Significance of urinary C-megalin excretion in vitamin D metabolism in pre-dialysis CKD patients

Serum 1,25(OH)2D and 24,25(OH)2D are decreased in CKD. Megalin in proximal tubular epithelial cells reabsorbs glomerular-filtered 25(OH)D-DBP complex to convert 25(OH)D to 1,25(OH)2D and 24,25(OH)2D. Urinary C-megalin excretion is increased via exocytosis from injured nephrons overloaded with megalin-mediated protein metabolism. This study investigated the significance of urinary C-megalin excretion in vitamin D metabolism in 153 pre-dialysis CKD patients. Urinary C-megalin was positively associated with urinary protein, β2MG and α1MG, and exhibited negative correlations with serum 25(OH)D, 1,25(OH)2D and 24,25(OH)2D. Multiple regression analysis showed that urinary C-megalin had a significantly negative association with 25(OH)D. Serum 1,25(OH)2D and 24,25(OH)2D, as well as 1,25(OH)2D/25(OH)D and 24,25(OH)2D/25(OH)D ratios, showed positive correlations with eGFR. Additionally, wholePTH was positively associated with 1,25(OH)2D/25(OH)D and 1,25(OH)2D/24,25(OH)2D, while FGF23 was positively associated with 24,25(OH)2D/25(OH)D and negatively with 1,25(OH)2D/24,25(OH)2D. Urinary C-megalin emerged as an independent factor positively associated with 1,25(OH)2D/25(OH)D and 1,25(OH)2D/24,25(OH)2D. Although 1,25(OH)2D and 24,25(OH)2D are decreased in CKD patient serum, our findings suggest that PTH and FGF23 retain their effects to regulate vitamin D metabolism even in the kidneys of these patients, while production of 1,25(OH)2D and 24,25(OH)2D from 25(OH)D is restricted due to either impairment of megalin-mediated reabsorption of the 25(OH)D-DBP complex or reduced renal mass.


Correlations of serum vitamin D metabolites with urinary C-megalin excretion in CKD patients.
We next examined the correlation of urinary C-megalin excretion with the serum vitamin D metabolites 25(OH)D (Fig. 3A), 1,25(OH) 2 D (Fig. 3B), and 24,25(OH) 2 D (Fig. 3C). Those results showed that each was significantly correlated in a negative manner (r = −0.292, p < 0.001, r = −0.223, p = 0.007 and r = −0.293, p < 0.001, respectively) with urinary C-megalin excretion in the present 153 CKD patients. These findings indicate that urinary loss of megalin via exosomes is associated with reduced serum levels of vitamin D metabolites in CKD.

Discussion
Results in the present study of 153 pre-dialysis CKD patients indicate that determination of urinary C-megalin excretion is clinically relevant for assessment of PTECs injury. That is based on its good correlation with urinary excretion of β 2 MG/Cr and α 1 MG/Cr (Fig. 1), and our finding that urinary loss of C-megalin might cause a reduction of serum 25(OH)D, based on its negative association with serum 25(OH)D (Table 3). Furthermore, it is likely that urinary C-megalin loss may also be involved in reductions of serum 1,25(OH) 2 D, and 24,25(OH) 2 D (Fig. 3), probably due to restricted transport of 25OH)D to PTECs mitochondria as a result of impaired megalin-mediated absorption of 25(OH)-DBP into PTECs. Additionally, our results showed that PTH and FGF23 likely retain a critical role in regulation of vitamin D metabolism from 25(OH)D to 1,25(OH) 2 D or 24,25(OH) 2 D even in the kidneys of CKD patients. When serum wholePTH findings were divided into quintiles, serum 1,25(OH) 2 D, and 24,25(OH) 2 D were shown to be decreased in the higher wholePTH quintiles, while serum 1,25(OH) 2 D/25(OH)D ratio was not significantly changed regardless of the quintile (Fig. 4). The present findings also revealed a significant and positive correlation of urinary protein with urinary C-megalin (Table 2). That finding is quite reasonable, since increased glomerular filtration of proteins, such as albumin and other low molecular weight proteins, which are taken up by PTECs via megalin, likely overloads the cellular endo-lysosomal system, leading to increased urinary C-megalin excretion by exocytosis from injured PTECs 10 . On the other hand, urinary C-megalin excretion, but not urinary protein excretion, was found to have a significantly negative association with serum 25(OH)D in a manner independent of eGFR, wholePTH, and FGF23 (Table 3). Interestingly, it has been speculated that renal dysfunction may accelerate vitamin D It is interesting to note that serum wholePTH, but not FGF23, was associated in a significantly negative manner with serum 25(OH)D ( Table 3). The present study showed that serum wholePTH associated in a negative manner with serum 25(OH)D, which was in good agreement with our previous study that serum PTH increases as serum 25(OH)D level decreases 26,27 . Therefore, reduced serum 25(OH)D due to impaired megalin-mediated absorption of 25(OH)D-DBP complexes may be responsible, at least in part for development of hyperparathyroidism as indicated by the finding of increased wholePTH, based on the suppressive effect of 25(OH)D on PTH synthesis at parathyroid gland 28 . Therefore, it is possible that urinary exosome megalin excretion has an influence to increase serum PTH, which stimulates 25(OH)D metabolism to 1,25(OH) 2 D, as compared to the effect of FGF23 to stimulate 25(OH)D metabolism to 24,25(OH) 2 D in CKD patients. Since the eGFR values of the present CKD patients were distributed between 5.0 and 58.8 mL/min/1.73 m 2 , it is reasonable to speculate that 25(OH) D metabolism is heavily affected by secondary hyperparathyroidism, based on our results indicating that serum PTH starts to increase at a level below 50 mL/min/1.73 m 2 and then in an eGFR-dependent manner thereafter 26 . The lack of association between eGFR and 25(OH)D shown in the present study (Table 3) can be explained by a previous finding showing that vitamin D is metabolized to 25(OH)D in the liver, but not the kidneys. Therefore, the reported decrease of 25(OH)D in CKD patients may be mainly explained by increased urinary loss of 25(OH) D due to failure of megalin-mediated reabsorption of 25(OH)D, altered intracellular handling of 25(OH)D along with phenotypic changes in PTECs, or enhanced degradation of 25(OH)D to 1,25(OH) 2 D by development of secondary hyperparathyroidism 29 .
In previous studies of CKD patients with eGFR <60 mL/min/1.73 m 2 , including ours 1,19 even though serum PTH and FGF23 were increased, either serum 1,25(OH) 2    increased serum 1,25(OH) 2 D in CKD mice 34 and rats 35,36 . Therefore, it is unlikely that the site of activation of 25(OH)D to 1,25(OH) 2 D might be too severely damaged to produce 1,25(OH) 2 D even in patients with CKD. These results suggest that impaired megalin-mediated absorption of the 25(OH)D-DBP complex plays an important role in the reduction of serum 1,25(OH) 2 D and 24,25(OH) 2 D in CKD, because of their restricted absorption of 25(OH)D into PTECs, although the resultant reduction of 1,25(OH) 2 D might decrease megalin expression in PTECs 37 . On the other hand, a previous study demonstrated a severe decrease in 25(OH) D-1-hydroxylase activity in mitochondria fractions from kidney tissues obtained from CKD stage 4 patients 38 , therefore it is likely that the reduction in 25(OH)D-1α-hydroxylase and 25(OH)D-24-hydroxylase activities resulting from PTECs injury also contributes to reduced serum 1,25(OH) 2 D and 24,25(OH) 2 D in CKD patients.
The present study has several limitations. First, the number of patients examined was relatively few, mainly because we enrolled those consecutively examined at a single institution. Another limitation is that the present cross-sectional study was limited to Japanese CKD patients and it remains unclear whether the results obtained can be extended to other ethnicities. Furthermore, as shown in Fig. 2, serum 25(OH)D levels were rather low in our Japanese population, because of the relatively low level of intake of dairy products and absence of vitamin D-fortified foods. Also, it was not possible to assess causality because of the cross-sectional design. A cohort survey of disease occurrence is necessary to directly evaluate the causality of megalin and its clinical effects on vitamin D metabolites. In

Ethics statement. This study was approved by the Ethics Committee of Osaka City University Graduate
School of Medicine (approval #3366). All study participants provided written informed consent for sampling of blood and urine, as well as examinations of clinical records. The research was done in accordance with the Declaration of Helsinki.
Subjects. CKD was defined by criteria proposed by the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines 39 . Among CKD patients regularly followed by nephrologists at the Department of Nephrology at Osaka City University Hospital, 153 examined from April to June 2017 were enrolled in the present study. The primary conditions related to CKD in our patients were hypertensive nephrosclerosis (n = 45), diabetic nephropathy (n = 26), IgA nephropathy (n = 19), membranous nephropathy (n = 17), autosomal dominant polycystic kidney disease (n = 12), focal and segmental glomerulosclerosis (n = 5), myeloperoxidase-anti-neutrophil cytoplasmic antibody (MPO-ANCA)-associated glomerulonephritis (n = 4), membranoproliferative nephropathy (n = 2), minimal change nephrotic syndrome (n = 1), and unknown (n = 22). Patients with advanced liver disease or taking vitamin D supplements or drugs known to affect vitamin D metabolism including vitamin D derivatives were excluded.
Measurements. Blood and urine samples were collected from all subjects in the morning after overnight fasting. Urine samples were kept on ice for 1 hour and then centrifuged at 1500 rpm for 10 minutes, as previously described 40,41 . All laboratory measurements were performed using routine assays with automated methods 19,26,42 . eGFR was calculated using the new Japanese coefficient for the abbreviated Modification of Diet in Renal Disease Study equation, including a correction factor for women of 0.739 43 . Serum calcium was corrected based on serum albumin, which was calculated as corrected calcium (cCa), as previously reported 26 . Serum wholePTH, which reacts with biologically active full-length PTH (1-84), was measured using a wholePTH assay (Scantibodies Laboratory, Inc. Santee, CA), which is a two-site immunoradiometric assay that exclusively measures PTH (1-84), with intra-and coefficients of variation (CVs) less than 2.3-6.1% and 2.9-8.9%, respectively 15,26,44,45 . Serum FGF23 was determined using a fully automated random access chemiluminescence immunoanalyzer device, the CL-JACK System [Kyowa Medex Co. Ltd., Tokyo, Japan; intra-assay CV 2.7-3.4%, inter-assay CV 1.9-6.3% (internal data)] 16,19 .
Quantification of urinary C-megalin was performed as previously described 8,9 . Briefly, 90 μL of urine was mixed with 10 μL of a solution containing 2 mol/L Tris-HCl, 0.2 mol/L EDTA, and 10% Triton X-100 (pH 8.0), then incubated for 1 minute at room temperature, followed by reactions between the captured monoclonal antibodies immobilized on ELISA plates and the carboxy-terminal domain of megalin. An alkaline phosphatase-labeled tracer monoclonal antibody was then added to the plate and measurements were conducted using a chemiluminescent immunoassay detection system. As surrogate markers of renal tubule injury, urinary concentrations of Cr, β 2 MG, and α 1 MG were determined using an automated instrument (7170 S; Hitachi High-Technologies Corp., Tokyo, Japan), with CRE-S (Denka Seiken Co., Ltd.), BMG-Latex (Denka Seiken Co., Ltd.), and αMi-Latex (Denka Seiken Co., Ltd.) kits, respectively, as previously described 8,9 . The urinary concentration of each marker was normalized to that of Cr, then expressed as g/g Cr (protein), µg/g Cr (β 2 MG), mg/g Cr (α 1 MG), and pmol/g Cr (C-megalin). Median (range) values were used for continuous variables with skewed distribution. Simple regression analysis was performed using a non-parametric Spearman's rank correlation test. Multiple regression analyses were performed after logarithmic transformation of eGFR, wholePTH, FGF23, urinary protein/Cr, urinary C-megalin/ Cr, urinary β 2 MG/Cr, urinary α 1 MG/Cr, serum 25(OH)D, 1,25(OH) 2 D/25(OH)D ratio, 24,25(OH) 2 D/25(OH) D ratio, and 1,25(OH) 2 D/24,25(OH) 2 D ratio, because of their transformation to an approximated normal distribution. Comparison of two regression slopes of 1,25(OH) 2 D and 24,25(OH) 2 D was performed as described previously 48,49 . Association between urinary C-megalin and quintiles of β 2 MG or α 1 MG values were analyzed by one-way ANOVA (analysis of variance), followed by Dunnett's test. Also, association between serum vitamin D metabolites and quintiles of wholePTH or FGF23 values were analyzed by one-way ANOVA (analysis of variance), followed by Dunnett's test. All statistical analyses were performed using the Stat View V system (Abacus Concepts, Berkeley CA) and JMP Pro 12 (SAS Corporation, Cary, North Carolina, United States) on a Windows computer. P values < 0.05 were considered to indicate statistical significance.