Books and publications on the
interaction of systems in real time by A. C. Sturt
and the Human Immune System
by A. C. Sturt
A system dynamics model has been developed as a working hypothesis for bacterial infection in humans, based on the competing rate processes of bacterial multiplication and antibody production. The active entities are treated in chemical terms, with antibodies manufactured in response to displacement of an equilibrium. Erratic cloning of themselves by bacteria causes resistance to antibiotics. A process by which resistance may develop is proposed, analogous to natural selection.
Erratic cloning occurs not only in vitro but during the course of an infection. It is the immune system which ultimately kills off resistant bacteria. Antibiotics cannot kill them, because of their specificity; their role is to give the immune system a chance to develop and manufacture a suitable antibody. Antibiotics which merely slow down the rate of multiplication of intruder bacteria may also play a role.
The model of competing rates implies that the number of antibodies produced to combat infection must exceed the numbers of invading bacteria produced by multiplication. This is a problem if there is an induction period in immune system response while it develops a specific new antibody, which may take a day or two. Vaccination eliminates induction periods by providing a ready-made antidote. Mother’s milk, because it contains antibodies, may work by providing cover until a child’s immune system becomes operational.
The model confirms procedures for combating bacterial infection. The paramount requirement is to keep the number of invading bacteria to a minimum, both by cleanliness and by isolation. This minimises the number of resistant bacteria produced by multiplication at every instant after intrusion, and makes it more likely that the immune system will be able to produce enough of the right antibodies in time. Procedures must not damage the host of bacteria already present, because they limit the space which intruding bacteria need to grow, as well as performing valuable functions in their own right. Treatment then is with antibiotics targeted against intruders until the immune system completes the elimination.
The model emphasises the importance of in vivo testing, because treatment is an interaction with the immune system; antibiotics which are effective in the laboratory may not work well in the patient because of unforeseen effects on the immune system, on other bodily processes, on the resident bacterial populations etc, which are all an essential part of the cure.
The model suggests limits to the silver bullet approach to bacterial infections, because the bacteria that do the damage are a new “species” which may differ significantly from those used as targets during development. What is hopeful about the model is that the immune system makes its own silver bullets to order, given the chance.
Problems caused by erratic cloning may also apply to viruses and fungi, although the dynamics will be different. For agriculture part of the answer may be a return to more differentiated, less monocultural production, which is the equivalent of medical isolation.
Two independent discoveries were mentioned at recent Royal Society lectures which suggest that it may be useful to consider bacteria in terms of system dynamics, as well as in the usual biological and medical contexts. The first discovery was made some time ago by the nuclear physicist Szilard, who found that a single bacterium, nurtured under laboratory conditions, could apparently “evolve” (1). The second discovery made in much more recent work was that the human immune system could undergo changes “to meet a threat in a day or two” (2). This also amounted to “evolution”, taking place over a surprisingly short timescale in a creature which measures even minor evolutionary changes in terms of tens of millennia i.e. ourselves. Taken together the two discoveries suggest possible hypothetical models which could shed light on these “evolutionary” systems and their interaction. The basic principles of the models are such that they can all be put to the test by experiment and measurement.
Szilard’s discovery suggests that bacteria follow a different model from the accepted view of evolution by natural selection. Natural selection depends on variation in a population, from whatever cause, which produces individuals that differ from the population at large. Such variation may have an exogenous cause, such as cosmic radiation, but it is much more frequently caused by faulty reproduction of interacting individuals; something is missed out which ought to be there, or something is added by mistake. When complex individuals produce offspring, they do not simply make clones of themselves.
Most variants which result from the process of natural selection are less suited to survival in the environment than the original, and progressively die out through competition. However, a few may be more suited and prosper to the extent that they become a distinct species in their own right. Differentiation between variants may be exaggerated by changing environmental conditions, and eventually a variant may even displace the original in the new environment. In systems terms, survival means more efficient capture of inputs and processing them to outputs as growth, energy, disposal of waste etc.
However, a single bacterium isolated in a test tube must unambiguously belong to a single species with no variation, whatever relation it bears to other bacteria outside. Feeding the bacterium allows it to divide, so as to form clones. These clones produce more clones. If we discount intervention from outer space, any variation in the population of bacteria formed from the single bacterium must have its cause in the internal processes of reproduction of the bacterium. It cannot be a process of selection by environmental conditions i.e. natural selection, because the environment is stable.
Thus if a single isolated bacterium evolves, it is because bacteria are rather erratic at cloning themselves. Some are much better than others, because they are known not to have changed over a period of a thousand years or more e.g. typhoid. Others seem to be “evolving” and producing new “species” every year. The effectiveness of the internal mechanisms which control cloning must vary from species to species.
The significance of this analysis is that we ourselves are a most suitable temperature-controlled environment for growth with plenty of inputs etc, and so it must be assumed that what happens in the laboratory medium also happens in the human body.
This is quite different from the conventional view of bacteria, for example, that they are continually probing our weaknesses to cause mischief. If the hypothesis of poor cloning is correct, it suggests that bacteria may be considered as chemical entities with the unusual properties of cloning themselves. They are no more malicious than, say, oxygen molecules which intervene to form a layer of oxide on metal surfaces because they are all bathed in it, an advantage for some processes which affect man, but a disadvantage for others e.g. soldering. Another example might be water molecules, which are immediately adsorbed onto available surfaces, a vital necessity for natural processes, but a distinct nuisance for engineers when they are trying to make adhesive bonds.
The point is that when changes occur which produce bacteria that are undesirable in humans, it is not the probing of our weaknesses, but the natural result of disturbing systems which were in equilibrium. The results may be exaggerated for us, because the systems in question are dynamic, so that apparent stability results from the balance of opposing forces. If bacteria do not multiply each time with precision, equilibria may be upset. The flaw may be only one in a million, but one bacterium soon grows to a million, given the inputs.
The process by which this could occur is shown in Figure 1. It is implicit in the preceding argument that populations of bacteria are highly likely to be heterogeneous i.e. not every bacterium is absolutely identical to every other in the population, even though they are all apparently the same “species”. If a population is treated with a bactericide solution, a large proportion of its bacteria may be killed, but a proportion will remain which comprises those more resistant to the bactericide (after 1st treatment). If the more resistant bacteria are then fed with inputs, they are likely to produce a population of bacteria which is more resistant to the bactericide at the concentration used. The average resistance to bactericide has increased.
If this resistant population is treated in its turn to more concentrated bactericide solution, the process repeats itself (2nd treatment). Most bacteria will be killed, but a proportion will remain which can tolerate the higher concentration of bactericide. If these are provided with nourishment, they will produce a population which is yet more resistant to the bactericide, until eventually a population of bacteria is produced which is completely resistant to the bactericide at the concentration first used. The bactericide is in effect being used to produce “natural” selection.
Bacteria as living entities grow to fill the available space, provided there are enough inputs. However, if there are too many bacteria competing for space and inputs, they do not necessarily form a hierarchy which we would recognize. The bacteria which we consider dangerous, may be just another competitor for bacteria which do not threaten us at all. Thus the scope for colonisation and expansion may be limited by the populations of other species of bacteria which are already present. If all populations are treated with bactericide indiscriminately, and if some survive the treatment and have access to inputs, the result could be to leave the field to bacteria which are resistant, quite the opposite of what was intended. The cure would be worse than the disease.
The human immune system appears to be a very large resource comprising different species of active entities which deal with intruders into the system by locking onto and destroying them. Each species of active entity or antibody deals with a specific species of intruder, where specificity is conferred by stereochemical linkages. Each individual antibody appears to be used only once, because it is itself destroyed in its attack on an individual of an intruder species. This is a process which is going on continuously under normal circumstances.
If this is so, there must exist a means of making more antibodies of a species as required, and the question then arises: how does the immune system know that more antibodies of a particular species need to be made? The mechanism that suggests itself is again a chemical analogy, this time of chemical equilibrium, in which chemical species A and B combine to form compound AB, but AB also decomposes to form A and B. Where the equilibrium settles depends on the rate at which the two processes occur.
If the concentration of one component, say B, is reduced by some exogenous source, the relative rates of formation and decomposition of AB restore the equilibrium.
In this case information is fed back to the locus of manufacture by displacement of the equilibrium, which manifests itself as a decrease of concentration of a specific antibody. So as the concentration of a specific antibody is consumed by successful attacks on intruders, more is produced to restore equilibrium. This is the body as chemical reactor with a definite size. The biological question, as opposed to the systems question, would then be how such an equilibrium came to be established in the first place.
From time to time intrusions of a species occur for which there is no ready made antibody, because the antibody system has not met that particular type of intruder before. In this case there is no equilibrium to restore, which is where the second discovery described at the beginning comes into play, the evolution of the antibody system. The immune system produces new variants in a matter of days, possibly by targeting the intruder, possibly on a stochastic basis, and they increase in number to some equilibrium concentration. In either case, after an interval of time a variant will be produced which has the specificity required to destroy the new intruder. Its concentration will tend to grow, but it will also be continually reduced as it attacks the intruding entities. This will cause more be made, which continues the process of combating the intruder; in effect there will be a new species of antibody in the repertoire.
Combining these models, it can be seen that the outcome of such encounters for the human body will depend on the relative rates of production of intruder and antibody. If the number of antibodies produced is sufficient to outstrip the number of intruders, health is maintained. However, if the number of intruders grows faster than the antibody can be discovered and produced, the intruder manifests itself as pathogenic, and ill health results. The actual ratio of antibodies to intruders required to prevent infection depends on the efficiency of the process i.e. rates of mixing may mean that it needs several antibodies to track down each intruder, even though only one actually completes the destruction.
If this model provides a reasonable working hypothesis, it suggests that most of the mutations which infect animals occur during the process of replication within animals themselves, since they are the most suitable culture media of all. Thus humans are the sources of most of the bacterial infections which are pathogenic to them. Pathogens may be transferred from other humans, or they may be produced by poor cloning within a human body during formation of a population of bacteria.
If bacteria spray off variants stochastically as they clone themselves during the course of an infection, almost all populations of bacteria are likely to be heterogeneous. They can therefore be differentiated by change of environment e.g. addition of dyes, bactericides etc. Differentiation shows that bacteria in the population have a statistical distribution of properties, which will not necessarily be the same for all additives used to discriminate between individual bacteria. The discriminant which ultimately matters is susceptibility to species of antibody in the immune system.
In these competing processes, the immune system almost always wins, which is why we are normally healthy. However, occasionally variants are produced with which the immune system cannot cope rapidly enough e.g. the delay in “evolving” a new antibody allows the bacterial variant to reproduce itself in numbers too large for the immune system to match.
These are the conditions under which an antibiotic is administered, but by the preceding argument it is unlikely that an antibiotic could kill off an entire pathogenic population, because of differentiation of resistance. The nature of the process is that there will always be some, however small a number, which are resistant to the antibiotic. These would then go on to multiply in the culture medium which is the body.
Nevertheless, there is no doubt that antibiotics produce the desired results in most circumstances, even if they may never have seen the variants before. But if they do not destroy the resistant bacteria, the question is: what does? The corollary is that there must be a further process at work which completes elimination of resistant bacteria. This can only be the immune system. Thus the function of the antibiotic may therefore be to give the immune system time to respond, say by producing a new species of antibody.
The most encouraging aspect of this hypothesis is that the immune system can win against “resistant” bacteria, given the chance. We do not need a perpetual supply of new magic bullets, because the immune system will make them. It is a matter of controlling competing rate processes. (Not that magic bullets are without applications if they are available).
A condition such as MRSA therefore probably means that there remains too large a number of pathogenic bacteria for the immune system to cope with after antibiotic treatment. It is not that every individual bacterium in the population is of a new resistant species, just an unusually high proportion. In principle the shape of the distribution curve could be confirmed by laboratory tests. In the case of MRSA, methicillin is in effect being used as the discriminant.
“Friendly” bacteria may also play a part in limiting infection. The body may be regarded as a vessel which contains at least 500 species of bacteria, all competing for space i.e. inputs and freedom from increasing predation. This must be true in any system or it would continue to grow for ever. The result is that these bacteria co-exist in equilibrium, apparently without causing harm, and possibly performing essential functions for the body. Under normal circumstances, if something happens to disturb the equilibrium of these species, the system tends to return to a stable state again.
However, an intruder bacterium has to force its way into this system to establish itself, which could have two consequences. First, it is possible that a boost to the equilibrating system may limit the rate of growth of the intruder species by denying it “space”, for example by addition of “friendly” bacteria. Secondly, antibiotics may damage the populations of equilibrating bacteria to such an extent that space is inadvertently made for the intruder.
In competing rate processes, the initial conditions are likely to determine the entire outcome. There may be induction periods which influence all subsequent development. If bacterial reproduction and immune system production occur at equivalent rates, the effect is neutral, but if the immune system has an induction period while it finds the right antibody, it can never match bacterial reproduction. To compete it has to find a faster rate from somewhere. There is the possibility that the balance can be redressed in favour of the immune system by antibiotics or other agents which cause induction periods in the onset of growth of bacteria, or reduce the rate of division of bacteria , say division every hour rather than 40 minutes, which has a considerable effect on numbers of intruders at any particular time. This is the sort of competition of which the outcome is entirely calculable, if the rate constants are known.
The model would also explain the mechanism by which vaccines protect. They provoke the evolution of the correct species of antibody to add to the collection, so that when real infection occurs the immune system has a head start with no induction period. Antibodies passed on with mother’s milk would operate in a slightly different way; they would provide cover while the offspring’s own immune system found and made its own antibodies. Since antibodies in this model form such a small proportion of chemicals by weight of the human body, the concept of “overloading” the immune system would appear to be meaningless. Vaccines could also work in conjunction with antibiotics, neither being conclusive, but both contributing to the effectiveness of the immune system to finish the job.
first discovery –
single bacterium “evolves”
best adapted survive
cannot apply to single bacterium
must be sloppy cloning
humans are temperature controlled incubators
bacteria are like chemicals, not malicious
but disturb equilibrium of systems
laboratory simulation of “natural selection”
resistant populations formed
competition among bacterial populations
human immune system
different species of active entities
manufactured as required
suggests dynamic equilibrium
restored after disturbance
develops new variants
rates of production of intruders and antibodies
humans are source of pathogens
immune system usually wins
immune system does the job
defeats “resistant” bacteria, given the chance
populations of bacteria compete
intruders denied space
initial rates determine the outcome
retard bacterial multiplication
Copyright A. C. Sturt 2005