Discovery of Autoantibodies Targeting Nephrin in Minimal Change Disease Supports a Novel Autoimmune Etiology

Significance Statement

Although corticosteroids are an effective first-line therapy for minimal change disease, relapse, steroid dependence, and intolerance are common in this podocytopathy of unknown etiology. The efficacy of B cell–targeted therapies in some patients suggests an autoantibody-mediated etiology. This study describes the novel discovery in both adults and children with minimal change disease of autoantibodies targeting nephrin, a critical component of the podocyte slit diaphragm that ensures integrity of the glomerular filtration barrier. This observation aligns with the established proteinuric effect of antinephrin antibodies demonstrated in animal models. These findings identify an important autoimmune mechanism in a subset of patients with minimal change disease and provide a framework for the application and development of precision medicine strategies in this condition.

Visual Abstract

Abstract

Background

Failure of the glomerular filtration barrier, primarily by loss of slit diaphragm architecture, underlies nephrotic syndrome in minimal change disease. The etiology remains unknown. The efficacy of B cell–targeted therapies in some patients, together with the known proteinuric effect of anti-nephrin antibodies in rodent models, prompted us to hypothesize that nephrin autoantibodies may be present in patients with minimal change disease.

Methods

We evaluated sera from patients with minimal change disease, enrolled in the Nephrotic Syndrome Study Network (NEPTUNE) cohort and from our own institutions, for circulating nephrin autoantibodies by indirect ELISA and by immunoprecipitation of full-length nephrin from human glomerular extract or a recombinant purified extracellular domain of human nephrin. We also evaluated renal biopsies from our institutions for podocyte-associated punctate IgG colocalizing with nephrin by immunofluorescence.

Results

In two independent patient cohorts, we identified circulating nephrin autoantibodies during active disease that were significantly reduced or absent during treatment response in a subset of patients with minimal change disease. We correlated the presence of these autoantibodies with podocyte-associated punctate IgG in renal biopsies from our institutions. We also identified a patient with steroid-dependent childhood minimal change disease that progressed to end stage kidney disease; she developed a massive post-transplant recurrence of proteinuria that was associated with high pretransplant circulating nephrin autoantibodies.

Conclusions

Our discovery of nephrin autoantibodies in a subset of adults and children with minimal change disease aligns with published animal studies and provides further support for an autoimmune etiology. We propose a new molecular classification of nephrin autoantibody minimal change disease to serve as a framework for instigation of precision therapeutics for these patients.

Diffuse podocytopathy with minimal changes (minimal change disease [MCD]) is an important and common pathologic diagnosis in adults and children with nephrotic syndrome. It is characterized by minimal changes by light microscopy, yet extensive injury to glomerular podocytes with diffuse foot process effacement and loss of slit diaphragms by electron microscopy, in the absence of electron-dense deposits.¹ The consequence of these alterations is massive proteinuria secondary to failure of the glomerular filtration barrier, whose integrity is critically dependent on the specialized junctional slit diaphragm protein complex linking the interdigitating podocyte foot processes.²

Nephrin is an essential structural component of the slit diaphragm,² as illustrated by genetic mutations in NPHS1, that cause complete lack of nephrin cell surface localization, underlying congenital nephrotic syndrome of the Finnish type (CNF).^3,4 In contrast to congenital nephrotic syndrome with an established genetic basis, the cause of noncongenital nephrotic syndrome in both children and adults remains largely unknown. There is strong evidence supporting immune dysregulation with a potential causative circulating factor; however, its identity has remained elusive.^5,6 Glucocorticoids are effective at inducing remission; however, relapse, steroid dependence, and intolerance are common, often requiring alternative immunosuppressive agents.⁷ In those patients with steroid-dependent nephrotic syndrome who progress to end stage kidney disease (ESKD) and require kidney transplantation, the disease can promptly recur in the allograft,¹ a devastating and difficult-to-treat complication.

The recent discovery that anti-CD20 B cell–targeted therapies are effective in children with frequently relapsing or steroid-dependent nephrotic syndrome^8–10 and in adults¹¹ suggests a potential autoantibody-mediated etiology. However, this possibility is hard to reconcile with the traditional view of MCD lacking IgG deposition on renal biopsy.¹² Although diffuse podocyte-associated IgG is described in MCD, it is minimal compared with that seen in membranous nephropathy, and given the absence of electron-dense deposits by electron microscopy, it is generally attributed to nonspecific protein resorption of little significance.¹³

Antibodies targeting the essential structural slit diaphragm component nephrin have been shown to cause massive proteinuria when administered in animal models,^14–16 and when they arise as alloantibodies following kidney transplantation in children with CNF and complete nephrin deficiency.^17,18 In both animal models^14,15 and cultured podocytes,⁶ anti-nephrin antibodies cause a redistribution of nephrin that is identical to that observed in renal biopsies of patients with nephrotic syndrome.^19,20 This redistribution of nephrin away from the slit diaphragm along with separation of intercellular junctions between adjacent podocytes has been proposed to explain the proteinuria in these patients; however, the cause of this redistribution remains unknown.

The phenomenon of direct autoantibody-mediated disruption of a junctional adhesion complex is well described for the blistering skin condition pemphigus.²¹ In this disease, autoantibodies directly bind desmogleins (dsgs), critical crosslinkers in the desmosomal cell adhesion complex that are analogous to nephrin in the slit diaphragm complex, and cause redistribution of dsgs away from the cell surface with consequent loss of desmosomal integrity and ultimately cell-cell interactions.²¹

Taken together these observations led us to hypothesize that autoantibodies against nephrin may underlie noncongenital MCD by interfering with the integrity of the slit diaphragm complex.

Methods

Clinical Samples (Kidney Tissue and Serum/Plasma)

Renal biopsies were independently assessed by collaborating renal pathologists across four institutions: Brigham and Women’s Hospital (BWH), Massachusetts General Hospital (MGH), Boston Medical Center (BMC), and the Mayo Clinic. Serum/plasma was obtained from patients attending those institutions as either discarded samples originally collected for clinical analysis (BWH/BMC/Mayo clinic) or archival samples from the Kidney Disease Biobank (courtesy of Dr. Sushrut Waikar, Partners Healthcare, in accordance with Partners Healthcare institutional review board approval for patients attending BWH or MGH and were consented for serum/plasma collection at the time of renal biopsy). Histologic studies were performed on archival kidney tissue that was received for routine clinical evaluation and included diffuse podocytopathies, other nephrotic conditions, and non-neoplastic renal parenchyma from tumor nephrectomies. Medical record review and histologic and serological studies were approved by the respective institutional review boards for those institutions. Genetic testing was only performed for the patient with post-transplant recurrent disease but not for the other patients enrolled outside the NEPTUNE cohort.

Similarly, sera were obtained from patients with biopsy proven MCD from the Nephrotic Syndrome Study Network (NEPTUNE) longitudinal study²² during active disease (urine protein creatinine ratio [UPCR] >3 g/g) and where available, in remission. Complete remission was defined as UPCR <0.3 g/g and partial remission as a >50% reduction in proteinuria. Steroid-dependent nephrotic syndrome was defined as steroid-sensitive nephrotic syndrome with two or more consecutive relapses during tapering or within 14 days of stopping steroids. No specific exclusion criteria were applied.

Healthy control sera were randomly selected from Partners Healthcare Biobank, specifically excluding those subjects with any renal or autoimmune disease. Sera from nephrotic patients were evaluated for anti-human phospholipase A2 receptor (hPLA₂R) antibodies at the MGH immunopathology laboratory using a commercial ELISA and indirect immunofluorescence test (IIFT) (Euroimmun). Samples were coded to preserve patient anonymity.

Whole Genome Sequencing of the NEPTUNE Cohort

Whole genome sequencing with a goal median depth of 30× was performed using Illumina HiSeq. A standard pipeline, GotCloud, was applied for sequence alignment and variant calling.^23,24 The variant analysis focused on 68 genes implicated in Mendelian nephrotic syndrome. To screen pathogenic variants in the 68 previously implicated Mendelian nephrotic syndrome genes, we employed a pipeline similar to one that has been previously reported.²⁵ The pathogenicity variants were ultimately classified according to the American College of Medical Genetics and Genomics standards and guidelines.²⁶ The analyzed genes were as follows: ACTN4, ADCK4, ALG1, ANLN, ARHGAP24, ARHGDIA, AVIL, CD151, CD2AP, CDK20, CFH, COL4A3, COL4A4, COL4A5, COQ2, COQ6, CRB2, DGKE, DLC1, EMP2, FAT1, HNF1B, IL15RA, INF2, ITGA3, ITGB4, ITSN1, ITSN2, JAG1, KANK1, KANK2, KANK4, LAGE3, LAMB2, LMX1B, MAGI2, MTTL1, MYH9, MYO1E, MYO5B, NEIL1, NPHS1, NPHS2, NUP107, NUP205, NUP93, NXF5, OCRL, OSGEP, PAX2, PDSS2, PLCE1, PMM2, PODXL, PTPRO, SCARB2, SGPL1, SMARCAL1, TNS2, TP53RK, TPRKB, TRPC6, TTC21B, UMOD, WDR73, WT1, XPO5, and ZMPSTE24.

Human Glomerular Extract

Human glomerular extract (HGE) was prepared as previously described by Beck et al.²⁷ Briefly, glomeruli were isolated from human kidneys deemed nonsuitable for transplantation (that had been authorized for use in medical research) obtained from New England Donor Services, by graded sieving followed by isolation of glomerular proteins in radioimmunoprecipitation assay (RIPA) buffer (Boston BioProducts). IgG was precleared from tissue lysate by incubation with Protein G Plus agarose beads (Santa Cruz). Only kidneys with <20% global glomerulosclerosis, on routine wedge biopsy, were used for glomerular isolation.

Routine Renal Biopsy Processing

After biopsy acquisition, renal cortex was immediately allocated for light (10% neutral-buffered formalin), immunofluorescence (Zeus transport media), and electron microscopy (Karnovsky’s fixative) processing. For routine clinical immunofluorescence, 4μm cryosections were fixed in 95% ethanol for 10 minutes and incubated with FITC-conjugated polyclonal rabbit F(ab)₂ anti-human IgG antibody (Dako; F0315) diluted 1:20. FITC-conjugated sheep anti-human IgG1, IgG2, IgG3, and IgG4 (binding site: AF006, AF007, AF008, and AF009, respectively) diluted 1:20 were used for IgG subclass evaluation. Albumin was detected using FITC-conjugated polyclonal rabbit anti-human albumin (Dako; F0117) diluted 1:30. Sections were mounted using Dako fluorescence mounting medium (Dako; S3023) with a #1.5 coverslip. Immunofluorescence images were acquired on an Olympus BX53 microscope with an Olympus DP72 camera at 150 ms exposure.

Confocal Microscopy

For confocal microscopy, 4μm cryosections of human kidney biopsies were fixed in 95% ethanol for 10 minutes and subsequently blocked for 1 hour at room temperature (RT) with PBS supplemented with 0.2% fish gelatin, 2% BSA and 2% FBS. All antibodies were diluted in this blocking solution and incubated for 1 hour at RT. Nephrin was detected using 1 μg/ml primary polyclonal sheep anti-human nephrin (R&D Systems; AF4269) followed by a secondary Alexa Fluor^TM 568-conjugated donkey anti-sheep IgG (Invitrogen; A21099). Synaptopodin was detected using anti-synaptopodin (N terminus) guinea pig polyclonal antiserum (Progen; GP94-N) diluted 1:1000 followed by a secondary Alexa Fluor^TM 568-conjugated goat anti-guinea pig IgG (Invitrogen; A11075) antibody. Podocin and Wilms’ tumor 1 (WT1) were detected using a primary polyclonal rabbit anti-human podocin (Millipore Sigma; P0372) and a primary monoclonal rabbit anti-human WT1 clone SC06-41 (Invitrogen; MA5-32215) diluted 1:500 and 1:300, respectively, followed by a secondary Alexa Fluor^TM 568-conjugated donkey anti-rabbit IgG (Invitrogen; A10042). IgG immune deposits were detected using a primary monoclonal mouse anti-human IgG antibody (Abcam; ab200699) diluted 1:750, followed by a secondary Alexa Fluor^TM 488-conjugated donkey anti-mouse IgG (Invitrogen; A21202). All secondary Alexa Fluor^TM-conjugated antibodies were diluted 1:500. Sections were mounted using Vectashield antifade mounting medium (Vectashield, H-1000) with a #1.5 coverslip and images were acquired on a Leica TCS SPE microscope.

Structured Illumination Microscopy

Structured illumination microscopy (SIM) imaging was performed on 4μm fixed, frozen human kidney biopsy sections processed according to the aforementioned protocol for confocal microscopy. All images were collected using an OMX V4 Blaze (GE Healthcare) microscope equipped with three water-cooled PCO.edge sCMOS cameras, 488 nm and 568 nm laser lines, and 528/48 nm and 609/37 nm emission filters (Omega Optical). Images were acquired with a 60×/1.42 Plan-Apochromat objective lens (Olympus) with a final pixel size of 80 nm. Z stacks of 4–8 μm, were acquired with a 0.125 μm z-spacing, and 15 raw images (three rotations with five phases each) were acquired per plane. Spherical aberration was minimized for each sample using immersion oil matching.²⁸ Super-resolution images were computationally reconstructed from the raw data sets with a channel-specific, measured optical transfer function, and a Wiener filter constant of 0.001 using a CUDA-accelerated three-dimensional SIM reconstruction code.²⁹ Axial and lateral chromatic misregistration was determined using a single biologic calculation slide, prepared with human kidney tissue stained with a primary mouse anti-human IgG monoclonal antibody (Abcam; ab200699), followed by both secondary Alexa Fluor^TM 488-conjugated donkey anti-mouse IgG (Invitrogen; A21202) and Alexa Fluor^TM 568-conjugated goat anti-mouse IgG (Invitrogen; A11031) antibodies on the same tissue cryosection. Experimental data sets were then registered using the imwarp function in MATLAB (MathWorks).³⁰

Generation of Recombinant hPLA₂R

Separate plasmids encoding the extracellular subdomains of human nephrin (amino acids 1–1059), comprising the eight Ig-like C2-type domains and a single fibronectin type III domain, and hPLA₂R, comprising the N-terminal ricin domain, fibronectin type II domain, and eight C-type lectin domains, both with C-terminal polyhistidine (6XHIS) tags, were generated by standard cloning techniques. The correct sequences were confirmed by whole plasmid sequencing (MGH DNA core). HEK293-F cells (Thermo Fisher) were transfected with 0.5 μg plasmid per 10⁶ cells using 1.5 μg polyethylenimine. The plasmid and polyethylenimine were pre-incubated for 20 minutes in Freestyle media (Thermo Fisher) at one-tenth the final volume and then added dropwise to the cells. After 3–5 days, provided the cell viability was >95%, the cell culture media was harvested by centrifugation (300 × g for 10 minutes). Imidazole was added to a final concentration of 10 mM and the media was filter sterilized (0.2 μm) on ice. Nickel NTA resin (Qiagen) was washed three times with 10 mM imidazole in PBS and then incubated with the filtered media overnight (O/N) at 4°C on a roller mixer (Thermo Fisher). The nickel NTA resin was then washed three times with 10 mM imidazole in PBS and the recombinant proteins were eluted with 300 mM imidazole in PBS. The purity of the eluted fractions was confirmed by SDS-PAGE with a 4%–12% Bis-Tris gel (Invitrogen), pooled together and concentrated to 1 ml using an Amnicon centrifugation filter with a 10K molecular weight cutoff (Millipore). The resultant protein was run over a Sephadex^ΤΜ 300 column and 0.5-ml fractions were collected. The purity of the eluted fractions was confirmed by SDS-PAGE on a 4%–12% Bis-Tris gel (Invitrogen) and the concentration determined by measuring absorbance at 280 nm using a NanoDrop spectrophotometer (Thermo Fisher). Immunoreactivity of the purified nephrin was confirmed by Western blot analysis, under reducing conditions using a primary sheep anti-human nephrin antibody (R&D Systems) followed by a secondary horseradish peroxidase (HRP)-conjugated donkey anti-sheep IgG antibody (Jackson ImmunoResearch), and of the purified hPLA₂R under nonreducing conditions using serum from a patient with known anti-PLA₂R antibodies (determined by commercial ELISA and IIFT [Euroimmun]) diluted 1:1000 and a secondary HRP-conjugated donkey anti-human IgG antibody (Jackson ImmunoResearch).

ELISA

Nunc MaxiSorp^TM ELISA plates (Thermo Fisher) were coated with 100ng per well of either recombinant extracellular domain of human nephrin or hPLA₂R diluted in coating buffer (BioLegend) and incubated O/N at 4°C. Uncoated control wells were used to determine nonspecific binding (in the absence of antigen) for each patient sample, and this allowed for background subtraction. The plates were washed three times with 300 μl PBS containing 0.05% Tween 20 (PBST). Plates were blocked with 300 μl of SuperBlock (Thermo Fisher) for 1 hour at RT and then incubated O/N at 4°C with 100 μl of patient samples diluted 1:100 in SuperBlock containing 0.1% Tween 20 (SuperT). Samples with an initial high titer were subsequently diluted to 1:200 or 1:400. Plates were washed a further five times with 300 μl of PBST, followed by incubation with 100 μl of biotin-conjugated goat anti-human IgG Fc, highly cross-absorbed antibody (Thermo Fisher) diluted to 0.75 μg/ml in SuperT, shaking at 500 rpm for 1 hour at RT. Plates were washed five times with PBST followed by incubation with 100 μl of HRP-conjugated avidin (BioLegend) diluted to 1:2000 in SuperT, shaking at 500 rpm for 30 minutes at RT. Following five final washes with PBST, 100 μl of TMB substrate (BioLegend) was added and the plates incubated for 10 minutes at RT. Then 100 μl of stop solution (BioLegend) was added and the absorbance at 450 nm was measured. Background subtraction was performed by subtracting the average OD of duplicate uncoated wells from the average OD of duplicate antigen-coated wells for each individual patient sample.³¹ Anti-nephrin antibody titers were then determined using a standard curve derived from a serial two-fold dilution series of a positive patient sample (MCD15+) in which a 1:100 dilution was arbitrarily defined as containing 1000 U/ml.

Immunoprecipitation and Western Blot

One volume of patient serum or plasma was mixed with five volumes of RIPA buffer containing HGE, or 100 ng recombinant extracellular domain of human nephrin, and incubated O/N at 4°C. IgG-antigen complexes were precipitated with Protein G Plus agarose beads (Santa Cruz) for 2 hours at 4°C. The beads were collected by centrifugation and washed three times with Tris-buffered saline supplemented with 0.2% Tween 20 (TBST) and a final wash with distilled water. Proteins were eluted from the beads and denatured under reducing conditions by heating at 95°C for 5 minutes in 1× Laemmli buffer (Bio-Rad) containing 2.5% β-mercaptoethanol. Samples were loaded into precast 7.5% Mini-PROTEAN Tris-glycine gels (Bio-Rad) and electrophoresed at 100 V for 90 minutes in the presence of Novex Tris-Glycine SDS running buffer (Thermo Fisher). Proteins were transferred to polyvinylidene fluoride membranes (EMD Millipore) using the Pierce Power Blotter system (Thermo Fisher) for 10 minutes at 25 V, 1.3 A. Membranes were blocked for 1 hour at RT in TBST containing 5% skimmed milk (w/v) followed by incubation O/N at 4°C with 1 μg/ml polyclonal sheep anti-human nephrin antibody (R&D Systems; AF4269) diluted in TBST with 2% skimmed milk (w/v). All other antibodies were diluted in TBST containing 5% skimmed milk and incubated for 1 hour at RT. Membranes were washed with TBST three times for 5 minutes each, followed by a secondary HRP-conjugated donkey anti-sheep IgG antibody (Jackson ImmunoResearch; 713-035-147) diluted 1:20,000. Human IgG heavy chain was detected using HRP-conjugated donkey anti-human IgG antibody (Jackson ImmunoResearch; 709-035-149) diluted 1:10,000. Finally, membranes were washed three times with TBST and incubated with SuperSignal^TM West Pico PLUS or Femto chemiluminescent substrate (Thermo Fisher) for 3 minutes and images were acquired on a Universal Hood III gel dock system (Bio-Rad).

Results

Circulating Nephrin Autoantibodies Are Present during Active Disease in a Subset of Adults and Children from the NEPTUNE Longitudinal Study Cohort with Biopsy Proven MCD

To first determine whether circulating autoantibodies against nephrin are detectable in the sera of patients with biopsy proven MCD and no underlying genetic basis (lacking known pathogenic variants in established Mendelian nephrotic syndrome genes), we evaluated serum obtained from the Nephrotic Syndrome Study Network (NEPTUNE) longitudinal cohort study²² consisting of 41 (66%) children and 21 (34%) adults (Supplemental Table 1). We developed an indirect ELISA using a recombinant, affinity-purified extracellular domain of human nephrin (hNephrin_G1059) (Supplemental Figure 1) and established a threshold for anti-nephrin antibody positivity, based on the maximum titer in a healthy control population (n=30) (Figure 1A). Evaluation of the earliest serum sample obtained during active disease (UPCR >3 g/g) revealed that 18 (29%) of 62 patients, with an equal number of adults and children, were positive for autoantibodies against nephrin (Figure 1A). Control sera from 53 (98%) of 54 patients who tested positive for anti-hPLA₂R antibodies, by clinically validated ELISA and IIFT assays, were negative for anti-nephrin antibody (Figure 1A).

Figure 1.

Circulating autoantibodies against nephrin are present in a subset of patients with MCD from the NEPTUNE study cohort and correlate with disease activity. (A) Antibodies against the extracellular domain of recombinant human nephrin (hNephrin_G1059) were measured by indirect ELISA. Antigen-specific binding was determined by subtracting the average OD_450nm of duplicate uncoated wells (nonspecific background) from the average OD_450nm of duplicate hNephrin_G1059 coated wells for each individual patient sample. A relative antibody titer was then determined from a standard curve that was generated using a single positive patient sample with a 1:100 dilution defined as containing 1000 U/ml. The threshold for anti-nephrin antibody positivity (187 U/ml) was defined as the maximum antibody titer in a healthy control population (n=30) without kidney disease, as the cohort was not normally distributed and to maximize specificity. The earliest serum sample available during active disease (UPCR >3 g/g on the day of sample collection) was positive for anti-nephrin antibodies in 18 (29%) of 62 patients with biopsy proven MCD from the NEPTUNE cohort. Fifty-three (98%) of 54 nephrotic control patients with anti-hPLA₂R antibodies, as determined by clinical ELISA and IIFT assays (Euroimmun), were negative for anti-nephrin antibodies. The intra- and inter-assay coefficient of variances for the anti-nephrin antibody ELISA were 5.56% and 14.36%, respectively. The antibody titers for the NEPTUNE patients and controls are given in Supplemental Table 1. (B) Eleven of the 18 NEPTUNE patients who were anti-nephrin antibody positive during active disease (blue bar) had a subsequent serum sample available during complete remission (UPCR <0.3 g/g on the day of sample collection) which was anti-nephrin antibody negative (red bar) in all patients. Dotted line indicates threshold for positive antibody titer (187 U/ml). The t test was used to compare differences between the active and remission samples, **P<0.01; ***P<0.001. (C) The same serum samples evaluated by ELISA from the NEPTUNE cohort (B) were evaluated for their ability to immunoprecipitate nephrin from HGE derived from non-diseased human kidnies. In keeping with the ELISA results, only serum obtained during active disease (indicated by an arrow with an “A” [blue colored] above it) immunoprecipitated nephrin, whereas serum obtained during remission (indicated by an arrow with an “R” [red colored] above it) did not. Total IgG was comparable between active and remission samples for each patient.

The patients’ clinical characteristics (Table 1) and the median time from enrollment to complete remission were similar between the anti-nephrin antibody positive and negative groups (4.4 months versus 5.4 months, respectively; P=0.72) (Figure 2). However, the relapse-free period was shorter for the anti-nephrin antibody positive group compared with the antibody negative group, although this finding did not reach conventional levels of statistical significance (median time to relapse 6.0 months versus 21.57 months, respectively; P=0.09) (Figure 2).

Table 1.

Comparison of the clinical and demographic data of patients from the NEPTUNE cohort, grouped based on anti-nephrin antibody status.

	Anti-Nephrin Antibody Positive	Anti-Nephrin Antibody Negative	P value
Patients, n (% of total patients)	18 (29)	44 (71)
Age (yr)	20 (5.5–37.5)	14 (6.25–17.75)	0.11
Male, n (%)	12 (67)	23 (52)	0.40
Female, n (%)	6 (33)	21 (48)	0.40
Disease onset, n (%)
Adulthood	9 (50)	12 (27)	0.14
Childhood (<18 yr)	9 (50)	32 (73)	0.14
Race, n (%)
White	8 (44)	24 (55)	0.58
Black	4 (22)	12 (27)	0.76
Asian	5 (28)	2 (4.5)	0.02*
Multi	1 (6)	4 (9)	1.00
Unknown	0 (0)	2 (4.5)	1.00
Hispanic/Latino, n (%)	5 (28)	9 (20)	0.51
No remission, n (%)	0 (0)	2 (4.5)	1.00
Partial remission, n (%)	3 (17)	2 (4.5)	0.14
Complete remission (CR), n (%)	15 (83)	40 (91)	0.40
Relapse after CR, n (% of CR)	9 (60)	25 (63)	1.00
UPCR (g/g) at anti-nephrin antibody assay	9.1 (5.4–13.79)	7.6 (5.20–10.48)	0.34
Peak serum creatinine (mg/dl)	0.89 (0.58–1.6)	0.7 (0.52–1.02)	0.17

The threshold for a positive anti-nephrin antibody (α-nephrin antibody) level was based on a randomly selected, healthy control population with no renal disease. Complete remission was defined as a UPCR <0.3 g/g. Partial remission was defined as a >50% reduction in proteinuria that did not fall below 0.3 g/g. The continuous variables are presented as median (interquartile range). Statistical analysis was performed using the Mann–Whitney U test for continuous variables and Fisher’s exact test for categorical variables (*P<0.05).

Figure 2.

Kaplan–Meier estimates for complete remission (CR) and relapse-free period in the NEPTUNE cohort. (A) Time to CR from enrollment was similar between the anti-nephrin antibody (Ab) positive (median time 4.4 months) and negative groups (median time 5.4 months) by a log-rank test (P=0.72). (B) The relapse-free period following CR was shorter in the anti-nephrin antibody positive group (median time 6 months) compared with the antibody negative group (median time 21.57 months), although this finding did not reach statistical significance by a log-rank test (P=0.09). CR was defined as UPCR <0.3 g/g. Relapse when UPCR >3 g/g after reaching CR.

Circulating Nephrin Autoantibodies Are Reduced or Completely Absent during Treatment Response

A subsequent serum sample was available during either complete (UPCR <0.3 g/g) or partial (>50% reduction in proteinuria) remission from 12 of the 18 anti-nephrin antibody positive patients, in whom we observed a complete absence or a significant reduction of nephrin autoantibodies, respectively (Figure 1B, Supplemental Figures 2 and 3). In keeping with the ELISA results, only serum obtained during active disease or partial remission immunoprecipitated nephrin from healthy donor kidney-derived HGE, whereas serum obtained during complete remission did not (Figure 1C).

Podocyte-Associated Punctate IgG Is Present in a Subset of MCD Biopsies and Specifically Colocalizes with Nephrin

To further investigate a potential pathogenic role of these nephrin autoantibodies, we next sought to establish whether they are present within kidneys of patients with MCD. One limitation of the NEPTUNE cohort is that biopsy material from these patients was not available for further evaluation and so we turned to our own institution and collaborators for biopsy and serum samples.

For many years, we have observed a delicate punctate staining for IgG in a subset of patients with MCD (MCD+) by routine immunofluorescence staining that is distinct from the background (Supplemental Figure 4A). It is much more subtle when compared with the prominent IgG staining observed in membranous nephropathy (Supplemental Figure 4A), and although this feature has been previously described,¹³ its significance has not been fully established. We therefore hypothesized that this subtle IgG may represent autoantibodies targeting nephrin. To limit the possibility of staining artifacts, we routinely use a directly conjugated FITC anti-human IgG F(ab)₂ antibody which we have independently validated with a distinct unconjugated anti-human IgG antibody, an anti–light chain antibody, and isotype-specific anti-IgG antibodies that all show an identical staining pattern (Figure 3). Importantly, we observe a complete lack of concurrent glomerular albumin staining, indicating that this feature is IgG selective and does not reflect nonspecific protein resorption (Figure 3).

Figure 3.

Routine clinical epifluorescence microscopy, periodic-acid Schiff (PAS) stained light microscopy and electron microscopy (EM) images of IgG-positive MCD (MCD7+). Top row: The diffuse punctate staining seen with FITC-conjugated IgG antibody is not seen by albumin staining. Note the positive albumin staining in proximal tubule reabsorption droplets. Middle row: Staining for IgG subtypes confirms no restriction of the punctate staining, and, in this case, more immunoreactivity for IgG1 and IgG2 compared with IgG3 and IgG4 subtypes was observed. Bottom row: Staining for κ and λ light chains shows equal intensity and an appearance mirroring the IgG staining. PAS staining shows minimal light microscopic changes and EM demonstrates diffuse podocyte foot process effacement. Scale bars: Immunofluorescence and PAS: 20 μm. Scale bar: EM: 1 μm.

We utilized confocal microscopy to further evaluate this punctate IgG in renal biopsies that were received and processed by us over the last 3 years. We observed two predominant patterns of IgG distribution: glomerular basement membrane–associated fine punctate or curvilinear structures and more apically located punctate and vaguely vesicular clusters with the latter being more common. These disparate staining patterns may reflect different stages of antibody binding and/or redistribution. In all the MCD+ biopsies evaluated, we observed specific colocalization of nephrin with the punctate IgG and not the background (Figure 4A, Supplemental Figure 4B, Supplemental Figure 5), which was further corroborated in the control biopsies lacking this punctate IgG (Figure 4A, Supplemental Figure 4B). Antigen specificity was evidenced by a clear spatial association of the IgG with the podocyte slit diaphragm–associated nephrin but not with the podocyte foot process–associated synaptopodin by confocal microscopy (Figure 4, A and B, Supplemental Figure 4, B and C) and by super-resolution SIM which achieves an even higher spatial resolution (Figure 4, C and D, Supplemental Figure 6).²⁸ Furthermore, in those biopsies exhibiting the granular redistribution of nephrin away from the slit diaphragm, as previously described in MCD,^19,20 the IgG did not colocalize with the three intracellular podocyte specific proteins: synaptopodin (foot process–associated), podocin (slit diaphragm–associated), and WT1 (nuclear) (Supplemental Figure 7).

Figure 4.

Punctate IgG present in a subset of MCD renal biopsies colocalizes with the critical podocyte slit diaphragm protein nephrin. (A) Representative confocal microscopy images of glomeruli in IgG-positive MCD (MCD1+/MCD7+) and IgG-negative MCD (MCD5−), stained for IgG (green) and the podocyte slit diaphragm protein nephrin (red). There is a clear overlap (yellow) of the punctate IgG (not the background) with nephrin in the IgG-positive MCD+ biopsies (white arrows) but not in IgG-negative MCD biopsies.The right panel shows magified images of the boxed areas for each corresponding image in the left panel. Scale bar: 10 µm. (B) Representative confocal microscopy images of glomeruli in IgG-positive MCD (MCD1+/MCD7+) and IgG-negative MCD (MCD5−), stained for IgG (green) and the podocyte cytoskeletal protein synaptopodin (red). There is no discernable overlap between the punctate IgG (indicated by white arrows) and synaptopodin in any of those patients. Magnified images for each boxed area in the biopsies are shown in the corresponding panel on the right. . Scale bar: 10 µm. (C) Super-resolution SIM images of 0.125µm individual Z-slices showing en face views of the podocyte junction from a representative IgG-positive MCD renal biopsy (MCD1+) in which the nephrin remains glomerular basement membrane–associated, forming a curvilinear pattern. The left image shows colocalization (yellow) of IgG (green) with the slit diaphragm protein nephrin (red), in contrast to mutual exclusivity with the foot process–associated synaptopodin (red) shown in the right image, indicating intimate spatial association with nephrin along the podocyte slit diaphragm. (Full image stack is shown in Supplemental Figure 6.) Scale bar: 1 µm. (D) Super-resolution SIM images of 0.125µm individual Z-slices from a representative IgG-positive MCD renal biopsy (MCD7+) in which nephrin is redistributed to a more granular pattern. The left image shows colocalization (yellow) of IgG (green) with the slit diaphragm protein nephrin (red), in contrast to mutual exclusivity with the foot process–associated synaptopodin (red) shown in the right image, indicating a continued close spatial association of the IgG with the redistributed nephrin. (Full image stack is shown in Supplemental Figure 6.) Scale bar: 1 µm.

Circulating Autoantibodies to Nephrin Are Present in Patients with MCD and Punctate IgG on Renal Biopsy (IgG-Positive MCD)

To confirm that patients with MCD and punctate IgG on renal biopsy (MCD+) do indeed have circulating autoantibodies against nephrin, we evaluated serum or plasma that was available specifically during active disease for nine of them. As expected, all nine patients were serologically positive for anti-nephrin antibodies by ELISA, in contrast to 12 control patients lacking punctate IgG on renal biopsy, who were all serologically negative (Figure 5A, Supplemental Table 2). A follow-up serum or plasma sample was available for four of the nine MCD+ patients, which showed a significant reduction in antibody titer concordant with treatment response (Figure 5B, Supplemental Figure 8). These findings were corroborated by immunoprecipitation (Supplemental Figure 9), and for all the patients in this study who were serologically positive for circulating nephrin autoantibodies they did not cross react with human PLA₂R (Supplemental Figure 10).

Figure 5.

Circulating autoantibodies against nephrin are exclusively present in patients with renal biopsy IgG-positive MCD. (A) All of the patients with MCD and IgG deposition on biopsy (n=9) were serologically anti-nephrin antibody positive, whereas all of the control subjects lacking IgG deposition on biopsy, consisting of diabetic nephropathy (n=2), amyloidosis (n=1), IgG-negative FSGS (n=2), IgG-negative TL (n=1), normal (n=1), disease-free region of tumor nephrectomy (n=2), and IgG-negative MCD (n=3), were serologically anti-nephrin antibody negative (n=12). The Mann–Whitney U test was used to compare differences between the groups, ***P<0.001. (B) Serum/plasma samples were obtained from patients with biopsy proven IgG-positive MCD (MCD+) during active disease (within 7 days of presentation with nephrotic syndrome) and follow-up samples were obtained during complete (MCD4+, MCD7+) or partial (MCD8+) remission on the day of sample collection. For MCD3+, the follow-up serum sample was obtained approximately 3 weeks after entering a period of sustained complete remission. The threshold for a positive anti-nephrin antibody titer of 187 U/ml (indicated by dotted line) was based on the upper limit of a healthy control population. Anti-nephrin antibodies in serum/plasma were undetectable or significantly reduced to below the threshold for positivity (red bar) during clinical remission compared with those during active disease (blue bar). Complete remission was defined as UPCR <0.3 g/g or urinary albumin creatinine ratio <0.2 g/g. Partial remission was defined as a >50% reduction in proteinuria (UPCR) that did not fall below 0.3 g/g. The t test was used to compare differences between the active and remission samples, **P<0.01; ***P<0.001.

Pretransplant Nephrin Autoantibodies Are Associated with Massive Proteinuria Recurrence in the Allograft

Finally, to highlight a potential role of pretransplant nephrin autoantibodies in post-transplant disease recurrence, which generally shows morphologic features indistinguishable from MCD, we identified a 27-year-old patient with childhood onset, steroid-dependent MCD and no underlying genetic basis (as determined by clinical whole exome sequencing) who progressed to ESKD requiring kidney transplantation (detailed clinical history is given in Supplemental Clinical Case). In keeping with a potential pathogenic role for anti-nephrin autoantibodies, she developed massive proteinuria early post-transplant, that in contrast to CNF was associated with high pretransplant levels of nephrin autoantibodies (Figure 6).^17,18 Considering a dilution factor of 0.67 of the initial plasmapheresate relative to pre-pheresis patient plasma,³² detection of anti-nephrin autoantibodies above the threshold in the plasmapheresate indicated presence of circulating anti-nephrin antibody in the patient at the time of proteinuria recurrence (Figure 6). Having reached sustained remission, no circulating nephrin autoantibodies were detectable in this patient 1 year post-transplant (Figure 6).

Figure 6.

Pretransplant nephrin autoantibodies are associated with early post-transplant massive proteinuria recurrence in a patient with steroid-dependent MCD that progressed to ESKD. (A) Clinical course of patient with childhood onset, steroid-dependent MCD that progressed to ESKD with subsequent biopsies showing FSGS. She developed early, massive proteinuria recurrence, following a cadaveric, pediatric donor kidney transplant that rapidly responded to treatment with plasmapheresis (purple lines) and rituximab (green lines) shown above the graph. Full clinical details are available in the Supplemental Clinical Case. (B) Two pretransplant serum samples (day −617, day −528) and the initial plasmapheresate (shaded box) on day 13 (d13 (PP), indicated by arrow in (A)) tested positive by ELISA for anti-nephrin antibodies (dotted line indicates threshold for positivity of 187 U/ml). Serum samples evaluated following treatment response at day 27 (d27, indicated by arrow in (A)) and day 365 (UPCR <0.3 g/g) were negative for anti-nephrin antibodies. (C) Similarly, the pretransplant serum and plasmapheresate immunoprecipitated nephrin from HGE (derived from disease-free human kidneys) in keeping with the ELISA findings.

Discussion

MCD is a podocytopathy of unknown etiology, affecting both adults and children, with evidence supporting a role for B cells.^9–12 In this study, we report the novel discovery of autoantibodies in a subset of patients with noncongenital, childhood and adult-onset MCD that target the fundamental structural slit diaphragm component nephrin. This important observation, together with previous work demonstrating that anti-nephrin antibodies directly interfere with nephrin homophilic interactions in vitro³ and cause proteinuria in rodent models, associated with an identical redistribution of nephrin to that observed in MCD,^14,15 supports a model of nephrin autoantibody-mediated MCD that shares striking similarities to pemphigus. In this blistering skin condition, autoantibodies target dsgs, the fundamental structural proteins of the desmosomal cell adhesion complex that links adjacent keratinocytes,²¹ and is analogous to nephrin in the specialized slit diaphragm junctional complex between adjacent podocytes. These dsg autoantibodies directly interfere with cell adhesion through redistribution, clustering, and endocytosis of dsgs, which disrupts the integrity of the desmosome.²¹ Furthermore, pemphigus exhibits a rapid response (within days to weeks) to glucocorticoid treatment that cannot be explained by reduced IgG synthesis alone and may be due to compensatory dsg synthesis in keratinocytes.³³ Similarly, most patients with MCD respond rapidly (within weeks) to glucocorticoids, which have also been shown to upregulate nephrin cell surface expression in cultured human podocytes.³⁴ Based on these observations we speculate that the subtle punctate IgG that specifically colocalizes with nephrin in the MCD kidney biopsies and whose significance has not been fully appreciated, represents in situ binding of nephrin autoantibodies. This targeted binding may be sufficient to disrupt nephrin homophilic interactions leading to early loss of slit diaphragm integrity, and the redistribution of IgG along with its target nephrin may explain this subtle punctate staining pattern, in contrast to the much more intense staining seen with membranous nephropathy labeling large IgG immune complex aggregates that progressively accumulate along the base of the podocyte foot processes.

Fortunately, progression of MCD to ESKD is rare; however, in those patients that do progress or in those with an initial diagnosis of primary FSGS that progresses more commonly, the disease can rapidly recur in the allograft. A role for nephrin autoantibodies in early post-transplant massive proteinuria is illustrated by our patient with steroid-dependent MCD that eventually progressed to ESKD with subsequent biopsies showing FSGS. Early post-transplant massive proteinuria occurred in the presence of circulating nephrin autoantibodies identified both prior to transplantation and at the time of disease recurrence, which were successfully treated with plasmapheresis/rituximab, leading to sustained remission associated with their disappearance. These findings are entirely in keeping with previous studies in CNF patients who develop alloantibodies to nephrin in association with disease recurrence in the allograft and respond to plasmapheresis/rituximab.^18,35 Importantly, the critical distinction is that in our patient, the nephrin autoantibodies were present both prior to the transplant and at the time of disease recurrence whereas in CNF they arise as a direct consequence of the transplant because of alloimmunization to nephrin.¹⁸ FSGS and MCD share some important similarities, such as indistinguishable ultrastructural changes and response to B cell therapies,³⁶ and together with this illustrative case, we would speculate that our findings of nephrin autoantibodies may also extend to a subset of patients with a diagnosis of primary nongenetic FSGS.

We identified circulating nephrin autoantibodies in almost one-third of patients from the NEPTUNE cohort with active proteinuria at the time of sampling. However, this is likely to be an underestimate of the prevalence of nephrin autoantibody MCD in this cohort for two reasons: (1) we used a highly stringent threshold based on the maximum titer in a healthy control population to avoid false positives and because the population was not normally distributed, and (2) essentially all the patients (97%) had initiated therapy prior to the earliest available serum sample and this may have been sufficient to reduce the titer to just below the threshold for positivity in some of them. Nonetheless, our discovery highlights the heterogeneity of MCD and allows us to propose a new molecular classification based on the presence of this biomarker. This will serve as the foundation for future prospective studies that will be essential to establish both the prognostic significance of these new molecularly defined subgroups of MCD and to investigate the etiology in those patients who do not have nephrin autoantibodies.

In conclusion, our discovery provides valuable insights into the etiology of a disease that has been poorly understood for decades. Although we do not provide direct mechanistic proof in this observational study, our findings are entirely in keeping with the established pathogenic effect of anti-nephrin antibodies in animal models and support an autoimmune etiology in a subset of MCD patients. Although the prognostic and therapeutic significance of nephrin autoantibodies in MCD remains to be established in future prospective studies, our findings provide a mechanistic rationale for the consideration of B cell–targeted therapy and allows us to molecularly define those patients who stand to benefit most from the development of new therapeutic strategies for nephrin autoantibody MCD.

Disclosures

L. Beck reports consultancy agreements with Alexion and Novartis; has research funding with Pfizer; has honoraria with UpToDate, Inc.; is a co-inventor on and receives royalties related to the US patent “Diagnostics for Membranous Nephropathy”; is scientific advisor or has membership with Kidney International Reports editorial board and Kidney Medicine editorial board; and has other interests/relationships with advisory boards for Alexion, Ionis Pharmaceuticals, Novartis, and Visterra. A. Chandraker reports consultancy agreements with Allovir, eGenesis, Immucor, Natera, and Shire; has research funding with Allovir, Amgen, CSL, Hansa, and Natera; has honoraria with Natera; and is scientific advisor or has membership with American Society of Transplantation (development chair and Transplant Therapeutics Consortium). D. Fermin reports ownership interest with 10X Genomics, AstraZeneca, and Moderna. A. Greka has financial interest in Goldfinch Biopharma, which was reviewed and is managed by Brigham and Women’s Hospital, Mass General Brigham, the Broad Institute, and Harvard University in accordance with their conflict of interest policies; reports consultancy agreements with Biogen, Blackstone, Bristol-Myers Squibb, Cyclerion Pharmaceuticals, Kyn Therapeutics, Lycia Therapeutics, Maze Therapeutics, MPM, Novartis, Ono Pharmaceuticals, and Q32Bio; and has other interests/relationships with NephCure Kidney International and Rare Kidney Disease Foundation. J. Henderson reports consultancy agreements and research funding from Pfizer. J. Henderson, K. Keller, H. Rennke, A. Watts, and A. Weins have filed a patent “Methods for identifying and treating patients with antibody-mediated acquired primary or recurrent idiopathic nephrotic syndrome.” H. Rennke reports consultancy agreements with Ionis Pharmaceuticals; and has honoraria with Wolters Kluwer royalties book: H. Rennke and B. Denker, Renal Pathophysiology: The Essentials, 5th ed. L. Riella reports research funding with Bristol-Myers Squibb, CareDx, Natera, and Visterra; has honoraria with CareDx; and is scientific advisor or has membership with CareDx. I. Rosales reports consultancy agreements with eGenesis; is a scientific advisor or has membership with eGenesis and Philippine Journal of Pathology; and has other interests/relationships via support from the International Society of Nephrology fellowship program. M. Sampson reports consultancy agreements with Janssen Pharmaceuticals and Maze Therapeutics; and is scientific advisor or has membership with Natera. S. Waikar reports consultancy agreements with Allena, BioMarin, CVS, GlaxoSmithKline, Johnson & Johnson, Mallinckrodt, Mass Medical International, Metro Biotechnology, Oxidien, Pfizer, Regeneron, Roth Capital Partners, Sironax, Strataca/3ive, Venbio, and Wolters Kluwer; has research funding with Vertex; is a scientific advisor or has membership with Kantum (scientific advisory board); and has other interests/relationships as expert witness for litigation related to GE Healthcare product Omniscan, expert witness for litigation related to Fresenius product Granuflo, expert witness for litigation involving cisplatin toxicity, expert witness for litigation related to Gilead product tenofovir, and expert witness for litigation related to DaVita laboratory testing. A. Weins reports consultancy agreements with Expansion Technologies, Inc. and Goldfinch Biopharma. All remaining authors have nothing to disclose. A. Greka is supported by National Institutes of Health (NIH) grants DK099465 and DK095045. M. Sampson is supported by NIH grants R01DK119380 and R01DK1088085. S. Waikar is supported by NIH grants U01DK085660, U01DK104308, and UG3DK114915.

Funding

A. Watts was supported by the National Institutes of Health T32 training fellowships T32HL007627 and T32DK007527. A. Weins was supported by Harvard Medical School’s Eleanor Miles Shore Fellowship and is the recipient of a Nephcure/Kidney International NEPTUNE/CureGN ancillary study award. G. Lerner was supported by the National Institutes of Health grant DK007053-45S1. The National Institutes of Health Nephrotic Syndrome Study Network Consortium (NEPTUNE), grant U54-DK-083912, is a part of the Rare Disease Clinical Research Network, supported through a collaboration between the Office of Rare Diseases Research, the National Center for Advancing Translational Sciences, and the National Institute of Diabetes and Digestive and Kidney Diseases. Additional funding and/or programmatic support for this project has also been provided by the University of Michigan, the NephCure Kidney International, and the Halpin Foundation.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2021060794/-/DCSupplemental.

Supplemental Figure 1. Generation and validation of recombinant human nephrin and PLA₂R.

Supplemental Figure 2. Clinical course of nephrin autoantibody positive patients from the NEPTUNE cohort.

Supplemental Figure 3. Clinical/Serological data for a single patient from the NEPTUNE cohort with partial response.

Supplemental Figure 4. Immunofluorescence staining for IgG alone or together with nephrin or synaptopodin in renal biopsies.

Supplemental Figure 5. Colocalization of IgG and nephrin by immunofluorescence in biopsy IgG-positive MCD (MCD+) cases.

Supplemental Figure 6. Super-resolution SIM showing the two patterns of anti-nephrin IgG distribution in MCD+.

Supplemental Figure 7. Colocalization of the IgG with nephrin but not with other podocytes markers in MCD+.

Supplemental Figure 8. Clinical course of the MCD+ patients with circulating nephrin autoantibodies.

Supplemental Figure 9. Serum from nephrin autoantibody positive patients immunoprecipitates recombinant nephrin or nephrin from HGE.

Supplemental Figure 10. Antibodies to nephrin in patients with MCD do not cross react with hPLA₂R.

Supplemental Table 1. Clinical characteristics of the NEPTUNE and control patients evaluated for circulating nephrin autoantibodies.

Supplemental Table 2. Clinical characteristics of the cohort from our institutions evaluated for punctate IgG in the renal biopsies together with nephrin autoantibodies in sera where available.

Supplemental Clinical Case. Description of a single patient with steroid-dependent MCD progression to ESKD and massive proteinuria recurrence post-transplant associated with pretransplant nephrin autoantibodies shown in main Figure 6.