Cellular binding and uptake of fluorescent glucose analogs 2-NBDG and 6-NBDG occurs independent of membrane glucose transporters

Abstract

The classical methods for determining glucose uptake rates in living cells involve the use of isotopically labeled 2-deoxy-D-glucose or 3-O-methyl-D-glucose, which enter cells via well-characterized membrane transporters of the SLC2A and SLC5A families, respectively. These classical methods, however, are increasingly being displaced by high-throughput assays that utilize fluorescent analogs of glucose. Among the most commonly used of these analogs are 2-NBDG and 6-NBDG, which contain a bulky 7-nitro-2,1,3-benzoxadiazol-4-yl-amino moiety in place of a hydroxy group on D-glucose. This fluorescent group significantly alters both the size and shape of these molecules compared to glucose, calling into question whether they actually enter cells by the same transport mechanisms. In this study, we took advantage of the well-defined glucose uptake mechanism of L929 murine fibroblasts, which rely exclusively on the Glut1/Slc2a1 membrane transporter. We demonstrate that neither pharmacologic inhibition of Glut1 nor genetic manipulation of its expression has a significant impact on the binding or uptake of 2-NBDG or 6-NBDG by L929 cells, though both approaches significantly impact [³H]-2-deoxyglucose uptake rates. Together these data indicate that 2-NBDG and 6-NBDG can bind and enter mammalian cells by transporter-independent mechanisms, which calls into question their utility as an accurate proxy for glucose transport.

1. Introduction

Glucose is a common catabolic fuel for the majority of mammalian cell types. Because the rate of glucose uptake is a key indicator of metabolic status, it is frequently assessed in studies across fields of biomedical science, including those focused on normal physiology and disease conditions such as cancer [1]. In vivo evaluation of glucose uptake is dominated by positron emission tomography using the tracer 2-[¹⁸F]-fluoro-2-deoxy-D-glucose (FDG-PET), which has become a mainstay for detection of tumor metastasis in medical oncology [2]. Biochemical methods for routine laboratory measurements of glucose uptake in living cells similarly rely on isotopic labeling and are considered the gold standard for this assay.

The most common in vitro assays for glucose uptake utilize radiolabeled forms of either 3-O-methyl-D-glucose (3-OMG) or 2-deoxy-D-glucose (2-DG) [3]. 3-OMG is not further metabolized and has the advantage of measuring only the transport process; however, the rapid exchange of glucose requires very short incubation times to accurately measure transporter activity, limiting its application in some studies. It has, therefore, become more common to utilized 2-DG to measure glucose uptake. This analog is transported and undergoes the first metabolic step of phosphorylation, which traps the radiolabeled 2-DG-6-phosphate in most cells. In nearly all cases, the transport step is significantly slower than the phosphorylation step, which allows radiolabeled 2-DG accumulation to serve as an accurate measure of transporter activity [4].

Despite their time-tested utility, the classical methods cited above are increasingly being displaced by fluorescence-based assays that utilize analogs of glucose in which one of the hydroxyl groups is replaced with a fluorophore that is itself similar in size or larger than glucose [5]. Among the various fluorescent analogs of glucose that are commonly used for these assays, 2-NBDG has emerged as the most popular in the literature [6][7][8]. This molecule has a bulky 7-nitrobenzofurazan fluorophore attached to D-glucosamine in place of the endogenous 2-hydroxy group. Though early work with 2-NBDG clearly demonstrate that it is taken up by bacteria and mammalian cells alike, very little effort has been expended to demonstrate that it enters cells by a mechanism that accurately mimics actual glucose uptake. Studies that have shown some impact of D-glucose on the rate of 2-NBDG uptake in a single cell line have been deemed sufficient to demonstrate its broader viability as a measurement of glucose uptake across all mammalian cell lines [9]. However, given the known mechanism of hexose transport across the plasma membrane, which involves both electrostatic and steric selectivity, it seems intuitively unlikely that 2-NBDG or any other fluorescent analog is actually transported in similar fashion [10][11].

One of the challenges in determining the mode by which 2-NBDG or other fluorescent analogs are taken up by mammalian cells is the complexity of glucose uptake mechanisms available to many cell types. Three distinct classes of eukaryotic sugar transporters have been characterized: (1) the recently discovered SWEET family, primarily responsible for intra- and intercellular transport; (2) the well-characterized passive glucose transporters of the GLUT/SLC2A family; and (3) the active sodium-glucose linked symporters of the SGLT/SLC5A family [12][13]. Fourteen different members of the SLC2A family and six members of the SLC5A family exist within the human genome, many of which are broadly distributed in expression across different tissues [14][15]. The overlapping expression of these transporters in any given cell type make the determination of which one is predominant in glucose uptake a challenging matter to dissect experimentally. The development of selective inhibitors that are able to distinguish among different glucose transporters has provided some help in this respect, as has the advent of genetic technologies capable of selective silencing the expression of single genes [16][17]. For the most part, however, the specificity regarding which transport system is being used for glucose absorption by a given cell type is rarely assessed in any significant detail when glucose uptake is being evaluated.

Of the various SLC2A family members that have been characterized in the literature, SLC2A1/GLUT1 stands out as perhaps the best understood due to its broad expression across different cell types and well-defined transport kinetics [11][18]. Furthermore, the protein structure for this transporter has been defined by X-ray crystallography, yielding additional insights into the mechanism by which it transports glucose across the plasma membrane [19]. The prominent role that GLUT1 plays in the hypoxic response and cancer metabolism has spurred further interest in its expression and regulation, as well as the generation of selective inhibitors that can block its transport function without significantly affecting other physiologically important glucose transporters [20][21][16]. In addition to their therapeutic potential, these drugs present helpful tools for dissecting the role of GLUT1 in basic research applications.

In this study, we systematically evaluated the ability of the mouse Slc2a1/Glut1 to bind and transport 2-NBDG and its structural isomer, 6-NBDG, in L929 fibroblasts. Compared to many other mammalian cell types, the L929 mouse fibroblast line is relatively simple in its mode of glucose uptake due to the fact that expresses only the Glut1 facilitative glucose transporter [22]. We leveraged this simple system to evaluate the impact of pharmacologic inhibition and genetic manipulation of Glut1 on 2-NBDG and 6-NBDG transport kinetics. Our results suggest that L929 cells bind to both glucose analogs and actively accumulate 2-NBDG, as expected, but do so by undefined mechanisms that largely fail to involve Glut1 or any other glucose transporter. These findings suggest that the uptake of fluorescent glucose analogs should be interpreted with great caution since they may not represent an accurate proxy for glucose transport in mammalian cells.

2. Materials and Methods

2.1. Chemical reagents

The small molecule inhibitors cytochalasin B, BAY-876, and WZB-117 were obtained from Sigma-Aldrich (St. Louis, MO). Cytochalasin B (1 mg/mL) was dissolved in ethanol; BAY-876 (100 μM) and WZB-117 (1 mM) were dissolved in DMSO. 2-NBDG and 6-NBDG (2- or 6-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-D-glucose) were obtained from Cayman Chemical (Ann Arbor, MI) and dissolved in ethanol to 20 mM. AlexaFluor-647 labeled dextran (m.w. ~10,000) was obtained from ThermoFisher (Waltham, MA) and dissolved in water at 10 mg/mL. All stocks of diluted compounds were stored at −20°C and protected from light prior to use in assays. The rabbit monoclonal antibody for Glut1 was obtained from Abcam (Cambridge, MA) and the mouse monoclonal antibody for β-actin was obtained from Cell Signaling Technology (Danvers, MA)

2.2. Cell culture

L929 fibroblasts were obtained from the lab of Dr. Larry Louters (Calvin University, Grand Rapids, MI). The L929/GB-M6 stable transgenic line is a monoclonal derivative of L929 cells that has been previously described [23]. HK2 immortalized human kidney cells were obtained from American Type Culture Collection (ATCC, Manassas, VA). Both lines were tested in-house for mycoplasma contamination by PCR-based methods and found to be negative. The L929 and L929/GB-M6 cell lines were maintained in low glucose (1 g/L) DMEM supplemented with GlutaMAX (Thermo-Fisher) and 5% fetal bovine serum. The HK2 cell line was maintained in standard Roswell Park Memorial Institute 1640 (RPMI-1640) medium supplemented with GlutaMAX (Thermo-Fisher) and 5% fetal bovine serum (Atlanta Biologicals, Flowery Branch, GA). Cells were split 2–3 times per week to maintain log-phase growth. Detachment of cells from culture plates was performed with TrypLE Express reagent (Thermo-Fisher) after washing cells with sterile DPBS (ThermoFisher). For all assays, cells were seeded at the specified densities after counting with a standard hemocytometer.

2.3. Quantification of relative cellular ATP levels

L929 cells were seeded to white-wall, clear-bottom 96-well culture plates (Greiner, Kremsmünster, Austria) in the indicated culture media at a density of 1.0 × 10⁴ cells/well. At 18–24 hours after plating, media was removed from the plate by inverted shaking and cells were washed with DPBS to remove residual media. Fresh media containing the indicated drugs was replaced in a volume of 100 μL per well and cells were then incubated for two hours. The relative concentration of ATP was measured using a Cell Titer Glo assay (Promega, Madison, WI), which produces an ATP-dependent luminescent signal proportional to cellular ATP concentrations. Luminescent values were captured with a Synergy H1 plate reader (BioTek, Winooski, WT) and exported to Microsoft Excel (Redmond, WA) for data analysis. Average luminescent values from quadruplicate measurements of each condition were normalized against the average value of control cells treated with vehicle only. Processed data were exported to Prism 6 software for line fitting and calculation of IC₅₀ values.

2.4. Flow cytometry Analysis of NBDG Binding or Uptake

L929 cells were plated to 12-well dishes at a density of 1.5 × 10⁵ cells/well and allowed to adhere for 18–24 hours under normal culture conditions. The following day, media was aspirated and cells were treated for the indicated times at 37°C with 0.3 mL/well of Hank’s Balanced Salt Solution (HBSS, ThermoFisher) containing the 2-NBDG or 6-NBDG analogs or AlexaFluor-647 labeled dextran at the concentrations indicated for each figure. After treatment, cells were rinsed with PBS and detached from the plate in 0.2 mL of TrypLE Express (Thermo-Fisher). Cells were subsequently diluted with 1 mL PBS and filtered to achieve a single cell suspension before pelleting in polystyrene cytometry tubes (BD Biosciences, San Jose, CA).

In circumstances where only binding of 2-NBDG or 6-NBDG was being assessed, cells were first detached from plates and incubated with the compounds at 4°C to prevent uptake. Assessment of total cellular binding sites was performed in L929 cells that were first fixed in 0.5 mL Cytofix solution (BD Biosciences, San Jose, CA) at 4°C with constant shaking. After 25 minutes of fixation, the cell suspension was diluted 1:1 with cold PBS and pelleted again. The resulting pellet was resuspended in 0.5 mL Cytoperm solution (BD Biosciences) and incubated at 4°C with constant shaking for 30 minutes to permeabilize cells. Cells were pelleted again and then incubated in 100 μL of Cytoperm solution containing 2-NBDG or 6-NBDG.

After treatment and detachment, cells were incubated on ice in HBSS prior to analysis with a FACScalibur flow cytometer (Becton Dickinson Company, Franklin Lakes, NJ). Laser intensity for the FL1 channel was set to center the negative control cell population (untreated L929 cells) distribution at a mean fluorescent intensity (MFI) of 10¹ units. No fewer than 2.0 × 10⁴ cells were captured for each sample, and each condition was measured in triplicate to obtain an averaged MFI for each cell population.

2.5. Plate Reader Analysis of NBDG Uptake

L929 cells were seeded to black-wall, clear-bottom 96-well culture plates (Greiner) in the indicated culture media at a density of 1.0 × 10⁴ cells/well. At 18–24 hours after plating, media was removed from the plate by inverted shaking and cells were washed with DPBS to remove residual media and then incubated in HBSS (100 μL/well). Cells were then treated for the indicated times at 37°C with 50 μM 2-NBDG or 6-NBDG analogs dissolved in HBSS. At the end of the time course, buffer was shaken off the plate and cells were washed three times with cold PBS before measuring fluorescence with a Synergy H1 plate reader. Fluorescent values for each well were obtained with monochromator settings at Ex/Em:465/540 nm and exported to Prism 6 for data analysis. Average fluorescent values for from quadruplicate biological samples were normalized against the average value of untreated cells.

2.6. Radiolabeled 2-DG uptake assay

Cells were plated to 24-well dishes (Greiner) at a density of 8.0 × 10⁴ cells/well and allowed to adhere and equilibrate to culture conditions for 18–24 hours prior to the specified treatment. Uptake was measured using the radiolabeled glucose analog [1,2-³H]-2-deoxyglucose ([³H]-2-DG). Briefly, the media was replaced with 0.3 mL of glucose-free HEPES buffer [pH 7.4] (140 mM NaCl, 5 mM KCl, 20 mM HEPES, 2.5 mM MgSO₄, 1 mM CaCl₂, 2 mM sodium pyruvate) supplemented with 1.0 mM (0.3 μCi/mL) [³H]-2-DG. After a 15-minute incubation, cells were washed twice with cold glucose-free HEPES. The cells were digested in 0.25 mL of 0.3 M NaOH prior to measuring the [³H]-2-DG uptake. Quadruplicate [³H]-2-DG uptake values were averaged and tested for significance using student’s T-test.

2.7. siRNA-mediated knockdown of Glut1

The siLentFect transfection reagent was obtained from BioRad (Hercules, CA). Lyophilized control and experimental siRNAs for Glut1 knockdown experiments were obtained from Integrated DNA Technologies (IDT, Coralville, IA) and resuspended in the provided resuspension buffer to 10 μM concentration before aliquoting and freezing at −20°C. The universal negative control siRNA (DS-NC1) from IDT was used as a negative control for knockdown assays. Predesigned siRNAs targeted to the 3’-UTR (mm.Ri.Slc2a1.13.1) or the coding sequence (mm.Ri.Slc2a1.13.2) of the mouse Glut1 mRNA were used to knock down Glut1 in L929 cells.

L929 cells were plated to 6-well plates (Greiner) at a density of 5.0 × 10⁵ cells/well and allowed to adhere and equilibrate to culture conditions for 18–24 hours prior to transfection. Transfection solutions were prepared by diluting siLentFect reagent (2.5 μL/mL) and siRNA (50 nM) in equal volumes of serum-free DMEM media and then combining the two solutions in a single tube with gentle mixing by vortex. The solutions were allowed to complex for 20 minutes in a sterile tissue culture hood at room temperature before being added to cells (1 mL/well). Cells were incubated for 4 hours in the serum-free transfection complex before an additional 1 mL of DMEM containing 10% FBS was added to each well. Cells were incubated for a total of 24 hours in transfection media prior to replating for 2-DG uptake, immunoblotting, or flow cytometry assays. All functional assays were carried out at 48 hours after the initial siRNA transfection.

2.8. Protein extraction and immunoblotting

Cells in 6-well plates were rinsed with cold PBS and harvested into 200 μL of lysis buffer (20 mM Tris [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium pyrophosphate, 1 mM sodium glycerophosphate, 1 mM sodium orthovanadate, 0.5% NP40, 0.1% Brij35, 0.1% sodium deoxycholate) supplemented with mammalian cell protease inhibitor cocktail (Sigma-Aldrich). Lysates were homogenized by brief sonication at 30% power on ice and cleared by centrifugation at 10,000 rcf for 10 min at 4°C. Concentration of each lysate was determined by Bradford assay. Equal amounts of protein lysate (50 μg) were separated by reducing polyacrylamide gel electrophoresis and transferred overnight to nitrocellulose membrane using a wet transfer tank (TE62 model; Hoefer, Holliston, MA). Membranes were blocked with 3% non-fat dry milk in Tris-buffered saline containing 0.05% Tween-20 (TBST), and then probed overnight at 4°C with primary antibodies diluted 1:1,000 in TBST with 3% bovine serum albumin (BSA). After washing off unbound primary antibody, the membranes were incubated for 1 hour with goat anti-rabbit-DyLight-800 and goat anti-mouse-DyLight-680 secondary antibodies (Cell Signaling Technology). Membranes were imaged with an Odyssey scanner (LiCor, Lincoln, NE) to produce images that were processed with Odyssey Infrared Imagining software (version 3.0.25) to ensure that signal was in the linear range prior to export as TIFF files.

2.9. Data fitting and statistical analysis

The results of this study were obtained from exploratory experiments aimed at demonstrating the uptake of 2-NBDG and 6-NBDG can occur independent of a membrane transporter in mammalian cells. Each experiment was repeated a minimum of three times to ensure that results could be replicated. All data points used for fitting data to IC₅₀ curves represent the average of quadruplicate values. Regression analysis was carried out in GraphPad Prism 6 software with a 4-parameter, variable-slope dose-response equation to calculate IC₅₀ values, which were reported only in cases where the goodness of fit met a minimum threshold of R² > 0.95. The statistical significance for pairwise comparisons between control and experimental conditions was determined by a two-tailed student’s t-test using Microsoft Excel. Statistical significance is reported for comparisons with p-value < 0.05 as indicated in the figure legends.

3. Results

3.1. L929 cells as a model for glucose uptake

Prior studies have indicated that mouse L929 fibroblasts exclusively express Glut1 as a means of absorbing glucose [22]. To provide support for this contention, we analyzed the sensitivity of L929 cells to the small-molecule inhibitor BAY-876, which is highly selective for Glut1 (IC₅₀ = ~2 nM) over other GLUT1/SLC2A family members [16]. In complete media, the impact of BAY-876 on cellular ATP production is not obvious due to compensation by mitochondrial catabolism of alternative fuels such as pyruvate and glutamine, both of which are found in the DMEM media (Figure 1A). However, in the presence of a saturating dose of the Complex I inhibitor rotenone (1 μM), which completely ablates mitochondrial oxidative phosphorylation, blockade of glucose uptake by BAY-876 completely attenuates cellular ATP production in L929 cells in a dose-responsive fashion (Figure 1A). Regression analysis of the BAY-876 dose-response curve provides an IC₅₀ = 2.5 nM, matching the published IC₅₀ value for Glut1 and consistent with the sole expression of this transporter in L929 cells [16]. Identical treatment of HK2 cells, which are a human cell line derived from the proximal tubule of the kidney, produces an IC₅₀ value that is roughly 100-fold higher in magnitude (IC₅₀ = 199.3 nM), consistent with their co-expression of GLUT1 and GLUT2 (Figure 1B) [24]. The incomplete effect of BAY-876 on cellular ATP production in HK2 cells, which reaches a maximum of about 60%, also provides evidence that these cells express SGLT2, a member of the SLC5A family of sodium-glucose cotransporters that are insensitive to BAY-876 [25]. Together these data demonstrate that L929 fibroblasts are a unique model that allows that analysis of glucose uptake mechanisms exclusively mediated by Glut1.

Figure 1 – Kinetics of 2-NBDG and 6-NBDG uptake by L929 fibroblasts.

(A-B) Murine L929 fibroblasts or human HK2 renal epithelial cells were treated with a two-fold serial dilution (0.2–200 nM) of the selective Glut1 inhibitor BAY-876 for two hours in the presence or absence of the oxidative phosphorylation inhibitor rotenone (1 μM). Cellular ATP content was determined by luminescent assay and normalized to cells treated with vehicle only (0.1% DMSO). Data were fitted to a four-parameter variable slope equation to determine the IC₅₀ values for BAY-876. Error bars represent the standard deviation of quadruplicate values for each concentration of BAY-876. (C-D) Representative flow cytometry histograms (FL1 channel) for live, adherent L929 cells treated with 100 μM 2-NBDG or 6-NBDG for the indicated times. (E) Mean fluorescent intensity values extracted from flow cytometry of L929 cells treated with 100 μM 2-NBDG or 6-NBDG is plotted as a function of incubation time. Error bars represent the standard deviation of triplicate values from separate cytometry runs at each timepoint. (F) Fluorescent intensity of L929 cells treated as above in 96-well format. Absolute values were determined on a fluorescent plate reader and plotted as a function of incubation time with 50 μM 2-NBDG or 6-NBDG. Error bars represent the standard deviation of quadruplicate wells for each treatment time. (G) Representative flow cytometry histograms (FL4 channel) for live, adherent L929 cells treated with fluorescent-labeled dextran (AlexaFluor-647, 10 μg/mL) for the indicated times. (H) Mean fluorescent intensity values extracted from flow cytometry of L929 cells treated with 10 μg/mL AlexaFluor-647 dextran is plotted as a function of incubation time. Error bars represent the standard deviation of triplicate values from separate cytometry runs at each timepoint.

3.2. Kinetics of 2-NBDG and 6-NBDG uptake by L929 cells

Having established L929 cells as a model for glucose uptake by Glut1, we proceeded to evaluate the ability of these cells to bind and absorb the fluorescent glucose analogs 2-NBDG and 6-NBDG. Rather than using a plate reader for these assays, however, we chose to utilize flow cytometry as a means to quantitate and evaluate cellular uptake of these compounds. Though lower in throughput, the advantage of this approach is that it provides population-level uptake dynamics in single cells rather than a normalized population average based on cell number estimates. This is especially important in cases where cells within a population exhibit heterogenous uptake rates during different phases of the cell cycle [26]. The value of this methodology became immediately obvious in differentiating between 2-NBDG versus 6-NBDG uptake in L929 cells.

To establish the binding and uptake kinetics for these compounds, we performed a time-course incubation of live, adherent L929 cells with a fixed concentration (100 μM) of each glucose analog before detachment and evaluation by flow cytometry (Figure 1C,D). Analysis of 2-NBDG binding and uptake as a function of time produces a progressively biphasic distribution. The smaller peak of cells with lower fluorescence remains at a fixed fluorescent intensity as a function of time, while the larger peak continues to increase in fluorescent intensity over the one-hour duration of the assay (Figure 1C). These data suggest that 2-NBDG binds to all L929 cells, but only a certain proportion (60–75%) of these cells progressively accumulate the compound by transport across the plasma membrane [26]. In contrast, it appears that L929 cells bind but fail to take up the 6-NBDG analog (Figure 1D). The flow cytometry histogram for 6-NBDG demonstrates a single population of cells that rapidly increases in fluorescence upon exposure to the compound but fails to increase in fluorescent intensity after 15–30 minutes of exposure. A plot of population mean fluorescent intensity (MFI) for each compound as a function of time clearly demonstrates the difference in kinetics between 2-NBDG and 6-NBDG (Figure 1E). These population averages from flow cytometry provide comparable data to similar assays performed in 96-well plate format (Figure 1F), demonstrating the validity of this approach in comparison to the more common high-throughput format of a plate reader.

The time-dependent dynamics of binding and uptake for both NBDG analogs suggest a receptor-mediated process, which would be consistent with transport by Glut1. These kinetics are distinct from those of mass transport by endocytosis, which are typically linear as a function of time. This distinction can be demonstrated in L929 cells using fluorescent labeled dextran as a substrate. Flow histograms for dextran uptake demonstrate the presence of a single population of cells that increase in fluorescent intensity over the course of the assay (Figure 1G). Regression analysis of population mean fluorescent intensity as a function of time produces a linear uptake rate consistent with the uptake of dextran by endocytosis rather than receptor-mediated transport (Figure 1H).

3.3. Binding kinetics of 2-NBDG and 6-NBDG in live and fixed L929 cells

In an attempt to calculate the binding constant for 2-NBDG and 6-NBDG in L929 cells, we analyzed the fluorescent intensity of detached cells as a function of compound concentration. Binding was performed for one hour at 4°C to prevent uptake by Glut1 or by endocytosis, both of which are dramatically attenuated at this temperature. The utility of this assay, unfortunately, was limited by the solubility of both NBDG compounds, which can achieve maximum of 400 μM in aqueous solution.

Flow cytometry histograms of binding to 2-NBDG and 6-NBDG demonstrate a progressive increase in L929 cellular fluorescence as a function of compound concentration (Figure 2A, B). Both compounds produced a single population of cells at 4°C, consistent with the idea that binding—but not uptake—was occurring in contrast to the data at 37°C (Figure 1C). A plot of mean fluorescent intensity as a function of NBDG concentration produces a linear relationship for both compounds, which indicates a failure to fully saturate binding sites on the cell surface (Figure 2C). Higher concentrations of compound cannot be examined due to the limitation in their solubility in aqueous solution.

Figure 2. Saturation binding of 2-NBDG and 6-NBDG to L929 fibroblasts.

(A-B, D-E) Representative flow cytometry histograms for live (A-B) or fixed (D-E) L929 cells treated for one hour with the indicated concentration (12.5–400 μM) of 2-NBDG or 6-NBDG. Incubations with each compound were performed in suspension at 4°C to allow for binding but to prevent kinetic uptake of the compounds. (C, F) Mean fluorescent intensity values extracted from flow cytometry of live (C) or fixed (F) L929 cells is plotted as a function of 2-NBDG or 6-NBDG concentration. Error bars represent the standard deviation of triplicate values from separate cytometry runs at each timepoint.

Because Glut1 is known to transit between the plasma membrane and intracellular vesicles through the process of endocytic recycling [27], we sought to determine whether fixed and permeabilized L929 cells display enhanced binding to the NBDG analogs due to the accessibility of additional binding sites. Binding assays performed at 4°C were repeated under these conditions and cells were once again analyzed by flow cytometry (Figures 2D, E). The plot of data from these experiments demonstrates a modest increase in fluorescence at all concentrations of 2-NBDG and 6-NBDG, which is consistent with the presence of increased numbers of binding sites (Figure 2F). Due to the inability of these assays to reach full saturation, however, the difference in total binding capacity cannot be fully assessed for either compound. These data are consistent with the possibility that both 2-NBDG and 6-NBDG bind to Glut1, but do not demonstrate conclusively that this is the case.

3.4. Pharmacologic inhibitors of Glut1 fail to block binding or uptake of 2-NBDG or 6-NBDG

In an effort to more clearly demonstrate that Glut1 is responsible for the binding and uptake of 2-NBDG and 6-NBDG, we turned to a panel of pharmacologic compounds that are known to block transport of glucose by Glut1. These include the fungal toxin cytochalasin B and two small-molecule inhibitors, WZB-117 and BAY-876 [28][29][16]. Each of these inhibitors is thought to bind at a distinct site in the central channel of Glut1, which therefore provides a range of different mechanisms by which to interrupt the binding of glucose and its fluorescent analogs [21][30]. All three of these compounds potently inhibit 2-deoxyglucose (2-DG) uptake in L929 cells (Figure 3A). In contrast, neither 2-NBDG nor 6-NBDG effectively blocks 2-DG uptake in L929 cells when used at 100 μM concentration.

Figure 3 – Pharmacologic inhibitors of Glut1 fail to inhibit uptake of 2-NBDG or 6-NBDG.

(A) The relative rate of [³H]-2-deoxyglucose uptake by L929 cells was measured in the presence of 2-NBDG (100 μM), 6-NBDG (100 μM) or the known Glut1 inhibitors cytochalasin B (Cyt B, 20 μM), BAY-876 (100 nM), or WZB-117 (1 μM). The average uptake rates were normalized to cells treated with vehicle only (0.1% DMSO). Error bars represent standard deviation of quadruplicate values. [*, p-value < 0.05]. (B,D) Representative flow cytometry histograms for live, adherent L929 cells treated for 30 minutes with 50 μM 2-NBDG or 6-NBDG in the presence of glucose (25 mM) or the indicated Glut1 inhibitors. (C, E) Mean fluorescent intensity values extracted from flow cytometry of L929 cells treated as indicated above for panels B, D. L929 cells without NBDG treatment serve as negative controls [(−) con] as indicated in each flow histogram. Error bars represent the standard deviation of triplicate values from separate cytometry runs for each treatment. [*, p-value < 0.05].

To determine whether the three Glut1 inhibitors or glucose itself affect the binding or uptake of 2-NBDG and 6-NBDG, saturating doses of each were included with each of the fluorescent NBDG analogs (50 μM) during incubation with live, adherent L929 cells for 30 minutes at 37°C. Cells were then detached and evaluated by flow cytometry as before. The overlaid flow cytometry histograms for assays with 2-NBDG show little variation in population fluorescence except in the case of cytochalasin B, which shows a very modest decrease (Figure 3B). Quantitative evaluation of mean fluorescent intensity for triplicate assays confirms this modest effect of cytochalasin B is statistically significant; however, neither glucose nor the small-molecule inhibitors blocked binding or uptake of 2-NBDG, which suggests that Glut1 independent mechanisms may be involved in its uptake.

Parallel assays for 6-NBDG similarly indicate that Glut1 may not be the primary binding site for this compound (Figure 3D). Not only did all four compounds fail to decrease 6-NBDG binding, but two of them (cytochalasin B and WZB-117) actually increased the fluorescent intensity of L929 cells (Figure 3E). Because it was not possible to determine the binding constants for either NBDG compound in L929 cells, however, these data cannot exclude the possibility that they bind at distinct site from the inhibitors or with very high affinity that cannot easily be competed away. As such, these data do not completely rule out Glut1 as the binding site for 2-NBDG or 6-NBDG.

3.5. Genetic overexpression of Glut1 fails to impact 2-NBDG or 6-NBDG binding or uptake

Because the pharmacologic approach provided reason to doubt that Glut1 is responsible for binding and uptake of 2-NBDG or 6-NBDG, we sought to obtain additional evidence to support or refute these data. We reasoned that overexpression of Glut1 in L929 cells would provide more binding sites for these compounds, and thus we sought to evaluate the uptake of 2-NBDG and 6-NBDG in a monoclonal line of L929 cells (L929/GB-M6) that we had previously developed for bioluminescent resonance energy transfer (BRET) analysis of Glut1 oligomerization [23]. These cells constitutively express mouse Glut1 with nanoluciferase fused to the N-terminus (Nluc-Glut1) at low levels that are comparable to endogenous Glut1, and also express a mCherry-Glut1 fusion protein at supraphysiological levels upon induction with doxycycline (Figure 4A). Accumulation of high levels of mCherry-Glut1 in response to doxycycline saturates the endogenous turnover mechanism for Glut1 in these cells, leading to accumulation of both endogenous Glut1 and Nluc-Glut1. The Nluc-Glut1 fusion protein is functional in glucose transport as demonstrated by 2-DG uptake assays in the basal or glucose-starved state, which enhances uptake more dramatically in the L929/GB-M6 line in comparison to L929 cells (Figure 4B) [31]. Addition of doxycycline to the L929/GB-M6 line, which induces mCherry-Glut1 expression, provides a modest increase in 2-DG uptake as well.

Figure 4. Genetic overexpression of Glut1 fails to promote uptake of 2-NBDG or 6-NBDG.

(A) L929/GB-M6 transgenic cells were treated for 24 hours with the indicated dose of doxycycline [Dox, ng/mL] prior to harvesting protein lysates and evaluating Glut1 expression by immunoblot. Arrows indicated the expected size of each Glut1 isoform prior to glycosylation. Doxycycline induces expression of the mCherry-Glut1 fusion protein along with the constitutively expressed Nanoluciferase-Glut1 protein (Nluc-Glut1) in a dose-dependent fashion. Blotting for β-actin was included as a control to demonstrate equal loading for each sample. (B) Rates of [³H]-2-deoxyglucose uptake by parental L929 cells and L929/GB-M6 transgenic cells were compared under basal conditions (5 mM glucose, black bars) and after one hour of starvation in glucose-free media (gray bars). L929/GB-M6 uptake rates were measured +/− doxycycline treatment (250 ng/mL) for 24 hours prior to the assay. Error bars represent standard deviation of quadruplicate values. [*, p-value < 0.05]. (C,D) Representative flow cytometry histograms of parental L929 cells and L929/GB-M6 transgenic cells treated +/− doxycycline (250 ng/mL) in the FL1 channel (C) and the FL3 channel (D). The FL1 channel used for NBDG uptake is unaffected by overexpression of mCherry-Glut1 while the FL3 channel demonstrates that dox treatment specifically induces expression of this protein in the L929/GB-M6 cell line. (E, F) Representative flow cytometry histograms for live, adherent L929 cells or L929/GB-M6 cells treated for 30 minutes with 50 μM 2-NBDG or 6-NBDG. Both cell lines were treated +/− doxycycline (250 ng/mL) for 24 hours prior to the assay. (G, H) Mean fluorescent intensity values extracted from flow cytometry of L929 cells or L929/GB-M6 cells treated with 50 μM 2-NBDG or 6-NBDG for the indicated times. Both cell lines were treated +/− doxycycline (250 ng/mL) for 24 hours prior to the assay. Error bars represent the standard deviation of triplicate values from separate cytometry runs for each treatment.

Because the mCherry-Glut1 fusion protein is naturally fluorescent, we were concerned that it might increase the background fluorescence of L929/GB-M6 cells and thus interfere with evaluation of 2-NBDG and 6-NBDG binding assays. A careful analysis of the cell line in different channels of our flow cytometer, however, demonstrates that induction of mCherry-Glut1 fails to produce any measurable background signal in the FL1 channel (Ex/Em: 488/530 nm) used to measure NBDG fluorescence (Figure 4C). In contrast, mCherry-Glut1 fluorescence can be detected in the FL3 channel (Ex/Em: 488/670 nm), clearly demonstrating that the fusion protein is only expressed in the L929/GB-M6 line in response to doxycycline (Figure 4D).

Having validated the cellular model for Glut1 overexpression, we next sought to determine whether increasing the available number of Glut1 binding sites would alter the binding and uptake kinetics for either fluorescent glucose analog. Both the parental L929 cells and the L929/GB-M6 cells were treated +/− doxycycline (250 ng/mL) for 24 hours prior to analysis. Live, adherent cells of each line were then incubated with 2-NBDG or 6-NBDG (50 μM) for 5, 15, or 30 minutes at 37°C to evaluate binding and uptake. Somewhat surprisingly, we found no difference in the binding or uptake of either compound in both the basal and doxycycline-induced L929/GB-M6 cells compared to the parental L929 line (Figure 4E, F). Quantitative analysis of mean fluorescent intensity for each line was essentially indistinguishable both for 2-NBDG and 6-NBDG over the entire 30-minute time-course (Figure 4G, H). These data once again call into question whether Glut1 actually serves as a binding site for either compound and whether the uptake of 2-NBDG observed in L929 cells occurs via Glut1.

3.6. Genetic knockdown of Glut1 fails to impact 2-NBDG or 6-NBDG binding or uptake

Though neither pharmacologic inhibition nor genetic overexpression assays are consistent with a role for Glut1 in the binding or transport of 2-NBDG or 6-NBDG, it remains possible that these analogs bind at sites distinct from the pharmacologic inhibitors with very low affinity that fails to saturate available Glut1 sites in L929 cells. To address the possibility, we sought to deplete L929 cells of endogenous Glut1 to determine whether loss of these binding sites would affect the kinetics of 2-NBDG and 6-NBDG binding or uptake. This was accomplished by transfecting L929 cells with two distinct siRNAs that are targeted to the 3’UTR (Glut1-siRNA-A) or coding sequence (Glut1-siRNA-B) of the mouse Glut1 transcript. These were compared to a universal control siRNA lacking homology to transcripts in the mouse genome.

L929 cells treated with either or both of the two Glut1-targeted siRNAs show ~75–80% depletion of endogenous Glut1 at 48 hours after transfection (Figure 5A, B). Note that Glut1 is variably N-glycosylated in L929 cells, which accounts for the lack of a single, discrete band when the endogenous protein is evaluated by immunoblotting. We found that siRNA-B was slightly more effective at knocking down Glut1. This efficiency of knockdown translated into a more effective decrease in basal 2-DG uptake by L929 cells, though both individual siRNAs and the combination produced statistically significant decreases in uptake (51%, 73%, and 74%) (Figure 5C).

Figure 5. Genetic knockdown of Glut1 expression fails to decrease uptake of 2-NBDG or 6-NBDG.

(A) L929 cells were transfected with negative control or Glut1-targeted siRNAs (50 nM) and incubated for 48 hours prior to harvesting protein lysates and evaluating endogenous Glut1 expression by immunoblot. Arrows indicate molecular weight sizes on the blot based on a molecular weight ladder run with protein lysates. The endogenous protein is variably N-glycosylated and therefore runs as a smear rather than a single band. Blotting for β-actin was included as a control to demonstrate equal loading for each sample. (B) The relative amount of endogenous Glut1 protein in samples from panel A was determined by densitometry analysis of each lane. Total signal from the Glut1 was divided by β-actin signal for each lane before normalization to the untransfected control sample. (C) L929 cells were transfected with negative control or Glut1-targeted siRNAs (50 nM) and incubated for 24 hours prior to replating at equal density to 24-well plates. The rates of [³H]-2-deoxyglucose uptake were then measured 48 hours after initial transfection for each condition. Error bars represent standard deviation of quadruplicate values. [*, p-value < 0.05]. (D, E) Representative flow cytometry histograms for live, detached L929 cells treated for 60 minutes at 4°C with 100 μM 2-NBDG or 6-NBDG after 48 hours of siRNA-mediated knockdown of Glut1. (F) Mean fluorescent intensity values derived from flow cytometry plots of live, adherent L929 cells treated for the indicated times with 50 μM 2-NGDG or 6-NBDG. Error bars represent the standard deviation of triplicate values from separate cytometry runs at each timepoint. [*, p-value < 0.05].

Having demonstrated the efficacy of our knockdown strategy, we next evaluated the binding efficiency of 2-NBDG and 6-NBDG to L929 cells that had been transfected with control or Glut1-targeted siRNA for 48 hours. Binding assays were carried out as before at 4°C for 1 hour in the presence of either NBDG compound at 100 μM concentration. The overlay of histograms from these assays demonstrated no difference in fluorescent intensity between untransfected cells and those transfected with any of the three siRNAs (Figure 5D, E).

To more carefully evaluate binding and uptake kinetics, we next exposed live, adherent L929 cells with depleted levels of Glut1 (siRNA A+B, 2–5 nM each) to a 50 μM concentration of 2-NBDG or 6-NBDG over a 60-minute time-course prior to detachment and analysis by flow cytometry. A plot of mean fluorescent intensities as a function of time for demonstrates kinetics that are similar to what was previously observed for untransfected L929 cells (Figure 5F). Uptake of 2-NBDG was slightly decreased in cells transfected with Glut1-targeted siRNA compared to those transfected with the control siRNA, though this effect was only statistically significant at the 45 and 60-minute timepoints. No statistical difference in binding of 6-NBDG was detected between these two conditions. These data indicate that Glut1 likely plays a relatively minor role in uptake of 2-NBDG into L929 cells, while it seems to play no role at all in the binding of 6-NBDG.

4. Discussion and conclusions

In this study we evaluated the uptake of 2-NBDG and 6-NBDG into L929 cells, which represent a uniquely simple model owing to their exclusive expression of Slc2a1/Glut1 as a mode for glucose transport. The simplicity of this system was complemented by our methodological approach of relying primarily on flow cytometry for a readout of NBDG binding and uptake. This methodology has a significant advantage over plate readers or other fluorimetry devices in that it reveals population-level dynamics based on single-cell analysis. Others have used this approach as well with similar bimodal results for 2-NBDG uptake, though the mechanism of uptake was not considered in their case [26].

Our initial findings are consistent with prior reports of binding and uptake for both 2-NBDG and 6-NBDG. The logarithmic uptake kinetics for 2-NBDG in L929 cells are consistent with a saturable receptor-mediated process, though quantification of binding capacity proved impossible due to the limited solubility of this compound [9][8]. In contrast, the kinetic parameters for 6-NBDG, which rapidly saturate as a function of time, suggest that it binds at the cell surface, but is not internalized. This is consistent with prior studies suggesting that 6-NBDG has utility as a probe for glucose transporter presence in neurological tissues [32][33]. In the absence of more careful analysis, it could easily be assumed that these two molecules represent complementary tools for determining the glucose uptake capacity and cell surface transporter availability in L929 cells.

The initial reason for concern regarding this assumption comes from our finding that neither glucose nor Glut1 inhibitors have an appreciable effect on 2-NBDG or 6-NBDG binding or uptake in L929 cells, although there are two exceptions to this general statement. First, we found that cytochalasin B has variable effects on 2-NBDG and 6-NBDG uptake; it slightly decreases 2-NBDG uptake in L929 cells and actually increases 6-NBDG binding. This effect on 2-NBDG uptake has been previously noted in other studies [8][7]. Because cytochalasin B is also known to modulate the actin cytoskeleton of mammalian cells, however, the effect of this compound on both 2-NBDG and 6-NBDG may also reflect Glut1-independent effects involving receptor trafficking and localization. Second, we also observed that the small molecule WZB-117 affected 6-NBDG binding, though it actually promoted rather than decreased the affinity of 6-NBDG for L929 cells. Though a clear mechanism for this observation remains elusive at this point, it is possible that binding of WZB-117 locks Glut1 into a structural conformation that is conducive to 6-NBDG binding [21]. The lack of inhibition, however, still calls into question whether both NBDG compounds could bind or enter cells at sites other than Glut1.

Another prior study has also addressed the fact that binding and uptake of 6-NBDG by astrocytes is poorly competed by glucose or pharmacologic inhibitors of glucose transport [32]. This study, however, largely dismissed the possibility that 6-NBDG enters cells by a mechanism other than through glucose transporters by relying on an alternative binding site hypothesis or attributing the effect to different affinities in binding. Here we tested both of these hypotheses by using a variety of inhibitors that bind with high affinity at distinct sites on Glut1 and by genetic overexpression and knockdown of Glut1 in a cell line that exclusively uses this transporter to absorb glucose into cells. Our data demonstrate very little change in 2-NBDG or 6-NBDG binding or uptake when levels of Glut1 are genetical modulated, indicating that neither of these hypotheses present a likely explanation for the negative pharmacologic data. The Glut1 knockdown data is particularly compelling, as it showed no impact on 6-NBDG binding and a very small impact on 2-NBDG uptake after 30–60 minutes.

At a minimum, our study provides two important points regarding uptake of 2-NBDG. First, it demonstrates that 2-NBDG cannot enter cells by transmembrane transport through Glut1. Though it remains possible that the internalization of 2-NBDG we observed reflects binding of 2-NBDG to Glut1 followed by receptor-mediated endocytosis, actual transport across the membrane by Glut1 is ruled out by the inability of any of its high-affinity inhibitors to significantly impact accumulation. The receptor-mediated endocytic mechanism of import, while theoretically possible, is distinct from the physiological transmembrane uptake of glucose and its radiolabeled analogs 2-DG and 3-OMG and therefore fails to represent true glucose uptake.

Second, these data also suggest that there are mechanisms for 2-NBDG and 6-NBDG to bind and enter cells independent of any glucose transporter. This is implied by the fact that L929 cells exclusively express Glut1 and continue to robustly bind and accumulate 2-NBDG or 6-NBDG even after this protein is inhibited or genetically depleted. It therefore seems likely that transport of 2-NBDG into L929 cells is actually facilitated by another integral membrane transport protein outside of the GLUT, SGLT, or SWEET families of glucose transporters. Similar conclusions have recently been reached for 2-NBDG in murine T cells, which further casts doubt on the reliability of either fluorescent glucose analog [34]. Because it remains unclear what role the NBD fluorophore—as opposed to glucose itself—plays in binding and uptake of these analogs, we urge great caution in the use of 2-NBDG or 6-NBDG as a means to measure glucose uptake rates in mammalian cells.

Highlights.

• Glucose uptake rates are a key metabolic parameter for cell growth

• Fluorescent glucose analogs are commonly used to monitor glucose uptake

• Murine L929 fibroblasts import glucose exclusively via the Glut1 transporter

• Multiple methods of Glut1 inhibition fail to block fluorescent analog uptake

• Fluorescent glucose analogs enter cells by transporter-independent mechanisms

Abbreviations (footnote)

2-DG

2-deoxyglucose

2-NBDG

2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-D-glucose

DMEM

Dulbecco’s Modified Eagle Medium

DPBS

Dulbecco’s phosphate buffered saline

DMSO

dimethyl sulfoxide

GLUT

glucose transporter

MFI

mean fluorescent intensity

PBS

phosphate buffered saline

rcf

relative centrifugal forces

RPMI-1640

Roswell Park Memorial Institute 1640

SGLT

sodium-dependent glucose transporter

siRNA

small interfering RNA

SLC2A

solute carrier family 2A