| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Mycobacteria Research Laboratories, Department of Microbiology, Colorado State University, Fort Collins, Colorado
Correspondence: Ian Orme, Ph.D., Department of Microbiology, Colorado State University, Fort Collins CO 80523. E-mail: iorme{at}lamar.colostate.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT BACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
Key Words: BCG • animal models • T cells
|
BACILLUS CALMETTE-GUERIN (BCG): THE CURRENT VACCINE |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
After its acceptance as a vaccine for tuberculosis (TB) in humans, the BCG Pasteur Lille was widely distributed throughout the world to other institutions. As a result, many daughter substrains came into being, and recent molecular analysis has revealed numerous changes, gene deletions, as well as a suspicion that continued passage has attenuated BCG to almost complete avirulence [2 3 4 5 ].
|
THE NEED FOR NEW VACCINES |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
As BCG became more widely accepted, various countries began to conduct controlled, clinical trials. At first, the news was good. For instance, in a trial in 1935 in native North Americans, there was an 82% reduction in deaths from TB [6 ]. Similar efficacy was seen in children in Chicago at high risk of disease [7 ] and in a study in the United Kingdom [8 ].
Other trials did not give such good results, however. Two trials in Georgia showed no protection after BCG, although in both, the incidence of disease was low (only a few cases) [1 ]. In contrast, in a study in a region of southern India, in which TB was prevalent, over 265,000 people were tested, and again the effects of BCG were zero [9 ].
This famous study began a process of complete re-evaluation of BCG as a vaccine. Although it is accepted that BCG does have some protective effect in children, particularly against forms such as meningeal TB, it does not prevent the emergence of pulmonary TB, particularly as the individual approaches adulthood [10 11 12 13 ].
Although there is a concerted effort ("directly observed therapy") to get drugs to the people who need them the most, as Kaufmann pointed out recently [14 ], an inexpensive vaccine that induces good immunity would be equally effective regardless of whether the infection was caused by drug-susceptible bacilli or by multiple drug resistance (MDR)-TB. In the long run, vaccination will be far more cost-effective than chemotherapy. Although not politically correct, there is a further caveat: namely, that large pharmaceutical companies probably dread the idea of finding a new, highly effective TB drug, given that (1) the cost of developing such drugs can be over $100 million, and (2) their marketing experts would remind them that the great majority of patients with active disease tend to be poor people, mostly in developing nations.
|
WHAT SHOULD A NEW VACCINE DO? |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
|
It may well be impossible to block these mechanisms in any way, and so the immediate outcome resides on the ability of the macrophage to kill the bacillus. These cells flatten out across the alveolar surface and are tightly opposed; hence, if the bacillus cannot be killed, it begins to replicate and can potentially erode into the interstitium, thus creating an infectious site [16 ]. Because of this process, most scientists believe that a vaccine that can prevent infection is not possible.
Initial disease
A vaccine that can rapidly detect sites of infection and mobilize
the immune response, thus preventing any further bacterial
proliferation and potential dissemination, is the most practical
likelihood. However, there may be physiologic limitations to this
process. Even in mice highly immune to TB, the kinetics of influx of
effector-memory T cells into the lungs after rechallenge are not
especially rapid [17
], and as a result, although about a
tenfold reduction in the bacterial load can be observed, some bacterial
proliferation seems inevitable.
To get memory cells to the site, gradients created by chemokines and other molecules that regulate local inflammation must be followed [18 ]. This takes time and, at least partially, involves aspects of the lung-innate response unaffected by vaccination. As we discuss elsewhere in this review, such multiple mechanisms may exist.
Post-exposure or immunotherapeutic vaccines
The concept here is to produce a vaccine that can be given to
people with active disease, promoting increased clearance of the
bacilli from the lungs.
The drawback is that even if an antigen-specific, T-cell pool could be
induced, the location of the bacterial infection may preclude any
beneficial effect. The core of the host response is to surround sites
of infection with a granuloma response. These structures can grow
large, and within them, only a small percentage of macrophages actually
contain a viable bacillus [19
, 20
]. Thus,
not only will the T cell have to find these structures, it will also
need the necessary integrins to burrow into the structure, find
infected macrophages, and release interferon-
(IFN-). This may
indeed occur if the integrity of the granuloma is poor, and the
structure is breaking down, but in situations where the granuloma is
stable, it is hard to envisage how this could happen.
Vaccines against latency
I have expressed elsewhere my skepticism [21
] that
there actually is latent TB, and I do not think that there is an
adequate animal model for this as yet, even if it does indeed exist.
Semantics aside, everyone would probably agree that the scenario
envisaged here—lesions that are not sterilized completely by host
immunity or drug therapy—allows the survival of a few bacteria that
have the potential to give rise to reactivation disease at a later
time.
It is probable that if low-grade, chronic disease equates to latent disease, it is monitored by memory T cells in a sentinel or surveillance mode [22 ]. Thus, a good starting point would be to target this memory cell population in order to increase the pool size and also extend longevity. This strategy may not only involve CD4 cells; it is tempting to speculate that CD8 T cells, which we have shown recently to be scattered toward the outer edges of granulomas in the lung [23 ], are also performing such a role.
|
WHAT ARE THE TARGET T-CELL POPULATIONS? |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
Macrophages present antigens to CD4 T cells via class II major histocompatibility complex (MHC) molecules. They also secrete interleukin (IL)-1, stimulating IL-2 and IL-2 receptor (IL-2R) expression by the T cell. However, the key cytokine is IL-12. Mycobacterial infection is a potent inducer of this molecule from host macrophages, and it is pivotal in driving the T-helper cell type (Th1) response to TB [26 ].
Once committed to this pathway, the effector CD4 cell subset begins to
secrete the key effector cytokine IFN- [27
,
28
]. This activates the infected macrophage to produce
oxygen and nitrogen radicals and to assemble proton pumps on the
phagosomal membrane, thus reducing the phagosomal pH
[29
30
31
].
The role of radicals in the containment of the infection is controversial. For some time, it was generally believed that oxygen radicals did not play any role, but more recent information using phox-knockout (KO) mice indicates a brief period during aerosol infection when lack of superoxide causes a transient rise in susceptibility [32 ]. Similarly, although blocking nitric oxide (NO) generation in intravenously (i.v.) infected mice causes a large increase in the bacterial load [33 , 34 ], the effects in the lungs of aerosol-infected mice are greatly reduced in models of M. tuberculosis and M. avium infection [35 , 36 ]. When taken in concert with the information [37 ] that clinical isolates of M. tuberculosis show a wide range of resistance to NO (some are completely oblivious), it must be concluded that the role of NO may be overstated.
Having said that, the role of NO may not be mycobactericidal but may be more in terms of regulation of the granuloma formation instead, as recent studies in NOS2-KO mice have suggested [35 , 36 ]. Perhaps this reflects its effects on local blood vessels.
Of course, the central target of the vaccine is CD4 memory T cells. To study these cells, the usual model is infection of mice followed by a prolonged period of chemotherapy to clear the bacilli [38 ]. Prior to rechallenge, these cells are small (low forward-scatter on flow analysis), resting T cells that express a CD44hi CD45RBlo cell-surface phenotype [39 ]. Following rechallenge, they mediate accelerated resistance to the infection rapidly [38 , 40 ].
Having said that, this is still an under-researched area, especially considering the pivotal importance of these cells to the ultimate goal of new TB vaccines. Their phenotype may not be stable, and a recent study suggests they may undergo reversion to a CD45RB+/hi phenotype [41 ]. Also, they may actually comprise two distinct subsets, based on their expression of the chemokine receptor CCR7 and their predilection to home to inflammatory sites within tissues or lymph nodes [42 ]. In my laboratory, this is an important issue, because I tend to rely on the CD44hi45lo phenotype to identify memory CD4 cells and use a footpad-challenge assay in which I measure cells with this phenotype accumulating in the draining, popliteal lymph node.
Perhaps the most important issue, however, is the longevity of the memory T-cell population. In fact, a gradual disappearance of these cells over time may be a major factor in unsuccessful BCG clinical trials. There is plenty of evidence now that memory T cells "turnover" from time to time, and this may be a homeostatic way of retaining these cells. Therefore, it is quite possible that if the stimuli for this process are too intense or too frequent, then this will have the opposite effect and diminish the memory T-cell pool.
Left unperturbed, the longevity of memory may be considerable. There is
plenty of data indicating that people born in the first half of this
century in first-world countries were skintest-positive and remain so
to this day. (Presumably, this is mediated by CD4+ CCR7neg
memory cells.) However, what if the individual is exposed to other
stimuli constantly, such as environmental mycobacteria? In my
laboratory, I have evidence that repeated exposure of BCG-vaccinated
mice to dead M. avium gradually diminishes the size of the
memory T-cell pool. To date, it is unknown exactly how this might
occur; most of the literature describes studies in CD8 responses
[43
44
45
46
], and there, memory CD8 T-cell turnover can be
driven nonspecifically by an IFN
ß/IL-15 axis. If a similar
mechanism exists for CD4 memory cells, then it can be speculated that
BCG vaccination may induce similar protection in everyone, but
depending on where you live and to what you are exposed, it may
directly influence the longevity of your memory T-cell population.
CD8 T cells
The role of CD8 T cells in immunity to TB remains elusive. Many
years ago, I [24
] was able to show that an enriched, CD8
T-cell population from mice immunized with a high dose of BCG prolonged
the survival of irradiated mice exposed to an acute, high-dose,
aerosol-challenge infection, and more recent work by Flynn and her
colleagues [47
] showed that ß2-microglobulin-deficient
mice were highly susceptible to i.v. infection.
Using the same model, I have found that early resistance to aerosol infection does not seem directly attributable to lack of CD8 T cells, because direct comparison to CD8-KO mice indicates that these animals are not more susceptible until the lung infection is well into the chronic phase [48 , 49 ]. What causes the early rise in the lung bacterial load in ß2M-KO mice remains unknown, although class-1b-restricted natural killer (NK) cells or NKT cells might be candidates. Another molecule expressing ß2M, the CD1 molecule, is not involved, at least in the mouse model [48 ], although it seems to be important in human studies [50 , 51 ].
After i.v. infection, CD8 T cells accumulate in the lung lesions quite quickly [52 ]. We have seen a similar pattern after aerosol infection, but in our study [23 ], immunohistochemistry revealed that most CD8 cells tend to accumulate around blood vessels and bronchioles rather than infiltrate the lesions. Furthermore, this study also showed that during the chronic phase of the infection when the granulomas are large and stable, there are clear distinctions in distribution between CD4 and CD8 T cells, with the CD4 cells found in large aggregations within the structure and CD8 cells, more scattered with a tendency to be found toward the outside of the lesion (Fig. 2 ).
|
If this is the case, what are these cells doing, and what antigens are
they recognizing? CD8 T cells could be activating the infected
macrophages via secretion of IFN-, or they could be killing them.
This could be via direct lysis or by the delivery of apoptotic signals
[49
]. Viable bacilli could be then taken up by adjacent
macrophages or even killed directly by granulysin [53
].
Logically, apoptosis seems more likely. Lytic mechanisms would cause considerable, local tissue damage, and it should be remembered that cytolytic T lymphocyte (CTL) activity and bacterial damage by granulysin have only been demonstrated in vitro. In our study, mice lacking the perforin molecule look identical to controls during the chronic stage of the disease [49 ]. In the same study, we also observed that lack of CD95 (Fas) allowed regrowth of the infection during chronic disease, indicating that apoptotic mechanisms are indeed operative during this phase of the disease process.
Antigen presentation to CD8 T cells is restricted by class I MHC
molecules, which requires processing of the antigen via a cytoplasmic
rather than endosomal route. Therefore, it is possible that certain
macrophages are able to control the infection well into the
chronic-disease stage, but then this control is lost (maybe the cell is
too far away from the aggregates of IFN--secreting CD4 cells in the
granuloma, or their cytokine production has diminished). At this point,
the bacilli grow unchecked in the cell, damaging the phagosome and
allowing leakage into the cytoplasm. These are detected by the CD8
cells in the granuloma, which induce lysis or apoptosis in this cell,
releasing the bacilli to be taken up by immunocompetent, adjacent
macrophages.
If we wish to target these CD8 cells with a vaccine, we need to know what they are recognizing; as yet, this is unknown. Even then, would this be beneficial? If the end result is lysis, then amplifying this mechanism may actually worsen the lung pathology.
|
INNATE MECHANISMS: COULD THEY BE TARGETED? |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
can be detected before this
event. Moreover, mice lacking CD4 cells or transgene mice that have no
T-cell receptor (TCR) capable of recognizing mycobacterial antigens
live much longer than IFN--KO mice (unpublished results). The
explanation for these observations is that, in fact, the mouse has
several, probably redundant, innate mechanisms of immunity in the lungs
[15
].
Yet again, it is accepted that these mechanisms are poorly understood.
In humans, there is compelling evidence that CD3+4-8- T cells
recognize mycobacterial lipoproteins presented by CD1 molecules
[50
, 51
]. Pattern-recognition Toll
receptors on macrophages may also be involved [54
,
55
], and this stimulation may be the source of IL-12
needed to drive early sources of IFN-. NK or NKT cells may be one
target of the IL-12, and indeed, these cells are known to be a potent
source of IFN- (Fig. 3
).
|
and differs from regular
CD4 cells by lower expression of its TCR (unpublished results).
Finally, the role of 
T cells remains elusive. We have suggested
[16
] that these play a regulatory role in cell
trafficking rather than a directly protective role, given that
-KO mice have a similar bacterial load but a considerably
different, much more pyogenic lung-tissue pathology
[56
]. It is interesting that cells from humans
seem to strongly recognize pyrophosphate-containing chemical moieties
[57
58
59
60
61
62
], whereas in animal models (mouse, cow), the
targets seem to be proteins.
Would stimulation of these innate mechanisms be beneficial? It seems not. Not only do innate mechanisms lack any form of memory component, but also strong expression of innate immunity may completely preclude any proper emergence of acquired immunity. This would be acceptable if it could be ensured that innate immunity would completely sterilize the site of infection, but this seems unlikely to happen.
|
THE FORGOTTEN LYMPHOCYTE: WHAT ABOUT B CELLS? |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
However, should the role of B cells be discounted altogether? Perhaps not. B-cell KO mice show no changes in bacterial load after aerosol infection [65 ], but their lung granulomas are noticably smaller [66 ]. As we have shown recently [23 ], a simple explanation for this is that an appreciable number of lymphocytes within lung TB granulomas are indeed B cells. Why they are there is unknown; it may be a simple mistake because blood-vessel adhesion molecules at lesion sites would attract B-cell migration as well as T cells (the B cells think they are entering lymphoid tissue).
I would propose that the reason a role for B cells should not be completely discounted is because they may contribute to the continued stimulation of memory immunity. This is a contentious area; some argue that stimulation of memory cells is needed, and indeed, a recent study showed reduced T-cell memory in B-KO mice [67 ]. Other studies, such as those parking memory cells in MHC-deficient mice, argue the opposite [68 , 69 ].
If B cells are indeed needed, then it is believed that the antibody response to mycobacterial antigens results in deposition over time of immune complexes that may end up being captured by follicular, dendritic cells and periodically presented to passing memory T cells. If this is the case, then it can be experimentally tested to see if despite their capacity to deal with active disease, B-KO mice cannot generate persistent memory, measured as reduced resistance to rechallenge.
|
BCG: IS IT GETTING A BAD RAP? |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
An appreciation of this depends on a "glass half empty" or "glass half full" belief. Certainly, a percentage of scientists fall into the half-empty group, believing that BCG is essentially worthless and needs to be completely replaced. Conversely, the half-full membership points to the fact that in some of the trials, BCG did work well indeed, whereas in others it did not. So one question, still unanswered, is what is BCG doing when it does have a positive effect?
In fact, the stark reality is that this question has never been adequately answered. If the working hypothesis is believed, that BCG induces memory immunity, but this immunity is gradually lost, then perusal of the literature shows that recent, serious studies of this parameter in models can be counted on two hands [17 , 38 39 40 41 , 70 ]. The classical studies showing that BCG induces lymphocytes and that these can transfer resistance to naive recipients were done decades ago [71 , 72 ], long before there were CD markers and flow cytometers, knowledge of cytokines/chemokines barely existed, and no one had an inkling that there were CD1 molecules capable of presenting mycobacterial lipids and class-1b molecules.
In fact, it can be hypothesized that BCG can switch on a variety of
mechanisms, as summarized in Figure 4
. Central to these is a potent, memory T-cell response
[73
]. In areas in which BCG is ineffective, or
protection is of limited longevity, one possible explanation is that
this memory T-cell pool is interfered with gradually or consumed,
leading to an eventual loss of resistance and full susceptibility to
disease by the teenage years. This may be a result of nonspecific
stimulation of memory T-cell turnover and division, which may force a
percentage of the pool into apoptosis, thus eventually reducing the
pool size to a frequency whereby it is no longer effective. Although it
is not yet understood how the size of the overall CD4 memory pool is
regulated, it has been shown for CD8 memory cells that turnover can be
induced via the production of IFNß and IL-15 after nonspecific
activation of macrophages [46
, 74
].
|
|
IS REPLACING BCG THE BEST STRATEGY? |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
However, relying on "protection" alone may be limiting our options in screening potential vaccines. For example, we have made several observations whereby a particular test candidate does not induce a significant reduction in the day-30, bacterial-load assay. Such an example is the Ag85A-DNA vaccine, but if left alone, guinea pigs given this vaccine show excellent long-term survival [78 ]. Moreover, and perhaps much more importantly, the necrotic pathology that kills unprotected guinea pigs was replaced in the vaccinated pigs by a protective lymphocytic granuloma. Because a similar pathology occurs in human patients, this may be a positive effect, and this teaches us that a vaccine should not be completely discounted just because it cannot "beat BCG" in the 30-day assay.
It is also evident that testing a new vaccine against BCG is going to be difficult to achieve, simply because most children in areas of the world in which TB is endemic are given BCG as neonates. As a result, much of the experimental effort in my laboratory is to find ways to boost BCG as a vaccine.
This can be done in two ways. In the first method, the assumption is made that BCG is indeed effective (glass half-full), but memory immunity diminishes. If BCG-vaccinated mice are given BCG when young, their resistance to aerosol infection declines gradually as they approach old age. However, if given Ag85A protein-boosting in mid-life, their resistance to aerosol when elderly is equivalent to that of a young mouse [79 ].
A second way is to simply add a highly immunogenic vaccine to the BCG inoculum. In a recent study, guinea pigs were vaccinated with a mixture of BCG and an immunogenic fusion protein (M. tuberculosis 72f). Whereas the BCG, control guinea pigs began to die after about 300–350 days, four out of five pigs given the admixture remain alive (500 days, at the time of writing).
|
TYPES OF NEW VACCINES |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
This approach avoids the need for adjuvants, which are problematic in terms of other types of TB-vaccine candidates, and because a living mycobacterium is not involved, it can be safely put into an immunocompromised host. Once inoculated, experiments using a luciferase reporter gene suggest that the plasmid remains stable for some prolonged period of time, and the risk of genomic integration (a big worry) seems to be rare.
DNA vaccines seem to work well because they induce a type of pan-immune response; whatever the response of choice, DNA seems to induce them. After presentation by dendritic macrophages present in the muscle mass, a strong Th1 response can be observed, as well as strong CD8 CTL responses. (It is interesting that if given by "gene gun", which drives particles or liquid droplets into the dermis, Th2 responses are favored.) In addition, certain promotor systems, such as ubiquitin [81 ], may favor Th1 immunity and resulting protection in mice against TB infection.
Thus far, the majority of models tested have been viruses, but an increasing number of studies using bacterial, protozoal, and helminthic models are also appearing. In the TB field, several mycobacterial antigens have been targeted, including Ag85A, the phosphate-binding protein family PstS, the 36-kDa protein, ESAT-6, and the heat-shock proteins (hsp)60 and hsp70 [78 , 82 83 84 85 86 ].
The two most effective vaccines are against Ag85A and hsp60, although the results regarding the latter have unfortunately become controversial. In terms of the Ag85A vaccine, it reduces the day-30 bacterial load in mice. In guinea pigs, the load is appreciably not changed, but there is excellent, long-term survival, and the necrotic lung pathology seen in controls is completely prevented [78 ]. However, as a post-exposure vaccine in mice already infected with M. tuberculosis, the Ag85A DNA has no protective effect [87 ].
In mice given the hsp60 DNA (derived from M. leprae), the vaccine is equally, highly effective if given before or during TB infection and in both models causes a massive reduction in the bacterial load [84 ]. In addition, in a "Cornell style" mouse model in which the bacterial load is reduced to apparent sterility by chemotherapy, and then residual "latent" bacilli are reawakened by cortisol suppression of immunity, administration of three injections of the DNA vaccine totally prevented reactivation.
In our study, a DNA vaccine encoding the M. tuberculosis hsp60 gene gave different results [88 ]. The vaccine was completely unprotective in the mouse and guinea pig models, and in both cases, we noticed that many animals exhibited respiratory distress a few weeks after aerosol infection, requiring their euthanasia. Subsequent histologic analysis of the lungs of these animals revealed a severe necrotic, airway-inflammatory reaction; this was not seen in the vector controls or in animals vaccinated with Ag85 DNA. Given the close homology between the DNA vaccine genes involved, I cannot explain these divergent results as yet.
Subunit vaccines
Two studies published almost simultaneously in 1986 were the first
to suggest that proteins secreted or expelled from the bacillus had the
potential to be the key, protective antigens early during the host
reponse [89
, 90
], thus challenging the
prevalent view at the time that constitutive proteins were the primary
targets [91
, 92
]. Gradually, this
hypothesis has become widely accepted [93
], and several
laboratories were able to show protection in animal models using
proteins from the culture-filtrate fraction soon thereafter
[78
, 94
95
96
97
].
This fraction contains many immunogenic proteins, some potent enough to induce some degree of immunity by themselves or as an admixture (Ag85 complex, ESAT-6, for example). However, in the more stringent guinea pig, none have been found as yet to be capable of inducing protection and long-term prevention of pathology seen in BCG controls.
Subunit vaccines need to be delivered in adjuvants, and it may be that the primary limitation is a failure to drive a sustained Th1 response rather than to find a truly effective protein candidate. Several adjuvants have been tried, and the most effective to date include monophosphoryl-lipid A (MPL), QS21 (a saponin), and dimethyl-dioctadecyl ammonium bromide (DDA).
This continues to be an active area of research. One recent advance is to identify immunogenic peptides or epitopes using cloned, human cell lines from PPD-positive individuals and then create fusion proteins using this information. Some of these fusion proteins are protective in the mouse and guinea pig models, and addition to a BCG vaccine as an admix significantly prolongs survival of guinea pigs when compared with BCG alone.
Another approach is to separate complex mixtures of proteins such as
filtrates using isoelectric focusing and then in a second dimension,
using gel electrophoresis; then electroelute each separated protein,
and test whether they stimulate IFN- secretion by T cells from
infected mice. Those that are the most active are then identified by a
proteomics approach using liquid chromatograph-mass spectrometry, and
then predicted sequences are matched to known M.
tuberculosis proteins using commercial software such as SEQUEST.
Using this approach, several novel targets have been identified
[98
] and are now undergoing evaluations in animal
models.
Auxotrophic vaccines
There are two concepts at work here. The first is that many people
at risk of TB are also HIV-positive, and hence, even an attenuated
bacterium such as BCG may cause fatal disease in these individuals. The
second concept is to try to adapt the organism to the extent that it
could itself be used as a vaccine, because the majority of people
exposed to M. tuberculosis are actually able to express
strong resistance to it.
To date, the major adaptation applied, in both cases, is auxotrophy. Several have been developed and tested now [99 , 100 ]. The initial results are encouraging in that protection can be conferred, and yet the vaccines themselves are cleared gradually, even in completely immunodeficient mice. If there is a drawback to this approach, it is the concern that the lack of persistence of the vaccine inoculum fails to adequately drive a memory T-cell response so that if the vaccine-challenge interval in the animal models is extended, protection measured tends to diminish. However, we are optimistic that this problem can be resolved.
Others
There are other categories of new vaccines being tested but with
varying degrees of success. One such area is recombinant vaccines,
which usually involve insertion of genes into the existing BCG vaccine.
This has resulted in several innovative strategies [101
,
102
], but when tested against BCG controls in mice, there
is no evidence that the vaccine is more potent. In fairness, however,
the point should be made that a true comparison has yet to be
performed, because no one has attempted as yet to see if recombinant
BCGs (rBCGs) could be better in long-term studies in guinea pigs, which
might be a better form of evaluation of this class of vaccines.
In this regard, a recent study shows for the first time that protection in excess of that given by BCG can indeed be attained [103 ]. In that study, a BCG overexpressing the Ag85 protein was constructed, which gave about a half-log improvement in bacterial-load reduction compared with the BCG control in the guinea pig model. This is an exciting result, and it remains to be seen if this could result in better, long-term protection.
Another area of interest is to use avirulent mycobacteria as vaccines. There is some precedent for this; M. microti was tested in comparison with BCG in a trial in the United Kingdom [8 ], and protection was similar, although it should be noted that TB rates in children in the UK were already significantly falling during that time. There is nothing to suggest, however, that M. microti has any particular advantage over BCG and could be used to replace it.
A second candidate in this category is M. vaccae. In our animal models, M. vaccae was completely devoid of any immunological reactivity, but despite this, it has been seriously considered by some as a potential immunogen. At the end of the day, however, it did not fare well in a controlled, clinical trial (another validation that results in mice can predict results in humans) [104 ]. Having said that, one recent study did suggest some clinical improvement in patients receiving M. vaccae [105 ].
|
VACCINE SAFETY TESTS |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
Obviously, the purpose of these stringent requirements is to minimize the risk to the vaccine. However, as Brennan and his colleagues probably correctly argue [106 ], given the devastating problem that TB has become, one must accept that testing new vaccines in humans may indeed incur some risk.
|
ACKNOWLEDGEMENTS |
|---|
Received January 9, 2001; revised March 1, 2001; accepted March 2, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTBACILLUS CALMETTE-GUERIN (BCG):...THE NEED FOR NEW...WHAT SHOULD A NEW...WHAT ARE THE TARGET...INNATE MECHANISMS: COULD THEY...THE FORGOTTEN LYMPHOCYTE: WHAT...BCG: IS IT GETTING...IS REPLACING BCG THE...TYPES OF NEW VACCINESVACCINE SAFETY TESTSREFERENCES |
|---|
