Urinary cMet as a prognostic marker in immunoglobulin A nephropathy

Abstract The prediction of prognosis in patients with immunoglobulin A nephropathy (IgAN) is challenging. We investigated the correlation between urinary cMet (ucMet) levels and clinical parameters and examined the effects of cMet agonistic antibody (cMet Ab) in an in vitro IgAN model. Patients diagnosed with IgAN (n = 194) were divided into three groups representing undetectable (Group 1), below‐median (Group 2) and above‐median (Group 3) levels of ucMet/creatinine (ucMet/Cr). Stained kidney biopsy samples were graded according to cMet intensity. Primary‐cultured human mesangial cells were stimulated with recombinant tumour necrosis factor (TNF)‐α and treated with cMet Ab. Our results showed that ucMet/Cr levels positively correlated with proteinuria (P < .001). During the follow‐up, patients in Group 3 showed a significantly lower probability of complete remission (CR; uPCr < 300 mg/g) than those in groups 1 and 2, after adjusting for blood pressure, estimated glomerular filtration rate, and proteinuria, which influence clinical prognosis (HR 0.60, P = .038); moreover, ucMet/Cr levels were also associated with glomerular cMet expression. After TNF‐α treatment, the proliferation of mesangial cells and increased interleukin‐8 and intercellular adhesion molecule‐1 expression were markedly reduced by cMet Ab in vitro. In conclusion, ucMet/Cr levels significantly correlated with proteinuria, glomerular cMet expression, and the probability of CR. Further, cMet Ab treatment alleviated the inflammation and proliferation of mesangial cells. Hence, ucMet could serve as a clinically significant marker for treating IgAN.


| INTRODUC TI ON
Immunoglobulin A nephropathy (IgAN) is the most common form of primary glomerulonephritis, [1][2][3][4] occurring in approximately 40% of Korean adults. 5,6 Deposition of IgA-containing immunocomplexes in the glomerular mesangium results in mesangial hypercellularity, mesangial expansion, glomerular inflammation and structural changes. IgAN is a common cause of chronic kidney disease and end-stage renal disease (ESRD) worldwide. 1,[7][8][9] Approximately 25%-30% of patients with IgAN progress to ESRD, around 20-25 years after initial diagnosis. 10,11 The clinical features and natural progression of IgAN are diverse, 7,12,13 and only minor haematuria may be seen throughout the patient's lifetime, sometimes with proteinuria. However, rapid decline of renal function may lead to ESRD after diagnosis. Hence, it is important for nephrologists to identify and prioritize high-risk patients with poor prognosis. The established prognostic factors of IgAN currently include proteinuria, blood pressure, renal function and biopsy findings at the time of diagnosis. [12][13][14][15][16][17][18] The development of methods to predict disease prognosis without invasive renal biopsy has been clinically challenging. Among the prognostic tools employed, large amounts of urine samples may be easily, simply and repeatedly obtained by a non-invasive method; therefore, urine samples have been used to actively identify novel biomarkers. cMet, a transmembrane tyrosine kinase receptor of hepatocyte growth factor (HGF), is involved in cell growth, survival and regeneration. 19,20 The HGF/cMet pathway regulates the progression of various diseases by reducing oxidative stress, inflammation, apoptosis and fibrosis. 21,22 The role of urinary cMet (ucMet) as a biomarker in diabetic nephropathy has recently been identified. 23 In addition, animal and cell experiments have demonstrated the effects of attenuating kidney fibrosis and acute kidney injury (AKI) using agonistic monoclonal antibodies of cMet (cMet Ab). [23][24][25] In glomerular endothelial cells (GECs) and proximal tubular epithelial cells (PTECs), cell markers and the expression region of cMet are well merged. 23,24 However, the application of cMet as a clinical marker of IgAN, and the ability of cMet Ab to ameliorate IgAN, is yet to be explored.
Therefore, in the present study, we investigated the correlation between ucMet levels at the time of diagnosis and clinical manifestations in IgAN, along with the effects of ucMet levels on clinical outcomes. Furthermore, we verified if treatment with cMet Ab reduced inflammation and mesangial proliferation in in vitro IgAN models.

| Study population and data collection
The present study was approved by the Institutional Review Board Demographic and clinical characteristics at the time of kidney biopsy, including comorbidities, blood pressure, serum creatinine (sCr), estimated glomerular filtration rate (eGFR), spot urine protein-to-creatinine ratio (uPCr) and other laboratory findings (serum albumin, uric acid, total cholesterol and IgA), were collected from electronic medical records. eGFR was calculated using isotope dilution mass spectrometry and a traceable, modified, Modification of Diet in Renal Disease equation. The use of therapeutic agents post-diagnosis, including angiotensin-converting-enzyme inhibitors or angiotensin II receptor blockers, statins and immunosuppressive agents, was also investigated.
Kidney tissue samples were evaluated by light, electron and immunofluorescence microscopy and diagnosed by a renal pathologist.

| Measurement of urine soluble cMet levels
ucMet levels were measured using an enzyme-linked immunosorbent assay (KHO 2031; Thermo Fisher Scientific Inc) according to the manufacturer's instructions. All measurements were performed in a blinded manner, and in duplicate. Sample urine creatinine levels were measured (Roche C702, CREJ2), adjusted, and expressed as the ucMet-to-urine Cr ratio (ucMet/Cr). The association between the ucMet/Cr levels and several laboratory and histological findings was analysed.

| Clinical outcomes
Patients were divided into three groups based on the ucMet/Cr levels as follows: Group 1, undetectable levels of ucMet; Group 2, below-median levels of ucMet/Cr; and Group 3, above-median levels of ucMet/Cr Then, we analysed the correlation between ucMet/Cr and clinical outcome in the three groups of the patients with IgAN who were followed up for at least 3 months after the biopsy was performed. In the present study, clinical outcome was defined as complete remission (CR) following the KDIGO guideline, uPCr < 300 mg/g.

| Immunohistochemistry of kidney biopsy samples
Unstained slides of the tissue samples obtained from the study population were used. Paraffin-embedded kidney tissue samples TA B L E 1 Baseline characteristics and demographics based on the urine cMet/creatinine level   and examined by light microscopy. In each slide, a minimum of 10 fields were assessed for glomeruli (at 400×) and tubules (at 200×), in a blinded fashion by a kidney pathologist. The cMet intensity score was graded semi-quantitatively from 0 to 3 as follows: 0, absence of or faint staining; 1, mild staining; 2, moderate staining; and 3, strong staining. (1.0 μg/mL, R&D Systems) was also used. The cMet Abs concentration used in this study was similar to previous reports. [23][24][25] Mesangial cell proliferation was quantified using a colorimetric MTS cell proliferation assay kit (Promega) according to the manufacturer's protocols.  Note: The data are expressed as the proportion (%), mean ± SD or median (IQR).

| Baseline characteristics and histologic findings by ucMet group
The baseline characteristics of each of the three groups divided by ucMet/Cr levels are described in Table 1

| Renal outcomes
The hazard ratios (HRs) of CR between Groups 1 + 2 and

F I G U R E 2
Comparison of the CR probability according to ucMet/Cr level. A, Patients in Group 3 had a significantly lower probability of reaching CR compared to patients in Groups 1 + 2 (Log rank P = .010). B, Four combinations of proteinuria and ucMet/ Cr levels were made, and the probabilities of reaching CR were compared. In patients with proteinuria >1 g/d, higher ucMet/Cr levels resulted in lower CR (Log rank P < .001) was also significant after adjusting for factors such as systolic BP, eGFR and proteinuria, which influence clinical prognoses (model 2; HR 0.60, 95% CI 0.37-0.97, P = .038).
Four combinations of uPCr and ucMet/Cr levels were made at the point of diagnosis. In groups 1 to 4, it was noted that patients were progressively older, with increased blood pressure, poorer renal pathology and renal function, raised levels of total cholesterol, and decreased levels of serum albumin (Table S1). In addition, many patients were treated with renin-angiotensin system blockers and immunosuppressive agents. Compared to patients with proteinuria <1 g/d and undetectable or low ucMet/Cr levels, increasing ucMet/ Cr levels did not affect the probability of CR (model 2; Group 1 vs 2, HR 0.61, 95% CI 0.29-1.29, P = .194) ( Table 3; Figure 2B).
Conversely, although ucMet/Cr levels were low or undetected, patients with proteinuria ≥1 g/d were 63% less likely to develop CR

| cMet expression in kidney tissues and ucMet/ Cr
cMet staining in human kidney tissues (n = 18) and representative images according to the quantified scores are shown in Figure 3.
After measuring and scoring the cMet intensity in the glomeruli, the results were divided based on the ucMet/Cr levels of the patients. It was seen that as ucMet/Cr levels increased, Met intensity in the glomeruli also increased ( Figure 3A). On the other hand, Met intensity in the tubules was not correlated with ucMet/Cr levels ( Figure 3B).

| In vitro model for IgAN
First, we identified mesangial cells using desmin staining in a kidney biopsy slide sample of an actual IgAN patient and then stained cMet, which confirmed that the mesangial proliferation and expansion and cMet expression were merged well ( Figure 4A).
Next, mesangial cells were treated with TNF-α in a primary culture and observed for 24 hours. Their proliferation was con- however, this expression was significantly decreased on treatment with cMet Ab ( Figure 4D). Recent studies have reported that several markers in serum and urine samples, such as angiotensinogen, 27 epidermal growth factor, 28 fibroblast growth factor-23, 29 galactose-deficient IgA1, 30 autoantibodies against galactose-deficient IgA1, 31 and kidney injury molecule-1, 32 can be used to predict the early progression of IgAN.

| D ISCUSS I ON
Recently, matrix metalloproteinase-7 levels in urine samples have been shown to predict the progression of IgAN in addition to existing histopathological scores and clinical information. 33 However, these studies were performed in a relatively small number of patients and did not confirm the expression of markers in the kidney tissue, and the correlation with the level of markers. In addition, the mechanism of action of the markers was not proven.
In the present study, treatment with TNF-α imparted a proliferative and inflammatory phenotype to human mesangial cells and A dose-dependent decrease was observed on treatment with cMet Ab (N = 3/ group). D, ICAM-1 expression was also increased on TNF-α stimulation; it was seen to reduce after cMet Ab treatment. Magnification: 800× (bar = 50 μm). The data shown are representative of three independent experiments. All data are presented as the mean ± SEM. *P < .05 (unpaired t test); **P < .01 (unpaired t test); ***P < .001 (unpaired t test) normal control group, patients with gastric cancer expressed significantly lower levels of cMet which decreased with time after diagnosis, indicating that soluble cMet possesses antitumor potential. 39 In diabetic nephropathy, higher urine cMet levels are correlated with poorer outcomes. 23 There was also a significant increase in the expression of cMet in patients with AKI. 25 This is in line with the findings of our study, in which urine cMet levels correlated significantly with proteinuria.
However, rather than leading to disease progression, an increased expression of cMet may induce protection and recovery.
Previously, HGF was shown to ameliorate renal damage and reduce acute inflammatory responses in AKI models. 40  we propose that cMet can be used as a marker of survival and recovery, rather than just an indicator of damage. Hence, ucMet may prove to be of clinical significance in patients with IgAN.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed in this study are included in this published article (and its supplementary information files Table S1). All other data supporting the presented findings are available from the corresponding author upon request.