Treatment of Tuberculosis, A Persistent Threat to Global Health

Part 1: Tuberculosis a Persistent Threat

00:00:05.13 00:00:06.29 Hello. My name is John McKinney.
00:00:07.00 00:00:10.09 I'm a professor with the Global Health Institute School of Life Sciences
00:00:10.10 00:00:13.06 at the Swiss Federal Institute of Technology
00:00:13.07 00:00:14.21 at EPFL in Lausanne.
00:00:14.22 00:00:18.05 Today I'm going to talk to you about a once and future plague:
00:00:18.06 00:00:21.10 tuberculosis -- a disease that has been with humanity
00:00:21.11 00:00:23.27 for about as long as humanity has existed,
00:00:23.28 00:00:25.07 but as you'll see, continues to be
00:00:25.08 00:00:28.27 one of the leading causes of morbidity and mortality even today.
00:00:28.28 00:00:32.14 So, I'd like to begin by reminding you
00:00:32.15 00:00:36.16 of how human history, in recent years, has differed
00:00:36.17 00:00:38.14 from human history at any time in the past.
00:00:38.15 00:00:40.21 That's shown here, where I have graphed
00:00:40.22 00:00:44.26 the total global human population versus calendar year
00:00:44.27 00:00:49.00 going back to about 10,000 BC -- about the time that agriculture
00:00:49.01 00:00:52.04 was invented and humanity began to congregate in cities.
00:00:52.05 00:00:55.00 As you can see, for most of human history and prehistory,
00:00:55.01 00:00:57.11 there was little change in the global human population.
00:00:57.12 00:01:00.16 That started to change quite dramatically a few hundred years ago,
00:01:00.17 00:01:04.13 when the human population started to undergo rapid growth
00:01:04.14 00:01:05.25 which continues to this day.
00:01:05.26 00:01:08.13 So, as you can see, there's rapid exponential growth
00:01:08.14 00:01:09.27 right up to the present time.
00:01:09.28 00:01:13.03 So, that is quite unprecedented in human history.
00:01:13.04 00:01:16.16 Even more strikingly, in the last 200 years or so,
00:01:16.17 00:01:19.04 the rate of urbanization of humanity
00:01:19.05 00:01:21.25 has outpaced even this incredible rate of population growth.
00:01:21.26 00:01:25.02 So, looking back about 200 years ago to the year 1800,
00:01:25.03 00:01:28.15 less than half of the human population lived in cities.
00:01:28.16 00:01:32.03 Now, more than half of humanity is congregated in cities.
00:01:32.04 00:01:35.25 Why do these factors matter? Population growth and urbanization.
00:01:35.26 00:01:40.16 Well, from a pathogen's point of view, our bodies are merely substrate for them.
00:01:40.17 00:01:44.16 So, having a lot of substrate congregated together in a small space
00:01:44.17 00:01:47.23 to maximize transmission potential
00:01:47.24 00:01:49.26 is optimal, obviously.
00:01:49.27 00:01:54.03 So, in a sense, things have never been so good as they are today
00:01:54.04 00:01:56.12 for pathogens that infect human beings.
00:01:56.13 00:02:01.19 Now, pathogens have learned how to exploit many different routes of entry into our bodies.
00:02:01.20 00:02:05.23 Some pathogens -- the malaria parasite being a notable example --
00:02:05.24 00:02:09.10 have learned how to exploit insect vectors to become directly injected
00:02:09.11 00:02:11.11 into the blood stream of our bodies --
00:02:11.12 00:02:13.07 a kind of optimal route of entry, if you like.
00:02:13.08 00:02:16.21 Others enter the blood stream through scratches, abrasions, and so on.
00:02:16.22 00:02:20.08 But, the fact is that most pathogens, including most bacterial pathogens,
00:02:20.09 00:02:24.05 enter through mucosal surfaces, including the respiratory tract,
00:02:24.06 00:02:28.21 the digestive tract, and the urogenital tract.
00:02:28.22 00:02:32.16 These pathogens have learned to exploit the food that we eat,
00:02:32.17 00:02:34.12 the water that we drink, and so on,
00:02:34.13 00:02:37.08 as their means of getting from one host to the next.
00:02:37.09 00:02:41.10 The tubercle bacillus that causes tuberculosis is a notable example
00:02:41.11 00:02:44.24 of a pathogen that has learned how to be transmitted through the air that we breathe.
00:02:44.25 00:02:47.04 So, I'm sure you all remember, when you were a kid,
00:02:47.05 00:02:50.17 your mom used to tell you to cover your mouth when you coughed or sneezed.
00:02:50.18 00:02:52.16 As usual, mom was right.
00:02:52.17 00:02:56.14 The reason that she told you to cover your mouth or nose when you sneezed,
00:02:56.15 00:03:00.19 is because when you cough or sneeze, you produce huge numbers of airborne particles
00:03:00.20 00:03:03.17 like this. And if this individual, who is coughing,
00:03:03.18 00:03:08.11 had tuberculosis, many of these airborne particles would contain viable tubercle bacilli,
00:03:08.12 00:03:11.08 transiting from this individual to the next host.
00:03:11.09 00:03:15.17 Now, when these bacteria are respired into the airways,
00:03:15.18 00:03:18.09 most of these bacteria will impinge on the upper airways,
00:03:18.10 00:03:20.29 where they will become trapped in the mucous lining the airways
00:03:21.00 00:03:24.13 that is secreted by these goblet cells -- the bare patches here.
00:03:24.14 00:03:27.17 Those trapped bacteria will then be swept up and out
00:03:27.18 00:03:34.18 by the mucociliary elevator -- the ciliary action of these ciliated epithelial cells
00:03:34.19 00:03:38.16 where they will end up in the mouth and then be swallowed and destroyed in the stomach.
00:03:38.17 00:03:41.10 So, in fact, in order to initiate infection,
00:03:41.11 00:03:44.11 the tubercle bacillus has to penetrate all the way down
00:03:44.12 00:03:48.10 into the terminal ramifications of the respiratory tree --
00:03:48.11 00:03:50.29 all the way down into the lung alveoli,
00:03:51.00 00:03:55.06 where they can implant, be phagocytosed by resident alveolar macrophages
00:03:55.07 00:03:57.25 (the cell type that they parasitize within the lung),
00:03:57.26 00:03:59.08 and initiate replication.
00:03:59.09 00:04:03.12 So, that's the sequence of events that leads to just about every new case
00:04:03.13 00:04:05.24 of tuberculosis infection.
00:04:05.25 00:04:07.17 What happens following these initial events, though,
00:04:07.18 00:04:10.17 is enormously variable from individual to individual,
00:04:10.18 00:04:14.19 which to my mind is one the chief mysteries and challenges
00:04:14.20 00:04:17.00 of tuberculosis, and is beautifully encapsulated, I think,
00:04:17.01 00:04:21.05 in this quote from a classic paper by the epidemiologist George Comstock
00:04:21.06 00:04:22.19 at Johns Hopkins,
00:04:22.20 00:04:24.12 in which he stated that, "following infection,
00:04:24.13 00:04:27.28 "the incubation period of TB" -- that is the interval between
00:04:27.29 00:04:32.11 exposure to the pathogen and the development of overt signs and symptoms of disease
00:04:32.12 00:04:35.26 can range "from a few weeks to a lifetime."
00:04:35.27 00:04:38.22 In other words, TB is a classic example of a persistent,
00:04:38.23 00:04:40.23 even a life-long infection,
00:04:40.24 00:04:45.18 in which disease may not develop for months, years, or even decades
00:04:45.19 00:04:47.06 after initial exposure.
00:04:47.07 00:04:49.14 It's very common in endemic countries
00:04:49.15 00:04:52.04 for individuals to be infected in childhood
00:04:52.05 00:04:55.27 and not to develop disease until immunity wanes with old age.
00:04:55.28 00:05:01.18 So, let's go through in some detail the steps that lead to the establishment
00:05:01.19 00:05:05.07 of a latent TB infection and reactivation from latency.
00:05:05.08 00:05:07.25 As I've indicated already, the principal route of entry
00:05:07.26 00:05:11.06 of the tubercle bacillus into the human body is through the airways.
00:05:11.07 00:05:16.20 And the bacteria, in order to infect, must travel all the way down into an alveolus,
00:05:16.21 00:05:21.16 where they can implant and initiate infection, replicating inside macrophages.
00:05:21.17 00:05:24.09 That's shown here in this micrograph from a human lung,
00:05:24.10 00:05:28.00 So, if you look carefully, you'll see that there are clusters here
00:05:28.01 00:05:31.25 of sort of bright pink staining rods.
00:05:31.26 00:05:34.20 These are tubercle bacilli that have been stained with a special stain
00:05:34.21 00:05:36.21 that stains only this type of bacterium.
00:05:36.22 00:05:39.11 As you'll see, they're clustered within groups, and that's because
00:05:39.12 00:05:41.26 they're contained within macrophages of the host.
00:05:41.27 00:05:44.20 Now, the macrophage, of course, is one of the front-line defenses
00:05:44.21 00:05:47.03 of the innate immune response in humans.
00:05:47.04 00:05:51.00 And the macrophage is perfectly capable of destroying most bacteria
00:05:51.01 00:05:52.20 that happen to enter the lung.
00:05:52.21 00:05:55.08 But, the tubercle bacillus, being a professional pathogen,
00:05:55.09 00:05:59.01 has learned actually how to exploit the macrophage as its niche
00:05:59.02 00:06:02.05 for replication and persistence in human hosts.
00:06:02.06 00:06:05.25 One of the strategies that it uses to do this
00:06:05.26 00:06:10.00 is to block the normal pathway of progression from the phagosome
00:06:10.01 00:06:12.09 that initially contains the bacterium
00:06:12.10 00:06:15.19 to the lysosome, which is the digestive compartment of the cell.
00:06:15.20 00:06:18.21 Now most bacteria, when they're phagocytosed by a macrophage
00:06:18.22 00:06:23.22 will end up within the confines of a closely apposed vacuolar membrane
00:06:23.23 00:06:27.15 (“the phagosome”) that undergoes a series of maturation steps,
00:06:27.16 00:06:31.18 leading to the sequential acquisition of different molecular markers,
00:06:31.19 00:06:36.11 gradual acidification, and finally, fusion with the lysosome.
00:06:36.12 00:06:38.16 This, of course, is the digestive compartment of the cell --
00:06:38.17 00:06:43.08 chock full of hydrolytic enzymes that break down most bacteria that enter the macrophage.
00:06:43.09 00:06:48.13 TB manages to avoid this fate, using mechanisms that have yet to be discovered.
00:06:48.14 00:06:53.00 It somehow blocks the progression of that nascent phagosome
00:06:53.01 00:06:55.21 along the phagosome-lysosome fusion pathway,
00:06:55.22 00:06:58.09 which means that, inside the parasitized macrophage,
00:06:58.10 00:07:01.25 the tubercle bacillus is residing within a phagosome
00:07:01.26 00:07:04.16 that does not acquire markers of maturation,
00:07:04.17 00:07:10.05 that fails to undergo acidification beyond about pH 6.5 (close to neutral),
00:07:10.06 00:07:14.14 and which fails, most importantly, to fuse with lysosomes --
00:07:14.15 00:07:15.29 this degradative compartment of the cell.
00:07:16.00 00:07:20.06 So, the bacterium actually lives within the phagosome, inside of the macrophage
00:07:20.07 00:07:23.06 and replicates within this niche.
00:07:23.07 00:07:28.03 So that's how infection initiates.
00:07:28.04 00:07:29.26 The bacteria replicate within a macrophage
00:07:29.27 00:07:32.09 until the macrophage is overwhelmed and lyses,
00:07:32.10 00:07:36.13 releasing those bacteria, which are then phagocytosed by other macrophages
00:07:36.14 00:07:40.08 that have immigrated to the site, drawn there through chemotactic signals
00:07:40.09 00:07:42.19 that are released by the infected macrophage.
00:07:42.20 00:07:45.24 Now, again, normally this is a defensive mechanism of the host
00:07:45.25 00:07:49.09 to bring macrophages to the site of battle where they can destroy the invader,
00:07:49.10 00:07:51.18 but in the case of the tubercle bacillus,
00:07:51.19 00:07:55.22 this is merely providing it with an additional niche within which it can replicate.
00:07:55.23 00:07:59.11 So, this process of growth within macrophage,
00:07:59.12 00:08:03.03 lysis of the macrophages, release of the bacteria,
00:08:03.04 00:08:08.03 phagocytosis by immigrating macrophages, and again, intracellular replication...
00:08:08.04 00:08:10.24 This cycle continues for a period of about a week or two.
00:08:10.25 00:08:15.03 And at the point, the bacteria escape from this initial primary lesion
00:08:15.04 00:08:19.14 that's been formed, initially into the lymph nodes that drain the region
00:08:19.15 00:08:21.13 of the lung that's affected by infection.
00:08:21.14 00:08:25.20 And then, they reach the bloodstream, probably carried there by the thoracic duct,
00:08:25.21 00:08:27.02 although we really don't know the details.
00:08:27.03 00:08:31.29 This process of dissemination through the lympho-hematogeneous route
00:08:32.00 00:08:35.16 is critical for establishing infection at secondary sites
00:08:35.17 00:08:39.10 within the lungs, as well as within extrapulmonary organs.
00:08:39.11 00:08:43.06 So, the lung is, of course, the primary target of tuberculosis.
00:08:43.07 00:08:44.26 That's its favorite niche.
00:08:44.27 00:08:49.08 And that is the location from which the bacterium disseminates to secondary hosts.
00:08:49.09 00:08:52.29 But, in fact, in about 10% of all cases of TB, there is involvement
00:08:53.00 00:08:55.25 of one or more extrapulmonary organs, as well.
00:08:55.26 00:08:58.28 And, TB, being a protean bacteria,
00:08:58.29 00:09:02.06 can adapt to almost any environment, and in fact can cause disease
00:09:02.07 00:09:04.18 of the spleen, the liver, the bone, the eye... you name it.
00:09:04.19 00:09:08.13 Any organ system can be susceptible to tuberculosis.
00:09:08.14 00:09:12.08 But, the lung is for sure the most important organ affected.
00:09:12.09 00:09:14.23 And, by seeding through the hematogeneous route,
00:09:14.24 00:09:19.05 all parts of the lung become infected with the tubercle bacillus,
00:09:19.06 00:09:22.19 which goes again through this process of infection of macrophages
00:09:22.20 00:09:25.08 and formation of what are called, now,
00:09:25.09 00:09:27.24 secondary lesions that are seeded by the bloodstream
00:09:27.25 00:09:29.04 rather than the airway.
00:09:29.05 00:09:31.08 Now, at about this point, the host normally
00:09:31.09 00:09:36.04 wakes up to the fact that it has been infected and mounts an adaptive immune response
00:09:36.05 00:09:37.26 against the organism.
00:09:37.27 00:09:41.06 And this is signaled by the arrival, in these lesions,
00:09:41.07 00:09:45.26 of T lymphocytes that recognize specifically antigens of the tubercle bacillus.
00:09:45.27 00:09:49.07 And, in about 95% of individuals who are infected,
00:09:49.08 00:09:52.22 that immune response succeeds in arresting the further progression
00:09:52.23 00:09:53.27 of the disease at this stage.
00:09:53.28 00:09:57.21 In about 5% of infected individuals, and in particular those
00:09:57.22 00:09:59.18 who are immune compromised for any reason,
00:09:59.19 00:10:03.11 disease will continue to progress soon after infection.
00:10:03.12 00:10:05.02 We call this primary tuberculosis.
00:10:05.03 00:10:07.24 But, the fact is, the overwhelming majority of individuals
00:10:07.25 00:10:10.24 control infection after this dissemination stage,
00:10:10.25 00:10:14.10 at a point where they are probably not even aware they are infected.
00:10:14.11 00:10:17.25 They have no obvious signs or symptoms of infection.
00:10:17.26 00:10:21.18 Now, latent infection can last a lifetime
00:10:21.19 00:10:24.20 and, in fact, in old studies from the 20th century,
00:10:24.21 00:10:28.20 it was shown that latently infected individuals who have died of other causes
00:10:28.21 00:10:33.18 will yield viable tubercle bacilli upon culture of lung homogenate.
00:10:33.19 00:10:36.13 So, we know that tubercle bacillus, once it has infected you,
00:10:36.14 00:10:39.05 is capable of persisting for a lifetime.
00:10:39.06 00:10:44.16 Most of those infections will remain latent throughout the remainder of the individual's life,
00:10:44.17 00:10:48.21 but in a minority of cases -- about 5-10% over the course of a lifetime --
00:10:48.22 00:10:52.15 individuals with latent TB infection will reactivate.
00:10:52.16 00:10:54.27 For reasons we really don't understand,
00:10:54.28 00:10:57.24 in a minority of latently infected individuals,
00:10:57.25 00:11:00.14 reactivation will occur at some time later in life.
00:11:00.15 00:11:04.14 So, that's about a 5-10% cumulative lifetime risk of reactivation
00:11:04.15 00:11:06.07 in people who are latently infected.
00:11:06.08 00:11:10.17 When reactivation occurs, it usually occurs in the upper, as opposed to lower,
00:11:10.18 00:11:13.21 lung fields. We really don't understand the reason for this,
00:11:13.22 00:11:16.28 but it illustrates the importance of this lymphohematogeneous
00:11:16.29 00:11:21.00 dissemination of infection that occurs as an early event of tuberculosis.
00:11:21.01 00:11:23.27 Infection normally occurs in the lower airways.
00:11:23.28 00:11:26.23 That's because this is, by far, the best ventilated part of the lung
00:11:26.24 00:11:29.05 when we are standing or sitting upright.
00:11:29.06 00:11:33.20 But, reactivation and progressive TB usually occur in the upper airways,
00:11:33.21 00:11:36.16 and those lesions are seeded by the bloodstream.
00:11:36.17 00:11:41.04 Now, let's look for a moment at the architecture of a tuberculous lesion
00:11:41.05 00:11:48.14 because it's a very common kind of tissue architecture for a chronic inflammatory process.
00:11:48.15 00:11:51.07 In an early granuloma, such as the one shown here,
00:11:51.08 00:11:55.05 the core of the lesion may have begun to undergo a process called
00:11:55.06 00:11:59.07 caseating necrosis. This is simply the death and breakdown of infected macrophages,
00:11:59.08 00:12:00.14 as I described before.
00:12:00.15 00:12:06.02 This core of caseation necrosis is surrounded by a mantle of intact macrophages,
00:12:06.03 00:12:10.24 which, if they were properly stained, would reveal large numbers of tubercle bacilli,
00:12:10.25 00:12:13.09 which are, of course, growing inside those macrophages.
00:12:13.10 00:12:17.06 Further out, there is another layer now of lymphocytes
00:12:17.07 00:12:20.12 that have arrived on the scene to try to control infection.
00:12:20.13 00:12:23.27 So, you can see there's a kind of striation of the lesion --
00:12:23.28 00:12:27.00 a core of caseation necrosis, a macrophage zone
00:12:27.01 00:12:29.21 where the bacteria are replicating, and a lymphocyte zone
00:12:29.22 00:12:33.00 where the adaptive immune response is taking place.
00:12:33.01 00:12:36.04 Now, somewhat later, as infection progresses,
00:12:36.05 00:12:39.20 that process of caseation necrosis can become quite extreme,
00:12:39.21 00:12:42.01 so that a large core within the lesion develops
00:12:42.02 00:12:44.26 where there is essentially no intact cells whatsoever.
00:12:44.27 00:12:48.08 There's still a mantle of heavily parasitized macrophages
00:12:48.09 00:12:51.23 surrounding that core, but now further out, in place of lymphocytes,
00:12:51.24 00:12:56.09 we largely see fibroblasts laying down a matrix of collagen and fibrin
00:12:56.10 00:12:59.19 and other extracellular matrix fibers.
00:12:59.20 00:13:03.16 It's as though the host is trying to contain the infection.
00:13:03.17 00:13:07.18 But, at the same time, this containment of infection within the granuloma
00:13:07.19 00:13:11.23 may prevent, we think, the development of an optimal immune response
00:13:11.24 00:13:13.24 against the pathogen.
00:13:13.25 00:13:16.04 So, it's thought that tuberculosis granuloma
00:13:16.05 00:13:20.18 may play a dual role in infection - both for protection and for pathogenesis.
00:13:20.19 00:13:23.21 Now, going back to our model,
00:13:23.22 00:13:27.27 if that core of caseation necrosis undergoes complete liquefaction,
00:13:27.28 00:13:30.28 this can lead to the excavation of an actual cavity
00:13:30.29 00:13:33.25 within the lung, as diagrammed here.
00:13:33.26 00:13:36.03 This is a large hole - literally a hole in the lung
00:13:36.04 00:13:40.20 where the tissue has completely broken down and has exited via the airway here.
00:13:40.21 00:13:43.19 This is very easy to see on a chest x-ray.
00:13:43.20 00:13:47.02 For example, here. This is a tuberculosis cavity
00:13:47.03 00:13:50.16 in the lung of a 19-year old, otherwise healthy male
00:13:50.17 00:13:53.22 who unfortunately was infected with multi-drug resistant tuberculosis,
00:13:53.23 00:13:58.04 requiring removal of this entire lung as a last ditch effort to save this patient.
00:13:58.05 00:14:00.24 So, this kind of extreme pathology
00:14:00.25 00:14:03.02 can indeed be life threatening.
00:14:03.03 00:14:09.22 And, within cavitary lesions, where the tissue has liquified as I described,
00:14:09.23 00:14:12.07 the tubercle bacillus, it's thought can, for the first time,
00:14:12.08 00:14:14.13 undergo extracellular growth.
00:14:14.14 00:14:17.21 And, as shown here, where the bacteria are stained pink with Kenyon stain,
00:14:17.22 00:14:20.14 they can reach absolutely enormous numbers
00:14:20.15 00:14:23.25 within the liquified tissue inside the cavity.
00:14:23.26 00:14:27.29 This is a prime breeding ground, obviously, for the development of drug resistance --
00:14:28.00 00:14:31.03 a problem that I'll come back to a bit later in the talk.
00:14:31.04 00:14:36.24 Now, once a cavity like this actually erodes into an airway,
00:14:36.25 00:14:39.07 the bacteria can enter the airway, and then when the patient
00:14:39.08 00:14:44.19 coughs, sneezes, talks, sings... does anything that expels air from the airways,
00:14:44.20 00:14:49.19 the bacteria can then exit the airway outward and enter the environment.
00:14:49.20 00:14:51.26 And, of course, then this is when transmission occurs.
00:14:51.27 00:14:56.18 So, that's the complete life cycle, if you like, of the tubercle bacillus
00:14:56.19 00:14:56.29 within the human host.
00:14:57.00 00:15:01.03 And, again, I want to emphasize that there can be a period of many decades
00:15:01.04 00:15:05.29 separating the initial infection step and the transmission of infection,
00:15:06.00 00:15:09.24 which makes it very, very difficult for this pathogen to be dislodged from communities
00:15:09.25 00:15:11.01 once it has taken hold.
00:15:11.02 00:15:14.18 We can look at this in another way.
00:15:14.19 00:15:17.24 So far, we've been talking about persistence of the tubercle bacillus
00:15:17.25 00:15:19.15 at the level of an infected patient.
00:15:19.16 00:15:21.23 But, this persistence within individuals
00:15:21.24 00:15:25.08 translates into persistence within human populations.
00:15:25.09 00:15:30.05 And this is nicely illustrated using the so-called SEIR model of epidemic dynamics,
00:15:30.06 00:15:34.15 in which susceptible human populations are broken down into four compartments,
00:15:34.16 00:15:37.02 if you like. A compartment of susceptibles
00:15:37.03 00:15:40.12 (that is, individuals who have not yet been infected, but who could be infected),
00:15:40.13 00:15:45.10 a population of exposed individuals (these are individuals who have acquired the infection
00:15:45.11 00:15:48.29 but are not yet transmitting it), an infectious compartment
00:15:49.00 00:15:52.01 (describing those individuals who both have the pathogen
00:15:52.02 00:15:55.15 and are actively transmitting it), and a recovered population
00:15:55.16 00:16:00.06 (indicating individuals who have gotten over the illness and are no longer infectious).
00:16:00.07 00:16:04.18 Now, if we look at a typical acute and transient infection
00:16:04.19 00:16:08.02 like the measles virus, which has been very well studied epidemiologically,
00:16:08.03 00:16:11.28 what we can see is that the transit through these four compartments
00:16:11.29 00:16:15.01 is very rapid and unidirectional.
00:16:15.02 00:16:18.18 So, an individual who is susceptible and becomes exposed to measles
00:16:18.19 00:16:21.18 will incubate the infection for a period of about two weeks
00:16:21.19 00:16:24.14 before he or she becomes infectious to others around them.
00:16:24.15 00:16:28.00 That period of infectiousness lasts for only about a week or two
00:16:28.01 00:16:32.19 before the individual either dies of the infection or hopefully recovers from infection.
00:16:32.20 00:16:35.12 Transit through these compartments is unidirectional
00:16:35.13 00:16:38.05 because the acquired immunity that develops during infection
00:16:38.06 00:16:41.16 is solidly protective against re-infection.
00:16:41.17 00:16:43.22 So, if you've had measles once, you're not going to get measles again.
00:16:43.23 00:16:48.23 In contrast, a persistent, chronic infection, like TB
00:16:48.24 00:16:50.05 shows very different dynamics.
00:16:50.06 00:16:53.04 First of all, when a susceptible individual becomes exposed to TB,
00:16:53.05 00:16:58.03 the period of time that elapses between exposure and the development of infectiousness
00:16:58.04 00:17:02.24 can last anything from months to, as I've said already, decades.
00:17:02.25 00:17:05.03 In other words, a latently infected individual can continue
00:17:05.04 00:17:07.13 to harbor the pathogen for a lifetime.
00:17:07.14 00:17:10.18 Furthermore, once an individual becomes infectious to others,
00:17:10.19 00:17:14.25 that period can last again, for a period anything from months to decades
00:17:14.26 00:17:17.08 before either death or recovery occurs.
00:17:17.09 00:17:22.02 In the pre-chemotherapeutic era, it was very common for individuals to have
00:17:22.03 00:17:26.10 relapsing bouts of tuberculosis over periods of many years or even decades.
00:17:26.11 00:17:30.15 This, obviously, maximizes opportunities for transmission in a community.
00:17:30.16 00:17:35.00 Furthermore, an individual who has recovered from tuberculosis
00:17:35.01 00:17:38.16 is not protected against reinfection.
00:17:38.17 00:17:41.23 So, unlike in the case of measles, where an individual
00:17:41.24 00:17:44.11 who has had measles is not going to get measles again
00:17:44.12 00:17:48.24 an individual who has had TB can be reinfected again one or many times.
00:17:48.25 00:17:52.15 So, the acquired immunity that develops in the course of infection
00:17:52.16 00:17:56.23 not only does not necessarily clear the pathogen, it does not prevent re-infection either.
00:17:56.24 00:18:00.15 This causes the epidemic dynamics of TB to be radically different
00:18:00.16 00:18:04.25 from those of sort of garden variety acute infections like measles.
00:18:04.26 00:18:10.14 And this translates into an epidemiological parameter that is startlingly different
00:18:10.15 00:18:12.00 between these two categories of diseases.
00:18:12.01 00:18:14.07 This is called the critical community size.
00:18:14.08 00:18:17.28 This is simply a measure of how large a population
00:18:17.29 00:18:20.25 of contiguous interacting individuals is required
00:18:20.26 00:18:24.12 to maintain a pathogen permanently without burnout
00:18:24.13 00:18:26.08 or reintroduction from another source.
00:18:26.09 00:18:29.22 In the case of measles, this number is about 300,000.
00:18:29.23 00:18:35.06 In other words, a minimum population of 300,000 is required to sustain measles indefinitely
00:18:35.07 00:18:36.26 without burnout or reintroduction.
00:18:36.27 00:18:40.13 In the case of TB, this number is on the order of 100-200.
00:18:40.14 00:18:45.09 Very small. And this explains, I think, why tuberculosis is such an ancient pathogen.
00:18:45.10 00:18:50.12 It's well adapted to surviving in small, scattered populations
00:18:50.13 00:18:52.21 of its host -- in this case, human beings --
00:18:52.22 00:18:55.00 which, of course, describes the state of humanity
00:18:55.01 00:18:58.02 until the very recent history, as I described earlier.
00:18:58.03 00:19:03.29 And I believe that it's this persistence... this unusual tenacity of the pathogen
00:19:04.00 00:19:08.02 that accounts for its enormous continued impact on global human health.
00:19:08.03 00:19:10.07 According to the World Health Organization,
00:19:10.08 00:19:15.10 of the 57 million or so people who die every year of various causes --
00:19:15.11 00:19:17.10 so this is of all causes combined, world-wide --
00:19:17.11 00:19:21.10 about 1 in 4 individuals die of 1 or another infectious disease.
00:19:21.11 00:19:26.29 Among infectious diseases, the big three are tuberculosis, HIV/AIDS, and malaria,
00:19:27.00 00:19:30.23 which together account for something like 6-7 million deaths every year.
00:19:30.24 00:19:37.20 It's worth noting, in fact, that tuberculosis is the leading cause of death world-wide
00:19:37.21 00:19:39.29 among individuals who are infected with HIV/AIDS.
00:19:40.00 00:19:42.21 These two diseases synergize in a really sinister way.
00:19:42.22 00:19:46.10 Obviously, being immunocompromised due to HIV/AIDS infection
00:19:46.11 00:19:50.08 pre-disposes an individual to acquire all kinds of infectious diseases,
00:19:50.09 00:19:55.12 including tuberculosis, but it's become clear that having active tuberculosis
00:19:55.13 00:20:00.23 also somehow accelerates progression from HIV infection to full-blown AIDS,
00:20:00.24 00:20:02.19 so it's a real synergy between these two diseases,
00:20:02.20 00:20:05.08 the convergence of which, in developing countries,
00:20:05.09 00:20:08.08 has spelled a real disaster for public health, globally.
00:20:08.09 00:20:15.23 So, 2 million deaths attributed to tuberculosis is obviously a large number,
00:20:15.24 00:20:18.04 but it's really only the tip of the iceberg...
00:20:18.05 00:20:20.10 what I like to call the iceberg of pathogenesis,
00:20:20.11 00:20:23.18 in terms of its impact on global human health,
00:20:23.19 00:20:25.22 which is illustrated in this pyramid diagram here.
00:20:25.23 00:20:28.22 It's a bit like an iceberg if you think about it.
00:20:28.23 00:20:31.24 We tend to focus on the bit of the iceberg that shows above the waterline
00:20:31.25 00:20:32.27 -- the tip of the iceberg,
00:20:32.28 00:20:35.26 but it truly is only part of the problem.
00:20:35.27 00:20:38.23 And in fact, the bulk of the problem -- the bulk of the iceberg if you like --
00:20:38.24 00:20:43.14 that is below the water line is in fact what's going to sink your ship if you run into it.
00:20:43.15 00:20:45.22 So, it's much the same in the case of tuberculosis.
00:20:45.23 00:20:48.15 We tend to focus on what we can see -- what's obvious --
00:20:48.16 00:20:52.15 which is the 2 million or so deaths attributable to TB each year.
00:20:52.16 00:20:53.29 But, in fact, that number is quite small,
00:20:54.00 00:20:58.00 compared to the number of active cases of tuberculosis at any given time,
00:20:58.01 00:21:00.17 which is in the range of 16-20 million.
00:21:00.18 00:21:04.00 That's a large number of individuals at any given time who are out there,
00:21:04.01 00:21:05.15 sick, transmitting the disease.
00:21:05.16 00:21:09.00 But even that number is dwarfed by the number of individuals
00:21:09.01 00:21:12.24 who currently harbor the tubercle bacillus, albeit in a latent form.
00:21:12.25 00:21:14.25 That number approaches 2 billion.
00:21:14.26 00:21:19.18 So, at this point in history, about 1 in 3 people on the planet
00:21:19.19 00:21:24.07 harbor the tubercle bacillus in their tissues and are at risk of developing disease.
00:21:24.08 00:21:29.00 Now, if these individuals who have latent tuberculosis and are not infectious
00:21:29.01 00:21:32.02 continue to harbor the infection only in a latent form,
00:21:32.03 00:21:36.05 this would be a sort of interesting medical curiosity, but not much more than this.
00:21:36.06 00:21:38.13 The problem, as I've already indicated,
00:21:38.14 00:21:41.18 is that individuals who are latently infected are at significant risk
00:21:41.19 00:21:42.26 for the remainder of their lives
00:21:42.27 00:21:47.06 of reactivating and developing full-blown, infectious tuberculosis.
00:21:47.07 00:21:50.10 The lifetime risk for an individual with latent TB
00:21:50.11 00:21:54.08 is about 10%. That number goes up to about 10% per annum
00:21:54.09 00:21:57.18 for individuals who become co-infected with HIV/AIDS.
00:21:57.19 00:22:01.06 When you think of the implications of this, it's rather staggering
00:22:01.07 00:22:02.29 in terms of the future of global public health.
00:22:03.00 00:22:09.12 What it means is that, even if today we could intervene with some magic intervention
00:22:09.13 00:22:11.18 that blocked transmission and new infection from happening... let's say
00:22:11.19 00:22:14.20 a transmission-blocking vaccine, for example...
00:22:14.21 00:22:16.23 Of course, we don't have a tool like that, but let's say we did.
00:22:16.24 00:22:21.12 Even so, we could expect to see, over the course of the next 50 years or so,
00:22:21.13 00:22:27.13 something like 200 million or more new cases of tuberculosis arising around the world,
00:22:27.14 00:22:32.06 due to the reactivation of infections that already exist today.
00:22:32.07 00:22:36.28 So, clearly, there's an enormous need to tackle this problem
00:22:36.29 00:22:39.25 of latent tuberculosis and reactivation.
00:22:39.26 00:22:43.23 But the stark reality is that we currently have no tools whatsoever
00:22:43.24 00:22:48.06 that are both effective and practical to intervene against latent TB.
00:22:48.07 00:22:50.29 The only tool we currently have to use against latent TB
00:22:51.00 00:22:54.14 is 9 months of chemoprophylaxis with a drug called isoniazid,
00:22:54.15 00:22:57.21 and I don't need to explain why I think it's impractical to treat
00:22:57.22 00:23:00.04 2 billion people on a global basis
00:23:00.05 00:23:02.23 with 9 months of drug therapy. It's not going to happen.
00:23:02.24 00:23:06.10 So, this is an enormous unmet need -- something I'll talk about again later --
00:23:06.11 00:23:07.29 in global health.
00:23:08.00 00:23:10.17 And it has an impact on us,
00:23:10.18 00:23:13.06 here, in developed countries like the United States.
00:23:13.07 00:23:16.01 We don't think a lot about TB because most
00:23:16.02 00:23:18.15 of the burden of TB globally is in other countries,
00:23:18.16 00:23:19.29 particularly developing countries.
00:23:20.00 00:23:22.26 But, in fact, what's happening in developing countries
00:23:22.27 00:23:26.05 has a big impact on our health right here in the United States.
00:23:26.06 00:23:30.29 These are data from the CDC, where I've plotted the incidents per year
00:23:31.00 00:23:35.04 in the United States of tuberculosis among individuals born in the United States
00:23:35.05 00:23:37.28 (shown in blue) versus individuals born elsewhere
00:23:37.29 00:23:40.03 who have immigrated to the United States (in red).
00:23:40.04 00:23:43.10 So, following an upsurge in TB cases in the 1980s,
00:23:43.11 00:23:46.23 public health measures were instituted to prevent transmission
00:23:46.24 00:23:49.06 to identify individuals who were infected,
00:23:49.07 00:23:54.18 and case contacts were contacted and new infections
00:23:54.19 00:23:56.13 were detected early and treated and so on.
00:23:56.14 00:23:59.16 So this was quite an effective means -- these public health measures --
00:23:59.17 00:24:02.15 of bringing down the incidents of tuberculosis
00:24:02.16 00:24:04.20 among individuals born here in the United States.
00:24:04.21 00:24:09.06 But, as you can see, the numbers simultaneously, among individuals who were born abroad,
00:24:09.07 00:24:09.28 haven't budged at all.
00:24:09.29 00:24:11.27 And what we think is happening is that
00:24:11.28 00:24:16.24 almost all of these cases are occurring as a result of a reactivation of TB
00:24:16.25 00:24:20.07 from infections that were acquired in childhood.
00:24:20.08 00:24:23.01 So, individuals living in endemic countries acquired the infection in childhood,
00:24:23.02 00:24:26.13 they then immigrated with a latent infection to the United States
00:24:26.14 00:24:29.03 and reactivate and develop tuberculosis.
00:24:29.04 00:24:31.24 What these data, I think, clearly show
00:24:31.25 00:24:35.07 is that failure to control tuberculosis anywhere in the world
00:24:35.08 00:24:39.20 translates into a failure to control TB everywhere in the world,
00:24:39.21 00:24:41.28 including in the United States.
00:24:44.22 00:24:48.29 Now, one of the myths I'd like to dispel about TB is that it's an old people's disease,
00:24:49.00 00:24:54.10 and that it mainly preys on those whose immune systems have waned
00:24:54.11 00:24:59.08 due to old age or to immunosuppressive therapy, and so on.
00:24:59.09 00:25:03.29 In fact, the reality is that around the world, tuberculosis is overwhelmingly a disease
00:25:04.00 00:25:05.07 of young adults.
00:25:05.08 00:25:09.04 And that's illustrated in the graphs shown here,
00:25:09.05 00:25:12.00 where, as you can see, looking on this side
00:25:12.01 00:25:14.20 at the age distribution along the y-axis
00:25:14.21 00:25:17.15 versus number of TB deaths along the x-axis,
00:25:17.16 00:25:22.15 TB deaths are overwhelmingly concentrated among adults in the age group 15-59.
00:25:22.16 00:25:27.27 It's very similar, in fact, to the demographic distribution of AIDS cases around the world,
00:25:27.28 00:25:30.10 which again, primarily afflicts young adults.
00:25:30.11 00:25:34.29 This is in stark contrast to most other infectious diseases, shown here.
00:25:35.00 00:25:39.22 These are the data compiled for all infectious diseases, except tuberculosis and AIDS.
00:25:39.23 00:25:42.18 And, as you can see, in contrast to tuberculosis,
00:25:42.19 00:25:46.22 most infectious diseases prey overwhelmingly on the very young.
00:25:46.23 00:25:48.02 Why does this matter?
00:25:48.03 00:25:51.05 Well, it matters because TB and AIDS,
00:25:51.06 00:25:54.07 since they remove precisely those individuals
00:25:54.08 00:25:56.08 in whom society has already invested
00:25:56.09 00:25:58.16 and those individuals who are sort of the backbone
00:25:58.17 00:26:01.06 of the socioeconomic structure in societies
00:26:01.07 00:26:05.01 have a disproportionate impact. Yes, they kill large numbers of people,
00:26:05.02 00:26:10.24 but they also kill that segment of the population that populations can ill afford to lose.
00:26:10.25 00:26:15.13 This is particularly true in developing countries where social security systems,
00:26:15.14 00:26:18.06 for example, may not exist, and where young adults
00:26:18.07 00:26:22.00 usually have dependents in both the young and the old age categories.
00:26:22.01 00:26:25.01 In fact, it might surprise you to know that tuberculosis
00:26:25.02 00:26:28.18 continues to be the third leading cause of death worldwide
00:26:28.19 00:26:32.19 among young adults, age 15-59. HIV/AIDS being first
00:26:32.20 00:26:35.12 and ischemic heart disease being second.
00:26:35.13 00:26:42.18 TB is also the third leading cause of morbidity among the young adult age group,
00:26:42.19 00:26:45.28 measured in terms of disability-adjusted life years,
00:26:45.29 00:26:47.10 rather than deaths.
00:26:47.11 00:26:50.01 So, this is a metric that was devised by Murray and Lopez
00:26:50.02 00:26:52.21 at Harvard University in the mid-1990s
00:26:52.22 00:26:56.09 which attempts to capture the disease burden caused by a condition
00:26:56.10 00:26:58.24 apart from simple mortality.
00:26:58.25 00:27:02.08 So, I think it's quite interesting in this context, for example...
00:27:02.09 00:27:05.14 just as an aside... that unipolar-depressive disorders,
00:27:05.15 00:27:07.24 which do not show up in the mortality charts at all,
00:27:07.25 00:27:11.19 have now become the second leading cause of disability-adjusted life years lost
00:27:11.20 00:27:16.24 worldwide, and are slated to become the leading cause within the next few decades.
00:27:16.25 00:27:21.04 This matters a great deal -- how we quantify the burden of disease --
00:27:21.05 00:27:26.13 because, of course, the allocation of resources follows the perceived need.
00:27:26.14 00:27:29.03 So, when we focus on morbidity, rather than mortality,
00:27:29.04 00:27:31.05 we target rather different conditions.

Part 2: Tools for Tuberculosis Control: Not Just a Problem of Implementation

00:00:03.27 00:00:07.27 In the last lecture, we talked about the global problem of tuberculosis.
00:00:07.28 00:00:13.12 I gave you some statistics on the ongoing morbidity and mortality of this global epidemic.
00:00:13.13 00:00:18.23 Now, I want to talk to you about the tools that we currently have for tuberculosis control.
00:00:18.24 00:00:22.08 Clearly, whatever it is that we're doing, it's not good enough,
00:00:22.09 00:00:24.09 because it continues to be one of the leading causes
00:00:24.10 00:00:26.10 of human morbidity and mortality worldwide.
00:00:26.11 00:00:29.27 The question is: is this because the tools that we have are inadequate
00:00:29.28 00:00:34.27 or is it just a matter of implementing more adequately the tools that we do, in fact, have?
00:00:34.28 00:00:39.28 It has been opined by many that the tools that we have, if properly administered,
00:00:39.29 00:00:42.18 could do the job. I personally don't believe that.
00:00:42.19 00:00:45.03 I think that the tools we have are not adequate to the task,
00:00:45.04 00:00:48.04 and I hope to convince you of that in this next segment of the lecture.
00:00:48.05 00:00:53.13 So, let's go back to the statistics that I described before,
00:00:53.14 00:00:56.03 in terms of this iceberg of pathogenesis
00:00:56.04 00:00:59.16 indicated here, where we see that there are large numbers of individuals
00:00:59.17 00:01:01.05 who die each year of tuberculosis
00:01:01.06 00:01:06.04 (about 2 million), 16 million individuals or so who are currently ill with the disease,
00:01:06.05 00:01:10.26 and about 2 billion individuals who are infected and at significant risk
00:01:10.27 00:01:13.04 of developing tuberculosis in the future.
00:01:13.05 00:01:15.18 That's a huge global burden of disease, obviously,
00:01:15.19 00:01:19.09 that begs the question: what about the tools that we currently
00:01:19.10 00:01:21.19 are using to prevent tuberculosis?
00:01:21.20 00:01:22.27 I'm sure, as many of you are aware,
00:01:22.28 00:01:27.16 we do have a vaccine for tuberculosis -- the so-called BCG vaccine
00:01:27.17 00:01:31.20 which was first introduced in 1921, and is still widely administered.
00:01:31.21 00:01:36.01 We also have multi-drug therapy for TB that can be effective
00:01:36.02 00:01:39.03 and that's been available since the early 1950s,
00:01:39.04 00:01:44.09 and yet, every year, about 2 million people die of this disease,
00:01:44.10 00:01:47.10 and 16 million people are sick with it.
00:01:47.11 00:01:50.29 So, this begs the question, why aren't these tools doing a better job?
00:01:51.00 00:01:52.29 Let's start first with the BCG vaccine.
00:01:53.00 00:01:57.19 Here, I've sketched out a brief history of the development
00:01:57.20 00:02:00.16 of the so-called BCG or Bacille Calmette-Guerin
00:02:00.17 00:02:02.21 vaccine for prevention of tuberculosis.
00:02:02.22 00:02:04.27 The history begins in 1902,
00:02:04.28 00:02:07.25 when a microbiologist Edmond Nocard isolated a virulent strain
00:02:07.26 00:02:10.17 of Mycobacterium bovis from a tuberculous cow.
00:02:10.18 00:02:14.05 M. bovis is virtually identical to Mycobacterium tuberculosis,
00:02:14.06 00:02:16.22 the primary cause of tuberculosis in humans.
00:02:16.23 00:02:20.29 He then gave that strain to a couple of scientists at the Pasteur Institute --
00:02:21.00 00:02:23.03 Albert Calmette and Camille Guerin --
00:02:23.04 00:02:26.27 who attenuated it by serial passage in axenic culture
00:02:26.28 00:02:30.21 between 1908 and 1919 (a technique that was, of course, first developed
00:02:30.22 00:02:33.14 by Louis Pasteur in developing the rabies vaccine).
00:02:33.15 00:02:39.15 In 1921, the BCG vaccine was given for the first time to a human being --
00:02:39.16 00:02:42.03 a child who was born to a mother who was tuberculous
00:02:42.04 00:02:45.00 in a family that was, in fact, tuberculous
00:02:45.01 00:02:47.19 and deemed at high risk of developing the disease.
00:02:47.20 00:02:50.09 The child suffered no adverse consequences, and in fact,
00:02:50.10 00:02:54.17 grew to a healthy adulthood without a bout of tuberculosis.
00:02:54.18 00:02:57.29 From 1921 to the present, BCG has been administered
00:02:58.00 00:03:00.23 to more than 3 billion individuals worldwide.
00:03:00.24 00:03:04.15 So, on the surface of things, it would seem to be a highly successful...
00:03:04.16 00:03:07.12 at least a very widely administered vaccine.
00:03:07.13 00:03:12.16 And in fact, BCG, as an immunogen, is highly impressive
00:03:12.17 00:03:17.14 and has many characteristics that are ideal for a vaccine against tuberculosis,
00:03:17.15 00:03:18.29 which I'd like to outline for you here.
00:03:19.00 00:03:22.16 The first -- and this is an important one -- is that BCG can be given
00:03:22.17 00:03:25.01 at the time of birth or at any time thereafter.
00:03:25.02 00:03:28.23 In fact, it can be given multiple times if that's desirable for boosting purposes.
00:03:28.24 00:03:34.10 So, it's one of the few live, attenuated vaccines that's quite safe to give to newborn infants.
00:03:34.11 00:03:35.24 Why is that important?
00:03:35.25 00:03:41.12 Well, you want to use a TB vaccine in countries where TB is endemic, obviously,
00:03:41.13 00:03:43.15 and this overwhelmingly means developing countries --
00:03:43.16 00:03:46.18 poor countries, where an individual's best chance of seeing
00:03:46.19 00:03:48.25 a healthcare provider is, in fact, at birth.
00:03:48.26 00:03:51.24 If you don't manage to vaccinate the child at birth,
00:03:51.25 00:03:55.22 you have to wait a few months... you're going to lose a lot of those kids.
00:03:55.23 00:03:56.28 They will simply not be vaccinated.
00:03:56.29 00:03:59.15 So, that's a very attractive trait.
00:03:59.16 00:04:03.19 Second, BCG is very safe -- even when administered to newborns.
00:04:03.20 00:04:08.07 In fact, the case fatality rate is about 0.19 per million,
00:04:08.08 00:04:11.15 which is much lower than other live, attenuated vaccines,
00:04:11.16 00:04:14.11 such as the smallpox vaccine that had been widely administered.
00:04:14.12 00:04:18.08 Third, BCG is very inexpensive to produce.
00:04:18.09 00:04:21.03 It's been said it costs about a dime a dose to produce this stuff.
00:04:21.04 00:04:22.18 Why is this important?
00:04:22.19 00:04:26.13 Well, as I said, we want to use a vaccine against tuberculosis
00:04:26.14 00:04:29.13 primarily in developing countries that are poor.
00:04:29.14 00:04:33.04 So, any vaccine that's going to be accessible to individuals in developing countries
00:04:33.05 00:04:36.01 has to be inexpensive to produce and distribute.
00:04:36.02 00:04:41.19 The next point is BCG is the most stable live vaccine that's currently in use.
00:04:41.20 00:04:46.01 It does not require refrigeration in order to maintain its immunogenic properties.
00:04:46.02 00:04:51.01 Why is that important? Well, again, we want to administer this vaccine
00:04:51.02 00:04:55.24 to populations living in remote, rural areas -- that's overwhelmingly where TB occurs.
00:04:55.25 00:04:59.16 And so, it's very important to be able to use the vaccine
00:04:59.17 00:05:02.16 in environments where the cold chain does not reach.
00:05:02.17 00:05:06.04 If we had to refrigerate the vaccine, we could not reach nearly as many people.
00:05:06.05 00:05:12.13 Next, BCG can be administered orally, and it seems to be about equally immunogenic
00:05:12.14 00:05:14.14 whether it's given orally or by other routes
00:05:14.15 00:05:16.09 (for example, the percutaneous route).
00:05:16.10 00:05:17.18 Why is that an advantage?
00:05:17.19 00:05:21.03 Again, we want to administer this vaccine in developing communities,
00:05:21.04 00:05:23.09 where needles are not always available, and if they are
00:05:23.10 00:05:26.28 available, they're not always safe, because needles often
00:05:26.29 00:05:29.22 in developing country scenarios are reused.
00:05:29.23 00:05:33.24 That leads to risk of transmission of HIV/AIDS, hepatitis C, and so on.
00:05:33.25 00:05:40.20 BCG is a very immunogenic vaccine
00:05:40.21 00:05:44.04 in the sense that it elicits both cellular and humoral responses.
00:05:44.05 00:05:47.18 So, you get antibodies and T cell responses to this vaccine.
00:05:47.19 00:05:49.01 It's quite potent, in that regard.
00:05:49.02 00:05:53.26 And, it elicits immunity at both the systemic and mucosal level.
00:05:53.27 00:05:59.06 So, as an immunogen, it's a fantastically broad eliciting
00:05:59.07 00:06:01.20 of immune responses that you get with BCG.
00:06:01.21 00:06:03.28 Lastly, and this is a huge advantage,
00:06:03.29 00:06:08.28 BCG seems to confer long-lived memory without the necessity of boosting.
00:06:08.29 00:06:12.06 So, it's very common for individuals who are vaccinated in childhood
00:06:12.07 00:06:18.29 still to be reactive to the skin test that we use to detect vaccination many decades later
00:06:19.00 00:06:22.02 even without boosting. Memory, obviously, is very important
00:06:22.03 00:06:24.19 for a vaccine if you want protection that is long lived.
00:06:24.20 00:06:27.27 So, this is a really formidable list of advantages that
00:06:27.28 00:06:32.00 BCG has to offer. In fact, it's going to be tough with any new vaccine
00:06:32.01 00:06:33.06 to match these advantages.
00:06:33.07 00:06:37.04 And, in fact, BCG from my perspective, has only 1 really major drawback,
00:06:37.05 00:06:39.10 which is that it does not prevent tuberculosis.
00:06:39.11 00:06:43.16 Admittedly, that's a rather major drawback for a vaccine
00:06:43.17 00:06:45.15 that's designed specifically for that purpose.
00:06:45.16 00:06:50.06 Actually, it's a little unfair to BCG to say that it does not prevent tuberculosis.
00:06:50.07 00:06:52.24 It does not consistently prevent tuberculosis.
00:06:52.25 00:06:57.11 And this has been shown in a large number of clinical trials,
00:06:57.12 00:07:02.23 in which the protective efficacy of the BCG vaccine in preventing adult tuberculosis
00:07:02.24 00:07:04.21 has been measured. As we can see...
00:07:04.22 00:07:08.07 Look in different regions. In some trials, such as this one shown here,
00:07:08.08 00:07:12.03 (a case controlled study) the vaccine efficacy was better than 80%.
00:07:12.04 00:07:14.10 Really quite high.
00:07:14.11 00:07:16.11 In other trials that are operationally just as good,
00:07:16.12 00:07:18.07 we see no protection whatsoever.
00:07:18.08 00:07:23.09 And, this pattern of sort-of variable responsiveness does not seem
00:07:23.10 00:07:30.04 to depend on the type of strategy that is used to study the effect of BCG vaccination.
00:07:30.05 00:07:36.24 And, in fact, most compellingly, in what we regard as the gold standard of clinical trials
00:07:36.25 00:07:39.23 to test a vaccine or a drug of any new intervention
00:07:39.24 00:07:43.01 (that is, randomized, double-blinded, placebo control trials),
00:07:43.02 00:07:47.19 we find, again, this enormous variability from trial to trial
00:07:47.20 00:07:50.19 in the apparent efficacy of BCG in preventing TB.
00:07:50.20 00:07:53.09 So, in some trials, for example, this one in Haiti,
00:07:53.10 00:07:56.29 BCG efficacy approached 80% -- really quite good.
00:07:57.00 00:08:00.22 We'd be pretty happy if we had a vaccine that was consistently 80% effective.
00:08:00.23 00:08:05.16 But, in other trials, you can see that the efficacy of BCG was much lower.
00:08:05.17 00:08:08.22 And in fact, in the largest trial that has yet been completed to date,
00:08:08.23 00:08:11.01 that was carried out in India many years ago,
00:08:11.02 00:08:15.03 which involved more than 300,000 individuals enrolled in the trial,
00:08:15.04 00:08:17.12 the efficacy of the vaccine was nil.
00:08:17.13 00:08:19.03 No protection whatsoever.
00:08:19.04 00:08:22.08 And I want to emphasize that operationally, these are all equally good trials
00:08:22.09 00:08:27.02 that meet the criteria for randomized, double-blind, placebo-controlled trials.
00:08:27.03 00:08:28.05 That's about as good as it gets.
00:08:28.06 00:08:32.19 So, we don't understand... There are a number of theories
00:08:32.20 00:08:35.22 out there, but we truly don't understand at this point
00:08:35.23 00:08:39.09 why it is that BCG sometimes works and sometimes does not work.
00:08:39.10 00:08:41.11 There are a number of competing theories.
00:08:41.12 00:08:45.07 My personal suspicion is that there are probably multiple factors at work here.
00:08:45.08 00:08:48.23 But, one thing is absolutely clear, and that is that we desperately need
00:08:48.24 00:08:52.08 a vaccine that is more consistently effective than BCG.
00:08:52.09 00:08:54.15 But, I'm not sure how we're going to get there at this point.
00:08:54.16 00:08:59.18 Now, most of what we know about the immune responses
00:08:59.19 00:09:03.17 to TB or to BCG vaccination come from studies
00:09:03.18 00:09:05.19 done in animal models of infection --
00:09:05.20 00:09:09.14 in particular, the mouse model of infection -- where we can infect
00:09:09.15 00:09:11.20 these animals with small numbers of tubercle bacilli
00:09:11.21 00:09:16.13 by the respiratory route. And I think the kinetics of infection and reinfection that we see
00:09:16.14 00:09:19.19 in these models provide us with a clue as to why
00:09:19.20 00:09:22.29 the immune responses that develop against TB are not more effective.
00:09:23.00 00:09:28.06 So, in this schematic, what I have shown is the
00:09:28.07 00:09:31.29 infection dynamics in a single animal
00:09:32.00 00:09:38.03 of a typical acute, transient pathogen (I've used Listeria as my example
00:09:38.04 00:09:40.22 because it's a very well studied acute pathogen)
00:09:40.23 00:09:45.03 versus Mycobacterium, as emblematic of a chronic, persistent pathogen.
00:09:45.04 00:09:47.08 So, focusing first on Listeria, what you can see
00:09:47.09 00:09:51.17 is that after entry into the body, the organism replicates rapidly,
00:09:51.18 00:09:56.13 reaching a peak and engendering a rapid adaptive immune response to the pathogen.
00:09:56.14 00:10:00.22 This leads, in many cases, to the total clearance of the pathogen
00:10:00.23 00:10:03.07 from the animal, and the animal is truly self-cured
00:10:03.08 00:10:06.01 in the sense that if you were now to immune-suppress this animal,
00:10:06.02 00:10:08.04 you would not see any relapse.
00:10:08.05 00:10:11.29 So, the immunity that develops appears to sterilize the animal completely.
00:10:12.00 00:10:15.04 That acquired immunity is also quite effective
00:10:15.05 00:10:18.22 in preventing reinfection with the same pathogen, as shown here.
00:10:18.23 00:10:22.23 If you take an animal that has self-cured of Listeria infection
00:10:22.24 00:10:27.03 and attempt to reinfect it, what you see is essentially an abortive infection
00:10:27.04 00:10:30.14 in which the pathogen never really takes hold and is effectively cleared out.
00:10:30.15 00:10:36.00 In contrast, when we look at the dynamics of Mycobacterium infection
00:10:36.01 00:10:41.09 in animal models, what we see is that there is a much longer period of slow growth
00:10:41.10 00:10:45.20 in the tissue. So, one striking feature of this pathogen is that it grows very slowly in vivo,
00:10:45.21 00:10:48.16 with a doubling time of about 24 hours.
00:10:48.17 00:10:52.10 So, perhaps for this reason, perhaps for other reasons... we really don't understand
00:10:52.11 00:10:56.15 why... it takes a very long time for the adaptive immune response to rally
00:10:56.16 00:10:59.19 in a TB infection. But when it finally does,
00:10:59.20 00:11:03.10 this leads to a curtailment of bacterial growth in the lungs,
00:11:03.11 00:11:06.11 but it does not lead to a diminution of bacterial numbers
00:11:06.12 00:11:11.17 in the lungs. So, what ensues is a kind of stalemate between the pathogen and the host,
00:11:11.18 00:11:15.22 in which bacterial growth is contained, but in which, in contrast to say, Listeria,
00:11:15.23 00:11:18.29 the pathogen is not cleared from the tissue.
00:11:19.00 00:11:23.16 So, there's a failure, if you like, of adaptive immunity to clear the host of infection,
00:11:23.17 00:11:26.21 which of course leads to a state of latency in humans.
00:11:26.22 00:11:29.25 That failure of adaptive immunity is also apparent
00:11:29.26 00:11:34.11 if we look at the kinetics of reinfection of a previously infected animal.
00:11:34.12 00:11:38.17 And what we see in this case is that the initial kinetics of growth
00:11:38.18 00:11:42.08 in the tissues of the experienced animal are not altered at all.
00:11:42.09 00:11:46.17 So, the initial growth rate is exactly the same as it would be in a naive,
00:11:46.18 00:11:51.05 inexperienced animal because there are already pre-existing
00:11:51.06 00:11:54.04 populations of antigen-specific T-cells that are capable
00:11:54.05 00:11:56.09 of recognizing the tubercle bacillus.
00:11:56.10 00:11:59.17 We see that these infiltrate the lesions about a week or two earlier
00:11:59.18 00:12:01.09 than they would in naive animals,
00:12:01.10 00:12:05.06 leading to a curtailment of further growth somewhat earlier in the establishment
00:12:05.07 00:12:09.29 of a plateau that's about 1 log or so lower than it would be in a naive animal.
00:12:10.00 00:12:13.03 But, there's no elimination of infection in this case either,
00:12:13.04 00:12:15.06 and there's no prevention of infection.
00:12:15.07 00:12:20.23 So, the kinetics of infection and reinfection in a persistent pathogen
00:12:20.24 00:12:25.17 like TB are radically different than they are in a transient pathogen, like Listeria.
00:12:25.18 00:12:29.13 And, if you think about it, this has enormous implications for vaccine development.
00:12:29.14 00:12:32.25 The reason is that, overwhelmingly, the vaccines that we use
00:12:32.26 00:12:35.13 around the world today are predicated on principles
00:12:35.14 00:12:37.27 that were laid down by Louis Pasteur in the 19th century.
00:12:37.28 00:12:40.13 Namely, you take the pathogen of interest,
00:12:40.14 00:12:43.12 you attenuate it to the point where it is still immunogenic,
00:12:43.13 00:12:46.28 but no longer causes disease, and that's your vaccine.
00:12:46.29 00:12:50.10 And most of the vaccines that we use currently today --
00:12:50.11 00:12:52.27 the hepatitis B subunit being a notable exception --
00:12:52.28 00:12:57.04 are based on the live, attenuated pathogen principle.
00:12:57.05 00:13:01.10 Now, that's an effective approach if you're talking about a pathogen like Listeria,
00:13:01.11 00:13:04.05 where the adaptive immune responses that develop in response
00:13:04.06 00:13:07.22 to natural infection are adequate to prevent reinfection.
00:13:07.23 00:13:12.11 But, clearly, this is not an adequate strategy for pathogens like the tubercle bacillus,
00:13:12.12 00:13:17.25 where that adaptive immunity is helpless to stop secondary infection.
00:13:17.26 00:13:21.11 And, in recent years, in clinical studies, largely in South Africa,
00:13:21.12 00:13:26.14 it's become very clear that, in fact, reinfection of previously experienced individuals
00:13:26.15 00:13:29.18 is a major source of tuberculosis disease globally.
00:13:29.19 00:13:34.11 So, to develop a new vaccine that's more effective than BCG,
00:13:34.12 00:13:38.07 we are somehow going to have to do better than nature.
00:13:38.08 00:13:40.13 We're going to have to go beyond Louis Pasteur's principles
00:13:40.14 00:13:44.10 and come up with a vaccine that is more protective than natural infection.
00:13:44.11 00:13:48.00 The exciting thing is that right now there are efforts underway in many labs
00:13:48.01 00:13:50.02 around the world to develop such a vaccine.
00:13:50.03 00:13:53.09 But to some extent, I think it's true that we're still flying the dark.
00:13:53.10 00:13:57.22 We really don't understand what we want to achieve with a vaccine,
00:13:57.23 00:14:00.09 either in terms of the immune responses that we elicit
00:14:00.10 00:14:03.20 or in terms of getting around the defense mechanisms of the pathogen.
00:14:03.21 00:14:07.04 So, this is an enormous area for future research.
00:14:07.05 00:14:12.22 So, we've hit upon the problems of the BCG vaccine
00:14:12.23 00:14:14.25 as a tool to prevent tuberculosis.
00:14:14.26 00:14:17.16 Let's talk now about drug therapy as a means of preventing
00:14:17.17 00:14:20.26 or treating the disease tuberculosis.
00:14:20.27 00:14:26.04 So, we do have effective therapy for TB, and we have had since 1952.
00:14:26.05 00:14:30.11 The story of development of anti-TB drugs began with the discovery of
00:14:30.12 00:14:34.06 streptomycin, a natural product, in the year 1947 by Selman Waksman,
00:14:34.07 00:14:39.22 followed rapidly by the development of pyrazinamide and isoniazid (both in 1952),
00:14:39.23 00:14:45.16 ethambutol in 1958, and most recently, rifampin, which was introduced in 1967.
00:14:45.17 00:14:49.23 So, I might point out that since the development of rifampin,
00:14:49.24 00:14:54.03 the further development of the anti-TB armamentarium has really faltered.
00:14:54.04 00:14:57.03 We haven't had a new frontline drug for treatment of tuberculosis,
00:14:57.04 00:15:01.14 despite the overwhelming need, since rifampin was introduced in the 1960s.
00:15:01.15 00:15:06.08 And, to some extent, I think it's true that the failure, if you like,
00:15:06.09 00:15:10.28 of the pharmaceutical industry to develop new interventions against TB
00:15:10.29 00:15:12.11 since introduction of rifampin
00:15:12.12 00:15:16.29 is kind of a case of it being a victim of its own success.
00:15:17.00 00:15:20.07 So, it was believed at the time these drugs were developed
00:15:20.08 00:15:23.18 that a cure had been found that would reverse the tide, so to speak,
00:15:23.19 00:15:25.04 of global tuberculosis.
00:15:25.05 00:15:30.04 And this early optimism was reflected, I think, quite nicely in this quote
00:15:30.05 00:15:34.15 from Selman Waksman in his optimistically entitled book,
00:15:34.16 00:15:38.23 The Conquest of Tuberculosis, published in 1964,
00:15:38.24 00:15:43.01 in which he opined that, “tuberculosis was on its way to being reduced to a
00:15:43.02 00:15:46.24 “minor ailment of man”. He predicted that it would be essentially eliminated
00:15:46.25 00:15:51.04 as a significant cause of human illness and death in his lifetime.
00:15:51.05 00:15:54.26 Well, Selman Waksman has been gone for a long time now,
00:15:54.27 00:15:55.23 but tuberculosis is still with us.
00:15:55.24 00:16:00.16 So, I think it's appropriate to ask, why did anti-TB chemotherapy
00:16:00.17 00:16:02.22 fail to live up to its early promise?
00:16:02.23 00:16:08.16 The proper answer to that question is, of course, multi-factorial and complex.
00:16:08.17 00:16:10.18 But I think there's no doubt that the major problem
00:16:10.19 00:16:15.17 confronting TB chemotherapy is that it just takes too long to treat this disease.
00:16:15.18 00:16:19.07 So, this slide outlines the so-called short course chemotherapy regimen
00:16:19.08 00:16:22.24 that's currently recommended by the Centers for Disease Control
00:16:22.25 00:16:24.15 for treatment of tuberculosis.
00:16:24.16 00:16:28.25 It consists of two phases: an induction phase lasting for 2 months,
00:16:28.26 00:16:32.16 in which 4 or 5 of these drugs are given on a daily basis,
00:16:32.17 00:16:37.04 followed by a continuation phase lasting from 4-7 months,
00:16:37.05 00:16:41.27 in which these 2 drugs, isoniazid and rifampin, are given intermittently.
00:16:41.28 00:16:47.02 My point here is that, to properly cure tuberculosis,
00:16:47.03 00:16:52.24 requires administration of multiple drugs for a period of 6-9 months.
00:16:52.25 00:16:57.07 It's very difficult, as you will know if you've ever had to take an antibiotic,
00:16:57.08 00:17:02.22 to adhere to an antibiotic regimen for a period of even a week or two.
00:17:02.23 00:17:06.12 So, there's a problem here of patient adherence or compliance.
00:17:06.13 00:17:11.05 It's simply very difficult to get patients to complete a regimen like this.
00:17:11.06 00:17:16.19 And the problem becomes more and more acute, if you like, as treatment is extended.
00:17:16.20 00:17:19.18 That was shown very nicely in a study from Christopher Murray,
00:17:19.19 00:17:23.00 carried out in Botswana in the late 1980s.
00:17:23.01 00:17:28.11 And, in this graph, I've plotted the likelihood that a given patient will default on therapy
00:17:28.12 00:17:30.26 in percentile terms on the y-axis,
00:17:30.27 00:17:34.12 versus the days that patient is on chemotherapy on the x-axis.
00:17:34.13 00:17:37.03 And, as you can see, it's a constantly rising graph.
00:17:37.04 00:17:40.18 So, the longer a patient is on chemotherapy, the more likely it is
00:17:40.19 00:17:43.18 that the patient is going to default on chemotherapy.
00:17:43.19 00:17:45.12 And, intuitively, that makes a lot of sense.
00:17:45.13 00:17:49.23 Once patients begin therapy, they start to feel better almost immediately
00:17:49.24 00:17:52.03 within a month of taking their meds.
00:17:52.04 00:17:55.12 In most cases, the individual feels that they have been cured.
00:17:55.13 00:17:58.01 So, at that point, it's very difficult to get them to comply
00:17:58.02 00:17:59.25 with a regimen that is so demanding.
00:17:59.26 00:18:01.22 It requires them to take multiple drugs every day,
00:18:01.23 00:18:04.25 many of which have unpleasant side effects,
00:18:04.26 00:18:06.26 such as gastric upset and so on.
00:18:06.27 00:18:12.18 So, the problem is simply that most patients, in fact, do not complete chemotherapy.
00:18:12.19 00:18:17.08 In fact, in some studies, as many as 90% of all patients on chemotherapy
00:18:17.09 00:18:21.05 have defaulted on chemotherapy unless that chemotherapy was closely supervised
00:18:21.06 00:18:26.06 by an attending medical doctor or a nurse or a social worker.
00:18:26.07 00:18:30.23 The most problematic aspect, in some ways,
00:18:30.24 00:18:33.26 of the high rates of default on chemotherapy
00:18:33.27 00:18:36.28 requiring re-treatment and so on, is that it has led to the emergence
00:18:36.29 00:18:40.15 of drug-resistant and even multi-drug-resistant strains of tuberculosis
00:18:40.16 00:18:44.04 which are now a threat to global health in every part of the world.
00:18:44.05 00:18:46.27 As we are hearing these days, even in the United States,
00:18:46.28 00:18:49.27 with the outbreak of extreme multi-drug-resistant tuberculosis
00:18:49.28 00:18:50.28 at multiple sites.
00:18:50.29 00:18:56.10 So, here I'm focusing on data taken from the world's two most populous countries,
00:18:56.11 00:19:00.04 China and India, both of which have very high rates of tuberculosis
00:19:00.05 00:19:03.04 and very high rates of drug resistance.
00:19:03.05 00:19:05.19 In the case of China, in Henan Province,
00:19:05.20 00:19:10.10 where the TB incidence is about 100 per 100,000 population per year,
00:19:10.11 00:19:11.18 (quite a high incidence),
00:19:11.19 00:19:17.12 they find already that about 1/3 or more
00:19:17.13 00:19:21.00 of all TB cases that present in clinic are resistant to one or more
00:19:21.01 00:19:25.22 TB drugs, and fully 15% of these cases are multi-drug-resistant,
00:19:25.23 00:19:29.14 meaning that they are resistant to at least isoniazid and rifampin,
00:19:29.15 00:19:32.13 which currently are our two best front-line TB drugs.
00:19:32.14 00:19:36.02 In India, the rate is rather higher than in China,
00:19:36.03 00:19:38.07 400 per 100,000 in this particular state.
00:19:38.08 00:19:42.04 The rates of resistance are somewhat lower,
00:19:42.05 00:19:46.14 but the overall numbers are very similar to China because of the higher rates of disease.
00:19:46.15 00:19:51.00 So, in this case, about 1 out of 4 TB cases is already resistant to 1 or more drugs,
00:19:51.01 00:19:54.23 and almost 10% of those cases are multi-drug-resistant.
00:19:54.24 00:19:59.12 And this is presenting an enormous challenge in terms of global public health
00:19:59.13 00:20:02.19 because in many cases now, there are no drugs left for treatment
00:20:02.20 00:20:05.04 of extreme multi-drug-resistant TB,
00:20:05.05 00:20:08.05 which can be a virtual death sentence for the person who is so-afflicted.
00:20:08.06 00:20:12.18 And, of course, this individual can also transmit these resistant strains
00:20:12.19 00:20:15.28 to individuals in his or her surroundings.
00:20:15.29 00:20:22.02 So, the question arises why it's so difficult to treat tuberculosis.
00:20:22.03 00:20:24.12 A model that has been around for some time now
00:20:24.13 00:20:26.18 that was first proposed by Denis Mitchison
00:20:26.19 00:20:29.19 (which I therefore call the Mitchison hypothesis)
00:20:29.20 00:20:33.25 suggests that it's difficult to treat tuberculosis because the population of bacteria
00:20:33.26 00:20:36.25 in the lungs of the patient is heterogeneous.
00:20:36.26 00:20:40.27 So, he suggested that there are different tissue environments
00:20:40.28 00:20:45.04 within the infected human lung
00:20:45.05 00:20:48.11 that lead to different physiologies on the part of the tubercle bacillus.
00:20:48.12 00:20:51.08 He suggests that in some compartments, the bacteria
00:20:51.09 00:20:53.02 confront a favorable environment
00:20:53.03 00:20:56.10 where they are able to undergo continuous and rapid growth.
00:20:56.11 00:21:00.14 And these bacteria are most susceptible to the drugs isoniazid and streptomycin,
00:21:00.15 00:21:03.09 which are quite effective against rapidly growing bacteria,
00:21:03.10 00:21:07.10 but poorly effective against slowly growing or non-growing bacteria.
00:21:07.11 00:21:11.09 He suggested that there are acidic compartments
00:21:11.10 00:21:13.28 that the bacteria reside in, where the drug pyrazinamide,
00:21:13.29 00:21:17.22 which acts only under acidic conditions would be most effective.
00:21:17.23 00:21:21.01 That some bacteria are in compartments where they can undergo only
00:21:21.02 00:21:24.11 spurts of metabolism -- so intermittent metabolism and growth --
00:21:24.12 00:21:28.02 where only rifampin would be effective against them.
00:21:28.03 00:21:31.14 And, most importantly, he suggested that there are tissue environments
00:21:31.15 00:21:33.06 where the bacteria are altogether dormant --
00:21:33.07 00:21:35.24 so these are unfavorable environments for growth.
00:21:35.25 00:21:40.24 And dormant bacteria are not terribly susceptible to any of the drugs that we currently use
00:21:40.25 00:21:42.23 to treat tuberculosis.
00:21:42.24 00:21:45.09 And that is not too surprising, if you look at
00:21:45.10 00:21:48.29 the targets of the drugs that we currently use to treat tuberculosis.
00:21:49.00 00:21:53.15 It's a variety of different molecular processes that are targeted by these drugs.
00:21:53.16 00:21:56.24 They all have one thing in common -- cell wall biogenesis,
00:21:56.25 00:22:00.23 RNA synthesis, protein synthesis, and so on
00:22:00.24 00:22:04.02 are all processes that are primarily important for cell growth and division.
00:22:04.03 00:22:08.04 These processes are much less important for the survival of cells
00:22:08.05 00:22:11.23 that have exited the cell division cycle and entered into a state of dormancy.
00:22:11.24 00:22:15.13 So, it's quite plausible, on first principle,
00:22:15.14 00:22:17.23 that dormancy of a subpopulation of bacteria
00:22:17.24 00:22:20.23 may explain the recalcitrance of TB to therapy.
00:22:20.24 00:22:26.27 And, remarkably, there's actually some clinical evidence that would support this hypothesis.
00:22:26.28 00:22:31.02 This comes from a study that was quite remarkable
00:22:31.03 00:22:34.20 published in 1956 by Vandiviere and colleagues
00:22:34.21 00:22:39.15 in which they sampled lung lesions taken from patients who had been on chemotherapy,
00:22:39.16 00:22:42.11 in many cases for a year or more,
00:22:42.12 00:22:46.13 and they cultured the bacteria out of lesions within the lungs of these patients.
00:22:46.14 00:22:49.01 Now, they classified these lesions into two categories,
00:22:49.02 00:22:50.17 according to whether they were closed
00:22:50.18 00:22:54.29 and appeared to be dormant (that is, these were lesions that were more or less inactive
00:22:55.00 00:22:57.28 but persistent) or open and active,
00:22:57.29 00:23:00.23 where the bacteria were clearly undergoing luxury and growth.
00:23:00.24 00:23:04.17 From the active lesion, they found that overwhelmingly,
00:23:04.18 00:23:06.13 the bacteria that they isolated from these lesions
00:23:06.14 00:23:09.23 were resistant to the drugs that the patient was being treated with,
00:23:09.24 00:23:12.06 which explains why the bacteria were not cleared.
00:23:12.07 00:23:14.11 They were simply resistant mutants.
00:23:14.12 00:23:18.03 And the colonies of these bacteria appeared in what is considered
00:23:18.04 00:23:20.12 a normal timecourse for the tubercle bacillus,
00:23:20.13 00:23:24.14 which is 1-2 months. As I indicated before, it's a very slow growing bacterium.
00:23:24.15 00:23:29.20 So, 1-2 months is about what you expect for actively metabolizing bacteria to form a colony.
00:23:29.21 00:23:33.18 In contrast, when they cultured the bacteria from these closed lesions, they found
00:23:33.19 00:23:36.04 that in most cases, they were still drug sensitive,
00:23:36.05 00:23:41.14 (even though they were coming from individuals who were undergoing active chemotherapy
00:23:41.15 00:23:44.09 with drugs that these bacteria were susceptible to),
00:23:44.10 00:23:47.28 and that the outgrowth of these bacteria was delayed quite substantially.
00:23:47.29 00:23:51.27 These observations suggested that the bacteria in these lesions
00:23:51.28 00:23:54.14 might have been in a state of dormancy
00:23:54.15 00:23:58.00 which protected them from the killing action of these drugs.
00:23:58.01 00:24:00.08 And, if it's true, as Mitchison proposed,
00:24:00.09 00:24:04.00 that these dormant bacteria are always present as a subpopulation,
00:24:04.01 00:24:07.14 even in individuals with active tuberculosis,
00:24:07.15 00:24:11.12 this could explain why it's so difficult to treat this disease.
00:24:11.13 00:24:16.07 And, it would suggest that for the development of new and more effective drugs against TB,
00:24:16.08 00:24:22.00 we should perhaps be focusing on targeting processes that are required for survival
00:24:22.01 00:24:26.26 of non-replicating cells, rather than active cell growth and division, as we have in the past --
00:24:26.27 00:24:29.11 a completely different set of molecular targets.
00:24:29.12 00:24:34.24 Now, given the obvious need that I've described to you, you might think
00:24:34.25 00:24:39.26 that tuberculosis is a major priority for the pharmaceutical industry
00:24:39.27 00:24:42.04 and for governments, even in the rich world.
00:24:42.05 00:24:46.15 Because the need -- the human need -- is so overwhelming and so obvious.
00:24:46.16 00:24:49.16 That's a reasonable conjecture, but you'd be dead wrong if you thought
00:24:49.17 00:24:53.00 that, and the reason for that is illustrated graphically here,
00:24:53.01 00:24:57.07 where I've plotted data from a recent human development report from the United Nations.
00:24:57.08 00:25:01.14 On the y-axis, I've plotted TB incidents per 100,000 population.
00:25:01.15 00:25:07.26 On the x-axis, the human development index ranking, according to the United Nations.
00:25:07.27 00:25:11.29 So, each dot in this graph represents one country.
00:25:12.00 00:25:16.19 And, human development index indicates human development
00:25:16.20 00:25:18.13 in terms of income, health, and so on.
00:25:18.14 00:25:21.14 So, a low number, according to the human development index
00:25:21.15 00:25:23.28 is a good thing, and a high number is a bad thing.
00:25:23.29 00:25:27.17 And I have not drawn a regression curve through the data here,
00:25:27.18 00:25:29.20 because 1) it's really not necessary.
00:25:29.21 00:25:32.29 As you can see, the correlation is quite clear --
00:25:33.00 00:25:36.05 where human development is low, as out here,
00:25:36.06 00:25:39.21 the rate of tuberculosis in the population is extremely high.
00:25:39.22 00:25:43.01 Where human development is high, the rate of tuberculosis is low.
00:25:43.02 00:25:45.03 Here's the United States, represented by the blue dot
00:25:45.04 00:25:48.08 right here. So, there's a very clear and striking relationship
00:25:48.09 00:25:51.07 between tuberculosis and poverty.
00:25:51.08 00:25:56.19 And, in fact, it's the case that 95% of TB cases
00:25:56.20 00:25:59.16 (people who are ill with TB) occur in developing countries,
00:25:59.17 00:26:02.28 and 99% of TB deaths occur in developing countries.
00:26:02.29 00:26:06.15 So, this is overwhelming a disease of the poor.
00:26:06.16 00:26:09.27 And for this reason, pharmaceutical industries and governments of rich nations
00:26:09.28 00:26:13.18 have traditionally not seen a lot of incentive to invest
00:26:13.19 00:26:16.01 in development of new interventions and strategies
00:26:16.02 00:26:18.19 to prevent and to treat tuberculosis.
00:26:18.20 00:26:21.06 I'm happy to say that this is beginning to change.
00:26:21.07 00:26:24.25 I think in developed countries, we are waking up to the reality that we live
00:26:24.26 00:26:30.18 in a global community that is shrinking all the time with advances in transportation
00:26:30.19 00:26:33.26 and increased migration from country to country,
00:26:33.27 00:26:36.20 such that health in developing nations, in fact,
00:26:36.21 00:26:39.07 does develop our own personal well-being and health.
00:26:39.08 00:26:43.21 And we are seeing, largely led by non-governmental organizations,
00:26:43.22 00:26:45.16 such as the Bill and Melinda Gates Foundation,
00:26:45.17 00:26:50.24 a renewed interest in confronting global inequities and inequalities in health,
00:26:50.25 00:26:53.12 tuberculosis being one of the most striking examples.
00:26:53.13 00:26:55.11 So, I'm quite hopeful for the future that
00:26:55.12 00:26:59.08 there will be increased investment in these areas
00:26:59.09 00:27:02.13 and increased attention paid to the development of new vaccines and drugs
00:27:02.14 00:27:04.02 against problems like tuberculosis.
00:27:04.03 00:27:06.17 As I've tried to indicate, however,
00:27:06.18 00:27:10.19 the scientific challenges for developing new interventions against TB
00:27:10.20 00:27:12.26 are really quite formidable.
00:27:12.27 00:27:15.27 And this, I think, is where those of us who work in academia
00:27:15.28 00:27:17.08 have an important part to play.
00:27:17.09 00:27:21.06 Precisely because there is not as much investment in these areas
00:27:21.07 00:27:22.24 on the part of the pharmaceutical industry,
00:27:22.25 00:27:26.29 as you would see in a field like, say, cancer, cardiovascular diseases, and so on.
00:27:27.00 00:27:28.25 Those of us who work in academia, and who are
00:27:28.26 00:27:31.28 free from the constraints of profit motives and the like
00:27:31.29 00:27:33.28 have a special role to play.
00:27:33.29 00:27:38.00 We can drive a lot of the basic research and development in areas
00:27:38.01 00:27:41.17 like target development, antigen identification, and so on
00:27:41.18 00:27:45.02 that the pharmaceutical industry, then, can use,
00:27:45.03 00:27:48.28 if we can hand them, sort of, a ready-made package of information,
00:27:48.29 00:27:51.25 for the development of new interventions.
00:27:51.26 00:27:53.10 So, I predict that what we're going to see in the coming years
00:27:53.11 00:27:55.22 is a growing number of public-private partnerships,
00:27:55.23 00:27:59.05 where the basic research and development will be done
00:27:59.06 00:28:02.25 in academia, and the actual product development and clinical trials
00:28:02.26 00:28:04.22 will be done in the pharmaceutical industry.
00:28:04.23 00:28:09.10 An alliance between the public and private sectors
00:28:09.11 00:28:17.23 in service of global health in the world's most under-served populations.

Part 3: Phenotypic Heterogeneity and Antibiotic Tolerance

00:00:00.00 In the last talk, we discussed some of the problems
00:00:07.25 confronting development of new tools to prevent and to treat tuberculosis.
00:00:12.25 In this next segment, I really want to focus on the problem of antibiotic therapy
00:00:16.22 and delve into this issue of why the antibiotics we currently use to treat tuberculosis
00:00:21.09 are not more effective than they are.
00:00:23.28 I'll remind you that we discussed in the last talk
00:00:27.24 the "Mitchison hypothesis" that's been with us for some time, proposing
00:00:31.10 that TB chemotherapy is problematic because the bacteria
00:00:35.15 inhabit different tissue compartments
00:00:38.06 and therefore adopt different physiologies that make them
00:00:41.13 refractory to one or more anti-TB drugs.
00:00:44.17 This idea, which has been around for quite a while now
00:00:48.25 might explain why the kinetics of killing with the antibiotics that we use
00:00:53.02 against tuberculosis are, in all cases, biphasic or multi-phasic.
00:00:58.08 In this case, this is an experiment in which chronically infected mice
00:01:01.24 were treated with one of our front line drugs, isoniazid
00:01:05.14 via mono-therapy. What you can see,
00:01:08.13 looking on the y-axis at the number of viable bacteria
00:01:13.03 still within the lungs (that's a log scale)
00:01:15.02 versus the days on chemotherapy
00:01:17.29 is that killing of bacteria by isoniazid (or INH as I'll call it)
00:01:21.20 is not a log-linear process.
00:01:23.24 So, initially, there's a period of rapid cell killing --
00:01:27.17 rapid being a relative term here.
00:01:29.10 It's actually very slow compared to the action of most antibiotics against most bacteria.
00:01:33.21 But, the remarkable thing is that after about 4 weeks of chemotherapy,
00:01:37.15 the killing essentially stops and the bacterial population stabilizes at a level about
00:01:42.05 0.1 to 1% of the initial population.
00:01:45.18 So there's a very high frequency of survivors in the face of chemotherapy.
00:01:49.15 We call these the "persisters."
00:01:52.00 We call them persisters to distinguish them from
00:01:55.02 the resisters, which are bacteria that have evolved through mutation
00:01:58.08 resistance to the antibiotics
00:02:00.16 Because, in fact, this is a phenotypic, metastable state
00:02:02.29 that is not inherited by the progeny of these cells
00:02:05.27 So, if you were to take these bacteria out of the tissues,
00:02:08.20 re-grow them in culture, re-infect animals,
00:02:11.12 and treat them again, you would get exactly the same curve back.
00:02:13.28 They are not stably, heritably resistant mutants.
00:02:17.14 This is a purely phenotypic phenomenon that we know very little about.
00:02:20.26 But, clearly, if we could convert this biphasic curve to a simple
00:02:25.19 log-linear kill curve, that would be a major advance
00:02:28.25 towards accelerating the clearance of infection
00:02:31.05 with chemotherapy.
00:02:33.18 Now, we can postulate that the reason for these biphasic kinetics is due
00:02:39.13 to the existence, as Mitchison proposed, of a dormant
00:02:42.05 subpopulation of bacteria that is killed with slow kinetics
00:02:45.21 compared to the bulk bacterial population.
00:02:48.06 If we perform a simple mathematical model in silico for this process
00:02:55.05 on the basis of differential killing dynamics
00:02:58.27 of a "fast" population (that is a population that's killed with rapid kinetics
00:03:03.13 that represents the bulk of the bacteria -- maybe 99.9% of the population)
00:03:07.24 versus a "slow" population (the "persisters" that's killed with very slow kinetics),
00:03:12.21 and then calculate the composite of these two curves, as shown by the blue line here,
00:03:17.03 what we get is an almost perfect match to the experimental data.
00:03:20.13 Now, of course, the fact that a model matches the experimental data
00:03:23.27 does not prove that it is true, but it at least
00:03:26.00 says that the idea is plausible.
00:03:29.01 So, according to the Mitchison hypothesis then,
00:03:32.15 the persistence of TB in the face of chemotherapy
00:03:36.20 is due to a set of environmental inputs leading to a phenotypic switch
00:03:41.26 (in this case a state of dormancy in a subset of organisms),
00:03:45.08 which causes them to become these drug-tolerant persisters.
00:03:49.02 So, there's a simple flow of information from the environment
00:03:52.00 to the physiology of the bacteria,
00:03:54.08 and this is what's responsible for the recalcitrance
00:03:56.20 of these bacteria to killing with antibiotics.
00:03:59.29 It's a very attractive idea, and as I indicated in the last talk,
00:04:05.07 there's actually some clinical evidence in favor of this idea,
00:04:07.23 but there are some awkward bits about this model as well.
00:04:11.10 One of the most obvious and glaring being
00:04:13.28 that the persister phenomenon is not peculiar to the in vivo environment.
00:04:17.13 In fact, even if we grow a clonal population of bacteria in an axenic culture
00:04:23.18 -- just in a tissue culture flask in vitro --
00:04:26.11 and then subject them to a drug like isoniazid,
00:04:29.01 we see that the killing curve is at least biphasic, if not multiphasic,
00:04:33.11 just as it is in vivo. Now, the kinetics of killing
00:04:35.15 are quite different in vitro, as compared to in vivo.
00:04:38.20 I don't want to say that they're the same.
00:04:40.02 But, the biphasic shape of the kill curve in vitro is essentially the same as what we see in vivo,
00:04:46.29 suggesting that the persister phenomenon, although it may be
00:04:50.13 modulated by the tissue environment, is not absolutely dependent,
00:04:54.05 perhaps, on the tissue environment.
00:04:56.03 In a way, this is good news, because it means that we can study this
00:04:59.22 phenomenon in a test tube, rather than in an animal,
00:05:03.13 and that, of course, greatly accelerates the research on understanding
00:05:06.21 this persister phenomenon.
00:05:08.01 So, the fact that there are no obvious environmental inputs
00:05:13.22 in this axenic culture for generating a heterogeneity of populations
00:05:19.00 suggests that, instead, purely stochastic factors...
00:05:22.08 random fluctuation from cell to cell in gene expression, inter-division time, and so on
00:05:28.17 could play an important role in generating the persister phenomenon.
00:05:32.17 So, we would like to postulate that stochastic processes may, in fact, underlie
00:05:37.17 the fundamental phenotypic switch to the drug-tolerant persister fate.
00:05:41.05 Now, this does not, of course, preclude a role for environmental factors.
00:05:44.24 In fact, environmental inputs could have a modulatory effect --
00:05:49.20 a stimulatory effect on the rate of switching -- even if the underlying switch mechanism
00:05:55.06 is essentially stochastic. And there are excellent examples of this from model organisms
00:06:00.03 that have been well studied.
00:06:01.14 Probably the best studied of which is the switch
00:06:03.29 that occurs at the lac operon in Escherichia coli.
00:06:06.28 So, I'm sure all of you have studied the lac operon at one time or another
00:06:11.05 in your undergraduate studies. I'll just remind you of
00:06:14.20 the components of the system: the lac operon encodes a beta-lactamase,
00:06:19.11 encoded by the LacZ gene, a permease (LacY)
00:06:23.11 that brings lactose into the cell,
00:06:26.22 and the beta-galactosidase can then act on that lactose that's brought into the cell
00:06:32.17 to convert it in part to allolactose, which is the true
00:06:35.05 inducer of the system.
00:06:37.04 Allolactose then binds to the LacI repressor
00:06:39.23 and causes it to fall off the DNA, thereby derepressing
00:06:44.05 expression of the lac operon.
00:06:46.09 So, again, the sequence of events is lactose enters
00:06:48.27 through the LacY permease, that lactose is then converted to allolactose
00:06:53.25 by beta-galactosidase, that then causes derepression by binding to the LacI repressor
00:06:59.14 and causing it to fall off the DNA, leading to derepression of the operon.
00:07:04.13 So, it's a very simple, lovely architecture, but it obviously can't work
00:07:08.13 as I've drawn it here, because in an uninduced cell
00:07:12.18 where the lac operon is repressed, there's no LacY.
00:07:15.23 There's no permease to bring the lactose into the cell.
00:07:18.12 And even if the lactose could enter through, say, simple diffusion,
00:07:22.04 there's no beta-galactosidase to make the true inducer, which is allolactose.
00:07:27.14 The reason why this system works is that, in fact,
00:07:30.02 it's not perfect. It's leaky. In a subset of cells, there's always some
00:07:34.02 expression of the lac operon, and this is simply because repression
00:07:39.10 depends on the continued interaction of the LacI protein and the LacO operator
00:07:44.20 on the DNA. And like any intermolecular interaction,
00:07:48.14 it breathes a bit. Sometimes the LacI just falls off randomly,
00:07:51.26 and when this happens, there's some transcription of the lac operon,
00:07:55.19 leading to some production of LacZ and LacY.
00:07:58.06 And that subset of cells in which these events occurred are now primed
00:08:02.20 to respond if lactose happens to be added to the culture at this time.
00:08:06.28 So there's always a subset of cells, even in a clonally homogeneous population --
00:08:12.01 a genetically homogeneous population that are primed, if you like, purely by accident
00:08:17.24 to respond to lactose if it should happen to enter the environment.
00:08:21.23 The key, as I said, is LacI and the interaction of LacI with its operator.
00:08:29.28 Now, it's important to note that, although this is essentially a stochastic switch
00:08:34.08 that occurs when LacI happens to fall off the operator,
00:08:38.13 the rate at which actual induction can occur and switching to the Lac on state
00:08:42.28 is modulated by environmental factors, most notably by the presence
00:08:46.26 of other sugars like glucose in the environment.
00:08:49.22 Glucose is a preferred sugar over lactose, so when that is present
00:08:52.17 in the environment, even with addition of lactose
00:08:55.04 the rate of switching is modulated because glucose blocks
00:08:58.28 the uptake of lactose, and it also represses the CRP transcription factor,
00:09:05.14 which is required for expression of the lac operon.
00:09:08.10 So, here we have an example of what is fundamentally a stochastic switch
00:09:12.13 to a state that's primed for responsiveness to lactose,
00:09:16.24 but where that phenotypic switch can be modulated
00:09:20.03 by the presence of other environmental factors like glucose.
00:09:23.10 What this kind of architecture then leads to
00:09:27.05 is a bi-stability of the response of the bacterial population to an inducer like lactose.
00:09:34.05 When lactose is added those cells that are primed to respond,
00:09:37.12 respond completely and rapidly,
00:09:39.21 and they induce the lac operon to its full state.
00:09:43.02 In those cells in which LacZ and LacY are still absent when lactose is added,
00:09:49.13 they're essentially blind to it because they can't take up the lactose,
00:09:52.07 and they can't convert it to allolactose,
00:09:53.22 and so they stay in the uninduced state until, by random chance,
00:09:58.13 LacI falls off of its operator and allows those cells to become primed.
00:10:01.28 Because this is essentially a positive feedback loop, once cells have entered a state
00:10:06.11 in which they're capable of being induced, they are very rapidly induced to the full state.
00:10:10.13 So, in a situation like this that is shown graphically here,
00:10:13.19 if you look at a population of cells after addition of an inducer like lactose,
00:10:18.16 where the cells have been engineered to express GFP once the lac operon is turned on
00:10:22.22 (that's the green fluorescent protein, a marker that can be used on a single cell basis),
00:10:28.03 what you find is that the population of induced cells consists of two discrete subpopulations.
00:10:34.10 These white cells are cells that were not primed to respond to lactose
00:10:38.22 at the time this inducer was added, and therefore they have not.
00:10:42.01 So, the lac operon is still off in these cells.
00:10:44.15 In contrast, these green cells were primed to respond, and they have
00:10:47.29 by completely inducing the lac operon.
00:10:50.23 There's no intermediate state.
00:10:53.08 So, in fact, cells are on or they are off -- they are not partially on or off.
00:10:58.20 So this is a bistable phenomenon,
00:11:01.15 rather reminiscent perhaps of the persister switch leading to a subpopulation of cells
00:11:05.28 that are refractory to chemotherapy,
00:11:07.22 although in that case, we do not understand the underlying mechanisms.
00:11:10.27 And this bi-stability of the population that I've shown you,
00:11:15.23 where simultaneously in a clonal population, two very different cell types can coexist,
00:11:20.07 I think underscores the inadequacy of the traditional approaches
00:11:24.04 that many of us have been taking to study phenomena in microbiology,
00:11:28.26 cell biology, and so on,
00:11:30.18 based on averaging the behavior of cells, using what I call batch culture methods.
00:11:35.04 So that would typically be to take a population of cells growing
00:11:39.25 in a shaker flask, and then to pool the cells, to grind them up,
00:11:43.12 extract (say) an enzyme, and measure enzyme activity,
00:11:46.14 or the amount of DNA, or this sort of parameter.
00:11:49.00 These approaches are very good for giving you an average value,
00:11:52.29 which I'll call alpha, of some parameter in the population.
00:11:56.07 But, these approaches reveal nothing whatsoever about the underlying population structure,
00:12:00.25 which can be quite heterogeneous, and which, for the same value of alpha
00:12:04.25 can represent very different underlying population structures.
00:12:08.02 And that's illustrated in a simplified form graphically here,
00:12:11.11 where I have schematized on the y-axis number of cells
00:12:15.22 occupying a particular phenotypic state
00:12:18.04 versus on the x-axis some particular cell parameter.
00:12:22.16 That could be expression of a particular enzyme, a particular cell growth time,
00:12:27.26 interdivision time... any parameter than can be quantified.
00:12:30.19 So, my point here is that in these four different populations that I've illustrated,
00:12:34.25 the average value (alpha) is exactly the same for these four populations, and yet,
00:12:39.27 clearly, these are very different populations of cells.
00:12:43.23 So, this average value alpha could represent a simple, narrow Gaussian distribution,
00:12:49.11 as illustrated in this panel, or a broad Gaussian distribution, as indicated here.
00:12:55.05 It could even, at one extreme, represent a bimodal distribution of cells,
00:13:00.10 in which the average value, in fact, is not represented by any individual cell
00:13:04.29 within the population whatsoever.
00:13:06.14 So that average value is extremely misleading vis-a-vis the actual phenotype
00:13:10.19 of individual cells within the population.
00:13:13.24 This bimodal or multimodal distribution can also occur with the
00:13:19.17 interesting phenotype, if you like, being represented by a small minority of the population.
00:13:23.29 This would be representative, for example, of the persister phenomenon,
00:13:27.04 in which the cells in this small bump here are the ones which we're actually interested in,
00:13:32.22 and yet, if we were to take a batch culture approach to studying this phenomenon,
00:13:35.25 this is the signal we would get --
00:13:38.12 the signal from the cells that we're not interested in --
00:13:40.17 and we would learn nothing about the behavior of this subpopulation of cells,
00:13:44.26 which is in fact the target that we want to study.
00:13:46.23 So, we and many others now are taking new approaches to
00:13:51.15 the study of bacterial behavior at the single cell level --
00:13:55.04 that is, approaches that will allow us to capture the full complexity of cellular behavior
00:14:00.04 and the variation between cells and behavior of particular characteristics.
00:14:06.18 One of the approaches in my lab that we're taking is the use a combination
00:14:10.25 of microfluidic culture of bacteria
00:14:12.23 and time-lapse microscopy to study the behavior of large numbers of
00:14:18.00 individual cells before, during, and after imposition of a stress,
00:14:22.08 such as antibiotics.
00:14:23.22 We do this, as I said, using microfluidic devices
00:14:26.27 that we fashion using soft lithography,
00:14:29.23 and which are illustrated schematically here.
00:14:32.24 They consist, in this case, for example,
00:14:36.09 of a simple glass coverslip on which the bacteria are seeded,
00:14:39.09 which is the overlaid with a semipermeable membrane,
00:14:42.13 so that the bacteria are growing right underneath the membrane,
00:14:45.18 sandwiched between the membrane and the coverslip,
00:14:47.27 We this overlay this with a block of material that's been fashioned by soft lithography
00:14:53.17 with an input port for media to flow in, an output port for media to flow out,
00:14:59.00 and flow channels cut through it for the media to flow across the surface of the membrane.
00:15:03.21 So this is essentially a sandwich, through which we can pump media
00:15:08.08 that feeds the bacteria. They grow happily on the coverslip,
00:15:11.22 fed by the material diffusing through this membrane.
00:15:14.12 And, using a fully motorized and computer-controlled microscope
00:15:18.16 that can visit as many as a hundred different points on a coverslip
00:15:22.17 and remember where it's been
00:15:23.26 (so it can keep coming back to the same points time and time again),
00:15:26.24 we can collect, over weeks of observation, even months of observation,
00:15:31.07 in fact, the detailed behavior of a bacterial population at single-cell resolution.
00:15:38.09 So, first and most importantly, we find that
00:15:41.27 when we grow bacteria in these microfluidic devices
00:15:45.01 and then expose them to a drug like isoniazid,
00:15:47.09 the kill curve that we see is clearly biphasic.
00:15:50.02 There's a period of rapid killing, followed by a period in which killing essentially stops.
00:15:53.27 So, the phenomenon that we're trying to study,
00:15:56.28 this biphasic killing response to the drug isoniazid
00:16:01.18 can readily be studied at single-cell resolution using these microfluidic devices.
00:16:05.12 So the beauty of this approach is that it allows us to study the genealogy, if you like,
00:16:11.01 of individual cells. We can follow them by recording their behavior
00:16:14.12 through multiple cell divisions
00:16:15.17 before we impose a stress, such as the drug isoniazid.
00:16:19.17 We then expose them to the drug for whatever period of time we desire,
00:16:24.29 which of course leads to the death of most of the cells
00:16:27.23 with the exception of the rare persisters.
00:16:30.05 But then we can readily identify these persisters that failed to lyse
00:16:34.13 and then recover growth after we wash out the drug,
00:16:36.25 and now we have a complete record through time of
00:16:40.11 the behavior of these cells and their ancestors.
00:16:44.07 So, we can now ask, is there anything that we can identify that's different
00:16:47.28 about these cells, either before, during, or after the exposure of the drug
00:16:52.08 that causes them to be persisters when 99.9% of their siblings are not.
00:16:57.15 So, using this type of approach, it was possible
00:17:00.00 for the first time to test a hypothesis that was proposed many years ago
00:17:03.26 (in fact, more than 60 years ago now)
00:17:06.01 by Joseph Bigger in a classic paper, where he looked at the response
00:17:10.17 of staphylococcus to treatment with penicillin
00:17:14.19 in a test tube, and for the first time observed this
00:17:18.16 persister phenomenon and coined the term persister.
00:17:21.16 He suggested -- it was purely hypothetical --
00:17:24.20 it was quite a brilliant insight, in fact,
00:17:26.12 that the persisters might be insensitive to the penicillin
00:17:30.01 because they might be temporarily in a dormant or non-dividing phase,
00:17:34.20 similar to bacteria that have exited the cell cycle due to, say, nutrient starvation.
00:17:38.26 And, this would make sense, because penicillin kills bacteria only when
00:17:42.19 they are undergoing active cell growth and division.
00:17:45.01 So that was Bigger's hypothesis back in the 1940s.
00:17:47.25 It had to wait 60 years, in fact, for experimental validation using newly developed tools --
00:17:53.22 for time-lapse microscopy and microfluidics.
00:17:57.03 And this was reported in a paper published in the year 2004
00:18:01.03 by Balaban and colleagues
00:18:02.20 (there's the reference for those who are interested),
00:18:04.26 in which they used microfluidic cultivation of Escherichia coli
00:18:08.17 to study, at single-cell resolution,
00:18:10.28 the response of these bacteria to the drug ampicillin, a beta-lactam.
00:18:15.10 So, the persister in this case is circled here in red.
00:18:18.15 So, I'll ask you, when the movie plays,
00:18:20.11 to focus on that cell.
00:18:23.26 As you can see, the cells around it are actively growing and dividing,
00:18:27.08 but that pair of cells is not growing.
00:18:30.06 The ampicillin is added, and the cells die,
00:18:32.24 but when it's washed out, these cells are capable of resuming growth.
00:18:37.08 And, in fact, if you follow them long enough, they eventually resume rapid growth
00:18:41.02 that is indistinguishable from the growth rate of their siblings
00:18:46.09 who died in response to the drug.
00:18:47.22 So, using this type of approach, Balaban and colleagues
00:18:51.06 were able to demonstrate that Bigger's hypothesis was, in fact, correct --
00:18:54.17 that the persisters comprise a pre-existing [subpopulation], and that's the important point.
00:18:59.05 This is not an adaptation to the drug. It's a pre-existing subpopulation of
00:19:02.25 bacteria that are temporarily in a slow or non-dividing state,
00:19:07.18 and are therefore protected from the killing action of this drug.
00:19:10.23 The actual mechanisms that are responsible for generating these persisters
00:19:16.14 are likely to be stochastic, but they have not been identified.
00:19:20.12 Likewise, for reactivation of these persisters after washout of the drug.
00:19:24.02 So that's a major challenge for the future is to try to understand the molecular processes
00:19:27.25 that allow these cells to become persisters.
00:19:30.15 But, at least at the cellular level, in the case of E. coli,
00:19:33.13 it seems to be very clear that Bigger, in fact, was correct.
00:19:37.08 Persisters are a pre-existing subpopulation of non-growing cells.
00:19:41.26 So, we began to collaborate with this group to study,
00:19:46.17 using these same approaches, the persister phenomenon in Mycobacteria,
00:19:50.22 focusing at first on the drug, isoniazid.
00:19:52.28 We had a rather strong bias at the outset that the persister phenomenon in Mycobacteria
00:19:58.05 was going to be very similar for the reasons I've already described,
00:20:01.06 and which Mitchison proposed -- mainly that
00:20:03.21 drugs like isoniazid kill bacteria only when they are actively growing and dividing,
00:20:08.14 and therefore, it would make sense if the persisters would be these cells,
00:20:12.01 that are in a temporarily non-growing state.
00:20:14.18 So, that was our assumption when we began these experiments...
00:20:17.19 or our hypothesis, rather,
00:20:19.23 when we began these experiments.
00:20:20.28 But, the beauty of doing science is that our ideas about the world
00:20:24.26 are never as interesting and complex as the world itself.
00:20:27.12 And so, we found, in fact, very quickly
00:20:30.04 that this idea was wrong, and that what had been discovered in E. coli, in fact,
00:20:33.09 apparently does not apply to Mycobacteria.
00:20:36.15 That's shown here -- a typical experiment in which we have grown
00:20:41.19 Mycobacteria, in this case in a microfluidic device.
00:20:44.21 The bacteria have been engineered to express GFP, as in the
00:20:47.11 previous E. coli experiments.
00:20:49.08 That's why they glow green.
00:20:50.18 So, what you're going to be looking at is bacteria grown in 7H9
00:20:54.12 initially, which is just basic broth with no drug added.
00:20:57.25 We'll then switch to isoniazid-containing medium, and that will be shown by
00:21:01.07 the INH appearing in the corner there
00:21:03.02 and look at the response of cells to the drug.
00:21:06.03 We will then wash out the drug.
00:21:08.01 7H9 will reappear.
00:21:09.08 And we will look at the regrowth of the rare persisters that survived exposure to the drug.
00:21:14.28 So, if we can roll the film, you see
00:21:17.20 cells are actively growing and dividing in the absence of drug.
00:21:21.18 As soon as we add the drug, they cease growth
00:21:23.20 within minutes of addition of the drug.
00:21:25.26 It's really an extremely rapid response.
00:21:28.01 And then the cells undergo lysis.
00:21:30.04 So, they stop growing very rapidly, then there's a delay of about a day or so
00:21:36.13 where cells began undergoing lysis.
00:21:39.07 This particular experiment involved recording the bacteria for a period of about 1 week.
00:21:43.19 So, this is a very compressed time course here.
00:21:46.25 Now, remarkably, after prolonged exposure to the drug,
00:21:51.13 the subset of surviving cells starts to grow and divide again --
00:21:53.13 at slower rates than in the absence of drug --
00:21:55.07 but, very appreciable rates.
00:21:57.05 Some of the progeny of those divisions die, some survive.
00:21:59.23 Those that survive long enough
00:22:01.21 can then, when the drug is washed out,
00:22:03.16 reactivate growth, and in fact, they do so immediately,
00:22:07.03 and they resume very rapid cell growth
00:22:11.04 within just a few minutes of washout of the drug.
00:22:14.12 So, the response of the bacteria, in this case, is very different
00:22:17.10 than what was seen by Balaban and colleagues in the case of E. coli
00:22:20.09 responding to the beta-lactam, ampicillin.
00:22:23.07 Now, using this type of approach, we can, as I said,
00:22:27.07 study the responsive bacteria, at single-cell resolution.
00:22:30.17 And this has made it very clear to us that the behavior of these cells
00:22:34.13 is even more heterogeneous than we might have thought.
00:22:36.25 One example of this is in the movie shown here
00:22:40.19 where you're going to see a typical cell,
00:22:43.19 but one that we think is behaving in an extremely interesting way
00:22:47.19 that we think is likely to be very informative about how the drug
00:22:50.18 is actually acting on these cells.
00:22:52.10 So, we can roll the film.
00:22:56.01 What we see again is that cells are growing and dividing in 7H9.
00:22:58.08 We add the drug, they arrest
00:23:00.10 very rapidly. They swell a bit, and then
00:23:02.11 they begin to undergo lysis
00:23:03.27 and disappear from the frame.
00:23:05.27 But, focus on the cell over here.
00:23:08.08 It's going to do something rather extraordinary.
00:23:10.20 It arrests, but then, in the continued presence of the drug,
00:23:14.03 it eventually resumes very rapid cell growth.
00:23:17.08 In fact, it's growing as fast now, as if there's no drug present at all.
00:23:20.19 But clearly, this was not a drug-resistant cell
00:23:22.28 because it did eventually die.
00:23:24.22 In fact, it appears that division is inhibited in the cell
00:23:27.14 even though cell growth is not.
00:23:29.13 So, my point in showing this sort of rare occurrence
00:23:34.07 is that there's an enormous amount of heterogeneity in the behavior
00:23:37.17 of individual cells within a bacterial population
00:23:40.21 that simply isn't captured by the kinds of batch culture methods we typically use
00:23:44.12 to study bacterial behavior.
00:23:46.17 My personal feeling is that the future of microbiology is
00:23:49.12 going to be to focus on the individuality of bacterial behavior,
00:23:53.26 rather than to stick with the averaging methods we have used in the past.
00:23:57.11 I think this is going to be a particularly important approach for us to take
00:24:01.15 when we want to study the behavior of important subpopulations
00:24:04.21 like the persisters.
00:24:06.15 So, there are a number of points here that I want to make,
00:24:10.13 in terms of comparing and contrasting between E. coli.
00:24:13.20 The first and most obvious one
00:24:15.07 is that, in contrast to E. coli, where cells
00:24:18.02 that turn out to be persisters after drug exposure
00:24:20.23 are already in a very slow or non-dividing state
00:24:24.17 prior to the addition of drug, this simply isn't true in Mycobacteria.
00:24:27.19 So, in fact, if we quantify the growth rate prior to drug addition
00:24:31.18 of cells that die, over here, versus cells that survive, over here,
00:24:35.20 in a particular experiment
00:24:37.05 (these are data from one experiment),
00:24:38.15 we find that there's absolutely no difference in the average growth rate
00:24:41.21 or in the spread of growth rates
00:24:43.03 between these two populations.
00:24:44.17 So, Bigger's hypothesis apparently is true for E. coli
00:24:49.21 responding to ampicillin, but it is not true for Mycobacteria responding to isoniazid.
00:24:54.05 We do not, at this point, know whether this is a difference
00:24:57.10 in the species that we're looking at or a difference in the drugs that we're looking at.
00:25:00.21 This remains a challenge for the future.
00:25:02.26 So, that's the first take home lesson from this sort of compare and contrast...
00:25:08.20 that in E. coli, persisters are rare, slow-growing cells
00:25:11.19 and they exist prior to antibiotic exposure,
00:25:14.26 whereas in Mycobacterium, the persisters are growing and dividing
00:25:18.28 at completely normal rates prior to antibiotic exposure,
00:25:21.22 so it's a very different phenomenon.
00:25:23.16 The second point of contrast is that in E. coli,
00:25:26.10 the non-persisters (those cells that actually die in the presence of the drug)
00:25:30.02 continue to grow at nearly normal rates after exposure to the antibiotic,
00:25:33.29 so the antibiotic doesn't stop them at all.
00:25:35.23 They keep on elongating until they lyse.
00:25:37.28 And that cell lysis occurs actually quite rapidly after exposure
00:25:41.18 to the drug.
00:25:43.02 In contrast, in Mycobacteria
00:25:44.27 the non-persisters -- that is the bulk of the population --
00:25:48.04 arrest growth really within minutes after exposure to this drug, isoniazid.
00:25:53.08 So, it's a very rapid response to the drug.
00:25:55.06 But then, cell lysis begins only after quite a prolonged delay of about 24 hours or so
00:26:00.11 after initial exposure to the drug.
00:26:02.23 So, it's a completely different, really opposite response
00:26:05.10 to drug exposure.
00:26:07.14 In E. coli, after washout of the drug in a microfluidic device...
00:26:12.07 In the continued presence of the drug,
00:26:17.14 the persisters do not resume growth
00:26:19.01 as long as we continue to pump it through the microfluidic device.
00:26:21.20 In fact, they do not resume growth until after washout of the drug.
00:26:25.22 Whereas in Mycobacterium, the persisters, as you saw,
00:26:29.07 resume growth and division after prolonged (that is, many days)
00:26:32.10 of antibiotic exposure, although some of the progeny
00:26:35.20 of those divisions, in fact, do not survive.
00:26:37.26 And when we washout the drug in a microfluidic device,
00:26:41.19 we find that E. coli persisters reactivate with slow and stochastic kinetics
00:26:46.27 after antibiotic washout.
00:26:48.13 So, it's almost as though it's a process of radioactive decay determining
00:26:53.01 whether a cell can reactive at a given time or not
00:26:56.03 after drug washout.
00:26:57.06 Whereas, in Mycobacteria, those persisters reactivate almost immediately
00:27:01.01 or very rapidly after drug washout and with uniform kinetics.
00:27:05.04 So, it seems very clear that the drug is actually keeping in
00:27:08.01 this non-dividing state. As soon as that pressure is removed,
00:27:11.16 the bacteria start to grow again.
00:27:13.08

Part 4: Targeting M. tuberculosis Carbon Metabolism In Vivo

00:00:04.09 In the last lecture, I told you about some of our
00:00:06.07 studies to try to understand the persister phenomenon in Mycobacterium tuberculosis.
00:00:11.01 That is an attempt to understand why the drugs that we currently use
00:00:13.24 to treat this disease are not more effective than they are.
00:00:16.19 In this last lecture, I want to tell you about some of the work we're doing in my lab
00:00:20.16 to try to identify new and hopefully better targets
00:00:23.23 for development of new drugs against tuberculosis
00:00:26.05 that we hope will allow us to kill the bacteria more rapidly and overcome
00:00:30.18 this persistence problem.
00:00:31.20 Now, our segue into this area has been to study the basic metabolism
00:00:36.22 of M. tuberculosis when it's living in the lung,
00:00:39.08 as opposed to living in a test tube, ex vivo.
00:00:41.06 And our segue into this area came from some really beautiful biochemical studies
00:00:45.13 that were published more than half a century ago now
00:00:48.23 and had since been forgotten, but were rediscovered by us a few years ago.
00:00:53.03 And here's an important lesson for the students who might be watching this lecture...
00:00:56.02 Pay attention to the pre-Pubmed literature because there are real gems there,
00:01:00.26 particularly in fields like tuberculosis and malaria that have
00:01:03.14 undergone periods of dormancy among the research community.
00:01:07.02 So, in these studies, which were essentially biochemical studies,
00:01:11.13 the authors focused on characterizing the physiology of bacteria
00:01:16.21 that were grown in the in vivo environment.
00:01:19.04 And they did this by using as their starting material
00:01:21.28 the lungs of infected mice -- that is mice that harbored M. tuberculosis --
00:01:26.14 which was then homogenized and fractionated
00:01:29.26 using the rather crude methods that they had available at the time
00:01:33.04 (density-gradient centrifugation, in this case)
00:01:35.28 to yield a population of more-or-less pure, in vivo-grown mycobacteria
00:01:40.08 which they then subjected to a battery of biochemical and physical tests.
00:01:44.21 And these studies, which were published over the period of about a decade,
00:01:49.24 culminated in the conclusion that when the bacteria are grown
00:01:53.06 in their natural environment (that is, the mammalian lung),
00:01:55.27 they're very, very different physiologically than they are
00:01:59.24 when the same bacteria are grown under the artificial conditions
00:02:02.21 that we use to study them ex vivo.
00:02:04.29 So, these studies emphasize the importance of studying
00:02:08.07 the bacteria in their natural environment
00:02:11.14 if you want to gain a full understanding of their actual physiology during infection.
00:02:15.12 Now, among the observations that they made,
00:02:18.07 a set of observations that particularly piqued my curiosity
00:02:21.13 is summarized down here.
00:02:23.11 First, they found that the ability to metabolize carbohydrates
00:02:27.18 was essentially shut off in these bacteria.
00:02:30.01 That's an interesting observation because, even to this day,
00:02:33.17 we still continue to cultivate bacteria for in vitro studies
00:02:36.26 on carbohydrate-containing media.
00:02:39.22 But, apparently, this is not what the bacteria are using
00:02:42.05 when they're actually growing in the mammalian lung.
00:02:44.12 Conversely, they found that the ability of these bacteria to metabolize
00:02:48.01 fatty acid substrates was apparently switched on
00:02:51.05 when they were grown in the mammalian lung.
00:02:53.18 Concomitant with that physiologic switch, if you like,
00:02:57.11 or metabolic switch, they found that expression of the enzymes of two pathways --
00:03:01.13 the fatty acid beta-oxidation cycle, and the glyoxylate cycle --
00:03:05.08 were strongly switched on when these bacteria were grown in vivo,
00:03:08.24 as opposed to in vitro.
00:03:10.22 Now, most of the work in my lab that I wanted to tell you about today
00:03:13.25 has been focused on the glyoxylate cycle, although we study
00:03:16.16 both of these pathways in my lab.
00:03:18.27 We elected to target the glyoxylate cycle for analysis first
00:03:22.22 for the simple reason that this pathway was lost in evolution.
00:03:26.22 So, although it's widely found in bacteria and in fungi,
00:03:30.02 it is not found in any vertebrate, including humans.
00:03:33.03 And, although the work we do in our lab is largely basic research,
00:03:37.06 we always try to point it in a direction that might
00:03:39.17 have potential applications in the future.
00:03:41.14 If you're thinking ahead to the possibility of drug development,
00:03:45.05 obviously it's most attractive to target molecules
00:03:48.15 like the enzymes in the glyoxylate cycle,
00:03:51.07 which are present in the bacteria that you want to kill
00:03:54.01 but absent in the host cells that you want to spare.
00:03:57.03 The risk of toxicity is much, much lower thereby.
00:03:59.28 So, to remind you, the fatty acid beta-oxidation cycle is this
00:04:05.01 iterated cycle, shown here, in which long-chain fatty acids
00:04:08.18 are sequentially broken down through a series of steps
00:04:11.24 into these two-carbon acetyl-CoA units.
00:04:15.14 So, two-carbon units at a time, in a sense, are bitten off the fatty acid chain.
00:04:20.00 These acetyl-CoA units can enter metabolism --
00:04:23.10 either the citric acid cycle or the glyoxylate cycle.
00:04:27.24 Now, I'll remind you that the citric acid cycle, in fact, plays two essential roles
00:04:31.24 in metabolism of carbon substrates.
00:04:36.03 The role in generating reducing potential in the form of NADH and FADH,
00:04:40.19 which is responsible for respiration and the generation of ATP,
00:04:45.00 the energy currency of the cell.
00:04:46.28 But also, an important role in biosynthetic pathways
00:04:50.24 because many of these carbon intermediates, such as alpha-keto-glutarate here,
00:04:54.00 are constantly being siphoned off to the biosynthesis of
00:04:58.23 small molecules like amino acids, hemes, nucleotides, and so on.
00:05:05.05 So, there's a constant drain of carbon to these biosynthetic pathways.
00:05:09.03 And this creates a problem for the cell
00:05:11.16 because with each turn of a citric acid cycle, two carbons
00:05:14.28 enter from acetyl-CoA up here, and two carbons leave as CO2 here and here.
00:05:21.15 So, the balance sheet is 0: two carbons in, two carbons out.
00:05:24.12 There's no net gain or loss of carbon.
00:05:26.07 But, as I said, there's this constant drain of carbon to biosynthetic pathways.
00:05:29.16 So, the cell faces an anaplerotic problem, as it's called --
00:05:33.05 a filling up problem.
00:05:34.14 It has to find some way to replenish these carbon intermediates
00:05:38.14 that are drained to biosynthetic pathways,
00:05:40.11 else the Krebs cycle will run down, energy metabolism will fail,
00:05:43.24 and eventually the cell will die.
00:05:45.12 Now, a variety of different pathways are used for carbon anaplerosis.
00:05:49.19 Which pathway is used depends largely on which type of substrate
00:05:53.20 the bacteria, in fact, are metabolizing.
00:05:55.27 Most bacteria, when they're growing on fatty acid substrates
00:05:59.27 utilize a simple subroutine within the citric acid cycle
00:06:04.22 called the glyoxylate cycle -- so this is the pathway
00:06:07.07 that's required for anaplerosis on fatty acid substrates.
00:06:10.17 As I said, it's a very simple pathway, comprising just two
00:06:14.00 dedicated enzymes: isocitrate lyase (or ICL, as I'll call it)
00:06:17.23 and malate synthase (or MLS, as I'll call it).
00:06:20.21 ICL takes the isocitrate from the citric acid cycle,
00:06:25.21 and it splits it to form succinate and glyoxylate,
00:06:29.08 from which the glyoxylate cycle gets its name.
00:06:31.24 Malate synthase then takes that glyoxylate and condenses it
00:06:35.15 with a second molecule of acetyl-CoA
00:06:37.21 to make malate, down there.
00:06:39.12 The important point here is that the carbon-losing steps in the citric acid cycle
00:06:43.22 are bypassed by the glyoxylate cycle,
00:06:46.16 and with each turn of the subroutine, four carbons
00:06:49.15 coming from acetyl-CoA here and acetyl-CoA here
00:06:53.15 are fixed in the cycle in the form of malate.
00:06:57.00 So, it's a way of replenishing carbon intermediates
00:07:00.19 that are lost to biosynthetic pathways.
00:07:02.29 We wanted to know, what's the role of this pathway in tuberculosis metabolism
00:07:08.22 in physiology during an infection.
00:07:10.14 We took a genetic approach to this problem
00:07:12.26 using tools that had only just been developed for genetics in M. tuberculosis,
00:07:17.14 and we asked what would happen if we eliminated
00:07:20.15 the genes encoding isocitrate lyase in M. tuberculosis.
00:07:24.05 Now, this turned out to be a somewhat more complicated undertaking
00:07:27.14 than we had originally envisaged
00:07:28.24 because, interestingly, it turns out that mycobacteria have
00:07:32.10 not just one, but actually two isocitrate lyase enzymes
00:07:36.04 in their genomes -- a rather unique arrangement.
00:07:39.03 We have called them (rather unimaginatively)
00:07:40.23 ICL1 and ICL2.
00:07:42.15 As shown here, encircled, they are not linked on the bacterial chromosome,
00:07:47.03 and these two enzymes are very different in size.
00:07:49.27 ICL1 is quite a bit smaller than ICL2.
00:07:52.12 In terms of evolution, ICL1 looks, for all the world,
00:07:56.04 like a typical eubacterial isocitrate lyase.
00:07:59.06 ICL2 is a real outlier. It looks more like a plant or fungal isocitrate lyase,
00:08:05.22 and we have no idea how it got there.
00:08:07.10 There's an interesting evolutionary problem there that has yet to be understood.
00:08:11.11 So, the question we wanted to address is what happens if we delete these genes
00:08:16.04 from the chromosome of M. tuberculosis.
00:08:18.04 What impact would those mutations have on the ability of these bacteria
00:08:22.03 to metabolize, to grow, and to persist in the lungs of a mammalian host?
00:08:26.16 So, we deleted ICL1. We deleted ICL2.
00:08:30.07 And then, we made a strain in which both of those genes were deleted together.
00:08:33.19 We then took that mutant ICL1 strain for starters
00:08:38.18 and put it into the lungs of mice
00:08:41.01 and monitored the growth and persistence of the bacteria in the lungs of those mice.
00:08:44.29 So, in this graph, I'm plotting on the y-axis the logarithm (base 10)
00:08:49.20 of the colony forming units or viable units in the lungs of mice
00:08:53.01 versus weeks post-infection.
00:08:55.17 What we found is that the ICL1-deficient mutant
00:08:59.06 actually grew quite well during the acute phase of infection --
00:09:02.12 the first couple of weeks of infection --
00:09:04.03 but it was somewhat defective for persistence
00:09:06.29 during the chronic phase of infection after bacterial numbers
00:09:10.04 had more-or-less stabilized
00:09:11.07 under pressure from the host immune response.
00:09:13.28 There's the curve for wild-type bacteria.
00:09:16.17 There's the curve for the ICL1-deficient strain.
00:09:19.05 So it's quite a significant attenuation, specifically during the chronic phase of infection.
00:09:22.28 In contrast, we found that when we knocked out ICL2 alone
00:09:27.13 and deleted that gene,
00:09:28.19 the mutant bacteria showed no defect whatsoever
00:09:32.23 for either acute phase growth or chronic phase persistence
00:09:35.13 in the lungs of mice.
00:09:36.28 As it turns out, that doesn't mean that ICL2 is irrelevant.
00:09:40.10 In fact, it's an extremely important gene for physiology of M. tuberculosis in vivo.
00:09:44.02 But, the function of this gene, as we found, is heavily buffered by the presence of ICL1.
00:09:48.28 And we know that's true because when we deleted both genes simultaneously
00:09:53.08 from the chromosome of the bacteria, something really remarkable happened.
00:09:56.13 We found that the double mutant, lacking both ICL1 and ICL2,
00:10:01.14 as shown by the red symbols here,
00:10:03.11 was completely unable to establish infection,
00:10:05.22 and, in fact, was totally cleared from the lungs within 2 weeks of infection.
00:10:10.04 Just to put this in context of chemotherapy,
00:10:12.23 using the best cocktail of drugs that we currently have for treatment of tuberculosis,
00:10:16.23 if we were to infect mice and immediately put them on chemotherapy,
00:10:20.13 it would require about 8 weeks of continuous chemotherapy
00:10:25.17 (even TB is not that difficult to cure)
00:10:27.06 to, in fact, eliminate the organisms from the lungs of these mice.
00:10:31.26 So, this is a spectacularly rapid clearance of bacteria from the lungs of these animals.
00:10:36.07 Clearly, these enzymes are very important for growth and survival in vivo.
00:10:40.07 Now, we also found that these enzymes are very important for growth and survival
00:10:47.10 in a simplified model of tuberculosis,
00:10:49.27 in which we take macrophages from the mouse, we culture them
00:10:52.23 ex vivo, and we then infect those macrophages with the bacteria.
00:10:57.06 As you will recall from the first lecture in this series,
00:11:00.05 the tubercle bacillus is an intracellular pathogen.
00:11:03.25 It actually makes its living by invading into macrophages and replicating within
00:11:08.04 a vacuolar compartment inside those macrophages.
00:11:11.10 So, we can replicate this process ex vivo.
00:11:13.27 It's a much simpler model and a more manipulable model
00:11:16.24 than the mouse for our experiments.
00:11:18.23 What we find in what are called non-activated, resting macrophages
00:11:23.09 that have not been immunologically activated
00:11:26.10 is that wild-type bacteria in blue grow quite well
00:11:28.27 inside these macrophages. The ICL-deficient strain
00:11:32.01 fails to grow and slowly loses viability.
00:11:34.16 If we now activate those macrophages using the cytokine interferon-gamma,
00:11:40.29 which is thought to be the key immune effector for activation
00:11:44.17 of macrophages during infection,
00:11:46.02 what we find in the case of wild-type bacteria
00:11:48.20 is that this leads to bacteriostasis.
00:11:51.23 It's literally a curtailment of growth inside the macrophages.
00:11:54.26 In contrast, the ICL-deficient bacteria are very rapidly killed in activated macrophages.
00:12:01.09 So, the phenotype that we see in a whole-animal model
00:12:03.29 can, to some extent, be replicated in this simpler, tissue culture model of infection,
00:12:08.19 which makes it much more easy to study the processes involved.
00:12:12.12 Now, we can get even simpler than this,
00:12:15.12 which we need to do for drug development purposes.
00:12:17.23 And we can study the behavior of the bacteria in an in vitro system.
00:12:24.13 The aim of these studies is to try to develop inhibitors
00:12:30.02 against ICL1 and ICL2 that would simultaneously block the function
00:12:34.07 of both enzymes.
00:12:36.11 Now, this essentially becomes a structural problem.
00:12:39.01 I told you before that the two enzymes are very different from each other
00:12:41.29 evolutionarily. The question really is, what about their molecular shape?
00:12:45.24 We now know from studies in many labs
00:12:48.06 that very different primary amino acid sequences can lead to essentially
00:12:51.26 the same molecular shape, and that's what we want
00:12:54.25 to target with drug development.
00:12:56.24 So, going to a purely in vitro system here,
00:12:59.01 we have obtained, through collaboration with the crystallography experts
00:13:03.26 in Jim Sacchettini's lab at Texas A&M
00:13:06.21 the 3-dimensional structure of the isocitrate lyase 1
00:13:10.17 molecule at atomic resolution. Here, I'm showing just the alpha-carbon backbone
00:13:14.23 tracing of this enzyme, shown in blue.
00:13:17.12 ICL2 has been a difficult molecule to crystallize.
00:13:21.03 We do not yet have a definitive structure for ICL2.
00:13:24.18 But, the theoretical structure for ICL2, which we have modeled
00:13:28.11 in silico is shown here in red.
00:13:33.08 So, my point is that, although there are large areas of divergence
00:13:36.01 between the two enzymes -- for example, here in red, there's a domain
00:13:39.05 present in ICL2 that's altogether lacking in ICL1.
00:13:42.27 If we just focus on the active site of the enzyme, shown here,
00:13:46.09 it's very similar. In fact, it overlaps in space almost perfectly between ICL1
00:13:51.08 and ICL2. And this gave us our first confidence that we might, in fact, be able to
00:13:55.22 develop dual-specific inhibitors that would simultaneously
00:13:59.29 block the action of both ICL1 and ICL2,
00:14:02.18 which would be necessary to achieve the maximal
00:14:05.10 therapeutic effect. So, towards that goal,
00:14:09.24 Jim's lab and my lab and the laboratory of our collaborator David Russell at Cornell
00:14:13.19 University entered into a partnership with GlaxoSmithKline,
00:14:17.20 where they have carried out a drug discovery program
00:14:20.08 targeting the isocitrate lyase enzymes of Mycobacterium tuberculosis.
00:14:24.24 Now, as I said, we have mouse and tissue culture models that we can use
00:14:29.02 to study the function of these enzymes.
00:14:32.00 But even the tissue culture model is still a rather cumbersome model.
00:14:34.16 What we'd really like to have
00:14:36.03 is a whole-cell assay that we can use to look at the activity of
00:14:39.28 ICL inhibitors against bacteria grown under appropriate conditions.
00:14:43.13 We've succeeded in developing such a system,
00:14:46.24 and that's shown here.
00:14:47.24 So, in axenic culture, we can grow bacteria,
00:14:50.24 either on carbohydrate substrates, shown in the far panel,
00:14:54.10 or on fatty acid substrates, as shown here,
00:14:56.24 with and without the addition of ICL inhibitors, like this protoinhibitor,
00:15:00.23 3-nitropropionate -- not a drug-like molecule, but a very useful tool compound.
00:15:05.17 When we grow the bacteria on carbohydrate substrates
00:15:08.13 we find that whether we add the inhibitor or not, we get robust growth.
00:15:12.22 The inhibitor has a slight inhibition of growth on the bacteria, but it's quite minimal.
00:15:17.24 In contrast, when we grow the bacteria on fatty acid substrates,
00:15:21.21 as shown in this panel, we see that although the bacteria
00:15:25.16 in the absence of inhibitor grow on fatty acids just fine,
00:15:27.24 their ability to do so is completely blocked by the addition of this inhibitor.
00:15:32.06 In fact, more recently, we found that these bacteria
00:15:34.15 are not only failing to grow, as indicated by this spectrophotometric assay,
00:15:39.05 but they are in fact being killed. And we still don't understand why it is
00:15:43.09 that inhibition of isocitrate lyase leads, in fact, to rapid cell death
00:15:48.16 in the presence of fatty acid substrates,
00:15:51.00 but we think that this probably explains why this bacteria in vivo
00:15:55.00 undergoes such rapid cell death and clearance.
00:15:57.06 Now, we can, in fact, use at the next level of complexity, our macrophage
00:16:03.18 model of infection to look at the activity of compounds against bacteria
00:16:07.26 when they're living inside their host cell.
00:16:10.15 And this is obviously a very important model for us,
00:16:12.19 because a stumbling block for many drugs that we might develop against TB
00:16:15.28 is their ability to reach the bacteria when they are living inside a vacuole
00:16:21.01 inside a macrophage.
00:16:22.17 So, just to remind you, if we compare the behavior of ICL-deficient
00:16:26.14 and wild-type bacteria in non-activated macrophages,
00:16:30.04 we find that wild-type bacteria grow luxuriantly,
00:16:33.05 whereas ICL-deficient bacteria fail to grow.
00:16:36.17 We can phenocopy this behavior of the bacteria
00:16:40.01 by simply growing wild-type bacteria
00:16:42.18 inside the macrophages, but either in the presence or the absence
00:16:46.13 of the ICL inhibitor 3-nitropropionate.
00:16:48.29 Again, in the absence of inhibitor, the bacteria grow just fine,
00:16:52.08 as indicated by the blue symbols.
00:16:54.18 When we add 3-nitropropionate, this completely blocks the ability
00:16:57.03 of these bacteria to grow inside macrophages.
00:16:59.10 So, we now have enzyme assays that we can use to look at inhibitors,
00:17:04.11 using pure recombinant enzymes,
00:17:06.15 as were used for the crystallographic studies.
00:17:09.16 We have whole-cell assay in axenic culture
00:17:13.14 as I showed previously, and we now have a macrophage culture system as well
00:17:17.13 that we can use to study inhibitors and their activity against the tubercle bacillus.
00:17:22.02 Now, it turns out that ICL1, although biologically it's an extremely attractive target
00:17:28.28 for the reasons I've described already,
00:17:30.27 is a very difficult target for structural reasons.
00:17:34.13 So, this is a 3-dimensional depiction of the X-ray crystallographic structure
00:17:38.29 of ICL that was obtained by Jim Sacchettini and Vivek Sharma
00:17:42.24 in his lab at Texas A&M.
00:17:45.01 This depicts a cut-away showing the active site of the enzyme.
00:17:49.02 It's a very shallow groove in the surface of the molecule.
00:17:52.03 And when the drug actually binds to the enzyme,
00:17:55.21 this protein flap here folds down over the molecule,
00:17:59.19 essentially locking it in place.
00:18:01.19 This means that the active site, particularly when it contains a ligand,
00:18:05.10 is extremely space constrained, and this is a problem for drug development
00:18:08.12 we found because it means that there's very little room to tinker with the molecule
00:18:12.06 to try to add and move groups to achieve a higher potency drug
00:18:17.02 by promoting different interactions with the enzyme itself.
00:18:20.23 So, it turns out that we can isolate inhibitors of ICL,
00:18:25.00 but as soon as you start to tinker with their structure, they tend to lose activity,
00:18:28.13 rather than becoming more potent.
00:18:30.00 There's also the fact that there are two of these molecules in the cell,
00:18:34.06 and although I think it might be feasible to develop a dual-specific inhibitor
00:18:38.08 like 3-nitropropionate, again, it adds to the complexity of the problem.
00:18:42.15 So, it's not at all clear now that the efforts on the part of Glaxo Smith Kline
00:18:47.10 to develop ICL inhibitors, in fact, in the end are going to bear fruit.
00:18:50.20 That is, fruit in this scenario means a drug
00:18:55.14 that we can actually use to treat tuberculosis.
00:18:57.23 So, more recently, we have focused our sights on
00:19:02.12 looking at the second unique step in the glyoxylate cycle.
00:19:06.28 That is the enzyme malate synthase,
00:19:09.04 the 3-dimensional structure of which has also been solved
00:19:12.00 by Claire Smith in Jim Sacchettini's lab, using x-ray crystallography.
00:19:16.21 Now, malate synthase is a much more attractive target from a
00:19:20.13 structural point of view for two reasons.
00:19:21.27 First, there's very clearly only one of these enzymes
00:19:25.13 in Mycobacterium tuberculosis
00:19:26.26 And second, the active site is configured very differently in this enzyme.
00:19:30.28 So, this little acidic patch here shows the opening into
00:19:35.12 the active site of malate synthase.
00:19:37.19 And if we now look at a cutaway view of that active site,
00:19:40.07 what we see is that, in fact, the active site, right here, where catalysis takes place
00:19:44.00 is reached by this long, hydrophilic tunnel.
00:19:47.19 This gives us a lot of elbow room, so to speak,
00:19:50.14 for tinkering with molecules to try to create new contacts to the enzyme
00:19:55.18 that would increase the interaction, and therefore the potency of a drug,
00:20:03.08 based on an inhibitor of malate synthase.
00:20:05.11 This is a very attractive feature of the enzyme,
00:20:10.03 and our current efforts in my lab, Jim Sacchettini's lab, and at GlaxoSmithKline
00:20:13.25 are to target malate synthase for the same kind of drug discovery program
00:20:17.26 that we've been pursuing for isocitrate lyase.
00:20:20.18 Our hope is, of course, that if we can stop up this hole
00:20:22.27 in malate synthase, we may also be able to stop TB in its tracks.
00:20:26.21 So, to sum up what I've told you,
00:20:30.08 there are old problems that have been around throughout the history of TB
00:20:36.12 that continue to be problems today, but I think
00:20:38.14 we now see on the horizon the hope and the prospect of developing new tools
00:20:43.16 to attack these old problems.
00:20:45.21 Before I close, I just want to identify three areas
00:20:48.18 where I think there's a particularly urgent need for development of new tools,
00:20:51.27 and I think in all cases, academic scientists, like myself,
00:20:55.16 have an important role to play in driving discovery efforts.
00:20:58.19 First, as I hope I've made clear,
00:21:01.14 there's a tremendous need for a new vaccine against tuberculosis.
00:21:05.13 This is probably the greatest unmet need of all.
00:21:07.28 Ideally, we would like a prophylactic vaccine that is consistently effective
00:21:12.09 to replace the current BCG vaccine.
00:21:15.15 So this would be ideally a vaccine that would be given in childhood --
00:21:18.28 hopefully at birth -- that would prevent development of tuberculosis
00:21:22.19 later in life, when these individuals reach adulthood.
00:21:26.02 Obviously, that's a huge scientific and medical challenge.
00:21:29.01 But, a prophylactic vaccine like this is, presumably, not going to do anything
00:21:34.04 to address another problem that already exists, and that's
00:21:37.02 the problem of these 2 billion people who are currently latently infected.
00:21:41.04 There's no reason to think that using an effective prophylactic vaccine
00:21:45.15 on these people would be effective in preventing reactivation.
00:21:48.23 As I said, we can look forward to hundreds of millions of new cases of TB
00:21:52.07 due to reactivation of infections that are already established
00:21:57.07 at this time.
00:21:58.23 So, I think there's also a tremendous need for a post-exposure vaccine,
00:22:01.28 if you like, that could be administered to individuals that are lately infected already
00:22:05.22 that will effectively prevent them from reactivating,
00:22:09.11 either by putting the bacteria into a permanently suppressed state,
00:22:12.12 or hopefully inciting the immune system to a state that's capable
00:22:16.10 of completely eradicating these bacteria.
00:22:18.22 There's a tremendous need for new diagnostics, which is an area
00:22:22.22 that I didn't really talk about.
00:22:24.11 But, it's an extremely important area, where we're starting to see
00:22:27.17 some activity. I would say, in terms of diagnostics,
00:22:30.28 there are 2 things that we need. First, we need an ability to identify active cases of TB
00:22:37.16 with high specificity, and also with high sensitivity.
00:22:41.13 We need to be able to identify the people who are sick and need treatment.
00:22:44.19 But we also need to be able to distinguish between individuals
00:22:48.08 who have active tuberculosis and those who have latent tuberculosis,
00:22:52.13 who might be handled in a somewhat different way, clinically.
00:22:55.13 So, that's an important distinction that any new diagnostic needs to be able
00:22:59.01 to make.
00:23:00.16 Also, we need a diagnostic that is capable of distinguishing between
00:23:03.13 tuberculosis infection and BCG vaccination.
00:23:06.17 Undoubtedly, BCG will continue to be used for a long time to come.
00:23:10.03 And it is a very widely used vaccine in developing countries.
00:23:13.26 One of the major problems with BCG vaccination,
00:23:17.09 besides the fact that it is not always effective,
00:23:19.25 is that it compromises what is currently the only diagnostic that we
00:23:23.25 have for tuberculosis, namely the tuberculin skin test
00:23:27.14 that was originally devised by Robert Koch in the 19th century.
00:23:31.28 So, it's a very antiquated tool.
00:23:33.10 A diagnostic that, with great sensitivity and specificity,
00:23:37.22 could distinguish active TB from latent TB from BCG vaccination
00:23:42.06 would be an extremely powerful new tool in the TB armamentarium.
00:23:46.12 Lastly, as I've emphasized, there's a tremendous need for development
00:23:52.03 of new drugs. We need new drugs for the treatment of the multi-drug-resistant cases
00:23:56.01 that already exist in the world.
00:23:57.21 There are many essentially untreatable cases that have emerged in recent years.
00:24:02.13 So, there's always going to be a need for a pipeline
00:24:06.01 for new drugs to emerge, simply for treatment of drug-resistant TB.
00:24:09.12 But, more important, I would say, we need a new drug
00:24:12.26 that does something none of our current drugs accomplish,
00:24:16.00 and that is an ultra-short chemotherapeutic regimen for tuberculosis.
00:24:19.13 And this is to deal with the human problem of adherence
00:24:23.08 or compliance. If we could develop
00:24:24.18 a new drug that would radically shorten the time for treatment
00:24:27.27 of TB from 6-9 months to 6-9 days,
00:24:31.27 this would have an enormous impact on the rates of adherence and compliance
00:24:37.02 with chemotherapy and the cure rates that we achieve with chemotherapy.
00:24:41.06 This is a very daunting scientific challenge, but I'm actually quite
00:24:44.07 hopeful that we will, in the not too distant future
00:24:47.13 see new drugs coming into the clinic and onto the market
00:24:51.25 that achieve at least faster, if not ultra-short, therapeutic regimens.
00:24:57.28 Lastly, we need a drug... and it might be the same, or it might be a different drug
00:25:02.13 that we can use prophylactically. Again, this gets to the problem of this latent
00:25:06.02 population of individuals for whom we currently have no effective
00:25:10.08 and practicable tools for intervention.
00:25:12.15 If we could achieve an ultra-safe and ultra-short course chemotherapeutic regimen
00:25:18.12 to give to people who are already latently infected to prevent them
00:25:22.06 from reactivating, we could finally do something
00:25:26.01 to confront and deal with this huge reservoir of 2 billion
00:25:30.21 latently infected individuals worldwide, who in fact
00:25:32.24 are responsible for the majority of active cases that we are going to see in the future.
00:25:38.00 Thanks very much. I hope I've managed to
00:25:41.27 outline for you some of the problems that we confront in TB.
00:25:45.26 Clearly, they're not just scientific problems. They're problems of delivery of medicines
00:25:51.01 and the public health goods. I hope I've also made it clear to you that
00:25:56.07 there is an important role for academic labs to play in development
00:26:00.09 of these new interventions against tuberculosis.
00:26:02.09 And I've tried to give you a couple of examples of how my lab has tried to
00:26:05.24 contribute to this process. Thanks very much.
00:26:08.17

Videos in this Talk

Part 1: Tuberculosis a Persistent Threat
Audience:
- Researcher
- Educators of H. School / Intro Undergrad
- Educators of Adv. Undergrad / Grad
Part 2: Tools for Tuberculosis Control: Not Just a Problem of Implementation
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad
Part 3: Phenotypic Heterogeneity and Antibiotic Tolerance
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad
Part 4: Targeting M. tuberculosis Carbon Metabolism In Vivo
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad

Speaker: John McKinney

Total Duration: 1:49:15

Recorded: June 2007

All Talks in Microbiology

Talk Overview

Human population growth and urbanization have accelerated dramatically in recent centuries, providing unprecedented opportunities for microbes that use our bodies as vehicles for their own propagation and transmission. These conditions have led to the emergence of virulent new pathogens and the increased prevalence of “classic” scourges, such as Mycobacterium tuberculosis. This tenacious microbe is transmitted via infectious aerosols produced by individuals with pulmonary tuberculosis. Infection is lifelong and symptomatic tuberculosis may develop following a period of clinical latency lasting for months, years, or decades. The first part of this lecture provides an overview of the natural history of TB infection and the global impact of TB on human health.

Tuberculosis remains one of the most important causes of human disease and death despite the introduction of vaccination in 1921 and chemotherapy in 1952. Although these interventions are inexpensive and widely available their impact is limited. The effectiveness of vaccination is unclear; in clinical trials, the protection conferred by vaccination has been variable and generally poor. Although chemotherapy can be highly effective, multiple drugs must be administered for 6-9 months to provide a reliable cure; the majority of tuberculosis patients are unable or unwilling to complete such a demanding regimen unless closely supervised. The second part of this lecture will discuss the challenges facing development of more effective vaccines and drugs for prevention and treatment of tuberculosis.

The principal obstacle to successful treatment of tuberculosis is the lengthy duration of current regimens, which require administration of multiple drugs for 6-9 months. The requirement for prolonged therapy is attributed to sub-populations of bacillary “persisters” that are refractory to antimicrobials. The persisters are not drug-resistant in the conventional (heritable) sense and it is a mystery why they are spared whilst their genetically identical siblings are killed. The third part of this lecture describes recent work in our laboratory using microfluidics and time-lapse microscopy to analyze the behavior of drug-stressed bacteria at single-cell resolution. These studies challenge conventional views of how antimicrobials kill (or fail to kill) bacteria.

All pathogens must acquire and assimilate nutrients from their hosts in order to grow and multiply — our tissues are literally their food — yet surprisingly little is known about this fundamental aspect of the pathogenic lifestyle. Accumulating evidence suggests that M. tuberculosis might utilize fatty acids as its principal carbon and energy source during infection. The fourth part of this lecture describes work in our laboratory that is focused on identifying the metabolic pathways that are essential for growth and persistence of M. tuberculosis in vivo. Some of these pathways are potentially interesting targets for antimicrobial drug development, as they are not found in human cells.

Speaker Bio

John McKinney

John McKinney is Professor of Bacteriology at the École Polytechnique Fédérale de Lausanne (EPFL) in Switzerland. McKinney received B.A. (biology) and B.Sc. (genetics) degrees from The Evergreen State College (1987) and a Ph.D. from The Rockefeller University (1994). After completing his postdoctoral studies in the laboratory of William Jacobs at the Albert Einstein College of… Continue Reading

Part 1: Tuberculosis a Persistent Threat

Part 2: Tools for Tuberculosis Control: Not Just a Problem of Implementation

Part 3: Phenotypic Heterogeneity and Antibiotic Tolerance

Part 4: Targeting M. tuberculosis Carbon Metabolism In Vivo

Talk Overview

Speaker Bio

John McKinney

Leave a Reply Cancel reply

Funders

Start Here

Explore

Talks By Topic

Talks by Series

Career

Courses

Educator

About Us

Talk Overview

Speaker Bio

John McKinney

Playlist: Pathogens

Related Resources

Reader Interactions

Leave a Reply Cancel reply

Footer

Funders

Start Here

Explore

Talks By Topic

Talks by Series

Career

Courses

Educator

About Us