System-based approaches as prognostic tools for glioblastoma

Abstract

Background: The evasion of apoptosis is a hallmark of cancer. Understanding this process holistically and overcoming apoptosis resistance is a goal of many research teams in order to develop better treatment options for cancer patients. Efforts are also ongoing to personalize the treatment of patients. Strategies to confirm the therapeutic efficacy of current treatments or indeed to identify potential novel additional options would be extremely beneficial to both clinicians and patients. In the past few years, system medicine approaches have been developed that model the biochemical pathways of apoptosis. These systems tools incorporate and analyse the complex biological networks involved. For their successful integration into clinical practice, it is mandatory to integrate systems approaches with routine clinical and histopathological practice to deliver personalized care for patients.

Results: We review here the development of system medicine approaches that model apoptosis for the treatment of cancer with a specific emphasis on the aggressive brain cancer, glioblastoma.

Conclusions: We discuss the current understanding in the field and present new approaches that highlight the potential of system medicine approaches to influence how glioblastoma is diagnosed and treated in the future.

Keywords: Apoptosis; Computational model; Glioblastoma; Molecular signatures; Network model; Numerical simulation; Precision oncology; Prognostic biomarker; Systems biology; Systems medicine.

Fig. 1

Newly-diagnosed tumors (n = 31) expressed higher protein concentrations of Apaf-1, Procaspase-3, Procaspase-9, SMAC and XIAP compared to specimens collected from recurrent patients (n = 15) in the GBM cohort. a Representative images of Western blot experiments. Each lane contains a unique patient tumor sample from newly-diagnosed or recurrent tumors as indicated. β-actin served as a loading control. b-f Normalized protein levels were converted to absolute concentrations (in μM, as required for inputting into APOPTO-CELL) by linear regression with known concentrations in HeLa cells [13, 17, 47]. Reference concentrations were previously determined in HeLa cell extracts with titrated concentrations of recombinant proteins [47]. Prior to pooling together protein quantifications for the de novo patients with those reported in [17], batch-effects in the measurements were removed. For each protein, the median concentration from the de novo newly-diagnosed samples was aligned to the median concentration measured in the newly-diagnosed specimens from [17]. Protein concentrations measured in tumor samples from de novo recurrent patients were also batch-corrected, but the scaling constants were computed based on median-aligning the newly-diagnosed samples only. Statistically significant differences between protein expression in newly-diagnosed vs. recurrent samples were examined by Mann-Whitney U tests

Fig. 2

Assessment of the prognostic significance of single proteins regulating caspases-dependent apoptosis. a-e Kaplan-Meier estimates for Apaf-1 (a), Procaspase-9 (c), SMAC (d) and XIAP (e) showed no statistical significant differences in PFS curves among patients grouped by protein expression (>median vs. ≤median, in black and gray, respectively). Patients expressing higher concentrations of Procaspase-3 (>median) had longer PFS compared to those with low levels (≤median), (log-rank P = 0.049, b)

Fig. 3

APOPTO-CELL model as a personalized risk assessment tool. a and b Patient-specific temporal profiles for substrate cleavage predicted by APOPTO-CELL (n = 46, a). The substrate cleavage reached at 15 min was deemed as the primary readout from the model simulations (b). Patients who did not cleave an amount of substrate of at least 80% were categorized as apoptosis-resistant (in red) whereas those above this threshold were classified as apoptosis-sensitive (in blue). c Association between apoptosis susceptibility predicted by APOPTO-CELL (SC ≤ 80% vs. SC > 80%, in red and blue, respectively) and type of tumor sample (newly-diagnosed and recurrent, light and dark shades, respectively). d-f Kaplan-Meier estimates of PFS in GBM patients categorized as apoptosis-resistant (n = 9, in red) or apoptosis-sensitive (n = 37 in blue) by APOPTO-CELL for the whole cohort (d) and stratified by type of tumor sample (newly-diagnosed and recurrent in e and f, respectively). P-values were determined by log-rank tests

Fig. 4

APOPTO-CELL can conduct in silico clinical trials for targeted apoptosis sensitization with SMAC mimetics. a-c Patient-specific dose-response curves simulated by APOPTO-CELL depicting the relationship between apoptosis susceptibility and pharmacological intervention. Apoptosis susceptibility is represented by the amount of simulated substrate cleavage reached at 15 min from the simulation start. Left hand-side of each plot before gap highlights basal apoptosis susceptibility (i.e. no administration of SMAC mimetics). Concentrations of SMAC mimetics tested in silico where selected to span the physiological doses administered in real-world clinical trials (1 nM - 1 μM). Patients were deemed “responsive to standard therapy” if classified as apoptosis-sensitive in simulations without any SMAC mimetics intervention (n = 37, a). Conversely, patients predicted to have apoptosis impairment in basal settings were deemed “responsive to only standard therapy and SMAC mimetics” (n = 3, b) or “non-responsive to standard therapy and SMAC mimetics” (n = 6, c) if administration of SMAC mimetics could induce (or not) re-sensitization, respectively