Dynamics of Mitochondrial Redox and Energy Networks: Insights from an Experimental–Computational Synergy - Systems Biology of Metabolic and Signaling Networks - page 132

Under metabolic stress, however, an imbalance between production and scav-
enging can take ROS over a threshold resulting in strong coupling between
mitochondria through RIRR (Figs. 5.4 and 5.8 ) (Aon et al. 2003 ; Zorov
et al. 2000 ). When subjected to these challenging stressful conditions, the mito-
chondrial network spontaneously organizes into a synchronized cluster with a
dominant low-frequency high-amplitude oscillation that spans the whole cell
(Aon et al. 2004a , 2006b ). The RIRR mechanism is effective locally, but long-
range synchronization is due to the attainment of criticality by ~60 % of
mitochondria that form a cluster, the “spanning cluster”, across the cardiomyocyte.
Mitochondria belong to the “spanning cluster” when ROS (more specifically, O 2 . )
attain a threshold in the matrix, after which ΔΨ m depolarization ensues, triggering a

similar response in neighboring mitochondria.
Underlying the inverse power law behavior observed experimentally in the
power spectrum is the inverse relationship found in the model simulations of
the double log plot of amplitude versus frequency (Fig. 5.11 ). Two key factors
contribute to this dependence—the superoxide dismutase (SOD) activity and the
balance between the rate of ROS production and scavenging. In the oscillatory
domain, an increase in the SOD rate results in longer periods and higher amplitude
oscillations (Fig. 5.11 ).
We hypothesized that if, according to the experimental results, the mitochondrial
network were exhibiting a mixture of frequencies, then we should be able to simulate
the inverse power law behavior obtained by either PSA or RDA (see Box 5.4 ). To
test this hypothesis we simulated five different oscillatory periods ranging from
70 ms to 300 ms and one long period (1 min) oscillation (Fig. 5.11 ). A combination
of 80 % short-period and 20 % long-period oscillations allowed us to simulate the
inverse power law behavior observed experimentally by either PSA or RDA
(Fig. 5.12 ). This result demonstrated that mixing a relatively few (six) periods of
limit-cycle type of oscillation is enough to explain our experimental data. Using a
similar approach, we were also able to simulate the transition from physiological to
pathophysiological behavior (Aon et al. 2006b ). This transition, according to
simulations, and in agreement with experimental data, is effected when at least
60 % of the mitochondrial network dynamics is dominated by the long period, high-
amplitude

ΔΨ m oscillations (Fig. 5.12 ) ((Aon et al. 2006b ), and their Supplementary

Material).

Box 5.3: Fast Fourier Transform (FFT) Is a Key Mathematical
Procedure Utilized in Power Spectral Analysis (PSA)
FFT of a time series enables the statistical determination of the power
(equivalent to the amplitude squared) of each frequency component of a
signal. When this analytical procedure was applied to the time series of
ΔΨ m from the mitochondrial network, a large number of frequencies in
multiple timescales became evident (Aon et al. 2006b , 2008b ). Thus,

(continued)