Galactose-Deficient IgA1 B cells in the Circulation of IgA Nephropathy Patients Carry Preferentially Lambda Light Chains and Mucosal Homing Receptors

Significance Statement

IgA nephropathy (IgAN) is associated with mesangial deposition of aberrantly glycosylated IgA1 containing λ light chains and the association of upper respiratory or digestive tract infection with macroscopic hematuria. We found that peripheral blood Gd-IgA1⁺ cells from IgAN patients express predominantly λ light chains and CCR9 and CCR10, compared with healthy controls. Furthermore, Gd-IgA1⁺ cell populations in peripheral blood are enriched with plasmablasts/plasma cells. Therefore, IgAN is associated with an increased number of migratory Gd-IgA1-λ⁺ cells predestined for homing to upper respiratory and digestive tract mucosal tissues, where their final maturation and Gd-IgA1-λ secretion may be stimulated during upper respiratory or digestive tract infections.

Abstract

Background

IgA nephropathy (IgAN) primary glomerulonephritis is characterized by the deposition of circulating immune complexes composed of polymeric IgA1 molecules with altered O-glycans (Gd-IgA1) and anti-glycan antibodies in the kidney mesangium. The mesangial IgA deposits and serum IgA1 contain predominantly λ light (L) chains, but the nature and origin of such IgA remains enigmatic.

Methods

We analyzed λ L chain expression in peripheral blood B cells of 30 IgAN patients, 30 healthy controls (HCs), and 18 membranous nephropathy patients selected as disease controls (non-IgAN).

Results

In comparison to HCs and non-IgAN patients, peripheral blood surface/membrane bound (mb)-Gd-IgA1⁺ cells from IgAN patients express predominantly λ L chains. In contrast, total mb-IgA⁺, mb-IgG⁺, and mb-IgM⁺ cells were preferentially positive for kappa (κ) L chains, in all analyzed groups. Although minor in comparison to κ L chains, λ L chain subsets of mb-IgG⁺, mb-IgM⁺, and mb-IgA⁺ cells were significantly enriched in IgAN patients in comparison to non-IgAN patients and/or HCs. In contrast to HCs, the peripheral blood of IgAN patients was enriched with λ⁺ mb-Gd-IgA1⁺, CCR10⁺, and CCR9⁺ cells, which preferentially home to the upper respiratory and digestive tracts. Furthermore, we observed that mb-Gd-IgA1⁺ cell populations comprise more CD138⁺ cells and plasmablasts (CD38⁺) in comparison to total mb-IgA⁺ cells.

Conclusions

Peripheral blood of IgAN patients is enriched with migratory λ⁺ mb-Gd-IgA1⁺ B cells, with the potential to home to mucosal sites where Gd-IgA1 could be produced during local respiratory or digestive tract infections.

IgA nephropathy (IgAN) is the most common cause of primary glomerulonephritis in many countries.^1,2 Current data indicate that it is a disease of autoimmune character in which polymeric (p)IgA1 with altered O-glycans in the hinge region serves as an autoantigen recognized by anti-glycan antibodies, predominantly of the IgG1 subclass, resulting in the formation of circulating immune complexes (CICs).^3–8 Alteration of O-glycans in the α1 chains of the hinge region reduces the galactose moiety (Gd-IgA1).^9–11 Large immune complexes, which are not effectively catabolized, are deposited in the kidney mesangium, where they stimulate the proliferation of mesangial cells and the production of an extracellular matrix, leading to the glomerular damage.^12–15 Thus, production of Gd-IgA1 plays the pivotal role in the pathogenesis of IgAN.

Immunohistochemical evaluations of mesangial deposits in IgAN revealed the pronounced dominance of light (L) chains of the λ isotype.^16–18 Furthermore, the pIgA1-λ from patients with IgAN exhibited increased binding to human mesangial cells in vitro in comparison to pIgA1-λ from healthy controls (HCs).¹⁹ However, the reason for this preferential deposition of IgA1 with λ L chains remains enigmatic. Infections of the upper respiratory and digestive tracts are often associated with macroscopic hematuria in IgAN patients.^14,20,21 Such infections may induce the production of multiple cytokines, including the proinflammatory cytokine IL-6, reported to be involved in the pathogenesis of IgAN.²² Indeed, IgAN patients display elevated levels of IL-6 in the blood and urine^23,24 and in vitro the IL-6 and its analog leukemia inhibitory factor increase production of Gd-IgA1 by cell lines isolated from IgAN patients.^22,25,26

Here we characterized, for the first time, the phenotypes of surface Gd-IgA1 positive memory B cells, lymphoblasts, and plasmablasts (PBs) in the peripheral blood of IgAN patients, non-IgAN disease controls, and HCs with respect to the expression of λ L chains and migration/homing receptors. We observed that, in contrast to non-IgAN patients and HCs, surface/membrane bound (mb)-Gd-IgA1⁺ cells from peripheral blood of IgAN patients express predominantly λ L chains and also exhibit preference for the expression of cellular receptors involved in their homing to upper respiratory and digestive tracts. In contrast, total mb-IgA⁺, mb-IgG⁺, and mb-IgM⁺ cells are preferentially positive for κ L chains in all characterized groups. Nevertheless, in IgAN patients λ L chain subpopulations of mb-IgG⁺, mb-IgM⁺, or mb-IgA⁺ cells are significantly higher in comparison to non-IgAN patients and/or HCs, indicating altered differentiation, maturation, and behavior of B cells in IgAN.

Methods

Study Subjects

The IgAN patients were recruited from the Nephrology, Rheumatology, and Endocrinology Department, University Hospital Olomouc, Czech Republic in June, July, and August of 2020. All IgAN patients visiting the department in the stated period were sampled and involved in this study. The diagnosis of IgAN was based on the biopsy-proven predominant mesangial deposition of IgA by immunofluorescence as a standard procedure. Baseline clinical data, including gender, age, blood pressure, eGFR, urinary albumin-creatinine ratio, 24-hour proteinuria, hematuria, and treatment with angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, were obtained from review of medical records. The disease control group consisted of membranous nephropathy patients from the Nephrology, Rheumatology, and Endocrinology Department, University Hospital Olomouc, Czech Republic and from the Department of Internal Medicine, University Hospital Ostrava, Ostrava, Czech Republic. The other control group consisted of healthy subjects from Department of Transfusion Medicine clients. All study subjects provided informed consent. This study was approved by the ethics committee of the University Hospital Olomouc and University Hospital Ostrava. The characteristics of the IgAN group and the control groups are described in Table 1 and Supplemental Figure 1.

Table 1.

Clinical and biochemical characteristics of study subjects

Characteristic	IgAN	Healthy	Non-IgAN
Number of subjects	30	30	18
Males/Females	18/12	17/13	12/6
Age (yr)	49.9±15.1	42±8.9	58±11.7
Time since diagnosis (yr)	12±11	ND	7±5.4
Blood pressure, SBP/DBP mean values (mmHg)	131/77	120/80	145/86
Serum IgA concentration (mg/ml)	2.7±1.2	1.5±0.7	1.32±0.6
Gd-IgA1, relative to standard Gd-IgA	0.75±0.2	0.41±0.16	0.5±0.2
eGFR (ml/min per 1.73 m2)	52.2±24	ND	95±15
Serum creatinine (μmol/l)	148±84	ND	139±194
Serum urea (mmol/l)	12.7±15	ND	8.2±3.4
Proteinuria (g/d)	1.2±1.5	ND	1.05±0.7
Urinary albumin-creatinine ratio (mg/mmol)	66.7±79.2	ND	43.4±51.4
Hematuria (Ery/μl)	27.1±56.9	ND	5.4±5.7
Disease activity stabilized/progressiona	17/13	ND	18/0
Ongoing glucocorticoid therapy, treated/all	6/30	0/30	9/18
ACEi/ARB, treated/all	29/30	0/30	18/18

Serum concentrations of Gd-IgA1 and total IgA are shown in Supplemental Figure 1. Non-IgAN cohort is membranous nephropathy patients. Glucocorticoid therapy of IgAN is prednisone. Results are presented as counts for categorical data or mean±SD (apart from blood pressure) for continuous data.

ND, not done; SBP, systolic blood pressure; DBP, diastolic blood pressure; Ery, erythrocytes; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker.

^aProgression of disease was based on a decrease in eGFR of >10%/year.

Reagents

All chemicals, except where mentioned, were purchased from Thermo Fisher Scientific (Waltham, MA), including tissue culture media, media supplements, and some antibodies.

The murine mAb specific to Gd-IgA1 (35A12) was provided by the Department of Nephrology, School of Medicine, Fujita Health University in Japan. The 35A12 mAb was preferred over the more commonly used KM55²⁷ as the second one is available in BSA-containing buffer, acting as the stabilizer which, nevertheless, should be removed before biotin conjugation required for flow cytometry analyses. This procedure led to substantial losses of antibody processed for flow cytometry analysis purposes. 35A12 and KM55 mAb were compared in preliminary flow cytometry and ELISA experiments as documented in Supplemental Figure 2.

Determination of Gd-IgA1 and Total IgA

Levels of total serum IgA were determined by ELISA according to previously described protocols.^28,29 The Gd-IgA1 concentration was measured using the 35A12 antibody specifically recognizing terminal N-acetylgalactosaminyl (GalNAc) residues.³⁰ ELISA plates were coated with goat anti-human IgA F(ab’)₂ at final concentration 1 µg/ml overnight at 4°C (Jackson ImmunoResearch). Plates were blocked with 1% BSA/PBS/0.05% Tween 20 for 3 hours at room temperature and diluted sera applied into wells. Bound IgA was desialilated by treating with sialidase A (Arthrobacter ureafaciens; ProZyme, Agilent Technologies, Hayward, CA) at concentration 2 mU/ml in 100 mM sodium phosphate buffer (pH 6.0) for 3 hours at 37°C. Gd-IgA1 was detected with mouse anti-human Gd-IgA1 IgG3 antibody 35A12 at 20 µg/ml in blocking buffer. Bound 35A12 antibody was detected with rabbit anti-mouse IgG antibody conjugated with horseradish peroxidase diluted 1:4000 in blocking buffer (Sigma Aldrich). After development with O-phenylenediamine-H₂O₂ substrate, the absorbance was read at 490 nm. Results were expressed as the ratio of individual sera absorbance relative to standard myeloma Gd-IgA1 absorbance.

PBMC Isolation

PBMCs were isolated by the standard protocol⁸ with minor modifications. Briefly, 25 ml of EDTA-treated blood was mixed with sterile PBS (Biosera, Nuaille, France) at 3:2 ratio. The blood was applied by overlaying on the Histopaque 1077 (Sigma, St. Louis, MO) and centrifuged at 400 × g at room temperature (RT) for 30 minutes. The obtained PBMCs were washed twice with PBS for 20 minutes at RT using centrifugation at 200 × g.

Staining of Surface Gd-IgA1 in PBMCs for Flow Cytometry Analysis

Freshly isolated or stimulated PBMCs were stained in four steps. The first step consisted of the cells’ treatment with 0.05% subtilisin A protease in Tris-Cl buffer, pH 7.4, 2 hours at RT.²⁷ Next, Fc receptors were blocked using 10% heat-inactivated human serum in PBS for 10 minutes at RT, followed by incubation with 4 μg of mouse anti-human Gd-IgA1 biotin-labeled mAb in 5% FBS in PBS, for 1 hour at 37°C. During the next step, fluorophore-labeled mAbs to cell surface markers were applied to the cells with fluorophore-labeled streptavidin, all diluted in 5% FBS in PBS for 30 minutes, at RT in the dark. All further washing steps were performed using 1 ml of PBS followed by centrifugation at 250 × g for 5 minutes. Surface markers were labeled using the following mAbs: anti-CD19 PE-AlexaFluor 610, anti-IgG BV750, anti-IgM BV605, anti-IgD BV510, anti-CD199 (CCR9) BV421, anti-CD197 (CCR7) BV785, anti-CD38 BV650, anti-CD138 BV711 (Sony Biotechnology, Tokyo, Japan), anti-CD49d (α4) PEcy5, anti-CCR10 PerCP cy5.5 (Becton Dickinson, Franklin Lakes, NJ), anti-IgA-FITC (Jackson ImmunoResearch, West Grove, PA), streptavidin-PE (SouthernBiotech, Birmingham, AL), and anti-λ L chain Pacific Blue (BioLegend, San Diego, CA). Gd-IgA1-specific antibody (35A12) was conjugated with biotin using the Easy Biotin Labeling Kit (Abcam, Cambridge, UK). All gating steps were performed according to a set of fluorescence minus one (FMO) controls (Supplemental Figures 3 and 4).

Flow Cytometry Analysis

All analyses were performed by the Sony Spectral Analyzer (Sony Biotechnology) equipped with blue (488 nm) and violet (405 nm) lasers and SP6800 software. More detailed analyses were performed by FlowJo software v.10.7 (BD Life Sciences, Ashland, OR).

Statistical Analyses

The Wilcoxon rank sum test was used to compare subject and control cell populations. All statistical analyses were done by GraphPad Prism v.8 software (GraphPad Software, La Jolla, CA).

Results

Phenotypic Characteristics of mb-Gd-IgA1⁺ B cells from Peripheral Blood

Mb-Gd-IgA1 molecules on lymphocytes of B cell lineage in peripheral blood were detected in PBMCs from 30 IgAN patients, 30 HCs, and 18 non-IgAN patients. Cells were initially stained for mb-Gd-IgA1 followed by staining for mb-IgA, mb-IgG, mb-IgM, and mb-IgD, and λ L chains. In parallel, B cell subpopulation-specific markers CD19, CD38, and CD138 were determined to distinguish B cells (CD19⁺), CD19⁺ CD38⁺ CD138⁺ cells, categorized as plasma cells (PCs) irrespective of the absence of typical PC morphologic features, and PBs/PCs (CD19⁺ CD38⁺). Figure 1A represents an example of gating strategy. Figure 1B demonstrates that the percentage of mb-Gd-IgA1⁺ cells was relatively low and did not differ significantly among IgAN, HC, and non-IgAN groups (Figure 1B). mb-Gd-IgA1⁺ cells represent 0.9%–1.7% of the total mb-IgA⁺ cell population.

Figure 1.

Gd-IgA1⁺ cells from IgAN patients preferentially express λ L chains. (A) Gating strategy used for Gd-IgA1⁺ cell detection. Gd-IgA1⁺ cells were characterized by their size (lymphocyte gate Lymphocytes), followed by gating for single cells (exclusion of doublets and clumps) (Single cells), CD19⁺ cells (CD19⁺), followed by surface IgA⁺ cells gating (IgA⁺), and finally Gd-IgA1⁺ gating (Gd-IgA1⁺). (B) The populations of surface Gd-IgA1⁺ cells from total IgA⁺ cells in IgAN patients (30 subjects), non-IgAN patients (18 subjects diagnosed with membranous nephropathy) and HCs (30 subjects). (C) The population of λ L chain-positive cells from total Gd-IgA1⁺ cells. The significantly higher population of λ L chains on Gd-IgA1⁺ cells from IgAN patients was detected. (D) The population of λ L chain-positive cells from total IgA⁺ cells. The significantly higher population of λ L chains was detected on IgA⁺ cells from IgAN patients versus HCs and non-IgAN patients. (E) The proportion of κ and λ L chain expression on Gd-IgA1⁺ and total IgA⁺ cells of IgAN patients. In IgAN patients IgA⁺ cells are predominantly κ L chain-positive whereas Gd-IgA1⁺ cells are predominantly positive for λ L chains. Means and standard deviations are shown. P values were calculated using Wilcoxon rank sum test, *P<0.05, ***P<0.001, ****P<0.0001.

Parallel expression of IgA1 and λ L chains has been discussed for IgAN in many publications.^17,31,32 Thus, we analyzed the percentage of λ L chain-positive Gd-IgA1⁺ cells in the blood from IgAN, HC, and non-IgAN subjects. IgAN patients exhibited a significantly higher proportion of λ L chain-positive Gd-IgA1⁺ cells (>68%) in comparison to HCs (<40%), and the non-IgAN group (<22%) (Figure 1C and Supplemental Table 1A). The trend of a higher proportion of λ L chains in IgAN patients was confirmed also for the population of total IgA⁺ cells, although at substantially lower absolute values (35% in IgAN group, 29% in HC, and 19% in non-IgAN) (Figure 1D and Supplemental Table 1A). Thus, the proportions of κ versus λ L chain-positive IgA⁺ and Gd-IgA1⁺ cells in IgAN patients and HCs or non-IgAN patients were different. In the group of IgAN patients, total IgA⁺ cells were preferentially positive for κ L chains (65%), whereas Gd-IgA1⁺ cells were preferentially λ L chain-positive (68%) (Figure 1E and Supplemental Table 1A). A detailed overview of cell numbers in populations and subpopulations generated by gating strategy is shown in Supplemental Table 2.

λ L Chain Expression on B cells of Other Ig Isotypes

Considering the significant increase in the λ L chain-positive populations of Gd-IgA1⁺ and total IgA⁺ cells in IgAN patients in comparison to HCs and non-IgAN patients, we analyzed λ L chain expression on total CD19⁺ cells and on CD19⁺ cells positive for mb-IgG, mb-IgM, and mb-IgD (Figure 2).

Figure 2.

Analysis of λ L chain expression on CD19⁺ cells and CD19⁺ of individual Ig isotypes. (A) The population of CD19⁺ cells from IgAN patients contains over 25% of λ L chain-positive cells, which was moderately but significantly higher than in HCs and non-IgAN controls (membranous nephropathy). (B) The above difference among IgAN patients, HCs, and non-IgAN patients in the CD19⁺ λ⁺ subpopulation is predominantly attributable to IgA⁺, followed by IgG⁺ cells. No differences in λ subpopulations were detected among IgAN patients and HCs for IgM⁺ and IgD⁺ CD19⁺ cells. Comparison of cells from IgAN and non-IgAN patients identified a significantly higher proportion of λ L chain-positive cells in IgAN patients for IgM, IgA, and IgG isotype. Interestingly, IgD⁺ cells exhibited a significantly higher population in the non-IgAN group when compared with the IgAN group. (C) The proportion of λ subset in individual Ig isotype-positive cells was compared for IgAN and non-IgAN patients and HCs. Only IgG⁺ CD19⁺ cells from IgAN patients exhibit the significantly higher proportion of λ⁺ cells (40%) in comparison to HC (>30%). The non-IgAN (membranous nephropathy) group exhibited the lowest proportion of λ L chain-positive cells in IgM, IgG, and IgD isotypes. P values were calculated using Wilcoxon rank sum test; *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Graphs show means and SDs.

The expression of the surface λ L chains on CD19⁺ cells of IgAN patients, HCs, and non-IgAN patients was evaluated. Our results demonstrated that total CD19⁺ cell populations from IgAN patients displayed a significantly higher percentage of λ⁺ cells in comparison to HCs and non-IgAN patients (Figure 2A). Considering individual Ig isotypes, not only mb-IgA⁺ cells but also mb-IgG⁺ cells from IgAN patients exhibit a significantly increased percentage of λ⁺ population than HCs and the non-IgAN group (Figures 1D and 2B). On the other hand, neither mb-IgM⁺ nor mb-IgD⁺ cells differ in λ L chain proportions in IgAN patients and HCs. Interestingly, the non-IgAN group exhibited a significant increase in λ⁺ mb-IgD⁺ cells (Figure 2B). The percentage of λ L chain-positive subsets for individual mb-IgG⁺, mb-IgM⁺, and mb-IgD⁺ is presented in Figure 2C, indicating the general tendency of IgAN patients toward higher proportion of λ L chain-positive B cells. A detailed percentage of λ versus κ subsets in individual mb-Ig⁺ populations and percentage of λ⁺ mb-Ig⁺ cells from the total CD19⁺ population is presented in Supplemental Table 1.

Analysis of mb-Gd-IgA1⁺ Relative to Maturation/Differentiation Stadium of B Cells

The broad panel of fluorophore-labeled detection antibodies used in these experiments allowed us to determine the mb-Gd-IgA1⁺ cells’ maturation and differentiation stadium based on identification of PBs/PCs, PCs, and memory cells (CD19⁺ CD38⁻ CD138⁻).

In the population of mb-Gd-IgA1⁺ cells, the percentage of PBs/PCs was significantly higher in the IgAN than in the non-IgAN group (Figure 3A). In contrast, no significant differences were detected for the PB/PC populations of total mb-IgA⁺ cells from IgAN patients, HCs, and non-IgAN patients (Figure 3B). Furthermore, the PC population of mb-IgA⁺ cells is significantly higher in IgAN patients in comparison to HCs, and also non-IgAN patients (Figure 3B). This confirms our previous observations that IgA⁺ cells from IgAN patients are more highly differentiated than IgA⁺ cells from HCs.³³ There is a substantial difference between the mb-Gd-IgA1⁺ and mb-IgA⁺ cells’ maturation phenotype proportions in IgAN patients (Figure 3). Only 16% of total mb-IgA⁺ cells are PBs or PCs (Figure 3B), whereas about 33% of mb-Gd-IgA1⁺ cells are PBs or PCs, showing that Gd-IgA1⁺ cells display a more differentiated pattern than do IgA⁺ cells (Figure 3). We did not detect significant deviations in the proportion of κ versus λ L chain-positive mb-IgA⁺ and mb-Gd-IgA1⁺ (Figures 1 and 2) relative to maturation stadium (data not shown).

Figure 3.

Gd-IgA1⁺ populations comprise more PCs and PBs over memory cells in comparison to total IgA⁺ cells. Cells were stained for the expression of CD19, Gd-IgA1, IgA, CD38, and CD138 to separate Gd-IgA1⁺ and IgA⁺ CD19⁺ cells into PBs/PCs (CD38⁺), PCs (CD38⁺, CD138⁺), and memory cells (CD38⁻, CD138⁻). (A) Only a modest difference of borderline significance between IgAN patients and HCs was detected in the Gd-IgA1⁺ PBs/PCs and no difference was detected for PC populations. Gd-IgA1⁺ PB/PC populations and Gd-IgA1⁺ PC populations were significantly increased in the IgAN group compared with the non-IgAN group. Phenotype distribution of Gd-IgA1⁺ cells from IgAN patients’ blood is shown in the pie chart. (B) In the population of total IgA⁺ CD19⁺ cells, no difference between IgAN patients and HCs was detected for PB/PC populations. In contrast IgAN patients have a significantly higher population of IgA⁺ PCs, but still this population consists of <1% of total IgA⁺ blood cells. Phenotype distribution of IgA⁺ cells from IgAN patients’ blood is shown in the pie chart. P values were calculated using Wilcoxon rank sum test; *P<0.05, **P<0.01, ***P<0.001. Graphs show means and SDs.

λ⁺ mb-Gd-IgA1⁺ Cells from IgAN Patients Are Prone to Homing to Mucosal Tissues

Before the development of a technique for the selective detection of Gd-IgA1 cells, we reported that total mb-IgA⁺ cells from peripheral blood of IgAN patients comprise a significantly higher proportion of cells positive for integrin α4β1 and a significantly lower proportion of cells positive for α4β7 integrin, in comparison to HCs. These cells are homing to mucosal tissue of the upper respiratory or digestive tract.³³ Here we analyzed whether κ and λ subsets of mb-Gd-IgA1⁺ cells differ in the populations homing to mucosal tissues by measuring the expression of the integrin α4 subunit and chemokine receptors CCR7, CCR9, and CCR10 (Figure 4). About 50%–60% of all mb-Gd-IgA1⁺ cells were positive for α4. In comparison to HCs, fewer α4⁺ cells were detected in the blood of IgAN patients. This difference was seen for both κ⁺ and λ⁺ subsets, although the statistical significance was achieved only in the λ⁺ subset (Figure 4A). In contrast, a substantial and significantly higher proportion of mb-Gd-IgA1⁺ λ⁺ cells from IgAN patients express CCR10, CCR9, and CCR7 (Figure 4, B and C). Analysis of mb-Gd-IgA1⁺ κ⁺ subsets did not confirm significant differences in CCR9, CCR10, and CCR7.

Figure 4.

Homing pattern of Gd-IgA1⁺ λ⁺ cells. Cells were stained for the expression of CD19, Gd-IgA1, α4, CCR7, CCR9, CCR10, and λ L chains to identify (A) α4⁺, (B) CCR7⁺, (C) CCR10⁺, and (D) CCR9⁺ subsets of CD19⁺ Gd-IgA1⁺ λ L chain-positive subsets in the IgAN and HC groups. Populations of the κ L chain-positive subsets exhibited no significant differences between IgAN patients and HCs for α4⁺, CCR7⁺, CCR9⁺, or CCR10⁺ subsets. In contrast, analysis of λ L chain-positive subsets identified significantly higher populations of CCR7⁺, CCR9⁺, and CCR10⁺ cells in IgAN patients in comparison to HCs. In contrast, the α4⁺ population was significantly higher in HCs. (E) In contrast to previous analysis of the population size (A–D), the measurement of surface expression of α4, CCR7, CCR9, and CCR10 (mean fluorescence intensity, MFI) on κ and λ L chain-positive cell subsets identified significant changes only for CCR9. P values were calculated using Wilcoxon rank sum test; *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Graphs show means and SDs.

Discussion

We demonstrated significant λ L chain expression preference in the peripheral blood of IgAN patients in comparison to HCs and non-IgAN patients, in subsets of total B cells, mb-IgA⁺, mb-Gd-IgA1⁺, and mb-IgG⁺ cells. We did not observe such differences in the populations of mb-IgM⁺ and mb-IgD⁺ cells between IgAN patients and HCs, but we detected a significant decrease of mb-IgM⁺, mb-IgA⁺, and mb-IgG⁺ cells the non-IgAN group. Importantly, the predominance of λ L chains is in agreement with reports concerning mesangial deposits in IgAN patients.^{16,18,31,34–36} Moreover, the ratio of κ/λ L chains in IgA1 molecules in sera of IgAN patients was significantly decreased compared with HCs,³² indicating that IgAN patients have more serum IgA1 with λ L chains. The L chain ratio was also analyzed for different isotypes of antibodies from sera of IgAN patients. In all measured Ig isotypes (IgA, IgG, IgM), the ratio of κ/λ was decreased in IgAN patients when compared with HCs.³⁷ Although we showed that CIC and mesangial deposits in IgAN patients may share ids, no disease-specific id was identified.³⁸ This observation, together with an abundance of λ L chains in serum and CIC Gd-IgA1 and on mb-Gd-IgA1⁺ cells, could be a consequence of the action of the immune system to eliminate potentially autoreactive B cells in the bone marrow by L chain receptor editing.³⁹

In this study, the percentage of the λ⁺ subset of mb-Gd-IgA1⁺ cells was approximately two times higher in IgAN patients than in HCs, and three times higher than in non-IgAN patients, reaching over 60% of total mb-Gd-IgA1⁺ cells. It is relevant that the percentage of peripheral blood mb-Gd-IgA1⁺ cells is low in IgAN patients and in HCs and non-IgAN patients. These findings agree with previous reports^29,40 that healthy individuals also have Gd-IgA1 Igs in serum, although at lower concentrations as indicated in Supplemental Figure 1.

Based on the comparison of peripheral blood mb-IgA⁺ and mb-Gd-IgA1⁺ B cell distribution into memory cells, PBs/PCs, and PCs, we identified a distinct pattern for mb-IgA⁺ and mb-Gd-IgA1⁺ cells in IgAN patients, HCs, and non-IgAN patients. In the mb-IgA⁺ cells a very rare population of CD138⁺ categorized as PCs was identified in IgAN patients, HCs, and non-IgAN patients, resembling physiologic blood cell counts characterized by the absence or sporadic detection of PCs. Nevertheless, the population of mb-IgA⁺ PCs was significantly higher in the IgAN group than in the HC and non-IgAN groups. In contrast, the PC subpopulation of the mb-Gd-IgA1⁺ cells was substantially higher in the IgAN and HC groups, reaching in IgAN patients almost 7% of total mb-Gd-IgA1⁺ cells. On the other hand, the PB/PC subpopulations of mb-IgA⁺ cells were equally represented among the IgAN, HC, and non-IgAN groups and only moderate enrichment of PBs/PCs was detected for mb-Gd-IgA1⁺ cells in IgAN patients.

Immunosuppressive therapy and particularly glucocorticoids are frequently used in IgAN patients with persistent proteinuria and rising serum creatinine, irrespective of maximal antiproteinuric therapy. Prednisone was reported to significantly reduce proteinuria and levels of serum IgA, Gd-IgA1, and IgA-IgG immune complexes.²⁸ Because our cohort included six IgAN patients with ongoing prednisone therapy, we performed preliminary analysis showing that prednisone in the IgAN group reduced only the IgA⁺ PB/PC population. Interestingly, in the non-IgAN controls, glucocorticoids reduced Gd-IgA1⁺ λ⁺, IgA⁺ λ⁺, and IgA⁺ PB/PC populations (Supplemental Table 3 and Supplemental Figure 5). Although preliminary, this analysis indicates that B cell subpopulations together with serum Gd-IgA1, IgA, and CICs could be explored as novel biomarkers of therapy response.

IgAN is characterized by the association of upper respiratory or digestive tract infections with macroscopic hematuria,^14,20,21 indicating that mucosal infections or the associated inflammation could stimulate enhanced production of Gd-IgA1.⁴¹ Therefore, we analyzed homing receptors, including mucosal targeting integrins and chemokine receptors. Interestingly, mb-Gd-IgA1⁺ λ⁺ cells from IgAN patients were substantially and significantly enriched with CCR9⁺ and CCR10⁺ subsets. The modest dominance of α4⁺ cells was observed in HC samples. α4⁺ is the common subunit in α4β1 (VLA-4) and α4β7 (LPAM-1) involved in targeting leukocytes to the upper respiratory tract and Waldeyer’s nasopharyngeal lymphoid ring and the digestive tract.⁴² Because there is no direct correlation between the localization of B cells in different tissues and the surface expression of adhesion molecules,⁴³ modest differences in α4 expression detected in this study seem to be less important in comparison to substantial differences in CCR9⁺, CCR10⁺, and CCR7⁺ mb-Gd-IgA1⁺ subsets. Thus, we propose that peripheral blood of IgAN patients is substantially enriched with mb-Gd-IgA1⁺ λ⁺ cells homing to the upper respiratory and digestive tracts and also to secondary lymph nodes. Nevertheless, this hypothesis requires further experimental confirmation.

This is, to our knowledge, the first study comparing populations of Gd-IgA1⁺ and Gd-IgA1 λ L chain ⁺ cells in the peripheral blood of IgAN patients, HCs, and non-IgAN controls as a potential source of Gd-IgA1 molecules, a crucial etiopathological component of IgAN. Because IgAN patients exhibit substantial differences among various geographical and racial cohorts, further studies are needed to generalize our observations. Furthermore, more detailed characterization of the biology of Gd-IgA1-λ⁺ cells, their mucosal tissue distribution, and their response to various inflammatory and maturation stimuli could provide a new direction in the development of causative therapy for IgAN.

Disclosures

J. Mestecky received royalties from Elsevier/Academic Press. J. Orsag reports consultancy with Astellas, Bayer, and Pfizer (PFE). K. Takahashi reports research funding with Mitsubishi Tanabe Pharma Corporation. All remaining authors have nothing to disclose.

Funding

The research was supported by the Ministry of Health, Czech Republic conceptual development of research organization grant MHCZ-DRO, FNOL, 00098892 (awarded to M. Raska and J. Zadrazil) and by the Ministry of School, Youth, and Sport, Czech Republic grant CEREBIT CZ.02.1.01/0.0/0.0/16_025/0007397 (awarded to M. Raska) and European Regional Development Fund Project ENOCH, CZ.02.1.01/0.0/0.0/16_019/0000868 (awarded to M. Raska and P. Kosztyu).

Author Contributions

J. Mestecky and M. Raska conceptualized the study; D. Galuszkova, J. Jemelkova, P. Kosztyu, K. Matousovic, J. Orsag, N. Petejova, K. Zachova, and J. Zadrazil were responsible for investigation; M. Raska was responsible for project administration and provided supervision; Y. Ohyama and K. Takahashi were responsible for resources; K. Zachova wrote the original draft; and K. Matousovic, J. Mestecky, Y. Ohyama, M. Raska, K. Takahashi, and J. Zadrazil reviewed and edited the manuscript.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2021081086/-/DCSupplemental.

Supplemental Table 1A. λ⁺ or κ⁺ subsets in individual Ig isotype mb-Ig⁺ cells.

Supplemental Table 1B. Percentage of λ⁺ mb-Ig⁺ subsets of total CD19⁺ cells.

Supplemental Table 2. Cell counts in individual analyzed PBMC populations.

Supplemental Table 3. Glucocorticoids-related changes in Gd-IgA1⁺, IgA1⁺, λ⁺, CD38⁺ populations of CD19⁺ cells.

Supplemental Figure 1. Serum Gd-IgA1 and IgA levels in IgAN patients and HC.

Supplemental Figure 2. Comparison of 35A12 and KM55 mAbs detection of Gd-IgA1.

Supplemental Figure 3. Gating strategies used for generation of Figures 2–4.

Supplemental Figure 4. FMO controls for multicolor staining.

Supplemental Figure 5. Effect of prednisone therapy on Gd-IgA1⁺ and λ⁺ cells.