Key Role for the Organic Anion Transporters, OAT1 and OAT3, in the in vivo Handling of Uremic Toxins and Solutes

In vitro data indicates that the kidney proximal tubule (PT) transporters of uremic toxins and solutes (e.g., indoxyl sulfate, p-cresol sulfate, kynurenine, creatinine, urate) include two “drug” transporters of the organic anion transporter (OAT) family: OAT1 (SLC22A6, originally NKT) and OAT3 (SLC22A8). Here, we have examined new and prior metabolomics data from the Oat1KO and Oat3KO, as well as newly obtained metabolomics data from a “chemical double” knockout (Oat3KO plus probenecid). This gives a picture of the in vivo roles of OAT1 and OAT3 in the regulation of the uremic solutes and supports the centrality of these “drug” transporters in independently and synergistically regulating uremic metabolism. We demonstrate a key in vivo role for OAT1 and/or OAT3 in the handling of over 35 uremic toxins and solutes, including those derived from the gut microbiome (e.g., CMPF, phenylsulfate, indole-3-acetic acid). Although it is not clear whether trimethylamine-N-oxide (TMAO) is directly transported, the Oat3KO had elevated plasma levels of TMAO, which is associated with cardiovascular morbidity in chronic kidney disease (CKD). As described in the Remote Sensing and Signaling (RSS) Hypothesis, many of these molecules are involved in interorgan and interorganismal communication, suggesting that uremia is, at least in part, a disorder of RSS.

creatinine and urate [11][12][13] . Since it is well-established that both the Oat1KO and Oat3KO animals display the expected defects in the handling of drugs and toxins 12,[14][15][16][17][18][19][20][21][22] , such alterations in the plasma and/or urine concentrations of endogenous metabolites in the Oat1KO and Oat3KO animals are expected to reflect changes in their handling due to the absence of the transporters. Here, using a more comprehensive metabolomics analysis, levels of over 90 reported uremic toxins and uremic retention solutes were analyzed from the plasma of Oat1KO and Oat3KO animals, as well as from Oat3KO mice treated with probenecid, an inhibitor of OAT-mediated transport ("chemical double knockouts") 23 .

Results
Despite clearly functioning as drug transporters, considerable in vitro and in vivo data indicate that OAT1 and OAT3 also handle many endogenous metabolites, including a number of uremic toxins and/or retention solutes 5,[8][9][10]24 . In this study, levels of uremic toxins and uremic solutes were analyzed from the plasma of the Oat1KO and Oat3KO, as well as from Oat3KOs treated with probenecid ("chemical double knockouts"). Among the more than 600 metabolites examined, ~90 of them are water-soluble or protein-bound uremic toxins/retentions solutes identified by a number of groups (Supplemental Table S1) 1,2,[25][26][27][28][29] . Partial least squares discriminant analysis of these uremic toxins/retention solutes revealed separation of both Oat1KO and Oat3KO groups from their wildtype controls (Supplemental Figure 1).
While adding a few molecules which accumulate in the plasma of the Oat1KO (e.g., cysteine, kynurenate and S-adenosylhomocysteine), the metabolomics analysis of the plasma from the Oat1KO largely agreed with previous studies (Table 1) 12,13 . For example, kynurenine, methionine and orotate, whose concentration was found to be altered in previous metabolomics analyses 12,13 , also accumulated in the plasma of the Oat1KOs studied here (Table 1). On the other hand, analysis of the Oat3KO revealed considerably more alterations than previous attempts 30 , including significant increases in the plasma concentration of numerous uremic toxins/retention solutes, among them trimethylamine N-oxide (TMAO), indoxyl sulfate, p-cresol sulfate and 3-carboxy-4-methyl-5-propyl-2-furanpropanoate (CMPF) ( Table 1). Taken together with in vitro data (Supplemental Table S2), the new in vivo metabolomics data showing the plasma accumulation of a number of uremic toxins provide strong support for the notion that Oat1 and Oat3 independently play an important role in the uptake and handling of a wide variety of water soluble and protein-bound uremic toxins/retention solutes 12,13,30,31 . Moreover, certain compounds (e.g., kynurenine, indolelactate, indoxyl sulfate, methionine, creatinine, p-cresol sulfate and putrescine) appear to be in vivo substrates for both OATs.
However, the most significant increases in plasma concentration of uremic toxins were observed in the Oat3KOs treated with probenecid-a so-called "chemical double knockout" (Tables 2 and 3). Thus far it has not been possible to create an OAT1/OAT3 double knockout mouse in part because of the chromosomal proximity of these two genes which essentially precludes recombination events in a genetic cross between the Oat1KO and Oat3KO. While probenecid does not exclusively inhibit OAT1 and OAT3 in vitro, by treating the Oat3KO with this drug, at least for organic anion transport across the basolateral membrane of the renal PT cell, a "chemical double knockout" was essentially created. Since, in the Oat3KO, the basolateral transport of organic anions by PT cells is dependent upon the remaining OAT (i.e., OAT1), treatment with probenecid should inhibit basolateral organic anion uptake mediated by OAT1 and lead to the accumulation of these molecules in the plasma, potentially allowing for identification of the uremic toxins with which the transporters interact. As above, partial least squares discriminant analysis of uremic toxins/retention solutes revealed separation between the Oat3KO and probenecid-treated-Oat3KO groups based on the metabolite profile of the uremic toxins/retention solutes ( Fig. 1) and analysis of the plasma samples derived from these "chemical double knockout" models revealed profound plasma accumulation of several known and/or suspected uremic toxins/ solutes compared to the untreated Oat3KO (Table 2). For example, kynurenate was only slightly elevated in the plasma of Oat3KO mice, however, following treatment of the Oat3KO with probenecid, the plasma level of this uremic toxin significantly increased almost 5-fold compared to that seen in the Oat3KO (Table 2). This accounted for nearly 80% of the total fold-increase in the plasma seen in the Oat3KO (i.e., total of fold-increase seen in both the Oat3KO and the probenecid-treated Oat3KO). Thus, since, as described above, probenecid is inhibiting OAT1-mediated basolateral uptake from the plasma in the proximal tubule of the Oat3KO, this data strongly suggests that the in vivo PT uptake and handling of kynurenate is largely mediated by OAT1 (Fig. 2). A similar phenomenon was seen with other toxins/retention solutes, including cysteine, S-adenosylhomocysteine, kynurenine, putrescine, spermidine, orotate and hypoxanthine (Tables 1-3). On the other hand, p-cresol sulfate appears to be handled more or less equally by both OAT1 and OAT3 (Fig. 2), with ~50% of the total fold increase in the plasma of the Oat3KO following treatment with probenecid, a finding that is in agreement with in vitro data (Supplemental Table 2). In contrast, while 3-carboxy-4-methyl-5-propyl-2-furanpropanoate (CMPF) and trimethylamine-N-oxide (TMAO) were both elevated in the plasma of Oat3KO mice, treatment of the Oat3KO with probenecid led to relatively small additional accumulations of these metabolites compared to the Oat3KO which were not statistically significant, indicating that the majority of the observed plasma accumulation of these metabolites in the Oat3KO is likely due to the absence of OAT3 (Fig. 2). This suggests that CMPF and TMAO are more likely to interact with OAT3 than OAT1.
While data supports the notion that CMPF is more a substrate of OAT3 rather than OAT1 25,32 , recently published in vitro data seems to suggest that OAT3 does not directly transport TMAO (albeit this assay was performed at a single high concentration) 25 . This would seem to be consistent with TMAO's more cationic character; however, an early physiological study found that the renal excretion of TMAO could be inhibited by probenecid, a finding which was, at that time, attributed to probenecid-mediated blockage of the movement of TMAO into the urine via organic anion transporters localized to apical membrane of PT cells 33 . Nevertheless, the finding here of increased plasma levels of TMAO in the Oat3KOs would appear to indicate that renal uptake of endogenous TMAO from the plasma is, directly or indirectly, mediated via this basolateral organic anion transporter. It is important to note that OAT3 has been shown to have a greater capacity than OAT1 to interact with cationic compounds, possibly including TMAO 34,35 .
We have analyzed the in vivo roles of renal proximal tubule organic anion transporters, OAT1 and OAT3, individually and in combination, in mediating the uptake and excretion of uremic toxins and uremic retention solutes. Considerable new metabolomics data from the Oat1KO, Oat3KO and a "chemical double knockout" (Oat3KO treated with probenecid) when analyzed in the context of limited previous data clearly links the OATs to the handling of numerous uremic toxins/retention solutes. When considered in light of existing in vitro uremic toxin interaction data for these transporters (Supplemental Table 2), it is clear that OAT1 and OAT3 are required for effective renal elimination of many of the "best known" uremic toxins and retention solutes (Table 3).

Discussion
OAT1/SLC22A6 and OAT3/SLC22A8 are not only both expressed in the basolateral membrane of renal PT cells, but they share a high degree of sequence similarity. In fact, they are each other's closest homolog and share a great deal of similar functionality 35,36 . Nevertheless, the data from OAT knockout mice suggests that, while these transporters handle many uremic toxins, certain toxins/solutes prefer OAT1 (e.g., kynurenate, kynurenine, orotate), while others prefer OAT3 (e.g., TMAO, CMPF) with some appearing to be capable of interacting more or less    Table 2 for Km, Ki, IC50 and PMIDs.
equally with both (e.g., p-cresol sulfate) ( Fig. 2; Tables 1-3). Therefore it was of interest to determine if there is an additive or synergistic effect of simultaneously blocking transport mediated by OAT1 and OAT3. Thus, in this study, a "chemical double" knockout was generated by treating the Oat3KO with probenecid. For the majority of the metabolites altered in the Oat3KO, it seems that the additional inhibition of OAT1 by probenecid did not significantly alter the plasma levels of the toxin/solute, suggesting that, while the transporters might share some ligands, for the large part the transporters appear to have a significant level of substrate selectivity. Nevertheless, there were a group of toxins/solutes which only show a significant level of accumulation in the  . The data represents the percentage of the total fold change in the overall plasma concentration for each of these toxins/retention solutes in the Oat3KO (grey boxes) and in the Oat3KO treated with probenecid (black boxes). As described in the text, since in the Oat3KO, treatment with probenecid inhibits OAT1-mediated renal uptake, this allows one to see the contribution of each organic anion transporter to the handling of these metabolites (please also see Table 3).
Although we cannot absolutely exclude minor effects of probenecid on other transporters, such as URAT1 (RST) and MRPs, based on available data it seems likely that OAT1 is likely to be the major transporter affected by treatment of the Oat3KO with probenecid. For example, probenecid exhibits a significant inhibition potency towards OAT1 (and OAT3) with K i s in the low micromolar range (i.e., 1.3-32 µM depending on the substrates and model system used), while the inhibition potency towards other transporters has been found to be much higher 37 . For example, the K i for OAT2 is 766 µM and MRP4 is >1000 µM, while that for the apical uptake transporter, OAT4, is ~55 µM and for the apical extrusion transporter, MRP2, it is ~45 µM 37 .
Furthermore, early pharmacokinetic studies in rodents demonstrated that the effective unbound plasma concentration of probenecid following a single intraperitoneal dose of probenecid at a concentration of 250 mg/kg was roughly equivalent to that of a human receiving a typical oral dose of 0.5-2.0 g of probenecid 38 , which is in the range of 3-50 µM 39 . This concentration is well within the range of that needed to inhibit OAT1 in the Oat3KO, but well below that needed to inhibit OAT2 or MRP4. Although it approaches micromolar concentrations for inhibiting OAT4 and MRP2, these transporters are found on the apical membrane of the proximal tubule cell and are involved in reabsorption from the tubular lumen. Since the endogenous metabolites we are discussing are those which are accumulating in the plasma of the Oat3KO treated with probenecid, the inhibition of these transporters is likely to be less of a concern.
Moreover, the likelihood that the changes we are observing are due to probenecid-mediated inhibition of organic anion transporters other than OAT1, such as URAT1 (RST) seems low since metabolomics studies of the knockout and extensive in vitro studies of URAT1 (RST) in relation to uremic toxins do not support the view that it transports any uremic toxin other than uric acid 4,40 , and the many in vivo studies of OAT1 and OAT3 knockouts are quite consistent with in vivo probenecid data. Moreover, if basolateral OAT1 and OAT3 are blocked by a combination of gene deficiency and probenecid, apical extrusion transporters such as MRPs would not significantly come into play for organic anion uremic toxins in the blood.
Since both of these transporters mediate the pharmacokinetics of a wide variety of drugs, the data raises the possibility that drugs that compete for transport via OAT1 and OAT3 with the uremic toxins have the potential to induce alterations in the levels of the toxins 41 . For example, the accumulation of neurotransmitter metabolites and drugs in the brain in the setting of chronic kidney failure has been suggested to be due to reduced OAT3-mediated brain-to-blood transport due to the increased concentration of uremic toxins such as indoxyl sulfate 42 . Taken together, this could provide a partial explanation for some of the off-target effects of some drugs, including those associated with metabolic syndrome, particularly in the setting of chronic disease and long-term drug treatment 5,8,9 . For example, the furan fatty acid metabolite 3-carboxy-4-methyl-5-propyl-2-furanpropanoic acid (CMPF) is a uremic toxin whose renal elimination is mediated by the OATs. This metabolite has recently been shown to act on pancreatic β-cells leading to reduced insulin biosynthesis and represents a potential link to glucose intolerance 32 . Thus, one can see how competition for OAT-mediated clearance of CMPF and a drug could result in a drug-metabolite interaction at the level of the renal transporter affecting the concentrations, not only of the drug, but of the endogenous metabolite as well, which in the case of CMPF could lead to alterations in glucose tolerance with its effects on the pancreas.
In further support of this notion, it was recently demonstrated that administration of either ketoprofen or diclofenac (non-steroidal anti-inflammatory drugs (NSAIDs) capable of inhibiting OAT1-and OAT3-mediated transport), significantly decreased the renal clearance of indoxyl sulfate thereby increasing systemic exposure to this uremic toxin 43 . It was suggested that this NSAID-mediated increase in the concentration of indoxyl sulfate, might contribute to the progression of indoxyl sulfate-induced cardiovascular disease and, at least in part, explain the pathogenesis of analgesic nephropathy 44 . Taken together, the major point is that partial blockade of OAT1-or OAT3-mediated transport of uremic toxins by competing drugs has the potential to alter the concentrations of certain toxins and lead to the distal cascade effects of the toxin.
Our study also supports the relevance of the remote sensing and signaling hypothesis to the pathophysiology of uremia and CKD 5,8,9,[45][46][47][48] . This hypothesis emphasizes the centrality of multispecific SLC and ABC transporters to inter-organ and inter-organismal small molecule communication throughout the body and thus the importance of the transporters to modulation of metabolism and signaling 5,8,9,30,45,47,49,50 . In this regard, it is worth noting that some of the uremic toxins are derived from metabolites generated by the microbiome in the gut. For example, indoxyl sulfate, ultimately derived from metabolism of tryptophan by the gut bacteria, is absorbed from the intestine into the blood as indole, which is then sulfated in the liver to indoxyl sulfate which is transported into various organs and/or excreted by the kidney 51 . This amounts to interorganismal movement of a small organic anion capable of inducing toxicity or activating signaling pathways via nuclear receptor signaling. For example, indoxyl sulfate, kynurenine and kynurenate signal via the aryl hydrocarbon receptor (AHR) [52][53][54] , which is expressed in most tissues, including those affected in uremia such as the liver, intestine and kidney 55,56 . Kynurenate can also activate the orphan G-couple protein receptor, GPR35 57 , which is largely expressed in tissues affected in uremia, such as the gut and CNS 58,59 . CMPF, on the other hand, has been implicated in a glucose sensing mechanism involving OAT3, which is expressed in the pancreas (among other non-renal tissues), leading to altered metabolism 32 . Of note, metabolic reconstructions of "omics" data from the Oat1KO implicated regulation by OAT1 and OAT3 of a number of biochemical pathways reported to be altered in uremia (e.g., polyamine and purine metabolism) 11 . Although much more work needs to be done in this regard, it is possible that high levels of uremic toxins inhibit the normal metabolic functions of the OATs via uremic toxin-metabolite competition at the level of the transporter; certainly there is evidence for this in the case of uremic toxin-drug competition 25,43 .
The results support the roles of organic anion transporters in the handling of a number of uremic toxins/retention solutes that follow interorgan and interorganismal (gut microbiome-host) communication pathways that may include the activation of key signaling pathways affecting a wide variety of metabolic processes, including those which are known to be affected in uremic syndrome pathophysiology 46 . This general view is also supported by metabolic reconstructions of "omics" data from the Oat1KO and Oat3KO mice 11,30,49 , where it has been demonstrated that Oats regulate biochemical pathways of interorgan and interorganismal communication such as those involving gut microbiome metabolites, energy metabolism, purine metabolism, antioxidants and lipid metabolism and many other pathways altered in CKD. Taken together, this suggests that uremia is in part a disease of disordered remote sensing and signaling 5,8,9,11,[45][46][47] .

Methods
Animals. All experimental protocols were approved by The University of California San Diego Institutional Animal Care and Use Committee (IACUC). The animals were handled in accordance with the Institutional Guidelines on the Use of Live Animals for Research; all experiments involving the use of animals were also conducted in accordance with the Institutional Guidelines on the Use of Live Animals for Research. Adult (n = 3) wildtype, Oat1-and Oat3-deficient male mice received a single, daily intraperitoneal (i.p.) injection of either 200 mg/kg water-soluble probenecid (Invitrogen, Carlsbad, CA) [0.02 mg/µL in PBS (10 µL/g of body weight)] or PBS (sham-treated control) for three days. The animals were housed separately under a 12-h light-dark cycle and were provided access to food (standard diet) and water ad libitum. On day three, two hours before samples were taken, the mice were given a final i.p. injection of either probenecid or PBS. Two hours after this last treatment, blood was collected, and plasma was isolated and stored at −80 °C until being shipped for analysis as described below.