The cycle that ends the cycle
February 25, 2011 § Leave a comment
Senescence is the sign of a cell that’s given up. Cells that have reached the “Hayflick Limit” — usually 40-60 divisions for non-transformed cells in cell culture — can no longer divide; this observation, which was a surprise at the time, led to the coining of the term cellular senescence.
One reason that cells become senescent is that the telomere region can’t be perfectly copied during replication, so telomeres get shorter with every successive division. Another route to senescence is via DNA damage. Both telomere shortening and DNA damage activate the DNA damage response pathway, which includes our friend p53 (“the guardian of the genome”) and the cyclin-dependent kinase p21. But senescence is a lot more complicated than “just” a response to DNA damage. Microarray profiling shows that the expression of many genes changes when cells become senescent, and that different genes show changes in different cell types. Senescence is an important anti-cancer mechanism, holding back the progression of precancerous prostate cells and melanomas.
One of the puzzles of senescence is how senescence becomes irreversible; there has been no very obvious explanation for this to date. Jeremy Purvis pointed out a recent paper (Passos et al. 2010. Feedback between p21 and reactive oxygen production is necessary for cell senescence. Mol Syst Biol. 6 347 doi: 10.1038/msb.2010.5) that provides evidence for a rather vicious kind of vicious cycle: DNA damage (or telomere shortening) leads to the production of reactive oxygen species (ROSs), which leads to DNA damage.
A couple of papers before this one had drawn a connection between ROSs and senescence (and even suggested that there might be a positive feedback loop for ROS production), but how DNA damage might trigger ROS production was unclear. To try to sort this out, Passos et al. first looked at the kinetics of ROS production in cells becoming senescent. Whether the senescence was induced by DNA damage or by telomere shortening, what they saw was a delayed response: ROS production increased about a day after the cells became senescent. At the same time, mitochondrial membrane potential went down, indicating some kind of functional damage to the mitochondria. To ask how the DNA damage signaling pathway might be involved in triggering ROSs they knocked down the expression or function of p53, p21 and several other genes using siRNA, small molecules or antibodies, while monitoring DNA damage and reactive oxygen species production. Whichever gene they targeted, the number of DNA damage foci and the production of ROSs always moved in the same direction — apparently confirming that there was a tight connection between the two events, but not really helping to define the pathway. An important clue from this set of experiments was that inhibiting TGF-beta — which is known to be able to induce production of ROSs, perhaps via mitochondrial damage — not only reduces ROS production but also reduces the number of DNA damage foci. If DNA damage induces TGF-beta, this could lead to ROS production and hence DNA damage, completing the cycle and perhaps helping to explain why the relationship between DNA damage and ROS production is so hard to tease apart.
How might you get from p53 and p21 to TGF-beta? Passos et al. combed through several collections of “interactome” data, looking for potential paths (reasonably short ones) from p21 to TGF-beta. They found several, and all of the ones that looked most interesting went through GADD45A. This protein is a direct target of p21 which signals to a mitogen-activated protein kinase implicated in TGF-beta induction, MAPK-14 (which can also trigger ROS production, by the way). Their in silico analysis led them to define an eight-step pathway as the most likely path between DNA damage and TGF-beta (hence, ROS production). When they looked at microarray analyses they found that all the putative members of the pathway were coordinately regulated during the transition to senescence in fibroblasts, giving some support to the idea that this set of gene products represents a real pathway.
So there’s a plausible route from DNA damage to ROS production, and we know that ROSs can cause DNA damage. This is nice, but can we really get a self-perpetuating cycle out of the behavior of this circuit? The authors used a stochastic model to ask whether the amount of ROSs produced by DNA damage would be enough to cause enough DNA damage to maintain ROS production. Using realistic assumptions, the answer is yes: the kinetics of the production of DNA damage foci, p21 activation, and several other parameters seem to be quite well explained by a model of a linear pathway from DNA damage to p21 to TGF-beta to ROSs, and then back to DNA damage. And to check the relevance of their findings in vivo, the authors looked at telomerase mutant mice. Late generations of such mice usually have a significantly shortened lifespan; given the authors’ hypothesis, knocking out p21 should increase their lifespan, and it does. More important for the ROS part of the hypothesis, the p21 knockout mice also show decreases in signatures of oxidative damage in various situations where senescence is important.
So can senescence be prevented by blocking this loop? Yes and no. Yes, if you block the p21 to ROS pathway within one week after the DNA damage (using either an inhibitor of MAPK14 or a free radical scavenger to mop up ROSs); no, if you wait longer. Rescued cells can re-enter the cell cycle and start growing again. It seems that the ROS/DNA damage cycle is essential to initiate the commitment to senescence, but after a certain length of time another mechanism kicks in and the ROS/DNA damage cycle isn’t needed any more. No clues yet about what that mechanism might be. ROS to DNA damage is an important — perhaps crucial — feedback loop, but it is not the only one. And remember, while you might be keen to block this DNA-damage-causing cycle if you’re concerned about aging, you might want to accelerate it if you’re worried about cancer. Everything’s a balancing act….
Passos JF, Nelson G, Wang C, Richter T, Simillion C, Proctor CJ, Miwa S, Olijslagers S, Hallinan J, Wipat A, Saretzki G, Rudolph KL, Kirkwood TB, & von Zglinicki T (2010). Feedback between p21 and reactive oxygen production is necessary for cell senescence. Molecular systems biology, 6 PMID: 20160708