Structural analysis of urinary light chains and proteomic analysis of hyaline tubular casts in light chain associated kidney disorders

Urine sediment

Morning urine samples were collected from 17 patients with multiple myeloma or monoclonal gammopathy of renal/undetermined significance. Demographics, clinical course of the patients and their predominant paraprotein are depicted in Table 1. After centrifugation of the urine at 3,000 g for 10 min, the supernatant was frozen at −80 °C for later use in fluid analysis. The pellet was re-suspended in 2.5 mL cell culture media (RPMI 1640 containing 10% calf serum) and 100 µL were applied to the funnel of a Shandon cyto-centrifuge. Following a spin at 1200 RPM for 3 min cytoslides were air dried for at least two hours and then either immediately processed as described below or wrapped in aluminum foil and frozen at −25 °C for further analysis.

Hematoxylin/Eosin staining was carried out as used in routine hematology staining of blood smears. Slides were covered in Depex mounting media and a cover slip was applied for microscopic observation.

Immunofluorescence for anti kappa/lambda staining (confocal microscopy) was carried out after fixation in acetone for 4 min. The cell and cast containing area was surrounded by a fat pen (Dako) and the staining area was rewetted using PBS. For kappa/lambda staining goat (diluted 1:10000, A0191 and A0193 from Dako) or rabbit anti human kappa/lambda polyclonal antibody (Ab) (P0212, Dako) was applied (1:5000 in PBS) and incubated for two hours at room temperature following two washes in PBS each for 10 min, the donkey anti goat or goat anti rabbit (Alexa 488, dilution 1:1000 in PBS) was applied for 1 h at room temperature. Before washing the slide 20 µL of DAPI solution was spotted onto the Ab containing area. Following the final wash Vetashield mounting media and cover slip were applied and the slides were recorded under a Zeiss invert confocal microscope using ZEN program for picture recording.

Urinary cast isolation

Urine sediment generated as described above was air dried from patients presenting with AL-amyloidosis and cast nephropathy or presenting with cast nephropathy alone. Cytopreparations were viewed under phase contrast light microscope (Labovert FS, Leitz, Germany) and individual casts were picked using the tip of specifically gas flame pulled glass Pasteur pipettes under microscopic observation. Five casts were placed into a 400 µL PCR tube containing 10 µL PBS. This represented the work up starting material for proteome analysis.

Substructural fractionation on a sucrose gradient

Cells and cell nuclei and large aggregates were pelleted by centrifugation at 1500 RPM. Five hundred µL of resultant supernatant containing cast fragments and micro-aggregates and exosomes were loaded on to a discontinuous sucrose gradient in Ultra-clear centrifuge tubes (Beckman, 344062) at 4 °C. Following two hours centrifugation at 40000 RPM using an SW60 rotor in an L-80 ultracentrifuge the gradient was immediately fractionated in 200 µL aliquots using a peristaltic pump starting at the bottom of the tube. Individual fractions were subjected to immunoblotting using antibodies for kappa and lambda light chains (Dako as above) and aquaporin 1 (AQP1; Millipore).

Immunoblotting

Twenty µL of individual fractions of urine were loaded onto a 12% SDS PAGE gel and run under denaturing conditions. Following the transfer onto nitrocellulose the proteins were stained using kappa and lambda specific antibodies (Dako) as well as Ab against AQP1 (Millipore). The specific protein staining was developed by HRP labeled goat anti rabbit affinity purified antibodies (Dako). After each staining procedure the blot was washed using TPBS for 10 min twice. Chemiluminescence reagent was applied to the blot which was visualized using Fusion software at the luminescence recorder (Fusion Fx, Vilber Lourmat). Pictures were further processed using Photoshop version 6. Densitometric evaluation of individual signals of dimers vs. monomers were performed using the Fusion software (Fusion Fx, Vilber Lourmat).

Uromodulin ELISA test

Uromodulin concentration was measured in the same sucrose density gradient fractions as used for immunoblotting by a commercially available ELISA assay (BioVendor, Brno, Czech Republic). The Elisa was performed as suggested in the company’s test manual. In brief, fractions were diluted in sample dilution buffer and incubated together with the standard series. Following the incubation with biotinylated detection antibody and streptavidin-HRP conjugate the signal was developed with TMB substrate and read by an ELISA reader. Concentrations were calculated according the standard curve. The detection range of the ELISA assay is 0.5 to 25 ng/mL.

Mass spectrometry

Statistical analyses

We present quotients as means ± standard deviation (SD). Differences of lambda and kappa dimer vs. monomer ratio were analyzed with paired t-test. Data management and analysis was conducted by Microsoft Excel (Microsoft, Redmont, WA, USA). The statistical testing performed was two-sided and a p-value ≤ 0.05 was considered significant.

Ethics approval and consent to participate

This study involves human participants. Urine samples were obtained from adult study participants and/or their legal guardians older than 18 years who have given written informed consent. All methods of experiments involving human participants were carried out according to relevant guidelines. All experimental protocols involving human participants received approval of the Ethics Committee of the Medical University of Vienna (EK 1043/2016).

Cast morphology and constituent analysis

Hematoxylin/Eosin staining of tubular casts revealed various staining intensities as well as colour variation. Colour variation results from different contents, which either are more basophilic or acidophilic depending on constituents. As demonstrated in Figs. 1A, 1C, 1D three casts are more eosinophilic when compared to one in Fig. 1B that shows a fainter staining. The thinner cast at Fig. 1C is originating from a different stretch of the nephron than the other cast fragments, such as the the loop of Henle. Three tubular cells are incorporated in this thin cast.

In order to further study the constituents of tubular casts we sought for the incorporation of light chain components. As already reported in earlier work tubular casts contain the light chain (LC) paraprotein such as also measured in serum. As shown in Figs. 2A and 2B the lambda light chain is incorporated at the tubular cast which has a thinner diameter at the left side upward directed end when compared with the other part. This picture further confirms the assumption that casts can originate from thinner stretches of the nephron.

Analysis of casts by electron microscopy has identified various fine structures in previous studies (Uribe-Uribe & Herrera, 2006) and the most dominant variant was represented by the hyaline form, however, the biochemical proteinaceous constituents of hyaline casts in light chain associated kidney disorders has not been the subject of investigation before. Only uromodulin was documented as light chain interaction partner. Therefore we sought by mass spectrometry to analyze the proteome of isolated urinary casts from patient 14 who had no detectable uromodulin in the sucrose density fractionation. In these experiments we identified the patient’s lambda LC and histones in particular H2B sub-variants (Table 2) as a dominant additional constituents of the cast proteome. It is of interest that a urinary protease such as cathepsin B was also contained in this complex (Table 2). In repeated proteome analyses (n = 3) H2 variants were similarly highly present. Depending on the cast type chosen for analysis more or less cellular proteins were recovered. However, this study was focused on hyaline casts. Of particular note is the observation that as shown in (Fig. S1) patient 14 had received immunomodulatory therapy 2.5 days before the onset of AKI stage 3 according to KDIGO. His urine sediment was filled with hyaline casts. The liberation of histones from apoptotic plasma cells and the filtration into the urine might have represented the main provocation of cast formation in the renal tubules as uromodulin was not detectable. His lambda light chain analysis in the density gradient fractions (Fig. 3A) showed to some extent presence of aggregated LC signals in density fractions 6-10 representing most likely complexes but there was almost no presence of LCs in exosome containing fractions 2-6. Presence of exosomes was verified by immunoblotting of the same fractions with Ab against AQP1 a tubular cell membrane protein (Fig. 3B). By a second approach using urine SDS PAGE analysis and immunoblotting lambda LCs appeared predominantly as dimers in urine and made up 78% in the densitometric analysis (Fig. 4, lane 14). No amyloid deposition had been observed in the kidney biopsy of this patient although the duration of disease had been lasting for already 5 years (Table 1). The patient underwent high cut-off dialysis and as demonstrated in Fig. S1 his renal function improved.

Analysis of urinary light chain structure

As already documented in earlier work not the concentration but rather the structure and morphology represents the underlying cause for manifestation of renal light chain disease. For this reason we analyzed the light chain dimerization in 17 myeloma or monoclonal gammopathy patients (Table 1). As demonstrated in Fig. 4 patients positive for lambda LCs showed patterns with higher quantities of LC dimers than monomers. Furthermore, eight patients out of 11 positive for lambda LC developed CKD stages 4–5 according to KDIGO (Table 1). Patients 3, 8, 9, 10, 13 and 14 were highly positive for lambda dimers in the immunoblots (Fig. 4). Dimerization of lambda LC could be entirely resolved into monomers by adding DDT as reducing agent in the sample buffer (Fig. 4B).

Patients positive for kappa LCs showed patterns with higher quantities of LC monomers than dimers (Fig. 5). The densitometric analysis evaluated ratio of lambda dimers vs. monomers was significantly higher (2.12 ± 0.75) when compared with the ratio of kappa dimers vs. monomers (0.64 ± 0.47), p = 0.00001.

We further intended to elaborate the light chain aggregates or complex formation resulting in renal light chain disease by density gradient centrifugation in patients 13, 14 and 15. In the urine sediment of amyloidosis patient 13 the light chain was found in both the soluble urine proteins (Fig. 6A, fractions 11–14) and to a lesser extent in the membrane fractions (Fig. 6A, lanes 1–4). Membrane exosome fractions were identified by AQP1 blotting of the same fractions (Fig. 6B). This was similar for patient 15 despite the fact that his disease was linked to kappa LC which was going along with AKI stage 3 (Fig. 7A). Membrane fractions represented by lanes 1–4 containing exosomes and cast particles were identified by presence of AQP1 (Fig. 7B).

The Uromodulin protein (UMOD) has been shown earlier to represent the binding partner for LCs in forming casts. In this study we did not find uromodulin protein in the casts of patient 14, which was confirmed by absence in sucrose gradient density fractions and urine. In contrast, there was a small peak concentration in heavy membrane fraction 2 in patients 13 and 15 (Figs. 6C and 7C). This fractions corresponded with presence of exosomes and cast fragments in the gradient sedimentation experiments. Fractions 11–14 correspond to soluble molecules in urine. In patients 13 and 15 urine uromodulin concentration was high (Figs. 6C and 7C).