Regression/eradication of gliomas in mice by a systemically-deliverable ATF5 dominant-negative peptide

Abstract

Malignant gliomas have poor prognosis and urgently require new therapies. Activating Transcription Factor 5 (ATF5) is highly expressed in gliomas, and interference with its expression/function precipitates targeted glioma cell apoptosis in vitro and in vivo. We designed a novel deliverable truncated-dominant-negative (d/n) form of ATF5 fused to a cell-penetrating domain (Pen-d/n-ATF5-RP) that can be intraperitoneally/subcutaneously administered to mice harboring malignant gliomas generated; (1) by PDGF-B/sh-p53 retroviral transformation of endogenous neural progenitor cells; and (2) by human U87-MG xenografts. In vitro Pen-d/n-ATF5-RP entered into glioma cells and triggered massive apoptosis. In vivo, subcutaneously-administered Pen-d/n-ATF5-RP passed the blood brain barrier, entered normal brain and tumor cells, and then caused rapid selective tumor cell death. MRI verified elimination of retrovirus-induced gliomas within 8-21 days. Histopathology revealed growth-suppression of intracerebral human U87-MG cells xenografts. For endogenous PDGF-B gliomas, there was no recurrence or mortality at 6-12 months versus 66% mortality in controls at 6 months. Necropsy and liver-kidney blood enzyme analysis revealed no adverse effects on brain or other tissues. Our findings thus identify Pen-d/n-ATF5-RP as a potential therapy for malignant gliomas.

Keywords: ATF5; apoptosis; brain cancer; cell penetrating peptide; d/n- ATF5.

Figure 1. GFP-d/n-ATF5 C-terminally truncated fusion protein (GFP-d/n-ATF5-Tr) promotes the same level of apoptosis as full-length GFP-d/n-ATF5 protein in C6 glioma cells

C6 cells were transfected with pQC-X-I-eGFP, pQC-d/n-GFPATF5, or pQC-GFPATF5-tr. The percentages (mean ± SEM, n = 4; total of approximately 200 cells scored per condition) of condensed apoptotic nuclei in GFP + transfected cells were determined 2 days later. Student's t-test; GFP+ cells versus GFP-d/n-ATF5+ cells or GFP-d/n-ATF5-tr cells, (*p < 0.05); GFP-d/n-ATF5+ cells versus GFP-d/n-ATF5-tr cells, (Not Significant).

Figure 2. Purity and molecular properties of bacterially expressed and purified 6xhistidine-Flag-Tagged Penetratin-Flag-D/N-ATF5-tr (Pen-d/n-ATF5-RP) and 6xhistidine-Flag-Tagged Penetratin-Flag-Control (Pen-control-RP) peptides

A. Coomassie stained SDS-PAGE of purified Pen-d/n-ATF5-RP and Pen-control-RP (5 μg per lane). Molecular weight markers are shown on the left, and a linear scheme of each peptide is shown above each lane. Purification was as described in Methods. B. Deconvoluted mass spectra from LC-high Resolution mass spectrometry of purified Pen-d/n-ATF5-RP. The most abundant species is the 12,948.88 Da monomer form without formyl-methionine followed by the formyl-methionine 13,127 Da monomer form (isoform). The spectrum also reveals a small amount of the 25,897.5 Da dimer. C. Stability of Pen-d/n-ATF5-RP in Human Serum. Pen-d/n-ATF5-RP (36 μM) was incubated with human serum (25% v/v in PBS) at 37° C for 0 to 48 h. Aliquots were withdrawn at various times and the Pen-d/n-ATF5-RP peptide was resolved by SDS-PAGE, transferred to PVDF membrane and probed with anti-Flag antibody. The anti-Flag signal was detected by near IR using LiCor software and densitometry of the band at the expected size of Pen-d/n-ATF5-RP and quantified using Image J. Values are mean ± SEM, n = 3).

Figure 3. Uptake and retention of Pen-d/n-ATF5-RP by cultured glioblastoma cells

A. Confocal images of C6 rat glioblastoma cells incubated for 4 hours with either 200 nM Pen-control-RP (left) or Pen-d/n-ATF5-RP (right). Cells were washed, fixed and stained with anti-Flag (red) and DAPI (blue). Scale bar = 2 μm. B. Rat C6 and human U87 glioblastoma cells were incubated for the indicated times with 3 μM Pen-d/n-ATF5-RP, washed, fixed and immunostained with anti-Flag (green) and DAPI (blue). Scale bar = 5 μm.

Figure 4. Pen-d/n-ATF5-RP promotes apoptosis of C6 glioblastoma cells

C6 cells were treated with 3 μM Pen-d/n-ATF5-RP or 3 μM Pen-Control-RP, or were untreated. The percentage (mean ± SEM; n = 4 in 2 independent experiments; approximately 200 cells scored) of condensed apoptotic nuclei in cells was determined 5 days later. Student's t-test; Pen-d/n-ATF5-RP versus Pen-Control-RP cells or nontreated, (*p < 0.05); Pen-Control-RP cells versus nontreated cells, (p = 0.29).

Figure 5. Pen-d/n-ATF5-RP enters the mouse brain and causes targeted apoptosis of glioma cells

A.-F. Representative brain sections stained with Flag antibody to indicate presence of Pen-d/n-ATF5-RP or HA to identify presence of tumor-inducing retrovirus (red); TUNEL to identify apoptosis (green) and DAPI to localize nuclei (blue). A. Murine brain tumor 24 h post-treatment (16 h after last injection) with Pen-d/n-ATF5-RP (52 days post-retrovirus injection). B. Normal contralateral cerebral hemisphere of the same mouse in A.. C. Murine brain tumor 24 h post-injection with saline (59 days post-retrovirus injection). Presence of Pen-d/n-ATF5-RP within cells is confirmed in the treated mouse A.,B. versus saline control C. by increased Flag antibody staining. Glioma cell-specific induction of apoptosis by Pen-d/n-ATF5-RP is illustrated by increased TUNEL staining (green) in A. as compared to B. and C.. D. TUNEL and DAPI staining of a tumor-containing brain section 160 days post-retrovirus injection and 3 days after injection of Pen-control-RP. Note HA+ cells identifying tumor cells and absence of TUNEL staining. E. Staining as in D. of a tumor-containing section (143 days post-retrovirus injection) and 3 days [3D] after Pen-d/n-ATF5-RP treatment. Note the presence of TUNEL staining in HA+ tumor cells and fragmented appearance of the staining as compared to A. and D.. F. Staining as in D. of a tumor-containing section 150 days after retrovirus injection and 2 days after 2 treatments of subcutaneous Pen-d/n-ATF5-RP injections at seven days [7D] after the first injection. Note the qualitative similarity of staining pattern to E. with fragmented PDGF-B-HA and TUNEL staining. Scale bars equal 20 μm.

Figure 6. Example of MRI and histopathology of a mouse glioma treated with Pen-Control-RP peptide

A. Post-contrast 3D FLASH MRI coronal image of the cerebrum of a control mouse that was not injected with PDGF-B-HA/sh-p53 retrovirus. B. Post-contrast 3D FLASH MRI coronal image of mouse cerebrum showing a bilateral tumor (white contrast) 246 days after PDGF-B-HA/shp53 retrovirus injection and prior to treatment with Pen-Control-RP peptide. C. Post-contrast 3D FLASH MRI image of the same mouse brain 40 days after subcutaneous treatment with Pen-Control-RP peptide (as described in the text) reveals persistence of the tumor (arrows). D. H&E stained sections of the same mouse brain at tumor-containing areas 1 and 2 shown by arrows in panel C.. The mouse was sacrificed 116 days after the second treatment with Pen-Control-RP peptide due to moribund behavior. Presence of tumor is indicated in both sections by hyperchromatic nuclei and higher cellularity. E. Immunostaining for HA tag in sections from areas 1 and 2 shown in Panel C. reveals presence of virally-delivered PDGF-B-HA in induced tumor cells. F. Immunostaining of sections from areas 1 and 2 shown in Panel C. reveals a high index of Ki67+/dividing cells indicative of tumor. Scale bars in D-F are 20 μm.

Figure 7. Pen-d/n-ATF5-RP promotes rapid and long-term regression/eradication of mouse glioma as indicated by MRI and histology

A. Post-contrast 3D FLASH MRI scans of a mouse brain before and at various times after treatment (as described in text) with Pen-d/n-ATF5-RP. Pretreatment shows image of cortex 243 days after PDGF-B-HA/shp53 retrovirus injection. Yellow arrows indicate location of the bilateral tumor. Post-treatment images of the same position of the mouse cortex are at the indicated times after the second administration of Pen-d/n-ATF5-RP. Yellow arrows in post-treatment images show location of original tumor. B. H&E image of the same mouse brain harvested 192 days after the second Pen-d/n-ATF5-RP treatment. Region 1 represents the location of the section as shown in the final time point in A. and at which the tumor was present before treatment. Note the absence of hyperchromatic nuclei and higher cellularity that characterize gliomas. C. Ki67 staining in region 2 (from Panel A/176 days post-treatment). Note the absence of Ki67+/proliferating cells seen in gliomas. D. HA/DAPI staining of section from region 1. Note the absence of cells expressing exogenous PDGF-B-HA. E. GFAP/DAPI staining of section region 1. Note clusters of GFAP+ cells consistent with the presence of a glial scar where the tumor was formerly present. Lack of HA staining of a nearby section confirmed the absence of tumor cells. Diagonal green stripes are due to tissue folds. Scale bar is 20 μm.

Figure 8. Long-term survival and tumor presence outcomes for PDGF-BHA/shp53 glioma-bearing mice treated with Pen-d/n-ATF5-RP

A. Survival of glioma (PDGF-B/shRNA-p53 induced)-bearing mice (verified by MRI) with or without treatment with Pen-d/n-ATF5-RP (subcutaneous delivery as described in the text). Of the nine control mice, four mice were treated with Pen-Control-RP peptide and five were untreated. The experimental endpoint was 200 days after initial tumor detection by MRI. Survival analysis achieved by Log-rank (Mantel-Cox) test showed a p-value = 0.0194. B. MRI outcomes for tumor-bearing mice before and after subcutaneous treatment with Pen-d/n-ATF5-RP as described in the text. Fisher Exact Match test showed p = 0.0002 with Positive Predictive Value 95% confidence interval of 0.6306 to 1.000. The latter times range from 176-225 days after tumor treatment (183-230 days after tumor detection). C. Brain histopathological outcomes for tumors in control and Pen-d/n-ATF5-RP treated mice. In all cases, MRI verified the presence of tumors prior to treatment. Fisher Exact Match test showed p < 0.0001 with Positive Predictive Value 95% confidence interval of 0.0 to 0.3694. The brains of animals described in A. were harvested either after death (6 controls), after the experimental endpoint (4 Pen-d/n-ATF5-RP treated animals; 2 Pen-Control or 1 non-Treated animal) or after sacrifice for non-tumor related health problems (2 Pen-d/n-ATF5-RP treated animals). For treated animals, histological analysis was carried out 260-547 days after tumor initiation (183-392 days after Pen-d/n-ATF5-RP administration and 190-397 days after initial tumor detection). Brain sections were prepared as described in Methods and were stained with H&E and immunostained for Ki67 and HA (to identify PDGF-B-HA+ tumor cells). The presence/absence of tumors was based on observations of hyperchromatic nuclei, high cellularity, elevated Ki67 staining and HA immunostaining.