C3aR-initiated signaling is a critical mechanism of podocyte injury in membranous nephropathy

Abstract

The deposition of antipodocyte autoantibodies in the glomerular subepithelial space induces primary membranous nephropathy (MN), the leading cause of nephrotic syndrome worldwide. Taking advantage of the glomerulus-on-a-chip system, we modeled human primary MN induced by anti-PLA2R antibodies. Here we show that exposure of primary human podocytes expressing PLA2R to MN serum results in IgG deposition and complement activation on their surface, leading to loss of the chip permselectivity to albumin. C3a receptor (C3aR) antagonists as well as C3AR gene silencing in podocytes reduced oxidative stress induced by MN serum and prevented albumin leakage. In contrast, inhibition of the formation of the membrane-attack-complex (MAC), previously thought to play a major role in MN pathogenesis, did not affect permselectivity to albumin. In addition, treatment with a C3aR antagonist effectively prevented proteinuria in a mouse model of MN, substantiating the chip findings. In conclusion, using a combination of pathophysiologically relevant in vitro and in vivo models, we established that C3a/C3aR signaling plays a critical role in complement-mediated MN pathogenesis, indicating an alternative therapeutic target for MN.

Keywords: Chronic kidney disease; Complement; Nephrology.

Figure 1. The GOAC recapitulates podocyte injury in vitro.

(A) Schematic representation of the GOAC. (B) Light microscope image, depicting the formation of the barrier (72 hours). Scale bar: 400 µm. (C) Representative image of albumin test (albumin-FITC, 40 mg/mL, yellow) confirming permselectivity. (D) Podocytes in vitro express PLA2R (red). Scale bar: 25 µm. DAPI, blue. (E) Measurement of FITC-albumin in GOAC exposed to anti-PLA2R MN serum (n ≥ 4/group). Additional results in Supplemental Figure 2. (F) PLA2R1 RNA expression following knockdown on hAKPC cells; control: scramble hAKPC cells. (G) Western Blotting showing reduced PLA2R (150 kDa) expression after KD in hAKPC-derived podocytes. β-Actin: 42 KDa; n = 3/group. (H) Measurement of FITC-albumin in GOAC filtrate with PLA2R–knocked-down hAKPC-derived podocytes (groups: regular media, healthy serum and PLA2R-Ab+ MN serum, n ≥ 4/group.) (I) Antibody neutralization by IdeS in MN serum prevents GOAC leakage (n ≥ 4/group). Additional results in Supplemental Figure 5. (J)Complement neutralization by heat-inactivation successfully prevents GOAC injury by MN patient serum (n ≥ 4/group). Additional results in Supplemental Figure 6. (K) C5b-9 (MAC) formation on GOAC after serum exposure. Healthy serum with or without protein S (no MAC, top 2 panels) vs. MN serum (MAC, third panel) in podocin⁺ cells (red). Protein S prevents MAC formation (lower panel). Yellow arrows: MAC/podocin overlap. Magnified views of GEC (unstained) and podocin⁺ cells (red) with or without MAC (green); DAPI, blue. Scale bar: 100 µm. (L) MAC neutralization, while preventing C5b-9 formation, does not prevent leakage (n = 4/group). Additional results in Supplemental Figure 6. (M) GOAC exposure to healthy serum (1%), heat-inactivated anti-PLA2R (0.5%) + healthy serum (0.5%), heat-inactivated anti-PLA2R (0.5%) + C6-deficient healthy serum (0.5%), heat-inactivated- anti-PLA2R (0.5%) + anti-PLA2R (0.5%), or heat-inactivated anti-PLA2R (0.5%) + anti-PLA2R (0.5%) + protein S. (n ≥ 4/group). All statistical values determined by 1-way ANOVA with the exclusion of E and G (2-tailed Student’s t test). *P < 0.05 **P < 0.01; ***P < 0.001; ****P < 0.0001.

Figure 2. C3aR/C3aR signaling plays a key role in PLA2R⁺ MN-induced injury.

(A) Representative picture of immunofluorescence staining confirming expression of C3aR (red, top panel) in podocytes in vitro. DAPI nuclear staining: blue. Magnification, ×20. (B) Box plot graph of fluorescein absorbance (485 nm/535 nm, 0.1 second) in filtrate collected from channel F 60 minutes after addition of albumin. Exposure of GOAC to 50 nM C3a for 24 hours increased albumin leakage compared with the control group. Exposure to the chip to 50 nM C3aR antagonist was able to prevent the deleterious C3a effect, confirming the role of C3a in driving injury in the GOAC. *P < 0.05, n ≥ 4/group. (C) Box plot graph of fluorescein absorbance (485 nm/535 nm, 0.1 second) in filtrate collected from channel F 60 minutes after addition of albumin. Addition of 50 nM C3aR antagonist to GOAC was able to prevent leakage caused by MN serum, further confirming the involvement of C3a in podocyte damage. ****P < 0.0001, n ≥ 4/group. (D) Measurement of C3ar RNA expression by qPCR in control hAKPC-derived podocytes and hAKPC-derived podocytes in which C3ar was knocked down, confirming an 80% decrease in its expression. n = 3, Ctrl hAKPC; n = 3, KD hAKPC. (E) Western blotting analysis of C3aR in hAKPC-derived podocytes and hAKPC-derived podocytes in which C3ar was knocked down confirmed an efficient decrease in protein expression. Measured density for C3aR bands (54 KDa) was normalized against β-actin, showing a significantly decreased protein level. ***P < 0.001. β-Actin: 42 KDa; n = 3/group. (F) Box plot graph of fluorescein absorbance (485 nm/535 nm, 0.1 second) in filtrate collected from channel F 60 minutes after addition of albumin confirming a statistically significant decrease in MN sera–induced albumin leakage in GOAC generated with hAKPC^{C3aR KD} podocytes compared with control. ****P < 0.0001; n ≥ 4/group. All statistical values determined by 1-way ANOVA, with the exclusion of E , which used 2-tailed Student’s t test.

Figure 3. Effect of anti-PLA2R serum on podocyte phenotype.

(A and B) C3aR (A) and PLA2R1 (B) RNA expression in podocytes exposed to 0.5% healthy or MN serum for 72 hours. Red dotted line: hPOD in media with no serum, normalization group (n = 3/group). (C) Western Blotting bands for C3aR (54kDa) and PLA2R1 (150 kDa) and respective beta actin (42 kDa) exposed to healthy or anti-PLA2R serum (n = 3/group). (D) Western blotting analysis of C3aR (54kDa) in hPOD monolayers exposed to MN sera, confirming an increased expression after 72 hours. Bands were normalized against β-actin (42 kDa), showing a significantly increased expression. n = 3/group. (E) Western blotting analysis of PLA2R (150 kDa) in hPOD monolayers exposed to MN sera, showing a statistically significant increase in protein expression at 72 hours. Bands were normalized against β-actin (42 kDa), showing a significantly increased expression (n = 3/group). (F) Western Blotting bands for ROMO1 (10 kDa) and caspase-3 (35 kDa) and respective β-actin (42 kDa) exposed to healthy or anti-PLA2R serum (n = 3/group). (G) Western blotting analysis of oxidative stress, measured by expression of ROMO1 (10 kDa), in hPOD monolayers exposed to MN sera. Bands were normalized against β-actin (42 kDa), showing a significantly increased protein level (n = 3/group). (H) Western blotting analysis of apoptosis, measured by expression of caspase-3 (35 kDa), in hPOD monolayers exposed to MN sera. Bands were normalized against β-actin (42 kDa), showing an increased expression (n = 3/group). (I and J) hPOD monolayers exposed to healthy (I) or MN serum (J) confirmed an increase in ROMO1 (red). DAPI, blue. Scale bar: 25 µm. (K and L) Representative immunofluorescence staining of hPOD monolayers exposed to healthy (K) or MN serum (L) confirmed an increase in caspase-3 (red). DAPI, blue. Scale bar: 25 µm. All statistical values determined by 2-tailed Student’s t test. *P < 0.05, **P < 0.01.

Figure 4. C3aR antagonism effectively prevents changes in podocyte phenotype by MN serum.

(A) Western blotting analysis of C3aR in hPOD monolayers exposed to MN sera or MN sera + C3aRA. The antagonist efficiently prevented C3aR (54 kDa) increase. Bands were normalized against β-actin (42 KDa), showing a significantly decreased protein level. n = 3/group. (B) Western blotting analysis of PLA2R (150kDa) in hPOD monolayers exposed to MN sera or MN sera + C3aRA. The antagonist efficiently prevented PLA2R increase. Bands were normalized against β-actin (42 kDa), showing a significantly decreased protein level. n = 3/group. (C) Western blotting analysis of apoptosis, measured by expression of caspase-3 (35 kDa), in hPOD monolayers exposed to MN sera or MN sera + C3aRA. The antagonist significantly decreased apoptosis in hPOD. Bands were normalized against β-actin (42 kDa), showing a significantly decreased protein level. n = 3/group. (D) Western Blotting bands for Synaptopodin (100kDa) and ROMO1 (10 kDa) and respective β-actin (42 kDa) after exposure to healthy serum, anti-PLA2R serum, or anti-PLA2R serum + C3aRA. (n = 3/group). (E) Western blotting analysis of synaptopodin (100 kDa) in hPOD monolayers exposed to MN sera or MN sera+ C3aRA. The antagonist efficiently prevented synaptopodin loss. Bands were normalized against β-actin (42 kDa), showing a significantly increased protein level. Bands are shown in D. n of replicates/group: 3. (F) Western blotting analysis of oxidative stress, measured by expression of ROMO1 (10 kDa), in hPOD monolayers exposed to MN sera or MN sera+ C3aRA. The antagonist significantly decreased ROMO1 expression in hPOD. Bands were normalized against β-actin (42 kDa), showing a significantly decreased protein level. n = 3/group. Bands are shown in D. All statistical values determined by 1-way ANOVA. *P < 0.05; **P < 0.01, ***P < 0.001.

Figure 5. DAF expression and C3a/C3aR signaling affect disease severity in murine anti-THSD7a antibody–induced MN.

(A) Box plot graph of fluorescein absorbance (485 nm/535 nm, 0.1 second) in filtrate collected from channel F 60 minutes after addition of albumin, confirming a statistically significant increase in anti-THSD7A⁺ sera–induced albumin leakage in GOAC compared with control. Statistical values determined by 1-way ANOVA. *P < 0.05; **P < 0.01; n ≥ 4/group. (B) C3b and synaptopodin staining in representative glomeruli from WT and DAF^–/– mice at 8 weeks after injection of serum from patients with MN with anti-THSD7a antibodies. Scale bar: 50 μm. (C) Quantification of C3b deposition (BALB/c WT, n = 3; DAF^–/–, n = 3). Glomerular fluorescence intensity was quantified measuring MFI in all glomeruli of a kidney section. DAF and C3b expression was then divided by the average MFI of glomeruli stained with isotype (normalized data) using ImageJ software. Statistical values determined by Mann-Whitney U test. ****P < 0.0001. (D) DAF and synaptopodin staining in representative glomeruli from WT mice at 0, 2, and 8 weeks after injection of serum from patients with MN with anti-THSD7a antibodies. Scale bar: 50 μm. (E) Quantification of DAF immunofluorescence (0 weeks, n = 3; 2 weeks, n = 2; 8 weeks, n = 2). Glomerular fluorescence intensity was quantified relatively to isotype using ImageJ software. Statistical values determined by Kruskal-Wallis test. ****P < 0.0001. (F) Urinary albumin/creatinine (ACR) levels in DAF^–/– BALB/c male mice injected with serum from patients with MN with anti-THSD7a antibodies. One group of mice was treated with C3aR antagonist starting from the day of serum injection (THSD7a sera, n = 3; THSD7a sera + C3aRA, n = 3). Statistical values determined by 2-way ANOVA. *P < 0.05, **P < 0.001.

Figure 6. Working model.

Representative scheme of the working model for the presented work.