Significance of Urinary Full-Length and Ectodomain Forms of Megalin in Patients With Type 2 Diabetes

The 867–base pair (bp) fragment of human megalin cDNA encoding the NH₂ terminus of the LBD1 (Gln26–Ala314) was amplified by PCR with the forward primer 5′-CGAGGATCCCAAGAATGTGACAGTGCGCA-3′ and the reverse primer 5′-GGTCTCGAGCTAGGCAGAGCACAGAGTCAT-3′. The 640-bp fragment of human megalin cDNA encoding the CT (His4447–Val4655) was also amplified with the forward primer 5′-GTTGGATCCCACTATAGAAGGACCGGCT-3′ and the reverse primer 5′-GGTCTCGAGGCTGGTATAGCTATACTTCAG-3′. The PCR products, digested with BamHI and XhoI, were subcloned into the corresponding sites of a pGEX-6P-1 expression vector (GE Healthcare UK Ltd., Buckinghamshire, U.K.) with a glutathione S-transferase (GST) tag at the 5′ end. GST-fusion proteins were expressed in the Escherichia coli BL21 strain, and recombinant megalin LBD1 and CT proteins were released by cleavage with PreScission protease (GE Healthcare UK Ltd.) according to the manufacturer's instructions. These recombinant proteins were used as antigens to raise monoclonal antibodies (mAbs), as described previously (19). The specificity of the mAbs toward human megalin was determined by immunoblotting as described previously (20).

The 867-bp fragment of the human megalin cDNA encoding the LBD1 was amplified by PCR with the forward primer described above and the reverse primer 5′-GGTGGATTCCTAGGCAGAGCACAGAGTCAT-3′. The PCR products, digested with BamHI, were subcloned into the corresponding site of the aforementioned pGEX-6P-1 into which CT cDNA had been subcloned. Preparation of GST-LBD1-CT fusion proteins and purification of LBD1-CT proteins were carried out as described above. The LBD1-CT proteins were used as the reference samples for calibration of the ELISA.

The capture mAbs (5 μg/mL) were immobilized on the ELISA plates (LumiNunc F16 Maxisorp Surface plate; Thermo Fisher Scientific, Inc., New York, NY) (100 μL/well in 50 mmol/L carbonate buffer, pH 9.5) at 4°C overnight. The plates were washed with Tris-buffered saline (25 mmol/L Tris-HCl, 137 mmol/L NaCl, 2.68 mmol/L KCl, pH 7.4), blocked with 1% casein and 0.05% Tween 20 (200 μL/well) in Tris-buffered saline containing 0.1% NaN₃ at 4°C overnight, and stored at 2–8°C. The tracer mAbs were digested by pepsin to prepare F(ab’)₂ fragments that were reduced to Fab’ fragments by 2-mercaptoethylamine. Reduced Fab’ fragments were conjugated to alkaline phosphatase (Roche Diagnostics GmbH, Mannheim, Germany) as described previously (21). Samples of urine (90 μl) were mixed with 10 μL of solution A (2 mol/L Tris-HCl, 0.2 mol/L EDTA, 10% [volume for volume] Triton X-100, pH 8.0), incubated for 1 min at room temperature for the COOH (C)-megalin assay or for 3 h at 50°C for the NH₂ (amino [A])- and full-length (F)–megalin assays, and reacted with alkaline phosphatase–labeled tracer mAbs in the ELISA plates. A chemiluminescent immunoassay detection system with CDP-Star substrate and Emerald-II enhancer (Applied Biosystems, Carlsbad, CA) was used according to the manufacturer’s instructions (ELISA-Light System; Applied Biosystems). Urinary megalin concentrations were standardized by adjustment to urinary creatinine concentrations.

Urine samples and clinical data were collected from adult residents who had participated in public medical examinations at Tagami-machi (Niigata-ken, Japan) from 2007 to 2009 and volunteers at Denka Seiken Co., Ltd. (Tokyo, Japan) in 2004. Normal control individuals (n = 160; 20–76 years of age; male/female = 76/84) who satisfied standard medical criteria as defined in Supplementary Table 1 were selected from these populations. The criteria included normal values for estimated glomerular filtration rate (eGFR) (60–104 mL/min/1.73 m²) and urinary albumin creatinine ratio (<10 mg/g), which had been found in the general population with a mortality hazard ratio of ≤1.02 (22). Patients with type 2 diabetes (n = 68), including 43 men and 25 women (33–83 years of age), in Niigata, Tsukuba, and Juntendo university hospitals were also recruited into this study. The first- or second-void midstream urine samples were collected from the participants, on the basis of the results of a pilot study that had shown no significant differences between the first-void (preprandial) and the second-void (postprandial) urinary levels of A- and C-megalin in 19 randomly selected patients with type 2 diabetes (Supplementary Fig. 1). In addition, other pilot studies have indicated that urinary A- and C-megalin levels were stable during a 12-h incubation at 37°C (Supplementary Fig. 2) or after long-term (2–4 years) storage at −80°C (Supplementary Fig. 3). Urine samples (45 mL) were mixed with 633 μL of protease inhibitor mixture (120 μL of 10% [weight for volume] NaN₃, 45 μL of 5 mg/mL aprotinin, 450 μL of 500 mmol/L benzamidine, 13.5 μL of 11 mg/mL pepstatin A, and 4.5 μL of 50 mg/mL leupeptin), frozen at −80°C, and thawed within 1 year, just before analysis. Collection of urine samples and evaluation of clinical data evaluation were approved by the ethical committees of each university in accordance with the principles embodied in the Declaration of Helsinki, and all participants provided written informed consent.

Urine specimens from patients with type 2 diabetes (n = 10) and normal control subjects (n = 10) were centrifuged at 17,000 × g for 15 min at 4°C to precipitate urinary sediment, including whole cells, large membrane fragments, and other debris. Aliquots of the supernatants were further centrifuged at 200,000 × g for 1 h at 4°C to precipitate exosomes (23) or other membrane vesicles (24). All the fractions of precipitants and supernatants were assayed by megalin sandwich ELISAs. The concentrations in the precipitants were standardized to the original sample volumes.

Normal human renal cortex tissue was obtained at a site distant from renal cell carcinoma in a nephrectomized kidney. This procedure was approved by the Ethical Committee of Niigata University Graduate School of Medical and Dental Sciences, and the tissue donor provided written informed consent. The tissue was homogenized in 200 mmol/L Tris-HCl buffer (pH 8.0) containing 1% Triton X-100, 20 mmol/L EDTA, 10 mmol/L benzamidine, 5% aprotinin, and 1 mmol/L phenylmethylsulfonyl fluoride. After incubation for 60 min at 4°C, the suspension was centrifuged for 60 min at 100,000 × g to obtain the supernatant lysates.

Specimens of urine (900 μl) were mixed with 100 μL of solution A and subjected to gel filtration on a 160 × 1.5-cm Sephacryl S-300 column (GE Healthcare UK Ltd.) in 200 mmol/L Tris-HCl buffer (pH 8.0) containing 0.1% Triton X-100, 20 mmol/L EDTA, and 150 mmol/L NaCl. The fractionation was carried out under conditions of 1 mL/tube recovery at 10 cm/h of liner flow rate. Lysates (10 μg) of normal human renal cortex were also applied to the column for fractionation. All the fractions were assayed by megalin sandwich ELISA. Molecular weights of the fractionated samples were determined by using gel filtration calibration kits (GE Healthcare UK Ltd.).

Plasma concentrations of creatinine were measured by an enzymatic method in the clinical laboratories in Niigata, Tsukuba, and Juntendo university hospitals. Urinary concentrations of creatinine, albumin, N-acetyl-β-d-glucosaminidase (NAG), α₁-microglobulin, and β₂-microglobulin were measured by an automated instrument (7170S; Hitachi High-Technologies Corp., Tokyo, Japan) with reagent kits CRE-S (Denka Seiken Co., Ltd.), ALB-TIA ×1 (Denka Seiken Co., Ltd.), N-assay L NAG NITTOBO (Nittobo Medical Co., Ltd., Tokyo, Japan), αMi-Latex (Denka Seiken Co., Ltd.), and BMG-Latex (Denka Seiken Co., Ltd.), respectively. Concentrations of each urinary marker were normalized to those of urinary creatinine. The eGFR was calculated with an equation validated for the Japanese population (25).

In urine samples collected from normal control individuals (n = 160), A-megalin, C-megalin, albumin, NAG, α₁-microglobulin, and β₂-microglobulin were measured as described above. Normalization transformation of the values was carried out by the Box-Cox transformation method. Outliers were rejected using the Grubbs-Smirnov test; the upper value of the 95% confidence interval was defined as the normal cutoff level for each marker.

Comparisons of three or more groups were performed by one-way ANOVA followed by the Tukey test (for the same distribution) and the Kruskal-Wallis H-test followed by the Dunn test (for different distribution). Spearman rank correlation test was used for single regression analysis in which dummy variables, 1 and 0, were assigned for male and female, respectively.

We raised mAbs against recombinant proteins encoding the NH₂-terminal LBD1 (mAbs A5 and A12) and the COOH-terminal CT (mAbs C25 and C37) of human megalin. The specificity of the mAbs for human megalin was confirmed by immunoblotting (Supplementary Fig. 4). Three types of megalin sandwich ELISA systems were developed to measure LBD1 (the A-megalin assay), CT (the C-megalin assay), and F-megalin in human urine. mAbs A12 and A5 (for the A-megalin assay), C25 and C37 (for the urinary C-megalin assay), and A5 and C37 (for the F-megalin assay) were used as capture and tracer antibodies, respectively. Calibration standard curves for these ELISA systems using recombinant megalin protein as a reference are shown in Supplementary Fig. 5. The accuracies of these systems were all within a 10% coefficient of variation.

To identify the forms of megalin in human urine, we used urine protein lysates fractionated by gel filtration. As representatively shown in Fig. 1A, C- and F-megalin assays detected the same high-molecular-weight fraction with a single peak at ∼600 kDa that corresponds to the full-length form of megalin in urine samples from patients with type 2 diabetes. However, the A-megalin assay detected not only the high-molecular-weight fraction but also another smaller but more abundant fraction with a peak at ∼550 kDa lacking the CT. In normal urine (Fig. 1B), the high-molecular-weight peak fraction was rarely detected by the C- and F-megalin assays, whereas the smaller fraction was detected by the A-megalin assay. These data demonstrate that there are primarily two forms of megalin in human urine: full-length and ectodomain forms. The full-length form appears to be less abundant in normal individuals but increased in patients with type 2 diabetes.

Next, we used normal human kidney–derived protein lysates similarly fractionated. As shown in Fig. 1C, the A- and F-megalin assays detected the same fraction with a single peak at the high-molecular-weight region, which is likely to be the full-length form of megalin. In contrast, the C-megalin assay detected not only the same high-molecular-weight peak fraction but also smaller fractions with three peaks at ∼20, 40, and 100 kDa. Collectively, normal human kidney contains smaller COOH-terminal megalin fragments in addition to its full-length form.

The ectodomain form of megalin detected in urine samples was not found in kidney lysates, and small COOH-terminal megalin fragments detected in kidney lysates were not found in urine samples. These data suggest that the ectodomain form may be shed into urine by cleavage of the full-length form in PTC membranes, and the small COOH-terminal megalin fragments found in kidney lysates may also be derived from these cleavage reactions.

In the following analyses, we chose the C-megalin assay to evaluate the full-length form of megalin in urine, as this assay virtually detects only this form in urine and does not require heat activation. In addition, we used the A-megalin assay to evaluate primarily the ectodomain form of megalin in urine, because the ectodomain form was highly abundant compared with the full-length form in urine.

To investigate the origins or mechanisms of shedding of the full-length and ectodomain forms of human megalin into urine, we used urine samples, fractionated by centrifugation, from normal control individuals and patients with type 2 diabetes. As shown in Supplementary Fig. 6A, the A-megalin assay detected its targets predominantly in the soluble urinary fraction but less in the insoluble fractions. In contrast, the C-megalin assay detected its targets in both the soluble and insoluble fractions (Supplementary Figure 6B). In addition, the targets of the A-megalin assay were stoichiometrically correlated with C-megalin assay targets in the insoluble fractions precipitated by low-speed centrifugation and ultracentrifugation, respectively (Supplementary Fig. 6C and D). These data indicate that the ectodomain form of megalin is primarily in the soluble fraction, whereas the full-length form is present both in the soluble and insoluble urinary fractions.

The urinary profiles of 160 normal control individuals are shown in Supplementary Table 2. The normal upper cutoff levels for urinary A-megalin, C-megalin, albumin, NAG, α₁-microglobulin, and β₂-microglobulin are 211.3 pmol/g creatinine, 523.5 fmol/g creatinine, 9.2 mg/g creatinine, 5.3 IU/g creatinine, 9.0 mg/g creatinine, and 207.9 μg/g creatinine, respectively. In the normal control individuals, urinary C-megalin levels were weakly associated with age (r_s = 0.33; P < 0.001). When the normal control individuals were stratified into two groups at the median age (42 years), the cutoff levels of the C-megalin assay were 434.7 and 608.0 (fmol/g creatinine) for the younger (20–41 years; n = 78) and the older group (42–76 years; n = 82), respectively, although the levels were much lower than those of patients with type 2 diabetes. Urinary A-megalin levels were not associated with age.

We then investigated the relationship between urinary megalin excretion and albuminuria stratification in patients with type 2 diabetes (n = 68). Urinary C-megalin levels tended to increase with increasing albuminuria in these patients (Fig. 2A and Supplementary Table 3). Notably, the levels in patients with normoalbuminuria were significantly higher than those in normal control individuals; 39% of normoalbuminuric patients (15 of 39) showed urinary C-megalin levels over the normal cutoff level.

Moreover, the urinary C-megalin assay was found to show a significant association with reduced eGFR (<60 mL/min/1.73 m²), whereas traditional markers of PTC injury such as NAG, α₁-microglobulin, and β₂-microglobulin did not show such an association with eGFR (Table 1 and Supplementary Fig. 7). Of note, the Spearman rank correlation coefficient of the urinary C-megalin assay against eGFR (<60 mL/min/1.73 m²) was higher than that of urinary albumin, even in the normo- and microalbuminuric stages (Table 1).

On the other hand, urinary A-megalin levels were significantly higher in normo- and microalbuminuric patients than in normal control individuals but not in patients with macroalbuminuria (Fig. 2B and Supplementary Table 3). However, this is not because the A-megalin assay is extensively inhibited by other proteins, such as albumin, in urine (Supplementary Fig. 8). Urinary A-megalin levels also showed no association with eGFR (<60 mL/min/1.73 m²) (Table 1).

The urinary A- and C-megalin assays were further evaluated for association with other clinical parameters in 52 of 68 patients with type 2 diabetes (Supplementary Table 4), who were cared for in the same Niigata University Hospital with the same laboratory examination standards used. The patients showed no features of bleeding, and took no medicines such as vitamin D analogs, phosphate binders, or erythropoiesis-stimulating agents before and during the study. Single regression analysis, performed with the Spearman rank correlation method (Table 2), showed that urinary C-megalin levels were associated positively with urinary albumin, urinary NAG, and plasma inorganic phosphate (Pi) and negatively with blood hemoglobin. However, urinary C-megalin levels were not associated with age or sex. Moreover, plasma Pi levels were not associated with urinary albumin, NAG, α₁-microglobulin, or β₂-microglobulin levels. An association of blood hemoglobin levels was found with urinary NAG levels, possibly because blood hemoglobin and urinary NAG levels appear to be confounding factors for aging. In contrast, urinary A-megalin levels were negatively associated with fractional excretion of sodium (FENa), a renal sodium handling marker.

We identified full-length and ectodomain forms of megalin in human urine with newly established sandwich ELISA for C- and A-megalin, respectively. The urinary full-length form of megalin was increased in correlation with the severity of DN in patients with type 2 diabetes. Notably, the urine levels in normoalbuminuric patients were significantly higher than those in normal control individuals, suggesting that the urinary C-megalin assay is sensitive enough for early diagnosis or prediction of the development of DN. The assay also showed a significant association with eGFR <60 mL/min/1.73 m², whereas urinary NAG, α₁-microglobulin, and β₂-microglobulin did not show such an association with reduced eGFR. Of note, the C-megalin assay showed a better association with reduced eGFR than did albuminuria.

The molecular mechanisms by which full-length megalin is shed into urine may be complicated because it is found in both soluble and insoluble urinary fractions. It is not simply derived from PTCs detached into urine because COOH-terminal megalin fragments (∼20, 40, and 100 kDa) seen in the kidney tissue are not detected in urine samples. A part of the shed form in the insoluble urinary fractions may be derived from exosomes (23) or other membrane vesicles such as microparticles and microvesicles (24). The endocytic machinery is directly regulated by the oxygen-sensitive pathway (26). The CT of megalin interacts with many intracellular adaptor proteins that act as connectors to motor proteins (9). Therefore, PTC hypoxia may alter the molecular interactions and cause urinary shedding of full-length megalin.

In contrast, urinary excretion of megalin ectodomain appears to be associated with distinctive mechanisms involved in early stages of DN. Megalin undergoes RIP similarly to the Notch and amyloidal precursor protein families (27,28). Megalin is subjected to metalloprotease-mediated ectodomain shedding that produces a 35–40-kDa, membrane-bound megalin COOH-terminal fragment (MCTF), and the MCTF is processed by γ-secretase activity to produce a soluble megalin intracellular domain (MICD) involved in signal transduction in vitro (10). In this study, we identified for the first time that the small CT fragments of megalin (∼20 and 40 kDa), which may correspond to MICD and MCTF, respectively, are present in human kidney tissue. It is thus likely that RIP is a mechanism for urinary shedding of the megalin ectodomain. Decreased FENa in diabetic patients may compensatorily induce RIP of megalin, which increases the release of the ectodomain and the production of MICD that is known to suppress the expression of Na⁺/H⁺ exchanger isoform 3 in PTCs (29).

Low-molecular-weight protein markers of PTC injury such as α₁-microglobulin and β₂-microglobulin are filtered by glomeruli and reabsorbed by PTCs via megalin (30,31). In the early stage of PTC injury, megalin may be dissociated from the cell membranes and excreted into urine, but its endocytic function may still be preserved. In that stage, urinary excretion of megalin’s endocytic ligands is not likely to be increased. In addition, in advanced stages with reduced GFR, their urinary excretion may not be increased because their glomerular filtration is decreased. In addition, the release of NAG, a lysosomal enzyme, into urine may not be as sensitive to PTC injury as megalin, which is abundantly expressed at the apical PTC membranes.

Similarly, albumin is excreted into urine by its increased glomerular leakage and/or decreased PTC reabsorption via megalin and cubilin (9). In fact, a substantial number of diabetic patients are known to show normoalbuminuria with reduced GFR (32,33). In such patients, hypertension and dyslipidemia may be involved in the development of intrarenal arteriosclerosis or glomerular ischemia to reduce GFR and albuminuria (34,35), but concomitant PTC hypoxia may cause increased urinary megalin excretion.

Anemia is recognized to play an important role in the development of CVD in patients with CKD (36). Dysfunction of PTCs is known to impair the production of erythropoietin and is involved in the development of renal anemia (37,38). Because of the association of blood hemoglobin levels with urinary C-megalin levels, it is speculated that the C-megalin assay reflects the pathologic role of PTCs involved in the development of renal anemia. The positive association of plasma Pi levels with urinary C-megalin levels is also intriguing, because hyperphosphatemia is also associated with the development of CVD and high mortality in patients with CKD (39). This mechanism needs to be studied further because megalin is known to downregulate sodium-phosphate cotransporter IIa in PTCs (40).

Collectively, the urinary full-length form of megalin could be a novel biomarker associated with the severity of DN and related disorders in patients with type 2 diabetes. The urinary megalin ectodomain may be a distinctive RIP-related biomarker that appears to be involved in the mechanisms of early DN. These two urinary forms of megalin should be further investigated to confirm their clinical usefulness for DN by prospective studies.

This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, and Culture of Japan (19590941 and 21591023) and a Grant for Promotion of Niigata University Research Projects.

S.O., H.Ku., Y.H., and S.S. are employees of Denka Seiken Co., Ltd. No other potential conflicts of interest relevant to this article were reported.

S.O. and A.S. analyzed the data and wrote the manuscript. M.H., R.K., H.Ka., K.Y.-K., H.Ku., and H.S. analyzed the data. N.I. and T.T. analyzed the data and contributed to discussion. Y.S., I.N., K.Y., Y.T., F.G., Y.H., and S.S. analyzed the data and reviewed and edited the manuscript. A.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented at the 42nd Annual Meeting of the American Society of Nephrology, San Diego, California, 27 October–1 November 2009; the 7th International Congress on Uremia Research and Toxicity, Nagoya, Japan, 12–14 May 2011; and the 44th Annual Meeting of the American Society of Nephrology, Philadelphia, Pennsylvania, 8–13 November 2011.