Andrew J. Curtright and Stephen T. Abedon*
Department of Microbiology, The Ohio State University, Mansfield, Ohio USA
Received Date: September 01, 2011; Accepted Date: October 20, 2011; Published Date: November 08, 2011
Citation: Curtright AJ, Abedon ST (2011) Phage Therapy: Emergent Property Pharmacology. J Bioanal Biomed S6:002. doi:10.4172/1948-593X.S6-002
Copyright: © 2011 Curtright AJ, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Visit for more related articles at Journal of Bioanalysis & Biomedicine
Antibiotics; Antibacterial agents; Emergent properties; Phage therapy; Pharmacology
“In complexity there is simplicity.” (35 Google hits)
Biological control, or biocontrol, is the application of specific types of organisms, to environments, to combat other types or organisms. One of its earliest forms involved the introduction of pathogenic organisms into naïve human populations, that is, biowarfare [1,2]. A less belligerent and perhaps even older approach is the application of diseases or predators into populations of nonhuman targets, such as the keeping of cats for the sake of rodent control . Alternatively, biocontrol includes the application of the viruses of bacteria – known as bacteriophages or phages – into populations of unwanted bacteria, so-called phage therapy . In this article we consider explicitly the use of phages to combat bacterial diseases associated with humans or animals, phage therapy sensu stricto, reserving the term biocontrol for all other phage uses as antibacterial agents .
A large number of reviews have addressed phage therapy generally as well as historically; see Abedon and Thomas-Abedon  along with Abedon et al.  for references. Here we focus less on the actual practice of phage therapy, or its history, and more on how the use of phages to combat bacterial infections can be considered in pharmacological terms. Since phage therapy, as well as biocontrol in general, is a form of applied ecology, we additionally integrate these ideas into a more general biological framework, focusing on the concept of emergent properties. This article also can be viewed as an extension of a number of recent efforts to consider the biology of phage therapy from a pharmacological perspective [5,6,8,9].
Overall, the use of phages as antibacterial agents can be contrasted with that of antibiotics. An urgency to these considerations exists because, for various reasons, antibiotics as antibacterial agents appear to possess less utility [10,11] or benignness [12,13] than once had been expected. For example, “at least 25,000 patients in Europe die per year because their bacterial infections are not treatable with available antibiotics” (p. 68) . As possible solutions to this world-wide ‘antibiotic crisis’, phages – as semi-autonomous genetic entities which, to a first approximation, would appear to be too complex to develop as antibacterial “drugs” – might not seem promising. Conversely, in practice phages appear to be both efficacious and relatively easily developed, including within the context of human treatment [7,14]. In this article it is this paradox in particular that we seek to address. That is, how does what would appear to be an extreme in pharmacological complexity instead beget a relative pharmacological simplicity? To address this question we begin by defining the concept of emergent properties, indicating how this notion is relevant to pharmacology in general and phage therapy pharmacology in particular.
In drug development, emergent properties can be thought of as contributing to the high failure rate of new chemical entities in clinical trials. The behaviour(s) of most candidate drugs in humans is not predicted by their activities in the assays used in typical drug discovery efforts, explaining, in large part, the high failure rate of new drugs in clinical testing. Berg et al. , p. 204.
The goal of pharmacological research is the development of safe and effective pharmaceuticals that are useful in the treatment or prevention of disease. Though to a degree these efforts can be theory driven, in practice pharmacology requires substantial empirical testing, including in animals along with human patients. Animal testing as well as subsequent clinical trials are important because pharmacological data from chemical assays, in vitro models, or even complex systems biology approaches – given less than complete knowledge – cannot effectively capture all emergent properties associated with the systems drugs interact with [16-18]. As a consequence, whether or not a candidate drug becomes clinically useful is dependent upon factors that are not readily predictable, but instead are discovered, often at great expense, in the course of both animal and human testing. A key bottleneck in pharmacological development thus is the potential for drugs in combination with biological systems to display emergent properties.
Emergent or “nonreducible” properties  can be defined as higher-level aspects of systems that, qualitatively, are not directly exhibited by lower-level components of the same systems. These properties emerge because the resulting higher-level features are not apparent within lower-level aspects until those aspects interact as a more cohesive whole. Salt , quoting Harré , and considering emergent properties as an ecological concept, provides this definition: “…the property of the whole is produced by properties of the parts but is not qualitatively similar”
Ecological properties in particular emerge from an organism’s interaction with its environment, and among environmental components are those aspects that enter into an organism’s body from the outside world. The field of ecophysiology in fact considers explicitly how extraorganismal factors impact the organism internally, that is, physiologically. Body-drug as well as body-toxin interactions equivalently can also be viewed as inherently ecological and thereby at least potentially possessing emergent properties. Drugs, as extraorganismally sourced “environmental” components, in other words, often have substantially different properties viewed chemically within the laboratory, or even within the context of complex in vitro or in silico models, versus physiologically within experimental animals or human patients.
Appreciation of emergent properties typically requires an empirical component inasmuch as emergent properties, essentially by definition, are difficult to predict; this especially is experiential observation biased towards more-complete assemblages, such as approximations of whole ecosystems, versus analyses of system components as they may be studied in relative isolation. Thus, in Salt’s own words (p. 145): “‘An emergent property of an ecological unit is one which is wholly unpredictable from observation of the components of that unit.’ The corollary is: ‘An emergent property of an ecological unit is only discernable by observation of that unit itself.’” Furthermore, the idea of emergence can be contrasted with that of collective . Pharmacologically, additive drug interactions might be deemed collective (‘sum of the parts’) whereas, for example, synergistic or antagonistic interactions between drugs – which typically represent far more complex phenomena, i.e., more, or less, than the sum of the parts – we could describe instead as emergent.
Drug-body interactions can be difficult to predict in part because of the complexity of animal bodies but also because of the variance in characteristics that exist between individuals. Thus, new understandings of body systems can ‘emerge’ upon animal or clinical testing. Antimicrobial drugs take this issue to an additional level. Here not only is complexity and variability inherent in the larger organism but so too the properties of target parasites, or pathogens, are complex. Thus, in addition to it being difficult to fully anticipate an antimicrobial drug’s impact on the body, such as its toxicology, or the body’s impact on the drug (e.g., it’s rate of clearance), but so too one must study living microorganisms, rather than just their components (e.g., genes), to ascertain a drug’s antimicrobial properties even in greatly simplified environments such as test tubes or Petri dishes. Yet additional complexity can emerge with antibiotic-bacterial interactions such as the now well-appreciated ability of bacterial biofilms to resist antibiotics that constituent bacteria, growing as individual cells, are quite sensitive to , or the more general ability of bacteria to evolve biofilm-independent mechanisms of antibacterial resistance [11,23]. Therefore, while pharmacology to a substantial extent can involve the elucidation of drug-body emergent properties, this idea can be even more applicable to consideration of the pharmacology of antimicrobial drugs.
An additional level of emergence can occur when the “drug” itself is also somewhat complex. One sees this with the use of live-attenuated vaccines (e.g., MMR) as well as whole-killed vaccines, particularly against the highly variable influenza virus. These reagents are not just interacting with living organisms – directly with ourselves and indirectly with would-be infecting pathogens – but themselves are organisms or at least their intact derivatives. Our ability to predict the biological properties of these entities such as from their chemical components, as a consequent, can be fairly limited. Vaccine development, like pharmacology in general, therefore possesses a substantial empirical component: At best we apply educated guesses to drug design, and then always brace ourselves for the emergence of unexpected properties upon animal and clinical testing. Another category of live as well as naturally attenuated “drugs”  are a group of viruses, phages, which can be used as antibacterial agents in the guise of what is known as phage therapy.
Phages are viruses that infect members of domain Bacteria, which contains all known pathogenic bacteria . The majority of phages are both lytic and tailed, the former a description of how phage virions are released from infected bacteria and the latter a description of virion morphology. A tail is a capsid appendage that is responsible for virion adsorption along with certain aspects of DNA translocation into bacterial cells. Furthermore, all tailed phages have double-stranded DNA genomes. In lytic infections, mature phages are released from infected bacteria via degradation of the bacterial cell envelope. In the course of this lysis the infected bacterium is physiologically and to a large extent also structurally destroyed. See Figure 1 for summary of selected characteristics of various phage types.
The problem of temperance: Many lytic phages are also temperate. A temperate phage is able to display lysogenic as well as productive cycles. In a productive cycle, phage progeny are produced and subsequently released, such as through the above-described lytic process. In a lysogenic cycle, phages persist mostly as genetic material, termed prophages, replicating intracellularly in tandem with their bacterial hosts.
Prophages, through a process of induction, can also display productive cycles. Indeed, a majority of the infections initiated by temperate phages give rise to lytic rather than lysogenic cycles. Temperate phages nonetheless are problematic for phage therapy for a number of reasons, and consequently should be avoided for these purposes. These reasons include the display what is known as superinfection immunity, which in practical terms means that bacteria that are lysogenically infected – and consequently described as lysogens – can no longer be killed by the same phage types. It is as though the majority of an applied antibacterial were to succeed in killing the encountered bacteria but with some fraction instead providing an ‘antidote’ or ‘vaccine’ to the bactericidal activity.
In addition, many temperate phages carry genes that encode bacterial virulence factors including such infamous exotoxins as diphtheria toxin, cholera toxin, and Shiga toxin . One can mostly avoid virulence factor genes by limiting phage choice for phage therapy to what can be described as “professionally lytic” phages . These, defining the term narrowly, are phages that not only are “obligately lytic” (that is, unable to display temperate cycles) but which also have not recently descended from temperate phages. In phage therapy, the phages of choice tend to be professionally lytic as well as tailed. See Figure 1 for summary.
Phage environmental microbiology: Any description of phages would be remiss if it did not mention their ubiquity. This issue is relevant to phage therapy pharmacology in the sense that most substances that we live our lives awash in tend to be relatively benign, with phages seemingly not an exception. It has been estimated based on combinations of known phage prevalence, along with considerations of likely ratios of phages to bacterial hosts within environments, that perhaps 1031 or 1032 phage virions exist on Earth at any given moment, e.g., as discussed in Abedon . This means that any time that one goes swimming in either the ocean or a less than oligotrophic lake, then a million or so phages are also bathing us for every ml surrounding our bodies . A liter of water can easily contain a billion individual virions! Further, food, including especially bacteria-fermented products, e.g., [29,30], tend to be full of phages, as too can be animal bodies, in particular colons . We are born into a world of phages and develop through our adulthoods experiencing a constant phage barrage. Indeed, animals have been so exposed presumably for as long as animals have been in existence, with phages an unseen and at times underappreciated but nevertheless omnipresent aspect of our world.
Phage therapy is the application of phage virions to bacteriacontaining environments with the goal of bacterial control. In most instances with phage therapy “control” translates as “kill” and the bactericidal nature of obligately lytic phages can be well suited to this task. Phage therapy, strictly, is a form of biocontrol, with one organism (here, phages) employed to control another, typically pathogenic or nuisance organism (here, bacteria). We concentrate in this article on issues relevant to the phage therapy of animals, particularly as may be applied to humans [7,14]. Others have considered, by contrast, the phage-mediated biocontrol, for example, of foods [32,33], of food animals such as to reduce loads of zoonotic pathogens , or of plants .
Active versus passive treatments: Phage treatments, whether phage therapy or biocontrol in general, can be distinguished into two basic types, those that are dependent on phage in situ replication to achieve bacterial control and those that do not require such replication. The former has been dubbed an “active” treatment after the idea that it requires “active” phage replication to be effective; see Abedon and Thomas-Abedon  for discussion. By contrast, phage therapy that does not require active phage replication has been dubbed passive or, alternatively, inundative phage therapy.
Though seemingly a somewhat intimidating approach to antibacterial therapy, in fact inundative treatment is the phage equivalent of conventional antibiotic therapy, i.e., where enough drug is supplied via traditional approaches to dosing that bacterial infections are brought under control. In any case, unless phages are either engineered to kill bacteria without producing infectious phage particles  or otherwise are unable to reproduce upon infecting bacteria due to limitations in their ‘spectrum of activity’, then phages should still display some degree of in situ “self amplification” or “auto dosing”  in the course of their antibacterial action.
Phages as emergent property pharmaceuticals: Note that this property of phage replication, in the course of their antibacterial action, can be considered to be emergent since it is a phage aspect that would not be easy to predict were we to analyze phage components, such as individual capsid proteins, solely in isolation. Indeed, numerous genes are found in the genomes particularly of larger phages for which no function has been assigned, e.g., Hendrix , but which in many cases presumably can contribute to phage population growth in some manner. Nevertheless, the phage property of replication and subsequent population growth is an expected consequence of phagebacterial interactions given our understanding of phage biology. Pharmacologically, therefore, phage population growth is not an emergent property in the sense of requiring animal testing to discover any more than discovery of a chemical antibiotic’s basic antibacterial ability would require such testing.
By contrast, an unexpected absence of phage replication or antibacterial action, in situ, could be described as pharmacologically emergent, that is, a pharmacodynamic property that emerges only upon characterization of the larger system. Here this larger system would be phages plus target bacteria plus the body environment rather than phages plus target bacteria within a simpler laboratory environment. See, for example, the results of Bull et al.  as well as Abedon , along with references therein cited, for a broader discussion of what can ‘go wrong’ in terms of bacterial killing during phage therapy animal testing. To clarify: Phage replication when in the presence of target bacteria certainly can be viewed as a phage ‘behavior’ that emerges from the lower-level aspects of phage biochemistry, but that replication is less emergent in pharmacological terms because it is something that is expected to occur prior to the performance of animal or clinical testing.
Phages within the context of phage therapy thus may be considered to be emergent property pharmaceuticals, given the impressive complexity from which their population growth and even antibacterial activity in fact emerge. However, and a prominent theme within this article, the chemical complexity of phages as “drugs” – that is, the multitude of genes and gene products that together make up phages – in certain ways conspire to simplify rather than complicate their pharmacological properties. Such pharmacological simplification, as we will consider, may be most prominently observed from the perspective of phage safety.
Phages are used therapeutically mostly “off label”: In principle phage therapy can be used against any bacterial disease, though in practice the diseases that are targeted tend to be ones against which alternative treatments are less available. This, particularly in Western medicine, can be against antibiotic-resistant bacteria. In the former Soviet Union, by contrast, phage therapy has been used against bacterial infections in general, owing to a tradition of phage use that developed in that country prior to widespread antibiotic availability or, for that matter, before the development of a robust understanding of phages as biological entities. In Poland phages have been employed on an “experimental” basis to treat chronically infected wounds that have not responded to conventional antibiotic treatment. What is fascinating about these efforts is that they seemingly have been quite successful, and, especially in the former Soviet republic of Georgia as well as Wroclaw, Poland, have been ongoing. Reviews by Kutter and colleagues [7,14,40] have explored these various international aspects of phage therapy use.
Though phage therapy has been employed extensively outside of North America, the technique has not been subject to a great deal of double-blind clinical trials. Phages consequently, as a class of “drugs”, have been employed almost entirely “off label”. Widespread double-blind clinical testing of phage therapy likely will only happen given substantial increases in the availability of resources to the field. In addition, phage therapy, as a field, appears to be hampered by a combination of (i) unfamiliarity with phage use by Western physicians – which is not to say that there is a lack of Western physicians who are enthusiastic of the idea, but instead that obtaining and then using phage therapeutics is difficult in environments where their use effectively is not even off label – (ii) questions of intellectual property rights and particularly the patenting of technologies that have been in existence for nearly 100 years; and (iii) the fact that in those countries most familiar with phage therapy the therapeutic use of phages already represents a standard of care, a situation that for ethical reasons can deter double-blind study.
Notwithstanding these considerations, it is our opinion that it would be helpful to the field to adopt a more pharmacologically rigorous perspective on phage therapy research and development. While various issues of phage therapy pharmacology have been considered elsewhere, as cited in the introduction, here we explicitly discuss their relevance or helpfulness to the idea of employing phages as antibacterial biocontrol agents and do so within a context of constraints that are imposed upon pharmacological research and development by drug emergent properties. See also Table 1 for discussion of advantages of using phages as antibacterial drugs including in comparison to chemical antibiotics, with many of these advantages discussed further in Loc-Carrillo and Abedon .
|Bactericidal||Reduced potential for bacterial development of resistance|
|Auto “dosing”||In situ activity increases numbers (though only given favorable bacterial densities)|
|Low inherent toxicity||Virions consist of only proteins and DNA|
|Low normal flora impact||Low likelihood potential for superinfection by endogenous flora, e.g., C. difficile|
|Narrow resistance evolution||Selection for resistance limited mostly to within populations of targeted bacteria|
|Lack of cross-resistance||Antibiotic-resistant bacteria tend to retain phage sensitivity|
|Rapid discovery process||Phages with large therapeutic windows are often easy to isolate|
|Potential for modification||Phages can be easy to molecular characterize and manipulate|
|Use with other agents||Versatility in formulation development and combination with other drugs|
|Dosing versatility||Phage formuations can take many forms and can be delivered via many routes|
|Biofilm clearance||Certain phages, unlike most chemical antibiotics, can be relatively good at this|
|Favorable pharmacokinetics||Delivery to targets or persistence in situ often is either good or improvable|
|Single-dose potential||Can provide dosing convenience; auto dosing consequence|
|Low-dosage potential||Of possible economic or safety utility; auto dosing consequence|
|Single-hit killing kinetics||Nonetheless effectively multi-hit since phages still multiply adsorb bacteria|
|Engineered lower toxicity||Particularly elimination of bacterial lysis, but auto dosing advantage is thus lost|
|Low environment impact||Narrow specture of anti-bacterial activity, lability, low inherent toxicity|
|Not antibiotics||Fewer societal concerns with use, such as in agriculture; avoids antibiotic allergies|
|Natural products||Potential appeal to natural medicinals market|
|Relatively low cost||As drugs, reasonable production costs|
|Public perception||Public perception of use of phages as antibacterials seemingly is positive|
1Advantages are as seen relative particularly to the use of chemical antibiotics as antibacterial agents.
2Line break between upper and lower portions of list is indicative of what we feel are greater (upper) versus lesser (lower) advantages to phages as antibacterial drugs.
3Ratio of drug dosage that gives rise to toxic effects to that which gives rise to efficacy. The larger the ratio, the safer the drug and/or the easier it is to work with.
Table 1: Advantages Associated with Phage Use as Antibacterials1.
Toxicity and side effects
Toxicity or side effects can be described as negative pharmacodynamic consequences, contrasting the directly sought after positive pharmacodynamic effects such as bacterial control. Pharmacodynamically, therefore, the key phage characteristic that contributes to the potential for phage therapy to serve as an alternative to antibiotic treatment, other than the phage ability to kill bacteria, is in fact the generally low toxicity of especially professionally lytic phages. The advantage of low-toxicity pharmaceuticals is not just their safety but also a freedom from patient monitoring to prevent toxicity as well as an ability to lengthen intervals between dosing, which together can greatly reduce costs while increasing patient convenience. In addition, low toxicity can allow in situ levels to exceed a drug’s minimum effective densities, resulting in antibacterials with increased potential to control infections. As a result, it means that the phage ability to replicate in situ in the course of killing bacterial targets can be a dosing ally, allowing phage densities to increase precisely at their site of action and do so without simultaneously serving as a toxic hindrance.
Pharmacokinetics, contrasting pharmacodynamics, describes the impact of the body on drugs. Low toxicity can result in a reduction in certain pharmacokinetic concerns since higher doses, or phage self amplification, can be used to counter mechanisms of drug loss or dilution in the course of what are known, pharmacokinetically, as drug absorption and distribution. More generally, toxicity and side effects represent the most troubling consequence of pharmaceutically emergent properties and a relative absence of such aspects would presumably simplify pharmacological development. It is somewhat surprising, therefore, how phages, with a multitude of their own emergent properties, appear to lack a majority of the side effects observed with seemingly simpler, small-molecule antibacterial drugs. We thus begin our discussion of pharmacology as it applies to phages by focusing on phage safety since much of how phages are or could be used as pharmaceuticals derives from this property.
Toxicity as well as phage-immune system interactions: Drugs can be mutagenic or teratogenic, they can activate or inactivate receptors, they can be cytotoxic or mimic hormones, etc. Given appropriate phage choice, which generally means employing only professionally lytic phages for phage therapy (Figure 1), then what are presented to bodies by phages are mostly proteins and DNA that display little toxicity to body tissues. An alternative perspective is that in order for phages to display an evolved toxicity to human tissue they would have to be able to interact with human cells similarly to how they interact with target bacterial cells. Inasmuch as the processes inside a human cell are vastly different from the ones occurring inside a bacterial cell, such tissue toxicity generally is not even a potential concern nor has it demonstrated itself to be upon empirical observation; for review of phage-human- or animal-body interactions, see Merril .
Figure 1: Types of phages. Phage therapy, as indicated, is preferably achieved employing professionaly lytic phages, that is, phages that are both obligately lytic and otherwise not substantially genetically similar to temperate phages. The arrow between “Temperate” and “Virulent mutant” is meant to indicate descent while “Virulent mutant” as well as “Professionally lytic” are both types of “Obligately lytic” phages.
Though not necessary acting directly as toxins, protein-based drugs still can interact with host immune systems, in some cases to disastrous effect . More common is simply the production of anti-drug antibodies (ADAs). ADAs are usually benign, but when protein-based drugs resemble endogenous molecules then severe complications can arise, stemming from ADA cross reactivity and subsequent autoimmune response . As phages typically are not expected to be similar to endogenous molecules, and in fact are typical components of animal environments, such complications would be expected to be less likely in phage therapy application. Low likelihoods of complications stemming from phage-immune system interactions could be due, perhaps, either to bodies acquiring a tolerance to ubiquitous phages found in their environments or, instead, phages that are normally found as animal commensals evolving to minimize their potential to destructively impact their overall environment, i.e., our bodies. The latter possibility, we stress, is dependent on phage choice, that is, the use in phage therapy of professionally lytic phages and/or phages that have been shown to not carry bacterial virulence factor genes.
In practice there has been little documentation that phages used therapeutically will produce substantial side effects, or even of reductions in phage therapy effectiveness due to phage interactions with adaptive immune system . At least in part this general lack of evidence of negative phage-body interactions may stem from typical phage application practices, which is topical for chronic infections or systemic use over only relatively short spans against more acute infections. That is, the impressive extent to which phages appear to safely serve as antibacterial agents upon application to bodies may not be a wholly unexpected consequence of phage-patient interaction .
Bacterial dysbiosis: Besides direct interaction with animal tissues, including immune systems, a second means by which an antibacterial can negatively impact treated individuals is via the destruction of nontarget bacteria, particularly beneficial members of normal flora. The loss of these non-target bacteria can result in antibiotic-associated superinfections such as vaginal yeast infections or Clostridium difficileassociated colitis. Generally this kind of side effect is more likely to occur when broad-spectrum rather than narrow-spectrum antibacterials are employed. Most phages, however, display narrow spectra of activity , a property that may be of utility even among chemical antibacterials . Indeed, as Blaser  notes (p. 394), “We… need new, narrow-spectrum antibacterial agents to minimize collateral effects on the microbiota. …and, importantly, better diagnostics that rapidly identify the problematic agent.”
In considering the narrowness of phage spectra of activity, one may conceptualize the different steps of phage-bacterial interactions as “sieves” that generally result in prevention of non-targeted cell death. The first sieve eliminates all cells which do not carry the cognate receptor for the phage. The second eliminates all cells that are capable of resisting phage DNA uptake. Another blocks elimination of those cells that possess restriction enzymes against phage DNA or other mechanisms of active resistance by bacteria to phage infections [44,46,47], and so forth. Even when multiple phages are combined to form cocktails, the result typically is still an overall spectrum of activity that is no wider than that of the narrowest spectrum antibiotics currently employed clinically. Phages thus may be applied even prophylactically (below) with little concern of generation of bacterial dysbioses.
Drug interactions: Many drugs are metabolized in the liver by a class of enzymes known as cytochromes. When pharmaceuticals, including antibiotics, interact with these enzymes, inhibition or activation may occur. The result is that potentially harmful drug interactions may occur including increased toxicities such as due to build up of drug densities within bodies following less effective drug removal or, alternatively, loss of drug potency due to too rapid elimination . The characteristics of such drug-drug interactions are difficult to predict due to the emergent properties of the hepatic system as a whole, including drug transport and gene regulation, which are not obvious during in vitro toxicity or metabolism screenings. A concern with taking quinolones, for instance, is that their interaction with cytochrome P450 isoform CYP1A2 decreases its ability to metabolize other drugs, including antidepressants, and antipsychotics . If patients are not carefully monitored, then harmful levels of the interacting drug can accumulate.
For phages, to at least a first approximation, such potentially harmful drug-drug interactions should not be a concern since these same liver enzymes will not play as large a role in phage elimination any more than they do in the elimination of body proteins and DNA in general. Another way of stating this is that body metabolism of phage particles should have little impact on normal mechanisms of body homeostasis. Phages thus may be more readily used in conjunction with other drugs and not just other antibacterials but other drugs that a patient may already be consuming. That is, we conjecture that phagedrug interactions do not substantially give rise to emergent properties in terms of the functioning of these other drugs, though with the caveat that this issue to our knowledge has not been systematically studied.
Release of bacterial toxins: Most phages, via a complex series of steps, bring anti-cell wall hydrolases into the immediate vicinity of bacterial peptidoglycan layers. The resulting bacterial lysis results in the solubilization of bacteria-associated molecules, such as endotoxin or certain exotoxins, that can be directly or indirectly toxic to body tissues. One consequence of this is that, particularly with phage systemic application, there can be a need for substantial purification of phage virions , i.e., as protein-based drugs in general require purification prior to use, e.g., Magalhäes et al. .
This issue is also a concern with regard to the lysis of bacteria as that can occur in situ, during phage therapy, which in fact is the root of occasional efforts to engineer phages that are bactericidal without being lytic . Phages, though, are not unique in being lytic antibacterials since that characteristic tends to be shared with smallmolecule antibiotics such as penicillins that target cell walls. As a narrow-spectrum lytic antibacterial, however, the number of bacteria lysed by phages is fewer. Indeed, much of the complexity of phages as bacterial lysing agents can be seen as contributing to the selectivity of their lytic behavior.
This latter advantage should be relevant only to the extent that potentially lysed non-target bacteria are found in locations in which their lysis could be potentially harmful to the body, particularly within the blood such as during the treatment of septicemias. It is helpful, though, to recall that it is the bacteria being treated that ultimately are the problem, more so than the impact of antibacterials employed to eliminate them. That is, the elimination of systemically infected bacteria is not necessarily going to occur without harming the integrity of those bacteria and thereby releasing at least some bacterial toxins, though as noted phages can be engineered to reduce this damage .
Safety in numbers: Phages only replicate when they are efficacious and the result, from a pharmacodynamic standpoint, is a wide breadth of possible dosing ranges during phage treatment. This range extends from low doses in anticipation of in situ phage self amplification – which has a low likelihood of releasing toxic bacteria-associated molecules unless bacterial targeting is successful – to high doses if rapid bacterial lysis is not a concern and/or if phage self amplification is less likely. Notwithstanding this potential for a wide breadth of dosing ranges, a standard approach towards phage therapy dosing, for both safety and economic reasons, should be the use of relatively low phage densities, e.g., such as no more than ~108 phages/ml, with either repeated dosing or application of greater phage densities should phage access to bacteria prove limiting. This suggestion is rather than striving under all circumstances to inundate bacteria using substantial excesses of phages, such as >>108 phages/ml.
The relatively low toxicity of phages in combination with their potential for self amplification provides substantial leeway in terms of levels of dosing, with neither “too low” phage densities (e.g., <107/ml) nor “too high” phage densities (e.g., >109/ml) necessarily resulting in poor phage pharmacodynamic effects. Less latitude exists in terms of what peak phage densities in the vicinity of target bacteria are required to achieve effective bacterial killing. Because phages display low toxicities, however, generally the problem is more one of not achieving sufficient phage densities rather than providing excessive doses. Note that pharmacodynamically sufficient peak phage densities, that is, minimally effective phage densities delivered to or generated within the immediate vicinity of target bacteria, may be assumed, as a rule of thumb, to be approximately 108 or more active phage particles per ml [6,9,52].
Types of treatment
Treatments with drugs can be differentiated into a number of different approaches including prophylaxis, empirical treatment, and pathogen-driven therapy. In this section we introduce these various concepts in terms of phage therapy.
Prophylaxis: The utility of prophylaxis as a treatment approach is that it can be easier to prevent an infection rather than treat an ongoing disease. Furthermore, in situ-generated bacterial lysis products should be much less abundant given the associated (or expected) lower bacterial densities. Thus, phage safety should be further enhanced in primary prophylaxis, increasing the reasonableness of such an approach and indeed potentially allowing for phage use under circumstances where application of chemical antibiotics might be avoided due to the possibility of bacteria-unrelated side effects. In addition, phages tend to not interfere with wound healing, e.g., Soothill .
To the extent that prophylaxis is against bacteria existing at low densities, then active treatment, that is, in situ phage self amplification, may not be possible as threshold bacterial densities required for successful phage auto dosing will not be present [6,9,52]. Prophylactic treatment also will often be empirical (below) since the pathogen to which an individual might be exposed in many instances will not be known. The use of phages prophylactically nonetheless has occurred, with success, both experimentally in animals and clinically particularly in terms of war wounds; see, for example, Abedon et al.  for limited review.
Empirical therapy: Empirical therapy, as we employ the term here, is the initiation of treatment prior to full diagnosis, particularly in terms of the diagnosis of antibiotic susceptibility of an infectious agent (synonym: presumptive treatment). Although phages lack the same dysbiosis or superinfection concerns of antibiotics, their spectrum of activity generally is much narrower, making successful empirical therapy by specific phages less probable. Phage cocktails – formulations containing combination of potentially efficacious phage isolates – may help to overcome this concern, however, to the extent that a cocktail is sufficiently well formulated to cover most bacteria that would be encountered per infection type in a given geographic area.
In Georgia and the rest of the former Soviet Union phage cocktails have been continuously modified in response to changing spectra of locally prevalent bacterial pathogens . Though the practitioners of this approach of ongoing formulation modification certainly would appear to have found it to be beneficial, an obvious concern is that such a strategy could be inconsistent with standard approaches to drug development or regulatory approval (such as by the U.S. Food and Drug Administration). Alternatively, continuous surveillance along with periodic formulation revision is the hallmark of influenza vaccine implementation .
Pathogen-directed therapy: In pathogen-directed therapy the infectious etiology is not just identified prior to the initiation of treatment but also is characterized in terms of its antibacterial susceptibility. Because of phage specificity, pathogen-directed therapy is the type of treatment that best suits phage therapy. Such an approach, as typically practiced such as in Poland, usually employs what can be described as the phage bank approach, where isolated bacterial pathogens are tested for susceptibility to pre-prepared phage stocks . Indeed, periodically bacteria can be retested, given extended phage treatment such as of chronic infections, with the phages employed in the treatment of specific individuals revised if there is evidence of bacterial resistance . It is also possible to isolate phages from environmental samples against specific bacterial strains that have been isolated from patients, though obviously this can introduce additional delays as well as costs. The development of phage cocktails can be viewed as employing just such techniques, only peformed well in advance of patient treatment.
Mechanisms of action
Antibiotics possess a wide array of mechanisms of action. Penicillins inhibit the formation of the cell wall , for example, while quinolones inhibit DNA gyrase  and topoisomerase  in bacteria. By contrast, phage infections are multistep processes with phages genetical killing of bacteria generally occurring well prior to their physiologically termination via lysis [59-61]. In this section, for the sake of illustration of the diversity of phage antibacterial actions, we provide brief overviews of general modes of phage-mediated bacterial killing. See Figure 2 for illustration of the relationship of these different approaches to bacterial killing to passive versus active treatments.
Figure 2: Phage impact on bacteria. “Phages” refers to intact phage particles whereas “Enzybiotics” are purified phage products. Treatment failure is associated with substantial bacterial survival (black outcomes with white text). Passive or inundative treatment occurs when sufficient phage numbers are provided that substantial bacterial losses are achieved without reliance on phage population growth (darker gray endpoints with white text , all but “Productive infection” and “Phage population growth”, give rise to obligately passive treatments). Active treatment, indicated as a white outcome with black text, cannot substantially eradicate target bacterial populations unless sufficient phage population growth occurs in situ. Phage infections in which bacteria die and no phage progeny are produced we define as abortive. Note that CRISPR systems do not necessarily result in bacterial survival in combination with phage death but rather can also give rise to abortive infections .
Note to editor, the reference  has been temporarilty placed here to preserve numbering order since it is found in the legend to Figure 2.
Lysis from without: Certain phages, at high multiplicities of adsorption, can display the emergent property of being able to lyse bacteria without first infecting . The result can be viewed as a type of abortive infection, that is, a phage infection that results in the death of both adsorbed bacterium and adsorbing phage(s). Claims of lysis from without should always include empirical evidence that phages can lyse bacteria in a manner that (i) occurs soon after phage adsorption, (ii) does not result in phage production, and (iii) is more likely, in terms of bacterial lysis, given higher versus lower phage multiplicities of adsorption, e.g., 50 versus only 5. By contrast, and despite numerous publications which appear to suggest otherwise, simply showing that phages kill bacteria following high multiplicity adsorption does not represent evidence for lysis from without, and this is so even if phage progeny production also does not occur.
Abortive infection: When phage infection productivity is completely blocked but bacteria nonetheless die, the result can be described as an abortive infection. These can be a consequence of expression of specific bacterial anti-phage resistance genes [44,46]. They also can result from an absence in bacteria of specific factors that are required for productive phage infection in combination with a presence of those factors necessary to support phage-mediated bacterial killing. As with lysis from without, phage therapy mediated by phages that kill bacteria solely abortively by necessity represents a form of passive rather than active treatment.
Though bacterial genome sequence determinations, including those of bacterial plasmids, in principle can allow identification of abortive infection-effecting genes, the interactions of bacterial and phage gene products can be hard to predict. Thus, whether presumptive bacterial abortive-infection genes will give rise to functional abortiveinfection systems that block the infection of specific phage types must be confirmed experimentally, as too what phages are susceptible to these systems. In practice, as with elucidation of other phageassociated emergent properties, one simply tests phage isolates against bacterial strains to determine bacterial susceptibility. Due to abortive infection systems, though, it is possible for otherwise obligately lytic phages to inhibit bacterial replication while at the same time failing to productively infect those same bacteria. The potential for phage replication in the presence of target bacteria, rather than simply killing bacteria, therefore also must be experimentally determined, such as via plaque assays or broth-culture lysis (following low-multiplicity phage application), if both active and empirical therapy are sought (Figure 2).
Lysis from within: The more typical scenario resulting in death of a bacterium, in the course of phage therapy, occurs via a phage lytic cycle. The resulting lysis is described as a lysis from within to distinguish it from the above-noted lysis from without. Given multiple mechanisms by which phages can mortally impact infected bacteria during productive infections (above), the cell lysis event may be viewed simply as an additional layer of bacterial killing. Nonetheless, lysis from within can be an important aspect of phage therapy since bacterial killing without associated progeny release, that is, release as mediated by the lysis from within of infected bacteria, is effectively an abortive infection and therefore cannot support active phage therapy.
Enzybiotics: Certain phage molecules that possess antibacterial activity can act from without rather than from within, that is in the absence of phage presence rather than in the course of phage infection. The harnessing of purified forms of these molecules as a means of antibacterial treatment has been described as enzybiotics [64, 65]. The two most prominent enzybiotics are phage endolysins and phage extracellular polymeric substance (EPS) depolymerases.
Endolysins are the enzymes that effect lysis from within and typically, as noted, are not what actually kills bacteria though certainly are key to bacterial destruction along with phage progeny release from infected bacteria. Endolysins can, however, be used in a purified form to kill especially Gram-positive bacteria, which lack the protection afforded by the Gram-negative outer membrane. This process also can be described as lysis from without, though it is a form of lysis from without that differs from the more traditionally described phenomenon discussed above . Of interest, whereas penicillin binding proteins, i.e., the targets of penicillin and related antibiotics, can be easily mutated to no longer bind penicillin while still performing their peptidoglycan synthesis function , the actual bonds of peptidoglycan are not as easily modified. The result is a low potential for bacterial populations to evolve endolysin resistance.
EPS depolymerases affect particularly bacterial biofilms though are not directly bactericidal [67,68]. Instead, they exert their impact on extracellular polymers associated with retaining biofilm integrity. Thus, upon lysis, enzymes can be released from phage infections that can both increase phage penetrating ability into biofilms and otherwise potentially make biofilm bacteria more susceptible to antibacterial treatment (including, potentially, to antibiotic treatment). See Figure 2 for partial comparison of enzybiotic therapy to phage therapy.
Phage virions as antibacterial delivery agents
Most targets of antibiotics are bacterial molecules that are found inside of cells. This typically means that an antibacterial must be able to pass through the cell wall and/or cell membrane in order to take effect. The first molecule contacted by phages during the infection process, however, is found on the outside of the cell and phages subsequently have an inherent ability, within the limits of their host ranges, to penetrate through bacterial cell walls and membranes. Consequently, there is no need to design or even explicitly select phages for possession of this specific ability. Rather, a phage that has been shown to replicate in the presence of a specific bacterial type too must possess a bacterial cell-envelope penetrating ability. In this section we consider this phage ability in terms of virion modularity, that is, phage possession of a delivery component (virion proteins) in combination with one or more bactericidal mechanisms (encoded by phage DNA).
Naturally occurring emergent-property antibacterial agents: It is readily feasible to design or discover molecules that penetrate into body compartments as well as into bacteria, that interfere with bacterial metabolic processes, that are relatively lacking in side effects, that can target particular species of bacteria, or that possess reasonable stability during storage as well as following administration to patients. Identifying molecules that have all of these properties at once, however, can be quite difficult. This is why isolating natural antibacterial products has been a standard approach to antibacterial discovery since the introduction of antibiotics to clinical medicine. That is, antibiotics traditionally are natural, microorganism-produced molecules with an evolved ability to penetrate into and then either kill bacteria or reversibly inhibit their growth. Once discovered, antibacterial molecules are then chosen based upon their other desirable properties, as listed above. Numerous antibiotic features that are useful in terms of pharmacological functioning thus to a large extent are emergent properties, that is, evident only upon application to bodies. As discussed above, however, the bacteria-killing aspects of antibiotics themselves do not generally require antibiotic-body interactions to become manifest so are not similarly emergent in the larger scheme of drug pharmacology. Indeed, useful antibiotics almost by definition are chemicals that display antibacterial activities in a manner that is somewhat independent of whether they have been administered to a body environment.
Phages, too, are evolved antibacterial agents though substantially more complex than naturally occurring chemical antibacterials. Perhaps most notably, though, phages are both antibacterial delivery systems (virions) and antibacterial synthesizing agents (phage infections). As a consequence, they can be naturally benign until they are presented with target bacteria and, of course, are even capable of increasing in number once bacteria have been acquired. As a consequence, and despite their underlying complexity, professionally lytic phages tend to be highly safe antibacterials. In addition, from phage interactions with bacteria emerge the relatively straightforward phage properties of bacterial lysis and production of new, progeny phages, though the details of these infections themselves can be somewhat complex . Natural selection, in part as a consequence of phage modularity, thus has resulted in substantial constraints on phage emergent properties, which, for professionally lytic phages, typically are limited to bacterial acquisition, phage progeny production, and bacterial lysis from within. These constraints exist despite or perhaps even as a consequence of the substantially greater complexity that phages display relative to naturally occurring antibiotics.
Another way of making this point is to describe professionally lytic phages as possessing relatively complex genetic programs whose primary function is the production of new phages rather than the a general interference with metabolisms, other than the metabolisms of bacteria they are directly infecting (and even then, only as such interference is useful towards the endpoint of making and then release new, progeny virions). Furthermore, when not actively infecting bacteria, the primary phage function is one of bacterial acquisition, which is accomplished in a mostly non-metabolically active manner. Phage killing power thus is focused both metabolically and spatially on target bacteria, resulting, with professionally lytic phages, in only chance interference with animals physiologically, and seemingly little likelihood of even inadvertent adverse consequences given the nonxenobiotic nature of phage molecules. Antibiotics, by contrast, exist as somewhat general disrupters of metabolic processes, resulting in the evolution of less targeted processes and thereby a greater potential for disruption of human and animal physiology.
Modularity as a simplifying characteristic: Phage properties can be distinguished into those that operate extracellularly versus those that operate instead intracellularly. The extracellular aspects determine formulation stability, most phage pharmacokinetic characteristics, and target bacterium acquisition properties including host range specificity. Indeed, a phage virion in principle is simply a bacterium-acquisition devise whose sole function is the delivery of the intracellular acting agents to bacterial cytoplasms. To reach bacterial targets, by contrast, chemical antibiotics must inherently cross membranes by being relatively small and hydrophobic, have targets that are found external to membranes such as cell walls, be able to pass through porins found in the Gram-negative outer membranes, or otherwise must be associated with supplied bacteria-penetrating mechanisms. Bacteria-penetrating properties, however, are not necessarily benign within the body environment to the extent that they can also give rise to eukaryotic cell penetration, or simply plasma membrane disruption. Thus, the phage potential to penetrate bacterial envelopes without first dissolving in bacterial membranes likely contributes greatly to their relative dearth of negative pharmacodynamic aspects. That is, phages as antibacterial agents may be relatively safe in part because they are large and complex rather than in spite of these properties.
Once DNA has been delivered to a bacterium’s cytoplasm, the potential to encode and produce antibacterial agents are almost limitless. Indeed, not only can phages encode bacteria-toxic substances that further phage reproduction but so too phages can easily be engineered to produce antibacterial agents that have nothing to do with phage propagation, such as restriction endonucleases , all while preserving their inherent delivery mechanisms. By contrast, what is difficult to engineer are antibacterial activities that do not also interfere with the production of new phage virions. Fortunately, though, there often is little need to improve upon the antibacterial activity of naturally occurring phages as phages with novel properties typically may be isolated in the course of standard laboratory enrichment protocols.
In addition to this separation of extracellular and intracellular properties, phages are capable of varying their properties in response to different situations such as differences in target bacterium genotype and physiology. A result of these various phage emergent properties is that it is generally unnecessary to fully characterize every phage in terms of the mechanistic underpinnings of their antibacterial activity. A phage must be demonstrated to be safe through a combination of avoiding temperate phages and otherwise steering clear of bacterial virulence factors such as via bioinformatic analysis, but otherwise can be treated as essentially bacteria-killing black boxes which vary most importantly in terms of their individual antibacterial spectra of activity. Phages thus can be viewed as possessing assorted emergent properties that we need not fully nor, especially, mechanistically appreciate to effectively employ phages as antibacterial agents.
Modularity and the thwarting of bacterial resistance mechanisms: The intricacies of the phage infection cycle requires that many reactions successfully proceed and as a result there exist many opportunities for things to go wrong. Bacteria as a consequence have many steps available to them that they might interfere with to effect a resistance to specific phage types. Similarly, the specificity required for these various phage actions typically will limit which bacteria are susceptible to specific phage isolates. Analogous to mechanisms of antibiotic resistance, bacterial resistance to phage-mediated killing can be a consequence of both phage inactivation and modification of phage targets [44,46], though the targets which bacteria modify to prevent antibiotic-mediated killing, or factors involved in antibiotic inactivation, are typically not the same as for resisting phages. The result, minimally, is that bacterial evolution of antibiotic resistance typically is not expected to translate into cross resistance to phages.
Bacterial resistance to phages that is acquired mutationally typically involves loss or alteration of bacteria-encoded phage receptor molecules which are found on the surfaces of bacterial cells. However, and a crucial difference between phages and antibiotics as antibacterial agents, is that adsorption blocks act only on the extracellular, bacterial acquisition step of phage adsorption, which is highly variable among phages . Restoration of antibacterial ability by phages consequently can require solely modification of the extracellular step, with little need to modify the intracellular infection/bactericidal program. Such adsorption affinity modification often can be attained via simple laboratory adaptation experiments . Alternatively, and more conceptually similar to rational-based drug design, phages can be genetically engineered so that they display affinities for new phage receptors . In actual practice the solution to the occurrence of bacterial resistance often is far simpler: One simply isolates new phages, ones which vary both in extracellular and intracellular characteristics.
To summarize, (i) phages possess narrow spectra of activity, that is, their host range [44,73]; (ii) they are affected by resistance mechanisms that are not thought to share specificities with antibioitic-resistance mechanisms; (iii) it is extremely difficult for bacteria to acquire a general resistance to phages or even resistance to a majority of possible phages, and, as a consequence, (iv) it often is relatively straightforward to isolate from the environment new, professionally lytic phage isolates that are specific for target bacterial pathogens while at the same time are unaffected by a given spectrum of bacterial anti-antibiotic as well as anti-phage resistance mechanisms. Substantial exploitation of naturally occurring anti-resistance properties found among phages in phage therapy, however, is highly dependent on a regulatory framework that allows for or even encourages the development of what potentially could be an embarrassing wealth of both safe and diverse antibacterial agents, that is, rather than focusing predominantly on a relatively few phages that have been subject to the full spectrum of regulatory inquiry. Indeed, if we simply redefined antibiotics to include phages, and then took advantage of the diverse potential for phages to overcome bacterial resistance mechanisms to both phages and antibiotics, the antibioticresistance crisis likely would be much smaller in scope .
Pharmacokinetics is the study of the impact of bodies on drugs where body can be defined to include not just animal cells and tissues but also associated microflora. Typically this impact is differentiated into various mechanisms that together describe barriers to the achievement of effective densities at a drug’s site of action. Drugs thus are subject to absorption, distribution, metabolism, and excretion (together, ADME). These are drug movement into the blood, particularly from the gut though also by more direct routes; drug movement from the blood to other body compartments; drug chemical modification; and movement of intact drug out of the body, respectively.
Antimicrobial drugs thus are administered to patients and in some manner reach bacterial targets where they can have positive effects. All the while, however, phages are also blocked in their movement, diluted, chemically destroyed, removed from the body, or otherwise sequestered away from bacterial targets. For additional discussion of these ideas as they pertain to phage therapy, see Abedon and Thomas-Abedon  along with Abedon . In this section we consider additional, more specific aspects of pharmacokinetics as this subject can be applied to phages. See Figure 3 for summary of pharmacokinetics in general, with associated pharmacodynamics, and as compared with those of phage therapy.
Figure 3: Pharmacokinetics basics (panel A) as compared with phage therapy basics (panel B). Phage therapy as typically practiced often differs in two basic ways from the ‘life cycle’ of most drugs (panel A). First, absorption, that is, uptake into the blood, is not an issue for the sake of positive pharmacodynamic effects unless systemic treatment is sought. Second, most drugs, unlike phages, do not self amplify in situ following contact with target tissues.
Topical versus systemic administration: Topical phage administration, which includes application to both wounds and lungs, does not require absorption nor necessarily substantial distribution. With phages that are topically applied to bacteria, one instead can use the term penetration , as in phage penetration into bacterial biofilms. Such biofilm penetration by phages [67,68] may be more effective than the impact of chemical antibiotics . Indeed, we can apply the phrase “active penetration” to describe the phage process , reflecting the possibility that phage-induced bacterial lysis, particularly of bacteria found in the outer layers of bacterial biofilms, may be aiding phage penetration. This lysis either supplies phages directly to underlying bacteria or allows nutrients to diffuse to already established phage infections.
Oral phage delivery too can involve phage penetration without accompanying absorption or distribution. High stomach acidity can inactivate phages, however, and thus “penetration” through gastric juices should always be a concern given oral phage administration. Poorly understood, though, is the issue of penetration of phages to gastrointestinal bacteria once past the stomach. In particular, the question of whether or when phages can consistently display efficient active therapy within the gastrointestinal lumen is somewhat open and efforts to reduce bacterial densities using relatively low phage densities and/or few dosings are unlikely to be effective unless active treatment is in fact possible. Thus, not only is phage inactivation in the course of passage through the stomach a concern but so too phage dilution upon mixing with food, drink, or gastric secretion as dilution can also have the effect of reducing overall phage densities .
Phage application to the gastrointestinal tract of animals can result in uptake into blood, a process that has been dubbed bacteriophage translocation . Unfortunately, little is understood of the mechanism of this translocation, except that it is not always demonstrable [76-78]. When such absorption is anticipated following oral administration, then the application can be described as per os. By design, most antibiotics employed systemically can be delivered in this manner and since oral dosing can be highly convenient for patients, including in a non-clinical setting, it would be helpful were phages similarly deliverable. Consequently, a clear limitation to the potential of phage therapy as a general strategy of bacterial disease control is the substantial uncertainty that surrounds the phage facility to reach systemic circulation following oral delivery.
More direct approaches of administration to systemic circulation of course also exist. These, however, can be more problematic due to a combination of greater invasiveness, lower convenience, and a need for greater phage purification . Contrasting per os, though, delivering phages directly to systemic circulation may be more reliably achieved. In addition, passive treatment also is possible with direct administration whereas per os phage delivery would appear to be wholly dependent on active treatment, that is, phages entering systemic circulation in low numbers that are then bolstered upon contact with target bacteria. Human treatment using these more direct approaches to systemic delivery have been used though such efforts are not nearly as common as more local applications . A reasonable assumption is that the phage therapy niche as an alternative to chemical antibiotics may be biased towards local administration, probably including oral administration for gastrointestinal treatments, with phage treatment of systemic infections perhaps limited to circumstances in which antibiotic use is less permissible, for instance against antibioticresistant bacteria such as methicillin-resistant Staphylococcus aureus (MRSA) or vancomycin-resistant enterococci (VRE) . Indeed, a number of studies have either demonstrated in humans [79,80], shown in animals , or proposed  the phage therapy or biocontrol  of MRSA. Phage-based treatment of MRSA may even be less expensive than treatment with chemical antibiotics .
Distribution: Distribution is the movement of drugs out of the blood and into other body tissues. Achievement of this movement is a function of drug molecular properties and though it probably is reasonable to speculate that the phage potential for such movement occurs for reasons that are somewhat independent of the mechanisms employed by small-molecular drugs, otherwise the means by which phages achieve such movement is poorly understood [9,85]. What is crucial to keep in mind, however, is that issues of drug distribution towards positive pharmacodynamic effects only matter to the extent that drug application is systemic and intended to reach tissues other than blood. Another key issue is that disease, including bacterial infections, can increase tissue permeability such as to phages; see Abedon  for discussion. Also relevant is that the pharmacokinetic property of self amplification to at least some degree can compensate for inefficiencies in phage distribution: If phages can reach infections then they can have some impact on those infections and if their ability to replicate is sufficiently robust then they may be able to achieve a substantial impact.
Metabolism: Metabolism in a pharmacokinetic sense describes the chemical modification of drugs. Though this often is a means to the end of drug elimination from the body, in some cases metabolism instead can convert a prodrug into its active form. With chemical pharmaceuticals most metabolism occurs in the liver. There cytochrome P450 family enzymes make lipophilic xenobiotics, that is, molecules not found within the body, more polar and thus more easily excreted through the renal system . As phages consist mostly of proteins and DNA, they generally are not xenobiotic. In addition, in terms of phage pharmacokinetics, phage sequestration such as due to immune system action need not result in subsequent chemical modification to have phage levels reduced through metabolism so long as phages otherwise have been permanently separated from their bacterial targets.
Phages must be resupplied on an ongoing basis to the vicinity of target bacteria for phage therapy to succeed in the face of elimination from tissues. If effective phage densities are, for example, 108/ml, and phages are inactivated at a rate of, say, 1% per minute, then 106 new phages must be added to the system per minute just to sustain phage densities at effective levels. With active treatment not only must phages be sustained at effective levels but so too must those relatively high phage densities be attained in the first place, all of which must occur in the face of ongoing phage losses. Nonetheless, if phages are easily delivered, or otherwise can replicate sufficiently robustly in situ, then inactivation by immune system action need not be a substantial issue. Note that phage self amplification in situ too may be described as an action of metabolism [6,9] though of course not one associated with the action of liver enzymes.
Excretion: Phage elimination from bodies is mostly due to immune system-mediated sequestration or inactivation, as opposed to excretion, which is elimination of a drug in the same form as it was administered. Phages do have some access to urine, which may be useful for treatment of urinary tract infections following systemic phage administration . Other than phage delivery to such sites, excretion is not terribly relevant to phage therapy and particularly less so given topical rather than systemic application.
The impact of a standard chemical drug is difficult to predict due to the emergent properties that are associated with the complexity of the organisms they are treating, such as ourselves. With small-molecule antimicrobial drugs, that complexity is further increased owing to the multifaceted nature of the parasitic organisms being targeted. This means that negative impacts on target pathogens (i.e., positive pharmacodynamic effects) along with negative impacts on the body itself can be difficult to predict. When phages are properly chosen, by contrast, their negative impact on bodies can be both minimal and somewhat predictable, and this predictability can occur in spite of a lack of specific knowledge of the biochemical processes that phages encode. The result can be a much simpler path towards antibacterial development than one typically sees with chemical antibacterials, where toxicity testing often serves as a major road block .
The pharmacologically emergent properties of phages thus are mostly limited to their negative impact on target bacteria, that is, especially professionally lytic phages can possess biases towards pharmacologically desirable rather than undesirable effects. Phages in particular, and like antibiotics, are naturally occurring antibacterial agents whose positive pharmacodynamic effects have been honed by billions of years of organic evolution. Furthermore, to a degree the phage potential to self amplify, in situ, can obviate many pharmacokinetic issues that otherwise can negatively impact their potential to achieve minimally effective antibacterial densities in the vicinity of target bacteria. The upshot is an emergent property laden but nonetheless non-xenobiotic pharmaceutical that can be more predictable than traditional antibiotics at the ‘higher levels’ of antibacterial characterization, such as in terms of animal or clinical testing, even as lower-level, especially biochemical understanding of the phage impact on target bacteria can be vastly more difficult to achieve.