Introduction

Gut microbiota plays an important role in regulating the production of a plethora of microbial metabolites with key functions in the host’s metabolic regulation1. p-Cresyl sulfate (pCS), the most well-studied and damaging protein-bound uremic toxin2,3, is one of the aforementioned metabolites, which is synthesized through a multistep process involving both gut microbial and host factors4,5, as will be explained below.

Dietary aromatic amino acids tyrosine and phenylalanine are converted into p-cresol (pC)6,7,8, predominantly in the distal part of the colon through bacterial fermentation9,10,11. Within the human gut microbiota, specific bacterial families, such as Fusobacteriaceae, Enterobacteriaceae, Clostridiaceae, and Coriobacteriaceae, are recognized as strong pC producers. pC undergoes sulfation to yield pCS by the sulfotransferase 1A1 (SULT1A1), highly expressed in enterocytes of the colonic mucosa12, prior to release into the portal vein1,13. Unmetabolized pC can also be transformed into pCS in the liver14,15 by SULT1A1, which is highly abundant in hepatocytes as well12 (Fig. 1). Although pC is mainly converted into its sulfated metabolite, a small fraction is glucuronated into p-cresyl glucuronide (pCG)16,17by host epithelial UDP-glucuronosyltransferases4. Consequently, changes in diet composition, as well as gut microbiota, may interfere with pCS production14,18,19. Eventually, pCS is cleared by the kidney, ending up in the urine, making it a urinary marker of renal disease progression6,20,21 (Fig. 1).

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Schematic representation of p-cresyl sulfate (pCS) metabolism. Tyrosine (Tyr) and phenylalanine (Phe) are dietary aromatic amino acids that are transformed into p-cresol (pC) in the distal colon through bacterial fermentation. Once formed, pC is predominantly converted into pCS via sulfation by sulfotransferase 1A1 (SULT1A1), which is abundantly expressed in colonic enterocytes, before being released into the portal circulation. Alternatively, pC that escapes intestinal metabolism can be sulfated in the liver, where SULT1A1 is also highly expressed. Finally, pCS is eliminated by the kidney and excreted in urine. Created in https://BioRender.com. Based on Blachier and Andriamihaja (2022)14.

Regarding biological effects, accumulation of pCS under renal impairment conditions22 causes toxicity in several tissues, primarily affecting the renal system, as well as the liver. Mainly, pro-apoptotic and proinflammatory effects have been described23,24,25,26,27,28, along with an induction of oxidative stress29. Besides, toxic effects in the cardiovascular system have also been reported20,30,31,32,33,34,35,36,37.

Furthermore, pCS is the predominant alkylphenol in ruminant milk38,39. Recently, Potts and Peterson have described pCS as a small molecule whose presence in bovine milk may alter its somatosensory attributes40. This fact, coupled with the potential presence of the enzyme arylsulfatase in milk, which transforms pCS into pC41, associated with off-flavors described as barnyard and cowshed in milk42,43, may cause an impairment of the somatosensorial properties of milk, which is directly linked to milk quality, rendering it unacceptable to consumers44,45.

Knowledge of the factors influencing bioavailability, tissue distribution, and secretion into milk of pCS is of particular relevance, considering the toxic effects associated with this molecule and the decline in milk quality related to the presence of this metabolite in that biological fluid.

ABCG2, also known as breast cancer resistance protein (BCRP), is a membrane transporter expressed in the apical membrane of cells of several tissues, including the jejunum, liver, and kidney46; biological barriers, such as the blood-brain, blood-testis and blood-placental barriers47,48; and the alveolar epithelial cells of the lactating mammary gland49. It acts as an efflux pump, extruding drugs as well as natural and endogenous compounds from the cells, thereby affecting their absorption, distribution, and elimination50,51,52. Consequently, ABCG2 influences the pharmacokinetics and systemic exposure, reducing plasma levels of its substrates47. Notably, being upregulated during lactation53, it is the only ABC transporter involved in the secretion of its substrates into milk49.

Previous studies using membrane vesicles expressing the human ABCG2 variant suggested that pCS is an in vitro substrate of this transporter54. Notwithstanding, vectorial transport and in vivo interactions remain unknown. Accordingly, this study aimed to assess the in vitro interaction of pCS with different species variants of ABCG2, as well as to evaluate the role of this transporter in the pharmacokinetics, tissue distribution, and milk secretion of pCS, in order to correlate in vitro and in vivo outcomes.

Results

In vitro transport of p-cresyl sulfate and p-cresol

The polarized cell line Madin-Darby canine kidney (MDCK-II) was used to determine the role of ABCG2 in the in vitro transport of pCS. Parental cell line and its subclones transduced with the murine Abcg2 (mAbcg2) and human, ovine, and bovine ABCG2 (hABCG2, oABCG2 and bABCG2, respectively) were grown to form confluent monolayers, which were used to evaluate the vectorial transport of pCS (Fig. 2). In addition, relative efflux ratios were calculated (Table 1). A similar transport pattern was observed in the parental MDCK-II cell line, with equal apical to basolateral (AB) and basolateral to apical (BA) vectorial translocation (Fig. 2a), which is reflected in the relative efflux transport ratio (BA/AB) at 4 h, 1.04 ± 0.10 (Table 1). In contrast, polarized MDCK-II cells overexpressing ABCG2 exhibited a markedly enhanced BA transport and a significantly reduced AB transport (Fig. 2c, e, g, i) when compared with parental cells. This resulted in significantly higher relative efflux transport ratios at 4 h in comparison with the parental cell line (3.83 ± 0.31 for murine Abcg2, p < 0.001; 2.31 ± 0.28 for human ABCG2, p = 0.003; 15.47 ± 4.79 for ovine ABCG2, p = 0.009; and 2.21 ± 0.43 for bovine ABCG2, p = 0.012) (Table 1).

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Transepithelial transport assay of p-cresyl sulfate (10 µM) in the presence or absence of Ko143 (1 µM), the ABCG2 specific inhibitor, in parental MDCK-II cells (a and b, respectively) and its subclones transduced with murine (mAbcg2) (c and d, respectively), human (hABCG2) (e and f, respectively), ovine (oABCG2) (g and h, respectively) and bovine (bABCG2) (i and j, respectively) variants of the transporter. Initially, medium of both compartments was replaced with fresh culture medium with 10 µM p-cresyl sulfate, containing or not the inhibitor. Aliquots were collected at 1, 2, 3 and 4 h on the opposite side where the potential substrate had been added. All samples were stored at − 20 °C until being analyzed by ultra-performance liquid chromatography. p-Cresyl sulfate detected in the opposite compartment is expressed as the percentage of the total compound added at the beginning of the experiment. (n ≥ 4). Results are shown as mean ± S.D.

Table 1 Relative efflux transport (BA/AB) ratio, apically directed translocation percentage (%BA transport) divided by basolaterally directed translocation percentage (%AB transport), at 4 h for p-cresyl sulfate (10 µM) in parental MDCK-II cells and its subclones transduced with murine (mAbcg2), human ABCG2 (hABCG2), ovine (oABCG2) and bovine (bABCG2) variants, either in the absence (-Ko143) and the presence of the ABCG2 specific inhibitor Ko143 (1 µM) (n ≥ 4).
Full size table

The specificity of ABCG2-mediated transport was checked using Ko143, the specific inhibitor of ABCG255. In all subclones, ABCG2-mediated transport was completely inhibited (Fig. 2d, f, h, j), and similar relative efflux transport ratios between parental and transduced cells were obtained (Table 1). These findings indicate that pCS is an effective in vitro substrate of ABCG2.

In the case of pC, no difference was observed in vectorial translocation between AB and BA in parental and mAbcg2 transduced cells, with an efflux ratio at 4 h of (0.96 ± 0.05 for MDCK-II cells vs. 0.99 ± 0.16 for murine Abcg2 transduced cells, n = 4–5, p = 0.740), indicating that pC is not an Abcg2 substrate.

Plasma pharmacokinetics assays in Wild-Type and Abcg2−/− male mice

Plasma pCS concentrations were measured in wild-type and Abcg2−/− male mice and expressed as a function of time following administration of pC, the parental molecule, for the purpose of assessing whether in vitro ABCG2-mediated transport of pCS is mirrored in vivo. After oral dosing of pC, the maximum plasma concentration (Cmax) of pCS was detected at the first sampling time point (30 min) in both groups of mice (Fig. 3), being in Abcg2−/− mice more than 2.5-fold higher than in the wild-type counterparts (2.69 ± 0.61 µg/mL vs. 1.06 ± 0.23 µg/mL; p < 0.001). Statistically significant differences were also observed at 90 min (0.48 ± 0.19 µg/mL in wild-type mice vs. 1.07 ± 0.30 µg/mL in Abcg2−/− mice; p = 0.011) and 120 min post-administration (0.47 ± 0.08 µg/mL in wild-type mice vs. 0.68 ± 0.17 µg/mL in Abcg2−/− mice; p = 0.047). pCS plasma levels obtained at the 240 min time point were below the limit of quantification (LOQ), so they were not included in Fig. 3.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Plasma concentrations of p-cresyl sulfate (pCS) after oral administration of p-cresol (10 mg/kg) in wild-type and Abcg2−/− male mice (n = 4–6). Plasma samples were collected at 30, 45, 60, 90, 120, 150, 180, and 240 min. pCS concentrations were quantified by ultra-performance liquid chromatography analysis. Results are presented as individual data and mean ± S.D. pCS plasma levels at 240 min were below the limit of quantification. (*) p ≤ 0.05: significant differences between both groups of mice.

The estimated area under the plasma concentration-time curve (AUC) corroborated significant differences in plasma concentration-time profiles between both groups of mice. The AUC parameter from 0 min to 240 min post-administration for Abcg2−/− mice was almost 1.6-fold higher than in the wild-type counterparts (3.04 ± 0.09 µg·h/mL vs. 1.91 ± 0.15 µg·h/mL; p = 0.022).

Our findings further corroborate the fact that Abcg2 modulates the plasma pharmacokinetic profile of pCS.

Tissue distribution assays in Wild-Type and Abcg2−/− male mice

pCS concentration was analyzed in a range of tissues, including liver, kidney, small intestine, small intestine content, spleen, brain, testis, and heart, 2 h after administration of a single oral dose of 10 mg/kg pC (Fig. 4).

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Tissue concentration (µg/g tissue) of p-cresyl sulfate in wild-type and Abcg2−/− male mice, measured 2 h after oral administration of the parental molecule, p-cresol at a dose of 10 mg/kg body weight (n = 5). Data are shown as individual values and mean ± S.D. (*) p ≤ 0.05: significant differences between both groups of mice.

pCS was detected in all tissues analyzed. When Abcg2 was lacking, tissue accumulation of pCS in almost all organs studied was significantly higher in comparison with wild-type mice. Notably, the Abcg2−/− group of mice showed a 2.8-fold higher accumulation in liver (4.84 ± 1.73 µg/g in Abcg2−/− vs. 1.69 ± 0.56 µg/g in wild-type, p = 0.025), a 1.5-fold higher concentration in kidney (16.39 ± 4.27 µg/g in Abcg2−/− vs. 10.60 ± 1.43 µg/g in wild-type, p = 0.021), almost 2.3-fold higher levels in small intestine (15.23 ± 6.90 µg/g in Abcg2−/− vs. 6.63 ± 2.76 µg/g in wild-type, p = 0.042), a 2.6-fold higher accumulation in spleen (14.07 ± 4.40 µg/g in Abcg2−/− vs. 5.38 ± 4.21 µg/g in wild-type, p = 0.009) and an almost 2-fold higher concentration in testis (13.49 ± 1.43 µg/g in Abcg2−/− vs. 7.02 ± 0.40 µg/g in wild-type, p = 0.030).

Additionally, the small intestine content retrieved from wild-type mice showed a 2.2-fold higher accumulation in comparison with Abcg2−/− mice (5.26 ± 2.41 µg/g in wild-type vs. 2.41 ± 0.79 µg/g in Abcg2−/−, p = 0.037). These results indicate that Abcg2 mediates pCS excretion into the intestinal lumen, thereby contributing to its elimination.

The observed differences in tissue distribution substantiate that Abcg2 influences pCS biodistribution.

Milk secretion assays in Wild-Type and Abcg2−/− lactating female mice

To assess whether Abcg2 contributes to pCS secretion into milk, the parental molecule was orally given at 10 mg/kg to lactating wild-type and Abcg2−/− female mice. Blood and milk samples were collected 2 h after administration (Fig. 5).

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Plasma (a), milk concentration (b), and milk-to-plasma ratio (c) of p-cresyl sulfate in wild-type and Abcg2−/− female lactating mice following oral administration of the precursor molecule, p-cresol, at a dose of 10 mg/kg (n = 9–11). Plasma and milk samples were collected 2 h after oral administration and concentrations were determined by ultra-performance liquid chromatography. Data are shown as individual values and mean ± S.D. (*) p ≤ 0.05: significant differences between both groups of mice.

pCS plasma concentrations (Fig. 5a) were similar in the two groups of mice (0.73 ± 0.40 µg/mL in Abcg2−/− vs. 0.46 ± 0.27 µg/mL in wild-type, p = 0.053). By contrast, milk from the wild-type mice (Fig. 5b) showed a 3.2-fold higher accumulation compared to their Abcg2−/− counterparts (1.47 ± 0.73 µg/mL in wild-type vs. 0.46 ± 0.21 µg/mL in Abcg2−/−, p = 0.003). Furthermore, the milk-to-plasma ratio of pCS (Fig. 5c) in wild-type mice was almost 6-fold higher than in Abcg2−/− mice (4.31 ± 3.29 vs. 0.74 ± 0.40; p = 0.012).

Based on the aforementioned data, we determine that Abcg2 is actively involved in the secretion of pCS into milk, providing strong evidence of its in vivo interaction with the transporter.

Discussion

Alkylphenols represent key flavor-active compounds in dairy products, existing largely in conjugated forms within ruminant milk38. Of them, pC and its conjugates are the principal alkylphenols detected in ruminant milk, with the sulfated metabolite, pCS, as the predominant form39. Given its role in the somatosensory attributes of milk products40, and its biological effects derived from its accumulation in the bloodstream and tissues as a consequence of being a uremic retention solute56,57,58, it is of the utmost importance to study the pharmacokinetics and biodistribution of pCS and whether it is influenced by the presence of the efflux transporter ABCG2.

Accordingly, although previous studies have already suggested the active transport of pCS via transporters such as P-glycoprotein59 and hABCG254, this study shows for the first time the in vitro interaction of pCS with mAbcg2, oABCG2, and bABCG2 and confirms the in vitro interaction with hABCG2. Likewise, it provides first evidence of the impact of murine Abcg2 on plasma concentrations, systemic exposure, and milk secretion of pCS.

Regarding in vitro assays, in vitro transepithelial transport assays demonstrated that pCS is efficiently transported by all variants of ABCG2 (Fig. 2; Table 1), as well as other gut microbiota-derived compounds such as the ellagic acid-derived metabolite, urolithin A60, and the lignan-derived metabolites enterodiol61 and enterolactone62. These results differ from those observed for the parent compound (pC), which has not been identified as an in vitro substrate of ABCG2. Interactions with ABCG2 are strongly influenced by the physicochemical properties of molecules, with hydrophobicity playing a key role. In this case, sulfation of pC to form pCS increases not only its hydrophilicity, but also its topological surface area63,64. This physicochemical behavior has also been reported for other compounds, such as albendazole65,66 and thiabendazole67, in which only the metabolites, rather than the parent compounds, function as in vitro substrates of ABCG2.

Differences observed between murine and human ABCG2-transduced subclones (Fig. 2; Table 1) were consistent with those reported for other gut-derived metabolites, such as enterodiol61 and urolithin A60, and for other endogenous ABCG2 well-known substrates, such as lumichrome68 and riboflavin, its precursor69, and 6-sulfatoxymelatonin, the main phase II metabolite of melatonin70. It should also be noted that, similar to human ABCG2-transduced subclones, cells transduced with bABCG2 exhibit lower transport ratios at 4 h than murine ABCG2-transduced subclones. In contrast, cells transduced with oABCG2 display markedly higher transport ratios compared to those transduced with other transporter variants (Table 1). These disparities were in line with those previously reported for the gut-derived metabolite, urolithin A60; for the hydroxylated melatonin metabolite, 6‐hydroxymelatonin70; and for some drugs67,71,72. These divergences may reflect variations in the expression and in the affinity or selectivity of a species-specific transporter for the compound. However, we cannot rule out possible differences in transduction efficiency73,74.

Results obtained for the human variant of the transporter are consistent with the ones previously reported, in which in vitro studies were carried out in membrane vesicles transduced with human ABCG254. Vesicular membrane transport models have been used since the 1950s to study the transport of substances across cell membranes. Nevertheless, they are structurally simpler than the polarized MDCK-II cells used in the current study75, which more closely reflect the physiological transport of potential substrates of ABCG276, further reinforcing that pCS is an in vitro substrate of human ABCG2.

Regarding our positive outcomes with cells overexpressing the ruminant variants of the transporter, they support experiments conducted by Kilic et al. (2005) that confirmed the appearance of pCS in ovine and bovine milk without addressing the underlying excretion mechanism at that moment39. Studying the in vitro transport mediated by the ovine and bovine variants of ABCG2 allowed us to analyze its activity more exhaustively, paving the way for further research into the presence of pCS in milk mediated by ABCG2 and the consequences of its consumption.

Regarding in vivo assays, pCS concentrations were determined in plasma and several tissues of male wild-type and Abcg2−/− mice after the oral administration of pC, the precursor metabolite, at a dose of 10 mg/kg. Although pCS exhibits strong plasma protein binding14,22,77, the plasma protein precipitation undertaken in the sample preparation enables measurement of its total plasma concentration.

After the sulfation of pC by SULT1A1, which is highly expressed in hepatocytes and enterocytes12, as was previously noted, pCS was rapidly absorbed in both groups of mice. However, plasma concentrations remained higher in Abcg2−/− mice compared to the wild-type mice at 30 min, 90 min, and 120 min time points, resulting in increased systemic exposure in Abcg2−/− mice (Fig. 3), as evidenced by a higher AUC0-240 min in that group of mice. Similarly, in the case of the antibacterial agent nitrofurantoin76 and the anti-inflammatory drugs tolfenamic acid78 and meloxicam79, following oral administration, plasma levels of these compounds were higher in Abcg2−/− mice compared to their wild-type counterparts. This supports the idea that Abcg2 contributes to pCS plasma bioavailability.

Concerning tissue distribution, pCS was demonstrated to be extensively distributed across multiple tissues (Fig. 4). With regards to the small intestine, likewise for other Abcg2 substrates such as melatonin70and the carcinogenic heterocyclic amines 2-amino-1-methyl-6- phenylimidazo[4,5-b]pyridine (PhIP)80,81and 2-amino-3-methylimidazo[4,5-f]quinoline (IQ)82, higher concentrations of pCS were found in the intestine of the Abcg2−/− mice in comparison with their wild-type counterparts. In contrast, concentrations in the small intestine content retrieved from wild-type mice were higher than from Abcg2−/− mice. This indicates that Abcg2 may reduce pCS oral bioavailability by altering its intestinal absorption, as reflected by decreased concentrations in the enterocytes of the small intestine and elevated concentrations in the intestinal lumen of the wild-type mice.

In the kidney, the Abcg2−/− mice presented a 1.5-fold higher pCS concentration than the wild-type mice (Fig. 4). This suggests reduced renal elimination of this compound and its subsequent accumulation in the kidneys of knockout animals, supporting the fact that Abcg2 participates in the renal elimination of pCS, its main route of excretion4,14,15. The transporter Abcg2 also plays a key role in the urinary elimination of other sulfated metabolites83, such as trans-resveratrol sulfate84, enterolactone sulfate85, 6-sulfatoxymelatonin70, and indoxyl sulfate86. The fact that pCS is utilized as a urinary marker of renal disease progression6,20,21, together with being a human in vitro substrate of ABCG2, may be of significance for human clinical applications, as well as in the case of 6‐sulfatoxymelatonin70. As was stated before, an accumulation of pCS has been associated with renal injury. Pro-apoptotic effects have been described for this molecule23,24,25, along with the increase in the expression of cytokines and proinflammatory genes in renal tubular cells, the activation of the renin-angiotensin-aldosterone system and the TGF-β1 pathway, and the induction of epithelial-mesenchymal transitions, leading to renal fibrosis26,27,28. Besides, high levels of pCS lead to a decrease in Klotho expression, contributing in this way to the senescence of renal cells87. Alterations in the ABCG2 expression or functionality may condition renal distribution of its substrates, including pCS, which could have an impact on its effects48,88, due to its role as an urinary marker, resulting in potential misinterpretations of the renal damage.

Additionally, a higher accumulation of pCS was observed in the liver and spleen of the Abcg2−/− mice when compared to the wild-type mice (Fig. 4). An impairment of the normal function of the liver29,89 and the immune response90,91 has been shown in the presence of clinically significant levels of pCS. In the liver, pCS induces oxidative stress, glutathione depletion, cellular necrosis, bile acid transport disorders, and apoptosis29; whereas it suppresses Th1-type cellular immune response91, and although it induces macrophage activation, it hampers antigen processing, resulting in a compromised adaptive immune response90.

Finally, no significant differences in pCS heart accumulation were found between Abcg2−/− and wild-type mice (Fig. 4). Similar pCS concentrations may be due to the inherently low expression of Abcg2 in cardiac tissue with respect to pharmacologically relevant tissues such as the liver, kidney, or the gastrointestinal tract, where Abcg2 is more highly expressed92.

Taking these findings together, Abcg2, which is expressed on the apical membrane of cells of tissues and biological barriers and mediates the egress of its substrates from the cell46,47,48, affects the accumulation of pCS in the liver, kidney, spleen, and testis after its oral administration. Consequently, Abcg2−/− mice exhibit a more elevated drug exposure compared with the wild-type group, potentially leading to increased toxicity. Alterations in Abcg2 activity may result in changes in these toxicological effects due to pCS.

In light of these results, Abcg2 has a meaningful impact on the pharmacokinetics and tissue distribution of pCS as the activity of the transporter may be reduced by many molecules, such as natural and dietary compounds comprising soy isoflavones93,94 and flaxseed95, the primary source of dietary lignans96, both polyphenols97, along with the concomitant administration of multiple drugs commonly used in the treatment of human diseases98,99,100,101. As well, genetic single-nucleotide polymorphisms (SNPs) could also give rise to a dysfunction of the transporter. This is the case of Q141K, which presents a significant incidence within the Asian population102.

With respect to milk, the involvement of Abcg2 in the active secretion of pCS was also assessed after the oral administration of pC at a dose of 10 mg/kg to lactating wild-type and Abcg2−/− lactating mice.

Unlike in male mice, plasma levels of pCS in the female wild-type group were comparable to their Abcg2−/− counterparts, which may be due to the higher interindividual variability compared to male mice or to the more elevated Abcg2 expression and activity in the liver from male mice in comparison to female mice, as reported in previous studies103. Albeit the lack of significant differences in plasma concentrations between wild-type and Abcg2−/− female mice, pCS concentrations in milk (Fig. 5b) were 3.2-fold higher in the wild-type group in comparison to the Abcg2−/− mice, as well as the milk-to-plasma ratio (Fig. 5c), being in the wild-type group of lactating mice almost 6-fold higher than in the Abcg2−/− lactating mice. These divergences in milk concentrations of pCS and milk-to-plasma ratio have also been observed for other gut microbial metabolites that are Abcg2 substrates, such as enterodiol61 and enterolactone61,62. Moreover, the administration of other natural substrates likewise resulted in elevated concentrations of them in the milk of wild-type relative to Abcg2−/− mice, such as riboflavin69 and its byproduct, lumichrome68, melatonin and its metabolites70, biotin69, and bile acids104. Taken together, these findings indicate that Abcg2 plays a key role in the active secretion of pCS into milk, thereby influencing its milk concentration.

Any factor that could modify the expression or the activity of the ABCG2 transporter could have an impact on the concentration of ABCG2 substrates secreted into milk49, as is the case for pCS, reported in this work. This metabolite is known to modify somatosensory flavor attributes in milk40 and, in the presence of an arylsulfatase, could be transformed into pC41, responsible for unpleasant barnyard42,43 and cow-shed aroma in milk41. Thereby, alterations in the ABCG2 transporter could vary pCS milk concentrations, which may consequently influence milk quality.

The identification of ABCG2 as a key determinant in regulating levels of pCS in plasma, tissue, and milk, together with the impact of gene regulation (ABCG2) and environmental factors, including dietary modifications105 and gut microbiota alterations106,107, on pCS production and accumulation, underscores their potential clinical significance.

Based on the above, we can conclude that the ABCG2 transporter interacts with the gut-derived metabolite pCS. This work shows for the first time that ABCG2 is clearly involved in the active in vitro transport of pCS by murine, human, ovine, and bovine variants. Moreover, in vivo experiments allowed us to demonstrate that Abcg2 affects pharmacokinetics, biodistribution, and secretion into milk of pCS. Thus, changes in ABCG2 expression or function may influence plasma and milk concentrations, along with tissue distribution of pCS, potentially affecting its biological effects.

Methods

Chemicals

pCS, pC, and albendazole-2-amino sulfone were purchased from LGC Standards (Teddington, Middlesex, UK). Lucifer Yellow and Ko143 were acquired from ThermoFisher (Waltham, MA, USA) and Tocris (Bristol, UK). The buffer 4-(2‐hydroxyethyl) − 1‐piperazineethanesulphonic acid (HEPES), utilized in in vitro assays, anthranilic acid, and the charcoal activated used to prepare the milk matrix blanks were purchased from Sigma Aldrich (St. Louis, MO, USA). For in vivo assays, isoflurane (Isovet®) was obtained from B. Braun VetCare (Barcelona, Spain), oxytocin (Falcipart) from SYVA (León, Spain) and heparin (ROVI®) from Laboratorios Farmacéuticos ROVI, S.A. (Madrid, Spain).

Cell cultures

The polarized cell line Madin-Darby canine kidney (MDCK-II) and its murine Abcg2 and human ABCG2 transduced subclones, previously generated81,108 and provided by Dr. A. H. Schinkel from the Netherlands Cancer Institute (Amsterdam, The Netherlands), were used for in vitro studies. Transduced subclones with the ovine and bovine variants of ABCG2 were previously generated by our research group109,110. Cell culture conditions have been previously described109. Concisely, cells were cultured at 37 °C in an atmosphere with 5% CO2 in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with GlutaMAX™ (Life Technologies Limited, Paisley, UK), penicillin (50 units/mL) and streptomycin (50 µg/mL) (Life Technologies Corporation, Grand Island, NY, USA) and 10% (v/v) fetal bovine serum (MP Biomedicals, Solon, OH, USA). Cells were trypsinized when a subconfluent state was reached for subculturing.

Transepithelial transport assays

Transepithelial transport assays were carried out as previously described76, with slight modifications. Cells were seeded at 1.0 × 106 cells/well as a monolayer on microporous membrane filters (3.0 μm pore size, 24 mm diameter; Transwell® 3414; Costar, Corning, NY). Cells were grown for three days, and culture medium was replaced daily.

At the beginning and the end of the assay, monolayer tightness was checked using Millicell®ERS 3.0 Digital Voltohmmeter (Merck Millipore, Burlington, MA, USA). Likewise, monolayer confluence was tested at the end of the experiment via Lucifer Yellow permeability assay111. Transport proficiency was constantly checked by testing a typical ABCG2 substrate (nitrofurantoin 10 µM)76.

Two hours before the start of the assay, medium at both the apical and basolateral side of the monolayer was replaced with 2 mL of prewarmed Hanks’ balanced salt solution (Sigma-Aldrich) supplemented with HEPES (25 mM), the transport medium, either with or without the inhibitor Ko143 (1 µM)68,76, to verify the specificity of ABCG2 on the potential transport. The experiment began (t = 0 h) by replacing the medium of the apical or basal compartments with fresh transport medium containing 10 µM pCS, either with or without inhibitor, or 10 µM pC. Cells were incubated at 37 °C in 5% CO2 and 100 µL aliquots were collected at 1, 2 and 3 h on the opposite side from where pCS or pC had been added. This volume was replaced with fresh transport medium. Finally, 600 µL aliquots were taken at 4 h time from both compartments of each well. All samples were stored at − 20 °C until being analyzed by ultra-performance liquid chromatography (UPLC), as described below.

The appearance of the compounds studied in the opposite compartment is presented as the fraction of total compound added at the beginning of the experiment and expressed as percentage. The relative efflux transport ratio was calculated as the basal to apical directed transport percentage divided by the apical to basal directed transport percentage at 4 h.

Animals

Animals were housed and handled in accordance with institutional and ARRIVE guidelines and European legislation (Directive 2010/63/EU). Experimental procedures were conducted in compliance with relevant guidelines and regulations and were approved by the Animal Care and Use Committee of the University of León and the Junta de Castilla y León (ULE_009_2023). Abcg2−/− and wild-type mice were used, all of > 99% FVB/N genetic background, generated and kindly supplied by Dr. A. H. Schinkel (The Netherlands Cancer Institute)112. All animals, aged from 9 to 14 weeks and weighing 20–36 g, were kept in a temperature-controlled environment with a 12:12 light/dark cycle and received a standard diet and water ad libitum.

Plasma and tissue distribution

For in vivo assays, male mice were lightly anesthetized with isoflurane and administered an oral dose of pC, the parental molecule, via oral gavage. A dose of 10 mg/kg was selected, as it represents the minimum required to detect the sulfated metabolite in biological samples using the chromatographic method described in the Ultra-performance liquid chromatographic analysis section. For these administrations, pC dissolved in saline solution was dosed at 300 µL per 30 g of body weight.

Blood samples were collected at different time points (30, 45, 60, 90, 120, 150, 180 and 240 min). Each mouse contributed two samples under isoflurane anesthesia (retro-orbital and cardiac puncture at different time points). Organs were harvested at the 120-minute time point. All animals were euthanized by cervical dislocation at the end of the experimental procedure. Heparinized blood samples were centrifuged immediately at 3000 g for 15 min68. Plasma and organs were stored at − 20 °C until UPLC analysis. Four to six animals were used for each time point. The AUC was estimated as a pooled parameter using the linear trapezoidal rule, since only two samples were collected from each individual animal.

Milk secretion experiments

Milk secretion experiments were performed using female mice mated with males. Females were separated from males upon detection of a vaginal plug or after five days of cohabitation. Pregnant females were housed individually, and the day of pup birth was recorded. Experiments were conducted between days 10 and 14 postpartum. Four hours prior to the start of the experiment, pups were separated from their mothers. pC at a dose of 10 mg/kg and dissolved in saline solution, was orally administered to lactating wild-type and Abcg2−/− female mice, at 300 µL per 30 g of body weight. Milk secretion was stimulated by subcutaneous injection of oxytocin (200 µL of a 1 IU/mL solution) into lactating mice 10 min before sample collection. Subsequently, 2 h after pC administration, blood and milk samples were collected under isoflurane anesthesia by retro-orbital plexus puncture and gentle nipple pinching using capillaries, respectively. At the end of the experiment, animals were euthanized by cervical dislocation. Heparinized blood samples were centrifuged immediately at 3000 g for 15 min to collect plasma. Milk and plasma were also kept at − 20 °C until UPLC analysis.

Nine to eleven animals were used for each group.

Sample Preparation

Tissue samples were homogenized with a solution of potassium phosphate monobasic at a concentration of 0.3 M with an adjusted pH 3.2; 1 mL of solution per 0.1 g of organ was used. A volume of 10 µL of the appropriate internal standard solution was added to each 100 µL of milk, plasma, or tissue homogenate. Anthranilic acid (10 µg/mL) served as the internal standard for plasma and tissue samples, while albendazole-2-aminosulfone (5 µg/mL) was used for milk samples. In addition, 600 µL of acetonitrile was added to precipitate proteins. After 10 min of being horizontally vortexed, samples were centrifuged at 14,000 g for 15 min at 4 °C. Supernatants were evaporated to dryness at 40 °C under a stream of nitrogen. Samples were reconstituted in 100 µL of cold methanol and injected into the UPLC system. However, samples for in vitro assays were directly injected into the UPLC system.

Ultra-performance liquid chromatographic analysis

Sample analysis was performed on a Waters ACQUITY UPLC® H-class system coupled with a UV photodiode array detector. Chromatographic separation was undertaken on an Acquity UPLC® BEH C18 column (1.7 μm particle size, 2.1 × 50 mm, 130 Å, Waters Corporation®, Milford, MA, USA) maintained at 40 °C. The binary mobile phase consisted of 0.1% formic acid aqueous solution (solvent A) and acetonitrile with 0.1% formic acid (solvent B) as the organic phase. Chromatography was performed using a flow rate of 0.7mL/min with the following linear gradient: solvent B at 5% (0–3.75 min), increased to 80% (3.75–3.80 min), increased to 95% (3.80–5.40 min) decreased to 20% (5.40–5.70 min) and held at 20% (5.70–13 min). For pC, the gradient started at 15% of solvent B, while the remaining conditions were identical. The UV absorbance for pCS was measured at 218 nm and for pC at 220 nm, and samples were maintained at 4 °C throughout the analysis.

For proper quantification of pCS levels in milk samples, a milk analyte-free matrix was prepared as previously described with minor modifications113. A portion of 5 g of milk powder was resuspended in 50 mL of distilled water by vortexing until complete homogenization. Following this, 100 mg of activated charcoal was added into 1 mL of milk solution. The mixture was horizontally vortexed for 2 h at room temperature and centrifuged at 4000 g for 20 min at 4 °C. The supernatant was transferred to clean eppendorf tubes and centrifuged again at 14,000 g for 20 min at 4 °C. The supernatant was transferred again to clean eppendorf tubes and kept at − 20 °C until use.

Standard samples of pCS were prepared in the appropriate analyte-free matrix to generate a defined concentration range between 0.0195 and 10 µg/mL for culture, 0.156–20 µg/mL for plasma and tissue samples, and 0.156–10 µg/mL for milk samples. Coefficients of correlation for culture and milk samples were above 0.98, whereas in plasma samples they ranged between 0.95 and 0.99, and in tissue samples between 0.91 and 0.99. Similarly, standard solutions of pC were prepared to obtain a concentration range of 0.0195–10 µg/mL for culture samples, showing correlation coefficients above 0.99.

The LOQ and limit of detection (LOD) were calculated as described by Taverniers et al.114. In the case of pCS, LOQ was 0.009 µg/mL and LOD 0.004 µg/mL for cell culture samples; LOQ was 0.110 µg/mL and LOD 0.057 µg/mL for plasma samples; LOQ was 0.138 µg/mL and LOD 0.070 µg/mL for milk samples, and for tissues, LOQ was 0.074–0.296 µg/mL and LOD 0.035–0.141 µg/mL. Regarding pC, LOQ was 0.005 µg/mL and LOD 0.004 µg/mL for culture samples.

Statistical analysis

Statistical analyses were conducted using SPSS Statistics software (v. 29.0; IBM, Armonk, NY, USA). Data normality was assessed with the Shapiro-Wilk test. Comparisons between groups were performed using a two-tailed unpaired Student’s T-test for normally distributed data or the Mann-Whitney U test for non-normally distributed data. Differences were considered statistically significant at p ≤ 0.05.