Highlights from the Literature

Benefit from procarbazine, lomustine, and vincristine in oligodendroglial tumors is associated with mutation of IDH

In 2012 the Radiation Therapy Oncology Group (RTOG) Trial 9402 and the European Organization For Research and Treatment of Cancer (EORTC) Trial 26951 both showed that patients with anaplastic oligodendroglial tumors with 1p/19q codeletion survived significantly longer with radiotherapy (RT) and PCV chemotherapy compared to RT alone. However, a subset of patients with intact 1p/19q also appeared to benefit from chemoradiotherapy, and significantly more patients lived ≥10 years after chemoradiotherapy than RT, suggesting that 1p/19q status alone did not identify all responders. In this study the investigators examined whether IDH mutations or the G allele of rs55705857, a germ-line polymorphism associated with a six-fold risk of developing IDH-mutated glioma, were associated with improved outcome after chemoradiotherapy. Seventy-four percent of the 210 evaluable patients had IDH mutations and 31% of the 245 evaluable patients had the G allele. Both were associated with longer progression-free survival after chemoradiotherapy. Mutant IDH was associated with longer overall survival (9.4 v 5.7 years; hazard ratio [HR], 0.59; 95% CI, 0.40 to 0.86; P = .006). Patients with wild-type IDH tumors had a poor outcome and did not benefit from chemoradiotherapy (survival for chemoradiotherapy 1.3 years v RT, 1.8 years; HR, 1.14; 95% CI, 0.63 to 2.04; P = .67). There were also no differences in 10-year survival rates (chemoradiotherapy, 6% v RT, 4%). Patients with 1p/19q codeleted IDH mutated tumors lived longer after chemoradiotherapy than RT alone (14.7 v 6.8 years; HR, 0.49; 95% CI, 0.28 to 0.85; P = .01), as did noncodeleted IDH mutated tumors (5.5 v 3.3 years; HR, 0.56; 95% CI, 0.32 to 0.99; P < .05). Overall, patients with codeleted mutated tumors had the longest survival, those with noncodeleted mutated tumors had an intermediate survival, and those with neither had the shortest survival. This important study suggests that the presence of IDH mutations identifies a subgroup of patients who respond to chemoradiotherapy, regardless of their 1p/19q status, although the underlying mechanisms remain to be defined.

Hypoxia and oxygenation induce a metabolic switch between pentose phosphate pathway and glycolysis in glioma stem-like cells.

Recently, the implications and better understanding of cancer energy metabolism has gained signficant interest, with ultimate hopes that targets of altered metabolism will offer new therepeutic options. Otto Warburg was the first to present the concept that glucose metabolism in cancer cells is fundamentally different from normal tissue. Warburg observed excessive “fermentation” of glucose to lactate in cancer cells even when sufficient oxygen was present for normal respiration to take place. Although he mistakenly contributed the altered metabolism to irreversibly damaged mitochondria and hence “the origin of cancer cell”, nevertheless his contributions were pivotal to early understanding of cancer metabolism. Now we can appreciate that metabolic reprograming is possibly an adaptive response to meet the challenges of rapidly proliferating tumor cells. The Warburg effect represents a tumor cell-specific phenomenon that could potentially be exploited for therapeutic targeting while sparing normal tissue. It describes the phenomenon that tumor cells predominantly use glycolysis for energy production, resulting in increased lactate formation, even when sufficient oxygen is available (aerobic glycolysis). Energetically, glycolysis is far less efficient than oxidative phosphorylation, and the causes and functional advantages of the Warburg effect in cancer metabolism still remain largely elusive. Currently, there is sufficient evidence that glioblastoma cell metabolism differs significantly from that of normal brain tissue. One advantage is that high substrate flux through glycolysis facilitates effective shunting of carbon into subsidiary biosynthetic pathways, such as the pentose phosphate pathways (PPP). The PPP produces ribose-5-phosphate for DNA/RNA synthesis and NADPH for fatty acid synthesis and glutathion reduction, and represents an alternative anabolic pathway to the preparatory phase of glycolysis downstream of hexokinase.

In a recent elegant publication by Kathagen et al, the authors studied adaptive mechanisms in glioblastoma stem like cells in response to acute and chronic hypoxia as well as oxygenation. By microarray gene expression analyses they identified an as yet unknown oxygen level-dependent metabolic switch between the PPP and glycolysis. While PPP enzymes were found to be highly expressed under normoxic conditions, acute hypoxia entailed downregulation of PPP enzymes concomitant with upregulation of glycolysis enzymes, most significantly those of the preparatory phase. The most strongly inversely regulated enzymes of glycolysis and the PPP were aldolase C and glucose-6-phosphate dehydrogenase, respectively. Metabolic flux analyses using [1,2-¹³C₂]-D-glucose confirmed the switch from the PPP to glycolysis in response to hypoxia. Interestingly, despite their inducibility by hypoxia, enzymes of the preparatory phase of glycolysis downstream of hexokinase were found to be generally downregulated in glioblastomas compared with normal brain (by REMBRANDT and tissue microarray analyses), whereas enzymes of the pay-off phase of glycolysis and the PPP were upregulated. In glioblastoma tissue in situ, expression of PPP enyzmes was particularly strong in highly proliferative areas but downregulated in hypoxic pseudopalisades, whereas enzymes of the preparatory phase of glycolysis displayed an inverse expression pattern. Functional studies revealed that normoxia and elevated PPP enzyme expression were associated with increased cell proliferation, whereas hypoxia and elevated glycolysis enzymes were associated with enhanced cell migration.

These findings extend Warburǵs observation that tumor cells predominantly utilize glycolysis for energy production, by suggesting that PPP activity is elevated in rapidly proliferating tumor cells but suppressed by acute severe hypoxic stress, favoring glycolysis and tumor cell migration to protect cells against hypoxic cell damage. With regard to metabolic targeting, these findings imply that individual tumors are metabolically heterogeneous, with different metabolic pathways dominating in distinct regions. This result implies that simultaneous therapeutic strategies may be required to target the different “metabolic compartments”. Deciphering the interplay of the PPP with other relevant metabolic pathways and enzymes, such as hexokinases (eg HK2), pyruvate kinase izoenzymes (PKMs) is important, in addition to understanding the role altered GBM metabolism plays in resistance to therapy.

Mutational analysis reveals the origin and therapy-driven evolution of recurrent glioma

Diffuse gliomas inevitably recur following surgery and maximal safe resection. Treatment approaches for tumor recurrence in GBM is a major clinical problem for this disease. While advances in the molecular characterization of genomic potentially targetable signatures of newly diagnosed lower grade gliomas have been made, little is known regarding the molecular characterization of recurrent gliomas, and a critical question is whether the mutational profile of the primary tumor is mirrored at recurrence. The authors (Johnson et al) performed exome sequencing on 23 matched primary recurrent glioma pairs to characterize concordance of genomic alterations, and also correlated results with treatment history. Interestingly, a substantial proportion of mutations identified in the primary tumor were lost at recurrence. Overall, an average of 33 mutations were identified in the 23 primary tumors, and collectively only 54% of these were also detected at recurrence. Key driver mutations in gliomas including IDH1, TP53, and ATRX were retained at recurrence in most (but not all) recurrent tumors. The remaining mutations were found in only 1 of the matched primary-recurrent pairs (referred to in the article as “private” mutations). In addition, the authors confirmed findings of others in cases from patients with prior temozolomide treatment. Six of 10 temozolomide-treated tumors were hypoermutated at recurrence and tended to show driver mutations affecting the RB and mTOR pathways. Overall, the findings highlight an unexpected complexity of recurrent gliomas, offering a note of caution in the reliance of the genomic signatutr of the primary tumor as a means to determine therapy in the recurrent setting.

Mathematical modeling to test and optimize radiation dose schedule in glioma

The authors of this recent paper focussed on modeling radiation response to optimize radiation schedules that improve efficacy in a mouse model of PDGF-driven glioma. Their model considered two separate populations of glioma cells: stem-like resistant cells (SLRCs) and differentiated-sensitive cells (DSCs). One feature of their model hypothesized plasticity of radiation sensitivity and specifically, that radiation causes a proportion of radiosensitive cells to revert to the radioresistant state. The authors use a mouse model of PDGF-B-induced tumors in Nestin-tv- a;E2f1-Luc mice using the established RCAS system. Their mathematical model assumes that a subset of cells can switch from an SLRC state to a DSC state and vice versa. While the intricacies of the mathematics in this paper are beyond the expertise of this reviewer to evaluate, the authors build on an established linear quadratic model, which is accepted in the radiobiology literature due to its close agreement with experimental results. Mathematically, they account for a variety of biological processes, including a per-Gy production of lethal DNA kesions, kinetics of clonal expansion of cells, kinetics of quiescence, and kinetics of interconversion of cells from the 2 states. Within constraints relevant to clinical care, the developed an initial iteration of an optimum radiation schedule (“optimum-1” and shoed that mice lived significantly longer on this schedule compared to standard therapy (2 Gy per day for 5 days). They then consider hyper- and hypofractionation schedules, neither of which show improvements in humans compared to standard therapy. While their original model predicted good outcomes for the hyperfractionated schedule, experimentally this was not observed, and the authors used this information to improve their model, by including information that the fraction of cells rapidly acquiring resistance, now depended on the time elapsed since the previous dose of radiation. Since the 2 models (original vs. new time-dependent model) had different predictions on SLRC enrichment following radiation, the authors test this experimentally and found the time-dependent model more accurate. A new schedule (“optimum-2”) was then developed from the time-dependent model and while the median survival was improved compared to optimum-1, the difference was not statistically significant. The authors speculate that the optimum-1 and optimum-2 schedules increase mouse survival by converting glioma cells from the fast-growing radiosensitive population to the slow-growing radioresistant population. Finally, they do a “thought experiment” assuming a model with no reversion of cell types, but find that this model would contradict their experimental findings, adding validity to the notion that ionizing radiation encourages rapid reversion of a subset of glioma cells to a radioresistant stem-like state. What are the implications from this line of investigation? Overall, while much additional work would be required to translate these findings to the clinic, this work opens the door to consider the plasticity of cell phenotypes (rapidly growing non-stem cell like cells versus slow growing stem-like cells) to optimize radiation dose schedules in glioma.