Discovering Amino Acid Transporter Inhibitors: History
Please note this is an old version of this entry, which may differ significantly from the current revision.

Amino acid transporters are abundant amongst the solute carrier family and have an important role in facilitating the transfer of amino acids across cell membranes. Because of their impact on cell nutrient distribution, they also appear to have an important role in the growth and development of cancer. Naturally, this has made amino acid transporters a novel target of interest for the development of new anticancer drugs. 

  • inhibitors
  • amino acids
  • cancer
  • amino acid transporters
  • solute carriers

1. Introduction

The Solute Carrier (SLC) gene family encodes structurally diverse groups of membrane proteins that facilitate the transport of nutrients, metabolites, vitamins, and minerals, as well as several small molecular drug substances across cell membranes. Solute carriers are the second largest family of membrane proteins, with a current number of 448 members across 66 different subfamilies [1]. Despite their implications in various diseases, SLCs are surprisingly unexplored as pharmacological targets compared to more popular targets such as ion channels, G protein-coupled receptors, and kinases [2,3]. One disease where SLCs have gained some interest is cancer, where they function as facilitators of cancer cell growth or, on the contrary, as tumor cell suppressors [4,5,6,7,8]. Due to their high proliferation, cancer cells need substantial amounts of nutrients to fuel their energetic and biosynthetic demands, making SLCs important metabolic gatekeepers. In the 1920s, Warburg and colleagues discovered that cancer cells take up immense amounts of glucose when compared to cells at rest [9]. The high glucose consumption was used for glycolysis as well as lactic acid fermentation, even in the presence of oxygen, coining the terms “aerobic glycolysis” and the “Warburg effect”. This uptake of glucose is mainly facilitated by the increased expression of GLUT1 (SLC2A1) in various cancer cells [10]. Hence, GLUT1 is central in the diagnosis and detection of tumors, where the radiolabeled glucose analog 18F-fluorodeoxyglucose (FDG), which is a substrate of GLUT1, is used to detect and quantify the glucose utilization of tumors [11]. Other SLCs have also been shown to be useful in cancer detection. For instance, the sodium/iodide symporter (NIS, SLC5A5) is responsible for the uptake of radioisotopes of iodide in thyroid cancers [12]. It has also been shown that 11C-sarcosine might have a use as a new PET imaging probe due to its accumulation in prostate cancer tissue [13]. This uptake of sarcosine was first believed to be mainly PAT- (SLC36) mediated, but it appears that SNAT2 (SLC38A2) also plays an important role in sarcosine uptake in prostate cancer cells, at least in vitro [14].
The human genome encodes at least 60 SLCs that can transport amino acids, found across 11 different SLC subfamilies. Some of these have also been shown to have an importance in the metabolism in cancer cells, given the role of amino acids in protein synthesis, activation of the cell growth regulator mTORC1 (mammalian target of rapamycin complex 1), and various other metabolic pathways (as reviewed in [15]). In particular, the role of glutamine transporters has been examined. Where the Warburg effect is a term for the rewired glucose metabolism, the term “glutamine addiction” points to the increased glutamine consumption of cancer cells [16,17]. Glutamine has an important role in providing its γ-nitrogen for nucleotide synthesis, converting it to glutamic acid, which in turn can donate its α-nitrogen to the synthesis of non-essential amino acids [18]. Furthermore, glutamine can be used to replenish the tricarboxylic acid (TCA) cycle with metabolites, which in some cancer cells is sparsely supplied by glucose-derived pyruvate due to the Warburg effect. This anaplerotic reaction of glutamine is known as glutaminolysis, where glutamine is first converted into glutamate and then into α-ketoglutarate, which is a substrate of the TCA cycle. Glutaminolysis can then provide cancer cells with critical precursors for synthesizing non-essential amino acids, nucleotides, and lipids (as reviewed in [19]). It is also believed that glutamine has an important role in being an exchange substrate for amino acid antiporters, such as LAT1 (SLC7A5), where it is effluxed to drive the uptake of essential amino acids such as leucine [20]. This, of course, has made glutamine metabolism and uptake an interesting possible target in the discovery of anticancer drugs. Various SLCs can transport glutamine, but specifically, ASCT2 (SLC1A5) has gained interest as a target for inhibiting glutamine uptake, given its general overexpression in tumor tissue and its use as a prognostic biomarker [21]. However, as pointed out by Bröer and colleagues, ASCT2 is an amino acid exchanger and, therefore, needs to trade the uptake of glutamine with the efflux of another amino acid substrate [22]. The SLC38 transporters SNAT1 (SLC38A1) and SNAT2 (SLC38A2) are symporters that utilize a Na+ gradient to drive the uptake of glutamine as well as other amino acids in cells. Bröer et al. refer to these amino acid transporters as “loaders” because they facilitate the net uptake of amino acids by coupling to an ion gradient. On the other hand, transporters like LAT1 and ASCT2 are called “harmonizers” because of their role as exchangers to balance out the amino acid pool within the cell by trading the accumulated “loader” substrates like glutamine and alanine for other essential amino acids [22]. 

2. Discovering Amino Acid Transporter Inhibitors

When studying molecules that inhibit SLCs, it is important to be familiar with key transporter terminology [26]. The movement of a substrate across a membrane, where the transporter moves from the outward open to the inward open conformation, is called a translocation cycle. The transporter turnover rate is the number of molecules moved by a single transporter per second [27]. Substrates are compounds that bind to and are translocated by the solute carrier. Substrates are generally characterized by their Michaelis constant (Km), which reflects the substrate concentration where half of the maximal transport velocity is achieved. However, when two substrates compete for binding, one can be considered a “competitive inhibitor” of the translocation of the substrate in question. This is due to competitive binding to the substrate binding pocket of the transporter. Calling such substrates inhibitors might be misleading as it only indicates that the substrate can inhibit the translocation of other substrates. True inhibitors are molecules that bind to the transporter without being translocated themselves. This can be due to competitive inhibition at the substrate binding site. Binding to this site is called orthosteric, whereas allosteric refers to a site other than the substrate binding site. Un-competitive inhibition is when the inhibitor only binds the substrate-transporter complex. Non-competitive binding is when the inhibitor equally binds the free transporter and substrate-transporter complex. If the inhibitor binds the free transporter and the substrate-transporter complex with unequal affinity, this is called mixed-inhibition. Another commonly used term is ‘ligand’, which generally refers to any small molecule that is able to bind to a macromolecule like a transporter. Whether the molecule is translocated or not is often not addressed when the term ligand is used. To compare inhibitors, affinity or potency are used, which are terms that are generally more defined in receptor pharmacology. Affinity reflects the binding of the inhibitor to the transporter and is usually quantified through the dissociation constant (Kd), which is the inhibitor concentration where 50% of the transporters are bound by the inhibitor. Potency reflects the ability of the compound to inhibit the transporter function and is usually quantified by the IC50, which is the inhibitor concentration that leads to 50% inhibition of the transport activity. The IC50 is highly dependent on the substrate concentration used in the assay to determine this value. An absolute value that is easier to compare is the inhibition constant (Ki), which is independent of the substrate concentration. The relation between the IC50 and the Ki is dependent on the different modes of inhibition (i.e., competitive, un-competitive, non-competitive, and mixed), which is seen in the so-called Cheng-Prusoff equations [28]. Generally, the binding of the inhibitor to the transporter is also indicative of its ability to inhibit, which is why affinity is sometimes used synonymously with potency. Selectivity refers to the ability of the compound to selectively bind to a given transporter without binding to other related transporters. Often, selectivity is a matter of differences in affinity relative to the concentration of inhibitor the transporter is exposed to, as compounds tend to bind to multiple targets at higher concentrations. For instance, the selective serotonin uptake inhibitor sertraline binds the serotonin transporter (SERT, SLC6A4) with a Kd of 0.29 nM but also binds the norepinephrine transporter (NET, SLC6A2) with a 1400 times lower affinity (Kd = 420 nM) [29]. So, the typical goal in the development of SLC inhibitors is to find a compound that is non-translocated, has great inhibitory potency, and is selective to limit any unwanted effects due to binding to other targets. It is generally more difficult to find transporter inhibitors with similarly low potencies as compounds targeting receptors, which makes potencies in the submicromolar range a great accomplishment in the transporter field. Since some inhibitors might work by inhibiting the transporter at an intracellular site, the inhibitor’s ability to cross the plasma membrane by preferably passive diffusion is an important property of transporter inhibitors. Inhibitors with greater lipophilicity have higher permeabilities across lipid barriers but at the detriment of aqueous solubility. Higher lipophilicity can lead to an increased potency since lipophilic compounds prefer to stick to protein cavities rather than being fully exposed to water. However, lipophilicity is also commonly associated with decreased selectivity, increased toxicity, increased protein binding, and increased metabolic clearance (reviewed in [30]). So, in addition to having a high potency and selectivity, the ideal inhibitor should have sufficient solubility and oral bioavailability (>85%), low metabolism and protein binding, and be cleared really by glomerular filtration without any secretion or re-absorption.
Several strategies and methodologies can be pursued to discover SLC inhibitors. The more traditional approach is through high-throughput screening (HTS). Here, a large library of compounds is screened for their inhibition using various functional in vitro assays, like an uptake assay of a radiolabeled or fluorescent substrate. A nice overview of available cell-based assays for the study of SLCs can be found in the following review [31]. Of course, this approach requires a large library of chemically diverse compounds as well as a scalable assay that can screen many compounds at a time. An example of this approach has recently been used to discover inhibitors of SNAT2 by screening a library of 33,934 compounds [32]. However, a more focused screen can be implemented if resources are limited or if only low-throughput assays are available. This approach usually screens structural analogs of already known substrates or inhibitors in the pursuit of more potent ligands. Another and more modern way to narrow down the search is to implement in silico methods. Broadly speaking, computer-aided drug design (CADD) can be categorized into two main categories: ligand-based drug design and structure-based drug design ([33] for review). The ligand-based approach uses known ligands to generate a model that can help predict novel ligands. In its simplest form, this can be a similarity search that finds chemically similar molecules to known ligands. More sophisticated is a pharmacophore model, which is a spatial representation of the key chemical features of a ligand and its possible interactions with its biological target. The pharmacophore model can then be used to query virtual compound libraries to find other compounds that fit the pharmacophore model and are possible ligands. A ligand-based approach has been used to study the human proton-coupled amino acid transporter 1 (hPAT1) and helped identify a novel scaffold for hPAT1 substrates [34]. Structure-based drug design, on the other hand, focuses on the structure and binding site of the biological target and thus relies on a solved 3D structure of the protein (e.g., through X-ray crystallography or cryo-EM). If no structure is available, homology modeling can be used, where a homologous protein structure is used as a template to model the protein of interest. Using the protein structure, a virtual compound library can be screened by docking molecules into the binding site. The resulting docking score can then be used to predict which compounds are likely to be ligands. This often requires a protein structure that includes a ligand to define the binding site; otherwise, this needs to be defined through other means. Virtual screening through molecular docking is known to be most effective when screening ultra-large compound libraries [35]. To make docking of larger libraries more efficient, machine learning can be included to guide the docking, which has recently been demonstrated in the discovery of P-glycoprotein inhibitors [36]. Structure-based drug design can also implement molecular dynamic simulations that simulate both the movement of the protein and ligand, contrasting docking that usually treats the protein as being rigid. One instance of structure-based drug design for amino acid transporters can be seen for LAT1, where docking to a homology model was used to find possible inhibitors that could be tested experimentally [37]. It is uncommon to use only one of these approaches exclusively, as they often work best in collaboration. For instance, in silico methods can be used to generate a more focused screen, and structure-based drug design can also be used to generate pharmacophore models that can be compared to ligand-based ones.
However, there are some obvious concerns and limitations that must be taken into consideration that make it more difficult to develop pharmacological agents that inhibit SLCs when compared to other protein targets. For most SLCs and also for AA transporters, limited knowledge about their ligands or other compounds that might interact with these proteins is available; some are even described as “Orphan” meaning that the endogenous substrate is not known. This limits the chemical tools available for studying these proteins [2]. Moreover, there is a lack of structural information on SLCs due to the difficulty in crystallizing membrane proteins. However, in recent years, several SLCs have been deorphanized, and there has been an increase in the amount of available SLC crystal structures, several of which are amino acid transporters [38]. The recent development of AI has also allowed AlphaFold2 to accurately predict the structure of proteins that lack crystal structures, which can be very helpful for structure-based drug design against SLCs [39]. So, this is a prime age for exploring amino acid transporters and SLCs in general as pharmacological targets and helping expand researchers' knowledge of their function, structure, and chemical probes.

This entry is adapted from the peer-reviewed paper 10.3390/pharmaceutics16020197

This entry is offline, you can click here to edit this entry!
Video Production Service