Abstract

Solute carrier family 6 member 19 (SLC6A19) inhibitors are being studied as therapeutic agents for phenylketonuria. In this work, a potent SLC6A19 inhibitor (RA836) elevated rat kidney uremic toxin indoxyl sulfate (IDS) levels by intensity (arbitrary unit) of 13.7 ± 7.7 compared with vehicle 0.3 ± 0.1 (P = 0.01) as determined by tissue mass spectrometry imaging analysis. We hypothesized that increased plasma and kidney levels of IDS could be caused by the simultaneous inhibition of both Slc6a19 and a kidney IDS transporter responsible for excretion of IDS into urine. To test this, we first confirmed the formation of IDS through tryptophan metabolism by feeding rats a Trp-free diet. Inhibiting Slc6a19 with RA836 led to increased IDS in these rats. Next, RA836 and its key metabolites were evaluated in vitro for inhibiting kidney transporters such as organic anion transporter (OAT)1, OAT3, and breast cancer resistance protein (BCRP). RA836 inhibits BCRP with an IC50 of 0.045 μM but shows no significant inhibition of OAT1 or OAT3. Finally, RA836 analogs with either potent or no inhibition of SLC6A19 and/or BCRP were synthesized and administered to rats fed a normal diet. Plasma and kidney samples were collected to quantify IDS using liquid chromatography–mass spectrometry. Neither a SLC6A19 inactive but potent BCRP inhibitor nor a SLC6A19 active but weak BCRP inhibitor raised IDS levels, whereas compounds inhibiting both transporters caused IDS accumulation in rat plasma and kidney, supporting the hypothesis that rat Bcrp contributes to the excretion of IDS. In summary, we identified that inhibiting Slc6a19 increases IDS formation, while simultaneously inhibiting Bcrp results in IDS accumulation in the kidney and plasma.

SIGNIFICANCE STATEMENT This is the first publication to decipher the mechanism for accumulation of indoxyl sulfate (IDS) (a uremic toxin) in rats via inhibition of both Slc6a19 and Bcrp. Specifically, inhibition of Slc6a19 in the gastrointestinal track increases IDS formation, and inhibition of Bcrp in the kidney blocks IDS excretion. Therefore, we should avoid inhibiting both solute carrier family 6 member 19 and breast cancer resistance protein simultaneously in humans to prevent accumulation of IDS, a known risk factor for cardiovascular disease, psychic anxiety, and mortality in chronic kidney disease patients.

Introduction

Solute carrier family 6 member 19 (SLC6A19), sodium-dependent neutral amino acid transporter B0AT1, is expressed in the apical membrane of intestinal and renal epithelial cells and is responsible for the absorption of > 95% of free neutral amino acids such as Phe and Trp in the small intestine as well as the reuptake of these amino acids in the renal proximal tubule cells (Kleta et al., 2004; Bröer et al., 2011). Hartnup disorder, a disease caused by mutations in the SLC6A19 gene (Levy, 2001), is characterized by intestinal malabsorption of neutral amino acids and renal aminoaciduria. Patients may remain clinically asymptomatic if their niacin intake is adequate (Bröer, 2009). Transgenic knockout (KO) mice lacking Slc6a19 (Javed and Bröer, 2019) have increased excretion of Phe and other neutral amino acids in their urine but are otherwise healthy. These results suggest that inhibition of SLC6A19 to reduce Phe absorption in gut and increase Phe excretion in urine may represent a novel approach for the treatment of phenylketonuria (PKU) (Belanger et al., 2018), a disorder where patients have elevated plasma and brain Phe levels resulting in neurotoxicity and permanent central nervous system damage (Yadav et al., 2020).

In 1966, Tada et al. (1966) showed that after oral administration of 80 mg/kg (body weight) of L-Trp, Hartnup patients had strikingly elevated levels of the uremic toxins indoleacetic acid and indican in their urine compared with healthy individuals. No accumulation of these uremic toxins suggests effectively cleared via renal elimination. Excretion of uremic toxins depends on glomerular filtration and tubular secretion via a multitude of transporter proteins expressed in renal proximal tubules. For example, indoxyl sulfate (IDS), a protein-bound uremic toxin, is an endogenous metabolite of Trp normally excreted into urine. Chronic kidney disease (CKD) patients, with dysfunctional urinary excretion pathways, are known to accumulate IDS in plasma. These patients show a marked increase in IDS levels compared with healthy individuals (13 μM to 65 μM or higher in end-stage renal disease patients) (Mutsaers et al., 2011; Duranton et al, 2012). Several studies have hypothesized that high concentrations of uremic toxins perpetrate interactions with drug metabolizing enzymes and transporters, which could increase drug-related adverse events and the risk of cardiovascular disease, psychic anxiety, and mortality in CKD patients (Chao and Chiang, 2015; Prokopienko and Nolin, 2018; Brydges et al., 2021; Lu et al., 2021; Takkavatakarn et al., 2021).

Understanding the molecular mechanisms behind transporter-mediated tubular secretion of uremic toxins is crucial for determining why their levels become elevated. On the basolateral membrane of the proximal tubules, human and rat organic anion transporters (OAT)1 and OAT3 have been shown to play major roles in the renal uptake of IDS (Deguchi et al., 2004, 2005; Nigam and Granados, 2023). Oat1-knockout mice had dramatically higher plasma concentrations of IDS, highlighting the role of OAT1 in the uptake process from the blood to the kidney (Eraly et al., 2006; Jansen et al., 2019). Analysis showed that uremic toxins, including IDS, inhibited OAT1 and OAT3 at clinically relevant concentrations and contribute to the decline in renal drug clearance in patients with CKD (Hsueh et al., 2016). On the apical membrane of the proximal tubule epithelial cell, breast cancer resistance protein (BCRP or ABCG2) (Takada et al., 2018; Nigam and Granados, 2023), multidrug resistance-associated protein 4 (Mutsaers et al., 2011) and other transporters (Rosenthal et al., 2019) were shown to eliminate uremic toxins. Recent publications (Takada et al., 2018; Grangualy et al. 2021) indicated that BCRP mediated ATP-dependent transport of IDS. Abcg2-knockout mice with CKD accumulated IDS in plasma and kidney and had much less urinary excretion of IDS than in wild-type CKD mice. Thus, the hypothesis that BCRP significantly influences CKD progression aligns with the observed correlation between the decrease in BCRP function and age of dialysis onset in humans (Matsuo et al., 2016).

Recent reviews on uremic toxins (Lowenstein and Nigam, 2021; Nigam and Granados, 2023) indicated that IDS can modulate metabolism and signaling as a part of an extensive “remote sensing and signaling” network involving drug transporters and drug metabolizing enzymes.

In our work, we discovered elevation of IDS levels in rat plasma and kidney after administration of a potent SLC6A19 inhibitor, RA836. We hypothesized (Fig. 1) that increased IDS levels were the consequence of inhibition of Slc6a19 in small intestine. This action resulted in an increase in unabsorbed Trp, which is metabolized by the gut microbiome to indole, and further metabolized to IDS in the liver (Banoglu et al., 2001; Banoglu and King, 2002). We further hypothesized that elevation of IDS in kidney and plasma could be potentiated if kidney transporters involved in renal excretion of IDS are also inhibited. We demonstrated both in vitro and in vivo that inhibition of both Slc6a19 and Bcrp contributed to the accumulation of IDS in the plasma and kidney in rats.

Fig. 1.

Elucidation of the hypothesis that inhibition of SLC6A19 and BCRP transporters simultaneously leads to an increase of indoxyl sulfate in kidney.

Materials and MethodsMaterials

Compounds such as RA836 and its metabolites (M1 and M2), RA636, RA235, RA733, and 3-(trifluoromethyl)benzamide were synthesized in-house. The synthetic procedures and compound structure are not described herein.

Indoxyl-4,5,6,7-d4 sulfate potassium salt (D4-IDS) with 97% purity and 13C6, 15N-L-isoleucine with 98% purity were purchased from Sigma Aldrich. 2,5-Dihydroxybenzoic acid (DHB) MALDI-MS matrix was purchased from Sigma Aldrich with a purity >99%. Estrone-3-sulfate sodium salt, Ko143 hydrate, AMP, ATP, IDS, 7-n-butyl-6-(4-hydroxyphenyl) [5H]-pyrrolo[2,3-b]pyrazine (RP107), 3-(trifluoromethyl)benzoic acid, 3-(trifluoromethyl)benzylamine, trifluoroacetic acid, and acetonitrile were of high-performance liquid chromatography (HPLC) grade and obtained from Sigma Aldrich (St. Louis, MO).

In Vitro SLC6A19 Inhibition Assay

To measure the IC50 of compounds in inhibiting human SLC6A19 and rat Slc6a19, an in vitro stable isotope uptake assay was used as described previously (Danthi et al., 2019). Briefly, Madin-Darby canine kidney (MDCK) cells stably expressing human SLC6A19 (hSLC6A19) + human transmembrane protein 27 were incubated at 37°C for 20 minutes with designed concentration (2×) of test compounds in assay buffer (137 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, 10 mM glucose, pH 7.2) following which 2 mM 13C6,15N-L-isoleucine in assay buffer was added and incubated for 20 minutes more. The solution was then aspirated, and cells were washed. Plates were air-dried and subjected to a freeze–thaw cycle to lyse cells. The clarified cell extracts were analyzed using a Thermo Fisher LX-4 chromatograph coupled to an ABsciex API-4000 triple quad mass spectrometer (Sciex, Framingham, MA) to quantitatively measure peak areas of 13C6,15N-L-isoleucine. Similarly, for inhibition of rat SLC6a19, Madin-Darby canine kidney (MDCK) cells stably expressing rSlc6a19 + human transmembrane protein 27 were used.

In Vivo Rat Studies

Sprague–Dawley (SD) male rats were housed in accordance with local standards with free access to commercial chow and tap water. All animal experiments were approved by the local Sanofi institutional animal care and use committee and conducted in line with principles from the Guide for Animal Care and Use of Laboratory Animals (https://www.ncbi.nlm.nih.gov/books/NBK54050/). Male SD rats were administered by oral gavage once daily with RA836 at 200 mg/kg (n = 5/per group) or vehicle (n = 3/per group) for 3, 7, and 14 days. Following a designated number of dosing days, blood samples were collected from all animals via the saphenous vein into K2 EDTA-coated tubes over a specified range of time. The samples were placed on wet ice prior to centrifugation at 1500 × g (+4°C) for 10 minutes. Plasma samples were transferred to labeled microtubes and frozen at approximately –80°C before analysis. Kidney samples were collected from all animals at necropsy and were snap frozen in liquid nitrogen and kept at –80°C before analysis.

Blood samples for clinical chemistry on days –1, 3, 6, 7, 9, 13, and 14 were collected from the saphenous vein into tubes and processed to serum within 1 hour and 30 minutes of blood collection. Each blood sample was centrifuged at approximately 10,000 rpm for 5 minutes. Resulting serum samples were analyzed. Twenty clinical chemistry parameters were evaluated including creatinine, urea nitrogen, and total bilirubin.

In designated studies, male rats (n = 5/per group) were fed either a diet without tryptophan (L-Amino Acid Rodent Diet without Added Tryptophan-A10033Yi) or a control diet with tryptophan (L-Amino Acid Rodent Diet-A10021Bi). The specified diets were administered ad libitum from day 1 to day 15. Beginning on day 10, vehicle or RA836 at 200 mg/kg were administered by oral gavage daily through day 14. On day 15, animals were killed and necropsied for the collection of blood and kidney samples.

Biomarker Evaluation by Tissue MALDI-MS Imaging (tMSI)

Immediately after euthanasia, rat kidneys were dissected and snap frozen in liquid nitrogen and stored at –80°C. For MALDI MS imaging experiments 10-μm thick rat kidney cryosections were cut with a cryostat (CM 1950 cryostat, Leica Biosystems, Nussloch, Germany) and thaw-mounted on glass microscope slides (ground edge, Superfrost) and subsequently desiccated at room temperature (RT) over night until matrix application or stored at –80°C under vacuum until further use.

MALDI matrix application was performed as previously described (Hinsenkamp et al., 2016). Briefly, 2,5-DHB matrix was dissolved at a concentration of 60 mg/ml in acetonitrile/ddH2O/trifluoroacetic acid (50/50/0.2, v/v/v). Matrix deposition onto the tissue was performed by spray coating using a SunCollect sprayer (SunChrom, Friedrichsdorf, Germany), by applying increasing flow rates (10, 15, 20, 25, 25 μl/min) at a velocity speed of 300 mm/min. tMSI data acquisition was performed as elsewhere described (Garikapati et al., 2019). Briefly, after successful matrix deposition (confirmed by applied matrix density), tMSI data were acquired using a high-resolution atmospheric-pressure matrix-assisted laser desorption/ionization ion source (“AP-SMALDI 10,” TransMIT GmbH, Giessen, Germany), coupled to a Fourier transform orbital trapping mass spectrometer (Q Exactive Plus, Thermo Scientific GmbH, Bremen, Germany) and equipped with a nitrogen gas laser (LTB Lasertechnik GmbH, Berlin, Germany) with a wavelength of 337 nm, operating at a repetition rate of 60 Hz.

All MALDI MSI analysis was performed by measuring the high mass resolution mode of the QExactive Plus MS (280,000 at m/z 200) with a lateral resolution (50 × 50 or 100 × 100 μm pixel size) and a step size of 50 or 100 μm. The mass spectrometer was operated in positive ionization mode within the (m/z) range of m/z 80–1000 with a target voltage of +5.0 kV and a heat capillary temperature of 250°C and an S-lens RF level 65. Automatic gain control was disabled, and the injection time was fixed at 500 milliseconds (microscans = 1). A total of 45 laser pulses per pixel were accumulated in the C-trap before being analyzed in the orbitrap analyzer. To ensure high mass accuracy (<1 ppm) an internal lock mass calibration was performed on known DHB ion clusters (m/z 431.03735).

Data Processing and Image Visualization

AP-SMALDI tMSI raw data sets (.raw) were converted to centroid labeled imzML files using the RAW2IMZML software (DOI: 10.1101/2021.02.23.432420). For ion image visualization and statical evaluation, imzML data were loaded and processed using SCiLS Laboratory MVS (version 7.02.10901) (Bruker Daltonics, Bremen, Germany). Normalization of signal intensities was performed using the total ion count normalization procedure and target intensities were exported as arbitrary unit (a.u.). Graph design and statistical analysis were performed using GraphPad Prism Software (Microsoft Windows Version 9.0.0, San Diego, CA).

Metabolite Profiling of RA836Samples Preparation for MetID

In vitro metabolite profiling was performed for RA836 by incubating 5 μM compound with 1 × 106 cells of cryopreserved hepatocytes from human or rat for 2 h at 37°C, 5% CO2 shaken at 300 rpm. The reaction was quenched with 600 μl of ice-cold acetonitrile and then centrifuged at 1800 × g for 20 minutes at 25°C. The supernatant was evaporated using N2 stream at 35°C in an EvapoREX device (EVX-192, Apricot Designs) to an approximate volume of 200 μl. The samples were then analyzed for metabolites using liquid chromatography–mass spectrometry (LC-MS).

In vivo metabolite profiling was performed on plasma, kidney, and urine samples taken after oral administration of RA836 to SD rats at 200 mg/kg (five animals/per group). The plasma and kidney homogenates or urine were pooled at equal volume from individual animals and treated with 3 volumes of acetonitrile, followed by centrifugation at 13,000 rpm for 10 minutes. The resulting supernatant was transferred to a tube and concentrated under a stream of nitrogen at 30°C to obtain concentrated extract (∼200 μl). The samples were centrifuged at 13,000 × g for 10 minutes prior to LC-MS analysis.

LC-MS/UV Method for MetID

All LC/UV/MS analyses were performed using an ACQUITY UPLC system (Waters, Milford, MA), a diode-array detector (PDA UV detector) coupled to a LTQ Orbitrap XL (Thermo Fisher Scientific Inc., San Jose, CA). Liquid chromatographic separation of metabolites was achieved on a Kinetex C18 column (1.7 μm, 100 Å, 2.1 × 100 mm; Phenomenex, Torrance, CA) at 35°C. The mobile phase consisted of A: 10 mM ammonium acetate in water, pH 5.0 and B: acetonitrile. The LC gradient initiated at 5% mobile phase B for 5 minutes and was increased linearly to 70% mobile phase B at 27 minutes and then to 95% mobile phase B at 28 minutes, then held at 95% mobile phase B for another 5 minutes with a flow rate of 0.3 ml/min. The column was re-equilibrated for 8 minutes prior to the next injection. The chromatographic effluent flowed through a PDA for UV spectra acquisition; and was then delivered to a LTQ Orbitrap XL mass spectrometer. Prior to the data acquisition, the Orbitrap was calibrated for mass accuracy on the day of analysis using Pierce ESI negative ion calibrate solution (product no. 88324, Thermo Scientific), followed by cleaning the ESI source with methanol. Orbitrap was set at ESI negative with resolution of 60,000. The resulting UV and MS data files were processed with the Qual browser module of Xcalibur 2.1 (Thermo Fisher Scientific Inc., San Jose, CA).

Quantification of Compounds and Indoxyl Sulfate Levels in Biological Samples

The concentrations of test compounds and IDS in plasma or tissues were determined using an liquid chromatography-tandem mass spectrometry (LC-MS/MS) system consisting of a SCIEX Triple Quad 6500 MS/MS system (AB Sciex LLC, Toronto, Canada) coupled to a Shimadzu Nexera UHPLC using Kinetex C8 column (100 Å, 1.7 μm, ID: 2.1 mm × 30 mm; Phenomenex) at 45°C. The mobile phase was composed of a gradient of solution A (95% acetonitrile with 5% 10 mM ammonium acetate buffer) and solution B (5% acetonitrile with 95% 10 mM ammonium acetate buffer, pH7.0) at the flow rate of 0.50 ml/min. The tissue samples were first homogenized with 25% acetonitrile in water. The plasma and tissue homogenate samples were purified by protein precipitation with four volumes of internal standard solution (30 ng/ml of RP107 and 1.00 μg/ml of indoxyl-4,5,6,7-d4 sulfate in acetonitrile). The clear supernatant was further diluted with eight volumes of 25% acetonitrile in water before injection to LC-MS/MS system. The analytes were monitored at negative multiple reaction monitoring mode. The monitoring parameters (m/z of precursor and product ions and collision energy) were as follows: IDS (212.1 → 80.0, collision energy (CE) = –25); RA836 (450.0 → 393.0, CE = –40); [D4]-IDS (216.1 → 80.0, CE = –25) and RP107 (266.0 → 222.0, CE = –47). Peak analyses were performed using Analyst software version 1.6.2 (AB Sciex LLC).

In Vitro BCRP, OAT1, and OAT3 Transporter Inhibition Studies

Mammalian cell membrane (HEK293) derived inside-out vesicle transporter inhibition assay method was used for assessment of human BCRP and rat Bcrp.

The assays used to measure the potency for inhibiting human BCRP and rat Bcrp were performed using a vesicular transport (hBCRP, rBcrp) assay according to the general Solvo protocol (Solvo Biotechnology, Szeged, Hungary). Briefly, the inside-out membrane vesicles, derived from HEK293 cells expressing the rBcrp or hBCRP transporters, were incubated with designed concentration of compounds in the presence or absence of 4 mM ATP. Incubations on membrane vesicles were carried out at 32°C for 1 minute (rBcrp and hBCRP) in 0.1 M Mg(NO3)2, 10 mM Hepes-Tris, pH 7.4, using 12.5 μg of total protein/well for rBcrp and hBCRP. The tritium labeled probe substrate solution estrone-3-sulfate (E3S) (Biotrend GmbH, Cologne, Deutschland) exist in the mixture in 25 μM total concentration (including labeled and nonlabeled bile salts). The reaction was stopped by the addition of ice-cold wash buffer (100 mM Tris-HCl, 1 M KNO3) and immediate filtration via glass fiber filters mounted to a 96-well plate (filter plate). Liquid scintillator (Microscint-40; Perkin Elmer, Waltham, MA) was added to the dried residual amount of substrate on the filter plate and light emission signal was determined by MicroBeta liquid scintillation counter measurement (Perkin Elmer).

The calculation of ATP-dependent transport and % inhibition for vesicular transport inhibition assays used the equation shown below: in which AccATP represents probe substrate elevation with ATP in cpm, AccAMP represents probe substrate elevation with AMP in cpm, TCPM represents cpm in dosing solution, V represents volume per well (μl), CCsub represents probe substrate concentration (μM), Prot represents total protein per well (mg), and t represents incubation time (min). Results are obtained using the four-parameter logistic model according to Ratkowsky and Reedy (1986). IC50 values are calculated by BIOST@T-SPEED LTS version 2.0 or higher (SAS).

The inhibition potential toward human OAT1 and OAT3 was determined using cryopreserved, transiently transfected HEK293 cells (TransportoCells, Corning). Transfected and mock control cells were seeded on 96-well microplates (Costar #3628, Corning) at a density of 100,000 cells per well in Dulbecco’s modified Eagle’s medium (GIBCO, ThermoFisher) containing 10% fetal calf serum (GIBCO, ThermoFisher) and 1% nonessential amino acids supplementation (GIBCO, GlutaMax-I, ThermoFisher). Media was changed after 4 hours attachment phase and fresh media was added. After 24 hours medium was aspirated and stimulation buffer containing Hanks’ balanced salt solution (HBSS) and 500 μM glutaric acid (ACROS, ThermoFisher) was added for 30 minutes. The stimulation buffer was aspirated and 0.03–30 μM test compound and 3 μM test substrate [14C]para-aminohippuric acid (PAH) (ABCR, Germany) for OAT1, or 2 μM test substrate [3H]-estrone-3-sulfate (E3S) (ARC) for OAT3 including labeled and nonlabeled substrate was added in 50 μl HBSS (Invitrogen) containing 5 mM HEPES (Invitrogen) for 10 minutes incubation time for [14C]PAH at 37°C and 5 minutes for [3H]E3S. The reaction was stopped by adding 150 μL ice cold HBSS and the cells were washed twice with 200 μl ice cold HBSS. Wash solution was aspirated and 100 μl liquid scintillator (Microscint-40, Perkin Elmer) was added to the cells and light emission signal (cpm) was determined by MicroBeta liquid scintillation counter measurement (Perkin Elmer) and compared with a [14C]>PAH, or [3H]E3S calibration curve. Specific transporter activity was determined, normalized to protein content as measured using BCA protein assay (Pierce, ThermoFisher) and expressed as pmol/min × mg protein. The substrate uptake [pmol/well] was calculated from the radioactivity detected for each concentration as mean and divided by total activity. The uptake in both cell lines (hOAT1-HEK293 or hOAT3-HEK293 and mock-HEK293) is normalized by protein amount in each well [mg/well] to calculate pmol/mg protein. With consideration of incubation time (10 minutes for hOAT1 and 5 minutes for hOAT3) uptake into each cell line is given as pmol/min/mg protein for each inhibitor and substrate concentration. Net uptake by hOAT1, or hOAT3, respectively, was calculated by subtraction of background uptake (uptake by hOAT1-HEK293 or hOAT3-HEK293 minus uptake by mock-HEK293).

The IC50 values were determined by plotting the percentage of residual net uptake (in %) versus inhibitor concentration (in μM). Results are obtained using the four-parameter logistic model (Ratkowsky and Reedy, 1986). IC50 values are calculated by BIOST@T-SPEED LTS version 2.0 or higher (SAS).

Plasma Protein Binding Determination

The plasma protein binding was measured using a rapid equilibrium dialysis (RED) device (Brouwer et al., 2000). Briefly, in triplicate, 300 μl of rat plasma spiked with 1 μM final concentration of compound was loaded into RED donor chamber, 500 μl of 0.01 M phosphate buffer (pH 7.4) was loaded into RED receiver chamber. The RED plate was sealed with parafilm and incubated for 4 hours with 5% CO2 at 37°C with gentle shaking at 250 rpm. A total of 15 μl from each chamber was quenched with six volumes solution containing internal standard, then the mixture was centrifuged at 3100 RPM for 10 minutes to pellet precipitated protein. The supernatant was analyzed using an LC-MS/MS system. The % Bound = 100*[1 – (concentration on buffer side)/(concentration on plasma side)]% was calculated. The results need to meet a criterion of having a CV% of less than 20%.

ResultsSlc6a19 Inhibition Resulted in IDS Elevation in Rat Kidney

RA836 inhibited human SLC6A19 and rat Slc6a19 with IC50 of 7 nM and 28 nM, respectively, in an in vitro stable isotope uptake assay. RA836 selectively inhibited SLC6A19 over other related transporters such as hLAT1, hSGLT1, hPEPT1, and hACS1 (IC50 >10 μM) (unpublished data).

SD male rats fed a normal diet were orally administered RA836 daily at 200 mg/kg for 3, 7, or 14 days. The kidney tissues were examined using tMSI technology. Untargeted Biomarker Discovery revealed a complex pattern of signaling molecules and exhaustive analysis of ions (m/z = 100–1000 with mass accuracy ≤1 ppm) found putative peaks whose intensities were more than 2.5-fold higher than in vehicle-treated rats (Supplemental Fig. 1 and Supplemental Table 1). Among them, IDS reflected the highest increase (intensity a.u. 13.7 ± 7.7) observed in kidney of RA836-treated rats compared with kidneys of vehicle-treated (control, intensity a.u. 0.3 ± 0.1) rats on day 14 (Fig. 2A). To quantitatively measure IDS in kidney homogenates, an LC-MS/MS assay was developed using D4-IDS as an internal standard with a lower limit of quantitation of 0.05 μg/g. As shown in Fig. 2B, the concentration of IDS in the kidney of RA836-treated rats was 3.01 ± 1.25 μg/g on day 3 and 4.66 ± 1.71 μg/g on day 14. This was approximately 3- to 4-fold higher (P = 0.03) than kidneys from vehicle-treated rats that had IDS concentrations of 0.458 ± 0.414 μg/g on day 3 and 1.11 ± 0.597 μg/g on day 14.

Fig. 2.

Indoxyl sulfate accumulation in rat kidney in rats treated with RA836 (n = 5) compared with in control rats (n = 3) detected by tMSI (A) and quantitated by LC-MS/MS; *P-value = 0.03 (B).

To evaluate renal function and kidney histopathology, clinical chemistry parameters after SD rats treated with RA836 daily at 200 mg/kg were compared with these rats treated with vehicle. No test article-related changes were noted in blood urea nitrogen or creatinine. Histopathological changes in the kidney were noted and consisted of a higher incidence and severity of minimal to moderate epithelial degeneration and regeneration within the thick ascending tubules and minimal to mild epithelial hypertrophy in the S3 segment of the proximal tubules from day 3; a modest elevation in severity scores was observed on day 14, as compared with day 3.

Trp-Free Diet Resulted in Lower Levels of IDS in Rat Plasma and Kidney

We hypothesized that RA836 could block the intestinal absorption of dietary Trp, and metabolism in the gut of excess Trp to indole following conjugation with sulfate in liver would lead to increased IDS formation. Rats were fed a Trp-free or a normal Trp-containing diet for 9 days to establish Trp levels in plasma. Rats continued on their Trp-free or control diets and were administered vehicle control or RA836 (200 mg/kg) for 5 days starting on day 10. At necropsy, IDS in rat plasma and kidney were analyzed by LC-MS/MS and tMSI as shown in Fig. 3. The results demonstrated that rats fed a normal diet had higher plasma and kidney IDS levels than rats fed a Trp-free diet when dosed with RA836 (P = 0.02) but no statistical difference if dosed with vehicle control.

Fig. 3.

Indoxyl sulfate in rat plasma and in rat kidney (Y-axis as log 2, Mann Whitney test, ***P-value < 0.0002, *P-value < 0.02). Male rats were fed Trp-free (–Trp) (n = 4/per group) or normal diets plus daily (+Trp) (n = 5/per group) PO dosed with either vehicle as a control or RA836 (200 mg/kg) for 5 days, plasma IDS were quantitated by LC-MS/MS (A) and kidney IDS were by tMSI (B).

Characterization of In Vitro and In Vivo Metabolites of RA836

To understand whether any key metabolites of RA836 contribute to the observed IDS accumulation, metabolic profiling of RA836 was performed in vitro in human and rat cryopreserved hepatocytes and in vivo in rat plasma, kidney, and urine. After incubation of RA836 with rat hepatocytes for 2 hours, 26.3% of RA836 remained and M1 and M2 were formed at 31.2% and 16.8%, respectively, as measured using a semiquantitative LC-UV method, which % was calculated as (UV area of metabolite)/(sum of UV area of parent and all identified metabolites)*100%. The major metabolism pathways and metabolites are described as shown in Fig. 4, in which the hydrolysis pathway resulted in the formation of M1 and 3-(trifluoromethyl)benzylamine; the N-dealkylation pathway resulted in the formation of M2, and 3-(trifluoromethyl)benzoic acid (via oxidation of an aldehyde intermediate). The in vivo plasma, kidney, and urine samples from SD rats orally administered RA836 at 200 mg/kg were also used for metabolic profiling. In rat plasma and kidney, RA836 was the major component (>65% by UV) and the hydrolysis metabolite, M1, was the major metabolite identified (14% in plasma and 20% in kidney). N-dealkylation metabolite M2 and three other metabolites (N-demethylation, oxidation, and amide hydrolysis) were observed at 1%–2%. In rat urine collected over 8–24 hours, RA836 was the minor component (<1% by LC-MS, calculated as (MS area of metabolite)/(sum of MS area of parent and all identified metabolites)*100%), M1, M2, and other secondary metabolites were detected at higher amounts than the parent. M1 and M2 were synthesized for characterization of transporter interaction.

Fig. 4.

Major in vitro and in vivo metabolism pathways of RA836.

Inhibition of Renal Transporters OAT1, OAT3, and BCRP

The inhibition of renal OAT1 and OAT3 and BCRP by RA836 and its metabolites was investigated in addition to their ability to inhibit hSLC6A19.

M1, M2, 3-(trifluoromethyl)benzylamine, and 3-(trifluoromethyl)benzoic acid were not inhibitors of hSLC6A19 (IC50 values >10 μM) (Table 1) or rSlc6a19 (Supplemental Table 2).

TABLE 1

Inhibition of transporters by RA836 and its major metabolites

The uptake of [14C]PAH (hOAT1) or [3H]-estrone-3-sulfate (hOAT3) substrate was examined in the presence of RA836 or its metabolites using HEK293 cells transiently expressing hOAT1 and hOAT3. All compounds were tested to maximum concentration based on the limit of solubility. RA836 had no effect on either hOAT1 or hOAT3 at concentrations up to 30 μM; metabolites M1 and M2 had no effect on hOAT1 at concentrations up to 30 μM, but both inhibited hOAT3 with IC50 values of approximately 15 μM (Table 1). Both the 3-(trifluoromethyl)benzylamine and 3-(trifluoromethyl)benzoic acid metabolites inhibited hOAT1 and hOAT3 with IC50 of approximately 10 μM.

The inside-out membrane vesicles derived from HEK293 cells expressing hBCRP transporters were used to determine IC50 for inhibition of 1 μM of [3H]E3S uptake in the presence or absence of 4 mM ATP. IC50 values of compounds were summarized in Table 1. RA836 was found to be a highly potent inhibitor of hBCRP with an IC50 of 0.045 μM. M1 and M2 were potent inhibitors of hBCRP with IC50 values of less than 1 μM. Neither 3-(trifluoromethyl)benzylamine nor 3-(trifluoromethyl)benzoic acid had any effect on hBCRP activity up to 30 μM.

To study the impact of BCRP inhibition on IDS levels, structural analogs with broad IC50 range (0.01–100 μM) in hBCRP were evaluated using rBcrp inside-out membrane vesicles derived from HEK293 cells expressing rBcrp transporters in a similar manner. The results showed consistent inhibition potential between human and rat (Supplemental Table 3) with RA836 inhibiting rBcrp with an IC50 of 0.005 μM.

Characterization of SLC6A19 and BCRP Inhibition on IDS Levels

To test the impact of SLC6A19 and BCRP inhibition on the formation and excretion of IDS, a set of structural analogs of RA836 possessing similar physical chemical properties but various activities against SLC6A19 and BCRP were identified and characterized in vitro. As listed in Table 2, RA733, was found to be a potent inhibitor of both hSLC6A19 and hBCRP. RA636, the closest structural analog of RA836, is not an inhibitor of hSLC6A19 (IC50 >10 μM) but inhibited hBCRP with an IC50 of 0.06 μM. In contrast, RA235 was a potent inhibitor of SLC6A19, but a weak inhibitor of BCRP (IC50 of 15.1 μM). RA362 was a weak inhibitor of both SLC6A19 and BCRP. There were no species differences in SLC6A19 or BCRP inhibition between rat or human with this structural series (Supplemental Tables 2 and 3).

TABLE 2

Characterization of structural analogs of RA836 in in vitro transporter assays and in vivo in rat

In Vivo Rat Pharmacokinetics (PK) and in Vitro Rat Plasma Protein Binding

Based on the exposure of RA836 achieved in the studies with observed IDS accumulation, rat PK studies for four analogs were conducted. All compounds were administered orally to male rats as a single dose of 100 mg/kg, except RA235 which was dosed at 30 mg/kg in solution formulation. Plasma concentrations of compound at 0.167, 0.25, 0.5, 1, 2, 4, 6, 8, and 24 postdose were quantitatively measured using LC-MS/MS methods. The mean plasma compound–time profiles are presented in Fig. 5. Pharmacokinetic parameters were analyzed using Phoenix WinNonlin software (version 8.2). The Cmax,tmax, and t1/2 values are listed in Table 2. The in vitro rat plasma protein binding (1 μM final concentration) was measured for the tested compounds in triplicates with CV%<20% and %bound is shown in Table 2. RA836, RA733, and RA636 were highly protein bound (>99%) and therefore the free fraction was estimated at 99.95% (based on the detection limit). The unbound Cmax were calculated and summarized in Table 2. The estimated unbound Cmax of 0.01–2.2 μM, was in the range of in vitro IC50 of inhibiting SLC6A19 activity, as well as of the in vivo concentration of RA836, enabling us to test the hypothesis.

Fig. 5.

Mean plasma compound concentration–time profiles after single PO dose of the compound to male SD rats: RA836 (100 mg/kg, n = 3); RA733 (100 mg/kg, n = 3); RA636 (100 mg/kg, n = 3); RA235 (30 mg/kg, n = 3); and RA362 (100 mg/kg, n = 3).

In Vivo Quantitation of Plasma IDS and Kidney IDS after Oral Administration of Test Compounds to Rats

Further in vivo rat studies were conducted at doses designed to achieve similar exposures for each compound that would distinguish between the potent (IC50 <0.1 μM) and weak inhibitors (IC50 >10 μM) of rSlc6a19 and/or rBcrp. SD male rats were dosed orally with vehicle, RA733 at 100 mg/kg/day, RA636 at 50 mg/kg/day, RA235 at 300 mg/kg twice a day, or 600 mg/kg/day, or RA362 at 75 mg/kg twice a day or 150 mg/kg/day for 7 days. The compound and IDS concentrations in plasma at selected time points were measured in the same LC-MS/MS run. The observed Cmax and calculated unbound Cmax of all compounds were summarized in Table 3. The expected concentrations for all compounds were achieved. The mean plasma IDS-time profiles for all compounds are presented in Fig. 6 for a direct comparison, although the study with RA235 was conducted separately. Only RA733-treated rats demonstrated increased plasma IDS levels (4700 ± 961 ng/ml), no significant increase in plasma IDS was observed in rats treated with vehicle or other compounds that inhibited only Slc6a19 or only Bcrp. For RA235-treated rats, we further looked at all studies including PK and pharmacodynamics (PD) studies and compared plasma IDS in rats dosed with RA235 versus control, no significant increase in plasma IDS was seen in any studies (unpublished data). To define the effect in plasma IDS due to compound, we used the mean maximal plasma concentrations of IDS (at tmax) compared with its own baseline to obtain the change in plasma IDS, as calculated using (IDSmax – IDSbaseline), which baseline was calculated using all individual animal data at first sampling time point of 0.5 hours. The ratios of [(IDSmax – IDSbaseline)compound/(IDSmax – IDSbaseline)vehicle)] were obtained and results are presented in Table 3 (middle column). RA636, RA235, and RA362, which do not simultaneously inhibit both transporters, showed less than three times fold difference in plasma IDS, whereas RA733 (dual rSlc6a19 and rBcrp inhibitor) had 27 times higher plasma IDS compared with vehicle-treated animals.

TABLE 3

Summary of compound and IDS concentrations in plasma at tmax and IDS in kidney at necropsy (at t24h) after PO administration of compound to male SD rats

Fig. 6.

Mean plasma IDS concentration–time profiles after PO dose of the compound to male SD rats: RA733 (100 mg/kg, n = 5); RA636 (50 mg/kg, n = 5); RA235 (600 mg/kg, n = 5); RA362 (150 mg/kg, n = 5); and vehicle (0 mg/kg, n = 6).

In the same study, kidneys from all treated groups were collected at necropsy (at 24h). LC-MS/MS assays were developed for quantitative determination of test article and IDS in kidney homogenates. The impact of drug treatment on kidney IDS levels was calculated as the ratio of (IDScompound/IDSvehicle) and results are shown in Table 3 (right column). Similar to what was observed in plasma, RA733 (dual rSlc6a19 and rBcrp inhibitor) treatment led to increases in kidney IDS concentration, while RA636, RA235 and RA362 treatment caused minimal increased or even decreased kidney IDS compared with vehicle-treated rats.

Discussion

Inhibition of SLC6A19 represents a potential therapeutic approach for treatment of PKU (Belanger et al., 2018). Hartnup patients who lack SLC6A19 activity have malabsorption of Trp from the diet and have increased urine levels of IDS. That is because the conversion of excess Trp by colonic bacteria in gut to indole, then further metabolized by liver enzymes to IDS, and finally excreted in urine (Levy, 2001; Javed and Bröer, 2019). However, no elevation of IDS in plasma or kidney has been reported in this population. An inhibitor of SLC6A19 is expected to interfere with the absorption of Trp from the diet, thus increasing IDS excretion in the urine. In our research, we identified a potent SLC6A19 inhibitor that caused a significant elevation of IDS in rat plasma and kidney. IDS is a uremic toxin that has been associated with increased cardiovascular disease, psychic anxiety, and mortality in CKD (Brydges et al., 2021; Lu et al., 2021; Takkavatakarn et al., 2021). Recent reviews indicated IDS could be part of an extensive “remote sensing and signaling” network that includes drug transporters and drug metabolizing enzymes. This network facilitates communication between organs and organisms, helping to maintain and restore homeostasis in healthy and disease states. (Lowenstein and Nigam, 2021; Nigam and Granados, 2023). Therefore the mechanism behind the increase in IDS was investigated.

We hypothesized that elevation of IDS in kidney is derived from the breakdown of unabsorbed Trp that is metabolized via enzymatic reactions in gut and liver (Fig. 1). Furthermore, IDS accumulation might arise from an inability to effectively excrete IDS into urine. We investigated the formation of IDS through a Trp-free diet coupled with administration of RA836, an SLC6A19 inhibitor. The IDS levels in plasma and kidney were significantly lower when rats were fed with Trp-free diet than these fed with normal diet (Fig. 3), supporting our hypothesis that IDS may be derived from metabolism of unabsorbed Trp.

In the kidney, IDS is transported across the basolateral membrane of the proximal tubule by OAT 1 and OAT3 (Eraly et al., 2006; Jansen et al., 2019; Nigam and Granados, 2023); transport out of the proximal tubule on the apical membrane is mediated by BCRP (Takada et al., 2018; Nigam and Granados, 2023). We therefore studied whether RA836 or its metabolites might inhibit OAT1, OAT3, or BCRP. RA836 and metabolites M1 and M2 were weak inhibitors of OAT1 and OAT3 but potent inhibitors of BCRP. RA836 had an IC50 of 0.045 μM on BCRP activity (Table 1), which is one of the most potent inhibitors of BCRP that has been reported (Mao and Unadkat, 2015; Toyoda et al., 2019).

BCRP is an ATP-binding cassette efflux transporter. Recent studies suggested that BCRP mediates physiological secretion of urate, so dysfunction increases the risk of hyperuricemia and gout (Woodward et al., 2009; Matsuo et al., 2016). Comparing the differences in IDS production between normal and dysfunctional kidney conditions or between the wild-type and Abcg2-knockout mice further support that Bcrp acts as a transporter for the excretion of IDS in the kidney (Takada et al., 2018). Similarly, in a rat model of CKD decreases in rBcrp gene expression correlated with disease severity (Lu and Klaassen, 2008). To our knowledge, there is no report on the elucidation of IDS production and excretion using small molecule inhibitors.

RA636, a weak inhibitor of SLC6A19, was administered to rats for comparison with RA836, a potent SLC6A19 inhibitor. After RA636 treatment, the plasma IDS levels over time were similar to those in vehicle control-treated rats (Fig. 6). IDS levels were also measured in kidney and showed no elevation. These results demonstrated the role of Slc6a19 in the formation of IDS by blocking Trp absorption in gut, in the absence of exogenous Trp, IDS levels remained similar to vehicle-treated rats. RA235, which is a potent inhibitor of SLC6A19 (IC50 of 0.007 μM) but weak inhibitor of BCRP (IC50 of 15.1 μM), did not show an increase of IDS in rat plasma or kidney (Fig. 6 and Table 3) because the formed IDS was effectively eliminated through urine. However, RA733, a potent inhibitor of both SLC6A19 (IC50 of 0.008 μM) and BCRP (IC50 of 0.028 μM), significantly elevated rat plasma IDS by 27 times and increased kidney IDS by 2 times, which confirmed the initial observation with RA836, a dual inhibitor for SLC6A19 and BCRP. Therefore, the results support our hypothesis that the increase in IDS arises from simultaneous inhibition of both SLC6A19 and BCRP (Fig. 1).

Our studies demonstrate for the first time that small molecule inhibitors of Slc6a19 increase the formation of IDS, and inhibition of both Slc6a19 and Bcrp simultaneously increase plasma and kidney IDS in rats.

Acknowledgments

The authors are grateful to Mark Czekaj, Bradford Hirth, and Kristen Terranova for compound synthesis and Yan He for preparing formulation to support in vivo experiments; to Oanh Smicker and Lauren Bernotsky for conducting in vitro SLC6A19 assay; to James Murray for in vivo study protocol; to Matthias Schiell for technical LC support. The authors are grateful to Judit Hajagos-Toth and Marko Andric from Solvo for conducting rat Bcrp inhibition assays. We are grateful to Adam Belanger and Yew Nelson for involvement of some discussions. The authors are grateful to Joshua Dekeyser and John Darbyshire for their critical reviewing and their advice. This work was supported by Sanofi and received no external funding. This work was not previously presented in any meeting.

Data Availability

The authors declare all data supporting the findings are available in this paper and its Supplemental Material.

Authorship Contributions

Participated in research design: Wang, Munteanu, Marker, Luo, Kane, Poulton, Sedic, Jayyosi, Riedel, Fretland.

Conducted experiments: Wang, Muntean