Conditioned medium from Thp-1 macrophages stimulated by ADO enhances the malignant phenotypes of PDAC cells.
To explore the effect of ADOR-activation macrophages on the pancreatic cancer cell lines, we collected the supernatant of Thp-1 macrophages activated with ADO in a concentration gradient. Then, the supernatant was mixed with a fresh medium containing 10% FBS in a 2:1 proportion to prepare the conditioned medium (CM). Subsequently, the CM was used for culturing pancreatic cancer cell lines, and the related phenotypes were tested respectively. The results showed that the migration of pancreatic cancer cell lines PANC-1 and BxPC-3 was significantly increased compared with the control group (Fig. 1A-B). Meanwhile, the invasion of PANC-1 and BxPC-3 was enhanced considerably compared with the control group (Fig. 1C-D). Furthermore, the proliferation of PANC-1 and BxPC-3 was improved compared with the control group (Fig. 1E-F). At last, the western blot results exhibited that the expression of EMT and cell-cycle-related proteins in PANC-1 and BxPC-3 was also significantly elevated compared with the control group (Fig. 1G).
ADO promotes LIF secretion from Thp-1 macrophages.
Next-generation sequencing was performed to screen for potential cytokine secreted from the ADOR-activation macrophages that promote the malignant phenotype of pancreatic cancer cells. At first, the heat and volcano map exhibited 2560 up-regulated genes and 1307 down-regulated genes (Fig. 2A). Moreover, KEGG analysis for the sequencing data showed that the cytokine-cytokine receptor interaction signaling pathway was significantly enriched in ADOR-activation macrophages (Fig. 2B). Meanwhile, the GO analysis indicated that the cytokine activity of ADOR-activation macrophages was also markedly improved (Fig. 2C). Subsequently, cytokine-related genes in ADOR-activation macrophages were screened by taking the intersection from the two gene pools of KEGG and GO enrichment previously described.
Moreover, we removed those genes with primary copy numbers less than ten to ensure the practical value of cytokine used in subsequent studies. Then, the genes were sorted by the value of the fold change. (Fig. 2D). The ranking results indicated that IL23A, CXCL5, IL1A, LIF, and CCL20 were ADOR-activation macrophages' top five most highly expressed cytokine-related genes. However, only IL23A, CXCL5, LIF, and CCL20 exhibited the expression difference between cancer and cancer-adjacent tissue in PDAC (Fig. 2E). Furthermore, only the expression of LIF and IL1A was significantly correlated with the level of ADOR2AB in PDAC (Fig. 2F). Finally, based on the analysis results of the public database, LIF was the only one simultaneously satisfying the two criteria in Fig. 2E-F.
Activation of ADORA2B promotes LIF secretion from Thp-1 macrophages through the RAF-ERK-MEK signaling pathway.
Then, we explored the change of LIF secretion through ELISA assay in Thp-1 macrophages stimulated by ADO. It was revealed that the concentration of LIF in the supernatant of Thp-1 macrophages was significantly elevated after ADO stimulation (Fig. 3A). Meanwhile, the western blot results in Thp-1 macrophages showed that the LIF expression was significantly elevated after ADO stimulation (Fig. 3B). Based on the results of previous research, the four types of adenosine receptors on the surface of Thp-1 macrophages were demonstrated to be connected with the MAPK signaling pathway. Similarly, the KEGG enrichment of the present study also exhibited that the MAPK signaling pathway was remarkedly activated (Fig. 3C). Subsequently, we confirmed the activation of the RAF-MEK-ERK signaling pathway by western blot in Thp-1 macrophages stimulated by ADO (Fig. 3D).
Moreover, combined treatment of ADO, antagonists for ADOR, and RAF were performed on the Thp-1 macrophages. The western blot results indicated that ADO activates the RAF-MEK-ERK signaling pathway and elevates the expression of LIF in Thp-1 macrophages compared to the control group. Meanwhile, the use of antagonists for ADORA2B and RAF could reverse it. On the contrary, using the antagonists for ADORA1, ADORA2A, and ADORA3 had no significant effect on the RAF-MEK-ERK signaling pathway. It also did not influence the expression of LIF in ADOR-activation macrophages. (Fig. 3E). Finally, the results of ELISA also established that only antagonists for ADORA2B and RAF reduced the LLF concentration in the supernatant of Thp-1 macrophages after ADO stimulated. In contrast, the ADORA1, ADORA2A, and ADORA3 antagonists observed no similar inhibitory effect (Fig. 3F).
ADORA2B is overexpressed in PDAC tissues and correlates with poor prognosis.
Data from TCGA was analyzed to explore the oncogenic role of different types of AODR in PDAC. It has been indicated that the expression of ADORA2A, ADORA2B, and ADORA3 in tumor tissue was significantly higher than in normal tissue (Fig. 4A). Meanwhile, the expression ratio of ADORA3 was the highest among the four types of ADOR, followed by ADORA2B, ADORA2A, and ADORA1 (Fig. 4B). However, the TCGA data showed that only the expression levels of ADORA2A and ADORA2B significantly correlated with OS and PFS in PDAC patients (Fig. 4C). Next, IHC staining for ADORA2B was performed on 100 PDAC patients who received radical resection. Moreover, the staining tissue section was further analyzed through the HALO system for H-score (Fig. 4D). Then, a matched-pairs comparison of the ADORA2B H-score between normal and tumor tissue was conducted, exhibiting that ADORA2B expression in tumor tissue was significantly higher than in normal tissue (Fig. 4E). Next, the correlation between the expression level of ADORA2B and OS, PFS was analyzed, indicating that PDAC patients with higher ADORA2B expression level were more likely to have a poor OS and PFS (Fig. 4F).
Macrophages-derived LIF promotes malignant phenotypes of PDAC cells by activating the JAK-STAT signaling pathway.
To verify the promoting effect of LIF, rh-LIF with concentration gradient was used to treat pancreatic cancer cell lines, and the functional change was recorded. The results of phenotypic experiments exhibited that the migration capacity of pancreatic cancer cell lines PANC-1 and BxPC-3 was significantly increased compared with the control group (Fig. 5A-B). Moreover, the invasion of PANC-1 and BxPC-3 were considerably enhanced compared with the control group (Fig. 5C-D). Furthermore, the proliferation of PANC-1 and BxPC-3 was elevated after rh-LIF stimulation (Fig. 5E-F). Subsequently, the western blot for PANC-1 and BxPC-3 exhibited that EMT and cell cycle-related proteins were significantly up-regulated after rh-LIF stimulation (Fig. 4G). Last but not least, the western blot results also revealed that rh-LIF activated the JAK-STAT signaling pathway in PANC-1 and BxPC-3, while the addition of EC359 (antagonist for the LIF receptor) could revise it (Fig. 5H).
The ADORA2B and LIFR antagonists inhibit malignant phenotypes of PDAC cells in the coculture system.
Then, to verify the promoting effect of the ADO/ADORA2B/LIF axis previously described, a combined treatment of ADO, PSB603 (antagonists for ADORA2B), and EC359 (LIFR) was utilized in the co-cultured system consisting of Thp-1 macrophages and pancreatic cancer cells lines PANC-1 and BxPC-3 (Fig. 6A-B). As the results showed, the migration of PANC-1 and BxPC-3 was significantly increased after adding ADO, while this promoting effect could be inhibited by PSB603 and EC359 (Fig. 6C-D). Similarly, the invasion of PANC-1 and BxPC-3 was significantly strengthened after adding ADO, while it could be revised by PSB603 and EC359 (Fig. 6E-F). Moreover, the proliferation of PANC-1 and BxPC-3 has been elevated with the addition of ADO, while the promoting effect could be prevented by PSB603 and EC359 (Fig. 6G-H). Finally, the western blot indicated that the expression of EMT and cell cycle-related proteins in PANC-1 and BxPC-3 were significantly up-regulated after the addition of ADO. At the same time, it could be reversed by PSB603 and EC359 (Fig. 6I).
ADORA2B and LIFR antagonists inhibit the progression and metastasis of pancreatic orthotopic BxPC-3 tumors.
We further explored whether clodronate liposomes (macrophage scavenger), PSB603, and EC359 could inhibit the tumorigenicity of pancreatic orthotopic cancer cells in vivo. The schematic diagram shows the medication regimens (Fig. 7A). After eight weeks of treatment, the vivo bioluminescence imaging was performed and recorded (Fig. 7B). As the results showed, the fluorescence intensity of the tumor in the mice that only received ADO (group 2) was significantly more robust than the no-drug group (group 1). In contrast, the fluorescence intensity of the tumor in mice receiving ADO combined with clodronate liposomes, PSB603, and EC359 (group 3–5) was remarkedly lower than group 2 (Fig. 7B). Then, the mice were sacrificed, followed by the orthotopic tumor and liver destruction (Fig. 7C). Subsequently, the orthotopic tumors were weighed and rinsed in PBS. Meanwhile, the surface metastatic foci in the liver were counted manually. The analysis showed that tumor weight in group 2 was significantly higher than in group 1. Meanwhile, the tumor weight in groups 3–5 was lighter than in group 2 (Fig. 7D). Furthermore, more liver metastatic foci occurred in group 2 compared with group 1. Meanwhile, surface liver metastatic foci in groups 3–5 were notably fewer than in group 2 (Fig. 7E).