Expression of IGF2BP2 is positively correlated with glutamine metabolism-associated gene SLC1A5 in pancreatic cancer and predicts poor prognosis of pancreatic cancer patients

Reprogrammed energy metabolism is an important feature of tumors, and one common metabolic change in tumors is an increase in glutamine metabolism. Glutamine, as the most abundant circulating amino acid in blood, plays a unique nutritional role in the growth and proliferation of tumor cells [9, 25, 26]. First, to define the level of glutamine metabolism in pancreatic cancer, we acquired a list of 41 glutamine metabolism-associated genes (Supplementary Table 1), and computed a score for the glutamine metabolism signature (Gln signature) by calculating the levels of The Cancer Genome Atlas (TCGA) PDAC database. The tumors were classified as Gln signature-high or Gln signature-low, and survival analysis showed that the Gln signature was significantly associated with poor OS (Fig. 1A, Supplementary Fig. 1A).

It was further illustrated by the ROC curve and area under the curve (AUC at 3 years = 0.690, AUC at 5 years = 0.910) that the score is a good predictor of prognosis in pancreatic cancer patients (Fig. 1B). These results suggested that compared to pancreatic cancer patients with less active tumor glutamine metabolism, patients with highly active tumor glutamine metabolism had poorer survival.

RNA m6A modification is essential for tumor metabolism. Then, we tried to investigate whether the pancreatic cancer intrinsic glutamine metabolism has relationship between m6A modification-associated genes (Supplementary Table 1). We conducted a correlation analysis between m6A modified genes and score, and found that IGF2BP2 had the most significant correlation (Supplementary Fig. 1B). Further analysis revealed that IGF2BP2 is closely related to the glutamine metabolism-associated genes SLC1A5 (Fig. 1C). Data mining of GEPIA database also showed a positive correlation between expression of IGF2BP2 and SLC1A5 (Fig. 1D), further providing evidence that IGF2BP2 is closely related to SLC1A5. These results supported the notion that IGF2BP2 may be involved in mediating the uptake of glutamine by tumor cells.

We employed bioinformatics-based methods to investigate the role of IGF2BP2 and SLC1A5 in human pancreatic cancer. The expression of IGF2BP2 and SLC1A5 were upregulated in pancreatic cancer tissues compared with that in normal tissues based on the GEPIA bioinformatics tool (Fig. 1E). Then, both GEPIA database and the bioinformatics tool Kaplan-Meier Plotter showed that patients with high IGF2BP2 or SLC1A5 expression had poorer overall survival (OS) than those with low IGF2BP2 or SLC1A5 expression (Fig. 1F, Supplementary Fig. 1C). Moreover, the bioinformatics tool GEPIA validated patients with pancreatic cancer exhibiting increased IGF2BP2 or SLC1A5 mRNA levels had worse DFS (Supplementary Fig. 1D).

Then, in order to further uncover the upregulation of IGF2BP2 and SLC1A5 expression in pancreatic cancer, we examined the expression of IGF2BP2 and SLC1A5 in human pancreatic cancer tissues. Consistent with the above research results, specific immunoreactivity of IGF2BP2 and SLC1A5 observed during malignant transformation and was found significantly increased in from PanIN lesions to pancreatic cancer (Fig. 1G and H). In addition, the correlation analysis of the immunohistochemical scores reconfirmed the positive correlation between expression of IGF2BP2 and SLC1A5 in pancreatic cancer tissues (Fig. 1I). These observations were further confirmed in the widely available KPC model. Compared with normal pancreatic ductal cells, IGF2BP2 and SLC1A5 are increased in cancerous ductal cells (Fig. 1J). Meanwhile, the expression of IGF2BP2 and SLC1A5 was commonly increased in pancreatic cancer cell lines compared with the nonmalignant hTERT-HPNE (hereinafter referred to as HPNE) cells (Fig. 1K, Supplementary Fig. 1E). In addition, CCK8 experiments further showed that glutamine deficiency significantly inhibited the growth of pancreatic cancer cells (Fig. 1L). Collectively, these findings indicate that IGF2BP2 was closely correlated with the expression of glutamine metabolism gene SLC1A5, which is up-regulated in pancreatic cancer, and its upregulation predicted poor prognosis of patients with pancreatic cancer.

Fig. 1

IGF2BP2 is positively correlated with SLC1A5 and is associated with poor prognosis of pancreatic cancer. A. Kaplan-Meier survival analysis based on glutamine score in TCGA-PDAC cohort (P < 0.0001, log-rank test). B. ROC curve and AUC of glutamine level in patients with pancreatic cancer. (ROC = receiver operator characteristic. AUC = area under curve). C. Correlation between mRNA expression of IGF2BP2 and glutamine metabolism-associated genes in the TCGA database. D. Correlation between mRNA expression of IGF2BP2 and SLC1A5 in the GEPIA database. E. GEPIA database showing that the expression of IGF2BP2 and SLC1A5 in PANCREATIC CANCER tissues (T) and normal pancreas tissues (N). F. GEPIA database showing that overall survival of pancreatic cancer patients with diverse IGF2BP2 and SLC1A5 expression. G. Immunohistochemistry staining of IGF2BP2 and SLC1A5 in human pancreatic cancer. Scale bar, 2000 μm 20 μm. H. Immunohistochemical score statistical analysis. I. Correlation between the expression of IGF2BP2 and SLC1A5 based on immune score. J. Immunohistochemistry staining for KPC mice pancreatic tissue. Scale bar, 50 μm. K. The protein expression of IGF2BP2 and SLC1A5 were measured by western blot in the nonmalignant hTERT-HPNE (HPNE) cells and pancreatic cancer cell lines (PaTu8988, Mia-PaCa2, SW1990, PANC-1, CFPAC, Aspc-1). L. Viability of HPNE cells, PANC-1 cells and PaTu 8988 cells with or without glutamine analyzed by the CCK8 assay. *P < 0.05, ****P < 0.0001

IGF2BP2 promotes tumor proliferation in vitro and in vivo by regulating glutamine metabolism

To evaluate the biological role of IGF2BP2 in pancreatic cancer, we generated IGF2BP2-knockout (KO) cells by utilizing the CRISPR-Cas9 system and determined the protein levels of IGF2BP2 in IGF2BP2-KO PANC-1cells and IGF2BP2-KO PaTu 8988 cells (Fig. 2A, Supplementary Fig. 2A). As shown by CCK8 and Colony formation tests, IGF2BP2 knockout remarkably suppressed the proliferation and colony formation of PANC-1 and PaTu 8988 cells (Fig. 2B and C). To further validate whether IGF2BP2 promoted tumor progression in vivo, we subcutaneously injected PANC-1cells into nude mice. The growth of IGF2BP2-KO transfected subcutaneous tumors was hindered compared to that of the controls (PANC-1 cells) (Fig. 2D). Meanwhile, the average tumor volume and weight at sacrifice were markedly decreased in mice with IGF2BP2-KO compared with the control mice (Fig. 2E and F). Further, significantly decreased proliferation was observed for the IGF2BP2-KO transfected subcutaneous tumors, as evidenced by the fact that Panc-1-IGF2BP2-KO cell-derived xenografts exhibited lower levels of Ki67 proteins (Fig. 2G). We further found by CCK8 and ELISA assay that deletion of IGF2BP2 leads to a decrease in the level of glutamine uptake by pancreatic cancer cells (Fig. 2H and I), which affects the proliferative capacity of tumor cells. These observations suggest that IGF2BP2 promotes the growth of pancreatic cancer in vitro and in vivo by regulating glutamine metabolism.

Fig. 2

IGF2BP2 promotes proliferation and glutamine uptake in pancreatic cancer. A. Protein levels of IGF2BP2 in pancreatic cancer cell lines determined by western blot. B. Viability of PANC-1 cells and PaTu 8988 cells with or without IGF2BP2 knockout analyzed by the CCK8 assay. C. Representative images from the colony-forming assay and colony number analysis as indicated. D. Representative xenograft tumors after subcutaneous injection of PANC-1 cells transfected with KO-IGF2BP2 and KO-NC 42 days after inoculation. E. Tumor volume (n = 5) was recorded every week. F. Tumor weight (n = 5) of PANC-1 cells xenografts. G. Proliferation (Ki67) immunohistochemistry (IHC) staining of tumor sections. Scale bar, 50 μm. H. Viability of PANC-1 cells and PaTu 8988 cells with or without IGF2BP2 knockout analyzed by the CCK8 assay in the presence and absence of glutamine. I. Detection of intracellular glutamine levels through ELISA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns: not statistically significant

IGF2BP2 maintains SLC1A5 mRNA stability in an m6A-dependent manner

Recently, IGF2BP2 functions as an m6A reader, affecting mRNAs stability and translation of m6A-modified mRNAs [27]. To validate SLC1A5 as a potential target of IGF2BP2, IGF2BP2 was knocked out in pancreatic cancer cells, and our results showed that both protein and mRNA expression levels of SLC1A5 significantly inhibited by IGF2BP2 knockout (Fig. 3A and B, Supplementary Fig. 2B). Consistent with the decrease of protein and mRNA level, a shortening of mRNA half-life of SLC1A5 was observed upon knockout of IGF2BP2 (Fig. 3C).

To comprehensively profile genes with m6A modification mediated by IGF2BP2, we performed MeRIP-seq in PANC-1 cells with or without IGF2BP2 knockout. MeRIP-seq showed that PANC-1 cells had 11,499 unique peaks, while knockout of IGF2BP2 had 686 unique peaks (Fig. 3D). In addition, by analyzing the sequencing data, we found that GGAC was the most enriched motif found in m6A peaks identified from both IGF2BP2-KO cells and control PANC-1 cells, consistent with previous reports [28] (Fig. 3E). Moreover, from our MeRIP-seq data, we confirmed that m6A abundance in the SLC1A5 mRNA was diminished in response to IGF2BP2 knockout (Fig. 3F). These data together implicate an m6A-dependent regulatory mechanism.

Additionally, high-confidence m6A modification sites were predicted based on the m6A-seq data as well as the prediction from SRAMP (http://www.cuilab.cn/sramp) to determine the mechanism of m6A modification of SLC1A5 in pancreatic cancer (Fig. 3G), after which we designed primers at the very high-confidence m6A modification site (SLC1A5 # 1) and non m6A modification site (SLC1A5 # 2) respectively (Fig. 3H), and utilized Methylated RNA immunoprecipitation (MeRIP) to evaluate relative m6A abundance changes. MeRIP assays showed that upon knockout of IGF2BP2, the m6A abundance of SLC1A5 was significantly reduced at the modification sites, whereas the non-m6A modification sites were essentially unchanged (Fig. 3I). Moreover, the interaction between IGF2BP2 and SLC1A5 mRNA in both PANC-1 and PaTu 8988 cells was confirmed using the IGF2BP2-specifific antibody in RNA immunoprecipitation (RIP) assays (Fig. 3J). Collectively, these data establish that IGF2BP2 regulates SLC1A5 expression via m6A-dependent manner.

Fig. 3

IGF2BP2 regulates SLC1A5 expression in an m6A-dependent manner. A. Protein levels of SLC1A5 in pancreatic cancer cell lines after IGF2BP2 knockout determined by western blot. B. mRNA levels of SLC1A5 in pancreatic cancer cell lines after IGF2BP2 knockout determined by RT-PCR. C. mRNA half-life (t1/2) of SLC1A5 in PANC-1 cells and PaTu 8988 cells with or without IGF2BP2 knockout. D. MeRIP-seq of PANC-1 cell lines showed a number of m6A peaks. E. MeRIP-seq analysis identified m6A motif. F. IGV visualization based on MeRIP-seq reveals significant m6A modification sites in the mRNA of SLC1A5. G. SRAMP online software predicted the m6A modification site of the mRNA of SLC1A5. H. Scheme showing the design of primers for MeRIP-qPCR to validate m6A modifications on SLC1A5 mRNA. Primer#1 is a potential m6A site, while primer#2 is a non m6A site. I. MeRIP-qPCR validated m6A modification with primers #1 and #2. J. The interaction between IGF2BP2 and the mRNA of SLC1A5 was detected by the RIP assay. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

SLC1A5 promotes pancreatic cancer proliferation by activating the mTORC1 signaling pathway

SLC1A5, an amino acid transporter, is frequently overexpressed in tumors and enhances glutamine metabolism [29]. Loss-of-function assays was used to assess the role of SLC1A5, a target gene of IGF2BP2, in pancreatic cancer. We knocked down the expression of SLC1A5 in pancreatic cancer cells and verified the knockdown efficiency by RT-PCR analysis (Fig. 4A). The CCK8 assay showed that depletion of SLC1A5 significantly suppressed the proliferation of pancreatic cancer cells compared to controls (Fig. 4B). Furthermore, the colony-forming ability of pancreatic cancer cells was remarkably inhibited when SLC1A5was knocked down (Fig. 4C). Together, the above data mimicked the effect of IGF2BP2 knockout.

In recent years, it has been found that amino acids are essential for mammalian target of rapamycin complex 1 (mTORC1) activation and that is particularly sensitive to decreases in glutamine [30,31,32], so we conjecture that SLC1A5 may promote the proliferation of pancreatic cancer cells by activating the mTORC1 signaling pathway. Therefore, we next dissected the role of SLC1A5 on mTORC1 signaling in pancreatic cancer. Our experimental results showed that knockdown of SLC1A5 inhibited mTORC1 pathway activation as evidenced by reduced phosphorylation of mTOR, ribosomal protein S6 kinase β-1 (S6K1) and eukaryotic translation initiation factor 4E binding protein 1 (4E-BP1) (Fig. 4D, Supplementary Fig. 2C). Additionally, the addition of glutamine reversed the results (Supplementary Fig. 3A). Furthermore, we found that knockdown of SLC1A5 may reduce glutamine uptake by pancreatic cancer cells (Fig. 4E). We next dissected the role of IGF2BP2 on mTORC1 signaling in and the results showed that knockout of IGF2BP2 inactivated the mTORC1 pathway (Supplementary Fig. 3B and D-E). Taken together, our data show that a unique molecular association between SLC1A5 and mTORC1 signaling in PANCREATIC CANCER.

Fig. 4

SLC1A5 promotes pancreatic cancer cell proliferation and glutamine uptake through activation of the mTORC1 pathway. A. mRNA levels of SLC1A5 in PANC-1 and PaTu 8988 cells determined by RT-PCR. B. Viability of PANC-1 cells and PaTu 8988 cells with or without SLC1A5 knockdown analyzed by the CCK8 assay. C. Representative images from the colony-forming assay and colony number analysis as indicated. D. Depletion of SLC1A5 inactivated mTORC1 signaling by western blot. E. Detection of intracellular glutamine levels through ELISA. **P < 0.01, ***P < 0.001, ****P < 0.0001

IGF2BP2 is a promising therapeutic target in pancreatic cancer

The above data together demonstrated that IGF2BP2, through regulating glutamine metabolism pathways, is critical for promoting the growth of pancreatic cancer cells. Considering the critical role of glutamine cleavage in pancreatic cancer, we tested whether IGF2BP2 could be an attractive target for pancreatic cancer therapy.

To investigate the effect of IGF2BP2 on cell radiosensitization, we performed CCK8 experiments and colony formation experiments, which showed that although KO-IGF2BP2 or 6GY irradiation individually inhibited cell proliferation, their combination showed a more significant inhibitory effect compared to either KO-IGF2BP2 or 6GY irradiation alone (Fig. 5A and B). Additionally, we established a xenograft mouse model to confirm that IGF2BP2 has the same effect in vivo. In accordance with the cell-based results, IGF2BP2-KO cells treated with 6GY irradiation markedly repressed tumor growth in mice, as reflected by the significant inhibition of the tumor volume and weight when compared to the control group (Fig. 5C and D).

Currently, gemcitabine is still the first-line drug for the treatment of pancreatic cancer, but its chemotherapy resistance creates significant resistance to the treatment. Therefore, enhancing sensitivity to gemcitabine may contribute to enhancing the effects of therapy and improving the prognosis of pancreatic cancer. We hypothesized that the combination of knockout IGF2BP2 and gemcitabine might achieve a better effect in pancreatic cancer treatment. To this end, PANC-1 and PaTu 8988 cells with or without IGF2BP2 knockout were treated with gemcitabine. As expected, combination treatment resulted in more significant inhibition on the growth of pancreatic cancer cells (Fig. 5E).

To determine whether IGF2BP2 has a role in mediating gemcitabine sensitivity in pancreatic cancer cells in vivo, we established a subcutaneous xenograft tumor model with PANC-1 cells with or without KO-IGF2BP2, followed by intraperitoneal administration of gemcitabine 2 times a week. Targeting IGF2BP2 or administration of gemcitabine alone modestly reduced tumor size and weight as compared to control. Additionally, combination treatment resulted in more significant suppression of PANC-1 xenograft tumor growth (Fig. 5F and G). Our data collectively show that knockout IGF2BP2 can not only enhance the radiosensitivity of pancreatic cancer, but also enhance the chemosensitivity of pancreatic cancer, providing a new direction for improving the therapeutic effect of pancreatic cancer.

Fig. 5

Knockout of IGF2BP2 enhances sensitivity of pancreatic cancer cells to radiotherapy and chemotherapy. A. Viability of PANC-1 cells and PaTu 8988 cells with or without IGF2BP2 knockout under 6 Gy radiation therapy and 0 Gy radiation therapy analyzed by the CCK8 assay. B. Representative images from the colony-forming assay and colony number analysis as indicated. C. The effects of 6 Gy radiation treatment on the growth of subcutaneous PANC-1 cells with or without IGF2BP2 knockout xenografts. D. Tumor weight (n = 5) of PANC-1 cells xenografts. E. PANC-1 and PaTu 8988 cells with or without IGF2BP2 knockout were treated with gemcitabine at different concentrations for 48 h, and cell viability was then measured by CCK8 assay. F. Effect of gemcitabine treatment on the growth of subcutaneous PANC-1 cells with or without IGF2BP2 knockout xenografts. G. Tumor weight (n = 5) of PANC-1 cells xenografts