Engineering Tumour Cell-Binding Synthetic Polymers with Sensing Dense Transporters Associated with Aberrant Glutamine Metabolism

Increased glutamine uptake toward the elevated glutaminolysis is one of the hallmarks of tumour cells. This aberrant glutamine metabolism has recently attracted considerable attention as a diagnostic and therapeutic target. Herein, we developed glutamine-functionalized polymer to achieve a selective high affinity to tumour cells overexpressing glutaminolysis-related transporter ASCT2. In in vitro study, our developed polymer exhibited faster and higher cellular uptake in tumour cells than that in normal cells. Uptake inhibition study revealed the dominant contribution of ASCT2 to the polymer-cell interaction. Furthermore, the binding affinity of the polymer to tumour cells was estimated to be comparable to that of the potent ligand molecules reported in the literature. In in vivo study, the polymer showed prolonged retention at tumour site after intratumoral injection. This study offers a novel approach for designing tumour cell-binding synthetic polymers through the recognition of dense transporters related to tumour-associated metabolism.


Results
Design and synthesis of glutamine-functionalized polymers. A series of glutamine-functionalized polymers were synthesized by ring-opening polymerization of N-carboxyanhydride and subsequent modification of the side chain ( Supplementary Fig. S1-S4). Briefly, azide-functionalized poly(L-lysine) was first synthesized by ring-opening polymerization of N-carboxyanhydride of ε-trifluoroacetyl-L-lysine and deprotection of trifluoroacetyl group. Then, the side chain was modified to have glutamine moieties through γ-amide linkage. It should be noted that modification of γ-amide group of glutamine is expected to tolerate the transporter-glutamine interaction as reported in a previous study, where glutamine analogues with γ-amide modification strongly inhibited ASCT2 function 17 . Finally, dibenzocyclooctyne modified Cy5 fluorescence dye was reacted with azide terminus of the polymer by copper-free click chemistry, for the detection of polymers by fluorescence in subsequent experiments. The resulting polymers were termed as PLys(Gln)-n (Fig. 1a), where n is the mean degree of polymerization (DP). In this design, polymer length is a crucial factor for multivalent interaction with surface transporters. Accordingly, the polymers having the different length (DP was 30, 50, and 100) were synthesized. Meanwhile, control polymers (termed as PLys(α-Glu)-n (Fig. 1b), n is DP) were prepared in the same manner with minor modification: glutamine moieties were changed to α-glutamyl moieties to investigate the effect of chemical structure of the side chain on binding affinity. Quantitative modification of lysine residue and narrow molecular weight distribution were confirmed for all polymers by 1 H NMR and size exclusion chromatography, respectively (Supplementary Fig. S5-S12). ASCT2 Expression. We next evaluated the expression of ASCT2 in vivo and in vitro. Immunohistochemical analysis of a subcutaneous BxPC3 (human pancreatic cancer cell) tumour in a mouse with anti-human/murine ASCT2 antibody revealed that ASCT2 was distinctly overexpressed in the tumour compared to normal tissues (Fig. 2a, Supplementary Fig. S13), suggesting that ASCT2 can be a potential target transporter on cancer cells. Flow cytometric analysis also demonstrated the overexpression of ASCT2 on BxPC3 cells compared to the HEK293 (human embryonic kidney) cells (Fig. 2b,c). In addition, enzyme-linked immunosorbent assay (ELISA) (c) Illustration of interaction between the glutamine-functionalized polymer and cell surface. The polymer strongly interacts with tumour cell surface by multivalent interaction associated with high transporter density. In contrast, the polymer weakly interacts with normal cell surface because of low transporter density.
SCIentIfIC REPORTS | 7: 6077 | DOI:10.1038/s41598-017-06438-y revealed 4.8-fold higher expression of ASCT2 on BxPC3 cells compared to normal HEK293 cells ( Supplementary  Fig. S14). This overexpression ratio is in line with previous studies, where tumour cells showed approximately 2to 10-fold higher expression of ASCT2 compared to their counterparts 7,18,19 . Thus, we used these cell lines for the assessment of interaction between the developed polymers and ASCT2 in in vitro studies.

Cellular Uptake Analysis.
To examine the cellular interaction of PLys(Gln)-n with cultured tumour cells, the flow cytometric analysis was performed. The cellular uptake was quantified by measuring Cy5 fluorescence intensity from the cells treated with the polymers (Fig. 3a). A series of PLys(Gln)-n exhibited DP-dependent uptake behaviour; PLys(Gln)-100 showed the highest uptake in BxPC3 cells, which was 9.7-fold and 18-fold higher than that of PLys(Gln)-50 and PLys(Gln)-30, respectively. Similar DP-dependent interaction was also observed in HepG2 (human liver cancer) cells ( Supplementary Fig. S15), which overexpress ASCT2 ( Supplementary Fig. S14, ref. 20). According to a previous study, the interaction potency of multivalent polymeric ligand was exponentially enhanced by an increase of the polymer length 21 . Thus, this drastically high cellular uptake of PLys(Gln)-100 is probably due to the multivalent interaction between the polymer and the tumour cells.
To investigate the tumour specificity of PLys(Gln)-100, the time-dependent cellular uptake was evaluated using HEK293 cells as well as BxPC3 cells (Fig. 3b). PLys(Gln)-100 exhibited faster uptake in BxPC3 cells compared to HEK293 cells. After 8 h incubation, PLys(Gln)-100 uptake in BxPC3 cells was 4.7-fold higher than that in HEK293 cells. These results are consistent with the scheme that PLys(Gln)-100 might strongly interact with tumour cells overexpressing ASCT2 through the strong multivalent effect, while weakly interacting with normal cells because of less binding sites as illustrated in Fig. 1c.
To get insight about the effect of chemical structure on this cellular interaction, we investigated the time-dependent cellular uptake of PLys(α-Glu)-100. This polymer was used as a control polymer because it has same molecular weight and neutral charge with PLys(Gln)-100 with the different position of free amine group at the side chain. As shown in Fig. 3c, PLys(α-Glu)-100 revealed slightly higher uptake in BxPC3 cells compared to HEK293 cells. Considering that glutamate weakly interacted with ASCT2 22, 23 , it could be assumed that PLys(α-Glu)-100 having glutamate-like structure at the side chain might interact with ASCT2 to some extent and exhibit multivalent effect. Consistent with this assumption, a series of PLys(α-Glu)-n also showed DP-dependent uptake behaviour in cultured cells ( Supplementary Fig. S16). However, compared to PLys(Gln)-100, PLys(α-Glu)-100 exhibited slower uptake in BxPC3 cells and lower uptake ratio between BxPC3 and HEK293 cells. The relatively high affinity of PLys(Gln)-100 to tumour cells is probably due to its chemical structure at the side chain, which is similar to potent ASCT2 inhibitor having γ-amide modified glutamine structure 17, 24 . Transporter Specificity. To examine the transporter specificity of the developed polymers, cellular uptake was evaluated with several inhibitors for glutamine transporters: ASCT2, system A, system L, and system N transporters. For their inhibitors, we used O-benzyl-L-serine (BzlSer, ASCT2 inhibitor), 2-(methylamino)isobutyric acid (MeAIB, system A inhibitor), 2-amino-2-norbornanecarboxylic acid (BCH, system L inhibitor), and L-glutamine (Gln, system N inhibitor). As shown in Fig. 4a, only BzlSer exhibited strong concentration-dependent inhibitory effect, while the other inhibitors showed negligible inhibitory effect, indicating the dominant contribution of ASCT2 to the uptake of PLys(Gln)-100. Similar inhibitory effects were also observed for PLys(Gln)-50 ( Supplementary Fig. S17). Among above-mentioned glutamine transporters, ASCT2 has relatively high affinity with glutamine (Michaelis constant (K m ) is 20-60 μM 12, 23 , whereas the other transporters has K m in the range of approximately 150-1600 μM [25][26][27][28] ). Because of this relatively strong interaction between ASCT2 and glutamine molecule, ASCT2 could be a dominant target transporter for PLys(Gln)-n. Moreover, the overexpression of ASCT2 on BxPC3 cells as shown in Fig. 2c may also contribute to the dominant role of ASCT2 for tumour cell-selective interaction of PLys(Gln)-n. Cellular uptake of PLys(α-Glu)-100 was also inhibited by only BzlSer while its inhibitory effect was higher compared to that of PLys(Gln)-100 (Fig. 4b), suggesting that PLys(α-Glu)-100 modestly interacted with ASCT2. Subcellular Distribution of the Polymers. Subcellular distribution of the polymer was then investigated using confocal laser scanning microscopy (CLSM) to assess the internalization of the polymers. As shown in Fig. 5a, in the BxPC3 cells incubated with PLys(Gln)-100 at 37 °C, the polymer (red) was located at intracellular space, and colocalized (yellow) with the late endosome/lysosome indicator, Lysotracker Red (green). In contrast, when BxPC3 cells were incubated with the polymer at 4 °C to suppress the endocytosis, the polymer was located only on the cellular membrane, and did not colocalize with the late endosome/lysosomes. In addition, mean  fluorescence intensity from the cells incubated with the polymer at 4 °C was significantly lower compared to that of 37 °C incubation (Fig. 5c). These results indicate that PLys(Gln)-100 first attached to the cell membrane through the polymer-ASCT2 interaction, and the attached polymer was internalized by endocytosis. Note that similar behaviours in subcellular distribution were observed for PLys(α-Glu)-100 and PLys(Gln)-50 ( Fig. 5b,d, Supplementary Fig. S18), indicating that these polymers were also internalized by endocytosis after interacting with ASCT2 on cell surface.
Binding Affinities of the Polymers to ASCT2 on Tumour Cells. To evaluate the binding affinity of the polymers to ASCT2, cell-based competitive inhibition study was carried out using BzlSer. Because BzlSer is a competitive inhibitor for ASCT2 with an inhibition constant (K i ) of 0.9 mM 29 , apparent K d values of these polymers can be estimated by assuming a simple inhibition model 30 . PLys(Gln)-100 showed apparent K d value of 62 nM, which is comparable to the K d value of potent ligands 13 . By contrast, K d for PLys(Gln)-50 could not be estimated in this method since the inhibition curve of PLys(Gln)-50 indicated incomplete inhibition within the examined BzlSer concentration ( Supplementary Fig. S19). Meanwhile, PLys(α-Glu)-100 showed apparent K d value of 250 nM. The K d values of PLys(Gln)-100 and PLys(α-Glu)-100 were consistent with cellular uptake behaviour as shown in Fig. 3.
In Vivo Tumour Retention. Finally, to examine in vivo binding ability, the polymers were intratumorally injected to subcutaneous BxPC3 tumours in mice, and their retention in the tumour was evaluated by measuring fluorescence intensity at tumour site using in vivo imaging system (Fig. 6). PLys(Gln)-50 was most rapidly eliminated from the tumour because PLys(Gln)-50 had low binding affinity to ASCT2 on BxPC3 cells as discussed above. Compared with PLys(α-Glu)-100, PLys(Gln)-100 exhibited longer retention in the tumour. This prolonged retention of PLys(Gln)-100 can be attributed to its higher binding affinity to the tumour cells, which is in line with the apparent K d values.

Discussion
Previous studies revealed potential of aberrant glutamine metabolism as a target of cancer therapy and diagnosis 4,[7][8][9][10][11]31 , and this aberrant metabolism accompanies overexpression of the corresponding amino acid transporters including ASCT2 6-8 as we also demonstrated in in vitro and in vivo ( Fig. 2 and Supplementary Fig. S13). While previous research utilized increased cellular uptake of glutamine to visualize accelerated metabolism 10, 11, 31 , this study focused on these overexpressed transporters on tumour cells, and developed the polymer functionalized with glutamine molecules for exerting selective affinity to tumour cells through multivalent effect.
Although exact mechanism of interaction between ASCT2 and glutamine or glutamine-analogues is not fully understood, a previous study has proposed key structures of glutamine-analogues to get a high affinity to ASCT2: α-amine and α-carboxyl groups for charge interaction, γ-amide group for hydrogen binding, and modified side chain for hydrophobic interaction 17 . As most potent ASCT2 inhibitors have α-amine and α-carboxyl groups, these chemical structures appeared to be important for glutamine-ASCT2 interaction. Indeed, PLys(Gln)-100 that has α-amine and α-carboxyl groups at the side chain showed higher affinity both in in vitro and in vivo compared to PLys(α-Glu)-100 although the difference between these polymers is just the position of amine group at the side chain. In addition to α-amine and α-carboxyl groups, glutamine-like structure should be important for achieving ASCT2 selective interaction because amino acids transporters generally recognize these functional groups. For instance, α-amine and α-carboxyl groups are key components for the substrates of system L transporter 1 (LAT1) 32 , which is overexpressed on a variety of tumour cells [33][34][35] . Given that LAT1 is reported to be expressed on BxPC3 cells 36 , PLys(Gln)-100 was expected to interact with LAT1 as well as ASCT2 if amino acids transporters recognize just the side chain terminus of the polymers. However, as shown in Fig. 4a, LAT1 inhibitor BCH did not inhibit the PLys(Gln)-100 uptake in BxPC3 cells while ASCT2 inhibitor BzlSer strongly inhibited the uptake. This ASCT2-selective interaction of PLys(Gln)-100 suggests that ASCT2 may recognize not merely α-amine and α-carboxyl groups but whole glutamine-like structure (including γ-amide groups) of the side chain.
On the other hand, given that PLys(Gln)-100 and PLys(α-Glu)-100 has the same valency and density of the binding motifs, moderate interaction between PLys(α-Glu)-100 and ASCT2 might be attributed to the recognition of glutamate-like structure at the side chain by ASCT2. This explanation is consistent with previous reports, where glutamate did not interact with system L and system N glutamine transporters but weakly interacted with ASCT2 22,23,27,37 . In addition, this glutamate-ASCT2 interaction was reported to be pH-dependent; a decrease of pH could enhance the interaction capacity of glutamate with ASCT2 17,23 . In line with this report, PLys(α-Glu)-100 showed higher uptake in BxPC3 cells at pH 6.5 compared to that at pH 7.4 ( Supplementary  Fig. S20). However, this increased cellular uptake of PLys(α-Glu)-100 at pH 6.5 was still lower than that of PLys(Gln)-100, indicating that even at the intratumoral pH (~ 6.5) PLys(Gln)-100 should be the preferred structure for achieving high affinity to ASCT2.
As aforementioned, the fine-tuning of DP and chemical structure at the side chain permits the selective interaction of the polymer to dense ASCT2. Given that the overexpression ratio of ASCT2 between BxPC3 and  Fig. S14) is comparable to that in clinical studies 7,18,19 , the tumour-selective interaction of the polymer as shown in Fig. 3 strongly indicates the potential in vivo activity of our synthesized polymer for tumour-targeting ligand. In our synthetic strategy, the Cy5 fluorescence dye can be easily replaced with the other functional molecules including drugs, diagnostic agents, and functional macromolecules. Hence, the glutamine-based ligand would be potentially useful as a versatile platform for delivering such functional molecules to diseased sites related to aberrant glutamine metabolism. However, systemic application of this polymer still requires optimization. For example, PLys(Gln)-100 would be rapidly eliminated from the body because the molecular weight of the polymer is lower than the threshold of renal clearance 38 ; it would have few chances of interacting with tumour tissue after systemic administration. Thus, the design of glutamine-based ligand with suitable architecture for long blood circulation and appropriate binding affinity to the target cell is our future interests.
In summary, we developed the novel glutamine-functionalized polymer with a high affinity to tumour cells overexpressing glutaminolysis-related transporter ASCT2. The glutamine-functionalized polymer exhibited enhanced cellular uptake in the cultured cells by strong interaction with ASCT2. Even in in vivo environment, the glutamine-functionalized polymer showed prolonged retention at tumour site, indicating the potent binding affinity to tumour. To the best of our knowledge, this is the first report utilizing glutamine as a functional molecule showing a high affinity to tumour cells, inspired by tumour-associated metabolism. This study offers a fundamental molecular design to target dense transporters associated with tumour-specific aberrant metabolisms. Immunohistochemistry. To establish the BxPC3 xenograft model, the cells (5.0 × 10 6 cells) were subcutaneously inoculated to BALB/c nu/nu mice. When tumour volume reached approximately 150 mm 3 , the mice were perfused using saline and fixed using 4% paraformaldehyde. Tissues were excised and frozen in OCT compound (Sakura Finetek Japan Co., Ltd., Tokyo, Japan). The frozen samples were sectioned at 4-μm thickness and washed with PBS. After 1 h blocking treatment using 5% skim milk in PBS, the sections were incubated with rabbit anti-human/murine ASCT2 polyclonal antibody (1:100 dilution in PBS containing 5% skim milk, LSBio, Seattle, WA) overnight at 4 °C. Then, the sections were washed with PBS three times, incubated with Alexa Fluor 568-conjugated goat anti-rabbit IgG (1:800 dilution in PBS containing 5% skim milk, Thermo Fischer Scientific) at ambient temperature for 45 min. After washing with PBS three times, the sections were incubated with Hoechst 33342 (1:1000 dilution in PBS) for staining the nuclei, and mounted in Vectashield mounting media (Vector Laboratories, Burlingame, CA). The obtained tissue sections were observed using a fluorescence microscope (BZ-X710, Keyence, Osaka, Japan).

Materials. N-Carboxyanhydride
Flow cytometric analysis of ASCT2 expression. BxPC3 cells and HEK293 cells (2.0 × 10 5 cells) were transferred into 1.5 mL tube and washed with assay buffer (PBS containing 1% BSA). Then, the cells were resuspended in 100 μL of rabbit anti-human ASCT2 polyclonal antibody solution (1:50 dilution in the assay buffer, Santa Cruz Biotechnology, Dallas, TX) or isotype control rabbit IgG solution (1:250 dilution in the assay buffer, abcam, Cambridge, MA) and incubated for 1.5 h on ice. The cells were washed twice with 500 μL of the assay buffer and resuspended in 100 μL of DyLight 650-conjugated anti-rabbit IgG solution (1:500 dilution in the assay buffer, abcam). After 45 min incubation on ice in the dark, the cells were washed twice with 500 μL of the assay buffer and resuspended in the 500 μL of the assay buffer, followed by analysis using flow cytometer (Guava easy-Cyte 6-2 L, Merck Millipore, Billerica, MA) (ex/em = 642 nm/661 nm).

Confocal laser scanning microscopic observation.
To visualize the subcellular distribution of the polymers, BxPC3 cells treated with each polymer at 37 °C or 4 °C were observed using a confocal laser scanning microscope (LSM710, Carl Zeiss, Oberkochen, Germany). BxPC3 cells (5 × 10 4 cells/dish) were seeded into 35 mm glass-based dish (Asahi Glass Co., Ltd., Tokyo, Japan) and incubated at 37 °C for 24 h. For the treatment at 37 °C, the cells were washed with PBS, and incubated in 1 mL of PBS with 10% FBS and 1 μM of the polymer at 37 °C for 2.5 h. To stain late endosome/lysosome and nuclei, LysoTracker Red (Thermo Fischer Scientific) and SCIentIfIC REPORTS | 7: 6077 | DOI:10.1038/s41598-017-06438-y Hoechst 33342 were added to the solution. After additional 30 min incubation at 37 °C, the cells were washed twice with PBS, and observed in fresh medium using LSM710 equipped with an incubator (37 °C, 5% CO 2 in humidified atmosphere). For the treatment at 4 °C, the cells were cooled at 4 °C for 10 min and washed with cold PBS, followed by incubation in 1 mL of PBS with 10% FBS and 1 μM of the polymers at 4 °C for 2.5 h. Then, LysoTracker Red and Hoechst 33342 were added to the solution, and the cells were incubated at 4 °C for another 30 min. The cells were washed twice with cold PBS, fixed with 4% paraformaldehyde in PBS, and observed in fresh PBS using LSM710. Cy5, LysoTracker Red, and Hoechst 33342 were excited using laser light at 633 nm, 561 nm, and 405 nm, respectively. Cellular uptake at 4 °C. BxPC3 cells (5.0 × 10 4 cells/well) were seeded into 24-well plates and incubated at 37 °C for 24 h. The cells were cooled at 4 °C for 10 min and washed with cold PBS, followed by the incubation in 500 µl of the assay buffer (10% FBS in PBS) containing 1 μM of the polymers. After 3 h incubation at 4 °C, the cells were washed twice with cold PBS, detached with 150 μL of trypsin, and suspended with 300 μL of the assay buffer. The cells were analysed using flow cytometer, and the obtained Cy5 fluorescence intensities were expressed as a percentage of the fluorescence intensity obtained from cells treated at 37 °C.
Cell-based competitive inhibition. BxPC3 cells (5.0 × 10 4 cells/well) were seeded into 24-well plates and incubated at 37 °C for 24 h. For the inhibition of PLys(Gln)-100, the cells were cooled at 4 °C for 10 min, washed with cold PBS, followed by incubation at 4 °C for 2 h in 500 μL of the assay buffer (10% FBS in PBS) containing 1 μM of PLys(Gln)-100 and BzlSer in a concentration ranging from 30 mM to 0.04 mM. After the incubation, the cells were washed with cold PBS, treated with 150 μL of trypsin, and suspended with 300 μL of the assay buffer. The cells were analysed using flow cytometer, and half maximal inhibitory concentrations (IC 50 ) of BzlSer were calculated. Apparent dissociation constant (K d ) of the polymers were estimated using the obtained IC 50 values, following Cheng-Prusoff equation (1) 30, 39 : where K i is the apparent inhibition constant of BzlSer for ASCT2 (0.9 mM), and [L] is the polymer concentration in this assay. For the inhibition of PLys(Gln)-50 and PLys(α-Glu)-100, concentration of both polymers was changed to 3 μM.
In vivo tumour retention. To establish the BxPC3 xenograft model, the cells (5.0 × 10 6 cells) were subcutaneously inoculated to BALB/c nu/nu mice (female, 4 weeks old). When average tumour volume reached approximately 150 mm 3 , 20 μl of PLys(Gln)-100, PLys(Gln)-50, or PLys(α-Glu)-100 solution (40 μM in PBS) was intratumorally injected (n = 3). At 0.25, 1, 4, and 8 h after injection, the mice were imaged by in vivo imaging system (IVIS, Perkin Elmer, Waltham, MA) using 640 nm/680 nm ex/em filter. Average radiant efficiency of the tumour region was quantified using Living Image software. Statistical analysis. Statistical analysis was performed using Student's t-test for the comparison of the mean between two groups, and ANOVA with Tukey's multiple comparison test for the comparison of the mean among more than three groups. The p values less than 0.05 were considered as statistically significant.