Amino acid transporter LAT1 is expressed on cancer cell-derived exosomes with potential as a diagnostic and prognostic biomarker Authors: Yumiao Liu, Ryuichi Ohgaki, Hiroki Okanishi, Minhui Xu, Yoshikatsu Kanai Published: 2024 ABSTRACT AbstractL-type amino acid transporter 1 (LAT1) is upregulated in various cancers and contributes to the growth and proliferation of cancer cells. Previous clinicopathological studies have revealed associations between the high expression of LAT1 and the poor prognosis in multiple cancer types. However, in those studies, the expression of LAT1 has been evaluated solely by immunohistochemistry on resected tumors or tissue biopsies. Cancer cell-derived exosomes are attracting increasing attention as a resource of diagnostic, prognostic, and therapeutic biomarkers. In this study, we revealed the expression of LAT1 on exosomes from pancreatic cancer T3M-4 cells by western blotting, immunoprecipitation, and immuno-transmission electron microscopy. LAT1 was generally detected by western blotting and ELISA on exosomes from several pancreatic, biliary tract, and ovarian cancer cells. Notably, similar trends existed between the abundance of exosomal LAT1 and the cellular LAT1 expression levels. The expression of LAT1 on exosomes in vivo was verified using the peritoneal wash from intraperitoneal T3M-4 tumor-bearing mice. These results indicate that exosomal LAT1 released from cancer cells holds significant potential as a novel biomarker for diagnosis and prognosis.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-024-79425-9.Keywords: Exosome, Cancer, Biomarker, Amino acid transporter, LAT1, SLC7A5Subject terms: Cancer therapy, Tumour biomarkers, Biliary tract cancer, Pancreatic cancer, Carrier proteins, Membrane proteins Introduction Exosomes are small membrane vesicles produced by most eukaryotic cells1,2. Originating from multivecular bodies (MVBs), exosomes contain various cargo molecules derived from their parental cells, such as proteins, mRNAs, miRNAs, lncRNAs, lipids, and metabolites. An accumulating number of studies have demonstrated that cancer cell-derived exsosomes are involved in pathological processes both in autocrine and paracrine manners, including cancer stemness, immune escape, drug resistance, angiogenesis, and premetastatic niche formation3. Additionally, exosomes have been attracting growing interest as a resource of noninvasive cancer biomarkers, because they are secreted into body fluids, including blood, saliva, urine, milk, and ascites1,2. Many exosomal proteins have been reported and extensively studied as potential biomarkers for early cancer diagnosis, prognostic monitoring, and assessment of the treatment efficacy4,5. Cancer cells generally exhibit an increased nutrient uptake to fulfill their metabolic demands for rapid growth and proliferation. L-type amino acid transporter 1 (LAT1; SLC7A5)6, which preferentially transports large neutral amino acids, is known to be upregulated in various types of cancer cells7. Amino acids transported into cancer cells by LAT1 are not only directly utilized as building blocks for the protein synthesis, but also metabolically fuel the synthesis of macromolecules such as nucleotides, lipids, and ATP. LAT1 also contributes to the regulation of these multiple biosynthetic and metabolic processes through the uptake of leucine as an essential signaling molecule that activates of a serine-threonine kinase complex mTORC1 (mechanistic target of rapamycin complex 1)8–11. Higher expression of LAT1 is associated with the poorer prognosis of patients in multiple cancer types12,13, including the most refractory pancreatic and biliary tract cancers14–16. Intriguingly, LAT1 exhibits a limited tissue distribution, such as in brain blood barrier and placental barrier17–20. Based on these pathological characteristics, we have been proposed LAT1 as a promising novel target for cancer treatment, diagnosis, and prognosis21. It is especially of note that an LAT1 inhibitor nanvuranlat (JPH203 or KYT-0353)22 is under clinical evaluation as the first-in-class anti-cancer agent. In the recently conducted first phase II clinical trial against advanced biliary tract cancer, nanvuranlat by monotherapy demonstrated a significant improvement in progression-free survival of patients (UMIN000034080)23. Therefore, developing noninvasive methods to determine the expression levels of LAT1 in cancer tissues may lead to broad clinical applications not only for cancer detection and prognosis but also for companion diagnosis in the future cancer therapy with LAT1 inhibitors. However, previous clinicopathological studies on the prognostic values of LAT1 have been conducted solely with specimens of surgically resected tumor or tissue biopsy. Considering the broadly upregulated expression of LAT1 in cancer tissues, we hypothesized that LAT1 is expressed on cancer cell-derived exosomes. Indeed, several in vitro large-scale proteomics studies with bladder cancer, colorectal cancer, ovarian cancer, prostate cancer, melanoma, and neuroblastoma cells have reported LAT1 as a candidate cargo protein on cancer cell-derived exosomes24–30 However, because none of these studies have been conducted with a specific focus on the presence of LAT1 on exosomes, further detailed in vitro confirmation and in vivo validation remain to be performed. Moreover, it has not been investigated whether the abundance of exosomal LAT1 is associated with the expression level of LAT1 in the parental cancer cells. In this study, we first characterized the expression of LAT1 on exosomes secreted from pancreatic cancer T3M-4 cells. Then, the cellular and exosomal LAT1 expression levels were comparatively analyzed in vitro using multiple cancer cell lines. We further established an exosome ELISA method for detecting LAT1 and confirmed the in vivo expression of LAT1 on exosomes in the peritoneal wash of intraperitoneal tumor-bearing mice. Results LAT1 is expressed on exosomes isolated from cell culture supernatant of T3M-4 cells To examine whether LAT1 is expressed on cancer cell-derived exosomes in vitro, an exosome fraction was prepared from the cell culture supernatant (CCS) of pancreatic cancer T3M-4 cells by ultracentrifugation method. Nanoparticle tracking analysis demonstrated a homogeneous particle size distribution in the obtained exosome fraction (Supplementary Fig. S1). Mean and mode sizes of particles were confirmed to be within the acceptable size range of exosomes (30–200 nm)31, i.e., 178 ± 3.5 nm (mean) and 147 ± 4.8 nm (mode), respectively. LAT1 and two exosome markers, CD81 and CD9, were detected in the exosome fraction by western blotting, whereas non-exosome proteins, GM130 and calnexin, were not (Fig. 1a). CD81 and CD9 were considerably enriched in the exosome fraction compared to the crude membrane fraction, which represents membrane-associated proteins expressed in the host T3M-4 cells. Unlike CD81 and CD9, LAT1 was slightly less abundant in the exosome fraction than in the crude membrane fraction. A lower amount of LAT1 was also detected in 10 K × g precipitate, suggesting its expression in larger extracellular vesicles other than exosomes, such as microvesicles and apoptotic bodies1,2. The presence of LAT1 on exosomes was further evidenced by immunoprecipitation using anti-CD81 and CD9 antibodies (Fig. 1b). Immuno-transmission electron microscopy of the exosome fraction demonstrated that LAT1 is localized on the exosome-like membrane vesicles (~ 100 nm, Fig. 1c). The C-terminus of LAT1, which is recognized by the anti-LAT1 antibody used in this study7, is supposed to be orientated to the exosome lumen. Consistently, the immuno-gold signals of LAT1 were exclusively localized in the luminal side of membrane vesicles. These results demonstrate that LAT1 is expressed on the exosomes secreted by T3M-4 cells into CCS. Detection of LAT1 in exosome fraction isolated from T3M-4 cell culture supernatant. (a) Western blotting of LAT1, CD81, CD9, GM130, and calnexin in 10 K × g precipitate and exosome fraction (100 K × g precipitate) of T3M-4 CCS, as well as in crude membrane fraction of T3M-4 cells. The same amounts of proteins were loaded into each lane. (b) Immunoprecipitation of exosomes from T3M-4 CCS. Exosome fraction collected by ultracentrifugation was subjected to immunoprecipitation with anti-CD81 and CD9 antibodies. Beads pre-treated with Laemmli buffer were used for negative control samples (Neg. control). LAT1, CD81, and CD9 were detected by western blotting. (c) Transmission electron microscopy image of the exosome fraction immuno-gold labbeled with anti-LAT1 antibody. Gold particles (5 nm) were present in the luminal side of an exosome-like membrane vesicle. Scale bar = 100 nm. (d) Western blotting of LAT1 (left) and 4F2hc (right) in the crude membrane and exosome fractions of T3M-4 cells under non-reducing DTT(−) and reducing DDT(+) conditions. The bands representing LAT1-4F2hc heterodimer and their monomers are indicated with arrowheads. For the functional expression as an amino acid transporter, LAT1 requires a covalent assembly with a type II transmembrane protein 4F2hc (SLC3A2) via an extracellular disulfide bond6. By western blotting under reducing DTT(+) conditions, LAT1 and 4F2hc were detected exclusively as monomers at ~ 37 KDa and ~ 75 KDa, respectively, in both the crude membrane and exosome fractions. Under non-reducing DTT(−) condition, in addition to the bands of monomers, LAT1 and 4F2hc were detected in the major bands of identical molecular size at ~ 150 KDa, which correspond to LAT1-4F2hc heterodimers (Fig. 1d). These results indicate that LAT1 on exosomes from T3M-4 cells mainly exists in the form of heterodimeric complex with 4F2hc as it does on the plasma membrane. Abundance of LAT1 on exosomes reflects its cellular expression levels in cancer cell lines We then perfomed a comparative analysis on the amount of exosomal LAT1 using multiple cancer cell lines. In addition to T3M-4 cells, the exosome fractions were prepared from CCS of two additional pancreatic cancer (SUIT-2, CFPAC-1), three biliary tract cancer (HuCCA-1, HuCCT1, and OCUG-1), and one cervical cancer (HeLa) cell lines. Nanoparticle tracking analysis was conducted on the exosomes of these cell lines (Supplementary Fig. S1). The mean and mode sizes of particles were confirmed to be within the accepted size range of exosomes (30–200 nm): SUIT-2 (146 ± 3.8 nm [mean], 124 ± 3.3 nm [mode]), CFPAC-1 (170 ± 4.7 nm [mean], 129 ± 2.2 nm [mode]), HuCCA-1 (175 ± 4.9 nm [mean], 158 ± 5.3 nm [mode], HuCCT1 (170 ± 7.2 nm [mean], 132 ± 10.1 nm [mode]), OCUG-1 (172 ± 7.2 nm [mean], 137 ± 5.0 nm [mode]), and HeLa (203 ± 7.8 nm [mean], 146 ± 8.0 nm [mode]). Western blotting analysis of the exosome fractions revealed the presence of two exosome markers, CD9 and CD81, and the absence of an exosome-negative marker GM130, supporting the successful enrichment of exosomes. LAT1 was detected in the exosome fractions from all the analyzed cancer cell lines, while exhibiting a considerable variation in its abundance (Fig. 2a). For comparison, the cellular expression levels of LAT1 were analyzed using crude membrane fractions. Densitometric quantification of the immunoreactive bands in western blotting demonstrated that T3M-4, HuCCA-1, SUIT-2, and HeLa cells exhibit relatively higher expressions of LAT1 compared to CFPAC-1, HuCCT-1, and OCUG-1 cells in the crude membrane fractions (Fig. 2b). Interestingly, the exosomes derived from cancer cell lines with higher LAT1 expression, except SUIT-2 cells, tended to harbor higher amounts of LAT1 compared to those originating from cells with lower LAT1 expression. Given the considerable differences in signal intensities of CD9 and CD81 observed in exosome fractions across cell lines (Fig. 2a), we employed two normalization methods for quantifying exosomal LAT1, i.e., by the exosomal protein amount and fraction volume. Both normalizations yielded consistent results, suggesting that the observed variations in exosomal CD9 and CD81 likely reflect the intrinsic differences in their abundance on exosomes between cell lines as observed in previous studies32–35, rather than artificial disparities in the purity of exosomes among samples. These results indicate that LAT1 is contained in cancer cell-derived exosomes, largely reflecting its cellular expression levels in their parental cells. Comparison of the LAT1 expression in crude membrane and exosome fractions between cancer cell lines. (a) Western blotting of LAT1, CD81, CD9, and GM130 in the crude membrane and exosome fractions of pancreatic cancer (T3M-4, SUIT-2, and CFPAC-1; red), biliary tract cancer (HuCCA-1, HuCCT1, and OCUG-1; blue), and ovarian cancer (HeLa; green) cell lines. (b) Densitometric quantification of the LAT1 protein amount shown in (a) in the crude membrane (left) and exosome fractions (middle; normalized by exosomal protein mount, right; normalized by exosomal fraction volume). Data shown are mean ± SEM of three independent experiments (n = 3). Establishment of an ELISA method for detecting exosomal LAT1 We next established an ELISA method for more rapid and quantitative detection of LAT1 on cancer cell-derived exosomes. This assay system utilized the exosome capturing by T-cell immunoglobulin domain and mucin domain-containing protein 4 (Tim4), which specifically binds to phosphatidylserine displayed on the outer surface of exosomes36, followed by the detection with biotinylated anti-LAT1 antibody and HRP-conjugated streptavidin. Exosomes were lysed by a brief sonication before loading on to Tim4-coated plates so that the anti-LAT1 antibody could access its epitope in the exosomal lumen. Detection of LAT1 in the exosome fraction from T3M-4 CCS exhibited a linearity between 50 and 1,000 ng protein with the R2 value of 0.9960 (Fig. 3a). Quantification of the LAT1 expression on exosomes from CCS of different cancer cell lines by this ELISA method normalized by exosomal protein amount also revealed a clear tendency that the abundance of LAT1 in exosomes are associated with its expression levels in their parental cancer cells (Fig. 3b), in a good agreement with the densitometry of western blotting analysis shown in Fig. 2b. Quantification of LAT1 in exosome fractions from cancer cell lines by ELISA. (a) Linearity of the ELISA method for detecting exosomal LAT1 confirmed using serial dilutions of the exosome fraction from T3M-4 CCS. (b) The LAT1 protein amount in exosome fractions of pancreatic cancer (T3M-4, SUIT-2, and CFPAC-1; red), biliary tract cancer (HuCCA-1, HuCCT1, and OCUG-1; blue), and ovarian cancer (HeLa; green) cell lines quantified by ELISA. Data were normalized by protein amount and shown as mean ± SEM of three independent experiments (n = 3). Exosomes from peritoneal wash of tumor-bearing mouse containg LAT1 The expression of LAT1 on cancer cell-derived exosomes in vivo was investigated using a mouse tumor model constructed by intraperitoneal dissemination of T3M-4 cells. One month after grafting, bloody ascites were developed in the abdominal cavity in 80% of tumor-bearing mice. Tumor nodules were formed on the surface of peritoneum, diaphragm, mesentery, bowel, and reproductive organs (Fig. 4a). Histology and immunohistochemistry clearly demonstrated that LAT1 is selectively expressed in the tumor-forming T3M-4 cells but not in adjacent stromal cells of mouse origin (Fig. 4b, c). Detection of LAT1 in exosome fraction from peritoneal wash of tumor-bearing mice. (a) Gross appearance of the peritoneal cavities of control (left) and T3M-4 tumor-bearing (right) mice. Tumor masses of T3M-4 cells were indicated with white arrowheads. (b) and (c) Hematoxylin-eosin staining (b) and immunohistochemistry of LAT1 (c) on a T3M-4 tumor paraffin section. S, stromal tissue; T, tumor tissue. Scale bars = 100 μm. (d) Nanoparticle tracking analysis of exosome fractions preapared from peritoneal wash of control (left) and T3M-4 tumor-bearing (right) mice. Data shown are mean (black line) ± SEM (red error bars) of five technical replicates (n = 5). (e) Quantification of the LAT1protein amount in exosomes from peritoneal wash by ELISA. Statistical significance was tested between control (white, n = 7) and T3M-4 tumor-bearing (red, n = 8) mice by two-tailed unpaired Student’s t-test (* p < 0.01). Data were normalized by protein amount and shown as mean ± SEM. Exosome fractions were prepared by ultracentrifugation method from peritoneal wash of control and tumor-bearing mice. Nanoparticle tracking analysis resulted in a detection of single peak of exosomes from tumor-bearing mouse at the sizes of 155 ± 2.8 nm (mean) and 126 ± 5.9 nm (mode) with a total number of 9.75 ± 1.20 × 109 particles/mouse (Fig. 4d). Even though exosomes of the similar sizes were also isolated from control mouse, a drastically less number of 0.56 ± 0.14 × 107 particles/mouse was detected. When quantitatively analyzed by ELISA, the exosomal LAT1 expression normalized by protein amount was significantly higher in the peritoneal wash of tumor-bearing mice compared to that of control mice (Fig. 4e). These results demonstrate that LAT1 is present on the cancer-cell derived exosomes in vivo isolated from the peritoneal wash of T3M-4 intraperitoneal tumor model. Discussion The present study verified the expression of LAT1 on cancer cell-derived exosomes, by western blotting, immunoprecipitation, and immuno-transmission electron microscopy on exosomes secreted from pancreatic cancer T3M-4 cells (Fig. 1). In the subsequent experiment with pancreatic, biliary tract, and ovarian cancer cell lines, we confirmed the release of LAT1-containing exosomes from all the tested cancer cells (Figs. 2 and 3). These observations, together with the datasets from previous proteomics studies24–30, strongly suggest that LAT1 is contained as a cargo protein on exosomes derived from a wide range of cancer cells. Unlike representative exosome markers CD81 and CD9, LAT1 was not enriched in the exosome fraction of T3M-4 cells (Fig. 1a). This suggests that mature LAT1 expressed on the plasma membrane or on the organellar membranes during its biosynthesis, or both, is passively incorporated into the inner vesicles of MVBs and extracellularly released as exosomal LAT1. It is also of note that a lower but detectable amount of LAT1 was recovered in the 10 K ×g precipitate of T3M-4 CCS, suggesting its broad distribution on extracellular vesicles other than exosomes (Fig. 1a). Most importantly, the abundance of exosomal LAT1 showed a similar trend with their cellular expression levels in parental cells (Figs. 2 and 3). We also demonstrated the expression of LAT1 by ELISA on exosomes isolated from the peritoneal wash of T3M-4 tumor-bearing mice (Fig. 4). Considering the pathological and therapeutic significance of LAT1 in cancer21,37, the findings of this study emphasize the potential of exosomal LAT1 as a novel biomarker for cancer detection, prognosis, and prediction of treatment efficacy. From this perspective, the observation in SUIT-2 cells that exhibited considerably low LAT1 amount on exosomes despite the relatively high LAT1 expression at the cellular level requires careful consideration. The molecular mechanisms underlying this apparent inconsistency remain elusive. It is unlikely due to a specifically lower number of exosomes released from SUIT-2 cells compared to other cell lines, as evidenced by nanoparticle tracking analysis (Supplementary Fig. S1). Possible explanations for this phenomenon include defects in the incorporation of LAT1 as an exosomal cargo protein and selective distributions of LAT1 to sub-regions of the plasma membrane or organelle membranes that are not en route to the synthesis of exosomes in this specific cell line. To further validate the significance of exosomal LAT1 as a biomarker, it is important to investigate the prevalence of cell lines exhibiting trends similar to that in SUIT-2 cells and elucidate the underlying molecular mechanisms through the comparative analysis of exosome biogenesis and loading processes of exosomal LAT1 between cell lines. Exosomes contain several distinct classes of cargo molecules1,2. Because well-established molecular biological techniques allow us to amplify RNAs and DNAs, these intraexosomal cargos may generally propose biomarkers of higher sensitivity. On the other hand, their detection requires additional preapartive steps, such as exosome isolation, extraction from exosomes, and subsequent amplification. Many exosome surface proteins have also been reported as potential cancer biomarkers, covering a wide range of major cancer types5. Exosome surface proteins like LAT1 are advantageous, especially when considering the possibility of their direct detection in liquid biopsy. For example, a previous study reported that sandwich ELISA with anti-CD9 and anti-CD91 antibodies against patients’ sera showed a significantly higher sensitivity for detecting early-stage (stage-I, II) lung adenocarcinoma than existing biomarker CEA38. Similarly, the significance of exosomal developmental endothelial locus-1 (Del-1) for identifying early-stage breast cancer was demonstrated by ELISA using patients’ sera, achieving distinction from benign breast tumors and noncancerous diseases39. Considerable progress has been also made in the development of sensitive and rapid methodologies other than ELISA, for detecting exosome surface protein in lipid biopsy40–42. To the best of our knowledge, the present study provides the first evidence for the presence of LAT1 on cancer cell-derived exosomes prepared from the body fluid. However, the ELISA method established in this study still requires improvement for future clinical applications because cancer cell-derived exosomes had to be collected by ultracentrifugation from the peritoneal wash of ascites-developing mice constructed with T3M-4 cells, which showed a relatively high LAT1 expression on exosomes among the tested cell lines (Figs. 2 and 3). Accumulation of ascites often represents the advanced phase of cancers in or around the abdominal cavity43. Especially for early cancer diagnosis, ideal sources of exosomes are serum and plasma, which would contain lower amounts of exosomes than ascites. More sensitive detection methods are necessary to further manifest the potential of exosomal LAT1 as a biomarker for cancer. It is also noteworthy that we needed to disrupt the isolated exosomes by sonication for the detection by ELISA because the anti-LAT1 antibody used in this study was raised against the C-terminus of LAT1, which is located inside the exosomes7. This preparative step reduces the assay throughput and, furthermore, may result in large fluctuations between samples and measurements. Therefore, antibodies that recognize the extracellular regions of LAT1 are more suitable for detecting the exosomal LAT1. According to the three-dimensional structure recently revealed by cryo-EM, the extracellular part of LAT1 is largely covered with the ectodomain of 4F2hc, resulting in a limited exposure of the molecular surface of LAT144,45. We demonstrated that LAT1 on exosomes also exists mainly as heterodimers with 4F2hc (Fig. 1d). Therefore, a rational approach would be to use the LAT1-4Fhc heterodimer as the antigen rather than LAT1 alone, aiming to generate antibodies against the extracellularly exposed surface of LAT1. A previously reported rat monoclonal antibody against LAT1 might represent such antibodies, as the authors directly immunized animals with HeLa cells46. While the abovementioned technical challenges remain to be addressed, the potential of exosomal LAT1 for early cancer detection warrants further investigation. Previous immunohistochemical studies have reported the elevated expression of LAT1 in early-stage diseases of various cancer types, including lung, gastric, breast, pancreatic, biliary tract, and liver cancers14,16,47–51. Among them, the high LAT1 expression was even associated with poor prognosis of patients with stage I lung squamous cell carcinoma and pulmonary adenocarcinoma47,48 and stage IB-II non-scirrhous gastric cancer49. In addition, KRAS mutations have been reported to be involved in the elevated LAT1 expression that supports the proliferation of colorectal cancer cells in vitro52. In the small intestine and colon of an APC-deficient mouse model, a KRAS-activating mutation of G12D also induces the LAT1 expression and epithelial cell proliferation52. KRAS, one of the most common cancer-causing genes, is known as a major early driver gene in several cancer types, such as colorectal and pancreatic cancers, where its mutations are found in more than 40% and 90% of cases, respectively53. It is thus tempting to speculate that the upregulation of LAT1 occurs in the early phase of tumorigenesis and offers a possibility of detecting these malignant cancers at treatable stages. In summary, we demonstrated that cancer cell-derived exosomes contain LAT1 as a cargo molecule, highlighting its potential as a diagnostic and prognostic biomarker for cancer. It would also be interesting to explore whether LAT1 on exosomes is involved in any pathological processes of cancers in future studies. The origins of cell lines used in this study include most malignant cancer types, such as pancreatic and biliary tract cancers. One of the reasons for the poor prognosis of these cancers is their asymptomatic characteristics at early stages. Further studies for developing methodologies for detecting exosomal LAT1 in liquid biopsy may pave the way to early detection of these formidable cancers, allowing patients to access more effective therapeutic options and improve their prognosis. Methods Cell culture The following human cancer cell lines were purchased and used in this study: T3M-4 (RCB1021, RIKEN BRC Cell Bank), HuCCA-1 (JCRB1657, JCRB Cell Bank), SUIT-2 (JCRB1094, JCRB Cell Bank), CFPAC-1 (CRL-1918, ATCC Cell Bank), HuCCT1 (JCRB0425, JCRB Cell Bank), OCUG-1 (JCRB0191, JCRB Cell Bank), and HeLa (CCL-2, ATCC Cell Bank). All the cell lines were routinely maintained in the complete RPMI-1640 medium (Nacalai Tesque), which is supplemented with 10% (v/v) heat-inactivated FBS (Nichirei) and 100U/mL penicillin − 100 µg/mL streptomycin (Nacalai Tesque). Cells were grown in a humidified incubator at 37 °C with 5% CO2. Exosomes-depleted culture medium Culture medium depleted of FBS-derived exosomes was prepared as previously described54 with some modifications. RPMI-1640 medium containing 20% (v/v) FBS, and 200U/mL penicillin − 200 µg/mL streptomycin was ultracentrifuged at 100,000 × g at 4 °C. The supernatant was sterilized through a 0.22-µm PES bottle top vacuum filter (Corning) and diluted with RPMI-1640 medium at 1:1 volume ratio to give the final concentration of 10% (v/v) FBS and 100U/mL penicillin − 100 µg/mL streptomycin. Exosome isolation and purification from cell culture supernatants Cells were seeded in 10 cm dishes (at 2.0 × 106 cells/dish) and cultured for 24 h in the complete RPMI-1640 medium (10 mL/dish). After replacing the medium to the exosomes-depleted medium, cells were further cultured for 48 h. Isolation and purification of exosomes from cell culture supernatant were performed by ultracentrifugation method as follows54. The cell culture supernatants were subjected to sequential centrifugation steps at 300 × g for 10 min, 2,000 × g for 10 min, and 10,000 × g for 30 min at 4 °C. The supernatants were then ultracentrifuged at 100,000 × g for 70 min at 4 °C. The obtained pellets were resuspended in PBS and ultra-centrifuged again at 100,000 × g for 70 min at 4 °C to collect the final exosome pellets. Nanoparticle tracking analysis Measurements of particle size distribution and concentration were performed with a Nanosight LM10 unit (NanoSight) equipped with an sCMOS camera. Exosome pellets from cell culture supernatants and mouse peritoneal wash were suspended in PBS to give the optimal concentration of 108-109 particles/mL. Five videos of 60 s at camera level 13 were recorded from each sample. Data analysis was performed with NTA software version 2.3 (Detection Threshold: 8 Multi). The Stokes–Einstein equation was employed to determine the hydrodynamic diameters of particles. Western blotting Exosome pellets from cell culture supernatants were lysed by sonication in a lysis buffer (50 mM Tris-HCl [pH 7.4], 0.15 M NaCl, 1 mM EDTA, 1 mM PMSF, protease inhibitor cocktail [cOmplete EDTA-free, Roche Diagnostics], and 10% [v/v] glycerol) and mixed with Laemmli buffer. For the preparation of crude membrane fractions, cells were lysed in the same lysis buffer by sonication. After removing unbroken cells and debris by centrifugation at 1,000 × g for 5 min at 4 °C, the supernatants were further ultracentrifuged at 350,000 × g for 30 min at 4 °C. The obtained pellets were resuspended with the lysis buffer and mixed with Laemmli buffer. SDS-polyacrylamide gel electrophoresis and western blotting were performed as described previously55. The loading amount of protein was adjusted by MicroBCA assay. Primary antibodies and dilution factors used are as follows: anti-CD81 (SHI-EXO-M03, Cosmo Bio, 1:1,500), anti-CD9 (sc-13118, Santa Cruz Biotechnology, 1:200), anti-GM130 (G7295, Sigma-Aldrich, 1:300), anti-CD98 (4F2hc) (sc-9160, Santa Cruz Biotechnology, 1:1,000), and anti-calnexin (sc-11397, Santa Cruz Biotechnology, 1:1,000). A polyclonal rabbit anti-LAT1 antibody against the C-terminal peptide of human LAT1 (CQKLMQVVPQET) was synthesized previously7. ImageJ software (National Institute of Health) was used for densitometric quantification of the detected immunoreactive bands. Uncropped blot images are presented in Supplementary Fig. S2 and S3. Immunoprecipitation Exosome fraction from the cell culture supernatant of T3M-4 cells was prepared by ultracentrifugation and subjected to immunoprecipitation using Dynabeads Exosome Human CD81 Isolation Reagent and Human CD9 Isolation Reagent (Veritas) according to the manufacturer’s instructions. For negative controls, the beads were pre-treated with Laemmli buffer for 30 min at 37℃ to denature beads-conjugated anti-CD9 and CD81 antibodies. Immunoprecipitated exosomes were lysed in Laemmli buffer and analyzed by western blotting for LAT1, CD9, and CD81 as described above. Immuno-transmission electron microscopy Exosomes from cell culture supernatant of T3M-4 cells were rapid-frozen in liquid propane (-175℃), followed by substitution with 2% tannic acid in ethanol for 48 h at -80℃, 3 h at -20℃, and 1 h at 4℃. Then the samples were dehydrated with ethanol at 4 ℃ for 30 min three times, infiltrated with 1:1 volume ratio of ethanol and resin (LR white resin, London Resin Co. Ltd.) for 30 min once and with the resin for 30 min three times, and embedded at 50℃ overnight. 70 nm-thick sections were made by ultramicrotome (Ultracut UCT, Leica) and placed on nickel grids. Immunostaining was performed by incubation with anti-LAT1 antibody at 4℃ overnight, followed by 5-nm gold-conjugated anti-rabbit goat IgG (EMGAR5, BBI Solutions) at room temperature for 2 h. After fixation in 2% glutaraldehyde at room temperature for 3 min, the sections were negatively stained using 2% uranyl acetate for 5 min and Lead stain solution (Sigma-Aldrich) for 2 min. The samples were observed by a transmission electron microscope JEM-1400Plus (JEOL Ltd.) at an acceleration voltage of 100 kV. Digital images were captured by a CCD camera (EM-14830RUBY2, JEOL Ltd.). Exosome ELISA Exosome pellets from cell culture supernatants and mouse peritoneal wash were suspended in the lysis buffer and lysed by sonication. LAT1 on exosomes was detected using PS Capture Exosome ELISA Kit (Streptavidin HRP) (Wako) according to the manufacturer’s instructions. Anti-LAT1 antibody was biotinylated by Biotin Labeling kit – NH2 (Dojindo Molecular Technologies Comapny) and used as the detection antibody. Mouse xenograft model T3M-4 cells suspended in PBS were intraperitoneally injected into 4- to 6-week-old BALB/c-nu/nu male mice (1.0 × 107 cells, 0.5 mL/animal). Control mice were injected with 0.5 ml of PBS alone. After one month, 5 mL of ice-cold PBS was injected into the peritoneal cavity of mice anesthetized with intraperitoneal injection of a three-agent mixture anesthetics (midazolam 4.0 mg/kg, medetomidine 0.3 mg/kg, and vetorphale 5.0 mg/kg in saline). The peritoneal wash were collected with a syringe equipped with a 25-gauge needle, immediately frozen in liqN2, and stored at -80 °C. Tumors were excised from sacrificed mice, fixed by 4% (w/v) PFA in PBS at 4 °C overnight, and used for preparation of paraffin-embedded tissue blocks. All the animal experiments in this study complied with the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines, and were conducted in accordance with the protocols approved by the Osaka University Medical School Animal Care and Use Committee. Exosome isolation and purification from mouse peritoneal wash Purification of exosomes from peritoneal wash was performed as described previously54. Briefly, the 5 mL of peritoneal wash were thawed on ice, diluted by adding 5 ml of PBS, and sequentially centrifuged at 2,000 × g for 30 min and 12,000 × g for 45 min at 4 °C. The supernatants were ultra-centrifuged at 110,000 × g for 2 h at 4 °C. The obtained pellets were resuspended in PBS, filtered through a 0.22-µm filter (SLGP033RB, Merck-Millipore), and ultracentrifuged at 110,000 × g for 70 min at 4 °C. Then, the pellets were resuspended in PBS and ultracentrifuged again at 110,000 × g for 70 min at 4 °C to collect the final exosome pellets. Histological analysis and immunohistochemistry Histological analysis and immunohistochemistry with anti-LAT1 antibody were performed as described previously55. Bright field images of 2048 × 2048 pixel resolution with 8-bit intensity depth were acquired with BZ-9000 (KEYENCE) equipped with CFI Plan-Apochromat λ 20 × /NA 0.75 objective lens. Statistical analysis GraphPad Prism version 9.0 (GraphPad Software) was used for calculation. Two-tailed unpaired Student’s t-test was applied to compare the LAT1 in exosomes between control mice and T3M-4 tumor bearing mice. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Authors would like to thank Ms. M. Kurauchi and Y. Hayamizu for their technical assistance. This study was technically supported by the Center of Medical Innovation and Translational Research, Graduate School of Medicine, Osaka University. Author contributions Y.L.: Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft. R.O.: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, review, and editing. M.X.: Funding acquisition, Writing – review and editing. H.O.: Funding acquisition, Writing – review and editing. Y.K.: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review and editing. Funding This research was supported by Pancreas Research Fundation of Japan (to R.O.), JSPS Grants-in-Aid for Scientific Research ([22H02809] to Y.K.), and the Project for Cancer Research and Therapeutic Evolution from AMED [JP20cm0106151], [JP21cm0106151], and [JP24ama221121] to Y.K.) Data availability All the data analyzed and presented in this study are available from the corresponding authors upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Ryuichi Ohgaki, Email: ohgaki@pharma1.med.osaka-u.ac.jp. Yoshikatsu Kanai, Email: ykanai@pharma1.med.osaka-u.ac.jp. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement All the data analyzed and presented in this study are available from the corresponding authors upon reasonable request.