Organic anion transporting polypeptides: Pharmacology, toxicology, structure, and transport mechanisms

Abstract

Organic anion transporting polypeptides (OATPs) are membrane proteins that mediate the uptake of a wide range of substrates across the plasma membrane of various cells and tissues. They are classified into 6 subfamilies, OATP1 through OATP6. Humans contain 12 OATPs encoded by 11 solute carrier of organic anion transporting polypeptide (SLCO) genes: OATP1A2, OATP1B1, OATP1B3, the splice variant OATP1B3–1B7, OATP1C1, OATP2A1, OATP2B1, OATP3A1, OATP4A1, OATP4C1, OATP5A1, and OATP6A1. Most of these proteins are expressed in epithelial cells, where they mediate the uptake of structurally unrelated organic anions, cations, and even neutral compounds into the cytoplasm. The best-characterized members are OATP1B1 and OATP1B3, which have an important role in drug metabolism by mediating drug uptake into the liver and are involved in drug–drug interactions. In this review, we aimed to (1) provide a historical perspective on the identification of OATPs and their nomenclature and discuss their phylogenic relationships and molecular characteristics; (2) review the current knowledge of the broad substrate specificity and their role in drug disposition and drug-drug interactions, with a special emphasis on human hepatic OATPs; (3) summarize the different experimental systems that are used to study the function of OATPs and discuss their advantages and disadvantages; (4) review the available experimental 3-dimensional structures and examine how they can help elucidate the transport mechanisms of OATPs; and (5) finally, summarize the current knowledge of the regulation of OATP expression, discuss clinically important single-nucleotide polymorphisms, and outline challenges of physiologically based pharmacokinetic modeling and in vitro to in vivo extrapolation.

Significance Statement

Organic anion transporting polypeptides (OATPs) are a family of 12 uptake transporters in the solute carrier superfamily. Several members, particularly the liver-expressed OATP1B1 and OATP1B3, are important drug transporters. They mediate the uptake of several endobiotics and xenobiotics, including statins and numerous other drugs, into hepatocytes, and their inhibition by other drugs or reduced expression due to single-nucleotide polymorphisms can lead to adverse drug effects. Their recently solved 3-dimensional structures should help to elucidate their transport mechanisms and broad substrate specificities.

I. Introduction

Organic anions containing a carbon skeleton are a large group of endogenous (eg, intermediates of carbohydrate metabolism containing carboxylic acid, bile salts, bilirubin and its conjugates, fatty acids, and prostaglandins) and exogenous [eg, anionic drugs, contrast agents, indocyanine green, bromosulfophthalein (BSP), and toxins] compounds, which carry a net negative charge at physiologic pH. Bile salts are a major constituent of bile, which is produced by the liver.

The Greek myth of Prometheus illustrates that people have been fascinated by and interested in the liver since ancient times. The liver was important for soothsaying in many ancient cultures, as illustrated by an uncovered 4000-year-old Babylonian clay model of sheep liver. In modern scientific history, the study of the liver likely started at the turn of the 18th century, when it became known that the amount of bile recovered in feces was much lower than the amount produced by the liver and when the term “motus circularis bili” (enterohepatic circulation) was coined (Beuers and Boyer, 1994). Schiff reported in 1870 that absorption of bile in the intestine was necessary for adequate bile formation by the liver (Schiff, 1870) and in 1878, Tappeiner communicated that the ileum was the main site in the gut for the absorption of bile salts (Tappeiner, 1878). Hence, studying endogenous organic anions in the context of the liver goes back to the early 19th century.

Using the exogenous organic anion phenoltetrachlorophthalein, liver function was studied in the early 20th century (Rowntree et al, 1913). Experiments with a series of phenolphthalein derivatives lead to the conclusion that the liver and the kidneys are important organs for the clearance of exogenous compounds and that these organs may display selectivity toward exogenous compounds (Abel and Rowntree, 1909). This compound-dependent organ selectivity demonstrated the presence of specific clearance systems in different organs. Taken together, the abovementioned examples illustrate that biomedical research aimed at elucidating the mechanism(s) of how the liver handles organic anions started about 200 years ago.

Over time, studies were restrained to the functional identification of putative transport systems using organs, isolated cells, or membrane vesicles. In liver transport research, these studies typically involved anionic and cationic organic compounds (Petzinger, 1994). Identification and sequence determination of individual transport proteins made very slow progress until the method of expression cloning of transporters with Xenopus laevis oocytes was introduced (Hediger et al, 1987). This methodology was quickly adopted in transporter research and led to the molecular identification of numerous transport systems at an impressive pace (Romero et al, 1998). The sequence of the first organic anion transporter, namely rat organic anion transporting polypeptide (OATP) 1 (now rOATP1A1), was published in 1994 (Jacquemin et al, 1994). After the initial characterization of OATPs as multispecific drug transporters (Bossuyt et al, 1996a), ongoing research finally led to a general acceptance of the concept of multispecificity in the field of drug metabolism and pharmacokinetics (Fenner et al, 2012).

The aim of this review was to summarize the current state of research regarding the expression, function, and structure of OATPs. We will present a brief historical overview of the identification of OATPs, their phylogenic relationships and nomenclature, and their molecular characteristics, including tissue distribution. This will be followed by a discussion of their substrate specificity and how the organ-selective expression of some OATPs can have clinical implications for imaging. We will mention potential drug-drug interactions (DDIs) and methods to detect them using, for example, biomarkers that are selective for OATP1B1 and OATP1B3. We will review different experimental systems used to study the function of OATPs and their advantages and disadvantages. We will then discuss the recently solved 3-dimensional structures, compare and contrast them to results obtained by structure prediction algorithms, and examine how the available structures can help in elucidating the transport mechanisms. An overview of the regulation of OATP expression and clinically important single-nucleotide polymorphisms (SNPs) will be summarized, and challenges of physiologically based pharmacokinetic (PBPK) modeling and in vitro to in vivo extrapolation (IVIVE) will be discussed.

II. Identification and cloning of OATPs

Functional studies of hepatocellular organic anion transport demonstrated the involvement of protein-mediated processes, whereby the number of individual transport systems involved in this process remained elusive. In addition, different driving forces were identified. Many organic anions have an amphiphilic chemical nature (Berk et al, 1987; Tiribelli et al, 1990) and are tightly bound to albumin. This may have contributed to the difficulty of clearly identifying the molecular basis of the involved transport system(s) and led to the postulation of numerous transport systems with different driving forces (Berk et al, 1987; Tiribelli et al, 1990). Based on the demonstration of a high affinity chloride-dependent BSP transport mechanism (Potter et al, 1987), an expression cloning strategy with total rat liver mRNA in X laevis oocytes (Jacquemin et al, 1991) led to the cloning of the first OATP (Jacquemin et al, 1994). This protein was later renamed to rOATP1A1 (Hagenbuch and Meier, 2003). Expression of mRNA and sib selection of clones was performed with a buffer system without sodium, but with the presence of 5% (w/v) bovine serum albumin (BSA; molar ratio BSA/BSP 3.7) and the difference between the presence and absence of 100 mM chloride. Characterization of the cloned transport protein revealed that the transport system was not chloride dependent, operated in the absence of BSA and the K_m value increased with increasing BSA concentration in the uptake medium. While BSP uptake kinetics followed Michaelis-Menten type kinetics in the absence of BSA (K_m 1.5 μM), the addition of BSA yielded a sigmoidal dependence on BSP concentration (Jacquemin et al, 1994). This observation may be best explained by the presence of chloride-binding sites on albumin (Halle and Lindman, 1978) and that the addition of chloride displaces drugs bound to albumin (Wilting et al, 1980a, 1980b). The K_m for BSP in the absence of BSA was comparable with the reported values for isolated rat hepatocytes in the range between 6.1 and 16 μM (Schwenk et al, 1976; van Bezooijen et al, 1976; Laperche et al, 1981; Stremmel and Berk, 1986; Persico et al, 1988; Ziegler et al, 1994; Sorrentino and Bartoli, 1996). In addition to BSP uptake, rOATP1A1 mediated the transport of cholate and taurocholate, suggesting a broad substrate specificity. An antisense approach to block the expression of the cloned rOATP1A1 when total rat liver mRNA was injected in X laevis oocytes yielded strong evidence for additional non–bile acid–transporting organic anion uptake transporters in rat liver (Hagenbuch et al, 1996). The second OATP to be cloned was rOATP2A1, which is a prostaglandin transporter (Kanai et al, 1995). The first identified human OATP was OATP1A2, which was isolated by hybridization screening of a human liver cDNA library with a cDNA probe of rOATP1A1 (Kullak-Ublick et al, 1995). Interestingly, the highest mRNA level of this OATP was found in the brain. Subsequently, a large number of OATPs was identified from various species using different methods such as homology screening of libraries or in silico methodologies.

III. SLCO family members

After cloning the first few OATPs from different species by different research groups, some of these OATPs had identical names, for example, OATP2, because they were the second OATP cloned in that species. However, it became clear that the original rat OATP2 (Noe et al, 1997), the first mouse OATP2 identified, and human OATP2 (Hsiang et al, 1999) were not orthologs. Therefore, an amino acid sequence–based nomenclature was proposed in 2003 (Hagenbuch and Meier, 2003) and finalized in 2004 (Hagenbuch and Meier, 2004). Based on this nomenclature, the OATPs were classified into families and subfamilies. Figure 1 shows a phylogenetic tree of 173 OATP protein sequences available from primates, rodents, and several domestic animals. These OATPs are classified into 6 families, named OATP1, OATP2, OATP3, OATP4, OATP5, and OATP6. Families OATP1, 2, 4, and 6 have subfamilies, for example, OATP1A, OATP1B, and OATP1C. Individual OATPs are then numbered, for example, OATP1A1, OATP1A2, OATP1A3 (protein symbols are in capitals), and so on, with the corresponding gene symbols Slco1a1 (following the mouse genome nomenclature rodent genes start with a capital letter followed by lower case letters and are given in italics), SLCO1A2 (the human gene symbols are all in capital letters), Slco1a3, Slco1a4, and so on (see also Hagenbuch and Stieger, 2013). To distinguish OATP isoforms from different species, for example, human OATP1C1 versus rat OATP1C1, we recommend either spelling it out as “rat OATP1C1” or abbreviating it as rOATP1C1. As was outlined in 2004 (Hagenbuch and Meier, 2004), to keep the nomenclature unambiguous, the Human Genome Nomenclature Committee suggested keeping the “SLC” symbol for the gene symbols and adding an “O” instead of using SLC21. This would result in the gene symbol of OATP1A2 as SLCO1A2 and not as SLC211A2, which could be confused with a potential OATP211 family. Six OATP families contain 11 human OATPs encoded by 11 genes (Fig. 1). One splice variant of the SLCO1B3 gene with the pseudogene SLCO1B7 coding for OATP1B3–1B7 has been identified and characterized (Malagnino et al, 2018) (Table 1). Figure 2 should help to visualize better the multiple gene duplications that occurred in rodents in subfamily 1A. In subfamily 1B, there occurred a gene duplication in primates only, resulting in OATP1B1 and OATP1B3. Thus, all primates have these 2 proteins, but rodents and other animals only have a single OATP1B protein, which is called OATP1B2 (Slco1b2) in rodents and OATP1B3 or OATP1B4 (SLCO1B2) in other nonhuman species. However, we would like to emphasize that only primates have 2 OATP1B subfamily members and that the OATP1B3/OATP1B4 proteins in other species are orthologs to the ancestor of the primate OATP1B1/OATP1B3. Similar gene duplications in rats and mice resulted in OATP6B, OATP6C, and OATP6D family members, while primates, including humans and other animals, only have a single OATP6A1 (SLCO6A1) protein (Fig. 3). A recent search of the National Center for Biotechnology Information database revealed that an OATP4A1-like protein is present in deep-sea sponges (Steffen et al, 2023) and corals (Selwyn and Vollmer, 2023). Both sponges and corals are among the evolutionary oldest animals and appeared more than 500 million years ago (Worheide et al, 2012). In 2002, an OATP1C1-like transporter was cloned and functionally characterized from another evolutionary old species, the little skate (Cai et al, 2002), which is a member of the cartilaginous fish that appeared almost 450 million years ago. In contrast, no OATP homologs were found when searching the prokaryotes and plants, strongly suggesting that OATPs have evolved with the animals. This suggestion is supported by the fact that OATPs have been identified and partially characterized in essentially every animal species investigated. In the following paragraphs, we will briefly summarize the main properties of each of the 12 human transporters in the 6 OATP families. Information on the sequences of these OATPs, as well as of OATPs from species used in drug development, are given in Table 1. Information on mRNA expression levels of hepatocellular SLCO genes can be found in Table 2. Several laboratories have determined the protein expression levels of hepatocellular OATPs in human liver tissue samples. These results are summarized in Table 3.

Fig. 1.

Phylogenetic tree and classification of members of the OATP/SLCO superfamily. The human OATPs are listed in capitals, while the other members are indicated by their scientific names. Families 1 and 6 do not have straight 1-to-1 orthologs in all other species and are enlarged in Figs. 2 and 3.

Table 1.

Gene and protein names of OATPs with NCBI reference sequence identifiers

Gene	Gene Locus	NM Number	Protein	Reference
Homo sapiens (human)
SLCO1A2	12.p12.1	NM_021094	OATP1A2	Kullak-Ublick et al, 1995
SLCO1B1	12.p12.1	NM_006446	OATP1B1	Abe et al, 1999
SLCO1B3	12.p12.2	NM_019844	OATP1B3	Konig et al, 2000a
SLCO1B3–SLCO1B7	12.p12.2	NM_001371097		Malagnino et al, 2018
SLCO1C1	12.p12.2	NM_017435	OATP1C1	Pizzagalli et al, 2002
SLCO2A1	3q22.1-q22.2	NM_005630	OATP2A1	Lu et al, 1996
SLCO2B1	11q13.4	NM_007256	OATP2B1	Tamai et al, 2000
SLCO3A1	15q26.1	NM_013272	OATP3A1	Tamai et al, 2000
SLCO4A1	20q13.33	NM_016354	OATP4A1	Tamai et al, 2000
SLCO4C1	5q21.1	NM_180991	OATP4C1	Mikkaichi et al, 2004
SLCO5A1	8q13.3	NM_030958	OATP5A1	Isidor et al, 2010
SLCO6A1	5q21.1	NM_173488	OATP6A1	Suzuki et al, 2003
Macaca fascicularis (cynomolgus monkey)
SLCO1A2		AB173194		Wang et al, 2007
SLCO1B1		JX866725	OATP1B1	Shen et al, 2013
SLCO1B3		JX866726	OATP1B3	Shen et al, 2013
SLCO1C1		XM_045364828	OATP1C1	Ebeling et al, 2011
SLCO2A1		AB168630	OATP2A1	Osada et al, 2005
SLCO2B1		JX866727	OATP2B1	Shen et al, 2013
SLCO3A1		XM_005560546	OATP3A1	Jayakumar et al, 2021
SLCO4A1		XM_015430056	OATP4A1	Jayakumar et al, 2021
SLCO4C1		XM_005557441	OATP4C1	Jayakumar et al, 2021
SLCO5A1		XM_045398363	OATP5A1	Jayakumar et al, 2021
SLCO6A1		XM_045394048	OATP6A1	Jayakumar et al, 2021
Macaca mulatta (rhesus monkey)
SLCO1A2		NM_001261711	OATP1A2	Zimin et al, 2014
SLCO1B1		XM_028829310	OATP1B1	Zimin et al, 2014
SLCO1B3		XM_015151097	OATP1B3	Zimin et al, 2014
SLCO1C1		XM_01515124	OATP1C1	Zimin et al, 2014
SLCO2A1		NC_041755	OATP2A1	Zimin et al, 2014
SLCO2B1		NM_001257810	OATP2B1	Zimin et al, 2014
SLCO3A1		XP_014998543	OATP3A1	Zimin et al, 2014
SLCO4A1		XM_001087895	OATP4A1	Zimin et al, 2014
SLCO4C1		XM_001097443	OATP4C1	Zimin et al, 2014
SLCO5A1		XM_001099565	OATP5A1	Zimin et al, 2014
SLCO6A1		XM_001095830	OATP6A1	Zimin et al, 2014
Canus lupus familiaris (dog)
SLCO1A2		XM_038576975	OATP1A2	Janssen et al, 2004
SLCO1B2		NM_001166047	OATP1B2	Gui and Hagenbuch, 2010
SLCO1C1		XM_038438837	OATP1C1
SLCO2A1		NM_001011558	OATP2A1
SLCO2B1		XM_038568619	OATP2B1	Janssen et al, 2004
SLCO3A1		XM_038532751	OATP3A1	Janssen et al, 2004
SLCO4A1		XM_038457884	OATP4A1	Lindblad-Toh et al, 2005
SLCO4C1		XM_038532243	OATP4C1	Janssen et al, 2004
SLCO5A1		XM_535098	OATP5A1	Lindblad-Toh et al, 2005
SLCO6A1		XM_038660728	OATP6A1
Rattus norvegicus (rat)
Slco1a1		NM_017111	OATP1A1	Jacquemin et al, 1994
Slco1a3		NM_030837	OATP1A3	Saito et al, 1996
Slco1a4		NM_131906	OATP1A4	Noe et al, 1997
Slco1a5		NM_030838	OATP1A5	Abe et al, 1998
Slco1a6		NM_130736	OATP1A6	Dransfeld et al, 2005
Slco1b2		NM_031650	OATP1B2	Cattori et al, 2000
Slco1c1		NM_053441	OATP1C1	Wittmann et al, 2015
Slco2a1		NM_022667	OATP2A1	Kanai et al, 1995
Slco2b1		NM_080786	OATP2B1	Nishio et al, 2000
Slco3a1		NM_177481	OATP3A1	Adachi et al, 2003
Slco4a1		NM_133608	OATP4A1	Fujiwara et al, 2001
Slco4c1		NM_001002024	OATP4C1	Choudhuri et al, 2003
Slco5a1		NM_001107898	OATP5A1	Kindla et al, 2011
Slco6b1		NR_185094	OATP6B1	Suzuki et al, 2003
Slco6c1		NM_173338	OATP6C1	Suzuki et al, 2003
Slco6d1		NM_001410265	OATP6D1	Gaudet et al, 2011
Mus musculus (mouse)
Slco1a1		NM_013797	OATP1A1	Hagenbuch et al, 2000
Slco1a4		NM_030687	OATP1A4	van Montfoort et al, 2002
Slco1a5		NM_001267707	OATP1A5	Walters et al, 2000
Slco1a6		NM_023718	OATP1A6	Choudhuri et al, 2001
Slco1b2		NM_020495	OATP1B2	Ogura et al, 2000
Slco1c1		NM_021471	OATP1C1	Tohyama et al, 2004
Slco2a1		NM_033314	OATP2A1	Pucci et al, 1999
Slco2b1		NM_001252530	OATP2B1	Okazaki et al, 2004
Slco3a1		NM_023908	OATP3A1	Chan et al, 2011
Slco4a1		NM_148933	OATP4A1	Blackshaw et al, 2004
Slco4c1		NM_172658	OATP4C1	Cheng et al, 2005
Slco5a1		NM_172841	OATP5A1	Dettman and Justice, 2008
Slco6c1		NM_028942	OATP6C1	Lin et al, 2021
Slco6d1		NM_027584	OATP6D1	Gaudet et al, 2011

Fig. 2.

Phylogenetic tree and classification of the OATP/SLCO 1 family. While in subfamily C, a 1-to-1 relationship exists, there were gene duplication events in subfamily A in rodents (Muridae) and in subfamily B in primates.

Fig. 3.

Phylogenetic tree and classification of the OATP/SLCO 6 family. Subfamily A contains the human OATP6A1 and has 1-to-1 orthologs in all other species, while subfamilies B, C, and D are rodent specific.

Table 2.

mRNA expression levels of SLCOs in human livers

Transporter mRNA	Minimuma	Maximuma	Meana	SD	Fold Difference	nb	Method	Reference
SLCO1B1	(2.78E-2)c 7.73E-4	(1.29E-1)c 1.66E-2	(6.95E-2)c 4.83E-3	(5.44E-3)c 2.96E-5	22	17	PCR ΔΔCT against β-actin	Ohtsuki et al, 2012
SLCO1B3	(8.68E-4)c 7.52E-7	(8.68E-3)c 7.53E-5	(3.82E-3)c 1.46E-5	(3.93E-3)c 1.54E-5	100
SLCO2B1	(4.64E-3)c 2.15E-5	(1.92E-1)c 3.69E-4	(7.90E-2)c	(7.38E-2)c	17
SLCO1B1	(0.033) 1.02	(0.475) 1.38	(0.219) 1.16	(0.129) 1.09	1.35	9	PCR ΔCT Against mean of GAPDH, PPIA, HMBS, RPLP0, and RPS9	Kurzawski et al, 2019
SLCO1B3	(0.145) 1.11	(0.686) 1.61	(0.344) 1.27	(0.178) 1.13	1.45
SLCO2B1	(0.070) 1.05	(0.397) 1.32	(0.197) 1.15	(0.100) 1.07	1.25
SLCO1B1	0.55c	1.85c	0.5c		3.3	20	PCR ΔCT Against mean of GAPDH, PPIA, HMBS, RPLP0, and RPS9	Drozdzik et al, 2022
SLCO1B3	0.25c	2.23c	0.98c		8.8
SLCO2B1	0.53c	1.70c	0.99c		3.2

^aΔΔCT or ΔCT values against housekeeping genes listed in the methods column are given.

^bn = number of samples analyzed.

^cEstimated from the article; ΔCT values from articles are given in parenthesis.

Table 3.

Protein expression levels of OATPs in human livers

Transporter Protein	Minimum	Maximum	Mean	SD	Fold Difference (Maximum/Minimum)	n	Method	Reference
	fmol/mg membrane protein	fmol/mg membrane protein	fmol/mg membrane protein
OATP1B1	<LLQ	12.3	2.74	3.67		17	Targeted proteomics with enriched plasma membrane fraction (pmol/mg protein)	Ohtsuki et al, 2012
OATP1B3	0.967	2.51	1.70	0.45	2.6
OATP2B1	<LLQ	3.18	0.46	0.87
OATP1B1			7.2	0.31		12	Targeted proteomics with total membrane protein fraction (fmol/mg membrane protein)	Karlgren et al, 2012
OATP1B3			6.3	0.41
OATP2B1			4.0	0.41
OATP1B1	3.0a	14.0a	9.67	4.27	5.0	9	Targeted proteomics with tissue lysate (fmol/mg protein)	Kimoto et al, 2012
OATP1B3	2.4	11.8	6.29	2.82	4.9
OATP2B1	1.9	4.9	3.68	1.39	2.6
OATP1B1	3.1	14.9	10.6	3.04	4.8	4	Targeted proteomics with tissue lysate (fmol/mg protein)	Balogh et al, 2013
OATP1B3	2.4	11.8	5.90	2.97	4.9
OATP2B1	1.2	4.6	2.85	1.13	3.8
OATP1B1	0.6a	5.1a	2.0	0.9	7.0	64	Targeted proteomics with total membrane protein fraction (fmol/mg membrane protein)	Prasad et al, 2014
OATP1B3	0.2a	2.4a	1.1	0.5	8.0
OATP2B1	0.6a	3.6a	1.7	0.6	5.0
OATP1B1	12.1	44.4	23.2	9.4	3.7	12	Targeted proteomics with total membrane protein fraction (fmol/mg membrane protein)	Vildhede et al, 2014
OATP1B3	0.2	6.0	3.8	0.2	32
OATP2B1	0.6	4.3	1.2	0.6	7.5
OATP1B1	0.95a	9.02a	2.99a	2.25a	9.5	41	Targeted proteomics with total membrane protein fraction (fmol/mg membrane protein)	Prasad et al, 2016
OATP1B3	0.3a	3.21a	1.66a	0.76a	10.7
OATP2B1	0.54a	2.85a	1.54a	1.63a	5.3
OATP1B1			5.52	2.94		15	Targeted proteomics with total membrane protein fraction (fmol/mg membrane protein)	Wang et al, 2017
OATP1B3			0.41	0.27
OATP2B1			2.31	0.70
OATP1B1	9.1	53.7	17.7b		5.9	8	Targeted proteomics with crude membrane fraction (pmol/g tissue)	van Groen et al, 2018
OATP1B3	15.1	38.9	21.9b		2.6
OATP2B1	34.5	77.8	61.3b		2.3
OATP1B1	106	341	228	73.1	3.2	9	Targeted proteomics with tissue lysate (fmol/mg tissue)	Kurzawski et al, 2019
OATP1B3	47.9	254	116	60.7	5.3
OATP2B1	15.9	61.1	44.9	14.0	3.8
OATP1B1	3.33a	15.7a	8.13	2.82	4.7	50–52	Targeted proteomics with total membrane protein fraction (pmol/mg membrane protein)	Vildhede et al, 2020
OATP1B3	0.32a	8.92a	4.46	2.27	27.6
OATP2B1	0.32a	4.48a	2.55	0.83	14.0
OATP1B1	0.88	4.67	2.70	0.90	5.2	54	Shot gun proteomics with tissue lysate (pmol/mg protein)	Wegler et al, 2021
OATP1B3	0.45	2.48	1.17	0.31	5.7
OATP2B1	0.49	1.12	0.80	0.072	2.3
OATP1B1	0.41	2.76	1.16	0.56	6.7	29	Targeted proteomics with total membrane protein fraction (pmol/mg membrane protein)	Achour et al, 2021
OATP1B1			34	12		14	Targeted proteomics with scaling to total liver protein (pmol/mg membrane protein)	El-Khateeb et al, 2021
OATP1B3			31.8	14
OATP2B1			48.5	41
OATP1B1	876a	187a	438a		4.7	20	Targeted proteomics with tissue lysate (fmol/mg tissue)	Drozdzik et al, 2022
OATP1B3	277a	14.0a	86.0a		19.8
OATP2B1	205a	51.2a	130a		5.6

^aEstimated from the article.

^bMedian.

A. Family OATP1

OATP1A2 (SLCO1A2) is a glycoprotein of 670 amino acids (representing isoform 1) (Table 4) that was first cloned in 1995 (Kullak-Ublick et al, 1995). Twelve additional protein isoforms encoded by 30 different transcripts have been identified. However, to our knowledge, none of these additional isoforms are functional. Although OATP1A2 was originally cloned from a human liver cDNA library, its mRNA is most strongly expressed in the brain, followed by the lungs, liver, kidneys, and testes (Kullak-Ublick et al, 1995). At the protein level, OATP1A2 has been detected in brain capillaries, at the apical membrane of the distal nephron, in cholangiocytes (Gao et al, 2000; Lee et al, 2005), at the basolateral membrane of the pars plana in the ciliary body (Gao et al, 2005), in the retina, and in neurons (Gao et al, 2015). The initially reported localization in the small intestine (Glaeser et al, 2007) could not be confirmed later at the mRNA (Hilgendorf et al, 2007; Meier et al, 2007; Takada et al, 2014) or protein level (Drozdzik et al, 2014).

Table 4.

Splice variants of human SLCOs

Transcript Variant	Isoform	mRNA Length	Accession No.	Amino Acids	Functional	References
		bp
SLCO1A2
1	1	7750	NM_134431	670	Yes	Kullak-Ublick et al, 1995
2	1	7231	NM_021094	670	Yes
3	1	7256	NM_001386878	670	Yes
4	1	7114	NM_001386879	670	Yes
5	1	7109	NM_001386880	670	Yes
6	1	7309	NM_001386881	670	Yes
7	1	7258	NM_001386882	670	Yes
8	2	7317	NM_001386886	668
18	2	7387	NM_001386926	668
27	2	7469	NM_001386947	668
29	2	7581	NM_001386949	668
36	2	7275	NM_001386960	668
28	3 precursor	7067	NM_001386948	650
31	3 precursor	7179	NM_001386952	650
9	4	7214	NM_001386887	617
34	5 precursor	7172	NM_001386958	597
30	6 precursor	6884	NM_001386951	589
26	7 precursor	2097	NM_001386946	584
13	8	6925	NM_001386908	583
14	8	7330	NM_001386919	583
19	8	7122	NM_001386927	583
35	9 precursor	2426	NM_001386959	578
16	10 precursor	6827	NM_001386921	570
17	11	7046	NM_001386922	538
20	11	7142	NM_001386929	538
22	11	7485	NM_001386937	538
23	11	6839	NM_001386938	538
24	11	6976	NM_001386939	538
25	11	6981	NM_001386940	538
38	11	6934	NM_001386962	538
39	12	2152	NM_001386963	517
10	13	6818	NM_001386890	514
15	13	7223	NM_001386920	514
21	13	7015	NM_001386931	514
32	13	7168	NM_001386953	514
33	13	6685	NM_001386954	514
37	13	6960	NM_001386961	514
SLCO1B1
1	1	2787	NM_006446	691	Yes	Abe et al, 1999
SLCO1B3
1 (Lt-OATP1B3)	1	3013	NM_019844	702	Yes	Konig et al, 2000b
2 (Ct-OATP1B3)	2	2815	NM_001349920	674	Yes	Thakkar et al, 2013
SLCO1C1
1	1	3468	NM_001145946	730
2	2	3498	NM_017435	712	Yes	Pizzagalli et al, 2002
3	3	3228	NM_001145945	663
4	4	3305	NM_001145944	612
SLCO2A1
1	1	4067	NM_005630	643	Yes	Lu et al, 1996
SLCO2B1
1	1	4374	NM_007256	709	Yes	Tamai et al, 2000
2	2	4156	NM_001145211	687
3	3	3942	NM_001145212	565
SLCO3A1
1	1	5102	NM_013272	710	Yes	Tamai et al, 2000
2	2	2853	NM_001145044	692	Yes	Huber et al, 2007
SLCO4A1
1	1	2716	NM_016354	722	Yes	Tamai et al, 2000
SLCO4C1
1	1	5069	NM_180991	724	Yes	Mikkaichi et al, 2004
SLCO5A1
1	1	8991	NM_030958	848
2	2	8987	NM_001146008	687
3	3	3430	NM_001146009	793
SLCO6A1
1	1	2682	NM_173488	719
2	2	2630	NM_001289002	719
3	3	2496	NM_001289004	657
4	4	1871	NM_001308014	466

OATP1B1 (SLCO1B1) is 1 of the 2 “liver-specific” OATPs. It is a glycoprotein (Konig et al, 2000a), has 691 amino acids, and so far, only 1 isoform has been reported (Abe et al, 1999). Besides in the liver, OATP1B1 mRNA has been found in the placenta (Wang et al, 2012), in the hormone-independent breast cancer cell line MDA-MB-231 (Banerjee et al, 2012), and in several human cancers (Pressler et al, 2011; Svoboda et al, 2011). At the protein level, OATP1B1 is almost exclusively expressed at the sinusoidal or basolateral membrane of human hepatocytes.

OATP1B3 (SLCO1B3) is the other so-called liver-specific OATP. Two isoforms of this glycoprotein are known. Isoform 1 has 702 amino acids, being cloned from human liver cDNA, is glycosylated (Konig et al, 2000a), and is known as liver-type (Lt)-OATP1B3. Isoform 2, a 674-amino acid splice variant of Lt-OATP1B3, is missing the first 28 amino acids of the full-length protein. It is found in many cancer cell lines and human cancers and, therefore, called cancer-type (Ct)-OATP1B3 (Thakkar et al, 2013). While Lt-OATP1B3 is a multispecific endobiotic and xenobiotic uptake transporter, Ct-OATP1B3 is mainly expressed intracellularly in a lysosomal fraction (Haberkorn et al, 2022) and has a much-reduced function (Thakkar et al, 2013; Chun et al, 2017; Haberkorn et al, 2022).

We want to emphasize that currently available antibodies against OATP1B3 cannot distinguish between the 2 proteins because any antibody recognizing Ct-OATP1B3 also recognizes Lt-OATP1B3 because of their identical amino acid sequence (Thakkar et al, 2013). Only specific RT-PCR primers can detect and distinguish the 2 splice variants.

OATP1B3–1B7 (SLCO1B3–SLCO1B7) is a splice variant of the SLCO1B1 and SLCO1B7 genes and encodes a protein of 687 amino acids (Malagnino et al, 2018, 2019a). It is mainly expressed intracellularly in the liver and the small intestine, where it might mediate the transport of drugs into and out of the smooth endoplasmic reticulum, facilitating their conjugation (Malagnino et al, 2019a).

OATP1C1 (SLCO1C1) is the third member of the OATP1 family, isolated from a human brain cDNA library (Pizzagalli et al, 2002). Four isoforms have been reported that encode proteins with 612 to 730 amino acids, but so far, only isoform 2, encoding a 712-amino acid protein, has been shown to be functional. It is mainly expressed in the brain, the testes, and the heart. Functionally, it is a high-affinity thyroid hormone transporter with a narrow substrate specificity (Pizzagalli et al, 2002).

B. Family OATP2

OATP2A1 (SLCO2A1) is 1 of the 2 members in the OATP2 family, originally cloned from a human kidney library (Lu et al, 1996). A single isoform of this 643-amino acid protein, which is ubiquitously expressed with higher levels in the lungs and testes, has been reported (Fallon et al, 2018; Hau et al, 2022). OATP2A1 is a prostaglandin transporter with a narrow substrate specificity limited to prostanoids and related chemicals. It seems important to terminate prostaglandin signaling by mediating the uptake of prostaglandins for their eventual metabolism and, thus, inactivation (Nakanishi et al, 2021).

OATP2B1 (SLCO2B1) is the second member of the OATP2 family. Three isoforms are known to encode proteins with 565 to 709 amino acids. So far, only isoform 1, a 709-amino acid glycoprotein, is functional. OATP2B1, cloned in 2000 from a human brain cDNA library (Tamai et al, 2000), is ubiquitously expressed in many epithelia, including the small intestine, the liver, and the kidneys (Kinzi et al, 2021). Functionally, OATP2B1 transports a wide variety of endobiotics and xenobiotics, and its transport function is enhanced by an outside acidic pH, which would help with the absorption of drugs in the small intestine (Kinzi et al, 2021).

C. Family OATP3

OATP3A1 (SLCO3A1) is a single member of the OATP3 family. Transcript variant 1 (OATP3A1_v1), which codes for a 710-amino acid protein, was isolated from a human kidney library (Tamai et al, 2000). A second transcript variant encoding a 692-amino acid protein (OATP3A1_v2) was isolated from a human brain library (Huber et al, 2007). OATP3A1 is the OATP with the highest amino acid identity among the different species and is expressed, at least at the mRNA level, ubiquitously, with the strongest expression in the testis, heart, brain, lung, and spleen (Tamai et al, 2000; Huber et al, 2007). Functionally, OATP3A1 has not been extensively characterized. Initial uptake studies in OATP3A1-expressing human embryonic kidney (HEK) 293 cells demonstrated that it transports estrone-3-sulfate (E1S), benzylpenicillin, and prostaglandin E₂ (Tamai et al, 2000). When stably expressed in Chinese hamster ovary (CHO) cells, both splice variants transported in addition to prostaglandins also thyroxine, BQ-123, and vasopressin (Huber et al, 2007). A more recent study reported that the expression of OATP3A1 at the sinusoidal membrane of human hepatocytes is upregulated under cholestatic conditions, and it functions as a bile acid efflux transporter (Pan et al, 2018). Furthermore, a recent report characterized OATP3A1 as an amino acid transporter, particularly for l-tryptophan, l-tyrosine, and l-phenylalanine (Surrer et al, 2024).

D. Family OATP4

OATP4A1 (SLCO4A1) was the first of the 2 members of the OATP4 family and was cloned from a human kidney cDNA library (Tamai et al, 2000). A single transcript variant encoding a 722-amino acid protein is known. OATP4A1 is expressed ubiquitously with higher mRNA levels in the heart, the lungs, and the placenta. It can transport prostaglandins, some hormone conjugates [E1S and estradiol-17β-glucuronide (E17βG)], thyroid hormones, and certain bile acids (Tamai et al, 2000; Fujiwara et al, 2001). Besides its ubiquitous expression in normal tissues, OATP4A1 is also expressed in several cancer cell lines and tissues.

OATP4C1 (SLCO4C1) is the second member of the OATP4 family with a single transcript encoding a 724-amino acid protein. It was originally cloned from a human kidney cDNA library and characterized as a digoxin transporter localized to the basolateral membrane (Mikkaichi et al, 2004). In addition to digoxin, it also transports E1S, triiodothyronine, the dipeptidyl-peptidase 4 inhibitor sitagliptin, and l-arginine and its derivatives (Mikkaichi et al, 2004; Chu et al, 2007; Yamaguchi et al, 2010; Taghikhani et al, 2020). Furthermore, some uremic toxins are substrates of OATP4C1 (Toyohara et al, 2009; Taghikhani et al, 2019, 2020).

E. Family OATP5

OATP5A1 (SLCO5A1) is the only member of the OATP5 family in humans. Three transcript variants coding for proteins with 848, 687, and 793 amino acids, respectively, are known. These transcripts are expressed ubiquitously in essentially all tissues. The open reading frame of the 848-amino acid protein was isolated from a cDNA obtained from mature dendritic cells and expressed in X laevis oocytes or HeLa cells. It encoded a glycoprotein of about 105 kDa, which was able to homodimerize. Using immunofluorescence and a yellow fluorescent fusion protein, expression of OATP5A1 was seen in intracellular membranes as well as at the plasma membrane of HeLa cells. None of the general OATP substrates was transported (Sebastian et al, 2013). In a more recent study, the SLCO5A1 gene was identified in a GWAS analysis for juvenile myoclonic epilepsy, and further analysis of the gene in a Drosophila model suggested that OATP5A1 is involved in impulsivity and seizure susceptibility (Roshandel et al, 2023).

F. Family OATP6

OATP6A1 (SLCO6A1) has 4 known transcript variants that code for proteins of 719, 719, 657, and 466 amino acids, respectively. Its mRNA has been detected mainly in testis and in certain cancers (Suzuki et al, 2003; Lee et al, 2004; Fietz et al, 2013). However, no functional transport could be determined for this protein (Fietz et al, 2013).

IV. Substrate specificity

A. Endogenous compounds

The first OATP was cloned by following the transport of BSP during expression cloning. BSP was chosen as the lead substrate because this compound was previously postulated to be (in part) transported into hepatocytes by hitchhiking a hepatocellular bilirubin uptake system (Clarenburg and Kao, 1973; Scharschmidt et al, 1975). After BSP, conjugated and unconjugated bile acids were identified as OATP substrates, and the list of demonstrated endobiotic and xenobiotic OATP substrates swiftly increased. OATPs generally mediate the transport of large [molecular weight (MW) > 350 Da], bulky, and often anionic compounds. However, they can also accommodate cationic and neutral compounds. OATPs often transport the more hydrophobic compounds, while the members of the SLC22 gene family transport mainly small (MW < 350 Da), more water-soluble, anionic and cationic compounds. Substrates of OATPs constitute a large variety of endogenous molecules, including bile acids and their derivatives, hormones like thyroid hormones, sulfated and glucuronidated steroid metabolites, and peptides. In addition, various natural xenobiotics, including flavonoids and several toxins, as well as numerous other chemicals, such as synthetic xenobiotics and drugs, are substrates of OATP.

Several reviews have listed endogenous substrates (Hagenbuch and Stieger, 2013; Stieger and Hagenbuch, 2014). About a decade after the discovery of the first OATP, rOATP1A1 (Jacquemin et al, 1994), bilirubin was finally confirmed in vitro to be a substrate for OATP1B1 and OATP1B3 (Cui et al, 2001b; Briz et al, 2003). Understanding the molecular defect that leads to Rotor syndrome confirmed the role that the liver OATPs, OATP1B1 and OATP1B3, play in hepatocellular bilirubin uptake (van de Steeg et al, 2012). Rotor syndrome is a rare, benign syndrome, presenting with normal liver enzymes and a mild, predominantly conjugated hyperbilirubinemia, and (if assessed) a delayed excretion of anionic dyes (BSP) with a single plasma peak. In urine, such patients display elevated coproporphyrins (with about 65% coproporphyrin I). Patients with Rotor syndrome have homozygous mutations in the neighboring SLCO1B1 and SLCO1B3 genes, leading to a simultaneous deficiency of both OATP1B1 and OATP1B3. Glucuronidated bile salts are massively elevated in the plasma of these patients (Kimura et al, 2021). Owing to the limited number of patients analyzed, no clear data on plasma levels of unconjugated bile acids are available in this study. However, it should be pointed out that in mice with inactivated Slco1a and Slco1b genes, total bile acids and unconjugated bile acids were massively elevated, while conjugated bile salts were normal in plasma (van de Steeg et al, 2010). Similarly, mice with an inactivated Slco1b2 gene were reported with elevated total plasma bile acids, again due to elevated unconjugated bile acids with normal conjugated bile salts (Csanaky et al, 2011). A comparable study in rats with an inactivated Slco1b2 gene reported elevated plasma levels of conjugated and unconjugated bilirubin as well as total bile acids (Ma et al, 2020). Patients with mutations in the SLC10A1 gene leading to nonfunctioning Na⁺/taurocholate cotransporting polypeptide (NTCP) have elevated bile acids (Schneider et al, 2022), whereby the unconjugated bile acids are typically not elevated or only marginally elevated (Liu et al, 2017; Vaz and Ferdinandusse, 2017; Schneider et al, 2022). These findings clearly demonstrate that despite in vitro data, hepatocellular OATPs do not (efficiently) transport conjugated bile acids but seem to be relevant for the transport of unconjugated bile acids.

B. Exogenous compounds and drugs

Soon after cloning the first OATP from rat liver, it was realized that rOATP1A1 also mediates the transport of drugs, such as APD-adjmalinium (Bossuyt et al, 1996a) or rocuronium (van Montfoort et al, 1999). These compounds, interestingly, are cations. In addition, ouabain was also identified as an OATP substrate. Hence, OATPs can mediate the transport of anionic, neutral, and cationic compounds. The second identified OATP, rOATP1A4, mediates the transport of the cardiac glycoside digoxin (Noe et al, 1997), a finding corroborating OATPs as transporters of not only endogenous compounds but also drugs. The important role of transporters in drug disposition is well understood today (Iversen et al, 2022). Consequently, members of the OATP superfamily are now generally accepted as important drug transporters, and lists of drug substrates have been compiled in numerous reviews (Hagenbuch and Gui, 2008; Fahrmayr et al, 2010; Niemi et al, 2011; Emami Riedmaier et al, 2012; Grandvuinet et al, 2012; Nakanishi and Tamai, 2012; Obaidat et al, 2012; Roth et al, 2012; Tamai, 2012; Chu et al, 2013; Hagenbuch and Stieger, 2013; Kovacsics et al, 2017; Oswald, 2019; Kinzi et al, 2021; Nies et al, 2022). Besides these reviews, the DRUGBANK Online (https://go.drugbank.com) is a good source of substrates and inhibitors of OATPs and other transporters, and access is free for most academic researchers.

Polypharmacy is associated with DDIs (Khezrian et al, 2020; Wolff et al, 2021; Bettonte et al, 2022). Such interactions can occur both at the pharmacokinetic and pharmacodynamic levels. Pharmacokinetic DDIs involve not only interactions at the stage of drug metabolism but also those that frequently occur at the level of drug transporters, such as at OATPs (Yoshida et al, 2012; Yu et al, 2017; Taskar et al, 2020; Kikuchi et al, 2023; Armani et al, 2024). Consequently, the regulatory agencies recommend in their guidelines testing of new chemical entities on interactions (eg, by inhibition studies) with specific OATPs and other drug transporters during drug development.

Natural products, including toxins, are also known to interact with OATPs or are substrates of OATPs (Tamai, 2012; Stieger and Hagenbuch, 2014; Stieger et al, 2017; Oswald, 2019; Oyanna and Clarke, 2024). For example, fruit juices not only affect drug metabolism by inhibiting cytochrome P450 but also inhibit drug transport, both in the intestine and in other organs (Dolton et al, 2012). Therefore, OATPs significantly contribute to altered drug action caused by natural products. For example, one of the toxins of the deadly mushroom death cap, α-amanitin, is a substrate of OATP1B3 (Letschert et al, 2006), and its uptake can be inhibited by the antidot silibinin (Letschert et al, 2006; Wlcek et al, 2013).

C. Imaging agents

Imaging is a routine procedure for many indications in the diagnostic workup of patients complementing, for example, plasma parameters. Imaging comprises a wide variety of methods, both with and without exogenously applied agents such as contrast agents (Barentsz et al, 2006; Hussain et al, 2022). For liver diseases, typical plasma parameters are bilirubin and alanine aminotransferase. However, alanine aminotransferase is not completely specific to the liver and is also expressed in other cells (Lindblom et al, 2007). For the management of patients with advanced liver disease, scoring systems such as the Child–Turcotte–Pugh score or the model for end-stage liver disease score are important tools. These scores consider, among other parameters, plasma bilirubin levels, which monitor the transport and detoxification capacity of the liver. Lately, preoperative scintigraphic monitoring of the hepatobiliary transport of [^99mTc]-mebrofenin was demonstrated to improve the postoperative outcome of liver resections (Rassam et al, 2019), and we know that [^99mTc]-mebrofenin is a substrate of OATP1B1 and OATP1B3 among other hepatocellular transporters (Table 5). So-called liver function tests may provide additional important information on liver function in certain clinical situations (Stieger et al, 2012). These liver function tests can be divided into passive or active liver function tests (Hoekstra et al, 2013). Passive liver function tests involve the determination of liver enzymes in plasma or histologic analysis of liver biopsies, while active liver function tests measure hepatic plasma clearance of compounds or the metabolic transformation capacity of the liver. The clearance of BSP was a classic test to monitor liver function with an exogenous marker, but it was later replaced by indocyanine green clearance for safety reasons (Sakka, 2007). Indocyanine green is transported into hepatocytes by OATP1B3 and NTCP (Table 6) (de Graaf et al, 2011). For patients with advanced liver disease, it is important to discriminate between diffuse and focal liver lesions. This can be accomplished using imaging methods (Pastor et al, 2014). The 3 main methods for liver imaging are magnetic resonance imaging, positron emission tomography, or single-photon emission computed tomography (Pastor et al, 2014). These methods typically require the use of contrast agents. For example, magnetic resonance imaging agents are highly membrane impermeable, and many are OATP substrates. Tables 5 and 7 list OATP imaging agents established as OATP substrates using OATP-expressing in vitro assays. In contrast to these in vitro results, more recently, in vivo imaging methods were used to monitor transporter function in humans (Mann et al, 2016; Tournier et al, 2018). Fluorescent OATP substrates have, to the best of our knowledge, not yet reached routine applications in clinical diagnostics and are, therefore, not covered in detail in this review. A recent review compiled a list of fluorescent OATP substrates (Stieger, 2022) that may be useful for in vitro transport assays (Patik et al, 2015; Ozvegy-Laczka et al, 2023; Ungvari et al, 2023) or to establish high through-put assays to screen for OATP modulators (Gui et al, 2010). A proof-of-concept study provided evidence that the fluorescent bile salt derivative cholyl-lysyl-fluorescein might be a useful exogenous liver function marker in patients (Milkiewicz et al, 2000). However, no further clinical data are available, which might be due to the observation that this compound is a substrate neither of the major bile salt uptake system NTCP nor of the bile salt export pump (BSEP) but of OATP1B3 and the multidrug resistance–associated proteins MRP2 and MRP3 (de Waart et al, 2010). Along another line, fluorescein, which is a substrate of OATPs and MRP2, is currently being evaluated for assessing the degree of hepatic ischemia-reperfusion injury in a rat model by using physiologically based pharmacokinetic modeling in ex vivo perfused livers (Monti et al, 2024). Lately, indocyanine green application via a retrograde approach using an endoscopic nasobiliary access has been tested for monitoring of the hepatic parenchyma in patients in an explorative study (Gao et al, 2024). This method could, for example, provide an imaging tool for guiding laparoscopic segmental liver resection in patients having local intrahepatic bile duct stones. This is complemented by the demonstration of the usefulness of indocyanine green for the laparoscopic evaluation of complex forms of hepatolithiasis (Wang et al, 2024a) or in the functional characterization of patients with alterations in SLCO1B3 (Tanimoto et al, 2024).

Table 5.

PET and scintigraphic agents tested on hepatocellular organic anion transporters

Data are from individual transporters expressed heterologously in different cell lines. For inhibitor, radioactively labeled taurocholate was used as substrate.

	OATP1B1	OATP1B3	OATP2B1	Reference
[99m]Tc-mebrofenin	Substrate Km 1.68 mM	Substrate Km 2.57 mM	NT	Ghibellini et al, 2008; de Graaf et al, 2010; Neyt et al, 2016
(15R)-[11C]-Tic_me	Substrate Km 0.96 μM	Substrate Km 1.3 μM		Takashima et al, 2012
[11C]-SC-62807	Substrate Km 260 μM	Substrate Km 19.8 μM		Takashima et al, 2013
[99m]Tc-N-pyridoxyl-5-methyl-tryptophan	Substrate	Substrate	NT	Kobayashi et al, 2014
[18F]pitavastatin; [18F]pitavastatin derivative ([18F]PTV-F1)	Substrate	Substrate		Kimura et al, 2016
[99mTc]-DTPA-CDCA (chenodeoxycholate derivative)	Substrate Km 61.7 mM	Substrate Km 162 mM		Neyt et al, 2016
[99mTc]-DTPA-CA (cholate derivative)	Substrate Km 8.45 μM	Substrate Km 13.7 μM		Neyt et al, 2016
[99mTc]-DTPA-MEB (mebrofenin derivative)	Substrate Km 958 μM	Substrate Km 3547 μM		Neyt et al, 2016
LCA[18F]TD(lithocholate derivative)	Substrate Km 10 μM	Substrate		Testa et al, 2017
[11C]erlotinib	NT	NT	Substrate Km 0.324 μM	Bauer et al, 2018
[11C]Dehydroprava-statin	Substrate	Substrate		Kaneko et al, 2018
3α-[18F]fluorocholic acid	Inhibitor			De Lombaerde et al, 2018
3β-[18F]fluorochenodeoxycholic acid	Inhibitor			De Lombaerde et al, 2018
3β-[18F]fluoroglycocholic acid	Inhibitor			De Lombaerde et al, 2018
2β-[18F]fluorocholic acid	Inhibitor			De Lombaerde et al, 2018
7β-[18F]fluorocholic acid	Inhibitor			De Lombaerde et al, 2018
[131I]6-β-iodomethyl-19-norcholesterol	Substrate	Substrate	NT	Kobayashi et al, 2019
[11C]tariquidar	NT	NT	NT	Hernandez Lozano et al, 2020
[123I]m-iodobenzyl guanidine)	NT	NT	NT	Kobayashi et al, 2020
[125I]acetaminophen	Substrate	Substrate	Substrate	Sato et al, 2023
5 Mn complexes of 1,2-diaminocyclohexane-N,N,N′,N′-tetraacetic acid derivatives	2-complex substrates	NT		McRae et al, 2023

NT, not transported.

Table 6.

Dyes tested on hepatocellular organic anion transporters

Data are from individual transporters expressed heterologously in different cell lines.

	OATP1B1	OATP1B3	OATP2B1	Reference
BSP	Substrate Km 0.3 μM	Substrate Km 0.4 μM	Substrate Km 0.7 μM	Kullak-Ublick et al, 2001
Fluo-3	Substrate Km 2.03 μM	Substrate Km 3.1 μM		Cui et al, 2001a; Gui et al, 2008; Izumi et al, 2016
Oregon Green 488 taxol	Substrate	Substrate		Gui et al, 2010
Alexa Fluor 488 methotrexate	NT	NT		Gui et al, 2010
Fluorescein	Substrate	Substrate	Substrate Km 19.8 μM	Gui et al, 2010; De Bruyn et al, 2011; Kawasaki et al, 2020
Flutax-2	NT	Substrate	NT	Izumi et al, 2016
8-Fluorescein-cAMP	Substrate Km 2.9; 9.05 μM	Substrate Km 1.8 μM	Substrate	Bednarczyk, 2010; Izumi et al, 2016
Fluorescein methotrexate	Substrate Km 3.8; 0.23 μM	Substrate Km 7.9; 0.53 μM	Substrate	Gui et al, 2010; Patik et al, 2015; Szekely et al, 2020
Indocyanine green	NT	Substrate	NT	de Graaf et al, 2011
Fluorescein	Substrate Km 25.7; 19.1 μM	Substrate Km 38.6 μM	Substrate	Patik et al, 2015; Izumi et al, 2016
Six fluorescein derivatives of trimethyl lock quinone-cyanobenzothiazole and trimethyl lock quinone-luciferin	Substrate Km 0.3 to 0.8 μM	Substrate Km 0.2 to 1.3 μM		Mustafa et al, 2016
Dibromofluorescein	Substrate Km 4.16 μM	Substrate	Substrate Km 0.818 μM	Izumi et al, 2016; Kawasaki et al, 2020
Dichlorofluorescein	Substrate Km 5.29 μM		Substrate Km 6.79 μM	Izumi et al, 2016; Kawasaki et al, 2020
Oregon Green	Substrate Km 54.1 μM	Substrate	Substrate Km 87.1 μM	Izumi et al, 2016; Kawasaki et al, 2020
Alexa Fluor 405	Substrate Km 1.29 μM	Substrate Km 16.98 μM	Substrate Km 4.10 μM	Patik et al, 2018
Cascade blue hydrazide	Substrate Km 2.6 μM	Substrate Km 21 μM	Substrate Km 21 μM	Patik et al, 2018; Szekely et al, 2020
Live/dead blue	Substrate	Substrate	Substrate	Patik et al, 2018
Live/dead green	Substrate	Substrate	Substrate	Patik et al, 2018; Szekely et al, 2020
Live/dead violet	Substrate	Substrate	Substrate	Patik et al, 2018; Szekely et al, 2020
Zombie violet	Substrate	Substrate	Substrate	Patik et al, 2018; Szekely et al, 2020
Sulforhodamine 101	Substrate	Substrate	NT	Bakos et al, 2020
Pyranine	Substrate Km 27.8 μM	Substrate Km 92.2 μM	Substrate Km 65.6 μM	Szekely et al, 2020
Sulforhodamine 101	Substrate			Szekely et al, 2020
Eosin y			Substrate Km 0.694 μM	Kawasaki et al, 2020
5-Carboxyfluorescein			Substrate Km 8.56 μM	Kawasaki et al, 2020
6-Carboxyfluorescein			Substrate Km 9.92 μM	Kawasaki et al, 2020
5-Carboxy-2',7'-dichlorosulfonfluorescein			Substrate Km 5.25 μM	Kawasaki et al, 2020
5-Carboxy-2′,7′-dichlorofluorescein			Substrate	Kawasaki et al, 2020
Lucifer yellow				Szekely et al, 2020
Indocyanine green	Substrate	Substrate		Lee et al, 2021
Phenolsulfonphthalein	Substrate Km 11.3 μM	Substrate Km 7.0 μM	Substrate Km 5.1 μM	Wang et al, 2021
6,8-Dihydroxy-1,3-pyrenedisulfonic acid	Substrate	Substrate	Substrate	Ungvari et al, 2021
Ace (8-acetoxy-1,3,6-pyrenetrisulfonic acid)	Substrate Km 17.6 μM	Substrate Km 61.4 μM	Substrate Km 17.1 μM	Ungvari et al, 2021

NT, not transported.

Table 7.

Magnetic resonance imaging substrates tested on hepatocellular organic anion transporters

Data are from individual transporters expressed heterologously in different cell lines.

	OATP1B1	OATP1B3	OATP2B1
Gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)	Substrate Km 0.7 mM	Substrate Km 4.1 mM	NT	Leonhardt et al, 2010
Gd-EOB-DTPA	Substrate Km 1.17 mM	Substrate Km 0.532 mM		Nassif et al, 2012; Wu et al, 2019
Gd-BOPTA	Substrate	Substrate		Lee et al, 2021

NT, not transported.

D. Biomarkers

Exogenous markers used for imaging can be applied in a microdosing setup and, consequently, reduce the risk of adverse events caused by such markers. However, many currently used imaging agents impose radiation on the body, which is a health risk factor. Replacement of such markers by endogenous biomarkers, particularly for drug interaction studies, would eliminate such a risk. The heme metabolites coproporphyrin I and III have been identified as OATP substrates and as potential endogenous biomarkers for pharmacokinetic drug studies (Bednarczyk and Boiselle, 2016; Lai et al, 2016; Shen et al, 2016). The analysis of 20 endogenous biomarkers in the same individuals demonstrated that none of the additional analyzed endogenous biomarkers outperform coproporphyrin I in the in vivo evaluation of OATP1B-mediated DDI (Barnett et al, 2019). Multiple studies in humans have until now supported the usefulness of coproporphyrin I and III as endogenous biomarkers for OATPs (Muller et al, 2018; Shen et al, 2018; Yee et al, 2019; Kalluri et al, 2021; Neuvonen et al, 2021; Lai, 2023; Mochizuki and Kusuhara, 2023; Kikuchi et al, 2023; Kinzi et al, 2024a), including cancer patients (Chu et al, 2018; Muller et al, 2018; Wang et al, 2024b). Furthermore, Kikuchi et al (2023) highlighted using coproporphyrin I for DDI assessments and proposed a decision tree for guiding these assessments. Metabolomic analysis and genome-wide association studies have identified additional potential endogenous biomarkers, including hexadecanedioate, tetradecanedioate, and unconjugated and conjugated bilirubin (Yee et al, 2016). Identification and validation of endogenous biomarkers as endogenous transporter substrates is currently a very active research area (Mariappan et al, 2017; Li et al, 2021; Kimoto et al, 2022; Jin et al, 2024).

E. Substrate predictions

Bile salts and estrogen metabolites have both a steroid derived structure. For a better understanding of the broad and overlapping substrate specificity, multiple studies have used a quantitative structure-activity relationship approach to find the structural motives of OATP substrates. In silico modeling of OATP substrates is a frequently used tool for flagging or predicting potential clinically relevant DDIs early in preclinical development (Lane et al, 2022). Using this approach, a pharmacophore feature of OATP substrates containing an anionic head group and a large hydrophobic moiety was identified (Yarim et al, 2005). This is also a feature of steroid derivatives such as, for example, E1S, a prototypic OATP substrate. In quantitative structure-activity relationship studies, the feature of steroid derivatives yielded a rather populated structural cluster for OATP substrates (Türkova et al, 2019). It did, therefore, not come as a big surprise that OATP74D of Drosophila species mediates the transport of the steroid derivative ecdysone (Okamoto et al, 2018). Ecdysone carries no charge and is a prohormone of hormones stimulating metamorphosis and molting in insect larvae. The conversion of the prohormone to its active form occurs intracellularly, which requires the membrane presence of OATP74D in Drosophila species (Yamanaka, 2021).

V. Experimental systems

The aim of pharmacodynamics is to gain a deep understanding of the molecular mechanisms leading to a drug’s action in the body. In this context, a drug’s pharmacokinetics will govern the concentration of a drug at its molecular target. Hence, in transport research, a thorough understanding of the molecular mechanism(s), including determining kinetic parameters of drug transport across barriers, is the ultimate goal.

Transport processes can be studied at the level of an organism and at the level of an isolated organ, isolated cells, and cell lines stably expressing a transporter or subcellular components. The intestine, liver, and kidney are key organs in drug disposition. These organs are all composed of many different cell types, and consequently, working in vivo with humans or with experimental animals to study drug disposition would be ideal. Working with humans or experimental animals, as well as their isolated organs, allows us to study the complex interplay of different transport processes and drug metabolism in the different organs as well as different cell types. During drug development, phase I studies involving humans are performed to obtain, among others, information on the pharmacokinetic parameters of a drug (often newly developed) and are tightly regulated by regulatory authorities (Shen et al, 2019). However, the complexity of organisms and organs makes conclusions on an individual molecular process, such as drug uptake into hepatocytes by an individual OATP, very challenging, if not impossible. While invasive procedures are feasible in animals, this approach in humans is very restricted. However, it is possible to study intestinal drug absorption in the upper human intestine with a balloon method by determining the disappearance of the drug from the luminal side and the appearance of the drug in plasma (Lennernas et al, 1992). Here, it has to be taken into account that plasma concentrations are among many factors governed by the liver, which can alter the systemic concentration by a first-pass effect. Again, determining the uptake of drugs or xenobiotics into human hepatocytes is impossible in situ unless the drug under investigation is available as a molecule amenable to imaging. At the systemic level, transport processes into hepatocytes can be indirectly monitored by clearance determinations in humans (Stieger et al, 2012). Experimental animals can be genetically modified by creating knockout strains for individual or several transporters or by creating so-called humanized animals expressing human instead of animal transporters (Scheer and Wolf, 2014; Durmus et al, 2016; Kinzi et al, 2024b).

Studies in experimental animals have demonstrated that the microclimate at the surface of different cell types involved in drug disposition varies between different organs. The apical pole of enterocytes is acidic owing to the acidic intestinal microclimate (Daniel et al, 1989). The same holds for the surface of the basolateral membrane of hepatocytes facing the space of Disse, which is slightly acidic (Ichikawa et al, 1994). Parenthetically, the apical membrane of kidney tubular cells faces a variable pH depending on the total acid load of the body.

For the liver, studies with isolated perfused organs from animals are routine and have been a valuable tool for the development of concepts and identification of individual processes of solute transport (Tirona et al, 1998). An early example of the value of the isolated perfused liver is the demonstration that this organ is the major site of albumin biosynthesis and secretion (Miller et al, 1951). Short-term experiments with human livers, for example, studying microcirculation and its alteration in liver disease, have been performed (Villeneuve et al, 1996a,b). Today, it is possible to maintain an ex situ perfused human liver and keep it alive and functional for several days (Eshmuminov et al, 2020; Eshmuminov and Clavien, 2023) and to study, for example, bile formation (Eshmuminov et al, 2021). Livers used for isolated perfusion experiments are mostly organs that are not suitable for transplantation. Because such livers may recover and can be transplanted (Clavien et al, 2022), it would be ethically questionable to use such organs for transport studies.

The application of human hepatocytes in drug development as well as, for example, in transport studies is routine (Li, 2007; 2008; Ramboer et al, 2013), but costly and human hepatocytes typically are scarce compared with isolated rodent hepatocytes. Hepatocytes can be used for transport experiments as well as for the determination of intracellular drug concentrations in suspension or cultured in a monolayer configuration. It should be kept in mind that primary cultured hepatocytes change the expression of transport proteins in a species-dependent manner (Swift et al, 2010; Godoy et al, 2013). Hepatocytes in suspension can be cryopreserved before use. This may, however, lead to altered transporter expression compared with native tissue as well as altered subcellular expression of transporters, which in turn affects clearance predictions (Kimoto et al, 2012; Lundquist et al, 2014). Instead of culturing hepatocytes in a single monolayer on dishes, hepatocytes can also be cultured in a so-called sandwich culture system, culturing the cells between layers of extracellular matrix (De Bruyn et al, 2013; Yang et al, 2016). This configuration allows us to determine the canalicular secretion of substrates transported by hepatocytes (Swift et al, 2010). In addition to changes in protein expression over time, typical culture systems are not assessed as functionally coordinated systems as they lack the 3-dimensional architecture with the different cell types of the liver. Culture systems arranged in microscale architectures are a possible solution to this issue (Khetani and Bhatia, 2008; March et al, 2015; Youhanna et al, 2022). Consequently, functional characterization of isolated transport systems is typically performed in heterologous expression systems.

There are numerous different heterologous expression systems using nonpolar as well as polarized cell models (Khan, 2013; Volpe, 2024). The latter can be used to simulate transepithelial cellular fluxes of substrates. To study transporters of the SLC superfamily, HEK293 cells (Thomas and Smart, 2005), CHO cells (Tihanyi and Nyitray, 2020; Pulix et al, 2021), and COS-1 cells (Wong et al, 1995) are widely used. These systems also allow comparisons of transport characteristics of genetic variants of transporters (Nakanishi and Tamai, 2012; Gong and Kim, 2013; Nies et al, 2013; Schulte and Ho, 2019). For the investigation of transcellular transport, MDCK and LLC-PK1 cell lines are used in a monolayer configuration or a Transwell setup (Youhanna and Lauschke, 2021). These cell lines, to some extent, mimic epithelial barriers in culture. Heterologous expression systems are used with transiently expressed transporters or as stably transfected cell lines. There are many possible approaches for transport measurement in heterologous expression systems (Dvorak et al, 2021). As OATPs and other SLC transporters relevant to drug disposition have commonly overlapping substrate specificities (Kalliokoski and Niemi, 2009; Karlgren et al, 2012; Brouwer et al, 2013), selection of probe substrates for in vitro DDI studies needs to be carefully considered. Numerous reviews have listed kinetic parameters for OATP-mediated drug transport determined in vitro in heterologous expression systems (Hagenbuch and Gui, 2008; Fahrmayr et al, 2010; Niemi et al, 2011; Emami Riedmaier et al, 2012; Grandvuinet et al, 2012; Nakanishi and Tamai, 2012; Gong and Kim, 2013; Kinzi et al, 2021).

Transporters can be assessed in subcellular organelles or vesicles. The latter are typically isolated as membrane vesicles because transport studies require 2 different compartments separated by a semipermeable barrier. Vesicles can be isolated from organs or cells overexpressing transport proteins. Vesicles allow experimental control of the conditions on both sides of membranes (Murer et al, 1984). Using vesicles isolated from heterologous expression systems is common for efflux systems like ABC transporters but, to our knowledge, is not used for OATPs. Finally, transporters can be studied as individual proteins. They can be purified by classic biochemical methods from organisms and reconstituted into liposomes, or after cloning, expression, and purification, proteins can be incorporated into liposomes of controlled lipid composition.

The abovementioned examples of the microenvironment of transporters and of different experimental systems illustrate that different systems and experimental conditions should be considered depending on the aim of the planned experiments to determine the functional properties of OATPs and other transport systems. Comparing the parameters obtained in these different systems is not straightforward, but because the K_m value of a transporter for a given substrate is an inherent property of that transporter, and such information is often available, we focused in this review on K_m values.

Tables 8 and 9 summarize a selection of affinities of OATPs to various substrates, for both endogenous compounds and xenobiotics and drugs. These tables includes in vivo data from rat studies with indocyanine green and BSP because the first OATP was identified using BSP as the lead substrate and because BSP, as outlined earlier, was used as a biomarker for liver function tests in humans (Sakka, 2007; Stieger et al, 2012) but was later replaced by indocyanine green owing to safety concerns. Comparing the affinity determined in in vivo models, the K_m value of ratOATP1A1 for BSP is comparable. However, it must be kept in mind that more than 1 OATP is expressed in hepatocytes at the basolateral membrane, the exact isoform being species specific. Consequently, extrapolation of kinetic parameters of individual OATPs to the in vivo situation must be done with caution. The examples listed in Tables 8 and 9 show that published K_m values of substrates vary widely for the same substrate and transporter. While this is not new, as it has been shown, for example, for P-glycoprotein (Bentz et al, 2013) as well as for OATP1B1 (Vaidyanathan et al, 2016), this situation is clearly unsatisfactory. In particular, it can be assumed that different binding sites for substrates are an inherent property of a given transporter and, as such, should be observed in all different experimental setups and systems.

Table 8.

K_m values of endogenous transporters determined in different experimental systems

Transporter Substrate	Perfused Rat Liver	Primary Cultured Rat Hepatocytes	Suspended Primary Rat Hepatocytes	Rat blLPM
Indocyanine green	17.0 μM femalea 23.7 μM malea
BSP	7.9 μMb	0.28 μM (ratio BSA:BSP 18.4:1)c 3.8 μM male; 6.1 μM femaled 0.31 μM (ratio BSA:BSP 18:1)e	7 μMf 6.5 μMg 6.2 μMh 22 μMi 88 nM (600 μM BSA)j 7.1 μMk 80 nM (600 μM BSA)k 3.42 μMl	53.1 μMm 1150 μMm 5.2 μM (electrogenic)n 20 μM (electroneutral)n

^aOtt, 1995.

^bPaumgartner, 1975.

^cWolkoff et al, 1987.

^dPersico et al, 1988.

^eMin et al, 1991.

^fSchwenk et al, 1976.

^gLaperche et al, 1981.

^hStremmel and Berk, 1986.

ⁱPotter et al, 1994.

^jSorrentino et al, 1994.

^kSorrentino and Bartoli, 1996.

^lShitara et al, 2009.

^mPotter et al, 1987.

ⁿTorres et al, 1993.

Table 9.

K_m values of recombinant transporters determined in different cell lines

Transporter Substrate	HeLa Stably Transfected	Xenopus laevis Oocytes	HEK 293 Stably Transfected	CHO Stably Transfected	MDCKII Stably Transfected
rOATP1A1
BSP	3.3 μMa	1.5 μMb 1.5 μMc
OATP1B1
BSP		0.4 μMd 0.3 μMe	0.28 μMd 0.14 μMf		2.4 μMg
Estrone-3-sulfate	94 nM and 5.34 μMi	67.5 nM and 7.00 μMh	0.46 μMi 0.286 μMd 12.5 μMf 0.38 μM and 36.1 μMj,k 0.12 μM and 10.6 μMj,l 0.16 μMm 59.3 nMn	0.23 μM and 45 μMo
Estradiol-17β-glucuronide		3.71 μMh	8.29 μMi 8.17 μMd 8.2 μMp 5.76 μMm
Cholate		13.6 μMq	11.4 μMf 47.1 μMr
Taurocholate			33.9 μMs,t 10.0 μMf 21.3 μMj,k 8.52 μMj,l 2.6 μMr
Fluvastatin			4.80 μMu	2.5 μMo	2.4 μMg
Bosentan			4.27 μMu	44 μMv
OATP1B3
BSP		3.3 μMw 0.4 μMe	3.3 μMf		2.2 μMg
Estradiol-17β-glucuronide		5.4 μMw	24.5 μMi 5.4 μMf
Cholate		41.8 μMx	42.2 μMr
Taurocholate		42.2 μMx	9.5 μMr
Fluvastatin					7.0 μMg
OATP2B1
BSP		0.7 μMe			3.4 μMg
Estrone-3-sulfate		0.10 and 29.9 μMy 6.3 μMe	9.04 μMg 8.09 μMz 10.2 μMo 20.9 μMaa
Fluvastatin				0.8 μMo	0.7 μMg

^aShi et al, 1995.

^bJacquemin et al, 1994.

^cKullak-Ublick et al, 1994.

^dIzumi et al, 2013.

^eKullak-Ublick et al, 2001.

^fCui et al, 2001a.

^gKopplow et al, 2005.

^hTamai et al, 2001.

ⁱHirano et al, 2004.

^jTransiently transfected.

^kHuang et al, 2013.

^lHong et al, 2015b.

^mMiyagawa et al, 2009.

ⁿShitara et al, 2012.

^oNoe et al, 2007.

^pKonig et al, 2000b.

^qAbe et al, 1999.

^rSuga et al, 2017.

^sTransiently transfected in HEK293-EBNA cells.

^tHsiang et al, 1999.

^uIzumi et al, 2015.

^vTreiber et al, 2007.

^wKonig et al, 2000a.

^xBriz et al, 2006.

^yShirasaka et al, 2012.

^zNozawa et al, 2004.

^aaHirano et al, 2006.

There are several possible reasons for this observed variability of transporter parameters. First, the exact experimental conditions used differ between the selected publications. In general, the determination of kinetic parameters of transport systems has to be performed under so-called initial uptake rates, whereby correction for unspecific binding, passive diffusion, and endogenous additional transport systems is needed (Murer et al, 1984; Brouwer et al, 2013; Zamek-Gliszczynski et al, 2013). Second, the experimental systems used include a variety of different cell lines. All expression systems used express a set of endogenous transporters (Shneider and Moyer, 1993; Goh et al, 2002; Hilgendorf et al, 2007; Ahlin et al, 2009). Transfection of cell lines may alter the expression of endogenous transporters in comparison with the wild-type cells (Ahlin et al, 2009). Third, data may be influenced by differences in unstirred water layers in the different systems (Thomson and Dietschy, 1977). Fourth, the addition of proteins to incubation buffers in case of poorly water-soluble compounds will lead to lower free, unbound concentrations, which necessarily need to be determined. Typically, in vitro transport experiments are performed in buffers and in the absence of plasma, which, in vivo, is the solute to which transporters are exposed at the extracellular side. A comparison of kinetic parameters of OATP1B1 and OATP1B3 expressed in HEK293 cells has shown a significant impact of the presence of plasma compared with standard buffers for both K_m as well as V_max values for model drugs (Bowman et al, 2020). Other studies have also found differences in transport in the absence and presence of plasma (Katneni et al, 2018; Bi et al, 2021). Fifth, the specific lipid composition of the plasma membrane is cell-type specific and may change under physiologic and pathophysiologic conditions (Harayama and Riezman, 2018). Transporters are membrane proteins spanning the residing membrane multiple times. Thus, membrane lipids may interact with membrane proteins, such as transporters, as tightly bound or annular lipids or at a distance (Brown, 2017). The effects of the lipid composition and of altering the membrane lipid composition on transporters have been studied both for uptake and efflux transporters. These findings have been summarized recently (Stieger et al, 2021). Taken together, the experimental system chosen for their investigation considerably influences some of the functional properties of OATPs (as well as of other transporters).

VI. Structural insight into OATP-mediated transport

The OATP transporters feature the fold and transmembrane topology of members of the major facilitator superfamily, which is characterized by 12 transmembrane helices, cytoplasmic amino and carboxy terminals, and a pseudosymmetry of the N-terminal and C-terminal bundle of 6 transmembrane domain (TM) helices (Fig. 4). Unlike other major facilitator superfamily proteins, OATPs contain extensive extracellular loops (ECLs) and a Kazal-like domain in ECL5. The function of this domain has remained unclear. Several disulfide bonds are predicted to covalently link external loops or secondary structure motifs of the Kazal domain. However, not all these disulfides are visible in experimental structures. Another unique feature of OATP transporters is the presence of a signature motif located at the external half of TM6 and encompassing part of the ECL3. This motif contains the sequence D-x-RW-(I/V)-GAWW-x-G-(F/L)-L, where x denotes any amino acid residue.

Fig. 4.

(A) OATP1B1 topology and conformations. The TM helices are numbered and arranged to highlight the pseudo–2-fold rotation symmetry (indicated by red dashed line and ellipse). The coloring is in rainbow from blue (N-terminus) to red (C-terminus). Extracellular loops (ECLs) are numbered and N-glycans indicated schematically, with the precise glycan unknown and possibly heterogeneous. (B) Ribbon diagrams of experimental OATP1B1 structures in outward-open (left) and inward-open (right) conformations. The coloring matches that of the topology diagram in (A).

Two groups have reported overexpression and purification of human OATP1B1 for structural studies. One group (Shan et al, 2023) used transient transfection of Expi293F cells, whereas the other group generated stable expression in Flp-In T-REx293 cells (Ciuta et al, 2023). The results of the structural studies are similar, but the transient transfection does not appear to lead to posttranslational modification of the protein. Human OATP1B3 was also overexpressed and purified from stably expressing cells.

At present, experimental structures of human OATP1B1 in distinct conformations and 1 structure of human OATP1B3 have been reported. All structures were determined using single particle cryoelectron microscopy and were published in 2023 by 2 research groups. One group reported structures of OATP1B1 determined in detergent solution and without protein binders but in 2 conformations and bound to distinct compounds (Shan et al, 2023). The other group used synthetic Fab fragments to determine the structures of nanodisc-reconstituted proteins. They reported one structure of E1S-bound OATP1B1 and another of OATP1B3 without bound drug but with evidence of bound bicarbonate (Ciuta et al, 2023).

Structures of OATP1B1 were captured in outward-open and inward-open conformations. These distinct states reveal access to drug-binding pockets from the cytoplasmic side of the membrane or from the extracellular side, which corresponds to the space of Disse. The published structures revealed 2 adjacent pockets. A central pocket is located at the interface of the 2 helical bundles, the N-terminal bundle (N-half) and the C-terminal bundle (C-half) of the transporter. A second pocket is located within the C-half, where a cavity exists with an apex suitable for binding negatively charged groups (Fig. 5).

Fig. 5.

Surface representation of experimental OATP1B1 structures in outward-open (left, PDB 8K6L) and inward-open (right, PDB 8PHW) conformations. Two cavities are apparent: a central cavity is located between the 2 halves of the transporter (N-half containing TM1–TM6 and C-half containing TM7–TM12). A side cavity is located in the C-half of the protein. While the central cavity is sufficiently large for drug substrates in both conformations, the side cavity is only properly formed in the inward-open conformation. The PDB codes of the shown structures are indicated in the legend. The figures were newly drawn, and the PDB ID code serves as a reference.

The structural data provided insight into the drug-binding pockets of OATP proteins and into a possible coupling mechanism to countertransport bicarbonate. Both studies reported an inward-facing OATP1B1 structure with a single molecule of E1S bound in a side pocket formed by helices of the C-half. Residues interacting with the negatively charged sulfate group of E1S include R633 (from TM12), which carries a positive charge and probably contributes to the selectivity of OATP1B1 for negatively charged substrates. In addition, the hydroxyl groups of 2 tyrosine residues (Y422 and Y425, both from TM9) also contact the sulfate group. The polycyclic scaffold of E1S is wedged between several hydrophobic residues, some of which had previously been identified by mutagenesis as important contributors to the substrate specificity (eg, L545) (Gui and Hagenbuch, 2009). The study by Shan et al (2023) also reported structures of OATP1B1 bound to bilirubin, simeprevir (a hepatitis C medication), and 2',7'-dichlorofluorescein. Unlike the E1S-bound state, these structures showed outward-open OATP1B1 conformations, and the bound compounds were either located in the central drug-binding pocket (bilirubin and simeprevir) or partly in the central and the side binding pockets (2',7'-dichlorofluorescein). It should be noted that the experimental electron microscopy density for these 3 compounds is less clear than that of bound E1S. Hence, the reported drug poses may represent only one of the several possible binding modes (Shan et al, 2023). This may be a consequence of higher substrate mobility in the central binding pocket compared with that in the narrower side pocket.

The study by Ciuta et al (2023) reported the structure of OATP1B3 in a substrate-free state and an inward-open conformation. Near the signature motif, the structure revealed a small ligand that was concluded to be bound bicarbonate (HCO₃⁻). The ion was wedged in a pocket formed by the 2 tryptophan side chains W258 and W259 of the signature motif, an arginine residue (R58) of TM1, and a histidine residue (H115) of TM3 (Fig. 6). Given that the histidine residue is only present in OATP proteins that have been demonstrated or suspected to be dependent on pH in their transport activity, it was inferred that the role of this residue was to shuttle protons to bound bicarbonate during the transport cycle (Ciuta et al, 2023). However, no experimental evidence for such proton transfer exists at present.

Fig. 6.

Ribbon representation of OATP1B3 structure (PDB 8PG0), with signature motif colored purple. (A) Overview with a bound bicarbonate ion shown as spheres. (B) Close-up view with side chains forming the bicarbonate pocket shown as sticks. The PDB code of the shown structure is indicated in the legend. The figures were newly drawn, and the PDB ID code serves as a reference.

VII. Transport mechanism and inhibition of OATPs by small-molecule compounds

Solute transport processes can be equilibrative or concentrative (Preusch, 2007). Equilibrative transport is defined as protein-mediated transport of a solute from the outside to the inside of a cell (cytoplasm) along its electrochemical concentration gradient, whereas concentrative refers to transport against an electrochemical gradient in the same direction.

Equilibrative transport uses the energy of the (electro)chemical gradient of a solute over the barrier to be crossed. In this situation, the flux is proportional to the concentration gradient and follows Fick’s first law (Fick, 1855). Equilibrative transport, therefore, requires a low and unbound (free) intracellular concentration of the substrate of a transporter. Unfortunately, free concentrations of OATP substrates are typically not known. As drugs transported by OATPs tend to be hydrophobic, the binding of drugs to plasma proteins will influence drug disposition and transport rates of the respective transporters for the cellular uptake (Weisiger, 1985; Wanat, 2020; Celestin and Musteata, 2021; Schulz et al, 2023). As typical OATP substrates are anions, transport kinetics may be affected by changes in membrane potential and/or surface charge density (Barts and Borstpauwels, 1985). An additional parameter influencing transport rates is the unstirred water layer (Korjamo et al, 2009). Given these uncertainties both at the level of transport theory and at the exact environment near the binding site(s) of a transporter, a detailed understanding of the molecular transport mechanism of OATPs is currently elusive.

As outlined earlier, OATPs transport predominantly anionic compounds such as bilirubin or the prototypic substrate BSP. Some OATPs also accept uncharged molecules or even cationic compounds as substrates. Publications of acid dissociation constant (pK_a) values of bilirubin list conflicting data ranging from smaller than 5.0 (Vega-Hissi et al, 2013) to higher than 8.0 (Mukerjee and Ostrow, 2010). Hence, bilirubin may be transported with 2 negative charges. The lead substrate that was used to clone the first OATP—BSP—carries 2 negative charges. Hepatocytes have an inside negative membrane potential of about −35 mV (Boyer et al, 1992). Thus, transporting negative charges into hepatocytes is energetically unfavorable unless the out-to-in concentration gradient remains steep. This may be the case for bilirubin, which is rapidly conjugated for canalicular excretion. While the concentration of free intracellular bilirubin is not known, it was estimated to be 70 nM in rat hepatocytes compared with 0.16 nM in plasma (resulting in uptake against a steep concentration gradient) (Levitt and Levitt, 2014). The outside of cells expressing OATPs is often the interstitial space, but it can also be the blood (eg, at the blood-brain barrier), the space of Disse (in the liver), or the outside of the body like in the intestinal lumen. In contrast to the cytoplasm, extracellular compartments do not have a constant pH but may be acidic or alkaline, depending on physiologic states, and thus, the charge of a substrate and the protonation state of the OATP might be modified.

A. Counterions

It is generally accepted that OATPs are sodium-independent transporters and work as exchangers, with the imported substrates being exchanged against another anion. It should be pointed out that some studies reported a partial sodium dependence of OATP1A2-mediated E1S (Bossuyt et al, 1996b) and dehydroepiandrosterone sulfate uptake (Kullak-Ublick et al, 1998) and of rOATP1B2-mediated phalloidin transport (Meier-Abt et al, 2004). However, neither of these studies excluded an indirect coupling to the Na⁺/H⁺ exchanger under sodium conditions. In the following, we will review studies aiming to identify anionic counterions involved in OATP-mediated transport reactions. The first counterion identified was bicarbonate (Satlin et al, 1997). Using HEK293 cells stably transfected with rOATP1A1 and an alkali-loading protocol, the authors observed a rapid intracellular acidification accompanying taurocholate uptake. Another counterion reported to be involved in OATP transport reactions is glutathione (GSH). Injecting GSH into X laevis oocytes expressing rOATP1A1 resulted in enhanced rOATP1A1-mediated uptake of taurocholate and leukotriene C4 (Li et al, 1998). Extending this study to rOATP1A4 expressed in X laevis oocytes, GSH, S-methylglutathione, S-sulfobromophthalein-glutathione, S-dinitrophenyl glutathione, and ophthalmic acid were identified as counterions stimulating the uptake of taurocholate (Li et al, 2000). In addition, the uptake of the cardiac glycoside digoxin was also stimulated by preloading with S-dinitrophenyl glutathione. Interestingly, S-dinitrophenyl glutathione, in contrast to GSH, did not stimulate rOATP1A1-mediated taurocholate uptake. In membrane vesicles, isolated with defined sidedness from HeLa cells stably expressing rOATP1A1, asymmetric GSH transport was demonstrated with a higher uptake rate for inside-out compared with right side–out vesicles. This suggests a higher efflux than uptake rate for GSH (Mittur et al, 2002). The observation that the uptake of E17βG and of conjugated or unconjugated cholate into X laevis oocytes expressing OATP1B3 (but not OATP1B1) was stimulated by the extracellular addition of GSH or GSSG (oxidized GSH) (Briz et al, 2006) could not be confirmed in a later study (Mahagita et al, 2007). It should also be pointed out that no data were presented, demonstrating a formal cotransport of GSH and bile salts. Rather, reciprocal stimulation of GSH and bile salts was shown (Briz et al, 2006). Similarly, OATP2B1 was not activated by the addition of extracellular GSH (Nozawa et al, 2004). Under physiologic conditions, a steep GSH gradient exists between cells (high concentrations) and the surrounding plasma (Ookhtens and Kaplowitz, 1998), which makes an inward cotransport with OATP1B3 substrates energetically unfavorable but would favor an exchange mechanism. Under physiologic conditions, the cytoplasmic concentration of free bile salts in hepatocytes is likely <1 μM (Weinman and Maglova, 1994).

B. Dependence of transport on external pH

Several OATP proteins show pH dependency in their transport reactions. For example, OATP2B1-mediated uptake increases in a pH-dependent manner by lowering extracellular pH (Kobayashi et al, 2003; Nozawa et al, 2004). OATP2B1 shows a broad tissue distribution and is also expressed at the apical membrane of enterocytes in the intestine (Kobayashi et al, 2003), where the apical surface of these enterocytes represents an acidic microenvironment (Daniel et al, 1989; Sanderson, 1999). OATP2B1 transport kinetics are pH-sensitive. For example, a 3.2-fold increase in the uptake of E1S was observed at pH 5.0 compared with that seen at pH 7.4. The microclimate in the space of Disse facing the hepatocellular OATP1B1, OATP1B3, and OATP2B1 is also an acidic microenvironment (Ichikawa et al, 1994). The pH dependence of 13 rat and human OATPs expressed in X laevis oocytes or mammalian cells was studied. Except for OATP1C1, all transporters were stimulated by an acidic extracellular pH (Leuthold et al, 2009). Importantly, the level of pH stimulation was dependent on the substrate under investigation. For example, prostaglandin E₂ transport by OATP2B1 was not stimulated by lowering the extracellular pH, whereas the transport of thyroxine was. OATP2B1-mediated pemetrexed transport was strongly stimulated by a lower extracellular pH, whereas the transport of bromosulfophthalein was pH insensitive (Visentin et al, 2012). The effect of lowering extracellular pH on statin transport by OATP2B1 varied with the statin tested (Varma et al, 2011). A kinetic analysis of various OATPs showed that the transport stimulation by a low extracellular pH was due to an increase of the affinity (a decrease in K_m) for the substrate at low pH, while V_max remained unchanged (Leuthold et al, 2009). This was used to argue against a mechanism whereby protons act as cosubstrates in a symport mechanism. A detailed investigation of changing the magnitude of the pH gradient on the stimulation of rOATP1A1-mediated uptake found no correlation of transport stimulation with the magnitude of the pH gradient, which was interpreted as a modification of the protonation state of the intracellular part of the transporter (Marin et al, 2003).

C. Substrate-binding sites and conformational changes during the transport cycle

Given that OATPs are major facilitators, they likely adopt conformations and undergo conformation changes during the transport cycle, similar to other, well-studied members of this protein family (Drew et al, 2021). These conformations include, at the minimum, an outward-open, an inward-open, and an occluded conformation. The conversion of these conformations is made possible by the rigid-body motion of the 2 helical bundles (N-terminal bundle and C-half). This has been referred to as the “rocker switch” mechanism, which would expose the drug-binding pockets alternatingly to the cytoplasm or the external medium (Sauve et al, 2023).

Among the intriguing features of OATPs is that they display more than 1 binding site with differing affinities to certain substrates. For example, OATP1B1 expressed in HEK293 cells displays 2 apparent K_m values (67.5 nM and 7.0 μM) for E1S, which was interpreted as a low-affinity and a high-affinity binding site (Tamai et al, 2001). The 2 different binding sites of OATP1B1 were also observed when the transporter was expressed in X laevis oocytes (Tamai et al, 2001) or upon expression in CHO cells (K_m values of 0.23 and 45 μM, respectively) (Noe et al, 2007) or HEK293 cells (0.22 and 312 μM, respectively) (Gui and Hagenbuch, 2009). In contrast, OATP1B1 showed classic Michaelis-Menten kinetics with a single binding site for E17βG, and unlike OATP1B1, OATP2B1 displayed a single binding site for E1S. Multiple binding sites are not restricted to human OATPs. Bovine OATP1A2 expressed in HEK293 cells also has 2 K_m values of 0.25 and 46.6 μM for E1S transport, but apparently only 1 binding site for taurocholate (Liu et al, 2013).

Unlike highly specific transporters such as glucose transporters, the drug-binding pockets of OATP1B1 may change in size and shape during the conversion to the occluded conformation. This has profound implications for our understanding of specificity. The 2 drug-binding cavities observed in the structure of OATP1B1 are in line with, and can indeed rationalize, earlier conclusions that OATP1B1 has more than 1 drug-binding pocket and that transport of 1 substrate can be stimulated by the presence of another substrate (Stieger and Hagenbuch, 2014) because the central pocket appears to provide sufficient space for 2 drug molecules. However, it is possible that the shape and size of the drug-binding cavity changes as OATP1B1 switches from an outward-open to an occluded conformation, where less space and distinct surface properties of the cavity might exist. This can influence the binding of substrates and inhibitors. The situation is reminiscent of the well-studied multidrug transporter P-glycoprotein (ABCB1), where an occluded conformation reveals a much smaller volume than the drug-accepting, inward-open conformation (Alam et al, 2019; Nosol et al, 2020). Hence, while the drug-binding pockets and cavities revealed in the OATP1B1 and OATP1B3 structures can rationalize how the protein accepts 1 or more structurally distinct substrates, a key element to understanding the molecular basis of the multidrug specificity, DDIs, and inhibition, is to obtain experimental structural data of a drug-bound, occluded conformation at sufficiently high resolution.

There are additional open questions. One insufficiently clarified aspect is whether a driving force and a possible counter-ion are required for OATP1B1-mediated transport. The presence of a bound bicarbonate molecule in the structure of OATP1B3, while suggestive, is not direct evidence of HCO₃⁻ serving as a strictly coupled counterion, in part because its binding site is remote from the drug-binding pockets. It is conceivable that bicarbonate may have a regulatory role in transport. Furthermore, it is unclear whether OATP1B1-mediated transport is concentrative. Given that the substrates are either metabolized in the hepatocyte or transported into bile, it is not clear whether substrates are transported against a concentration gradient. To address these questions, future experimental studies in vitro, particularly proteoliposome-based transport assays using purified OATP1B1 protein, will be required.

Considering the mechanistic uncertainties, it is conceivable that OATP proteins have multiple modes of action. Depending on the number of substrates moved in 1 cycle, transport might be electrogenic or electroneutral, and more than 1 counterion might be involved. As was found for other drug transporters, the slow rate of these proteins makes experimental investigations challenging. Nevertheless, the ability to functionally purify them is a promising step toward a more complete mechanistic understanding.

D. Modulation of transport

Multidrug transporters can be inhibited or stimulated by a wide variety of compounds, and the function of OATPs is indeed modulated by compounds that are either substrates themselves or act as inhibitors that are not translocated. In the following, we summarize some of the observations reported in the literature. Human OATP4C1 transports 3,5,3'-triiodo-L-thyronine (T3) and digoxin. Interestingly, T3 does not inhibit digoxin transport, and digoxin is a very poor inhibitor of T3 transport, suggesting distinct binding sites or translocation pathways for the 2 compounds (Mikkaichi et al, 2004). The same group later reported partial reciprocal inhibition of digoxin and E1S (Yamaguchi et al, 2010). The observation that rOATP1A4-mediated transport of taurocholate can be stimulated by the coadministration of E17βG but not the transport of digoxin may suggest distinct binding sites on this transporter (Sugiyama et al, 2002). However, these observations should be interpreted with caution because OATP2B1 expressed in X laevis oocytes displays 2 distinct binding sites for E1S (Shirasaka et al, 2012) but only 1 binding site if expressed in HEK293 cells (Tamai et al, 2001). OATP2B1-mediated transport of dehydroepiandrosterone sulfate and E1S is stimulated by the simultaneous presence of prostaglandin A₁ or A₂ at the extracellular side of the transporter (Pizzagalli et al, 2003). A later study found stimulation of OATP2B1-mediated E17βG transport by testosterone in HEK293 cells (Karlgren et al, 2012). It should be noted that these 2 findings do not allow an unambiguous conclusion because 2 different cell lines and 2 different substrates were used. Using cell lines expressing OATP2B1 and E1S or dehydroepiandrosterone sulfate as substrates, differing inhibition data were observed for some inhibitors (Grube et al, 2006): Both substrates were stimulated by androsterone, hydroxyprogesterone, and pregnenolone. Estrone inhibited E1S transport but stimulated dehydroepiandrosterone sulfate, which was also observed for β-estradiol. Progesterone stimulated the transport of both substrates at low concentrations, an effect that was attenuated at high concentrations (Grube et al, 2006). Inhibition experiments with dietary or herbal components also have different effects on OATPs: OATP1B1-mediated dehydroepiandrosterone sulfate transport is stimulated by rutin but inhibited by the 2 other flavonoids, biochanin A and luteolin (Wang et al, 2005). While E1S transport by OATP1B3 was stimulated by the external addition of epigallocatechin gallate, E1S transport by OATP1A2, OATP1B1, and OATP2B1 was inhibited by the same compound (Roth et al, 2011). The interaction of the milk thistle compound silibinin with OATP1B1 is complex: the low-affinity binding site for E1S is stimulated at low concentrations of silibinin but inhibited at high concentrations. Thus, silibinin appears to be an inhibitor of the high-affinity binding site (Wlcek et al, 2013). Among drugs, clotrimazole stimulates OATP1B3-mediated E17βG transport (Gui et al, 2008) and fendiline stimulates OATP2B1-mediated E1S transport (Karlgren et al, 2012). These examples of modulators of the transport activity of OATP clearly demonstrate a very complex interaction of compounds with OATPs. They do suggest, but not prove, that OATPs may have more than 1 substrate and/or inhibitor binding site as well as sites for allosteric modulation.

E. Modeling drug binding

In the absence of experimental structural data, it is tempting to use structure prediction software combined with computational ligand docking to explore the interaction of OATP proteins with suspected or proven substrates, modulators, and inhibitors. Prediction software, most prominently AlphaFold, has become very successful at predicting the structures of most proteins. The predicted structures of the human OATP proteins (Jumper et al, 2021) are shown in Fig. 7. Notably, AlphaFold does not predict the same conformation of the 11 OATP members. Rather, the proteins display distinct conformations, including inward-open, occluded, outward-open, and intermediate conformations.

Fig. 7.

AlphaFold (Jumper et al, 2021) models of human OATP proteins in ribbon representation and colored in rainbow (blue to red from N-terminus to C-terminus). The predictions feature distinct conformations, which are summarized under the models. The figures were newly drawn from publicly available PDB codes of the indicated proteins.

Docking programs may be used to evaluate the binding of compounds to these predicted structures, and many recent publications have used such approaches. However, there are multiple observations that highlight the limitations of this approach for OATP proteins: first, the structures of OATP1B1 and OATP1B3, as predicted by AlphaFold, would not allow the experimentally observed ligands to be docked without conformational changes, because clashes between side chains and bound E1S or bicarbonate exist. This suggests that small conformational changes not only in the binding pockets but, possibly, also further away in the protein scaffold occur with drug binding, complicating the computational approaches. Second, even with experimental structural data, ligand docking in combination with molecular dynamics (MD) simulation can be challenging and yield misleading results. Both the study by Shan et al (2023) and that by Ciuta et al (2023) used MD simulation combined with docking to evaluate the binding of substrates to OATP1B1, including the binding of statins. However, while the results can provide general indications of possible binding modes, the exact binding poses are ambiguous, and alternative poses may be possible. Third, given that an occluded conformation is likely of key importance during the transport cycle, an accurate structural model of this conformation is required for MD simulation and docking of ligands. The occluded conformation of active transport proteins is generally believed to represent the state in which the specificity of the transport process is established. Hence, experimental structural data revealing an occluded conformation are likely required before reliable docking can be pursued.

The limitations of protein structure prediction software combined with ligand docking algorithms in drug discovery or even just for the understanding of drug–protein interactions have been outlined in recent reviews and commentaries (Karelina et al, 2023; Scardino et al, 2023; Schapira et al, 2024; Terwilliger et al, 2024) (https://www.science.org/content/blog-post/docking-alphafold-structures-oops). It should be mentioned, however, that progress with prediction software may, in the future, allow reasonably accurate structural predictions of how ligands, drugs, modulators, and so on might interact with protein targets. A recent release of AlphaFold 3 claims to make progress in this regard (Abramson et al, 2024). However, at present, it is too early to evaluate whether algorithms can reliably predict how small-molecule compounds interact with drug transporters, and experimental structural studies are required to reveal the basis of functionally relevant drug-OATP interactions.

VIII. Regulation of expression

The expression of OATPs is regulated at the transcriptional, posttranscriptional, and posttranslational levels (Hagenbuch and Stieger, 2013; Brouwer et al, 2022; Nies et al, 2022). At the transcriptional level, several nuclear receptors, including the aryl hydrocarbon receptor, the constitutive androstane receptor (CAR), the farnesoid X receptor (FXR), the hepatocyte nuclear factor (HNF) 4α, the liver X receptor α, the pregnane X receptor (PXR), and the small heterodimer partner (SHP) are involved in regulating OATPs.

A. Transcriptional regulation

Early studies demonstrated that besides nuclear receptors, the liver-enriched HNF1α controlled the expression of the liver-specific SLCO1B1 and SLCO1B3 genes (Jung et al, 2001). The same group also demonstrated that the promoter of SLCO1B3 was transactivated by FXR (Jung et al, 2002). In contrast, FXR repressed the SLCO1B1 promoter indirectly by inducing the expression of SHP, which leads to the decreased activation of HNF1α via HNF4α (Jung and Kullak-Ublick, 2003). However, Meyer zu Schwabedissen et al (2010) reported that OATP1B1 expression was stimulated in Huh-7 cells and primary human hepatocytes by the FXR agonist chenodeoxycholic acid. Furthermore, this report also demonstrated the activation of OATP1B1 expression by liver X receptor α. Activation of PXR and CAR had no effect on OATP1B1 expression. These findings are supported and extended by Niu et al (2019), demonstrating that in sandwich-cultured human hepatocytes, PXR did not induce the expression of SLCO1B1, SLCO1B3, SLCO2B1, and SLCO4C1 mRNA. Moscovitz et al (2018) evaluated the selective induction of drug-metabolizing enzymes and transporters in sandwich-cultured cryopreserved human hepatocytes. They found that the PXR activators rifampin and PF-06282999 increased SLCO1B1 and SLCO2B1 mRNA about 3-fold. The CAR agonist CITCO [6-(4-chlorophenyl)imidazo(2,1-b)(1,3)thiazole-5-carbaldehyde O-(3,4-dichlorobenzyl) oxime] induced SLCO1B1 mRNA about 2-fold, as did the aryl hydrocarbon receptor agonist omeprazole. The tyrosine kinase inhibitor pazopanib, which was suggested to induce the expression of CYP3A4 presumably through activation of PXR, induced the expression of SLCO2B1 mRNA more than 8-fold and of SLCO1B1 mRNA more than 3-fold (Moscovitz et al, 2018). Treating Huh-7 and HepaRG cells with FXR activators or antagonists resulted in increased and decreased transcription of SLCO1B3-1B7 (Malagnino et al, 2019b). Knauer and colleagues (2010) reported that the promotor of the SLCO2B1 gene contains a functional binding site for HNF4α, leading to increased expression of OATP2B1 when HNF4α is activated. In a recent study, Yadav et al (2022) demonstrated that isotretinoin, a drug approved to treat severe acne and neuroblastoma, acting via binding to nuclear receptors, induces the expression of SHP. This then leads to the downregulation of SLCO1B1 mRNA in human hepatocytes (Yadav et al, 2022). Although they predicted a 55% decrease in OATP1B1 activity, there were no changes in plasma levels of coproporphyrin I in samples obtained from 3 previously conducted clinical studies. This led the researchers to conclude that SLCO1B1 mRNA was decreased significantly by isotretinoin, but this decrease did not translate to in vivo DDIs. Thus, in vitro and in vivo predictions must be cautiously taken (Yadav et al, 2022).

The expression of SLCO1A2 mRNA was increased almost 10-fold in breast cancer tissue compared with that in adjacent normal breast tissue (Meyer zu Schwabedissen et al, 2008). The PXR ligand rifampin was able to increase SLCO1A2 mRNA as well as OATP1A2 transport function. The group identified a PXR response element in the SLCO1A2 promoter and demonstrated that in cells treated with PXR siRNA, the induction of mRNA and function was abolished (Meyer zu Schwabedissen et al, 2008). This clearly established that SLCO1A2 expression can be regulated via PXR. Besides being regulated by nuclear receptors and transcription factors, the expression of OATPs can be regulated by growth factors, cytokines, and other chemicals.

B. Posttranscriptional regulation

OATPs are also regulated at the posttranscriptional level. A few studies report that microRNAs regulate OATP mRNA and protein expression. For OATP1B1, miR-206 overexpression in Huh-7 cells resulted in decreased expression of SLCO1B1 mRNA and protein measured using western blots and immunofluorescence (El Saadany et al, 2019). In addition, overexpression of miR-206 resulted in decreased uptake of E1S into the cells. Reporter assays suggested that binding of miR-206 to the 3' untranslated region of the SLCO1B1 mRNA resulted in the downregulation of mRNA and protein. In addition, SLCO1B1 mRNA and protein levels correlated negatively with miR-206 levels in liver tissues obtained from 20 patients undergoing partial hepatectomy. Similar findings that miR-206 expression decreased OATP1B1 protein expression were reported using HepG2 cells (Xu et al, 2020). However, in this study, SLCO1B1 mRNA was not affected. Studies investigating the mechanism of upregulation of OATP1B3 in prostate cancer cells by the drug abiraterone resulted in the identification of miR-579-3p binding to the 3'-untranslated region of SLCO1B3 mRNA (Barbier et al, 2021). The presence of miR579-3p on SLCO1B3 mRNA reduced OATP1B3 expression, while downregulation of miR579-3p by abiraterone resulted in increased expression of the Lt-OATP1B3. OATP2B1 is also regulated by microRNAs. Two independent studies reported that miR-24 regulates the expression of OATP2B1. Liu et al (2020) found that overexpression of miR-24 resulted in decreased expression of SLCO2B1 mRNA and protein in HEK293 cells that were transfected with an OATP2B1 expression plasmid that contained the 3'-UTR. In addition, in Caco-2 cells, the intrinsic expression of OATP2B1 decreased after overexpression of miR-24. In human livers, the expression level of miR-24, which also represses HNF4α expression, inversely correlated with the expression of SLCO2B1 mRNA and protein, and in HepaRG cells, expression of OATP2B1 and HNF4α decreased after transfection with an miR-24 precursor (Tajiri et al, 2020).

C. Posttranslational regulation

At the posttranslational level, OATPs can be regulated by glycosylation, phosphorylation, and ubiquitination, and this has recently been reviewed in detail (Lee et al, 2020b; Brouwer et al, 2022; Nies et al, 2022). N-glycosylation occurs at 3 asparagine residues in ECL2 and ECL5 of OATP1B1. When all 3 sites were mutated, the protein was retained in the endoplasmic reticulum, and transport function was markedly reduced (Yao et al, 2012). OATP1B3 has 6 N-glycosylation sites in ECL2 and ECL5. Mutations of individual sites do not affect the function or expression of OATP1B3; however, when all 6 sites are mutated or after treatment with tunicamycin, surface expression and turnover number are reduced, suggesting potential functional effects in patients with incomplete or defective glycosylation (Liang et al, 2024a). Interestingly, in human liver samples from normal subjects, OATP1B1, OATP1B3, and OATP2B1 were N-glycosylated, but N-glycosylation decreased in patients with steatosis and even more in patients with nonalcoholic steatohepatitis (Clarke et al, 2017). For OATP1A2, glycosylation at N124 and N135 is essential for plasma membrane expression and transport, while for OATP2B1, N176 and N538 are important (Kataoka et al, 2024). Phosphorylation was shown for several OATPs. When OATP1A2 was expressed in COS-7 cells, protein kinase (PK) C activation with the phorbol ester phorbol 12-myristate 13-acetate (PMA) decreased E1S uptake and cell surface expression of OATP1A2 (Zhou et al, 2011). PMA treatment increased the internalization of OATP1A2 but did not affect its recycling back to the plasma membrane. Furthermore, the authors could show that the OATP1A2 internalization occurred through clathrin-dependent but not caveolin-dependent pathways. For OATP1B1 expressed in COS-7, HEK293, and MDCK cells, the PKA inhibitor KT5720 decreased OATP1B1 membrane expression and uptake function. The PKA activator 8-bromo-cAMP stimulated OATP1B1 membrane expression and E1S uptake (Sun et al, 2008). PKC activation with the phorbol ester PMA, in contrast, reduced E1S uptake into HEK293 cells stably expressing OATP1B1. This reduced uptake was due to rapid internalization upon PMA treatment (Hong et al, 2015a). Crowe et al (2019) compared the phosphorylation status of wild-type OATP1B1 with the polymorphic OATP1B1∗5 that has a V174A mutation. Both wild-type and mutant OATP1B1 showed the same membrane expression, but OATP1B1∗5 had reduced uptake function for E17βG when stably expressed in HEK293 cells. Both proteins were phosphorylated, but OATP1B1∗5 showed increased phosphorylation. Hayden et al (2021) expressed OATP1B1 in HEK293 cells and used LC-MS/MS proteomics to demonstrate that the protein was phosphorylated at 23 different tyrosine residues. Inactivation of LYN, a kinase expressed in hepatocytes, resulted in decreased OATP1B1-mediated E17βG transport, indicating that LYN was the major kinase that phosphorylated OATP1B1. When OATP2B1-expressing MDCKII cells were treated with PMA, reduced OATP2B1-mediated E1S uptake was observed (Kock et al, 2010), and it could be demonstrated that this was due to internalization of the transporter.

The intracellular loops in OATPs between TMDs 6/7 contain lysine residues that may be subject to ubiquitination. When cotransfected with ubiquitin in HEK293 cells, OATP1B1 and OATP1B3 were ubiquitinated (Alam et al, 2017). However, the effect of this ubiquitination on OATP-mediated transport was not reported. Treatment of primary cultured rat hepatocytes with hepatocyte growth factor (HGF) leads to an increase of E17βG uptake and rOATP1A1 protein levels. At the same time, hepatocyte growth factor suppressed the ubiquitination of rOATP1A1 (Iwakiri et al, 2008). This is suggestive of ubiquitin-induced alterations of transport capacity in hepatocytes. Furthermore, immunoprecipitation of flagged OATP1B1 and OATP1B3 followed by proteomic analysis revealed the association of ubiquitin-related enzymes (Powell et al, 2023).

In addition to these more classical posttranslational modifications, for some OATPs, protein–protein interactions were shown to affect the function and expression (Zhang et al, 2020). OATP1B3 can form homo-oligomers and hetero-oligomers, whereby the individual transporter and not the homo-oligomer is sufficient for transport function (Zhang et al, 2017). The OATP1B1 sequence contains leucine heptad motifs within TMs. Mutation experiments with these domains have demonstrated their relevance both in function and in oligomerization (Ni et al, 2021). Additionally, OATPs interact with PDZK1 and PDZK2 (Stieger and Hagenbuch, 2014). For example, OATP1B1 requires interaction with PDZK1 for efficient expression at the plasma membrane (Wang et al, 2023).

Liver disease can also affect the expression level of drug transporters such as OATPs (Atilano-Roque et al, 2016; Thakkar et al, 2017; Evers et al, 2018; Chu et al, 2022; Drozdzik et al, 2022; Marin et al, 2024). Drug-induced induction of OATPs remains currently controversial (Rodrigues et al, 2020). Therefore, all these different ways of regulation can lead to increased or decreased OATP-mediated transport, and drugs or diseases that affect any of these regulation mechanisms could result in modified drug disposition.

IX. Genetic polymorphisms and their impact on drug disposition and adverse events

The SNP database (https://www.ncbi.nlm.nih.gov/snp/) contains over 900 million variants from over 200,000 subjects (ALFA Project Release 3 as of 3 August, 2023). When searching for variants of the SLCO family members, 518,505 entries can be found. However, only a limited number of these variants have been functionally characterized and shown to have a role in drug disposition. A good source for information regarding such clinically important SNPs is the NIH-funded PharmGKB (https://www.pharmgkb.org) (Davis and Long, 2001; Hewett et al, 2002). It is a searchable database that contains information on how genetic variation can affect drug response in patients. In the following sections, we will summarize the entries that seem clinically relevant.

A. Polymorphisms in the SLCO1A2 gene

For SLCO1A2, there are 23 variant annotations listed, 5 of them with some clinical relevance. Three of these 5 variants affect rocuronium disposition. The rs7967354 variant is an intron variant where the GG genotype is associated with decreased dose requirements for rocuronium (Ahlstrom et al, 2021). The same study also identified another intron variant, rs11045995, where the GG genotype is associated with lower dose requirements for rocuronium. In this study, 918 female patients who underwent breast cancer surgery were included. The rs3764043 variant, an insertion of A at position −189_−188 of the SLCO1A2 promoter, was found in 17 of 30 patients undergoing elective surgeries and was associated with reduced rocuronium dose (Costa et al, 2017). The rs4149009 variant, a miRNA-binding site polymorphism located in the 3'-UTR of SLCO1A2, was associated with delayed methotrexate elimination for the GG genotypes when 141 Chinese children with acute lymphoblastic leukemia were screened (Wang et al, 2018). Another promoter polymorphism, rs4148978, was identified a position −1032 of the SLCO1A2 promoter. The GG genotype was associated with increased imatinib clearance in 18 of the 34 patients with chronic myeloid leukemia (Yamakawa et al, 2011). In addition to these 5 variants with clinical relevance, numerous polymorphisms and mutants have been functionally characterized, including base changes in exons that lead to amino acid replacements (Zhou et al, 2013, 2015; Wang et al, 2020).

B. Polymorphisms in the SLCO1B1 gene

Several clinically relevant SLCO1B1 polymorphisms have been identified and functionally characterized. Based on the “star allele nomenclature,” introduced in 2001 (Tirona et al, 2001), 49 ∗ alleles or haplotypes have now been defined with OATP1B1∗1 or SLCO1B1∗1 as the wild type (Cooper-DeHoff et al, 2022; Ramsey et al, 2023). Several of those initial polymorphisms, when expressed in HeLa cells, showed reduced uptake of the OATP1B1 model substrates E1S and E17βG. In particular, OATP1B1∗5 (rs4149056) resulted in an amino acid change p.V174A, which was characterized by reduced uptake of OATP1B1 model substrates E1S and E17βG (Tirona et al, 2001), as well as reduced expression on the plasma membrane when expressed in HEK293 cells (Wagner et al, 2020). Similarly, expressing OATP1B1∗15 in HEK293 cells also resulted in decreased plasma membrane expression (Wagner et al, 2020). Importantly, some of the SLCO1B1 haplotypes result in increased (∗14/∗14) or poor function (∗5/∗5; ∗5/∗15; ∗15/∗15) (Cooper-DeHoff et al, 2022). These haplotypes might need dose adjustments to avoid adverse effects, such as statin-associated musculoskeletal symptoms, sometimes observed in patients treated with the wrong statin doses (Cooper-DeHoff et al, 2022). Besides their effects on statin disposition, some of the additional SLCO1B1 variants were indicated in potential toxicities with drugs like methotrexate, mycophenolate, cyclophosphamide, docetaxel, and doxorubicin (rs4149056, rs4149081, and rs11045879). However, the effect of many of these additional variants on drug disposition and their clinical importance still needs to be confirmed.

C. Polymorphisms in the SLCO1B3 gene

For SLCO1B3, several polymorphisms were characterized functionally after expressing the mutated plasmids in HeLa cells. The uptake of the OATP1B3-specific model substrate CCK-8 (cholecystokinin octapeptide) and of atorvastatin and rosuvastatin was measured. Transport of CCK-8 and rosuvastatin was lower than wild-type OATP1B3 for p.M233I (rs7311358), p.H520P (rs559692629), p.V560A (rs12299012), but rosuvastatin uptake was not affected (Schwarz et al, 2011). In lung allograft recipients treated with mycophenolic acid, the variant rs4149117 (p.S112A), which was reported to result in increased substrate uptake (Letschert et al, 2004), was associated with an increased risk for acute rejection and allograft failure (Tague et al, 2020). Additional studies demonstrated that this variant was associated with carboplatin/paclitaxel toxicity (Mbatchi et al, 2015), with an increased risk of prostate cancer-specific mortality (Wright et al, 2011). Docetaxel disposition (Chew et al, 2012) and sunitinib efficacy (Kloth et al, 2018) were affected by rs4149117. Furthermore, the intron variant rs11045585 was associated with docetaxel toxicity (Kiyotani et al, 2008).

D. Polymorphisms in the SLCO1C1 gene

For SLCO1C1, only a single clinical variant, rs3794271, is listed in PharmGKB. This variant was initially associated with the response to antitumor necrosis factor medications (Acosta-Colman et al, 2013). This association was confirmed in a study with 130 patients with psoriasis from Spain (Julia et al, 2015). However, analyzing genotype data from 1750 UK patients treated with tumor necrosis factor inhibitors did not result in any evidence of a significant association (Smith et al, 2016). Thus, whether the variant rs3794271 indeed is associated with a clinical outcome with respect to rheumatoid arthritis is still unclear.

E. Polymorphisms in the SLCO2A1 gene

Contradictory study outcomes have also been reported for the single variant rs34550074 (resulting in p.A396T in OATP2A1) in the SLCO2A1 gene. While one study reported a significant association with thiazide response in hypertensive patients (Ware et al, 2017), several studies regarding prostaglandin response in patients with open-angle glaucoma are contradictory. No indication of an association of rs34550074 with intraocular pressure response to prostaglandin analogs was reported in a study with 267 subjects from the United States (McCarty et al, 2012). However, a positive association was found between rs34550074 and the response to latanoprost in a study with 89 Han Chinese patients with glaucoma (Zhang et al, 2016). A recent publication again reported no significant correlation of this polymorphism with open-angle glaucoma (Gowtham et al, 2024). Thus, like other genetic variants, the clinical significance of rs34550074 remains to be established.

F. Polymorphisms in the SLCO2B1 gene

There are 3 variants with clinical annotations for SLCO2B1. The variant rs3781727 is a 3'-UTR variant that was reported to affect the oral absorption of voriconazole (Lee et al, 2020a). Variant rs12422149 was reported to result in a poor response to rosuvastatin, while variant rs2306168 did not affect rosuvastatin (Kim et al, 2017; Lehtisalo et al, 2023). Thus, as indicated earlier, over half a million SNPs have been identified in the SLCO genes, but only a small number have been associated with clear clinical outcomes.

X. Role in drug disposition, physiologically based pharmacokinetic modeling, and IVIVE

Drug disposition involves the absorption, distribution, metabolism, and excretion, also termed ADME, of a drug (Caldwell et al, 1995). These processes involve multiple crossings of plasma membranes by drugs and hence the involvement of drug transporters. Physiologically based pharmacokinetic (PBPK) modeling aims to mathematically predict in detail the absorption, distribution, and elimination of a compound in the body such as in humans in drug development (Schmitt and Willmann, 2005). ADME ultimately determines both the pharmacodynamics as well as adverse events of drugs. PBPK was introduced by Teorell (1937) in the 1930s and is now included in the application of drugs and approved by the FDA (Sun et al, 2024). PBPK uses mathematical modeling of all physical and physiologic processes, which determine the pharmacokinetics of a compound as detailed as possible. For this, the organism is subdivided into single organs. It is aimed at describing the disposition of a compound in each of the organs and compartments by taking into account in detail physical and physiologic processes. Commonly considered processes are distribution with the blood flow, partitioning between plasma and organs, processes for crossing organ boundaries or into the cellular space, metabolism, and excretion processes (Schmitt and Willmann, 2005). As outlined earlier, OATPs display a wide tissue distribution in the body and are also expressed in most, if not all, human epithelial tissues and at blood-organ boundaries (Hagenbuch and Stieger, 2013). In terms of drug disposition, these tissue boundary locations are strategic because they can control the access of drugs to different body compartments. Consequently, the access of drugs to their targets, which are OATP substrates, may partly be controlled by the SLCO transporter superfamily. It is necessary for a drug to reach its target (receptor) for obtaining a pharmacodynamic response (Rang, 2006). In the case of a receptor expressed at the cell surface, the drug needs to reach the interstitial space facing this receptor, and in the case of an intracellular target, the drug needs to enter the cell by crossing the plasma membrane. According to the free drug hypothesis, the action of a drug is elicited by free, unbound drugs at the target site. For drug targets expressed at the plasma membranes, the free drug concentration may, in most cases, be very close to its free drug concentration in plasma. This is most likely not the case for intracellular drug targets, for drug targets behind tight barriers like the blood-brain barrier, as well as for drug metabolism, which occurs inside cells (Zhang et al, 2019). Determination of intracellular drug concentrations in animals typically involves killing the animals. Determination of free drug concentrations in human cells in situ requiring invasive procedures is, at best, very challenging (Zhang et al, 2019). Determination of drug concentrations in tissues, for example, in the interstitial space, is possible in humans by microdialysis (Muller et al, 1995), but again, this is an invasive procedure. However, positron emission tomography is a valuable tool that allows us to also obtain information about intracellular drug accumulation in humans (Ghosh et al, 2022). Taken together, knowledge of intracellular and tissue concentration is still not easily obtained, and methods also include extrapolations from PBPK modeling (Guo et al, 2018; Galetin et al, 2024).

The pharmacodynamic action of a drug depends on the concentration of the drug at its molecular target. The concept that the pharmacodynamic response of a drug is tightly linked to its pharmacokinetic properties was pioneered by Dr Levy (Levy, 1966; Nagashima et al, 1969). This concept of the interplay between pharmacokinetics and pharmacodynamics is nowadays widely accepted (Danhof, 2015) and has been extended into pharmacokinetic pharmacodynamic modeling (Schwinghammer and Kroboth, 1988; Meibohm and Derendorf, 1997; 2002). One of the frequent reasons for the attrition of new chemical entities during drug development is an inappropriate pharmacokinetic property in humans, which is discovered early during the clinical phase of drug development (Prentis et al, 1988; Waring et al, 2015). A possible solution to this issue is modeling of drug concentrations in the body by PBPK and IVIVE, which is being constantly refined (Santos et al, 2023; Sugiyama and Aoki, 2023; Li et al, 2024). The basic concept in drug development of IVIVE comprises the quantitative extrapolation of experimental data obtained in vitro as well as data obtained by PBPK to predict the interactions (pharmacokinetic and/or pharmacodynamic) of compounds in humans (Lin, 1998). This approach may increase the safety of new drugs, for example, for first-in-human use, may be used for assessing DDIs (Russell et al, 2024) and can reduce the number of animals needed to develop a new drug (Jaroch et al, 2018). For successful IVIVE, information about the chemical and physical properties of the drug is needed, as well as detailed physiologic knowledge about the targeted species, typically humans (Rostami-Hodjegan, 2012). This information needs to be complemented by data such as expression levels about potential binding proteins, transport proteins involved in the partitioning of the drug (and its metabolites) into tissues, and enzymes involved in the metabolism of the drug and its excretion. Based on the free drug concept, for any biological system, the free drug concentration at its site, along with the interaction of any relevant biological system, should be known. These parameters are subject to interindividual variability (eg, expression levels of the proteins of interest) as well as the physiologic and/or pathophysiologic condition of the individual (Rostami-Hodjegan and Tucker, 2007). The drug information must be supplemented by information about the biological system to which the drug is applied, such as human subjects. These parameters involve information about blood flow, tissue volumes, expression levels of proteins (typically transporters and metabolizing enzymes) involved in drug disposition, and, last but not least, the drug target.

Understanding how a drug moves in the body includes information on the involved transporters, the direction of the transport (uptake versus efflux), kinetic and thermodynamic (driving force) parameters of the individual transporters, tissue distribution, and expression levels of the transporters (Beringer and Slaughter, 2005; Ho and Kim, 2005; Govindarajan and Sparreboom, 2016; Mao et al, 2018). In addition, the endogenous substrates of the transporters of interest should be known. Furthermore, knowledge of the modulation of transporters by the properties and composition of the surrounding membranes is very helpful. It is particularly important to understand the role of transporters in understanding DDIs and their potential effect on altered pharmacokinetics and pharmacodynamics of victim drugs and adverse drug actions (Konig et al, 2013; Gessner et al, 2019). The latter aspect led to guidance documents by the regulatory authorities on the assessment of drug transporters in potential DDIs during drug development (Prueksaritanont et al, 2013; Rollison et al, 2024). Consequently, transport and inhibition experiments are now routine during drug development (Elsby et al, 2022; Rollison et al, 2024). In this context, the broad substrate specificity and the often overlapping substrate specificity between individual OATPs make the incorporation of OATPs in IVIVE models challenging (Izumi et al, 2017). Often, OATP substrates are rather hydrophobic, that is, poorly soluble in aqueous solutions and consequently protein bound in plasma, which poses an additional challenge for incorporating OATPs in IVIVE models. This may be overcome by using albumin as surrogate for protein binding in transport experiments as well as by modifying IVIVE models accordingly (Kim et al, 2019; Miyauchi et al, 2022). Among the drugs used in the study by Izumi et al (2017), there were several statins. The rate-limiting step in hepatic statin elimination is the intrinsic hepatic clearance and not the intrinsic metabolic clearance, and at least in rats, their uptake was the rate-limiting step (Watanabe et al, 2010). It is well-known that with increasing serum concentrations of statins, there is an increased risk for myopathies (Thompson et al, 2003). Because OATP2B1 is expressed in human muscles (Knauer et al, 2010), OATPs may also directly contribute to the development of myopathy in muscles. Hence, reduction of the intrinsic hepatic clearance as a consequence of reduced OATP-transport function in hepatocytes leads to an increased risk of statin-induced myopathy (Link et al, 2008), which may be exacerbated by muscle OATP(s) in susceptible individuals. The study by the SEARCH Collaborative Group (Link et al, 2008) also demonstrates that while all hepatocellular OATPs can mediate the transport of statins, OATP1B1 is the key culprit for simvastatin-induced myopathy.

A general problem when using in vitro information for IVIVE approaches is not only the large variability of reported K_m values (Tables 8 and 9) but also the reported large variability of transporter protein expression levels (correlating to the number of transporters), as exemplified in Table 3. This may, at least in part, also be due to the fact that there is a considerable variability of quantitative transporter protein determination between individual laboratories (Harwood et al, 2013, 2023). For transport protein quantification, frequently a method that determines the total amount of transporter protein in a tissue sample is used, and not the functional portion of the transporters at the plasma membrane, which is required for IVIVE.

While the abovementioned example demonstrates the challenge of successfully incorporating OATPs into IVIVE models, pharmacogenetics and pharmacogenomics were instrumental in substantiating the concept of the in vivo relevance of OATPs in drug disposition. Applying pharmacogenetic strategies has proven to be very useful (Konig et al, 2006; Katz et al, 2008; Kalliokoski and Niemi, 2009; Sissung et al, 2014; Zhou et al, 2017).

Detailed knowledge of the function and expression of OATPs may help to increase the precision of IVIVE in predicting adverse drug events (Kimoto et al, 2022; Liang et al, 2024b). Because OATP substrates are typically rather hydrophobic, they are often, to a considerable extent, bound to proteins. Consequently, this will increase the difficulty of predicting free intracellular (ie, not bound to proteins) drug concentrations, and consequently, predicting metabolism in IVIVE in such situations is not straightforward.

XI. Concluding remarks

Since the cloning of the first OATP, research conducted by basic scientists and clinicians has resulted in the recognition of several OATPs, particularly liver-expressed OATP1B1 and OATP1B3, as important drug uptake transporters. These 2 OATPs are involved in the disposition of numerous drugs and thus can lead to adverse drug interactions due to either DDIs or genomic variabilities. While the recent elucidation of the structure of these 2 OATPs resulted in a better understanding of the function of these proteins and where their substrates might interact, additional experiments will be needed to explain the exact transport mechanisms and the broad substrate specificity. Furthermore, we still do not know enough about members of the OATP3, OATP4, OATP5, and OATP6 families. Research into the physiologic, pathophysiologic, and potential pharmacologic role of these lesser-studied OATPs is urgently needed for a better understanding of the function and relevance of the whole OATP/SLCO family and how they potentially can be used to target drugs to certain tissues or organs.

Conflict of interest

The authors declare no conflicts of interest.

Associate Editor: Joanne Wang