Role of compartmentalization on HiF-1α degradation dynamics during changing oxygen conditions: a computational approach

Abstract

HiF-1α is the central protein driving the cellular response to hypoxia. Its accumulation in cancer cells is linked to the appearance of chemoresistant and aggressive tumor phenotypes. As a consequence, understanding the regulation of HiF-1α dynamics is a major issue to design new anti-cancer therapies. In this paper, we propose a model of the hypoxia pathway, involving HiF-1α and its inhibitor pVHL. Based on data from the literature, we made the hypothesis that the regulation of HiF-1α involves two compartments (nucleus and cytoplasm) and a constitutive shuttle of the pVHL protein between them. We first show that this model captures correctly the main features of HiF-1α dynamics, including the bi-exponential degradation profile in normoxia, the kinetics of induction in hypoxia, and the switch-like accumulation. Second, we simulated the effects of a hypoxia/reoxygenation event, and show that it generates a strong instability of HiF-1α. The protein concentration rapidly increases 3 hours after the reoxygenation, and exhibits an oscillating pattern. This effect vanishes if we do not consider compartmentalization of HiF-1α. This result can explain various counter-intuitive observations about the specific molecular and cellular response to the reoxygenation process. Third, we simulated the HiF-1α dynamics in the tumor case. We considered different types of mutations associated with tumorigenesis, and we compared their consequences on HiF-1α dynamics. Then, we tested different therapeutics strategies. We show that a therapeutic decrease of HiF-1α nuclear level is not always correlated with an attenuation of reoxygenation-induced instabilities. Thus, it appears that the design of anti-HiF-1α therapies have to take into account these two aspects to maximize their efficiency.

Figure 1. Molecular network of HiF-1α regulation.

Sketch of the molecular network considered in the model.

Figure 2. Experimental and simulated HiF-1α degradation curves.

Blue: experimental degradation curve determined by Moroz et al. (2009) after the addition of cycloheximide. Solid line: mathematical bi-exponential law determined with experimental data. Dotted lines: uncertainty on the bi-exponential law. Red: simulated degradation curve. The total HiF-1α level (free and complexed, cytoplasmic and nuclear) is plotted as the percentage of its initial value.

Figure 3. Accumulation of HiF-1α under hypoxia.

Nuclear (red curve) and cytoplasmic (blue curve) equilibrium levels of HiF-1α are plotted as functions of the pVHL/HiF-1α complex formation rate (). varies over a [1 100] range (A), and a [0.1 10] range (B).

Figure 4. HiF-1α accumulation and localization as a function of hypoxia intensity and pVHL export rate.

A. The total (nuclear+cytoplasmic) HiF-1α equilibrium level, normalized to its value obtained with the default parameters, is plotted as a function of the pVHL-export rate () and of the pVHL/HiF-1α complex formation rate (). To the left of the red dotted line, total HiF-1α significantly increases with respect to the default value. B. Ratio between nuclear and cytoplasmic HiF-1α levels as a function of and . To the left of the red dotted line, HiF-1α accumulates in the nucleus.

Figure 5. Simulation of the temporal evolution of HiF-1α during a hypoxic event.

From the red curve to the black curve: simulations for increased levels of hypoxia, with  = 10, 5, 1.

Figure 6. Effect of a hypoxia/reoxygenation sequence on nuclear HiF-1α level.

A. The level of nuclear HiF-1α is plotted during two 30 h hypoxic events of different intensities, followed by reoxygenation. Red:  = 10. Black:  = 1. The dotted blue line corresponds to the HiF-1α concentration in normoxia. B. Response to reoxygenation in our compartmentalized model (red curve) and in a non-compartmentalized model of HiF-1α regulation (black curve). The nuclear HiF-1α was normalized to its value at normoxia

Figure 7. Consequences of a pathological overexpression of HiF-1α.

A. Accumulation of HiF-1α in mutated cells, compared to normal cells. The ratio [HiF-1α]/[HiF-1α] is plotted as a function of (HiF-1α synthesis rate) and (pVHL export rate). The black dotted square indicates the location of the peritoneal cancer studied by Yoshikawa et al. (2006) , characterized by a 2-fold increase of HiF-1α protein and a 3-fold increase of HiF-1α mRNA. B. Effects of reoxygenation on normal cells, and mutated cells. Reoxygenation after hypoxia is simulated, and the evolutions of nuclear HiF-1α concentration are plotted in the case of normal ( = 1000, red curve), and mutated cells ( = 5000, black curve). The dotted lines represents the normoxic equilibrium levels in each case.

Figure 8. Consequences of pVHL mutations.

Accumulation of HiF-1α in tumor cells with mutated pVHL. The value of R =  [HiF-1α]/[HiF-1α] is plotted as a function of (pVHL/HiF-1α complex formation rate) and (pVHL export rate). The white line limits the space of the RCC cancer cells studied by Wiesener et al. (2001) .

Figure 9. Effects of siRNA therapy in pVHL mutated cells.

Nuclear Accumulation of HiF-1α in the tumor case, without or with therapy. We calculated the value of  =  ([HiF-1α]/[HiF-1α]) (in the nucleus) as a function of (complexation rate) and (pVHL export rate). A. Without siRNA therapy (HiF-1α synthesis rate  = 1000). B. With siRNA therapy ( = 200).

Figure 10. Effects of siRNA therapy on HiF-1α level evolution after reoxygenation.

Reoxygenation after a hypoxic event is simulated for the HiF-1α synthesis rate  = 1000 (black curve) and  = 200 (red curve). The blue dotted line represents the normoxic level for a normal cell.

Figure 11. Effects of an increased proteasomal activity on nuclear HiF-1 level of -mutated cells.

The value of  =  [HiF-1α]/[HiF-1α] is plotted as a function of (pVHL export rate) and (HiF-1α synthesis rate). Left: non-treated cells. Right: result after a 10-fold increase of pVHL-dependent degradation of cytoplasmic and nuclear HiF-1α degradation rates ( and ).

Figure 12. Effects of an increased proteasomal activity on HiF-1α evolution after reoxygenation.

Reoxygenation after a hypoxic event was simulated for non-treated cells (black curve), and with a 10-fold increase of pVHL-dependent HiF-1α degradation rate (red curve). The cell considered is a tumor cell, with a deregulation of HiF-1α expression ( = 3000) The blue dotted line represents the normoxic level for this cell.