Original Investigation
JNEPHROL 2000; 13: 291-316
Membranous glomerulonephritis (MGN)
Principal discussant: Walter H. Hörl 1, Pathologist's comment: Dontscho Kerjaschki 2 - 1 Division of Nephrology and Dialysis, University Hospital, Vienna - Austria 2 Institute of Clinical Pathology, School of Medicine, Vienna - Austria
CASE PRESENTATION
Case 1. A 63-year-old female was admitted to our department in June 1997 because of profound peripheral oedema. She had been well three months before admission and her prior medical history was unremarkable. The course of proteinuria and creatinine clearance is shown in Figure 1.
Fig. 1 - Case 1. 18 month course of proteinuria and creatinine clearance. Improvement following steroid and cyclophosphamide therapy is depicted.
Renal investigation revealed a proteinuria of 8.5 g/day with a
serum albumin of 23 g/L. Urine analysis did not show hematuria or leukocyturia. Serum
creatinine was within the normal range, however, creatinine clearance was reduced to 57
ml/min. Profound hypercholesterolemia and hypertriglyceridemia were found. Hematological
parameters and liver tests were normal. Immunology showed no changes in immunoglobulin
levels and there was no monoclonal band in serum. Rheumatoid factor levels and ANCA were
negative. Complement factors (C3, C4) and anti-DNA antibodies were normal; however, a
marginal increase in antinuclear antibodies was found (1:20). A percutaneous renal biopsy
was performed (see comment of the Pathologist)
Serology for HBV, HCV and HIV was negative. Tumor markers were found to be
negative. Chest X-ray, abdominal echography as well as gynecologic evaluation including
mammography were unremarkable. There was no evidence of occult gastrointestinal bleeding.
No possible etiologic agents (drugs or toxins) could be identified. It was concluded that
the patient suffered from idiopathic membranous glomerulonephritis.
Initially, the patient received an ACE inhibitor (Ramipril at 1.25 to 2.5
mg/day), an HMG-CoA reductase inhibitor (simvastatin at 10 to 20 mg/day, then switched to
atorvastatin at 10 mg/day) and diuretics. The patient received furosemide, during the
first month together with xipamide. As a result of dietary sodium restriction, 24-hour
sodium excretion was generally below 150 mmol. Because of persistence and severity of the
nephrotic syndrome (proteinuria was 12.6 g/24 h) three months later, cytotoxic therapy was
started. The patient was treated with cyclophosphamide (initial dose of 2 mg/kg/day) and
prednisolone (initial dose of 1.5 mg/kg/day). Under this therapy proteinuria continuously
decreased achieving levels below 1 g/day after six months. Complete remission was observed
9 months after admission (Fig. 1). Cyclophosphamide was discontinued 9 months and
prednisolone 12 months after initiation of therapy.
Case 2. This 53-year-old female was referred to our hospital in
January 1994 for evaluation of proteinuria and peripheral oedema. With the exception of an
infection of the lower urinary tract about 20 years before, her medical history was
unremarkable. A 24-hour urine collection contained 2.6 g of protein. Laboratory
examination demonstrated hypoalbuminemia and marked hypercholesterolemia. Serum creatinine
was normal, however, GFR was found to be slightly reduced. A thorough workup did not
reveal evidence of an infectious, immunologic or malignant disease. A renal biopsy
revealed the diagnosis of a membranous glomerulonephritis (see comment of the
Pathologist). During the first two months proteinuria increased to 4.1 g/day and
immunosuppression with cyclosporin A (2 to 3 mg/kg/day) was initiated because of severe
nephrotic syndrome and complaints of the patient. However, CSA therapy failed to reduce
proteinuria (urinary protein excretion was in a range of 6 g/day and one day exceeded 10
g/day) so that after another four months cytotoxic drug therapy was initiated
(chlorambucil at 0.2 mg/kg/day); after another month cyclophosphamide at 2 mg/kg/day (due
to side effects of chlorambucil) and steroids (three doses of dexamethasone at 100 mg IV,
followed by oral prednisolone at 0.5 mg/kg/day) were given for a total of five months.
Since nephrotic syndrome persisted, probucol was recommended for six months (1 g/day). As
shown in Figure 5,
Fig. 5 - Case 2. Clinical course followed for 70 months. To achieve first complete remission, treatment with probucol has been added to steroids and immunosuppressors. Instead, IVIG therapy did not improve clinical data at recurrence.
proteinuria continuously decreased achieving levels below detection limit
approximately two years after admission as a result of the combination of steroids,
cytotoxic drugs and/or probucol. Remission of the nephrotic syndrome for about four years
was observed. Then, however, nephrotic range proteinuria recurred. Intravenous
immunoglobulin (IVIG) therapy for three months was started due to the dose of
immunosuppression given before, but was without any effect. Several years after diagnosis
of membranous nephropathy, the patient acquired primary hypothyroidism
(thyroid-stimulating hormone level was > 40 mU/L and exceeded shortly 100 mU/L).
Sonographic and immunologic evaluation showing elevated levels of antibodies to thyroid
peroxidase, thyroglobulin and TSH receptor suggested autoimmune thyroiditis. This
diagnosis was confirmed by a needle biopsy.
Case 3. This 49-year-old male with end stage renal disease due to
primary membranous glomerulonephritis (renal biopsy performed 14 years before) received a
living-related donor kidney graft from his haploidentical brother in April 1997. The
post-transplant course of proteinuria and serum creatinine is depicted in Figure 9.
Fig. 9 - Case 3. The post-transplant course of proteinuria and serum creatinine is depicted. Failure of additional therapy to control clinical parameters is observed.
Initially, basal immunosuppressive therapy consisted of FK506 (oral dose
adjusted to a level of 7.5 to 15 ng/ml), mycophenolate mofetil (2 g/day) and steroids
(prednisolone was administered according to a local tapering regimen). Due to clinically
suspected early allograft rejection two months post transplantation, the patient received
steroid bolus therapy (three doses of 100 mg/day dexamethasone). Several weeks later,
FK506 was switched to cyclosporin A (oral dose adjusted to a level of 100 to 150 ng/ml)
because of tacrolimus-associated neurotoxicity. Stable kidney graft function with serum
creatinine levels between 1.5 and 1.8 mg/dl allowed steroid withdrawal one and a half year
post-transplant. Initially, urine protein excretion was within the normal range. In
November 1998 (one and a half years after transplantation), however, the patient presented
with peripheral oedema, 6 kg weight gain and new onset hypertension. Laboratory
examination demonstrated an increase of serum creatinine to 2 mg/dl and a proteinuria of 4
g/day. Under treatment with a HMG CoA reductase inhibitor (simvastatin, 20 mg/day), serum
cholesterol was below 250 mg/dl and serum triglycerides were within the normal range.
Diuretic therapy with furosemide and an ACE inhibitor (enalapril, 10 mg/day) was started.
Because of a further increase in urine protein loss, 4 months later, renal allograft
biopsy was performed, which gave the diagnosis of recurrent membranous glomerulonephritis
(see comment of the Pathologist). An increase in proteinuria to about 10 g/day led
to the reinstitution of prednisolone, however, without any effect. Four months after renal
biopsy, therapy with IVIG was administered for three months. The last outpatient control,
however, revealed heavy proteinuria (15.9 g/24 h) and a further increase of serum
creatinine (2.5 mg/dl) and decrease in creatinine clearance (58 ml/min).
COMMENT OF THE PATHOLOGIST
Dontscho Kerjaschki: Institute of Clinical Pathology, School of Medicine, Vienna, Austria
Case 1 (Fig. 2).
Fig. 2 - Case 1. By silver stain, glomerular basement membrane presents discrete spike formation on its outer aspect (200X).
Light Microscopy. The majority of the biopsy core consisted of medulla with a small
portion of renal cortical tissue containing only 4 glomeruli. All glomeruli showed uniform
changes. The glomerular basement membranes (GBM) were slightly thickened. On silver stain
they showed discrete spike formation on their outer aspects or a reticular pattern of the
GBM material if sectioned tangentially. There was no mesangial matrix increase or
mesangial hypercellularity. Extraglomerular blood vessels showed no pathologic changes.
The interstitium was focally slightly fibrotic with a few atrophic tubular cross sections
in areas of fibrosis. Immunohistochemistry. Immunohistochemistry on formalin fixed
paraffin sections with antibodies against IgG, IgM, IgA, C3, C1q and fibrin revealed a
strong granular signal only for IgG, C3 and C1q along the thickened capillary walls of all
glomeruli. No immune deposits were detectable in other regions of glomeruli (Fig. 3 WEB
see WEB Edition). Electron Microscopy. Two glomeruli were examined. The most striking
changes were densely packed immune complex type deposits along the GBM. Most of the
deposits were electron dense, in a subepithelial position and separated by basement
membrane like material extending from the underlying GBM. Some of the deposits were
completely surrounded by GBM material and showed an unhomogenous loss of electron density.
The foot processes of podocytes were completely effaced (Fig. 4, WEB).
In conclusion, renal biopsy revealed the diagnosis of a membranous
glomerulonephritis stage II-III (according to Churg and Ehrenreich).
Case 2 (Fig. 6)
Fig. 6 - Case 2. By light microscopy a normal glomerulus is evident. Renal interstitium does not show abnormalities. (PAS, 100X).
Light Microscopy. Two biopsy
cores with 18 glomeruli and 2 glomerular scars have been examined. Glomeruli looked normal
in HE and PAS stains. In silver stained sections the GBM showed only minor irregularities
in their structure and spikes could be visualised only occasionally. The mesangium was
delicate, without hypercellularity. Arterioles and small arteries showed low-grade intimal
hyaline thickening. Larger arteries had no significant pathologic changes. In a few areas
of slight interstitial fibrosis without inflammatory infiltrates tubuli were atrophic but
unremarkable in all other areas. Immunohistochemistry. Immunohistochemistry on formalin
fixed paraffin sections with antibodies against IgG, IgM, IgA, C3, C1q and fibrin showed
unevenly distributed, finely granular deposits of IgG and C3 along GBM. No other deposits
of Ig or complement were present (Fig. 7, WEB). Electron Microscopy. Three glomeruli were
examined. Small electron dense deposits were seen in subepithelial position, leading to
depressions of the underlying surface of the GBM. Only occasionally true spike-formation
could be demonstrated between the unevenly distributed deposits. The podocytic foot
processes were highly but not totally effaced (Fig. 8, WEB).
In conclusion, renal biopsy revealed a membranous glomerulonephritis stage
I-II (according to Churg and Ehrenreich).
Case 3 (Fig. 10).
Fig. 10 - Case 3. A reticular pattern of the glomerular basement membrane and presence of spikes are clearly shown by silver stain (200X)
Light Microscopy. The biopsy was adequate (2 glomeruli and 3 arterial
cross-sections). All glomeruli showed identical changes, i.e. low grade expansion of
mesangial matrix and slight thickening of GBM. In silver stains the GBM appeared
perforated on tangential sections. There was no glomerular hypercellularity. Signs of
acute rejection were present neither in arteries nor in tubulo-interstitial space. Fibrous
intimal thickening and slight interstitial fibrosis indicated chronic allograft
nephropathy. Immunohistochemistry. Immunohistochemistry with antibodies against IgG, IgM,
IgA, C3, C1q and fibrin revealed finely granular deposits positive for IgG, C3 and C1q
along GBM of all glomeruli (Fig. 11, WEB). No glomeruli were available for Electron
Microscopy.
In conclusion, percutaneous renal allograft biopsy showed recurrent
membranous glomerulonephritis.
OVERWIEW ON THE PATHOLOGY OF MEMBRANOUS GLOMERULOPATHY (MGN)
Dontscho Kerjaschki: Institute of Clinical Pathology,
School of Medicine, Vienna, Austria
MGN is a major cause for nephrotic range proteinuria and
chronic renal insufficiency. The characteristic glomerular lesion of MGN was originally
recognized as a histopathologic entity by the characteristic feature of diffuse thickening
of glomerular capillary walls, to form thick "membranes". Later it was shown
that this GBM thickening occurs in advanced stages of MGN that usually follows a
stereotype pattern of development (1), as briefly reviewed here. It is now firmly
established that this alteration is caused by immune complexes that accumulate
subepithelially on the outer aspect of the glomerular basement membrane (GBM). However,
subepithelial deposits and consequent GBM changes are not defining a single etiologically
homogeneous disorder (2, 3).
The identification of immune deposit formation as the characteristic
morphological feature was the basis for the differential diagnosis of MGN from other renal
diseases presenting with isolated high grade proteinuria, such as minimal change nephrotic
syndrome (MCNS) and focal segmental glomerular sclerosis (FSGS). As many authors believe
that a focused therapeutic interference with immune complex deposition could cure the
disease, the pathogenesis of their formation was and is a subject of many investigations.
Major contributions to our current understanding of the evolution of immune complex
mediated renal disease come from Heymann nephritis (4-7). The systematic evaluation of
this model eventually led to the identification of a unique constitutive glomerular
antigenic target for immune deposit forming antibodies (4). However, in humans the
situation appears to be more complex. The search for antigens of human MGN led for example
to the detection of microbial antigens (8-11) as well as several others, as briefly
reviewed in the following, but not to the detection of a common etiologic agent. As a
document of our inability to unravel the pathogenetic and etiologic factors, the majority
of MGN cases are still assigned "primary" or "idiopathic".
The idiopathic or primary form of MGN accounts for approximately 75% of all
cases of MGN. For the remaining 25% a broad spectrum of associated diseases has been
reported (12). The conditions most frequently related to MGN are neoplasms, infections,
autoimmune diseases and certain drugs. It is not always easy to judge whether there is a
pure coincidence of otherwise unrelated disorders or potentially important pathogenic
links between two diseases.
Although the wide range of different conditions associated with MGN
indicates an etiological heterogeneity, the morphological changes in humans are remarkably
common. This is especially true for the idiopathic MGN, while secondary forms of the
disease occasionally tend to a greater variability of morphological changes (13, 14).
PATHOMORPHOLOGIC CHANGES
Light Microscopy
The morphologic changes of the earliest stages of MGN may
be undetectable on sections routinely stained with Hematoxilin and Eosin (HE). And even in
more advanced stages HE is often unsuitable to unequivocally demonstrate the pathognomonic
changes of MGN due to its inability to discriminate between GBM and cellular structures
layered beneath it (Fig. 12, WEB). In light microscopy in many cases we are not able to
visualize the immune deposits themselves but rely on the reactive structural changes of
the GBM induced by immune complexes. Therefore special stains highlighting the GBM
structure like Methenamin Silver (MS) and Periodic Acid Schiff (PAS) or trichrome stain
are more useful than HE.
The earliest change detectable on silver stains is a slight irregularity on the GBM which
is smooth and absolutely homogeneous in normal kidneys (Fig. 13, WEB). This mottled
appearance is best seen on tangential sections and represents slight indentations of the
GBM by immune complexes adhering to its surface. The most specific and easily recognized
change of the GBM is the formation of so called spikes (Fig. 14, WEB). These are
projections of GBM material between immune complexes that lead to a saw tooth like
appearance of the GBM. This pattern is pathognomonic of fully developed membranous GN and,
if present, warrants an unequivocal diagnosis based on light microscopy alone (Fig. 15,
WEB). The disease progression (for definition of stages see below) results in a diffuse
thickening of GBM. This can be easily visualised with PAS or even HE stain. In idiopathic
MGN the morphologic changes described above are generally diffuse and global. That means
that all glomeruli, and within one glomerulus all segments of the capillary tuft, are
involved. Usually there are no additional signs of glomerular disease such as changes in
the mesangium or significant hypercellularity. A focal and segmental pattern of GBM
changes, a marked glomerular hypercellularity or any type of prominent mesangial injury
are suggestive of a secondary form of MGN. Likewise a prominent tubular interstitial
injury is not a feature of idiopathic GN if it exceeds a degree attributable to loss of
nephrons due to glomerular obsolescence. The only tubular interstitial change constantly
associated with MGN is the accumulation of protein droplets in tubular epithelial cells.
This is neither demonstrable in all cases nor specific for MGN but is the usual finding in
many states of high grade proteinuria. Although in most cases the GBM changes detected
with silver and PAS stain are sufficient for a reliable diagnosis, immunohistochemistry
and electron microscopy can be required in early stages of the disease and are the only
way to directly visualize distribution and extent of immune deposits.
Immunohistochemistry
The causative agent of the disease leading to subsequent
GBM damage and functional impairment of the glomerulus are the immune complexes. Their
localization and composition can be evaluated by immunohistochemistry with antibodies
against immunoglobulins and complement factors. It is known that the major constituent of
the immune complexes is IgG together with complement factor 3 (C3). Other immunoglobulins
like IgM and IgA, have been demonstrated as well as C1q and other complement components.
Membrane attack complex (C5b-9) which has been shown to play a crucial role in the
development of proteinuria in Heymann nephritis (6) can be detected in deposits of humans
as well (15). Moreover, the degree of urinary excretion of C5b-9 has
been linked to increased disease activity (16) and inferior clinical outcome (17). Several
studies investigating the IgG subclasses within the deposit surprisingly demonstrated IgG
4 as the most abundant subclass in the majority of cases of idiopathic MGN (18). In
contrast, IgG 3 has been reported to be the predominant subclass in 50% of SLE associated
MGN cases (19).
The immunoreactants show a diffuse, finally granular pattern along the GBM. Occasionally
the distribution may be focal and segmental in early stages of the disease or some forms
of secondary MGN. If the deposits are very small the staining pattern appears linear (Fig.
16, WEB). The immunohistochemical pattern represents a negative image of the GBM stained
with silver methanamin where the staining defects represent non-stained immune complexes
surrounded by argyrophilic basement membrane material (Fig. 17, WEB).
Electron Microscopy
The crucial technique for the identification of immune complex
deposits and the exact definition of their interaction with GBM and other constituents of
the glomerulus is electron microscopy (EM). EM revealed that a wide spectrum of GBM
changes seen in light microscopy is related to the subepithelial deposition of immune
complexes, their influence on GBM and their eventual resolution followed by restoration of
the GBM. Ehrenreich and Churg (1) categorized this evolution of changes into 4 classical
stages (Fig. 18, WEB):
Stage 1: Initial immune complex
deposition
Stage 2: Response of GBM with spike
formation between IC
Stage 3: Incorporation of deposits into
GBM
Stage 4: Resolution of immune complexes
and restoration of GBM
This classification has proven to be very useful in describing
the sequence of events after the initial development of immune complexes. It assumes,
however, a single hit injury and the more complex patterns of injury in the relapsing
course of the disease cannot be so easily categorised within 4 stages.
MGN Stage 1
This stage (Fig. 19, WEB) is defined by the accumulation of
predominantly small electron dense deposits between the GBM and the podocytic processes
covering the outer surface of the capillary wall. The immune complexes are closely
attached to the GBM, but initially do not alter its outer contour. Podocytic foot
processes being displaced by deposits or lying in their close vicinity are flattened and
fused. Sometimes pseudo villous transformation of the podocytic cytoplasm occurs in this
stage. In this stage the GBM seems unchanged at the light microscopy level, even in silver
stained sections. Immunohistochemistry, however, reveals a typical granular staining
pattern for Igs and complement factors.
MGN Stage 2
This stage (Fig. 20, WEB) is characterized by the appearance
of GBM-like material extending between the deposits and progressively embedding them.
These projections appear as so-called spikes on silver stains in light microscopy. Their
molecular composition and the spatial distribution of laminin and a structurally altered
Collagen IV differ from that of the normal GBM (20). The length of these projections may
exceed, depending on the size of deposits, the thickness of the underlying lamina densa by far. Some deposits are
already entirely surrounded by basement membrane material. The outer surface of the GBM is
still irregular, formed by deposits with intervening projections. In most cases the
deposits are predominantly electron dense with only few areas of unhomogeneously electron
lucent material indicating partial resolution of the immune complex.
MGN Stage 3
The characteristic pattern of stage 3 (Fig. 21, WEB) can be
described as further evolution of the changes observed in stage 2. Now the majority of
deposits are completely engulfed by basement membrane material. There is a continuous
layer of newly formed GBM on top of the deposits. These appear incorporated within a now
considerably thickened GBM. Podocytic foot processes are still fused and flattened. In
this stage the electron density as an indicator of the age of the deposit is highly
variable. In some cases the majority of deposits is still electron dense, in others most
of them show highly irregular and unhomogeneous predominantly electron lucent pattern. Now
the marked thickening of GBM can be easily appreciated under light microscopy. In silver
stains a reticular pattern reflects the so called incorporation of deposits into the GBM.
In some cases of otherwise classical stage 3 disease, electron dense subepithelial
deposits can be observed - obviously presenting newly formed immune complexes indicating a
relapsing course of the disease (Fig. 22, WEB). This type of repeated injury contributes
to a rather complex morphological pattern that sometimes cannot unequivocally be
classified as stage 3.
MGN Stage 4
Progression of stage 3 MGN leads to stage 4 Fig. 23, WEB), as
defined by Churg and Ehrenreich. Here the GBM is thickened and structured irregularly but
contains no more electron dense deposits and only occasionally electron lucent remnants of
degraded immune complexes. Consequently there is no more granular staining pattern
detectable by immunohistochemistry.
The four stages describe a sequence of glomerular injury induced by a single generation of
immune complexes ultimately leading to restoration of the basement membrane and remission
of proteinuria. Although formation of deposits is potentially reversible, the disease can
proceed to end-stage renal disease in some patients. It is thought that the continuous
deposition of nearly formed immune complexes prevents reparation of the GBM and is thus
associated with a progressive course of the disease (21).
Clinico-Pathologic Correlation
Many studies have correlated the morphological stage with
proteinuria and renal function. Although several reports demonstrate a better prognosis
for stage 1 and 2 MGN (1, 22-24), others could not confirm the predictive value of the
morphologically defined stages (25, 26). The degree of proteinuria, however, mainly
depends on the presence and number of electron dense deposits and the associated podocytic
foot process damage (21). The decrease of GFR has been attributed to an impaired hydraulic
permeability of thickened GBM and by progressive loss of functioning glomeruli (27).
Secondary MGN
Secondary forms of membranous GN have been reported in
association with a wide spectrum of different diseases which can be categorized into four
major groups: neoplasms, infections, autoimmune diseases and drugs (Fig. 24, WEB).
Neoplasms and MGN
Three lines of evidence support a causal relationship between
epithelial neoplasms and MGN:
1) Temporal association of MGN and cancer. Simultaneous occurrence or proteinuria following malignancy within one year
(28).
2) Remission of proteinuria and disappearance of the glomerular lesion
after removal of the tumor (29, 30).
3) Tumor associated antigens can be detected in the kidney (31).
The presence of both a tumor related antigen and antibody
reactive to it within the immune complexes in the glomerulus has not been demonstrated yet
(32). In most cases morphology of tumor associated MGN does not differ significantly from
its idiopathic form.
Infections and MGN
Several types of bacterial and viral infections are
complicated by MGN. The prototype of infection related glomerular disease is MGN
associated with hepatitis B. This type of secondary MGN is associated with a carrier state
and the incidence is higher in children than in adults (33). In addition to epidemiologic
data, the demonstration of the major viral antigens (HBsAG, HBcAG and HBeAG) in
subepithelial deposits favors but does not prove a causal relationship between viral
infection and MGN (8, 10). Similar to hepatitis B, viral antigen was shown in cases of MGN
associated with hepatitis C infection (9). These patients showed no co-infection with
hepatitis B, and they had no glomerular cryoglobulin deposits that are usually found in
membranoproliferative GN, a glomerular disease more commonly associated with hepatitis C.
MGN has also been linked to helicobacter pylori infection by one study, demonstrating bacterial antigen in
glomerular basement membranes (11). These data, however, have not yet been confirmed by
others.
MGN and autoimmune Disease
The most common form of secondary MGN in Europe is membranous
lupus nephritis. MGN in systemic lupus erythematosus (SLE) can closely resemble idiopathic
MGN (SLE class V, according to the WHO classification) (19, 34). A membranous pattern of
injury is very often combined with some other type of glomerular injury in SLE (Fig. 25,
WEB). Typically, in addition to subepithelial deposits, there is subendothelial and
mesangial deposition of immune complexes, usually accompanied by a prominent glomerular
hypercellularity (Fig. 26, WEB). These additional features are highly suggestive, but not
truly pathognomonic for SLE-MGN. Association with MGN has also been reported for other
autoimmune diseases, including rheumatoid arthritis, mixed connective tissue disease,
Graves' disease and others (35-37), however these associations are by far less common than
with SLE.
Drugs and MGN
For two drugs the causative relation to MGN is particularly
well established, i.e. gold salts and D penicillamine (38). Both are primarily applied to
patients with rheumatoid arthritis which by itself, may be associated with MGN, in absence
of gold or penicillamine treatment. Nonetheless, the rapid resolution of clinical and
morphological symptoms after cessation of therapy strongly argues for a causative role of
both drugs. In addition, MGN has also been reported in patients receiving penicillamine
for treatment of other diseases, like hepatitis and scleroderma (39, 40). In both
drug-induced MGN the pattern of glomerular injury closely resembles the idiopathic form of
the disease.
MGN in Transplanted Kidneys
Membranous GN belongs to the most common types of immune
complex mediated GN in renal allografts (41) (Fig. 27, WEB). It may represent a recurrence
of the patient's original disease, or a de novo glomerulonephritis, and it is not always possible to distinguish
between these two possibilities, as this would require a biopsy proven diagnosis of the
native kidney disease that is not always available.
Despite this diagnostic limitation, de
novo MGN seems to be far more common than
recurrence of the disease. In a recently published study 8 out of 30 patients with MGN in
the native kidney developed recurrence associated with an unfavourable outcome (42),
however no risk factors for recurrence were established in this study.
Both, recurrent and de novo MGN sometimes present morphological features different from
idiopathic MGN (43) (Fig. 28, WEB). Distribution of immune deposits is often segmental,
transplant glomerulopathy and glomerular hypercellularity sometimes accompany the
membranous pattern of injury (Fig. 29, WEB).
Occurrence of MGN in renal transplants is of particular
interest for the study of the pathogenesis of human MGN, because it shares several
features with experimental rat Heymann nephritis, in which inoculation with allogeneic
renal tissue (Fx1A) induces a form of MGN. However, the antigen(s) responsible for the
development of MGN in transplants are still elusive.
Pathogenetic concepts of MGN
Because molecular analysis of MGN is limited by the scant availability of human material, we and many others have chosen to study Heymann nephritis (HN) (44), an experimental rat disease which faithfully duplicates MGN. HN was originally induced by active immunization of rats with fractions of renal cortex (active HN), resulting in typical subepithelial immune deposits and heavy proteiuria (Fig. 30, WEB). Most experiments mentioned in this review were performed in rats with passive HN (45, 46). In this variant of HN a nephritogenic antibody was injected intravenously into rats, resulting in formation of subepithelial immune deposits within a few minutes, and followed by proteinuria 5-6 days later.
Formation of Immune Deposits in Heymann Nephritis
The antigenic target of circulating nephritogenic
(auto)antibodies in HN has been defined as a large membrane glycoprotein, designated
megalin/ gp330 (47). Determination of its amino acid sequence has revealed a deduced
molecular weight of 515 kD, and therefore, the original name "gp330" was
appropriately replaced by "megalin", to emphasize its size (48). Structural
similarities were found with the lipoprotein receptor-related protein LRP/a2-macroglobulin
receptor, a member of the LDL-receptor family (49, 50) (Fig. 31, WEB). Similar to LRP (51)
megalin/gp330 is a polyspecific receptor for a rapidly increasing number of newly
recognized diverse ligands (52, 53). Examples are Ca++ (53), apolipoproteins E, J (clusterin) and B 100 (54-56), uPAI
complexes (57, 58), cationic nephro- and ototoxic antibiotics (59), lactoferrin (60), and
vitamin-carrier protein complexes (61) (Fig. 32, WEB).
For the pathogenesis of immune deposit formation it is of
relevance that megalin/gp330 is expressed in clathrin coated pits on the base of
footprocesses of podocytes, where initial immune complexes are formed (62, 63) (Fig. 33,
WEB). Indirect evidence indicates that megalin/gp330 in this location forms complexes with
a 44 kD protein (4) which was designated variably as C14 (64),a 2-macroglobulin
receptor associated protein (a2-MRAP), or receptor associated protein (RAP)
(65, 66). Antibodies specific for RAP were found to induce pHN-like small immune deposits
(67, 68).
Binding Sites of Immune Deposit forming Antibodies on RAP
RAP has attracted interest because it serves as chaperon, it blocks all binding sites of megalin/gp330 and LRP. In addition, a high-affinity heparin binding site was also localized (69) and - most interesting in the context of this review - a single epitope was discovered that participates in the formation of immune deposits in HN (67). This epitope was narrowed to 14 amino acids, and specific antibodies directed against a corresponding synthetic peptide were found to induce small immune deposits similar to those found in HN (70).
Binding Sites of Immune Deposit forming Antibodies on Megalin
Despite the large size of megalin/gp330, the number of
pathogenic, i.e. immune deposit-forming, epitopes appears to be quite restricted. A
binding site for IgG eluted from glomeruli of pHN rats was localized to the 5th
cystine-rich repeat of the second LDL-receptor like domain of megalin/gp330 (48, 71), and
active and passive HN were induced with a ~ 130 amino acid long recombinant fusion protein
comprising this region of megalin (72) (Fig. 34, WEB). The relevant epitope on the 5th
cystine-rich repeat of the second LDL-receptor like domain was recently reconstructed in a
3-D molecular model and broken up into individual synthetic peptides (unpublished work).
It is realistic to expect that within the near future the molecular essentials of immune
deposit formation will be understood sufficiently to design specific immunotherapies for
HN, which eventually could serve as blueprint for human NM.
Unexpectedly, recent evidence has indicated that the epitope
on the 5th cystine-rich repeat of the second LDL-receptor like domain also
serves as major binding site for several ligands of megalin/gp330, in particular for apoE
containing lipoprotein particles. Thus, the immune deposit forming antibodies are
inhibitory for some ligands. These findings have also indicated that megalin/gp330 serves
as endocytic receptor for lipoproteins in podocytes, and blocking results in accumulation
of these ligands (73).
Mechanisms of Proteinuria in HN
Proteinuria Depends on C5b-9 Formation
Activation of complement plays a direct role in glomerular injury and proteinuria in pHN. The experimental basis for this view are the findings that C5b-9 membrane attack complex was localized within the immune deposits and that depletion of complement by cobra venom factor was associated with lack of proteinuria, while formation of immune deposits was not inhibited (reviewed in 74, 75). Apparently podocytes are in command of a defense mechanism to escape from the potentially dangerous membrane insertion of C5b-9: they retrieve C5b-9 and/or inactivated C5b-9-S protein complexes from immune deposits by endocytosis, followed by transport into multivesicular bodies, and finally discharge into the urinary space (6). This transepithelial transport of C5b-9 also accounts for the appearance of C5b-9 fragments in the urine of patients with MGN.
Formation of Radical Oxygen Species (ROS) and Proteinuria
Proteinuria in HN critically depends on the formation of ROS, because interventional application of oxygen radical scavengers drastically reduced proteinuria, while formation of immune deposits was not affected (76, 77) (Fig. 35, WEB). This raised the question whether ROS could be produced by intrinsic glomerular cells, especially because "professional" ROS-producing inflammatory cells were absent from glomeruli in pHN. In keeping with this hypothesis C5b-9 "activated" podocytes were found to express relatively large amounts of the membrane enzyme cytochrome b558(78), an essential component of the neutrophil respiratory burst NADPH oxido-reductase complex (79) (Fig. 36, WEB). These findings have raised the question whether ROS-mediated chemical modifications of GBM matrix proteins could be detected in proteinuric pHN rats.
Lipid Peroxidation Contributes to Proteinuria
The detection of lipoproteins within immune deposits, and the podocytes as potential source for intraglomerular ROS production has directed our attention to the role of lipid peroxidation (LPO), a common means of ROS-induced tissue damage (80, 81) causing generation of highly reactive compounds, such as malondialdehyde (MDA) and 4-hydroxynonenal that crosslink proteins by Schiff-base formation with lysyl residues. LPO and adduct formation in vascular walls were recognized as important components of atherosclerotic lesions, and monoclonal antibodies specific for MDA-lysyl Schiff bases were generated in these studies (80, 82). Using these antibodies as markers for LPO it was found that MDA-adducts accumulated in glomeruli of proteinuric HN rats while they were barely detectable in non-proteinuric controls (7) (Fig. 37, WEB). Collectively, the data obtained so far has provided evidence that LPO occurred within glomeruli with proteinuria in pHN, and that adduct formation was found in the GBM matrix.
LPO-Adduct Formation Occurs on GBM Type IV Collagen
Exposure of the GBM to LPO products raised the question
whether adduct formation affected all GBM matrix proteins randomly, or showed selectivity
for certain GBM components. We have provided evidence that type IV collagen was a major
target of LPO attack. Immunoprecipitation experiments using H3-cyanoborohydride radiolabeled Schiff bases detected polymers of
type IV collagen and indicated that crosslinking occured in GBMs of proteinuric rats (7)
(Fig. 38, WEB).
Since NC1 domains of type IV collagen are particularly rich in
cationic lysyl residues (83) it was of interest to determine whether they could serve as
targets for adduct formation with MDA. Indeed it was found that NC1 domains purified from
GBMs of proteinuric rats, contained MDA adducts in high density (7).
Is Proteinuria a Functional Consequence of LPO-Adduct Mediated Cross Linking of the GBM Matrix?
These findings raised the possibility that MDA and other
dialdehydes dimerize the fibrous type IV collagen "backbone" of the GBM via
cross linking of NC1 domains (Fig. 39, WEB). As a consequence, the porosity of the GBM
could be increased. This view is in agreement with previous in
vitro experiments in which isolated GBMs were
chemically cross linked by the bifunctional compound dimethylmalonimidate (84). Evidence
was obtained that relatively low amounts of cross links caused increase in permeability
for albumin and immunoglobulins, while massive cross linking resulted in clogging of the
GBM filter (85).
It is intriguing that lesions in the NC1 domains of GBM type
IV collagen were previously found also in other glomerular diseases which cause
proteinuria, i.e. Goodpasture's syndrome (86, 87) and the hereditary Alport disease (88).
Synoptically, these findings suggest a common final pathway towards proteinuria for
pathogenetically unrelated glomerular lesions (Fig. 40).
Fig. 40
The LPO Inhibitor Probucol Reduces Proteinuria
Is the formation of LPO-adducts on GBM matrix proteins of pathogenic relevance for proteinuria, or just an epiphenomenon? This question was answered in an interventional approach by treating rats three weeks before induction of HN with the potent inhibitor of lipid peroxidation, probucol (89, 90) (Fig. 41, WEB). This regime resulted in reduction of proteinuria by ~80% of that of untreated HN rats, and it was found to be independent of the cholesterol lowering effect of probucol (7). These results indicated that LPO-induced formation of adducts in the GBM are a major cause of damage of the glomerular filtration barrier, and could be causally related to proteinuria.
Amelioration of Proteinuria by Probucol Also in Human MGN ?
In a small clinical trial the efficacy of probucol in proteinuria in MGN was studied and the data suggest that indeed there is a reduction of proteinuria in a majority of cases after 3 months of treatment (manuscript submitted). The patient's proteinuria returned to starting levels 6 months after probucol treatment. These preliminary, though encouraging results call for a full scale clinical trial.
Clinical Aspects of MGN
Walter H. Hörl: Division of Nephrology and Dialysis, University Hospital, Vienna, Austria
Membranous nephropathy is the most common cause of
nephrotic syndrome in Caucasian adults. It is likely to be autoimmune in nature (94).
Spontaneously complete or partial remission can occur in 25-65% of the patients. Incidence
of end-stage renal disease is approximately 40% after 15 years. Therapeutic results,
however, are still unsatisfactory (Fig. 42, WEB). As mentioned before, etiologic agents
include immune diseases (e.g. systemic lupus erythematosus, rheumatoid arthritis),
infections (e.g. hepatitis B,C) or parasitic diseases, drugs and toxins (e.g. gold,
mercury) and miscellaneous (e.g. tumors, sarcoidosis) (95). Subepithelial immune-deposit
formation in membranous nephropathy is HLA linked. There is an increase in relative risk
for membranous nephropathy in patients with HLA-DR3, HLA-B8, HLA-B18 (96-98) and in
Japanese patients with HLA-DR2 (99). DR may relate to progression of the disease (98).
Clinical characteristics of membranous nephropathy are
summarized in Table I.
TABLE I - CLINICAL CHARACTERISTICS OF PATIENTS
SUFFERING FROM MEMBRANOUS NEPHROPATHY (21, 24, 26, 147-176)
- Manifestation in patients with a median age of 40 years (two peaks sug_gested between 30 and 40 as well as between 50 and 60)
- More common in men than women (2-3:1 ratio)
- Latent or prodromal period of weeks or months
- Medical attention because of proteinuria (~ 20%) or oedema (80%)
- Type of proteinuria: nonselective, usually in the 5- to 10-g/day range, quite variable day to day (e.g. due to protein intake, posture, exercise, hemodyna_mic variables)
- Microhematuria in up to 50% of the adults
- Hypertension in up to 30%
- Renal function well preserved in early membranous nephropathy
- Uremia very uncommon at presentation
Recently, Marx and Marx (22) studied 120 patients with
idiopathic membranous nephropathy for a median of five years (range 1-24 years) (Fig. 43,
WEB). Of these patients 19% developed end-stage renal disease (ESRD) or deterioration of
renal function. Proteinuria >3.5 g/day persisted in 34%, and 47% were in complete or
partial remission (Fig. 44, WEB). Renal survival was 91 ± 3% at five years and 75 ± 6%
at ten years. The predictors for the primary outcome (ESRD) were histological stage III-IV
and nephrotic syndrome, the predictors for the secondary outcome (ESRD or patient death)
were histological stage III-IV, nephrotic syndrome and comorbidity (22).
Paraneoplastic membranous glomerulonephritis is characterized
by a marked male preponderance, mean age over 50 years and a full-blown nephrotic syndrome
is always present. Of the patients with carcinoma and membranous glomerulonephritis, 40%
to 45% clinically manifest the nephrotic syndrome prior to the diagnosis of their tumor,
40% simultaneously and 15% to 20% following diagnosis of their tumor as recently discussed
in the nephrological forum of Kidney International by Pierre Ronco (100). There was no
evidence for malignancy in our patients with nephrotic syndrome and membranous
nephropathy.
Differential diagnosis of idiopathic membranous nephropathy
includes:
1) primary renal diseases
a) minimal change nephropathy (the most common cause of
nephrotic syndrome in children)
b) focal glomerulosclerosis (the most common cause of
nephrotic syndrome in USA)
c) membranoproliferative glomerulonephritis type I and II
2) secondary renal lesions include
a) diabetes mellitus and
b) amyloidosis
Non-disease-specific therapy is summarized in Table II.
TABLE II - THERAPEUTIC PROBABILITIES IN
PATIENTS WITH IDIOPATHIC MEMBRANOUS NEPHROPATHY (FOR REVIEW 2, 95)
Non specific therapy
- protein restriction (0.8 g/kg body weight)
- angiotensin-converting enzyme (ACE) inhibitors
- nonsteroidal anti-inflammatory agents
- HMG-CoA reductase inhibitors
Disease-specific therapy
- methylprednisolone (intravenous pulse)
- methylprednisolone/chlorambucil (177-179)
- methylprednisolone/cyclophosphamide (180, 181)
- cyclosporine A (182-184)
- intravenous immunoglobulin
- probucol
Since the course of idiopathic membranous nephropathy is often
benign, only symptomatic therapy is recommended (101).
Patient 1 responded rapidly to methylprednisolone and
cyclophosphamide therapy. Multiple therapeutic immunosuppressive interventions, however,
were less successful in patient 2. Therapeutic studies of idiopathic membranous
nephropathy have been summarized by Hogan et al (102).
Recently, Haas et al (103) investigated the effects of
short-term probucol treatment on proteinuria in patients with idiopathic membranous
nephropathy (Fig. 45, WEB). Fifteen patients (9 male, 6 female, mean age 48 ± 14 years)
with biopsy proven membranous nephropathy and proteinuria >3.5 g/dl were recruited from
five Austrian centers. Median time from kidney biopsy to study inclusion was 6 months
(range 1-53). Four patients had stage I, eight stage II and three stage III disease
according to Churg and Ehrenreich. Seven patients were pretreated with immunosuppressive
agents, 12 patients received ACE inhibitors, and seven patients had additional
antihypertensive therapy. Serum creatinine was <2 mg/dl in all patients (Fig. 46, WEB).
After a run in period of two months without treatment, probucol was administered orally (2
x 500 mg/day) for three months followed by a wash out phase without treatment for one
month. Then lovastatin (10 mg/day, and after one month 20 mg/day) was given for three
months (Fig. 47, WEB).
Probucol therapy caused a significant reduction in urinary
protein excretion (from 6.3 ± 2.2 to 5.5 ± 4.1 g/day). There was a decrease in protein
excretion in 10 patients (group A), but protein excretion did not change or increase in
five patients (group B). In group A protein excretion decreased by probucol therapy to 54
± 19% of baseline, and increased to 75 ± 28% during the wash out period. No effect of
lovastatin therapy was observed in this patient group (Fig. 48, WEB). Two patients of
group A restarted probucol therapy for a further six months. Proteinuria decreased to 19%
(1.7 g/day) and 34% (1.3 g/day). In group B patients protein excretion increased to 141 ±
44% during probucol therapy, and to 130 ± 67% during lovastatin treatment. According to
ACIMEN criteria, partial remission defined as a decrease of proteinuria <50% occurred
in four patients by probucol but in none by lovastatin therapy (103) (Fig. 49, WEB).
Complete remission (proteinuria <0.3 g/day) was obtained in
our second patient, probably as the result of the combination of cyclosporin A, high-dose
steroids plus chlorambucil and cyclophosphamide, and finally probucol plus vitamin E (Fig.
50, WEB).
Several years after complete remission of the nephrotic
syndrome, relapsing membranous nephropathy occured in our second patient. A recent study
of Suki et al (104) demonstrates relapses in one-third of the patients.
In 1982 Iwaoka et al (105) described a 54-year old women who
had proteinuria due to stage II membranous nephropathy. A fall in T3 and T4 and
concomitant elevation of TSH levels within a month was found (this was also the case with
our patient). Positive antithyroglobulin and antimicrosomal antibodies were compatible to
chronic thyroiditis. In a second kidney biopsy, glomerular deposits of thyro-globulin,
thyroid microsomal antigens and IgG were demonstrated. It was concluded that membranous
nephropathy in this patient was caused by immune complexes by thyroid constituents (105).
Several other cases of immune complex-type nephritis in which components of the thyroid
gland were considered to be antigens have been reported (106-112). Apart from the
classical membranous type, other glomerular lesions (crescentic glomerulonephritis,
rapidly progressive glomerulonephritis) were observed (113). Akikusa et al (114) described
the association of progressive systemic sclerosis with two additional immune-mediated
diseases: Hashimoto's thyroiditis and membranous nephropathy. Minimal change nephrotic
syndrome and acute renal failure were observed in a patient with aged onset
insulin-dependent diabetes mellitus and autoimmune thyroiditis (115) (Fig. 51, WEB).
Nishiki et al (116) reported a case of steroid-sensitive
nephrotic syndrome in association with sarcoidosis, immunoglobulin A nephropathy and
chronic thyroiditis. Hemophagocytic syndrome was found in a patient with Hashimoto's
thyroiditis and membranous nephritis (117). Weetman et al (118) investigated the
prevalence of proteinuria in patients with Graves' disease and chronic autoimmune
thyroiditis. Using the urine protein creatinine index, proteinuria was found in 29.8% of
patients with autoimmune thyroid disease. In the study of Cahen et al (3) 82 consecutive
Caucasian adults (52 males, 30 females, aged 17-86 years) with membranous
glomerulonephritis were prospectively evaluated for possible aetiological factors 1-4
weeks after renal biopsy. Presumed causes were identified in 17 patients (21%). Chronic
thyroiditis was the cause of membranous nephropathy in three cases (Fig. 52, WEB).
Rodriguez et al (119) described a patient with thyroiditis diagnosed after ten years of
the evolution of membranous nephropathy (in our case approximately 5 years after evolution
of membranous nephropathy). It was concluded that the association of autoimmune
thyroiditis and glomerular diseases is not very frequent but the most common association
is with membranous nephropathy. The frequency of this association, however, may be
underestimated. Autoimmune thyroiditis was diagnosed in our patient five years after the
evolution of the membranous nephropathy and the clinical manifestation of the autoimmune
disease was probably the reason for relapsing membranous nephropathy. Recently, a 3-year
prospective follow-up study of 42 patients with biopsy-confirmed glomerulonephritis was
presented (120). The prevalence of thyroid peroxidase antibodies remained unchanged,
although the majority of the patients had been treated with immunosuppressive agents.
Thyroid antibodies, however, were not restricted to membranous nephropathy, and were
notably found in 4 out of the 8 patients with vasculitis (Fig. 53, WEB).
Recurrent and de novo membranous nephropathy after renal transplantation
Recently, Hariharan et al (41) initiated a retrospective
study through the Renal Allograft Disease Registry to evaluate the prevalence and impact
of recurrent and de novo diseases after renal transplantation (Fig. 54, WEB). A
total of 167 representing 3.4% of the cases of recurrent and de novo disease were
diagnosed by renal biopsy in a total of 4913 renal transplants performed in 6 different US
transplant centers (Fig. 55, WEB). The authors found focal segmental glomerulonephritis in
57 patients, immunoglobulin A nephritis in 22 patients, membranoproliferative
glomerulonephritis in 18 cases and membranous nephropathy in 16 cases.
The diagnosis of membranous nephropathy occured in 9.6% of
recurrent and de novo disease types after a mean period of 664 days (88-23, 65)
after the transplant. The percentage of graft failure was 44% in patients with
post-transplant membranous nephropathy and kidney half-life for these patients with
individual recurrent and de novo disease was 1193 days (Fig. 56, WEB). The median
kidney survival rate for all patients with and without recurrent and de novo
disease was 1360 versus 3382 days (41). Eight of the 16 (50%) patients with membranous
nephropathy had documented recurrent disease, defined by the histological identification
of the same disease in the native as well as the transplanted kidney biopsies. The 1-, 2-,
3-, 4-, and 5-year kidney survival rates for post-transplant membranous nephropathy were
93%, 84.9%, 65.9%, 44%, and 44%, respectively (Fig. 57, WEB). The relative risk for graft
failure because of various recurrent or de novo disease, also estimated by Cox
Proportional Hazard model, was significant for all patients (Fig. 58, WEB). There was a
1.9 times increase in the risk of graft failure in patients with recurrent and de novo
disease as compared to patients without (Fig. 59, WEB). This difference was not
significant for membranous nephropathy (p<0.13), in contrast to patients with hemolytic
uremic syndrome-thrombotic thrombocytopenic purpura (relative risk at 5.36),
membranoproliferative glomerulonephritis (relative risk at 2.37) or focal segmental
glomerulonephritis (relative risk at 2.25) (41). The authors found a significant increase
in the relative risk for graft failure among patients undergoing cadaver transplant
(1.98), those with panel reactive antibody >20% (1.38), and for cold ischemia time
>20 hours (1.13). Recurrence of membranous nephropathy after renal transplantation has
been reported both anecdotally and in a few series ranging from 7 to 57% (for review
(42)). Cosyns et al (42) calculated the actuarial risk for recurrence by evaluating a
total of 30 renal transplant patients with membranous nephropathy. The authors found that
the rate of recurrence reached 29% at 3 years, plateauing up to 10 years (Fig. 60, WEB).
The percentage of recurrence of membranous nephropathy after renal transplantation may be
underestimated, since the disease may be present in the graft without overt proteinuria
(96) and graft biopsy may not be routinely performed under these conditions (Fig. 61,
WEB). The risk for graft loss was 38% in the eight patients with recurrence and 11% in the
22 patients without recurrence (difference not statistically significant) (Fig. 62, WEB).
Duration of membranous nephropathy in the native kidneys, duration of pretransplant
hemodialysis therapy, the presence of HLA-DR3, graft origin (living related donor or
cadaver donor) or immunosuppressive therapy were not different between renal transplant
patients with and without recurrence (42). Reviewing the currently available literature on
recurrence of membranous nephropathy, the calculated risk for graft loss was found
to be 72% at 5 years. In 10 of the reported 33 patients (37%), hemodialysis had to be
resumed 24 months after renal transplantation (range 1-57 months) (42).
The rate of recurrent membranous glomerulonephritis is
summarized in Table III.
TABLE III -
| First author (reference) | Rate of recurrence |
| Morzycka et al (185) | 57% |
| Berger et al (123) | 4% |
| Montagnino et al (134) | 33% |
| O'Meara et al (128) | 10% |
| Schwarz et al (133) | 10% |
| Couchoud et al (122) | 26% |
| Marcen et al (186) | 50% |
| Cosyns et al (42) | 29% |
| Hariharan et al (41) | 50% |
Rates for recurrent membranous nephropathy in renal
transplants differ between 4% (123) and 57% (185). More recent studies between 1995 and
1999 found a recurrence rate between 26% and 50% (41,42,122,186).
It has been suggested that a short duration of membranous
nephropathy in the native kidneys is a risk factor for post-transplant recurrence (121).
In the study of Cosyns et al (42), however, duration of the original disease was 9 years
(range 3-18) in patients with recurrence and 6 years (range<1-22) in those without.
Duration of the disease in the native kidneys was 14 years in our patient presented here.
Prolonged time on hemodialysis before undergoing renal
transplantation does not protect against posttransplant recurrence of membranous
nephropathy. In the study of Cosyns et al (42) pre-transplant hemodialysis was performed
for 33 ± 26 months in the patients with recurrence as compared to 29 ± 20 months in
those without recurrence. Since our patient received the kidney from a living related
donor (brother), he was never dialyzed before renal transplantation.
A strong association between idiopathic membranous nephropathy
and HLA-DR3 has been reported (98). In the study of Cosyns et al (42) HLA-DR3 was present
in 5/8 patients (62%) with recurrence as compared to 5/21 patients (21%) without
recurrence (p=0.08). In an other study the frequency of HLA-DR3 was twice as high in the
patients who had recurrence as compared to those who did not (40% versus 21%) (122).
HLA-DR3 and HLA-B8 (which was initially mentioned as a risk factor for membranous
nephropathy) were found in our patient. Membranous nephropathy may more likely recur in
living related donor grafts than in cadaver grafts, probably as a result of a familial
predisposition to the disease (121). Immunological status of the recipient (HLA-DR3
allele) and the characteristics of the transplantation itself (living related donor) as a
risk factor was also suggested by the data from Berger et al (123), Agarwal et al (124),
Lieberthal et al (125), Iskandar and Jennette (126), Obermiller et al (121) and First et
al (127). In O'Meara et al's series (128) recurrence of the disease was significantly
greater in the living related donor group (11.9%) than in the cadaveric group (1.3%). In
the study of Cosyns et al (42), however, recurrence was equally frequent in cadaver (7/26)
and living related donor grafts (1/4). This is a clinically important observation since
Agarwal et al (124) argued that recurrent membranous nephropathy is common in males and in
living related renal transplants. The authors stated that patients with living related
renal transplants develop recurrent membranous nephropathy earlier than patients with
cadaver renal transplants. Of 23 cases, 11 allografts were from living related renal
transplants. It was concluded that delaying transplantation should be considered in such
patients to avoid early recurrence of disease, and that a cadaver source should be
preferred to a living donor (124).
Discussing allograft membranous nephropathy Davison and
Johnston (129) concluded that recurrent membranous nephropathy is uncommon, the antigens
involved are unknown, but are presumably the same as those involved in the native
nephropathy, the clinical and pathological features are indistinguishable from the
idiopathic form.
It has been shown that first recurrence is a risk factor. Two
sequencial grafts developed recurrences in a patient within a year of the transplants
(130). Recurrent membranous glomerulonephritis has a worse prognosis than de novo
membranous glomerulonephritis. Recurrent membranous glomerulonephritis can appear within
four months after transplantation in 28% of the patients, presenting with severe
proteinuria (121, 123, 131, 132). Of those that recurred within four months, 75% were in
kidneys with HLA-matched living related donors (121, 123). The case report and literature
review by Josephson et al (130), however, illustrate that membranous glomerulopathy can
recur in both poorly and well-matched allografts, and that the course of recurrent
membranous glomerulopathy is unpredictable.
Based on early reports of recurrent membranous nephropathy in
allografts it was suggested that the majority of recurrences occur in patients on
azathioprine.
A beneficial effect of cyclosporin A was shown by Schwarz et
al (133) who found recurrence only in the group of renal transplant patients not treated
by cyclosporin A. Marcen et al, however, observed in agreement with earlier findings of
Montagnino et al (134) recurrence of membranous nephropathy only in patients on
cyclosporin A. Cosyns et al (42) concluded from their study that cyclosporin therapy does
not prevent recurrence of nephropathy after renal transplantation. Proteinuria was not
affected by the replacement of azathioprine by cyclophosphamide (42, 127).
Methylprednisolone pulses are unsuccessful (123). ACE inhibitors may decrease proteinuria
(130). The development and progression of recurrent and de novo disease do not seem
to be influenced by the use of newer immunosuppressive agents such as mycophenolate
mofetil and tacrolimus (41).
Preliminary results of the efficacy of high-dose intravenous
human gamma-globulin therapy in patients with idiopathic membranous nephropathy (135, 136)
encouraged us to treat our patient (under immunosuppression with low-dose prednisolone,
cyclosporin A and mycophenolate mofetil) with intravenous immunoglobulin. Since several
reports of acute renal failure after the administration of intravenous immunoglobulin have
been published and sucrose has been suspected to be the cause of acute renal failure
(137), sucrose-free immunoglobulin was given to our patient in a dose of 0.4 g/kg body
weight twice/week during the first month, once/week during the second month, and
twice/month during the third month (Fig. 63, WEB). This therapy, however, was without any
effect with respect to proteinuria or serum creatinine (proteinuria between 10 and 15
g/day, serum creatinine between 2.1 and 2.5 mg/dl).
Recently, a case of membranous nephropathy in a donor
allograft kidney was presented showing gradual disappearance of electrondense deposits and
histological alteration of glomerular lesions within 16 months after renal transplantation
(138). Parker et al (139) also reported a case of transplantation of a donor kidney with
preexisting membranous nephropathy and good renal function three years after renal
transplantion. Membranous lesions of Heymann nephritis resolved by transplantation of the
kidneys to syngenic recipient rats (140). It was suggested that transplantation of a
kidney with membranous nephropathy results in clearance of the endogenous immune complexes
and histological repair of the glomerular basement membrane.
Risk factors discussed for the manifestation of membranous
nephropathy, recurrence and de novo disease (positive and negative results) can be
summarized as follows (Tab. IV).
TABLE IV - RISK FACTORS DISCUSSED FOR MANIFESTATION OF
MEMBRANOUS NEPHROPATHY, RECURRENCE AND DE NOVO DISEASE IN RENAL ALLOGRAFTS
(POSITIVE AND NEGATIVE RESULTS SUMMARIZED)
HLA-DR3, HLA-B8, HLA-B18, HLA-DR2 (Japan), HLA-DR4
Male
Duration of original disease (not confirmed)
Short-term dialysis therapy (not confirmed)
Living related donor
First recurrence
Type of immunosuppression (not confirmed)
Most cases of membranous glomerulonephritis (83% - 90%) in
allografts are due to de novo disease (133, 141). The prevalence is about 1.3% to
2.1% overall among transplant patients (133, 142). The frequency may be increasing
slightly, since 1.3% were observed in a study published in 1982 (142) and 2.1% in a study
published in 1994 (133). One possible explanation is that membranous nephropathy is
predominantly a condition of middle aged and older patients. The mean age of patients
being transplanted is increasing (129). Children may have a higher risk for de novo
membranous glomerulonephritis (9%) (143, 144).
De novo membranous glomerulonephritis in allografts is
typically a late complication. The risk of the de novo disease increases with time
after renal transplantation reaching 5.3% at 8 years (143). Proteinuria in the nephrotic
range is observed in 20% to 70% of these patients (143, 145). Approximately 20% to 30% of
the patients are not proteinuric. The presence of de novo membranous
glomerulonephritis in a first graft increases the risk in a second graft (57%, four out of
seven cases (146)). One series found an association with HLA-DR4 in recipients (43), no
other risk factors have been identified. The pathogenesis has so far not been established,
no therapy is effective.
Conclusion
Membranous nephropathy is the most frequent cause of nephrotic syndrome in the adult. Predictors of primary and secondary outcome are the histological stage at initial biopsy, the nephrotic syndrome and co-morbidity. Disease specific therapies are currently not available because the pathogenesis of MGN is not fully understood and apparently only partially reflected in currently available experimental models. The appearance of MGN in transplanted kidneys (both de novo and as a recurrence) is relatively frequent.
Reprint requests to: Walter H. Hörl,
M.D., Ph.D.
Universität KliniK für Innere Medizin III Klinische Abteilung fur Nephrologie und
Dialyse Wahringer gurtel 18-20 1090 Vienna, Austria
QUESTIONS AND ANSWERS
DR. C. PONTICELLI (Division of Nephrology and Dialysis, Policlinic Hospital, Milan - Italy): To start the discussion I have a comment and a question.
The comment is that there still is much confusion about the natural course of idiopathic membranous nephropathy. This is due to the fact that most of the available papers reporting a so called natural outcome actually put together patients with a nephrotic syndrome and with non nephrotic proteinuria, often they put together treated and non treated patients, but, most important, the mean follow-up of these patients is quite short. Even the last paper you mentioned in Kidney International had a medium follow-up of 5 years, and some patients had just one year of follow-up. This may be incorrect for a disease which usually progresses in long term, and when we look just at patients presenting with a nephrotic syndrome followed for at least 10 years, most papers report that the kidney survival is around 50%. I think this is important to point out because, of course, we have to decide whether or not to treat the patient on the basis on his natural prognosis. Looking at the prognosis I have a question for both of you. You mentioned the importance of stage of glomerular damage, but many analyses pointed out that the most important histological factor is the presence of interstitial fibrosis in membranous nephropathy, as in many other glomerular diseases. Do you agree with this?
DR. D. KERJASCHKI: As a snap shot of the disease, yes. If you see massive interstitial fibrosis, this is a kidney which is on its way out. When it comes to a pathogenetic relation it is very difficult to say Òwhich is the hen and which is the eggÓ. I think this is true in many cases of interstitial fibrosis. I donÕt believe that interstitial fibrosis in membranous nephropathy is an independent disease.
DR. G. D'AMICO (Division of Nephrology and Dialysis, San Carlo Hospital, Milan - Italy): We just made a survey of the literature on the prognostic factors in membranous nephropathy. It was evident that the most powerful risk factor appears to be tubulo-interstitial damage. In many of the statistical analysis I was really impressed by the fact that it appears to be even more powerful as an indicator of prognosis than the stage of the glomerular histological lesion. But I have a question for Dontscho Kerjaschki: I must confess that I am still not convinced that in the human disease we have to do with antibodies to a component of the podocytes of the glomerulus. In my opinion, the fact that there are many immune complex diseases that can give a pure or almost pure membranous nephropathy, including lupus nephritis and some infectious diseases, suggests deposition of exogenous immune complexes more than of antibodies to megalin, or some other component of the glomerulus. In lupus nephritis, in humans, we have a situation which is similar to that of experimental models of injection of immune complexes. According to the charge and to the size of the immune complex you can have the deposition in the mesangium, on the subendothelial side, or on the subepithelial side of the glomerular basement membrane. In lupus nephritis we can have a prevalent mesangial, subendothelial, or subepithelial deposition: in this human disease, even if the deposition is exclusively or prevalently subepithelial, we still think to deposition of immune complexes.
DR. D. KERJASCHKI: Obviously there are many different ways of getting an immune deposit into a subepithelial position. One way is the in situ hypothesis, as you know. It is very tempting to believe that what you see in man is similar to what happens in rat membranous nephropathy in which the in situ hypothesis is established beyond any doubt. The concept of circulating immune complexes was at the very beginning of the entire development of renal immunopathology. However, this view was discounted for several reasons that may be justified or not. One of the major reasons was that nobody really was successful in producing immune complexes experimentally which would not be trapped subendothelially or mesangially, but would find their way directly into the subepithelial space and accumulate there. This does not mean, however, that this does not occur. There is also the possibility of the implantation of antigens. If you have antigens that have certain physical properties, such as cationic charge or some sort of affinity to a matrix component, they may as well cross the glomerular basement membrane and serve as splendid antigens. However, several authors agree that megalin is not relevant in human membranous nephropathy. Just as a point of illustration, there was a very recent article coming out of McCluskeyÕs group. They found that patients with Hashimoto autoimmune thyroiditis have high antibody titers against megalin in their circulation; however, they did not find a relation to immune deposits. We looked at two cases of membranous nephropathy that were associated with autoimmune thyroiditis. On paraffin sections (again there is a limitation with the technology and the human material available) by immunocytochemical staining we could not detect megalin in glomeruli. In conclusion, I would not be surprised if human membranous nephropathy is actually a syndrome involving several antigens. At some point, the different diseases may converge into a common pathway that we recognize as membranous nephropathy. Thus, I think it is of value to study this putative common mechanism that includes features such as proteinuria, which in a significant fraction of the patients could be similar to what we see in rat membranous nephropathy. The antigen business has to be dealt with separately.
DR. D. ROCCATELLO (Immunopathology Center, L. Einaudi Hospital, Turin - Italy): I would like to come back to the interstitial lesions. I would like to know exactly what to your mind is the weight of interstitial changes, I mean also interstitial infiltration, both as independent prognostic factor and as predictors of response to therapeutic approaches.
DR. D. KERJASCHKI: I would pass this question on directly to Dr. MŸller who is really an expert in this field.
DR. G.A. MÜLLER (Georg-August-Universität, Zentrum Innere Medizin, Göttingen - Germany): I want to make a comment on this. Already in 1990 we published a paper in which we analysed retrospectively 300 patients with membranous nephropathy. It was very clear that patients who didnÕt show any sign of interstitial fibrosis had a kidney survival of 80% after 10 to 15 years. However, when there were signs of interstitial fibrosis the kidneys lost their function after 5 years. So it was clear for us that when we saw a biopsy from a patient with interstitial fibrosis we did not treat him with any immunosuppressive drug anymore, we only treated the hypertension. This was clear also from the study. Regardless of the therapy regimen we used, when there were signs of interstitial fibrosis the kidney was lost.
DR. C. PONTICELLI (Division of Nephrology and Dialysis, Policlinic Hospital, Milan - Italy): It depends on the degree of interstitial fibrosis, because some patients who were treated with our regimen with chlorambucil could maintain a stable renal function, and also in a small number of cases complete remission of proteinuria occured. Clearly, if interstitial fibrosis involves a major part of the kidney, therapy may be futile.
DR. G.A. MÜLLER (Georg-August-Universität, Zentrum Innere Medizin, Göttingen - Germany): I just want to say that these were not end-stage kidneys that were biopsied. All patients had creatinine levels about 2 mg/dl at the time of biopsy, not higher. Most of them were about 1.5 Ð 1.6 mg/dl.
DR. F.P. SCHENA (Division of Nephrology and Dialysis,
Policlinic Hospital, Bari - Italy): I have two questions concerning probucol therapy
for Walter Hšrl: You showed two groups in your slides: responsive and unresponsive
patients. It is evident that when you stop probucol, proteinuria increases and it is
evident that proteinuria is dose dependent on probucol. At the same time it is evident
that probucol may induce the reduction of ROS. However, the immune complexes continue to
grow. For this reason I believe that probucol should be combined with immunosuppressive
therapy. My first question is: What is your comment about responsive and unresponsive
patients to probucol? What is the cause of unresponsiveness to probucol?
The second question is: In the 3rd patient you described, at the end you administered intravenous
immunoglobulins. Before using these intravenous immunoglobulins, did you try
cyclophosphamide in association with steroids or cyclosporin?
DR. W.H. Hörl: I'll start with the second question,
it is easier to answer. Knowing the literature we did not add cyclophosphamide, we did
also not replace cyclosporin A by cyclophosphamide. I discussed this patient with several
well known experts on membranous nephropathy in Europe, including Claudio Ponticelli. I
tried to find out the best I really could do for him, knowing that he had this association
of HLA DR3 and HLA B8, was male, had a kidney from his brother, was never on dialysis and
had all predictive factors for further recurrence of the disease in the next transplant.
Its serum creatinine was 2.5 mg/dl and his proteinuria 15-16 g/day, so we decided not to
combine cyclophosphamide, particularly in association with mycophenolate mofetil which the
patient has had as additional immunosuppressive therapy. Although the literature shows
that there is no convincing effect of high dose steroid or pulse steroid therapy, we gave
the patient 3 times 250 mg methylprednisolone (without any effect). Even though there are
no convincing data in the literature on high dose immunoglobulin therapy, we tried this
treatment protocol but, disappointingly, we did not see any benefit.
With respect to your first question I believe you are right
that we should add probucol as additional therapy. I have no explanation on the response.
Probably Dontscho can speculate better than I on the reason why some patients are
responders and others are non responders to probucol therapy. What we did, based on animal
experience and data he presented, we gave with encouraging results probucol to some of our
patients with nephrotic syndrome and membranous nephropathy. We decided to start a pilot
study, based on the protocol, shown to you. Of course, the data were disappointing also
for us but nevertheless we should perform a multicenter trial. The problems I see are the
potential cardiac side effects of probucol. You know there are some concerns about cardiac
arrhythmias in patients treated with probucol. So we probably need a better antioxidant
drug.
DR. M.R. DAHA (University Medical Center, Department of Nephrology, Leiden - The Netherlands): Apparently there is a difference between the experimental model and the patients with membranous nephropathy for the response to probucol. Maybe in the experimental model you do not really have a very long established disease with large destruction of most epithelial cells, so you can still recover renal function some way or another, while in the patients you are past the point of no return. My question is more on the clinical side. Maybe you are dealing with patients that have a more severe type of disease, that is not reversible or not treatable; therefore you must compare their response to that of rats which have developed end-stage Heymann nephritis.
DR. D. KERJASCHKI: You are addressing the question that we really have not found a very good adequate and faithful human model for rat Heymann nephritis, as it looks. I think, given all these points you have mentioned, the effect probucol apparently has on a subset of patients and on their proteinuria is reasonably encouraging because it shows you that there may be something to be cured and be influenced and repaired in these basement membranes, even when the lesion is pretty much advanced. So, I am not so pessimistic. In view of the alternatives, I think probucol may be an interesting supplement to what is available for these patients at this point.
DR. W.H. Hörl: Coming back to what Schena said, the second patient, responding to these mixed therapies, did also take vitamin E, so she was not included in our protocol. It could be due to the combined effect of all these compounds that she responded so beautifully for several years. And also coming back to what has been just said before, of course we have several months silent period or a period with minimal prodromic symptoms, not leading the patient to a nephrologist, and this may be the big difference between animal and patient studies.
DR. D. KERJASCHKI: May I make a quick comment on the efficacy of vitamin E and probucol? My friends in arteriosclerosis research tell me that these compounds are completely incomparable in their efficiency. Probucol is extremely efficient for scavenging lipid peroxidation because it is a very efficient oxygen radical scavenger which, on top, is an extremely hydrophobic substance. So, every molecule of probucol you administer to a patient will go to a cell membrane or a lipoprotein core, and will not persist anywhere else. This is why it has a specific protection just of lipids and membranes from peroxidation. Vitamin E is much less efficient. So, there is no real replacement at this point for probucol.
DR. E. ALEXOPOULOS (Hippokration General Hospital, Dept. Internal Medicine, Tessaloniki - Greece):Let me confess, before all, that I donÕt use cytotoxic drugs for the treatment of membranous nephropathy. I am still using corticosteroids. My comment is related to the previous question on the role of the interstitial space. So, when we analysed our results with the use of steroids on membranous nephropathy we have seen that the patients who responded completely or partially to corticosteroids had the largest number of interstitial infiltration by monocytes, whereas those who did not respond at all had very low numbers of interstitial infiltrates. The degree of interstitial fibrosis did not differ actually between the responders and non responders in our experience. When we expanded this study and we looked at the MCP1 in the urine we found a very good correlation between the number of monocytes in the interstitial space and the MCP1 in the urine. Now we try to treat patients again with steroids in a prospective manner by using a cut off point of MCP1 in the urine and decide to treat or not to treat with steroids.
DR. W.H. Hörl: Thank you, I certainly believe that there is a subgroup of patients with membranous nephropathy and nephrotic syndrome which would respond beautifully to high dose steroids alone. Looking at the whole group of patients the results as presented in several papers by Claudio Ponticelli are better if the combination of steroids with cytotoxic drugs was used. This is also my experience. I personally prefer cyclophosphamide instead of chlorambucil. I have had two earlier chiefs: Prof. Heidland, in WŸrzburg, did never use chlorambucil and my ex-chief, Prof. Schollmeier in Freiburg, loved it. So I had the experience with both kinds of treatment and can confirm that chlorambucil is very effective in some patients. We have had discussions with our hematologist, Prof. Lechner, and he argued that chlorambucil could be a leukaemia-inducing substance. I have, however, no experience or information whether or not this really occurs in patients treated with chlorambucil. Claudio, you could probably comment on that.
DR. C. PONTICELLI (Division of Nephrology and Dialysis, Policlinic Hospital, Milan - Italy): Yes, we looked at our patients treated with chlorambucil. We looked particularly at the risk of cancer in patients treated with chlorambucil and we found, as reported in a J Am Soc Nephrol article a few years ago, that the risk of cancer is the same as that we have encountered in the general population of the same age. Let me remind you that in the last trial performed in Italy comparing chlorambucil with cyclophosphamide we could not find any difference in the efficacy. Yes, there is a trend to lower numbers of side effects with cyclophosphamide and you can now use with satisfaction cyclophosphamide instead of chlorambucil if you believe in this type of treatment.
DR. G.A. MÜLLER (Georg-August-Universität, Zentrum Innere Medizin, Göttingen - Germany): Dr. Hšrl, some of your data on probucol were very encouraging. Could you tell us something about the side effects and if you performed any re-biopsy in some of these patients? What was the result?
DR. W.H. Hörl: No, side effects were not observed. Our second patient received 2 g/day and she took this drug for more than 6 months. She wanted to continue on this treatment, since she believed that the drug cured her nephrotic syndrome. I reduced the dose to 500 mg twice a day. She took this drug for more than 1 year without any side effects. Also all other 15 patients enrolled in the study did not complain about any side effects. This drug was used years before to treat hyperlipidemia. However, I think the cardiac side effects were so severe that this drug was removed from the market in many countries. Rebiopsy is not indicated in patients with improvement of the disease.
DR. L. NAKOPOULOU (Dept. of Pathology, University of Athens Medical School, Athens - Greece): I would like to go back to fibrosis. I think the fibrosis is very nicely related to the grade of membranous glomerulonephritis. For example, in grades 1 and 2 usually we did not find fibrosis; fibrosis, however, occurred in grades 2 or 3. And, of course this is a main point for treatment.
DR. D. SANTORO (Division of Nephrology and Dialysis, Policlinic Hospital, Messina - Italy): It has been reported that focal and segmental glomerulosclerosis is associated with primary membranous glomerulonephritis in 25% of cases. In your experience, is the prognosis different for these patients and are these patients less responsive to treatment or do you use a different treatment?
DR. D. KERJASCHKI: We have seen 2 or 3 ÒmixedÓ cases out a few hundred cases of membranous nephropathy. I donÕt think that this is a very frequent event, at least in our hands. My understanding is that the severity of the disease is dictated by the segmental focal sclerosis and not so much by the membranous component of the disease. I believe this is what clinicians have treated.
DR. P. RONCO (Unit de Recherche de Nephrologie, Hôpital Tenon, Paris - France): Although IgG4 is a minor subclass in the blood, most patients with membranous nephritis seem to have IgG4 deposits. On the other hand IgG4 does not activate complement or activates complement to a very low extent, whereas complement activation is a key player in the pathogenesis of Heymann nephritis. So, I have three questions: First, what is the role of IgG4 in human membranous glomerulonephritis? Second, where do they come from, how are they deposited? If I remember, they may be cationic. And third, has anybody tried to isolate IgG from kidneys from patients at autopsy or from biopsy sections and to look at the specificity of IgG and most specifically of IgG4?
DR. D. KERJASCHKI: It is my understanding that IgG4 appears when an immune process is cooking for a very long time. Secondly, IgG 4 comes in all sorts of variations, there are also cationic variations. Thirdly, several people have tried to elute IgG from sections or from whole isolated glomeruli, among them ourselves too. It is very difficult to recover enough biologically active antibody to do anything, because you need acid dilution and so on. There are several more contemporary approaches to find out what the antigens are, and we are kind of optimistic to see what comes up there.
DR. L. GESUALDO (Division of Nephrology and Dialysis, Policlinic Hospital, Bari - Italy): It has been demonstrated that in anti-GBM disease the cellular immunity plays an important role. It was demonstrated that if you knock out the cellular immunity you donÕt get disease. So I have a speculative question: What is the role of cellular-mediated immunity in membranous GN? Actually I have been interested to see if there is an epitope on megalin that is best recognized by the antibody. Do you think that there are some epitopes that can be recognized by the cellular immune system?
DR. D. KERJASCHKI: Let me start answering the second question: We donÕt only think, but we know there is such a case. In collaboration with a group in Australia we are examining the situation right now and apparently, we have located a T-cell epitope in a small region of 130 kd in megalin. We are looking now for the B-cell epitopes. They will be available, if everything works well, in the next couple of months. This could be the prelude for a feasibility study of inducing tolerance and to see if you can vaccinate. We have tried to produce CD8 positive cells by intranasal immunization with synthetic peptides and recombinant products, because it has been reported that this brings up tolerance; however we got the most beautiful high titered antibody you could imagine, and no tolerance at all. To the first part of your question: The fact is that you barely find any lymphocytes or inflammatory cells in glomeruli in membranous nephropathy, unless you have something superimposed. The same is true also for Heymann nephritis.
DR. A. PANI (Division of Nephrology and Dialysis, San Michele ÒG. BrotzuÓ Hospital, Cagliari - Italy): As you have pointed out MGN is a HLA associated disease. In a Chinese population, D8, D18, and the Sardinian population DQ2 are strongly associated to it. But in my opinion HLA is not associated to progression of renal damage. What do you think about it? What is the role of HLA molecules?
DR. W.H. Hörl: As I mentioned, this issue is controversially discussed in the literature. Bill CouserÕs opinion is that HLA, DR3, B8 and B18 are favourable for membranous nephropathy. The literature on the recurrence of membranous nephropathy in renal transplant is controversial. A tendency for a higher level of recurrence is observed in patients with HLA-DR3. Due to a small number of patients in many studies, statistical analysis is not significant. If you put all the published literature together, you may have a significant effect for HLA-DR3.
DR. G.A. MÜLLER (Georg-August-Universität, Zentrum Innere Medizin, Göttingen - Germany): It is not surprising that it is also HLA B8, for there is a linkage disequilibrium and a normal haplotype is A1, B8 DR3. So this is a combined haplotype which is associated with membranous nephropathy, but there is even a higher association to a properdin factor BF F1 which is also located on the short arm of the chromosome six. As far as I remember the association is near about 80%.
DR. J. EGIDO (Fundacion Jimenez Diaz, Renal Unit, Madrid - Spain): As you know oxygen radicals have been implicated in many glomerular diseases a long time ago. Even mesangial cells stimulated with immune complexes produce oxygen radicals. Both in puromycin and adriamycin models oxygen radicals were implicated, they were treated with a good result with scavengers. My question is, do you believe that oxygen radicals are indeed a common phenomenon of inducing proteinuria, or do you believe that this is more particularly related to membranous diseases? Second question: Is there any evidence that potentially harmful amounts of oxygen radicals produced in the glomeruli could damage NC1 domains of collagen IV?
DR. D. KERJASCHKI: These are two complex questions,
let me start with the second one.
There is evidence that in vitro produced oxygen
radicals by Fenton-reaction are able to selectively interfere with the structure of NC1
domains that were built from recombinant pieces. Radicals unfold the NC1-hexamer and
expose the antigenic epitopes seen by GoodpastureÕs antibody. If you increase the
concentrations of the radicals, this whole structure will fall apart and the antigen will
degrade and disappear. So, I think this is just one point of illustration on what
radicals, without any lipid peroxidation, can do to matrix proteins. The lipid
peroxidation helps to cross link the matrix proteins, including the NC1 domains, but
radicals by themselves also can form copolymers and cross-links. They can cleave proteins.
We have published a study in the Am J Pathol where we performed a very simple and naive
experiment with EHS sarcoma matrix that is reasonably well defined. We incubated it with
different concentrations of oxygen radicals and just looked at what happens. You can
dissolve 30-40 % of this matrix, just by radicals in a strict pecking order: Entactin
falls into pieces first, then comes apparently intact laminin out of the matrix. However,
this laminin is damaged, its tryptophans are oxidized and it will not re-associate
anymore. Then the solubilized laminin is degraded. Then all of a sudden, type IV collagen
crawls out of the matrix. This is intact type IV collagen by rotary shadowing and by SDS
PAGE. However, it lacks the 7S domains which hook the molecules together. So, the radicals
by themselves can do a lot of damage to a matrix.
Secondly, you have lipid peroxidation in the lesions I have
mentioned, which supports chemical cross linking. Thirdly, and I did not talk about this
at all, you have matrix proteases around, and lipid peroxidation and oxygen radical damage
make proteins superb substrates. Some of the metalloproteases are activated by oxygen
radicals. So, you are creating a real soup in the GBM. The amazing thing in Heymann
nephritis is that, by just cutting down lipid peroxidation, you abolish proteinuria, and
you somehow stop almost the entire process of damage. How these events are interconnected
we do not know, but at least the animal model tells us that stopping this apparently
crucial process will hold the entire chemical avalanche at least for a while.
Your first question was whether lipid peroxidation is confined
only to membranous nephropathy in rat and, maybe, man. The answer is ÒnoÓ. There is a
study which was published with Jim Neale when he was in our Lab. He showed cross linking
of type IV collagen and sensitivity to probucol also in puromycine nephrosis which is a
non immunological disease. Moreover, we have studied a mouse knock out model that is
called the mpv17 mouse. This is an interesting animal system in which one of the
preferential integration sites (designated mpv17 locus) for retroviruses was inactivated.
This mpv17 mouse is healthy at birth, but it develops heavy proteinuria and will die in
renal fibrosis. The disease looks like minimal change or focal sclerosis, with flattening
of foot processes, heavy albuminuria and terminates in interstitial fibrosis. We found
that isolated glomeruli of mpv17 mice produce 3 times more oxygen radicals when compared
to normals. Lipid peroxidation products accumulated in glomeruli by a factor of 5 times
more than in normal. More interesting, we could prevent this disease by feeding the
animals probucol, as long they were eating probucol pellets. When we treated sick animals
with a mini-pump which releases ROS-scavengers for exactly two weeks, the animals remained
healthy, just for these two weeks. The moment the pump was empty, proteinuria started and
the foot processes became flat. So, this is another example which, to our knowledge, is
the "cleanest" system in which oxygen radicals produce proteinuria, lipid
peroxidation and flattening of foot processes.
DR. D. ROCCATELLO (Immunopathology Center, L. Einaudi Hospital, Turin - Italy): Since the presence of anti-thyroid antibodies does not mean necessarily that thyroiditis and membranous nephropathy are genetically related, should we systematically search for thyroid antigens at the histologic level?
DR. D. KERJASCHKI: Searching for antigens at the histologic level is a very complicated story. It depends on the type of tissue you have, on the preservation, on the methods you use, and on your tools, such as antibodies and what not. A biochemical approach would be much more straightforward, if there is a possibility to isolate individual glomeruli and have a technique which is sensitive enough in producing immunoprecipitates or immunoblots. Searching for antigens by immunocytochemistry is just one aspect, and you definitely need an alternative confirmation of whatever you find. But, you have to be aware of the fact that even when you identify an antigen within the deposits, there is no real proof that this is the pathogenic antigen. We do not have an assay in humans. Apparently in rat it is easy, you purify a protein or a fragment, such as megalin, and you go back to a rat and immunize it and you see if it develops disease. In man there is no such thing. So, this is one of the problems why people resorted to animal models and then sometimes they have a hard way finding it back to humans.
References (when available, each reference has been linked to PubMed)