Diffuse intrinsic pontine glioma (DIPG) is a poor-prognosis pediatric brain tumor with a median survival of less than 1 year. No effective therapy is currently available, and no therapeutic advances have been made in several decades. We have previously identified BMI-1 as a potential therapeutic target in DIPG and have shown that BMI-1 is highly expressed in DIPG tumors regardless of histone 3 subtype. In the present study, we show that the modulation of BMI-1 leads to DNA damage, M phase cell-cycle arrest, chromosome scattering, and cell death. Interestingly, EZH2 inhibition did not alter these effects. Furthermore, modulation of BMI-1 sensitizes DIPG patient-derived stem-like cells to ionizing radiation (IR). Treatment of DIPG stem-like cells with PTC596, a BMI-1 modulator, and IR impairs the kinetics of DNA damage response (DDR). Both DDR foci formation and resolution were delayed, resulting in further reduction in cell viability compared with either treatment alone. In vivo, treatment of mice bearing DIPG xenografts with PTC596 leads to decreased tumor volume and growth kinetics, increased intratumoral apoptosis, and sustained animal survival benefit. Gene expression analysis indicates that BMI-1 expression correlates positively with DIPG stemness and BMI-1 signature. At the single-cell level, the analysis reveals that BMI-1 pathway is upregulated in undifferentiated cells and positively correlates with stemness in DIPG tumors.

Implications:

Together, our findings indicate that BMI-1 modulation is associated with mitotic abnormalities, impaired DDR, and cell death, supporting the combination of BMI-1 modulation and radiation as a promising novel therapy for children with DIPG.

Brain tumors are the most common solid tumors of childhood and are the leading cause of cancer-related deaths in children (1). Diffuse intrinsic pontine glioma (DIPG) is a universally fatal brainstem tumor with survival less than 1 year despite multimodal therapy (2). This underscores an urgent need for novel therapies to improve outcome in this vulnerable population. The Polycomb repressive complex 1 (PRC1) and 2 (PRC2) are large multimeric complexes involved in gene silencing through modifications of chromatin structure. BMI-1 is a subunit of PRC1 required for monoubiquitination of histone H2A at lysine 119 (H2AK119Ub). BMI-1–associated E3 ubiquitin ligase activity represses multiple gene loci, including INK4A/ARF locus encoding for two tumor suppressors, p16INK4A and p14ARF (3, 4). BMI-1 is a proto-oncogene implicated in development, stemness of normal and malignant cells, self-renewal, and many cancer types (reviewed in ref. 4). We have previously shown that BMI-1 is highly expressed and a potential therapeutic target in DIPG (5). Selective posttranscriptional inhibition of BMI-1 using a small molecule, PTC209, repressed BMI-1 production, leading to the inhibition of DIPG patient-derived neurosphere cell proliferation, cell-cycle signaling, self-renewal, telomerase expression and activity, and DIPG cell migration (5). PTC596 is a novel small molecule, recently identified as a potent repressor of BMI-1 (6). PTC596 modulates BMI-1 by posttranslational phosphorylation ultimately leading to protein level reduction (6). PTC596 has completed phase I clinical trials in adults with advanced solid tumors (NCT02404480) and is currently being tested in newly diagnosed children with DIPG and high-grade gliomas (NCT03605550). In this study, we investigated the correlation of BMI-1 expression with DIPG stemness in bulk tumor and at the single-cell level. We also examined the mechanism of action and the in vitro and in vivo effects of PTC596 in DIPG. Our findings support BMI-1 modulation in combination with radiotherapy as a promising novel therapy for children with DIPG.

Patient samples and cell culture

All patient specimens were collected after obtaining written-informed consent from patients and families in accordance with approved Institutional Review Board (IRB) studies. DIPG tumors with their matched normal tissue were obtained from patients (CCHMC-PBTR cohort) as previously described (7). Primary normal human foreskin fibroblasts (HFF) strain (ATCC CRL-2091) was purchased from the American Type Culture Collection. HFF cells were cultured in DMEM (Gibco) supplemented with 10% FBS. CCHMC-DIPG-1 and CCHMC-DIPG-2 are primary DIPG neurosphere cells derived by aseptically dissociating brain tumor tissues obtained after autopsy from patients consented under the Pediatric Brain Tumor Repository (PBTR) study (IRB-approved protocols 2013-1245) at Cincinnati Children's Hospital Medical center (CCHMC). The primary patient-derived DIPG cells SU-DIPG-IV, SU-DIPG-XIII, CCHMC-DIPG-1, and CCHMC-DIPG-2 were cultured in neurosphere stem cell media as described elsewhere (5). HFF were characterized at their original source. All cells were expanded upon receipt or establishment for 2 to 3 passages and used within 1 to 2 months after thawing the cryopreserved cells. SU-DIPG-IV and XIII were authenticated by our collaborator M. Monje (Stanford University). CCHMC-DIPG-1 and 2 cells were generated in the Drissi laboratory with no additional authentication. Using the Universal Mycoplasma Detection Kit (ATCC; 30-1012K), the cell lines used in this study were confirmed to be negative for mycoplasma contamination.

Reagents

PTC596 was provided by PTC Therapeutics and was reconstituted in DMSO for in vitro studies. Doses were optimized based on the required time points, 100 nmol/L for short-term (24–48 hours) and 60 nmol/L for long-term treatment (72 hours and beyond). For in vivo studies, 0.5% HPMC with 0.1% Tween 80 in distilled water (vehicle) was used to reconstitute PTC596. EPZ011989 was provided by Epizyme Inc. and reconstituted in DMSO. Colchicine (Sigma-Aldrich) was reconstituted in sterile water.

RNA sequencing

RNA was extracted using the RNeasy Plus Mini Kit (Qiagen) according to the manufacturer's instructions and quantified using Qubit RNA BR assay kit (Invitrogen). TruSeq RNA Access kit was used for RNA library preparation. HiSeq X platform was used for sequencing. For RNA sequencing (RNA-seq) of the CCHMC-PBTR cohort, pseudoalignment and quantification of RNA sequencing data were performed by Kallisto 0.44.0 (8) with reference transcriptome build GRCh38.94. The quantification was normalized by Bioconductor package DESeq2 (9) and converted to normalized gene count and log transformed.

For single-cell RNA-seq (scRNA-seq), publicly available dataset GSE102130 (Filbin Cohort) was analyzed and visualized as described (10). The relative expression of BMI-1 was represented by transcript per million (TPM). Only the cells with nonzero BMI-1 TPM values are represented as colored dots.

In utero electroporation DIPG mouse model

The in utero electroporation (IUE) DIPG mouse model is described elsewhere (11). The RNA-seq data from the extracted tumors (n = 11) and control brainstem (n = 3), as well as the cell lines derived from these tumors were kindly provided by T. Phoenix (University of Cincinnati). The RNA-seq data were processed following the same workflow as above, with reference GRCm38.94. The log2 fold change was calculated comparing expression of each DIPG tumor sample with the mean expression of the control brainstems.

Glinsky signature and stemness score

Glinsky score was calculated as previously described (12). Stemness scores for human samples were calculated as previously described (10). Briefly, for a given set of genes reflecting an expression signature of a specific cell type or biological functions, the relative expression of each gene for each sample |i $| was calculated and centered. The signature scores of a set of genes (⁠|{G_j}) $| were calculated to quantify the relative expression of the genes compared with the average relative expression (⁠|Er $|⁠) of a control set using the formula |S{C_j}\ = \ average\ ( {Er( {{G_j}})})-\ average\ [ {Er( {{G_{j,contrl}}} )} ]$|⁠. All analyzed genes were binned into 25 bins by aggregated expression levels. For each signature gene, 100 genes were randomly selected from the same expression bin as the control set. The stemness score was defined as oligodendrocyte progenitor cells (OPC)-shared gene set minus the maximal expression score of the two differentiation programs gene sets (Astrocytic differentiation and Oligodendrocytic differentiation).

For the mouse DIPG dataset, we selected the mouse homolog genes to human using (ftp://ftp.ncbi.nlm.nih.gov/pub/HomoloGene/build68/homologene.data) to calculate the Glinsky scores and stemness scores following the same methods as outlined above.

For the scRNA-seq (Filbin cohort), the Glinsky scores of individual cells were calculated from the relative gene expression levels calculated using the above-mentioned method.

Cell-cycle, proliferation, and apoptosis assays

Cell cycle was analyzed as previously described (5). Data were analyzed using FlowJo v.10 (FlowJo) software. Cell proliferation was measured using WST-1 assay (Takara Bio) as per the manufacturer's instructions. WST-1 reagent was added to each well at a final concentration of 1:10, incubated for 1.5 hours at 37°C, and absorbance was measured at 450 nm with 650 nm as the reference wavelength. Apoptosis was measured using the Annexin V Apoptosis Detection kit (#88-8007, Affymetrix) according to the manufacturer's instructions.

Western blotting

Immunoblot assays were performed as previously described (5). Antibodies used were against mouse Bmi-1 (Abcam), human BMI-1, H2AK119Ub, H3 S10-P, Cyclin B1, Cleaved caspase-3, H3K27me3, Total H3, Total H2A, β-Actin (Cell Signaling Technology), and H2AX S139-P (Millipore). Bands visualized with ECL were captured using Azurec500 imaging system (Azure Biosystems) and quantified using Image Studio Lite (LI-COR). For the λ-Phosphatase experiments, 15 μg of protein lysates were treated with λ-Phosphatase (NEB) according to the manufacturer's protocol and then immunoblot was performed. Cytoplasmic and nuclear fractions were isolated as described elsewhere (13).

Immunofluorescence

Immunostaining was performed as described previously (14). Primary antibodies were used against γH2AX (1:500, Cell Signaling Technology), or cleaved caspase-3 (1:400, Cell Signaling Technology) with corresponding secondary antibodies (Alexa-Fluor 488– or 594–conjugated donkey anti-rabbit, or anti-mouse (1:500, Jackson ImmunoResearch) as applicable. For nuclear staining, cells were embedded with mounting media with DAPI (Vector Laboratories H1200). Images were captured with 60X oil objective on Nikon Eclipse Ti confocal microscope. Quantifications were performed using ImageJ software.

Microtubule staining

SU-DIPG-IV cells grown on coverslips were treated with DMSO or PTC596 (100 nmol/L) followed by microtubule staining using a live cell permeable ViaFluor 488 dye (1:500, Biotium) for 30 minutes at 37°C in cell culture incubator. For nuclear staining, cells were mounted with DAPI (Vector Laboratories H1200). Images were captured using Nikon Eclipse Ti confocal microscope and processed by NIS-element software (Nikon). Metaphase cells were identified by their characteristic polar spindle assembly formation and assessed for chromosome scattering as previously defined (15).

Metaphase immunofluorescence and centromere FISH

For metaphase spreads, cells treated with DMSO or PTC596 (48 hours) were synchronized with colchicine (0.1 μg/mL) for 2 hours and collected for swelling in hypotonic solution (0.2% KCl, 0.2% Tri-sodium citrate) for 10 minutes. Cells (150,000 cells per mL) were cytospun to obtain metaphase spreads, followed by fixation with 4% PFA and washing in 1X PBS (3 times). IF-FISH was performed as described above using CENPB-Cy3 pan-centromere and primary antibody against γH2A.X (1:500, Cell Signaling Technology). DAPI was used to visualize the chromosomes.

γH2AX foci quantification

IMARIS (Version9.2; BITPLANE) software was used to quantify γH2AX foci from z-stacked IF images. Data are represented as box and whisker plots, denoting the 50th percentile and 10th–90th percentile values, respectively. Fifty to 75 nuclei from more than three different fields were counted in two independent experiments.

Premature sister chromatid separation

CCHMC-DIPG-1 and SU-DIPG-IV cells were treated with DMSO or PTC596 (60 nmol/L) for the indicated time points after which cells were enriched in metaphase using 10 μg/mL Colcemid (Gibco Laboratories) for 20 minutes at 37°C, hypotonized with 0.075 mol/L KCL at room temperature, fixed with 3 parts methanol: 1 part acetic acid, and dropped on slides to obtain metaphase spreads. The slides were then stained with Giemsa Stain (Harleco). A metaphase was scored as positive for premature sister chromatid separation (PSCS) if all or the majority of sister chromatids in the metaphase spread demonstrated sister chromatid separation.

Immunohistochemistry

Formalin-fixed paraffin-embedded sections were used for IHC as described previously (14). Primary antibody was used against cleaved caspase-3 (1:1,000; Cell Signaling Technology) and BMI-1 (1:500; Cell Signaling Technology) in 2% goat serum in TBST overnight at 4°C. Tissue was incubated with biotinylated anti-rabbit secondary antibody (1:500), followed by signal-amplification using ABC Kit (Vector Laboratories). Color was developed using DAB (Vector Laboratories) and counterstained with Harris Hematoxylin (Sigma). Finally, mounting was performed with Permount (Fisher Scientific), and images were taken using Nikon eclipse 80i microscope. Hematoxylin and eosin staining was performed using hematoxylin-1 and eosin-Y (Thermo Scientific).

Soft-agar assay

Assay protocol was primarily followed as previously described (5). Briefly, neurospheres were dissociated, and 2,000 single cells were seeded per well in culture medium containing 0.24% Noble agar (Difco) over a solidified 0.4% agar layer. Cells were seeded in triplicates with DMSO or PTC596, or after drug wash-off following a pretreatment. In case of experiments with radiation, doses from 2 to 8 Gy were delivered after solidification of the top agar layer with cells. After 12 days, colonies were stained with 0.05% crystal violet solution (Sigma) for an hour, and blue colonies were quantified.

Mouse DIPG xenograft model

NBSGW mice (NOD.Cg-KitW-41J Tyr+Prkdcscid Il2rgtm1Wjl/ThomJ; 6–7 weeks old) were subcutaneously injected in the flank with 200,000 CCHMC-DIPG-1 cells per mouse. Mice were weighed and distributed in two groups (control and PTC596). After tumor establishment (day 10 after implantation) with an average volume of 50 to 100 mm3, mice were administered with PTC596 (12.5 mg/Kg) twice a week by oral gavage for 4 weeks. Tumors were measured by slide calipers taking two longest tumor-diameters (length and width) perpendicular to each other, and volumes were calculated by using the formula: (π/6) × d3, where d = mean diameter. All animal procedures were conducted according to our Institutional Animal Care and Use Committee protocol (#IACUC2018-0071, CCHMC).

Statistical analyses

Data from at least two independent experiments with individual technical replicates wherever applicable were collected. Representative images or blots are shown. Results are shown as mean ± SD. GraphPad Prism 8.0.1 was used to perform statistical analysis. One- or two-way ANOVA followed by a post-hoc Dunnet's or Tukey test, wherever applicable, was used to analyze the data. The log-rank test was used for the in vivo survival analysis. RNA-seq data analyses were performed using R3.6.1. The difference between tumor and normal tissue scores was tested by Wilcoxon Rank Sum test. The associations between continuous-scale data were calculated by Pearson correlation coefficient. Two-group comparisons or categorical association were analyzed by two-sided Fisher exact test. P < 0.05 was considered significant.

Spatial variability of BMI-1 expression in DIPG tissue specimens

We have previously shown that BMI-1 is highly expressed in 4 of 5 DIPG tumor tissue samples and 5 of 5 patient-derived primary DIPG stem-like cells lines evaluated (5). We expanded our evaluation of BMI-1 level in a larger cohort of post autopsy DIPG tissue samples (n = 26). Surprisingly, BMI-1 was not consistently upregulated in all DIPG samples analyzed at both RNA and protein levels (Fig. 1A and B). BMI-1 is a stem cell factor in both normal and malignant cells (16–18). We hypothesized that the observed lack of uniformity in BMI-1 upregulation might be the result of a variation in stemness levels in individual DIPG samples collected. DIPGs are heterogeneous and highly infiltrative tumors characterized by intratumoral histopathologic variation (7, 19). Indeed, IHC staining of BMI-1 showed intratumoral variability of BMI-1 levels in several DIPG tumor tissue samples evaluated (Supplementary Fig. S1A). More importantly, this differential BMI-1 expression was observed in tumors with both high and low BMI-1 expression values (log2 fold change compared with matched normal) calculated from bulk RNA-seq (Supplementary Table S1).

Figure 1.

Spatial variability of BMI-1 expression in DIPG tissue specimens. A,BMI-1 expression in patient tumors and their respective matched normal (CCHMC PBTR cohort). B, Immunoblot analysis of BMI-1 in patient DIPG tumors (T) and their matched normal tissue (N). Band intensities were quantified, normalized to β-actin, and represented as values relative to respective matched control. C,Bmi-1 expression in murine DIPG tumors and control brain stem. D, Immunoblot analysis of Bmi-1 levels in cell lines derived from mouse DIPG tumors and normal neural stem cells. Band intensities were quantified, normalized to β-actin, and represented as values relative to mouse neural stem cells (NSC). E, Glinsky score and its correlation with Bmi-1 expression in murine DIPG tumors and control brain stem (F). G, Stemness score and its correlation with Glinsky score in murine DIPG and control brain stem (H). I, Glinsky score and its correlation with stemness score in patient tumors and matched normal (CCHMC-PBTR cohort, J). K, scRNA-seq profile of DIPG patient tumors (Filbin-cohort) represented using stemness (y axis) and lineage score (x axis) and scored (colored dots) for cell-cycle program, (L) BMI-1 expression, and (M) Glinsky score (> median). R (Pearson's coefficient) and P values are indicated.

Figure 1.

Spatial variability of BMI-1 expression in DIPG tissue specimens. A,BMI-1 expression in patient tumors and their respective matched normal (CCHMC PBTR cohort). B, Immunoblot analysis of BMI-1 in patient DIPG tumors (T) and their matched normal tissue (N). Band intensities were quantified, normalized to β-actin, and represented as values relative to respective matched control. C,Bmi-1 expression in murine DIPG tumors and control brain stem. D, Immunoblot analysis of Bmi-1 levels in cell lines derived from mouse DIPG tumors and normal neural stem cells. Band intensities were quantified, normalized to β-actin, and represented as values relative to mouse neural stem cells (NSC). E, Glinsky score and its correlation with Bmi-1 expression in murine DIPG tumors and control brain stem (F). G, Stemness score and its correlation with Glinsky score in murine DIPG and control brain stem (H). I, Glinsky score and its correlation with stemness score in patient tumors and matched normal (CCHMC-PBTR cohort, J). K, scRNA-seq profile of DIPG patient tumors (Filbin-cohort) represented using stemness (y axis) and lineage score (x axis) and scored (colored dots) for cell-cycle program, (L) BMI-1 expression, and (M) Glinsky score (> median). R (Pearson's coefficient) and P values are indicated.

Close modal

Furthermore, reanalyzing previously published scRNA-seq data (Filbin cohort), ref. (10) also showed differential BMI-1 expression within tumor samples (Supplementary Fig. S1B and S1C). This suggests that traditional bulk RNA-seq of DIPG tumor tissue is not representative of the tumor transcriptomic heterogeneity, particularly in DIPG with documented intratumoral clonal diversity (7, 20).

The intertumor heterogeneity of BMI-1 expression is probably due to the variability in tumor stem cell content in the DIPG samples collected. BMI-1–driven gene expression signature also called the Glinsky score was shown to be uniformly predictive of metastasis and therapy failure in several cancer types including pediatric brain tumors (12, 21). We, therefore, sought to evaluate the correlation between Glinsky score, BMI-1 expression, and stemness score in DIPG tissue samples collected from patients. We validated this approach, in four in utero electroporation models of murine DIPG, harboring (i) PDGFB overexpression and (ii) PdgfraD842V expression, combined with dominant-negative Trp53 (DNP53) and H3.3 WT or H3.3K27M mutation (11). These murine DIPG tumors showed increased expression of BMI-1 at the mRNA (P = 0.0055) and protein levels compared with normal brainstem tissue (Fig. 1C and D). The Glinsky scores were significantly higher than control brainstems (P = 0.0055), which, as expected, positively correlated with Bmi-1 expression (Pearson's correlation coefficient, R = 0.84, P = 0.00017) for both H3.3 WT and K27M tumors (Fig. 1E and F). Because the Glinsky signature represents a stem-cell like expression profile (12), we calculated the stemness score as defined in (10), which was also significantly higher in these tumors compared with their normal brain stem counterpart (P = 0.0055) and positively correlated with the Glinsky score (R = 0.8, P = 0.00055; Fig. 1G and H) and with Bmi-1 expression as well (R = 0.71, P = 0.0049; Supplementary Fig. S1D) in both H3.3 WT and K27M tumors. Interestingly, unlike in human tumors (Supplementary Fig. S1A), BMI-1 expression was relatively homogeneous in murine DIPG tumors (data not shown).

Using this approach in our DIPG patient cohort comprising all H3K27 subtypes (CCHMC-DIPG cohort), we demonstrate that these DIPG tumors have significantly higher Glinsky scores compared with their respective matched normal tissue (P < 0.0001) and positively correlate with their stemness scores (R = 0.56, P = <0.0001; Fig. 1I and J). Finally, our analysis of the Filbin cohort scRNA-seq data indicates that BMI-1 is primarily expressed in undifferentiated cells resembling the cycling stem cell population (Fig. 1K and L). Interestingly, this population of cells was also enriched with high Glinsky signature (>50th percentile), further validating the use of the Glinsky score (Fig. 1M).

Together, these data indicate that the bulk evaluation of BMI-1 expression in a given DIPG tumor sample might not be an accurate representation of BMI-1 expression in the whole tumor and that BMI-1, although differentially expressed within a DIPG tumor, is enriched in the stem cell compartment. Moreover, the positive correlation between BMI-1 mRNA expression, stemness, and the Glinsky score suggests a possible role of BMI-1 in DIPG tumorigenesis.

BMI-1 modulator PTC596 is a potent inhibitor of DIPG cell growth

We have previously shown that PTC-209, an investigational compound that targets the posttranscriptional regulation of BMI-1 leading to decreased BMI-1 protein levels, induces G1–S cell-cycle arrest, limited apoptosis, and DIPG cell growth inhibition (5). Similarly, BMI-1 knockdown also resulted in the inhibition of DIPG cell growth (data not shown). Here, we evaluated the mechanism of action and therapeutic efficacy of a novel small molecule PTC596 (PTC Therapeutics) in DIPG. Unlike PTC209, PTC596 modulates BMI-1 by posttranslational phosphorylation ultimately leading to protein level reduction (6). Regardless of H3.1/H3.3 mutation status, PTC596 selectively inhibited the growth of all tested DIPG cells with IC50 ranging from 23 to 37 nmol/L compared with 85 nmol/L in normal primary HFF (Fig. 2A; Supplementary Table S2). Using the soft-agar colony formation assay, treatment of DIPG cells with 60 nmol/L of PTC596 completely abolished the colony-forming ability of all DIPG cells tested (Fig. 2B). These results indicate that PTC596 is a more potent inhibitor of DIPG cell growth than PTC209 with IC50 in the nanomolar range.

Figure 2.

BMI-1 modulator, PTC596 selectively inhibits DIPG cell growth regardless of histone subtype. DIPG cells treated with PTC596 were assayed for (A) cell growth after 72 hours using WST-1 assay and (B) colony formation ability using soft-agar assay. C, Immunoblot analysis of BMI-1 and H2AK119Ub following 24-hour treatment with the indicated concentrations of PTC596 in DIPG cells. β-Actin and total H2A served as loading controls. Arrows denote phosphorylated BMI-1 (BMI-1-P). For H2AK119Ub, band intensities were quantified, normalized to their respective loading controls, and represented as values relative to DMSO. For BMI-1, modified, unmodified, and total BMI-1 (modified and unmodified) bands were quantified and represented as percentage of total BMI-1. D, Immunoblot analysis of BMI-1 in CCHMC-DIPG-1 and SU-DIPG-IV cells treated with PTC596 (100 nmol/L for 24 hours) with or without λ-Phosphatase treatment. β-Actin served as loading controls. Arrow denotes BMI-1-P. Error bars represent SD from two independent run in triplicates.

Figure 2.

BMI-1 modulator, PTC596 selectively inhibits DIPG cell growth regardless of histone subtype. DIPG cells treated with PTC596 were assayed for (A) cell growth after 72 hours using WST-1 assay and (B) colony formation ability using soft-agar assay. C, Immunoblot analysis of BMI-1 and H2AK119Ub following 24-hour treatment with the indicated concentrations of PTC596 in DIPG cells. β-Actin and total H2A served as loading controls. Arrows denote phosphorylated BMI-1 (BMI-1-P). For H2AK119Ub, band intensities were quantified, normalized to their respective loading controls, and represented as values relative to DMSO. For BMI-1, modified, unmodified, and total BMI-1 (modified and unmodified) bands were quantified and represented as percentage of total BMI-1. D, Immunoblot analysis of BMI-1 in CCHMC-DIPG-1 and SU-DIPG-IV cells treated with PTC596 (100 nmol/L for 24 hours) with or without λ-Phosphatase treatment. β-Actin served as loading controls. Arrow denotes BMI-1-P. Error bars represent SD from two independent run in triplicates.

Close modal

It has previously been shown that PTC596 induces BMI-1 hyperphosphorylation, protein depletion, and loss of function (6, 22). It remains unclear whether the BMI-1 loss of function, as determined by the levels of monoubiquitination of histone H2A at lysine 119 (H2AK119Ub), is caused by the posttranslational modification or BMI-1 protein depletion. Treatment of all three histone subtypes of DIPG cells for 24 hours with various concentrations of PTC596 led first to dose-dependent increase in BMI-1 phosphorylation concurrent with a reduction in H2AK119Ub levels before any apparent decrease in BMI-1 protein levels (Fig. 2C and D). In contrast to previous studies (6), these results indicate that PTC596-induced BMI-1 phosphorylation is the primary cause of the inhibition of BMI-1 canonical function.

Of note, we have previously shown that total BMI-1 protein levels are variable but high in all DIPG patient-derived cell lines tested compared with normal astrocytes, irrespective of H3.1/3K27 mutation status (5). Interestingly, the extent of the increase in total BMI-1 protein levels is cell line specific and DIPG subtype-independent.

PTC596-induced BMI-1 modulation is associated with M phase accumulation, DNA damage, and apoptosis

We have previously shown that PTC-209 has essentially cytostatic effect on DIPG cells (5). Because PTC596 seemed to affect BMI-1 function through a different mechanism, we sought to delineate the mode of action of PTC596 in DIPG cells. Treatment of DIPG cells with 60 nmol/L of PTC596 for 72 hours led to cell growth inhibition (Supplementary Fig. S2A) and apoptosis which was sustained 6 days after drug removal (Fig. 3A). Importantly, this treatment minimally affected the growth of normal primary human fibroblasts (HFF) and did not induce apoptosis (Fig. 3A; Supplementary Fig. S2A). These results indicate that although both PTC-209 and PTC596 repress PRC1-dependent function of BMI-1 and inhibit DIPG cell growth, in contrast to PTC-209, PTC596 induces BMI-1 hyperphosphorylation and sustained apoptosis at nanomolar concentrations in DIPG cells, but not in primary human normal cells.

Figure 3.

PTC596 induces γH2AX foci formation and apoptosis in M phase. A, Scheme representing the experiment design of PTC596 treatment, drug wash-off, and collection time points to show cleaved caspase-3 expression (green) in HFF, CCHMC-DIPG-1, and CCHMC-DIPG-2 cells and their respective quantification. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). DAPI (blue) represents nuclei. B, Immunoblot analysis of BMI-1, H2AK119Ub, and γH2AX in DIPG cells treated with the indicated doses of PTC596 for 48 hours. C, Representative IF images and cell-cycle analysis (with the indicated % of cells) of CCHMC-DIPG-1 cells treated with PTC596 (100 nmol/L) for 24 and 48 hours. DNA damage (γH2AX in green) and apoptosis (cleaved caspase-3 in red) were evaluated and their respective quantification. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). DAPI (blue) represents nuclei. D, Immunoblot analysis of Cyclin B1, H3 S10-P, BMI-1, H2AK119Ub, γH2AX, and cleaved caspase-3 in CCHMC-DIPG-1, SU-DIPG-IV cells treated with PTC596 (100 nmol/L) for the indicated time points. β-Actin and total H2A served as loading controls. Arrow indicates BMI-1-P. E, Immunoblot analysis of BMI-1 and H2AK119Ub in the cytoplasmic (Cyto.) and nuclear (Nucl.) fractions of CCHMC-DIPG-1 cells treated with PTC596 (100 nmol/L for 24 hours). β-Actin and Total H3 are the loading controls. Arrow indicates modified BMI-1. Band intensities of modified (BMI-1-P), unmodified (BMI1), total BMI-1 (modified and unmodified), and H2AK119Ub were quantified and represented as bar graphs. Error bars represent the SD from two independent experiments. F, Representative IF images of CCHMC-DIPG-1 cells showing DAPI (blue), Nestin (red), and BMI-1 (green), treated with DMSO or PTC596. White arrows indicate cytoplasmic localization of BMI-1. The percentage of cells with cytoplasmic BMI-1 is quantified. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). G, Scheme representing the experiment design to evaluate γH2AX kinetics in CCHMC-DIPG-1 cells treated with DMSO, DMSO + IR, PTC596, and PTC596 + IR (top plot). Quantification of γH2AX foci/cell in DMSO or PTC596 treated cells with IR for the indicated time points (bottom plot). Two independent experiments were performed, and data from one experiment are shown as a representative result. H, Representative IF images of γH2AX (green) in CCHMC-DIPG-1 cells treated with DMSO or PTC596 with IR at the indicated time points. DAPI (blue) represents nuclei. WB, Western blot. *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001.

Figure 3.

PTC596 induces γH2AX foci formation and apoptosis in M phase. A, Scheme representing the experiment design of PTC596 treatment, drug wash-off, and collection time points to show cleaved caspase-3 expression (green) in HFF, CCHMC-DIPG-1, and CCHMC-DIPG-2 cells and their respective quantification. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). DAPI (blue) represents nuclei. B, Immunoblot analysis of BMI-1, H2AK119Ub, and γH2AX in DIPG cells treated with the indicated doses of PTC596 for 48 hours. C, Representative IF images and cell-cycle analysis (with the indicated % of cells) of CCHMC-DIPG-1 cells treated with PTC596 (100 nmol/L) for 24 and 48 hours. DNA damage (γH2AX in green) and apoptosis (cleaved caspase-3 in red) were evaluated and their respective quantification. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). DAPI (blue) represents nuclei. D, Immunoblot analysis of Cyclin B1, H3 S10-P, BMI-1, H2AK119Ub, γH2AX, and cleaved caspase-3 in CCHMC-DIPG-1, SU-DIPG-IV cells treated with PTC596 (100 nmol/L) for the indicated time points. β-Actin and total H2A served as loading controls. Arrow indicates BMI-1-P. E, Immunoblot analysis of BMI-1 and H2AK119Ub in the cytoplasmic (Cyto.) and nuclear (Nucl.) fractions of CCHMC-DIPG-1 cells treated with PTC596 (100 nmol/L for 24 hours). β-Actin and Total H3 are the loading controls. Arrow indicates modified BMI-1. Band intensities of modified (BMI-1-P), unmodified (BMI1), total BMI-1 (modified and unmodified), and H2AK119Ub were quantified and represented as bar graphs. Error bars represent the SD from two independent experiments. F, Representative IF images of CCHMC-DIPG-1 cells showing DAPI (blue), Nestin (red), and BMI-1 (green), treated with DMSO or PTC596. White arrows indicate cytoplasmic localization of BMI-1. The percentage of cells with cytoplasmic BMI-1 is quantified. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). G, Scheme representing the experiment design to evaluate γH2AX kinetics in CCHMC-DIPG-1 cells treated with DMSO, DMSO + IR, PTC596, and PTC596 + IR (top plot). Quantification of γH2AX foci/cell in DMSO or PTC596 treated cells with IR for the indicated time points (bottom plot). Two independent experiments were performed, and data from one experiment are shown as a representative result. H, Representative IF images of γH2AX (green) in CCHMC-DIPG-1 cells treated with DMSO or PTC596 with IR at the indicated time points. DAPI (blue) represents nuclei. WB, Western blot. *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001.

Close modal

Next, we investigated the mechanism of PTC596-induced apoptosis. Cellular survival depends on the integrity of the signaling pathways to repair DNA double-strand breaks (DSB) and restore genomic integrity. BMI-1 has been implicated in the regulation of DNA damage response (DDR) and in DNA DSB repair (23–26). We hypothesized that PTC596-induced apoptosis is caused by disruption of DDR and DNA damage repair triggered by BMI-1 inhibition. Treatment of DIPG cells with PTC596 led to a dose- and time-dependent manner increase in the levels of H2AX-S139 phosphorylation and γH2AX foci formation, indicative of DSBs and DDR activation (Fig. 3B and C; Supplementary Fig. S2B and S2C). Moreover, this increase correlated with a decrease in G1, an increase in G2–M and sub-G1 cell populations, and apoptosis. These effects were amplified over time (Fig. 3C; Supplementary Fig. S2C). Furthermore, 24-hour treatment with PTC596 led to a dose-dependent increase in H3 S10-P levels, indicating accumulation of cells in M phase (Supplementary Fig. S2D). To determine the early molecular events preceding apoptosis, we evaluated the accumulation of cyclin B1, H3 S10-P, cleaved caspase-3, BMI-1 phosphorylation, and H2AK119Ub at 1, 2, 4, 8, and 24 hours after treatment. We observed an accumulation of cyclin B1 as early as 1 to 2 hours and H3 S10-P at 4 hours, BMI-1 phosphorylation at 4 hours, and reduction in H2AK119Ub levels occurring at 24 hours in CCHMC-DIPG-1 and 4 hours in SU-DIPG-IV. We noted γH2AX accumulation and apoptosis at 24 hours and 8 hours in CCHMC-DIPG-1 and SU-DIPG-IV, respectively (Fig. 3D). These results suggest that PTC596 treatment leads to spindle assembly checkpoint (SAC) activation leading to M phase arrest, followed by inhibition of BMI-1 canonical function co-occurring with DNA damage and apoptosis (Supplementary Fig. S2E). We did not observe γH2AX accumulation and apoptosis in HFF cells treated in the same conditions (data not shown). Interestingly, treatment of DIPG cells with colchicine, which blocks mitotic cells in metaphase, also led as expected to M phase accumulation followed by BMI-1 phosphorylation and inhibition, γH2AX increase, and cell death (data not shown). These results suggest that M phase arrest is the trigger of BMI-1 phosphorylation which compromises its role in DDR leading to apoptosis in DIPG cells. It was previously shown that during M phase, BMI-1 dissociates from chromatin owing to its phosphorylation (27). However, the fate of BMI-1 following its phosphorylation in M phase was not reported. We evaluated the levels of unmodified and modified BMI-1 post-PTC596 treatment in the cytoplasmic and nuclear fractions. Compared with DMSO controls, the majority of modified BMI-1 was present in the cytoplasmic compartment (Fig. 3E). Furthermore, using IF to visualize BMI-1 cellular localization, we observed a significant increase in cells with BMI-1 localized in the cytoplasm (Fig. 3F).

Together, these findings imply that PTC596 induces SAC, thereby preventing mitotic exit followed by BMI-1 phosphorylation and translocation to the cytoplasm leading to the inhibition of its canonical function (H2AK119Ub levels) and its role in DNA damage sensing and resolution. This inhibition coincides with DNA damage and apoptosis in DIPG cells but not in normal cells, suggesting a survival role of BMI-1 in M phase in DIPG cells involving DDR.

These results prompted us to investigate the effect of BMI-1 inhibition in DDR kinetics. H2AX phosphorylation, leading to γH2AX, is an early step of DDR. γH2AX is a robust marker of DSBs that we used to monitor DDR and DSB repair. The extent of DNA damage and the kinetics of γH2AX foci formation and resolution were evaluated. Upon treatment of DIPG cells with PTC596 and 5 Gy of ionizing radiation (IR), both the appearance (at 10 minutes) and disappearance (at 24 hours) of γH2AX foci were delayed compared with the IR only–treated cells, indicative of delayed sensing and repair of IR-induced DNA damage in the presence of PTC596 (Fig. 3G and H). Moreover, the size of unresolved γH2AX foci (at 24 hours) in DIPG cells treated with PTC596 and IR was larger compared with those of IR only–treated cells (Fig. 3H), indicative of persistent unrepaired DNA damage.

PTC596 sensitizes DIPG cells to ionizing radiation

We have previously shown that the inhibition of BMI-1 function enhances sensitivity to radiomimetic drug Bleocin in DIPG (5). We evaluated the radiosensitization potential of PTC596 in DIPG cells, as radiotherapy remains the only standard therapy to treat DIPG, prolonging survival by only 2 to 3 months (28). We considered two approaches to assess the radiosensitization potential of PTC596 in DIPG cells. We evaluated the effect of continuous and discontinuous exposure to PTC596. Using two proliferation assays, WST-1 and soft-agar colony formation assay, pretreatment of DIPG neurospheres for 2 days with a sublethal dose of PTC596 (30 nmol/L) followed by drug withdrawal and IR sensitized all DIPG cells tested to 2 and 5 Gy, affecting cell proliferation and colony formation ability (Fig. 4A and B; Supplementary Fig. S3A). Of note, the effect of the combination with 2 Gy was similar to 4 Gy. Hence, reducing the IR dose by 50% produced the same effect on DIPG cells. These effects were augmented in the continuous-treatment setting (Supplementary Fig. S3B). The combination index was highly synergistic even at lower doses of IR (Fig. 4C). Together, these results provide a proof of concept that PTC596-induced BMI-1 dysfunction has the potential to act synergistically with IR to sensitize DIPG cells to radiotherapy, at least in part, by impeding the DDR and repair leading to apoptosis.

Figure 4.

PTC596 enhances radiosensitivity of DIPG cells. (A) Scheme representing the experiment design to assess the sensitization of DIPG cells to IR after a single treatment of PTC596 (30 nmol/L for 2 days) by evaluating cell growth by WST-1 assay and (B) colony formation by soft-agar assay with its respective quantification. C, Scheme representing the experiment design to assess the combination potency of IR and PTC596 in DIPG cells by evaluating the combination index (CI) using the CalcuSyn software. CI < 1 indicates synergism, CI  =  1 indicates additive effect, and CI > 1 indicates antagonism. Error bars represent SD from at least two independent experiments run in triplicates. WB, Western blot. *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001.

Figure 4.

PTC596 enhances radiosensitivity of DIPG cells. (A) Scheme representing the experiment design to assess the sensitization of DIPG cells to IR after a single treatment of PTC596 (30 nmol/L for 2 days) by evaluating cell growth by WST-1 assay and (B) colony formation by soft-agar assay with its respective quantification. C, Scheme representing the experiment design to assess the combination potency of IR and PTC596 in DIPG cells by evaluating the combination index (CI) using the CalcuSyn software. CI < 1 indicates synergism, CI  =  1 indicates additive effect, and CI > 1 indicates antagonism. Error bars represent SD from at least two independent experiments run in triplicates. WB, Western blot. *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001.

Close modal

Sensitivity of DIPG cells to PTC596 is independent of EZH2 activity

Next, we evaluated the dependency of the role of BMI-1 in DDR, DNA repair, and cell survival on PRC2 activity. Nearly 80% of DIPG tumors harbor H3K27M mutation in histone genes H3F3A and HIST1H3B encoding the replication-independent histone variant H3.3 and the canonical histone H3.1, respectively. Previous studies have shown that H3K27M mutations are disease drivers, and the presence of these mutations correlates with a global decrease in the levels of the repressive mark H3K27me3 (29–35). Trimethylation of H3K27 is catalyzed by H3K27-specific histone metyltransferase, enhancer of zeste homologue 2 (EZH2), a subunit of PRC2. The sequential histone modifications induced by PRC2 and PRC1 allow stable silencing of gene expression. Moreover, PRC2 has been implicated in DDR and DNA repair (36–38). Using a potent EZH2 inhibitor, EPZ011989 (39), alone or in combination with PTC596, our results indicate that EZH2 inhibition does not interfere with PTC596-induced decrease in H2AK119Ub levels, DNA damage induction, and apoptosis regardless of H3K27 subtype. (Fig. 5A–F; Supplementary Fig. S4). Together, these results indicate that EZH2 activity is dispensable for PTC596-induced DNA damage and apoptosis in DIPG cells. To our knowledge, this is the first report providing evidence supporting the effectiveness of PTC596 in all DIPG subtypes regardless of EZH2 activity.

Figure 5.

EZH2 activity is dispensable for PTC596-induced DNA damage and apoptosis. Scheme representing the experiment design to assess the effect of EZH2 inhibition on PTC596-induced (A) DNA damage and (D) apoptosis. Immunoblot analysis of H3K27me3, H2AK119Ub, and γH2AX (B) before (day 6) and (C) after adding PTC596 (day 8). Total H3 and H2A served as loading controls. E, Representative IF images of cleaved caspase-3 expression (red) at day 9 in CCHMC-DIPG-1 and SU-DIPG-IV cells. F, Quantification of the results in E. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001. DAPI (Blue) indicates nuclei. WB, Western blot.

Figure 5.

EZH2 activity is dispensable for PTC596-induced DNA damage and apoptosis. Scheme representing the experiment design to assess the effect of EZH2 inhibition on PTC596-induced (A) DNA damage and (D) apoptosis. Immunoblot analysis of H3K27me3, H2AK119Ub, and γH2AX (B) before (day 6) and (C) after adding PTC596 (day 8). Total H3 and H2A served as loading controls. E, Representative IF images of cleaved caspase-3 expression (red) at day 9 in CCHMC-DIPG-1 and SU-DIPG-IV cells. F, Quantification of the results in E. Error bars represent the SD from 4 different fields (n = 150–400 cells/field). *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001. DAPI (Blue) indicates nuclei. WB, Western blot.

Close modal

PTC596 induces chromosome scattering

In addition to γH2AX staining in metaphase of DIPG cells treated with PTC596, we noticed the presence of non-paired sister chromatids reminiscent of PSCS (Supplementary Fig. S2E). As shown above, shortly after treatment of DIPG cells with PTC596, cells accumulated in M phase and exhibited increased levels of cyclin B1, indicative of SAC induction and prolonged metaphase arrest (Fig. 3D). Extended metaphase arrest has been shown to cause cohesion fatigue, defined as uncoordinated loss of chromatid cohesion, resulting in asynchronous sister chromatid separation and chromosome scattering. Sustained SAC activation holds the cells in extended mitotic arrest leading to further cohesion loss, spindle defects, and cell death (15, 40). DIPG cells treated with 100 nmol/L of PTC596 showed a time-dependent increase in mitotic index as early as 1 to 2 hours after treatment (Fig. 6A). Tubulin staining of live cells after 1- and 2-hour treatment with 100 nmol/L of PTC596 indicates the presence of scattered chromosomes, failing to aggregate at the metaphase plate, with many cells showing multipolar spindles. These mitotic abnormalities were not common in DMSO-treated cells (Fig. 6B). Furthermore, when treated with 60 nmol/L PTC596 for an extended time, we observed a significant increase in percentage of cells with PSCS, consistent with previous observations (15, 40), indicating cohesion fatigue (Fig. 6C). Together, these data suggest that PTC596 treatment induces prolonged metaphase, cohesion loss, and chromosome scattering, leading to mitotic abnormalities, DNA damage, and cell death.

Figure 6.

PTC596 causes mitotic abnormalities and cell death in DIPG. A, Representative IF images of H3 S10-P (red) of CCHMC-DIPG-1 cells treated with PTC596 (100 nmol/L) at the indicated time points and its respective quantification. Error bars represent the SD from at least three different fields (n = 500–700 cells/field). *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001. B, Representative IF images of tubulin (green) in M phase SU-DIPG-IV cells treated with DMSO or PTC596 (100 nmol/L) at 1 and 2 hours. DAPI (blue) indicates chromosomes. White arrow heads indicate scattered chromosomes, and red arrow heads indicate multipolar spindle assemblies. Quantification of metaphase cells with scattered chromosomes. Error bars represent the SD from at least 5 different fields (n = 30–40 M phase cells). C, Representative images of G-band staining of metaphase spreads of CCHMC-DIPG-1 and SU-DIPG-IV cells treated with DMSO or PTC596 (60 nmol/L) at the indicated time points. In lay shows a representative paired sister chromatid. Quantification of the percentage of metaphase spreads with sister chromatid separation (n = 10–100 metaphase spreads were analyzed).

Figure 6.

PTC596 causes mitotic abnormalities and cell death in DIPG. A, Representative IF images of H3 S10-P (red) of CCHMC-DIPG-1 cells treated with PTC596 (100 nmol/L) at the indicated time points and its respective quantification. Error bars represent the SD from at least three different fields (n = 500–700 cells/field). *, P < 0.0332; **, P < 0.0021; ***, P < 0.0002; and ****, P < 0.0001. B, Representative IF images of tubulin (green) in M phase SU-DIPG-IV cells treated with DMSO or PTC596 (100 nmol/L) at 1 and 2 hours. DAPI (blue) indicates chromosomes. White arrow heads indicate scattered chromosomes, and red arrow heads indicate multipolar spindle assemblies. Quantification of metaphase cells with scattered chromosomes. Error bars represent the SD from at least 5 different fields (n = 30–40 M phase cells). C, Representative images of G-band staining of metaphase spreads of CCHMC-DIPG-1 and SU-DIPG-IV cells treated with DMSO or PTC596 (60 nmol/L) at the indicated time points. In lay shows a representative paired sister chromatid. Quantification of the percentage of metaphase spreads with sister chromatid separation (n = 10–100 metaphase spreads were analyzed).

Close modal

PTC-596 treatment delayed tumor growth kinetics in patient-derived DIPG xenografts by inducing in-tumor apoptosis

To evaluate the in vivo effect of PTC596, primary patient-derived DIPG cells, CCHMC-DIPG-1 were subcutaneously injected in the flank of NBSG-W mice. We previously observed that these cells form aggressive and fast growing tumors in mice (14). Tumors were established at 10 days after implantation reaching volumes ranging from 50 to 100 mm3 at which point PTC596 treatment (12.5 mg/kg twice a week) delivered by oral gavage was started. To monitor PTC596 toxicity in mice, we weighed both treated and untreated mice. We did not observe a noticeable weight difference between the two groups of mice, and no dehydration or clinical symptom of sickness was observed in the treated group, indicating that this PTC596 regimen is not toxic (Fig. 7A). Tumor growth kinetics showed the majority (4/6) of treated mice had a slower tumor growth rate compared with the untreated group, eventually leading to smaller tumor volumes compared with vehicle-treated mice (Fig. 7B and C). In order to evaluate the long-term effect of PTC596 on tumor growth, PTC596 treatment was discontinued after 4 weeks in a subset of mice (n = 3). In the absence of treatment, pretreated mice showed faster tumor growth kinetics (Fig. 7B). Nonetheless, when compared with vehicle-treated mice, the overall mean tumor growth kinetics (up to reaching the exclusion criteria of >1,500 mm3) in PTC596-treated group was still significantly delayed (Fig. 7D). Consequently, the mice showed significantly better survival even after treatment discontinuation (Supplementary Fig. S5). Mice treated with PTC596 had significantly increased in-tumor apoptotic bodies compared with vehicle-treated mice. However, the levels of apoptotic bodies returned to comparable levels to the control group after treatment discontinuation (Fig. 7E). We also observed increased in-tumor cleaved caspase-3 levels, indicative of apoptosis (Fig. 7F). Furthermore, in-tumor evaluation of PTC596 activity showed reduced BMI-1 and H2AK119Ub levels, demonstrating in-tumor activity of PTC596. Taken together, these results demonstrate the in vivo therapeutic potential of PTC596 to inhibit tumor growth by inducing apoptosis in DIPG xenografts.

Figure 7.

PTC-596 treatment delays tumor growth kinetics and induces in-tumor apoptosis in patient-derived xenograft model of DIPG. A, Average weight of mice treated with vehicle or PTC-596 (12.5 mg/kg, twice a week). Error bars, SD. B, Tumor growth kinetics of vehicle (black) and PTC-596 (gray) treated mice. Each line denotes tumor growth in each mouse. C, Average tumor volume (mm3) in vehicle or PTC-596 treated mice at 33 days after implantation (23 days after treatment). D, Average tumor growth kinetics of vehicle (black) and PTC-596 (gray) treated mice until reaching a tumor size of 1,500 mm3 (≥6 times of initial volume). E, Quantification of in-tumor apoptotic bodies in mice continuously treated with PTC596 (top plot) and after drug discontinuation (bottom plot) compared with vehicle treatment. F, Immunoblot analysis of in-tumor BMI-1, cleaved caspase-3, and H2AK119Ub from vehicle and PTC596-treated mice. β-Actin and Total H2A served as loading control. Number of mice per group (n) and the P values are indicated (*, P < 0.05).

Figure 7.

PTC-596 treatment delays tumor growth kinetics and induces in-tumor apoptosis in patient-derived xenograft model of DIPG. A, Average weight of mice treated with vehicle or PTC-596 (12.5 mg/kg, twice a week). Error bars, SD. B, Tumor growth kinetics of vehicle (black) and PTC-596 (gray) treated mice. Each line denotes tumor growth in each mouse. C, Average tumor volume (mm3) in vehicle or PTC-596 treated mice at 33 days after implantation (23 days after treatment). D, Average tumor growth kinetics of vehicle (black) and PTC-596 (gray) treated mice until reaching a tumor size of 1,500 mm3 (≥6 times of initial volume). E, Quantification of in-tumor apoptotic bodies in mice continuously treated with PTC596 (top plot) and after drug discontinuation (bottom plot) compared with vehicle treatment. F, Immunoblot analysis of in-tumor BMI-1, cleaved caspase-3, and H2AK119Ub from vehicle and PTC596-treated mice. β-Actin and Total H2A served as loading control. Number of mice per group (n) and the P values are indicated (*, P < 0.05).

Close modal

DIPG accounts for 10% to 15% of all new pediatric brain tumor diagnoses with a median overall survival of less than 1 year (2). Although radiotherapy does improve neurological function and extend survival by 2 to 3 months, no effective chemotherapeutic regimens are currently available. Hence, there is an urgent need for novel therapies to improve outcome in patients with this devastating tumor. We have previously shown that BMI-1 is a potential therapeutic target in DIPG (5), setting the stage for the present preclinical studies and the ongoing clinical testing of the BMI-1 inhibitor, PTC596, in DIPG (NCT03605550). Our results indicate that PTC596 is a potent inhibitor of DIPG cell growth and a promising radiosensitizer. The present studies demonstrate the vulnerability of DIPG cells to PTC596-induced mitotic abnormalities. We showed that the treatment of DIPG cells with PTC596 leads to an extended M phase arrest at metaphase causing cohesion fatigue and chromosome scattering as early as 2 hours after treatment. Our data support that PTC596 impairs mitotic progression and the induction of metaphase abnormalities characterized by premature chromatid separation, scattered chromosomes, failure to aggregate at the metaphase plate, and multipolar spindles. Persistent DNA damage and apoptosis were observed several hours after mitotic arrest suggesting that the acquisition of DNA damage and cell death are the consequence not the cause of the mitotic abnormalities. Importantly, we did not observe these events in normal human diploid cells.

Several studies have shown that abnormal mitosis is sufficient to generate DNA damage and cell death in M phase (reviewed in ref. 41). Prolonged mitosis can cause DNA damage through partial apoptotic response with limited DNase activation leading to DNA breaks (42). Induction of DNA breaks during mitosis leads to a partial activation of DDR by recruiting and activating early components of DDR such us γH2AX, thereby marking the damage sites for repair in the subsequent G1 (reviewed in ref. 43). However, DIPG cells did not survive mitotic arrest and were induced to die in M phase, probably caused by the extent of unrepaired DNA damage. In this regard, PTC596 can be viewed as an inhibitor of M phase progression inducing mitotic abnormalities that lead to extensive unrepaired DNA damage and cell death. Mitotic cells are significantly more radiosensitive compared with interphase cells (44, 45). Indeed, our results suggest that PTC596 treatment sensitizes cells to IR, probably through M phase arrest. Furthermore, our data indicate that the chromosome scattering occurred 2 hours after treatment with 100 nmol/L of PTC596 without affecting the spindle formation (Fig. 6B) contrasting with previously reported results suggesting that PTC596 inhibits microtubule polymerization in pancreatic ductal adenosarcoma (PDA; ref. 46). This discrepancy can be explained by the difference between DIPG and PDA biology. Our results indicate that DIPG cell survival requires high levels of BMI-1, as the pharmacologic and the genetic inhibition of BMI-1 protein production inhibits DIPG cell growth. However, BMI-1 deletion had no effect on cell viability of pancreatic tumor cells (46). Furthermore, the PTC596 doses used to inhibit PDA cell growth and tubulin polymerization (3–10 μmol/L) are far greater than the doses we used in DIPG cells (0.06–0.1 μmol/L). As indicated in the present studies, the IC50 of PTC596 in normal human diploid cells is 85 nmol/L, raising the concern of affecting the growth of normal human cells in pediatric patients with the use of high doses of PTC596. Lastly, we did not observe any striking effects on tubulin polymerization using the range of PTC concentrations outlined here.

Our results suggest that BMI-1 is involved after cohesion fatigue and chromosome scattering events, as the observed BMI-1 modification occurred after the induction of the observed mitotic abnormalities. We do not rule out the possibility of PTC596 causing mitotic abnormalities through an unknown mechanism involving BMI-1 phosphorylation, the SAC, and the anaphase promoting complex (APC/C). Because PTC596-induced phosphorylation/inactivation of BMI-1 precedes the acquisition of unrepaired DNA damage and apoptosis, hyperphosphorylated BMI-1 could inactivate its role in DDR. Phosphorylated BMI-1 was shown to be dissociated from chromatin during mitosis linking BMI-1 phosphorylation to M phase (27). Our data support this previous finding and indicate that BMI-1 translocates, mostly in its modified version, to the cytoplasm. These findings suggest that the phosphorylation of BMI-1 is inhibitory of its function in DDR, thereby leading to unrepaired DNA damage and cell death in DIPG cells. The exact mechanism of action of PTC596 and BMI-1 phosphorylation remains an area of active investigation. Previous studies suggested that PTC596 induces binding of CDK1 to BMI-1 and CDK1-mediated phosphorylation of BMI-1 at two sites (22). Future studies should determine the exact role of BMI-1 phosphorylation in M phase and identify the phosphorylation sites and the kinases responsible of BMI-1 phosphorylation.

In conclusion, PTC596-based monotherapy, although effective, may not be sufficient to treat patients with DIPG as we observed regrowth, once PTC596 treatment is completed or discontinued in our in vitro and in vivo models. Therefore, combination therapies including PTC596 should be considered.

M. Fouladi reports other from PTC Therapeutics (This work was initiated, designed and executed in the laboratory of Rachid Drissi which received unrestricted partial funding through a sponsored research agreement between Cincinnati Children's Hospital and PTC Therapeutics. No personal compensation or incentive was provided. PTC596 was provided by PTC Therapeutics.) during the conduct of the study; other from PTC Therapeutics (PTC Therapeutics is supporting the conduct of a phase I pediatric clinical trial for which Dr. Fouladi is the PI) outside the submitted work. R. Drissi reports non-financial support and other from PTC Therapeutics (This work was initiated, designed and executed in the laboratory of Rachid Drissi which received unrestricted partial funding through a sponsored research agreement between Cincinnati Children's Hospital and PTC Therapeutics. No personal compensation or incentive was provided. PTC596 was provided by PTC Therapeutics.) during the conduct of the study. No potential conflicts of interest were disclosed by the other authors.

S. Senthil Kumar: Formal analysis, investigation, methodology. S. Sengupta: Formal analysis, investigation, methodology. X. Zhu: Resources, formal analysis, methodology. D.K. Mishra: Resources, data curation, formal analysis, methodology. T. Phoenix: Resources. L. Dyer: Resources, formal analysis, methodology. C. Fuller: Resources, formal analysis. C.B. Stevenson: Resources. M. DeWire: Resources. M. Fouladi: Resources. R. Drissi: Data curation, formal analysis, investigation, methodology, writing-original draft.

This work was supported by CancerFree KIDS Pediatric Cancer Research Alliance to R. Drissi, Pray Hope and Believe Foundation to R. Drissi, PTC Therapeutics to R. Drissi, The Cure Starts Now Foundation to R. Drissi, and the Division of Oncology, and the Brain Tumor Center, Cincinnati Children's Hospital Medical Center, OH. We thank M. Monje (Stanford University) for kindly providing us with SU-DIPG-IV and XIII and XXI DIPG cell lines, and Audra Birri (Human Genetics, Cincinnati Children's Hospital Medical Center) and Sai Reddy (University of Cincinnati) for technical assistance. We thank PTC Therapeutics, Inc. and Epizyme, Inc. for providing PTC596 and EPZ011989, respectively.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Smith
MA
,
Seibel
NL
,
Altekruse
SF
,
Ries
LAG
,
Melbert
DL
,
O'Leary
M
, et al
Outcomes for children and adolescents with cancer: challenges for the twenty-first century
.
J Clin Oncol
2010
;
28
:
2625
34
.
2.
Cooney
T
,
Lane
A
,
Bartels
U
,
Bouffet
E
,
Goldman
S
,
Leary
SES
, et al
Contemporary survival endpoints: an international diffuse intrinsic pontine glioma registry study
.
Neuro Oncol
2017
;
19
:
1279
80
.
3.
Jacobs
JJ
,
Scheijen
B
,
Voncken
JW
,
Kieboom
K
,
Berns
A
,
van Lohuizen
M
. 
Bmi-1 collaborates with c-Myc in tumorigenesis by inhibiting c-Myc-induced apoptosis via INK4a/ARF
.
Genes Dev
1999
;
13
:
2678
90
.
4.
Bhattacharya
R
,
Mustafi
SB
,
Street
M
,
Dey
A
,
Dwivedi
SKD
. 
Bmi-1: at the crossroads of physiological and pathological biology
.
Genes Dis
2015
;
2
:
225
39
.
5.
Kumar
SS
,
Sengupta
S
,
Lee
K
,
Hura
N
,
Fuller
C
,
DeWire
M
, et al
BMI-1 is a potential therapeutic target in diffuse intrinsic pontine glioma
.
Oncotarget
2017
;
8
:
62962
75
.
6.
Nishida
Y
,
Maeda
A
,
Kim
MJ
,
Cao
L
,
Kubota
Y
,
Ishizawa
J
, et al
The novel BMI-1 inhibitor PTC596 downregulates MCL-1 and induces p53-independent mitochondrial apoptosis in acute myeloid leukemia progenitor cells
.
Blood Cancer J
2017
;
7
:
e527
.
7.
Hoffman
LM
,
DeWire
M
,
Ryall
S
,
Buczkowicz
P
,
Leach
J
,
Miles
L
, et al
Spatial genomic heterogeneity in diffuse intrinsic pontine and midline high-grade glioma: implications for diagnostic biopsy and targeted therapeutics
.
Acta Neuropathol Commun
2016
;
4
:
1
.
8.
Bray
NL
,
Pimentel
H
,
Melsted
P
,
Pachter
L
. 
Near-optimal probabilistic RNA-seq quantification
.
Nat Biotechnol
2016
;
34
:
525
7
.
9.
Love
MI
,
Huber
W
,
Anders
S
. 
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
2014
;
15
:
550
.
10.
Filbin
MG
,
Tirosh
I
,
Hovestadt
V
,
Shaw
ML
,
Escalante
LE
,
Mathewson
ND
, et al
Developmental and oncogenic programs in H3K27M gliomas dissected by single-cell RNA-seq
.
Science
2018
;
360
:
331
5
.
11.
Patel
SK
,
Hartley
RM
,
Wei
X
,
Furnish
R
,
Escobar-Riquelme
F
,
Bear
H
, et al
Generation of diffuse intrinsic pontine glioma mouse models by brainstem targeted in utero electroporation
.
Neuro Oncol
2020
;
22
:
381
92
.
12.
Glinsky
GV
,
Berezovska
O
,
Glinskii
AB
. 
Microarray analysis identifies a death-from-cancer signature predicting therapy failure in patients with multiple types of cancer
.
J Clin Invest
2005
;
115
:
1503
21
.
13.
Nabbi
A
,
Riabowol
K.
Rapid isolation of nuclei from cells in vitro
.
Cold Spring Harb Protoc 2015
: 
2015
:
769
72
.
14.
Sengupta
S
,
Sobo
M
,
Lee
K
,
Senthil Kumar
S
,
White
AR
,
Mender
I
, et al
Induced telomere damage to treat telomerase expressing therapy-resistant pediatric brain tumors
.
Mol Cancer Ther
2018
;
17
:
1504
14
.
15.
Daum
JR
,
Potapova
TA
,
Sivakumar
S
,
Daniel
JJ
,
Flynn
JN
,
Rankin
S
, et al
Cohesion fatigue induces chromatid separation in cells delayed at metaphase
.
Curr Biol
2011
;
21
:
1018
24
.
16.
Lessard
J
,
Sauvageau
G
. 
Bmi-1 determines the proliferative capacity of normal and leukaemic stem cells
.
Nature
2003
;
423
:
255
60
.
17.
Fasano
CA
,
Dimos
JT
,
Ivanova
NB
,
Lowry
N
,
Lemischka
IR
,
Temple
S
. 
shRNA knockdown of Bmi-1 reveals a critical role for p21-Rb pathway in NSC self-renewal during development
.
Cell Stem Cell
2007
;
1
:
87
99
.
18.
Molofsky
AV
,
He
S
,
Bydon
M
,
Morrison
SJ
,
Pardal
R
. 
Bmi-1 promotes neural stem cell self-renewal and neural development but not mouse growth and survival by repressing the p16Ink4a and p19 Arf senescence pathways
.
Genes Dev
2005
;
19
:
1432
7
.
19.
Buczkowicz
P
,
Bartels
U
,
Bouffet
E
,
Becher
O
,
Hawkins
C
. 
Histopathological spectrum of paediatric diffuse intrinsic pontine glioma: diagnostic and therapeutic implications
.
Acta Neuropathol
2014
;
128
:
573
81
.
20.
Vinci
M
,
Burford
A
,
Molinari
V
,
Kessler
K
,
Popov
S
,
Clarke
M
, et al
Functional diversity and cooperativity between subclonal populations of pediatric glioblastoma and diffuse intrinsic pontine glioma cells
.
Nat Med
2018
;
24
:
1204
15
.
21.
Bakhshinyan
D
,
Venugopal
C
,
Adile
AA
,
Garg
N
,
Manoranjan
B
,
Hallett
R
, et al
BMI1 is a therapeutic target in recurrent medulloblastoma
.
Oncogene
2019
;
38
:
1702
16
.
22.
Kim
MJ
,
Cao
L
,
Sheedy
J
,
Risher
N
,
Dumble
M
,
Lee
CS
, et al
Abstract 5517: PTC596-induced Bmi1 hyper-phosphorylation via Cdk1/2 activation resulting in tumor stem cell depletion
.
Cancer Res
2014
;
74
:
5517
.
23.
Liu
J
,
Cao
L
,
Chen
J
,
Song
S
,
Lee
IH
,
Quijano
C
, et al
Bmi1 regulates mitochondrial function and the DNA damage response pathway
.
Nature
2009
;
459
:
387
92
.
24.
Ginjala
V
,
Nacerddine
K
,
Kulkarni
A
,
Oza
J
,
Hill
SJ
,
Yao
M
, et al
BMI1 is recruited to DNA breaks and contributes to DNA damage-induced H2A ubiquitination and repair
.
Mol Cell Biol
2011
;
31
:
1972
82
.
25.
Ismail
IH
,
Andrin
C
,
McDonald
D
,
Hendzel
MJ
. 
BMI1-mediated histone ubiquitylation promotes DNA double-strand break repair
.
J Cell Biol
2010
;
191
:
45
60
.
26.
Facchino
S
,
Abdouh
M
,
Chatoo
W
,
Bernier
G
. 
BMI1 confers radioresistance to normal and cancerous neural stem cells through recruitment of the DNA damage response machinery
.
J Neurosci
2010
;
30
:
10096
111
.
27.
Voncken
JW
,
Schweizer
D
,
Aagaard
L
,
Sattler
L
,
Jantsch
MF
,
van Lohuizen
M
. 
Chromatin-association of the Polycomb group protein BMI1 is cell cycle-regulated and correlates with its phosphorylation status
.
J Cell Sci
1999
;
112
:
4627
39
.
28.
Hargrave
D
,
Bartels
U
,
Bouffet
E
. 
Diffuse brainstem glioma in children: critical review of clinical trials
.
Lancet Oncol
2006
;
7
:
241
8
.
29.
Schwartzentruber
J
,
Korshunov
A
,
Liu
XY
,
Jones
DTW
,
Pfaff
E
,
Jacob
K
, et al
Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma
.
Nature
2012
;
482
:
226
31
.
30.
Wu
G
,
Broniscer
A
,
McEachron
TA
,
Lu
C
,
Paugh
BS
,
Becksfort
J
, et al
Somatic histone H3 alterations in pediatric diffuse intrinsic pontine gliomas and non-brainstem glioblastomas
.
Nat Genet
2012
;
44
:
251
3
.
31.
Wu
G
,
Diaz
AK
,
Paugh
BS
,
Rankin
SL
,
Ju
B
,
Li
Y
, et al
The genomic landscape of diffuse intrinsic pontine glioma and pediatric non-brainstem high-grade glioma
.
Nat Genet
2014
;
46
:
444
50
.
32.
Fontebasso
AM
,
Papillon-Cavanagh
S
,
Schwartzentruber
J
,
Nikbakht
H
,
Gerges
N
,
Fiset
PO
, et al
Recurrent somatic mutations in ACVR1 in pediatric midline high-grade astrocytoma
.
Nat Genet
2014
;
46
:
462
6
.
33.
Taylor
KR
,
Mackay
A
,
Truffaux
N
,
Butterfield
YS
,
Morozova
O
,
Philippe
C
, et al
Recurrent activating ACVR1 mutations in diffuse intrinsic pontine glioma
.
Nat Genet
2014
;
46
:
457
61
.
34.
Buczkowicz
P
,
Hoeman
C
,
Rakopoulos
P
,
Pajovic
S
,
Letourneau
L
,
Dzamba
M
, et al
Genomic analysis of diffuse intrinsic pontine gliomas identifies three molecular subgroups and recurrent activating ACVR1 mutations
.
Nat Genet
2014
;
46
:
451
6
.
35.
Bender
S
,
Tang
Y
,
Lindroth
AM
,
Hovestadt
V
,
Jones
DTW
,
Kool
M
, et al
Reduced H3K27me3 and DNA hypomethylation are major drivers of gene expression in K27M mutant pediatric high-grade gliomas
.
Cancer Cell
2013
;
24
:
660
72
.
36.
Campbell
S
,
Ismail
IH
,
Young
LC
,
Poirier
GG
,
Hendzel
MJ
. 
Polycomb repressive complex 2 contributes to DNA double-strand break repair
.
Cell Cycle
2013
;
12
:
2675
83
.
37.
Zeidler
M
,
Varambally
S
,
Cao
Q
,
Chinnaiyan
AM
,
Ferguson
DO
,
Merajver
SD
, et al
The Polycomb group protein EZH2 impairs DNA repair in breast epithelial cells
.
Neoplasia
2005
;
7
:
1011
9
.
38.
Wu
Z
,
Lee
ST
,
Qiao
Y
,
Li
Z
,
Lee
PL
,
Lee
YJ
, et al
Polycomb protein EZH2 regulates cancer cell fate decision in response to DNA damage
.
Cell Death Differ
2011
;
18
:
1771
9
.
39.
Campbell
JE
,
Kuntz
KW
,
Knutson
SK
,
Warholic
NM
,
Keilhack
H
,
Wigle
TJ
, et al
EPZ011989, A potent, orally-available EZH2 inhibitor with robust in vivo activity
.
ACS Med Chem Lett
2015
;
6
:
491
5
.
40.
Stevens
D
,
Gassmann
R
,
Oegema
K
,
Desai
A
. 
Uncoordinated loss of chromatid cohesion is a common outcome of extended metaphase arrest
.
PLoS One
2011
;
6
:
e22969
.
41.
Ganem
NJ
,
Pellman
D
. 
Linking abnormal mitosis to the acquisition of DNA damage
.
J Cell Biol
2012
;
199
:
871
81
.
42.
Orth
JD
,
Loewer
A
,
Lahav
G
,
Mitchison
TJ
. 
Prolonged mitotic arrest triggers partial activation of apoptosis, resulting in DNA damage and p53 induction
.
Mol Biol Cell
2012
;
23
:
567
76
.
43.
Bakhoum
SF
,
Kabeche
L
,
Compton
DA
,
Powell
SN
,
Bastians
H
. 
Mitotic DNA damage response: at the crossroads of structural and numerical cancer chromosome instabilities
.
Trends Cancer
2017
;
3
:
225
34
.
44.
Giunta
S
,
Belotserkovskaya
R
,
Jackson
SP
. 
DNA damage signaling in response to double-strand breaks during mitosis
.
J Cell Biol
2010
;
190
:
197
207
.
45.
Stobbe
CC
,
Park
SJ
,
Chapman
JD
. 
The radiation hypersensitivity of cells at mitosis
.
Int J Radiat Biol
2002
;
78
:
1149
57
.
46.
Eberle-Singh
JA
,
Sagalovskiy
I
,
Maurer
HC
,
Sastra
SA
,
Palermo
CF
,
Decker
AR
, et al
Effective delivery of a microtubule polymerization inhibitor synergizes with standard regimens in models of pancreatic ductal adenocarcinoma
.
Clin Cancer Res
2019
;
25
:
5548
60
.

Supplementary data