S-Methyl-L-Ergothioneine to L-Ergothioneine Ratio in Urine Is a Marker of Cystine Lithiasis in a Cystinuria Mouse Model

Cystinuria, a rare inherited aminoaciduria condition, is characterized by the hyperexcretion of cystine, ornithine, lysine, and arginine. Its main clinical manifestation is cystine stone formation in the urinary tract, being responsible for 1–2% total and 6–8% pediatric lithiasis. Cystinuria patients suffer from recurrent lithiasic episodes that might end in surgical interventions, progressive renal functional deterioration, and kidney loss. Cystinuria is monitored for the presence of urinary cystine stones by crystalluria, imaging techniques or urinary cystine capacity; all with limited predicting capabilities. We analyzed blood and urine levels of the natural antioxidant L-ergothioneine in a Type B cystinuria mouse model, and urine levels of its metabolic product S-methyl-L-ergothioneine, in both male and female mice at two different ages and with different lithiasic phenotype. Urinary levels of S-methyl-L-ergothioneine showed differences related to age, gender and lithiasic phenotype. Once normalized by L-ergothioneine to account for interindividual differences, the S-methyl-L-ergothioneine to L-ergothioneine urinary ratio discriminated between cystine lithiasic phenotypes. Urine S-methyl-L-ergothioneine to L-ergothioneine ratio could be easily determined in urine and, as being capable of discriminating between cystine lithiasis phenotypes, it could be used as a lithiasis biomarker in cystinuria patient management.


Cystine Calculi Detection by X-ray In Vivo Imaging
At 3 and 6 months of age, isoflurane anesthetized mice were subjected to X-ray imaging for lithiasis detection in an IVIS Lumina XR Series III (Caliper Lifescience-Vertex Techniques, Hopkinton, MA, USA) following the manufacturer's imaging parameters with a calibration curve of cystine stones of known weights. Mice with a cystine stone present at 3-months of age were catalogued as Early Stone Formers (ESF) and those with a cystine lithiasis onset between 3 and 6-months of age were catalogued as Late Stone Formers (LSF). Those without a lithiasic phenotype were considered non-stone formers (NSF).

Sample Collection
Mice were individually housed in metabolic cages for 4 days with the first day as an adaptation period. Mice weight, water and food intake, and excreted urine were monitored daily. 24 h urine samples were collected and kept at −80 • C until further analysis with 50 µL 10% thymol in isopropanol as preservative. On the last day, blood was obtained by intracardiac puncture with EDTA coated syringes and transferred into Microvette EDTA-tubes (Sarstedt, Nümbrecht, Germany) on isoflurane anesthetized animals and centrifuged at 3000 rpm for 10 min and 4 • C in a minifuge after 10 min incubation at room temperature. Plasma was then separated into a new tube and kept on ice. Plasma absorbance at 414 nm was then determined to quantify hemolysis with a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and only those with OD < 0.2 were considered for further analysis. Blood and plasma samples were stored at −80 • C and centrifuged erythrocytes (RBCs) were also collected and stored at −80 • C.

L-Erg and S-Met-L-Erg Determination
L-Erg in blood, plasma and RBCs was measured as described by Sotgia et al. [25,26] while plasma creatinine as described by Zinellu et al. [27].
For L-Erg and S-Met-L-Erg analysis in urine, thawed urine samples removed from debris were 1:10 diluted in Milli-Q water supplemented with 3×-deuterated L-Erg and S-Met-L-Erg (50 ng/mL final concentration each). The L-Erg, S-Met-L-Erg and their respective 3×-deuterated isoforms used as standards where in-house synthesized following a previously described route [28]. Both, diluted samples and standards were then filtered through eXtremeFV PVDF 0.2 µm filter vials (Thomson Instrument Company, Oceansite, CA, USA), and L-Erg and S-Met-L-Erg quantified from 20 µL by LC-MS/MS using an LPG-3400SD LC System (Dionex, Sunnyvale, CA, USA) coupled to an LTQ-XL ESI tandem mass spectrometer (Thermo, Burlington, MA, USA). Samples and standards were kept at 15 • C in the autosampler and were injected into a ZORBAX Eclipse Plus C18 (3.5 µm, 75 × 4.6 mm; Agilent, Sta. Clara, CA, USA) maintained at 35 • C. Solvent A was 0.05% formic acid in ultrapure water, and solvent B was acetonitrile in 0.05% formic acid. Chromatography was carried out at a flow rate of 0.9 mL/min under isocratic conditions (99%A:1%B) for 3 min. Mass spectrometry was carried out under positive ion, electrospray ionization mode, using multiple reaction monitoring (MMR) for quantification of specific target ions. Source voltage was set at 3.0 kV, and capillary temperature was kept at 375 • C. Nitrogen sheath gas flow was 90 (a.u.), auxiliary gas flow 10 (a.u.) and sweep gas flow was 6 (a.u.). Alphagaz 2 helium (Air Liquide, Paris, France) was used as collision gas. Precursor to product ion transitions for each compound were as follows: L-Erg: 230.0 → 186.0; d3-L-Erg: 233.0 → 189.0; S-Met-L-Erg: 244.0 → 200.0; d3-S-Met-L-Erg: 247.0 → 203.0. In all cases, isolation width (m/z) and CID collision energies were 2.0 and 20%, respectively. Creatinine concentrations in 24 h urine samples were determined with Creatinine Assay Kit (Sigma, Sant Louis, MO, USA) as indicated by manufacturer after filtering through 10 kDa MWCO spin filters (Amicon Ultra 0.5 mL, Millipore, Burlington, MA, USA).

Statistical and Prediction Analysis
RStudio and ggplot2, xlsx, dplyr, ggpubr, plotROC, stringr, tidyr, ggsignif, and plyr R packages under R version 4.0.3 were used for data analysis and figure preparation. Mann-Whitney statistical test was used for statistical analysis and statistical significance was considered if p ≤ 0.05.

L-Erg and S-Met-L-Erg in Male Mice Blood and Urine
We first analyzed L-Erg content in blood and urine, S-Met-L-Erg concentration in urine and L-Erg and S-Met-L-Erg excretion of 3 months-old WT and Slc7a9 −/− male mice (shown in Figure 1, 3 months). No statistically significant differences were seen although a lower content of S-Met-L-Erg urine concentration and excretion in Slc7a9 −/− male mice compared to WT mice have been detected at this age.

L-Erg and S-Met-L-Erg in Female Mice Blood and Urine
Gender related L-Erg differences in RBC levels have also been described in female rats [30]. We then investigated as above L-Erg concentration in blood, plasma and RBCs, and L-Erg and S-Met-L-Erg concentration and excretion in urine in 3-and 6-months-old WT and Slc7a9 −/− female mice (shown in Figure 2). At 3 months of age (shown in Figure 2, 3 months), we only detected significant 20% Age related differences in RBCs' L-Erg concentration were reported in humans [29] and in rats [30]. To account for this results in mice, we analyzed as before blood L-Erg levels and urine L-Erg and S-Met-L-Erg levels in male mice at 6 months of age (shown in Figure 1, 6 months). We observed statistically 33% and 41% significant reductions at 6-months of age in S-Met-L-Erg concentration and excretion, respectively, in Slc7a9 −/− male mice compared to WT mice (shown in Figure 1B,C, 6 months).
When both ages were compared, only L-Erg levels in RBCs and urine were significatively 2-fold higher at 6 months of age in both WT and Slc7a9 −/− male mice (shown in Figure 1A, RBC and urine).

L-Erg and S-Met-L-Erg in Female Mice Blood and Urine
Gender related L-Erg differences in RBC levels have also been described in female rats [30]. We then investigated as above L-Erg concentration in blood, plasma and RBCs, and L-Erg and S-Met-L-Erg concentration and excretion in urine in 3-and 6-months-old WT and Slc7a9 −/− female mice (shown in Figure 2). female mice compared to WT (shown in Figure 2B,C, 3 months). At 6 months of age, reduced L-Erg levels were observed in plasma (31%) and RBCs (41%), and S-Met-L-Erg urine levels (27%) of Slc7a9 −/− female mice compared to WT (shown in Figure 2, 6 months).
We also compared both ages looking for age-related differences in female mice. We only observed an age-related significant decrease in L-Erg concentration in plasma in Slc7a9 −/− female mice (shown in Figure 2A, plasma).  At 3 months of age (shown in Figure 2, 3 months), we only detected significant 20% and 31% lower S-Met-L-Erg urine concentration and excretion, respectively, in Slc7a9 −/− female mice compared to WT (shown in Figure 2B,C, 3 months). At 6 months of age, reduced L-Erg levels were observed in plasma (31%) and RBCs (41%), and S-Met-L-Erg urine levels (27%) of Slc7a9 −/− female mice compared to WT (shown in Figure 2, 6 months). We also compared both ages looking for age-related differences in female mice. We only observed an age-related significant decrease in L-Erg concentration in plasma in Slc7a9 −/− female mice (shown in Figure 2A, plasma).

Sex-Related Differences in L-Erg and S-Met-L-Erg in Blood and Urine
We then looked for gender-related differences in WT and Slc7a9 −/− (KO) female mice (shown in Figure 3), as a sex-related difference in L-Erg concentration in blood had been reported in rats [30].

Sex-Related Differences in L-Erg and S-Met-L-Erg in Blood and Urine
We then looked for gender-related differences in WT and Slc7a9 −/− (KO) female mice (shown in Figure 3), as a sex-related difference in L-Erg concentration in blood had been reported in rats [30].
Significative sex-related reduced L-Erg levels could be found in blood, plasma, RBCs, and urine of Slc7a9 −/− female mice (43%, 36%, 48% and 30.8% lower, respectively) when compared to Slc7a9 −/− male mice (shown in Figure 3A). Urine S-Met-L-Erg concentration in female mice showed significative differences in both WT and Slc7a9 −/− mice at both tested ages, being about 40% lower at 3 months of age and about 50% lower at 6 months of age than corresponding male mice (shown in Figure 3B).  Significative sex-related reduced L-Erg levels could be found in blood, plasma, RBCs, and urine of Slc7a9 −/− female mice (43%, 36%, 48% and 30.8% lower, respectively) when compared to Slc7a9 −/− male mice (shown in Figure 3A). Urine S-Met-L-Erg concentration in female mice showed significative differences in both WT and Slc7a9 −/− mice at both tested ages, being about 40% lower at 3 months of age and about 50% lower at 6 months of age than corresponding male mice (shown in Figure 3B).
Except for WT mice at 3-months of age, L-Erg and S-Met-L-Erg urine excretions showed significant sex-related differences in both WT and Slc7a9 −/− mice, being female mice L-Erg and S-Met-L-Erg excretion 41-72% lower than the excretion of male mice (shown in Figure 3C).

L-Erg and S-Met-L-Erg Urine Concentration in Lithiasic Mice
One of the hallmarks of cystinuria, which the Type B mouse model (Slc7a9 −/− ) reproduce [23], is the presence of cystine stones in cystinuria patients from both sexes. According to the age at which the stones were first detected, we classified lithiasic Slc7a9 −/− mice as early stone formers (ESF) if the stone is first detected by X-ray analysis at 3 months of age, late stone formers (LSF) if first detected at 6 months of age, and non-stone formers (NSF) if not detected at this age.
Therefore, we analyzed, as above, L-Erg concentration in blood, plasma, RBCs and urine, S-Met-L-Erg levels in urine and L-Erg and S-Met-L-Erg excretions in stone and non-stone former (NSF) Slc7a9 −/− mice of both genders at 3 and 6 months of age (shown in Figure 4). L-Erg concentration in RBCs was significatively reduced in stone-former male mice at 6 months of age (shown in Figure 4A, RBC). Besides, L-Erg 24 h urine concentration was significatively higher in both ESF female and LSF male mice at 6 months of age and in ESF female mice at 3 months of age (shown in Figure 4A, urine). S-Met-L-Erg was lower in ESF females at 3 months of age and in stone former males at 6 months of age (shown in Figure 4B). L-Erg excretion was significant lower only in ESF males at 6 months of age and S-Met-L-Erg excretion in stone former mice from both sexes at both tested ages, 3 and 6 months of age (shown in Figure 4C).
As S-Met-L-Erg is a subproduct of L-Erg metabolism which levels are affected by those of L-Erg, we investigated if the urine concentration ratio between both compounds showed differences related to the lithiasic phenotype. As shown in Figure 5A, the ratio in stone former mice (SF) was a significatively 2-to 3-fold lower than in NSF Slc7a9 −/− mice from both sexes at any age. This result strongly suggests that the ratio between the 24 h urine concentrations of S-Met-L-Erg and L-Erg could be used to differentiate the lithiasic phenotype in mice. We investigated this possibility by the means of a receiver operator characteristic (ROC) curve for 95 mice (54 lithiasic and 41 non-lithiasic) (shown in Figure 5B). The area under the ROC curve (AUC) of 84.9% (95% CI: 76.98-92.76%) and the negative predictive value of 78.04% and positive predictive value of 83.33% at the Youden's cut-off value of 0.065 suggest the possibility of using this ratio as a lithiasic biomarker. Furthermore, the performance of the urinary ratio of S-Met-L-Erg to L-Erg was not severely affected by age (shown in Figure 5C) or sex (shown in Figure 5D).

OCTN1 as a Cystine Lithiasis Modulator
As L-Erg is transported by OCTN1 (Slc22a4) and as Slc22a4 −/− mice lack L-Erg in the kidneys [14], we crossed the Slc22a4 −/− mice with our cystinuria mice model (Slc7a9 −/− ) looking for differences in the rate of lithiasic mice related to OCTN1 loss (shown in Figure 6 and Table 1). We found a higher number of lithiasic female mice independently of Slc22a4 genotype except at 48 weeks of age in Slc7a9 −/− Slc22a4 +/− (Het, heterozygous for Slc22a4) where no age-related differences could be found. Generally, the rate of lithiasic mice was higher for the double KO (Slc7a9 −/− Slc22a4 −/− ) mice independently of sex, except at 48 weeks of age, in which no differences in the percentage of lithiasic mice could be related to the loss of Scl22a4. This result suggests that OCTN1 (Slc22a4) might modulate cystine lithiasis in mice by changing the cellular availability of the molecules it transports, especially L-Erg, as it is the main molecule transported known today. As S-Met-L-Erg is a subproduct of L-Erg metabolism which levels are affected by those of L-Erg, we investigated if the urine concentration ratio between both compounds showed differences related to the lithiasic phenotype. As shown in Figure 5A, the ratio in stone former mice (SF) was a significatively 2-to 3-fold lower than in NSF Slc7a9 −/− mice from both sexes at any age. This result strongly suggests that the ratio between the 24 h urine concentrations of S-Met-L-Erg and L-Erg could be used to differentiate the lithiasic phenotype in mice. We investigated this possibility by the means of a receiver operator characteristic (ROC) curve for 95 mice (54 lithiasic and 41 non-lithiasic) (shown in Figure  5B). The area under the ROC curve (AUC) of 84.9% (95% CI: 76.98-92.76%) and the negative predictive value of 78.04% and positive predictive value of 83.33% at the Youden's cut-off value of 0.065 suggest the possibility of using this ratio as a lithiasic biomarker. To better understand this result, we analyzed the amount of L-Erg and S-Met-L-Erg in the 24 h urine of these mice at 3 months of age. L-Erg was detectable in 24 h urine to similar levels in Slc7a9 −/− Slc22a4 +/+ (WT for Slc22a4) mice when compared to wild-type mice (Slc7a9 +/+ Slc22a4 +/+ ), but the amount of S-Met-L-Erg was 50% lower (p = 0.0022) (shown in Table 2). Knocking out Slc22a4 in the Slc7a9 −/− mice produced a significative reduction of L-Erg levels in urine (p = 0.0382 for Slc22a4 +/− and p = 0.0138 for Slc22a4 −/− ) and, although we could still detect S-Met-L-Erg, its 24 h urine concentration was below the quantification limits of the method used. As expected, double KO mice (Slc7a9 −/− Slc22a4 −/− ) showed lower concentration of L-Erg than Slc7a9 −/− Slc22a4 +/− mice.
Furthermore, the performance of the urinary ratio of S-Met-L-Erg to L-Erg was not severely affected by age (shown in Figure 5C) or sex (shown in Figure 5D).

OCTN1 as a Cystine Lithiasis Modulator
As L-Erg is transported by OCTN1 (Slc22a4) and as Slc22a4 −/− mice lack L-Erg in the kidneys [14], we crossed the Slc22a4 −/− mice with our cystinuria mice model (Slc7a9 −/− ) looking for differences in the rate of lithiasic mice related to OCTN1 loss (shown in Figure 6 and Table 1). We found a higher number of lithiasic female mice independently of Slc22a4 genotype except at 48 weeks of age in Slc7a9 −/− Slc22a4 +/− (Het, heterozygous for Slc22a4) where no age-related differences could be found. Generally, the rate of lithiasic mice was higher for the double KO (Slc7a9 −/− Slc22a4 −/− ) mice independently of sex, except at 48 weeks of age, in which no differences in the percentage of lithiasic mice could be related to the loss of Scl22a4. This result suggests that OCTN1 (Slc22a4) might modulate cystine lithiasis in mice by changing the cellular availability of the molecules it transports, especially L-Erg, as it is the main molecule transported known today.  Table 1 shows the number of mice per condition and age.   Table 1 shows the number of mice per condition and age. Table 1. Number of mice per condition and age used to assess the percentage of lithiasic mice in Figure 6.    [14], suggests that (i) another transporter might be involved in the absorption of L-Erg from the diet, (ii) OCTN1 is needed for L-Erg transport into cells, and (iii) S-Met-L-Erg is mainly produced intracellularly.

Discussion
L-Erg is particularly enriched in RBCs in comparison with plasma in WT mice; and L-Erg excretion is low, supporting its avid retention by the body [31], in agreement with previously published findings in humans [32][33][34]. But we have seen differences in either urine L-Erg concentration or excretion when comparing cystinuric mice with their WT counterparts.
In this sense, we have observed in 6-month-old cystinuric male mice a 15% reduction in L-Erg excretion without any change in 24 h urine concentration. L-Erg transporter, OCTN1, expression is upregulated in injured tissues to promote L-Erg uptake and help in reducing damage caused by oxidative stress [22]. Based on this, and due to the antioxidant and cytoprotective properties of L-Erg, this lower excretion might be an adaptive response against a lower antioxidant capacity of cystinuric kidneys, but the excretion was even lower for S-Met-L-Erg, a S-methyl derivative of L-Erg. No information on its biological source is available apart from being detectable in human blood and urine [32], and in many mouse tissues except spleen and heart [35]. The fact that its 24h urine concentration is low enough to avoid our quantification method in Slc22a4 −/− mice suggests that it might be produced inside the cells perhaps in an unknown intracellular transmethylation reaction mediated by methyl-S-transferases. Based on our results, we hypothesized a reduced methylation capability of the kidney in cystinuria that could explain the observed reduced urine levels and excretion of S-Met-L-Erg in cystinuric mice.
The other known products of L-Erg metabolism are hercynine and L-ergothioneine sulfonate, both stable oxidative products of L-Erg. We did not quantify the concentration or excretion of these products, but assuming that the production of these compounds is not compatible with the formation of S-Met-L-Erg, we hypothesize that the levels of either hercynine or L-ergothioneine sulfonate or both are higher in cystinuric mice, which again would point out to a hampered oxidation regulation in cystinuric kidneys. This hypothesis is supported by the correlation observed in the levels of hercynine or L-ergothioneine sulfonate and L-Erg in kidneys [35]. The reduce glutathione levels observed in the liver of cystinuric mice might predispose to injuries caused by oxidative stress [36] and can be speculated that ergothioneine, as an antioxidant, might be being used by tissues to compensate a reduced redox metabolism.
Age-related L-Erg changes in RBCs has been observed in humans, where it declines gradually at 19-50 years from the maximum at 18 years of age [29]. We have tested L-Erg in mice at 3 and 6 months of age, which corresponds to middle age humans (20-30 years) [37]. Although a reduction in L-Erg levels were expected from human observations, the levels in RBCs at 6 months of age in comparison with those at 3 months of age were higher in mice, replicating the observations made in rats, where no age-related decline in L-Erg levels in RBCs were detected [30], suggesting specie-specific characteristics.
No gender-related differences in L-Erg plasma concentration has been reported in middle age humans [38], although it had been reported in rats [30], pointing out again to a putative specie-specific characteristic. Gender biased expression in human kidneys has been demonstrated for 67 genes [39], and 114 genes in rats [40] and glutathione metabolic process was among the functional categories of gender biased genes in human kidneys [39]. To explain the observed gender-related differences in L-Erg and S-Met-L-Erg levels in mice, it is tempting to speculate that OCTN1 expression might be gender biased in mice. We did not observed significative changes in L-Erg levels in blood, in contrast to the lower levels present in the blood of mice and patients with CKD [22], except for the increase in cystinuria male mice at 6 months of age.
Some authors have noted an increasing incidence of pediatric nephrolithiasis, at least in the U.S. [41,42], with the consequence of an increasing need of noninvasive methodologies for diagnosis and follow-up. Currently, patient's cystine lithiasis might be controlled by analyzing the urinary crystal volume and number of cystine crystals, or the cystine capacity, or by imaging techniques. Crystalluria analysis, which is only available in selected clinical laboratories, shows an AUC of 66% [16], bellow the AUC of 84% displayed by the S-Met-L-Erg to L-Erg ratio described here.
One of the practical limitations of crystalluria is urine processing time [16]. Such limitation could be avoided quantifying urine parameters such as cystine capacity, supersaturation, and concentration, or as we propose from our results, measuring S-Met-L-Erg and L-Erg urinary levels. The AUCs of cystine capacity, cystine supersaturation and cystine concentration are 69%, 70% and 70%, respectively [17], clearly below the 84% of the S-Met-L-Erg to L-Erg ratio. The good performance of the S-Met-L-Erg to L-Erg ratio indicates that it could be a better biomarker of cystine lithiasis than the currently existing ones.
The abovementioned methods show a limited capability to detect cystine stones in clinical settings and no reliable biomarker of cystine lithiasis can be used by clinicians to manage cystinuria patients [16]. The use of the urinary S-Met-L-Erg to L-Erg ratio in a clinical setting would not be much different of the routine amino acid determination which is nowadays mostly done by mass spectrometry [43], opening the possibility of measuring it from dry spot urine or frozen samples, once the protocol is set up in clinical diagnosis laboratories. A more reliable and easier to determine cystine lithiasis biomarker could allow clinicians to detect cystine stones earlier and prescribe appropriate treatments. An earlier detection would also mean a smaller cystine stone, leading to shorter treatment times and a lower risk of urinary tract obstruction, which could also have an impact on patients' quality of life. Furthermore, a cystine lithiasis biomarker that does not rely on cystine could also be included in clinical assays to monitor the performance of future cystinuria treatments to prevent cystine stone formation.
The results presented here also suggest that OCTN1 (Slc22a4) might be a modulator of cystine lithiasis. The changes we observed in the levels of the antioxidant L-Erg together with the lower reduced and total glutathione levels described in cystinuric mice [36] and the reduction in cystine lithiasis when mice is treated with antioxidants [44,45], support the hypothesis of the involvement of redox mechanisms in cystinuria. Based on our results, it is also tempting to speculate that a lower activity or a defective OCTN1 might be behind those patients showing highly lithiasic phenotypes.
The differences observed in S-Met-L-Erg to L-Erg ratio between lithiasic and nonlithiasic cystinuric mice and its good performance in the cystinuria mouse model support considering urinary S-Met-L-Erg to L-Erg ratio as a good biomarker of cystine lithiasis. Its use in a clinical outcome, once verified, could clearly improve cystinuria patient management and, therefore, patients' quality of life.

Conclusions
L-ergothioneine and S-methyl-L-ergothioneine urine excretions showed significant sex-related differences in both WT and Type B cystinuria (Slc7a9 −/− ) mice.
In a mouse model of Type B cystinuria (Slc7a9 −/− ), the ratio of the urine concentrations of S-methyl-L-ergothioneine to L-ergothioneine discriminated between lithiasic and nonlithiasic mice regardless of sex and age.
The urinary S-methyl-L-ergothioneine to L-ergothioneine ratio performes as a good biomarker of cystine lithiasis in a Type B cystinuria (Slc7a9 −/− ) mouse model.

Patents
The results presented are part of the patent WO2021018774A1, entitled "Ergothioneine, S-methylergothioneine, and uses thereof".