IgA1 immune complex-mediated activation of MAPK/ERK kinase pathway in mesangial cells is associated with glomerular damage in IgA nephropathy

Abstract

IgA nephropathy (IgAN) is the most common primary glomerulonephritis worldwide and has significant morbidity and mortality as 20–40% of patients progress to end-stage renal disease (ESRD) within 20 years after disease onset. We aimed to gain insight into the molecular mechanisms involved in IgAN progression. A systematic evaluation of renal biopsy specimens from IgAN patients revealed that the MAPK/ERK signaling pathway was activated in mesangial areas of patients presenting with >1 g/day proteinuria and elevated blood pressure, but was absent in biopsy specimens from IgAN patients with modest proteinuria (<1 g/day). ERK activation was not associated with elevated serum levels of galactose (Gal)-deficient IgA1 or IgG specific for Gal-deficient IgA1. In in vitro studies with human mesangial cells, ERK activation controlled pro-inflammatory cytokine secretion and was induced by patients’ large-molecular-mass IgA1-containing circulating immune complexes. Moreover, we show that IgA1-dependent MAPK/ERK activation required renin-angiotensin system (RAS) activity. Finally, RAS blockers were more efficient in reducing proteinuria in IgAN patients exhibiting substantial mesangial activation of MAPK/ERK. Together, these results suggest that MAPK/ERK activation alters the mesangial cell-podocyte cross-talk, leading to renal dysfunction in IgAN. Assessment of MAPK/ERK activation status in diagnostic renal biopsies could therefore serve as a biomarker to predict the efficacy of RAS blockers in IgAN.

Introduction

IgA nephropathy (IgAN), also known as Berger’s disease, is characterized by the deposition of IgA1-containing immune complexes in the glomerular mesangium^{1, 2}. This condition, initially thought to be benign, is now recognized as a major health problem because 20 to 40% of patients progress to end-stage renal disease (ESRD) within 20 years after disease onset³.

Impaired kidney function is frequently associated with a rise in blood pressure which is, by itself, a risk factor for chronic kidney disease (CKD)⁴. Substantial proteinuria is also a recognized risk factor for CKD progression,⁵ and magnitude of proteinuria combined with the estimated glomerular filtration rate (eGFR) permits assessment of risk for progression towards ESRD⁶.

Mesangial cells participate in the regulation of glomerular hemodynamics and are critical for renal glomerular function. These contractile cells fulfill multiple functions, including the secretion of extracellular matrix proteins, growth factors and cytokines, and the uptake of macromolecules and immune complexes^{7, 8}. In this context, change in physiology of mesangial cells seems to be an early event in progressive glomerular injury leading to modified composition of extracellular matrix (ECM) and glomerular sclerotic changes⁹. However, little is known about the molecular pathways involved in altered physiology of mesangial cells during IgAN progression.

We have previously identified the transferrin receptor 1 (TfR1/CD71) as an IgA1 receptor and shown that IgA1 glycosylation and size (both altered in IgAN patients) are essential for IgA1 binding to TfR1 ¹⁰. In vitro stimulation of human mesangial cells (HMC) by high-molecular-mass IgA1 induces TfR1 expression, which initiates a positive feedback loop that could increase IgA1 mesangial deposition¹¹.

Herein we describe molecular mechanisms triggered by mesangial IgA1 deposition involved in IgAN progression. An immunohistochemical evaluation of renal biopsies revealed a correlation between IgA1-induced renal injury (i.e., high proteinuria and increased blood pressure) and mesangial activation of the MAPK/ERK pathway. This pathway was involved in secretion of inflammatory cytokines by human mesangial cells (HMC) activated by patients’ large-molecular-mass IgA1-containing immune complexes (800–900 kDa). MAPK/ERK activation did not correlate with serum galactose (Gal)-deficient IgA1 levels (which are elevated in IgAN patients), suggesting that aberrant O-glycan composition is not sufficient for IgAN progression. Finally, our data showed that renin-angiotensin system (RAS) was involved in IgA1-dependent HMC activation. Patients with mesangial MAPK/ERK activation responded with a significant decrease in proteinuria when treated with RAS blockers (angiotensin converting enzyme [ACE] inhibitor or angiotensin II type 1 receptor blocker). This observation supports the notion that IgA1-dependent mesangial cell activation interacts with downstream pathways following RAS activation. Altogether these results suggest that MAPK/ERK activation alters the mesangial cell-podocyte crosstalk, leading to renal dysfunction in IgAN. Our data also indicated that MAPK/ERK activation may serve as new biomarker of IgAN that can predict responsiveness to RAS blockers in IgAN patients.

Results

The MAPK/ERK pathway is activated in mesangial areas of renal biopsy specimens from IgAN patients with proteinuria (>1 g/day) and elevated blood pressure

To characterize the molecular mechanisms involved in IgAN progression, we first identified signaling pathways activated in the mesangial cells. The mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) signaling pathway plays a critical role in regulation of cell proliferation, hypertrophy, cytokine secretion and extracellular matrix synthesis¹². ERK includes two highly homologous kinases, denoted as ERK1 (p44, MAPK3) and ERK2 (p42, MAPK1). Immunohistochemical (IHC) analysis of phospho (p)-ERK1/2(Thr202/Tyr204) (p-ERK1/2) expression in remnant renal biopsy specimens from IgAN patients revealed two distinct patterns, depending on the magnitude of proteinuria: in the tissues from patients excreting >1 g/day, p-ERK1/2 expression was found in more than 20% of mesangial cells and in glomerular endothelial cells and/or podocytes in more than half of the glomeruli (Figure 1A). This pattern clearly differed from that observed in tissues from IgAN patients with modest proteinuria (<1 g/day) and tissues from healthy subjects undergoing donor nephrectomy that showed different degrees of podocyte and endothelial cells positivity but only few mesangial cells with the staining (<20% of stained cells) (Figure 1A). In all biopsy specimens, collecting ducts were stained, as previously described ¹³. Staining with an isotype control immunoglobulin did not show mesangial staining (Figure 1A). In mesangial and glomerular endothelial cells, p-ERK1/2 was observed mostly in the nucleus, which was observed less frequently in patients with modest proteinuria. Further comparative analysis using a semi-quantitative mesangial score of p-ERK1/2 staining further confirmed its correlation with proteinuria (Figures 1B and 1C). Podocyte- and endothelial-cell score (independently of mesangial-cell score) did not correlate with proteinuria (Supplementary Figure 1). Mesangial p-ERK1/2 expression was also associated with high mean blood pressure (MBP) (Figures 1D and 1E). Other parameters such as age, hematuria, eGFR_MDRD or gender were not associated with p-ERK1/2 staining (Figure 1F–I and Table 1). The strong correlation between proteinuria and MBP (r=0.5514; p=0.0006) did not allow analysis of a possible independent contribution of p-ERK1/2 to each of these parameters (Supplementary Figure 2). Association between MAPK/ERK activation and proteinuria was also observed when patients treated with RAS blockers at the time of diagnostic biopsy were excluded from the analysis (Supplementary Figure 3). Thus, IgAN patients with >1 g/day proteinuria and elevated blood pressure had activated MAPK/ERK in the mesangium, mostly in nuclei. Because phosphorylated ERK is known to translocate to the nucleus of cells where it can phosphorylate and regulate a number of transcription factors¹⁴, these observations provide a link between mesangial cell activation and functional alteration of glomeruli in IgAN.

Figure 1. Mesangial activation of MAPK/ERK pathway correlates with increased proteinuria and MBP.

(A) Immunostaining of p-ERK1/2 in renal biopsy tissues from IgAN patients and healthy controls. Patients were divided into two groups according to the p-ERK1/2 mesangial score (low <1.1 and high >1.1). (B) Proteinuria in patients presenting a high or low mesangial score for p-ERK1/2 staining. (C) Correlation between p-ERK1/2 mesangial staining score and proteinuria. (D) MBP distribution in patients presenting a high or low mesangial score for p-ERK1/2 staining. (E) Correlation between p-ERK1/2 mesangial staining score and MBP. (F) Age distribution in patients presenting a high or low mesangial score for p-ERK1/2 staining. (G) Hematuria in patients presenting a high or low mesangial score for p-ERK1/2 staining. (H) eGFR_MDRD values in patients presenting a high or low mesangial score for p-ERK1/2 staining. (I) Sex distribution (males in filled histograms and females in open histograms) in patients presenting a high or low mesangial score for p-ERK1/2 staining. All data are mean ± s.e.m. *P < 0.05, **P < 0.01, *** P < 0.001

Table 1.

Clinical data

Group	Low (<1.1) p-ERK1/2 mesangial staining	High (>1.1) p-ERK1/2 mesangial staining
Number	15	20
Age (yr)	38.53 (23–69)	37.10 (20–77)
Sex ratio (M/F)	9/6	16/4
MBP (mm Hg)*	95.1 (65.7–120.0)	105.2 (81.0–133.3)
Proteinuria (g/d) ***	1.66 (0.53–7.12)	3.67 (1.32–8.00)
Hematuria (x105 RBC/ml)	2.4 (0–10)	103 (0–104)
eGFRMDRD	62.8 (12–114)	72.7 (8–287)
CKD stage (1/2/3/4/5)	4/3/6/1/1	5/5/5/2/3
RAS blockers (w/wo)	4/11	4/16
M0/M1	0/15	3/17
E0/E1	5/10	8/12
S0/S1	6/9	5/15
T 0/T1/T2	8/4/3	10/5/5
Follow-up interval (mo)	22.3 (0–76)	25.3 (0–91)

Numbers are expressed as mean (range),

^*p<0.05,

^***p<0.001

MAPK/ERK pathway activation does not correlate with pathology variables included in the Oxford classification

The lesions observed in renal biopsy specimens from IgAN patients have been recently classified in the Oxford classification¹⁵. In this classification, four major pathology variables are evaluated in the MEST score that includes mesangial proliferation (M), endocapillary proliferation (E), glomerulosclerosis (S) and tubular atrophy and interstitial fibrosis (T). Mesangial hypercellularity, segmental glomerulosclerosis and tubular atrophy/interstitial fibrosis were associated with a worse prognosis (50% decrease in eGFR or ESRD achieved within 5 years of follow-up) independently of clinical assessment¹⁵. We performed analyses to determine whether p-ERK1/2 expression correlates with parameters based on the Oxford classification MEST score. The results revealed that p-ERK1/2 staining was not associated with any of the four variables (Figure 2).

Figure 2. MAPK/ERK pathway of mesangial activation is not associated with histological variables of the MEST score from the Oxford Classification.

Distribution of mesangial proliferation (M0/M1) (A), endocapillary proliferation (E0/E1) (B), glomerulosclerosis (S0/S1) (C) and tubular atrophy and interstitial fibrosis (T0/T1/T2) (D) scores in IgAN patients presenting low or high-mesangial score for p-ERK1/2 labeling.

MAPK/ERK pathway is not activated in renal mesangium in biopsy specimens from non-IgAN proteinuric glomerulonephritis patients

To further evaluate the significance of the correlation between mesangial p-ERK1/2 and proteinuria, we investigated p-ERK1/2-staining pattern in renal biopsy specimens from patients with other forms of glomerulonephritis (membranous glomerulonephritis [MGN] or systemic lupus erythematosus [SLE] nephritis; n=6). In renal biopsy specimens from proteinuric MGN patients, the pattern of p-ERK1/2 labeling was similar to that found for IgAN patients with modest proteinuria, whereas MAPK/ERK pathway was activated in the mesangium of SLE patients (Figure 3). IgA deposits were present in mesangia of all SLE patients, including the one with quiescent disease. Therefore, proteinuria by itself does not induce mesangial activation of MAPK/ERK pathway.

Figure 3. MAPK/ERK pathway of mesangial activation in other proteinuric glomerulonephritis.

Immunostaining of p-ERK1/2 in renal biopsy tissues from patients with systemic lupus erythematosus or MGN showing that proteinuria can be observed without p-ERK1/2 expression in mesangial cells in patients with MGN.

MAPK/ERK pathway activation is not associated with serum levels of Gal-deficient IgA1

IgAN is an immune complex-mediated glomerulonephritis characterized by increased levels of Gal-deficient IgA1 in the circulation and glomerular deposits^{16, 17}. We therefore explored a possible link between abnormal IgA1 O-glycosylation (evaluated by lectin ELISA, as previously described^18–20) and mesangial ERK1/2 phosphorylation. As expected, IgAN patients had elevated serum levels of Gal-deficient IgA1 compared to that for healthy controls (Figure 4A). However, there was no difference in the serum levels of Gal-deficient IgA1 between IgAN patients with and without mesangial p-ERK1/2 staining, even after neuraminidase treatment of purified serum IgA1 to remove terminal sialic acid (Figures 4A and 4B). Gal-deficient IgA1 neo-antigens present in the serum of IgAN patients are recognized by circulating glycan-specific IgG ²¹. Therefore, we compared the levels of IgG specific for Gal-deficient IgA1 between patients with and without mesangial activation of MAPK/ERK pathway. The results showed that serum levels of IgG antibodies and IgG-IgA1 complexes did not differ between these two groups of patients, although the differences between IgAN patients and healthy controls were significant (Figures 4C and 4D). Therefore, circulating levels of Gal-deficient IgA1, IgA1-IgG complexes and IgG antibodies specific for Gal-deficient IgA1 did not discriminate the mesangial activation of ERK in IgAN patients’ renal biopsy specimens. This finding suggests that additional factors are involved in IgAN progression^{21, 22}.

Figure 4. Activation of the MAPK/ERK pathway in mesangial cells is not associated with serum levels of aberrantly glycosylated IgA1, IgG-IgA immune complex, or anti-Gal-deficient-IgA1 IgG antibodies.

(A) Serum levels of Gal-deficient IgA1 (evaluated by ELISA using HAA lectin) in healthy subjects, in all IgAN patients, and in IgAN patients presenting high or low levels of p-ERK1/2 mesangial score. (B) Detection of Gal-deficient IgA1 following neuraminidase treatment. (C–D) Detection of IgG-IgA complexes and IgG anti-Gal-deficient IgA1 in the serum of patients presenting a high or low p-ERK1/2 mesangial score.

Pathogenic circulating macromolecular IgA1 complexes activate MAPK/ERK pathway activation

Factors that could be involved in mesangial activation of MAPK/ERK are high-molecular-mass circulating IgA1-containing immune complexes (IgA1-CIC)²³. Larger complexes (800–900 kDa) stimulate proliferation of cultured HMC, whereas smaller complexes (700–800 kDa) inhibit proliferation²⁴. In contrast, uncomplexed IgA1 does not affect cellular proliferation²⁴. Therefore, we evaluated the capacity of IgA1-CIC from an IgAN patient and Gal-deficient IgA1 purified from the plasma of a patient with multiple myeloma (IgA1 Mce) to stimulate ERK activation in cultured HMC. As shown in Figure 5A, phosphorylation of ERK1/2 was intense in HMC after addition of high-molecular-mass stimulatory CIC (IgA1-CIC-S, MW >800 kDa) and was observed to a lesser degree with small-molecular-mass inhibitory CIC (IgA1-CIC-I, MW <800 kDa) or with uncomplexed IgA1. In time-course studies, ERK1/2 was rapidly phosphorylated after incubation of HMC with a high-molecular-mass Gal-deficient IgA1 myeloma protein (pIgA1 Dou) previously shown to induce HMC activation¹¹ as well as after incubation with PDGF-BB (Figures 5B and 5C) (used as a positive control). These data strongly suggest that the mesangial p-ERK1/2 in proteinuric IgAN patients results directly from mesangial cell activation by deposited IgA1 complexes.

Figure 5. Activation of human mesangial cells (HMC) by pIgA1 induced MAPK/ERK and PI3K/Akt/mTOR signaling pathways.

(A) Circulating immune complexes (CIC) from an IgAN patient and uncomplexed pIgA1 myeloma protein activate MAPK/ERK1/2 pathway in HMC. Serum-starved HMC were stimulated for 15 min with PDGF-BB, IgA1-containing circulating immune complexes (CIC) from an IgAN patient (CIC-S, large-molecular-mass stimulatory CIC; CIC-I, small-molecular-mass inhibitory CIC), and uncomplexed naturally Gal-deficient pIgA1 myeloma protein (pIgA1 Mce, IgA1). Cell lysates were immunoblotted with anti-p-ERK1/2 and re-probed for actin as control for equal protein loading. (B, C) Time-course stimulation of HMC with pIgA1 and PDGF-BB. Serum-starved HMC were treated with or without 1 μM wortmannin (W) or 25 nM rapamycin (R) for 30 min and then activated with pIgA1 (Dou) (pIgA1) (B) or PDGF-BB (C) for the indicated lengths of time. Cell lysates were immunoblotted with anti-p-Akt, anti-p-mTOR and anti-p-ERK1/2 antibodies as described in Methods section. Anti-ERK1/2 antibody was used as control for equal protein loading. Data are representative of at least three independent experiments. (D) Circulating immune complexes (CIC) from an IgAN patient and uncomplexed pIgA1 myeloma protein activated Akt1/FKHRL1 pathway in HMC. Serum-starved HMC were stimulated for 15 or 60 min with PDGF-BB, IgA1-containing CIC from an IgAN patient (CIC-S, large-molecular-mass stimulatory CIC; CIC-I, small-molecular-mass inhibitory CIC) and uncomplexed Gal-deficient pIgA1 myeloma protein (pIgA1 Mce, IgA1). Cell lysates were immunoblotted with anti-p-FKHRL1 antibodies (p-FKHRL1) and re-probed with actin as control for equal protein loading.

Another pathway important for cellular proliferation is the phosphoinositide 3-kinase (PI3K) / protein kinase B (Akt) / mammalian target of rapamycin (mTOR) pathway. MAPK/ERK and PI3K/Akt/mTOR are the major pathways involved in cellular proliferation but their downstream effectors can differ²⁵. We assessed the induction of Akt phosphorylation on Ser 473 (p-Akt) following HMC stimulation. In pIgA1 (Dou)-stimulated HMC, Akt was phosphorylated in a time-dependent manner and pre-incubation of HMC with wortmannin (a specific inhibitor of PI3K) blocked Akt phosphorylation (Figure 5B). Phosphorylation of mTOR, one of the major downstream effectors of PI3K/Akt pathway²⁶, was induced by mesangial cell activation and rapamycin (a pharmacological inhibitor of mTOR) blocked mTOR activation (Figures 5B and 5C). FKHRL1, a signaling molecule downstream of the PI3K/Akt pathway, is involved in the regulation of cellular proliferation²⁷. IgA1-CIC-S purified from an IgAN patient increased FKHRL1 phosphorylation to a greater extent than did IgA1-CIC-I or uncomplexed polymeric IgA1 (Mce) (Figure 5D). Thus, the MAPK/ERK and PI3K/Akt/mTOR pathways are induced after HMC stimulation by IgA1-CIC from IgAN patients.

MAPK/ERK and PI3K/Akt/m-TOR pathways are involved in IL-6 secretion and HMC proliferation, respectively

We and others have shown that HMC activation by pIgA1 (Dou) leads to cellular proliferation and release of the pro-inflammatory cytokine IL-6 from these cells^{11, 28}. Therefore, we examined involvement of the MAPK/ERK and PI3K/Akt/mTOR signaling pathways in these HMC responses. HMC were pre-incubated with or without pharmacological inhibitors, stimulated with pIgA1 (Dou), and cellular proliferation was analyzed. Pre-incubation with either wortmannin or rapamycin blocked pIgA1-induced cellular proliferation (Figure 6A), whereas PD98059 (ERK1/2 inhibitor) had no significant effect (Figure 6A). In contrast, inhibition of ERK kinases reduced IL-6 secretion from HMC stimulated with pIgA1 (Dou) (Figure 6B). An opposite effect (i.e., an increased IL-6 secretion) was observed in wortmannin- or rapamycin-treated cells (Figure 6B). These results revealed that the stimulation of HMC proliferation by high-molecular-mass pIgA1 depends on the PI3K/Akt/FKHRL1 pathway but not on the MAPK/ERK pathway. In contrast, the MAPK/ERK pathway activated in HMC stimulated by high-molecular-mass IgA1 is mainly involved in pro-inflammatory cytokine secretion.

Figure 6. Activation of MAPK/ERK and PI3K/Akt/mTOR signaling pathways are involved in IL-6 secretion and HMC proliferation, respectively.

(A) Activation of PI3K/Akt/mTOR pathway regulates IgA1-dependent HMC proliferation. Serum-starved HMC were pre-incubated with inhibitors of PI3 kinase (wortmannin; 1 μM), mTOR (rapamycin; 25 nM) or ERK (PD98059, 10 μM) for 2 h and then activated with pIgA1 (Dou; 0.5 mg/ml) for another 48 h before an 18-h incorporation of [³H]-thymidine. Non-stimulated (NS) HMC served as control. Data are representative of three independent experiments (error bars, SD). * p < 0.05, ** p < 0.01, *** p < 0.001 (two-tailed t test). (B) Blocking of MAPK/ERK pathway inhibits IL-6 secretion by pIgA1-stimulated HMC. Serum-starved HMC were pre-incubated with inhibitors for 2 h and then activated with or without pIgA1 (0.5 mg/ml) overnight. Non-stimulated (NS) HMC served as control. Supernatants were then collected and IL-6 secreted in cultures was quantified by ELISA. Data are means +/− SD of three independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001 (two-tailed t test).

pIgA1-induced HMC activation is dependent on TfR1 signaling

Transferrin receptor 1 (TfR1/CD71) was previously identified as an IgA1 receptor and its role in IgAN pathogenesis has been characterized¹⁰. TfR1 is overexpressed and co-localized with deposited IgA1 in patients’ renal biopsy specimens²⁹. After exploring the biochemical pathways involved in IgA1-induced HMC activation, we assessed the possible involvement of TfR1 in IgA1-induced stimulation of these pathways in HMC. We have previously characterized a monoclonal antibody (mAb) A24 specific for TfR1 that targets proliferating cells, blocks cellular division, and induces cellular apoptosis^{30, 31}. Pre-incubation of HMC with mAb A24 blocked pIgA1 (Dou)-induced IL-6 secretion (Figure 7A) as well as cellular proliferation (Figure 7B). Based on these results, we conclude that the PI3K/Akt/mTOR and the ERK1/2 pathways leading to HMC proliferation and cytokine secretion involve TfR1 engagement by pIgA1.

Figure 7. pIgA1-dependent HMC stimulation is abrogated by TfR1 blocking.

(A and B) The anti-TfR1 mAb A24 inhibited IL-6 secretion (A) and cellular proliferation (B) in pIgA1-stimulated HMC. HMC were pre-incubated for 2 h with mAb A24 (50 μg/ml) before activation with pIgA1 (1 mg/ml). Cytokine secretion and cell proliferation were assessed after overnight and 48-h incubation, respectively, as described in Methods section. Data are the means +/− SD from three independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001 (two-tailed t test). (C) HMC were pretreated for 30 min with wortmannin (1 μM), rapamycin (25 nM), or soluble TfR1 (sTfR, 200 μg/ml) before stimulation with pIgA1 (0.5 mg/ml) for 10 min. Cell lysates were immunoblotted with anti-p-Akt1 (p-Akt1) and equal loading of proteins was determined by blotting with anti-Akt1 antibody. (D) PDGF-BB-dependent HMC stimulation was not abrogated by soluble TfR1. HMC were pretreated for 30 min with wortmannin (1 μM), rapamycin (25 nM), or soluble TfR1 (200 μg/ml) before stimulation with PDGF-BB (10 ng/ml) for 10 min. Cell lysates were immunoblotted with anti-p-Akt1 (p-Akt1) and equal loading of proteins was determined by blotting with anti-Akt1 antibody.

This conclusion was further examined by pre-incubation of the cells with the TfR1 ectodomain (sTfR1). This treatment completely abrogated IgA1-dependent signaling in HMC (Figure 7C). By contrast, PDGF-BB-induced Akt phosphorylation was not blocked by sTfR1 (Figure 7D). Therefore, pIgA1-induced signaling in HMC likely involves TfR1.

pIgA1 but not other TfR1 ligands promote calcium signaling in mesangial cells

In vitro stimulation of HMC by pIgA1 myeloma protein (IgA1 Dou) or by purified IgA1 from IgAN patients induced TfR1 expression and initiated a positive feedback loop promoting increased mesangial IgA1 deposition¹¹. In contrast, the other TfR1 ligands induced a decrease in receptor expression. Therefore, we compared the ability of TfR1 ligands, i.e., pIgA1 (Dou), Apo-Tf, and Fe-Tf, to induce signaling in mesangial cells. Because calcium mobilization is a common event involved in cell activation by most receptors, we evaluated calcium signaling as a marker of HMC activation by these ligands. pIgA1 was the only TfR1 ligand that induced calcium mobilization in serum-starved HMC (Figure 8). PDGF-BB, as the positive control, also induced calcium mobilization. Therefore, in growth-arrested HMC, pIgA1 was the only TfR1 ligand that induced HMC signaling.

Figure 8. Calcium signaling is triggered by pIgA1 (but not by other TfR1 ligands) in HMC.

The cells grown on glass coverslips were loaded with Fura-2/AM and intracellular calcium was measured as the ratio of Fura-2/AM intensities (excitation, 340 and 380 nm; emission, 515 nm; F/F0) before and after exposure to pIgA1, Fe-Tf, Apo-Tf (1: 0.5 μg/ml, 2: 5 μg/ml, 3: 50 μg/ml, 4: 250 μg/ml, 5: 500 μg/ml) or PDGF-BB (6: 10 ng/ml). The maximal calcium mobilization was measured after stimulation of the cells with ionomycin (i: 10 ng/ml) as an internal control.

pIgA1-induced mesangial cell activation involves the renin-angiotensin system (RAS)

Treatments simultaneously targeting high blood pressure and proteinuria often decrease the progression of CKD, thus delaying the onset of ESRD. One of the most prevalent treatments of proteinuria is based on angiotensin II blockade. Because IgA1-induced phosphorylation of mesangial ERK1/2 was observed in patients with high blood pressure and >1 g/day proteinuria, we examined the effect of angiotensin II blockade on ERK1/2 phosphorylation induced by pIgA1. As shown in Figure 9A, exposure of HMC to angiotensin II (100 nM) induced ERK1/2 phosphorylation to the same extent as did pIgA1 (Dou). Pre-incubation of the cells with an angiotensin II type 1 receptor (AT1R) blocker, losartan, inhibited this ERK1/2 phosphorylation in angiotensin II- and pIgA1-stimulated cells (Figure 9A). As a consequence, losartan also blocked IL-6 secretion from the cells (Figure 9B). Therefore, pIgA1-induced p-ERK1/2 activation is apparently dependent on AT1R activation and blockade of this pathway thus provides a new rationale for the use of RAS blockers in the treatment of IgAN.

Figure 9. pIgA1-induced HMC activation is dependent on ATR1 signaling.

Serum-starved HMC were pre-incubated with AT1R antagonist losartan (10 μM) overnight and then activated overnight with or without pIgA1 (1 mg/ml) in the presence or absence of angiotensin II (ANGII, 10 nM). (A) The presence of p-ERK1/2 (p-ERK1/2) was examined by immunoblotting of cell lysates. Blotting with anti-actin was used to control for equal protein loading. (B) IL-6 in cell culture supernatants was quantified by ELISA (data are the mean +/− SD from three independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001; two-tailed t test).

Reduced proteinuria due to RAS blocker therapy is more effective in IgAN patients with a high mesangial score for MAPK/ERK activation

To further evaluate the pathological or clinical relevance of the ability of RAS blockers to suppress MAPK/ERK activation in mesangial cells, we tested whether the decrease in proteinuria induced by RAS blocker therapy (ACE inhibitor or angiotensin II type 1 receptor blocker) was associated with the status of p-ERK1/2 mesangial score. Patients without RAS blocker therapy (ACE inhibitor or angiotensin II type 1 receptor blocker) scored for mesangial expression of p-ERK1/2 at the time of biopsy were given RAS blocker therapy for at least 28 days ³² were studied. As shown in Figure 10A, treatment reduced proteinuria to a greater degree in patients who presented higher scores for MAPK/ERK activation at the time of diagnosis. Interestingly, the efficacy of RAS blockers to decrease MBP did not differ among patients presenting high or low levels of p-ERK1/2 activation (Figure 10B). Altogether these data suggest that the effect of RAS blockers in the control of proteinuria may depend on mesangial MAPK/ERK activation, whereas control of MBP would be independent of this pathway. We conclude that immunostaining of renal biopsy specimens for p-ERK1/2 at the time of diagnosis may predict the efficacy of RAS blockers in proteinuric IgAN patients. Further studies with large cohorts of patients are necessary to validate these data in larger populations.

Figure 10. Mesangial score for p-ERK1/2 predicts decrease in proteinuria (but not MBP) induced by RAS blockers therapy.

(A) Baseline and last follow-up visit values of proteinuria after beginning RAS blockers therapy in patients with low or high scores of p-ERK1/2 staining. (B) Baseline and last follow-up visit values of MBP after beginning RAS blockers therapy.

Discussion

IgAN is characterized by IgA1-containing immune-complex deposits in the glomerular mesangium that are frequently associated with a mesangial cell proliferation, mesangial matrix expansion and intra-glomerular and interstitial infiltration of inflammatory cells³³. Although several reports have addressed the initial steps involved in the disease pathogenesis (such as the characterization of IgA1 structural abnormalities and the mechanisms underlying the formation of circulating IgA1 complexes)^{16, 17, 21, 22} the molecular mechanisms involved in disease progression are still unknown.

We aimed to identify molecular pathways involved in IgAN progression. We screened a series of remnant renal biopsy tissues from IgAN patients to identify these molecular pathways. Our experiments revealed that activation of the MAPK/ERK pathway (identified by ERK1/2 phosphorylation) occurs in a subgroup of IgAN patients presenting with high blood pressure and >1 g/day proteinuria, both clinical features frequently associated with disease progression. More importantly, the mesangial MAPK/ERK score correlated with degree of decrease in proteinuria induced by RAS blockers therapy. Further multi-center studies with a large cohort of proteinuric IgAN patients will be needed to validate the use of p-ERK1/2 immunostaining to help predict the therapeutic efficacy.

A surprising observation in this study was that IgAN patients with less proteinuria had no expression of mesangial p-ERK1/2 in spite of bona fide IgA deposits in the mesangium. Indeed, our in vitro data unequivocally demonstrate that IgA1-containing immune complexes induce pIgA1 induces ERK1/2 phosphorylation in HMC. Additional parameters therefore influence the response of mesangial cells in vivo. Among these parameters, the genetic background would be a prime candidate^34–37 through expression of various, yet unidentified, factors, whether inhibitory or co-stimulatory. Identification of these factors could be a major challenge for nephrologists for further progress in the understanding this common renal disease.

MAPK/ERK signal transduction pathway is implicated in several cellular functions, such as cytokine production, cell survival and inflammation^38–40. Masaki et al have previously investigated the status of MAPK/ERK activation in renal tissue from a cohort of patients with various glomerulopathies and showed a correlation between the status of MAPK/ERK and cell proliferation. Unfortunately, the absence of IgAN patients in this study precludes the comparison with our data¹³.

Our study revealed that IgA1-dependent stimulation of HMC induced a rapid phosphorylation of the major molecular effectors of the MEK/ERK cascade (namely ERK1/2), indicating that this pathway is activated in IgA1-stimulated mesangial cells. The MEK/ERK inhibitor PD098059 inhibited IgA1-induced IL-6 secretion by HMC but was less effective in blocking HMC proliferation. IL-6 itself can promote HMC activation⁴¹. Because PD98059 has only a partial effect on HMC proliferation we suggest that IgA1-induced HMC proliferation is not dependent on the autocrine secretion of IL-6. In different forms of glomerulonephritis, kidney inflammation is controlled by secretion of pro-inflammatory cytokines by both resident and infiltrating cells^42–44. Several findings from animal models indicate that the inflammation cascade can be triggered by activation of resident kidney cells. Thus, activation of the IgA1-induced MEK/ERK pathway in HMC could be important for the development of renal inflammation triggered by pro-inflammatory cytokines in IgAN. This postulate can be supported by the observation that in proteinuric IgAN patients p-ERK1/2 is observed in cell nuclei where these factors could regulate gene expression involved in inflammation. The cross-talk between the four different known MAPK pathways in IgAN remains to be elucidated and would provide important information on the inflammatory aspect of the pathogenesis of this disease.

Mesangial cells are located between glomerular capillaries and are therefore in close proximity with other glomerular cells (i.e., endothelial cells and podocytes). Although additional factors that influence IgAN progression remain to be identified, our data show an association between mesangial expression of p-ERK1/2 and alteration of the filtration barrier (manifested by >1 g/day proteinuria), suggesting a cross-talk between mesangial cells and podocytes. This cross-talk may involve pro-inflammatory cytokines produced by pIgA1-activated HMC (please cite PMID 18256312) that would modulate the glomerular filtration barrier, either directly or indirectly through recruitment of inflammatory cells. Therefore, in pathological conditions such as IgAN, mesangial MAPK/ERK activation would lead to altered mesangial cell biology and to the release of soluble factors that could induce podocyte de-differentiation and podocyte loss. Further studies are necessary to clarify the downstream target of MAPK/ERK involved in the control of mesangial-cell activation and cytokine secretion by mesangial cells.

A strong candidate for such a downstream target is NF-κB. Soluble IgA1 aggregates and aberrantly glycosylated IgA1 induce NF-κB activation^{45, 46} and angiotensin II subtype 1 receptor (AT1R) expression⁴⁷ in cultured mesangial cells, suggesting a cross-talk between these two systems. Interestingly, angiotensin is also able to directly promote NF-κB activation and IL-6 secretion⁴⁸. In addition, ACE inhibitors reduce macrophage infiltration through the control of monocyte chemoattractant protein-1 (MCP-1) synthesis that may promote the development of glomerulonephritis,⁴⁹ providing additional evidence of implication of NF-κB in glomerulonephritis.

We have previously identified the transferrin receptor (TfR1 or CD71) as a new pIgA1 receptor and characterized its implication in IgAN. Here we show that IgA1-dependent HMC signaling is dependent on TfR engagement. In addition, pre-incubation of HMC with an anti-TfR antibody (mAb A24) blocked IgA1-dependent mesangial cell function, as evidenced by the blockade of IL-6 secretion and mesangial cell proliferation. These results confirm that the pathways leading to both cellular responses originate from TfR1 engagement by pIgA1. These findings also underscore the critical requirement for identification of the first signaling events after IgA1-dependent TfR1 engagement to control HMC cell responses and disease progression. Interestingly, of all the known TfR1 ligands, only pIgA1 was able to induce cell activation as evidenced by calcium mobilization in growth-arrested HMC. The structural requirement to achieve TfR1-induced HMC activation by pIgA1 is still unknown. However, the multimeric nature of high-molecular-mass IgA1 that allows the formation of large TfR1 aggregates could be involved in this process.

Angiotensin II (AngII) plays a central role as a mediator of glomerular hemodynamic adaptation and injury. It has been suggested that AngII-induced mesangial cell contraction with efferent arteriolar vasoconstriction initiates intraglomerular hypertension and may lead to enhanced matrix formation and renal fibrosis after increased synthesis of TGF-β⁵⁰. Several studies have shown that angiotensin II type 1 receptor blockers (ARBs), are particularly effective to reduce proteinuria through their direct action on AT1R-expressing cells in endothelium and vascular smooth muscle expressing the receptor^{51, 52}. Therefore blocking the renin-angiotensin-aldosterone system would be beneficial beyond its effect on blood pressure⁵². For example, the glomerular efferent arteriole is particularly sensitive to angiotensin II⁵³. Therefore, ARBs would ease the excessive pressure in the glomerular capillaries and, consequently, reduce proteinuria.

ARBs are especially efficacious in controlling proteinuria in patients with glomerulonephritis associated with alterations in mesangial cell function (such as diabetes and IgAN)^{54, 55}. We show here that patients with a high mesangial score for p-ERK1/2 activation at the time of diagnostic biopsy had greater decrement in proteinuria after treatment. These data suggest that glomerular MAPK/ERK activation could be used to predict which IgAN patients will respond to anti-proteinuric, RAS-suppression, therapy. Altogether our results suggest that MAPK/ERK activation controls the crosstalk between mesangial cells and podocytes though activation of AT1R. We propose a new molecular mechanism by which AT1R would lead to proteinuria by controlling MAPK/ERK activation in mesangial cells and cross-talk with podocytes, slowing progression to ESRD. The link between AT1R and TfR1 remains to be elucidated.

In summary, our experiments revealed a new molecular mechanism in which MAPK/ERK activation controls the cross-talk between mesangial cells and podocytes. We postulate that blocking this activation pathway in mesangial cells will prevent progression to ESRD in IgAN patients. Moreover, we propose that activation of MAPK/ERK pathway in the mesangium may serve as a biomarker for renal injury. Treatment with angiotensin II type 1 receptor blockers and ACE inhibitors to dampen this pathway may slow progression of IgAN to ESRD.

Methods

Cells and chemical agents

Human mesangial cells (HMC) were obtained from Lonza Verviers (SPRL, Belgium). Cells were cultured at 37°C in a humidified environment containing 5% CO₂ in RPMI 1640-Glutamax^® medium supplemented with 20% FCS, 50 U/ml penicillin, 50 μg/ml streptomycin and insulin-transferrin-selenium-A (all from Invitrogen SARL, Cergy Pontoise, France). Experiments were performed with cells using passages 4 to 8. For Figure 5, HMC were cultured^{11, 24}. Human naturally Gal-deficient IgA1 myeloma proteins (Dou) and (Mce) were purified^{56, 57}. Naturally Gal-deficient pIgA1 myeloma protein (Mce) has been shown to mimic aberrantly glycosylated IgA1 in IgAN patients (cite refs number 1–4 from the response to reviewers) and the heterogeneity of its O-glycosylation has been characterized by high-resolution mass spectrometry (cite refs number 5, 6 from the response to reviewers) as well as by the patterns of binding to selective lectins (cite refs number 7, 8 from the response to reviewers). Circulating IgA1-containing immune complexes from serum of an IgAN patient were purified by size-exclusion chromatography and pooled as stimulatory or inhibitory complexes ²⁴(please add Novak et al NDT 2011). This serum sample from an IgAN patient was selected as a representative sample based on medium to high serum levels of Gal-deficient IgA1 and presence of both stimulatory and inhibitory immune complexes²⁴(please add Novak et al NDT 2011). PDGF-BB was purchased from Sigma (St. Louis, MO, USA). Monoclonal anti-human TfR antibody (mAb A24) was produced and purified in our laboratory³¹. Wortmannin was obtained from Alexis biochemicals (Paris, France); rapamycin and PD98059 were purchased from Calbiochem (Fontenay sous Bois, France). Soluble human TfR1 (kindly provided by P. Bjorkman, CalTech, CA, USA) was expressed in a lytic baculovirus/insect cell expression system⁵⁸.

ELISA

IL-6 production was determined by standard ELISA assay following the manufacturer’s instructions (Human IL-6 DuoSet ELISA, R&D Systems, Lille, France).

Stimulation of HMC

HMC were serum-starved (arrested in G0 phase of the cell cycle) before stimulation experiments. Briefly, HMC culture medium was changed to RPMI 0.5% FCS and replaced at 48 h, 36 h, 24 h, and 4 h before the stimulation. Cells were stimulated with human naturally Gal-deficient IgA1 myeloma proteins (Dou and Mce), circulating IgA1-containing immune complexes from IgAN patients or PDGF-BB as indicated in figure legends. Chemical inhibitors (wortmannin, PD98059 and rapamycin) or TfR blocking agents (A24 and sTfR) were pre-incubated for 30 min before the stimulation.

HMC were lysed in Lysis buffer (50 mM HEPES, 50 mM NaF, 100 mM NaCl, Triton-X-100 1%, SDS 0.1% supplemented with protease inhibitors: leupeptin and aprotinin (Sigma-Aldrich) at 2 μg/ml). After disruption with a scraper, cell lysates were centrifuged at 4°C at 12,000 rpm for 15 min to pellet the cellular debris. Protein (supernatant) content was measured using the Bio-Rad Protein assay, then denatured and reduced in Laemli sample buffer (62 mmol/L Tris-HCl, pH 6.8, 10% β-mercaptoethanol, 10% glycerol, 2% SDS) and boiling for 5 min. Thirty μg of proteins were loaded on a 7.5% SDS-PAGE gel. In Figure 5, HMC were lysed using lysis buffer (50 mM HEPES, 150 mM NaCl, 1 mM EDTA, 2.5 mM EGTA, 10% glycerol, 0.1% Tween-20, 1 mM DTT) supplemented with protease and phosphatase inhibitor cocktails (Sigma) (please cite Novak et al NDT 2011). The protease inhibitor cocktail contained 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF), aprotinin, leupeptin, bestatin, pepstatin A, and E-64. Phosphatase inhibitor cocktails included cantharidin, bromotetramisole, microcytin LR, sodium orthovanadate, sodium molybdenate, sodium tartarate, and imidazole. In addition, we used phenylmethanesulfonyl fluoride (1 mM), NaF (1 mM), and β-glycerophosphate (10 mM) (Sigma). The resultant lysates were separated by SDS-PAGE using 10% polyacrylamide gel (Bio-Rad, Hercules, CA, USA).

Proteins were transferred onto polyvinylidene difluoride (PVDF) membrane (Millipore, Saint-Quentin en Yvelines, Belgium) and immunoblots were incubated with the following antibodies: anti-phospho-mTOR (Ser 2448 residue), anti-phospho-Akt (Ser 473 residue), anti-phospho-ERK1/2 (the murine mAb reacts with ERK1 and ERK2 when dually phosphorylated at Thr202 and Tyr204 of Erk1 (Thr185 and Tyr187 of Erk2), and singly phosphorylated at Tyr204), and ERK1/2, Akt and mTOR (all from Cell Signaling Technology), followed by the appropriate secondary antibody-horseradish peroxidase conjugate (Pierce, France) and visualized by chemiluminescence with the ECL kit (Pierce).

Proliferation assay

HMC were trypsinized, resuspended in RPMI 1640 with 0.5% FCS, added in triplicate at the concentration of 1x10⁴ cells/well in 96-well tissue culture plates (Falcon, Oxnard, CA, USA), and starved for 48 h before experiments. Proliferation of quiescent HMC was induced using pIgA1 at 0.5 mg/ml or PDGF-BB at 50 ng/ml for 24 h. For the inhibition experiments, mAb A24 (50 μg/ml), wortmannin (1 μM), rapamycin (25 nM), and PD98059 (10 μM) were added 2 h before addition of IgA1. Proliferation was measured over 18 h, after pulses with 1 μCi/well [³H]thymidine (Amersham Life Science, Buckinghamshire, UK). Cells then were washed, trypsinized for 2 h at 37°C, and harvested on filters with a 96-wellHarvester (Pharmacia), and the incorporation of [³H]thymidine was measured with a β-plate microscintillation counter (LKB; Pharmacia).

Analysis of O-glycosylation by lectin assay using Helix aspersa agglutinin (HAA)

The glycan content of serum IgA1 and CICs-containing IgA1 was tested by ELISA using HAA, a lectin that binds to terminal N-acetylgalactosamine (GalNAc) residues²². Briefly, microtitration plates were coated overnight at 4°C with 10 μg/ml of F(ab′)2 fragment of goat anti-human IgA in BBS. The wells were blocked with blocking buffer for 1 h at 37°C. Serum samples adjusted to 10 μg/ml of IgA1 and 500 μg/ml of total proteins, respectively, were added in duplicates in dilution buffer overnight at 4°C. Samples were treated or not for 3 h at 37°C with neuraminidase from Vibrio cholerae (Roche, Meylan, France) at 10 mU/ml in 50 mM sodium acetate buffer with 4 mM calcium chloride and 10 μg/ml BSA, pH = 5.5, to allow desialylation of the captured IgA1. After washing, samples were incubated for 3 h at 37°C with biotinylated HAA diluted at 1/500 in PBS, 0.05% Tween 20, 1% BSA, 0.1% sodium azide, pH 7.6. The bound biotinylated HAA lectin was then detected with streptavidin-AP (Sigma, Saint Quentin Fallavier, France) incubated for 30 min at 37°C followed by the addition of streptavidin-AP substrate (BD, Le Pont de Claix, France) and by the reading of OD at 405 nm. The lectin reactivity was expressed as OD units. IgG-IgA complexes and IgG anti-Gal deficient IgA1 detection by ELISA were performed as previously described⁵⁹. (please provide brief details on antigens and capture etc.)

Measurement of free intracellular calcium

Cells were loaded in glass-bottom microwell dishes (Matek Corporation, Ashland, MA, USA) with 1 μM of Fura2-AM (Invitrogen) and pluronic acid (F-127, Invitrogen) and incubated at 37°C for 30 min. Cells were then washed with Ringer’s solution (in mM: 145 NaCl, 5.4 KCl, 2 CaCl2, 1 MgCl2, 10 glucose, 10 N-(hydroxyethyl)piperazine-N′-2-ethanesulfonic acid (HEPES) and 0.1% BSA, pH 7.5 with NaOH). For acquisition, cells were excited by wavelengths of 340 and 380 nm. Fluorescence emission of several cells was simultaneously recorded at a frequency of 1 Hz using a dual-excitation fluorometric imaging system (TILL-Photonics, Gräfelfingen, Germany) controlled by TILL-Vision software. Signals were computed as relative ratio units of the fluorescence intensity of the different wavelengths (340/380 nm). Data analysis: Individual ratios were then analyzed using Origin 7.5 software normalizing the ratio values with first value ((F/F0)-1). Area under curve was calculated for each stimulation.

Patients’ renal tissues: immunohistochemistry and analysis

Renal biopsy specimens (n=35) from patients with IgAN (defined by the predominant mesangial deposition of IgA by immunofluorescence) were obtained from the Nephrology Department of Hôpital Bichat-Claude Bernard, Paris, France. Biopsies were classified according to the Oxford classification (which scores biopsy findings for mesangial proliferation (M), endocapillary proliferation (E), glomerulosclerosis (S) and tubular atrophy and interstitial fibrosis (T), the MEST score ⁶⁰). All biopsy specimens contained at least 8 glomeruli available for scoring. Biopsy specimens with more than 80% of global glomerular sclerosis were excluded. Normal-kidneys tissue specimens were obtained from living-donor-nephrectomy kidneys prior to transplantation. All patients and controls signed an informed consent document.

Immunohistochemistry was performed in 4-μm paraffin sections using anti-p-ERK1/2 (Thr202/Tyr204) rabbit monoclonal antibodies (# 4370, Cell Signaling Technology, Danvers, MA, USA). Semi-quantitative mesangial score for p-ERK1/2 staining was determined for each glomerulus (0: absence of mesangial cells labeling, 1: <20% positive mesangial cells, 2: 20–50% positive mesangial cells, 3: >50% positive mesangial cells stained positively) and the mean score for each patient was considered as the patient’s p-ERK1/2 mesangial staining score. Scoring was performed by two independent pathologists in a blinded fashion. Glomeruli were analyzed at 400x magnification.

Statistical Analyses

Differences in proteinuria, MBP, age, hematuria, eGFR_MDRD, HAA binding, IgA1-IgG complexes and IgG anti-Gal-deficient IgA1 between IgAN patients with high and low mesangial MAPK/ERK activation were analyzed through Mann-Whitney or unpaired t-test with Welch’s correction depending on Gaussian distribution of variables. MBP and PU correlation with p-ERK1/2 mesangial staining score were performed with non-parametric Spearman correlation. Sex ratio, diagnosis repartition and MEST score repartition were analyzed with Fischer’s exact test or χ² when there were 3 or more conditions. Increase of HAA binding to IgA1 after neuraminidase treatment, and proteinuria, eGFR_MDRD and MBP changes were analyzed through 2-way ANOVA for repeated measures. Statistical analyses were performed using a two-tailed t test. Results are presented as mean +/− SD unless indicated otherwise in figure legends. A P value of less than 0.05 was considered significant. All statistical analyses were performed with GraphPad Prism version 5 (GraphPad Software, San Diego, CA, USA).