| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Renal Section, Division of Medicine and
British Heart Foundation Cardiovascular Medicine Unit, National Heart and Lung Institute, Imperial College School of Medicine, Hammersmith Hospital, London, United Kingdom; and
Biogen, Inc., Cambridge, MA 02142
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The role of selectins in initiating leukocyte migration into glomeruli remains unclear, with little apparent role for E-selectin, little data for L-selectin, and conflicting data for P-selectin (10, 11, 12, 13, 14, 15). There has been more extensive work on LFA-1 and ICAM-1 in GN, although much of this has focused on neutrophil migration, for which LFA-1, Mac-1 (another ICAM-1 ligand), and ICAM-1 appear to be important (14, 15). Monocyte trafficking is of particular interest since monocytes are the dominant glomerular leukocyte in progressive GN and probably play a major part in the irreversible scarring, which ultimately leads to end stage renal failure (16). Using a blocking mAb approach, Mac-1 was not found to be important for monocyte migration into glomeruli by Wu et al. (17), but Kawasaki et al., with a similar approach, demonstrated that both LFA-1 and ICAM-1 were critical to monocyte trafficking in a model of nephrotoxic nephritis (NTN) in the Wistar-Kyoto (WKY) rat (18). In this model, i.v. injection of a small quantity of heterologous (rabbit) anti-rat glomerular basement membrane (GBM) antiserum results in a severe, progressive form of crescentic GN (19). The cellular infiltrate in glomeruli is principally composed of monocytes, although there is also a population of CD8+ve cells, which do not bear typical T cell or NK cell markers (20), and whose prior depletion results in abrogation of nephritis (19). Further evidence in support of a key role for LFA-1 and ICAM-1 in leukocyte migration into glomeruli was provided by Nishikawa et al. in a model of experimental autoimmune glomerulonephritis in the WKY rat (21).
Very late Ag-4 (VLA-4), a heterodimer of 
4 and
ß1 integrin subunits, is another important monocyte and
lymphocyte cell surface integrin that has two major ligands: VCAM-1 and
the CS-1 peptide of fibronectin (FN) (22, 23, 24, 25). The
importance of VLA-4 in monocyte migration into glomeruli has not been
fully elucidated. In an early study examining PMN influx in the first
6 h of NTN in the Long-Evans rat, using various mAb that block
integrins, Mulligan et al. (10) showed a partial role for
VLA-4. PMN from rats express VLA-4 at levels sufficient to play a role
in PMN recruitment (26), which explains this effect. In a
model of NTN in Lewis rats, Wu et al. (27) duplicated the
effect on glomerular PMN migration (20% reduction in the first 24
h) using the same anti-VLA-4 mAb (TA-2). However, the mAb did not
attenuate proteinuria or monocyte migration into glomeruli up to
72 h after administration of anti-GBM Ab. It should be noted
that TA-2 binds to the 4 integrin chain and also
inhibits function of the 4ß7 integrin, a
leukocyte receptor for the mucosal addressin cell adhesion molecule
(MadCAM-1). To date, however, there is no evidence of either MadCAM-1
expression in the kidney, or recruitment of
4ß7-expressing leukocytes to this organ.
The only other study to examine directly the role of VLA-4 in renal
disease was in the mercuric chloride model of interstitial nephritis,
and it demonstrated a reduction in interstitial lymphocyte infiltration
in animals receiving anti-VLA-4 mAb (28, 29). A very
recent study in the same model suggested that anti-VCAM-1 mAb also
inhibits leukocyte recruitment into the renal interstitium
(29). Significantly, leukocyte influx into glomeruli is
not a feature of the mercuric chloride model.
We wished to clarify the importance of VCAM-1 and VLA-4 in monocyte migration into glomeruli in inflammatory GN, beyond the first wave of PMN influx. We first quantified and examined the time course of expression of VCAM-1 on endothelial cells in a model of NTN in the WKY rat. We achieved this by immunohistochemistry and by using a dual radiolabeled mAb technique. We then compared the roles of VLA-4 and VCAM-1 in the development of GN by giving blocking mAb up to day 7 of nephritis, by which time severe crescentic glomerular injury has commenced.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Male WKY rats aged between 6 and 8 wk (weight 200–250 g) were purchased from Charles River Laboratories (Margate, U.K.) and fed standard rat chow and water ad libitum. They were housed individually for 24 h in metabolic cages for urine collection prior to sacrifice. All procedures were performed according to United Kingdom Home Office regulations.
Antibodies
Blocking murine IgG1 anti-rat 4 integrin mAb
TA-2 (30, 31) was purchased from PharMingen (San Diego,
CA). Blocking murine IgG2a anti-rat VCAM-1 mAb (5F10) (32, 33) and isotype-matched controls P1.17 (IgG2a) and 1E6 (IgG1)
were purified as described (30, 33) and contained <5.3
EU/mg protein.
Nephrotoxic nephritis model
Heterologous nephrotoxic serum (NTS) was prepared in rabbits by standard methods. Briefly, rat glomeruli were isolated, sonicated, and lyophilized to provide a rat GBM preparation with minimal (<5%) tubular contamination. Albino rabbits were immunized with rat GBM in CFA, then boosted at monthly intervals with GBM in IFA until high titer anti-GBM antiserum was obtained at test bleeds. A total of 0.1 ml of NTS was then injected i.v. into WKY rats, leading to the development of NTN. The rats were sacrificed at various stages of nephritis up to day 28, according to the experimental protocols.
Assessment of renal disease
Urine was collected in metabolic cages for 24 h prior to harvest (days 6–7). Urinary albumin concentration was determined by rocket immunoelectrophoresis, and total albumin excretion over 24 h was calculated (34). Peripheral blood was taken at harvest for determination of serum creatinine concentration using Olympus reagents and an Olympus AU600 analyzer (Olympus, Eastleigh, U.K.). Urinary creatinine concentration was also measured (Bayer RA-XT, Newbury, U.K.) to permit calculation of creatinine clearance.
At harvest, rats were perfused with 50 ml PBS by aortic cannulation and the kidneys removed. One portion of kidney was fixed in formaldehyde and embedded in paraffin. Sections were cut, stained by hematoxylin-eosin and periodic acid-Schiff, and examined in blinded fashion by an experienced renal histopathologist (H.T. Cook). The percentage of glomeruli containing crescents was determined by examining 50 consecutive glomeruli per section. Similarly, the mean number of apoptotic bodies contained within glomeruli (excluding areas of necrosis) was calculated from examination of 50 consecutive glomeruli per section.
Computer-aided image analysis of glomerular fibrinoid necrosis
To quantify fibrinoid necrosis within glomeruli, paraffin-embedded kidney sections were stained using a standard MSB trichrome technique (Martius Yellow, Brilliant Crystal Scarlet, and Aniline Blue) which stains fibrin a red/orange color. Sections were then examined under x200 magnification using an Olympus BX40 microscope (Olympus Optical, London, U.K.) mounted with a Photonic Science Colour Coolview digital camera (Photonic Science, East Sussex, U.K.). Images were captured and analyzed using Image-Pro Plus software (Media Cybernetics, Silver Spring, MD), and color segmentation was performed for each slide individually, defining pixels that contained appropriate red/orange coloration. For each slide, 50 consecutive glomeruli were defined as areas of interest, and the percentage of each glomerular cross-sectional area comprising fibrin was calculated. The final value for each slide was derived by taking the mean of the results from the 50 glomeruli.
Immunohistochemistry
The mAb used were 5F10, ED1 (anti-CD68-like molecule in cytoplasm of rat monocytes/macrophages), ED3 (anti-sialoadhesin), HIS19 (anti-rat class II MHC, RT1b), PC10 (anti-proliferating cell nuclear Ag) and clone 6 (anti-inducible nitric oxide synthase (iNOS), Transduction Laboratories, Lexington, KY). Apart from 5F10 and clone 6, all mAbs were purchased from Serotec (Oxford, U.K.). Kidney tissue was snap frozen in isopentane and then immersed in liquid nitrogen and subsequently stored at -70°C. Five-micrometer sections were cut on a cryostat and endogenous peroxidase was blocked using 0.03% hydrogen peroxide in methanol. Sections were blocked with normal rabbit serum and incubated overnight at 4°C with 5F10, ED3, or HIS19. Sections were washed and bound Ab was detected using the avidin-biotin complex method (Dako, Glostrup, Denmark) with the chromogenic substrate 3,3'-diaminobenzidine (DAB). Sections were counterstained with hematoxylin and mounted. For ED3 and HIS19, positive cells were counted in at least 20 glomeruli by HTC in blinded fashion.
For both ED1 and anti-iNOS Abs, paraffin-embedded sections were first heated in 10% citrate buffer in a microwave oven. Sections were then treated as for frozen tissue with blocking of endogenous peroxidase, blocking with normal sheep serum, and incubation overnight with primary mAb at 4°C. Sections were washed, incubated with peroxidase-conjugated sheep anti-mouse Ig Fc (Amersham, Buckinghamshire, U.K.), and bound Ab was detected with 3,3'-diaminobenzidine. Sections were counterstained with hematoxylin, and mounted and positive cells were counted in at least 50 consecutive glomeruli by HTC in blinded fashion.
FACS analysis of glomerular leukocytes
Glomeruli were isolated from kidneys by sequential sieving, counted, then enzymatically digested with trypsin, collagenase, and DNase, as previously described (35), to yield a single cell suspension. Cells were washed in HBSS (Sigma, St. Louis, MO) with 5% FCS (HBSS-FCS), suspended at a concentration of 5–6 x 106/ml, then incubated with primary Ab at 4°C for 45 min. The Abs used were: MRC OX1 (anti-CD45, rat leukocyte common Ag), MRC OX8 (anti-CD8), MRC OX42 (anti-CD11b/c), and MRC OX23 (anti-human complement factor H) as control (all from Serotec), plus BSA3 (anti-CD11a) and BSA4 (anti-CD18), (both from British Biotechnology, Oxford, U.K.). After washing, cells were incubated with fluorescein-conjugated rabbit F(ab')2 anti-mouse Ig (Dako) with 5% normal rat serum for 30 min at 4°C. Cells were then washed again in HBSS-FCS and underwent FACS analysis (Coulter EPICS, Luton, U.K.) to quantify positively-stained cells after subtraction of background and control mAb binding. The number of positively stained cells per glomerulus was then calculated for each of the animals for each of the mAbs used. Positive controls receiving NTS alone in previous experiments gave comparable glomerular cell numbers to the contemporaneous controls of this experiment, assuring reproducibility of results.
Radiolabeled Abs
Anti-VCAM-1 (5F10) and isotype-matched control mAb (P1.17, IgG2a) were radiolabeled with 125Iodine (125I) and 111Indium (111In), respectively (Amersham), as previously described (36, 37). To quantify endothelial VCAM-1 expression, a mixture of approximately 25 µg each of radiolabeled 5F10 and P1.17 was injected i.v. into control rats (n = 5) or rats with NTN of variable duration (days 1, 2, 5, 7, 14, 21, and 28, n = 5 per group). The activity of injected material was calculated precisely by weighing syringes before and after injection. After exactly 5 min of circulation time, rats were perfused out via aortic cannulation and both kidneys and liver harvested. This short circulation time minimizes the quantity of mAb able to cross the endothelial barrier to bind subendothelial ligand. A short circulation time also reduces the possibility of VCAM-1 shedding from the endothelial cell surface with subsequent redistribution to the spleen, as observed by Harrison et al. (37). Thus organ uptake of anti-adhesion molecule mAb, after subtraction of control mAb binding, reflects endothelial Ag expression in a quantitative manner (37). The organs were weighed, then placed in an automated gamma counter along with isotope standards (to calculate channel crossover) and a known volume of injected mAb mixture (to calculate injected dose). Correction was made for background and crossover between channels, then data were expressed as counts per minute per gram of tissue, and as percentage of injected dose (%ID) per gram of tissue. The %ID/g for control mAb 111In-P1.17 was subtracted from the %ID/g for 125I-5F10 to yield the final corrected %ID/g.
Treatment protocols
The day of induction of NTN by NTS injection was designated day 0. Animals were killed on day 7 by being bled out while under isoflurane anesthesia. Rats received mAb by i.p. injection on days -1, 1, 3, and 6. TA-2 and its control 1E6 were used at a dose of 2.5 mg/kg. 5F10 and its control P1.17 were used at a dose of 5 mg/kg, after a pilot study demonstrated little effect at a dose of 2.5 mg/kg. Studies consisted of six animals receiving anti-VCAM-1 or anti-VLA-4 and five animals receiving the IgG1 or IgG2a control mAb. To ensure that the effects of mAb were not mediated by leukocyte depletion, peripheral blood was taken from animals at harvest into EDTA-tubes and cell counts performed on a Sysmex SE 9000 analyzer (Sysmex, Milton Keynes, U.K.).
Statistics
Data are expressed as mean ± SEM. Comparison between test and control groups for in vivo blocking mAb experiments was by two-tailed Mann-Whitney U test. For radiolabeled mAb uptake data, comparison between groups was by analysis of variance with Kruskal-Wallis test (nonparametric), with multiple comparisons between control and other time points made by Dunnett posttest. Differences were considered significant if p < 0.05. Statistical calculations were performed using Prism software (GraphPad Software, San Diego, CA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
NTN was induced in WKY rats, and kidneys were harvested at various
time points for analysis of VCAM-1 protein expression by
immunohistochemistry with the mAb 5F10. Normal WKY rat kidney expressed
little VCAM-1, although the parietal epithelial cells of Bowman’s
capsule were noted to express some VCAM-1 constitutively (Fig. 1
). VCAM-1 was not seen on the luminal
surface of normal glomerular endothelial cells. Occasional interstitial
cells were positive for VCAM-1, as were some peritubular capillaries
and very occasional basolateral surfaces of tubular epithelial cells.
Larger venules and arterioles expressed low levels of VCAM-1 along the
endothelial surface. At day 4 of nephritis, little difference was seen
in the expression pattern of VCAM-1 compared with baseline, although
cellular proliferation was seen in glomeruli at this time point. From
previous work (both FACS analysis and immunohistochemistry), glomerular
leukocytes peak in number at around day 3–4 in this model. By day 6,
however, up-regulation of VCAM-1 was evident, with substantial amounts
of protein present within glomeruli, particularly on cells within
crescents and cells within the tuft. Glomerular endothelial cell VCAM-1
expression was noted at this time point and beyond. VCAM-1 was also
expressed more densely on basolateral surfaces of tubular epithelial
cells, and in some peritubular capillaries, particularly by day 11 of
nephritis (Fig. 1). No VCAM-1 staining was seen on apical surface of
tubular epithelial cells at any time point.
|
Immunohistochemistry is an insensitive method of detecting
endothelial cell surface proteins. Quantitative comparisons between
different sections/animals are also unreliable. To quantify kidney
endothelial VCAM-1 expression through the course of nephritis, we chose
to use the technique of dual mAb radiolabeling, as has been used in pig
and mouse models of cutaneous inflammation and of systemic endothelial
activation (36, 37, 38). Anti-VCAM-1 mAb 5F10 was labeled with
125I, control mAb P1.17 was labeled with 111In,
and a mixture of the two mAbs was injected i.v. into normal and
nephritic rats. Uptake of both mAbs onto endothelium was measured for
kidneys and liver after a short circulation time of 5 min, and
subtraction of control mAb counts from anti-VCAM-1 counts allowed
reproducible quantification of endothelial Ag expression. By this
technique, normal rat kidneys expressed low levels of endothelial
VCAM-1, which was unchanged at day 2 of nephritis (Fig. 2). Between days 2 and 5, kidney
endothelial VCAM-1 was significantly up-regulated by approximately
3-fold (p < 0.01 for all time points after day 5
inclusive) and remained elevated until day 28. For comparison, liver
was examined as an unaffected organ, and there was no significant
change in endothelial VCAM-1 at any time point of nephritis compared
with the control. It should be noted that this technique does not
distinguish between glomerular and interstitial vascular endothelial
VCAM-1 expression. To ensure that we were detecting glomerular VCAM-1,
rats at day 6 of NTN were injected with 0.5 mg of cold mAb (control or
anti-VCAM-1) and perfused after a 5-min circulation time, as for
radiolabeled mAb studies. Indirect immunohistochemistry for mouse Ig on
frozen kidney sections from these rats demonstrated minimal specific
binding of control mAb but extensive glomerular binding of
anti-VCAM-1 mAb (Fig. 1
).
|
Having demonstrated substantial up-regulation of VCAM-1 within the kidney during NTN, the effects of preventive therapy with blocking mAb to either VCAM-1 or to its ligand, VLA-4, were determined. A 7-day study was performed, with mAb being given on days -1, 1, 3, and 6. The effects of mAb therapy on leukocyte migration into glomeruli was examined both by immunohistochemistry and by FACS analysis of glomerular leukocytes. Macrophage infiltration was first examined by ED1 immunohistochemistry and the results are shown in Figs. 3A and 4; neither anti-VLA-4 nor anti-VCAM-1 therapy led to any reduction in glomerular macrophages. Similarly, FACS analysis of glomerular leukocytes (Fig. 3B) showed that treatment with either anti-VLA-4 or anti-VCAM-1 led to no reduction in glomerular leukocyte numbers, including the important CD8+ve population of cells, as compared with controls. Of note, mAb therapy was not associated with any significant change in numbers of circulating neutrophils, lymphocytes, or monocytes (data not shown).
Effects of anti-VCAM-1 or anti-VLA-4 mAb therapy on renal injury
The effects of mAb therapy on renal injury as reflected by
albuminuria are shown in Figure 5A. An 82% reduction in
albuminuria was seen after treatment with anti-VLA-4 compared with
the control (p < 0.01). Anti-VCAM-1 mAb, by contrast,
reduced albuminuria by only 15% (not significant). In terms of
glomerular fibrinoid necrosis, an important precursor to crescent
formation (39, 40), anti-VLA-4 therapy led to a 74%
reduction (p < 0.05), whereas anti-VCAM-1 had no
significant effect (Figs. 4 and 5B). Anti-VLA-4 therapy also
reduced glomerular crescent formation by 58% (p <
0.01), and anti-VCAM-1 had a small effect (21% reduction,
p = 0.05) (Fig. 5C). By day 7 of the NTN
model, renal failure had not yet developed, and therefore the results
for serum creatinine and creatinine clearance for all four groups of
animals were within normal limits, with no differences observed between
groups (data not shown).
|
|
Simple counting of glomerular macrophage numbers does not allow
firm conclusions to be drawn regarding monocyte migration since it has
recently been shown that macrophages can proliferate within the kidney
(41), and emigration or apoptosis are other potential
fates awaiting glomerular leukocytes (42). We therefore
examined intraglomerular apoptosis by counting intraglomerular
apoptotic bodies (excluding necrotic areas). This technique correlates
well with results obtained by TUNEL or propidium iodide staining
(42), and demonstrated (Fig. 6A) that no difference in the
extent of intraglomerular apoptosis was evident at day 7 between any of
the experimental groups. Proliferation of leukocytes was assessed by
immunohistochemistry for intraglomerular expression of the
proliferating cell nuclear Ag (Fig. 6B), which demonstrated
no significant difference between groups.
|
).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We first determined whether VCAM-1 was expressed within nephritic glomeruli in the NTN model in the WKY rat. The immunohistochemistry confirmed the results of others (43, 44), showing that normal glomeruli express very little VCAM-1, except upon the parietal epithelial cells of Bowman’s capsule. No glomerular endothelial cell VCAM-1 was evident by this technique, although one cannot exclude the presence of small amounts of endothelial VCAM-1 below the detection limit of immunohistochemistry. However, as others have observed, glomerular endothelial VCAM-1 is up-regulated during GN, in both humans and animal models (10, 44, 45, 46), and was detected after day 4 in our model. Of note, VCAM-1 was also expressed elsewhere in inflamed glomeruli, with crescent cells and probably mesangial cells showing substantial staining with the anti-VCAM-1 mAb.
Immunohistochemistry does not offer a reliable means of quantifying endothelial adhesion molecule expression, and we approached this issue by measuring the binding to endothelium of differentially radiolabeled mAb in vivo. The technique of radiolabeling mAb, injecting tracer amounts i.v., harvesting organs after a short circulation time, and then counting isotopic activity has been accepted as a useful way of quantifying changes in endothelial adhesion molecule expression (36, 38, 47). In this study, 125I-labeled anti-VCAM-1 localization to kidney was determined, and standardized for injected dose and organ weight. Importantly, localization of 111In-labeled isotype-matched control mAb was then subtracted to correct for mAb penetration into tissues beyond endothelium (particularly if the endothelium is injured/leaking) and incomplete organ perfusion.
The results suggest that there is constitutive expression of
endothelial VCAM-1 within normal kidney, in that the corrected
percentage of injected dose localizing to kidney for anti-VCAM-1
mAb was not zero (
0.5%), although competition with excess unlabeled
mAb was not carried out. Of note, this technique is not able to
distinguish between the various vascular compartments of the kidney,
and the precise location of basal VCAM-1 expression is unclear. In
agreement with the immunohistochemistry results, VCAM-1 up-regulation
was not apparent in kidney by day 2 of nephritis, but had risen 3-fold
by day 5 and remained at this level until day 28. No interstitial
damage or inflammation is apparent in this model until day 10 at the
earliest, so it is highly likely that the majority of the early
up-regulation of VCAM-1 occurred in glomeruli. There may be a later
shift in expression to peritubular capillaries when tubulointerstitial
inflammation becomes widespread, coincident with a fall in total
glomerular endothelial surface area because of diffuse
glomerulosclerosis.
It is interesting to consider the significance of the different expression patterns of ICAM-1 and VCAM-1 during glomerulonephritis. Our previous work shows that leukocyte (principally monocyte) migration into glomeruli starts during day 1 and peaks at around day 3–4 in the NTN model in WKY rats (62). We (data not shown) and others (18, 48, 49) have shown a substantial amount of ICAM-1 to be present on the endothelium of normal rat glomeruli, which is rapidly (within 1 h) up-regulated after injection of NTS. Kawasaki et al. (18), in a model of NTN similar to ours, demonstrated convincingly that LFA-1 and ICAM-1 are critical for leukocyte migration into glomeruli, and that blockade of these molecules substantially reduced glomerular damage. Others have confirmed the importance of the LFA-1/ICAM-1 axis in directing both PMN and monocyte migration into glomeruli in various models of nephritis (10, 21, 48, 50) including ICAM-1 knockout mice (51). Knockout studies are not possible for VCAM-1, however, since VCAM-1 knockout mice are nonviable through failure of placentation and abnormal cardiovascular development (52, 53). Thus, leukocyte recruitment into glomeruli appears largely ICAM-1 dependent, and the absence of VCAM-1 up-regulation with the appropriate time course is consistent with this.
In the only study to address the role of VCAM-1 in nephritis, Escudero et al. gave 5F10 to rats with mercuric chloride-induced nephritis and observed inhibition of interstitial cellular infiltrates (29), glomerular leukocyte recruitment not being a feature of this model. Previous examination of the role of VLA-4 in NTN has been limited to the very early stages of nephritis (10, 27). Monocyte migration into glomeruli, studied up to day 3 in NTN by Wu et al. (27), was found to be unaffected by blocking mAb to VLA-4. They also observed no reduction in proteinuria in animals receiving anti-VLA-4. Our data on VCAM-1 expression would suggest that looking only at the first 3 days of nephritis might miss any VCAM-1/VLA-4-dependent leukocyte migration since endothelial VCAM-1 is not up-regulated until day 3 at the earliest in our model of NTN.
Therefore, our study is the first to analyze the pathophysiological
role of both VCAM-1 and VLA-4 in experimental GN beyond the first wave
of PMN-dominated leukocyte migration. To study these adhesion
molecules, we used the well-characterized adhesion blocking murine mAb
5F10 (which binds an epitope on domains 1 and 2 of rat VCAM-1)
(33) and TA-2 (which binds the rat 4
integrin chain) (30). Both immunohistochemistry of
glomeruli and FACS analysis of isolated glomerular cells demonstrated
that neither mAb reduced leukocyte recruitment into glomeruli. It is
unlikely that higher doses of 5F10 would be effective at reducing
recruitment for three reasons. First, no dose-response effect was
observed when the 5F10 dose was doubled from 2.5 to 5 mg/kg (pilot
study, data not shown). Second, 5F10 at lower doses than 5 mg/kg has
proved highly effective at blocking leukocyte migration in vivo in
other studies, including migration into kidney interstitium (29, 33). Third, 5F10 administration led to a significant effect on
macrophage activation markers (see below) suggesting that the dose used
was biologically effective. Therefore, we can suggest that VCAM-1 has
little role in early leukocyte recruitment into glomeruli in this
model, and the significance of its up-regulation is unknown. We have
not formally excluded the possibility that anti-VCAM-1 mAb may
itself function as a proinflammatory ligand by presenting mAb Fc domain
to adjacent cells. In vivo studies with Fab or F(ab')2
fragments are not feasible in view of the enormous quantities of mAb
required, but such an "Fc" effect has not been described in the
many studies that have demonstrated efficacy of whole mAb (both
anti-VCAM-1 and anti-ICAM-1) in blocking both leukocyte
recruitment and tissue inflammation in other models (10, 18, 21, 29, 33, 50). It is of particular note that inhibition of either
VCAM-1 or 4 integrin suppresses leukocyte recruitment
into the interstitium in mercuric chloride-induced nephritis (28, 29). The cell adhesion pathways utilized by leukocytes migrating
into glomeruli thus appear to differ from those used for interstitial
trafficking.
Despite the absence of inhibition of leukocyte recruitment, the TA-2
dose of 2.5 mg/kg was effective at reducing renal injury (reflected by
albuminuria) by over 80% at day 6–7 of nephritis, whereas 5F10 had no
significant effect at 5 mg/kg. The discrepancy between the results with
anti-VLA-4 and anti-VCAM-1 is interesting. Although VCAM-1 and
VLA-4 mAbs show similar efficacy in some in vivo systems of leukocyte
recruitment (29, 33), in others anti-VLA-4 is an
effective inhibitor whereas anti-VCAM-1 is not (54).
Presumably alternative VLA-4 ligands are operative within glomeruli in
NTN, and the most obvious candidate is the N-terminal V region of FN,
the rat homologue of the CS1 peptide of the CSIII domain of FN
(24). FN is present within the mesangium of normal
glomeruli (55), and during GN different isoforms are
up-regulated, with different patterns in acute leukocyte-rich and
chronic fibrotic stages of disease (56, 57, 58). In
particular, Alonso et al. (57) showed that the V120
protein isoform of FN, containing the rat VLA-4 ligand, was
up-regulated in glomeruli in an immune complex model of crescentic GN
in rats. We have confirmed, by RT-PCR, that mRNA encoding V120 FN is
present in normal and nephritic glomeruli in the NTN model of WKY rats
(data not shown). There are other, weaker, VLA-4 ligands such as
thrombospondin, and other domains within FN, but their functional
relevance is unclear since they bind only to a conformation of the
VLA-4 integrin present on highly activated cells (25). It
remains possible that there are unidentified VLA-4 ligands within
glomeruli. Finally, there is no evidence that kidney endothelium
expresses MadCAM-1, which would be another possible explanation for the
efficacy of anti-4 integrin, but not anti-VCAM-1
(by blocking 4ß7 integrin).
Together with the effects on albuminuria, anti-VLA-4 mAbs, but not
anti-VCAM-1 mAbs, were effective at reducing glomerular crescent
formation and fibrinoid necrosis. Our observations suggest that VLA-4
is implicated in tissue injury, mediated by non-VCAM-1 ligands. The
dominant glomerular leukocyte in NTN in the WKY rat is the
monocyte/macrophage, which expresses VLA-4 (25). Our
hypothesis, therefore, is that VLA-4 expressed on macrophages binds a
ligand within glomeruli, which signals to the macrophage. This signal
results in the macrophages developing a toxic phenotype, leading to
tissue damage, fibrin deposition, and crescent formation. Classical T
lymphocytes (expressing CD3 and TCR ß or 
) are not present
in the WKY rat NTN model in significant numbers within the first week,
a point of relevance since activated T lymphocytes express VLA-4, and
mAb effects in NTN could theoretically be mediated via such a cell
population. CD8+ leukocytes are critical to renal injury in
this model, as shown by depletion studies, but these cells are not
classical T cells, do not express NK cell markers (49),
and a subset of them coexpresses the macrophage marker
ED1,4 suggesting that at least
some are macrophages. Others have recently demonstrated CD8 expression
by rat alveolar and peritoneal macrophages, which differs from T cell
CD8 in the ligand-binding domain of the -chain, but has a signaling
function (59, 60).
To examine the state of macrophage activation within glomeruli, we
performed immunohistochemistry for several markers of macrophage
activation, including sialoadhesin, class II MHC, and iNOS. It was
surprising that, despite the significant reduction of injury,
anti-VLA-4 was associated with a significant increase in glomerular
cells expressing class II MHC and sialoadhesin. A nonsignificant rise
in iNOS-expressing cells was seen as well. The anti-VCAM-1 Ab
caused an even greater increase in class II and sialoadhesin expression
within glomeruli, suggesting that although it has no role in disease
pathogenesis, glomerular VCAM-1 is functional. The possibility of mAb
Fc-mediated effects on tissue activation has not been formally excluded
but, as noted above, such a proinflammatory phenomenon has not been
described in previous anti-adhesion molecule studies using intact
mAb (10, 18, 21, 29, 33, 50). It is clear that there is
dissociation between "activation markers" and glomerular
inflammation in this model. It is also becoming apparent that
macrophages do not simply operate in binary states of "quiescent"
or "activated." For example, recent work on LPS-priming of murine
macrophages suggests that there are two states inducible by LPS, with
reciprocal regulation of TNF- and nitric oxide production according
to cytokines expressed during priming (61). It is,
therefore, not possible simply to correlate the presence of
"activation markers" with macrophage-mediated injury—a greater
understanding of the various macrophage phenotypes is required before
we can meaningfully interpret such data.
In conclusion, we have demonstrated that VCAM-1 is up-regulated in the glomeruli of rats with experimental GN, but with a delayed time course compared with ICAM-1. The expression of VCAM-1 is probably not critical to monocyte migration into glomeruli. Monoclonal Abs to VLA-4 significantly attenuate renal injury, although without reduction in either leukocyte numbers or classical markers of macrophage activation. Alternative ligands for VLA-4 are implicated in mediating glomerular injury at a point after transendothelial migration of monocytes. Future studies will address the role of VLA-4 in the processes underlying progressive glomerulosclerosis and chronic renal failure.
|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Andrew Allen, Renal Section, Division of Medicine, Imperial College School of Medicine, Hammersmith Hospital, Du Cane Road, London W12 0NN, U.K. E-mail address:
3 Abbreviations used in this paper: GN, glomerulonephritis; NTN, nephrotoxic nephritis; WKY rat, Wistar-Kyoto rat; GBM, glomerular basement membrane; FN, fibronectin; MadCAM-1, mucosal addressin cell adhesion molecule; NTS, nephrotoxic serum; %ID/g, percentage of injected dose per gram of tissue; VLA-4, very late Ag-4; PMN, neutrophil.
Received for publication October 20, 1998. Accepted for publication February 16, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
4 Integrins in vivo. J. Clin. Invest. 94:1722.
4 integrin. J. Immunol. 153:2313.[Abstract]
4 integrin and to VCAM-1. J. Am. Soc. Nephrol. 9:1881.[Abstract]
-induced eosinophil accumulation in rat skin is dependent on 4 integrin/vascular cell adhesion molecule-1 adhesion pathways. Blood 90:4144. receptor-dependent neutrophil adhesion and complement-dependent proteinuria in acute glomerulonephritis. J. Exp. Med. 186:1853. and CD8ß induce nitric oxide production and associated killing of the parasite Leishmania major. J. Immunol. 160:6004.This article has been cited by other articles:
|
|
 |
|
  S. A. Asgeirsdottir, J. A. A. M. Kamps, H. I. Bakker, P. J. Zwiers, P. Heeringa, K. van der Weide, H. van Goor, A. H. Petersen, H. Morselt, H. E. Moorlag, et al. Site-Specific Inhibition of Glomerulonephritis Progression by Targeted Delivery of Dexamethasone to Glomerular Endothelium Mol. Pharmacol., July 1, 2007; 72(1): 121 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Timoshanko, A. R. Kitching, T. J. Semple, S. R. Holdsworth, and P. G. Tipping Granulocyte Macrophage Colony-Stimulating Factor Expression by Both Renal Parenchymal and Immune Cells Mediates Murine Crescentic Glomerulonephritis J. Am. Soc. Nephrol., September 1, 2005; 16(9): 2646 - 2656. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Ikezumi, L. Hurst, R. C. Atkins, and D. J. Nikolic-Paterson Macrophage-Mediated Renal Injury Is Dependent on Signaling via the JNK Pathway J. Am. Soc. Nephrol., July 1, 2004; 15(7): 1775 - 1784. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Bellingan, P. Xu, H. Cooksley, H. Cauldwell, A. Shock, S. Bottoms, C. Haslett, S. E. Mutsaers, and G. J. Laurent Adhesion Molecule-dependent Mechanisms Regulate the Rate of Macrophage Clearance During the Resolution of Peritoneal Inflammation J. Exp. Med., December 2, 2002; 196(11): 1515 - 1521. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. T. Cook, S. B. Khan, A. Allen, G. Bhangal, J. Smith, R. R. Lobb, and C. D. Pusey Treatment with an Antibody to VLA-1 Integrin Reduces Glomerular and Tubulointerstitial Scarring in a Rat Model of Crescentic Glomerulonephritis Am. J. Pathol., October 1, 2002; 161(4): 1265 - 1272. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. C. Kluth, C. V. Ainslie, W. P. Pearce, S. Finlay, D. Clarke, I. Anegon, and A. J. Rees Macrophages Transfected with Adenovirus to Express IL-4 Reduce Inflammation in Experimental Glomerulonephritis J. Immunol., April 1, 2001; 166(7): 4728 - 4736. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Karkar, J. Smith, and C. D. Pusey Prevention and treatment of experimental crescentic glomerulonephritis by blocking tumour necrosis factor-{{alpha}} Nephrol. Dial. Transplant., March 1, 2001; 16(3): 518 - 524. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Y. Larbi, A. R. Allen, F. W. K. Tam, D. O. Haskard, R. R. Lobb, P. M. R. Silva, and S. Nourshargh VCAM-1 has a tissue-specific role in mediating interleukin-4-induced eosinophil accumulation in rat models: evidence for a dissociation between endothelial-cell VCAM-1 expression and a functional role in eosinophil migration Blood, November 15, 2000; 96(10): 3601 - 3609. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
