A Positive Feedback DNA-PK/MYT1L-CXCR1-ERK1/2 Proliferative Signaling Loop in Glioblastoma

Abstract

Glioblastoma is the most common primary brain tumor in adults. Our previous studies revealed a functional interplay of myelin transcription factor 1-like (MYT1L) with the DNA-dependent protein kinase (DNA-PK) in the regulation of p21 transcription. However, the contributing role of this functional interplay in glioblastoma remains largely unknown. Here, we used cell lines with normal DNA-PK (HEK293 and M059K) or deficient DNA-PK (M059J) as a model system to demonstrate the importance of the DNA-PK-dependent activation of MYT1L in controlling the transcription of CXC chemokine receptor 1 (CXCR1) in a positive-feedback proliferative signaling loop in glioblastoma with numerous conventional techniques. In normal DNA-PK cells, MYT1L acted as an oncogene by promoting cell proliferation, inhibiting apoptosis, and shortening a cell cycle S phase. However, in DNA-PK-deficient cells, MYT1L functioned as a tumor suppressor by inhibiting cell proliferation and inducing a G1 arrest. The enforced expression of MYT1L promoted CXCR1 transcription in DNA-PK-normal cells but attenuated transcription in DNA-PK-deficient cells. Bioinformatics analysis predicted a MYT1L-binding sequence at the CXCR1 promoter. The functional dependence of MYT1L on DNA-PK in CXCR1 transcription was validated by luciferase assay. Although the expression of CXCR1 was lower in M059J cells as compared to M059K cells, it was higher than in normal brain tissue. The CXCR1 ligands interleukin 8 (IL-8) and GRO protein alpha (GROα) expressed in M059J and M059K cells may signal through the extracellular signal-regulated kinase 1/2 (ERK1/2) pathway that can be blocked by CXCR1 siRNA. Our findings demonstrate the existence of a positive feedback DNA-PK/MYT1L-CXCR1-ERK1/2 proliferation loop in glioblastoma cells that may represent a pharmacological target loop for therapeutic intervention.

Keywords: CXCR1; DNA-PK; MYT1L; glioblastoma; proliferative signaling loop.

Figure 1

An oncogenic role of MYT1L in glioblastoma cells with normal DNA-PK activity. M059K cells grown to 90% confluency were transfected with either 40 nM MYT1L siRNA or 40 nM negative control siRNA. (A) At 24 h after transfection, cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (B) At 72 h after transfection, whole cellular lysates were prepared and subjected to Western blot analysis of MYT1L; actin served as a loading control. (C,D) At 72 h after transfection, cells were harvested for apoptosis (C) and cell cycle (D) analyses as described in “Methods”; all the analyses were performed in triplicate. (E) HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells were replated in 96-well plates, and cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (F) Whole cellular lysates were prepared from HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to Western blot analysis of MYT1L; actin served as a loading control. (G–J) HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells were harvested for cell cycle (G,H) and apoptosis (I,J) analyses as described in “Methods”; all the analyses were performed in triplicate. (K) Whole cellular lysates were prepared from HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to Western blot analysis of AKT1, BAX, BCL2, caspase 3, CDK2, CDK4, CDK6, cyclin A2, cyclin D1, cyclin E1, ERK1/2, p21, p27, pAKT1/2/3, and pERK1/2; actin served as a loading control. PE indicates fluorescent dye Propidium Iodide (PI) bound to DNA, FITC indicates FITC-Annexin V bound to cell membrane phospholipid phosphatidylserine. * indicates p < 0.05. Note: the whole cellular lysates used in “(F)” and “(K)” are exactly the same biological samples.

Figure 1

An oncogenic role of MYT1L in glioblastoma cells with normal DNA-PK activity. M059K cells grown to 90% confluency were transfected with either 40 nM MYT1L siRNA or 40 nM negative control siRNA. (A) At 24 h after transfection, cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (B) At 72 h after transfection, whole cellular lysates were prepared and subjected to Western blot analysis of MYT1L; actin served as a loading control. (C,D) At 72 h after transfection, cells were harvested for apoptosis (C) and cell cycle (D) analyses as described in “Methods”; all the analyses were performed in triplicate. (E) HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells were replated in 96-well plates, and cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (F) Whole cellular lysates were prepared from HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to Western blot analysis of MYT1L; actin served as a loading control. (G–J) HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells were harvested for cell cycle (G,H) and apoptosis (I,J) analyses as described in “Methods”; all the analyses were performed in triplicate. (K) Whole cellular lysates were prepared from HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to Western blot analysis of AKT1, BAX, BCL2, caspase 3, CDK2, CDK4, CDK6, cyclin A2, cyclin D1, cyclin E1, ERK1/2, p21, p27, pAKT1/2/3, and pERK1/2; actin served as a loading control. PE indicates fluorescent dye Propidium Iodide (PI) bound to DNA, FITC indicates FITC-Annexin V bound to cell membrane phospholipid phosphatidylserine. * indicates p < 0.05. Note: the whole cellular lysates used in “(F)” and “(K)” are exactly the same biological samples.

Figure 1

An oncogenic role of MYT1L in glioblastoma cells with normal DNA-PK activity. M059K cells grown to 90% confluency were transfected with either 40 nM MYT1L siRNA or 40 nM negative control siRNA. (A) At 24 h after transfection, cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (B) At 72 h after transfection, whole cellular lysates were prepared and subjected to Western blot analysis of MYT1L; actin served as a loading control. (C,D) At 72 h after transfection, cells were harvested for apoptosis (C) and cell cycle (D) analyses as described in “Methods”; all the analyses were performed in triplicate. (E) HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells were replated in 96-well plates, and cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (F) Whole cellular lysates were prepared from HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to Western blot analysis of MYT1L; actin served as a loading control. (G–J) HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells were harvested for cell cycle (G,H) and apoptosis (I,J) analyses as described in “Methods”; all the analyses were performed in triplicate. (K) Whole cellular lysates were prepared from HEK293-MYT1L, HEK293-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to Western blot analysis of AKT1, BAX, BCL2, caspase 3, CDK2, CDK4, CDK6, cyclin A2, cyclin D1, cyclin E1, ERK1/2, p21, p27, pAKT1/2/3, and pERK1/2; actin served as a loading control. PE indicates fluorescent dye Propidium Iodide (PI) bound to DNA, FITC indicates FITC-Annexin V bound to cell membrane phospholipid phosphatidylserine. * indicates p < 0.05. Note: the whole cellular lysates used in “(F)” and “(K)” are exactly the same biological samples.

Figure 2

A tumor suppressive role of MYT1L in DNA-PK-deficient glioblastoma cells. M059J cells grown to 90% confluency were transfected with either 40 nM MYT1L siRNA or 40 nM negative control siRNA. (A) At 24 h after transfection. (B) Western blot analysis of MYT1L, 72 h after transfection; actin served as a loading control. (C,D) At 72 h after transfection, cells were harvested for apoptosis (C) and cell cycle (D) analyses as described in “Methods”; all the analyses were performed in triplicate. (E) M059J-MYT1L and M059J-GFP cells were replated in 96-well plates, and cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (F) Western blot analysis of MYT1L; actin served as a loading control. (G,H) M059J-MYT1L and M059J-GFP cells grown to subconfluency were harvested for cell cycle (G) and apoptosis (H) analyses. (I) Western blot analysis of AKT1, BAX, BCL2, caspase 3, CDK2, CDK4, CDK6, cyclin A2, cyclin D1, cyclin E1, ERK1/2, p21, p27, pAKT1/2/3, and pERK1/2; actin served as a loading control. PE indicates fluorescent dye Propidium Iodide (PI) bound to DNA, FITC indicates FITC-Annexin V bound to cell membrane phospholipid phosphatidylserine. * indicates p < 0.05; ** indicates p < 0.01. Note: the whole cellular lysates used in “(F)” and “(I)” are exactly the same biological samples.

Figure 2

A tumor suppressive role of MYT1L in DNA-PK-deficient glioblastoma cells. M059J cells grown to 90% confluency were transfected with either 40 nM MYT1L siRNA or 40 nM negative control siRNA. (A) At 24 h after transfection. (B) Western blot analysis of MYT1L, 72 h after transfection; actin served as a loading control. (C,D) At 72 h after transfection, cells were harvested for apoptosis (C) and cell cycle (D) analyses as described in “Methods”; all the analyses were performed in triplicate. (E) M059J-MYT1L and M059J-GFP cells were replated in 96-well plates, and cell growth was measured daily with the MTT assay kit in triplicate as described in “Methods”. (F) Western blot analysis of MYT1L; actin served as a loading control. (G,H) M059J-MYT1L and M059J-GFP cells grown to subconfluency were harvested for cell cycle (G) and apoptosis (H) analyses. (I) Western blot analysis of AKT1, BAX, BCL2, caspase 3, CDK2, CDK4, CDK6, cyclin A2, cyclin D1, cyclin E1, ERK1/2, p21, p27, pAKT1/2/3, and pERK1/2; actin served as a loading control. PE indicates fluorescent dye Propidium Iodide (PI) bound to DNA, FITC indicates FITC-Annexin V bound to cell membrane phospholipid phosphatidylserine. * indicates p < 0.05; ** indicates p < 0.01. Note: the whole cellular lysates used in “(F)” and “(I)” are exactly the same biological samples.

Figure 3

Transcriptional control of CXCR1 by MYT1L and DNA-PK. (A) Whole cellular lysates were prepared from BJ-5ta, A-172, M059J, M059K, IMR-5, IMR-32, SH-SY5Y, SK-N-AS, SK-N-BE(2), SK-N-MC, and SK-N-SH cells and subjected to Western blot analysis of MYT1L; GAPDH served as a loading control. (B) Total RNA was isolated from BJ-5ta, A-172, M059J, M059K, IMR-5, IMR-32, SH-SY5Y, SK-N-AS, SK-N-BE(2), SK-N-MC, and SK-N-SH cells and subjected to the qRT-PCR analysis of CXCR1 that was performed in triplicate. (C) Total RNA was isolated from HEK293-MYT1L, HEK293-GFP, M059J-MYT1L, M059J-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to the qRT-PCR analysis of CXCR1 that was performed in triplicate. (D,E) Whole cellular lysates were prepared from HEK293-MYT1L, HEK293-GFP, M059J-MYT1L, M059J-GFP, M059K-MYT1L, and M059K-GFP cells and subjected to Western blot analysis of CXCR1 and CXCR2; actin and GAPDH served as a loading control. Note that the images of blots have been cropped and aligned in the same order as panels “(C)” and “(E)”. (F) A diagram of wild-type and mutant CXCR1 promoter/reporter constructs. (G) HEK293, M069J, and M059K cells were cotransfected with either an empty vector or pGL3-wtCXCR1-luc or pGL3-mtCXCR1^{−2114/−2103}-luc reporter plasmid in combination with pCMV6-MYT1L and pRL-TK; at 24 h after transfection, the luciferase activity was measured in duplicate as described in “Methods”. * indicates p < 0.05.

Figure 4

GROα- and IL-8-CXCR1/ERK1/2 signaling in glioblastoma. (A) Total RNA was isolated from HEK293-GFP (Passage# 3), HEK293-MYT1L (Passage# 5), M059K-GFP (Passage# 4), M059K-MYT1L (Passage# 6), M059J-GFP (Passage# 5), and M059J-MYT1L (Passage# 5) cells and subjected to the qRT-PCR analysis of IL-8 that was performed in triplicate. (B,D,F) Total RNA was isolated from HEK293-GFP, HEK293-MYT1L, M059K-GFP, M059K-MYT1L, M059J-GFP, and M059J-MYT1L cells and subjected to the qRT-PCR analysis of GROα that was performed in triplicate. (C,E,G) HEK293-GFP, HEK293-MYT1L, M059K-GFP, M059K-MYT1L, M059J-GFP, and M059J-MYT1L cells grown on a glass cover slide were subjected to immunofluorescence staining for GROα as described in “Methods”. (H) Total RNA was isolated from the brain normal tissue, M059J, M059K, SK-N-BE(2), and A-172 cells and subjected to the qRT-PCR analysis of IL-8 that was performed in triplicate. (I) Whole cellular lysates were prepared from the brain normal tissue, M059J, M059K, and A-172 cells and subjected to Western blot analysis of GROα, CXCR1, CXCR2, pERK1/2, ERK1/2, pAKT1/2/3, and AKT1; GAPDH served as a loading control. (J) M059K cells were transfected with 20 nM or 80 nM negative control-A or CXCR1 siRNA; at 48 h after transfection, whole cellular lysates were prepared and subjected to Western blot analysis of CXCR1, pERK1/2, ERK1/2, pAKT1/2/3, and AKT1; GAPDH served as a loading control. (K) M059K cells were transfected with either 80 nM negative control-A or 80 nM CXCR1 siRNA; at 24 h after transfection, cell growth was measured with the MTT assay kit in triplicate as described in “Methods”. * indicates p < 0.05.

Figure 4

GROα- and IL-8-CXCR1/ERK1/2 signaling in glioblastoma. (A) Total RNA was isolated from HEK293-GFP (Passage# 3), HEK293-MYT1L (Passage# 5), M059K-GFP (Passage# 4), M059K-MYT1L (Passage# 6), M059J-GFP (Passage# 5), and M059J-MYT1L (Passage# 5) cells and subjected to the qRT-PCR analysis of IL-8 that was performed in triplicate. (B,D,F) Total RNA was isolated from HEK293-GFP, HEK293-MYT1L, M059K-GFP, M059K-MYT1L, M059J-GFP, and M059J-MYT1L cells and subjected to the qRT-PCR analysis of GROα that was performed in triplicate. (C,E,G) HEK293-GFP, HEK293-MYT1L, M059K-GFP, M059K-MYT1L, M059J-GFP, and M059J-MYT1L cells grown on a glass cover slide were subjected to immunofluorescence staining for GROα as described in “Methods”. (H) Total RNA was isolated from the brain normal tissue, M059J, M059K, SK-N-BE(2), and A-172 cells and subjected to the qRT-PCR analysis of IL-8 that was performed in triplicate. (I) Whole cellular lysates were prepared from the brain normal tissue, M059J, M059K, and A-172 cells and subjected to Western blot analysis of GROα, CXCR1, CXCR2, pERK1/2, ERK1/2, pAKT1/2/3, and AKT1; GAPDH served as a loading control. (J) M059K cells were transfected with 20 nM or 80 nM negative control-A or CXCR1 siRNA; at 48 h after transfection, whole cellular lysates were prepared and subjected to Western blot analysis of CXCR1, pERK1/2, ERK1/2, pAKT1/2/3, and AKT1; GAPDH served as a loading control. (K) M059K cells were transfected with either 80 nM negative control-A or 80 nM CXCR1 siRNA; at 24 h after transfection, cell growth was measured with the MTT assay kit in triplicate as described in “Methods”. * indicates p < 0.05.

Figure 4

GROα- and IL-8-CXCR1/ERK1/2 signaling in glioblastoma. (A) Total RNA was isolated from HEK293-GFP (Passage# 3), HEK293-MYT1L (Passage# 5), M059K-GFP (Passage# 4), M059K-MYT1L (Passage# 6), M059J-GFP (Passage# 5), and M059J-MYT1L (Passage# 5) cells and subjected to the qRT-PCR analysis of IL-8 that was performed in triplicate. (B,D,F) Total RNA was isolated from HEK293-GFP, HEK293-MYT1L, M059K-GFP, M059K-MYT1L, M059J-GFP, and M059J-MYT1L cells and subjected to the qRT-PCR analysis of GROα that was performed in triplicate. (C,E,G) HEK293-GFP, HEK293-MYT1L, M059K-GFP, M059K-MYT1L, M059J-GFP, and M059J-MYT1L cells grown on a glass cover slide were subjected to immunofluorescence staining for GROα as described in “Methods”. (H) Total RNA was isolated from the brain normal tissue, M059J, M059K, SK-N-BE(2), and A-172 cells and subjected to the qRT-PCR analysis of IL-8 that was performed in triplicate. (I) Whole cellular lysates were prepared from the brain normal tissue, M059J, M059K, and A-172 cells and subjected to Western blot analysis of GROα, CXCR1, CXCR2, pERK1/2, ERK1/2, pAKT1/2/3, and AKT1; GAPDH served as a loading control. (J) M059K cells were transfected with 20 nM or 80 nM negative control-A or CXCR1 siRNA; at 48 h after transfection, whole cellular lysates were prepared and subjected to Western blot analysis of CXCR1, pERK1/2, ERK1/2, pAKT1/2/3, and AKT1; GAPDH served as a loading control. (K) M059K cells were transfected with either 80 nM negative control-A or 80 nM CXCR1 siRNA; at 24 h after transfection, cell growth was measured with the MTT assay kit in triplicate as described in “Methods”. * indicates p < 0.05.

Figure 5

DNAPK/MYT1L-CXCR1 signaling loop in glioblastoma. (A,B) In glioblastoma cells with normal DNAPK activity, under non-stimulated conditions, the ERK1/2 pathway is inactivated, tumor suppressor MYT1L binds to CXCR1 promoter and acts as a repressor that blocks CXCR1 transcription (A); once ERK1/2 pathway activated by IL-8/GROα binding to receptor CXCR1, the activated ERK1/2 kinase may phosphorylate coactivator DNAPK that, in turn, phosphorylates MYT1L, the phosphorylated MYT1L recruits histone acetyltransferases (HATs) and transcription factors (TFs) to CXCR1 promoter, eventually resulting in CXCR1 transcription (B). (C) In glioblastoma cells with loss-of-function DNAPK, transcriptional repressor MYT1L can not be activated by DNAPK-dependent phosphorylation; as a result, the transcription of CXCR1 is reduced.