D-Asparagine is an Ideal Endogenous Molecule for Measuring the Glomerular Filtration Rate

Abstract

Introduction

An ideal endogenous molecule for measuring glomerular filtration rate (GFR) is still unknown. However, a rare enantiomer of serine, d-serine, is useful in GFR measurement. This study explored the potential of other d-amino acids for kidney function assessment.

Methods

This was a cross-sectional observational study of 207 living kidney transplant donors and recipients, for whom GFR was measured using clearance of inulin (C-in). Associations between levels of d-amino acids and GFR were analyzed using multivariate factor analysis. Fractional excretion (FE), a ratio of the clearance of a substance to C-in as a standard molecule, was calculated to monitor the excretion ratio after glomerular filtration. Dissociation from an ideal FE of 100% was assessed as a bias. Proportional bias against C-in was calculated using Deming regression.

Results

Multivariate analysis identified the blood level of d-asparagine to reflect GFR. Means of blood d-asparagine and clearance of d-asparagine (C-d-Asn) were 0.21 μM and 65.0 ml/min per 1.73 m², respectively. Inulin-based FE (FE_in) of d-asparagine was 98.67% (95% confidence interval [CI]: 96.43–100.90%) and less biased than those of known GFR markers, such as FE_in of creatinine (147.93 [145.39–150.46]; P < 0.001) and d-serine (84.84 [83.22–86.46]; P < 0.001). A proportional bias of C-d-Asn to C-in was −7.8% (95% CI, −14.5 to −0.6%), which was minor compared to those of clearance of creatinine (−34.5% [−37.9 to −31.0%]) and d-serine (21.2% [13.9–28.9]).

Conclusion

D-Asparagine acts similar to inulin in the kidney. Therefore, d-asparagine is an ideal endogenous molecule that can be used for GFR measurement.

Graphical abstract

A precise assessment of GFR is essential in clinical practice and drug development. The GFR is used for donor selection in kidney transplants, chronic kidney disease (CKD) stage classification, drug administration design, and as a metric of outcome or adverse effect in clinical trials.^1,2 The evaluation of GFR should be based on ideal endogenous molecules. The measurement of GFR using exogenous inulin is the gold standard; however, it is labor-intensive, time-consuming, and expensive.² An ideal molecule for GFR measurement is entirely excreted into urine after glomerular filtration; however, such a molecule has not yet been identified. Therefore, the current evaluation of GFR using endogenous molecules has several limitations. Estimated GFR (eGFR) based on blood levels of creatinine and cystatin C is convenient and useful for CKD screening; however, their precision is significantly lower than that of the standard methods for measuring GFR currently used. The clearance of creatinine (C-cre) is accurate but overestimates GFR because of the major proportional bias derived from the tubular secretion of creatinine.³ The bias of C-cre can be reduced when GFR is assessed in combination with d-serine, one of the d-amino acids.⁴

d-Amino acids are emerging biomarkers of kidney disease and function.⁵ Unlike their abundant chiral forms, the l-amino acids, the presence of d-amino acids has long been undetected and their functions were largely unknown until recently.6, 7, 8, 9, 10, 11, 12 Among d-amino acids, higher blood levels of 4 kinds of d-amino acids, that is, d-serine, d-asparagine, d-proline, and d-alanine, are associated with earlier progression to end-stage kidney diseases, thereby requiring kidney replacement therapy in patients with CKD.¹³ In particular, d-serine has been reported as a marker of GFR. The plasma level of d-serine correlates with the GFR measured by the clearance of inulin (C-in).¹⁴ In addition, the clearance of d-serine (C-d-Ser) is closely related to C-in and can serve as a measure of GFR.⁴ The bias of C-cre is due to the tubular secretion of creatinine, which can be mitigated by measuring GFR in conjugation with C-d-Ser. However, the GFR assessment is still imperfect. An ideal evaluation of GFR requires the measurement of additional kidney biomarkers that are entirely excreted into the urine.

As candidates for additional kidney biomarkers, we tested 3 d-amino acids besides d-serine, namely d-asparagine, d-alanine, and d-proline. In this cross-sectional observational study of a prospective cohort, we investigated which of the remaining d-amino acids is ideal for assessing GFR. In addition, we analyzed the suitability of the identified d-amino acids as kidney biomarkers by monitoring their urinary excretion dynamics after glomerular filtration.

Methods

Study Design and Participants

This is a cross-sectional observational study from a prospective cohort. Participants were adults aged ≥20 years who were potential living kidney donors, postkidney donor candidates, and kidney transplant recipients. We recruited 210 participants from 3 centers in Japan, and blood and urine samples were collected while measuring C-in at the Kansai Medical Hospital between July 2019 and January 2021. Exclusion criteria included cases with a urine output of <20 ml per 30 min. This study was conducted in compliance with the Declaration of Helsinki, the Ethical Guidelines for Medical Research Involving Human Subjects, and the Principles of the Declaration of Istanbul as outlined in the “Declaration of Istanbul on Organ Trafficking and Transplant Tourism.” Approval for all facilities was obtained from the Central Ethics Review Committee of Osaka University (#16330). Written informed consent was obtained from all the participants. This study adhered to the STROBE guidelines.

Clearance and FE

Clearance of inulin was calculated from serum and urine inulin concentrations and urine volume using standardized methods described previously.¹⁵ Inulin (1%; Fuji Yakuhin, Saitama, Japan) was administered intravenously using an infusion pump under fasting, medication-suspended, and hydrated conditions. The infusion rate was 300 ml/h for 0 to 30 minutes and 100 ml/h for 30 to 120 minutes. To maintain urine output during clearance measurements, the participants were given 500 ml of water 30 minutes before the start of inulin administration and 60 ml of water at each urine collection time point. The participants urinated completely 30 min after the start of the administration and urine was collected every 30 minutes (at 60, 90, and 120 minutes after the start of the administration). Clearance was calculated as follows: Ux (mg/dl) × V (ml/min)/Sx (mg/dl), where x is a substrate, such as inulin, creatinine, or d-serine, Ux is the urinary concentration of x, V is the flow rate, and Sx is the serum concentration of x. Each clearance was corrected for body surface area using the DuBois and DuBois formula.¹⁶ The FE was calculated as follows: Ux (mg/dl) × Sy (mg/dl)/Sx (mg/dl) × Uy (mg/dl), where x is a substrate substance, such as inulin or creatinine, y is a reference substance, Ux is the urinary concentration of x, Sy is the serum concentration of y, Sx is the serum concentration of x, and Uy is the urinary concentration of y. The mean clearance values at each time point (30–60, 60–90, and 90–120 minutes) were used for analysis. Blood was collected at the midpoint between each urine collection time (45, 75, and 105 minutes after dosing). Serum and urine inulin concentrations were colorimetrically determined using Diacolor inulin kit (Toyobo, Osaka, Japan). The measurements were continuously calibrated using a reference control material. Serum and urine creatinine levels were measured enzymatically (Determiner L CRE, Hitachi Chemical, Tokyo, Japan), and serum cystatin C was measured using an immunologic turbid metric assay (Nescoat GC Cystatin C, Alfresa Pharma, Osaka, Japan). The serum creatinine and cystatin C assays were continuously calibrated using human pooled serum (L-Consera EX, Nissui Pharmaceutical, Tokyo, Japan) and cystatin C standard (Nescoat Cystatin C Standard, Alfresa Pharma), respectively. The measurements were surveyed by the Japanese Association of Medical Technologist's quality control survey project and data standardization project system as external clinical laboratory quality control. eGFR was calculated from the equations developed by the Japanese GFR equation based on serum creatinine³ and serum cystatin C.¹⁷

Quantification of d-Amino Acids

The preparation of samples and quantification of amino acid enantiomers by a 2-dimensional high-performance liquid chromatography system were performed as previously described.^18,19 Twenty-fold volumes of methanol were added to the sample and an aliquot (10 μl of the supernatant obtained from the methanol homogenate) was placed in a brown tube. After drying the solution under reduced pressure, 20 μl of 200 mM sodium borate buffer (pH 8.0) and 5 μl of fluorescence labeling reagent (40 mM 4-fluoro-7-nitro-2,1,3-benzoxadiazole in anhydrous acetonitrile) were added and then heated at 60 °C for 2 minutes. An aqueous solution of 0.1% (v/v) trifluoroacetic acid (75 μl) was added, and 2 μl of the reaction mixture was subjected to the 2-dimensional high-performance liquid chromatography.

The enantiomers of the amino acids were quantified using the 2-dimensional high-performance liquid chromatography platform. The fluorescence-labeled amino acids were separated using a reversed-phase column (Singularity RP column, 1.0 mm i.d. × 50 mm; provided by KAGAMI Inc., Osaka, Japan), with the gradient elution using aqueous mobile phases containing MeCN and formic acid. To determine d-amino acids and l-amino acids separately, the fractions of amino acids were automatically collected using a multiloop valve and transferred to the enantioselective column (Singularity CSP-001S, 1.5 mm i.d. × 75 mm; KAGAMI Inc.). In addition, d-amino acids and l-amino acids were separated in the second dimension by the enantioselective column. The mobile phases are the mixed solution of MeOH-MeCN containing formic acid, and the fluorescence detection was carried out at 530 nm with excitation at 470 nm using 2 photomultiplier tubes.

Target peaks were quantified by scaling the standard peak shape.¹⁹ In this method, the shape of a peak was used for the identification of the substrate, whereas the magnitude of the intensity was used for quantification. From the chromatogram of a sample, target shapes of amino acid enantiomers were identified based on the elution time and shape of the peak. The peak shape obtained by the standard amino acid enantiomer was superimposed on the obtained peak sections, and the magnification constant best fitted to the target peak was identified. The concentration of the target enantiomer was calculated by using an identified magnification constant and the calibration lines. The peak shape method potentiated quantification within a few seconds. The fully-automatic 2-dimensional high-performance liquid chromatography system required <10 minutes for the measurements of each d-amino acid, including separation, identification, and quantification steps. The d-amino acid ratio was defined as the percentage of d-amino acids to the sum of l-amino acids and d-amino acids.

Statistical Analysis

Data were expressed as mean ± SD, or as count and ratio (%). The discriminable distributions of a variable based on predictive variables were analyzed using supervised orthogonal partial least square analysis. Bee swarm and violin plots were used to analyze the distributions. The bee swarm plots showed data distribution for each group using dot plots of individual data points. The violin plots used kernel density estimation to calculate data probability density of the data, which was expressed as the width of the shaded area. Continuous variables between 2 groups were compared using a 2-tailed Student’s t-test, and between multiple groups were compared using 1-way analysis of variance with Tukey’s post- hoc multiple comparison test. A comparison between GFR methods was performed using Deming regression through the evaluation of slope and regression line interception. Agreement between GFR methods was visualized using Bland-Altman plots. We compared the usefulness of those equations and the eGFR equations in the whole data set. The coefficients of determination (R²), bias, root-mean-square error, and accuracy were calculated as metrics for comparison. Equation bias was expressed as the mean of the absolute value of the difference between measured GFR and C-in. Precision was assessed as the interquartile range or SD for the true difference between the measured GFR and C-in. The root-mean-square error for C-in calculated using the equation was the square root of (sum of squared differences/n). Accuracy was expressed as the percentage of participants whose measured GFR was within <30% and 15% of C-in (P₃₀ and P₁₅). The 95% CIs were calculated using bootstrap resampling with a normal distribution approximation. Each method’s bias against C-in was compared using a signed-rank test. Accuracy between methods for C-in was compared using McNemar’s test. Sample size analysis was performed based on the assumption that clearance of new biomarkers agrees with GFR better than C-cre. Supposing that the proportional bias of C-cre against C-in was 30 and the bias from new biomarkers was reduced to 70% with a common standard deviation of 18, the number needed for the test based on an error probability of 0.05 and a power of 0.85 was 146. Because we set 20% as a margin, the total number of participants required was 175. Statistical significance was defined as P < 0.05. Statistical analyses and data visualization were performed using JMP pro 15.0 (SAS Institute, Cary, NC).

Results

Characteristics of Participants

The demographic characteristics of the participants are shown in Table1 and Supplementary Table S1. After excluding 3 participants with low urine output or sample collection failure, a total 207 participants were eligible. The participants comprised 129 potential living kidney donors, 29 postdonors, and 49 transplant recipients. In the overall cohort, the mean values of serum creatinine, cystatin C, and plasma d-serine were 0.87 mg/dl, 1.09 mg/l, and 2.21 μM, respectively. The intra-patient coefficients of variability for C-in was 14.7 ± 13.2 ml/min per 1.73 m².

Table 1.

Characteristics of the participants

Characteristics	Total (N = 207)
Participants
Potential living donor	129 (62.3)
Post-donor	29 (14.0)
Transplant recipient	49 (23.7)
Age, yr	60.3 ± 12.2
Male	90 (43.5)
Body mass index, kg/m2	23.3 ± 3.5
Body surface area, m2	1.64 ± 0.18
Diabetes	10 (4.8)
Hypertension	84 (40.6)
Hyperlipidemia	39 (18.8)
Hemoglobin, g/dl	12.7 ± 1.68
Serum creatinine, mg/dl	0.87 ± 0.35
Serum cystatin C, mg/l	1.09 ± 0.40
Plasma d-serine, μM	2.21 ± 0.78
Plasma d-asparagine, μM	0.21 ± 0.097
Urinary d-asparagine, μM	2.01 ± 1.54
Plasma d/l-asparagine ratio, %	0.483 ± 0.219
Urinary d/l-asparagine ratio, %	16.8 ± 9.30
d-serine clearance, ml/min per 1.73 m2	55.54 ± 17.13
d-asparagine clearance, ml/min per 1.73 m2	64.97 ± 21.92
Inulin clearance, ml/min per 1.73 m2	66.43 ± 20.42
Creatinine clearance, ml/min per 1.73 m2	97.28 ± 30.36
eGFR_cre, ml/min per 1.73 m2	68.22 ± 19.49
eGFR_cys, ml/min per 1.73 m2	70.74 ± 21.52

Cre, creatinine; Cys, cystatin; eGFR, estimated glomerular filtration rate.

Data, n (%) or mean ± SD. eGFR_cre and eGFR_cys denote creatinine- and cystatin C-based estimated glomerular filtration rates (eGFRs), respectively, based on the Japanese formula.^3,17

Blood Level of d-Asparagine Reflects GFR

To assess C-in, we first explored the potentials of the 3 d-amino acids (d-asparagine, d-proline, and d-alanine). For this purpose, we performed an orthogonal partial least square to analyze the discriminable capacities of d-amino acids on C-in. Models were developed in combination with clinical parameters, the 3 d-amino acids, d-serine, and some representative l-amino acids in a fraction of participants representing the essential background characteristics of the total cohort (Supplementary Table S2). This model selected the blood ratios of d-asparagine in both plasma and serum, followed by the plasma d-asparagine level, as influential factors for estimating C-in (Figure 1a and b). Similar to serum levels of creatinine and cystatin C, the percentage of d-asparagine and plasma d-asparagine level strongly correlated with C-in (Figure 1c–e, and Supplementary Figure S1a–g). The ratios and levels of d-asparagine in plasma and serum were almost identical (Supplementary Figure S2), and we used plasma d-asparagine values in the analysis.

Figure 1.

Blood level of d-asparagine reflects inulin clearance. (a) Orthogonal partial least squares (OPLS) analysis of d-amino acids on inulin clearance. Models were developed using plasma levels of d- and l-amino acids and clinical parameters. (b) Score plot of OPLS colored according to inulin clearance. The circle represents a 95% confidence interval. (c–e) Correlations between inulin clearance; (c) plasma ratio of d-asparagine; (d) level of d-asparagine; and (e) l-asparagine. r, Pearson’s correlation coefficient.

d-Asparagine is Nearly Excreted Completely Into Urine After Glomerular Filtration

A key condition of the kidney biomarker is that the clearance of the substance be identical to that of inulin. To analyze the suitability of d-asparagine as a GFR marker, we measured the urinary FE of d-asparagine. FE, which is the ratio of the clearance of a substance to the clearance of a standard molecule, such as inulin, is used to monitor the excretion ratio of the substance after glomerular filtration. The FE of the ideal molecule for measuring GFR is 100%. The mean clearances of inulin, creatinine, d-serine, and d-asparagine were 66.4, 97.3, 55.5, and 65.0 ml/min per 1.73 m² of body surface area, respectively (Table 1). The FE based on various standard molecules is shown in Table 2 and Figure 2a. FE_in of d-asparagine was 98.7% (95% CI, 96.4%–100.9%). The FE of other GFR markers were 147.9% (145.4%–150.5%) for FE_inCre and 84.8% (83.2%–86.5%) for FE_ind-Ser. Cystatin C is not excreted into urine, because FE_creCysC was reported to be less than 0.5%.²⁰ The bias of FE_in was calculated as the difference from 100%. The bias of FE_in of d-asparagine was 12.9% (11.6%–14.3%), which was less than that of FE_inCre (47.9% [45.4%–50.5%]; P < 0.001) and FE_ind-Ser (16.9% [15.6%–8.1%]; P = 0.007; Figure 2b). The bias of FE_in of d-asparagine was constant in the range of GFR analyzed in this study, and the subgroups were defined by sex, transplantation-related status, age, and stages for CKD (Figure 2c–e; Supplementary Figures S3 and S4). Approximately 100% of d-asparagine, like inulin, is excreted after glomerular filtration.

Table 2.

Fractional excretion of inulin, creatinine, d-asparagine, and d-serine based on various standard molecules

Substrate for FE	Inulin-based FE	Creatinine-based FE	d-Asparagine-based FE
Inulin	100.00 (Ref)	68.84 (67.81–69.87)	104.68 (102.42–106.93)
Creatinine	147.93 (145.39–150.46)	100.00 (Ref)	152.60 (149.75–155.45)
d-Asparagine	98.67 (96.43–100.90)	66.97 (65.68–68.25)	100.00 (Ref)
d-Serine	84.84 (83.22–86.46)	57.75 (56.70–58.75)	87.22 (85.54–88.90)
Cystatin C	–	< 0.520	–

FE, fractional excretion.

FE values for the substrate were calculated based on inulin, creatinine, and d-asparagine. Data, % (95% confidence interval).

Figure 2.

d-Asparagine is nearly excreted completely into urine after glomerular filtration. (a) Bee swarm and violin plots of FE, where FE_ind-Asn is the inulin-based FE of d-Asn. Outlines of the violin plot illustrate probability density. (b) Biases for FE of d-asparagine, d-serine, and creatinine based on inulin. Bias is defined as the absolute difference between FE and 100%. (c–e) Correlations between the bias of each FE and inulin clearance. The black line is the regression line obtained by the least squares method. FE, fractional excretion.

d-Asparagine Clearance Eliminates the Need for Adjustments to the GFR Measurement

The nearly complete excretion ratio of d-asparagine suggests that C-d-Asn would be identical to C-in and could be used as a measure of GFR. Therefore, we analyzed the performance of C-d-Asn as a measure of GFR without adjusting for the coefficients. The C-d-Asn was significantly and strongly correlated with C-in (R = 0.88, P < 0.0001). The C-d-Asn measured C-in with a proportional bias of −7.8% (95% CI, −14.5% to −0.6%) and a constant bias of 6.51 (2.55–11.78), whereas this proportional bias was minor compared to that of C-cre (−34.5% [−37.9% to −31.0%]) and C-d-Ser (21.2 % [13.9%–28.9%], Table 3, Figure 3a–c; Supplementary Figure S5). The Brand–Altman plot confirmed the smaller difference in C-d-Asn in the measurement of C-in than that of C-cre and C-d-Ser (Figure 3d−f; Supplementary Figure S6). The bias of C-d-Asn (8.38 [7.48–9.28]), which was expressed as the average of the absolute value of the difference from C-in, was less than that of C-cre (30.85 [29.01–32.69], P < 0.0001) and C-d-Ser (11.55 [10.51–12.58], P < 0.0001; Table 4). The C-d-Asn was also low in the interquartile range (8.77 [6.75–10.73]) and standard deviation (6.55 [6.53–6.57]; Supplementary Table S3). The C-d-Asn was also precise based on the root-mean-square error values (9.75 [8.87–10.79]), although this precision was less than that of the C-cre (7.22 [6.50–8.12]). In terms of accuracy, C-d-Asn agreed well with C-in (P₃₀, 95.7 [92.9–98.5]; Supplementary Figure S7). The minor proportional bias is the advantage of using C-d-Asn as a measure of GFR. Even without adjustment, C-d-Asn excelled in the precise and accurate measurement of GFR.

Table 3.

Coefficients of each clearance for inulin clearance

Clearance methods	Slope (95% CI)	Intercept (95% CI)
d-Asparagine clearance	0.922 (0.855–0.994)	6.506 (2.548 to 11.778)
d-Serine clearance	1.212 (1.139–1.289)	−0.867 (−4.829 to 2.832)
Creatinine clearance	0.655 (0.621–0.690)	2.697 (−0.321 to 5.801)

CI, confidence interval.

Slopes and intercepts were determined using Deming regression.

Figure 3.

Associations between d-asparagine and inulin clearances. (a–c) Scatter plots between inulin clearance and clearances of (a) d-asparagine, (b) d-serine, and (c) creatinine. Black lines, Deming regression lines; gray-dotted lines, identical lines. (d–f) Bland-Altman plots for clearances of (d) d-asparagine, (e) d-serine, and (f) creatinine based on inulin clearance. Solid black lines, mean of the differences; gray area, 95% limits of agreement.

Table 4.

Performance of each clearance method for inulin clearance

Clearance methods	R2	Bias (95% CI)	IQR (95% CI)	RMSE (95% CI)	P30, % (95% CI)	P15, % (95% CI)
d-Asparagine clearance	0.770	8.38 (7.48–9.28)	8.77 (6.75–10.73)	9.75 (8.87–10.79)	95.7 (92.9–98.5)	66.7 (60.2–73.1)
Creatinine clearance	0.873	30.85 (29.01–32.69)	18.36 (15.65–22.70)	7.22 (6.50–8.12)	13.5 (8.8–18.2)a	0a
d-Serine clearance	0.833	11.55 (10.51–12.58)	10.84 (9.45–12.38)	8.31 (7.64–9.24)	91.3 (87.4–95.2)a	43.5 (36.7–50.3)a

CI, confidence interval; IQR, interquartile range of difference; RMSE, root-mean-square error.

Bias, absolute value of residual. Accuracy was calculated as the ratio that differed from inulin clearance by less than 30% or 15% (P₃₀ and P₁₅).

^aStatistically significant versus d-asparagine clearance.

Discussion

In the present study, we identified d-asparagine as an ideal molecule for measuring GFR. d-Asparagine was selected as a marker that reflects GFR, and the dynamics of d-asparagine after glomerular filtration satisfied the conditions for GFR measurement. The C-d-Asn measured GFR without adjustment because the proportional bias of the C-d-Asn was minor. d-Asparagine, an endogenous molecule, enables precise and convenient measurement of GFR.

d-Asparagine is a natural product that is contained in food.⁹ It has been reported as a biomarker of kidney function.^13,14,21 A trace amount of d-asparagine was detected in human blood, and the blood level of d-asparagine correlated with eGFR and predicted the worst prognosis of CKD.¹³ Despite its low presence in the blood, d-asparagine is relatively abundant in urine.¹⁴ The mean d-asparagine ratio in blood and urine was 0.483% and 16.8%, respectively (Table 1). To the best of our knowledge, further information on d-asparagine, such as its physiological function and tissue distribution, is required.

The dynamics of d-asparagine after glomerular filtration are almost identical to those of inulin. Inulin fulfills the 4 key features of GFR measurement. These features include the following: (i) free filtration at the glomeruli, (ii) no reabsorption at the tubules, (iii) no secretion at the tubules, and (iv) stable blood levels.^22,23 d-Asparagine is a small molecule and is considered to pass through the glomerular filtration barrier.²⁴ The fact that the FE_in of d-asparagine is nearly 100% suggests 2 possibilities as follows: (i) d-asparagine is subject to neither reabsorption nor secretion in the tubules, or (ii) the reabsorption and secretion of d-asparagine is balanced. The blood level of d-asparagine is stable, because it is closely correlated with GFR. Thus, d-asparagine potentiates the unbiased measurement of GFR. In contrast, creatinine and d-serine are prone to proportional biases in the measurement of GFR because they are variably handled at the kidney. Creatinine is secreted from the tubules, whereas d-serine is reabsorbed in the tubules. In addition, the dynamics of d-asparagine are relatively unaffected by GFR, as seen in the minor bias of FE_in of d-asparagine over the wide range of GFR. The relatively independent dynamics of GFR are an overlooked requirement for GFR markers to potentiate a wide range of measurements. Stability in the dynamics, regardless of GFR, should be explored as the fifth key feature in future kidney biomarkers. Overall, d-asparagine dynamics are suitable for GFR measurements.

d-Asparagine provides versatile and essential information in medicine; it overcomes the problems associated with GFR measurement. The endogenous nature of d-asparagine greatly reduces the burden on patients and medical practitioners, as well as the medical cost of GFR measurement. Among the precise methods of GFR measurement based on endogenous molecules, C-d-Asn outperforms C-cre and C-d-Ser in accuracy. Stable dynamics, regardless of GFR, is also the basis for employing C-d-Asn for precise GFR measurement in patients with both good and poor kidney functions. The precise, unbiased, and stable features of C-d-Asn are key to its clinical use. In addition to being a GFR marker, d-asparagine is applicable as an intrinsic standard. An intrinsic standard is necessary for a substrate to assess its FE or to measure its urinary protein level using spot urine. For this purpose, creatinine has been used as a standard molecule owing to its relatively stable dynamics. Because the dynamics of d-asparagine are unaffected by GFR, it helps assess FE or urinary protein levels precisely. The measurement of d-asparagine is possible in various clinical environments because it requires only a small amount of frozen blood or urine sample (<100 μl). Therefore, d-asparagine can be readily applied in medical practice and research as an intrinsic standard for GFR and FE measurements.

The use of d-asparagine as a biomarker is new in the field of science and requires further biological and physiological investigation. The profiling of d-asparagine in diseases and aging provides fundamental knowledge for medical studies. Recently, the dynamics of d-amino acids were reported to change in viral infections, including COVID-19, and this reflects their severity.12, 25 Elucidating how diseases and aging processes affect the dynamics of d-asparagine, and whether d-asparagine reflects kidney function under these circumstances, will provide key information for the use of C-d-Asn in clinics. New physiological functions of d-amino acids are being unraveled. These include the homeostatic role of d-serine in the kidney and the protective effect of d-alanine against viral infection.^10,12 Therefore, unraveling the physiological function of d-asparagine may lead to a therapeutic opportunity for kidney diseases.

This study has some limitations. The cohort size was small, and the study cohort consisted of potential living kidney donors, postdonors, and transplant recipients and did not consist of patients with kidney disease. The use of C-d-Asn shares similar limitations with C-cre in terms of urine collection accuracy. The dominance of potential living kidney donors with relatively favorable kidney function may limit the evaluation in the rest of the cohort. Further studies are required to clarify the findings of this study. Currently, the measurement and quantification of d-amino acids are possible within 10 minutes. With the increasing needs of accurate GFR measurement, it is anticipated that the measurement system will gain further throughput and accessibility.

In conclusion, d-asparagine is an ideal endogenous molecule that can be used for accurate and unbiased measurement of GFR. The measurement of C-d-Asn was stable over a wide range of GFRs.

Disclosure

TK has equity in KAGAMI Inc. TK is an inventor on issued and applied patents (WO2020080484A1, PCT/JP2020/048977), which are related to this work. All other authors declare no competing interests.

Supplementary Material

Figure S1. Correlations between inulin clearance and chiral amino acid parameters or other kidney biomarkers.

Figure S2. Blood levels and ratios of d-asparagine in plasma and serum.

Figure S3. Subgroup analysis for bias of fractional excretion (FE).

Figure S4. Subgroup analysis for bias of fractional excretion (FE).

Figure S5. Scatter plots between inulin clearance and clearances of d-asparagine, d-serine, and creatinine.

Figure S6. Bland-Altman plots for clearances of d-asparagine, d-serine, and creatinine based on inulin clearance.

Figure S7. Performance of each clearance method for inulin clearance.

Table S1. Additional characteristics of the participants.

Table S2. Characteristics of the participants.

Table S3. Performance of each clearance method for inulin clearance.

STROBE Statement.