Small‐Molecule Inhibition of Glucose Transporters GLUT‐1–4

Abstract Glucose addiction is observed in cancer and other diseases that are associated with hyperproliferation. The development of compounds that restrict glucose supply and decrease glycolysis has great potential for the development of new therapeutic approaches. Addressing facilitative glucose transporters (GLUTs), which are often upregulated in glucose‐dependent cells, is therefore of particular interest. This article reviews a selection of potent, isoform‐selective GLUT inhibitors and their biological characterization. Potential therapeutic applications of GLUT inhibitors in oncology and other diseases that are linked to glucose addiction are discussed.


Introduction
Hyperproliferation is widely associated with dysregulated energy metabolism in order to fuel growth and cytokinesis. Thereby it is linked to variousd iseases such as cancer, [1] autoimmune diseases, [2] and fibrosis. [3] The alterede nergy metabolism of cancerw as first investigated by Otto Warburg in 1924. [4] He observed that cancerc ells perform glycolysis and ferment the thereby generated pyruvatet ol actate irrespective of oxygen availability ( Figure 1B). This phenomenonw as termeda erobic glycolysis, or the Warburg effect, and yields about 4mol ATPp er mol glucose. [1] Nonmalignant cells fuel pyruvatei nto the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS) in the presence of oxygen in order to generate approximately 36 mol ATPp er mol absorbed glucose( Figure 1A). The reason for this adapted metabolism remains am atter of debate (e.g.,s ee Liberti and Locasale [5] ).
Ah igh rate of glycolysis in cancer depends on key glycolytic enzymesa nd proteins, especially the glucose transporters GLUT-1/GLUT-3, hexokinase 2a nd pyruvate kinase 2. [6] The increased absorbance of glucose by overexpression of GLUTs contributed to the development of positron emission tomography using 18 F-labeled 2-deoxy-d-glucose, which is used to visualize tumors within the patient'sb ody. [7] Ta rgeting the first rate-limiting step of glycolysis promises to be an effective strategy to limit glucoses upply.T od ate, 14 dif-ferentG LUT isoformsa re known that are subdivided into three distinct protein classes according to their phylogenic homology (e.g.,s ee Barron et al. [8] ). Each GLUT isoform has au nique tissue distribution and substrate specificity and fulfills aspecific physiological function. TheG LUT isoforms GLUT-1 to -4 (class I) were investigated most intensively,w ith ap articularf ocus on GLUT-1. [9] However,n oG LUT inhibitor has been advanced to clinical studies. [10] This review gives an overview of the most promising GLUT inhibitors that were developed within the last 20 years with aview to the treatment of cancer and other disorders.

Discoveryoft he Most Potent GLUT Inhibitors
The fungal metabolite cytochalasin Bd ecreasedt he glucose supplyo fc ancerc ells, which originally spurred the interesti n Glucose addiction is observed in cancer and other diseases that are associated with hyperproliferation. The development of compounds that restrict glucoses upply and decrease glycolysis hasg reat potentialf or the developmento fn ew therapeutic approaches. Addressing facilitative glucose transporters (GLUTs), which are often upregulated in glucose-dependent cells, is therefore of particulari nterest. This article reviews a selection of potent, isoform-selective GLUT inhibitors and their biological characterization. Potential therapeutic applications of GLUT inhibitors in oncology and other diseases that are linked to glucose addictionare discussed. this mode of action. [11] Cytochalasin Bi nhibited the uptake of [ 14 C]2-deoxy-d-glucose ([ 14 C]2DG) in N1S1-67 cells and its incorporation in lactate with an IC 50 value below 4 mm in an oncompetitivem anner (Figure 2, Ta bles1and 2). [11,12] In human erythrocytes,c ytochalasin Bi nhibited the uptake of [ 14 C]2DG with an IC 50 value of 0.52 mm. [13] It targets GLUT-1 to -4 but not GLUT-7 (Table 2). [14] Furthermore, cytochalasin Bi nhibited the growth of murine B16F10 cells with aG I 50 value of about 0.4 mm as determined after four days of treatment (Table 2). [15] Although cytochalasin Bp otently inhibits actin polymerization, which restricts the therapeutic applicability of the natural product, it is often used as ac ontrol compound in metabolic studies. [16] Sincet he discoveryo fc ytochalasin B, several GLUT inhibitors with varying GLUT isoform selectivity have been described including natural products, non-natural small molecules and peptidea nalogues (e.g.,s ee Granchi et al. [10b] ). However,o nly af ew potent compounds (IC 50 < 1 mm)w ere identified (Figure2,T ables 1a nd 2).

GLUT Isoform Selectivity Profiles to Target Cancer
GLUTse xhibit at issue-specific distribution and selective expression in cancer. [8,32] GLUT-1 is expressed at high levelsi n most cancers, and GLUT-3 is predominantly found in the brain, arguing for the development of GLUT-1-selectivec ompounds. [22a] However,G LUT-3 is also overexpressed in numerous additional cancert ypes (beyond glioblastoma that originates from nervoust issue [32b] )s uch as breast and endometrial cancer, head and neck tumors, colon cancer,p ancreatic cancer, nonsmall cell lung cancera nd thyroid carcinomas. [33] Vander Heiden proposed that targeting GLUT-3 might enlarget he therapeutic window of glucoseu ptake inhibitors, as GLUT-3 is only expressedi nas mallf raction of somaticc ells (mainly neurons), but is overexpressed in many cancers. [6] BAY-876 is selective for GLUT-1, [22a] buti ti su nclear if the GLUT-1 isoform selectivity is necessary or sufficient. Chromopynone-1, glutor and glupin target the isoforms GLUT-1 and GLUT-3 and DLD-1 cells upregulate GLUT-1 and GLUT-3 after 24 ha nd 48 hw hen cultured under hypoglycemic conditions. [29,30] This adaptation mechanism was mimicked by the treatment with 0.5 mm glutor and 0.5 mm glupin, respectively. [29,30] GLUT4 mRNA stayed unaltered and GLUT2 mRNA was not detectable in these cells, indicating al ow relevance for this isoform under hypoglycemicc onditions. [29,30] Similar outcomes were observed previously in glucose-deprived neuronal rat cells and in MCF7 and HeLa cells cultured under reduced (2.5 mm)g lucose concentration. [34] The increased expression of GLUT-3 might be an atural rescue mechanism of neuronal cells to ensure glucose uptake in ah ypoglycemice nvironment. Because GLUT-3 has the highest affinity for glucosea mong the GLUT isoforms (K M (2DG) = 1.4 mm), [35] cells expressing GLUT-3 have an advantage in competing for glucosew ith the surrounding tissue. Hence, aG LUT-1-/-3-selective inhibitor may be necessary in order to completelyb lock the glucose uptake of cancer cells.

Combination Studies
Many cancers show high metabolicp lasticity,b ecause mitochondriaa re usually still functional andc an use alternative nutrientsf or energy production and biosynthesis. [37] To explore synergistic targeting of severalm etabolic pathways, the glucose uptake inhibitor glutor was combined with CB-839, [38] a small-molecule inhibitor that targetst he kidney glutaminase isoform which is overexpressed in many cancers,t os uppress the growth of HCT116 cells. [29] The combination of glutor with CB-839 decreased the GI 50 value of glutor from 428 nm (0 mm CB-839)b ya bout 40-fold to GI 50 = 10 nm (5 mm CB-839). [29] Inhibition of the glutaminase disrupts the supply of a-ketoglutarate to the TCA cycle andt herefore interferes with an alternative metabolic pathway for energy production.T he availability of the amino acid aspartate has also as trong impact on cell survival, as it influences the dependence of cell on glutamine and couldh ence offer another approachf or ac ombinatory treatment. [39] Combining chemotherapeutic agentsw ith glucose uptake inhibitors has already led to promising results. Ar easonc ould be that mostc hemotherapeutic agents elevate reactive oxygen species (ROS) levels and thereby influence redox status of the cancerc ell. [40] Treatmento fM CF-7 breast cancerc ells with WZB117 partially restored sensitivity of the cellst oward the chemotherapeutic agent adriamycin. [18j] WZB117 has also been successfully applied together with 5-fluorouracil on resistantc olon carcinomas (HCT116), which can be mostp robably explained with an observed GLUT-1 upregulation in 5-fluorouracil-resistant and-treated colon cells. [18p] The GLUT-1-selective inhibitor BAY-876 enhanced the response of cisplatin-treated esophageal squamous cell carcinomaw ith respect to cell proliferation. [25] Radiation of at umor acts through creating double-strand breaksi nD NA as well as through cellular water radiolysis, which creates ROS. [41] Hence, ac ombinatorial treatment of radiotherapy and glucose uptake inhibition might offer ap romising opportunity to target cancerm ore efficiently. An increase in GLUT-1 expression and higher glycolytic activity was observed upon radiotherapy treatment and in radiotherapy-resistant breast cancerc ells. [18l] The authors observedt hat simultaneous treatment of breast cancer cells with WZB117 sensitized the resistantc ells to radiotherapy. [18l] The simultaneous treatmento fh epatocellular carcinoma with 2DG and with kinase inhibitor sorafenib also showed promising resultsi nt argeting sorafenib-resistantp opulations in vitro and in vivo. [42] Overall,t he inhibition of glycolysis, for example, by glucose transporter inhibitors, seems to be highly effective to sensitize cancer to diverse treatment approaches.

Possible Applications beyond Oncology
Aerobic glycolysis and increased glucose dependence are also characteristic for inflammatoryd iseases (Figure 3). CD4 + Tcells switch from fatty acid b oxidation in the resting state to aerobic glycolysis after activation. Interestingly,G LUT-1-deficient CD4 + Tcells were unable to grow,p roliferate, survivea nd differentiate to Teffector cellsa fter activation. [2a] Tcells that upre- gulate aerobic glycolysisa re involved in the establishment of inflammatory boweld isease, graft-versus-host disease and systemic lupus erythematosus. [2,43] Notably,i ns ystemic lupus,a utoreactive CD4 + Tcells upregulate oxidative phosphorylation along with glycolysis, and combinatorialt reatment with 2DG and metformin showed promising results in mouse models. [2b] Also, HIV-infected patients hold al arge number of CD4 + Tcells, which overexpressGLUT-1. [44] Hyperplasia-associated diseases, such as psoriasis and fibrosis, exhibit uncontrolled cell proliferation and increased GLUT-1 levels, offering potential for modulation by treatment with glucose import inhibitors (Figure 3). [3,45] Age-related macular degeneration (AMD)isc haracterizedbyo cular neovascularization. Because increased levels of glycolysis have been observed in endothelial cells and AMD patients exhibit increased lactate/ pyruvater atios, restriction of glycolysis might be ap romising therapeutic approach to interferew ith the endothelial proliferation. [46] Intracellular bacteria and parasites may manipulate the host cell's metabolism to increase glycolysis ( Figure 3). The bacteria Chlamydia trachomatis, [47] Chlamydia pneumoniae, [48] Mycobacterium tuberculosis, [49] Brucella abortus [50] and Legionellap neumophila [51] have been reported to induce aW arburg-like phenotype of their host cells. Treatment with 2DG reduced the replication of L. pneumophila inside human macrophages. [51,52] The malaria parasite Plasmodium falciparum replicates inside erythrocytes and increases GLUT-1 expression of the host cells in order to fuel its own metabolism. [53] Recently,W ei et al. successfully applied WZB117t op lasmodium-infected erythrocytes, which induced oxidative stress and apoptosis. [18e] Viral infections also lead to an adaptation of the energy metabolism of the host cells towardaerobic glycolysis. Cells infected by rhinovirus increase GLUT-1 expression and release additional glucose from their glycogen storage.T reatmentw ith glycolysis inhibitor 2DG reverts the metabolism to lipogenesis. Thus GLUT inhibitors might open an alternative opportunity to address viral infection. [54] Furthermore,t he transformation of progenitor cells to differentiatedc ells often involves as witch in the metabolic phenotype of the cell (Figure 3). Izumi et al. recently demonstrated that the treatment of connective tissue with 2DG drives the differentiation to tendon cells and inhibits chondrogenesis, which is associated with poor tendon healing. [55] Applying small molecules in the field of directed differentiation might offer tremendous potential.

Summary and Outlook
Aerobic glycolytic phenotypes have been observed in multiple diseases.T argeting alteredg lucoseu ptake and metabolism with appropriate tool compounds, such as GLUT inhibitors, could yield new insighti nt he diseases and pave the way for novel therapeutics trategies. Eleven distinct compound classes that inhibitg lucose uptake with sub-micromolar potency and target the glucoset ransporters with different GLUT isoform selectivity have been reported.T hese compounds have been furtherc haracterizedb iologically and provide av aluablet ool compound platform to furtheri nvestigate glucose metabolism in differentdisease models.
For less-well-characterized GLUTi soforms of class II and III, selectivet ool compounds are not yet available. As these isoforms are also overexpressed in some cancers and other diseases, [56] this field of research could offer additional opportunities for the treatment of disease.