INTRODUCTION — Membranous nephropathy (MN) is among the most common causes of the nephrotic syndrome in nondiabetic adults, accounting for up to one-third of biopsied cases of nephrotic syndrome. (See "Overview of heavy proteinuria and the nephrotic syndrome", section on 'Etiology'.)
The term MN reflects the primary histologic change noted on light microscopy: glomerular basement membrane (GBM) thickening with little or no cellular proliferation or infiltration [1]. MN is most often primary (previously called idiopathic), although it has been associated with hepatitis B antigenemia, autoimmune diseases, thyroiditis, malignancies, and the use of certain drugs such as nonsteroidal antiinflammatory drugs (NSAIDs), gold, penicillamine, and captopril.
The epidemiology, pathogenesis, and etiology of MN will be reviewed here. The clinical manifestations, pathology, diagnosis, treatment, and prognosis of MN are presented separately:
●(See "Membranous nephropathy: Clinical manifestations and diagnosis".)
●(See "Membranous nephropathy: Treatment and prognosis".)
EPIDEMIOLOGY — MN accounts for approximately 20 to 30 percent of cases of nephrotic syndrome in White adults [2,3], and a rising incidence has been reported in China, perhaps related to environmental pollution [4]. (See "Overview of heavy proteinuria and the nephrotic syndrome", section on 'Etiology'.)
MN is seen in all ethnic and racial groups and in all sexes, but primary MN is more common in White males over the age of 40 years. MN in young females should raise the suspicion of systemic lupus erythematosus (SLE). MN is less commonly seen in children, in whom it is often associated with hepatitis B or, less commonly, autoimmune or thyroid disease [5].
PATHOGENESIS
Animal models of MN — Experimental models of MN suggest that the glomerular basement membrane (GBM) immune deposits develop in situ with the movement across the GBM of circulating immunoglobulin G (IgG) antibodies directed against endogenous antigens expressed on or near the podocyte foot processes or against circulating cationic or low-molecular-weight antigens that have crossed the anionic charge barrier in the GBM.
The pathogenic mechanisms leading to MN have been primarily elucidated from the rat model of Heymann nephritis, which closely resembles the human disease at both the clinical and histologic level [6,7]. In Heymann nephritis, circulating antibodies target the endocytic receptor megalin (gp330) on podocyte foot processes. The resultant subepithelial immune deposits activate complement, which leads to the assembly of C5b-9, the membrane attack complex, and its insertion into the podocyte plasma membrane [8,9]. Subsequent complement-mediated podocyte injury leads to two changes:
●Proteinuria via activation of signaling pathways in the podocyte that results in redistribution of actin and loss of slit diaphragm integrity [8-10]
●GBM expansion by the overproduction of type IV collagen and laminin by the injured podocytes [11-13]
Megalin is not expressed in the human glomerulus although other antigens have been implicated in human MN, such as the phospholipase A2 receptor (PLA2R), thrombospondin type-1 domain-containing 7A (THSD7A), neural epidermal growth factor-like 1 protein (NELL1), and neutral endopeptidase (NEP). (See 'Antigens in primary MN' below.)
It has not been possible to directly establish the pathogenicity of human anti-PLA2R antibodies in animal models, since PLA2R is not expressed by the podocytes of rodents and other experimental animals. A novel transgenic mouse model in which murine PLA2R is expressed by the podocyte has been generated [14]. Passive administration of rabbit anti-PLA2R antibodies to these mice induces the formation of small subepithelial deposits, C3 deposition, and features of the nephrotic syndrome. However, human anti-PLA2R antibodies do not react with the mouse PLA2R and cannot be utilized in this model. By contrast, passive transfer of anti-THSD7A serum from patients with MN into mice has been shown to induce features of MN including proteinuria, granular immune complexes containing human IgG, and subepithelial electron-dense deposits [15]. In addition, a heterologous model of MN was produced when mice were injected with rabbit antibodies against human and mouse THSD7A [16].
Antigens implicated in human MN
Antigens in primary MN
Phospholipase A2 receptor — The M-type PLA2R, a transmembrane receptor that is highly expressed in glomerular podocytes, has been identified as a major antigen in human primary (formerly known as idiopathic) MN [17]. In this study, circulating autoantibodies to PLA2R were identified in 26 of 37 (70 percent) patients with primary MN and could be associated with disease activity in patients for whom serial serum samples were available [17]. By contrast, there was no evidence of PLA2R antibodies in serum from eight patients with secondary MN due to systemic lupus erythematosus (SLE) or hepatitis B, from 15 patients with proteinuric conditions other than MN (such as diabetic nephropathy or focal segmental glomerulosclerosis), or from 30 healthy control individuals. The circulating anti-PLA2R antibodies were predominantly IgG4, the IgG subclass that is most abundant in the glomerular immune deposits in primary (but not secondary) MN. PLA2R colocalized with IgG4 in immune deposits of kidney tissue obtained by kidney biopsy from patients with MN, and anti-PLA2R antibodies could be eluted from this tissue. This was in contrast to the findings in secondary MN biopsies, in which there was no colocalization of IgG4 and PLA2R and from which no anti-PLA2R antibodies could be eluted.
Most if not all patients with PLA2R-associated MN have antibodies that target the N-terminal region of the PLA2R protein [18-20]. Two independent studies demonstrated that the dominant epitope for anti-PLA2R lies within the three most N-terminal domains [18,19], and one study identified a conformational 31-amino acid region of the ricin domain containing an internal disulfide bond that was largely able to inhibit the antibody-antigen interaction [18]. Subsequent studies provided evidence that as many as 80 percent of patients exhibit reactivity to epitopes within the 1st and 7th C-type lectin-like domains in addition to the N-terminal ricin domain (so-called epitope spreading) [21], while a smaller proportion of patients display reactivity to the 8th C-type lectin-like domain [22].
Anti-PLA2R antibodies have been identified in approximately 70 to 80 percent of patients with primary MN [23-28], and several studies have shown that anti-PLA2R seropositivity correlates with clinical activity. As examples, one study found that anti-PLA2R levels strongly correlated with clinical status [25]; another reported that lower anti-PLA2R titers were associated with a higher rate of spontaneous remission [26], and in two other studies, a decline in anti-PLA2R titers predicted the clinical response to immunosuppressive therapy [28,29]. Yet another study found that higher anti-PLA2R titers within two years of diagnosis predicted substantially greater progression of kidney function decline over the subsequent five years of follow-up [27]. However, this may have reflected increased disease activity at the time the serum was collected; lower titers may have identified individuals already undergoing immunological and clinical remission. Some studies have found that epitope spreading is associated with a lower spontaneous remission rate, a higher risk of chronic kidney disease, and a lower response rate to treatment with rituximab [21,30]. Others have argued that it is the higher antibody titer associated with epitope spreading, rather than the spreading per se, that determines the outcome and response to treatment [22]. (See "Membranous nephropathy: Treatment and prognosis", section on 'Treatment of resistant disease'.)
Staining the kidney biopsy specimen for PLA2R, either by immunofluorescence or immunohistochemistry, provides another assay by which to identify PLA2R-associated primary MN [23,31,32]. As an example, in one study that reported a relatively low sensitivity of circulating anti-PLA2R (57 percent), an additional 24 percent of patients who did not have circulating antibodies had the PLA2R antigen detected within immune deposits by immunofluorescence of the biopsy specimen [23]. This may occur as patients enter serological remission with still unresolved proteinuria and persistent immune deposits in glomeruli. Theoretically, it could also occur in the early stages of disease as anti-PLA2R antibodies are "soaked up" in the immune deposits and have not yet reached sufficiently high levels to be detected in the serum by existing immunoassays (figure 1) [33-35]. In general, tissue staining for PLA2R may be more sensitive (69 to 84 percent in various studies) than circulating anti-PLA2R in patients with primary MN [23,31,32,36,37]. Specificity is close to 100 percent; however, PLA2R has been detected in the immune deposits of some patients with secondary MN associated with hepatitis B virus (HBV) infection, neoplasms, nonsteroidal antiinflammatory drug (NSAID) use, or sarcoidosis but not lupus nephritis (LN) [31,32,36,37]. It is possible that this represents a coincidental association, and we regard such cases as having PLA2R-associated MN.
The detection of circulating anti-PLA2R and/or PLA2R kidney immune deposits in the majority of adult patients with idiopathic MN represents a large step forward in our understanding of the disease as investigators will now be able to test whether the pathogenetic mechanisms learned from experimental models are also involved in human disease. The sensitivity and specificity of anti-PLA2R autoantibodies in individuals with immunologically active idiopathic MN have also enabled the development of a serologic immunoassay for the noninvasive diagnosis of primary MN and monitoring of disease activity. It should be noted that, while the association of these anti-PLA2R autoantibodies with disease suggests a causal role, this has not yet been confirmed by transferring the disease to experimental animals; however, in vitro studies have shown that anti-PLA2R autoantibodies are able to induce complement-dependent injury of cultured human podocytes [38]. In addition, an important clue to the pathogenicity of anti-PLA2R autoantibodies was provided by a study that identified the autoantibodies in the banked serum of a substantial proportion of military recruits with PLA2R-associated MN weeks to months before the onset of clinical disease and in none of the banked samples of matched controls [39].
Thrombospondin type-1 domain-containing 7A — THSD7A is, like PLA2R, a transmembrane protein expressed on podocytes [40,41]. THSD7A may be the responsible antigen in approximately 3 percent of patients with primary MN and in 10 percent of primary MN cases that test negative for anti-PLA2R [42]. The association of THSD7A with MN was examined in a study of 154 patients with anti-PLA2R-negative idiopathic MN, 74 patients with anti-PLA2R-positive idiopathic MN, 76 patients with other glomerular disease, and 44 healthy controls [40]. Autoantibodies specific for THSD7A were identified in sera from 15 of 154 patients with anti-PLA2R-negative idiopathic MN but not in the sera from other individuals. In addition, the IgG that was eluted from the kidney biopsies of 1 of these 15 patients was specific for THSD7A, providing further support that THSD7A was the target antigen in these patients. THSD7A-associated MN has been found at a low frequency in American, European, and Chinese [43] cohorts, but it may be more prevalent in Japanese patients with primary MN [44].
THSD7A may also be involved in the pathogenesis of some cases of malignancy-associated MN [45]. As an example, in one case report of a patient with anti-PLA2R-negative MN and concomitantly diagnosed adenoneuroendocrine carcinoma of the gallbladder, the expression of THSD7A was detected by immunohistochemistry on tumor cells but not on normal gallbladder tissue [45]. The patient also had elevated plasma levels of anti-THSD7A antibodies, which the authors proposed were formed against abnormally expressed THSD7A by tumor cells. Treatment with chemotherapy led to the disappearance of THSD7A antibodies in the plasma within two weeks and a marked reduction in proteinuria. Other studies have found a higher rate of malignancy (ranging from 20 to 50 percent) among patients with THSD7A-associated MN [46,47]. However, in one study of 31 patients with THSD7A-associated MN, only two (6 percent) had a history of malignancy, and none were diagnosed with malignancy on follow-up [48]. (See 'Malignancy' below.)
Rare cases of primary MN with dual serological and/or tissue positivity for THSD7A and PLA2R have been reported [42,43,49]. The immunological explanation for this finding is unknown and the clinical features of cases with dual positivity are not different than other cases of primary MN.
In addition, approximately 15 to 20 percent of cases of suspected primary MN are both serologically and tissue negative for PLA2R and THSD7A, which indicates that there are as-yet undiscovered antigens in primary MN.
Other antigens
●Neural epidermal growth factor-like 1 – NELL1 may be the antigen responsible in approximately 16 percent of cases of PLA2R-negative primary MN [50]. This protein was initially identified by laser microdissection and mass spectrometry of glomeruli in 6 of 35 cases of PLA2R-negative primary MN; all controls including 23 cases of PLA2R-associated MN and 88 cases without MN were negative for NELL1. Immunohistochemistry for NELL1 showed positive staining and colocalization with IgG along the GBM in all six cases of NELL1-positive MN as well as 23 of 91 additional cases of PLA2R-negative primary MN, while no staining was observed in patients with PLA2R-associated MN or other glomerular diseases. Circulating anti-NELL1 antibodies were detected in all of five patients with NELL1 positive MN for whom serum was available but none of the control patients with PLA2R-associated MN, minimal change disease, or immunoglobulin A (IgA) nephropathy. Both the mass spectrometry data and the limited characterization of the circulating anti-NELL1 antibodies suggest that the predominant IgG subclass in this disease may be IgG1 rather than IgG4. These findings indicate that NELL1 positivity may define a distinct form of primary MN. Additional studies are required to determine the role of this protein in the pathogenesis of MN. One study revealed that up to one-third of patients with NELL1-associated MN had a concurrent malignancy [51].
Cases of NELL1-associated MN following lipoic acid supplementation have also been described [52]. In such patients, discontinuation of lipoid acid led to remission of MN, and no immunosuppression was required.
●Semaphorin 3B – Semaphorin 3B (Sema3B) has been identified as the target antigen in a unique form of PLA2R-negative primary MN that appears to involve mostly children and young adults [53]. Sema3B was originally identified by laser microdissection and mass spectrometry of glomeruli in 3 of 160 cases of PLA2R-negative primary MN; the protein was not detected in 23 cases of PLA2R-associated MN or 88 controls without MN. Immunohistochemistry for Sema3B revealed positive staining and colocalization with IgG along the GBM in all three cases of Sema3B-associated MN as well as 8 of 118 additional cases of PLA2R-negative primary MN from three validation cohorts; staining was negative in patients with PLA2R-associated MN and other glomerular disorders. Circulating antibodies against Sema3B were detected in four of the five patients with Sema3B-associated MN with available sera and in none of healthy controls or patients with other glomerular diseases. Among the 11 cases of Sema3B-associated MN, eight (73 percent) were pediatric patients (<18 years old) and three (27 percent) were adult patients; of the eight pediatric patients, five developed MN on or before the age of 2 years. Two of the pediatric patients were siblings, and a third patient had a father with MN, raising the possibility of an inherited component to the disease. Thus, the finding of Sema3B antibodies in children with nephrotic syndrome may prove to be a valuable clue to the presence of MN rather than more common causes of childhood nephrotic syndrome like minimal change disease or focal segmental glomerulosclerosis.
●Protocadherin 7 – Protocadherin 7 (PCDH7) was identified as a distinct target antigen in a subset (14 cases) of patients with MN whose kidney biopsy was negative for all other known antigens [54]. In this cohort, the median age was 66 years, and six patients had possible secondary associations, including Sjögren's syndrome/SLE, sarcoidosis, and malignancy. A unique feature of the histopathologic phenotype of PCDH7-associated MN may be decreased amounts of complement factor 3 as kidney biopsy revealed only trace to 1+ staining by immunofluorescence. The presence of circulating autoantibodies to PCDH7 was demonstrated, and in one case, IgG eluted from the kidney biopsy tissue was found to be reactive with PCDH7.
●Serine protease HTRA1 – The serine protease high-temperature requirement A1 (HTRA1) represents another target antigen in a small proportion of patients with MN [55]. In one cohort of 14 cases, the average age was 67 years, and there were only two potential secondary associations: one patient had concurrent antineutrophil cytoplasmic autoantibodies (ANCA)-associated vasculitis with crescentic glomerulonephritis, and another patient had been diagnosed with stage IV small cell lung cancer two years prior to kidney biopsy, at a time when there was no proteinuria. On histopathologic examination of the larger cohort, IgG4 was the predominant IgG subclass detected within the immune deposits on kidney biopsy, and levels of circulating anti-HTRA1 antibodies (also of the IgG4 subclass) suggested a correlation with proteinuria and clinical disease activity.
●Netrin G1 – Netrin G1 (NTNG1), a glycosylphosphatidylinositol (GPI)-anchored membrane protein expressed in neurons and healthy podocytes, was identified as a target antigen in three patients with MN who did not have any antibodies against other antigens and did not have any other autoimmune diseases [56]. All three patients had circulating IgG4-dominant anti-NTNG1 autoantibodies, enhanced NTNG1 expression in the kidney, and glomerular IgG4 deposits.
●Neutral endopeptidase – NEP, which is expressed on podocytes, is the probable target in a rare antenatal form of MN [57-59]. The transplacental passage of anti-NEP antibodies (from mothers genetically deficient in NEP who were alloimmunized during a prior pregnancy) caused MN with subepithelial immune deposits (anti-NEP and NEP) in the fetus/neonate. Nephrotic syndrome resolved several months after birth, with disappearance of the deposits, upon clearance of the maternal antibodies. Although noncomplement-fixing IgG4 was the predominant IgG subclass of anti-NEP in alloimmunized NEP-deficient mothers, the development of proteinuria in their babies correlated with the additional presence of complement-fixing IgG1 anti-NEP [58]. This finding, together with reports showing that C5b-9 is shed into the urine of patients with recent-onset MN [60,61], provides additional evidence that the observations in Heymann nephritis are relevant to the human disease.
●Intracellular antigens – In addition to the target antigens noted above, antibodies directed against other antigens expressed by podocytes may contribute to the pathogenesis of MN [62-64]. In one study, for example, serum IgG4 reactivity against aldose reductase, superoxide dismutase 2, and alpha-enolase, as well as the PLA2R and NEP, was measured in 186 patients with MN, 36 patients with focal glomerulosclerosis, and 60 patients with IgA nephropathy [62]. Elevated titers of IgG4 against the PLA2R, alpha-enolase, aldose reductase, and superoxide dismutase 2 were found in 60, 43, 34, and 28 percent of patients with MN, respectively, but not in patients with other glomerular diseases. Approximately one-half of the patients who were negative for antibodies against the PLA2R had an elevated titer for one of the other three antibodies. Although these antigens are predominantly intracellular and are probably not primarily responsible for MN, it has been proposed that podocyte injury causes the intracellular enzymes to translocate to the cell surface where they are accessible to the circulating antibodies, causing amplification of the immune injury and possibly aggravating the course of the disease. A subsequent study has suggested that the presence and higher titers of antibodies to the intracellular antigens superoxide dismutase 2 or alpha-enolase were independently associated with poor clinical outcome [65].
●Cationic bovine serum albumin – Antibodies to a cationic form of bovine serum albumin (BSA) are present in a small number of children with MN [66]. The BSA antigen, which was found within the immune deposits of biopsy specimens from these patients, is thought to be absorbed from the relatively underdeveloped pediatric intestinal tract in a partially digested or undigested form and then serve as a planted antigen within the glomerular capillary wall. Antibodies reactive with bovine, but not human, serum albumin were eluted from the kidney biopsy specimen in one case.
Possible antigens in secondary MN — Other components of the glomerular immune deposits have been identified in patients with secondary MN [67]. These include double-stranded DNA in SLE; exostosin-1 and -2 in a subset of class V LN [68]; protocadherin FAT1 in hematopoietic cell transplant recipients [69]; thyroglobulin in thyroiditis; hepatitis B antigen, treponemal antigen, and Helicobacter pylori in the relevant infections; and carcinoembryonic antigen and prostate-specific antigen in malignancy. Their pathogenicity is unproven.
Role of T cells — T helper cells activate different immune effector mechanisms and appear to play a role in the pathogenesis of glomerulonephritis and may also participate in the genesis of proteinuria in MN. The T helper subset Th1 tends to predominate in proliferative and crescentic forms of glomerulonephritis, whereas Th2 predominates in MN and minimal change disease [70,71]. (See "The adaptive cellular immune response: T cells and cytokines", section on 'Cytokine profiles and functions of CD4+ T helper cell subsets' and "Mechanisms of immune injury of the glomerulus", section on 'T cells'.)
In support of a pathogenetic role for Th2 in MN is the observation in a model of LN that deletion of the IL27RA gene (encoding a cytokine receptor integral for mounting a Th1 response) causes a shift toward a Th2 response and converts the diffuse proliferative pattern that is typically seen to a membranous pattern [72]. The site of the target antigen and subepithelial location of the immune deposits also favor antibody- and complement-mediated injury to the podocyte, rather than a direct cellular inflammatory lesion (which is typically associated with subendothelial immune deposits and glomerular endothelial injury) [73,74].
GENETICS — A genome-wide association study that included 75 French, 146 Dutch, and 335 British individuals identified single-nucleotide polymorphisms (SNPs) at two loci that are highly associated with primary (idiopathic) MN [75]. The two loci are within the genes for the phospholipase A2 receptor (PLA2R) on chromosome 2q24, and the human leukocyte antigen (HLA) complex class II alpha chain 1A (HLA-DQA1) on chromosome 6p21. PLA2R variants have also been found in other cohorts with idiopathic MN, although no single variant was consistently found that could explain the association with disease [76]. The PLA2R has been identified as a major antigen in primary MN. (See 'Phospholipase A2 receptor' above.)
When the French, Dutch, and British cohorts from the aforementioned genome-wide study were analyzed separately, the HLA-DQA1 allele was associated with disease in all three populations; the PLA2R1 allele was associated in the Dutch and British populations but not in the French, although this may have been related to the smaller sample size in the French cohort. Homozygous expression of the HLA-DQA1 allele conferred a higher risk for disease compared with the PLA2R1 allele (odds ratios of 20.2 versus 4.2, respectively). The homozygous expression of both alleles additively increased the risk of disease (odds ratio 78.5). These variants are also associated with MN in Chinese and Spanish patients [77,78]; in the Chinese population, homozygosity for the high-risk alleles was associated with a much higher prevalence of circulating anti-PLA2R than homozygosity for the low-risk alleles [77].
Other analyses have used high-density SNPs to impute molecular HLA haplotypes or have directly sequenced the HLA region. In individuals of European ancestry, primary MN was significantly associated with all the alleles in the ancestral MHC8.1 haplotype, including HLA-DQA1*0501, HLA-DQB1*0201, and HLA-DRB1*0301 [79]. Three studies regarding genetic risk for primary MN in Asian populations have identified a separate risk haplotype that includes HLA-DRB1*1501 [80-82], indicating that genetic risk alleles within the HLA region likely differ among ethnicities. Molecular modeling in a Chinese cohort identified putative amino acids in PLA2R that might engage the binding pocket of the corresponding major histocompatibility complex (MHC) class II risk molecules [80].
These collective findings were confirmed in a large genome-wide association study of 3782 cases of primary MN and 9038 controls of East Asian and European ancestry [83]. This study also identified two new loci that were highly associated with MN: one in NFKB1 and another in IRF4, both encoding important molecules in immune regulation. In addition, the study described an SNP in tight linkage disequilibrium to the previously identified top locus within the first intron of PLA2R1; the risk allele at this position may be linked to increased expression of PLA2R1 exclusively in kidney tissue. A detailed analysis of the HLA locus confirmed the associations with the molecular HLA haplotypes and additionally provided evidence that the major risk in both East Asian and European populations may lie exclusively in the DR-beta chain at positions that alter discrete amino acids in the binding groove.
ETIOLOGY — Approximately 75 percent of cases of MN in adults are primary (idiopathic). Secondary MN has been attributed to a variety of agents or conditions (table 1), largely due to observations that removal of the inciting agent or treatment of the condition led to resolution of the nephrotic syndrome.
Other than by associating a medication or disease state with MN, it is not possible to distinguish primary from secondary MN based solely upon the clinical manifestations. However, there are certain findings on electron microscopy and immunofluorescence that suggest secondary disease (see "Membranous nephropathy: Clinical manifestations and diagnosis", section on 'Primary versus secondary MN'). In addition, the identification of anti-phospholipase A2 receptor (PLA2R) antibodies in primary but not secondary forms of MN permits the discrimination of these two entities in most cases. (See 'Phospholipase A2 receptor' above and "Membranous nephropathy: Clinical manifestations and diagnosis", section on 'Diagnosis'.)
Primary MN — Primary MN includes forms of MN in which there is a humoral autoimmune response to a normal podocyte antigen in the absence of secondary features or etiologies of disease. Antigens implicated in primary MN include PLA2R, thrombospondin type-1 domain-containing 7A (THSD7A), neural epidermal growth factor-like 1 (NELL1), semaphorin 3B (Sema3B), the serine protease high-temperature requirement A1 (HTRA1), protocadherin 7 (PCDH7), and others. These are discussed in more detail elsewhere in this topic. (See 'Antigens in primary MN' above.)
Together, the antigens identified over the past decade account for approximately 90 percent of cases of primary MN, which means that it is likely that additional minor antigens will be added to the list in the future.
Secondary MN
Systemic lupus erythematosus — Approximately 10 to 20 percent of patients with lupus nephritis (LN) have MN, called class V LN. (See "Lupus nephritis: Diagnosis and classification".)
Some patients with lupus membranous nephropathy (LMN) present only with kidney disease and have no symptoms or serologic abnormalities suggestive of systemic lupus erythematosus (SLE), although such findings may arise several months after presentation with the nephrotic syndrome [84,85]. SLE should be suspected in any young woman with apparently idiopathic MN. In addition, there are some histologic findings on immunofluorescence (staining for IgG2, IgG3, IgA, immunoglobulin M [IgM], and C1q) and electron microscopy (subendothelial and subepithelial immune deposits and tubuloreticular structures in the glomerular endothelial cells) that are suggestive of underlying SLE [84,85]. PLA2R antibodies are typically negative in patients with class V LN. However, patients with PLA2R-associated primary MN may coincidently have a positive antinuclear antibody (ANA) if they also have SLE without kidney involvement. Additional features on biopsy, such as positive staining for PLA2R, lack of C1q staining, IgG4 predominance, exclusively subepithelial electron-dense deposits, and absence of tubular basement membrane staining will further help distinguish primary MN from class V LN. (See "Membranous nephropathy: Clinical manifestations and diagnosis", section on 'Primary versus secondary MN'.)
Positive immunostaining for the proteins exostosin 1 (EXT1) and exostosin 2 (EXT2) may help to identify a subset of patients with class V LN or another autoimmune disease such as Sjögren's syndrome. In one study, EXT1 and EXT2 within glomeruli were detected by mass spectroscopy and confirmed by immunohistochemistry in 21 of 224 cases of PLA2R-negative MN but none of 102 controls (including 47 cases of PLA2R-associated MN) [68]. Among the cases of EXT1/EXT2-positive MN, 71 percent were positive for autoimmune autoantibodies and 35 percent had a clinical diagnosis of SLE; kidney biopsy findings in these cases were also characteristic of systemic autoimmune disease. Analysis of a validation cohort of 48 patients revealed positive EXT1/EXT2 staining in 8 of 18 cases (44 percent) of class V LN and 3 of 16 cases (19 percent) of presumed primary MN associated with signs of autoimmunity, but only 1 of 14 cases (7 percent) of mixed class V and class III/IV LN. Among seven patients with EXT1/EXT2-associated MN who underwent testing, no circulating anti-exostosin antibodies were detected. Collectively, these findings identify EXT1 and EXT2 as markers of MN associated with systemic autoimmunity. However, their role in disease pathogenesis and as predictors of prognosis and response to treatment has yet to be defined.
Two additional proteins have been associated with class V LN as either biomarkers or target antigens:
●Neural cell adhesion molecule 1 (NCAM1) was identified as a target antigen in a subset of patients of class V LN or MN of identified type [86]. In this report, circulating antibodies to NCAM1 were found in several cases. The patients were predominantly young females (average age 34 years) with positive testing for ANA. A higher-than-expected proportion of this cohort exhibited neuropsychiatric manifestations of SLE.
●The type III transforming growth factor (TGF) beta receptor (TGFBR3) was reported as a distinct biomarker in a subset of patients with class V LN [87]. In this study, 11 of 199 cases (6 percent) of class V LN exhibited staining for TGFBR3 within the glomerular immune deposits, as opposed to 0 of 104 non-LMN cases. The cohort was nearly all female, with an average age of approximately 40 years. Circulating antibodies to TGFBR3 were not detected [87].
In case reports, LMN has been associated with a concurrent crescentic glomerulonephritis, a relationship that has also been described in primary MN. (See 'Crescentic glomerulonephritis' below.)
Drugs — Exposure to a variety of agents that are primarily used to treat rheumatoid arthritis has been implicated in the development of MN, including nonsteroidal antiinflammatory drugs (NSAIDs), penicillamine, parenteral gold salts, bucillamine, alemtuzumab, mercurial salts, elemental mercury, and possibly anti-tumor necrosis factor (anti-TNF) agents (etanercept, infliximab, or adalimumab) [88-97]. (See "Overview of the systemic and nonarticular manifestations of rheumatoid arthritis", section on 'Kidney disease' and "NSAIDs: Acute kidney injury" and "Overview of biologic agents and kinase inhibitors in the rheumatic diseases".)
The association of NSAIDs with MN was illustrated in a study of 125 patients with a biopsy diagnosis of MN [88]. Twenty-nine patients were taking an NSAID, and 13 (10 percent of the study population) fulfilled three criteria suggesting that the NSAID was responsible:
●No other apparent cause for the MN
●Resolution of proteinuria within 1 to 36 weeks of discontinuing NSAIDs
●No recurrence of proteinuria at follow-up (5 months to 13 years)
Many of the patients who developed MN had been treated with diclofenac, but probably any NSAID can be involved [88], including cyclooxygenase (COX) 2 inhibitors [90]. (See "NSAIDs: Acute kidney injury".)
The incidence of MN may be as high as 7 percent in patients treated with penicillamine and 1 to 3 percent in those treated with parenteral gold (the risk with oral gold [auranofin] appears to be lower) [98-100]. Tiopronin (2-mercaptopropionylglycine), an agent structurally similar to penicillamine that is used to treat cystine stones, has also been reported to cause MN, but the incidence is rare [101]. High-dose captopril, which also has a free thiol group, has also been associated with MN. (See "Cystinuria and cystine stones".)
The mechanisms responsible for drug-induced MN are uncertain. Human and experimental data suggest that induction of gold-specific and autoreactive T cells may lead to polyclonal B cell activation and autoantibody production in gold-induced MN [102,103]. Interestingly, it has been suggested that patients carrying the lupus-associated human leukocyte antigen (HLA) alleles DRB1*0301 (DR3) and DQA1*0501 are particularly susceptible to develop MN, as well as drug-induced lupus, after exposure to gold salts [104]. Of note, these HLA alleles have also been identified as risk alleles for primary MN in White European individuals. (See 'Genetics' above.)
Proteinuria generally develops within the first 6 to 12 months of drug therapy but can occur as late as three to four years [100]. Discontinuation of the drug leads to resolution of the proteinuria in virtually all cases [98,100]. However, studies with penicillamine, gold, and bucillamine indicate that protein excretion may continue to rise for the first 1 to 12 months (two months on average) after the cessation of therapy [100]. The mean time to resolution of the proteinuria is 9 to 12 months, although two to three years are required in some cases [92,98,100]. The prolonged time to recovery may be related, in part, to the slow rate of clearance of subepithelial immune deposits and to slow remodeling of the disorganized glomerular basement membrane (GBM) [12,73].
It should be noted that MN is not the only glomerulopathy seen with these drugs. Gold and NSAIDs can lead to minimal change disease [100], whereas penicillamine can induce an immune complex crescentic glomerulonephritis. Anti-TNF therapy has also been associated with the new onset of LN and pauci-immune necrotizing and crescentic glomerulonephritis [93].
Infections
Hepatitis B virus — MN due to hepatitis B virus (HBV) infection primarily occurs in children in endemic areas, many of whom are asymptomatic carriers with no history of active hepatitis [5,105-107]. The serum transaminases tend to be normal or only mildly elevated, and the serology is positive for HBV surface antigen, anti-core antibody, and usually e antigen. It appears that it is the e antigen and cationic anti-e antibody that are primarily deposited in the glomeruli [5,84,106].
HBV infection and SLE represent the only forms of MN that may be associated with hypocomplementemia [105].
Spontaneous resolution of the proteinuria is common in children with MN associated with HBV infection but not in adults, many of whom will have progressive disease [106]. (See "Kidney disease associated with hepatitis B virus infection".)
There is strong epidemiological and clinical evidence that HBV infection is responsible for secondary MN, especially in children:
●The prevalence of childhood MN closely parallels the geographic distribution of HBV.
●The frequency of HBV and MN in children with nephrotic syndrome in Taiwan declined from approximately 12 percent to less than 1 percent over a 20-year period following a universal vaccination program between 1984 and 2009 [108], and similar results were reported in China and South Africa.
●Antiviral therapy induces remission of proteinuria, especially in children and in individuals without advanced disease [109].
Most, but not all, studies have shown a low prevalence of anti-PLA2R antibodies and/or PLA2R staining of the immune deposits in hepatitis B-associated MN [17,24]. However, one retrospective study of 39 cases of hepatitis B-associated MN found that 64 percent of cases exhibited strong PLA2R staining of immune deposits [110]. Staining for the hepatitis B surface antigen colocalized with PLA2R within the deposits. Among the 6 cases for which serum was available, all were positive for anti-PLA2R antibodies.
In a series of 16 patients with MN and HBV infection, none of the patients who were positive for anti-PLA2R antibodies entered remission with antiviral therapy, whereas all of those who were negative for anti-PLA2R antibodies achieved complete remission with antiviral therapy alone [24].
The association between hepatitis B infection and PLA2R-positive MN is likely to be coincidental in most cases given the high prevalence of hepatitis B infection in certain populations [24].
Hepatitis C virus — MN may also be uncommonly associated with chronic hepatitis C virus (HCV) infection [111]. This is discussed separately. (See "Overview of kidney disease associated with hepatitis C virus infection".)
Syphilis — Congenital and secondary syphilis have been associated with MN [112-117]. Treponemal antigens have been identified in the glomeruli by immunofluorescence microscopy, and eluates from glomerular deposits contain antibodies specific for Treponema pallidum antigen [113,114,116]. Furthermore, effective treatment of syphilis can lead to resolution of the glomerular disease [113,115-117]. (See "Syphilis: Epidemiology, pathophysiology, and clinical manifestations in patients without HIV", section on 'Clinical manifestations'.)
Malignancy — Up to 5 to 20 percent of adults with MN, particularly those over the age of 65 years, have been reported to have a malignancy, most commonly a solid tumor (principally carcinoma of the prostate, lung, breast, bladder, or gastrointestinal tract), and less frequently, a hematologic malignancy, such as chronic lymphocytic leukemia [112,118-124].
Based upon the population examined, the risk of malignancy among those with MN can vary from 2 to 12 times higher than that observed in the general population after adjustment for age and sex [122,123]. The malignancy is presumed to be etiologically associated when there is a temporal relationship between the tumor and the MN, when removal of the tumor is followed by gradual remission of the proteinuria, and when recurrence of the malignancy is followed by return of the proteinuria. Ultimate proof of causality is the detection of tumor antigens in the GBM, but this is only rarely described [125-128].
One proposed mechanism is that deposition of tumor antigens in the glomeruli promotes antibody deposition and complement activation, leading to epithelial cell and GBM injury, and consequent proteinuria [125]. The production of antibodies produced against tumor antigens that also recognize similar or identical molecules present on podocytes may be an alternative mechanism for malignancy-associated MN. (See 'Thrombospondin type-1 domain-containing 7A' above.)
Despite these reports, many other cases of malignancy-associated MN may represent coincident disease processes, rather than a direct causal relationship, for the following reasons:
●The putatively associated tumors are common in males over the age of 50 years, the same population that tends to get MN.
●Remission of the nephrotic syndrome with removal of the tumor does not necessarily imply a therapeutic response, since there is a relatively high rate of spontaneous remission in MN (figure 2).
Regardless of whether the malignancy is a causal factor or simply coincident, in many cases it has already been diagnosed or is clinically apparent at the onset of proteinuria [122]. A diagnosis of MN preceding that of malignancy is more likely in older adults [118,122]. (See "Membranous nephropathy: Clinical manifestations and diagnosis", section on 'Screening for malignancy'.)
IgG4-related disease — IgG4-related disease is an increasingly recognized systemic syndrome of unknown etiology characterized by tumor-like swelling of involved organs, a lymphoplasmacytic infiltrate enriched in IgG4-positive plasma cells, and variable degrees of fibrosis that have a characteristic "storiform" pattern. In addition, elevated serum concentrations of IgG4 are found in 60 to 70 percent of patients with IgG4-related disease. The major kidney manifestation associated with IgG4-related disease is tubulointerstitial nephritis. (See "Pathogenesis and clinical manifestations of IgG4-related disease", section on 'Kidney disease'.)
However, IgG4-related disease may also be associated with MN [129-133]. This was described in nine patients with MN who also had other manifestations of IgG4-related disease (mostly pancreatitis, tubulointerstitial nephritis, or sialadenitis) [129]. Antibodies against the PLA2R, commonly present in patients with primary MN, were absent in all patients. In addition, none of the nine patients had evidence for SLE, hepatitis virus infection, or cancer.
Membranous-like nephropathy with masked IgG-kappa — A glomerulopathy has been described in 14 patients that is indistinguishable from MN by light microscopy and electron microscopy [134]. However, standard immunofluorescence microscopy was negative for the typical granular deposition of IgG. Only after the formalin-fixed tissue was digested with pronase were IgG deposits "unmasked." In addition, all IgG deposits consisted of IgG-kappa and not IgG-lambda. The etiology of this condition is unknown, but most of the 14 affected patients were young and had autoimmune phenomena such as inflammatory arthritis or hemolytic anemia. Another study found that the detection of serum amyloid P by immunostaining within the immune deposits may be a specific marker for this type of disease [135].
MN with light chain-restricted deposits — Cases of MN with light chain isotype-restricted deposits have been reported [124,136-138]. In the two largest series comprising 42 patients, a hematological malignancy was identified in approximately 25 percent of patients, few of whom had a detectable monoclonal protein in the serum or urine [124,137]. Systemic autoimmunity (two cases) and one case with hepatitis B and syphilis were also reported in one series; the remaining patients did not have a recognizable secondary etiology [124]. Immunofluorescence microscopy revealed kappa light chain restriction in 39 of 42 patients (93 percent) [124,137], and positive staining for PLA2R was reported in 7 of 27 patients (26 percent) in one series [124]. IgG1 subclass restriction was present in 70 percent of cases in which IgG subclass staining was available [124,137]. Patients with negative staining for PLA2R, positive staining for a single IgG subclass, and/or the presence of focal proliferation and/or crescents by light microscopy were more likely to have an underlying lymphoproliferative disorder [124].
In rare circumstances, circulating anti-PLA2R autoantibodies may be monoclonal in nature, representing a monoclonal gammopathy of renal significance. One report details a patient with monoclonal IgG3 kappa anti-PLA2R [136]. Clinical manifestations were similar to those in primary PLA2R-associated MN, although immunofluorescence of the biopsy revealed the additional presence of C1q and restriction to kappa light chain and the IgG3 subclass. The monoclonal autoantibody was also shown to cause recurrent disease in the allograft after transplantation. (See "Diagnosis and treatment of monoclonal gammopathy of renal significance" and 'MN after kidney transplant' below.)
Hematopoietic cell transplantation and graft-versus-host disease — Nephrotic syndrome occasionally arises in recipients of allogeneic stem cell or, less commonly, bone marrow transplants and is often temporally correlated with chronic graft-versus-host disease (GVHD).
MN is the most frequently reported underlying histology, although minimal change disease is also seen. The occurrence of MN is often associated with a decrease in immunosuppression. (See "Clinical manifestations and diagnosis of chronic graft-versus-host disease" and "Kidney disease following hematopoietic cell transplantation", section on 'Nephrotic syndrome'.)
Others — MN has been infrequently reported in association with a variety of other conditions in case reports or small series (table 1) [112,139,140]. These include hepatosplenic schistosomiasis, quartan malaria, sarcoidosis, Sjögren's syndrome, and chronic exposure to formaldehyde [107,112,141]. (See "Schistosomiasis and glomerular disease" and "Kidney disease in sarcoidosis" and "Kidney disease in primary Sjögren syndrome".)
A case of MN occurring shortly after administration of COVID-19 mRNA vaccines has also been reported [142]. (See "COVID-19: Issues related to acute kidney injury, glomerular disease, and hypertension", section on 'Glomerular disease'.)
MN after kidney transplant — MN may recur after kidney transplantation, especially if anti-PLA2R is positive at the time of transplantation, or it may occur de novo [143]. Alloimmunization against minor histocompatibility antigens or privately expressed donor epitopes may play a role in de novo MN [67]. Patients with de novo MN do not have PLA2R antigen deposits in glomerular capillaries and are nearly always anti-PLA2R antibody negative. (See "Membranous nephropathy and kidney transplantation".)
MN with other glomerular diseases — MN may be seen in conjunction with other glomerular diseases. These include:
●Diabetic nephropathy (see 'Diabetes mellitus' below)
●Crescentic (rapidly progressive) glomerulonephritis (see 'Crescentic glomerulonephritis' below)
●Focal segmental glomerulosclerosis (see "Focal segmental glomerulosclerosis: Epidemiology, classification, clinical features, and diagnosis")
●IgA nephropathy (see "IgA nephropathy: Clinical features and diagnosis", section on 'Associated conditions')
It is unclear if these disorders are causally related to MN or have developed concurrently. In addition, the mechanism may vary with the second disease. With focal segmental glomerulosclerosis, for example, the sclerotic lesions may reflect a response to injury induced by the membranous disease and/or secondary intraglomerular hypertension, rather than a separate disorder [144]. Patients with these sclerotic lesions have a much higher likelihood of progressive kidney function impairment than those with uncomplicated MN [145,146]. (See "Focal segmental glomerulosclerosis: Epidemiology, classification, clinical features, and diagnosis".)
Diabetes mellitus — MN (as well as other glomerular diseases) can occur in patients with diabetes with or without diabetic nephropathy [147,148]. In a biopsy study of 220 type 2 diabetic patients with proteinuria, nondiabetic kidney disease was present in 37.2 percent of patients with heavy proteinuria (average urine protein-to-creatinine ratio of 10.0 ± 7.3 g/g), and of those, 41.7 percent had MN [149]. In most cases, the diseases are not related, but porcine insulin may be the inciting antigen in some patients.
A possible pathogenetic role for porcine insulin was suggested in a report of seven patients with diabetes who developed MN [148]. Three of the patients had granular deposits of porcine insulin, as detected by anti-porcine insulin antibodies, as well as IgG and complement along the glomerular capillary wall. The nephrotic syndrome developed after the initiation of porcine insulin in all three patients, and in two of these patients, switching from porcine to human insulin led to amelioration of the proteinuria. The other four patients were thought to have idiopathic MN.
Alloimmune MN associated with hormone or enzyme replacement therapies — In addition to MN associated with porcine insulin (see 'Diabetes mellitus' above), MN has been documented in rare patients with hereditary enzyme deficiencies receiving enzyme replacement therapy (ERT) with recombinant human acid alglucosidase alfa (rhGAA) for lysosomal acid alpha-glucosidase deficiency (Pompe disease) and with recombinant aryl-sulfatase B (rhASB) for mucopolysaccharidosis type VI (MPS VI) [150,151]. In both cases, the recombinant enzyme was detected and colocalized with IgG in the glomerular immune deposits, and in the case of MPS VI, IgG eluted from the glomeruli was reactive with rhASB [151]. Whereas proteinuria improved in both cases with reduction or cessation of ERT, the patient with MPS VI relapsed when the life-saving ERT was resumed and required tolerance induction therapy. Although circulating alloantibodies to porcine insulin and recombinant enzymes occur commonly, it is unclear why the "foreign" protein localizes in the glomeruli to serve as a planted antigen in such rare cases. (See "Lysosomal acid alpha-glucosidase deficiency (Pompe disease, glycogen storage disease II, acid maltase deficiency)", section on 'Treatment'.)
Crescentic glomerulonephritis — MN is infrequently associated with a crescentic glomerulonephritis that in some patients may be related to antineutrophil cytoplasmic antibodies (ANCA) or, less often, anti-GBM antibodies [152-160]. This relationship has mostly been described in patients with idiopathic MN, but there are also case reports in patients with LMN [161]. (See 'Systemic lupus erythematosus' above.)
Crescentic glomerulonephritis and the associated nephritic sediment can be superimposed upon preexisting MN or occur in conjunction with MN. In a report of 10 patients, for example, 9 had concurrent disease at presentation, and 1 developed crescentic glomerulonephritis five years after the original presentation of MN [152].
Why these disorders are associated with MN is not clear. One possibility is the coincident production of ANCA or anti-GBM antibodies and the antibody responsible for MN in an individual predisposed to autoimmune disorders. An alternative mechanism is that GBM injury induced by the membranous lesion may expose previously "unseen" antigens, leading to anti-GBM antibody production.
A study from a major kidney pathology laboratory identified 14 patients with both MN and ANCA-associated necrotizing and crescentic glomerulonephritis; both conditions were diagnosed simultaneously in most of these patients [156]. In a review of over 13,000 native kidney biopsies processed by the same laboratory, idiopathic MN was present in 8.8 percent and ANCA-associated necrotizing and crescentic glomerulonephritis was present in 3.4 percent. Based upon these numbers, concurrent disease would have been expected by chance in 39 patients, well above the observed 14 cases. It was therefore concluded that concurrent disease represented a chance occurrence of two unrelated disease processes.
The findings were different in a retrospective review of 218 patients with MN; 10 (4.6 percent) had crescentic glomerulonephritis [152]. The incidence was much higher than would be expected for the random co-occurrence of the two disorders. ANCA were present in four of the nine patients who were tested. In one such case, the patient tested positive for both ANCA and anti-PLA2R [162].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Glomerular disease in adults".)
SUMMARY
●Epidemiology – Membranous nephropathy (MN) is among the most common causes of the nephrotic syndrome in nondiabetic adults, accounting for up to one-third of biopsy diagnoses. MN accounts for approximately 20 to 30 percent of cases of nephrotic syndrome in White adults, and a rising incidence has been reported in China, perhaps related to environmental pollution. (See 'Epidemiology' above.)
●Pathogenesis – The immune deposits that are characteristic of this disorder may develop in situ with the movement across the glomerular basement membrane (GBM) of circulating IgG antibodies directed against endogenous antigens expressed on or near the podocyte foot processes or against circulating cationic or low-molecular-weight antigens that have crossed the anionic charge barrier in the GBM. (See 'Pathogenesis' above.)
●Antigens implicated in primary MN – Antibodies to the M-type phospholipase A2 receptor (PLA2R), a major antigen expressed on podocytes, are specifically found in a high proportion of patients with primary (idiopathic) MN. Antibodies to thrombospondin type-1 domain-containing 7A (THSD7A), neural epidermal growth factor-like 1 (NELL1), semaphorin 3B (Sema3B), high-temperature requirement A1 (HTRA1), and protocadherin 7 (PCDH7) have been found to identify a smaller subset of patients with primary MN. (See 'Antigens implicated in human MN' above.)
●Etiology – MN is most often primary, associated with the presence of anti-PLA2R or anti-THSD7A antibodies, but it has been associated with a variety of conditions, including hepatitis B antigenemia, autoimmune diseases (eg, systemic lupus erythematosus [SLE]), malignancy, and the use of certain drugs such as nonsteroidal antiinflammatory drugs (NSAIDs), gold, and penicillamine. MN may also be seen in conjunction with other glomerular diseases such as diabetic nephropathy and crescentic glomerulonephritis. (See 'Etiology' above and 'MN with other glomerular diseases' above.)
