CV_a: Coefficients of Variations; HREC: Human Research Ethics Committee; ICS: Ion Chromatography System; LOD: Limit of Detection; LOQ: Limit of Quantification; NSAIDS: Nonsteroidal Anti-Inflammatory Drugs; OMIM: Online Mendelian Inheritance in Man; QC: Quality Control; SD: Standard Deviation; SLC: Solute Linked Carrier; STD: Standard.

Sulfate (SO₄^2-) is an important nutrient for human growth and development [1]. Sulfate conjugation (sulfonation) of glycosaminoglycans, such as heparan sulfate and chondroitin sulfate, is required to maintain the normal structure and function of tissues [2,3]. In addition, sulfonation detoxifies xenobiotics and certain pharmacological drugs, including acetaminophen [4,5]. Sulfonation also inactivates steroids and thyroid hormone, and contributes to the biotransformation of bile acids, catecholamines and cholecystokinin [6-10].

To date, more than 20 disorders have been linked to genetic defects that perturb sulfonation or sulfatase capacity [11]. For example, defects in several sulfatase genes are linked to metachromatic leukodystrophy (OMIM 250100), X-linked ichthyosis (OMIM 308100), chondrodysplasia punctata 1 (OMIM 302950), mesomelia-synostoses syndrome (OMIM 600383) and mucopolysaccharidoses types II (OMIM 309900), IIIA (OMIM 252900), IIID (OMIM 252940), IVA (OMIM 253000) and VI (OMIM 253200). In addition, mutations in the solute linked carrier 26A2 gene (SLC26A2) lead to under-sulfonation of proteoglycans in chondrocytes and cause a range of mild to lethal forms of chondrodysplasias: multiple epiphyseal dysplasia (OMIM 226900), diastrophic dysplasia (OMIM 222600), atelosteogenesis Type II (OMIM 256050) and Achondrogenesis Type IB (OMIM 600972) [12]. Collectively, these disorders highlight the physiological consequences of perturbed sulfate homeostasis.

In adults and children, approximately two thirds of daily sulfate requirements are met from the intracellular metabolism of sulfurcontaining amino acids [13,14]. Sulfite oxidase deficiency, a disorder with severe neurological consequences, perturbs the ability to generate sulfate from amino acid oxidation [15,16]. In addition, the human foetus lacks the ability to derive sulfate from amino acid oxidation and relies on sulfate supplied from maternal circulation via the placenta [17]. During human and animal gestation, maternal circulating sulfate levels increase by approximately 2-fold to meet the high fetal demands for sulfate [18,19]. Interest in human circulating sulfate level has expanded following studies that link hyposulfatemia to animal pathologies [11].

In humans, circulating sulfate levels are maintained at approximately 300 to 400 μmol/L [20]. The kidneys filter sulfate in the glomerulus and then reabsorb the majority (approximately 80%) of filtered sulfate back into circulation [21]. The first step of sulfate reabsorption is mediated by the SLC13A1 sulfate transporter on the apical membrane of the renal proximal tubule, and the second step via SLC26A1 on the basolateral membrane [22]. Loss of Slc13a1 and Slc26a1 in mice leads to renal sulfate wasting, hyposulfatemia, growth retardation, and behavioural abnormalities [21,23-25]. The depleted sulfate levels in circulation reduce intracellular sulfonation capacity, as shown in the Slc13a1 and Slc26a1 null mice [23,26]. In addition, genetic defects in SLC13A1 have been linked to hyposulfatemia, growth retardation and osteochondrodysplasias in dogs and sheep [27,28]. More recently, genetic variants in SLC13A1 were linked to renal sulfate wasting and hyposulfatemia in humans [29,30]. Furthermore, SLC13A1 gene expression is down-regulated by vitamin D depletion, hypokalaemia, metabolic acidosis, NSAIDS and glucocorticoids [31]. Collectively, the above studies have shown the dietary, genetic and physiological contributions to altering circulating sulfate levels and renal sulfate excretion. Those findings, together with animal studies showing the adverse physiological consequences of renal sulfate wasting and hyposulfatemia [21,23], warrant investigation of sulfate levels in humans.

It is remarkable that sulfate is not routinely measured in clinical settings, even though it has been recognised as an essential analyte in human physiology for more than six decades [32]. Since the 1950’s, several studies have tested various methods for measuring sulfate in biological fluids, including benzidine or barium precipitation, capillary electrophoresis, electrospray tandem mass spectrometry (ESMS-MS) and ion chromatography [20]. The benzidine precipitation assay has produced unaccountably high levels of sulfate, and this method has since been discontinued due to its carcinogenic potential. An early study comparing barium precipitation and ion chromatography showed these methods produce similar sulfate levels from the same samples [33]. More recently, Cole and Evrovski reported that the ion chromatography method has significantly less variation in serum sulfate measurements (304 ± 5 μmol/L, p=0.006, unweighted mean ± SD of data from six studies) when compared to nine studies that used the barium precipitation method (334 ± 43 μmol/L) [20]. In addition, the ion chromatography method has negligible interference by other anions, when compared to the barium precipitation method that can over-estimate sulfate levels by approximately 3-10% [34]. Based on these findings, we assessed the analytical validity of an ion chromatography method for measuring sulfate in plasma, serum and urine, including sulfate parameters on analytical sensitivity, linearity, recovery, imprecision, specificity, stability of extract, and interference.

Instruments and reagents

The following items were purchased from Thermo Scientific Dionex: Dionex ICS-2000 Ion Chromatography System (ICS-2000); Analytical IonPac AS18 4mm column (Part No. 060549); Guard column IonPac AG18 4mm (060551); Elugen™ hydroxide cartridge (053921); and an Anion standard (057590). Sodium thiosulfate (Na2S2O3.5H2O) was purchased from AJAX (Catalogue No. A517). Aliquots of Urine Control (Bio-Rad Lyphochek Quantitative Urine Control, Cat. No. 377) were stored at -20°C up to 12 months. Plasma control samples were prepared in-house and stored at -80°C up to 12 months. Deionized H₂O (dH₂O, Millipore system) was obtained on the day of analysis for dilution of controls and test specimens.

Blood and urine collection, processing and storage

Urine samples (no preservative added) were collected at approximately the same time as a blood sample in either a lithium heparin tube (for plasma) or a serum separator tube (for serum) from healthy volunteers and patients. The minimum and preferred volumes were 100 μL of urine, plasma or serum, that were stored at room temperature, 4°C or -20°C. The Chairperson of the Mater Health Services Human Research Ethics Committee (EC00332) assessed this research as meeting the conditions for exemption from the requirement of full HREC review and approval in accordance with section 5.1.22 (a) and (b) of the National Statement on Ethical Conduct in Human Research (2007) updated 2015.

Calibrators and procedure

The calibrator sulfate solutions were prepared on the day of analysis by diluting the Anion standard containing 1041.0 μmol/L sulfate: 1/2 (STD1, 500 μL Anion Standard+500 μL dH₂O, 520.5 μmol/L sulfate), 1/10 (STD2, 100 μL Anion Standard+900 μl dH₂O, 104.1 μmol/L sulfate) and 1/100 (STD3, 100 μL STD2+900 μL dH₂O, 10.4 μmol/L sulfate). Calibrator solutions, diluted samples of urine (20 μL urine+980 μL dH₂O) and plasma or serum (40 μL plasma/ serum+360 μL dH₂O), and sodium thiosulfate-spiked anion standard STD1 (final concentration 500 μmol/L sodium thiosulfate) and urine (20 μL urine+980 μL of 1 mmol/L sodium thiosulfate) were added to an auto-sampler vial and assayed by ion chromatography with suppressed conductivity detection using a Dionex ICS2000 with the program parameters shown in Supplemental Table 1. Data management and analyses were performed using Chromeleon^® software version 6.5 SP2.

Specificity

Typical elution profiles of sulfate and its separation from other inorganic anions by ion chromatography are shown in Figure 1. Sulfate was detected as a single peak in the chromatographs with a retention time of 7.00-7.40 minutes. Sulfate peaks in the chromatographs for serum, plasma and urine, correlate to the sulfate peak in the anion standard, and do not elute near other detectable anions, including chloride, nitrate and phosphate.

Figure 1: Analysis of sulfate by ion chromatography. Representative chromatographs of (A) anion standard, (B) plasma, (C) serum and (D) urine showing the separation of sulfate from other inorganic anions.

Analytical sensitivity

The limit of detection (LOD, 9 μmol/L) and quantification (LOQ, 27 μmol/L) were calculated from 3.3 and 10 times, respectively, the SDs of responses: LOD=3.3 σ/S and LOQ=10σ/S, where the slope of the calibration curve (S)=0.022 and the SD of y intercepts of regression lines (σ)=0.061 (Figure 2A). The LOD was low enough for the detection of sulfate, while the calculated LOQ was precise and accurate for the quantitation of sulfate in diluted serum, plasma and urine samples.

Figure 2: Linearity of sulfate detection and stability of sulfate in extracted samples. Individual data points are the mean of triplicate assays (n=15 to 30 samples assayed for each panel A-I). Day 0 corresponds to sulfate measurements on fresh extracts <2 hours of sampling.

Linearity and recovery

Measurement of sulfate standards ranging from 0-1041 μmol/L showed a linear response, with R2=1.000 (Figure 2A). Overall recovery of sulfate ranged from 95-105% for plasma and urine spiked with 50- 500 μmol/L sulfate and 10-100 mmol/L sulfate, respectively (Table 1). Cumulative retention of sulfate on the column between runs was not observed.

Table 1: Recovery of sulphate.

Imprecision

Within run and between run imprecision data are shown in Table 2. The within run imprecision coefficients of variations (CVa) were 0.8% for plasma and 5.0% for urine samples. The between run imprecision CVa were 6.1% and 4.3% for plasma and urine, respectively. To assess the inter-operator variability, two independent operators tested the same samples in triplicate within a 2-hour window (Table 3). The results [mean (SD)] were 213 (2.1) μmol/L for plasma and 3944 (171.8) μmol/L for urine, showing <5% analytical variability. In addition, CVa values were 0.0%, 2.0% and 5.7% for the inter-operator analyses of 520.5 μmol/L, 104.1 μmol/L and 10.4 μmol/L sulfate standards, respectively, indicating the greatest imprecision for samples at the lowest concentration.

Table 2: Imprecision of sulphate measurements.

Table 3: Inter-operator variability.

Stability of extracts

To determine the stability of the serum extracts, initial measurements at <2 hours of blood collection and serum extraction were compared to those from serum stored at room temperature for 1 day and at 4°C for 7 days (Figure 2B and 2C). Serum sulfate levels were unchanged at 1 and 7 days, showing R2=0.966 and 0.985, respectively. Urine stability was assessed after 1 day at room temperature, as well as 7 and 30 days at -20°C (Figure 2D, 2E and 2F). Urinary sulfate levels at all three time points were similar to measurements within 2 hours of urine sampling, showing R2=0.999. Plasma extracts were also stable when stored at 4°C for 7 days, with R2=0.985 (Figure 2G). Measurements of sulfate in plasma and serum were similar on the day of extraction, R2=0.992 and when stored at 4°C for 7 days, with R2=0.983 (Figure 2H and 2I).

Interference

Anion standard and urine samples were spiked with sodium thiosulfate to ascertain possible thiosulfate interference with sulfate detection. Interference (>5% of the peak response) was not observed in either sample at or near the retention times of other anions, including sulfate (Figure 3).

Figure 3: Thiosulfate interference was not observed at or near the retention times of other analytes. Representative chromatographs of *thiosulfate-spiked (A) anion standard and (B) urine showing the separation of sulfate from thiosulfate and other inorganic anions.

In this study, we validated an ion chromatography method for measuring free inorganic sulfate levels in plasma, serum and urine in a clinical setting. Our validation includes specificity, analytical sensitivity, linearity, recovery, imprecision, stability and interference. In addition, our data also show that sulfate levels are similar in plasma and serum extracts stored up to 7 days at 4°C. Collectively, this information demonstrates the clinical utility of the ion chromatography method and validates its use for establishing reference intervals of sulfate levels in healthy individuals. While the requirement for measuring sulfate in humans is largely underappreciated, its physiological significance has been highlighted in animal studies [21,23,26,35,36] that are relevant to renal sulfate wasting and decreased serum sulfate levels in humans with loss-of-function variants in the SLC13A1 sulfate transporter gene [29,30].

We show the ion chromatography method to consistently separate the sulfate peak from other anions and to produce a chromatograph of the seven anion standards that is similar to the manufacturer’s data. Earlier studies also demonstrated that other anions, including bromide, urate, citrate, oxalate, ascorbate, malate, succinate, iodide, chloride, sulfite, thiosulfate, borate and persulfate, do not interfere or co-elute with sulfate [33,34]. Similarly, we show distinct separation of sulfate from thiosulfate, which is relevant to elevated urinary thiosulfate excretion in those individuals with sulfite oxidase deficiency and molybdenum cofactor deficiency [37]. In addition, sulfoesters, including dehydroepiandrosterone sulfate and chondroitin sulfate, yield insignificant and distinct peaks from sulfate [33]. Furthermore, we show no difference in sulfate levels between plasma and serum, indicating no effect of the lithium heparin in plasma samples. Collectively, these findings demonstrate that the ion chromatography method produces a reliable elution profile for sulfate in plasma, serum and urine, which is distinct from other anions at physiological concentrations.

We show linear detection of sulfate within the range of 0-1041 μmol/L, with LOQ of 27 μmol/L. Free inorganic sulfate in human serum and plasma is approximately 300 to 400 μmol/L [20], which is well within the linear concentration range and above the LOQ when diluted 1:10. The required 40 μL of serum and plasma minimize the sampling volume of blood, which is relevant for sulfate measurements in neonates and children. For urine, sulfate concentrations are within the millimolar range, which requires dilution of 1:50 to achieve measurements within the linear range and above the LOQ. These findings demonstrate the relatively small volumes of plasma, serum and urine that are required for sulfate measurements using ion chromatography.

In our stability studies, we compared measurements from fresh extracts with those extracts stored at room temperature, 4°C or -20°C for periods of time that exceed the typical storage time of less than 2 hours at room temperature in a pathology setting. This is relevant to situations where samples may need to be stored or transported prior to analysis. Intracellular sulfate concentrations are low [38] and thereby are unlikely to contribute to extracellular levels. However, there is potential for removal of sulfate from endogenous molecules via sulfatases that could potentially lead to false elevated sulfate measurements. Our data indicate that sulfate levels in plasma, serum and urine after 1 day at room temperature, as well as longer term storage at either 4°C or -20°C, are not different to those levels found in fresh extracts. These findings extend a previous study showing that storage of serum at -70°C for 2, 8 or 60 days had no significant effect on the measured sulfate concentration [39], presumably due to the negligible effect of sulfatase activity in plasma, serum and urine. Together, these findings support the practicality of performing the analysis of plasma, serum and urine sulfate levels at a convenient time post sampling.

Sulfate has important roles in human physiology, and perturbed sulfate homeostasis has adverse health outcomes. Methodologies for measuring sulfate in human biological fluids have been reported in research settings over many years but none of these have been validated for their analytical utility in clinical settings. Despite this importance, and possibly due to the unreliability of previous methods, sulfate determination is rarely undertaken as a clinical investigation. In this study, we validated an ion chromatography method for measuring sulfate in serum, plasma and urine. Our study demonstrates that the ion chromatography method has utility for investigation of sulfate levels in human health.

This work was supported by Mater Pathology, Mater Research and a Mater Foundation Principal Research Fellowship to PAD.

The authors declare that they have no competing interests.