Transcriptional induction of NF-κB-inducing kinase by E2F4/5 facilitates collective invasion of GBM cells

Abstract

The prognosis of high-grade gliomas, such as glioblastoma multiforme (GBM), is extremely poor due to the highly invasive nature of these aggressive cancers. Previous work has demonstrated that TNF-weak like factor (TWEAK) induction of the noncanonical NF-κB pathway promotes the invasiveness of GBM cells in an NF-κB-inducing kinase (NIK)-dependent manner. While NIK activity is predominantly regulated at the posttranslational level, we show here that NIK (MAP3K14) is upregulated at the transcriptional level in invading cell populations, with the highest NIK expression observed in the most invasive cells. GBM cells with high induction of NIK gene expression demonstrate characteristics of collective invasion, facilitating invasion of neighboring cells. Furthermore, we demonstrate that the E2F transcription factors E2F4 and E2F5 directly regulate NIK transcription and are required to promote GBM cell invasion in response to TWEAK. Overall, our findings demonstrate that transcriptional induction of NIK facilitates collective cell migration and invasion, thereby promoting GBM pathogenesis.

Figure 1

Induction of NIK transcription directly correlates with GBM cell invasion. (A) Quantification of three-dimensional collagen invasion assay BT25, BT114, and BT116 patient derived GBM cells after 48 h. Invasion was conducted under basal, TWEAK (10 ng/mL), or TNFα (10 ng/mL) treated conditions. Data represented as mean±SD, two-way ANOVA p < 0.001. (B–E) RNA sequencing analysis of BT25 GBM cells with either TWEAK treatment (10 ng/mL for 4 h) or TNFα treatment (10 ng/mL for 30 min) on wild-type cells or NIK KO BT25 cells compared to wild-type BT25 cell gene expression. IPA software was used for pathway analyses. (F) RT-qPCR analysis of BT25 cells for MAP3K14, MAP3K8, IKKα, IKKβ, AKT. Cells were treated with either 10 ng/mL TWEAK or 10 ng/mL TNFα and analyzed at 0, 1, 4, 8, and 24 h. Data represented as mean±SD, two-way ANOVA. (G) RT-qPCR analysis was utilized to further examine the expression of NF-κB proteins and MAP3Ks in BT114 GBM cells. Cells were treated with either 10 ng/mL TWEAK or 10 ng/mL TNFα and analyzed at 0, 0.25, 0.5, 1, 4, 8, 16, and 24 h, represented by each group of shaded bars. RT-qPCR analysis of cells extracted from the collagen matrix at 0, 0.5, 4, 8, and 24 h post invasion. Statistical analysis by two-way ANOVA of MAP3K14 and MAP3K8 across treatment groups; MAP3K14-TWEAK 0 h vs. 4 h, 8 h, or 24 h ****p ≤ 0.0001, MAP3K14-TNF 0 h vs. 24 h *p < .05 (other time points are ns), MAP3K14-Invasion 0 h vs. 0.5 h, 1 h, 4 h, 8 h, or 24 h ****p ≤ 0.0001. MAP3K8-TWEAK 0 h vs. 0.25 h **p ≤ 0.01, 0 h vs. 0.5 h, 4 h, or 8 h ****p ≤ 0.0001, MAP3K8-TNF 0 h vs. 0.25 h, 0.5 h, 1 h, or 4 h ****p ≤ 0.0001.

Figure 2

NIK expression is upregulated in highly invasive GBM cells and promotes collective invasion. (A) Live-cell confocal microscopy was utilized to visualize the RFP reporter under the NIK promoter (pNIK-RFP) (red) under unstimulated conditions (NT). The monolayer (0 h) was pseudocolored white and overlaid to the reference invasion distance at 24 h and 48 h. (B) Graph representation of RFP intensity after 48 h and distance the cells invaded. Pearson’s correlation between RFP intensity and distance is r = 0.788 and p = 0.008, ***p ≤ 0.001. (C) Invasion of BT116 pNIK-RFP cells labeled with DiO (green) with no treatment (NT) or treated with 10 ng/mL TWEAK. (D) BT116 pNIK-RFP cells were used for a spheroid invasion assay. Spheroids were embedded in three-dimensional collagen matrix and either left untreated (NT) or treated with TWEAK (10 ng/mL). Spheres were allowed to invade for 72 h. Confocal microscopy was used to image spheroids, labeled DiO (green), RFP reporter (red), and DAPI (blue). (E) Outline of BT116 pNIK-RFP (red) spheroids embedded in three-dimensional collagen for 0–72 h, either untreated (NT) or treated with TWEAK (TW) for 48 h. (F) The ImageJ particle analysis function was used to quantify the number of cells invading into the three-dimensional collagen matrix at various time points. Data represented as mean±SEM, two-way ANOVA. (G) ImageJ was used to measure the radius of the spheroids at 0 h, 24 h, and 48 h. Data represented as mean±SEM. (H) Three-dimensional projection image of BT116 pNIK-RFP (Red) spheroids embedded in collagen and treated with TWEAK (TW) for 72 h. Cells were labeled with DiO (green) and DAPI (blue). White arrowheads with enhanced images of cells undergoing collective invasion. Spheroid invasion assays were conducted in n > 3 either untreated or treated with 10 ng/mL TWEAK.

Figure 3

Inhibition of NIK activity reduces GBM invasion. (A) BT116 pNIK RFP cells were labeled with DiO before being seeded on collagen matrix and allowed to invade for 48 h. Live cell confocal microscopy was used to obtain Z-stack images over 48 h. Cells either received no treatment (NT) or were treated with 10 ng/mL TWEAK, 10 ng/mL TNF-α, 10 ng/ml TWEAK with 100 μg/mL mangiferin, or 100 μg/mL mangiferin. (B) Quantification of cell invasion after 48 h. Data represented as mean±SEM, one-way ANOVA. (C) Immunoblot showing decreased RelB and p52 in the nucleic fraction of GBM cells treated with mangiferin.

Figure 4

E2F4 and E2F5 regulate NIK gene expression. (A) Cytoplasmic and nuclear E2F4 and E2F5 were examined in GBM cells by Western blot in unstimulated vs stimulated conditions (10 ng/mL TWEAK for 4 h) with RelB as a positive TWEAK-inducible control and Lamin A as a nuclear marker. (B) Immunoblot of NIK expression under MG132 (5 μM)- and TWEAK (10 ng/mL)-stimulated conditions in control and E2F4, E2F5, or E2F4 and E2F5 knockout cells. C) ChIP analysis performed in GBM cells treated with or without 10 ng/mL TWEAK by using E2F4, E2F5, or IgG antibodies. (D) qPCR analysis of NIK (MAP3K14) induction in double knockout E2F4 and E2F5 cells (BT25 and BT114). Data represented as mean±SD, unpaired student t-test. (E) Invasion assay of BT25 and BT114 double knockout E2F4/E2F5 cells. Data represented as mean±SEM, unpaired student t-test. (F) Invasion assay of BT25 and BT114 double knockout E2F4 and E2F5 cells rescued by NIK overexpression. (D–F) Comparisons were made using glioma cells in untreated (NT) vs. TWEAK-treated (10 ng/mL for 4 h) conditions. Data represented as mean±SEM, unpaired student t-test.