Therapeutic targeting of transcriptional elongation in diffuse intrinsic pontine glioma

Abstract

Background

Diffuse intrinsic pontine glioma (DIPG) is associated with transcriptional dysregulation driven by H3K27 mutation. The super elongation complex (SEC) is required for transcriptional elongation through release of RNA polymerase II (Pol II). Inhibition of transcription elongation by SEC disruption can be an effective therapeutic strategy of H3K27M-mutant DIPG. Here, we tested the effect of pharmacological disruption of the SEC in H3K27M-mutant DIPG to advance understanding of the molecular mechanism and as a new therapeutic strategy for DIPG.

Methods

Short hairpin RNAs (shRNAs) were used to suppress the expression of AF4/FMR2 4 (AFF4), a central SEC component, in H3K27M-mutant DIPG cells. A peptidomimetic lead compound KL-1 was used to disrupt a functional component of SEC. Cell viability assay, colony formation assay, and apoptosis assay were utilized to analyze the effects of KL-1 treatment. RNA- and ChIP-sequencing were used to determine the effects of KL-1 on gene expression and chromatin occupancy. We treated mice bearing H3K27M-mutant DIPG patient-derived xenografts (PDXs) with KL-1. Intracranial tumor growth was monitored by bioluminescence image and therapeutic response was evaluated by animal survival.

Results

Depletion of AFF4 significantly reduced the cell growth of H3K27M-mutant DIPG. KL-1 increased genome-wide Pol II occupancy and suppressed transcription involving multiple cellular processes that promote cell proliferation and differentiation of DIPG. KL-1 treatment suppressed DIPG cell growth, increased apoptosis, and prolonged animal survival with H3K27M-mutant DIPG PDXs.

Conclusions

SEC disruption by KL-1 increased therapeutic benefit in vitro and in vivo, supporting a potential therapeutic activity of KL-1 in H3K27M-mutant DIPG.

Diffuse intrinsic pontine gliomas (DIPGs) are among the most devastating childhood tumors, with a median survival of 9-12 months from diagnosis. Investigations into curative therapies have been extremely challenging, and a number of clinical trials involving different combinations of chemotherapeutic agents commonly used in glioma treatment have demonstrated no response in DIPG.^1,2 Given these dire circumstances, the identification of molecular mechanisms and efficacious therapeutic agents is of high importance for improving treatment outcomes for DIPG patients. DIPG with histone gene mutations includes substitution of lysine 27 with methionine (K27M) in H3F3A or HIST1H3B/C gene encoding histone H3.3 or H3.1 protein, respectively.^3,4 These mutations reduce the global levels of H3K27 tri-methylation (H3K27me3), which is mediated by polycomb repressive complex 2 (PRC2) methyltransferase and is required for the growth of DIPG by repressing neuronal differentiation and function.^5,6 In addition, increased co-localization of H3K27 acetylation (H3K27ac) and RNA polymerase II (Pol II) is strongly associated with active transcription (eg, of PDGFRA, TOP3A, and ACVR1) in H3K27M-mutant DIPG.^7–9 H3K27ac binds to the bromo- and extra-terminal domain (BET) protein BRD4 in the nucleosome and activates transcription.¹⁰ We have previously shown that targeted inhibition of BRD4 activity in DIPG results in a significant delay of tumor progression and prolonged animal survival in orthotopic DIPG patient-derived xenograft (PDX) models.⁷

The dynamic regulation of transcription elongation by Pol II is important for gene expression during the development of most metazoans. Pol II pauses transcription elongation transiently at promoter-proximal regions, which are 20-120 nucleotides downstream of the transcription start site (TSS) in the majority of transcribed genes.¹¹ The paused Pol II can be phosphorylated in its C-terminal domain and is released by the activity of positive transcription elongation factor b (P-TEFb), allowing transcription to resume. P-TEFb is composed of cyclin-dependent kinase 9 (CDK9) and its regulatory partner cyclin T1 (CCNT1), a functional component of the super elongation complex (SEC).^12–15 The SEC is composed of various factors including AF4/FMR2 4 (AFF4), P-TEFb, and eleven-nineteen lysine-rich leukemia (ELL).^12,15,16 AFF4 is a central component and required for the assembly and enzymatic activity of SEC. AFF4 binds CCNT1 in the SEC, leading to transcription elongation in cancer.^16–18 Previous studies of leukemia demonstrated that knockdown of AFF4 reduces the gene expression associated with leukemogenesis.^13,16 Dysregulation of the elongation stage of transcription is implicated in solid tumor, suggesting that mechanistic understanding of transcription elongation is therapeutically relevant for cancer therapy.^11,19 Given that devastating diseases, including DIPG, appear to be diseases of uncontrolled transcription elongation, targeting inhibition of transcriptional elongation can be an effective therapeutic strategy for DIPG.

We have recently shown that the peptidomimetic lead compound KL-1 has potent anti-tumor activity for breast cancer in vitro and in vivo.²⁰ KL-1 disrupts the interaction between AFF4 and P-TEFb, resulting in impaired release of Pol II from promoter-proximal regions, and reduces active transcription elongation.²⁰ DIPG is associated with transcriptional dysregulation driven by the H3K27 mutation^21–23 and is sensitive to transcriptional disruption by inhibiting BRD4 activity.^7,22 Here, we tested the effect of KL-1 disruption of the SEC in H3K27M-mutant DIPG to advance understanding of the molecular mechanism and as a new therapeutic strategy for DIPG.

Materials and Methods

Tumor Cell Sources

The primary human DIPG cell line SF8628 (H3.3K27M DIPG) was obtained from the University of California, San Francisco (UCSF, San Francisco, CA) medical center in accord with an institutionally approved protocol. Establishment of SF8628 from a surgical specimen and tumor cell modification for expression of firefly luciferase for in vivo bioluminescence imaging (BLI) has been described.^7,24–26 The DIPG-007 (H3.3K27M DIPG) cell line was kindly provided by Dr. Angel Montero Carcaboso (Hospital Sant Joan de Déu, Barcelona, Spain). SU-DIPG4 and SU-DIPG36 (H3.1K27M DIPG) cell lines were kindly provided by Dr. Michelle Monje (Stanford University, Stanford, CA). Normal human astrocytes (NHA) and isogenic human astrocytes expressing wild-type (Astro-WT) or K27M H3F3A transgene (Astro-KM) have been previously described.^5,24 Cell line KNS-42, with H3F3A G34V mutation (substitution of glycine 34 with valine),^24,27 was obtained from the Japanese Collection of Bioresources and was established from a 16-year-old male.

Cell Viability Assays

For determination of cell viability effects of KL-1, tumor cells were seeded in 96-well plates, at 2000 cells per well, and cultured in the presence of 0-100 µM KL-1 for 24 and 72 h with triplicate or quadruplicate samples for each incubation condition. Relative numbers of viable cells were determined using CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega). The 50% growth inhibition (IC₅₀) values were calculated using nonlinear least-squares curve-fitting. Inhibitor proliferation effects were analyzed on days 1, 2, 3, 4, and 6 in the presence of vehicle (DMSO) or IC₅₀ concentrations of KL-1.

Colony Formation Assays

Colony formation assays were performed by plating at 2500 cells/ml in 60-mm plates. The cells were allowed to adhere for 24 h and then treated with DMSO or IC₅₀ values of KL-1. Cells were incubated at 37°C for 2 weeks and colonies were counted following staining with 0.05% crystal violet.

Apoptosis Assay

For analysis of apoptosis, the BD Pharmingen FITC Annexin V Apoptosis Detection Kit I (BD Biosciences, San Jose, CA) was used according to the manufacturer’s instructions. Briefly, SF8628 and DIPG-007 cells were seeded into 6-well plates. The following day, cells were incubated with DMSO and KL-1 (20 µM and 40 µM) for 72 h. Cells were harvested using trypsin, washed three times with phosphate-buffered saline (PBS), and stained with propidium iodide (PI) staining solution and FITC Annexin V. Stained samples were analyzed using flow cytometry (BD LSRFortessa Analyzer cytometer) and the percentage of apoptotic cells (Annexin V-positive and Annexin V + PI double-positive) were quantified using FlowJo software (version 10.5).

Xenograft Studies

Six-week-old female athymic mice (rnu/rnu genotype, BALB/c background) were purchased from Envigo (Indianapolis, IN) and housed under aseptic conditions. Pontine injection of tumor cells was performed as previously described.^7,24–26 Each mouse was injected with 1 µL cell suspension (100,000 cells/µL) into the pontine tegmentum 4.5 mm deep from the inner base of the skull. Animals were randomized into two treatment groups: (1) control (vehicle, n = 7) and (2) KL-1 therapy (intraperitoneal [IP] injection of 40 mg/kg KL-1 for 15 consecutive days, n = 9). Mice with intracranial tumor began receiving the treatment on day 16 when consecutive BLI indicates a logarithmic tumor growth in all mice. Mice were monitored daily and euthanized at the endpoint, which included irreversible neurological deficit or a body condition score <2. All animal protocols were approved by the Northwestern University Institutional Animal Care and Use Committee.

Immunohistochemistry (IHC)

Brains were collected from the mice 3 h after completion of the last treatment (n = 2 for each treatment). Paraformaldehyde-fixed brains were paraffin-embedded and sectioned (10 µm) for hematoxylin and eosin (HE) and anti-Ki67 antibody (2 µg/mL) (Ventana, Tucson, AZ) staining. To assay the apoptotic response to treatment, TUNEL staining was performed using the DeadEnd Colorimetric TUNEL System (Promega) according to the manufacturer’s protocol for paraffin-embedded tissues.

Analysis of Drug Concentration in the Brain

Athymic mice were administered 40 mg/kg of KL-1 for 7 days, with brains resected and serum collected following mouse euthanasia at 1 hour after the seventh administration (n = 2). The brainstem and frontal lobe were dissected from the surrounding brain, the serum was collected by cardiac puncture, and the samples were snap frozen and stored at −80°C. KL-1 was extracted from homogenized tissues using a Bullet Blender (Next Advance, Troy, NY). Homogenates were extracted with organic solvent and were transferred to an autosampler for liquid chromatography-mass spectrometry (LC/MS) analysis (Shimadzu VP Series 10 System) for determination of KL-1 content (Integrated Analytical Solutions, Inc., Berkeley, CA). Brain penetration ratio was calculated as KL-1 brainstem concentration divided by serum concentration.

Statistics

The Kaplan-Meier survival curves were plotted with GraphPad Prism 8 (GraphPad Software, San Diego, CA) and the P values were calculated using the log-rank test. The two-sided Kolmogorov-Smirnov test was performed for the empirical cumulative density function (ECDF) curves and the P values were provided in pausing index analysis. For all other analyses, a two-tailed unpaired t-test was applied using the Prism software.

The details of materials and other methods are described in the Supplementary Material.

Results

AFF4 Depletion Suppressed the Cell Growth of H3K27M-Mutant DIPG

AFF4 binds CCNT1 and CDK9, which compose active forms of P-TEFb promoting transcriptional elongation in cancer.^16–18 These key components of the SEC are detected in many cancers and associated with their pathogenesis.^13,16,28 To assess the status of SEC proteins in DIPG cells, we analyzed the expression level of AFF4, CCNT1, and CDK9 using western blotting. Expression of AFF4 and its binding proteins, CCNT1 and CDK9, were detected in all H3K27M-mutant DIPG cell lines (H3.3K27M: SF8628, DIPG-007, H3.1K27M: SU-DIPG4, SU-DIPG36) (Figure 1A). NHA, Astro-WT and -KM, and H3G34V-mutant KNS-42 cells are also expressed the SEC proteins. Disruption of the SEC by depletion of AFF4 reduces Pol II transcription elongation activity and inhibits cell proliferation in cancer.²⁰ To address whether AFF4 activity is required for H3K27M-mutant DIPG cell growth, we used shRNAs to suppress the expression of AFF4 in SF8628 (Figure 1B) and SU-DIPG36 DIPG cells (Supplementary Figure S1A). AFF4 suppression was confirmed at the protein level and the effect of AFF4 depletion on cell viability was analyzed using the MTS assay. shRNA-mediated depletion of AFF4 significantly reduced the cell growth of SF8628 (Figure 1B) and SU-DIPG36 (Supplementary Figure S1A) cells relative to control cells with scrambled shRNA (SF8628: P < .0001 [shAFF4 #1-3] at day 6, SU-DIPG36: P = .0028 [shAFF4 #1], P = .0031 [shAFF4 #2], P = .0017 [shAFF4 #3] on day 5), suggesting AFF4 is required to maintain DIPG cell proliferation. To further support the SEC complex dependency of DIPG cell growth, we used shRNAs to deplete CDK9, the other key component of SEC, in SF8628 and SU-DIPG 36 DIPG cells (Supplementary Materials and Methods). The shRNAs-mediated depletion of CDK9 also significantly inhibited the DIPG cell growth (SF8628: P = .0006 [shCDK9 #1], P = .0008 [shCDK9 #2], P = .0004 [shCDK9 #3] at day 6, SU-DIPG36: P = .0003 [shCDK9 #1], P = .0006 [shCDK9 #2], P = .0001 [shCDK9 #3] on day 5, Supplementary Figure S1B).

KL-1 Decreased Cell Growth and Increased Apoptosis in H3K27M-Mutant DIPG Cells

The compound KL-1 is known to disrupt the interaction between AFF4 and P-TEFb and reduce the active transcription elongation, resulting in inhibited cell growth of breast cancer.²⁰ To determine the effects of KL-1 on AFF4 expression and tumor cell growth, we treated H3K27M-mutant DIPG (SF8628, DIPG-007, SU-DIPG4, SU-DIPG36, Astro-KM), H3 wild-type astrocytes (Astro-WT, NHA), and H3G34V-mutant glioma (KNS-42) cells with KL-1 (Figure 1C, Supplementary Figure S2). KL-1 induced a dose-dependent inhibition of AFF4 expression (Supplementary Figure S2A) and inhibited H3K27M-mutant DIPG cell growth with IC₅₀ at 72 h at 13 µM of SF8628, 20 µM of DIPG-007, 19 µM of SU-DIPG4, 19 µM of SU-DIPG36, and 12 µM of Astro-KM (Supplementary Figure S2B). KL-1 also suppressed the growth of H3 wild-type astrocytes (Astro-WT), H3G34V-mutant (KNS-42) glioma cells, and NHA with IC₅₀ at 72 h at 18 µM, 16 µM, and 18 µM, respectively (Supplementary Figure S2B). The IC₅₀ value of KL-1 significantly reduced cell growth relative to DMSO in the H3K27M-mutant DIPG cells (SF8628: P < .0001, DIPG-007: P = .0004, SU-DIPG4: P < .0001, SU-DIPG36: P = .0002 on day 6) (Figure 1C). KL-1 also inhibited the clonal growth of DIPG cells (SF8628, DIPG-007, SU-DIPG4, and SU-DIPG36: P < .0001) (Figure 1D). We also addressed whether KL-1 disruption of the SEC increases apoptosis in DIPG cells. The effects of KL-1 on early and late apoptosis in DIPG cells were analyzed using the Annexin V apoptosis assay (Figure 1E). SF8628 and DIPG-007 were treated with 20 µM and 40 µM KL-1 or DMSO for 72 h. KL-1 treatment increased Annexin V-positive cells dose dependently (SF8628: KL-1 20 µM 14.82%, 40 µM 28.93% vs DMSO 7.97%, DIPG-007: KL-1 20 µM 22.45%, 40 µM 36.9% vs DMSO 6.47%).

For addressing the specificity of KL-1 for inhibiting AFF4, we studied the effects of depleting of AFF4 on tumor cell response to KL-1 using doxycycline-inducible shAFF4 in SF8628 cells (Supplementary Materials and Methods, Supplementary Figure S3A-C). Doxycycline induced shRNA-mediated depletion of AFF4 leads to dramatic reduction in the anti-proliferative effect of KL-1 (Supplementary Figure S3D, E). Re-induced AFF4 by removing doxycycline rescues the anti-proliferative effect of KL-1 (Supplementary Figure S3F). These results support that the anti-tumor activity of KL-1 is specifically involved in the inhibition of AFF4.

KL-1 Altered Gene Expression in H3K27M-Mutant DIPG Cells

We have previously shown that KL-1 treatment up-regulates genes related to DNA repair and apoptosis, and suppresses genes involved in RNA splicing, proliferation, and MYC in HEK293T cells.²⁰ To analyze the effects of KL-1 inhibition on DIPG transcriptomes, we performed RNA-seq on the samples from SF8628 cells treated with 20 µM and 50 µM KL-1 or DMSO for 6 and 24 h. Unsupervised principal component analysis revealed KL-1-treated group separated from DMSO-treated group (Figure 2A), suggesting KL-1 alters transcriptional states in H3K27M-mutant DIPG cells. Differentially expressed genes included 208 up-regulated and 168 down-regulated genes in response to KL-1 treatment in SF8628 cells (Figure 2B). Gene set enrichment analysis (GSEA) showed KL-1 treatment up-regulated the hypoxia and metabolic pathways (eg, NDRG1, GDF15, PTGS2, ENO2, and ALDOC), whereas it down-regulated cellular processes including cell cycle (eg, CDCA7, UHRF1, and CDK1), transcription (eg, HIST1H2BB and HIST1H3J), and DNA repair (eg, PCNA and CCNF) in SF8628 cells (Figure 2B-D).

KL-1 Altered Genome-Wide Pol II Occupancy and Transcription in H3K27M-Mutant DIPG Cells

P-TEFb is required for the release of Pol II from promoter-proximal pausing and KL-1 disrupts the interaction between AFF4 and P-TEFb, resulting in increased Pol II chromatin occupancy. We analyzed KL-1 effects on Pol II occupancy and transcription by Pol II ChIP-seq in three DIPG cell lines (SF8628, SU-DIPG4, SU-DIPG36) treated with 20 µM and 50 µM KL-1 (Figure 3, Supplementary Figures S4 and S5). Consistent with a previous study,²⁰ KL-1 increased Pol II occupancy at the promoter regions of the genes involving DNA repair and transcription in three DIPG cells (Figures 2C and 3A, Supplementary Figures S4A, D and S5A, D). Genes associated with increased Pol II promoter binding were further split into two gene clusters by their expression pattern—gene cluster 1 had decreased expression and cluster 2 had increased expression in KL-1-treated samples comparing to DMSO-treated samples (Figure 3A, B, Supplementary Figures S4A, B and S5A, B). Gene Ontology (GO) pathway enrichment analysis showed PDGFRB signaling and/or cell differentiation pathways were enriched among cluster 1 genes whereas apoptosis and metabolic pathways were enriched among cluster 2 genes (Figure 3C, Supplementary Figures S4C and S5C). We next performed pausing index analysis as previously described²⁰ and found KL-1 treatment significantly increases Pol II pausing comparing to DMSO treatment (SF8628: P < 2.2E-16, Figure 3D, SU-DIPG4: P < 2.2E-16, Supplementary Figure S4E, SU-DIPG36: 20 μM KL-1 P = .034, 50 μM KL-1 P = .02123 vs DMSO, Supplementary Figure S5E). Taken together, our analysis suggests KL-1 inhibition increases genome-wide Pol II occupancy, resulting in suppressed active transcriptional elongation for multiple cellular processes involved in cell cycle, DNA repair, transcription, differentiation, and development of DIPG cells (Figure 4).

KL-1 Inhibited Tumor Growth and Prolonged Survival of Mice Bearing DIPG Patient-Derived Xenografts

Based on the biological effects of KL-1 in vitro, we hypothesized that KL-1 treatment suppresses tumor growth and increases survival in mice with orthotopic DIPG PDXs. Prior to conducting the efficacy study, we addressed the brain distribution of KL-1 compounds by systemic administration. We administered KL-1 by IP injection to the mice that were euthanized 1 h following KL-1 administration. Their brains were immediately resected, the brainstem and frontal lobe were dissected from the surrounding brain, and the serum was collected by cardiac puncture. LC/MS analysis of tissue extracts revealed a detectable KL-1 concentration in the mice frontal lobe (2430.00 ± 127.28 ng/ml, 28.86 ± 3.42% of serum concentration) and brainstem (2810.00 ± 212.13 ng/ml, 33.31 ± 3.18% of serum concentration) (Table 1), thereby supporting KL-1 access to the brain including the brainstem, a site of DIPG development. To determine the anti-tumor activity of KL-1, the mice were implanted with SF8628 cells in the pons and were treated with 40 mg/kg of KL-1 for 15 consecutive days. KL-1 treatment significantly inhibited tumor growth (P = .0225, Figure 5A) and extend survival of the mice with SF8628 DIPG PDXs compared to the control (DMSO treatment) group (P < .0001, Figure 5B). This in vivo efficacy study included euthanizing the mice at the end of treatment to obtain brainstem tumor samples to analyze tumor cell proliferation (Ki-67) and apoptosis (TUNEL). IHC analysis of Ki-67 staining revealed KL-1 treatment significantly reduced Ki-67-positive cells (36.52 ± 5.62%) relative to DMSO treatment (50.22 ± 4.07%) (P = .0075, Figure 5C). TUNEL staining results showed a higher proportion of positive cells in tumors from mice receiving KL-1 (4.75 ± 1.50%) relative to DMSO control (0.75 ± 0.50%) (P = .0026, Figure 5C). No TUNEL positivity was evident in normal brain surrounding tumors in mice receiving any of the KL-1 treatments.

Discussion

The SEC is required for transcription elongation through release of Pol II. Disruption of the SEC by depletion of AFF4, a central component of the SEC, reduces Pol II transcription elongation activity thereby reducing cell proliferation in cancer.²⁰ In this study, shRNAs-mediated depletion of AFF4 significantly reduced the cell growth of H3K27M-mutant DIPG cells (SF8628, Figure 1B; SU-DIPG36, Supplementary Figure S1A), indicating AFF4 activity is required to maintain DIPG cell proliferation. The shRNAs-mediated depletion of CDK9, the other key component of SEC, also significantly inhibited the DIPG cell growth (Supplementary Figure S1B). Indeed, targeted transcriptional elongation using P-TEFb/CDK9 inhibitor has been tested in the patients with advanced hematologic malignancies (NCT02745743).

Given that dysregulated transcription is central to the biology of many types of malignancy including DIPG,^22,23 inhibiting transcription elongation by SEC disruption can be an effective therapeutic strategy for DIPG. The peptidomimetic lead compound KL-1 disrupts the interaction between AFF4 and P-TEFb in the SEC, resulting in impaired release of Pol II from promoter-proximal regions and reduced active transcription elongation.²⁰ Our RNA- and Pol II ChIP-seq analysis identified KL-1 increases Pol II occupancy at the promoter-proximal region of multiple gene loci in DIPG cells (SF8628, SU-DIPG4, and SU-DIPG36) and reduces active transcription elongation by increase of genome-wide Pol II pausing in multiple cellular processes that promote cell proliferation and differentiation of DIPG cells (Figures 2-4, Supplementary Figures S4 and S5). Indeed, KL-1 treatment reduced cell proliferation and increased apoptosis in DIPG cells (Figure 1C-E, Supplementary Figure S2B). It would be noted that there is no difference of KL-1 effects on cell growth between H3K27M-mutant and H3 wild-type or H3G34V-mutant glioma cells, and between the DIPG cells with and without MYC expression. This could be due to the widely expressing functional components of the SEC (AFF4, CCNT1, CDK9) in these cell lines (Figure 1A). We also found KL-1 up-regulated the genes related to apoptosis, hypoxia, and metabolic pathways in three DIPG cell lines (Figure 2B-D, Supplementary Figures S4C, D and S5C, D). While KL-1 treatment induces apoptotic cell death, tumor cells would alter metabolism under hypoxic conditions and acquire resistance to therapeutic agents.^29-30 Hypoxia-related genes also promote tumor survival by affecting cell division, angiogenesis, metabolism, and stem cell formation.³¹ These results suggest that the stress response of DIPG cells under KL-1 treatment may alter hypoxia and metabolic pathways to enable to evade apoptosis for tumor survival. Combinatorial approach to targeting hypoxia and metabolic pathways with Kl-1 treatment would be further investigated in DIPG.

We have recently shown that KL-1 had a potent anti-tumor activity in breast cancer.²⁰ In contrast to other solid tumors, a fundamental limitation of brain cancer therapy is that <1% of most therapeutic agents administered systemically cross the blood-brain barrier (BBB) to access the brain.³² In the present study, LC/MS analysis of the samples from the mice treated with KL-1 for 7 days by IP injection revealed the concentration of KL-1 in the brainstem at 2810.00 ± 212.13 ng/ml (Table 1), which can be equivalent to 8.1 µM based on the molecular weight of KL-1. Because the mice bearing SF8638 DIPG PDXs were treated with KL-1 for 15 days, we anticipate that brainstem concentration of KL-1 after 15-day treatment can be higher than 8.1 µM and that close to the IC₅₀ value of 13 µM. The brainstem concentration of KL-1 is 33.31 ± 3.18% of serum concentration (Table 1) which is similar to the brain concentration of temozolomide, a first-line chemotherapeutic agent for glioblastoma, with ~30% of the plasma concentration.³³ These results suggest that KL-1 transports to the brain with favorable concentrations in the brainstem to a useful extent based on its in vivo anti-tumor activity (Figure 5A).

Despite the anti-tumor activity, the median survival of mice treated with 40 mg/kg of KL-1 as maximum tolerated dose was extended less than 10 days relative to the control animals (Figure 5B). We anticipate that even KL-1 inhibitor transports to the brain, high doses are needed to achieve therapeutic levels, and nonspecific distribution to the normal brain, which can lead to substantial toxicity. Therefore, we would further investigate a local drug delivery system such as convection-enhanced delivery (CED),^34–36 which enable to deliver high concentration of drug into the tumor site, as a potential alternative strategy for enhancing drug targeting, reducing systemic toxicity, and potentially increasing the efficacy of KL-1 treatment in vivo and for next-level clinical use.

The histopathological analysis of phase II HERBY trial indicated that Ki-67 index (percentage of Ki-67-positive nuclei) was the only independent biomarker associated with overall survival (OS) as a hazard ratio (HR) of 3.77 (95% CI: 1.55-9.16; P = .003) between Ki-67 index high (≥20%) vs low (<20%), and 10% increase Ki-67 index was associated with poor outcome as an HR of 1.53 (95% CI: 1.27-1.83; P < .0001).³⁷ The univariate analysis of Children’s Cancer Group (CCG)-945 trial also demonstrated that high Ki-67 index (>36%) was associated with worse progression-free survival (PFS) and OS, both in the overall cohort^38,39 and in patients with midline high-grade gliomas.⁴⁰ Although our IHC analysis revealed Ki-67 positivity decreasing from 50.22 ± 4.07% to 36.52 ± 5.62% by KL-1 treatment, it would not be clinically significant based on the analysis above.

In conclusion, our results demonstrate favorable brainstem distribution and promising anti-tumor activity of KL-1 support further development of targeting transcription elongation as a potential therapy for DIPG. Of note, an independent group recently observed similarly that transcriptional elongation is a therapeutic target in DIPG.⁴¹

Funding

Funding for this work was provided by Bear Necessities Pediatric Cancer Foundation and Rally Foundation for Childhood Cancer Research (R.H.), St Baldrick’s Foundation (R.H.), Alex’s Lemonade Stand Foundation for Childhood Cancer (R.H.), John McNicholas Pediatric Brain Tumor Foundation (A.M.S., A.S., and R.H.), and the National Institutes of Health (R01NS093079 to R.H., K08NS097264 to A.M.S., R01CA197313 and R21NS114431 to O.J.B., and R01CA214035-15 and R35CA197569 to A.S.).

Acknowledgments

We thank Dr. Angel Montero Carcaboso (Hospital Sant Joan de Déu, Barcelona, Spain) for use of DIPG-007 cell line. We thank Dr. Michelle Monje (Stanford University) for use of the SU-DIPG-4 and -36 cell lines. Histology, BLI, shRNA, and next-generation sequencing services were provided by the Mouse Histology and Phenotyping Laboratory, Center for Advanced Microscopy/Nikon Imaging Center, Skin Biology and Disease Resource-based Center (SBDRC), and NUSeq Core at Northwestern University, respectively.

Conflict of interest statement. The authors have declared that no conflict of interest exists.

Authorship statement. H.K. and R.H. designed the study. H.K. performed the majority of the experiments and H.K. and R.H. wrote the manuscript. H.K., T.S., and R.H. performed and analyzed the in vivo experiments. H.K., N.T., and Y.A. performed ChIP-sequence analysis. E.J.R. generated and sequenced the next-generation sequencing (NGS) libraries. Y.Z. and L.Z. performed all bioinformatics analyses and provided interpretation of the data. H.K., G.T.B., and F.D.E. performed and analyzed the apoptosis assay and interpreted the data. H.K., Y.T., and T.S. performed and analyzed the immunohistochemistry studies. A.M.S. and S.G. provided clinical supervision in the interpretation of data. A.K., O.J.B., and A.S. provided supervision in the interpretation of data. All authors commented on the manuscript and approved the included data.

References