Bacteria resistant to most or all available antibiotics are causing increasingly serious problems, raising widespread fears of returning to a pre-antibiotic era of untreatable infections and epidemics. Despite intensive work by drug companies, no new classes of antibiotics have been found in the last 30 years. There are hopes that the newfound ability to sequence entire microbial genomes and to determine the molecular bases of pathogenicity will open new avenues for treating infectious disease, but other approaches are also being sought with increasing fervor. One result is a renewed interest in the possibilities of bacteriophage therapy — the harnessing of a specific kind of viruses that attack only bacteria to kill pathogenic microorganisms (cf. Levin and Bull, 1996; Lederberg, 1996; Radetsky, 1996; Barrow and Soothill, 1997).
Phage therapy was first developed early in this century and showed much promise but also aroused much controversy. It has been little used in the West since the advent of antibiotics in the 1940s. However, extensive clinical research and implementation of phage therapy continued to be carried out in Eastern Europe over the last 50 years. The results of that work effectively complement the limited recent animal work in the West that is primarily cited in the recent articles, encouraging optimism that phage can indeed play an important role in dealing with infections involving increasingly drug-resistant microbes. We need to draw as much as possible on the largely-unknown body of knowledge that has accumulated in Poland, France and many parts of the former Soviet Union (FSU) as we again explore phage therapy, and to give credit where it is due for the many years of hard, careful work they have invested in the field. This paper is written primarily to put phage therapy in historical and ecological context and to explore some of the more interesting and extensive work in Eastern Europe, little of which has been accessible in English.
The Nature of Bacteriophages
Viruses are like space ships that are able to carry genetic material between susceptible cells and then reproduce in those cells, just as the AIDS virus, HIV, specifically infects human T lymphocytes which carry a particular surface protein called CD4. Each virus consists of a piece of genetic information, determining all of the properties of the virus, which is carried around packaged in a protein coat (Fig.1). In the case of bacteriophages, the targets are specific kinds of bacterial cells; they cannot infect the cells of more complex organisms because of major differences in key intracellular machinery as well as in cell-surface proteins. Most phages have tails, the tips of which have the ability to bind to specific molecules on the surface of their target bacteria (Fig. 2). The viral DNA is then injected through the tail into the host cell, where it directs the production of progeny phages — often over a hundred in half an hour. Each strain of bacteria has characteristic protein, carbohydrate and lipopolysaccharide molecules present in large quantities on its surface. These molecules are involved in forming pores, in motility, in binding of the bacteria to particular surfaces; each such molecule can act as a receptor for particular phages. Development of resistance to a particular phage generally reflects mutational loss of its specific receptor; this loss often has negative effects on the bacterium and does not protect it against the many other phage which use different receptors.
Each kind of bacteria has its own phages, which can be isolated wherever that particular bacterium grows – from sewage, faeces, soil, even ocean depths and hot springs. The process of isolation is easy. Just let the sample sit in an appropriate nutrient broth, separate off the liquid part, and pass it through a filter with pores so tiny that bacteria can’t get through. Then mix it (at several different dilutions) with a culture of the bacteria in question. Spread a few drops on a block of appropriate nutrient medium which is made firm with agar taken from seaweed. The next day, one sees a dense covering or lawn of bacteria with round clear spots, called plaques. Each plaque contains many million phage particles, all progeny of one phage which was immobilised there on the agar. That phage infected a cell, multiplied inside it, and caused it to burst. This released many phages, which infected nearby cells and repeated the process. One can stick a toothpick into one of these plaques, transfer it to a fresh culture of the bacteria in liquid medium, and grow up a homogeneous stock of descendants of that particular phage, whose properties can then be studied.
A century ago, Hankin (1896) reported that the waters of the Ganges and Jumna rivers in India had marked antibacterial action which could pass through a very fine porcelain filter; this activity was destroyed by boiling. He particularly studied the effects on Vibrio cholerae and suggested that the substance responsible was what kept cholera epidemics from being spread by ingestion of the water of these rivers. However, he did not further explore the phenomenon. Edward Twort (1915) and Felix d’Herelle (1917) independently reported isolating filterable entities capable of destroying bacterial cultures and of producing small cleared areas on bacterial lawns, seemingly implying that discrete particles were involved. They are jointly given credit for the discovery. It was d’Herelle, a Canadian working at the Pasteur Institute in Paris, who gave them the name “bacteriophages”– using the suffix phage not in its strict sense of to eat, but in that of developing at the expense of (d’Herelle, 1922, p. 21) – and who made them a major part of his life’s work. D’Herelle, a largely self-trained microbiologist, had just spent 10 years in Guatemala, Mexico and Argentina. There, he dealt with epidemics of dysentery, yellow fever and a coffee-killing fungus, isolated a bacterium from dying locusts to use in controlling locust plagues, and explored several interesting fermentation challenges – all good preparation for his later work with phages, as discussed in an interesting fashion by Summers (1998). At the Pasteur Institute, he was carrying out a careful study of vaccine preparation techniques using a model system – “B. typhimurium” in its natural host, mice; he felt strongly that meaningful data on immunity and pathogenicity could only be obtained when natural hosts were used. In his spare time, he was also doing research with dysentery patients – a frequent problem in wartime France. From the faeces of several of these patients, he isolated a filterable anti-Shiga “microbe” which multiplied through many serial passages on its host bacterium, and which could produce tiny clear circles on a plate of this “Shiga bacillus” (d’Herelle, 1917).
D’Herelle went on to carefully characterize bacteriophages as viruses which multiply in bacteria and worked out the details of infection by various phages of different bacterial hosts under a variety of environmental conditions, always working to combine natural phenomena with laboratory findings, to better understand immunity and natural healing from infectious disease (Summers, 1998). The Ninetieth Annual Meeting of the British Medical Association in Glasgow featured a very interesting discussion between d’Herelle, Twort and several other eminent scientists of the day on the nature and properties of bacteriophages (d’Herelle et al, 1922). The main issue at that time was whether the observed bacteriolytic principle was an enzyme produced by bacterial activity or a form of tiny virus with some sort of life of its own, as claimed by d’Herelle; this controversy continued for many years, splitting the rapidly-growing community of people working with phages.
D’Herelle summarized the early phage work in a 300-page book “The Bacteriophage” (1922). He wrote classic descriptions of plaque formation and composition, infective centers, the lysis process, host specificity of adsorption and multiplication, the dependence of phage production on the precise state of the host, isolation of phages from sources of infectious bacteria and the factors controlling stability of the free phage. He quickly became fascinated with the apparent role of phages in the natural control of microbial infections. He noted for example the frequent specificities of the phages isolated from recuperating patients for their own disease organisms and the rather rapid variations over time in their phage populations. He thus worked throughout his life to develop the potential of using properly selected phages as therapeutic agents against the most devastating health problems of the day. However, he initially focussed on simply understanding phage biology. Thus, the first known report of successful phage therapy came not from d’Herelle but from Bruynoghe and Maisin (1921), who used phage to treat staphylococcal skin infections.
After a year at the Pasteur Institute of Saigon, d’Herelle returned to tight physical conditions, personal conflict and intellectual controversy at the Pasteur Institute in Paris. He soon accepted an offer to move to the Netherlands, where he was provided better conditions for his work with the recovery from infectious disease and the properties of bacteriophages, published his first book and a number of papers, and received an honorary MD degree. In 1925, he became a health officer for the League of Nations, based in Alexandria, Egypt, with special responsibility for controlling infectious disease on ships passing through the Suez Canal and during some of the major Muslim pilgrimages. Phage therapy and sanitation measures were the primary tools in his arsenal to deal with major outbreaks of infectious disease throughout the Middle East and India. Throughout this period, he continued publishing on his research and clinical trials and assisting others who were willing to do so with phages and consultations, often undertaking extended travel at his own expense. One of the most extensive trials of phage therapy he helped set up was the Bacteriophage Inquiry of 1927-1936 (Summers, 1993), which led to “what seems to be convincing results, endorsed by august committees” yet still left many skeptics of phage therapy; these studies deserve closer scrutiny.
In 1928, d’Herelle was invited to Stanford to give the prestigious Lane Lectures; his discussion of “The Bacteriophage and its Clinical Applications” was published as a monograph (d’Herelle and Smith, 1930). He gave many lectures for medical schools and societies as he crisscrossed the country. He then went on to Yale to take up a regular faculty position, arranged with the support of George Smith, who had translated his first two books into English. He continued to spend summers in Paris working with the phage company he had established there, run by his son-in-law, in response to strong demands for phage preparations with careful quality control; this period is discussed particularly well by Summers (1998). He returned permanently to Europe in 1933, spending much time the following two years in Tiflis (Tbilisi), Georgia, helping to set up an international Bacteriophage Institute there, as discussed further below.
From early on, one major practical use of phages was for bacterial identification through a process called phage typing — the use of patterns of sensitivity to a specific battery of phages to precisely identify microbial strains. This technique takes advantage of the fine specificity of many phages for their hosts and is still in common use around the world. The sophisticated ability of phages to destroy their bacterial hosts can also have a very negative commercial impact; phage contaminants occasionally spread havoc and financial disaster for the various fermentation industries that depend on bacteria, such as cheese production and fermentative synthesis of chemicals (cf. Saunders, 1994)
Phage therapy was tried extensively and many successes were reported for a variety of diseases, including dysentery, typhoid and paratyphoid fevers, cholera, and pyogenic (pus-
producing) and urinary-tract infections. Phages were poured directly into lesions, given orally or applied as aerosols or enemas. They were also given as injections — intradermal, intravascular, intramuscular, intraduodenal, intraperitoneal, even into the lung, carotid artery and pericardium. The early strong interest in phage therapy is reflected in the fact that some 800 papers were published on the topic between 1917 and 1956; the results have been discussed in some detail by Ackermann and Dubow (1987). The reported results were quite variable. Many physicians and entrepreneurs became very excited by the potential clinical implications and jumped into applications with very little understanding of phages, microbiology or basic scientific process. Thus many of the studies were anecdotal and/or poorly controlled, many of the failures were predictable and some of the reported successes did not make much scientific sense. Often, uncharacterized phages at unknown concentrations were given to patients without specific bacteriological diagnosis, and there is no mention of follow up, controls or placebos.
Much of the understanding gained by d’Herelle was ignored in this early work, and inappropriate methods of preparation, “preservatives” and storage procedures were often used. On one occasion, d’Herelle reported testing 20 preparations from various companies and finding that not one of them contained active phages (Summers, 1998). On another occasion, a preparation was advertised as containing a number of different phages, but it turned out that the technician responsible had decided it was easier to grow them up in one large batch than in separate batches. Not too surprisingly, checking the product showed that one phage had outcompeted all the others and this was not, in fact, a polyvalent preparation. This was the origin of the phage T7, whose RNA polymerase now plays a major role in biotechnology (William Summers, personal communication). In general, there was no quality control except in a few research centers. Large clinical studies were rare and the results of those few were largely inaccessible outside of Eastern Europe.
In 1931, an extensive review of bacteriophage therapy was commissioned by the Council on Pharmacy and Chemistry of the American Medical Association (Eaton and Bayne-Jones, 1931). Its purpose was “(a) to present summaries and discussions of (1) the experimentally determined facts relating to the bacteriophage phenomenon, (2) the laboratory and clinical evidence for and against the therapeutic usefulness of bacteriophage and (3) the relation of so-called antivirus to materials containing bacteriophage, and (b) to serve as a basis for a survey of the status of some of the commercial preparations.” With 150 references, this report made a major effort to survey at least what they considered the most significant papers and reviews. In evaluating this report, it is important to realise how little was yet known then about bacteriophages. In fact, their first conclusion was “Experimental studies of the lytic agent called “bacteriophage” have not disclosed its nature. D’Herelle’s theory that the material is a living virus parasite of bacteria has not been proved. On the contrary, the facts appear to
indicate that the material is inanimate, possibly an enzyme.” In retrospect, the proof that phages are viruses looks solid and it is hard to see how they could have come to this conclusion, which clearly impacted all of their other findings. These included: “2.) Since it has not been shown conclusively that bacteriophage is a living organism, it is unwarranted to attribute its effect on cultures of bacteria or its possible therapeutic action to a vital property of the substance. 3.) While bacteriophage dissolves sensitive bacteria in culture and causes numerous modifications of the organisms, its lytic action in the body is inhibited or greatly impeded by blood and other bodily fluids. 4.) The material called bacterophage is usually a filtrate of dissolved organisms, containing, in addition to the lytic principle, antigenic bacterial substances, products of bacterial growth and constituents of the culture medium. The effects of all these constituents must be taken into consideration whenever therapeutic action is tested. 5) A review of the literature on the use of bacteriophage in the treatment of infections reveals that the evidence for the therapeutic value of lytic filtrates is for the most part contradictory. Only in the treatment of local staphylococcic infections and perhaps cystitis has evidence at all convincing been presented.”
This assessment clearly had a strong influence on the investment of the medical community in exploring phage therapy seriously, at least in the United States. Points are raised which still need to be considered, particularly in terms of the many trials described there in animals or humans which seemed to show little or no success and in terms of such potentially confounding explanations of the successes as the apparent strong stimulation of natural immune mechanisms by the bacterial debris in the lysates used. Then in the 1940′s, the new “miracle” antibiotics such as penicillin became became widely available, and phage therapy was largely abandoned in the western world.
Specific Problems of Early Phage Therapy Work
Today, many believe that phage therapy was proven not to work in the early part of this century. However, it appears that it simply was never given sufficient and appropriate trial, and reassessment is warranted. It is thus important to consider in some detail potential reasons for the early problems and the questions as to efficacy. These included:
1.Paucity of understanding of the heterogeneity and ecology of both the phages and the bacteria involved.
2.Failure to select phages of high virulence against the target bacteria before using them in patients.
3.Use of single phages in infections which involved mixtures of different bacteria. 4.Emergence of resistant bacterial strains. This can occur by selection of resistant mutants (a frequent occurrence if only one phage strain is used against a particular bacterium) or by lysogenization (if temperate phages are used, as discussed below).
5.Failure to appropriately characterize or titer phage preparations, some of which were totally inactive.
6.Failure to neutralize gastric pH prior to oral phage administration.
7.Inactivation of phages by both specific and nonspecific factors in body fluids.
8.Liberation of endotoxins as a consequence of widespread lysis of bacteria within the body (which physicians call the Herxheimer reaction). This can lead to toxic shock, and is a potential problem with chemical antibiotics as well.
9.Lack of availability or reliability of bacterial laboratories for carefully identifying the pathogens involved, necessitated by the relative specificity of phage therapy.
In making the choice to again explore the possibilities of phage therapy, we should also consider their many potential advantages, discussed in more detail below:
1.They are both self-replicating and self-limiting, since they will multiply only as long as sensitive bacteria are present and then are gradually eliminated from the individual and the environment.
2.They can be targeted far more specifically than can most antibiotics to the specific problem bacteria, causing much less damage to the normal microbial balance in the body. The bacterial imbalance or “dysbiosis” caused by treatment with many antibiotics can lead to serious secondary infections involving relatively resistant bacteria, often extending hospitalization time, expense and mortality. Particular resultant problems include infection by pseudomonads, which are especially difficult to treat, and Clostridium difficile, cause of serious diarrhea and membranous colitis (cf. Fékéty, 1995).
3.Phages can often be targeted to receptors on the bacterial surface which are involved in pathogenesis, so that any resistant mutants are attenuated in virulence.
4.Few side effects have been reported for phage therapy.
5.Phage therapy would be particularly useful for people with allergies to antibiotics. 6.Appropriately selected phages can easily be used prophylactically to help prevent bacterial disease in people or animals at times of exposure, or to sanitize hospitals and help protect against hospital-acquired (nosocomial) infections.
7.Especially for external applications, phages can be prepared fairly inexpensively and locally, facilitating their potential applications to underserved populations.
8.Phage can be used either independently or in conjunction with other antibiotics to help reduce the development of bacterial resistance.
Properties of Phages
One major source of confusion in the early phage work was the perception that all phages were fundamentally similar, though subject to adaptive change depending on the recent conditions of growth. One consequence of this was that often new phages were isolated for each series of studies, so that there was little continuity or basis for comparison. Phages specific for over 100 bacterial genera have now been isolated (Ackermann, 1996); they have been found virtually everywhere that they have been sought. However, only few have yet been well studied or classified (cf. Ackermann and DuBow, 1987)
A second early source of confusion affecting therapeutic uses was the question of whether the lytic principle termed “bacteriophage” simply reflected an inherent property of the specific bacteria or required regular reinfection by an external agent. During the 1930s and 1940s, it became increasingly clear that in some senses both were true — that there were in fact two quite fundamentally different groups of bacteriophages. Lytic phages always have to infect from outside, reprogram the host cell and release a burst of phage through breaking open, or lysing, the cell after a relatively fixed interval. Lysogenic phages, on the other hand, have another option. They can actually integrate their DNA into the host DNA, much as HIV can integrate the DNA copy of its RNA, leading to virtually permanent association as a prophage with a specific bacterium and all its progeny. The prophage directs the synthesis of a repressor, which blocks the reading of the rest of its own genes and also those of any closely-related lysogenic phages — a major advantage for the bacterial cell. Many prophages further aid their host by helping protect against various unrelated, lytic phages. Occasionally, a prophage escapes from regulation by the repressor, cuts its DNA back out of the genome by a sort of site-specific recombination and goes ahead to make progeny phage and lyse open the cell. Sometimes the cutting-out process makes mistakes and a few bacterial genes get carried along with the phage DNA to its new host; this process, called transduction, plays a significant role in bacterial genetic exchange. Such lysogenic phages are very bad candidates for phage therapy, both due to their mode of inducing resistance and to the fact that they can potentially lead to transfer of genes involved in bacterial pathogenicity; this is discussed in more detail below. However, their specificity often makes them very useful for phage typing in distinguishing between bacterial strains.
Key technical developments that helped clarify the general nature and properties of bacteriophages included: (1) the concentration and purification of some large phages by means of high-speed centrifugation and the demonstration that they contained equal amounts of DNA and protein (Schlesinger, 1933 a, b) and (2) visualization of phages by means of the electron microscopic (EM) (Ruska,1940; Pfankuch and Kausche, 1940). Soon after, Ruska (1943) reported the first attempts to use the EM for phage systematics; this has since become a key tool of the field (cf. Ackermann and DuBow, 1987). Each phage was found to have its own specific shape and size, from the “lunar lander”-style complexity of T4 and its relatives to the globular heads with long or short tails of lambda and T7 to the small filamentous phages that looked much like bacterial pili (Fig. 3., from Ackermann, 1996)
A much better understanding of the interactions between lytic phage and bacteria came from detailed one-step growth curve experiments expanding on the work of d’Herelle (1922) (Ellis and Delbrück, 1939, Doermann,1952). These demonstrated an eclipse period during which the DNA began replicating and there were no free phage in the cell, a period of accumulation of intracellular phage, and a lysis process which released the phage to go in search of new hosts. An example of this phage infection cycle is outlined in Fig. 4.
In the early 1940′s, developments occurred which were to have a major impact on the orientation of phage research in the United States and much of western Europe, strongly shifting the emphasis from practical applications to basic science. Physicist-turned-phage-biologist Max Delbrück began working with key phage biologists Alfred Hershey and Salvador Luria and formed the “Phage Group”, which eventually expanded dramatically with aid of the summer Phage Courses at Cold Spring Harbor, Long Island. These ran for many years starting in 1945 and regular phage meetings still continue there.
The influence of the Phage Group on the origins of molecular biology has been well documented (cf. Cairns et al, 1966; Fischer and Lipson, 1988; Summers, 1993b). Virulent phages had just the right balance of complexity and simplicity to tease out the key concepts of cell regulation at the molecular level. However, a major element of the rapid success of phage as model systems was that Delbrück convinced most phage biologists in the United States to focus on one bacterial host (E. coli B) and 7 of its lytic phages, building a very strong, tightly focussed community all working on the same set of problems, able to build effectively on each other’s work and communicate easily. The 7 phages were arbitrarily chosen and named T(type)1-T7. As it turned out, T2, T4 and T6 were quite similar to each other, defining the “T-even” family of phages, discussed in more detail below. These phages were key in demonstrating that DNA is the genetic material, that viruses can encode enzymes, that gene expression is mediated through special copies in the form of “messenger RNA”, that the genetic code is triplet in nature, and many other fundamental concepts. The negative side of this strong focus on a few phages growing under rich laboratory conditions, however, was that there was very little study or awareness of the ranges, roles and properties of bacteriophages in the natural environment, or of phages that infect other kinds of bacteria.
Rational Phage Therapy
The rapid, powerful developments in the understanding of phage biology had the potential to facilitate more rational thinking about the therapeutic process and the selection of therapeutic phages. However, there was generally little interaction between those who were so effectively using phage as tools to understand molecular biology and those still working on phage ecology and therapeutic applications. Many in the latter group were spurred on by concern about the increasing incidence of nosocomial (hospital-acquired) infections and of bacteria resistant against most or all known antibiotics. This is particularly true in Poland, France and the former Soviet Union where use of therapeutic phages never fully died out and where there has been some ongoing research and clinical experience. In France, Dr. Jean-François Vieu led the therapeutic phage efforts until his retirement some 10 years ago, he worked in the “Service des Entérobactéries” of the Pasteur Institute in Paris and, for example, prepared Pseudomonas phages on a case-by-case basis for patients. The experience there is discussed in Vieu (1975) and Vieu et al. (1979). Phage therapy was used extensively in many parts of Eastern Europe as a natural part of clinical practice, and there are now companies in Moscow and several other Russian cities making phage preparations for this purpose. However most of the research and much of the phage preparation came under the direction of key centers in Tbilisi, Georgia and in Wroclaw, Poland. I will thus focus on the work of these two groups.
Polish Academy of Sciences, Wroclaw
The most detailed publications documenting phage therapy have come from Stefan Slopek’s group at the Institute of Immunology and Experimental Medicine of the Polish Academy of Sciences in Wroclaw. This group published a series of detailed papers in the Archivum Immunologiae et Therapie Experimentalis (cf. Slopek et al, 1983, 1985, 1987), describing the results of phage treatments carried out from 1981 to 1986 with 550 patients. This set of studies involved ten Polish medical centers, including the Wroclaw Medical Academy Institute of Surgery Cardiosurgery Clinic. Children’s Surgery Clinic and Orthopedic Clinic; the Institute of Internal Diseases Nephrology Clinic and Clinic of Pulmonary Diseases. The patients ranged in age from 1 week to 86 years; in 518 of the cases, phage use followed unsuccessful treatment with all available antibiotics. The major categories of infections treated were long-persisting suppurative fistulas, septicemia, abscesses, respiratory tract suppurative infections and bronchopneumonia, purulent peritonitis and furunculosis. In a final summary paper (Slopek et al, 1987), the authors carefully analyzed the results with regard to such factors as nature and severity of the infection and monoinfection vs. infection with multiple bacteria. Rates of success ranged from 75 to 100 % (92% overall), as measured by marked general improvement of health, tendency to heal of local wounds and disappearance of titratable bacteria; 84% demonstrated full elimination of the suppurative process and healing of local wounds. Infants and children did particularly well; not surprisingly, the poorest results came with the elderly and those in the final stages of extended serious illness, with weakened immune systems and generally poor resistance.
The bacteriophages used all came from the extensive collection of the Bacteriophage Laboratory of the Institute of Immunology and Experimental Therapy; in the later studies, some of the specific phages used were named. All were virulent, capable of completely lysing the bacteria being treated. In the first study alone, 259 different phages were tested (116 for Staphylococcus, 42 for Klebsiella, 11 for Proteus, 39 for Escherichia, 30 for Shigella, 20 for Pseudomonas, and one for Salmonella); 40% of them were selected to use directly for therapy. All of the treatment was in a research mode, with the phage prepared at the Institute by standard methods and tested for sterility. Treatment generally involved 10 ml of sterile phage lysate orally half an hour before each meal, with gastric juices neutralized by (basic) Vichy water, baking soda or gelatin. In addition, phage-soaked compresses were generally applied three times a day where dictated by localized infection. Treatment ran for 1.5-14 weeks, with an average of 5.3; for intestinal problems, short treatment sufficed, while it was very long for such problems as pneumonia with pleural fistula and pyogenic arthritis. Bacterial levels and phage sensitivity were continually monitored, and the phage(s) being used were changed if the bacteria lost their sensitivity; therapy was generally continued for two weeks beyond the last positive test for the bacteria.
Few side effects were observed; those that were seen seemed directly associated with the therapeutic process. Pain in the liver area was often reported around day 3-5, lasting several hours; the authors suggested that this might be related to extensive liberation of endotoxins as the phage were destroying the bacteria most effectively. In severe cases with sepsis, patients often ran a fever for 24 hours about days 7-8 (Slopek et al, 1981a). Various other methods of administration were successfully used, including aerosols and infusion rectally or in surgical wounds. Intravenous administration was not recommended for fear of possible toxic shock from bacterial debris in the lysates (Slopek et al, 1981a). However, it was clear that the phages readily got into the body from the digestive tract and multiplied internally wherever appropriate bacteria were present, as measured by their presence in blood and urine as well as by therapeutic effects (Weber-Dabrowska et al, 1987). This interesting and rather unexpected finding has been replicated in other studies and systems (find references**).
Detailed notes were kept throughout on each patient. The final evaluating therapist also filled out a special inquiry form that was sent to the Polish Academy of Science research team along with the notes. The Computer Center at Wroclaw Technical University carried out the extensive analyses of the data. The authors used the categories established in the WHO (1977) International Classification of Diseases in assessing results. They also looked at the effects of age, severity of initial condition, type(s) of bacteria involved, length of treatment and other concomitant treatments. The papers include many specific details on individual patients which help give insight into the ways phage therapy was used, as well as an in-depth analysis of difficult cases.
Bacteriophage Institute, Tbilisi
The most extensive and least widely known work on phage therapy was carried out under the auspices of the Bacteriophage Institute at Tbilisi, d’Herelle’s institute in the former Soviet republic of Georgia. The work there will thus be discussed in some detail.
Georgia is an ancient and beautiful country of 5 million people, lying tucked in a valley between the High Caucasus mountains at the south of Russia, the Samtsxe-Dgavaxete range bordering on Turkey and Armenia, and the Black Sea. Through all the centuries of political upheaval at this crossroads of the ancient world, it has managed to keep its own culture and unique language, which is related only (and remotely) to Basque. It has been Christian since the third century, but prides itself strongly on its openness to all religions and cultures, its synagogue, mosque, and various Christian churches all clustered in the heart of old Tbilisi. Strong emphasis is placed on culture, intellectual pursuits and hospitality; the literacy rate is 100%, according to the 1996 UN Human Development Report on Georgia, and it has a long tradition of excellence in fields from music to mathematics to wine-making and cooking.
According to various Georgian physicians with whom I have spoken, phage therapy is part of the general standard of care there, used especially extensively in pediatric, burn and surgical hospital settings. Phage preparation was carried out on an industrial scale, employing 1,200 people just before the break-up of the Soviet Union. Tons of tablets, liquid preparations and spray containers of carefully-selected mixtures of phages for therapy and prophylaxis were shipped throughout the former Soviet Union (FSU) each day. They generally were available both over the counter and through physicians. The largest use was in hospitals, to treat both primary and nosocomial infections, alone or in conjunction with chemical antibiotics. They played a particularly important role when antibiotic-resistant organisms were found. The military is still one of the strongest supporters of phage therapy research and development, because phages have proven so useful for wound and burn infections as well as for preventing debilitating gastrointestinal epidemics among the troops.
The historical background of the institute is interesting, and reflects a relatively unknown period in d’Herelle’s caeer. The following material comes from a number of people at the Institute, from a recent article by Shrayer (1996) on d’Herelle in Russia, from Summers (1998) and from d’Herelle’s own work.
In 1917, George Eliava, of the Georgian Institute of Microbiology, noticed that the water of the Koura (Mtkvary) river in Tbilisi (Tiflis) had a bactericidal action – an observation that could be explained by d’Herelle’s bacteriophage discovery. Eliava spent several extended periods in Paris at the Pasteur Institute and was a very early and staunch collaborator of d’Herelle’s; several papers of his are cited in d’Herelle’s first book on phages (1922). The two developed the dream of founding an Institute of Bacteriophage Research in Tbilisi, to be a world center of phage therapy for infectious disease, including scientific and industrial facilities and supplied with its own experimental clinics. The dream quickly became a reality due to the support of Sergo Ordjonikidze, the People’s Commissar of Heavy Industry, despite KGB opposition to this “foreign project” and personal conflicts between Eliava and Beria, then the local KGB head. A large campus on the river Mtkvary was allotted for the project in 1926. For many years, d’Herelle sent supplies, equipment and library materials, most of which he paid for himself. In 1934 -1935 he and his wife spent two 6-month periods working in Tbilisi, during which time he visited Kamenski, the People’s Commisar of Health Care, in Moscow and turned down an invitation to move there. He also wrote a book on “The Bacteriophage and the Phenomenon of Recovery”, which was translated into Russian by Eliava and dedicated to Stalin. D’Herelle intended to eventually move to Georgia; in fact, the cottage built for his use still stands on the Institute grounds. However, in 1937 Eliava was arrested as a “People’s Enemy” by Beria, then head of the KGB in Georgia and soon to direct the Soviet KGB as Stalin’s much-feared henchman. Eliava was summarily executed without a trial, sharing the tragic fate of many Georgian and Russian progressive intellectuals of the time, and d’Herelle, disillusioned, never returned to Georgia or the USSR. However, their Institute survived, and is still functioning at its original site on the Mtkvary (which it now shares with the more modern Institutes of Molecular Biology and Biophysics and of Animal Physiology).
In 1938, the Bacteriophage Institute was merged with the Institute of Microbiology & Epidemiology, under direction of the People’s Commissary of Health of Georgia. In 1951, it was formally transferred to the All-Union Ministry of Health set of Institutes of Vaccines and Sera, taking on the leadership role in providing bacteriophages for therapy and bacterial typing throughout the former Soviet Union. Under orders from the Ministry of Health, hundreds of thousands of samples of pathogenic bacteria were sent to the Institute from throughout the Soviet Union, to isolate more effective phage strains and better characterize their usefulness. In 1988, the Scientific Industrial Union “Bacteriophage” was formed, centered in Tbilisi with Russian production facilities in Ufa, Khabarowsk and Nijnyi Novgorod. The industrial part was always run on a self-supporting basis. The institute’s government-supported scientific branch included the electron microscope facility, permanent strain collection, laboratories studying phages of the enterobacteria, staphylococci and pseudomonads and formulating new phage cocktails, and groups involved in immunology, vaccine production, work with Lactobacillus and other therapeutic approaches. It also carried out the very extensive studies needed for approval by the Ministry of Health in Moscow of each new phage strain, therapeutic cocktail and means of delivery.
This careful study of the host range, lytic spectrum and cross-resistance properties of the phages being used were a major factor in the reported successes of the phage therapy work carried out through the Institute. All of the phages used for therapy are lytic, avoiding the problems engendered by lysogeny. The problems of bacterial resistance were largely solved by the use of well-chosen mixtures of phages with different receptor specificities against each type of bacterium as well as of phages against the various bacteria likely to be causing the problem in multiple infections. The situation was further improved whenever the clinicians typed the pathogenic bacteria and monitored their phage sensitivity; where necessary, new cocktails were then prepared to which the given bacteria were sensitive. Not infrequently, using phage in conjunction with other antibiotics was shown to give better results than either the phage or the antibiotic alone.
The depth and extent of the work involved is very impressive. For example, 1n 1983-85 alone, the Institute’s Laboratory of Morphology and Biology of Bacteriophages carried out studies of growth, biochemical features and phage sensitivity of 2038 strains of Staphylococcus, 1128 of Streptococcus, 328 of Proteus, 373 of Ps. aeruginosa and 622 of Clostridium, received from clinics and hospitals in towns across the former Soviet Union. New broader-acting phage strains were isolated using these and other Institute cultures and included in a reformulation of their extensively-used Piophage preparation; it now inhibited 71% of their Staphylococcus strains instead of 58%,
76% of Pseudomonas instead of 55%,
51% of E. coli instead of 11%,
30% of Proteus instead of 3%,
60% of Streptococcus instead of 38%,
and 80% of Enterococcus instead of 3% (Zemphira Alavidze, personal communication.) In the years since, there have been continued improvements in the formulation based on further studies, and phages against Klebsiella and Acinetobacter have been isolated and developed into therapeutic preparations. One of the latest developments is their IntestiPhage preparation, which includes 23 different phages active against a range of enteric bacteria.
A good deal of work has gone into developing and providing the documentation to get approval from the Ministry of Health for specialized new delivery systems, such as a spray for use in respiratory-tract infections, in treating the incision area before surgery, and in sanitation of hospital problem areas such as operating rooms. An enteric-coated pill was also developed, using phage strains that could survive the drying process, and accounted for the bulk of the shipments to other parts of the former Soviet Union.
Much of the focus in the last 12 years has been on combating nosocomial infections, where multi-drug-resistant organisms have become a particularly lethal problem and where it is also easier to carry out proper long-term research. Clinical studies of the effectiveness of the phage treatment and appropriate protocols were carried out in collaboration with a number of hospitals, but little has been published in accessible form. Zemphira Alavidze and her colleagues who are currently doing most of the actual therapeutic development and clinical application have manuscripts in preparation which describe their work in institutions such as the Leningrad (St. Petersburg) Intensive Burn Therapy Center, the Academy of Military Medicine in Leningrad, the Kazan Trauma Center, the Kemerovo Maternity Hospital. Some of the most intensive studies were carried out in Tbilisi, at the Pediatric Hospital, the Burn Center, the Center for Sepsis and the Institute for Surgery. Special mixtures were developed for dealing with strains giving problems of nosocomial infections in various hospitals, and they were very effectively used in sanitizing operating rooms and equipment, water taps and other sources of spread of the infections (most of them involving predominantly Staphylococcus). (Table 1).
The Industrial Branch on the grounds of the Bacteriophage Institute had large vats for growing the selected phage, using appropriate nonpathogenic bacteria and broth they prepared themselves from high-quality beef. The resulting phage lysates were sterile filtered using ceramic filters which could themselves be sterilized in very hot ovens. The various different phages for each particular formulation were then combined and automatically packaged and sealed into 10-ml ampoules or otherwise prepared and packaged for administration. Approximate titers were determined by checking the dilution that would produce lysis after coinnoculation with specific numbers of bacteria of standard test strains, and each batch was also tested for any surviving bacterial contaminants. In those rare cases where injection was planned, the phages were concentrated and resuspended in physiological saline solution; testing in guinea pigs was added to the rest of the analytical regime, to make sure there were no residual bacterial surface fragments (endotoxins) that might cause problems if injected. (As mentioned above, phages have generally been reported to appear in the bloodstream and other body fluids rather shortly after being ingested or poured into a wound and to still be effective against systemic infections, so injection is usually not necessary.)
Injectable forms accounted for only about 5% of the phage production at its height at the Bacteriophage Institute. None are made there now due to such factors as the expense and complexities of keeping animals for the necessary toxicity controls in the difficult situation in Georgia since the dissolution of the Soviet Union. The extensive fighting in Abkhasia left 350,000 refugees in a country of 5 million people and cut off the major rail, road and power routes to Russia, leaving only one significant highway across the High Caucausus mountains. Power is still often available for only a few hours a day and heating is a serious problem in winter. The country has very little money available for science, but some research continues, despite virtually no funding for salaries or supplies. The conflict did provide an interesting opportunity for widespread phage use. Each soldier in the Georgian army carried a spray-on phage cocktail which they used to disinfect their wounds (Alavadze, Meipariani, Gvasalia, manuscript in preparation). The industrial plant was privatized a few years ago and put to other uses, so the phages currently used for therapy must be grown in large carboys, the appropriate mixtures made, and then transferred to vials and sealed by hand. (Photo) However, the checks for sterility and efficacy on the designated bacteria are still just as careful. Unfortunately, until the electron microscope is repaired and electricity made more predictable, the phage preparations can no longer be checked to be sure that the phage present are of the appropriate mixtures of morphotypes, or physical shape and size. The Institute scientists still continue to do the best they can under the circumstance, and many in Tbilisi feel they clearly owe their lives to the group’s efforts. Extensive therapeutic work still goes on in local surgical, burn, pediatric and infectious disease hospitals, and in local clinics for ambulatory patients, including one on the grounds of the Institute.
Recent Work in the West Related to Phage Therapy
Levin and Bull (1996) and Barrow and Soothill (1997) have provided good reviews of much of the work applying phage therapy in animals which has been carried out in Britain and the United States since interest in the possibilities of phage therapy began to resurface there in the early 80′s. The results in general are in very good agreement with the clinical work described above in terms of efficacy, safety and importance of appropriate attention to the biology of the host-phage interactions, reinforcing trust in the reported extensive eastern European results.
In Britain, H. W. Smith and M. B. Huggins (1982, 1983) carried out a series of studies on use of phages in systemic E. coli infections in mice and then in diarrheal disease in young calves. For example, they found that injecting 106 colony-forming units of a particular pathogenic strain intramuscularly killed 10/10 of the mice, but none died if they simultaneously injected 104 plaque-forming units of a phage selected against the K1 capsule antigen of that bacterial strain.This phage treatment was more effective than using such antibiotics as tetracycline, streptomycin, ampicillin or trimethoprim/sulfafurazole. Furthermore, the resistant bacteria that emerged had lost their capsule and were far less virulent. In calves, they found very high levels of protection even though they did not succeed in isolating phages specific for the K88 or K99 adhesive fimbriae, which play key roles in attachment within the small intestine. Still, the phage were able to reduce the number of bacteria bound there by many orders of magnitude and to virtually stop the fluid loss. The results were particularly effective if the phage were present before or at the time of bacterial presentation, and if multiple phages with different attachment specificities were used. Furthermore, the phage could be transferred from animal to animal, supporting the possibility of prophylactic use in a herd. If phage were given only after the development of diarrhea, the severity of the infection was still substantially reduced and none of the animals died (Smith et al, 1987). Levin and Bull (1996) carried out a detailed analysis of the population dynamics and tissue phage distribution of the 1982 Smith and Huggins study which can be helpful in assessing the parameters involved in successful phage therapy and its apparent superiority to antibiotics. They have gone on to do very interesting animal studies of their own (Levin and Bull, manuscript in preparation) and conclude that phage therapy is at least well worth further study.
Soothill (1994) carried out a series of very nice studies preparatory to using phages for infections of burn patients. Using guinea pigs, he showed that skin-graft rejection could be prevented by prior treatment with phage against Pseudomonas aeruginosa. He also saw excellent protection of mice against systemic infections with both Pseudomonas and Acinetobacter when appropriate phages were used (Soothill, 1992). In the latter case, as few as 100 phages protected against infection with 100 million bacteria — 5 times the LD50!
Merrill and coworkers (1996) have carried out a series of experiments designed to better understand the interactions of phages with the human immune system, and have started a company called “Exponential Biotherapies, Inc.” to explore the possibilities of phage therapy. Their published work has been with a lytic derivative of the lysogenic phage lambda. While this particular strain would be a poor choice for therapeutic use, as discussed above and below, they have gathered very interesting and important data about factors affecting interactions between phages and the immune system.
Most bacteria are not pathogenic; in fact, they play crucial roles in the ecological balance in various parts of our bodies, including the digestive system and all body surfaces. They actually help protect us from pathogens; this is one reason why the use of broad-spectrum antibiotics leaves us so vulnerable, and why more narrowly-targeted bactericidal agents would be highly advantageous. Furthermore, most of the serious pathogens are close relatives of non-pathogenic strains — so what are the differences that make particular strains so lethal? Studies clarifying the mechanisms of pathogenesis at the molecular level have progressed remarkably in recent years (cf. ** add refs). These have now been crowned by the determination of the complete DNA base sequence of (nonpathogenic) E. coli K12 and several other bacterial species and extensive cloning and sequencing of pathogenicity determinants. Generally, a number of genes are involved, and these are clustered in so-called “pathogenicity islands”, or “Pais”, which may be 50,000-200,000 base pairs long. They generally have some unique properties indicating that the bacterium itself probably acquired them as a sort of “infectious disease” at some time in the past, and then kept them because they helped the bacterium infect new ecological niches where there was less competition. Many of these Pais are carried on small extrachromosomal circles of DNA called plasmids, which also can be carriers of drug-resistance genes. Others reside in the chromosome; there, they often are found imbedded in defective lysogenic prophages which have lost some key genes in the process and cannot be induced to form phage particles. However, they sometimes can recombine with related infecting phages. Therefore, it makes sense to avoid using lysogenic phages or their lytic derivatives for phage therapy to avoid any chance of picking up and moving such pathogenicity islands.
For bacteria in the human gut, pathogenicity involves 2 main factors: (1) the production of toxin molecules, such as shiga toxin (from Shigella and some pathogenic E. coli) or cholera toxin. These toxins modify proteins in the target host cells and thereby cause the problems. (2) the acquisition of new cell-surface adhesins which allow the bacterium to bind to specific receptor sites in the small intestine, rather than just moving on through to the colon. They also all contain the components of so-called type-III secretion machinery, related to those involved in assembly of flagella (for motility) and of filamentous phages and instrumental in many plant pathogens. For all of the pathogenic enteric bacteria, the infection process triggers changes in the neighboring intestinal cells. These include degeneration of the microvilli, formation of individual “pedestals” cupping each bacterium and, in the case of Salmonella and Shigella, induction of cell-signaling molecules that trigger engulfment of the bacterium and its subsequent growth inside the cell.
Recently, E. coli O157 has been the subject of much concern, with contamination of such products as hamburgers and unpasteurized fruit juices leading to serious epidemics (cf. Grimm et al., 1995). Deaths have occurred, particularly in young children and the elderly, usually from hemorrhagic colitis (bloody diarrhea) or hemolytic-uremic syndrome, where the kidneys are affected. Antibiotic therapy has shown no benefit (cf. Greenwald and Brandt, 1997). We find that the version of O157 from the Seattle fast-food-chain epidemic, at least, is susceptible to several of our T4-related phages (Mark Mueller, Kutter et al., unpublished). It is interesting to consider their potential use in feedlots and meat-packing plants and in prophylaxis and therapy during outbreaks.
The T-Even Family of Phages
A substantial fraction of the phages in therapeutic mixes are relatives of bacteriophage T4, which has played such a key role in the development of molecular biology (cf. Karam, 1994). As discussed above, the name “T-even family of phages” is a historical accident reflecting the fact that T2, T4 and T6 out of the original collection of Delbrück’s “Phage Group” all turned out to be related. Large sets of T4-like phages have been isolated for study from all over the world — for example, from Long Island sewage treatment plants, animals in the Denver zoo, and dysentery patients in Eastern Europe (the latter often using Shigella as host). Members of the family are found infecting most enteric bacteria and their relatives (Ackermann and Krisch, Archives of Virology, in press). Most of the T-even phages studied to date use 5-hydroxymethylcytosine instead of cytosine in their DNA, which protects them against most of the restriction enzymes bacteria make to protect themselves against invading DNA and gives them a much more effective host range. The entire DNA base sequence of phage T4 is known (Kutter, Stidham et al., 1994) and we know a great deal about its infection process in standard laboratory conditions and about the methods it uses to so effectively target bacteria. We can potentially use some of that knowledge in developing more targeted approaches to phage therapy, particularly as more is learned about the similarities and differences in its extended family (cf. Monod et al., 1997; Kutter et al., 1996.) We know that different members of the T-even family use different outer membrane proteins and oligosaccharides as their receptors, and understand the tail-fiber structures involved well enough to potentially predict which phages will work on given bacteria and engineer phages with new specificities (cf. Henning and Hashemolhosseini, 1994; Krisch, personal communication.)
There have still been far too few studies of T4 ecology and its behavior under conditions more closely approaching the natural environment and the circumstances it will encounter in phage therapy, where the environment is often anaerobic and/or the bacteria experience frequent periods of starvation. The limited available information in that regard was summarized by Kutter, Kellenberger et al (1994). A variety of studies are shedding light on the ability of these highly virulent phages to coexist in balance with their hosts in nature. For example, they can reproduce in the absence of oxygen as long as their bacterial host had been growing anaerobically for several generations. We have found that they can also survive for a period of time in a sort of state of hibernation inside of starved cells and then allow their host to readapt enough when nutrients are again supplied to go on and produce a few phage. This is particularly interesting and important since bacteria undergo many drastic changes to survive periods of starvation which increase their resistance to a variety of environmental insults (cf. Kolter, 1992).
The T-even bacteriophages share a unique ability that contributes significantly to their widespread occurrence in nature and to their competitive advantage. They are able to control the timing of lysis in response to the relative availability of bacterial hosts in their environment. When E. coli cells are singly infected with T4, they lyse after 25-30 minutes at body temperature in rich media, releasing about 100-200 phage per cell. However, when additional T-even phages attack the cell more than 4 minutes after the initial infection, the cell does not lyse at the normal time. Instead, it continues to make phage for as long as 6 hours, with the exact time of eventual lysis affected by the multiplicity of superinfecting phage (cf. Doermann, 1948; Abedon, 1994). This delay is termed “lysis inhibition”.
Thus, for many reasons the T4-related family of phages make excellent candidates for therapeutic use in enteric and other gram-negative bacteria, and studies of their ecology and distribution are being carried out with these goals in mind both in Tbilisi and at The Evergreen State College. Developing this same sort of understanding of other phage families potentially useful in phage therapy is equally important, taking advantage of the many powerful tools now available. Work useful to this end is progressing in a number of labs around the world but is still in its infancy, particularly as one moves beyond the enteric bacteria.
It is clearly time to look more carefully at the potential of phage therapy, both through strongly supporting new research and examining carefully what is already available. As Barrow and Soothill conclude, “Phage therapy can be very effective in certain conditions and has some unique advantages over antibiotics. With the increasing incidence of antibiotic resistant bacteria and a deficit in the development of new classes of antibiotics to counteract them, there is a need to investigate the use of phage in a range of infections.” The stipulations of Ackermann (1987) are important here: “Blind treatment is clearly of no value; phages have to be tested just as antibiotics, and the indications have to be right, but this holds everywhere in medicine. However, phage therapy requires the creation of phage banks and a close collaboration between the clinician and the laboratory. Phages have at least one advantage….While the concentration of antibiotics decreases from the moment of application, phage numbers should increase. Another advantage is that phages are able to spread and thus prevent disease. Nonetheless, much research remains to be done … on the stability of therapeutic preparations; clearance of phages from blood and tissues; their multiplication in the human body; inactivation by antibodies, serum or pus; and the release of bacterial endotoxins by lysis… In addition, therapeutic phages should be characterized at least by electron microscopy.” While it seems premature to generally introduce injectible phage preparations in the West without further extensive research, their carefully-implemented use for a variety of agricultural purposes and in external applications could potentially help reduce the emergence of antibiotic-resistant strains. Furthermore, compassionate use of appropriate phages seems warranted in cases where bacteria resistant against all available antibiotics are causing life-threatening illness. They are especially useful in dealing with recalcitrant nosocomial infections, where large numbers of particularly vulnerable people are being exposed to the same strains of bacteria in a closed hospital setting. In this case, the environment as well as the patients can be effectively treated.
In 1925, Sinclair Lewis’s classic novel Arrowsmith, for which he won the Nobel prize in literature, played a significant role in raising popular interest in the possibilities of phage therapy and the potential scientific and ethical dilemmas involved (Summers, 1991). Today, the growing scientific, public and commercial interest in phage therapy is being reflected and fanned in a number of ways. For example, the BBC recently produced a Horizon documentary on phage therapy, The Virus that Cures, building on the ideas in Radetsky’s Discover article on Return of the Good Virus. Several companies are beginning to explore work with phage therapy. In addition, a nonprofit “PhageBiotics” foundation has been formed to help support communication, education and research in the field. Hopefully all of this attention will lead to increased support of badly-needed research in the field and to rapid progress in developing appropriate applications, providing at least one alternative to the growing problem of multi-drug-resistant bacteria.
Acknowledgments: Special thanks to Drs. Rezo Adamia, Zemphira Alavidze, Teimuraz and Nino Chanishvili, Taras Gabisonia, Liana Gachechiladze, Mzia Kutateladze, Amiran Meipariani and their colleagues at the Bacteriophage Institute, Tbilisi, for their hospitality and efforts to help me understand the extensive therapeutic work carried out there. Others who have been particularly helpful with information and communication include Dr. Marina Shubladze, pediatrician in Tbilisi for 10 years, now residing in Seattle; Nino Mzavia, Nino Trapaidze, Timur and Natasha Zurabishvili, who have worked in my laboratory on basic T4 biology; Hans-Wolfgang Ackermann (Laval University), Eduard Kellenberger (Basel), William Summers (Yale), Steve Abedon (Ohio State) and Bruce Levin (Emory); Mansour Samadpour, University of Washington; Kathy d’Acci, clinical lab director, St. Peter’s Hospital, Olympia); physicians Jess Spielholz, MD, and Robin Moore, ND; and, especially, the many colleagues and students involved in our laboratory at Evergreen, particularly Barbara Anderson, Pia Lippincott, Mark Mueller, Stacy Smith, Elizabeth and Chelsea Thomas, Burt Guttman and Jim Neitzel.