Crucial Role of AIM/CD5L in the Development of Glomerular Inflammation in IgA Nephropathy

Significance Statement

Apoptosis inhibitor of macrophage (AIM) protein is involved in various diseases. It removes dead cells in urinary tubules, which aids recovery in AKI. Moreover, dysfunctional AIM scavenging in macrophages is a well known cause of feline renal failure. This study clarified the role of AIM in IgA nephropathy (IgAN) in a mouse model. AIM-deficient IgAN mice (AIM^−/−gddY mice) showed IgA deposition similar to that of wild-type gddY mice but did not have glomerular accumulation of IgM/IgG/complement and subsequent regional inflammation, avoiding glomerular sclerosis, proteinuria, and hematuria. IgM/IgG-IgA immune complex formed by recombinant AIM restored the IgAN phenotype. Elucidating the role of AIM in IgAN may facilitate development of new IgAN therapies.

Visual Abstract

Abstract

Background

IgA nephropathy (IgAN) begins with aberrant IgA deposition in glomeruli, progresses to IgM/IgG/complement codeposition, and results in chronic inflammation and glomerular damage. However, the mechanism that drives such phlogogenic cascade has been unclear. Recently, apoptosis inhibitor of macrophage (AIM) protein was shown to modulate macrophages’ function in various pathologic conditions, thereby profoundly affecting the progression of renal disorders, including AKI. A spontaneous IgAN model, grouped ddY (gddY) mouse, revealed the requirement of AIM for the overall inflammatory glomerular injury following IgA deposition.

Methods

We established an AIM-deficient IgAN model (AIM^−/−gddY) using CRISPR/Cas9 and compared its phenotype with that of wild-type gddY with or without recombinant AIM administration. An IgA-deficient IgAN model (IgA^−/−gddY) was also generated to further determine the role of AIM.

Results

In both human and murine IgAN, AIM colocalized with IgA/IgM/IgG in glomeruli, whereas control kidneys did not exhibit AIM deposition. Although AIM^−/−gddY showed IgA deposition at levels comparable with those of wild-type gddY, they did not exhibit glomerular accumulation of IgM/IgG complements, CD45⁺ leukocyte infiltration, and upregulation of inflammatory/fibrogenic genes, indicating protection from glomerular lesions and proteinuria/hematuria. Recombinant AIM administration reconstituted the IgAN phenotype, resulting in IgM/IgG/complement IgA codeposition. Neither spontaneous IgM/IgG codeposition nor disease was observed in IgA^−/−gddY mice.

Conclusions

AIM may contribute to stable immune complex formation in glomeruli, thereby facilitating IgAN progression. Therefore, AIM deposition blockage or disassociation from IgM/IgG may present a new therapeutic target on the basis of its role in IgAN inflammation initiation.

IgA nephropathy (IgAN) was first described >50 years ago and is the most common causative disease of primary chronic GN worldwide. It has a poor prognosis, and 30%–40% of patients with IgAN progress to ESKD within 20 years of diagnosis.^1–3 One of the most important histologic characteristics broadly observed in the kidney of patients with IgAN is the deposition of polymeric IgA1 that forms an immune complex together with variable IgG and IgM, as well as complement 3 (C3) at the glomerular mesangial region.^4–9 Immunodeposits are presumed to cause persistent inflammatory immune responses, resulting in the proliferation of mesangial cells, macrophage infiltration, and matrix expansion and leading to further glomerular damage. The mechanisms underlying the development of the disease after IgA deposition are not yet fully understood, and no effective disease-specific treatments are established.

The grouped ddY (gddY) mouse is one of the best animal models for human IgAN. We previously established the gddY model using ddY mice that spontaneously exhibited mild IgAN-like clinical symptoms using selective intercrossing for >20 generations.¹⁰ All gddY mice develop proteinuria by 8 weeks of age followed by glomerular sclerosis and obvious renal failure with elevated serum creatinine levels by 24 weeks of age. Human patients with IgAN and gddY mice share many pathologic features, including glomerular deposits of an IgA-containing immune complex. Macrophages/monocytes are also known to be involved in IgAN pathogenesis.^11,12 We ascertained this in gddY mice, in which we found Gr1⁻ and CD115⁺ monocytosis in the peripheral blood. Administration of an anti-APRIL neutralization antibody, which reduced glomerular IgA deposition, could suppress monocytosis, confirming the involvement of macrophages/monocytes in IgAN progression.^13,14

An apoptosis inhibitor of macrophage (AIM; AIM/CD5L, encoded by the cd5l gene) is produced by resident tissue macrophages and was initially identified as an apoptosis inhibitor that supports the survival of macrophages against apoptosis-inducing stimuli.¹⁵ It is involved in various diseases, including arteriosclerosis, dyslipidemia, and autoimmune disorders, and acts by regulation of macrophages.^16–20 In the blood, AIM associates with IgM pentamers, which protect it from renal excretion and maintain high levels of circulating AIM (approximately 5 μg/ml in humans and 2 μg/ml in mice).¹⁶ Recently, we clarified the method by which AIM binds to an IgM pentamer through single-particle negative stain electron microscopy for direct observation of the complex. We found that the bona fide shape of IgM pentamer is a hexagon with one piece missing, wherein AIM stably associates via a disulfide bond and a charge-based interaction.²¹ Interestingly, during AKI, AIM dissociates from IgM pentamers in the blood. The IgM-free AIM passes through the glomerular membrane and accumulates on AKI-associated intraluminal dead cell debris that obstructs renal proximal tubules. Deposited AIM then promotes phagocytic removal of the debris, contributing to overall kidney tissue repair.²²

The involvement of AIM in kidney diseases via macrophages as well as its affinity for IgM prompted investigation into its possible roles in IgAN pathogenesis. The purpose of this study was to elucidate the role of AIM in the inflammatory cascade underlying IgAN.

Methods

Mice

The gddY mice were established by selective mating of early-onset ddY mice for >20 generations.¹⁰ AIM-deficient and IgA-deficient gddY mice were generated using the CRISPR/Cas9 system.²³ The plasmid pX330 (Addgene, Watertown, MA) with the guide sequence of the target gene inserted was microinjected into fertilized embryos of gddY mice. Mutations in the gene of interest were confirmed by examining DNA sequences of PCR-amplified fragments that contained the targeted area sequence. Homozygous knockout mice were obtained by intercrossing of heterozygous mice. Therefore, 23 bp including the start codon and 7 bp including the beginning of the Cα region were deleted in AIM-deficient and IgA-deficient gddY mice, respectively. AIM and IgA were absent from serum samples of the respective homozygous mice as confirmed by western blotting.

BALB/c mice were purchased from Sankyo Labo Service Corporation Inc. All mice were maintained at the animal facility of Juntendo University on a regular chow (Oriental Yeast, Tokyo, Japan) and water ad libitum in a specific pathogen–free room. The experimental protocol of this study was approved by the Ethics Review Committee for Animal Experimentation of Juntendo University Faculty of Medicine (permit no. 270110).

Human Subjects

Renal biopsy specimens were obtained at Juntendo University Hospital with the informed consent from patients and approval of the Research Ethics Review Committees of Juntendo University Hospital.

Generation of Anti-AIM Antibodies

We obtained a recombinant baculovirus encoding AIM protein plus 6xHistidine-Tag by homologous recombination with the pBlueBac4.5 transfer vector (V1995–20; Invitrogen, Waltham, MA) into which we subcloned an AIM cDNA fragment encoding the entire open reading frame. We transfected the recombinant baculovirus into egg cells of Trichoplusia (HighFive cells, B855–02; Invitrogen), and we purified recombinant apoptosis inhibitor of macrophage (rAIM) protein from the conditioned medium using a histidine-tag affinity column. We immunized rabbits four times with the purified AIM protein.¹⁵ Polyclonal anti-AIM serum has been used for many years in many publications.^16–18,22 Its specificity was fully confirmed by the fact that this antibody does not stain tissues and blood collected from AIM-deficient mice either by immunohistochemistry or by immunoblotting. As demonstrated in Supplemental Figure 1, the antibody did not crossreact with IgA, IgG, and IgM, and the antibody showed no recognition of any type of Ig.

Antibodies and Reagents

Antibodies and reagents used for histologic and biochemical analyses were as follows: primary antibodies included goat anti-mouse IgG (ab175661; Abcam, Cambridge, United Kingdom), goat anti-mouse IgA (ab97013; Abcam), goat anti-mouse IgM (ab175702; Abcam), horseradish peroxidase (HRP)–conjugated goat anti-mouse IgG-Fc antibody (A90–131P; Bethyl Laboratories, Montgomery, AL), HRP-conjugated goat anti-mouse IgM (A80–100P; Bethyl Laboratories), HRP-conjugated goat anti-mouse IgA (14–18–0; KPL, Milford, MA), rabbit anti-AIM (rab2 rabbit polyclonal for mice and human; customized in our laboratory), rat anti-mouse C3 (SC58926; Santa Cruz, Dallas, TX), rabbit anti-mouse C5b9 (bs-2673R-A488; Bioss, Woburn, MA), rat anti-mouse CD45 (550539; BD Pharmingen, Franklin Lakes, NJ), rabbit anti-human IgA (F0316; Dako, Santa Clara, CA), goat anti-human IgM (218; MBL, Nagoya, Japan), HRP-conjugated goat anti-human IgM (A80–100P), HRP conjugated with goat anti-rabbit IgG (32460; Thermo Fisher Scientific, Waltham, MA), goat anti-human IgG (109–476–098; Jackson, West Grove, PA), goat anti-human IgM (109–476–098), rabbit anti-human C3 (F0201; Dako), and rabbit anti-human C5b-9 (bs-2673R-A488). Secondary antibodies and related reagents included Alexa 488–conjugated anti-rabbit IgG (A21206; Invitrogen), Alexa 488–conjugated anti-rat IgG (A11006; Invitrogen), Alexa 555–conjugated anti-rabbit IgG (A31572; Invitrogen), Alexa 555–conjugated anti-rat IgG (A21434; Invitrogen), HRP-conjugated goat anti-rabbit IgG (32460), blocking solution (customized in our laboratory; 0.2% fish gelatin and 2% BSA with PBS), and Fluoromount (K024; Diagnostic Biosystems, Pleasanton, CA). Specimens were analyzed using confocal laser microscopy (FV-1000; Olympus Corporation, Tokyo, Japan).

Evaluation of Proteinuria, Serum Creatinine, and BUN Levels

Urinary albumin was measured by immunoassay (DCA 2000 system; SIEMENS Healthcare Diagnostics, Erlagen, Germany). Hematuria was examined by Uro-lab sticks SGL (2321; SIEMENS Healthcare Diagnostics). Serum creatinine and BUN levels were measured using an autoanalyzer (Fuji Dry-Chem 5500; Fujifilm, Tokyo, Japan).

AIM^−/−gddY Analysis and rAIM Injection Model

Serum, urine, and kidney samples were collected from AIM^−/−gddY, and the phenotype of the disease was compared with that in gddY mice. rAIM was purified as previously described²² and injected into the tail vein of AIM^−/−gddY mice; the phenotypes were then analyzed.

Histologic Analyses

For light microscopy, renal specimens were fixed in 15% formaldehyde and embedded in paraffin; 3-μm sections were then collected. Tissue sections were stained with periodic acid–Schiff and Azan stains. For the quantitative analysis of glomerular sclerosis, 20 glomeruli per cross-section were observed, and the following scores were assigned using a modified method on the basis of the report by Raij et al.²⁴: zero points, no glomerular sclerosis; one point, mild glomerular sclerosis (approximately 25%); two points, moderate glomerular sclerosis (approximately <50%); and three points, severe glomerular sclerosis (approximately >50%). Sclerosis scores were calculated as follows: [Σ(each score × number of glomeruli)]/20.

For fluorescent immunohistochemistry (mouse and human), renal specimens were mounted in OCT compound (4583; Sakura Fine Tek, Tokyo, Japan) and stored at −80°C. Specimens were sliced in 4-μm sections and fixed with acetone at −20°C for 4 minutes. Slides were blocked in a blocking solution (customized in our laboratory; described earlier in this paper) and then incubated with a primary antibody for 1 hour at room temperature or overnight at 4°C; slides were subsequently incubated with an appropriate fluorophore-conjugated secondary antibody for 30 minutes at room temperature. The specimens were subjected to analysis using confocal microscopy (FV10; Olympus). For quantitative analysis of CD45⁺ cells, the numbers of positive cells were counted for 15 glomeruli in 8- and 16-week-old gddY and AIM^−/−gddY mice, and the average number of positive cells per glomerulus was analyzed (n=5).

Quantitative PCR Assay

Glomeruli were isolated by the differential sieving method.²⁵ RNA was isolated using the QIAgen RNeasy kit (74104; Qiagen, Düsseldorf, Germany) in accordance with the manufacturer’s specifications. cDNA was generated using Taqman Fast Advanced Master Mix (4444556; Thermo Fisher Scientific). TaqMan probes were used for measurement of the expression of cd5l/Aim (Mm00437567_m1), IL-1β (Mm00434228_m1), IL-6 (Mm00446190_m1), Ccl2/MCP-1 (Mm00441242_m1), TNF (Mm00443258_m1), TGF-β (Mm01178820_m1), Acta2/αSMA (Mm01546133_m1), and collagen type I α2/COL1α2 (Mm00483888_m1). mRNA expression was adjusted relative to the expression of Gapdh (Mm99999915_g1). All probes were purchased from Life Technologies (Carlsbad, CA). The quantitative evaluation of mRNA was performed by the ΔΔCT method using a 7500 Fast Real-Time PCR system (Invitrogen).

Immunoprecipitation

We coupled rabbit anti-mouse IgA (GTX20289; GENETEX) and rabbit monoclonal anti-human IgA Ab (ACM-ab124716; Abcam) to CNBr-activated Sepharose 4B (17043001; GE Healthcare; Chicago, IL), and we immunoprecipitated the serum of each mouse strain, human IgAN, and healthy control. Further, we eluted samples by heating them at 100°C for 5 minutes with 2× SDS buffer and performed immunoblotting for anti-IgA, anti-IgM, and anti-AIM.

Statistical Analyses

Results are presented as mean ± SEM; n defines the number of biologic replicates. Differences between groups were assessed by unpaired two-tailed t test or two-way ANOVA. P values <0.05 were considered statistically significant. All analyses were performed using Prism software (version 7.0; GraphPad Inc., La Jolla, CA).

Results

AIM Codeposited with IgA in Murine and Human IgAN

In gddY mice, IgA codeposition with IgG and IgM and subsequent complement activation occur in glomeruli around 4–8 weeks of age; this results in chronic and persistent inflammation and glomerular lesions.¹⁰ In addition to the involvement of macrophages in the development of IgAN, marked accumulation of AIM was observed at glomeruli, colocalizing with IgA in gddY mice (Figure 1A). AIM deposition in glomeruli was present at multiple ages (multiple stages during the disease process) (Supplemental Figure 2). AIM staining in the glomeruli did not correspond to AIM expression by mesangial cells because no cd5l/Aim mRNA was observed in glomeruli isolated from gddY mice, but it was observed in the liver, which was the positive control. Thus, AIM in the glomerular region was derived from circulating AIM. AIM deposition in glomeruli was rarely detected in BALB/c mice compared with age-matched gddY mice (Figure 1A). Similar to our observations in gddY mice, AIM accumulation in glomeruli colocalized with IgA in human patients with IgAN but was not observed in kidneys from patients with minimal change nephrotic syndrome or membranous nephropathy (Figure 1B).

Figure 1.

No glomerular lesion was observed in AIM^−/−gddY mice. (A) Representative immunostaining images of IgA (red) and AIM (green) in the renal tissue of 8-week-old gddY, AIM^−/−gddY, and BALB/c mice (n=5). Scale bars: 50 μm. (B) Representative immunostaining images of IgA (red) and AIM (green) in the renal tissue of human patients with IgAN, membranous nephritis (MN), or minimal change nephritic syndrome (MCNS; n=4). Scale bars: 100 μm. (C) Serum BUN (s-BUN) and serum creatinine (s-Cre) levels (milligrams per deciliter) and urinary protein levels (micrograms per day) in 8- and 16-week-old gddY and AIM^−/−gddY mice (n=8) All data are expressed as mean±SEM and were analyzed using two-way ANOVA. *P<0.001; **P=0.01.

No Glomerular Lesion Was Observed in AIM^−/−gddY Mice

Because of the aberrant AIM deposition in glomeruli in human and murine IgAN, we investigated whether glomerular AIM was involved in IgAN pathogenesis: that is, whether AIM may be required for IgA deposition or may influence the progressive tissue destruction following IgA deposition. To assess this, we directly targeted the cd5l/Aim gene in gddY mice by using a CRISPR/Cas9 system. AIM-deficient (AIM^−/−) gddY mice harbored a 23-bp deletion around the ATG start codon in the cd5l/Aim gene, which resulted in a lack of AIM protein in the serum (Supplemental Figure 2). At 8 weeks of age, both wild-type (WT) and AIM^−/−gddY mice showed similar levels of IgA deposition in glomeruli, suggesting that AIM was not necessary for disease-initiating IgA deposition (Figure 1A). However, levels of proteinuria (8 weeks of age, P<0.001; 16 weeks of age, P=0.01) as well as the increased serum levels of serum creatinine and BUN (8 and 16 weeks of age, P<0.001) were significantly lower in AIM^−/−gddY mice, whereas they were prominent in WT gddY mice (Figure 1C).

Glomerular IgA Deposition Did Not Cause Glomerular IgM and IgG Codeposition or Complement Activation in AIM^−/−gddY Mice

We next assessed why AIM^−/− protected against renal injury, despite glomerular IgA deposition. Surprisingly, immunofluorescence analysis revealed that IgM, IgG, and C3 depositions were largely absent in AIM^−/−gddY mice (Figure 2A). Accordingly, the development of C5b–9, a membrane attack complex, was apparent in the glomerular region in WT gddY mice but not in AIM^−/−gddY mice at 8 and 16 weeks of age (Figure 2B) (8 weeks of age, P=0.05; 16 weeks of age, P=0.01). Consistent with these differences in immune responses in the glomerular region, WT gddY mice exhibited obvious glomerular injury by 8 weeks of age characterized by CD45⁺ leukocyte infiltration (P=0.05), glomerular sclerosis (P=0.05), interstitial fibrosis, mesangial cell proliferation, and extracellular matrix expansion, whereas AIM^−/−ddY mice did not exhibit these pathologic phenotypes (Figure 3, A and B). In accordance with the histologic observations, mRNA levels of inflammatory and fibrogenic genes expressed in glomeruli were significantly lower in AIM^−/− than in WT gddY mice when assessed by quantitative PCR (Figure 3C). Thus, although IgA deposition occurred as in WT gddY mice, subsequent inflammatory reactions and tissue destruction were markedly prevented in AIM^−/−gddY mice.

Figure 2.

Glomerular IgA deposition did not cause glomerular IgM and IgG codeposition or complement activation in AIM^−/−gddY mice. (A) Representative immunostaining images of IgA (green), IgG (blue), IgM (red), and C3 (green) in the renal tissue of 16-week-old gddY and AIM^−/−gddY mice (n=5). (B) Average percentage of membrane-attack complex; MAC (C5b-9)–positive area per glomerulus in 8- and 16-week-old gddY and AIM^−/−gddY mice (five glomeruli were analyzed per mouse; n=5; *P<0.05; **P<0.01). Scale bars: 50 μm.

Figure 3.

Glomerular sclerosis and inflammation were not induced without AIM. (A) Periodic acid–Schiff (PAS) staining showed less glomerular sclerosis and Azan staining revealed fewer glomerular and tubulointerstitial fibrotic changes in 8-week-old AIM^−/−gddY mice. Sclerosis scores were assigned in accordance with the method used in the previous report (15) (*P=0.05). Scale bars: 20 μm (PAS); 100 μm (Azan). (B) CD45⁺ cells were counted for 15 glomeruli in 8- and 16-week-old gddY and AIM^−/−gddY mice, and the average number of positive cells per glomerulus was analyzed (n=5). Data are expressed as mean ± SEM calculated using unpaired two-tailed t test. *P=0.01; **P=0.05. (C) Quantitative PCR analysis for mRNA expression of various inflammatory and fibrotic genes in gddY (white dots) and AIM^−/−gddY (black dots) mice (8 and 16 weeks old; n=3 per group). All data are expressed as mean ± SEM and were analyzed using unpaired two-tailed t test (B and C) or two-way ANOVA. *P=0.05; **P=0.01; ***P<0.001.

rAIM Restored Glomerular IgM/IgG Codeposition and Glomerular Damage in AIM^−/− Mice

The sole importance of AIM for the completion of the disease was determined by the following experiment. rAIM or PBS control was administered intravenously to AIM^−/−gddY mice (8 weeks old, n=8 per group), and the glomeruli were analyzed 2 hours later. As shown in Figure 4A, rAIM injection promoted massive accumulation of IgM, IgG, and C3 colocalized with AIM; this was not observed in the PBS control mice. Accordingly, proteinuria and hematuria were also significantly induced (Figure 4B) (P<0.001). These findings indicate that glomerular IgA deposition alone does not induce IgAN and that AIM is essential for binding of previously deposited IgA with IgM or IgG, leading to progression of tissue destruction. The codeposition of AIM with IgM, IgG, C3, and C5b-9 in human IgAN glomeruli also supports our hypothesis (Figure 5A, Supplemental Figure 4).

Figure 4.

rAIM restored glomerular IgM/IgG codeposition and glomerular damage in AIM^−/− mice. (A) Representative renal tissue immunostaining images at 2 hours after intravenous administration of 100 μg rAIM or PBS to AIM^−/−gddY mice (n=8 per group): IgG (blue), IgM (red), AIM (green), and C3 (green). Scale bars: 50 μmin the upper panels; 20 μm in lower panels. (B, left panel) The percentage ratio and percentage of mice with hematuria and proteinuria within 4 hours after the administration of rAIM or PBS. (B, right panel) Peak levels of hematuria before rAIM and PBS administration and within 4 hours after administration. All data are expressed as mean ± SEM. *P<0.001; two-way ANOVA. A/C, alubmin/creatinine.

Figure 5.

AIM was involved in IgA/IgM/IgG immune complex formation in human patients with IgAN. (A) Representative immunostaining images of IgG (blue)/IgM (red)/AIM (green) and C3 (green)/AIM (red) in the renal tissue of human patients with IgAN (n=4). Furthermore, we assessed whether spontaneous glomerular deposition of AIM or IgM/IgG occurred in the absence of IgA by evaluating IgA^−/−gddY mice. (B) Representative immunostaining images of IgG (green), IgM (red), and AIM (green) in the renal tissue of 8-week-old gddY and IgA^−/−gddY mice (n=5). Scale bars: 50 μm in A; 20 μm in B.

AIM Deposition Did Not Occur without Glomerular IgA Deposition

Furthermore, we assessed whether the spontaneous glomerular deposition of AIM or IgM/IgG occurred in the absence of IgA through direct disruption of the IgA gene in gddY mice via CRISPR/Cas9 (IgA^−/−gddY). As shown in Figure 5B, no deposition of IgM, IgG, or AIM was observed in the glomeruli of IgA^−/−gddY mice, clearly indicating that aberrant IgA deposition triggers the disease. In addition, it is likely that the formation of an immune complex of IgA, IgM, IgG, and AIM may primarily occur in situ because this immune complex with AIM was not detected in the serum of gddY mice either by immunoprecipitation or by immunoblotting of the IgM fraction (Figure 6, Supplemental Figure 5).

Figure 6.

No immune complex formation of IgA and IgM via AIM was detected in serum samples from gddY mice. Immunoprecipitation of a serum sample from gddY with anti-IgA (we used the serum of BALB/c and AIM^−/−gddY as controls) was performed, and immune complex was analyzed by immunoblotting with anti-IgM (left panel), AIM (center panel), and IgA (right panel) under nonreducing conditions. WB, western blotting.

Discussion

IgA with aberrant glycosylation plays an essential role (“first hit”) in the pathogenesis of IgAN. In particular, the levels of an IgA1 variant, which is galactose deficient in the hinge region (Gd-IgA1), are significantly increased in the blood of patients with IgAN, and the extent of this increase is correlated with disease prognosis.^26–30 However, healthy relatives of patients with IgAN exhibit significantly higher levels of Gd-IgA1 in blood than healthy nonrelatives, and autopsy results indicated that glomerular IgA deposits are present in approximately 10%–30% of individuals with normal renal function and no abnormal renal or urinary findings.^31,32 These findings indicate that increased Gd-IgA1 and glomerular IgA deposition are not entirely responsible for the progression of renal impairment and suggest that factors other than increased Gd-IgA1 deposition underlie disease progression. One factor currently under consideration is the formation of immune complexes of endogenous IgM and IgG autoantibodies that correspond with abnormal glycosylation of IgA. Blood concentrations of immune complexes containing abnormal IgA are likely correlated with the prognosis and severity of human and murine IgAN.³³ Furthermore, in human IgAN, the codeposition of glomerular IgM and IgG is correlated with the severity of tissue lesions.^4–8,32 Although the specific mechanism of immune complex formation has not been previously elucidated, the findings of this study may provide a critical clue. First, no deposition of IgM, IgG, or AIM was observed in glomeruli of IgA^−/−gddY mice, which indicates that IgA deposition is the initial trigger in IgAN. Second, the finding that AIM^−/−gddY mice were free from GN without codeposition of glomerular IgM and IgG demonstrates that AIM plays a key role in the mechanism underlying formation of nephritogenic IgA and IgM/IgG immune complexes. The restoration of the GN phenotype by administration of rAIM to AIM^−/−gddY also strongly suggests the direct involvement of AIM in immune complex formation.

It is known that AIM stably associates with IgM by fitting into the one missing part of the hexagon in circulating blood in both mice and humans.²¹ We also clarified that AIM associates with IgA in blood in both mice and humans (Figure 6, Supplemental Figure 7) and particularly binds to IgA-Fc in mice by immunoprecipitation (Supplemental Figure 6). The binding mechanism of AIM and IgA1 in humans could not be clarified with the same method as in mice, presumably because of aberrant glycosylation in IgA1 and polymeric formation of these IgA1. Thus, we need further evaluation under many different conditions to clarify the binding mechanism in humans in the near future. Although the entire mechanism of how AIM promotes the formation of immune complex remains unknown, the binding nature of AIM to both IgM²¹ and IgA (Figure 6, Supplemental Figures 6 and 7) suggests the linkage of IgA and IgM via AIM. Because it was previously demonstrated that AIM does not bind to IgG, the association mode of IgG remains unknown. One possibility is that circulating IgG/Gd-IgA1 immune complexes may directly deposit on glomeruli on the basis of the codeposition of IgG specific for Gd-IgA1 in all patients with IgAN, as recently reported by Rizk et al.³⁴ Another possibility is that aberrant IgM accumulation in this region activates the complement cascade, thereby inducing local inflammation. This inflammatory response might recruit IgG (i.e., IgG may colocalize, but not associate, with the AIM/IgA/IgM immune complex).

Our results suggest that the development of IgAN can be divided into two distinct phases: (1) deposition of IgA and (2) formation of the immune complex and complement activation, leading to glomerular damage. The first phase is independent of AIM, whereas AIM is significantly involved in the second phase. In human patients with IgAN, the immune complex is presumably generated in the serum in some degree.^29,35 However, we could not detect complexes of IgA and IgM containing AIM in the sera from patients with IgAN, even though we clearly found IgA codeposition with IgM/IgG and AIM (Supplemental Figure 7). Hence, the possibility of in situ immune complex formation in glomeruli via AIM must be carefully evaluated in the progression of human IgAN.

Our observation that aberrant AIM accumulation induces kidney injury is in agreement with our previous report that unexpected AIM accumulation on hepatocellular carcinoma (HCC) cells causes injury to the cancer cells.³⁶ This is mainly achieved by activation of the complement cascade, which begins with AIM deposition on HCC cells, thereby suppressing the function of complement inhibitors expressed on the cells through direct binding. Thus, in cancer cells, AIM deposition results in a beneficial outcome (i.e., prevention of HCC), whereas in normal cells, it leads to a detrimental outcome (i.e., initiation of IgAN). However, to the best of our knowledge, AIM does not typically accumulate on normal living cells (Figure 1). Hence, undesired pathogenic IgA deposition in glomeruli unexpectedly causes AIM to be identified as a source of pathogenesis during IgAN development.

Structurally, AIM belongs to the scavenger receptor cysteine-rich superfamily and functions as a soluble scavenger receptor. We recently demonstrated AIM deposition on dead epithelial cells obstructing the proximal renal tubules during AKI; here, AIM facilitates the removal of these cells by enhancing their engulfment by phagocytes, thereby promoting rapid repair of kidney injury.²² This suggests that AIM may induce rapid clearance of harmful immune complexes or injured mesangial cells, which may exhibit important roles in the repair process. It is certainly possible that AIM has two different activities for IgAN regulation during the initiation (detrimental) and progressive (beneficial) phases. However, as observed in AKI, AIM is functionally active when it dissociates from IgM but is inactivated when it associates with the large IgM pentamer. Interestingly, we recently demonstrated that the AIM-IgM binding affinity in cats is 1000-fold higher than that in mice, leading to abrogation of AIM function in cats and subsequent renal failure.³⁷ Therefore, it is also plausible that AIM is inactive within the immune complex in the glomeruli and does not exert its scavenging function in IgAN. Dissociation of AIM might thus provide a novel therapeutic strategy for IgAN through both destruction of immune complexes and induction of tissue repair by IgM-free and active AIM. Further studies involving identification of the mechanism by which glomerular IgM/IgG deposition occurs via AIM, as well as how it may be dissociated, could presumably confirm this hypothesis and facilitate the development of new therapies for IgAN.

In conclusion, our data provide new fundamental insights into how glomerular inflammation is triggered after IgA deposition. Our findings suggest that the blockage of AIM deposition or disassociation of AIM from IgM/IgG could serve as a new therapeutic target to take advantage of its role in the inflammatory initiation of IgAN.

Disclosures

All authors have nothing to disclose.

Funding

This study was supported by Grant-in-Aid for Scientific Research (KAKENHI) from Japan Society for the Promotion of Science grants JP16H06389 (to T. Miyazaki) and JP26461237 (to J. Nakata); AMED-CREST by Japan Agency for Medical Research and Development grant JP18gm0610009 (to T. Miyazaki); a Grant-in-Aid for Progressive Renal Diseases Research, Research on Rare and Intractable Disease, from the Ministry of Health, Labour and Walfare of Japan grant 20FC1045; and research grants from Kyowa Hakko Kirin, MSD Co., Ltd., and Boehringer Ingelheim.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2019100987/-/DCSupplemental.

Supplemental Figure 1. No crossreactivity of anti-AIM with Ig.

Supplemental Figure 2. AIM expression in gddY.

Supplemental Figure 3. AIM^−/−gddY mouse model establishment using the CRISPR/Cas9 system.

Supplemental Figure 4. C5b-9 codeposits with IgG/IgM and AIM in IgAN.

Supplemental Figure 5. No associations of IgA and IgM containing AIM were observed in serum fractionation samples from gddY or BALB/c mice.

Supplemental Figure 6. AIM associates with mouse IgA-Fc.

Supplemental Figure 7. No IgA/AIM/IgM immune complex was observed in serum samples from patients with IgAN.