Department of Microbiology – The Ohio State University
phage.org – phage-therapy.org – biologyaspoetry.org
An important question for those with ecological and evolutionary biological interests is when a given situation is of ecological relevance vs. when it might be more of evolutionary biological importance (Abedon, 2022b). Telling the difference can be important for all of us.
Ecology by definition is the interaction of organisms with their environments.
We can describe phage use as antibacterial agents, that is, phage therapy, as an example of community ecology, or more precisely an applied community ecology. This is community ecology because there is more than one species of organism involved, i.e., as making up an ecological community. Minimally this is the phage (species #1) and the targeted bacterium (species #2), but also of importance is the treated body (species #3).
By definition, bacterial resistance to phages is ecological, as it describes a specific type of interaction, in this case between at least two species, the phage and the bacterium. That the resistance ‘interaction’ is one of ‘non-‘ or ‘less-‘ contact by the bacterium with the phage antagonist is only a detail, just so long as this lack of interaction is phenotypic, i.e., as opposed to the phage and bacterium instead just happening to exist in different places.
Bacterial resistance to phages also of course can have evolutionary aspects.
Evolution by definition is a change in allele frequencies in at least one species, or, more precisely, changes in allele frequencies in one population, in either case as observed over time.
Often the changes in allele frequency that we care most about are consequences of the impact of natural selection, and natural selection under most circumstances has a strong ecological component. Indeed, natural selection in most cases can be defined as the impact of ecology on evolutionary biology (and hence, as an aside, the existence of the science of evolutionary ecology).
By definition yet again, changes in the frequency of phage-resistance alleles within a bacterial population is an evolutionary process and typically these changes are a consequence of natural selection. The selective agent would be that phage population that is negatively affecting a bacterial population, resulting in increases in the frequency of whatever bacterial alleles are conferring protection from this phage.
Of interest to phage therapy is that this ecology driving evolutionary biology can in turn drive ecology. Specifically, once the frequency of alleles conferring phage resistance are high enough within a targeted bacterial population, then the applied community ecology of phage therapy can be affected, e.g., phage therapy can stop working.
In addition, if the frequency of a phage-resisting allele is found to be 100% within the bacterial population, following phage treatment (i.e., a frequency of 1.0), then if nothing else this is indicative that bacterial survival – an ecological issue – in this instance likely is a function of the occurrence of phage resistance.
If the frequency of phage resistance instead is 0% following phage treatment (0.0), then if nothing else this is indicative that bacterial survival (again, an ecological issue) was not a function of the occurrence of phage resistance. In fact, any frequency of phage resistance below 100% within a targeted bacterial population means that phage-sensitive bacteria are persisting despite phage treatment. For an example of the latter, see, e.g., Box 2 of Abedon (2022c).
Phage-resistant bacteria may display reduced virulence against bodies or may be subsequently treated with a different phage. Consequently, in some ways phage-resistant bacteria are not necessarily that big of a deal as a midpoint of a phage treatment, and this can be particularly if a diversity of other treatment phages are available. Phage resistance is not desired nor welcome, of course, but evolution of phage resistance also is not a certain indication of phage therapy microbiological failure.
That, by the way, to a degree contrasts with the evolution of antibiotic resistance that can occur over the course of antibiotic treatments, which can indeed be associated with treatment failures with high likelihood. One difference is something called antagonistic pleiotropy – not to be confused with antagonistic coevolution (Abedon, 2022a)!!! – i.e., whether or not resistance alleles are otherwise costly to the carrying organism (Abedon, 2022d). If resistance is both easily attained and not ecologically costly, then, well, that can be problematic, particularly given only mono therapies. Another difference is the sheer abundance of diverse, typically safe-to-use phages that often can be available to phage therapists (Abedon and Thomas-Abedon, 2010).
In any case, the persistence of phage-sensitive bacteria despite phage treatment probably means that, for whatever reason, treatment phages are not able to successful infect targeted bacteria despite those bacteria being phage sensitive; again, see Box 2 of Abedon (2022c). This frankly should be viewed as a big deal as essentially by definition it implies a phage therapy microbiological failure, one that may or may not be easily rectified, or at least a lack of complete eradication of phage-sensitive bacteria. I mean, how does one deal with phages not being able to easily reach and/or kill the otherwise phage-sensitive bacteria they are targeting?
Perhaps, as an answer to that question, some other phage that happens to be able to do a better job of reaching and/or killing those otherwise phage-sensitive bacteria might be available for use. That certainly is possible, but at this point in time we really aren’t all that good at figuring out what might constitute a better phage for phage therapy, other than in terms of host range – though see for example Bull et al. (2002; 2019) – and particularly better than the phage that we started with, presumably assuming that the first phage tried we thought was the better phage for phage therapy, hence why it was used first.
Of course, we almost take it as a matter of faith that phage carriage of extracellular polymeric substance (EPS) depolymerases (Danis-Wlodarczyk et al., 2021b) will solve many problems of phage penetration to targeted bacteria. But whether that is actually true in all instances, e.g., such as phage distribution throughout lungs – yet again, see Box 2 of Abedon (2022c) – is in my opinion just not known.
How might use of a phage cocktail instead result in complete eradication of a phage sensitive bacteria when a monophage does not? Just better odds that at least one of the phages used will be particularly good at achieving this? As another aside (Danis-Wlodarczyk et al., 2021a; Abedon, 2022c), note that it can be helpful to just apply a phage or phages at higher or multiple doses before giving up on a given treatment strategy!
At any rate, not being able to eradicate bacteria from an infection even though those bacteria are sensitive to a given treatment can be a far greater problem than failure that get rid of bacteria that explicitly are not susceptible to a treatment protocol. That is, there exits a basic problem in the applied ecology of treatments if not even phage-sensitive bacteria can be removed in full, just as there is a basic problem for an antibiotic treatment if the antibiotic is unable to fully eliminate even antibiotic-sensitive bacteria, a.k.a., the concept of antibiotic tolerance. For a bit on the latter, see Appendix A1 of in fact yet yet again, Abedon (2022c).
So where exactly am I going with this? At the end of a phage treatment, it is important to know whether the frequency of phage-resistant bacteria among the targeted bacterial population is high (at or approaching 100%) rather than low (near 0%). But high precision in that measurement, e.g., more than just whole percentage-point differences, really is not all that important. Why not?
Especially ecologically, there likely is little difference between 0.1%, 0.01%, or even maybe 10% or 50% of the bacteria being phage resistant, as that will still leave an awful lot of phage-sensitive bacteria having escaped phages during treatment. At some probably higher frequency of phage resistance we might come to feel that the frequency of remaining phage-sensitive bacteria is less important, but exactly where that point lies is difficult to say. My gut feeling, though, is that at the point where we start having to do statistics to tell the difference, i.e., at a point where higher precision in measurements becomes important, the importance of differences in the frequencies of phage-resistant bacteria – 100% or a tiny bit less than 100% – probably are no longer all that relevant. (For consideration of the statistics of plating-based enumeration, see Abedon and Katsaounis, 2021.)
Ah, you are saying, clearly therefore I am leading up to claiming that if we are interested instead in the evolutionary biology phage resistance, then in that case we really should care about measuring resistance frequencies with higher precision. And you would be absolutely right! Except also maybe not.
The problem here is that a key word in the definition of evolution that we are using is “Change”, and by definition change cannot be measured using only a single data point, or in the case of quantifying evolutionary change, a single time point. Thus, no matter how precisely you measure the endpoint frequency of phage-resistant bacteria, that will not tell you that evolution has occurred in the course of phage therapy treatment, much less how much evolution.
Here is the basis of this latter point: At the start of an experiment, if your population of bacteria ever is going to contain phage-resistant members, then it likely already does contain those mutants (this, by the way, is why only-qualitative determinations that phage resistance is present, e.g., such as following phage treatments, are basically meaningless). Exceptional would be if the starting number of bacteria involved is so low that this number is, e.g., less than the inverse of the rate of mutation to phage resistance. Thus, for every time a bacterium divides, let’s say that there is a probability of 10-5 that a mutation to phage-resistance will occur. If so, then in a population of 106 bacteria, on average 10 phage-resistant bacteria will be expected to be present, more or less (Abedon et al., 2021).
That last part, “More or less”, is crucial, however, as the frequency with which mutations conferring phage resistance are expected to be present is predicted to somewhat “Fluctuate” about an average (Luria and Delbrück, 1943). In practice, this means that even if you precisely know bacterial rates of mutation to resistance to a given phage, you still will not know how many phage-resistant bacterial mutants will be present prior to the start of treatments. (As yet another aside, actually calculating mutation rates, vs. just mutant frequencies, is a not trivial thing to do.)
Without knowing the frequency of phage resistance prior to the start of treatments, then you are only really guessing whether evolution has occurred in the course of a phage treatment, no matter how precisely frequencies of phage resistance may be measured after a treatment is done.
In short, ecologically, the precision of measures of frequencies of bacterial phage resistance need not be all that high to possess high value in understanding the outcome of phage treatments. I mean, either phage-sensitive bacteria have persisted despite prior treatments or they have not, without a need to describe percentages with precision past the decimal point. Thus, 50.0% vs. 50.1%? Who cares? Indeed, 50% vs. 51%, who cares?
Alternatively, if one really cares about being precise in monitoring the evolution of phage resistance, then the most important place to emphasize that precision actually should be prior to the start of treatments, i.e., prior to initial phage application, and only then should one be measuring frequencies of phage resistance after treatments as well. But don’t forget that you need to have this information for explicitly that bacterial culture that is being treated, since evolutionarily all we really will care about is how a specific bacterial culture as a population changes in allele frequencies over time, and in phage therapy that bacterial population is precisely the one that you are treating.
Even so, how much more than order-of-magnitude precision do we really need in monitoring the evolution of phage resistance during phage treatments? Will we really care for example if the frequency of phage-resistant bacteria have changed from 10-5 to 10-5.5? And how hard would we have to try to be sure that such a relatively small change is actually real? I mean, seriously, except for the most hard-core evolutionary experiments, who would really care?
For what it is worth, when I look at the outcome of a phage treatment, if all of the targeted bacteria remaining are phage resistant, then I know what went wrong (clue: the bacteria have evolved resistance to the treatment phages, i.e., an evolutionary outcome). But when I look at the outcome of a phage therapy experiment and a substantial portion of the bacteria remaining are still phage sensitive, then more often than not I can only speculate as to what might have gone wrong, except again for those bacteria that have evolved phage resistance (Abedon, 2022c). Still, this latter scenario should be viewed at least as an ecologically relevant outcome.
But bottom line: Obtaining an additional decimal place or two in describing the frequency of phage-resisting alleles within the treated bacterial population generally will not greatly aid in improving the precision of our applied ecological speculation.
You would think that this essay came into existence as a natural outgrowth of the cited publications, particularly Abedon (2022c), but you would be wrong! On the other hand, I did wait a few months until that publication was published and available open access. Thanks for your interest!
Abedon, S. T. 2022a. A primer on phage-bacterium antagonistic coevolution, p. 293-315. In Bacteriophages as Drivers of Evolution: An Evolutionary Ecological Perspective. Springer, Cham, Switzerland. https://link.springer.com/chapter/10.1007/978-3-030-94309-7_25
Abedon, S. T. 2022b. Frequency-dependent selection in light of phage exposure, p. 275-292. In Bacteriophages as Drivers of Evolution: An Evolutionary Ecological Perspective. Springer, Cham, Switzerland. https://link.springer.com/chapter/10.1007/978-3-030-94309-7_24
Abedon, S. T. 2022c. Further considerations on how to improve phage therapy experimentation, practice, and reporting: pharmacodynamics perspectives. Phage 3:95-97. https://www.liebertpub.com/doi/full/10.1089/phage.2022.0019
Abedon, S. T. 2022d. Pleiotropic costs of phage resistance, p. 253-262. In Bacteriophages as Drivers of Evolution: An Evolutionary Ecological Perspective. Springer, Cham, Switzerland. https://link.springer.com/chapter/10.1007/978-3-030-94309-7_22
Abedon, S. T., and C. Thomas-Abedon. 2010. Phage therapy pharmacology. Curr. Pharm. Biotechnol. 11:28-47. https://pubmed.ncbi.nlm.nih.gov/20214606/
Abedon, S. T., and T. I. Katsaounis. 2021. Detection of bacteriophages: statistical aspects of plaque assay, p. 539-560. In D. Harper, S. T. Abedon, B. H. Burrowes, and M. McConville (ed.), Bacteriophages: Biology, Technology, Therapy. Springer Nature Switzerland AG, New York City. https://link.springer.com/referenceworkentry/10.1007/978-3-319-40598-8_17-1
Abedon, S. T., K. M. Danis-Wlodarczyk, and D. J. Wozniak. 2021. Phage cocktail development for bacteriophage therapy: toward improving spectrum of activity breadth and depth. Pharmaceuticals 14:1019. https://pubmed.ncbi.nlm.nih.gov/34681243/
Bull, J. J., B. R. Levin, T. DeRouin, N. Walker, and C. A. Bloch. 2002. Dynamics of success and failure in phage and antibiotic therapy in experimental infections. BMC Microbiol. 2:35. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC138797/
Bull, J. J., B. R. Levin, and I. J. Molineux. 2019. Promises and pitfalls of in vivo evolution to improve phage therapy. Viruses 11:1083. https://pubmed.ncbi.nlm.nih.gov/31766537/
Danis-Wlodarczyk, K., K. Dabrowska, and S. T. Abedon. 2021a. Phage therapy: the pharmacology of antibacterial viruses. Curr. Issues Mol. Biol. 40:81-164. https://pubmed.ncbi.nlm.nih.gov/32503951/
Danis-Wlodarczyk, K. M., D. J. Wozniak, and S. T. Abedon. 2021b. Treating bacterial infections with bacteriophage-based enzybiotics: in vitro, in vivo and clinical application. Antibiotics 10:1497. https://pubmed.ncbi.nlm.nih.gov/34943709/
Luria, S. E., and M. Delbrück. 1943. Mutations of bacteria from virus sensitivity to virus resistance. Genetics 28:491-511. https://pubmed.ncbi.nlm.nih.gov/17247100/