Virus Ecology and Evolution: from Virus Adaptation to Phage Therapy

Transcript of Part 3: Phage Therapy

00:00:14.24	Hi.
00:00:15.24	I'm Paul Turner from the Department of Ecology and Evolutionary Biology at Yale University,
00:00:20.00	and the Microbiology program at Yale School of Medicine.
00:00:22.14	Today, I'd like to talk about phage therapy.
00:00:25.18	It's an old idea for how to treat bacterial infections in humans and other organisms,
00:00:31.07	and it's an old idea that's gaining resurgence due to the rise of antibiotic resistance in
00:00:36.02	bacterial populations.
00:00:38.02	Now that antibiotics are failing widely, we're turning to this old idea as a possibly good
00:00:43.21	new one to combat the spread of bacterial pathogens and the rise of antibiotic resistance.
00:00:51.05	The key question is whether evolution thinking can improve virus applications such as phage
00:00:56.07	therapy.
00:00:57.07	So, what I'll present... our data that relate to this idea of, can you use phage therapy
00:01:03.10	effectively to treat bacterial infections?
00:01:06.02	But, as an evolutionary biologist, I'll talk about this in the context of understanding
00:01:11.02	evolutionary biology, and how we can more rationally choose candidates to use, which
00:01:17.11	it... within phage therapy so that we choose the right virus to employ in these types of
00:01:23.10	methods with an eye to evolution that is destined to happen in the future, and yet have the
00:01:28.18	method prevail through time.
00:01:32.05	As a reminder, the evolution of antibiotic resistance is a profound challenge to humans
00:01:38.04	and biomedicine, domesticated agriculture, both plants and animals... this is a rising
00:01:44.04	problem where we see that antibiotics... in the past, they were used very effectively
00:01:49.23	to treat and kill bacterial infections, but, lately, this is not happening at all well,
00:01:55.17	and that's because, largely, of an uncontrolled mass experiment that we've done around this
00:02:01.13	world.
00:02:02.13	When antibiotics were discovered in the '50s, we immediately started using them in a large
00:02:08.08	degree to treat infections in humans and also, prophylactically, to introduce them into domesticated
00:02:13.24	animal populations, to protect them from bacterial infections before they ever occurred.
00:02:18.23	Well, what happened is this imposed selection on bacteria to be resistant to the antibiotics,
00:02:25.13	and unfortunately, through time, what we've seen is that bacteria are able to not only
00:02:30.10	become resistant to some antibiotic that is used presently on them, they'll maintain the
00:02:35.07	resistance to that antibiotic when we roll out some new antibiotic that they subsequently
00:02:40.15	also become resistant to.
00:02:42.05	So, through time, you can think of bacteria creating this problem because they are collecting
00:02:47.07	antibiotic resistance genes at no fitness cost, and therefore we're running out of options
00:02:52.14	for antibiotics that we can use to treat them.
00:02:55.12	A particularly important kind of a bacterium that is showing the spread of antibiotic resistance
00:03:02.15	is Pseudomonas aeruginosa.
00:03:04.22	We see in this system that multidrug resistant bacterial genotypes are on the rise.
00:03:09.13	Now, this is a bacterium that all of us encounter in our homes and it's even in soil and waterways,
00:03:18.00	so it's a very prevalent bacterium on the planet, and people who have normal immunity
00:03:23.01	and strong health are able to simply fight off an infection from Pseudomonas aeruginosa
00:03:28.21	if it enters the body.
00:03:30.11	But there are very many people who cannot do this.
00:03:33.09	Cystic fibrosis patients, those with severe burns, and anyone who's immunocompromised
00:03:39.00	is at risk of seeing this very prevalent Pseudomonas aeruginosa bacteria, and have it enter the
00:03:44.17	body, and opportunistically cause an infection.
00:03:49.18	There's a main reason for this.
00:03:52.17	This is because Pseudomonas aeruginosa has a lot of what are called efflux pumps.
00:03:59.06	These are combinations or complexes of proteins that exist, that permeate the cell, spanning
00:04:06.23	the inner and the outer portion, and their function is to bring things actively in and
00:04:11.05	out of the cell, including antibiotics.
00:04:14.09	So, if antibiotics get into a Pseudomonas aeruginosa cell, it's actively pumped out
00:04:20.06	through an efflux pump.
00:04:22.04	And this works in a wide variety of antibiotics, but it's important to remember that efflux
00:04:27.22	pumps have a large number of functions for Pseudomonas aeruginosa.
00:04:31.16	In terms of pathogenesis, this protein complex is also important for host colonization, the
00:04:38.19	ability for bacteria to form a very tough layer called a biofilm that's very difficult
00:04:43.22	for antibiotics or anything else to get through, and they also function in allowing these bacteria
00:04:49.12	to evade or avoid host immune responses.
00:04:53.07	Now, these are chromosome encoded and they seem to be highly prevalent or, in other words,
00:04:59.03	genetically conserved in Pseudomonas aeruginosa genotypes.
00:05:02.21	And, a wide variety of antibiotics, as I said, get pumped out from efflux pumps, such as
00:05:08.24	these antibiotic classes -- macrolides, aminoglycosides, tetracyclines.
00:05:14.14	And here's a typical structure of an efflux pump, which is at least three proteins spanning
00:05:20.00	the outer and the inner portion of the cell.
00:05:23.07	Now, let's get back to phage therapy and how the use of phage therapy might be able to
00:05:28.21	help us overcome the problem of opportunistic pathogens such as Pseudomonas aeruginosa.
00:05:34.04	So, the idea of phage therapy is that you can use a phage, or a virus that is specific
00:05:39.18	to a bacterium, to kill the disease bacterium instead of some other type of a therapy.
00:05:45.13	So, this picture is showing phages or these virus particles in relation to a typical bacterial
00:05:50.20	cell in terms of their relative size.
00:05:54.17	Phage therapy is an old idea.
00:05:56.15	So, even before antibiotics were ever discovered by Alexander Fleming and used widely by the
00:06:02.04	human population, long ago, people discovered that phages... in the early 1900's, these
00:06:09.01	are some of the earliest viruses ever discovered and described... that these could be placed
00:06:14.06	in a human or an animal to in... to combat bacterial disease.
00:06:19.06	At that time, the Russians were the main ones who were developing the technology of using
00:06:23.20	phages to combat disease.
00:06:26.03	They were so enamored of the idea that they even outfitted their soldiers in World War
00:06:30.06	II with little vials of phage that they could use, in the field, to treat wounds and to
00:06:36.24	prevent those wounds from being colonized by bacteria, because, at that time, it also
00:06:42.16	was unlikely for soldiers to be able to even reach some place that they could get health
00:06:47.01	care and, in the case of an opportunistic bacterial infection that they encountered
00:06:51.09	in the field, this might invade the body and even kill them before they could reach any
00:06:55.02	medical help.
00:06:57.06	It is important to also know, even in the early discovery of phages, these experiments
00:07:02.05	placed phages in animals such as chickens and humans, which are both types of organisms
00:07:08.01	that can suffer from cholera bacteria infections, and, yes, we know that rehydration therapy
00:07:13.16	is very key to get people through a cholera infection, but phages are able to actually
00:07:18.11	get in the body and find these cholera bacteria, kill them, and destroy them, so that someone
00:07:24.20	can recover from cholera much faster, even in the absence of rehydration therapy.
00:07:30.17	So, now, I want to give you some of the details, or remind you of the details, of how a lytic
00:07:36.24	phage, one that is lethal to a bacterium, undergoes its normal growth cycle.
00:07:41.05	And I'll further explain how understanding how evolution of resistance to phages by bacteria
00:07:47.12	can help us, if we find the right phage, to overcome the problem of antibiotic resistance.
00:07:51.16	So, a phage encounters a cell -- a bacterial cell in this circumstance -- and, if it can
00:07:58.10	infect and get in...
00:07:59.10	I'm showing in this diagram how the phage enters the cell, hijacks the metabolism of
00:08:04.18	the cell, makes copies, the cell bursts open, and those phage particles can go on and infect
00:08:11.07	new cells.
00:08:12.07	So, in this diagram, you'll note that, at the exterior of the bacterial cell, I'm showing
00:08:17.23	these recognition proteins in black, and this is a protein binding that allows the phage
00:08:23.06	to infect the bacterium.
00:08:25.20	The normal circumstance, and one that's very understandable from the... in light of evolutionary
00:08:30.17	biology, is that this selection pressure of phage infection is going to favor mutants
00:08:37.16	of the bacteria that are able to escape.
00:08:40.02	In other words, they'll have phage resistance.
00:08:42.11	So, now, I'm showing this escape mutant as surrounded by blue proteins, and the phage
00:08:47.24	is unable to infect.
00:08:50.08	So, here's a key question: Suppose that if a phage interacts with bacteria and selects
00:08:58.00	for the bacteria to become resistant to the phage, suppose that, following along with
00:09:04.04	that, it made the bacteria somehow sensitive to antibiotics?
00:09:08.24	That would be a great thing for phage therapy, because, if the phage kills the bacteria,
00:09:15.02	that's what you intended to happen, but if evolution kicks in and the bacteria escape
00:09:20.04	the phage infection, if they are now suddenly vulnerable to antibiotics, especially antibiotics
00:09:26.03	that are normally useless in killing the bacteria, you have a double-edged sword.
00:09:31.07	You have not only the ability of the phage to infect and kill under normal circumstances,
00:09:35.20	but, even when the inevitable evolution happens where resistance to the phage occurs, you
00:09:42.00	still have sensitivity following along for these bacteria to become sensitive to antibiotics
00:09:48.12	and to die in that fashion instead.
00:09:50.22	So, what would this really indicate?
00:09:53.22	It would indicate a trade-off between phage resistance and drug sensitivity in bacteria.
00:09:59.03	Obviously, that would improve antimicrobial therapy options and, importantly, it would
00:10:04.09	extend the lifetime of current antibiotics in our arsenal.
00:10:08.00	So, these would be chemical antibiotics that are approved by, say, the US Food and Drug
00:10:12.18	Administration as safe and effective in people, but what we have as a biomedical problem is
00:10:18.06	that they are failing widely.
00:10:20.04	If we could intervene somehow and find a phage that prompts this kind of a genetic trade-off,
00:10:25.17	then this would lead us to use those currently useless antibiotics over again, and we wouldn't
00:10:30.23	have to continue searching for new antibiotics -- we can simply trot out the ones that are
00:10:36.06	failing currently.
00:10:37.19	So, we found just such a phage.
00:10:41.06	It's abbreviated as OMKO1 and it actually came from a lake in Connecticut.
00:10:47.06	We took the old school idea of go out into a natural environment and try and find a phage
00:10:52.23	that kills your target bacterium.
00:10:54.24	Well, we used this method with an eye to understanding evolutionary biology.
00:11:01.00	We found phages that are able to infect and kill multidrug-resistant Pseudomonas aeruginosa,
00:11:08.04	and when they fail to kill, because the bacteria have become resistant, we find that this sensitivity
00:11:14.00	to antibiotics that results from the trade-off is a great way to further kind of a rational
00:11:20.04	design behind phage therapy.
00:11:22.04	So, the diagram shows, when the phage is countering the efflux pump, the key thing is that we
00:11:28.18	found phages that bind to and enter cells because they are binding initially to the
00:11:34.08	outer-most exposed proteins in the efflux pump.
00:11:38.14	In this one example, that protein is abbreviated as OprM or outer membrane protein M, and OMKO1
00:11:47.13	is abbreviated as outer membrane knockout 1, meaning that when we challenged the virus
00:11:52.21	to infect knockout strains of the bacteria, the only one that had failed to infect is
00:11:59.01	when OprM is not made on the cell surface.
00:12:02.07	That indicates that it must be involved in the binding property of the phage.
00:12:06.05	So, now, this is wonderful, because we have this genetic trade-off.
00:12:11.10	The phage-sensitive bacteria efflux antibiotics but they're killed by the virus.
00:12:17.24	The phage-resistant mutants have an impaired drug efflux ability -- this makes them vulnerable
00:12:23.14	to antibiotics that are normally useless against this bacterium.
00:12:27.23	So, looking a little bit at, how do you measure this in the laboratory?
00:12:32.13	How would you even know what's going on?
00:12:34.16	This is a typical agar plate, where the surface of the plate is coated with a high density
00:12:39.13	of bacteria -- that's called a lawn -- and what you can do is you can take some commercially
00:12:44.12	available thing like a disc of paper that's been soaked in antibiotic, and you place that
00:12:50.19	on the agar surface, so if these bacteria are multidrug-resistant, including able to
00:12:56.03	grow in the presence of this drug that's leaching out from that paper, then they'll grow up
00:13:01.05	right next to the edge of the paper.
00:13:03.18	They completely shrug off the effect of the antibiotic -- it's... has no effect in controlling
00:13:08.24	their growth.
00:13:10.09	What we observed is, after we exposed those bacteria to this phage, they gain a mutation
00:13:16.12	that makes them resistant to the phage, and instead, now, they have a very different growth
00:13:21.06	ability in this assay.
00:13:22.23	Now, a much lower concentration of that drug is adversely affecting the growth of the bacteria,
00:13:31.10	and technically that's called a minimum inhibitory concentration.
00:13:35.07	That means that there's a lower level of this drug that is now impacting their growth in
00:13:40.14	this way.
00:13:41.14	So, now, I'll walk you through a table that shows and summarizes the key data from our
00:13:47.13	study.
00:13:48.16	Let's begin at the first line of this table.
00:13:50.15	In the column beginning here is an antibiotic, for example, and the class of antibiotic that
00:13:56.23	it's drawn from.
00:13:59.00	Next column is the minimum inhibitory concentration of these multidrug-resistant bacteria in the
00:14:05.02	presence of that drug.
00:14:06.23	And then, the column that follows is, how does that MIC change when the bacteria see
00:14:12.10	the phage and they become resistant to the phage, and how has it affected their MIC?
00:14:18.09	In these first two examples, you'll see that there was a huge impact -- the fold-increased
00:14:23.07	drug sensitivity is a very large number.
00:14:27.02	Through simply exposing the bacteria to the phage, they become resistant to the phage
00:14:32.20	and it makes them much more sensitive to the antibiotic.
00:14:35.18	Now, these first two examples are not by accident, because we know through prior scientific data
00:14:41.07	that tetracycline and erythromycin are examples of two popular antibiotics drawn from two
00:14:47.06	different antibiotic classes, for which efflux pumps provide the main mechanism of resistance
00:14:52.23	for Pseudomonas aeruginosa to remove those antibiotics from the cell.
00:14:56.23	So, we're not quite certain, in the terms of other antibiotics drawn from other antibiotic
00:15:01.20	classes, whether explicitly efflux pumps are always doing this, or only sometimes doing
00:15:07.04	it.
00:15:08.04	Well, when we looked at four more examples, drawn from different antibiotic classes, you
00:15:12.08	see mostly the same result, where there was an increased fold drug sensitivity.
00:15:18.10	Now, you may not be as impressed by these lower numbers, but I've shown you in asterisks
00:15:24.20	that the clinical resistance has reversed to sensitivity.
00:15:28.13	This means that a doctor, a physician, could actually administer these antibiotics successfully
00:15:34.14	against these phage-resistant bacteria, and that's because the bacteria are now sensitive
00:15:42.11	to these drugs at a clinically relevant level.
00:15:45.19	So, the last line on the table is more or less a control.
00:15:49.20	Here, we have ampicillin, which is known to be not efflux pumped out of the cell for resistance.
00:15:56.23	So, one would not expect these values to change when the bacteria see or do not see the phage.
00:16:03.09	And, indeed, there is no change in their fold-increased drug sensitivity.
00:16:07.16	So, these data, taken together, suggest that there are a wide variety of antibiotics drawn
00:16:14.04	from different antibiotic classes for which this phage exposure makes the bacteria sensitive
00:16:21.10	to antibiotics that are currently useless in controlling multidrug-resistant Pseudomonas
00:16:27.14	aeruginosa.
00:16:28.14	So, what we would like to do next is, at least... we know for now that the trade-off is... is
00:16:35.09	observed in a wide variety of Pseudomonas aeruginosa genotypes, and that gives us a
00:16:41.04	lot of... an indication, a strong indication, that this should be broadly useful in targeting
00:16:47.04	a wide variety of genotypes.
00:16:48.19	So, this worked not only in laboratory model strains PA01 and PA14, it worked in a wide
00:16:54.20	variety of clinical samples, including individuals who had suffered otitis, or ear infections,
00:17:02.02	diabetic foot ulcers, osteomyelitis, etc.
00:17:05.12	We took clinical samples from these individuals, put them into our assay, and we got data that
00:17:11.04	suggests that those clinically relevant bacteria are experiencing the same trade-off as in
00:17:16.24	the model strains.
00:17:17.24	And, even more interestingly, if you just simply take Pseudomonas aeruginosa from the
00:17:22.16	environment, and you can find multidrug-resistant forms of it in the environment, and estuaries,
00:17:27.16	and even in human homes, these also experience the exact same trade-off as the other strains.
00:17:35.01	So, ideally, then, we can now expand this out even further, look at a very, very large
00:17:41.04	repository of clinical isolates, and I would expect that these data would also be supportive.
00:17:47.06	But that's a goal for our near-future work.
00:17:50.13	Already, we've obtained the ability from the US FDA to use phage cocktails in chronically
00:17:58.09	infected human volunteers, where these could actually only be the single phage that we
00:18:04.00	found, or in combination with other phages.
00:18:06.15	And, by that, I mean these are human individuals who have suffered chronic infections with
00:18:12.03	MDR P. aeruginosa, and they have essentially run out of options for treatment.
00:18:17.05	There are no antibiotics that will help them.
00:18:20.07	In some cases, it's actually remarkable that these individuals are still alive, and they
00:18:24.07	are very willing to undergo experimental treatment if this is going to help them.
00:18:28.17	Well, we are identifying individuals at Yale New Haven Hospital and other locations, and
00:18:34.02	working with physicians and surgeons, in order to use this experimental treatment where these
00:18:39.24	volunteer patients are bravely looking to this option to see if it will improve their
00:18:44.16	health.
00:18:45.16	Already, we've had one successful event, where a man had an aortic arch replacement, which
00:18:51.11	is a pretty routine thing that is a surgery that leads to the replacement of a very key
00:18:56.21	portion of the heart, that kept this man alive, but, in any individual, and this is increasingly
00:19:02.13	seen in medical science... is that we have these routine surgeries that bring into the
00:19:06.24	human body artificial substrates, and these make great places for bacteria to form a biofilm,
00:19:13.12	a very tough infection that's hard to treat.
00:19:16.02	And, in many cases, these are multidrug-resistant bacteria.
00:19:19.15	Well, this man did not have... he did not have any other options and he volunteered
00:19:23.14	for treatment, and we placed the phage in his chest near the site of the infection,
00:19:29.06	and one administered treatment was enough to remove his bacterial biofilm, and he is
00:19:33.17	now completely off antibiotics and in much better health.
00:19:38.08	We talked about his one case as well as our general approaches to taking an evolutionary
00:19:44.11	eye to phage therapy in several radio shows, which you might want to listen to for further
00:19:49.05	information.
00:19:51.11	But our big goal is to test... test the safety and efficacy of this whole approach in a mouse
00:19:58.05	model.
00:19:59.05	In some ways, what we were able to achieve was a bit backwards.
00:20:02.12	We found some human patients that were very excited about this potential therapy, and
00:20:07.03	it did work safely, obviously, in the one case that we've tried, but now we'd like to
00:20:11.02	take a step back and see how broadly useful might this be for say lung pneumonia infections
00:20:17.04	in immunocompromised patients, or in cystic fibrosis patients, and it's very easy to create
00:20:22.00	a mouse model dataset in a laboratory to test the safety and efficacy of whether this works.
00:20:28.12	So, that's underway right now.
00:20:31.01	If that works, then we're at the stage where we can actually help humans on a grand scale,
00:20:38.15	ideally.
00:20:39.15	So, this would be a clinical trial that would happen in humans.
00:20:42.14	Perhaps we could use humans who are suffering hospital-acquired pneumonia, or cystic fibrosis-associated
00:20:48.18	pulmonary infections, and see if we can improve their quality of life by administering the
00:20:54.05	phage and the antibiotics simultaneously, and that might reduce, even just... for individuals
00:21:00.00	who are on chronic antibiotic therapy, which has a lot of bad consequences and side-effects,
00:21:05.22	even using the phage to help prevent the bacteria from infecting or removing them entirely,
00:21:12.09	it should expose those individuals to less of these chemicals that can adversely affect
00:21:17.07	their health, but are helping keep them alive.
00:21:19.07	I'd like to acknowledge the people who helped on the study.
00:21:23.00	Ben Chan was the first author on the paper that we produced in 2016, that looked at not
00:21:28.04	only the initial discovery of this phage but also a subsequent paper where we described
00:21:33.08	the case study of the individual who we helped with experimental therapy.
00:21:38.10	And two other key individuals who worked with me on this were Deepak Narayan, the surgeon
00:21:43.09	at Yale School of Medicine and Yale New Haven Hospital, who was able to find this patient
00:21:49.22	who bravely volunteered for the treatment and is also helping us administer the treatment
00:21:54.06	to other individuals who have run out of options.
00:21:56.23	Also, my long-term collaborator, John Wertz, who's at the Coli Genetic Stock Center at
00:22:02.06	Yale, was another key architect in the study.
00:22:05.11	And, without them and the help of my lab group, none of this work could have happened.
00:22:10.07	We're also thankful to those who funded our work externally, especially federal agencies
00:22:16.03	and private foundations.