Genetic Disruption of the Multifunctional CD98/LAT1 Complex Demonstrates the Key Role of Essential Amino Acid Transport in the Control of mTORC1 and Tumor Growth

Tumor cells face numerous stressors in their microenvironment and therefore have developed adaptive strategies to survive and grow. The limitation of essential nutrients is particularly acute in rapidly growing and hypoxic tumors. Interestingly, under conditions of limited oxygen perfusion, tumors cells induce, via hypoxia-induced transcription factors (HIF) and other adaptive mechanisms, increased expression of key nutrient transporters (1–4). An example of an overexpressed membrane nutrient transporter complex is the multifunctional heterodimer, CD98/LAT1, that controls import of essential amino acids (EAA).

CD98 is a single-transmembrane glycoprotein also known as the 4F2 antigen heavy chain (4F2hc). This protein binds to the cytoplasmic tail of β-integrin (5, 6) and regulates migration, adhesion-induced intracellular signaling, and anchorage-independent survival (7–9). It has been reported that CD98 overexpression induces malignant transformation of NIH3T3 and BALB3T3 cells (10, 11). Furthermore, CD98 knockout mouse embryonic stem cells display restricted teratocarcinoma formation (8). Moreover, a recent study has shown that CD98 regulates microenvironmental adaptation by amplifying cancer cell capacity to respond to extracellular matrix rigidity to facilitate their tumor growth in a mouse skin cancer model (12).

CD98 also interacts with LAT1 through a disulfide bond and acts as a chaperone by promoting LAT1 stabilization, trafficking, and functional insertion into the plasma membrane (6, 13). LAT1 is a 12-transmembrane spanning protein responsible for Na⁺-independent transport of large neutral EAA (Leu, Val, Ile, Phe, Trp, His, Met, Tyr). LAT1 is an obligatory exchanger with the uptake of one amino acid (AA) being coupled to the efflux of another AA (14, 15). The elevated bioenergetic need of rapidly dividing cells creates an increased demand for AAs to satisfy biomass increase. This nutritional demand imposes a constant stress in tumors growing in hostile, acidic, and low nourished microenvironments (16, 17). Thus, high expression and activity of LAT1, regulated by HIF2 as reported in lung tissues and kidney cancers, is implicated in the ability to sustain growth despite the challenges faced in the microenvironment (18). Numerous clinical studies have shown that the CD98/LAT1 complex is overexpressed and is a negative prognostic factor in different types of cancers, including prostate (19), non–small cell lung cancer (20), gliomas (21), and renal cell carcinomas (18, 22).

It is now well accepted that the CD98/LAT1 complex plays a key role in tumor growth and is therefore an attractive therapeutic target. However, to improve effective anticancer treatments, it is essential to clearly define which member plays the dominant protumoral role in this complex. This notion is far from being clarified and remains controversial in the literature (2). Despite the fact that CD98 knockout cells have reduced AA transport, Feral and colleagues and Cantor and Ginsberg reported that the protumoral action in a teratocarcinoma model or growth-promoting activities in T and B cells are mediated through integrins–CD98 signaling rather than the activity of LAT1 (8, 9). Considering the obligatory nutrient needs in tumors and, in particular, EAA, this is a surprising conclusion that we decided to challenge and investigate further.

Here, we report AA stress, mTORC1 activity, EAA transport rates, proliferation, and tumorigenicity in a colorectal and lung human adenocarcinoma cell lines (LS174T, A549), in which the corresponding genes for LAT1 and CD98 have been disrupted by zinc finger nucleases (ZFN) or knockdown by shRNA. Homozygous knockouts of each gene, CD98^KO and LAT1^KO, confirmed a clear interdependence of the two members of this complex. As anticipated, LAT1^KO cells display an in vitro and in vivo disruption of AA homeostasis leading to ATF4 induction, inhibition of mTORC1, and abolition of tumor growth. Full restoration of plasma membrane expression of CD98 failed to rescue growth of LAT1^KO cells, demonstrating the independence of CD98 in the growth phenotype. Finally, we confirmed and extended these results by specific inhibition of LAT1 with JPH203 (23, 24) in six cell lines from colon, lung, and renal cell carcinoma. Together, these findings clearly establish that EAA transport, but not CD98, is a key limiting step in tumor growth and therefore fully validate LAT1 as a major anticancer target. Undoubtedly, besides JPH203 (23, 24), enthusiasm for developing novel LAT1 (SLC7A5) inhibitors will emerge.

Human colon adenocarcinoma LS174T and HT29 cells were kindly provided by Dr. Van de Wetering (Utrecht, the Netherlands). The other cell lines from lung (A549, H1975) and renal cell carcinoma, (786-O, A498) were obtained from ATCC. These cell lines have been authenticated by DNA profiling using 8 different and highly polymorphic short tandem repeat loci (DSMZ). The cells, regularly checked for mycoplasma, were grown in DMEM (Gibco) supplemented with 7.5% FBS, penicillin (10 U/mL), and streptomycin (10 μg/mL). DMEM (0.3×) was obtained by mixing 2 volumes of DMEM lacking 5 EAA LEU, ILE, MET, PHE, TRP with 1 volume of regular DMEM. It is important to note that the final concentration of the EAA (except tryptophan) in DMEM (0.3×) still largely exceeds the physiologic EAA levels of human plasma (Supplementary Table S1).

LS174T cells were transfected with ZFNs designed by Sigma-Aldrich (Saint-Louis, MO). Transfection of the ZFN (CSTZFN-1KT, CompoZr Custom ZFN) targeting LAT1 (exon 5) or CD98 (exon 4) was performed with JetPRIME (Polyplus). Transfected cells were grown for 7 days to express the mutated forms and then CD98 surface expression was analyzed using flow cytometry. Negative and low expressing cells were sorted and plated in clonal conditions (250 individualized cells in 100 mm dishes). Each clone was picked and analyzed for CD98 or LAT1 expression by immunoblot and negative clones were re-cloned and further analyzed by DNA sequencing (Supplementary Table S2). Finally two independent clones for LAT1^KO, CD98^KO and LAT1^KO/CD98^KO double knockout were selected for this study.

We obtained prevalidated shRNA sequences targeting CD98 from Sigma (Sigma NM_002394, CCGGCTAGCTCATACCTGTCTGATTCTCGAGAATCAGACAGGTATGAGCTAGTTTTTG). Lentiviral particles (pLKO.1, Sigma) containing the shRNA were produced in HEK cells. Lentiviral infection of A549 cells was then performed, and puromycin selection was utilized to obtain a total population of shRNA-targeted cells. shControl (shCTRL) cells were created as a reference control (Addgene plasmid #1864, CCTAAGGTTAAGTCGCCCTCGCTCGAGCGAGGGCGACTTAACCTTAGG). shRNA efficacy was validated by comparison of shCD98 and shCTRL cells using qPCR and Western blotting.

Cells were lysed in 1.5× Laemmli buffer, and protein concentrations were determined using the Pierce BCA protein assay (23227 Thermo Scientific). Protein extracts (40 μg) were separated by electrophoresis on 10% SDS polyacrylamide gel and transferred onto polyvinylidene difluoride membranes (Millipore). Membranes were blocked in 5% nonfat milk in TN buffer (50 mmol/L Tris-HCl pH7.4, 150 mmol/L NaCl) and incubated with the following anti-human antibodies: rabbit LAT1 (1:1,000, KE026 TransGenic Inc.), rabbit CD98 (1:1,000, SC-9160 Santa Cruz Biotechnology), rabbit xCT (1:1,000, ab37185 Abcam), mouse GCN2 (1:250, sc-374609 Santa Cruz Biotechnology), mouse phospho-GCN2 (1:500, ab75836 Abcam), rabbit EIF2α (1:1,000, ab5369 Abcam), mouse phospho-EIF2α (1:1,000, ab32157 Abcam), rabbit ATF4 (1:1,000, 11815S CST), rabbit p70-S6K (1:1,000, 9202S CST), rabbit phospho-p70-S6K (1:1,000, 9202S CST), rabbit RPS6 (1:1,000, 2217S CST), and rabbit phospho-RPS6 (1:1,000, 2215S CST). Detection of tubulin was used as a protein loading control (1:10,000 MA5-16308, Thermo Scientific). Immunoreactive bands were detected with horseradish peroxidase anti-mouse or anti-rabbit antibodies (Promega) using the ECL system (Merck Millipore WBKLS0500). Analysis and quantification of immunoblots were performed using LI-COR Odyssey Imaging System.

Cells were trypsinized, washed, and incubated for 30 minutes with mouse anti-human CD98 (1:100, KS129 TransGenic Inc.; diluted in FACS buffer PBS/BSA 0.1%/EDTA 5 mmol/L) on ice. Cells were then washed and incubated for 30 minutes with anti-mouse PE-conjugated secondary antibody on ice. Cells were washed, resuspended in 500 μL of FACS buffer, filtered (40 μm), and analyzed using a fluorescence-activated cell sorter (BD Healthcare FACSCalibur Analyzer).

Cells were seeded (1 × 10⁵ cells) on Cel-Line Diagnostic Microscope Slides (30-256H-BLACK-CE24 Thermo Scientific). After 24 hours, cells were washed and fixed at room temperature for 20 minutes with 3% paraformaldehyde. Cells were permeabilized with PBS (Euromedex, ET330-A) containing 0.2% Triton X-100 (T8532 Sigma) for 2 minutes before being exposed to rabbit anti-human LAT1 (1:1000, KE026 TransGenic Inc.) and mouse anti-human CD98 (1:1000, KS129 TransGenic Inc.) for 1 hour at room temperature. Cells were washed 3 times with PBS and then incubated for 1 hour at room temperature with 1:1,000 dilution anti-mouse FluoProbe 488-labeled (FP-SA4110-T Interchim) and anti-rabbit FluoProbe 594-labeled (FP-SD5110-T Interchim) and mounted using ProLong Gold Antifade Reagent (P36934 Life Technologies). Images were captured on an SP5 confocal microscope (Leica) and analyzed using Leica LS AF software and ImageJ.

Cells (2.5 × 10⁵) were seeded onto 35-mm dishes, in triplicates per cell line. Cells were used for uptake experiments 24 hours after seeding. Culture media were removed and cells were carefully washed with prewarmed Na⁺-free Hank's Balanced Salt Solution (HBSS: 125 mmol/L choline chloride, 4.8 mmol/L KCl, 1.2 mmol/L MgSO₄, 1.2 mmol/L KH₂PO₄, 1.3 mmol/L CaCl₂, 5.6 mmol/L glucose, and 25 mmol/L HEPES), preincubated in 1.0 mL of prewarmed Na⁺-free HBSS at 37°C for 5 minutes before adding substrates for the uptake experiment. Cells were then incubated at 37°C for 1 minute in 750 μL of Na⁺-free HBSS containing 1.0 μmol/L of L-[¹⁴C]-leucine (0.03 μCie/mL; PerkinElmer). Subsequently, cells were washed three times with ice-cold Na⁺-free HBSS containing 1.0 mmol/L of nonradiolabeled leucine. Cells were then lysed with 50 μL of 0.1 N NaOH and mixed with 3.5 mL of Emulsifier-Safe cocktail (PerkinElmer). Radioactivity was measured using a β-scintillation counter. For the inhibition experiments, the uptake of 1.0 μmol/L L-[¹⁴C]-leucine is examined in the presence of BCH (1.0 mmol/L).

The different cell lines (2.5 ×10⁴ cells for 7 days, 5 × 10⁴ cells for 3 days) were seeded onto 6-well plates in triplicate per cell line and per condition. We measured proliferation by trypsinizing the cells and counting them daily with a Coulter Z1 (Beckman) during 3 or 7 days. The cell proliferation index was calculated as “fold increase” by standardizing each measurement to the cell number obtained 24 hours after seeding (day 0).

LS174T 3D cultures were prepared using the liquid overlay method. Briefly, 24-well culture plates were coated with 1.5% agarose prepared in sterile water. Cells (10,000) from a single-cell suspension were added per well. The plates were gently swirled and incubated at 37°C in 5% CO₂ atmosphere until cells were organized into 3D. Media were changed once every 3 days. After 9 days, 3D culture images were taken using an EVOS Cell Imaging Systems (Life Technologies), and the 3D culture surface area was measured using ImageJ.

LS174T-derived mutants (1,000 cells) were plated in 60-mm dishes and incubated at 37°C, 5% CO₂. Twenty-four hours after cell adherence, the media were replaced with regular 1× or 0.3× DMEM (Supplementary Table S1) supplemented with 7.5% serum and containing JPH203 for LAT1 inhibition experiment. Media were changed once every three days. Dishes were stained with 5% Giemsa (Fluka) for 30 to 45 minutes to visualize colonies.

The different cell lines (5 × 10⁴ cells) were seeded onto 6-well plates in triplicate for each cell line and JPH203 concentrations indicated. We measured proliferation as described above after 3 days. Cell number fold increases were calculated as described above.

The different LS174T stable cell lines (1 × 10⁶ cells) suspended in 300 μL of serum-free DMEM supplemented with insulin–transferrin–selenium (Life Technologies) were injected subcutaneously into the back of 8-week-old female athymic mice (Janvier). Tumor dimensions were measured twice a week using calipers, and the tumor volume was determined by using the formula: (4π/3) × L/2 × W/2 × H/2 (L, length; W, width; and H, height). When the tumor volume reached 1,000 mm³, mice were euthanized, and the tumors were excised. For protein analysis, tumors were lysed directly after harvesting. Tumors were incubated in cell extraction buffer (FNN0011 Thermo Scientific) supplemented with Halt protease inhibitor cocktail (78429 Thermo Scientific) and lysed using a Precellys homogenizer. Animal housing was done in compliance to the EU directive 2010/63/EU. Briefly, each cage contained 5 mice with an enriched environment. Food and water were given ad libitum, and the litter was changed on a weekly basis. Animal care met the EU directive 2010/63/EU ethical criteria. The animal experimentation protocol was approved by the local animal care committee (Veterinary Service and Direction of Sanitary and Social Action of Monaco; Dr. H. Raps, Centre Scientifique de Monaco, Monaco).

Data are expressed as mean ± SD. Each experiment was performed at least three times. Statistical analysis was done with the unpaired Student t test. Differences between groups were considered statistically significant when P < 0.05.

LAT1 or CD98 knockouts (KO) were created in the colon adenocarcinoma cell line LS174T using ZFNs. To avoid clonal effects, we always tested two independent clones of each disrupted gene. Lack of corresponding protein expression (Fig. 1A), genome analysis, and sequencing around the ZFN-targeted site demonstrated that the mutations introduced (see Supplementary Table S2) disrupted the different alleles. LAT1^KO cells exhibit an 80% decrease in total CD98 protein expression (Fig. 1A and Supplementary Fig. S1A), resulting in a 60% reduction of the functional plasma membrane expression (Fig. 1B). Furthermore, qPCR analysis revealed that the CD98 protein reduction in LAT1^KO cells is partially due to reduced mRNA expression (Supplementary Fig. S1B). Similar independent results were obtained in inducible LAT1 knockdown (KD) LS174T cells and LAT1^KO in the lung carcinoma A549 cell line (Supplementary Fig. S1C and S1D). Therefore, despite the fact that CD98 can heterodimerize with other transporters (LAT2, y⁺LAT1, y⁺LAT2, and xCT), LAT1 represents the major expressed light chain in the two tumor cell lines LS174T and A549. In contrast, LAT1 total protein expression is reduced in CD98^KD and CD98^KO cells (Supplementary Fig. S1A and S1C) and retained in the cytoplasm (Fig. 1C). This finding is in agreement with the notion that CD98 is required to control LAT1 trafficking and functional insertion in the plasma membrane.

We next investigated the impact of these gene knockouts on LAT1 activity by measuring the Na⁺-independent rate of leucine transport. Ablation of LAT1 fully abolished the leucine uptake, confirming that LAT1 is the only functional large neutral AA transporter expressed in LS174T (Fig. 1D). CD98^KO cells also show a 90% reduction in Na⁺-independent leucine transport (Fig. 1D). However, the residual 10% transport activity detected in CD98^KO cells is significant and sensitive to the LATs inhibitor BCH (Fig. 1D) and JPH203 (see below).

As LAT1 transport activity is severely decreased in both CD98 and LAT1 knockouts, we analyzed their effects on the two AA-sensing pathways: GCN2 and mTORC1 (reviewed in ref. 25). In nutrient-rich media (1× DMEM), LAT1^KO cells display an induction of the AA stress response pathway GCN2/EIF2a/ATF4, demonstrating AA deficiency (Fig. 2A). In contrast, minor changes in the mTORC1 pathway were observed through the phosphorylation of p70-S6K and RPS6, as only the phosphorylation of p70-S6K decreased while its target S6RP remained unchanged (Fig. 2A). We then challenged these cells with a media where the concentrations of LAT1 substrates are closer to physiological levels (0.3× DMEM; Supplementary Table S1). In this condition, while the wild-type (WT) LS174T cells do not present any modification in mTORC1 and GCN2 pathways, LAT1^KO cells display strong modifications in both pathways (Fig. 2A). Quantification of GCN2 activity using the ATF4 expression shows that this pathway is increased by 10-fold compared with the wild-type cells (Fig. 2B). In parallel, mTORC1 activity measured by the phosphorylation level of RPS6 decreased by 50% (Fig. 2C). Similar results were observed with the second cell line (A549) disrupted for LAT1 (Supplementary Fig. S2A). In contrast, although CD98^KO cells have a strongly reduced LAT1 activity (90%), no changes in the GCN2 and mTORC1 pathways are detectable in both 1× and 0.3× DMEM (Fig. 2A and B). Similar results were obtained for the GCN2 pathway with a 90% knockdown of CD98 in A549 cells (Supplementary Fig. S3A and S3B).

mTORC1 (26, 27) and more recently GCN2 (28) have been demonstrated to control cell proliferation. We therefore investigated the effect of either LAT1 or CD98 knockout on in vitro LS174T proliferation (Fig. 3A). In 1× DMEM, LAT1 ablation decreased LS174T proliferation by 60% (Fig. 3A and B, left), and a decreased nutrient challenge with 0.3× EAA-derived DMEM dramatically reduced their proliferation by more than 90% (Fig. 3A and B, right). Consistent with our results for mTORC1 and GCN2 activity, LS174T-CD98^KOand A549-CD98^KD cells are able to maintain their proliferation at the same level as WT cells (Fig. 3A and B and Supplementary Fig. S3C). Furthermore, we confirmed that this LAT1^KO-specific antiproliferative effect (70% reduction) was observed in 3D culture assays (Fig. 3C). Interestingly, whereas CD98 has been described to be important for anchorage-independent survival and growth (7), 3D cultures show that only CD98^KO cells are able to grow like WT cells (Fig. 3C, bottom).

These findings demonstrate an essential role of LAT1 for maintenance of tumor cell AA homeostasis and proliferation. We confirmed these results by either invalidating (Supplementary Fig. S2A–S2C) or inhibiting LAT1 with the specific inhibitor JPH203 (Supplementary Fig. S4A and S4B; refs. 23, 24) in six different cell lines from colon (LS174T, HT29), lung (A549, H1975), and renal cell carcinoma (786-O and A498).

We then focused on the surprising growth phenotype of CD98^KO cells. In particular, we wanted to understand how cells that have lost 90% of the LAT1 EAA transport activity (Fig. 1D) and should have also reduced activities from additional transporters like xCT (29) are able to proliferate in 0.3× media at the same rate as WT cells. First, we tested the possibility that CD98 disruption could have induced another transporter capable to compensate for the lack of EAA transport by LAT1. A good candidate was SLC6A14, which transports all neutral AA (30). However, this transporter was only detectable at very low levels by qRT-PCR and Western blotting, while its expression was not increased in CD98^KO cells. In addition we were not able to detect a Na⁺-dependent leucine transport activity (data not shown).

We then investigated the impact on proliferation of the residual LAT1 activity present in CD98^KO cells using JPH203. Growth dose response demonstrated an extreme sensitivity of CD98^KO cells to JPH203 with an IC₅₀ of 1.0 μmol/L, whereas the IC₅₀ for WT cells was 35-fold higher (Fig. 4A). Furthermore, 5.0 μmol/L of JPH203 completely abolished the growth of CD98^KO cells (Fig. 4A and B), indicating that CD98^KO cell growth is critically dependent on the LAT1 residual activity. These findings were confirmed in A549-CD98^KD cells, where 30 μmol/L of JPH203 abolishes growth while WT cells continue to grow (Supplementary Fig. S3D). We independently confirmed this finding by generating double knockout (dKO) LS174T cells: CD98^KO/LAT1^KO (Fig. 4C, lane dKO). dKO cells present the same stress response as the LAT1^KO cells: the GCN2/EIF2a/ATF4 pathway is induced while the mTORC1 activity is strongly decreased (Fig. 4C) and in vitro proliferation is also strongly inhibited (Fig. 4D). Interestingly, downregulation of mTORC1 is stronger in dKO cells than in LAT1^KO cells (Fig. 4C). This result can be easily explained by the fact that CD98 is able to heterodimerize with others transporters. Therefore, double CD98^KO, LAT1^KO likely abolish, besides EAA transport, additional AA transporter like xCT, explaining this deeper effect on the mTORC1 activity. Combined, these data demonstrate that maintenance of CD98^{KO and KD} cell proliferation is not due to an adaptation resulting from CD98 invalidation but to a residual LAT1 transport activity.

To investigate the impact of reduced CD98 expression in LAT1^KO cells, we designed an experiment that fully restored endogenous CD98 expression at the cell surface in LAT1^KO cells while maintaining the lack of EAA transport. For this, we transfected xCT cDNA in LAT1^KO cells. Like LAT1, xCT is a light chain of CD98 but has a completely different function. Although LAT1 is important for EAA transport, xCT transports cystine, an essential precursor of glutathione synthesis to counteract oxidative stress (31, 32). As expected, expression of xCT or of LAT1 as a control restored the plasma membrane expression of CD98 in LAT1^KO cells (Fig. 5A). However, only LAT1 reexpression was able to restore leucine transport activity (Fig. 5B). This restoration of AA transport activity in LAT1 rescue cells reestablished AA homeostasis. Indeed, the mTORC1 activity was completely recovered, while no induction of the GCN2 pathway was detectable (Fig. 5C). In LAT1^KO cells transfected with xCT (LAT1^KO xCT), the GCN2 pathway was still activated, but surprisingly, the induction of ATF4 was reduced by the xCT overexpression (Fig. 5C). The mTORC1 activity was also recovered in LAT1^KO xCT cells (Fig. 5C). With regards to the proliferation phenotype, as expected, LAT1 rescue completely restores the growth of LS174T cells, whereas xCT overexpression has only a minimal effect on proliferation that could reflect the increased expression of xCT (Fig. 5C and D). We confirmed these findings in the lung carcinoma A549 cell line. Transient transfection of LAT1 in A549 LAT1^KO cells rescued cell proliferation, whereas xCT had no effect (Supplementary Fig. S5). These results demonstrate that CD98 downregulation in LAT1^KO is not implicated in the severe growth phenotype of LAT1^KO and again highlight the key role EAA transport in cell growth. In addition, the complete growth phenotype rescue by LAT1 reexpression in LS174T- and A549-LAT1^KOillustrates that the phenotypes described were not due to ZFN-mediated off-target effects.

Furthermore, pharmacologic inhibition of LAT1 in six independent cancer cell lines [colon (LS174T, HT29), lung (A549, H1975), and kidney carcinomas (786-O, A498)] did not decrease CD98 expression (Supplementary Fig. S4A), while it mimicked the LAT1^KO growth phenotype. Indeed, JPH203 treatment increased the GCN2 pathway, reduced mTORC1 activity (Supplementary Fig. S4A), and inhibited proliferation (Supplementary Fig. S4B) in all of the cell lines tested. These findings further confirmed that the decreased expression of CD98 in LAT1^KO cells has no impact on their AA stress response and growth phenotype. Combined, these results demonstrate that LAT1 carries the dominant protumoral role in comparison with CD98 across multiple cancer types.

Finally, we wanted to validate our results demonstrating a strong impact of LAT1 removal on cancer cell proliferation using an in vivo xenograft model. LS174T-derived cells were injected subcutaneously into nude mice and tumor growth was monitored (Fig. 6A). Consistent with our in vitro results, WT and CD98^KO tumors grow at the same rate while LAT1 invalidation decreased tumor growth by 91% at day 20 (Fig. 6A). Following tumor growth experiments, protein analysis was performed in three independent tumors of each LS174T-derived cell line (Fig. 6B). Interestingly, CD98 expression is even more reduced in LAT1^KO tumors (93%) compared with in vitro cultures and LAT1 expression, as well, is more dramatically decreased in CD98^KO tumor (91%; Fig. 6B and C). Furthermore, as in vitro, AA-sensing pathways are altered specifically in LAT1^KO tumors: ATF4 is induced, whereas mTORC1 activity is decreased by more than 70% (Fig. 6B and C). Finally, genetic disruption of residual LAT1 activity in CD98^KO cells (dKO) disrupts mTORC1 activity and fully ablates the tumorigenic potential of LS174T cells (Supplementary Fig. S6A–S6C). These results demonstrate the key role of LAT1 for maintenance of tumor growth by sustaining tumor AA homeostasis and mTORC1 activity.

Growth factor signals transduced through the ERK/PI3K/Akt/mTORC1 pathway upregulate nutrient transporters (33). However, these transporters have kept the ability developed in bacteria to be upregulated in response to nutrient depletion, and in addition, some transporters are transcriptionally induced by HIFs in the nutrient-deprived tumor microenvironment (2, 3). Among the many AA transporters identified in human physiology (34, 35), at least three AA transporter systems have emerged as playing a major role in the control of growth and cancer aggressiveness if we consider their increased level of expression and correlation with poor disease prognosis. These are the bidirectional EAA transporter complex CD98/LAT1 (SLC5A7) induced by HIF2 (18) and apparently functionally “coupled” to the high-affinity glutamine transporter ASCT2 (SLC1A5; ref. 36) and the cystine transporter complex CD98/xCT (SLC7A11) essential for glutathione synthesis and resistance to chemotherapeutic agents (32).

In this report, investigating the multifunctional CD98/LAT1 transporter complex in the context of AA stress response, proliferation, and tumor growth, we first showed that LAT1^KO cells are capable to grow (50% reduction compared with WT) in the “super rich” nutrient culture media (DMEM). However, restoring an EAA concentration closer (but still above) to physiological values in the culture media with the exception of tryptophan (Supplementary Table S1) strongly induced an AA stress, restricted mTORC1 activity, and arrested in vitro proliferation in the two human cancer cell lines tested (LS174T, A549), as well as tumor growth for LS174T. However, this major growth defect associated to the single LAT1^KO might not seem surprising considering the concomitant 80% decrease in total CD98 protein and 60% to 65% reduction at the cell surface (Fig. 1). We therefore could have concluded that this drastic phenotype resulted from abrogation of the dual EAA transport activity of LAT1 and of β-integrin signaling. In fact, several arguments led us to fully reject this interpretation.

First, the single CD98^KO in LS174T cells also led to a concomitant severe reduction of LAT1 transport activity (10% residual) with, surprisingly, no detectable growth phenotype. Inhibition of the residual LAT1 activity of CD98^KO elicited AA stress and cell growth, illuminating LAT1 as the key protumoral element of the membrane heterodimer. However, this was a surprising result considering that CD98 (4F2hc) acts as a chaperone for multiple AA transporters, including xCT. The hypothesis that these cells might have acquired another AA transporter, such as SLC6A14 that is capable to transport several AAs, including EAA (37), could not be validated in our cellular models (data not shown). In contrast, pharmacologic inhibition of LAT1 (JPH203) or genetic invalidation of the residual 10% leucine transport activity (CD98^KO/LAT1^KO) was sufficient to induce an AA stress response and suppressed in vitro growth and tumorigenicity in CD98^KO cells.

Second, we fully restored CD98 expression in LAT1^KO cells. As CD98 requires a light chain as a co-chaperone to be expressed in the membrane, we ectopically expressed xCT in LS174T and A549 cells. Even under these conditions, which allow full surface reexpression of CD98 in LAT1^KO cells, we could not rescue the growth defect phenotype of LAT1^KO cells. The LS174T LAT1^KO CD98^high (xCT) cells display only a slightly attenuated AA stress response that might result from the overexpression of the xCT transporter responsible for antioxidant production.

Third, the pharmacologic targeting of LAT1 with JPH203 allowed us to explore six human cancer cell lines expressing different levels of LAT1 (Supplementary Fig. S4A). As expected, JPH203 did not reduce CD98 expression, with the expression being even slightly increased in all cell lines following the concomitant increased expression of LAT1. Interestingly however, targeting LAT1 alone with JPH203 drastically reduced cell proliferation associated with all detected markers of AA stress, including members of the GCN2 and mTORC1 pathways in all six cell lines (Supplementary Fig. S4A and S4B). This finding fully confirmed and extended the genetic disruption of LAT1.

Although not yet fully resolved, the field of mTORC1 activation by AA has rapidly progressed. Three variations of leucine-sensing mechanisms regulating mTORC1 activity have been uncovered. Han and colleagues demonstrated that a cytoplasmic detection of leucine occurs by the tRNA-charging enzyme leucyl-tRNA synthetase, which translocates to the lysosome and promotes mTORC1 activity (38). Wolfson and colleagues provided evidence that sestrin2 is responsible for cytoplasmic leucine sensing and mTORC1 activation (39). Finally, recent work by Milkereit and colleagues suggested a novel mechanism based on leucine lysosomal sensing (40). This last mechanism is particularly appealing in the context of this study, as LAT1 might be the first AA transporter to have two distinct chaperones, CD98 for plasma membrane and LAPTM4b for LAT1 lysosomal membrane expression (40). This finding of dual LAT1 expression could explain the extreme mTORC1 inhibition and growth defects of LAT1^KO cells, contrasting with the nonaffected CD98^KO cells. Intriguingly, we observed increased mTORC1 activity in A549 CD98^KD cells (Supplementary Fig. S3B), which agrees with a recent study on ES-derived fibroblasts CD98^KO cells (41). We propose that by suppression of CD98 the LAT1 expression ratio between lysosomal/plasma membranes may have increased and induced a stronger mTORC1 signal. However, even if mTORC1 is increased in certain cell lines when CD98 is knocked out or knocked down, these results demonstrate that AA homeostasis is drastically disrupted in LAT1^KO cells, whereas it is still maintained in CD98^KOor CD98^KD cells.

Our study highlights the dominant protumoral role of LAT1 in comparison with CD98. Indeed, although the tumoral impacts of integrin signaling are well described (42, 43), a loss of CD98 does not affect the integrin levels in the cell (Supplementary Fig. S1F and S1G), whereas a loss of LAT1 affects the foundation of cell proliferation by preventing essential nutrient import and mTORC1 activation. This dependency leads to cancer cell reliance on LAT1 activity to sustain tumoral AA homeostasis, mTORC1 activation, and tumor growth. Previous studies from our laboratory investigated the potential of disrupting different metabolism-related proteins as therapeutic targets (44–47). Use of the aggressive LS174 xenograft model in these previous studies underlined the problem of the functional protein redundancy in terms of cancer treatment. Indeed, dramatic reductions in tumor growth were obtained only by combined targeting of two carbonic anhydrases (44) or two lactate transporter isoforms (47). Intriguingly, here we show that in the same aggressive LS174 xenograft model, ablation of LAT1 alone is sufficient to abolish tumor growth, demonstrating an absence of redundancy for essential AA transport in certain tumors. Together, these results highlight the essential protumoral role of LAT1 and demonstrate that LAT1 is a promising individual target for future efforts in anticancer drug development.

Hitoshi Endou is the CEO at and has ownership interest (including patents) in J-Pharma. M.F. Wempe has ownership interest (including patents) in and is a consultant/advisory board member for J-Pharma. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Y. Cormerais, R. LeFloch, H. Endou, S.K. Parks, J. Pouyssegur

Development of methodology: Y. Cormerais, R. LeFloch, E. Tambutté, M.F. Wempe, J. Pouyssegur

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Cormerais, S. Giuliano, B. Front, E. Tambutté, H. Endou, J. Pouyssegur

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Cormerais, S. Giuliano, J. Durivault, L.R. de la Ballina, S.K. Parks, J. Pouyssegur

Writing, review, and/or revision of the manuscript: Y. Cormerais, B. Front, P.-A. Massard, M.F. Wempe, S.K. Parks, J. Pouyssegur

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Durivault, P.-A. Massard, L.R. de la Ballina, M.F. Wempe, M. Palacin, J. Pouyssegur

Study supervision: S.K. Parks, J. Pouyssegur

Other (suggested LAT1 mutations to eliminate transport activity without affecting CD98hc interaction and his laboratory generated the tagged versions of human light subunits of LAT1 used in this study): M. Palacin

The authors thank Ludovic Cervera and CytoMed, the IRCANs' Flow Cytometry Facility.

This work was entirely supported by the government of Monaco, including thesis (Y. Cormerais), master (P.A. Massard), and post-doctoral (S.K. Parks) fellowships. This project has also been, in part, supported by the GEMLUC, Ligue Nationale Contre le Cancer (JP, Equipe labellisée), IRCAN, University of Nice, and Centre A. Lacassagne. The materials of CytoMed were supported by the Conseil Général 06, the FEDER, the Ministère de l'Enseignement Supérieur, the Région Provence Alpes-Côte d'Azur and the INSERM, France.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.