Endogenous Oxalate Production in Primary Hyperoxaluria Type 1 Patients

Significance Statement

Primary hyperoxaluria type 1 (PH1) is a rare genetic disorder characterized by increased endogenous oxalate production (EOP). The metabolic pathways underlying oxalate synthesis have not been fully elucidated. Measurement of EOP can help evaluate PH1 drugs under development. By infusing stable isotopes of oxalate, glycolate, and glycine, we measured EOP and the contribution of glycolate to EOP and glycine production (to assess pyridoxine responsiveness) in patients with PH1 and in healthy volunteers. In this study, we provide a precise method to quantify oxalate kinetics that could serve as an additional tool to evaluate therapeutic efficacy and inform important clinical decisions (e.g., suitability for a kidney-alone transplant and prevent a liver transplant after pyridoxine or RNAi treatment).

Abstract

Background

Primary hyperoxaluria type 1 (PH1) is an inborn error of glyoxylate metabolism, characterized by increased endogenous oxalate production. The metabolic pathways underlying oxalate synthesis have not been fully elucidated, and upcoming therapies require more reliable outcome parameters than the currently used plasma oxalate levels and urinary oxalate excretion rates. We therefore developed a stable isotope infusion protocol to assess endogenous oxalate synthesis rate and the contribution of glycolate to both oxalate and glycine synthesis in vivo.

Methods

Eight healthy volunteers and eight patients with PH1 (stratified by pyridoxine responsiveness) underwent a combined primed continuous infusion of intravenous [1-¹³C]glycolate, [U-¹³C₂]oxalate, and, in a subgroup, [D₅]glycine. Isotopic enrichment of ¹³C-labeled oxalate and glycolate were measured using a new gas chromatography–tandem mass spectrometry (GC-MS/MS) method. Stable isotope dilution and incorporation calculations quantified rates of appearance and synthetic rates, respectively.

Results

Total daily oxalate rates of appearance (mean [SD]) were 2.71 (0.54), 1.46 (0.23), and 0.79 (0.15) mmol/d in patients who were pyridoxine unresponsive, patients who were pyridoxine responsive, and controls, respectively (P=0.002). Mean (SD) contribution of glycolate to oxalate production was 47.3% (12.8) in patients and 1.3% (0.7) in controls. Using the incorporation of [1-¹³C]glycolate tracer in glycine revealed significant conversion of glycolate into glycine in pyridoxine responsive, but not in patients with PH1 who were pyridoxine unresponsive.

Conclusions

This stable isotope infusion protocol could evaluate efficacy of new therapies, investigate pyridoxine responsiveness, and serve as a tool to further explore glyoxylate metabolism in humans.

Primary hyperoxalurias (PH) are rare inherited disorders of glyoxylate metabolism, characterized by increased endogenous oxalate production.¹ Oxalate is a nonfunctional end metabolite, primarily excreted by the kidneys. The increased urinary oxalate content leads to the formation of calcium oxalate crystals, which aggregate in the renal parenchyma and urinary tract, leading to nephrocalcinosis and urolithiasis, respectively. In combination with oxalate-induced tubular toxicity, these complications result in kidney failure in more than half of patients.² Once patients develop kidney failure, the deposition of oxalate in bodily tissues accelerates, resulting in systemic oxalosis, a devastating and life-threatening condition.

The estimated prevalence of PH is one per 58,000 people, as determined by whole genome sequencing.³ PH type 1 (PH1) is the most common and most severe subtype, accounting for 80% of patients. It is caused by a deficiency of the liver-specific, peroxisomal enzyme alanine:glyoxylate aminotransferase (AGT), which catalyzes the conversion of glyoxylate to glycine.² Over 150 mutations have been identified in AGXT, the gene encoding AGT (Human Gene Mutation Database, accessed January 2021).⁴ When AGT is deficient glyoxylate will accumulate, which is metabolized into oxalate and glycolate, leading to the characteristic increased formation of both compounds (Figure 1).

Figure 1.

A schematic representation of glyoxylate metabolism in HV and patients with PH1. A simplified representation of glyoxylate metabolism showing dominant metabolic pathways. HOGA, 4-hydroxy-2-oxoglutarate aldolase; GO, glycolate oxidase; LDH5, lactate dehydrogenase 5.

Patients with PH are treated with hyperhydration and alkali citrate supplementation to reduce calcium oxalate precipitation.⁵ Patients with specific AGXT mutations (c.508G>A and c.454T>A), approximately one-third of patients with PH1, respond to pyridoxine (vitamin B₆).⁶ In these patients, AGT is normally expressed and enzymatically active, at least in vitro, but is rendered functionally inactive due to its localization in the wrong subcellular compartment by mistargeting. Pyridoxine increases both the catalytic activity and peroxisome targeting of AGT in patients with mitochondrial mistargeting and results in a partial or complete normalization of oxalate excretion in responsive patients with PH1.⁷

Liver transplantation is the only curative treatment for patients with PH1. However, this procedure and the subsequent lifelong immunosuppression carry a substantial risk of morbidity and mortality. Promising novel treatments on the basis of RNA interference (RNAi) aim to reduce oxalate production in the liver by inhibiting either glycolate oxidase (GO) or lactate dehydrogenase A (LDHA).^8,9 Phase 3 trials are ongoing and recently Lumasiran, an RNAi drug targeting GO, was approved by both the European Medicines Agency and the Food and Drug Administration for PH1.¹⁰

Clinical trials in PH have used urinary oxalate excretion and plasma oxalate as surrogate outcome markers. However, intraindividual urinary oxalate excretion varies widely from day to day.¹¹ Also, 24-hour urine collections are cumbersome and fraught by collection inaccuracies. Furthermore, urinary oxalate excretion may not reflect endogenous oxalate production once calcium-oxalate crystals have accumulated in the various bodily tissues. This is due to the remobilization of oxalate stores, which results in hyperoxalemia and hyperoxaluria that may persist for years after the disease-causing enzymatic defect has been corrected by liver transplantation.¹² Finally, urine collections are impossible in patients with anuric ESKD. Both situations make it impossible to evaluate direct therapeutic efficacy in patients who present with renal failure, severely affecting clinical decision making.

Uncertainties remain about the metabolic pathways underlying hyperoxaluria. One of the crucial gaps in our understanding is the quantitative contribution of potential precursors to endogenous oxalate production. Previous studies suggest both glycine and hydroxyproline make only a minor contribution to oxalate synthesis.^13,14 One could thus assume glycolate is the main precursor, but this has not been studied. Understanding the molecular pathophysiology of PH1 could contribute to the development of treatments for all patients with PH.

Stable isotope tracers provide an ideal tool to investigate (deranged) metabolic fluxes and have found widespread application, including inborn errors of metabolism and cardiovascular disease.^15–17 We developed a stable isotope infusion protocol to study glyoxylate metabolism in humans, which entailed the development of a new gas chromatography–tandem mass spectrometry method.¹⁸ In this study, we applied this method to assess the oxalate rate of appearance (Ra) and the quantitative contribution of glycolate to oxalate synthesis in both patients with PH1 and healthy volunteers (HV). Glycine kinetics were assessed to evaluate (residual) AGT activity and thus pyridoxine responsiveness.

Methods

Subjects

Eight HV and eight patients with PH1 were included. Inclusion criteria for patients were: PH1 diagnosis confirmed by mutation analysis, age 18–65 years, and an eGFR >15 ml/min per 1.73 m² body surface area using the Chronic Kidney Disease Epidemiology Collaboration formula.¹⁹ In HV, previous medical history and plasma creatinine were assessed to ensure good health.

Stable Isotope Infusion Protocol

Study subjects maintained a low-oxalate diet 3 days before the study day and started fasting 12 hours preceding baseline assessment. All participants were asked to avoid the intake of well-known high-oxalate foods (e.g., spinach and rhubarb) and vitamin C supplements. During hospitalization they were allowed to eat our standard in-hospital diet. After admission to the metabolic ward, two intravenous catheters were inserted into an antecubital vein, one in each arm. One catheter was used to administer the stable isotopes, the other for repeated blood sampling. Baseline blood samples were obtained to assess background enrichments of oxalate, glycolate, glycine, and glyoxylate. Then three primed continuous infusions of [U-¹³C₂]oxalate, [1-¹³C]glycolate, and [D₅]glycine dissolved in 0.9% saline were started. The stable isotope infusion protocol and timing of sample collection are represented in Figure 2.

Figure 2.

Stable isotope infusion and sampling protocol. T0 (expressed in hours) is start of continuous infusion, bolus was given directly prior.

For the first patient and HV, the bolus dose and infusion rates for [U-¹³C₂]oxalate and [1-¹³C]glycolate were estimated on the basis of plasma oxalate and glycolate concentrations, total body water, and previous findings.²⁰ After analysis of the resulting high enrichments, we adjusted tracer dosages for the other HV and patients with PH1. For [U-¹³C₂]oxalate, the infusion rate was 0.06 µmol/kg per hour in HV and 0.2 µmol/kg per hour in patients with PH1 for a duration of 10 hours. The infusion rate for [1-¹³C]glycolate was 0.9 µmol/kg per hour in HV and 4.0 µmol/kg per hour in patients with PH1 for 6 hours. Priming bolus for both tracers and subject groups consisted of a 1–2.5-hour dose administered over a 2-minute period. For [D₅]glycine, the infusion rate was 20 µmol/kg per hour for both patients and HV, and the bolus consisted of a 2-hour dose.²¹ The study protocol was approved by the institutional review board of the Amsterdam University Medical Centers, location AcademicMedical Center (AMC). All subjects gave informed consent.

Sample Collection

Hourly blood samples were collected in ice-cooled (0°C) heparin tubes, with time of start of infusion being t=0. Samples were centrifuged within 30 minutes of collection for 10 minutes at 1700 × g at 4°C. Then 200 µl of plasma (in quadruplicate) was transferred in 2 ml glass screw-capped vials. Next, 30 µl of concentrated HCl (37%) was added, and mixed using a vortex. Immediate acidification (and deproteinization) after centrifugation prevents oxalate crystallization during cold storage and in vitro oxalogenesis from ascorbic acid. For glycine measurements residual plasma was transferred in plastic 2 ml cryo vials, without acidification. Samples were then stored at −20°C until analysis.

Chemicals

Clinical Trial Materials grade sodium glycolate (1-¹³C, 99%), sodium oxalate (U-¹³C₂, 99%), and glycine (D₅, 98%) were purchased from Cambridge Isotope Laboratories (Tewksbury, MA).

Method of Analysis

The enrichments of [1-¹³C]glycolate, [1-¹³C]oxalate, and [U-¹³C₂]oxalate were determined in plasma after derivatization of the analytes with ethylhydroxylamine and N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide. Analysis was performed using a 7890A gas chromatography (GC) coupled to a 7000 Triple Quadrupole Mass Spectrometer (Agilent Technologies, Amstelveen, the Netherlands). Preanalysis sample work-up, multiple reaction monitoring mode settings, mass spectrometry acquisition conditions, and assessment of validity were previously described in more detail by van Harskamp et al.¹⁸ Glycine enrichments and concentration were analyzed after derivatization with ethylchloroformate, using a previously developed method.^21,22 One sample per time point was prepared for both the assessment of [1-¹³C]glycine enrichment (using a Delta-V isotope ratio mass spectrometer coupled online with a Trace 1310 gas chromatograph and GC Isolink II [Thermo Fisher Scientific, Bremen, Germany]) and [D₅]glycine enrichment (after dilution of the samples, analyzed using a GC-MSD (5975c, Agilent, Amstelveen, the Netherlands). Glycine concentrations were determined in the t=0 samples after addition of [U-¹³C₂, 2,2-D₂, ¹⁵N]glycine as internal standard by Gas chromatography-mass selective detector (GC-MSD).

Calculations

Ra equals rate of disappearance once isotopic equilibrium has been reached. During fasting conditions, the Ra of oxalate in plasma of subjects without renal impairment reflects endogenous oxalate production, because no tissue storage and release of oxalate is expected as the kidneys are still able to excrete the excess amount of oxalate. This is also the case for glycolate, independent of kidney function. In all subjects, the Ra of oxalate, glycolate, and glycine was calculated with Steele’s steady-state equation (Equation 1).²³

$$\mathrm{Ra}=\mathrm{Rd}=\left(\mathrm{I}\times {\mathrm{E}}_{\mathrm{i}}\right)/\left({\mathrm{E}}_{\mathrm{p}}\times 100\right)-\mathrm{I}$$ Equation 1

where I is the tracer-to-tracee ratio (TTR) infusion rate (µmol/kg per hour); E_i and E_p represent the enrichment expressed as mole percent excess (MPE=TTR/[1+TTR]) of the tracer infusate and plasma, respectively; and Rd is the rate of disappearance.

The fractional synthetic rate (FSR) is calculated from the rate of tracer incorporation over time. The FSR of oxalate from glycolate was determined by measurement of the incorporation of the (M+1) labeled glycolate tracer (precursor) over time in oxalate (product) using Equation 2, as a measure of the relative contribution of glycolate to the oxalate pool per time unit.

$$\mathrm{FSR}(\%/\mathrm{h})=(\mathrm{Et2}-\mathrm{Et1})/\mathrm{Ep}\times 1/(\mathit{t}\mathrm{2}-\mathit{t}\mathrm{1})\times 100$$ Equation 2

where (Et2−Et1) represents the change in product (oxalate M+1) enrichment (expressed in MPE) between the time points t2 and t1 (expressed in hours), during which the precursor enrichment is in steady state. Ep is the precursor enrichment (glycolate M+1) during isotopic equilibrium. The absolute synthesis rate (ASR), that is, the contribution of glycolate as a precursor to oxalate production, was then calculated using the total oxalate pool size using Equation 3.

$$\mathrm{ASR}(\mathrm{mmol}/\mathrm{day})=(\mathrm{FSR}/100)\times (\mathrm{POX}\times 0.6\mathrm{L}/\mathrm{kgBW})\times 24$$ Equation 3

where ASR is absolute synthetic rate, POX denotes plasma oxalate concentration (µmol/L), and BW is body weight. The oxalate pool is estimated as POX×total body water (0.6 L/kg BW). The factual volume of distribution for oxalate is unknown, but the existence of an oxalate metabolic pool physiologically outside the extracellular fluid compartment has been reported.²⁴ Taking into account the fact that oxalate can deposit in virtually all tissues, we assumed oxalate is in rapid dynamic equilibrium for the purpose of this model. Finally, the quantitative contribution (%) of glycolate to oxalate production can be determined by dividing ASR (determined for M+1 oxalate dilution) by the endogenous production rate (Ra) of glycolate.

For the calculation of the contribution of glycolate to glycine synthesis in both patients with PH1 and HV and the glycolate to oxalate incorporation in HV, the above-mentioned FSR/ASR calculations were inappropriate because both precursor and product steady-state enrichments were reached during infusion. For this reason, we rewrote the hydroxylation Clarke and Bier model, in combination with the adjustments proposed by Thompson et al. (Equation 4).^25,26

The rate of glycolate conversion to glycine or oxalate (Q_gp; µmol/kg per hour) was derived using similar principles to those applied in Equation 1.

$${\mathrm{Q}}_{\mathrm{gp}}={\mathrm{Q}}_{\mathrm{p}}\times \left({\mathrm{E}}_{\mathrm{p}}/{\mathrm{E}}_{\mathrm{g}}\right)\times \left({\mathrm{Q}}_{\mathrm{g}}/{\mathrm{i}}_{\mathrm{g}}+{\mathrm{Q}}_{\mathrm{g}}\right)$$ Equation 4

where Q_p and Q_g are the product (either glycine or oxalate) and glycolate fluxes (µmol/kg per hour) estimated independently by Equation 1 and E_p and E_g are the respective enrichments of glycine or oxalate (M+1) and precursor enrichment. For subjects in whom Q_p (glycine flux) was unavailable (because the first four patients with PH1 and two HV did not receive [D₅]glycine infusion), mean values from other subgroup members were extrapolated and used.

Statistics

Integration of the peak area was performed using Mass Hunter Quantitative Analysis software version B.08.00 (Agilent Technologies, Amstelveen, The Netherlands). The slope and intercept of the enrichment curves (experimental TTR versus theoretical TTR) were determined by linear regression analysis. Data are expressed as mean (SD) and visualized with box–whisker plots, unless otherwise stated. Between-group differences were analyzed with nonparametric Kruskal–Wallis tests (three groups) or Wilcoxon (two groups), followed by post hoc tests with correction for multiple testing with the Bonferroni–Holm method. P<0.05 was considered statistically significant. All calculations were performed using SPSS version 26 (IBM Corp., Armonk, NY) and R version 3.5.1 (the R Foundation for Statistical Computing, Vienna, Austria).

Results

Eight patients with PH1 and eight HV were included. Demographic and clinical data are shown in Tables 1 and 2. On the basis of the results of mutational analysis, reported clinical response and active treatment with pyridoxine, patients were categorized as pyridoxine unresponsive (n=3) or responsive (n=5). Three patients with PH1 were homozygous for AGXT mutations associated with AGT function amendable to pyridoxine treatment. All three had shown clinical response and were on pyridoxine supplementation at the time of the study. Three patients with PH1 were compound heterozygous. Of these patients, only two had shown partial clinical response and were taking pyridoxine. The other had demonstrated no clinical response and was therefore classified as pyridoxine unresponsive. Two patients were homozygous for a null mutation (Ex5_11del) consistent with pyridoxine unresponsiveness and did not use pyridoxine. HV were, on average, older than patients. Both plasma oxalate and glycolate at baseline (t=0) were significantly higher in patients with PH1 as compared with HV. Plasma glycolate concentrations (upper reference limit: 22.0 µmol/L) at baseline varied widely between patients with PH1 (21.7–135.0 µmol/L). Mean intraindividual variation expressed as a coefficient of variation (CV_intra=SD/mean) was 33.1% and 52.3% for 24-hour urinary oxalate and glycolate excretion, respectively.

Table 1.

Demographic characteristics of study participants

Characteristics	PH1 B6−	PH1 B6+	HV
N	3	5	8
Age, yr	28.8±9.4	31.0±15.1	34.6±9.9
Sex, F, n, %	1, 33	4, 80	5, 62.5
Weight, kg	87.7±16.9	74.1±9.9	67.3±12.0
eGFR, ml/min per 1.73 m2	87.9±12.3	48.5±21.6	93.0±14.2
Metabolite profile
Plasma oxalate, µmol/L	10.9±3.3	12.6±4.4	2.8±1.5
Plasma oxalate CV intra, %	34.0±13.2	31.7±14.2	—
Plasma glycolate, µmol/L	80.9±41.1	49.4±28.5	5.2±3.8
Plasma glycolate CV intra, %	30.4±1.7	51.7±23.8	—
Urine oxalate, mmol/mmol creatinine	0.15±0.01	0.13±0.07	0.03±0.01
Urine oxalate CV intra, %	33.0±21.9	33.0±14.7	—
Urine oxalate, mmol/24 h	2.21±0.61	1.23±0.38	0.46±0.09
Urine oxalate CV intra, %	27.5±22.0	36.4±9.7	—
Urine glycolate, mmol/mmol creatinine	0.88±59.0	0.16±0.15	0.042±0.01
Urine glycolate CV intra, %	89.5±59.0	46.9±31.9	—
Urine glycolate, mmol/24 h	3.53±1.19	1.04±0.57	0.56±0.17
Urine glycolate CV intra (%)	57.2±14.0	46.9±21.5	—
Plasma glycine, µmol/L	268.9±88.5	271.5±91.0	219.0±68.0
Treatment
Pyridoxine, n, %	0, 0	5, 100	0, 0

Weight, eGFR, and serum and urine metabolite levels are presented as mean±SD. B6-, vitamin B6 unresponsive; B6+, vitamin B6 responsive; F, female; CV, coefficient of variation.

Table 2.

AGXT mutational analysis and pyridoxine treatment of patients with PH1

Patient	Allele 1	Allele 2	Pyridoxine (Vitamin B6), Dose, mg/d
PAT-001	c.33dupC	c.454T>Aa	Yes, 600
PAT-002	c.508G>A+c.1007T>Ab	c.454T>Aa	Yes, 100
PAT-003	c.508G>Aa	c.508G>Aa	Yes, 200
PAT-004	c.454T>Aa	c.454T>Aa	Yes, 50
PAT-005	c.454T>Aa	c.508G>Aa	Yes, 200
PAT-006	c.33dupC	c.508G>Aa	No
PAT-007	Ex5_11del	Ex5_11del	No
PAT-008	Ex5_11del	Ex5_11del	No

^aPyridoxine-sensitive mutation.

^bAGT double mutant, results in a serious pathogenic effect.

Oxalate Kinetics

Plasma enrichment of (M+2) oxalate during the 10-hour primed continuous infusion of [U-¹³C₂]oxalate are shown in Supplemental Figure 1. The first patient and HV received a significantly higher tracer dose (0.9 and 0.2 µmol/kg per hour, respectively) as compared with the other patients with PH1 and HV (0.6 and 0.06 µmol/kg per hour). Mean (M+2) oxalate enrichments (expressed as MPE) during plateau were 16.6% in patients with PH1 and 11.0% in HV (except for the first patient and HV who achieved mean plateau enrichments of 49.3% and 51.3%, respectively). Mean (SD) Ra of oxalate was 2.71 (0.54), 1.46 (0.23), and 0.79 (0.15) mmol/d in subjects with pyridoxine unresponsive PH1, pyridoxine responsive PH1, and HV, respectively (P=0.002) (Figure 3).

Figure 3.

Oxalate Ra. Oxalate rates of appearance of HV and patients with PH1 as assessed by dilution of (M+2) oxalate enrichments during the 10-hour primed, continuous infusion of [U-¹³C₂]oxalate expressed as (A) μmol/kg per hour and (B) mmol/d. Between-group differences were analyzed with nonparametric Kruskal–Wallis tests (three groups) followed by Wilcoxon pairwise comparisons with Bonferroni–Holm P value adjustment; (A) Kruskal–Wallis H test (P=0.003); (B) Kruskal–Wallis H test (P=0.002). B6+, patients with PH1 who were pyridoxine responsive; B6-, patients with PH1 who were pyridoxine unresponsive. *P<0.05; **P<0.01.

Glycolate Kinetics

An increase in plasma enrichment of (M+1) glycolate during the 6-hour infusion of [1-¹³C]glycolate was observed in all subjects until isotopic steady state was achieved (Supplemental Figure 2). Enrichments in two patients (PAT-001 and PAT-006) did not reach steady state and were therefore excluded from further calculations. Mean (M+1) glycolate enrichment was 45.2 MPE in patients with PH1 and 22.3 MPE in HV. Mean (SD) glycolate Ra was 9.6 (3.2), 9.0 (2.4), and 5.6 (1.3) mmol/d in subjects with PH1 (pyridoxine unresponsive), PH1 (pyridoxine responsive), and HV, respectively (P=0.03) (Figure 4).

Figure 4.

Glycolate Ra. Glycolate rates of appearance in HV and patients with PH1 as assessed by dilution of (M+1) glycolate enrichments during the 6-hour primed, continuous infusion of [1-¹³C]glycolate expressed as (A) μmol/kg per hour and (B) mmol/d. Between-group differences were analyzed with nonparametric Kruskal–Wallis tests (three groups) followed by Wilcoxon pairwise comparisons with Bonferroni–Holm P value adjustment; (A) Kruskal–Wallis H test (P=0.14); (B) Kruskal–Wallis H test (P=0.03). *P<0.05; **P<0.01.

Contribution of Glycolate to Endogenous Oxalate Production

After infusion with [1-¹³C]glycolate (precursor), the incorporation of the tracer in oxalate (measured as [M+1] oxalate) allows the calculation of the FSR of oxalate from glycolate. Mean (SD) FSR (%/h) was 7.7 (0.3) in patients with PH1 who were pyridoxine unresponsive versus 6.5 (2.4) in patients who were pyridoxine responsive (P=0.35) (Figure 5). The ASR of [1-¹³C]oxalate from glycolate (mean [SD] mmol/d) was slightly higher in patients with PH1 who were pyridoxine unresponsive (1.0 [0.5]) as compared with those who were pyridoxine responsive (0.7 [0.3]) but did not reach significance. In contrast, in HV the conversion of glycolate to oxalate was negligible. The mean contribution of glycolate to total endogenous oxalate production was 38.5% in pyridoxine unresponsive and 52.5% in patients with PH1 who were pyridoxine responsive as compared with only 1.3% in HV (P=0.007).

Figure 5.

Fractional synthetic rate and absolute synthetic rate of oxalate from glycolate and contribution of glycolate to endogenous oxalate production. (A) Fractional synthetic rate of oxalate from glycolate, (B) absolute synthetic rate of oxalate from glycolate, and (C) contribution of glycolate to total endogenous oxalate production in HV and patients with PH1 as assessed by dilution of (M+1) oxalate enrichments during the 6-hour primed, continuous infusion of [1-¹³C]glycolate. Between-group differences were analyzed with nonparametric Kruskal–Wallis tests (three groups) followed by Wilcoxon pairwise comparisons with Bonferroni–Holm P value adjustment; (A) Wilcoxon test (P=0.35); (B) Kruskal–Wallis H test (P=0.008); (C) Kruskal–Wallis H test (P=0.007). OX, oxalate; EOP, endogenous oxalate production; *P<0.05; **P<0.01.

Glycine Kinetics and Conversion of Glycolate to Glycine

Six HV and four patients with PH1 underwent a 6-hour primed, continuous infusion of [D₅]glycine. For plasma enrichments see Supplemental Figure 3. Mean (SD) Ra of glycine was 188.5 (38.0) in HV and 195.5 (39.8) µmol/kg per hour in patients with PH1 (P=0.52). The conversion rate of glycolate to glycine, which is catalyzed by AGT and therefore an adequate marker for the residual AGT activity in vivo, was calculated using both the incorporation of [1-¹³C]glycolate in glycine (measured as [M+1] glycine) and the glycine Ra as calculated after the [D₅]-glycine infusion (see Figure 6, Supplemental Figure 4). Two HV and four patients with PH1 did not undergo [D₅]glycine infusion. However, we did perform measurements of (M+1) glycine enrichments in these subjects, thus enabling the calculation of the rate of conversion of glycolate to glycine using the mean value of glycine Ra of the subjects who underwent the primed, continuous infusion of [D₅]glycine. Under the conditions of the study, 0.8 (0.3) µmol/kg per hour (mean [SD]) of glycolate was converted to glycine in HV, which accounted for <0.5% of total glycine turnover rate. In patients with PH1 (pyridoxine unresponsive and responsive), these conversion rates were 0.05 (0.02) and 1.11 (0.51) µmol/kg per hour, respectively, which also accounted for <0.5% of glycine turnover rate. The highest turnover rate, 1.8 µmol/kg per hour, was found in PAT-003, who is homozygous for c.508G>A and had the best clinical response to pyridoxine treatment.

Figure 6.

Glycine Ra and conversion rate of glycolate into glycine. Glycine rates of appearance (A) in HV and patients with PH1 as assessed by dilution of (M+2) glycine enrichments during the 6-hour primed, continuous infusion of [D₅]glycine and conversion rate of glycolate into glycine (B) as assessed by dilution of (M+1) enrichments during the 6-hour primed, continuous infusion of [1-¹³C]glycolate. Between-group differences were analyzed with nonparametric Kruskal–Wallis tests (three groups), followed by Wilcoxon pairwise comparisons with Bonferroni–Holm P value adjustment; (A) Wilcoxon test (P=0.29); (B) Kruskal–Wallis H test (P=0.04). *P<0.05; **P<0.01.

Discussion

In this study, we applied a novel method using [U-¹³C₂]oxalate, [1-¹³C]glycolate, and [D₅]glycine to assess oxalate kinetics in HV and patients with PH1. Our primed continuous infusion protocol yielded stable enrichment plateaus for all tracers administered (albeit not in all subjects), thus allowing us to assess endogenous rates of appearance of oxalate, glycolate, and glycine. Second, incorporation of the singly labeled glycolate tracer into newly formed oxalate enabled us to determine the contribution of this precursor to endogenous oxalate production. Simultaneously, this method enabled us to evaluate in vivo residual AGT activity by analyzing the incorporation of glycolate tracer into glycine. Our results show endogenous oxalate production in patients with PH1 who are unresponsive to pyridoxine was significantly higher than in both patients with PH1 who were pyridoxine responsive and HV. Furthermore, we found that glycolate is an important source of oxalate synthesis in patients with PH1, but not in healthy individuals. Finally, we found a significantly lower conversion rate of glycolate to glycine in patients who were pyridoxine unresponsive as compared with patients who were pyridoxine responsive, whose conversion rates approached those of HV.

Oxalate is mainly eliminated via renal excretion. Elder and Wyngaarden showed that the urinary recovery of injected [¹⁴C]-oxalate ranged between 89% and 100%.²⁷ Therefore, 24-hour urinary oxalate excretion is the best available assessment method of oxalate synthesis, and thus used as a surrogate endpoint for clinical trials in PH. Besides previously highlighted concerns, including high variability of 24-hour oxalate excretion, another constraint of the static measurement of concentrations in urine (or in plasma) is that these provide no information about the dynamics of the metabolic pathways involved, which limits their interpretation when assessing the functional consequences of inhibiting specific enzymes in glyoxylate metabolism (e.g., what is the metabolic fate of glyoxylate in PH patients treated with RNAi to inhibit LDHA). This is valuable information in the search for appropriate therapeutic targets in PH.

The direct measurement of oxalate synthesis in humans by means of [¹³C]oxalate infusion has not been reported previously. Huidekoper et al. (which includes some of our research group) were the first to attempt this, but the results of this pilot study including three patients with PH1 and three HV were never published in a peer-reviewed journal.²⁰ Our calculations show oxalate Ra in patients with PH1 would amount to 1.28–3.16 mmol/d, dependent on the extent of pyridoxine responsiveness. This is in accordance with previous estimates on the basis of daily urinary oxalate excretion rates in patients with PH1.^27,28 As expected, oxalate Ra in patients with PH1 was significantly higher than in HV. Despite small patient numbers, we were able to differentiate between patients with PH1 who were pyridoxine responsive and patients who were unresponsive, in the sense that oxalate Ra was lower in patients who were pyridoxine responsive.

Interestingly, oxalate Ra in HV was higher than expected. Mean daily oxalate Ra in our study was 0.79 mmol/d, which is almost double the known upper reference limit for urinary oxalate excretion in healthy subjects. This overestimation is most likely caused by dietary oxalate intake during the tracer infusion protocol. Dietary oxalate dilutes the tracer-to-tracee ratio of the oxalate pool, resulting in lower oxalate enrichments and consequently in higher oxalate rates of appearance, which is the sum of both the endogenous production and exogenous (dietary) sources. The contribution of dietary oxalate intake to daily oxalate excretion rate in healthy individuals has been reported to amount to 30%–70%, dependent on the bioavailability of ingested oxalate.²⁹ This will certainly have affected Ra calculations in healthy individuals. In patients with PH1, in whom endogenous production rates are higher and plasma oxalate pools significantly bigger, the contribution of dietary oxalate is likely to be less. Second, enteric cycling of oxalate tracer may also have contributed to the possible overestimation of the rate of oxalate appearance in HV.³⁰ Oxalate is known to be excreted into the gut lumen and is potentially reabsorbed (together with oxalate from the diet). Because the rate of oxalate cycling has not been quantified, it is hard to fathom its effect on measured enrichments. Finally, the administration of glycolate tracer may itself have stimulated oxalate production. This metabolic effect of tracer administration on calculations of oxalate Ra is probably minimal in healthy individuals due to the limited contribution to oxalate of this precursor molecule (see below).

Endogenous glycolate production has never been assessed before and the sources of glycolate are not entirely known. Glycolate Ra was significantly higher in patients with PH1 as compared with HV. However, endogenous glycolate Ra in patients with PH1 varied widely and were not directly associated with pyridoxine responsiveness. This is in line with the wide range of plasma glycolate concentrations and glycolate excretion rates found in patients with PH1, irrespective of the clinical response to pyridoxine. In our study we found the intraindividual variation (CV_intra) in 24-hour urinary excretion of glycolate was even higher than that of oxalate (mean CV_intra of 52.3% versus 33.1%, see Table 1). In one patient with PH1 who was pyridoxine responsive, plasma glycolate levels were in the high-normal range and glycolate Ra (5.7 mmol/d) resembled the mean glycolate Ra found in HV (5.6 mmol/d). This suggests the production of glycolate in patients with PH1 is complex and influenced by multiple factors. We hypothesize that the relative activities of two key enzymes in the glycolate-glyoxylate cycle (GR/HPR and GO) determine the production rate of glycolate in patients with PH1.

Glycolate is oxidized to oxalate via glyoxylate primarily by LDH, but the relative contribution to oxalate synthesis has previously not been investigated in patients with PH1. A study in rats using orally administered [1-¹⁴C]glycolate found that 1.8%–4.8% of the radioactive label was recovered as urinary [1-¹⁴C]oxalate.³¹ A marginal contribution of glycolate to oxalate production was also reported in two HV, in whom only 1% of intravenously administered [¹⁴C]glycolate was converted into oxalate over 24 hours.³² This is in line with our findings in HV. Our study is the first to investigate the contribution of glycolate to oxalate production in patients with PH1, and the first to assess the absolute quantitative contribution to oxalate production. The results showed that glycolate is responsible for at least half of the endogenous oxalate Ra in patients with PH1. This is possibly an underestimation because glycolate is not converted to oxalate itself. The intermediate glyoxylate, the product of glycolate conversion by GO and the direct precursor of oxalate, is possibly diluted from other sources, leading to an underestimation of the factual FSR and ASR. These findings are in line with the published ILLUMINATE-A data and support GO as a viable therapeutic target for substrate reduction therapy in PH1.¹⁰

A subgroup of subjects underwent infusion of [D₅]glycine. Glycine Ra did not differ between HV and patients with PH1 and were in line with previous reports.^33,34 These measurements, combined with the simultaneous infusion of [1-¹³C]glycolate, enabled us to assess the contribution of glycolate as a precursor of glycine. Because AGT catalyzes the final step in the conversion of glycolate to glycine (via glyoxylate), these measurements amount to the in vivo assessment of residual AGT enzyme activity. Although other human transaminases can convert glyoxylate into glycine, these enzymes are unable to compensate for the lack of AGT activity in patients with PH1.³⁵ In addition, in patients with PH1 with a null-mutation (resulting in complete absence of AGT), turnover rates (of glycolate to glycine) stayed close to zero as (M+1) enrichments of glycine remained under the detection limit of the analysis method, suggesting the role of other transaminases is negligible. More importantly, we found that patients with PH1 who were pyridoxine responsive had turnover rates similar to those in HV. This methodology provides a novel tool to assess pyridoxine sensitivity and could significantly contribute to the clinical management of patients with PH1 presenting with ESKD at time of diagnosis. In the current situation, pyridoxine sensitivity cannot be assessed reliably in patients who are oligo-anuric. Also, plasma oxalate levels poorly reflect the endogenous oxalate production because much of the produced oxalate may be stored in bodily tissues. Yet, assessing pyridoxine sensitivity is of importance because it might relieve the need for liver transplantation in such patients. More research is necessary, but we postulate this is the first step toward a tool to quantify residual AGT activity in vivo and thus to assess the need for liver transplantation in patients with PH1 who are oligo-anuric.

Limitations

Group sizes were relatively small due to the rarity of the disease studied and the cost of these experiments. Previously, the [1-¹³C]glycolate tracer had not been administered to human subjects, which increased both manufacturing and pharmacy costs. Due to strict national regulations to ensure the safety of study participants and aiming to apply our measurement method to evaluate therapeutic efficacy of RNAi drugs under development, we only used top-grade materials. For all patients with PH1, multiple 24-hour urine oxalate readings were readily available, but taking into account the diurnal intraindividual variability in urine oxalate excretion and that the analytical method used to quantify oxalate concentrations measures the sum of both the endogenous and infused oxalate molecules, we decided not to collect urine samples during the 10-hour hospitalization. In both groups we decided to conduct the study in the fed state, after an overnight fast and low-oxalate diet. In patients with PH1 we do not expect a major effect from diet, but in the HV group, an overestimation of endogenous oxalate Ra may have occurred as a result of unlabeled oxalate entering the oxalate pool from other sources than endogenous production. In future research, we will further investigate the influence of diet during our infusion protocol.

Possible Clinical Applications and Future

Direct assessment of endogenous oxalate production, the main feature of PH, could be an ideal parameter to evaluate therapeutic efficacy and can serve as a dynamic tool for new RNAi treatments. The first clinical trials in patients with PH1 using either GO or LDHA RNAi are being conducted and, with our method, we can now assess the effect of these drugs on endogenous oxalate production and identify other functional consequences (e.g., change in glycolate kinetics after GO inhibition and determine the metabolic fate of accumulated glyoxylate after LDHA inhibition), which will further improve our understanding of glyoxylate metabolism.

Preliminary studies are underway to validate our isotope method in patients with ESKD. This would be of great benefit because there is no clinical parameter available to monitor therapeutic response in these patients. It may also aid in choosing the optimal therapeutic approach (between either liver-kidney transplantation or sole kidney transplantation and eligibility for new, potentially costly, RNAi medication in the near future).

The protocol can be personalized to the specific needs. Both [D₅]glycine and [1-¹³C]glycolate infusions are required to assess pyridoxine responsiveness. In addition, [1-¹³C]glycolate infusion is mandatory to assess the contribution of glycolate to oxalate production, which is important in the case of GO-inhibition, but not necessary if the question is solely to evaluate the efficacy of these novel RNAi therapeutics to lower oxalate production. In the latter situation, sole infusion with [U-¹³C₂]oxalate may be considered, which will further tone down the costs and complexity of performing this analysis. The application for this specific reason (i.e., therapeutic efficacy of RNAi) involves a baseline measurement pre-RNAi, followed by a post-dose experiment once maximal effect is expected.

Moreover, we aim to apply our novel method in both patients with PH2 and PH3, to further explore glyoxylate metabolism, evaluate therapeutic efficacy and elucidate the contribution of (un)known sources of endogenous oxalate overproduction.

Disclosures

C. van den Akker reports having consultancy agreements with Baxter and Nutricia Early Life Nutrition; and reports receiving honoraria from Baxter, Nestlé Nutrition Institute, and Nutricia Early Life Nutrition. J. Groothoff reports consultancy agreements with UniQure; and reports being a scientific advisor or membership with Alnylam. J.W. Groothoff, M.J.S. Oosterveld, and S.F. Garrelfs report receiving an unconditional grant from Alnylam Pharmaceuticals and Dicerna Pharmaceuticals during the conduct of the study. H. Peters-Sengers reports receiving research funding from KOLFF and a personal grant from the Dutch Kidney Foundation. M.J.S. Oosterveld reports receiving honoraria from Apellis Pharmaceuticals. All remaining authors have nothing to disclose.

Funding

This work was funded by the Amsterdam UMC, a personal scholarship to S.F. Garrelfs, Dicerna Pharmaceuticals (unconditional grant), Alnylam Pharmaceuticals (unconditional grant), and Metakids (2019-04-UMD).

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2021060729/-/DCSupplemental.

Supplemental Figure 1. Plasma enrichments of (M+2) oxalate during the 10-hour primed continuous infusion of [U-¹³C₂]sodium oxalate in patients with PH1 (A) and HV (B).

Supplemental Figure 2. Plasma enrichments of (M+1) glycolate during the 6-hour primed continuous infusion of [1-¹³C]glycolate in patients with PH1 (A) and HV (B).

Supplemental Figure 3. Plasma enrichments of [D₅]-glycine during the 6-hour primed continuous infusion of [D₅]-glycine in patients with PH1 and HV.

Supplemental Figure 4. Plasma enrichments of (M+1) glycine during the 6-hour primed continuous infusion of [1-¹³C]glycolate in patients with PH1 (A) and HV (B).