), there was a trend toward more IgG in the glomeruli of
C1qa-/- mice compared with wild-type, but this
did not reach statistical significance. However, there was
significantly more IgG in the glomeruli of
C1qa/H2-Bf/C2-/- mice compared with wild-type
mice (Fig. 4b). Results were (AFU): wild-type, 17.6
(range, 12.5–24.9); C1qa-/-, 25.5 (range,
14.4–57); C1qa/H2-Bf/C2-/-, 30.5 (range,
18.4–65.4; p < 0.05 for
C1qa/H2-Bf/C2-/- compared with wild-type).
These data suggest that the increased disease in C1q-deficient mice may
reflect the presence of more murine IgG in the glomeruli.
|
We considered the possibility that the increased amount of murine IgG we observed in the glomeruli of C1q-deficient mice may reflect a difference in the immune response to rabbit IgG in these animals compared with wild-type. mice Serum levels of mouse anti-rabbit IgG of different subclasses on day 4 after induction of disease were measured by ELISA in C1qa-/- and wild-type mice (n = 7/group). Levels (expressed as the median) on day 4 were as follows for wild-type and C1qa-/-, respectively: IgG1, 11.4 µg/ml (range, 4.4–27.8) and 5.43 µg/ml (range, 4.3–10.7); IgG2a, 0.215 µg/ml (range, 0.05–0.8) and 0.3 µg/ml (range, 0.03–0.66); IgG2b, 10.2 µg/ml (range, 2.6–17.7) and 3.56 µg/ml (range, 2.3–11.4); and IgG3, 0.11 (range, 0.04–0.226) and 0.04 µg/ml (range, 0.03–0.07). The only statistically significant difference observed between complement-deficient and wild-type mice was in IgG3 anti-rabbit IgG, where there was significantly less in the serum of the C1qa-/- mice (p < 0.01), although total levels of this subclass were very low. This finding could not explain the increased disease in these mice.
The early phase of nephritis in C1q-deficient mice was characterized by both increased neutrophil influx as well as thrombosis
Groups of wild-type and
C1qa-/- mice (n =
5–7) were sacrificed every 24 h after induction of disease up to
day 4 to explore the events preceding the development of glomerular
thrombosis (Fig. 5). Although less severe
disease was seen in C1qa-/- mice in this
experiment than in the experiments shown in Figs. 1 and 2, the
experiment in Fig. 5 illustrates the histological changes that precede
glomerular thrombosis. There were significantly more neutrophils in the
glomeruli of C1qa-/- mice at all stages during
the development of disease, and between one and three
C1qa-/- mice at each time point showed
glomerular thrombosis, which was not seen in the wild-type mice.
Results (expressed as the median) for wild-type and
C1qa-/- mice were: day 1, 0.06
neutrophils/glomerular cross-section (range, 0.04–0.3) and 0.58
neutrophils/glomerular cross-section (range, 0.1–11); day 2, 0.12
neutrophils/glomerular cross-section (range, 0–0.22) and 1.63
neutrophils/glomerular cross-section (range, 0.1–10); day 3, 0.1
neutrophils/glomerular cross-section (range, 0.02–0.68) and 1.16
neutrophils/glomerular cross-section (range, 0.18–11.3); day 4, 0.12
neutrophils/glomerular cross-section (range, 0.04–0.24) and 1.16
neutrophils/glomerular cross-section (range, 0.4–2.28;
p < 0.05 for all except day 4, where p
< 0.01, and day 2, where p = 0.059). Thus, it appears
that thrombosis was preceded by increased glomerular neutrophil influx
in C1q-deficient mice.
|
This increased glomerular neutrophil number was not explained by
coincident differences in blood neutrophil numbers, as there was no
difference at baseline or on days 1 and 2. Measurements were made in
blood taken from mice on days 1 and 2 of the experiment shown in Fig. 5
(n = 5–7/group). Results (expressed as the median) for
C1qa-/- mice on days 1 and 2 were 895 (range,
590-2250) and 720 (range, 510-2620) neutrophils/µl, respectively.
Results for wild-type mice on days 1 and 2 were 710 (range, 130-1140)
and 590 (range, 200-1160) neutrophils/µl, respectively. There was no
difference in neutrophil counts in untreated
C1qa-/- mice and wild-type mice (data not
shown).
The early phase of the disease was characterized by an increased number of apoptotic cells in C1q-deficient mice
Glomerular apoptotic cells were counted 1 day after induction of disease to determine whether the disease developing in C1q-deficient mice was associated with increased numbers of apoptotic cells. Data were obtained from three experiments. In the first, one wild-type animal had 2 apoptotic cells/50 glomeruli, and five had none. Results in six C1qa-/- mice were 0, 2, 2, 4, 10, and 10 apoptotic bodies/50 glomeruli. The difference between these two groups was significant (p < 0.05). In two additional experiments there was a trend toward more apoptotic cells in C1qa-/- mice compared with wild-type mice, but the differences did not reach statistical significance. There was a trend for apoptotic cells to be associated with increased neutrophil numbers, but this correlation reached statistical significance in only one of the three experiments performed (Spearman’s rank correlation coefficient = 0.7544; p < 0.05).
Assessment of factors that may cause glomerular thrombosis
Glomerular thrombosis was the most striking feature of disease in
C1qa-/- and
C1qa/H2-Bf/C2-/- mice (Fig. 2). As an
additional and independent measure of thrombosis in the
complement-deficient mice, we compared expression levels of genes
encoding proteins involved in coagulation and fibrinolysis by RPA,
using kidney RNA from wild-type and C1qa-/-
mice on 1 and 4 days after induction of glomerulonephritis. No
differences were found in the expression of any of the above genes in
C1qa-/- compared with wild-type mice at 1 day
after the induction of disease (n = 6/group; data not
shown). On day 4 (n = 10–11/group) there was
significantly more PAI-1, Egr1, TF, and PAR-1 mRNA in
C1qa-/- mice than wild-type mice, and there was
significantly less VEGF RNA in C1qa-/- mice
compared with wild-type mice. There was no difference in the expression
of either uPA or TFPI. Results (expressed as the median) for
C1qa-/- and wild-type mice were: PAI-1, 84
arbitrary units (range, 45–100) compared with 28 arbitrary units
(range, 18–47; p < 0.0005); Egr-1, 95 arbitrary units
(range, 48–100) compared with 28.5 arbitrary units (range, 5–62;
p < 0.0005); TF, 79 arbitrary units (range, 25–65)
compared with 49.5 arbitrary units (range, 25–100; p
< 0.05); PAR-1, 59 arbitrary units (range, 22–100) compared with 13.5
arbitrary units (range, 12–37; p < 0.001); VEGF, 44
arbitrary units (range, 14–83) compared with 82.5 arbitrary units
(range, 23–100; p < 0.001); uPA, 62 arbitrary units
(range, 52.5–72) compared with 67 arbitrary units (range, 63.5–73.5);
and TFPI, 76 arbitrary units (range, 66–85) compared with 80 arbitrary
units (range, 66–85), respectively. These results, shown graphically
in Fig. 6, showed that differences in the
expression of genes encoding in procoagulant and antifibrinolytic genes
were observed at the same time as, but not before, thrombosis was
evident histologically.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our study also provides evidence against a major role for the complement system as an effector of inflammatory tissue damage in this model. Although C3 deposition was seen in the glomeruli of C1qa-/- mice, C3 was not deposited in the glomeruli of C1qa/H2-Bf/C2-/- mice. Because severe glomerular thrombosis and hypoalbuminemia developed in both groups of mice, this showed that C3 activation was not required for the development of disease, although C1qa/H2-Bf/C2-/- mice appeared to be protected from the development of renal failure. In this respect, the findings of the current study corroborate our previous observation that spontaneous proliferative glomerulonephritis developed in 64% of C1qa/H2-Bf/C2-/- mice by 8 mo of age (19), indicating that the autoimmune glomerulonephritis in these animals also develops largely independently of either classical or alternative pathway complement activation. However, previous comparable studies of experimental nephritis indicate that the genetic background of the mice studied may be critical in determining the complement dependence of the phenotype observed. For example, a study using CVF to induce complement depletion in a murine model of accelerated nephrotoxic nephritis suggested that complement played a role in mediating tissue injury in BALB/c, but not C57BL/6, mice (17).
The role of complement has been examined previously in other animal models of immune-mediated glomerular inflammation. A number of studies have been performed in complement-deficient mice using the model of heterologous nephrotoxic nephritis (23, 24, 25). This is a model of neutrophil-mediated acute inflammation caused by a single injection of heterologous nephrotoxic Ab. These previous studies have also yielded conflicting data for the role of complement in mediating the neutrophil influx that is the hallmark of this model. Two studies suggested a decrease in neutrophil numbers in mice with C3 or C4 deficiency (23, 25), but a third found no difference, although C3 deficiency did protect from proteinuria (24). Earlier studies in mice, using CVF to cause complement depletion, also suggested that complement did not play a major role in mediating the ingress of neutrophils (26, 27)
What is the mechanism leading to the development of severe,
complement-independent glomerular thrombosis in
C1qa-/- mice after the induction of accelerated
nephrotoxic nephritis in our present study? We considered a number of
possibilities. Firstly, we investigated whether there was a greater
initial deposition of nephrotoxic (rabbit) Ab in the kidney.
Immunofluorescence did not show any difference in glomerular binding of
nephrotoxic Ab in C1qa-/- compared with
wild-type mice. A second possibility was that an increased amount of
autologous mouse Ab bound to the rabbit Ab deposited in glomeruli of
C1qa-/- mice. Quantitative immunofluorescence
studies performed on day 1 showed that this was indeed the case (Fig. 4). We considered two possible explanations for this observation. It
could have occurred as a result of a heightened murine immune response
to the rabbit Ig, leading to greater deposition due to elevated
circulating Ab levels in the complement-deficient animals. However,
previous work has shown a defective humoral immune response in
C1qa-/- mice (28). In the present
study we did not detect significantly increased circulating levels of
mouse anti-rabbit IgG in C1qa-/- mice
compared with wild-type mice.
The second possible explanation for the increased glomerular IgG in
C1qa-/- mice was a defect in the clearance of
murine Ab bound to the target Ag (rabbit nephrotoxic immunoglobulin) in
the kidney. Complement has been shown to play a role in the
solubilization of aggregated immune complexes, but the alternative
pathway was thought to play a dominant role (29). However,
solubilization is delayed in C2-deficient serum (30), and
there is evidence that some solubilization occurs in the absence of the
alternative pathway (31). Moreover, these in vitro studies
used immune aggregates in the fluid phase, and different mechanisms may
operate when immune complexes are deposited in tissues, as in this
model. The presence of increased murine IgG in the glomeruli of
C1q-deficient mice may contribute to the increased renal injury by
ligation of Fc receptors on inflammatory cells. This is supported by
the observation of an increase in glomerular neutrophil numbers in
C1qa-/- mice compared with wild-type mice in
the phase preceding the development of thrombosis. This neutrophil
accumulation did not reflect differences in circulating neutrophils
between wild-type and C1qa-/- animals, at
baseline or following disease induction. Several recent reports have
supported a role for Fc receptors in mediating glomerular inflammation.
Accelerated nephrotoxic nephritis has been studied previously in
Fc
-chain deficient mice, lacking the functional activating receptors
FcRI and FcRIII. Animals lacking the Fc-chain were shown to be
protected from disease (32). This followed work with
lupus-prone New Zealand Black/White mice (33), showing
that Fc-chain deficiency also prevented the development of
spontaneous glomerulonephritis.
Results in other models of immune complex-mediated inflammation, such as the reverse passive Arthus reaction, have suggested that Fc receptors play a greater role than complement in certain contexts in causing tissue injury (34, 35). In the present study the observations support the hypothesis that C1q deficiency results in impaired processing of immune complexes, with persistence of a proinflammatory stimulus within the glomerulus, inducing Fc-mediated influx of neutrophils. However, there is also clear evidence that complement may contribute to inflammatory disease caused by immune complexes in other experimental situations (36, 37, 38). Other genetic factors may be important in determining the role of Fc receptors vs complement in the induction of inflammatory renal injury.
A role for complement in the clearance of apoptotic cells was suggested by the observation that C1q binds to apoptotic human keratinocytes in vitro (14). Further supportive evidence includes the presence of increased numbers of apoptotic cells in the glomeruli of C1q-deficient mice with spontaneous glomerulonephritis (15), and the defective clearance of apoptotic cells in the absence of complement components both in vitro (39) and in vivo (16). In the present study 1 day after induction of disease, we saw increased numbers of apoptotic cells in the glomeruli of C1qa-/- mice compared with wild-type mice. Abnormal clearance of apoptotic cells could lead to either autoimmunity or dysregulated inflammation. In addition, a procoagulant role for apoptotic cells has been demonstrated (40), and this may be relevant to the glomerular thrombosis demonstrated in the present study.
We explored the possibility that there was increased expression of genes involved in the coagulation and fibrinolytic pathways in C1qa-/- compared with wild-type animals, which might explain the development of glomerular thrombosis. Taken together, our results show that the exacerbated glomerular thrombosis in C1qa-/- mice can be accounted for by coordinated up-modulation of the procoagulant and the anti-fibrinolytic proteins. There was increased PAI-1, TF, PAR-1, and Egr-1 mRNA expression at 4 days in the kidneys of C1qa-/- mice compared with wild-type animals. In addition, there was decreased expression of VEGF mRNA, perhaps reflecting the greater endothelial cell damage. Previous work in models of crescentic nephritis in rabbits (41, 42) and studies of glomerulonephritis in man (43) have suggested that macrophages are the major cell type producing TF. However, there was little evidence of a significant mononuclear cell infiltrate in C1qa-/- mice in our model, suggesting that in this situation other cell types, perhaps endothelial cells, are responsible for TF production. There was no difference in uPA expression, and this is consistent with previous work suggesting that in murine glomeruli, tPA is the major plasminogen activator (44, 45). We did not find any differences in RNA levels at 1 day after induction of disease, suggesting that these changes developed as a result, rather than as a primary driver, of glomerular inflammation.
In summary, deficiency of either C1q or C4 in humans is very strongly associated with the development of SLE (11). In the present study we have shown that in our model of experimental glomerulonephritis severe disease developed in C1q-deficient, but not in C2-deficient, mice. This suggests an anti-inflammatory effect of a protein, either C1q or C4, lying proximal to C2 in the classical pathway of complement activation. Abnormal clearance of immune complexes or apoptotic cells may contribute to pathogenesis in C1qa-/- mice, but further work is needed to define the mechanisms underlying the protective effect of C1q or C4 and the relative importance of complement and Fc-mediated inflammatory processes in this model.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Kevin A Davies, Rheumatology Section, Division of Medicine, Imperial College School of Science, Technology, and Medicine, Hammersmith Hospital, Du Cane Road, London, U.K. W12 ONN. E-mail address: kevin.davies{at}ic.ac.uk
3 Abbreviations used in this paper: SLE, systemic lupus erythematosus; CVF, cobra venom factor; PAS, periodic acid-Schiff; AFU, arbitrary fluorescence units; PAR-1, protease-activated receptor-1; TFPI, tissue factor pathway inhibitor; TF, tissue factor; Egr-1, early growth response-1; VEGF, vascular endothelial growth factor; uPA, urokinase-plasminogen activator; PAI-1, plasminogen activator inhibitor-1; RPA, RNase protection assay.
Received for publication December 7, 2000. Accepted for publication March 29, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
receptor-dependent neutrophil adhesion and complement-dependent proteinuria in acute glomerulonephritis. J. Exp. Med. 186:1853. production by antigen-specific T cells. J. Exp. Med. 187:1789.RIII (CD16) deficient mice. Immunity. 5:181.[Medline]
This article has been cited by other articles:
|
|
 |
|
  M. A. Otten, T. W. L. Groeneveld, R. Flierman, M. P. Rastaldi, L. A. Trouw, M. C. Faber-Krol, A. Visser, M. C. Essers, J. Claassens, J. S. Verbeek, et al. Both Complement and IgG Fc Receptors Are Required for Development of Attenuated Antiglomerular Basement Membrane Nephritis in Mice J. Immunol., September 15, 2009; 183(6): 3980 - 3988. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R.J. Taylor, C. M. Turner, J. I. Elliott, J. McDaid, R. Hewitt, J. Smith, M. C. Pickering, D. L. Whitehouse, H. T. Cook, G. Burnstock, et al. P2X7 Deficiency Attenuates Renal Injury in Experimental Glomerulonephritis J. Am. Soc. Nephrol., June 1, 2009; 20(6): 1275 - 1281. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Tang, B. Lu, E. Hatch, S. H. Sacks, and N. S. Sheerin C3a Mediates Epithelial-to-Mesenchymal Transition in Proteinuric Nephropathy J. Am. Soc. Nephrol., March 1, 2009; 20(3): 593 - 603. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Giorgini, H. J. Brown, H. R. Lock, F. Nimmerjahn, J. V. Ravetch, J. S. Verbeek, S. H. Sacks, and M. G. Robson Fc{gamma}RIII and Fc{gamma}RIV Are Indispensable for Acute Glomerular Inflammation Induced by Switch Variant Monoclonal Antibodies J. Immunol., December 15, 2008; 181(12): 8745 - 8752. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Carlucci, J. Cortes-Hernandez, L. Fossati-Jimack, A. E. Bygrave, M. J. Walport, T. J. Vyse, H. T. Cook, and M. Botto Genetic Dissection of Spontaneous Autoimmunity Driven by 129-Derived Chromosome 1 Loci When Expressed on C57BL/6 Mice J. Immunol., February 15, 2007; 178(4): 2352 - 2360. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. J. Brown, S. H. Sacks, and M. G. Robson Toll-Like Receptor 2 Agonists Exacerbate Accelerated Nephrotoxic Nephritis J. Am. Soc. Nephrol., July 1, 2006; 17(7): 1931 - 1939. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. C. Pickering, J. Warren, K. L. Rose, F. Carlucci, Y. Wang, M. J. Walport, H. T. Cook, and M. Botto Prevention of C5 activation ameliorates spontaneous and experimental glomerulonephritis in factor H-deficient mice PNAS, June 20, 2006; 103(25): 9649 - 9654. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Trendelenburg, L. Fossati-Jimack, J. Cortes-Hernandez, D. Turnberg, M. Lewis, S. Izui, H. T. Cook, and M. Botto The Role of Complement in Cryoglobulin-Induced Immune Complex Glomerulonephritis J. Immunol., November 15, 2005; 175(10): 6909 - 6914. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Turnberg, M. Botto, J. Warren, B. P. Morgan, Mark. J. Walport, and H. T. Cook CD59a Deficiency Exacerbates Accelerated Nephrotoxic Nephritis in Mice J. Am. Soc. Nephrol., September 1, 2003; 14(9): 2271 - 2279. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Inal, M. Pascual, P. Lesavre, and J.-A. Schifferli Complement inhibition in renal diseases Nephrol. Dial. Transplant., February 1, 2003; 18(2): 237 - 240. [Full Text] [PDF]
|
|
|
|
 |
|
 B. A. O'Brien, Y. Huang, X. Geng, J. P. Dutz, and D. T. Finegood Phagocytosis of Apoptotic Cells by Macrophages From NOD Mice Is Reduced Diabetes, August 1, 2002; 51(8): 2481 - 2488. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
