The metabolism of 1,25(OH)2D3 in clinical and experimental kidney disease

Chronic kidney disease (CKD) results in calcitriol deficiency and altered vitamin D metabolism. The objective of this study was to assess the 24-hydroxylation-mediated metabolism of 25(OH)D3 and 1,25(OH)2D3 in a cross-sectional analysis of participants with a range of kidney function assessed by precise measured GFR (mGFR) (N = 143) and in rats with the induction and progression of experimental kidney disease. Vitamin D metabolites were assessed with LC–MS/MS. Circulating measures of 24-hydroxylation of 25(OH)D3 (24,25(OH)2D3:25(OH)D3) precisely decreased according to mGFR in humans and progressively in rats with developing CKD. In contrast, the 1,24,25(OH)3D3: 1,25(OH)2D3 vitamin D metabolite ratio increased in humans as the mGFR decreased and in rats with the induction and progression of CKD. Human participants taking cholecalciferol had higher circulating 1,24,25(OH)3D3, despite no increase of 1,25(OH)2D3. This first report of circulating 1,24,25(OH)3D3 in the setting of CKD provides novel insight into the uniquely altered vitamin D metabolism in this setting. A better understanding of the uniquely dysfunctional catabolic vitamin D profile in CKD may guide more effective treatment strategies. The potential that 24-hydroxylated products have biological activity of is an important area of future research.

Human study. We measured the vitamin D metabolome in participants of 2 cohort studies ( Supplementary   Fig. 1). All subjects were aged > 18 and had been recruited from Kingston Health Sciences Centre and the general population of Kingston, Ontario Canada. All participants gave informed consent according to the Declaration of Helsinki and both study protocols were approved by Queen's University and Associated Teaching Hospitals Research Ethics Boards. Demographic data as well as medications and supplement use were obtained by chart review and participant interview at the time of sample collection. The GFR Measurement Study included a convenience sample of 98 community dwelling participants ≥ 18 years of age. We excluded kidney transplant recipients (n = 14), those without biological sample (n = 9), participants taking calcitriol (n = 6) and those with substantial circulating vitamin D 2 (n = 3), as defined by > 15% of circulating total 25(OH)D, indicating consumption of plant-based vitamin D supplements. The metabolic products of vitamin D 2 are not captured by the assay used in this study. The second study was the Healthy Aging Study and included 78 healthy participants between the ages of 40 and 80 with targeted recruitment of 10 males and 10 females within each decade of life. We excluded 1 participant from the Healthy Aging Study with substantial circulating vitamin D 2 . Participants from two studies were pooled for a total convenience sample size of 143 participants. GFR was measured either by urinary inulin clearance (N = 51) or iohexol plasma clearance (N = 84), as previously described 8,9 . Inulin-based GFR was measured via urinary clearance averaging 3 1-h periods. Iohexolbased GFR determination measured plasma clearance hourly 2-4 h following iohexol administration, with a Brochner-Mortensen correction to account for the initial clearance phase 10 . GFR was corrected for body surface area (BSA). Nine participants had technical errors with iohexol and inulin measurement and creatinine-based CKD-EPI eGFR of less than 45 mL/min/1.73 m 2 but were included in grouped assessments as creatinine-based GFR approximations are more accurate at lower levels of GFR 11 . Serum creatinine, phosphate and calcium were measured at Kingston Health Sciences Centre Core Laboratory (Roche Plus Modular). Intact-PTH was measured using an ELISA (Immunotopics, Inc., Quidel). Urine microalbumin (LoQ 5.1 mg/L) was measured on Abbott ARCHITECT c systems c16000 with CV < 5%. Intact-FGF-23 was measured in duplicate via ELISA, in fifty-six samples (Kainos Laboratories ELISA kit), as per instructions.
Rat study. Animal procedures were in accordance with the Canadian Council on Animal Care and approved by Queen's University Animal Care Committee. Male Sprague-Dawley rats (n = 18, Hilltop Lab Animals Inc, PA, USA) were 15 weeks of age at the time of induction of CKD. Prior to intervention, animals were acclimatized for a week whilst housed individually and maintained on a 12-h light/dark cycle. All animals were provided with a diet containing 0.5% phosphate, 1% calcium, 0.05% magnesium, 0.2 mg/kg vitamin K, 1 IU cholecalciferol and 6% protein (Harlen Tekland WI, USA TD 150555). CKD was induced in 10 animals with the addition of 0.25% adenine to the diet weeks whilst the parallel group of non-CKD controls (n = 6) remained on the same non-adenine diet. If animals did not eat all administered food or reached 10% weight loss, they were supplemented with a high calorie supplement (Boost, Clear H2O, USA). Animals were blood sampled from the saphenous vein.
Serum creatinine and serum and urine calcium and phosphate were evaluated spectrophotometrically (Syner-gyHT Microplate Reader; BioTek Instruments, Winooski, VT). Creatinine was evaluated using the Jaffe method (QuantiChrom™ Creatinine Assay Kit, Bioassay Systems). Calcium was measured using the o-cresolphthalein complexone method at 540 nm, as previously described (Sigma-Aldrich, CAN) 2 . Free phosphate was measured using the malachite green (Sigma-Aldrich, CAN) method as described by Heresztyn and Nicholson 3 at 650 nm. Plasma levels of intact PTH and intact FGF-23 were measured using ELISAs (Immunotopics Inc., USA). The measurement of the serum vitamin D metabolites was identical to the procedures outlined above.
Statistical analysis. Data were analyzed using SAS software (V9.4), SAS/STAT software (V14.2) and GraphPad Prism (V8.4.2). The Shapiro-Wilk test was used to assess normality and several variables were logtransformed as a result. Participants were stratified into 15 mL/min/1.73 m 2 mGFR intervals. Groups were compared using a Mann-Whitney U-test or ANOVA. Pairwise comparisons were completed between mGFR groups using Fisher's LSD test in variables for which the global test was significant. Spearman's correlation coefficient was used to evaluate the association between vitamin D metabolites and demographic and laboratory measures, with subsequent partial associations consecutively controlling for age and mGFR. The threshold for significance for all p-values was 0.05. Vitamin D metabolites in rats were compared using repeated measures one-way ANOVA with post hoc test for differences between adjacent time points.   Table 2). The vitamin D metabolites and their ratios were highly associated with each other, even after adjustment for mGFR and age (Supplementary Table 1). Thirty-four (24%) participants self-reported taking cholecalciferol at the time of blood sampling (Fig. 2). The median dose was 1000 IU/day, ranging from 400 to 10,000 IU/day. We found higher levels of 1,24,25(OH) 3 D 3 in participants taking cholecalciferol with no elevation in 1,25(OH) 2 D 3 . The level of 25(OH)D 3 was higher overall in participants taking cholecalciferol regardless of kidney function, yet the 25-VMR ratio did not change.
Animal study. To evaluate the progression of metabolite changes longitudinally with the development of CKD, we measured the same vitamin D metabolite profile in rats with adenine-induced CKD and time-control healthy rats, as confirmed by elevations in serum creatinine (Fig. 3A). The longitudinal data demonstrated unchanged levels of 25(OH)D with progression of CKD (Fig. 3B) and a progressive decline in 24,25(OH) 2 D 3 (Fig. 3C) and the 25-VMR (Fig. 3D). The level of 1,25(OH) 2 D 3 and 1,24,25(OH) 3 D 3 decreased progressively with declining kidney function (Fig. 3E,F). However, aligned with our observation in humans, the ratio between 1,24,25(OH) 3 D 3 and 1,25(OH) 2 D 3 increased with the induction and progression of CKD (Fig. 3G). As expected, blood levels of phosphate, PTH and FGF-23 increased as CKD progressed ( Supplementary Fig. 1).

Discussion
This study reports 1,24,25(OH) 3 D 3 levels in humans across a spectrum of measured GFR. To address the limitations of a cross-sectional study, we evaluated the same parameters longitudinally in the circulation of rats with the induction and progression of CKD. 1,25(OH) 2 D 3 catabolism reflected by the 1,25-VMR increased whereas www.nature.com/scientificreports/ 25(OH)D 3 catabolism (25-VMR) decreased as mGFR declined, a finding that we also observed longitudinally in rats. The progressive reduction in the 25-VMR with decreasing mGFR in humans, and with the induction and progression of CKD in rats, supports the work of others 12 and indicates stagnation of 25(OH)D 3 catabolism . . The biological significance of 24-hydroxylated metabolites in the circulation is unclear; however, there is evidence that 1,24,25(OH) 3 D 3 is biologically active and 24,25(OH) 2 D 3 facilitates fracture repair [13][14][15][16] . Global reduction of 24-hydroxylation in CKD has been proposed to explain low levels of 24,25(OH) 2 D 3 , however, this theory requires careful further evaluation given that the 1,25-VMR does not trend similarly to the 25-VMR. The kidneys are the main site of origin of 1,25(OH) 2 D 3 in the circulation, yet increased or unchanged expression of CYP24A1 in kidney tissue has been reported [17][18][19][20] . While those finding are not consistent with the reduction in the 25-VMR, they are consistent with the 1,25-VMR.
Participants taking cholecalciferol had higher levels of 1,24,25(OH) 3 D 3 . One possibility is that cholecalciferol may have increased 1,25(OH) 2 D 3 production, and subsequent catabolism, despite low kidney function. If this hypothesis were true, this could support the use of cholecalciferol in patients with kidney disease. Alternatively, it could also suggest the presence of pathways that convert 25(OH)D 3 to 1,24,25(OH) 3 D 3 independently of a 1,25(OH) 2 D 3 intermediary as suggested by Martineau et al. 16 In a study of participants with moderate to severe CKD, cholecalciferol supplementation increased levels of 25(OH)D 3 substantially, but the change in 24,25(OH) 2 D 3 was more than proportional to the increase in 25(OH)D suggesting that supplementation increased delivery of 25(OH)D to CYP24A1 and/or increased CYP24A1 activity 21 . Levels of 1,25(OH) 2 D 3 did www.nature.com/scientificreports/   Study strengths include the LC-MS/MS assessment of vitamin D metabolites which allows for accurate assessment of these structurally very similar metabolites, a limitation of other methods 4,23 as well as the measurement of GFR, as opposed to estimation based on endogenous markers. The cross-sectional design of the human study limits interpretation, however the longitudinal rat data demonstrated the identical evolution of changes in vitamin D metabolites with progressing CKD. The participants were predominantly white limiting generalizability. Future studies should consider acute challenges of different metabolites to assess specificity and kinetics of transformation. We acknowledge that variability in metabolite-to-parent ratios may not solely reflect true enzymatic activity. For example, Hsu et al. demonstrated higher 25(OH)D clearance in Black individuals, despite lower a 25-VMR 24 . Further, given 1α-hydroxylated metabolites circulate at ~ 1000× lower concentration than their non-1α-hydroxylated counterparts, differences unrelated to 24-hydroxylation, for example intra-individual variation or assay variability, may meaningfully alter the 1,25-VMR, without altering other metabolites circulating at a higher-levels, and thus the circadian rhythm and the metabolism of 1,24,25(OH) 3 D 3 is an important area of future research before use as a potential diagnostic tool. The development of a vitamin D profile that includes VMRs and evaluates the response over time to supplementation may better define adequacy in this population and guide future therapy as suggested by Melamed et al. 25 .
In summary, the circulating 24-hydroxylated 1,25(OH) 2 D 3 ratio does not decrease like the 25-VMR as kidney function declines. The clinical implications of the divergence between catabolism of 25(OH)D 3 versus 1,25(OH) 2 D 3 requires further study. Understanding vitamin D catabolism and the potential that 24-hydroxylation products may have biological activity may help inform treatment strategies in the future.

Data availability
Data lies with the Principal Investigators and can be made available upon request.