Circular causality, another secret of life – on the occasion of Philip Ball's How Life Works

 

 guest blog post by Rolie Barth

 

Philip Ball has published an impressive study searching for the most important processes controlling the long way from genes and cells to genetic networks, tissues, organs and bodies. This implies he is talking about eukaryotic cells and their development and evolution. I start this blog with a short summary of the general picture he presents in his book. Then I discuss the concept of circular causality in more detail using a diagram to explain the multilayered nature of developmental processes. Finally, I will give an example of such a process.

If I understand Ball correctly, he is arguing for a view of life in which – in my own words – circular causality plays the leading role. Unfortunately, he himself did not mention this concept explicitly. But anyway, he frequently writes about top-down and bottom-up causality between the different layers of organization in an organism.
Genes play an important role in a multitude of complex processes, characterized not only by linear causality (from bottom to top, from genes to tissue) but also by feedback loops between the different levels of life: genes, networks, cells, tissues, bodies and groups of organisms in ecological systems. All these levels of organization are complex systems that are mainly characterized by (my words) the laws of biophysical phenomena whose dynamical behavior can be understood using the chaos theory. Even though Ball does not mention this theory (well, he does mention the catastrophe theory once), he often uses terminology from chaos theory, in particular: attractors. He describes attractors as valleys or basins in a landscape of possibilities. They are called attractors because the final state of a dynamical system is, as it were, drawn to one of those valleys (one of the possible final states).
To be honest, I might not have recognized this all-round scheme if I had not written about it in part II of my book De kosmos en het leven, een Meesterwerk, 2021 (The Cosmos and Life, a Masterpiece). 

Thinking about complex processes from the perspective of circular causality, the question of what is the most important part of the system actually becomes irrelevant. If we want to find an organizing center, I would talk about cell-centered life.
Considering the importance of the concept of circular causality one could speak about the third secret of life next to the DNA-code (first) and allostery (second, according to Jacques Monod). 

 

What is circular causality?

Many, many biological processes, as well as physical processes, are characterized by feedback loops between parts of a system. To clarify my point, think of the temperature control system of your home heating system. The room temperature is measured and when it is lower than the desired temperature the control unit sends a signal to the boiler to start burning. Thus heating the water in the radiators, resulting in an increase of the room temperature until it is higher than the desired temp. Then the control unit switches off your boiler, and so on. So, information about the room temperature is used by the control unit to regulate the boiler and thereby the room temperature – and thus the causal loop is closed.
Because this feedback system is trying to minimize the difference between actual and desired temperature we call this a negative feedback. You find this kind of controlling systems on many levels of your body, from gene regulation to homeostasis (all kind of processes like controlling body temperature and glucose concentration in the blood). Glucose concentration is maintained within some variation on a certain level by two feedback systems, one controlled by insulin to prevent glucose concentration getting too high and the second to watch that it doesn’t get too low, controlled by glucagon.  

Negative feedback is stabilizing all kinds of biological processes. But organisms also use positive feedback for example to generate neural impulses. Positive feedback means that a closed loop enhances an incoming signal up to a maximum level. The best known example is the singing around (buzzing) of a microphone signal, when some sound or even noise from the speakers is picked up by the microphone. Global warming due to increase of CO2-concentrations is a process with positive feedback too. More CO2 in the atmosphere increases the global temperature resulting in permafrost to melt partially. This leads to a release of greenhouse gasses CO2 and methane. Causally this is a self-amplifying process and therefore a positive feedback loop.
 

Combinations of feedback loops and moving particles

Both electronic systems and biological networks may have combinations of negative and positive feedback loops. The most simple version creates electrical signals or molecule concentrations that vary in a waveform during time. One of them is our internal biological clock. More complex combinations of positive and negative feedback loops may result in a kind of memory, switching the output level of the system to a permanent high level after an initial, short input signal.
Much more fascinating things happen when positive and negative feedback loops are part of systems with a large collection of moving particles, for example biomolecules in a growing tissue. In physical, chemical, biological and ecological systems interactions between ‘particles’ can give rise to all kind of patterns or structures. Examples are: sand ripples, mackerel clouds (physical), the Belousov–Zhabotinsky reaction (chemical), skin patterns, limb structures (biological) and large scale stripes and spots in semi-arid areas (ecological) etcetera. To illustrate what is meant we will take a look the way skin patterns can be described by mathematical models. These were first described by Alan Turing in 1952 and applied to biological processes by Hans Meinhardt and Alfred Gierer since 1974. Philip Ball discusses Turing patterns in chapter 8 (p. 307-328, printed version) which shows many parallels to my own book chapters 13 (p. 197-201) and 15 (p. 264-273).
Turing patterns like zebra stripes are examples of self-organization. At the start of such processes two or more species of molecules are randomly distributed in the fluid between cells of a developing tissue. So the initial situation is one without any pattern. Specific interactions between the two types of molecules, called morphogens, ánd differences in diffusion velocities may lead to all kinds of patterns (see p. 198 De kosmos en het leven).
In these systems, one kind of morphogen is working as an activator: it promotes the production of its own type of morphogens. The activator therefore has a self-amplifying effect. In addition, the activator promotes also the production of inhibitor morphogens. The activator therefore acts as a catalyst for production of new activators and inhibitor molecules. On the other hand, the inhibitor molecules de-activate the production of new activator molecules. So we see two kinds of feedback: (1) positive feedback of the activator (green, figure 2 below) and (2) negative feedback of the inhibitor (red). 

Figure 2

Pattern formation occurs because morphogens A and B diffuse through the fluid between cells, with different speeds. A random, local increase of concentration activator A, communicates via cellular mechanism to gene A to raise its production, ultimately leading to more activator molecules and more inhibitors in the intercellular space. That would make no difference except when the burst of inhibitors B spreads much faster than the activators A. The result of this is a slowing down of the production of gene A in cells further away from the initial fluctuation (see figure 3). This leads to a decrease in activator concentration in these areas. So the initial random fluctuation leads to the formation of a wave pattern in the intercellular space, finally resulting in a fixed and stable pattern, stripes, dots or more complicated shapes.

Looking at these processes two remarkable things show up. The two populations of many morphogens manifest a collective behavior in the form of waves, extending across a whole tissue. And these waves are changing the behavior of genes, far down in the cell nucleus. So there is a top-down causation from pattern to genes. Just like the traffic jam as a whole restricts your behavior as an individual car driver. And of course it also works the other way around – individual genes are influencing the wave form. And furthermore those waves are phenomena resulting from physical mechanisms. So, in these processes, gene expression is controlled by physics! In my book I suggested to call this extra-genetic regulation.

Fig. 3 Concentrations of activator (green) and inhibitor (red)
versus position at different times.
A random fluctuation in the activator concentration
starts up a process of wave formation, resulting in a stable pattern.

Mathematical models can be used to simulate the processes of limb and finger pattern-formation. Finger patterns correspond to a wave of a morphogen called SOX9. At positions where the concentration of SOX9 rises above a genetically defined threshold this morphogen induces cell differentiation: mesenchymal stem cells transform into chondrocytes (cartilage cells). So the wave pattern in combination with gene activity induces a condensation of this pattern. 

What does it all mean?

This kind of morphological processes has been recognized in many different stages of embryonic and fetal development. In combination with physical mechanisms, positive and negative feedback loops between genes and their products can lead to pattern formation starting from a structureless situation. One could wonder what is the central feature leading to pattern formation? The genes, the morphogens, the laws of motion, the feedback loops, the cells? The answer may be strange: none of them is thé central part of the system. All of them are indispensable, be it that positive feedback is the generating ‘force’ of the process.
One of the most important conclusions is that genes neither program for a pixel like description of biological patterns nor for a kind of positional information – where genes are coding at which position a stripe of a pattern should start and end (as theorized by Lewis Wolpert). No, genes are coding for processes of pattern formation. Genes are not creating patterns, they are providing suitable parameters of pattern formation processes, which are physical of origin and therefore universal. Mutations may tune gene interactions in such a way that patterns and structures in the course of generations are optimally robust and flexible.
In short: genes don’t code for pattern pixels, nor for pattern global positions but for a pattern process. This means Ball is right in saying that genes don’t code for a blueprint. But genes located on DNA strings, transfer much of the information needed for pattern formation to next generations. So genes are indispensable as the memory of life and as tools for the variation of life forms. 

In many chapters of his book, Philip Ball demonstrates that these kind of processes can be identified at all levels of biological organization: genetic networks, cell fate decision, tissue formation, pattern and structure formation.

Final conclusion: Philip Ball has written a very illuminating book about How Life Works. I read a lot of new interesting things about known biological processes. I agree with his frequently repeated statement that genes don’t code for a blueprint of structures and shapes. The analysis of complex systems driven by feedback loops shows that genes work in combination with physical mechanisms for the formation of biological structures. Ball is right to emphasize that there is no linear route from genes to pattern. Causal interactions are running up and down, from pattern to genes and the other way around. But these multilevel interactions do not deny the importance of genes but emphasize them. Unfortunately, Ball does not give much attention to this fact, sometimes actually denying this. Since the causality of these formation processes is circular in nature it is impossible and unnecessary to discuss which of the components is most important. A gene centered view is ruled out by this view of life. And because feedback loops are essential for so many biological processes, circular causality may be called the third secret of life.

In a future blog I hope to discuss some chapters of Ball’s book demonstrating the importance of circular causality during embryonic development. 

 

References

Allostery: Jacob, Francois en Jacques Monod, ‘Genetic regulatory mechanisms in the synthesis of proteins’, J Mol Biol. 3, 1961, p. 318-56.

Feedback loops: Xiao-Jun Tian et al., ‘Interlinking positive and negative feedback loops creates a tunable motif in gene regulatory networks,’ Physical Review E, 80, 2009, 011926-1-8.

Feedback loops: Dong-Eun Chang et al., ‘Building biological memory by linking positive feedback loops,’ PNAS 107-1, 2010, p. 175–180.

Morphogenesis: Allen Turing, “The chemical basis of morphogenesis”, Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 237, 1952, p. 37-72.

Morphogenesis: Hans Meinhardt and Alfred Gierer, ‘Applications of a theory of biological pattern formation based on lateral inhibition’, Journal of Cell Science 15, 1974, p. 321-346.