Embedded Files
What I have done and/or want to do
My theoretical approach is primarily based on dynamical systems modeling. Following the spirit of physics to "start with the simplest case," my research focuses on single-celled organisms like E. coli and yeast.
Theory of Cell Death
What is the difference between the "living" and "dead" states? This question is undoubtedly one of the most important themes in biology. Recently, many universal laws (though that might be an overstatement) have been discovered regarding the "living state," especially concerning exponentially growing bacteria. However, "death" remains full of mysteries.
Can we even clearly distinguish between a "living state" and a "dead state" to begin with? Suppose, for example, we want to determine if an E. coli cell is alive or dead. Since E. coli do not age or have a predetermined lifespan, we have to expose them to stress—like starvation, drugs, or high temperatures—to kill them. When we do, they first stop growing. But this doesn't necessarily mean they are "dead."
Except in obvious cases like cell rupture (a), determining cell death typically requires transferring the cells to a growth-permissive environment and observing whether they resume dividing. While this method often works, it is well known that cells exposed to stress exhibit a certain lag time before resuming growth. Even within a clonal population with identical genes, this lag can vary from hours to days, and could potentially span years (b). Therefore, to strictly determine cell death using this method, we would theoretically have to wait forever.
Is there any way to determine if a cell is alive or dead without relying on the "wait and see if it revives" approach? In other words, is it possible to scientifically identify the "Point of No Return"—the metaphorical River Styx—beyond which a cell can never be revived (c)?
---
(Update: Dec 13, 2024) Recently, I developed a theoretical framework that defines "death" based on the concept of controllability. (The actual paper is quite mathematical, so an article on Nazology might be easier to digest for Japanese readers. The main explanation starts after the references and the writer/editor introductions.)
Roughly speaking, my proposal is to define "death" based on whether or not an operation exists that can return the cell state to a predefined "representative point of the living state" (d). Consider a plant seed: it doesn't move and doesn't appear to eat. If we judged life and death purely by biological activity, we might conclude it is "dead." But we don't think of seeds that way in our daily lives. Why? Because we assume that if we give it water, the seed will generally germinate and bloom. If we define the "blooming state" as the "representative point of the living state," then "giving water" is the operation I'm talking about.
Furthermore, in this paper, I didn't just propose a definition of death; I also constructed a method to determine whether such a recovery operation exists within a cellular metabolic model. This has made it possible to quantify the "River Styx" that separates the "living state" and "dead state" in these models. Because a cell's metabolic state is represented as a point in a high-dimensional state space in dynamical systems theory, I call this boundary the "Hyper-surface of Styx" rather than a one-dimensional "river."
When I wrote the first part of this text, I thought to myself, "I'm writing down this Big Picture, but it's probably impossible to actually achieve." However, I feel like I have finally taken the first step. That being said, there are still a mountain of fundamentally difficult problems to solve at this stage, both regarding the extensibility of the theory itself and its connection to experimental data. If you are interested in these topics, I would love to discuss them with you!
(a) I include methods like PI staining in these "obvious cases."
(b) The average division time of certain deep-sea microbes is said to be on the order of years. Strictly speaking, this isn't a "lag," but the implication for the waiting time before division is the same.
(c) Simply put, for cells that have fallen into a state like VBNC (Viable But Non-Culturable) due to stress, I want to be able to look at a specific quantitative measure and know: will they eventually become culturable again, or will they never return and simply die after a while? (Assessing assimilation capacity is one possible approach, though...)
(d) There is a reason I use the rather roundabout phrase "representative point of the living state": it's to avoid the philosophical debate over "what exactly is life." I plan to write about the context behind this in another article sometime soon.
Theory of Dormancy
There is a state between the "living" and "dead" states called the "dormant state" (though, of course, the cell is still alive). Cells transition into this state upon exposure to starvation or drugs. In this state, chemical reaction rates slow down significantly, meaning there is naturally almost no growth, but the rate of death is also heavily suppressed. It is literally a "sleeping" state. If stress continues to be applied, cells in this state will gradually die, so this is (probably) considered to be one step before the aforementioned "Point of no return." Because cells in this state become highly tolerant to antibiotics, the pharmaceutical industry has taken a great interest in it, and many genes and metabolites associated with dormancy have been studied experimentally.
This dormant state raises various questions. For example:
There are many ways to induce dormancy, and the genes expressed differ depending on the method. But are the dormant states induced by these different methods fundamentally different? Or, once a cell enters a "sleeping" state, are there universal laws that emerge regardless of the induction method?
Is the dormant state an inherent property that naturally emerges in autocatalytic systems? Or should it be considered an evolutionary "function" acquired by bacteria?
"Hibernation" in bears and mice, or "anhydrobiosis" in tardigrades and sleeping chironomids, serve a similar function to dormancy in the sense of "lowering activity to survive harsh environments" (a). But how much commonality do these phenomena, seen across different species, truly share beyond just being "somewhat similar"?
Keeping these questions in mind, I have been conducting theoretical research on the dormant state. For example, I have reported that the lag time before growth resumption increases with the square root of the time spent in the dormant state [1], the conditions under which the lag time to resume growth increases or decreases evolutionarily [2], and that a transition to a dormant-like state can occur solely through the dynamics of metabolic reactions, even without gene regulation [3]. This area is currently my main focus.
(a) I am not an expert on hibernation or anhydrobiosis, so please correct me if I say something odd.
[1] YH and Kunihiko Kaneko, Phys. Rev. X, 7, 021049, (2017)
[2] YH and Namiko Mitarai, PLoS Comp. Biol., 17(2):e1008655, (2021)
[3] YH and Namiko Mitarai, Phys. Rev. Res. 4, 043223, (2021)
Energy to Live
As another approach to tackling cell death and survival, there is also a route from the perspective of energetics. In the 1960s, a biologist named S. J. Pirt proposed the following equation based on experimental data:
J(μ) = m + μ/Y
This equation asserts a linear relationship between the rate of nutrient consumption (J) and the growth rate (μ). The inverse of the proportionality coefficient, Y, is called the maximum yield.
This equation was derived by culturing cells in environments with various nutrient concentrations, obtaining data on nutrient uptake and growth rates, and fitting them with a linear function.
The relationship between J and μ makes intuitive sense—you have to eat a lot to grow fast (a). However, what is particularly interesting is that when the growth rate is extrapolated to zero, the y-intercept is non-zero in many cases.
This means that even if a cell stops growing, it absolutely must consume nutrients. In other words, it indicates that a certain amount of nutrient consumption is required simply to "stay alive."
Reflecting this implication, the intercept m was termed "maintenance energy."
Although maintenance energy is a phenomenological parameter, it is used in many contexts due to the clear concept that "if a cell cannot consume this much energy, it dies." For example, it is used as a criterion to theoretically predict whether a specific gene knockout will be lethal.
However, maintenance energy is ultimately just an extrapolation of a linear relationship to a growth rate of zero. Because it is defined as the "remainder" left over from growth, we do not know what specific intracellular processes actually determine its value.
Furthermore, if we actually try to bring cells close to a growth rate of zero, they transition into a "dormant state," which breaks the linear relationship. Therefore, it is also considered difficult to determine the maintenance energy by artificially creating a state of zero growth.
Still, even if it is merely an extrapolation, maintenance energy is arguably exactly what Schrödinger meant by "negative entropy." I can't help but wish there was some way to give it a more active definition, rather than just calling it the "remainder from growth" (b).
I wrote papers [1, 2] around this topic during my master's, but it proved to be quite difficult, and I've sort of given up on it for now. If anyone has any good ideas, please let's think about it together!
(a) The fact that the relationship becomes linear is not trivial at all. When it is linear, it probably means that the metabolic flux distribution remains almost unchanged and only its overall magnitude scales with the growth rate. In fact, there are cases where the relationship takes a strictly convex shape. (b) Then again, defining something as a "remainder" can sometimes lead to incredible universality—like heat in thermodynamics (d'Q = d'W - dU)—so perhaps leaving it as a 'remainder' is fine after all. But in that case, there needs to be an equivalent to the quasi-static process in thermodynamics, right?
[1] YH and Kunihiko Kaneko, Phys. Rev. E, 90, 042714 (2014)
[2] YH and Kunihiko Kaneko, Physical biology, 13.2 (2016)
Page updated
Google Sites
Report abuse