Simultaneous Inhibition of MEK and Hh Signaling Reduces Pancreatic Cancer Metastasis

Abstract

Pancreatic cancer, mostly pancreatic ductal adenocarcinoma (PDAC), is one of the most lethal cancer types, with an estimated 44,330 death in 2018 in the US alone. While targeted therapies and immune checkpoint inhibitors have significantly improved treatment options for patients with lung cancer and renal cell carcinomas, little progress has been made in pancreatic cancer, with a dismal 5-year survival rate currently at ~8%. Upon diagnosis, the majority of pancreatic cancer cases (~80%) are already metastatic. Thus, identifying ways to reduce pancreatic cancer metastasis is an unmet medical need. Furthermore, pancreatic cancer is notorious resistant to chemotherapy. While Kirsten RAt Sarcoma virus oncogene (K-RAS) mutation is the major driver for pancreatic cancer, specific inhibition of RAS signaling has been very challenging, and combination therapy is thought to be promising. In this study, we report that combination of hedgehog (Hh) and Mitogen-activated Protein/Extracellular Signal-regulated Kinase Kinase (MEK) signaling inhibitors reduces pancreatic cancer metastasis in mouse models. In mouse models of pancreatic cancer metastasis using human pancreatic cancer cells, we found that Hh target gene Gli1 is up-regulated during pancreatic cancer metastasis. Specific inhibition of smoothened signaling significantly altered the gene expression profile of the tumor microenvironment but had no significant effects on cancer metastasis. By combining Hh signaling inhibitor BMS833923 with RAS downstream MEK signaling inhibitor AZD6244, we observed reduced number of metastatic nodules in several mouse models for pancreatic cancer metastasis. These two inhibitors also decreased cell proliferation significantly and reduced CD45⁺ cells (particularly Ly6G⁺CD11b⁺ cells). We demonstrated that depleting Ly6G⁺ CD11b⁺ cells is sufficient to reduce cancer cell proliferation and the number of metastatic nodules. In vitro, Ly6G⁺ CD11b⁺ cells can stimulate cancer cell proliferation, and this effect is sensitive to MEK and Hh inhibition. Our studies may help design novel therapeutic strategies to mitigate pancreatic cancer metastasis.

Keywords: Ihh; MEK; hedgehog; metastatic niche; pancreatic cancer.

Figure 1

Activated hedgehog signaling in orthotopic mouse models of pancreatic cancer metastasis. (A) Expression of mouse Gli1 gene, an indicator for hedgehog signaling activity, in tumor-bearing pancreatic tissues, metastatic lymph nodes and liver, was detected by Taqman-based real-time PCR analysis. ** p < 0.005 by Student t-test of two independent means (n = 5/group). (B) The transcript levels of SHH and IHH in different tissues were detected by real-time PCR (Taqman-based) ** p < 0.005 by Student t-test of two independent means (n = 5/group). (C) Western blotting analysis of IHH protein levels in tumors of pancreas, lymph nodes and liver. Metastatic nodules were dissected (based on GFP-expression) from liver tissues for protein analysis to reduce the Ihh protein level from the tumor microenvironment. We performed protein analysis from three independent mice with liver metastases and three normal liver tissues with similar results. (D) The effect of smoothened inhibitor BMS833923 on liver gene expression patterns in MIA PaCa2-based orthotopic mouse model was shown following RNA sequencing analyses. While metastatic liver tissues have a very different gene expression pattern from naïve liver tissues, addition of BMS833923 every the other day for 3 weeks altered the gene expression pattern to reassemble that of naïve liver tissues. (E) The effect of smoothened inhibitor BMS833923 on pancreatic cancer growth was revealed by weekly with bioluminescent intensity from the cancer cells as well as the tumor weight difference after the mice were sacrificed. Five mice per group were used in this experiment (with power 95% alpha error 0.05). (F) Whole body imaging detection of GFP expression pancreatic cancer cells in pancreas, lymph nodes, lung and liver in MIA Paca2-based orthotopic mouse model following administration of BMS833923 or with vehicle control. White arrows indicate the site of pancreatic cancer cells (GFP positive). Scale bar: centimeter. Similar results were also observed in AsPC1-based orthotopic mouse model (Figure S4 and S5).

Figure 2

Simultaneous inhibition of MEK and Hh signaling in mouse models. (A) The number of metastatic nodules in the liver in MIA PaCa2-based orthotopic model in immune deficient NSG mice from four different treatments: (1) treated with smoothened inhibitor BMS833023; (2) treated with MEK inhibitor AZD6244; (3) the combination of AZD6244 and BMS833923; and (4) vehicle control (5) mice/group); (B) Effect of different treatments on the number of metastatic nodules in AsPC1-based mouse model (5 mice/group); (C) Kaplan-Meier survival curve for mice with different treatments in AsPC1-based mouse model, ** p < 0.05 was derived from Log-rank (Mantel-Cox) test; (D) Effect of different treatments on liver tissue weight, a measurement for liver metastasis, in MMC18-based model in immune-competent C57/B6 mice (5 mice/group). * p < 0.05 and ** p < 0.005 indicate significant or synergistic effect by BLISS independence analysis (see Statistical analyses for details).

Figure 3

Effects of simultaneous inhibition of MEK and Hh signaling on cell proliferation. (A) Effect of MEK/Hh signaling inhibition on Ki-67 positivity in metastatic liver tissues of MIA PaCa2-based model in immune-deficient NSG mice; (B) The rate of Ki-67 positivity in pancreatic cancer of MIA PaCa2-based model after different treatments; (C) Percentage of EdU labeling after different treatments in AsPC1-based model in immune deficient NSG mice; (D) Effect of different treatments on Ki-67 positivity in Pan02-based model in immune competent C57/B6 mice. The percentage of Ki67 positivity was derived from the average positivity of at least three independent mice, and the average was obtained from five fields of immunofluorescently stained slides. * indicates p < 0.05 and ** indicates a synergistic (more than additive) effect via BLISS independence analysis.

Figure 4

Regulation of CD45⁺ cell number by both MEK and Hh signaling pathways. (A) Representative immunofluorescent staining images of CD45⁺ cells in liver tissues after drug treatments. (B) Summary of A from 10 images in each group (from at least 3 mice) with the number of CD45⁺ cells per field under microscope (200×). (C) Summary of flow cytometry analysis of CD45+ cells in different treatment groups (from three mice/group). * indicates p < 0.05 and ** indicates a more than additive effect via BLISS independence analysis.

Figure 5

Effects of MEK/Hh signaling inhibition on Ly6G⁺CD11b⁺ cells in the metastatic liver tissues (MMC18-based model in immune competent C57/B6 mice). (A) Representative images of flow cytometry of CD45⁺CD11b⁺Gr1⁺ cells in different liver tissues; (B) Cell population composition in metastatic liver tissues, and the number was the average from five mice/group; (C) Representative images of liver tissues with or without treatment of Ly6G neutralizing antibodies. While IgG treated mice had many metastatic nodules (indicated by purple arrowheads), Ly6G neutralizing antibody treatment led to no visible or only a few metastatic nodules; (D) Ly6G neutralizing antibody treatment led to reduced liver weight in MMC18- based mouse model (five mice/group), suggesting a reduced liver metastasis; (E) Similar results were obtained using Pan02 cells (five mice/group). Ly6G neutralizing antibody treatment led to reduced liver weight. * p values were calculated using Student t test of two independent means.

Figure 6

Regulation of cancer cell proliferation by inhibition of MEK and Hh signaling in mouse models and in cultured cells. (A) Expression of Ki-67 was used as the readout of cell proliferation, and percentage of Ki-67 positive cells in the metastatic cancer cells was calculated under microscope from MMC18-based model in immune competent C57/B6 mice (five mice/group), * p value < 0.05. These results are similar to those from MIA PaCa2-based model in immune deficient NSG mice (see Figure 3A); (B) EdU incorporation of cancer cells in the presence or absence of Ly6G⁺CD11b⁺ cells in vitro. EdU incorporation of cancer cells was used to measure cell proliferation rate. With an increase in Ly6G⁺CD11b⁺ cells (ratio 0:1; 1:1; 1:2 with increased number of Ly6G⁺CD11b⁺ cells), EdU incorporation in the cancer cells was visualized (representative images shown above; summary from 3 independent experiments shown below), scale: 400×; (C) Effects of different compounds in reducing cell proliferation of cancer cells were assessed in vitro. Ly6G⁺CD11b⁺ cells were first treated with different compounds for 2 h before adding to the top chamber. The data were from three independent experiments. * indicates significant change (p < 0.05). While Ly6G⁺CD11b⁺ cells significantly stimulated cell proliferation of cancer cells (* indicates p < 0.005), treatment with MEK inhibitor AZD6244 (or Hh signaling inhibitor BMS833923) for 2 hours before being added onto the top chamber significantly reduced cancer cell proliferation (p < 0.05). Combination of AZD6244 with BMS833923 had a more than additive effect on cell proliferation (** indicates synergistic effect via BLISS independent analysis).