Introduction

Glutamine uptake through the transporter ASCT2 has been found stimulated in many cancer models (Fuchs and Bode 2005). This metabolic feature has prompted various attempts to inhibit glutamine entry as a device to hinder cancer cell proliferation. l-γ-Glutamyl-p-nitroanilide (GPNA) was proposed several years ago as an ASCT2 inhibitor (Esslinger et al. 2005) and has been subsequently widely used to this purpose (Hassanein et al. 2015; Ren et al. 2015; Wang et al. 2015; Bolzoni et al. 2016; van Geldermalsen et al. 2016). However, GPNA selectivity for ASCT2 had never been assessed in depth until, most recently, Broer et al. (2016) definitely demonstrated that the compound inhibits, besides ASCT2, other Na+-dependent carriers, such as several members of the SNAT family.

Besides ASCT2, also other transporters have been found overexpressed in human cancer cells in vitro, as well as in primary and metastatic human tumors in vivo. In particular, two of these carriers, LAT1 (coded by SLC7A5) and LAT2 (coded by SLC7A8) are highly expressed in a wide array of human tumors (Fuchs and Bode 2005; Kaira et al. 2008; Wang and Holst 2015; Barollo et al. 2016). LAT1 and LAT2, once complexed with the chaperone 4F2hc, account in many tissues for the activity of system l, a sodium-independent, non electrogenic, and exchange transport mechanism that operates the transmembrane fluxes of most essential amino acids. LAT1 and LAT2 have comparable operational features, both efficiently transport leucine, although LAT2 is endowed with a lower affinity for substrates (del Amo et al. 2008), and, through leucine transport, have an important regulatory role in the stimulation of mTORC1 activity (Nicklin et al. 2009; Chen et al. 2014; Milkereit et al. 2015).

Here, we show that GPNA inhibits the sodium-independent influx of leucine and lowers its cell content, indicating that the inhibitor hinders the activity of system l.

Materials and methods

Human oligodendroglioma Hs683 cells, provided by Prof. R. Kiss, University of Bruxelles, were grown in low-glucose (1 g/L) Dulbecco’s modified medium, DMEM (Euroclone), supplemented with 10% FBS (Lonza, Basel, Switzerland), 4 mM Gln, 25 mM HEPES, and antibiotics (100 U/mL penicillin, and 100 μg/mL streptomycin). Human cervix carcinoma HeLa cells, obtained from ATCC, human breast adenocarcinoma MCF7 cells, purchased from the IZSLER Cell Bank (Brescia, Italy), and human hepatocellular carcinoma Huh7 cells, a gift of Prof. G. Raimondo, University of Messina, were grown in high-glucose (4.5 g/L) DMEM supplemented with 10% FBS, 4 mM Gln, and antibiotics. Human lung alveolar carcinoma A549 cells, provided by Prof. L. Migliore, University of Pisa, were grown in Ham’s F12 medium supplemented with 10% FBS, 1 mM Gln, and antibiotics. Cells were incubated at 37 °C at 5% CO2; after thawing, all cells were used for less than ten passages.

Total cell RNA (1 μg) was isolated and reverse transcribed, and cDNA analyzed as previously described (Chiu et al. 2014). Primers were: 5′-GTGGAC TTCGGGAACTATCACC (SLC7A5, for), 5′-GAACAGGGACCCATTGACGG (SLC7A5, rev); 5′-AGGCTGGAACTTTCTGAAT (SLC7A8, for), 5′-ACATAAGCGACATTGGCAA (SLC7A8 rev); 5′-TGGTCTCCTGGATCATGTGG (SLC1A5 for), 5′-TTTGCGGGTGAAGAGGAAGT (SLC1A5 rev); 5′-CACCACAGGGAAGTTCGTATTC (SLC38A1 for), 5′-CGTACCAGGCTGAAAATGTCTC (SLC38A1 rev); 5′-ATGAAGAAGGCCGAAATGGGA (SLC38A2 for), 5′-TGCTTGGTGGGGTAGGAGTAG (SLC38A2 rev); 5′-GCAGCCATCAGGTAAGCCAAG (RPL-15, for), 5′-AGCGGACCCTCAGAAGAAAGC (RPL-15, rev). Data analysis was made according to the relative standard curve method (Bustin 2000).

Immunoblotting was performed as previously described (Chiu et al. 2012) using anti-LAT1 (rabbit, polyclonal, 1:1000, Cell Signaling Technology), anti-S6K1 phospho T389 (rabbit, monoclonal, 1:1000, Cell Signaling Technology), anti-S6K1 total (rabbit, monoclonal, 1:1000, Cell Signaling Technology), anti-ASCT2 (rabbit, monoclonal, 1:4000; Cell Signaling Technology) and anti-β-actin (rabbit, polyclonal, 1:1000, Sigma).

Intracellular leucine, isoleucine and glutamine were extracted with ice-cold absolute ethanol and determined with liquid chromatography coupled with mass spectrometry as previously described (Bolzoni et al. 2016).

For ASCT2 gene silencing, Hs683 cells were transfected with a scrambled (ON-TARGETplus Non-targeting Pool) or with a siRNA targeting ASCT2 (ON-TARGETplus SMARTpool, SLC1A5, Thermo Scientific DharmaFECT). 72 h after transfection, cells were rinsed in PBS and fresh medium was added. After 9 h, intracellular leucine was extracted, and its content determined.

For transport experiments, cells were seeded on 96-well plates at a density of 15 × 103 cells/well in normal growth medium. The initial influx of l-[3,4-3H]-glutamine (Amersham Biosciences) and l-[4,5-3H]-leucine (Amersham Biosciences) was measured following the method previously described (Bianchi et al. 2012). For l-glutamine transport, cells were rapidly washed with an Earle’s Balanced Salt Solution (EBSS, composition in mM: NaCl 117, KCl 5.3, CaCl2 1.8, MgSO4·7H2O 0.81, choline phosphate 0.9, glucose 5.5, supplemented with 0.02% Phenol Red, kept at pH 7.4 with 26 mM Tris–HCl) and transport assay (30 s) was performed in the same solution. For l-leucine, before transport determination, Na+-free EBSS, where N-methyl-d-glucamine chloride was used to replace NaCl, was used for the washing and the transport assay.

The non-saturable component of leucine influx was estimated measuring leucine uptake in the presence of 2 mM leucine.

For the kinetic analysis, l-leucine influx data, obtained at different concentrations of the amino acid, were fit to the equation:

$$v = \frac{{V_{\hbox{max} } \cdot [{\text{Leu}}]}}{{K_{\text{m}} + [{\text{Leu}}]}} + K_{\text{D}} \cdot [{\text{Leu}}].$$
(1)

For the kinetic analysis of GPNA inhibition activity, l-leucine influx data, obtained at different concentrations of the inhibitor, were fit to the equation for competitive inhibition:

$$v = v_{0} - \frac{{I_{\hbox{max} } \cdot [{\text{GPNA}}]}}{{\left[ I \right]_{0.5} + [{\text{GPNA}}]}},$$
(2)

where v 0 is the influx in the absence of inhibitor, I max the maximal inhibition and [I]0.5 the GPNA concentration at which the inhibition is half-maximal.

GraphPad Prism 5.0™ was used for all the statistical analyses, and p values <0.05 were considered statistically significant. Unless otherwise stated, Sigma–Aldrich was the source of all the chemicals, included GPNA.

Results

In a panel of human cancer cell lines, l-γ-glutamyl-p-nitroanilide (GPNA), an inhibitor of the sodium-dependent carrier ASCT2, inhibited most of glutamine influx in the presence of sodium (Fig. 1a). However, GPNA significantly inhibited glutamine transport also in the absence of sodium (Fig. 1b). To identify the sodium-independent transport system inhibited by GPNA, the saturable influx of leucine was determined in the same cell models in the absence of sodium. Leucine influx was different in the lines tested, with HeLa cells exhibiting the fastest influx and Huh7 cells the slowest (Fig. 1c). In all the cell lines GPNA significantly inhibited the sodium-independent leucine influx, with inhibitions ranging from almost 40% for Huh7 cells to over the 50% for Hs683 and MCF7 cells.

Fig. 1
figure 1

a, b l-[3,4-3H(N)]-Glutamine influx (25 μM; 5 μCi/mL; transport time, 30 s) was assayed in EBSS (a) or in Na+-free EBSS in the absence (control) or in the presence of l-γ-glutamyl-p-nitroanilide (GPNA, 3 mM) in Hs683, HeLa, A549, MCF7, and Huh7 cells. c The determination of Na+-independent influx of l-[4,5-3H(N)]-leucine (10 μM; 2 μCi/mL; transport time, 30 s) was performed in the same cells in the absence (control) or in the presence of GPNA, 3 mM. The saturable influx was obtained subtracting the non-saturable influx (see “Materials and methods”), determined in parallel under the same conditions, from the total influx. Data are expressed as nmol/mg prot/min. *p < 0.05, ***p < 0.001 versus control assessed with Student t test for unpaired data

Full size image

LAT1 and LAT2 system l transporters were consistently expressed, although at a variable degree, in all the cancer cell lines (Fig. 2a, b). Human oligodendroglioma Hs683 cells showed the highest relative LAT1 and the lowest LAT2 mRNA expression, while MCF7 breast cancer cells had the highest LAT2, but relatively low LAT1 mRNA expression, and hepatocellular carcinoma Huh7 cells had a low relative expression of both transporters.

Fig. 2
figure 2

a SLC7A5, encoding for LAT1 (left), and SLC7A8, encoding for LAT2 (right), expression was assessed by RT-PCR in the indicated cells incubated in growth medium. Data were normalized to the expression of RPL-15. b LAT1 protein expression was assessed by Western blot in Hs683, HeLa, A549, MCF7, and Huh7 cells. β-actin was used for loading control. A representative experiment, performed twice with comparable results, is shown

Full size image

The kinetic analysis of the Na+-independent Leu influx, performed in Hs683 cells (Fig. 3a), indicated that GPNA increased the K m for leucine from 56 ± 2 to 104 ± 15 μM, while the V max was not significantly modified (15.1 ± 2.10 nmol/mg/min, GPNA absent, versus 14.1 ± 1.33 nmol/mg/min, GPNA present). The diffusion constant K D was also comparable in the absence and in the presence of GPNA (49.5 ± 2.42 min−1, GPNA absent, versus 50.5 ± 3.33 min−1, GPNA present). The inhibition pattern was satisfactorily fitted with an equation for a competitive inhibition (Fig. 3b). At 10 μM [Leu], the maximal inhibition by GPNA was more than 65% of the uninhibited total influx with a half-maximal inhibitory concentration of 807 ± 70 μM. The inhibitory effects on the sodium-independent leucine influx of GPNA and of BCH—an amino acid analog that preferentially inhibits system l—are compared in Fig. 3c. BCH inhibited leucine influx by more than 80%, while GPNA-dependent inhibition was roughly 50% of the total leucine uptake.

Fig. 3
figure 3

a Kinetic analysis of Na+-independent leucine uptake by Hs683 cells. Cells were incubated for 30 s in Na+-free EBSS with l-[3,4-3H(N)]-Leu (2 μCi/mL; 1, 5, 10, 25, 50, 500 μM, left panel) in the absence (control) or in the presence of 3 mM GPNA. Grey lines represent the saturable component of leucine influx for control (dashed line) and GPNA (dotted line). Lines represent the best fit to Eq. 1 (R2 0.999 for control, 0.994 for GPNA inhibited influx). b Analysis of Leu uptake inhibition by GPNA. Cells were incubated for 30 s in Na+-free EBSS with l-[3,4-3H(N)]-Leu (2 μCi/mL;10 μM) in the absence or in the presence of GPNA (0.3, 1, 3 mM). The line represents the best fit to Eq. 2 (R 2 0.980). c Inhibition of Na+-independent leucine (10 μM) uptake by GPNA (3 mM) or BCH (1 mM). For a–c, data represent mean ± SD of five independent determinations each. *p < 0.05; **p < 0.01; ***p < 0.001 vs. control, assessed with Student t test for unpaired data

Full size image

We next investigated the intracellular content of the two system l substrates leucine and isoleucine upon 9 h of incubation with 3 mM GPNA. Under control conditions, the intracellular content of both leucine and isoleucine varied among the cell lines (Fig. 4a, b). Hs683 cells had the highest content of either leucine (14.4 ± 0.37 nmol/mg prot) or isoleucine (14.3 ± 0.85 nmol/mg prot), while Huh7 cells had the lowest (Leu = 4.3 ± 0.09 nmol/mg prot; Ile = 3.9 ± 0.59 nmol/mg prot). In all but Huh7 cells, the intracellular content the two amino acids significantly decreased in the presence of GPNA, with the highest inhibition (more than 40%) in Hs683 cells. The cell content of glutamine, measured in the same cells (Fig. 4c) also varied among the various cell lines, with MCF7 showing the highest intracellular levels and A549 cells the lowest. Only in Hs683 cells, GPNA significantly decreased intracellular glutamine.

Fig. 4
figure 4

The cell content of Leu (a), Ile (b) and Gln (c) in Hs683, HeLa, A549, MCF7, and Huh7 cells incubated in growth medium without (control) or with 3 mM GPNA for 9 h. Mean ± SD of three experiments are shown. *p < 0.05, **p < 0.01, ***p < 0.001 versus control; ns, not significant, as assessed with Student t test for unpaired data

Full size image

Under the same conditions, the abundance of the phosphorylated form of S6K1, an indicator of the activity of the kinase mTORC1, which is stimulated by intracellular leucine, was markedly lowered in all the cell lines (Fig. 5a). In all cells, but Huh7, changes in pS6K1 were not paralleled by changes in total S6K1 and, therefore, could be attributed to an effective decrease in mTORC1 activity. Interestingly, GPNA had inconsistent effects on LAT1 expression, with Hs683 and HeLa exhibiting sizable increases, while A549 and MCF7 showed a decreased expression of the transporter. As expected, rapamycin suppressed S6K1 phosphorylation in all the cell lines, confirming its absolute dependence upon mTORC1 activity (Fig. 4b).

Fig. 5
figure 5

a Western blot of S6K1 (phosphoT389 and total) and LAT1 in Hs683, HeLa, A549, MCF7, and Huh7 cells incubated in growth medium (control) or with 3 mM GPNA for 9 h. β-actin was used for loading control. b Cells were incubated in the absence or in the presence of rapamycin (100 nm) for 9 h. At the end of the incubation, the Western blot of S6K1 (phosphoT389 and total) was performed. Representative experiments, performed twice with comparable results, are shown

Full size image

To verify if the inhibition of ASCT2 by GPNA could be involved in the effects of the analog on the cell content of leucine, SLC1A5 was silenced in Hs683 cells, causing an almost complete suppression of ASCT2 expression both at mRNA (Fig. 6a) and protein level (Fig. 6b), and a significant inhibition of glutamine uptake that is exclusively detected in the presence of sodium (Fig. 6c). No evidence of a compensatory increase in the expression of the other sodium-dependent transporters SNAT1 (encoded by SLC38A1, Fig. 6e) or SNAT2 (encoded by SLC38A2, Fig. 6f) was detected. However, ASCT2 silencing did not change the cell content of leucine (Fig. 6d).

Fig. 6
figure 6

The expression of SLC1A5 (for ASCT2, a) SLC38A1 (for SNAT1, e) and SLC38A2 mRNA (for SNAT2, f) was assessed by RT-PCR in scramble- or anti-ASCT2-siRNA-transfected Hs683 cells 72 h after transfection. Data were normalized to the expression of RPL-15. b Western blot of ASCT2 in scramble- and anti-ASCT2-siRNA-transfected Hs683 cells. c 72 h after transfection, l-glutamine uptake (25 μM, 10 μCi/mL) was measured, in the presence and in the absence of sodium, in Hs683 cells transfected with scramble- or anti-ASCT2-siRNA as described under Materials and Methods. d 72 h after transfection, Hs683 cells, transfected with scramble- or anti-ASCT2-siRNA, were incubated for 9 h in fresh growth medium, and intracellular leucine was determined at the end of the incubation. For a, e, f, n = 4; for c, n = 5; for d, n = 3; for b, a representative experiment, performed twice, is shown. *p < 0.05, ns not significant, as assessed with a Student t test for unpaired data

Full size image

Discussion

This report demonstrates that GPNA is a competitive inhibitor of system l transport activity. This sodium-independent transport agency accounts for the cell uptake of most essential amino acids and is stimulated in many tumors (Kanai et al. 1998). Although the kinetics of the inhibitory effect on leucine influx have been studied in Hs683 cells, which are endowed with the highest expression of the system l transporter LAT1 when compared to the other cell lines tested, both LAT1 and LAT2 transporters are likely inhibited by GPNA. Indeed, the percentage inhibition of the Na+-independent leucine influx is substantially comparable in Hs683 and MCF7 cells, although the two cell lines exhibit specular differences in LAT1 and LAT2 mRNA expression. However, the approach used in this study is not sufficient to definitely discriminate if LAT1 or LAT2 are equally sensitive to GPNA or to exclude a possible sensitivity of the other system l transporters LAT3 and LAT4.

In the past few years, GPNA has been used as an experimental device to inhibit glutamine transport through the Na+-dependent transporter ASCT2 in several models of cancer cells (Hassanein et al. 2013; Indo et al. 2013; Ren et al. 2015; Takahashi et al. 2015; Wang et al. 2015; Bolzoni et al. 2016; van Geldermalsen et al. 2016). Therefore, the inhibition of cell growth by GPNA has been taken as an evidence for the essential metabolic role of the ASCT2 transporter in those models. Inhibition by GPNA has been also used to demonstrate that glutamine analogs, synthesized as potential probes of ASCT2 transport function for positron emission tomography, effectively interact with the transporter (Lieberman et al. 2011; Tang et al. 2016). At the light of the data presented here, these interpretations should be taken with caution. In particular, system l inhibition may contribute to the suppression of glutamine uptake by GPNA. Moreover, since the inhibition of system l by GPNA could directly hamper, besides glutamine uptake, the uptake of essential amino acids needed for protein synthesis and cell growth, the effects of GPNA on cell growth cannot be attributed to the sole inhibition of ASCT2.

According to the model of Nicklin et al. (2009), leucine influx may be inhibited as an indirect effect of GPNA inhibition of the glutamine influx mediated by the sodium-dependent ASCT2 transporter that would lower the amount of intracellular glutamine available for promoting the influx of leucine through system l exchange transporters LAT1 and LAT2. The involvement of this mechanism in the GPNA-mediated inhibition of leucine influx is highly unlikely, since, in the transport experiments, the inhibitor is only present during the assay (30 s), which, moreover, occurs in the absence of sodium. With the same argument, it is possible to exclude the involvement of other sodium-dependent transport systems, such as SNAT1 and SNAT2, recently described to be sensitive to GPNA inhibition (Broer et al. 2016).

On the contrary, it is possible that the significant depletion of the intracellular pool of leucine and isoleucine, observed in cells incubated for 9 h with GPNA, may involve the decrease in cell glutamine attributable to ASCT2 inhibition. Actually, in human oligodendroglioma Hs683 cells long term GPNA treatment significantly lowered also the cell content of glutamine (Fig. 4c). However, GPNA decreased leucine influx and lowered the cell content of the essential amino acid in all the cell lines tested, with the only exception of Huh7 cells. In three of these cell lines (HeLa, A549, and MCF7) cell leucine decreased in the absence of a significant depletion of intracellular glutamine. Even in Hs683 cells a contribution of ASCT2 inhibition to leucine depletion seems unlikely, since ASCT2 silencing does not significantly affect cell leucine while markedly inhibits total and sodium-dependent glutamine uptake (Fig. 6). Although a role for other GPNA-sensitive sodium-dependent transporters, such as SNAT1 and SNAT2, cannot be excluded, it should be noted that no compensatory induction of these transporters is detected in ASCT2-silenced Hs683 cells (Fig. 6d,e), at variance with the results reported in other cell models (Broer et al. 2016).

Interestingly, in all the cell lines tested, the long term incubation with GPNA markedly decreases, although at a variable degree, mTORC1 activity. Under the same conditions, the expression of LAT1 exhibited inconsistent changes, being increased in Hs683 and in HeLa cells and decreased in A549 and MCF7 lines. Therefore, the inhibition of mTORC1 was not associated to the changes in LAT1 expression and should be attributable to the inhibition of leucine influx by GPNA and/or to the partial depletion of the intracellular amino acid. Consistently, mTORC1 inhibition had been also observed in cells incubated with the system l inhibitor BCH (Ishizuka et al. 2008). However, given that also glutamine activates mTORC1 independently of leucine (Chiu et al. 2012; Jewell et al. 2015), GPNA-dependent inhibition of ASCT2-mediated glutamine influx may also contribute to kinase inhibition.

In conclusion, this report demonstrates that the ASCT2-inhibitor GPNA also inhibits the influx of essential amino acids through LAT1/2 and that this inhibition, depending on the expression level of LAT1/2, may affect the composition of intracellular amino acid pool and the activity of mTORC1. The demonstration of the metabolic relevance of ASCT2 in a cancer model should, therefore, rely on the genetic suppression of the transporter rather than on GPNA effects.