Membrane transporters in cell physiology, cancer metabolism and drug response

Plasma and organellar membrane transporters play a vital role in the import and export of a diverse range of endogenous and exogenous small molecules and ions, making them essential for whole-body, cell and organellar homeostasis. Given their importance for tissue and cell homeostasis, it's unsurprising that membrane transporters also play important roles in a variety of diseases, from rare inborn errors of metabolism to neurodegenerative disorders and, perhaps most notably, across a broad spectrum of cancer types.

In the context of cancer, changes in membrane transporter expression can be driven by a range of oncogenic mutations. These expression changes facilitate cell-intrinsic, pro-proliferative metabolic alterations and adaptations to, as well as metabolic crosstalk with, the tumour microenvironment. Several superfamilies of transporters exist, classed broadly by their mechanism of transport. The largest and most diverse of these is the Solute Carrier (SLC) transporter superfamily, comprising over 450 members that transport ions and metals, sugars, amino acids, nucleosides, and many other solutes across membranes (Meixner et al., 2020; Tweedie et al., 2020; Ferrada and Superti-Furga, 2022). The physiological function of SLCs, as well as their roles in cancer cell metabolism and how this may impact response to drugs, is the main focus of this Review. For consistency, transporters will be referred to by their SLC name, with any additional common nomenclature in brackets. Further information on the SLC nomenclature system has been reviewed by Hediger et al. (2013). 

Structurally, SLC transporters are a highly diverse group of multi-membrane pass proteins, with the Leu-T and the Major Facilitator Superfamily (MFS) comprising the two most common structural folds. Significant advances in membrane protein structural biology made over the past decade have hugely improved our understanding of how transport takes place at the molecular level – with mechanisms including Rocker Switch, Gated-Pore and Elevator, that have been reviewed elsewhere (Drew and Boudker, 2015; Colas et al., 2016; Dvorak and Superti-Furga, 2023). Functionally, SLCs are classed as either facilitative or secondary active transporters. Facilitative transporters allow the passage of a substrate down its concentration gradient, while secondary active transporters translocate substrates by utilising the electrochemical gradient of a second substrate, usually an ion, enabling the movement of the primary substrate against its concentration gradient (Crane, 1960) (Fig. 1). Co-transported substrates can either be translocated in the same (symporter) or opposing (antiporter) directions. Unlike active ATP-Binding Cassette (ABC) transporters, secondary active transporters indirectly rely on the energy status of the cell by utilising ion gradients established by ATP-hydrolysing pumps (Hediger et al., 2004) (Fig. 1). Examples are members of the sodium (Na⁺)-glucose co-transporter (SGLT) SLC5 family, which drive glucose import using the Na⁺ concentration gradient generated by ion pumps/transporters, such as the Na⁺/K⁺ ATPase (Crane, 1960; Wright et al., 2011).

The breadth of substrates makes the SLC family fundamental to many aspects of physiology and, as a result, implicates them in a number of inborn errors of metabolism (Box 1), with over 80 associated with monogenic disorders (Lin et al., 2015). The importance of SLCs is exemplified by SLC25A42, one of two known mitochondrial coenzyme A (CoA) transporters (Fiermonte et al., 2009). While synthesised in the cytoplasm (Martinez et al., 2014), CoA is required in the mitochondria for the formation of acetyl-CoA from pyruvate (Fiermonte et al., 2009), to fuel the tricarboxylic acid (TCA) cycle (Box 1) and energy metabolism, as well as for beta oxidation of fatty acids (Box 1). The movement of CoA across the inner mitochondrial membrane is, therefore, essential for fundamental metabolic functions. Indeed, children born with mutations in SLC25A42 display a wide array of symptoms, such as mitochondrial myopathy, developmental regression and epilepsy (Shamseldin et al., 2016; Almannai et al., 2018). Other non-silencing mutations in or around the gene appear to impact lipid metabolism. In fact, single-nucleotide polymorphisms near this gene have been associated with hypertension in some studies (Fox et al., 2011), likely due to the role of CoA in beta oxidation in the mitochondrion. Interestingly, SLC25A42 expression has been shown to be upregulated in foetuses undergoing transcriptional reprogramming to adapt to placental insufficiency (Box 1) (Chou and Wang, 2020), suggesting that SLC25A42 upregulation may contribute to the metabolic reprogramming required for foetal survival in sub-optimal nutritional environments.

Beyond inborn errors of metabolism, several SLCs have been associated with predisposition to or increased severity of multi-factorial diseases (reviewed previously by Lin et al., 2015), such as type II diabetes (Barragán-Álvarez et al., 2021), hypertension (Boedtkjer et al., 2011; International Consortium for Blood Pressure Genome-Wide Association Studies et al., 2011), obesity (Le et al., 2021; Hodges et al., 2022), depression (Santarelli et al., 2015), cancer (Nimmanon et al., 2017; Fang et al., 2021; Han et al., 2021; Mao et al., 2021) and aging (Crocco et al., 2018). For example, at synapses and in astroglia, the SLC1 (Xu et al., 2016) and SLC6 (Kristensen et al., 2011) families of amino acid transporters are responsible for the removal of neurotransmitters from the synaptic cleft, quenching postsynaptic effects after neurotransmission, and balancing excitatory and inhibitory neurotransmitters (Aykac and Sehirli, 2020). Reduced function of the microglial glutamate transporter SLC1A2 (also known as GLT1 or EAAT2) has been associated with neurodegenerative disorders, such as Amyotrophic Lateral Sclerosis (Rothstein, 1996; Yang et al., 2009), demonstrating the vital role of SLCs within the nervous system.

The diversity of substrate specificities and regulatory properties of SLCs allow cells to tailor transporter expression signatures to tissue- and cell type-specific metabolic requirements, as well as to their extracellular environment. Accordingly, SLCs are not only differentially but also dynamically expressed in varying cell types and tissues (Nishimura and Naito, 2008; Morioka et al., 2018; O'Hagan et al., 2018), even as early as in the developing embryo (Saunders et al., 2015; Schoels et al., 2021). Indeed, some tissue functions rely heavily on the import or export of specific metabolites or ions. For example, creatine (Box 1) is synthesised in the liver and kidneys (da Silva et al., 2009) but is mostly utilised in cell types with high energy demand, such as skeletal muscle and neuronal cells (Bessman and Geiger, 1981; Kazak and Cohen, 2020). The functioning of these cells, therefore, depends on the creatine transporter SLC6A8 (also known as CRT1), and reduced function of this transporter leads to symptoms such as intellectual disability and seizures (Cervera-Acedo et al., 2015; Tise et al., 2023).

In the kidney, the central role of filtering nutrients and salts from the blood means that SLC transporters are essential for the physiology of this organ. 380 SLCs are significantly expressed in renal tissues, with many being differentially expressed along the kidney tubules (Kanai et al., 1994; Lewis et al., 2021). Importantly, alternative transporter isoforms with varying substrate affinities may be preferentially expressed in specific locations or in response to altered diet of an organism. Indeed, in the earlier segments of the nephron, which are exposed to a higher concentration of a wide array of salts, the epithelial cells lining the tubules mostly express low-affinity and high-capacity SLCs to enable bulk removal of concentrated solutes. In the distal nephronal regions, these SLCs are replaced by higher affinity isoforms that ensure the removal of salts even at the low remaining concentrations (Giménez et al., 2002; Mount, 2014). Indeed, feeding mice a low-salt diet triggers an upregulation of NKCC2B, the low-capacity and high-affinity splice isoform of the Na⁺/K⁺/Cl⁻ co-transporter SLC12A1 (also known as NKCC2), in the renal cortex and medulla (Schiessl et al., 2013), demonstrating the importance of SLCs in adapting organ physiology to environmental conditions.

The expression level, activity and subcellular localisation of SLCs, in combination with substrate concentration gradients, determine the influx rate and, subsequently, the concentration of a substrate within the cell or a subcellular compartment. As a result, SLC activity must be tightly regulated – for example in response to substrate availability or requirement – to appropriately shape the metabolic architecture a cell adopts in response to external stimuli, such as hypoxia (Wojtal et al., 2014; Gorczyca et al., 2021). Regulation of SLCs can occur at the level of transcription (reviewed previously by Zhou and Shu, 2022) or post-transcriptionally through splicing modulation (Schiessl et al., 2013) or RNAi (reviewed previously by Yi and Yu, 2022). Short-term, dynamic changes can also occur at the post-translational level, where modifications including glycosylation, phosphorylation and ubiquitylation, can impact protein stability, folding and membrane insertion, as well as directly altering transporter kinetics (Wieman et al., 2007; Czuba et al., 2018). For example, the zinc (Zn²⁺) exporter SLC30A1 (also known as ZNT1) is glycosylated and endocytosed in the absence of Zn²⁺ while, by contrast, in the presence of Zn²⁺, a decrease in glycosylation increases its stability on the plasma membrane (Nishito and Kambe, 2019). SLCs have also been shown to be regulated by direct interaction with exogenous small molecules, including environmental chemicals, such as organophosphorus pesticides, that can activate or inhibit the transport of substrates (Chedik et al., 2019) (reviewed previously by Nicklisch and Hamdoun, 2020). In addition, when SLC transporters are bound to their substrates, they can undergo conformational changes that – as in the case of the Zn²⁺ transporter SLC39A4 (ZIP4) – may promote their removal from the plasma membrane by endocytosis, allowing substrate concentrations to directly regulate transporter numbers (Shimokawa et al., 2002; Zhang et al., 2020).

By facilitating the movement of metabolites and salts between the intra- and extracellular space, and by regulating intracellular compartmentalisation of compounds within organelles, SLCs inevitably play a fundamental and understudied role in regulating the metabolic network of the cell. According to the rules of mass action, the import of a metabolic substrate or product typically influences the direction and activity of metabolic reactions and pathways of which they are part (Song et al., 2020). Indeed, when glucose import is increased, such as by an upregulation of SLC2A1 (GLUT1) expression, hexokinase activity, as indicated by glucose-6-phosphate levels (Fujii and Beutler, 1985), and overall glycolytic activity, as shown by increased lactate levels, also increase (Fujii and Beutler, 1985; Freemerman et al., 2014; Williamson et al., 2018; Song et al., 2020).

Furthermore, SLC transporters on organellar membranes can influence the compartmentalisation of allosteric or orthosteric (Box 1) inhibitors relative to target enzymes. The metabolome is tightly regulated by both product inhibition and by inhibition or activation of enzymes and pathways by seemingly unrelated metabolites (Lindsley and Rutter, 2006; Li et al., 2010; Gerosa and Sauer, 2011; Macpherson and Anastasiou, 2017). The extent of small molecule-protein interaction has been clearly demonstrated in the ergosterol (Box 1) biosynthesis pathway in yeast, where 16 of the 21 proteins of this pathway were found to be bound to metabolites, along with a large number of protein kinases (Li et al., 2010). Indeed, it is estimated that more than 90% of enzymatic reactions in cells can be inhibited by one or more metabolites, the most impactful of which are orthophosphate, ATP, ADP and AMP (Alam et al., 2017). ATP, for example, can inhibit numerous metabolic enzymes, such as the rate-limiting glycolytic enzymes phosphofructokinase and pyruvate kinase (Larsson et al., 2000; Sander et al., 2019). These inhibitory interactions can be managed through compartmentalisation of metabolites within the cell (Zecchin et al., 2015), whereby the substrate breadth and subcellular localisation of SLC transporters enable them to shape the metabolic network by regulating the localisation of inhibitors relative to their targets (Bar-Peled and Kory, 2022). For example, ATP/ADP/AMP/P_i can be transported by one or more SLCs in the SLC17, SLC25, SLC29 and SLC35 transporter families (Ruprecht et al., 2014; Kunji et al., 2016; Thompson et al., 2016), many of which are localised to the mitochondrial inner membrane, allowing cells to compartmentalise energy equivalents and regulate their inhibitory properties. Phosphofructokinase can also be regulated by citrate, an intermediate of the TCA cycle. Citrate is transported across the inner mitochondrial membrane by SLC25A1, implicating this SLC in the regulation of glycolysis by connecting inhibitor and target (Nielsen, 1983).

Importantly, a large number of SLCs transport metal ions, many of which interact with and regulate ∼40% of all enzymes in cells (Andreini et al., 2009; Aulakh et al., 2022). Zn²⁺, for example, plays a role in regulating protein kinases and phosphatases (Nimmanon et al., 2017; Slepchenko et al, 2018) that function in signal transduction as well as fundamental metabolic pathways, like glycolysis (Tamaki et al., 1983; Fitzsimmons et al., 2018). Therefore, the many SLCs that can transport Zn²⁺ as well as other metal ions, such as iron, manganese and calcium, from the extracellular space or intracellular stores, play crucial roles in regulating enzymatic activity in the cell and, by extension, cellular metabolism and homeostasis.

As well as influencing metabolic reactions, SLCs can have a broader impact on cell physiology and metabolism by influencing transcriptional regulation (Hyde et al., 2007; Wieman et al., 2007; Menga et al., 2017). Zn²⁺ transporters from the SLC39 (ZIP) family have been associated with inflammation, cell movement and cell morphology owing to the regulation of Zn²⁺-dependent transcription factors, such as STATs, Snail proteins and NF-κB, when Zn²⁺ import is altered (Yamashita et al., 2004; Leung et al., 2012; Liu et al., 2013; Geng et al., 2018). In addition, purine import via SLCs promotes the interaction between the histone acetylation reader protein BRD4 and chromatin, highlighting the impact of SLC transporters on epigenetic regulation (Li et al., 2021).

SLC transporters play a central role in cellular interaction with the extracellular environment, and the SLC expression signature of a cell can heavily influence in vitro growth rates in culture media with differing compositions, as illustrated in a recent preprint (Chidley et al., 2023 preprint). In turn, the redox and energetic status of a cell can influence the SLC expression profile to enable cells to exploit the availability of extracellular nutrients. For example, metformin treatment alters the expression of SLCs transporting substrates ranging from Zn²⁺ and Cl⁻ to glycerol-3-phosphate and amino acids (Le et al., 2021), which underscores the metabolic impact of this drug. In fact, the significance of nutrient import in relation to cellular metabolic state is highly evolutionarily conserved. Unicellular organisms, such as yeasts and bacteria, also import non-essential metabolites, and different metabolic strategies may push cells to rely more on the extracellular metabolome. There is a clear logic behind this strategy: metabolite synthesis comes with the cost of resources, such as carbon, ATP or redox equivalents, including NAD⁺ or NADPH (reviewed previously by Keibler et al., 2016). Scavenging non-essential metabolites from the environment, therefore, increases the availability of metabolic precursors for alternative reactions (Barton et al., 2010). For example, respiratory-incompetent or NAD⁺-deficient Saccharomyces cerevisiae or Schizosaccharomyces pombe, despite being prototrophic in terms of amino acid production, show improved growth in medium supplemented with amino acids (Malecki et al., 2020; Vowinckel et al., 2021; Alam et al., 2023). Similarly, S. cerevisiae show increased tolerance to oxidative stress upon lysine supplementation, despite the fact that lysine is non-essential in the strains used. Lysine synthesis requires reduction, using NADPH at multiple steps, which diminishes the NADPH pool available for glutathione reductase, decreasing its capacity to remove reactive oxygen species (ROS) (D'Souza et al., 2014; Olin-Sandoval et al., 2019). S. cerevisiae was also shown to preferentially import metabolites if available and to downregulate the relevant biosynthetic pathways (Campbell et al., 2015). This underlines the significance of SLCs as main importers of extracellular metabolites that avoid costly ‘metabolic independence’.

Nutritional availability in different zones of a cell community, tissue or organ, inevitably leads to metabolic heterogeneity – even in a genetically homogenous community (Kondo et al., 2021). Even when cells are genetically capable of synthesising specific biomass components, if they lack the appropriate precursors to do so, they must rely on importing the final product through syntrophy (Box 1) (Andersen et al., 2015; Rodionova et al., 2015; Pacheco et al., 2019; Yu et al., 2022). A classic example of metabolic heterogeneity in a tissue is the retina: retinal pigmented epithelial (RPE) cells function as metabolic supporters for photoreceptors, which have a high turnover rate and energy demand but lack direct contact with the vasculature. Both RPE and photoreceptors rely on the expression of an array of SLCs to enable this supportive metabolic crosstalk. Recent work has shown that RPE cells preferentially take up proline and, therefore, have high expression levels of SLC6A20 (also known as SIT1, XT3 or XTRP3). RPE cells use this proline as an alternative nitrogen source for the synthesis of amino acids that are subsequently exported to photoreceptors in order to sustain their high metabolic rate (Takanaga et al., 2005; Liao et al., 2010; Strunnikova et al., 2010; Radeke et al., 2015; Chao et al., 2017; Liu et al., 2019; Li et al., 2020; Du et al., 2021).

Beyond the import of nutrients required for growth, SLCs allow cells to ‘sample’ nutrient availability: these extracellular metabolites can modulate intracellular metabolic signalling pathways, which enables cells to quickly adapt in response to changes in the extracellular metabolome (Hyde et al., 2007; Zhang et al., 2018; Pizzagalli et al., 2021). For example, reduced expression of the lactate transporter SLC16A1 (also known as MCT1) indirectly leads to AMPK activation (Carneiro et al., 2017), while inhibition of the plasma membrane amino acid transporter SLC7A5 (also known as LAT1) reduces mTOR activity (Imai et al., 2010; Ueno et al., 2016). SLCs on organellar membranes can also help regulate signalling pathways, impacting processes, such as protein synthesis (Tripathi et al., 2019). Even more directly, SLC transporters can take on a ‘transceptor’ role, i.e. that of a transporter and receptor (see reviews by Taylor, 2009; Kriel et al., 2011; Zheng et al., 2016; Pastor-Anglada and Pérez-Torras, 2018). For example, SLC38A9 (also known as URLC11) exports glutamine and/or arginine from the lysosome when lumen concentrations are high, such as in a fed state when cells are in amino acid-rich microenvironments. As well as exporting amino acids from lysosomal stores, the N-terminal domain of SLC38A9 provides a binding site for the Rag–Ragulator complex on the lysosome and promotes the activation of the Rag-GTPases (Shen and Sabatini, 2018). Its N-terminus has been shown to be exposed when SLC38A9 is transporting arginine, allowing activation of the Rag-GTPases and, in turn, of mTORC1 when arginine concentrations in the lysosome are high (Wu et al., 2006; Jung et al., 2015; Rebsamen et al., 2015; Wang et al., 2015b; Scalise et al., 2019; Lei et al., 2021). This positions SLC38A9 as a core sensory component of the mTORC1 regulatory machinery. Overall, SLC transporters play a key role in both sensing of and responding to the extracellular environment and are fundamental components of metabolic and signalling pathways.

Altered metabolism is a well-established hallmark of cancer, fuelling proliferation though the production of biosynthetic intermediates and maintenance of redox balance within a complex, often nutrient-depleted, microenvironment. Given the diversity of the SLC superfamily and their importance in shaping cellular metabolism, it is becoming increasingly appreciated that SLC transporters are central to this metabolic adaptation. As with their physiological roles, the contribution of SLCs to cancer are too numerous for us to comprehensively cover in this Review. Therefore, our aim is to cover the key thematic roles of transporters in cancer, illustrated with examples. Where available, we refer readers to additional publications for further reading.

Dividing cancer cells require exogenous carbon sources to generate ATP, biosynthetic intermediates and reducing power. Glucose is a major carbon source for cancer cells, and increased expression of SLC2 glucose transporter (GLUT) family members is among the best-known adaptations, as central contributors to the Warburg effect or aerobic glycolysis (Warburg, 1925; Warburg et al, 1927). In humans, SLC2A1 was the first of 14 GLUT family members to be identified and has been the most extensively studied to date (Kasahara and Hinkle, 1977; Carruthers et al., 2009; Thorens and Mueckler, 2010). A number of oncogenic and environmental alterations are associated with enhanced SLC2A1 expression in cancer, including mutation of EGFR or KRAS mutation, amplification of c-Myc and stabilisation of HIF1A (Box 1) (Ancey et al, 2018).

While metabolic changes in cancer are now known to be far more complex and heterogenous than can be encompassed by the Warburg hypothesis alone, the upregulation of SLC2A1 is a highly conserved feature of many tumour types (Brown and Wahl, 1993; Haber et al., 1998; Koh et al, 2017) and associated with decreased overall survival of cancer patients (Yu et al., 2017). Importantly, however, the enhanced glucose transport via GLUTs has been exploited for diagnostic and tumour-monitoring purposes with the development of FDG-PET imaging (Box 1) (Guerrero et al., 1990; Kelloff et al., 2005).

Increased glycolytic rates also lead to excess production of lactate, which can be transported into and out of the cell by the SLC16 family of monocarboxylate transporters. The highest capacity lactate exporters are SLC16A1 and SLC16A3 (MCT4), and their upregulation prevents intracellular acidosis.

Interestingly, lactate can be a carbon source, entering cells via SLC16A1, and the contribution of lactate to TCA cycle intermediates exceeds that of glucose in some tumour types (Hui et al., 2017). This adds to work by Sonveaux et al., who proposed a metabolic symbiosis model in some tumours, whereby glycolytic cancer cells export lactate via SLC16A3, which is subsequently imported via SLC16A1 for oxidative metabolism by cells within better-vascularised regions (Sonveaux et al., 2008). Furthermore, non-glucose carbon sources, including lactate, can be used even in well-perfused areas of non-small cell lung cancer patient tumours (Hensley et al., 2016).

In addition to lactate, monocarboxylate transporter substrates include pyruvate and ketone bodies, all of which are co-transported with a proton, rendering them sensitive to pH gradients across the membrane (Halestrap, 2013). Relative to SLC16A1 and SLC16A3, SLC16A7 (also known as MCT2) has been far less well studied in the context of cancer. However, SLC16A7 has the highest affinity for tested substrates of all members of the SLC16 family (Bröer et al., 1999), and a number of studies have demonstrated its increased expression in cancer, most notably in prostate carcinoma (Pértega-Gomes et al., 2012), where it has been suggested to be a putative biomarker. In breast cancer-derived lung metastases, SLC16A7 appears to be particularly influential as it drives cellular uptake of pyruvate. Downstream metabolism of this pyruvate promotes both the maintenance of mTORC1 signalling through increased activity of PHGDH (Rinaldi et al., 2021) and remodelling of the metastatic niche through hydroxylation of collagen (Elia et al., 2019). The increased production of α-ketoglutarate that results from pyruvate uptake has been implicated in driving these effects. As well as pyruvate, the uptake of adipocyte-derived β-hydroxybutyrate via SLC16A7 has been suggested to be pro-tumourigenic in breast cancer, as it has been linked to increased histone acetylation in the promotor region of tumour-promoting genes (Huang et al., 2017).

These studies highlight the importance of both metabolic and transporter plasticity to promote survival in dynamic nutrient states. As a result of their key roles in cancer cells, targeting GLUTs, MCTs and many other SLC transporters has been of major interest in recent years, with significant public and private funding deployed to improve understanding of this family (Wiedmer et al., 2022), as well as the development of several small-molecule inhibitors and clinical trials (Kopitz et al., 2016; Olszewski et al., 2022; Halford et al., 2023, see Table 1).

Exogenous amino acids provide an essential nitrogen source and contribute to energy production, redox homeostasis and cell signalling. Glutamine, the most abundant amino acid in human plasma, is considered conditionally essential in the context of cancer, since cancer cells are often glutamine dependent (Wei et al., 2021; Yang et al., 2016).

High uptake rates of glutamine, as well as several other amino acids, are facilitated by multiple SLC families, including SLC1, which consists of two transport systems: the excitatory amino-acid-transporters SLC1A1, SLC1A2 and SLC1A3 (EAAT3, EAAT2 and EAAT1, respectively) and SLC1A6 and SLC1A7 (EAAT4 and EAAT5, respectively) and the alanine-serine-cysteine transporters SLC1A4 and SLC1A5 (ASCT1 and ASCT2, respectively). Notably, SLC1A5 is a Na⁺-dependent neutral amino acid exchanger, whose expression and/or activity is upregulated in response to the oncogenic drivers c-Myc (Wise et al., 2008; Gao et al., 2009) and EGFR (Avissar et al., 2008; Lu et al., 2016; Tao et al., 2017), and has been correlated with a significantly poorer outcome in patients. Silencing of SLC1A5 was shown to reduce growth in breast (van Geldermalsen et al., 2016), lung (Hassanein et al., 2013) and prostate cancer (Wang et al., 2015a), as well as acute myeloid leukaemia (Willems et al., 2013) and, subsequently, there has been a push to identify specific inhibitors for this transporter (Schulte et al., 2016). While successful in inhibiting cancer cell growth in in vitro and in vivo preclinical models of colorectal cancer (Schulte et al., 2018), further work has suggested that the mechanism of action of some of these inhibitors is through the targeting of two other glutamine transporters, SLC38A2 (SNAT2) and SLC7A5 (Bröer et al., 2019). Amino acids are directly sensed by mTORC1 and, as such, their transport into cells not only provides biosynthetic intermediates but also drives signal transduction pathways that stimulate cell growth. SLC7A5, an amino acid antiporter (Singh and Ecker, 2018), promotes the maintenance of essential amino acid levels, often in exchange for glutamine. The subsequent promotion of mTORC1 signalling demonstrates the importance of SLC7A5 in driving proliferation of KRAS-driven colorectal cancer, both in vitro and in vivo (Najumudeen et al., 2021).

Altered cancer metabolism often leads to elevated levels of ROS, which can contribute to enhanced cancer progression and aggressiveness (Cheung and Vousden, 2022). ROS levels must, therefore, be kept in check, since uncontrolled elevation of ROS can damage DNA, proteins and lipids. Maintaining redox homeostasis is, therefore, critical for cancer cell survival. The cystine/glutamate antiporter SLC7A11 (also known as xCT) is a crucial regulator of redox homeostasis that functions by importing cystine into the cytoplasm in exchange for glutamate in a Na⁺-independent manner (Fig. 1). Cystine is the oxidised form of cysteine, a rate-limiting substrate for the biosynthesis of the antioxidant glutathione. While vital for maintaining redox balance, cystine is relatively insoluble, therefore, when import rates are high, it must be rapidly reduced to cysteine. This reduction reaction requires NADPH, which puts pressure on the pentose phosphate pathway (Box 1) (Liu et al., 2020). The subsequent increased dependence on glucose demonstrates that metabolic adaptation can often be coupled to novel vulnerabilities. Similarly, SLC7A11-driven depletion of the intracellular glutamate pool from cells that express high levels of this antiporter has also been demonstrated to increase sensitivity to glutamine deprivation in these cells (Ji et al., 2018).

Glutathione is a co-factor of glutathione peroxidase 4, an enzyme that eliminates toxic lipid peroxides, the build-up of which can induce ferroptosis (Box 1). Thus, SLC7A11 also promotes tumour survival through inhibition of ferroptosis (Badgley et al., 2020). Inhibition of SLC7A11 may, therefore, damage cancer cells by causing fatal ROS-mediated damage, making it an attractive pharmacological target.

As well as contributing to the metabolic rewiring of cancer cells to provide a growth or survival advantage, expression of some transporters or their isoforms has been suggested to directly contribute toward the tumourigenic potential of cells.

The mitochondrial pyruvate carrier (MPC), whose molecular identity remained elusive for many years, was identified by two groups in 2012 (Bricker et al., 2012; Herzig et al., 2012) and shown to be conserved from yeast to mammals. It is a heterodimeric complex formed of MPC1 and MPC2 (also known as SLC54A1 and SLC54A2, respectively) within the inner mitochondrial membrane (see Tavoulari et al., 2019; and the recent review by Tavoulari et al., 2023). By regulating the transport of pyruvate into the mitochondria, MPC sits at the interface between glycolysis and glucose oxidation, directly influencing the relative rates of each pathway. In several cancer types, decreased MPC expression correlates with increased aerobic glycolysis and reduced mitochondrial respiratory capacity (Li et al., 2017; Koh et al., 2018). In models of colorectal cancer, however, downregulation of MPC was sufficient to impair mitochondrial pyruvate oxidation. This promoted a glycolytic phenotype and the subsequent initiation of intestinal tumours both in mouse and fly models (Bensard et al., 2019). Supporting these studies, several reports have correlated reduced expression of the MPC with more aggressive cancer phenotypes and poorer patient prognosis (Schell et al., 2014; Zhong et al., 2015; Li et al., 2016; Wang et al., 2016). In contrast, within cholangiocarcinoma cells, increased MPC1 (SLC54A1) expression driven by the transcriptional coactivator PGC1α (officially known as PPARGC1A) enhanced pyruvate entry into the mitochondria. This facilitated an increase in oxidative phosphorylation at the expense of aerobic glycolysis and a subsequent rise of ROS, which was associated with increased metastatic potential of xenografted cholangioacarcinoma cells (Li et al., 2018). Similarly, pharmacological inhibition of the MPC in androgen-receptor-driven prostate cancer models reduced proliferation of cancer cells in vitro and in xenografts (Bader et al., 2019), suggesting that the tumourigenic effect of MPC depends on the metabolic context of the cell it is expressed in.

The mitochondrial glutamine importer was recently identified to be a variant of the above-discussed SLC1A5 and, thus, named SLC1A5_var. SLC1A5_var is directed to the mitochondria through an N-terminal targeting sequence, where it promotes glutamine anaplerosis (Box 1) to fuel the TCA cycle and generate ATP, as well as increased glutathione synthesis to contribute to the maintenance of redox balance (Yoo et al., 2019). The identification of SLC1A5_var sheds further light on the capacity of metabolic reprogramming to drive tumourigenicity.

The tumour microenvironment is complex and dynamic, with cancer cells co-existing alongside a variety of immune cells, cancer-associated fibroblasts and endothelial cells, often perfused by a poorly formed vasculature. Among various modes of crosstalk within this environment, metabolite exchange has been shown to be highly important and to have implications not only for proliferation and survival but also drug resistance. Transporters are, of course, central to this exchange.

Pancreatic cancer has a particularly dense stromal architecture that causes poor vascular perfusion, resulting in a nutrient-depleted environment. Pancreatic stellate cells (PSCs), however, provide tumour cells with non-essential amino acids derived from cancer cell-stimulated autophagic breakdown of proteins within PSCs (Sousa et al., 2016). In particular, alanine taken up by the cancer cells is used to fuel the TCA cycle. This exchange is mediated by specific transporters; SLC1A4 facilitating non-essential amino acid efflux from PSCs and upregulated SLC38A2 enabling their uptake into the cancer cells (Parker et al., 2020).

Metabolic crosstalk can also modulate immune cells. Local depletion of methionine that results from SLC43A2 upregulation in cancer cells impairs CD8⁺ T-cell function (Bian et al., 2020). By contrast, metabolic rewiring induced by the uptake of tumour-cell-derived lactate inhibits succinate secretion by T-cells, which promotes autocrine signalling and tumour cell cytotoxicity (Elia et al., 2022). While this study did not characterise specific transporters, they no doubt constitute a central node of this cancer-immune system metabolic interaction.

Transporter-driven metabolic interactions can extend beyond the local environment and impact whole-body metabolic homeostasis. For example, in fly models that had been fed a high-sugar diet, FGF signalling from malignant tumours in the eye epithelium promoted systemic muscle wasting (Newton et al., 2020). The subsequent release of amino acids from muscle tissue, in particular proline, was capitalised on by upregulation of an SLC36 family orthologue to drive tumour growth (Newton et al., 2020).

As with so many aspects of cancer, metabolism is heterogenous both at inter- and intra-tumoural level (Evers et al., 2019; Kim and DeBerardinis, 2019). It is clear that transporters play an important role in generating this metabolic heterogeneity. The human genome encodes a diverse range of SLC transporters (Fig. 2A). While the SLC and ABC transporter signatures in patient tumour samples largely cluster by tissue-of-origin (Fig. 2B), tumour samples express a far more varied complement of transporters than the corresponding normal tissue samples. This demonstrates substantial heterogeneity in transporter expression across tumours of the same type, probably originating from a combination of genetic variation in, e.g. oncogenic driver mutations, and non-genetic variations, such as proximity to blood vessels and subsequent nutritional availability. Given the established roles of transporters in cancer metabolism and their potential involvement in therapeutic response (see below), this heterogeneity could have significant implications for patients receiving cancer treatment.

As well as regulating endogenous metabolite transfer across lipid membranes, several families of transporters are better known for transporting exogenous substrates, affecting drug absorption, distribution, metabolism and excretion (ADME).

Whilst this article focuses on SLCs, the significance of ABC transporters in ADME warrants a brief discussion. The ABC family of transporters comprises 48 members (Moore et al., 2022) that actively use ATP hydrolysis to pump their substrates against their concentration gradient out of cells. Several, such as ABCB1 (also known as MDR1, P-glycoprotein), ABCG2 (also known as BCRP) or ABCC1 (also known as MRP1), have been widely studied due to their involvement in drug export and subsequent drug resistance. ABCB1, the best-characterized member of this family, has been associated with the transport of ∼200 compounds, including chemotherapeutic drugs, such as paclitaxel, irinotecan and olaparib (Vaidyanathan et al., 2016; César-Razquin et al., 2018). Given the number of ABCB1 interactions, the possibility of inhibiting the activity of this transporter has been explored to improve therapeutic efficacy of its substrates. Unfortunately, though three generations of ABCB1 inhibitors did improve the efficacy and toxicity profiles of chemotherapeutics during preclinical development, little success was seen in clinical trials (Robey et al., 2018).

Several SLC transporters from the SLCO (also known as OATP or SLC21), SLC22 (also known as OAT or OCT) and SLC47 (also known as MATE) families have also been well studied, particularly in the context of organs involved in drug absorption and metabolism, such as the intestine, liver and kidneys. Their activities, as well as those of a subset of ABC transporters, are recommended by the International Transporter Consortium to be investigated as part of the development, approval and safe use of all small-molecule drugs (Giacomini et al., 2018).

The role of SLC families beyond SLCO, SLC22 and SLC47 in the uptake of exogenous small molecules is less well understood. However, a number of studies have identified additional SLC transporters involved in drug uptake (Fig. 3A). Indeed, the chemical space occupied by many therapeutic compounds overlaps with that of endogenous metabolites (O'Hagan and Kell, 2017, 2020). Given that these transporters have evolved to regulate the transport of endogenous metabolites as well as exogenous natural products obtained from the diet, a broader role for transporters in drug uptake and efficacy warrants investigation. Expression of SLC16A1, for example, was shown to determine sensitivity to the glycolytic inhibitor 3-bromopyruvate (Birsoy et al., 2013; Skaripa-Koukelli et al., 2021), which is a candidate anti-cancer drug. Similarly, another monocarboxylate transporter, SLC16A7 (MCT2) facilitates the cellular uptake of MOG, leading to selective toxicity of DMOG (Box 1) in SLC16A7-expressing cells (Fets et al., 2018, 2022), which has implications for the use of this tool compound to inhibit prolyl hydroxylases. Several drugs, such as the chemotherapeutic melphalan (Begleiter et al., 1979), the key Parkinson's disease drug L-DOPA (Kageyama et al., 2000; Chien et al., 2018), the anticonvulsant gabapentin and its structural relative pregabalin (Dickens et al., 2013; Takahashi et al., 2018), are transported by SLC7A5. Sensitivity to the survivin inhibitor YM155, has also been shown to depend on its transporter SLC35F2, an orphan nucleoside transporter (Winter et al., 2014). YM155 has undergone numerous clinical trials in various cancers, with limited success (Li et al., 2019). It remains to be seen whether patient outcomes could be improved through stratification by SLC35F2 expression status.

Classically, most small-molecule drugs were thought to diffuse through the lipid bilayer of plasma membranes dependent on their physicochemical properties, with predictions of relative permeability being approximated by ‘Lipinski's rule of five’ (Box 1) (Lipinski et al., 1997). While the above examples clearly illustrate that membrane transporters can play a role in drug uptake (Birsoy et al., 2013; Winter et al., 2014; Fets et al., 2018; Girardi et al., 2020), the mechanism of entry into target cells is not well characterised for the majority of drugs, and the proportion of drugs that require a transporter for cell entry has been extensively debated (Kell and Oliver, 2014; Smith et al., 2014; Cocucci et al., 2017; Kell, 2021).

Estimating the fraction of drugs that enter cells via transporters relative to those passing directly through the membrane requires a systematic approach. Previous large-scale studies have used gene knockout libraries or CRISPR screens to determine how the disruption of individual transporter genes shifts drug sensitivity. One study used the yeast gene knockout collection to demonstrate that presence of transporters is associated with sensitivity to 18 of the 26 tested compounds (Lanthaler et al., 2011). Using the haploid cell line HAP1, an SLC-specific CRISPR knockout library targeting 394 human SLC genes tested the transporter-dependent effects of 60 different drugs. 101 SLC transporters were associated with sensitivity to 47 of the 60 drugs (Girardi et al., 2020). This screening approach was validated by treating the model with the aforementioned survivin inhibitor YM155, demonstrating that expression of SLC35F2 is required for cell sensitivity to the drug; therefore, surviving cells were enriched for guide RNAs targeting this transporter (Winter et al., 2014).

These efforts suggested that the efficacy of at least 70–80% of tested compounds is influenced by the presence of transporters, with their loss leading to increased drug resistance. However, since many of these transporters have essential roles in cellular metabolism, it is possible that metabolic changes in cells lacking specific transporters affect their survival or drug sensitivity in a manner that is independent of compound uptake (Barnes et al., 2020). By contrast, gene knock-out approaches combined with cytotoxicity screening, like those described above, are unable to fully account for potential redundancy in drug uptake between different transporters (Evers et al., 2018) and could, therefore, lead researchers to underestimate their importance (Jindal et al., 2019). The uptake of metformin, for example, has been shown to be mediated by four different transporters in Caco-2 cell lines (Box 1) (Han et al., 2015).

Transport of the antifolate gemcitabine (Box 1) has been associated with several SLCs from both the SLC28 (CNT) family of concentrative nucleoside transporters and the SLC29 (ENT) family of equilibrative nucleoside transporters. Expression of SLC29A1 (also known as ENT1) has been shown to determine the sensitivity of cancer cell lines to gemcitabine (Mackey et al., 1998; García-Manteiga et al., 2003). This indicates that transporter expression levels could have real clinical impact (Farrell et al., 2009). Perhaps unsurprisingly, given the number of potential transporters for gemcitabine, some studies have reported inconsistent results, highlighting the need for a better understanding of gemcitabine uptake (Kawada et al., 2012; Mohelnikova-Duchonova and Melichar, 2013; Carter et al., 2021).

Interestingly, however, pancreatic cancer-associated macrophages secrete pyrimidine nucleosides. Given that gemcitabine enters cancer cells via nucleoside transporters of the SLC28 and SLC29 family, this microenvironment-derived competition for the transporter decreases drug uptake and was shown to lead to drug resistance in pre-clinical models (Halbrook et al., 2019). In the case of transporter-dependent drugs, therefore, the metabolic context of the tumour and the presence or absence of competitive endogenous substrates can be crucial to determining the uptake of a compound and, consequently, its therapeutic efficacy.

However, by understanding the modes of regulation of transporters involved in uptake or efflux of therapeutic compounds, it may be possible to manipulate their expression or activity through pharmacological or dietary intervention to maximise efficacy of the drug in question. This concept has been demonstrated in a proof-of-principle study, in which a small drug-fragment library was screened in a pancreatic cancer cell line for capability to increase sensitivity to gemcitabine, identifying a small molecule that decreased expression of the efflux transporter ABCC2, rendering cells more sensitive to this chemotherapeutic agent (Grixti et al., 2017).

Understanding how a drug enters its target cell is of major importance since it has implications for predicting efficacy. If a drug requires a transporter for entry into a cell, then the level of expression of that transporter will dictate the intracellular concentration that the compound is able to reach (Fig. 3B). Our own work has illustrated this through the identification of SLC16A7 as a transporter for MOG and its analogues (Fets et al., 2018, 2022). MOG chemically stabilises the transcription factor HIF1A by inhibiting prolyl hydroxylases. While MOG-mediated inhibition of prolyl hydroxylases occurred in all tested cell lines, in those with high SLC16A7 expression, the active compound reached sufficient levels to bind several other, lower affinity α-ketoglutarate-dependent enzymes, thereby disrupting glutamine metabolism and causing toxicity (Fets et al., 2018).

Although DMOG is a tool compound and not clinically utilised, our study demonstrates that intracellular concentration can determine the functionality of a compound – too low and intended targets will not be sufficiently engaged, too high and the drug will interfere with unintended targets. This supratherapeutic effect occurs because, at higher concentrations, many ‘specific’ compounds interact with other targets as they reach selectivity limits (Fig. 3B) (Karaman et al., 2008). This means that transporter-dependent uptake could have implications for many targeted therapies, from inhibitors of tyrosine kinases to PARPs (Box 1) to PI3-kinases (Karaman et al., 2008; Lannutti et al., 2011; Knezevic et al., 2016). Similarly, high transporter expression could explain tissue-selective drug accumulation and associated toxicity in non-target tissues (Pagliarusco et al., 2011). Although relevant for all therapeutic compounds, the high inter-patient variability of transporter expression in cancer (Fig. 2B) means that a better understanding of how chemotherapeutic drugs interact with transporters could drastically improve the way we treat patients. If many drugs are, indeed, taken up via membrane transporters, the transporter signature of the tumour of an individual patient could dictate which compounds are able to reach therapeutically relevant intracellular concentrations. Identification of drug–transporter interactions could, therefore, provide functional biomarkers for the efficacy of therapeutic compounds in patients, allow stratification for clinical trials and, potentially, help to understand toxicities in non-target tissues.

Though a relatively understudied group of proteins for many years (César-Razquin et al., 2015), what is already known about transporters and, particularly, SLCs, puts them at the heart of many physiological as well as disease processes (Fig. 4) and makes clear that more research is needed to understand their pharmacological impacts. Over recent years, the central role of transporters in many aspects of cancer has also been acknowledged through systematic screens using CRISPR libraries tailored to SLC and ABC transporter subfamilies, as demonstrated in recent publications (Li et al., 2021) and preprints (Chidley et al., 2023 preprint). These approaches and the tools they generate will expedite our understanding of metabolite transport in both physiological and disease states, particularly cancer, and fast-track the identification of novel SLC targets for therapeutic intervention.

It is clear, however, that transporter expression in cancer is heterogeneous, and this could have important pharmacological implications for many small-molecule therapeutics. In the era of personalised medicine, understanding the influence of transporters on drug import could enable us to develop bespoke treatment regimens based not only on mutations that drive the tumour of a patient, but also tailored to the unique transporter expression profile of that tumour. Though this Review has mainly focused on cancer, the principles of transporter-mediated drug uptake could, of course, be applicable to many therapeutic interventions and should, therefore, be considered as part of drug development and toxicity studies. While the expression of membrane transporters may not be as heterogeneous outside of the cancer space, their impact on drug efficacy though SNPs or drug-drug interactions are still highly relevant to a broad range of pharmacological agents.

We thank members of the Drug Transport and Tumour Metabolism group for their feedback on the manuscript.