Reconstructing the molecular life history of gliomas

Abstract

At the time of their clinical manifestation, the heterogeneous group of adult and pediatric gliomas carries a wide range of diverse somatic genomic alterations, ranging from somatic single-nucleotide variants to structural chromosomal rearrangements. Somatic abnormalities may have functional consequences, such as a decrease, increase or change in mRNA transcripts, and cells pay a penalty for maintaining them. These abnormalities, therefore, must provide cells with a competitive advantage to become engrained into the glioma genome. Here, we propose a model of gliomagenesis consisting of the following five consecutive phases that glioma cells have traversed prior to clinical manifestation: (I) initial growth; (II) oncogene-induced senescence; (III) stressed growth; (IV) replicative senescence/crisis; (V) immortal growth. We have integrated the findings from a large number of studies in biology and (neuro)oncology and relate somatic alterations and other results discussed in these papers to each of these five phases. Understanding the story that each glioma tells at presentation may ultimately facilitate the design of novel, more effective therapeutic approaches.

Keywords: Glioma; Gliomagenesis; Oncogenesis; Senescence; Telomerase.

Fig. 1

Model of the molecular life history of gliomas, prior to becoming clinically manifest. The temporal sequence of events can be subdivided into five phases (I–V) represented in different colors. a The number of dividing cells (or proliferation rate) across each phase. Proliferation peaks towards the end of growth phases and dips going into senescence phases. b The tumor mass across each phase. Tumor mass increases exponentially during growth phases and logarithmically during senescence phases. c Telomere length across each phase. Telomere length over time follows a pattern that is inverse to tumor mass. d Cell doubling diagram indicating the growth barriers (senescence phases) and resulting selection bottleneck. e Somatic alterations associated with different phases in gliomagenesis. The timing of each event is indicated on the x-axis of panel C. Genomic instability events are accumulated during phase III–IV. Of note, this model is a simplified representation of true gliomagenesis. The x-axis is not drawn to scale, in part because the duration of the phases likely varies from cell to cell and between various tumor types. Furthermore, the position of the curves is arbitrary as cells in a tumor may not be in sync. BFB breakage–fusion–bridge, DM double minute, ALT alternative lengthening of telomeres

Fig. 2

Process diagram indicating the p16^INK4a/p14^ARF–RB–p53 pathway in normal conditions. Disruption of one or multiple components through mutation or copy number change may prevent or suppress the onset of senescence. Various stimuli use different routes to activate the senescence response, leaving compensatory mechanisms in place in case components fail. For example, if oncogene-induced senescence is repressed via CDKN2A/B inactivation, DNA damage and telomere shortening could still trigger replicative senescence via ATM and ATR. CDKs cyclin-dependent kinases (e.g., CDK2), MDMs murine double minutes (e.g., MDM2)

Fig. 3

Genomic instability related to telomere stress. a Schematic illustrating BFB cycles. Following a single BFB cycle, daughter cells are left with unequal DNA content, leading to a deletion in A and an amplification in B. BFB cycles may also involve fusion of non-sister telomeres (not shown). b Repetitive BFB cycles form palindromes demonstrating high intra-segmental homology. This can lead to intra-chromosomal fusions and formation of double minutes. c Following chromothripsis segments can be rearranged, lost or circularized. TSG tumor suppressor gene, OG oncogene, DDR DNA damage response, DM double minutes, BFB breakage–fusion–bridge

Fig. 4

a Integration of a clonal evolution and cancer stem cell model for gliomagenesis. This model assumes that sequential mutations and selection pressure drive the evolution of cancer stem-like cells. At the same time, these stem-like cells may give rise to more differentiated (i.e., phase IV) offspring that may divide further but rapidly become growth arrested. b According to this model these cells may be senescent and contribute to the cancer phenotype by eliciting a microenvironment response via SASP. SASP senescence-associated secretory phenotype; ILs interleukins; CXCLs chemokines (C–X–C motif); CCLs chemokines (C–C motif)