Daniels, David J.

Abstract BACKGROUND Diffuse midline gliomas harboring the H3K27M mutation are aggressive and universally fatal brain tumors that primarily occur in children. The blood-brain barrier (BBB) prevents many drugs from reaching the tumor at cytotoxic concentrations and efflux pumps found on the epithelial cells of the BBB rapidly pump drugs out of the brain. Convection enhanced delivery (CED) is a drug delivery technique that bypasses the BBB by directly injecting the drug into the tumor site under a pressure gradient. Further clinical implementation of CED requires understanding drug distribution, as optimal drug physico-chemical properties have never been evaluated. METHODS Sprague Dawley rats underwent a single injection of drug by CED to the brainstem with and without an efflux pump inhibitor. Animals were euthanized at 0, 2, 4, 8, 12 and 24 hours. Their brain drug concentration/distribution was analyzed by MALDI-MSI and plasma drug concentration was measured by LC-MS/MS. RESULTS Drug distribution and brainstem concentration were increased following BBB efflux pump inhibition compared to no pump inhibition controls. Additionally, efflux pump inhibition resulted in slower drug clearance for those drugs that are known pump substrates. CONCLUSIONS BBB efflux pump inhibition resulted in a larger volume of distribution, increased drug concentration and slower drug clearance following CED to the brainstem of known efflux substrates.

Abstract Peritoneal Carcinomatosis of anaplastic ependymoma is not a previously reported entity. The authors report on a child with multiple successfully treated brain and spine disease occurrences who subsequently develops carcinomatosis of the abdomen and no evidence of CNS recurrence. Ependymoma accounts for up to 10% of childhood CNS tumors diagnosed in the United States with a median age of 51–71 months. Typical locations are based on age. Disease is typically treated with surgical resection followed by radiation. The role of chemotherapy has not been proven but currently being examined with open clinical trials. We will describe patient’s presentations, clinical treatment and recurrence with subsequent treatment and outcome at time of meeting.

Abstract Diffuse midline gliomas (DMG) patients have a dire prognosis despite radiation therapy and there is an urgent need to develop more effective treatments. DMG are characterized by heterozygous mutations in select H3 genes resulting in the replacement of lysine 27 by methionine (K27M) that leads to global epigenetic reprogramming and drives tumorigenesis. We previously reported that pharmacological inhibition of aurora kinase (AKI) may represent a targeted approach for treating tumors with this mutation. Our analysis with both published dataset and patient samples showed that patients with higher aurora kinase A (AKA) expression were associated with worse survival. AKA phosphorylates H3S10 and H3S28 during mitosis. Intriguingly, phosphorylation of the H3S28 (H3S28ph) by AKA blocks PRC2 methyltransferase activity and decreases global H3K27me3 in certain stem cells. We propose that a similar mechanism occurs in H3K27M DMG tumors, where there is a reciprocal relationship between H3S28ph and H3K27me3. We found that AKI significantly decreases H3S28ph while increasing H3K27me3 specifically in H3K27M tumors. To further evaluate the link between the H3K27M mutation and H3 serine phosphorylation, we used CRISPR/Cas9-directed gene editing to silence H3S28ph by replacing serine with alanine (H3S28A) in DIPG cell lines. Ectopic expression of histone H3S28A leads to a prominent epigenetic changes in H3K27M tumors and is similar to AKA inhibition. Overall, this study highlights H3S28ph, one of the targets of AK, is a key driver of epigenetic changes in H3K27M tumors through both direct and indirect changes to H3K27me3 and H3K27ac across the genome.

Abstract Diffuse midline glioma (DMG) is a lethal brain tumor that typically occurs in children. Numerous studies have demonstrated the central role of the H3K27M mutation and secondary loss of H3K27 trimethylation (H3K27me3) in DMG tumorigenesis. Understanding how the H3K27M mutation alters the epigenetic landscape of the cell is necessary for revealing molecular targets that are critical to tumorigenesis. To investigate the epigenetic effects of H3K27M mutation in DMG, we developed revertant DMG cell lines with the mutant methionine residue reverted to wildtype (i.e., M27K). Revertant cells were analyzed for epigenetic changes and phenotypic differences in vitro and in vivo. H3M27K DMG cells grew in culture but displayed diminished proliferative capacity. H3M27K cells demonstrated total loss of H3K27M expression and restored trimethylation of H3K27 and H3K4. Furthermore, consistent with the hypothesis that the H3K27M mutation impacts H3 phosphorylation via expression of Aurora Kinase during mitosis, H3M27K cells demonstrated reduced expression of both Aurora Kinase A and phosphorylation of H3 serine residues 10 and 28. In line with the critical role of H3S10 phosphorylation in chromatin segregation, H3M27K cells also demonstrated restored chromosome segregation compared to H3K27M cells. In vivo data will be discussed. Revertance of the H3K27M mutation reduces tumorigenesis in DMG tumors. Isogenic H3M27K cells display reversal of key epigenetic changes associated with oncogenesis in DMG. The revertant H3M27K DMG model is a useful tool to investigate the downstream epigenetic reprogramming specific to H3K27M mutation in these tumors.

OBJECTIVE Pediatric Chiari I malformation decompression is a common neurosurgical procedure. Liposomal bupivacaine (LB) is a novel formulation that can have an impact on postoperative recovery for particular procedures, but its potential role in pediatric neurosurgery is largely unexplored. The authors sought to describe and assess their initial experience with LB in pediatric Chiari I malformation decompression to better define its potential role as an analgesic agent in a procedure for which the postoperative course is often remarkably painful. METHODS A retrospective review of all pediatric Chiari procedures performed at the authors’ institution between 2018 and 2020 was conducted. Patients were divided into those who were treated with a single intraoperative dose of LB (LB group) and those who were not (control group). Comparisons of total opioid use and pain control were made using chi-square and Wilcoxon rank-sum tests. RESULTS A total of 18 patients were identified, 9 (50%) in the LB group and 9 (50%) in the control group. Overall, there were 13 (72%) female and 5 (28%) male patients with a mean age of 15.9 years. No surgical complications were observed over a mean length of stay of 2.7 days. Within the first 24 hours after surgery, the LB group had significantly lower total opioid use than the control group (17.5 vs 47.9 morphine milligram equivalents, respectively; p = 0.03) as well as lower mean pain scores reported by patients using a 10-point visual analog scale (3.6 vs 5.5 for the LB vs control groups, p = 0.04). However, from the first 24 postoperative hours to discharge, total opioid use (p = 0.51) and mean pain scores (p = 0.09) were statistically comparable between the two groups. There were 2/9 (22%) LB patients versus 0/9 (0%) control patients who did not require opioid analgesia at any point during hospitalization. CONCLUSIONS The use of a single intraoperative dose of LB in pediatric Chiari I malformation surgery appears to be safe and has the potential to reduce pain scores and opioid use when administered during the first 24 postoperative hours. From that time period to discharge, however, there may be no significant difference in total opioid use or pain scores.

OBJECTIVEGlioblastoma (GBM) during infancy is rare, and the clinical outcomes of congenital GBM are not well understood. Correspondingly, the aim of this study was to present a long-term survivor case from the authors’ institution, and establish an integrated cohort of cases across the published literature to better understand the clinical course of this disease in this setting.METHODSThe authors report the outcomes of an institutional case of congenital GBM diagnosed within the first 3 months of life, and performed a comprehensive literature search for published cases from 2000 onward for an integrated survival analysis. All cases were integrated into 1 cohort, and Kaplan-Meier estimations, Fisher’s exact test, and logistic regression were used to interrogate the data.RESULTSThe integrated cohort of 40 congenital GBM cases consisted of 23 (58%) females and 17 (42%) males born at a median gestational age of 38 weeks (range 22–40 weeks). Estimates of overall survival (OS) at 1 month was 67%, at 1 year it was 59%, and at 10 years it was 45%, with statistically superior outcomes for subgroups in which patients survived to be treated by resection and chemotherapy. In the overall cohort, multivariable analysis confirmed resection (p < 0.01) and chemotherapy (p < 0.01) as independent predictors of superior OS. Gestational age > 38 weeks (p < 0.01), Apgar scores ≥ 7 at 5 minutes (p < 0.01), absence of prenatal hydrocephalus (p < 0.01), and vaginal delivery (p < 0.01) were associated with greater odds of surgical diagnosis versus autopsy diagnosis.CONCLUSIONSCongenital GBM can deviate from the expected poor prognosis of adult GBM in terms of OS. Both resection and chemotherapy confer statistically superior prognostic advantages in those patients who survive within the immediate postnatal period, and should be first-line considerations in the initial management of this rare disease.

Abstract BACKGROUND Central nervous system (CNS) germ cell tumors (GCT) are rare and complex pediatric neoplasms, the optimal management of which remains an area of active investigation. OBJECTIVE To present an updated cohort study, with particular attention to novel diagnostic methods and posttreatment clinical phenotypes. METHODS A single-institution cohort study of 80 primary, neurosurgically managed, CNS GCTs was conducted at Mayo Clinic, 1988-2017. RESULTS Postchemotherapy resection (eg, second-look surgery) was frequently required (27.0%), especially after adjuvant therapies for nongerminomatous GCTs (NGGCTs; 14 of 28 cases, excluding mature teratoma) and significantly associated with pineal lesions, as compared to neurohypophyseal or bifocal lesions (43.6% vs 5.9% vs 6.7%, P = .004), a finding that retained statistical significance after adjusting for index extent of resection and histology ( P = .04). Essentially every NGGCT case underwent at least 1 craniotomy, either on presentation, as second-look surgery, or following local recurrence. Mature teratomatous tissue was highly incident in second-look specimens (84.2%), even among lesions initially diagnosed as germinomas. Pretreatment cerebrospinal fluid (CSF) cell fraction analysis demonstrated an association between single lesions and neutrophil predominance, whereas nongerminomatous GCTs were associated with increased monocyte fractions. CONCLUSION CNS GCTs are clinically heterogeneous lesions, resulting in numerous opportunities for improved understanding and clinical management via novel diagnostic and therapeutic protocols. Samples from second-look surgeries for recurrent germinomas frequently demonstrate teratomatous tissue, suggesting possible underdiagnosis of mixed GCTs—particularly among pineal lesions. GCT subtypes demonstrate differential cell fraction distributions on CSF analysis, a novel and perhaps diagnostically helpful finding that requires validation in external cohorts.

OBJECTIVEFactors associated with syrinx size in pediatric patients undergoing posterior fossa decompression (PFD) or PFD with duraplasty (PFDD) for Chiari malformation type I (CM-I) with syringomyelia (SM; CM-I+SM) are not well established.METHODSUsing the Park-Reeves Syringomyelia Research Consortium registry, the authors analyzed variables associated with syrinx radiological outcomes in patients (< 20 years old at the time of surgery) with CM-I+SM undergoing PFD or PFDD. Syrinx resolution was defined as an anteroposterior (AP) diameter of ≤ 2 mm or ≤ 3 mm or a reduction in AP diameter of ≥ 50%. Syrinx regression or progression was defined using 1) change in syrinx AP diameter (≥ 1 mm), or 2) change in syrinx length (craniocaudal, ≥ 1 vertebral level). Syrinx stability was defined as a < 1-mm change in syrinx AP diameter and no change in syrinx length.RESULTSThe authors identified 380 patients with CM-I+SM who underwent PFD or PFDD. Cox proportional hazards modeling revealed younger age at surgery and PFDD as being independently associated with syrinx resolution, defined as a ≤ 2-mm or ≤ 3-mm AP diameter or ≥ 50% reduction in AP diameter. Radiological syrinx resolution was associated with improvement in headache (p < 0.005) and neck pain (p < 0.011) after PFD or PFDD. Next, PFDD (p = 0.005), scoliosis (p = 0.007), and syrinx location across multiple spinal segments (p = 0.001) were associated with syrinx diameter regression, whereas increased preoperative frontal-occipital horn ratio (FOHR; p = 0.007) and syrinx location spanning multiple spinal segments (p = 0.04) were associated with syrinx length regression. Scoliosis (HR 0.38 [95% CI 0.16–0.91], p = 0.03) and smaller syrinx diameter (5.82 ± 3.38 vs 7.86 ± 3.05 mm; HR 0.60 [95% CI 0.34–1.03], p = 0.002) were associated with syrinx diameter stability, whereas shorter preoperative syrinx length (5.75 ± 4.01 vs 9.65 ± 4.31 levels; HR 0.21 [95% CI 0.12–0.38], p = 0.0001) and smaller pB-C2 distance (6.86 ± 1.27 vs 7.18 ± 1.38 mm; HR 1.44 [95% CI 1.02–2.05], p = 0.04) were associated with syrinx length stability. Finally, younger age at surgery (8.19 ± 5.02 vs 10.29 ± 4.25 years; HR 1.89 [95% CI 1.31–3.04], p = 0.01) was associated with syrinx diameter progression, whereas increased postoperative syrinx diameter (6.73 ± 3.64 vs 3.97 ± 3.07 mm; HR 3.10 [95% CI 1.67–5.76], p = 0.003), was associated with syrinx length progression. PFD versus PFDD was not associated with syrinx progression or reoperation rate.CONCLUSIONSThese data suggest that PFDD and age are independently associated with radiological syrinx improvement, although forthcoming results from the PFDD versus PFD randomized controlled trial (NCT02669836, clinicaltrials.gov) will best answer this question.