In this study 41 consecutive patients with biopsy-proven THSD7A-associated MN were included. The PLA₂R1-associated MN cohort included 76 patients out of a 302-patients prospective cohort, which we have reported before (12). Since the major objective of the present study was to investigate patients with malignancy-associated MN, we included in this study all 16 (5.3%) out of the 302 PLA₂R1 antibody (PLA₂R1-ab) positive patients, who were diagnosed with a malignancy within 24 months of the MN diagnosis. In addition, we included 60 PLA₂R1-ab positive patients based on the following criteria: (a) 20 patients with persistent PLA₂R1-ab during follow-up, (b) 20 patients who became negative for PLA₂R1-ab during follow-up and had no relapse, and (c) 20 patients who became PLA₂R1-ab negative, but developed a relapse of PLA₂R1-ab during follow-up.

The study was approved by the local ethics committee of the chamber of physicians in Hamburg and was conducted in accordance with the ethical principles stated by the Declaration of Helsinki. An informed consent was obtained from all participating patients.

IgG subclass levels (IgG1-4) were measured using the SPAplus Assay from Bindingsite (Product Codes: IgG1 NK006.S, IgG2 NK007.S, IgG3 LK008.S, and IgG4 LK009.S) according to the manufacturer's instructions. Antigen specific pan-IgG and IgG4 were quantified by immunofluorescence test (IFT) as decribed before (12). A monoclonal anti-human IgG4–FITC antibody produced in mouse was used as secondary antibody for measurement of antigen-specific IgG4 (Sigma-Aldrich, F9890). In all patients PLA₂R1-ab were also measured by ELISA as described before (12).

Healthy parts of kidneys from patients who underwent nephrectomy because of a renal tumor were used for the preparation of the human glomerular extracts (HGE) as described before (2). Preparation of lung tissue was performed according to a similar procedure. Healthy parts of lung tissue, which had to be removed during tumor surgery, were used for the preparation of lung tissue extract (LTE). LTE was gathered after sieving of lung tissue using a 150 μm mesh. The tissue fraction >150 μm was homogenized using a Mixer Mill MM400 (Retsch, Haan, Germany) and used for further analyses, while the flow-through was collected to confirm the absence of the antigens. Membrane fractions of both glomeruli and lung tissue were solubilized in a RIPA buffer and contaminating immunoglobulin G was removed using Protein G resin, according to the manufacturer's protocol (GenScript, Piscataway, NJ, USA). Samples for recombinant protein, LTE or HGE were each gathered from one pool of protein/tissue for all performed experiments in the study.

Western blot analyses were largely performed according to a protocol previously described (12), representative blots are shown in Supplemental Figure 1. In detail, for detection of PLA₂R1- and THSD7A-ab by Western Blot, we used the respective recombinant proteins, HGE and LTE. Recombinant PLA₂R1 and THSD7A were produced in HEK293 cells upon transfection with the respective cDNA as described before (2). Western blot lanes were loaded with the following protein samples: 2 μg of HEK293 cell lysate containing recombinant PLA₂R1, 0.5 μg HEK293 cell lysate containing recombinant THSD7A, 1 μg of HGE and 6 μg of LTE per Western blot lane. Proteins were separated by electrophoresis in SDS-gels containing 5% acrylamide under non-reducing conditions, unless otherwise specified. Proteins were then transferred to methanol-activated PVDF membranes under semi-dry conditions using Transblot Turbo (Bio-Rad) at 25 V constant voltage for 25 min. Membranes were blocked in 5% dry milk with PBS-Tween (PBS-T, 0.1%) for 1 h at RT and then incubated in primary antibodies at 4°C overnight. Unless otherwise specified, all sera were diluted 1:100 in 0.5% dry milk in PBS-T. For specific detection of the PLA₂R1 and THSD7A antigens in the positive control of Western blot analyses the following commercially available antibodies were used: rabbit polyclonal antibody at a 1:1000 dilution in 0.5% dry milk in PBS-T (Atlas Antibodies) for PLA₂R1; rabbit polyclonal antibody at a 1:1,000 dilution in 0.5% dry milk in PBS-T (Atlas Antibodies) for THSD7A. A horseradish peroxidase–conjugated goat anti-rabbit IgG (Sigma-Aldrich, #A9169) was used as secondary antibody at a 1:20,000 dilution. For detection of subclass-specific IgG in Western blot analyses, horseradish peroxidase-conjugated monoclonal anti-human IgG1 (9054-05), IgG2 (9060-05), IgG3 (9210-05), IgG4 (9200-05), and IgG (9040-05) antibodies from mouse were used (IgG1, IgG2, and IgG3 at a 1:5,000 dilution; IgG4 at a 1:10,000 dilution; IgG at a 1:20,000 dilution, all SouthernBiotech, Birmingham, USA), diluted in 5% dry milk in PBS-T.

HGE, LTE and the respective recombinant proteins were deglycosylated using neuraminidase, N-glycopeptidase F and O-glycopeptidase (Roche Diagnostics, Mannheim, Germany). One μL of the enzyme solution was added to approximately 20 μl of tissue extract or recombinant protein and samples were incubated for 3 h at 37°C and subsequently analyzed by Western blot.

For immunofluorescence staining of PLA₂R1 and THSD7A 1 μm paraffin sections of kidney or lung tissue were deparaffinized and rehydrated with water. Antigen retrieval was obtained by boiling in citrate buffer pH 6.1 (both 15 min at constant 120°C). Unspecific binding was blocked with 5% horse serum (Vector) in phosphate-buffered saline (PBS) for 30 min at RT prior to incubation at 4°C o/n with primary antibodies (PLA₂R1: Atlas, rabbit primary antibody, 1:100; THSD7A: Atlas, mouse primary antibody, 1:100) in blocking buffer. Staining was visualized with fluorochrome-conjugated secondary antibodies (all affinity-purified from Jackson Immunoresearch Laboratories, 1:200) for 10 min RT in 2.5% horse serum in PBS. Nuclei were counterstained with DRAQ5 (Molecular Probes, 1:4,000). All negative controls were performed by omitting primary antibodies. Staining was evaluated with a confocal LSM 800 meta microscope using the LSM software (all Zeiss).

Remission of proteinuria was defined as proteinuria of <3.5 g/24 h and at least 50% reduction from the time of inclusion. Complete remission of proteinuria was defined as proteinuria <0.5 g/24 h. Continuous data are given as median and 1st−3rd quartile (Q1–Q3), categorical data as counts and percentages (%). Mann-Whitney U tests were performed to assess group differences for continuous and ordinal variables. For analyses of dichotomous data, Fisher's exact tests were performed. Statistical significance was defined as p < 0.05, two-tailed. Glomerular lesions were assessed by electron microscopy according to the classification of Ehrenreich and Churg (17).

We included 41 consecutive patients with THSD7A-associated MN and 76 patients with PLA₂R1-associated MN in the study (Table 1 and Supplemental Table 1). Data from some of those patients have been published in an earlier manuscript (12). Patients with THSD7A-associated MN were older than patients with PLA₂R1-associated MN. Advanced stages of glomerular lesions were more common in patients with PLA₂R1-associated MN. Differences in other clinical characteristics between the two patient groups were not statistically significant.

In 30 (73%) patients with THSD7A-associated MN study enrolment and first serum collection for THSD7A-antibody (THSD7A-ab) measurement were performed within 6 months of renal biopsy. In 30 (73%) patients no immunosuppressive treatment was started prior to study enrolment. In seven (17%) patients with THSD7A-associated MN a malignant disease was diagnosed within 2 years of diagnosis of MN (Tables 2, 3 and Supplemental Table 2). The total IgG4 levels were higher in patients with malignancy, compared to patients with no malignancy (0.70 vs. 0.25 g/l), however, the difference did not reach statistical significance (p = 0.06) and, moreover, the THSD7A-specific IgG4 levels were not different between the two groups (p = 0.65). Otherwise there were no statistically significant differences in the clinical or morphological characteristics of patients with THSD7A-associated MN with or without malignancy. For 25 patients follow-up sera for THSD7A-ab analyses were available. THSD7A-ab became negative in 10 (40%) patients and persisted in 15 (60%) patients (Supplemental Table 3). In patients who became negative for THSD7A-ab immunosuppression was more often used and remission of proteinuria was more common, but the difference was not statistically significant (p = 0.09). Doubling of serum creatinine levels occurred at the same rate in both groups, although patients with persistent THSD7A-ab had a better preserved renal function at baseline (not statistically significant, p = 0.09). Levels of all IgG subclasses were lower in patients with persistent antibodies, however, none of the differences was statistically significant.

All 76 patients with PLA₂R1-associated MN had a newly diagnosed, biopsy-proven MN and were enrolled in a prospective cohort of patients with PLA₂R1-associated MN. Study enrolment and the first serum collection for PLA₂R1-antibody (PLA₂R1-ab) measurement were performed within 6 months of renal biopsy. None of these patients had received an immunosuppressive treatment prior to study enrolment.

The aim of the present study was the identification of subclass differences in malignancy-associated MN compared to primary MN. To this end, all 16 (5.3%) patients, who had a diagnosis of malignancy made within 2 years of MN diagnosis and who participated in the aforementioned prospective cohort of 302 patients with PLA₂R1-associated MN, were included in this analysis. Patients with PLA₂R1-associated MN and no malignancy were significantly younger, had better preserved renal function and had less tubulointerstitial fibrosis compared to PLA₂R1-ab positive patients with malignancy (Table 2 and Supplemental Table 2). Proteinuria was higher in patients with malignancy (not statistically significant, p = 0.19), while neither total-IgG PLA₂R1-ab levels nor IgG4 subclass PLA₂R1-ab significantly differed between patients with and those without malignancy.

In addition to the 16 PLA₂R1-ab positive patients with malignancy, we included further 60 patients from our prospective cohort of patients with PLA₂R1-associated MN in this study. These patients were chosen to reflect a balanced selection of the cohort regarding their development of antibody level during follow-up: (a) 20 patients with persistent PLA₂R1-ab throughout the study follow-up, (b) 20 patients with stable remission of PLA₂R1-ab during follow-up (no relapse), and (c) 20 patients with PLA₂R1-ab relapse.

Patients who became negative for PLA₂R1-ab during follow-up achieved significantly more often a remission of proteinuria (p < 0.01, Supplemental Table 3). They also had 48% lower antibody levels at the beginning of the study, however, this difference was not statistically significant (p = 0.12). At baseline we found no statistically significant differences in any of the clinical characteristics between patients, in whom PLA₂R1-ab relapsed during follow-up, and patients who remained PLA₂R1-ab negative (Supplemental Table 4). While the rate of initial remission of proteinuria (upon negativity of PLA₂R1-ab) was not different between the two groups, patients in whom PLA₂R1-ab relapsed received more often immunosuppressive treatment and had more often a relapse of proteinuria during follow-up (p < 0.01 for both).

At baseline all patients had antigen-specific IgG4 subclass antibodies directed against PLA₂R1 or THSD7A (Figure 1). Antigen-specific IgG3 subclass was detected in almost all PLA₂R1-ab positive patients (99%), followed by IgG1 (75%) and IgG2 subclass (28%). In THSD7A-ab positive patients antigen-specific IgG3 and IgG1 subclass were present in the same number of patients (66%), while IgG2 was less common and found in 22% of the THSD7A-associated cases. IgG subclass patterns did not differ significantly between THSD7A-ab positive patients with or without immunosuppressive treatment prior to study inclusion.

We did not detect any statistically significant differences in antigen-specific IgG subclasses between primary and malignancy-associated MN (Table 2, Supplemental Table 2, Figures 1B,C, 2). Antigen-specific antibodies of IgG4 subclass were found in all patients in both groups, irrespective of the antigen. In PLA₂R1-ab positive patients, antigen-specific IgG1, IgG2, and IgG3 antibodies were positive in 72 vs. 88%, 28 vs. 25%, and 100 vs. 94% of patients with primary vs. malignancy-associated MN, respectively (Figure 1B). In THSD7A-ab positive patients antigen-specific IgG1, IgG2, and IgG3 were found in 65, 21, and 62% of cases with primary MN vs. 71, 29, and 86% of cases with malignancy-associated MN, respectively (Figure 1C). Seventeen out of the 23 patients with MN and malignancy received an immunosuppressive treatment in addition to the treatment of malignancy. Ten of these patients had a remission of proteinuria (five complete remission and five partial remission), five patients had no remission of proteinuria and two patients died during follow-up. Therefore, conclusive evidence on the effect of tumor treatment on disease development is not possible. Three out of six patients who received no immunosuppressive treatment had a remission of proteinuria (two complete remission, one partial remission), one patient had no remission while no follow-up data were available for the remaining two patients.

We next grouped patients with PLA₂R1- or THSD7A-associated MN depending on the clinical outcome (antibody negative during follow-up, remission of proteinuria, doubling of serum creatinine) and analyzed the respective IgG subclass patterns at baseline. There were no statistically significant differences in the IgG subclass patterns for PLA₂R1-ab (Figure 1D) or THSD7A-ab (Figure 1E) at baseline, which might discriminate patients with a favorable from those with an unfavorable outcome.

Since IgG4 is assumed to play a central pathogenic role in primary MN and to differentiate primary from malignancy-associated MN, we deeper analyzed the IgG4-specific findings in our cohort. 100% of the patients with PLA₂R1- or THSD7A-associated MN showed presence of antigen-specific IgG4 antibodies, irrespective of the presumed pathogenesis. We next measured the antigen-specific IgG4 levels in all patients and found no statistically significant difference between the antigen-specific pan-IgG, or IgG4 levels in patients with malignancy and those with no malignancy (Tables 1, 2). Furthermore, serum levels of all IgG subclasses were not statistically different between these patient groups.

In 20 patients with PLA₂R1-associated MN and 15 patients with THSD7A-associated MN autoantibodies persisted (antibodies were never negative in IFT or ELISA analysis) throughout the median study follow-up period of 15 months (Q1–Q3: 12–22 months) and 15 months (Q1–Q3: 12–27 months), respectively (Supplemental Table 3). In these patients there were no significant changes in any of the antigen-specific IgG subclasses between baseline and the end of the study follow-up (Figures 3A,B). In 20 PLA₂R1-ab positive patients autoantibodies relapsed after a median follow-up time of 21 months (Q1–Q3: 14–40 months) (Supplemental Table 4). At the time of relapse, PLA₂R1-specific IgG subclasses were not significantly different from the findings at baseline (Figure 3C).

In order to confirm whether these IgG subclass binding patterns are specific for the kidney, or may also apply to other organs, we searched for another human tissue where both antigens are stably expressed. Using immunofluorescence staining we, in fact, found that PLA₂R1 as well as THSD7A were expressed not only in podocytes, but also lung bronchioles (Figure 4). In human glomeruli PLA₂R1 was detected in the intracellular compartment and on the surface of podocytes, whereas THSD7A is expressed in an almost linear pattern at the basolateral side of the podocyte. In lung bronchioles PLA₂R1 is located within the epithelial cells, whereas THSD7A was also found at the basolateral side of the cells. When analyzed by confocal microscopy, PLA₂R1 and THSD7A showed only discreet overlap in the glomerulus, while no overlap was detected in the bronchioles (Figure 4, Supplemental Figure 2).

The presence of PLA₂R1 and THSD7A in lung tissue was further confirmed by Western blot analyses (Figure 5). In both cases the antigen-antibody interaction was only detected under non-reducing conditions, suggesting that the antibodies bound to the conformational structures of the antigens in both organs. We observed a different migrating pattern of both antigens derived from the lung and the kidney, which was abolished after deglycosylation, indicating tissue-specific differences in antigen glycosylation for both PLA₂R1 and THSD7A (Figure 6).

Binding of all antigen-specific IgG subclasses to glomerular and lung tissue was investigated for all THSD7A-ab positive patient sera at baseline (n = 41) as well as all 15 sera with persistent THSD7A-ab positivity during follow-up. All analyzed sera showed binding to glomerular and lung THSD7A antigen under non-reducing conditions and presented identical IgG subclass patterns in the kidney and the lung (Figure 7). To confirm these findings in patients with PLA₂R1-associated MN, we performed the analyses in a selection of PLA₂R1-ab positive sera at baseline: a) three patients, in whom PLA₂R1-ab had disappeared during follow-up, b) three patients with PLA₂R1-associated MN and malignancy, and c) four patients with persistence of PLA₂R1-ab, in whom we compared sera at baseline and during follow-up. Equivalently to the THSD7A-ab positive sera, all 14 PLA₂R1-ab positive sera showed identical IgG subclass pattern binding to lung and kidney tissue extracts (Figure 7).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.