Abstract
ABSTRACT. IgA nephropathy (IgAN) is characterized by predominant mesangial polymeric IgA1 (pIgA1) deposits, with increased plasma IgA1 levels. Plasma IgA levels are determined by the rate of IgA production, uptake by leukocytes, and removal by hepatocytes. Fcα receptor 1 (FcαR1) is a candidate molecule for the regulation of IgA levels, but reports of its expression in leukocytes in IgAN are conflicting. Increased binding of endogenous IgA to circulating granulocytes and monocytes in IgAN was demonstrated in this study. FcαR1 expression on leukocytes was increased, independently of plasma IgA levels. FcαR1 was not saturated in leukocytes, because of internalization of IgA after uptake. Further binding of exogenous IgA isolated from individual subjects was observed with leukocytes from the same subjects. Compared with cells from control subjects, granulocytes but not monocytes from patients with IgAN exhibited a greater binding capacity for exogenous IgA, predominantly pIgA. To circumvent the possibility that endogenous IgA might alter FcαR1 expression, granulocytes or monocytes derived from the HL-60 or U937 cell lines were used to explore the nature of IgA binding. A higher affinity for pIgA was demonstrated. Inhibition studies using unlabeled IgA, other serum proteins, or a specific FcαR1-blocking antibody suggested binding mechanisms other than FcαR1 for pIgA uptake by leukocytes. This study also suggested the migration and/or sequestration of “activated” leukocytes with predominant λ-IgA in the mononuclear phagocytic system or inflammatory tissues, after the initial binding of λ-pIgA. These immunologic abnormalities might contribute to the glomerulointerstitial injury in IgAN, in the presence of leukocytic infiltration.
IgA nephropathy (IgAN) is the most common form of glomerulonephritis throughout the world and is characterized by mesangial IgA deposits, mainly involving IgA1 (1). The predominance of polymeric IgA (pIgA) in the mesangium in IgAN is supported by studies of renal eluates. Monteiro et al. (2) demonstrated that 70% of the IgA eluted from 10 renal biopsies obtained from patients with IgAN exhibited characteristics of true pIgA. Recently, the subclasses related to the light chain composition of the mesangial IgA deposits have been studied, and a predominance of λ-IgA deposits has been observed (3). Furthermore, there are increased λ-IgA levels in patient sera (4,5) and increased numbers of IgA1-positive plasma cells in bone marrow (6), suggesting that IgA isotypic dysregulation might be confined to the bone marrow compartment.
Plasma levels of IgA molecules are determined by the rate of production by the bone marrow and mucosa-associated lymphoid tissues, uptake by leukocytes, and removal by hepatocytes and the spleen. Intravenous injection of radiolabeled human IgA1 in rodents revealed the greatest uptake by the liver, amounting to 5 and 16 times that by the kidney and the spleen, respectively (7). In vitro studies demonstrated that monocytes bound more IgA1 than did hepatocytes or isolated glomeruli. With the vast number of leukocytes in the circulation and in tissues, the nature and concentrations of IgA in the circulation and in other organs are likely to be determined by the pool of IgA bound to leukocytes and myeloid cells. It is also thought that the expression of IgA receptors on and the binding of IgA to circulating leukocytes might affect the mesangial uptake of IgA in IgAN. However, studies of IgA receptors on blood cells from patients with IgAN are few and the results are inconclusive (8,9). The tested hypothesis was that polymeric λ-IgA from patients with IgAN would demonstrate increased binding to leukocytes that might become activated or “primed” (10). These immunologic abnormalities might contribute to the glomerulointerstitial injury of IgAN in the presence of leukocytic infiltration.
Materials and Methods
Patients and Control Subjects
Thirty Chinese patients (14 male and 16 female patients) with clinical and renal immunopathologic diagnoses of primary IgAN were studied. The patients had been exhibiting symptoms for ≥12 mo, with proteinuria ranging from 0.4 to 2.9 g/d, and were between 19 and 45 yr of age (mean ± SD, 27.1 ± 5.2 yr). No significant renal impairment was documented for these patients, and their endogenous creatinine clearance values were ≥70 ml/min per 1.73 m2. Ten milliliters of heparinized blood were collected from each patient during clinical quiescence (no macroscopic hematuria and urinary erythrocyte counts of <10,000/ml in uncentrifuged urine samples). The plasma was separated by centrifugation at 3000 rpm for 10 min at room temperature. The plasma was aspirated and then stored frozen at −20°C until used for isolation of IgA with a jacalin-agarose affinity column. Plasma IgA levels were determined by nephelometry.
Thirty healthy subjects (16 male and 14 female subjects, comparable to the patients with respect to age and race) with normal renal function were used as healthy control subjects. No microscopic hematuria or proteinuria was documented for these control subjects, and their endogenous creatinine clearance values were >70 ml/min per 1.73 m2. Another 12 patients (nine patients with minimal-change nephropathy and three patients with membranous nephropathy, comparable to the patients with respect to age, race, and gender) with normal creatinine clearance values were used as disease control subjects. Plasma samples were similarly collected from these control subjects. The study was approved by the university and hospital ethics committees, and all patients and control subjects provided written informed consent for blood sampling.
Materials
R-phycoerythrin-conjugated monoclonal anti-human Fcα receptor 1 (FcαR1) (clone A59) and R-phycoerythrin-conjugated, isotype-matched, mouse IgG were purchased from Pharmingen (San Diego, CA). Anti-CD89 blocking antibody (clone My43) was obtained from Medarex Inc. (West Lebanon, NH). Horseradish peroxidase-conjugated rabbit monospecific anti-human κ light chain and anti-human λ light chain were obtained from Dakopatts (Copenhagen, Denmark). FITC-conjugated rabbit anti-human IgA F(ab′)2 and FITC-conjugated preimmune rabbit Ig F(ab′)2, used for flow cytometry, were obtained from Dako (Kyoto, Japan). FITC-conjugated anti-human IgA F(ab′)2, R-phycoerythrin-conjugated anti-human κ or λ light chain F(ab′)2, and FITC- and R-phycoerythrin-conjugated preimmune rabbit Ig F(ab′)2 were obtained from Pharmingen. Asialo-orosomucoid was prepared by desialylation of human orosomucoid with neuraminidase (0.03 U/mg protein), via incubation for 8 h at 37°C in 0.1 M sodium acetate buffer (pH 5.0). FITC-labeled IgA was obtained from Pierce (Rockford, IL); FITC-conjugated IgA was prepared from normal human serum IgA with a ratio of 3.1 mol FITC/mol IgA. The FITC-labeled IgA contained 9.5% high-molecular mass IgA, as determined by size-exclusion chromatography. Jacalin-bound proteins (JBP) and IgG were purified by using a jacalin-agarose affinity column (Pierce) and a protein G affinity column (Pharmacia, Uppsala, Sweden), respectively. All other chemicals were obtained from Sigma Chemical Co. (St. Louis, MO). Cell lines were obtained from the American Type Culture Collection (Rockville, MD) and included the monocytic line U937 and the promyelocytic line HL-60.
Isolation and Fast Protein Liquid Chromatography of JBP
JBP were fractionated at room temperature with a fast protein liquid chromatography (FPLC) system (Pharmacia), as described previously (11). Briefly, α2-HS glycoprotein (representing 24% of JBP in plasma globulins), monomeric IgA1 (mIgA1), pIgA1, and IgA1 immune complexes were separated by FPLC after jacalin affinity chromatography. After FPLC, the identity of IgA was confirmed by anti-IgA affinity chromatography and an IgA sandwich enzyme-linked immunosorbent assay (ELISA). Two pooled samples, from fractions 20 to 33 (pIgA1 fractions) and from fractions 34 to 50 (mIgA1 fractions), were prepared for further analysis. pIgA1 was high-molecular mass IgA (molecular masses between 350 and 1000 kD), and mIgA1 was low-molecular mass IgA (molecular masses of approximately 150 kD). The IgG contents in the fraction were measured with an anti-IgG ELISA. The pooled fractions were dialyzed and concentrated to 2 ml with Centriprep cartridges (Amicon, Beverly, MA) and were stored at −70°C until used. The purity of the IgA1 fractions was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting (11). IgG was prepared from plasma via ammonium sulfate precipitation, followed by protein G affinity chromatography.
Concentrations of total IgA, IgA1, and IgG in the pooled fractions from FPLC or in plasma were measured by ELISA, as described previously (11). For flow cytometric studies of the binding of IgA to leukocytes, the IgA concentrations of individually pooled mIgA and pIgA samples (from 30 patients with IgAN and 30 healthy subjects) were adjusted to comparable levels.
Determination of κ- and λ-IgA1 Concentrations and κ/λ Light Chain Ratios
The κ- and λ-IgA1 concentrations in different preparations were determined by ELISA, as described previously (3). Briefly, 96-well Immunlon 2 microtiter plates (Dynatech, Marnes la Coquette, France) were coated with 100 μl of rabbit anti-human IgA (Dakopatts), diluted 1:5000 in carbonate/bicarbonate buffer (pH 9.6), overnight at 4°C. The plates were then blocked for 1 h at 37°C with phosphate-buffered saline (PBS) containing 3% bovine serum albumin (BSA) and were washed three times with PBS containing 0.05% Tween 20 (PBS-Tween). One hundred-microliter samples of κ-IgA1 and λ-IgA1 standards (The Binding Sites, Birmingham, UK) diluted in PBS-Tween containing 0.5% BSA (PBS-Tween-BSA) were incubated in the wells for 2 h at 37°C; the wells were then washed three times with PBS-Tween. One hundred microliters of a 1:3000 dilution of horseradish peroxidase-conjugated rabbit anti-human κ light chain or anti-human λ light chain (Dakopatts) in PBS-Tween-BSA were added to the wells and incubated for 1 h at 37°C. After washing, o-phenyldiamine substrate (50 μl/well; Sigma) was added and the plates were incubated in the dark at 37°C for a minimum of 30 min. Finally, the reaction was terminated with the addition of 2 M sulfuric acid (50 μl/well), and the absorbances were measured at 490 nm by using a Spectra ELISA reader (SLT Lab Instruments, Salzburg, Austria). The κ-IgA1 and λ-IgA1 concentrations were determined from the absorbance values by referring to the corresponding standard curves, as described previously (11), and κ/λ ratios were calculated.
HL-60 and U937 Cell Lines
HL-60 and U937 cells were used to determine the intrinsic FcαR1 expression and IgA binding in leukocytes that had not been previously exposed to Ig or other human serum proteins. HL-60 and U937 cell lines were grown to the logarithmic phase and were collected by using 0.05% trypsin/0.02% ethylenediaminetetraacetate for 5 min at room temperature. HL-60 cells were induced to differentiate from promyelocytic to granulocytic development by culture for 6 d in complete RPMI 1640 medium with 1% DMSO (12).
Expression of FcαR1 (CD89)
Heparinized blood was washed free of plasma at 4°C by using cold PBS containing 1% fetal bovine serum and 0.1% NaN3. Nonspecific binding through FcγR was blocked by incubation of the washed blood cells with normal human IgG (10 mg/ml) at 4°C for 30 min. CD89 expression was measured by flow cytometry using R-phycoerythrin-conjugated monoclonal anti-human FcαR1 (clone A59; Pharmingen). R-phycoerythrin-conjugated, isotype-matched, mouse IgG was used as the control. The same protocol was used for the determination of CD89 expression on HL-60 and U937 cells. The stained cells were analyzed by using a Coulter Epics XL analyzer (Coulter Electronics, Miami, FL). At least 5000 fixed cells were analyzed for each sample. The granulocyte and monocyte populations were discriminated from other blood cells by using forward- and side-scatter parameters, as described previously (13). The fluorescence intensity was evaluated by assessing the percentages of positive cells, compared with isotypic controls. The results were expressed as positive cell percentages.
Binding of Endogenous and Exogenous IgA
Heparinized blood was washed free of plasma at 4°C by using cold PBS containing 1% fetal bovine serum and 0.1% NaN3. Endogenous bound IgA was measured by flow cytometry using FITC-conjugated anti-human IgA F(ab′)2. The binding of exogenous IgA to blood granulocytes or monocytes was determined by incubation of the IgA preparation with cold PBS-washed blood cells (plasma free); the bound IgA was measured by flow cytometry using FITC-conjugated anti-human IgA F(ab′)2. Background control staining was achieved by reaction with FITC-conjugated preimmune rabbit Ig F(ab′)2. Binding of exogenous IgA to HL-60 and U937 cells was determined in a similar manner. In inhibition experiments, blocking agents were added 20 min before the addition of exogenous IgA preparations. The stained cells were analyzed by using a Coulter Epics XL analyzer, as described above. Fluorescence intensity was evaluated by comparing the mean fluorescence channel or the percentage of positive cells with those of preimmune antibody-stained control samples. The results were expressed as mean fluorescence intensity (MFI) or cell percentages.
Flow Cytometric Analysis of κ- and λ-IgA
Surface-bound κ- or λ-IgA on granulocytes or monocytes was determined by using FITC-conjugated anti-human IgA F(ab′)2 and R-phycoerythrin-conjugated anti-human κ or λ light chain F(ab′)2. All staining was performed at 4°C with staining buffer (PBS with 1% fetal bovine serum and 0.1% NaN3). Background control staining was achieved by reaction with preimmune FITC- or PE-F(ab′)2 rabbit immunoglobulins. Binding of IgA isotypes to HL-60 and U937 cells was determined by using a similar method. The stained cells were analyzed by using a Coulter Epics XL analyzer. At least 5000 fixed cells were analyzed for each sample. The granulocyte and monocyte populations were discriminated from other blood cells by using forward- and side-scatter parameters. The fluorescence intensity was evaluated by comparing the mean fluorescence channel or the percentage of positive cells with those of preimmune Ig-stained control samples. The results were expressed as MFI or cell percentages.
Internalization of IgA by HL-60 and U937 Cells
Internalization of IgA by HL-60 and U937 cells was determined by using the method described by Stewart and Kerr (14). After addition of the exogenous IgA preparation, HL-60 and U937 cells were incubated at 37°C for 1 h. A duplicate set of cells was incubated at 4°C and used as control cells. The IgA isotypes on the HL-60 and U937 cells were determined by flow cytometry, as described above.
Statistical Analyses
The results are expressed as mean ± SD. For comparisons between the patient and control groups, the unpaired t test was used. Correlations between continuous variables were calculated by using Pearson’s correlation coefficient (r). All quoted P values are two-tailed. The results of IgA binding studies in HL-60 and U937 cells are expressed as the mean ± SD of five individual experiments.
Results
Plasma IgA Levels
The plasma IgA levels for patients with IgAN (3.36 ± 1.51 g/L) were significantly higher than those for healthy control subjects (2.0 ± 0.81 g/L, P = 0.0004). Similarly, the plasma κ- and λ-IgA levels for patients with IgAN were significantly higher than those for healthy control subjects (data not shown). The κ/λ ratio of total plasma IgA obtained from patients was lower than that for healthy control subjects (P = 0.036) (Figure 1). Similar findings were not documented with plasma IgG obtained from these subjects (data not shown). The plasma IgA levels and κ/λ ratios for the disease control subjects were similar to those for healthy control subjects (data not shown).
Figure 1. Percentages of granulocytes and monocytes expressing Fcα receptor 1 (FcαR1) (CD89) and κ/λ ratios of plasma IgA for 30 healthy control subjects (⧫) and 30 patients with IgA nephropathy (IgAN) (⋄). The horizontal bars represent mean values.
Quantitation of mIgA1 and pIgA1 in JBP
Two pooled fractions, i.e., the pIgA1 and mIgA1 fractions, were prepared for further analysis. pIgA1 was high-molecular mass IgA (molecular masses between 350 and 1000 kD), and mIgA1 was low-molecular mass IgA (molecular masses of approximately 150 kD). No IgG or IgM was detected in mIgA1 fractions, and IgG represented 0.1% of total protein in pIgA1 fractions, as measured by ELISA. The recovery of total IgA1 in the eluted JBP was 94.1 ± 2.1% of total IgA1 in the original plasma samples.
The concentrations of κ- or λ-IgA in the mIgA1 and pIgA1 fractions of the JBP samples were determined before adjustment to comparable levels of the IgA concentrations of the JBP samples from patients and healthy control subjects. Studies of IgA in the FPLC fractions demonstrated that mIgA and pIgA amounted to 92% and 8%, respectively, of total IgA for both healthy control subjects and patients with IgAN (Figure 2). mIgA fractions of JBP samples from patients with IgAN exhibited significantly higher levels of monomeric total IgA, κ-IgA1, and λ-IgA1, compared with healthy control subjects. However, pIgA fractions of JBP samples from patients with IgAN exhibited elevated levels of only polymeric λ-IgA1, compared with healthy control subjects. The κ/λ ratios of IgA fractions from patients were lower than those of IgA fractions from healthy control subjects. The κ/λ ratio of mIgA1 (but not pIgA1) in JBP samples was higher than that of total plasma IgA, which consisted of both IgA1 and IgA2.
Figure 2. Concentrations and κ/λ ratios of monomeric IgA (mIgA) and polymeric IgA (pIgA) in jacalin-bound protein (JBP) samples isolated from 30 healthy control subjects (⧫) and 30 patients with IgAN (⋄). The horizontal bars represent mean values.
For flow cytometric studies of the binding of IgA to leukocytes, the total IgA concentrations of pooled IgA samples (isolated by jacalin affinity chromatography) were adjusted to comparable levels. The mIgA1 and pIgA1 concentrations were similar in adjusted samples from patients and healthy control subjects. The κ/λ ratios of mIgA1 and pIgA1 in adjusted JBP samples from patients remained significantly lower, compared with samples from healthy control subjects.
Study of Leukocyte Subpopulations and the Occupation of IgA Receptors by Endogenous or Exogenous IgA1
We observed no differences in peripheral leukocyte counts between patients and healthy control subjects. The percentages of granulocytes (62.2 ± 8.4% versus 60.2 ± 5.4%) and monocytes (5.44 ± 0.34% versus 5.41 ± 0.62%) did not differ between the two groups of subjects. However, patients with IgAN demonstrated increased percentages of IgA-binding granulocytes and IgA-binding monocytes, compared with healthy control subjects (granulocytes, 2.43 ± 0.22% versus 1.98 ± 0.36%, P < 0.0001; monocytes, 2.94 ± 0.47% versus 2.46 ± 0.36%, P < 0.0001). In addition, patients with IgAN demonstrated increased percentages of granulocytes and monocytes expressing FcαR1 (CD89), compared with healthy control subjects (Figure 1). The MFI was significantly greater in monocytes from patients with IgAN (7.64 ± 1.06 versus 6.40 ± 0.99 for healthy control subjects, P < 0.0001), but similar measurements in granulocytes failed to reach statistical significance (5.95 ± 0.56 for patients with IgAN versus 5.80 ± 0.40 for healthy control subjects, P > 0.05). For both groups of subjects, no correlation was observed between plasma IgA levels and the percentages or fluorescence intensity of CD89-positive granulocytes or monocytes.
The occupation of IgA receptors by endogenous IgA1 was assessed as IgA1 bound to the surface of blood leukocytes. Increased levels of endogenous κ-IgA and λ-IgA bound to the surfaces of both granulocytes and monocytes from patients with IgAN, compared with healthy control subjects, were observed, but the κ/λ ratios of bound IgA1 did not differ (Figure 3). The κ/λ ratios of IgA1 bound to leukocytes were significantly lower than the κ/λ ratios of plasma IgA for the corresponding groups of subjects. The κ/λ ratios of IgA1 bound to monocytes were well correlated with those of plasma IgA for both groups of subjects (r = 0.8, P < 0.0001, for healthy control subjects; r = 0.88, P < 0.001, for patients with IgAN). However, such correlations were not observed for granulocytes from either group of subjects. Similar studies of IgG bound to the surfaces of blood leukocytes via FcγR revealed no differences in the amounts or κ/λ ratios between patients and healthy control subjects (data not shown).
Figure 3. The κ/λ ratios and κ and λ fractions of endogenous IgA bound to circulating granulocytes (a) and monocytes (b) from 30 healthy control subjects (⧫) and 30 patients with IgAN (⋄). The horizontal bars represent mean values. MFI, mean fluorescence intensity.
The binding of exogenous IgA to blood granulocytes or monocytes isolated from individual subjects was determined by incubation of IgA preparations from the same subjects with cold PBS-washed blood cells (plasma free); the bound IgA was measured by flow cytometry. The IgA preparations (isolated by jacalin affinity chromatography) were appropriately diluted to achieve comparable total IgA concentrations. The mIgA1 concentrations were similar in adjusted samples from patients and healthy control subjects, as were the pIgA1 concentrations. After incubation with exogenous IgA isolated from each studied subject, there was further binding by leukocytes from the same individual, as indicated by increased MFI. Although mIgA concentrations were 10-fold higher than those of pIgA in the JBP samples, there was increased binding of pIgA to blood leukocytes (compared with mIgA), amounting to 2-fold more for monocytes and 1.1- to 1.3-fold more for granulocytes (Figure 4). Although there were no differences in the absolute amounts or the κ/λ ratios of exogenous mIgA or pIgA bound to monocytes, granulocytes from patients with IgAN exhibited higher binding capacities for exogenous mIgA and pIgA, compared with cells from healthy control subjects. The κ/λ ratios of exogenous mIgA or pIgA bound to granulocytes did not differ between patients and control subjects.
Figure 4. The κ/λ ratios and concentrations of exogenous mIgA and pIgA bound to circulating granulocytes (a) and monocytes (b) from 30 healthy control subjects (⧫) and 30 patients with IgAN (⋄). The horizontal bars represent mean values. Leukocytes isolated from patients or subjects were incubated with IgA isolated from individual plasma samples, for determination of additional uptake of exogenous IgA.
FcαR1 expression on and endogenous and exogenous IgA binding to peripheral leukocytes from patients with IgAN were significantly greater than values for the disease control subjects. However, the values for disease control subjects did not differ from those for healthy subjects (data not shown).
Binding of IgA to HL-60 and U937 Cells
To minimize the potential effects of plasma IgA on the induction of FcαR1 expression on (9,15,16) and endogenous IgA binding to blood leukocytes, IgA binding was examined in leukocytes with no surface Ig and no previous exposure to human serum proteins. We used promyelocytic (HL-60) and monocytic (U937) cell lines; the former was induced to granulocytic development as described above. Flow cytometry revealed that both granulocytic HL-60 cells and monocytic U937 cells expressed FcαR1, with higher MFI in granulocytic HL-60 cells (6.57 ± 0.30 versus 5.09 ± 0.36, P < 0.0001, with MFI of 1.54 ± 0.05 for isotypic control antibody). When the cells were incubated with equivalent concentrations of FITC-mIgA and FITC-pIgA, greater binding of pIgA, compared with mIgA, was observed for both cell types. Identical experiments were performed with the addition of tenfold excess concentrations of unlabeled mIgA or pIgA (5 mg/ml) 20 min before the addition of exogenous FITC-IgA preparations (0.5 mg/ml) (Figure 5). The binding of FITC-mIgA was inhibited 98.2% in HL-60 cells and 96% in U937 cells with unlabeled mIgA, and similar values of 98% in HL-60 cells and 96% in U937 cells were obtained with unlabeled pIgA. Preincubation with unlabeled pIgA blocked the binding of FITC-pIgA by 98.7% in HL-60 cells and by 98.4% in U937 cells. However, preincubation with unlabeled mIgA blocked only 78% of the binding of FITC-pIgA to HL-60 cells and 71% of the binding to U937 cells, although a tenfold excess concentration of mIgA was added.
Figure 5. Binding of IgA to granulocytic HL-60 cells (a) and monocytic U937 cells (b) with no surface Ig and no previous exposure to human serum proteins. Tenfold excess concentrations of unlabeled mIgA blocked 78 and 71% of the binding of FITC-pIgA to HL-60 and U937 cells, respectively, but a tenfold excess concentration of unlabeled pIgA blocked >98% of the binding of FITC-mIgA. *P < 0.0001, compared with IgA binding in leukocytes incubated with FITC-IgA alone (0.5 mg/ml). CTL, control; Ab, antibody.
Binding of FITC-labeled human mIgA or pIgA to HL-60 or U937 cells was not inhibited by preincubation with 5 mg/ml human IgG, IgM, orosomucoid, or asialo-orosomucoid (Figure 6). Inhibitory studies using a specific FcαR1-blocking antibody (clone My43) (17) revealed distinctly different leukocyte binding by mIgA and pIgA. Preincubation with My43 at a concentration of 50 μg/ml blocked 82 and 83% of FITC-mIgA binding to granulocytic HL-60 cells and monocytic U937 cells, respectively. In contrast, preincubation with the same concentration of My43 blocked only 17 and 8% of FITC-pIgA binding to granulocytic HL-60 cells and monocytic U937 cells, respectively.
Figure 6. Inhibitory studies on the binding of mIgA and pIgA to granulocytic HL-60 and monocytic U937 cells with no surface Ig and no previous exposure to serum proteins. A specific FcαR1-blocking antibody (My43) at a concentration of 50 μg/ml blocked ≥80% of mIgA binding to either leukocyte subpopulation but reduced pIgA binding by ≤20%. IgG, IgM, orosomucoid (Oroso), and asialo-orosomucoid (ASOR) were unable to block mIgA or pIgA binding to these leukocytes. *P < 0.0001, compared with IgA binding to leukocytes incubated with FITC-IgA alone (0.5 mg/ml). CTL, control; Ab, antibody.
Determination of κ- and λ-IgA1 Concentrations Bound to Leukocytes and κ/λ Light Chain Ratios
The amounts of κ-IgA1 and λ-IgA1 in the appropriately adjusted mIgA and pIgA samples (from patients or control subjects) bound to granulocytic HL-60 cells or monocytic U937 cells were analyzed by flow cytometry. The IgA concentrations were measured from the fluorescence channel numbers by using standard curves, as described previously (11), and κ/λ ratios were then determined.
Whereas 92% of the IgA1 in the JBP in the binding study was mIgA1, pIgA1 amounted to 57 and 56% of total IgA1 bound to granulocytic HL-60 cells and monocytic U937 cells, respectively, indicating that pIgA1 had a higher affinity for leukocytes than did mIgA1. To determine whether there was selective binding of IgA to leukocytes, the κ/λ ratios of IgA1 in the incubated IgA samples were compared with the κ/λ ratios of the IgA1 bound to leukocytes (Figure 7). The κ/λ ratios of mIgA1 or pIgA1 bound to granulocytic HL-60 cells or monocytic U937 cells were significantly lower than the κ/λ ratios of mIgA1 or pIgA1 in the incubated IgA samples (P < 0.0001), suggesting preferential binding of λ-IgA1 (particularly polymeric forms).
Figure 7. Comparison of the κ/λ ratios of IgA1 (mIgA1 or pIgA1) in adjusted JBP samples and the κ/λ ratios of IgA1 (mIgA1 or pIgA1) bound to leukocytes. There were significant differences between the κ/λ ratios of IgA1 (mIgA1 or pIgA1) in the adjusted JBP samples and the κ/λ ratios of IgA1 (mIgA1 or pIgA1) bound to leukocytes, but a difference between patients and healthy control subjects was not observed. Open circles and vertical bars represent the mean and SD of the values.
Discrepancies between κ/λ Ratios of IgA Bound to Cultured HL-60 or U937 Cells and Circulating Leukocytes
Our study demonstrated discrepancies between the κ/λ ratios of IgA bound to cultured HL-60 or U937 cells versus circulating leukocytes. For granulocytic HL-60 cells or monocytic U937 cells with no surface Ig and no previous exposure to human serum proteins, the κ/λ ratios of mIgA and pIgA bound to the cell surface after incubation with IgA preparations averaged 0.70 ± 0.04 for mIgA and 0.77 ± 0.03 for pIgA (Figure 7). These ratios were significantly lower than those of endogenous IgA bound to circulating granulocytes or monocytes (Figure 3), suggesting that leukocytes in circulation demonstrated reduced surface λ-IgA, compared with “native” leukocytes, after their initial binding to plasma IgA. The changes in κ/λ ratios could result from selective internalization of λ-IgA into the cytoplasm or migration/sequestration of leukocytes with predominant λ-IgA in the mononuclear phagocytic system or inflammatory tissues. After incubation at 37°C for 1 h, internalization of mIgA and pIgA by HL-60 and U937 cells was evident by a decrease in the MFI of surface-bound IgA (Figure 8) and the finding of intracytoplasmic IgA in these cells (Figure 9). Both κ- and λ-IgA were detected in the cytoplasm after internalization (data not shown). The κ/λ ratios of surface-bound IgA remained unchanged before and after internalization, indicating that selective intracytoplasmic uptake of λ-IgA by these leukocytes was not likely (Figure 8).
Figure 8. The κ/λ (⋄) ratio and MFI (♦) of IgA1 (mIgA1 or pIgA1) bound to HL-60 and U937 cells. The granulocytic HL-60 cells and monocytic U937 cells had no surface Ig and had not been previously exposed to human serum proteins. After initial binding at 4°C, internalization occurred with further incubation at 37°C for 1 h. The internalization was accompanied by a reduction in IgA on the cell surface, but the κ/λ ratios remained unchanged.
Figure 9. Detection of intracytoplasmic mIgA and pIgA in granulocytes and monocytes after internalization. After initial binding of the mIgA or pIgA preparation at 4°C, internalization occurred with further incubation at 37°C for 1 h. Leukocytes were incubated, cytocentrifuged, fixed and permeabilized with 95% ethanol/5% acetic acid at −20°C, and stained with FITC-conjugated anti-human IgA F(ab′)2. Control experiments with preimmune rabbit Ig F(ab′)2 demonstrated no staining for IgA.
Discussion
The pathogenesis of IgAN remains unclear, but mesangial IgA deposits are detected early in allografts from healthy donors with no histologic evidence of IgAN (18). These findings tend to support the idea that the primary immunologic defect lies in the circulating IgA, rather than the kidney. In human subjects, the daily production of total IgA is greater than the production of all other Ig classes combined, but plasma IgA levels are less than one-fourth of plasma IgG levels. Plasma IgA levels reflect rapid turnover and catabolism of IgA in the human body. In human subjects, only negligible amounts of the total IgA produced in the bone marrow, spleen, and lymph nodes reach external secretion (19). Therefore, most of the IgA from the circulating pool is internally catabolized. The catabolism of IgA depends on two factors, i.e., IgA receptors and the immunochemical nature of the IgA molecules. There are three known IgA receptors, namely the FcαR1, the asialoglycoprotein receptor (ASGPR), and the pIg receptor (pIgR). FcαR1 and ASGPR are main candidate receptors in IgA catabolism and the clearance of IgA immune complexes from the circulation. FcαR1 binds IgA1 and IgA2 via the CH2 and CH3 domains of the Fc region (20), whereas ASGPR binds IgA1 via the terminal galactose of the O-glycans at the hinge region or via N-acetylgalactosamine present on any glycoprotein as a terminal carbohydrate. The former is expressed mainly by neutrophils, monocytes, macrophages, and eosinophils, and the latter is localized in hepatocytes. pIgR mediates the transepithelial transport of pIg (particularly pIgA) and is detected exclusively in human secretory epithelia. More importantly, recent data failed to document these IgA receptors in human mesangial cells (21). In vitro studies demonstrated that U937 cells bound more IgA1 than did hepatocytes or isolated glomeruli, and infusion studies in rodents revealed greater uptake by the liver than by the kidney or the spleen (7). With the vast number of leukocytes in the circulation and in tissues, the nature and concentrations of IgA in the circulation and in other organs are likely to be determined by the pool of IgA bound to leukocytes and myeloid cells. It is also thought that the expression of IgA receptors on and the binding of IgA to circulating leukocytes might affect the mesangial uptake of IgA in IgAN (9). Reduced binding to FcαR1 of mIgA from patients with IgAN was observed in a murine B cell line expressing myeloid IgA Fc receptors (22). In addition, structural changes in the hinge region of the IgA1 molecule might alter its uptake by different tissues in IgAN (23,24). Deficiencies of terminal galactose in the hinge region or reduction of N-acetylgalactosamine present on any glycoprotein as a terminal carbohydrate might have profound effects on the ASGPR recognition of IgA1 (and thus its catabolism). Oligosaccharides in the hinge region contain negatively charged sialic acid, which is large and bulky, compared with the protein backbone. Any change in the carbohydrate moieties would affect the tertiary structure as well as the electrostatic charge, both of which are pivotal in interactions with and recognition by other molecules, such as ASGPR and FcαR1 (24,25).
In this study, we examined the IgA receptor expression and IgA binding of circulating leukocytes in IgAN, because endogenous IgA binding to circulating leukocytes might affect the mesangial uptake of IgA (9). Previous studies demonstrated that plasma IgA concentrations always exceeded the IgA affinity of FcαR1 on purified polymorphs; therefore, these receptors would be expected to be occupied by IgA in the circulation (14). However, Stewart and Kerr (14) detected no cytophilic IgA on freshly purified polymorphs, because of endocytosis from the cell surface. Using flow cytometry, we detected IgA on the cell surface in 4.5 to 5.5% of circulating leukocytes. Patients with IgAN exhibited higher percentages of granulocytes and monocytes with bound IgA, compared with healthy control subjects. In comparisons of fluorescence intensity, the levels of endogenous κ-IgA and λ-IgA bound to the surface of either granulocytes or monocytes from patients with IgAN were higher than values for healthy control subjects. Interestingly, conflicting observations on FcαR1 reactivity on blood cells from patients with IgAN were recorded (8,9,26). Grossetete et al. (9) attributed the contradicting findings from their laboratory to crosslinking by a secondary antibody in indirect immunofluorescence studies. Using the same antibody against FcαR1 (A59) for direct immunofluorescence studies, we failed to confirm downregulation of FcαR1 on leukocytes in IgAN. Instead, we observed increased percentages of granulocytes and monocytes expressing FcαR1 among patients with IgAN, as reported by others (8,26). In both groups of subjects, we failed to establish a negative correlation between plasma IgA levels and the expression of FcαR1 on granulocytes or monocytes, as reported previously (9). Our study confirmed the previous finding that most of the FcαR1 on leukocytes were not occupied by endogenous IgA (9). After incubation at 4°C with IgA isolated from patients or control subjects (at comparable concentrations), further binding to IgA was observed with blood leukocytes from the same subjects. More importantly, granulocytes but not monocytes from patients with IgAN exhibited greater binding capacities for exogenous mIgA and pIgA, compared with cells from healthy control subjects. The reason for the discrepancy between the binding of exogenous IgA to granulocytes and monocytes is not clear; one explanation for the failure to demonstrate significant differences between patients and control subjects could involve the smaller number of monocytes (only one-tenth of granulocyte counts). Another possibility is a difference in binding affinity, because FcαR1 transcripts in monocytes from patients with IgAN have an insertion sequence, suggesting the existence of a new exon that is absent in granulocytes from patients with IgAN (27). Intriguingly, isolated blood leukocytes bound more pIgA than mIgA, although the concentration of mIgA in the IgA preparations was ten times that of pIgA. It could be argued that, if mIgA or pIgA binds to the myeloid leukocytes only via FcαR1, then predominant pIgA binding is not expected. Our findings and previous data on human circulating mononuclear cells (28) raise the possibility of a mechanism other than FcαR1 via which pIgA binds to myeloid leukocytes.
To circumvent the difficulties attributable to endogenous IgA on the leukocyte surface, which could alter FcαR1 expression (9,15,16), in the exploration of other mechanisms of IgA binding, granulocytes or monocytes derived from the HL-60 or U937 cell lines were used in our experiments. These cells, which did not exhibit any surface Ig and had not previously been exposed to serum proteins, readily expressed FcαR1 activity, with higher affinity for pIgA than for mIgA. The involvement of a binding mechanism other than FcαR1 in pIgA uptake by leukocytes was supported by inhibition studies using high concentrations of unlabeled proteins. Although the binding of FITC-mIgA by either granulocytic HL-60 cells or monocytic U937 cells was almost completely inhibited by unlabeled mIgA or pIgA, a similar inhibitory effect on the binding of FITC-pIgA was not observed with unlabeled mIgA. Our hypothesis was confirmed by experiments using a specific FcαR1-blocking antibody (clone My43). At a concentration that blocked ≥80% of mIgA binding to either leukocyte subpopulation, pIgA binding was reduced by ≤20%. The lack of inhibition by IgM, orosomucoid, or asialo-orosomucoid confirmed our previous studies demonstrating the absence of ASGPR and pIgR in blood leukocytes (21). Further studies of the recently reported potential mechanisms/receptors for IgA binding, including electrostatic charge (29), the Fcα/μ receptor (30), and the transferrin receptor (31), are warranted. The first two are mainly involved in the binding of pIgA, and the transferrin receptor predominantly binds mIgA.
We previously demonstrated that the mesangial IgA deposits in IgAN were predominantly pIgA1, with the λ light chain (3,11). In this study, we observed lower κ/λ ratios for mIgA1 and pIgA1 bound to granulocytes (isolated from blood or grown from the HL-60 line) or monocytes (isolated from blood or grown from the U937 line), compared with the κ/λ ratios of mIgA1 and pIgA1 in incubated IgA samples. The findings of preferential binding of polymeric λ-IgA1 in particular were similar to those in human mesangial cells (11). We observed significantly lower κ/λ ratios for IgA bound to cultured HL-60 or U937 cells, compared with those of circulating leukocytes, suggesting that circulating leukocytes had reduced surface λ-IgA, compared with native leukocytes, after their initial binding to plasma IgA. The changes in κ/λ ratios might result from selective internalization of λ-IgA or migration/sequestration of leukocytes with predominant λ-IgA in the reticulophagocytic system or inflammatory tissues. In contrast to the findings of Grossetete et al. (9), we documented internalization of IgA (mIgA and pIgA) in granulocytes and monocytes. Selective intracytoplasmic uptake of λ-IgA by leukocytes was not likely, because the κ/λ ratios of surface-bound IgA remained unchanged before and after internalization. These findings have important pathologic significance in IgAN, suggesting the migration and/or sequestration of leukocytes with predominant λ-IgA in the mononuclear phagocytic system or inflammatory tissues. Prominent infiltration of mononuclear cells is frequently detected in renal biopsies from patients with IgAN (10,32,33). The glomerular and interstitial infiltration by neutrophils increases with the histopathologic severity. One key factor responsible for attracting neutrophils to the kidney in IgAN is interleukin-8 (IL-8). Infiltrating mononuclear cells expressing IL-8 were detected in kidney tissues from patients with IgAN, and most were labeled with anti-monocyte/macrophage antibody (34). Data reported by Wada et al. (35) support the pivotal role of IL-8 in attracting neutrophils to the kidney in IgAN. IgA is involved in the initiation of intracellular oxidative metabolism in neutrophils (36). The production of reactive oxygen species (37) and the release of lysosomal proteolytic enzymes (38) by infiltrating “activated” neutrophils cause direct toxic effects on endothelial and mesangial cells in the glomeruli.
In conclusion, our findings revealed that the pool of IgA bound to leukocytes and myeloid cells determines the nature and concentration of IgA in the circulation and in other organs. Increased FcαR1 expression on and IgA binding to circulating leukocytes might affect the mesangial uptake of IgA in IgAN. A significant proportion of pIgA is bound to leukocytes not via known IgA receptors. Neutrophils primed by polymeric λ-IgA might contribute to the glomerulointerstitial injury in IgAN, in the presence of leukocytic infiltration.
Acknowledgments
This study was supported by the Research Grant Council of Hong Kong (Grant HKU 7263/01M). Loretta Chan was supported by Fresenius Medical Care, and Hong Guo was supported by the Ivy Wu Fellowship at the University of Hong Kong.
- © 2002 American Society of Nephrology
References
