Abstract
Background
Unresectable hypothalamic/optic pathway pilocytic astrocytoma (PA) often progresses despite multiple therapies. Identifying clinical and molecular characteristics of progressive tumors may aid in prognostication and treatment.
Methods
We collected 72 unresectable, non-neurofibromatosis type 1-associated hypothalamic/optic pathway PA to identify clinical and biologic factors associated with tumor progression. Tumors that progressed after therapy, metastasized, or resulted in death were categorized into Cohort B; those that did not meet these criteria were categorized into Cohort A. DNA methylation and transcriptome analyses were performed on treatment-naïve tumors, and the findings were validated by immunohistochemistry (IHC).
Results
The median follow-up of the entire cohort was 12.3 years. Cohort B was associated with male sex (M:F = 2.6:1), younger age at diagnosis (median 3.2 years vs 6.7 years, P = .005), and high incidence of KIAA1549-BRAF fusion (81.5% vs 38.5%, P = .0032). Cohort B demonstrated decreased CpG methylation and increased RNA expression in mitochondrial genes and genes downstream of E2F and NKX2.3. Transcriptome analysis identified transcription factor TBX3 and protein kinase PIM1 as common downstream targets of E2F and NKX2.3. IHC confirmed increased expression of TBX3 and PIM1 in Cohort B tumors. Gene enrichment analysis identified enrichment of MYC targets and MAPK, PI3K/AKT/mTOR, and p53 pathways, as well as pathways related to mitochondrial function.
Conclusions
We identified risk factors associated with progressive PA. Our results support the model in which the p53-PIM1-MYC axis and TBX3 act alongside MAPK and PI3K/AKT/mTOR pathways to promote tumor progression, highlighting potential new targets for combination therapy and refining disease prognostication.
Keywords: MAPK, metabolism, PI3K/mTOR, progressive pilocytic astrocytoma, signaling pathway, transcription network
Key Points.
Progressive hypothalamic/optic pathway pilocytic astrocytoma is associated with male sex, young age at diagnosis, and KIAA1549-BRAF fusion.
The p53-PIM1-MYC axis and TBX3 act alongside MAPK and PI3K/AKT/mTOR pathways to promote tumor progression.
Our results highlight characteristics of progressive disease and suggest potential new targets for combination therapy that requires further validation in prospective clinical trials.
Importance of the Study.
Unresectable hypothalamic/optic pathway pilocytic astrocytoma (PA) often progresses despite multiple therapies. Identifying activated signaling pathways may aid in prognostication and therapeutics. We retrospectively identified 72 unresectable, non-neurofibromatosis type 1-associated hypothalamic/optic pathway PA. Tumors that progressed after therapy, developed metastatic disease, or resulted in death were categorized into Cohort B; those that did not meet these criteria were categorized into Cohort A. DNA methylation profiling and transcriptome analysis were performed on treatment-naïve tumors, and the findings were validated by immunohistochemistry. The median follow-up of the entire cohort was 12.3 years. Cohort B was associated with male sex, young age at diagnosis, and KIAA1549-BRAF fusion. Methylation and transcription data suggest that p53-PIM1-MYC axis and TBX3 act alongside MAPK and PI3K/AKT/mTOR pathways to promote tumor progression. Our results highlight potential new targets for combination therapy and refine disease prognostication.
Pilocytic astrocytoma (PA) is the most common glioma in children.1 Surgery is curative when the tumor can be completely resected. However, tumors involving the hypothalamus/optic pathway are challenging to manage because a gross total resection is not feasible without significant morbidity. Therefore, unresectable hypothalamic/optic pathway PA is often treated with chemotherapy or radiation (RT). Natural history is variable as some patients achieve long-term stability after one or 2 chemotherapy regimens, while others continue to progress after multiple lines of chemotherapy and/or RT. Progression is predominantly local, but 2%–6% of patients have leptomeningeal disease either at diagnosis or progression.2,3 The 5-year progression-free survival (PFS) after first-line chemotherapy is 39%–45%, and after RT is 70%.4–6 Although RT is associated with superior local control compared to chemotherapy, RT is avoided as first-line therapy in young children due to neurocognitive, endocrine, and vascular effects.7,8 More recently, MEK (MEK-I) and BRAF inhibitors have been used as alternative therapies and demonstrated overall response rates of 30%–40%.9 Despite the recognition that unresectable hypothalamic/optic pathway PA is a chronic disease that can progress after multiple lines of therapy, little is known about the biological basis of tumor progression.
We hypothesized that progressive hypothalamic/optic pathway PA harbors distinct transcription networks and activated signaling pathways that might be prognostic and serve as therapeutic targets. To avoid confounding factors associated with cancer predisposition syndromes, we collected 72 cases of unresectable, non-neurofibromatosis type 1 (NF1)-associated hypothalamic/optic pathway PA. Histology was reviewed and confirmed in all cases. The cohort was divided into 2 groups based on the tumor’s response to therapy and disease burden. Patients who met any of the following criteria were included in Cohort B: (1) progressed after RT, (2) progressed after ≥2 lines of systemic therapy, (3) developed metastatic disease, and (4) died of disease. Patients who did not meet any of these criteria were included in Cohort A. Utilizing this cohort, we sought to identify prognostic variables and activated transcription networks and signaling pathways associated with Cohort B through an integrated clinical, histopathological, DNA methylomic, and transcriptomic analysis. Our results showed that hypothalamic/optic pathway PA in Cohort B was associated with male sex (M:F = 2.6:1), younger age at diagnosis (median 3.2 years vs 6.7 years, P = .005), and high incidence of KIAA1549-BRAF fusion (81.5% vs 38.5%, P = .0032). In addition, Cohort B demonstrated decreased CpG methylation and increased RNA expression in mitochondrial genes and genes downstream of E2F and NKX2.3 transcription factors. Transcriptome analysis identified transcription factor TBX3 and proto-oncogene serine/threonine protein kinase PIM1 as common downstream targets of E2F and NKX2.3 and potential drivers of tumor progression. Gene enrichment analysis identified enrichment of MYC targets and signaling pathways known to be implicated in PA, such as MAPK and PI3K/AKT/mTOR, as well as pathways related to mitochondrial biogenesis and oxidative phosphorylation. These results provide new insights into the biological basis of tumor progression in hypothalamic/optic pathway PA, allowing prospective identification of progressive tumors and supporting the development of novel targeted therapeutics.
Materials and Methods
Study Design and Patient Population
Seventy-two patients treated at St. Jude Children’s Research Hospital (St. Jude) from 1987 to 2020 with unresectable and sporadic/non-NF1-associated hypothalamic/optic pathway PA were identified through retrospective review. All patients had to have undergone a biopsy prior to any treatment to be included in the study. Patient demographics, treatment details, and outcomes were extracted from medical records. The cohort was divided into two groups based on progression after therapy and disease burden. Patients who met any of the following criteria were included in Cohort B: (1) progressed after RT, (2) progressed after ≥2 lines of systemic therapy, (3) developed metastatic disease, and (4) died of disease. Patients who did not meet any of these criteria were included in Cohort A. Systemic therapy was referred to either chemotherapy or targeted therapy. This study has been approved by St. Jude’s Institutional Review Board (#19-0038).
Histopathology Review, Immunohistochemistry, Reverse Transcription-Polymerase Chain Reaction, and Fluorescence In Situ Hybridization
Hematoxylin and eosin stained 5 μm sections of formalin-fixed paraffin-embedded (FFPE) tissue specimens were centrally reviewed by a board-certified neuropathologist specialized in pediatric central nervous system tumors to confirm the diagnosis of PA. The sections were evaluated for cell density, mitotic activity (per 1.96mm2, average of 9.8mm2, rounded to the nearest integer), Ki-67 proliferation index (% among tumor cells), and the presence of necrosis, vascular proliferation, infiltration into the adjacent brain parenchyma, and pilomyxoid histology.
The following antibodies were used for immunohistochemistry (IHC) in a Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory10: Ki-67 (DAKO, #M7240, clone MIB-1, diluted 1:200), histone H3 K27M (RevMab Biosciences, #31-1175-00, clone RM192, diluted 1:250), H3 K27 trimethylation (Cell Signaling Technology, #9733, clone C36B11, diluted 1:200), p53 (Zeta Corp, #Z2029M, clone DO-7, diluted 1:200), ATRX (Sigma, #HPA001906, diluted 1:600), BRAF V600E (Ventana, #790-4855, pre-diluted), IDH1 R132H (Dianova, #DIA H09L, clone R132H, diluted 1:50), TBX3 (Abcam, #ab99302, diluted 1:100, positive control: Placental trophoblasts, negative control: Placental stroma), PIM1 (Invitrogen, #710504, clone 19HCLC, diluted 1:50, positive control: Testicular germ cells, negative control: Testicular stroma), Tom20 (Santa Cruz, #sc-17764, clone F-10, diluted 1:50, positive control: Hepatocytes, negative control: Hepatic stroma), pS6 (Cell Signaling, #4858, clone D57.2.2E, diluted 1:400, positive control: Splenic lymphocytes, negative control: Splenic stroma), and phospho-ERK1/2 (pERK1/2, Cell Signaling, #4370, diluted 1:200, positive control: Splenic lymphocytes, negative control: splenic stroma). Only nuclear TBX3 and pERK1/2 immunoreactivity were counted as positive. Cytoplasmic pS6 immunoreactivity was counted as positive. PIM1 and Tom20 stained sections were scored semi-quantitatively using the following criteria: 3+ represents complete and intense cytoplasmic staining, 2+ represents incomplete, weak to moderate cytoplasmic staining, 1+ represents incomplete, faint to barely perceptible cytoplasmic staining, and 0 represents no staining. The H-score was calculated by multiplying the percentage of tumor cells on the tissue section by their corresponding intensity score according to the following formula: (1 × [% cells 1+] + 2 × [% cells 2+] + 3 × [% cells 3+]).
Reverse transcription-polymerase chain reaction with primers sets to detect 9 reported variants of KIAA1549-BRAF fusion was performed in a CLIA-certified laboratory (information available upon request). Chromosome 7q34 duplication (a marker for KIAA1549-BRAF fusion), CDKN2A deletion, and MYB rearrangement were detected by fluorescence in situ hybridization (FISH) in a CLIA-certified laboratory with probes developed in-house (information available upon request).
Genome-Wide DNA Methylation Profiling and Data Analysis
Genomic DNA extracted from tumor samples was used for genome-wide methylation profiling and CNV (copy number variant) analysis by the Illumina Infinium MethylationEPIC platform, as previously described.10–12 For comparison, publicly available reference profiles of brain tumors were downloaded from the Genomic Data Commons Data Portal (https://portal.gdc.cancer.gov/). Analysis of methylation profiles, including normalization, filtering, t-distributed stochastic neighbor embedding (t-SNE), identification of differentially methylated probes and regions, and CNV analysis was performed in R 4.1.0 using ChAMP 2.22.0, bumphunter, minfi 1.38.0, limma 3.48.0, conumee 1.26.0, and Rtsne 0.15 packages in Bioconductor 3.13 (http://bioconductor.org/), as previously described.10–15 Genes within differentially methylated regions (DMRs) were queried against MSigDB v7.4 to identify functional or pathway enrichment, as previously described.12
RNA Sequencing and Transcriptome Analysis
PureLink FFPE Total RNA Isolation Kit (Thermo Fisher Scientific) was used for total RNA extraction from FFPE tissue. Purified RNA was quantified on a Qubit 3 Fluorometer (Thermo Fisher Scientific) using Qubit RNA BR Assay Kit (Thermo Fisher Scientific), as previously described.11,12 Total RNA sequencing was performed using the Illumina TruSeq Stranded Total RNA protocol with at least 500 ng of total RNA. Libraries were prepared using the TruSeq Stranded Total RNA Sample Prep Kit (Illumina). All sequencing data were generated after 100 cycles of paired-end runs on an Illumina HiSeq 2500 or HiSeq 4000. The RNA-seq data were aligned to the human reference genome (build GRCh38). Fusion detection was performed using CICERO16 and Arriba v2.1.017 with default parameters and manually reviewed. Single nucleotide variants (SNVs), insertion, and deletion were identified using Bambino18 and CTAT-Mutations Pipeline (https://github.com/NCIP/ctat-mutations/wiki). The variants were annotated and ranked by putative pathogenicity using either CTAT or “PeCanPIE” 19 and manually reviewed.
Differentially expressed genes were analyzed using edgeR and DESeq2 in Bioconductor 3.13 (http://bioconductor.org/), NetBID 2.0.3, and SJARACNe 0.2.0 with default parameters and queried against MSigDB, as previously described.12 Only annotated protein-coding genes were considered. The read counts of gene expression data were preprocessed, normalized, and filtered to remove genes with very low expression values (bottom 5%) in more than 90% of samples using edgeR 3.34.0 and DESeq2 in R 4.1.0. Bayesian inference approach and SJARACNe-based network reconstruction were used to infer drivers (transcription factors or signaling factors) from the transcriptomics data by calculating the activity of drivers and gene sets in Ubuntu 18.04 with Python 3.6.1. Differentially expressed genes and differentially activated drivers were queried against MSigDB v7.4 using Fast Gene Set Enrichment Analysis (fgsea, https://www.biorxiv.org/content/10.1101/060012v3) in Bioconductor 3.13 to identify functional or pathway enrichment.
Radiology Review
MRI scans were reviewed by a pediatric neuro-radiologist to verify tumor location. The following parameters were extracted from MR images performed at the time of diagnosis: Primary tumor volume (measured using FLAIR (Fluid-Attenuated Inversion Recovery) sequence), extension to basal ganglia, and extension to medial temporal lobes.
Statistical Analyses
P values for comparison of differential gene expression and differential methylation were calculated from the hypergeometric distribution for (k−1, K, N−K, n), where k is the number of genes in the intersection of the query set with a set from MSigDB, K is the number of genes in the set from MSigDB, N is the total number of gene universe, and n is the number of genes in the query set. Only gene sets with a false discovery rate (q value) < 0.05 after correction for multiple hypothesis testing were considered. Fisher’s exact test and Wilcoxon rank-sum test were used to assessing frequency distributions for categorical and continuous variables, respectively. Kaplan–Meier method was used to estimate survival, and the logrank test was used to compare survival curves. Overall survival (OS) time was calculated from the time of diagnosis to the last follow-up or death. Progression was defined by radiographic criteria,20 visual deterioration, or change of therapy due to concern for progression. Time to the first progression was calculated from the start of the first therapy to progression, death, or last follow-up, whichever occurred first. Time to the second progression was calculated from the start of the second therapy to progression, death, or last follow-up, whichever occurred first. PFS after the first therapy and second therapy are represented by PFS1 and PFS2, respectively. All statistical analyses were carried out using Stata version 17 (StataCorp).
Results
Patient Characteristics
Seventy-two patients with unresectable, sporadic/non-NF1-associated hypothalamic/optic pathway PA treated at St. Jude were included in this retrospective analysis (Figure 1A). The median follow-up for the entire cohort was 12.3 years. Half of the patients (n = 36) met the criteria for Cohort B. As shown in Figure 1B, there was a male predilection in our study population (M:F = 45:27 or 1.7:1), which was exaggerated in Cohort B (M:F = 26:10 or 2.6:1), but did not reach statistical significance. Patients in Cohort B were significantly younger at diagnosis than those in Cohort A (median age at diagnosis: 3.2 years vs 6.7 years, P = .005, Figure 1C). Systemic therapies are reported in Supplementary Table 1. Most patients received chemotherapy as first-line therapy, followed by RT or MEK-I for subsequent progressions. Five patients were treated with MEK-I. The treatment course of Cohort B patients is listed in Supplementary Table 2. All patients who underwent RT as first-line therapy were older than three years of age at the time of treatment. The median age at the start of first-line therapy for patients undergoing RT was 9.2 years compared to 3.2 years for patients undergoing systemic therapy (P < .001). There was no difference in the median age of patients undergoing RT as first-line therapy in Cohort A vs. Cohort B (12 years vs 8 years, P = .167). The median number of total therapies (ie, lines of systemic therapies and RT) per patient was 3 in Cohort B compared to 1 in Cohort A. Although the study period spanned 33 years, there was no association between RT or chemotherapy use and the year of diagnosis. However, there was an association between MEK-I use and the year of diagnosis (P = .027), as more recently diagnosed patients were more like to be treated with MEK-I. As expected by study design, OS, PFS1, and PFS2 were significantly worse in Cohort B compared to Cohort A (all P < .02, Figure 1D–F).
Figure 1.
Study design and patient characteristics.
(A) Seventy-two treatment-naïve hypothalamic pilocytic astrocytomas were included in the study. The patients were divided into Cohort A or Cohort B based on treatment response and metastasis status. DNA methylome and transcriptome profiles were then performed. The findings were validated by immunohistochemistry, Reverse transcription-polymerase chain reaction (RT-PCR), or interphase fluorescence in situ hybridization (iFISH). (B) There was a male predilection in Cohort B tumors. (C) Age distribution (D–F) Kaplan–Meier curves of progression-free survival after the first (Progression-free survival 1) and second therapy (Progression-free survival 2) and overall survival. CNV, copy number variants; DMR, differentially methylated regions; SNV, single nucleotide variants.
Histopathology Review
H&E-stained sections were available for review for all Cohort A and B treatment-naïve tumors. The histology of Cohort A and B tumors was similar, and all showed comparable cell densities and architecture typical for PA. Notably, no tumors demonstrated pilomyxoid astrocytoma histology. There was no difference in mitotic activity (median = 0/1.96 mm2 in all, P = .649), Ki-67 labeling index (median = 0 to 1% in all, P = .821), or presence of necrosis (P = .240), vascular proliferation (P = .430), infiltrative growth (P = .804) when comparing Cohorts A and B (Supplementary Table 3).
Radiology Review
MRI scans at diagnosis were available for analysis in 23 Cohort A and 23 Cohort B patients. There was no difference in tumor volume (median tumor volume at diagnosis: 36 cc vs 20.9 cc, P = .087) or number of tumors with extension to basal ganglia (unilateral extension, P = .188; bilateral extension, P = .267) or extension to temporal lobes (unilateral extension, P = .171; or bilateral extension, P = .138) when comparing Cohorts A and B.
The Genetic Landscape of Progressive Hypothalamic/Optic Pathway PA
Twenty-seven Cohort B and 13 Cohort A treatment-naïve tumors had sufficient tissue for detailed molecular analysis. To confirm that the outcomes of patients with sufficient tissue for molecular analysis (n = 40) were similar to that of the entire study population (n = 72), we repeated the PFS1, PFS2, and OS analyses for the subgroup of patients with sufficient tissue. As expected, within this subgroup, PFS1, PFS2, and OS remained significantly worse for patients in Cohort B compared to Cohort A (Supplementary Figure 1).
With respect to molecular alterations, as shown in Figure 2, 81.5% (22/27) Cohort B tumors harbored a KIAA1549-BRAF fusion, whereas only 38.5% (5/13) Cohort A tumors harbored the fusion. On the other hand, 61.5% (8/13) Cohort A tumors harbored SNV in BRAF (n = 7) or FGFR1 (n = 1), whereas only 14.8% (4/27) Cohort B tumors harbored SNV in BRAF (n = 3) or FGFR1 (n = 1). The enrichment for KIAA1549-BRAF fusion in Cohort B tumors reached statistical significance (P = .0032) when compared with Cohort A. Alterations in NF1, TSC2, and PTPN11 were identified in single tumors, with concomitant BRAF or FGFR1 alterations. Alterations in IDH1, IDH2, histone H3 K27, TP53, ATRX, MYB, PIK3CA, and CDKN2A/B were not identified by IHC, iFISH, or RNA-seq in any of the 72 tumors.
Figure 2.
Oncoprint shows the distribution of gender and molecular alterations among the tumors.
Methylation Profiles of Progressive Hypothalamic/Optic Pathway PA
Twenty-four Cohort B and 13 Cohort A treatment-naïve tumors had sufficient tissue for genome-wide DNA methylation profiling by the Illumina Infinium MethylationEPIC array. As shown in Figure 3A, the methylation profiles of Cohorts A and B formed a single cluster on the t-SNE plot, supporting the notion that our cohort consisted of a homogeneous tumor type.
Figure 3.
DNA methylome analysis of Cohorts A and B tumors.
(A) t-SNE plot showing that the methylome profiles of Cohorts A and B tumors formed a single cluster distinct from other brain tumor types. AIDH, IDH-mutant diffuse astrocytoma; DNET, dysembryoplastic neuroepithelial tumor; G34R, H3 G34R-mutant diffuse hemispheric glioma; K27M, H3 K27M-mutant diffuse midline glioma; MYB, MYB-altered glioma; OIDH, IDH-mutant and 1p/19q-codeleted oligodendroglioma; RGNT, rosette-forming glioneuronal tumor; sDNET, septal dysembryoplastic neuroepithelial tumor. (B) Differentially hypomethylated nuclear-encoded mitochondrial genes in Cohort B tumors. (C) Differentially hypomethylated genes in Cohort B tumors were enriched for target genes of specific transcription factors.
Comparing the genome-wide DNA methylation profiles of Cohorts A and B tumors revealed that nine of the 30 genes differentially methylated in Cohort B tumors were nuclear-encoded mitochondrial genes, and the enrichment was statistically significant (Figure 3B). Among the products of the nine genes, NDUFA2 and NDUFA5 are parts of the mitochondrial electron transport chain. SSBP1 is required to maintain the copy number of mitochondrial DNA (mtDNA) and plays a crucial role during mtDNA replication and mitochondrial biogenesis.21 This finding suggests that Cohort B tumors may have increased mitochondrial biogenesis.
Gene set enrichment analysis of the differentially methylated genes between Cohorts A and B tumors additionally revealed enrichment of the downstream genes of 14 transcription factors (Figure 3C)—these formed candidates for further validation by transcriptome analysis.
Activated Drivers in Progressive Hypothalamic/Optic Pathway PA
Ten Cohort B and 10 Cohort A treatment-naïve tumors had transcriptome (RNA-seq) profiles sufficient for differential expression analysis. We first applied NetBID, a network-based Bayesian inference approach for drivers, SJARACNe, a scalable solution of ARACNe (Algorithm for the Reconstruction of Accurate Cellular Networks), and MSigDB to identify transcriptional and signaling drivers differentially activated in Cohort B tumors.12 As shown in Supplementary Figure 2, TBX3 and PIM1 showed statistically significant increases in activity and mRNA expression in Cohort B tumors compared to Cohort A, representing potential drivers of progressive disease. Gene set overlap matrix showed that TBX3 and PIM1 were downstream of transcription factors E2F, NKX2.3, CDPCR1, CDPCR3, HHEX, ZFP3, TATA-binding protein/TBP, ZNF512, and HOXC11 (Figure 4A). E2F and NKX2.3 co-regulate TBX3 and PIM1. PIM1 is additionally downstream of p53. This finding supports our methylation analysis (Figure 3C), as E2F and NKX2.3 are upstream of both differentially methylated and differentially expressed genes when comparing Cohort B tumors to Cohort A.
Figure 4.
Activated drivers and pathways in progressive hypothalamic/optic pathway PA.
(A) TBX3 and PIM1 were identified as top activated drivers in Cohort B tumors downstream of E2F, NKX2.3, and p53. DA, differentially activated. Expression of TBX and PIM1 was increased in Cohort B tumors (C, E) and low in Cohort A tumors (B, D; see also Supplementary Figure 3). (F) List of hallmark gene sets showing the activation of MYC, p53, E2F, and PI3K/AKT/mTOR pathways, oxidative phosphorylation, fatty acid metabolism, and glycolysis in Cohort B tumors. Scale bar: 100 μm.
To validate the expression of TBX3 and PIM1 at the protein levels in additional Cohorts A and B tumor specimens, we performed IHC on treatment-naïve tissue sections of 36 Cohort B and 19~31/36 Cohort A tumors. TBX3 expression, as measured by nuclear positivity, was increased in Cohort B tumors compared to Cohort A (Figure 4B vs 4C, Supplementary Figure 3, P =1.4 × 10−30). Similarly, PIM1 expression, as measured by H-score, was increased in Cohort B tumors compared to Cohort A (Figure 4D vs 4E, Supplementary Figure 3, P = 2.3 × 10−31).
Since PIM1 is known to phosphorylate and synergize with MYC to increase its stability and transcriptional activity,22–24 we searched genes upregulated in Cohort B tumors against the hallmark gene sets to identify evidence of MYC pathway activation. As shown in Figure 4F, significant enrichment of MYC targets was identified in Cohort B tumors, alongside the targets of E2F, p53, and PI3K/AKT/mTOR pathways. Multiple metabolic pathways known to be regulated by the MYC pathway, including oxidative phosphorylation, fatty acid metabolism, and glycolysis,25–28 were upregulated as well, further supporting the findings of our transcriptome and methylation analyses.
Activated Pathways in Progressive Hypothalamic/Optic Pathway PA
Our initial hallmark gene set search identified several signaling and metabolic pathways in Cohort B tumors worth further exploration and validation. Therefore, we searched the genes upregulated in Cohort B tumors against the MSigDB to identify pathways differentially activated in Cohort B tumors.12 We identified enrichment of mitochondrial and oxidative phosphorylation-related genes (Figure 5A) and evidence of MAPK (Figure 5D) and PI3K (Figure 5G) pathway activation. IHC was then applied to validate the findings in treatment-naïve Cohorts A and B tumors. Expression of Tom20 (Figure 5C–D, Supplementary Figure 4), a mitochondrial marker, nuclear phospho-ERK1/2 (Figure 5E–F, Supplementary Figure 4), a marker of MAPK pathway activation, and phospho-S6 (Figure 5H–I, Supplementary Figure 4), a marker of PI3K/AKT/mTOR pathway activation, was increased in Cohort B tumors compared to Cohort A (Tom20: P = 8.0 × 10−16; phospho-ERK1/2: P = 1.0 × 10−13; phospho-S6: P = 7.8 × 10−32), validating the findings identified at the transcriptome levels. Differences in the expression of senescence pathway-related genes were not observed between Cohorts A and B tumors (Supplementary Figure 5).
Figure 5.
Gene enrichment and proposed core signaling and transcription networks in progressive hypothalamic/optic pathway PA.
(A) Enrichment of mitochondrial-related genes in Cohort B tumors. (B–C) Levels of Tom20, a mitochondrial marker, was increased in Cohort B tumors and low in Cohort A tumors (see also Supplementary Figure 4). (D) Enrichment of MAPK pathway in Cohort B tumors (E–F) Nuclear phospho-ERK1/2 (pERK1/2), a marker of activated MAPK pathway, was increased in Cohort B tumors and low in Cohort A tumors (see also Supplementary Figure 4). (G) Enrichment of PI3K/AKT/mTOR pathway in Cohort B tumors (H–I) Levels of phospho-S6 (pS6), a marker of activated PI3K/AKT/mTOR pathway, was increased in cohort B tumors and low in cohort A tumors (see also Supplementary Figure 4). (J) p53, E2F, NKX2.3, TBX3, PIM1, and MYC all play roles in cell proliferation and the regulation of biosynthetic processes. TBX3, PIM1, and MYC are downstream of E2F. (K) Proposed transcription and signaling network in progressive hypothalamic/optic pathway PA. Scale bar: 100 μm.
MSigDB search further identified that p53, E2F, NKX2.3, TBX3, PIM1, and MYC regulate cell proliferation and biosynthetic process and that TBX3, PIM1, and MYC are targets of E2F (Figures 4A and 5J). Additionally, ribosomal components were upregulated in Cohort B tumors (Supplementary Table 4). The overall findings of our DNA methylome, transcriptome, and IHC analysis support the model in which the p53-PIM1-MYC axis acts alongside TBX3 and MAPK and PI3K/AKT/mTOR pathways to promote cell proliferation and multiple biosynthetic processes that support tumor progression in hypothalamic/optic pathway PA (Figure 5K).
Discussion
Managing unresectable hypothalamic/optic pathway PA remains a challenge. In our study, half of the patients fulfilled one of the following criteria: Progression after ≥ 2 lines of systemic therapy, progression after RT, development of metastatic disease, or progression resulting in death. Therefore, there is a need to identify new therapeutic vulnerabilities in progressive hypothalamic/optic pathway PA beyond the known alterations in the MAPK pathway. Furthermore, identifying prognostic biomarkers will improve patient risk stratification and guide optimal therapy.
This study sought to identify clinical, genetic, epigenetic, and transcriptomic parameters associated with progressive disease in hypothalamic/optic pathway PA and validate the findings by IHC. Patients in Cohort B presented at a younger age (Figure 1C), consistent with prior studies.29,30 We also identified a male predilection among patients in Cohort B (Figure 1B). We found that 81.5% of Cohort B tumors harbored a KIAA1549-BRAF fusion, while only 38.5% of Cohort A tumors harbored the fusion (Figure 2). Most (61.5%) of the Cohort A tumors had SNV in BRAF or FGFR1. Although gene fusions have been associated with a good prognosis and SNVs have been associated with a poor prognosis in pediatric low-grade neural tumors when all sites and histologies were considered,31 our series demonstrated the opposite phenomenon in hypothalamic/optic pathway PA, highlighting that tumor location may modify the prognostic value of BRAF fusions. However, it should be noted that there were only 13 patients in Cohort A with tissue available for analysis, and therefore, our findings should be considered hypothesis-generating.
Cohorts A and B tumors showed overlapping methylome profiles in our analysis (Figure 3A), supporting the histologic and epigenetic homogeneity of our cohort. We additionally showed that mitochondrial genes and genes downstream of E2F and NKX2.3 transcription factors were differentially methylated in Cohort B tumors (Figure 3B–C). This finding was further supported by transcriptome analysis (Figures 4A, F, and 5A–C).
Our transcriptome analysis and confirmatory IHC on tumor samples shed further light on the underlying drivers, signaling pathways, and transcription networks in progressive hypothalamic/optic pathway PA (Figures 4 and 5). We identified transcription factor TBX3 and PIM1 serine/threonine kinase as the potential drivers downstream of E2F and NKX2.3 in progressive hypothalamic/optic pathway PA and confirmed their upregulation at the protein levels by IHC. Both TBX3 and PIM1 kinase are overexpressed in many cancers, act synergistically with MYC, and activate the PI3K/mTOR pathway.32–36 PIM1 acts downstream of p53 and phosphorylates serine 62 of MYC to increase its stability, transcriptional activity, and carcinogenicity.22–24 TBX3 acts downstream of MYC and protects cells against MYC and p53-induced apoptosis.37,38 In addition, TBX3 and PIM1 activate the PI3K/mTOR pathway by suppressing PTEN39 and phosphorylating AKT,34 respectively. PIM1 inhibition hampers the tumorigenesis of MYC-driven cells, suppresses the MYC pathway, and sensitizes cancer cells to RT in vitro and in vivo.33,40 Inhibitors targeting PIM1 are being investigated in clinical trials (Clinicaltrials.gov: NCT04176198, NCT02370706).
Enhanced mitochondrial biogenesis and metabolism observed in Cohort B tumors are likely secondary to enhanced PI3K/AKT/mTOR, MAPK and MYC signaling. The PI3K/AKT/mTOR pathway is important in mediating glycolysis of cancer cells by activating phosphofructokinase 2 and hexokinase 2 and upregulating Glucose Transporter 1.41 Furthermore, AKT/mTOR signaling and MYC work synergistically to regulate ribosomal biogenesis through transcriptional control of ribosomal RNA and protein components.42,43 MYC additionally directly regulates mitochondrial biogenesis.25
This is a single-institution retrospective review with biases inherent to the study design. Patients who are referred to our institution commonly have progressive tumors. This may have increased the number of patients who are treated with RT or meet the criteria for Cohort B compared to other institutions.44 However, all biopsy specimens used for analysis were treatment-naïve; therefore, therapy modality should not affect the results of methylation, RNA-seq, and IHC analysis. Malignant transformation or therapy-associated secondary malignancy was not detected in any patients. None of the treatment naïve or subsequent specimens contained CDKN2A deletions, which have been associated with malignant transformations in pediatric low-grade gliomas.45
In summary, we compared the clinical, imaging, histologic, genetic, methylomic, transcriptional, and immunohistochemical profiles of hypothalamic/optic pathway PAs that differ in their outcomes. Cohort B included patients with progressive disease and was associated with male gender, enrichment of KIAA1549-BRAF fusion, and young age at diagnosis. In addition, Cohort B tumors demonstrated a core transcription network involving E2F, NKX2.3, TBX3, PIM1, p53, and MYC and showed increased activation of MAPK and PI3K/AKT/mTOR pathways, resulting in upregulation of multiple biosynthetic processes. Our findings shed light on the biological basis of tumor progression in hypothalamic/optic pathway PA and provide directions for developing effective combination therapies and refining prognosis.
Supplementary Material
Acknowledgments
The authors would like to thank Julie Justice in Department of Pathology and Emily Walker and Sai Prabhas Konduru in the St. Jude Hartwell Center for their assistance in the molecular characterization of the tumors.
Contributor Information
Xiaoyu Li, Department of Pathology, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA.
Daniel C Moreira, Department of Global Pediatric Medicine, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA; Department of Oncology, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA.
Asim K Bag, Department of Diagnostic Imaging, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA.
Ibrahim Qaddoumi, Department of Global Pediatric Medicine, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA; Department of Oncology, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA.
Sahaja Acharya, Department of Radiation Oncology, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA.
Jason Chiang, Department of Pathology, St. Jude Children’s Research Hospital, Memphis, Tennessee, USA.
Funding
SA receives support from Conquer Cancer ASCO Foundation and the Clark Foundation. JC receives support from the St. Jude Comprehensive Cancer Center (NCI grant P30CA021765), NCI Program Project P01CA096832, and American Lebanese Syrian Associated Charities.
Conflict of Interest
The authors declare no conflict of interest.
Data Availability
The methylation and RNA-seq data with associated de-identified clinical information are available on GEO (#GSE199362).
Authorship
Experimental design: DCM, IQ, SA, and JC. Implementation: XL, DCM, AKB, IQ, SA, and JC. Analysis and interpretation of the data: XL, DCM, AKB, SA, and JC. All authors have been involved in the writing of the manuscript and have read and approved the final version.
References
- 1. Ostrom QT, Cioffi G, Gittleman H, et al. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2012–2016. Neuro Oncol. 2019;21(suppl 5):v1–v100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Gnekow AK, Falkenstein F, von Hornstein S, et al. Long-term follow-up of the multicenter, multidisciplinary treatment study HIT-LGG-1996 for low-grade glioma in children and adolescents of the German Speaking Society of Pediatric Oncology and Hematology. Neuro Oncol. 2012;14(10):1265–1284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Chamdine O, Broniscer A, Wu S, Gajjar A, Qaddoumi I. Metastatic low-grade gliomas in children: 20 years’ experience at st. jude children’s research hospital. Pediatr Blood Cancer. 2016;63(1):62–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ater JL, Zhou T, Holmes E, et al. Randomized study of two chemotherapy regimens for treatment of low-grade glioma in young children: a report from the Children’s Oncology Group. J Clin Oncol. 2012;30(21):2641–2647. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Bouffet E, Jakacki R, Goldman S, et al. Phase II study of weekly vinblastine in recurrent or refractory pediatric low-grade glioma. J Clin Oncol. 2012;30(12):1358–1363. [DOI] [PubMed] [Google Scholar]
- 6. Cherlow JM, Shaw DWW, Margraf LR, et al. Conformal radiation therapy for pediatric patients with low-grade glioma: results from the children’s oncology group phase 2 study ACNS0221. Int J Radiat Oncol Biol Phys. 2019;103(4):861–868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Tsang DS, Murphy ES, Merchant TE. Radiation therapy for optic pathway and hypothalamic low-grade gliomas in children. Int J Radiat Oncol Biol Phys. 2017;99(3):642–651. [DOI] [PubMed] [Google Scholar]
- 8. Merchant TE, Conklin HM, Wu S, Lustig RH, Xiong X. Late effects of conformal radiation therapy for pediatric patients with low-grade glioma: prospective evaluation of cognitive, endocrine, and hearing deficits. J Clin Oncol. 2009;27(22):3691–3697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Fangusaro J, Onar-Thomas A, Young Poussaint T, et al. Selumetinib in paediatric patients with BRAF-aberrant or neurofibromatosis type 1-associated recurrent, refractory, or progressive low-grade glioma: a multicentre, phase 2 trial. Lancet Oncol. 2019;20(7):1011–1022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Chiang J, Diaz AK, Makepeace L, et al. Clinical, imaging, and molecular analysis of pediatric pontine tumors lacking characteristic imaging features of DIPG. Acta Neuropathol Commun. 2020;8(1):57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Chiang JCH, Harreld JH, Tanaka R, et al. Septal dysembryoplastic neuroepithelial tumor: a comprehensive clinical, imaging, histopathologic, and molecular analysis. Neuro Oncol. 2019;21(6):800–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Chiang J, Li X, Jin H, et al. The molecular characteristics of low-grade and high-grade areas in desmoplastic infantile astrocytoma/ganglioglioma. Neuropathol Appl Neurobiol. 2022;48(4):: e12801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Keenan C, Graham RT, Harreld JH, et al. Infratentorial C11orf95-fused gliomas share histologic, immunophenotypic, and molecular characteristics of supratentorial RELA-fused ependymoma. Acta Neuropathol. 2020;140(6):963–965. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Chiang J, Li X, Liu APY, et al. Tectal glioma harbors high rates of KRAS G12R and concomitant KRAS and BRAF alterations. Acta Neuropathol. 2020;139(3):601–602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Chiang J, Harreld JH, Tinkle CL, et al. A single-center study of the clinicopathologic correlates of gliomas with a MYB or MYBL1 alteration. Acta Neuropathol. 2019;138(6):1091–1092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Tian L, Li Y, Edmonson MN, et al. CICERO: a versatile method for detecting complex and diverse driver fusions using cancer RNA sequencing data. Genome Biol. 2020;21(1):126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Uhrig S, Ellermann J, Walther T, et al. Accurate and efficient detection of gene fusions from RNA sequencing data. Genome Res. 2021;31(3):448–460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Edmonson MN, Zhang J, Yan C, et al. Bambino: a variant detector and alignment viewer for next-generation sequencing data in the SAM/BAM format. Bioinformatics. 2011;27(6):865–866. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Edmonson MN, Patel AN, Hedges DJ, et al. Pediatric Cancer Variant Pathogenicity Information Exchange (PeCanPIE): a cloud-based platform for curating and classifying germline variants. Genome Res. 2019;29(9):1555–1565. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Fangusaro J, Witt O, Hernaiz Driever P, et al. Response assessment in paediatric low-grade glioma: recommendations from the Response Assessment in Pediatric Neuro-Oncology (RAPNO) working group. Lancet Oncol. 2020;21(6):e305–e316. [DOI] [PubMed] [Google Scholar]
- 21. Ciesielski GL, Bermek O, Rosado-Ruiz FA, et al. Mitochondrial single-stranded DNA-binding proteins stimulate the activity of DNA polymerase gamma by organization of the template DNA. J Biol Chem. 2015;290(48):28697–28707. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Wang J, Kim J, Roh M, et al. Pim1 kinase synergizes with c-MYC to induce advanced prostate carcinoma. Oncogene. 2010;29(17):2477–2487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Zhang Y, Wang Z, Li X, Magnuson NS. Pim kinase-dependent inhibition of c-Myc degradation. Oncogene. 2008;27(35):4809–4819. [DOI] [PubMed] [Google Scholar]
- 24. Kim J, Roh M, Abdulkadir SA. Pim1 promotes human prostate cancer cell tumorigenicity and c-MYC transcriptional activity. BMC Cancer. 2010;10:248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Popay TM, Wang J, Adams CM, et al. MYC regulates ribosome biogenesis and mitochondrial gene expression programs through its interaction with host cell factor-1. Elife. 2021;10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Casciano JC, Perry C, Cohen-Nowak AJ, et al. MYC regulates fatty acid metabolism through a multigenic program in claudin-low triple negative breast cancer. Br J Cancer. 2020;122(6):868–884. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Goetzman ES, Prochownik EV. The role for myc in coordinating glycolysis, oxidative phosphorylation, glutaminolysis, and fatty acid metabolism in normal and neoplastic tissues. Front Endocrinol (Lausanne). 2018;9:129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Tateishi K, Iafrate AJ, Ho Q, et al. Myc-driven glycolysis is a therapeutic target in glioblastoma. Clin Cancer Res. 2016;22(17):4452–4465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Gnekow AK, Walker DA, Kandels D, Picton S, Giorgio Perilongo, Grill J, Stokland T, Sandstrom PE, Warmuth-Metz M, Pietsch T, Giangaspero F, Schmidt R, Faldum A, Kilmartin D, De Paoli A, De Salvo GL; of the Low Grade Glioma Consortium and the participating centers. A European randomised controlled trial of the addition of etoposide to standard vincristine and carboplatin induction as part of an 18-month treatment programme for childhood (</=16 years) low grade glioma - A final report. Eur J Cancer. 2017;81:206–225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Mirow C, Pietsch T, Berkefeld S, et al. Children <1 year show an inferior outcome when treated according to the traditional LGG treatment strategy: a report from the German multicenter trial HIT-LGG 1996 for children with low grade glioma (LGG). Pediatr Blood Cancer. 2014;61(3):457–463. [DOI] [PubMed] [Google Scholar]
- 31. Ryall S, Zapotocky M, Fukuoka K, et al. Integrated molecular and clinical analysis of 1,000 pediatric low-grade gliomas. Cancer Cell. 2020;37(4):569–583.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Khan SF, Damerell V, Omar R, et al. The roles and regulation of TBX3 in development and disease. Gene. 2020;726:144223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Chen J, Tang G. PIM-1 kinase: a potential biomarker of triple-negative breast cancer. Onco Targets Ther. 2019;12:6267–6273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Leung CO, Wong CC, Fan DN, et al. PIM1 regulates glycolysis and promotes tumor progression in hepatocellular carcinoma. Oncotarget. 2015;6(13):10880–10892. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Bellon M, Lu L, Nicot C. Constitutive activation of Pim1 kinase is a therapeutic target for adult T-cell leukemia. Blood. 2016;127(20):2439–2450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Weirauch U, Beckmann N, Thomas M, et al. Functional role and therapeutic potential of the pim-1 kinase in colon carcinoma. Neoplasia. 2013;15(7):783–794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Carlson H, Ota S, Song Y, Chen Y, Hurlin PJ. Tbx3 impinges on the p53 pathway to suppress apoptosis, facilitate cell transformation and block myogenic differentiation. Oncogene. 2002;21(24):3827–3835. [DOI] [PubMed] [Google Scholar]
- 38. Willmer T, Peres J, Mowla S, Abrahams A, Prince S. The T-Box factor TBX3 is important in S-phase and is regulated by c-Myc and cyclin A-CDK2. Cell Cycle. 2015;14(19):3173–3183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Burgucu D, Guney K, Sahinturk D, et al. Tbx3 represses PTEN and is over-expressed in head and neck squamous cell carcinoma. BMC Cancer. 2012;12:481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Kirschner AN, Wang J, van der Meer R, et al. PIM kinase inhibitor AZD1208 for treatment of MYC-driven prostate cancer. J Natl Cancer Inst. 2015;107(2):dju407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Manning BD, Cantley LC. AKT/PKB signaling: navigating downstream. Cell. 2007;129(7):1261–1274. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. van Riggelen J, Yetil A, Felsher DW. MYC as a regulator of ribosome biogenesis and protein synthesis. Nat Rev Cancer. 2010;10(4):301–309. [DOI] [PubMed] [Google Scholar]
- 43. Chan JC, Hannan KM, Riddell K, et al. AKT promotes rRNA synthesis and cooperates with c-MYC to stimulate ribosome biogenesis in cancer. Sci Signal. 2011;4(188):ra56. [DOI] [PubMed] [Google Scholar]
- 44. Laithier V, Grill J, Le Deley MC, et al. ; French Society of Pediatric Oncology. Progression-free survival in children with optic pathway tumors: dependence on age and the quality of the response to chemotherapy--results of the first French prospective study for the French Society of Pediatric Oncology. J Clin Oncol. 2003;21(24):4572–4578. [DOI] [PubMed] [Google Scholar]
- 45. Mistry M, Zhukova N, Merico D, et al. BRAF mutation and CDKN2A deletion define a clinically distinct subgroup of childhood secondary high-grade glioma. J Clin Oncol. 2015;33(9):1015–1022. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The methylation and RNA-seq data with associated de-identified clinical information are available on GEO (#GSE199362).