| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Dept of Microbiology, Unit of Experimental Tuberculosis, Fundació Institut per a la Investigació en Ciències de la Salut Germans Trias i Pujol, and 2 Dept of Internal Medicine, Pneumology Service, Hospital Universitari Germans Trias i Pujol, Universitat Autònoma de Barcelona, Badalona, Spain
CORRESPONDENCE: P-J. Cardona, Unitat de Tuberculosi Experimental, Servei de Microbiologia, Hospital Universitari "Germans Trias i Pujol", Crta del Canyet s/n, 08916-Badalona, Catalonia, Spain. Fax: 34 934978895. E-mail: pcardona@ns.hugtip.scs.es
Keywords: Animal model, Mycobacterium tuberculosis, pathogenesis, tuberculosis
Received: June 16, 2004
Accepted September 3, 2004
Abstract
Mycobacterium tuberculosis-latent bacilli are microorganisms that adaptto stressful conditions generated by the infected host against them. By slowing metabolism or becoming dormant, they may counterbalance these conditions and appear as silent to the immune system. Moreover, the dynamic turnover of the infected cells provokes a constant reactivation of the latent bacilli when the environmental conditions are favourable, or an activation after being dormant in necrotic and fibrotic lesions for a long period of time. Since there is no in vivo nor in vitro evidence for quick resuscitation of dormant bacilli, the current authors strongly favour the possibility that latent tuberculosis infection can be maintained for no longer than
10 yrs, which is, nowadays, a time period very close to that considered for "primary" tuberculosis. This concept may also be helpful for newer epidemiological considerations regarding the real impact of reinfection in tuberculosis. 8,000,000 newly diagnosed cases and 2,000,000–3,000,000 deaths by TB every year 1. In this paper, the phenomenon of latency of M. tuberculosis is reviewed, focusing on the evidence for latent bacilli in clinical studies and experimental in vitro and in vivo models (table 1
).
|
Clinical evidence of latent tuberculosis infection
The first evidence of LTBI was obtained with the treatment with antibiotics. After stating that TB should be treated simultaneously with at least two drugs in order to avoid spontaneous mutations, latency was the only explanation forthe late reactivation of TB in patients who thoroughly followed the antibiotic treatments and the absence of drug resistance in isolated bacilli 2. Soon after, chemoprophylaxis assays provided some idea about the nature of latency. In a trial conducted by the International Union Against Tuberculosis Committee on Prophylaxis, a protection of 93, 69 or 32% was demonstrated among "completer-compliers", following a period of treatment with isoniazid against LTBI for 12, 6 and 3 months, respectively 3. Since susceptibility to antibiotics requires some level of metabolism and cell growth, this trial suggested that many bacilli involved in LTBI are constantly growing.
More evidence came from the natural history of pulmonary TB in humans 30. Historically, "primary" TB appears when there is a formation of primary complexes by a parenchymal lesion, usually at the base of lungs, and by hilar lymphadenopathy, caused by a very low dose of infection (usually 1–5 tubercle bacilli) 5. Anatomically, the basal and middle zones of the lungs are more prone to infection than the apical zones because of their volume, although sometimes ventilation seems to be favoured in the latter 6. This is also supported by the finding of single calcified primary lesions observed in necropsies, in which 66% of infections were located in the lower half of the lung and only 12% were supraclavicular 7. Haematogenous dissemination takes place after the initial infection 8, and the tuberculin test gives positive results 3–8 weeks after infection. This primary complex resolves spontaneously with no symptoms in 95% of infected people, but 5% develop the disease, which may be local (i.e. causing pleurisy when there is rupture into the pleural cavity) or systemic (i.e. causing meningeal or even miliary TB).
A "post-primary" cavitary form of TB located at the apical zone of the lungs 9 has also been accepted and has traditionally been associated with the reactivation of an old lesion containing latent bacilli 10. To support this natural history, it has been stated that bacilli lie dormant in a metastatic site, haematogenously seeded within a vulnerable region (e.g. in the upper pulmonary zone). Unlike infections in lower zones of the lung, the immune system would not be able to sterilise the infectious foci, thus maintaining the bacilli in a dormant state, even for life 5. After the wane of immunity with time, which takes place mostly in the elderly at an estimated rate of 5% per year until complete disappearance of immunity 32, 33, the tubercle bacilli resume multiplication and increase their concentration in the apical focus. Once immunity is restored, the interaction with high quantities of antigen could lead to extensive caseation necrosis, liquefaction and cavity formation 5.
This post-primary pulmonary TB also accepts the presence of dormant bacilli constantly related with the immune system, waiting to reactivate due to immunosuppression. Nevertheless, other authors, such as Canetti 34, were skeptical about this idea because, in most cases, the primary complex is sterile within 5 yrs. Since this author considered the metastatic foci as a part of the primary complex and, thus, suffered its same fate, and taking into account that the bacillary concentration would be even lower in the metastatic foci than in the original foci, it was believed that an exogenous reinfection would be the origin of post-primary pulmonary TB.
The first efforts to determine the nature of latent bacilli: the in vitro experiments
Soon after the clinical observations that led to the concept of LTBI, a new "miraculous" drug appeared: rifampicin. This new drug allowed a shorter period of treatment and is still the gold standard for the treatment of TB 35. In order to explain why the use of rifampicin in the chemotherapeutic treatment of the disease could sterilise lesions in a remarkably short period of time compared with other drugs, Mitchison and Dickinson 11 observed that the culture of M. tuberculosis at 8°C in the presence of isoniazid or rifampicin did not affect bacillary concentration, whereas the culture at 37°C revealed a similar bactericidal capacity for both drugs. An "intermittent" incubation was then designed to demonstrate the higher bactericidal capacity of rifampicin. In fact, when a culture of M. tuberculosis at 8°C was incubated at 37°C for 6 h·day–1, rifampicin showed a higher bactericidal capacity than isoniazid. This experiment was the foundation of the theories on bacillary populations in TB lesions, based on thespeed of growth of bacilli 36, 37. Bacilli with a high metabolism were highly susceptible to chemotherapy. A medium speed of growth or "spurts of growth" were observed in those populations under acidic pH, or in bacilli sensitive to pyrazinamide or rifampicin but not to isoniazid. Finally, latent bacilli showed no metabolic activity and, thus, were not sensitive to chemotherapy. Therefore, post-primary infection or reactivation may be caused by the re-stimulation of the metabolism of these latent bacilli.
Tuberculous lesions are characterised by a consistent intragranulomatous necrosis, by compact macrophage and lymphocytic rings, and, finally, by a fibrotic layer. Taking into account the strict aerobic nature of M. tuberculosis when cultured in artificial media, Wayne 13 hypothesised that latent bacilli might adapt to microaerobic and anaerobic environments. They were also encouraged by the studies carried out by Corper and Cohn 12, who kept several sealed cultures of human isolates at 37°C for 12 yrs and demonstrated a survival of 0.01%. Wayne and Lin 14 conducted many experiments in artificial media to demonstrate the capacity of bacilli to adapt to oxygen-restricted conditions. In those experiments, it was shown that after aprogressive introduction to a low oxygen pressure, M.tuberculosis changed its metabolism by enhancing some enzymes, mainly isocitrate lyase and glycine dehydrogenase to generate a reduced nicotinamide adenine dinucleotide, so as to obtain energy through a fermentative metabolism 14. However, this hypothesis did not solve the problem on how bacilli may survive the stress generated by surrounding activated macrophages, a low pH, and a high concentration of radical oxygen intermediates (ROI) and radical nitrogen intermediates (RNI) 38. Latent bacilli must adapt to such conditions generated in activated infected macrophages. This hypothesis can also be criticised regarding the relative importance of low oxygen pressure in the growth of M. tuberculosis in host tissues. The microaerophilic conditions generated by Wayne and coworkers 13, 14 in artificial media already exist in the macrophages physiologically 15. Therefore, M. tuberculosis is usually well adapted to grow inthese conditions in vivo. Conversely, data obtained in anaerobiosis showed that bacilli did not survive for more than a few months 14. Moreover, ex vivo experiments conducted with macrophages demonstrated that a high oxygen pressure induced a higher bacillary growth than a low oxygen pressure 39, thus explaining why most cases of lung TB develop at the apex of the lungs, where oxygen concentrations are higher 6.
Different authors have also demonstrated higher resistance against different stress conditions of bacilli in the stationary phase of their growth compared with bacilli growing exponentially 26. The stationary phase in a conventional culture is known to appear when bacteria are starving due to a lack of nutrients or the accumulation of toxins generated by bacterial metabolism 16. This resistance capacity observed in all bacteria might reflect some kind of adaptation to hostile conditions, such as the ones generated in partially activated macrophages. The study of both the genomic and the proteomic expression of M. tuberculosis under stress conditions, such as acidity, low oxygen pressure, heat, cold, hydroxide peroxide, or even stationary growth phase, showed an increase in the expression of the RNA polymerase sigma (Sig)F unit, which was also related to the accumulation of an
-crystalline-like 16-kDa protein in the cell wall 17. Interestingly, the presence of SigF during exponential growth was deleterious for M. bovis bacilli Calmette-Guérin (BCG) 18, perhaps because it directed the transcription of genes required for the stationary phase or a spore-like state when exponential growth was required. Whether the M. tuberculosis SigF protein primarily regulates a sporulation-like cascade (the same as Bacillus subtilis SigF and Streptomyces coelicolor SigF) is unknown 19.
Concerning an in vitro model of latency, Kaprelyants etal. 20 distinguished three major physiological states of bacilli states: 1) viable (cultureable) bacilli that may divide (i.e. forming a colony on an agar plate or proliferate in liquid medium); 2) dormant bacilli with a low metabolic activity that are unable to divide without a preceding resuscitation phase; and 3) nonviable (noncultureable) bacilli that cannot divide. Kaprelyants et al. 20 did not consider the bacilli obtained by Wayne and Sramek 21 as dormant because they maintained a high viability and developed sensitivity to metronidazole when anaerobic, thus indicating an active metabolism. Nevertheless, prolonged cultures in stationary phase induced a true dormancy, generating "noncultureable" cells that had to be "resuscitated" before resuming active growing 22. Mukamolova et al. 23 also described the resuscitation promoting factor (Rpf) in supernatants of growing Micrococcus luteus cultures. Rpf restored the ability to grow in M. luteus dormant cells 23. This same test was repeated in dormant M. tuberculosis cells (from a 4-month-old culture), and it was found that recombinant Rpf from M.luteus and supernatant of growing M. tuberculosis cultures resuscitated dormant cells 22. Zhang et al. 24 repeated the test in a 1-yr-old culture, discovering that some phospholipids and an 8-kDa protein were responsible for resuscitation. However, this latter test can be criticised because dormant cells were generated after physiological entry into a stationary phase, giving enough time to the bacilli to perfectly adapt tothis environment, as reported previously by Corper and Cohn 12. It is difficult to extrapolate this to the real circumstances experienced by the bacilli inside macrophages, if they were to suddenly suffer stressful conditions like low pH, or ROI and RNI 38. Therefore, if the test conducted by Corper and Cohn 12 showed 0.01% survival following 12 yrs of incubation, a minor survival rate of bacilli under stress may be extrapolated.
Nonacid-fast, cell wall-defective variants of tubercle bacilli were isolated from clinical specimens and mycobacterial cultures 40. Far from the controversy generated about whether they really were M. tuberculosis or an environmental contaminant 41, or if they were really able to revert to the parent form or even to multiply 42, these forms seemed to be induced by the administration of antibiotics, as is the case with many other organisms 43. Consequently, the lack of cellwall provides some advantage in the resistance against chemotherapeutic treatment by eliminating the targets to which the drugs are directed, although they would be more susceptible to changes in the environment and the chances to survive in an inflammatory area would be even lower. Hence, such bacilli may play a limited role in LTBI.
The search for latency in experimental models in animals
The Cornell model of latency was the first experimental evidence of the existence of latent bacilli in an in vivo model, and it is widely considered to be the experimental source of latent bacilli. Mccune and Tompsett 25 described "the persistence of drug-sensitive tubercle bacilli in tissues despite prolonged antimicrobial therapy". After infecting the animals, a 12-week course of drug administration with isoniazid and pyrazinamide was started, and a complete disappearance of cultureable M. tuberculosis was achieved in all the animals sacrificed at the end of treatment. The most interesting data were obtained after 12 weeks, when cultureable bacilli were recovered in two-thirds of infected mice. This percentage increased to 100% with 1 mg·day–1 of cortisone for 20 days 44. Shleeva et al. 23 established a parallelism between this model and the one they generated in vitro after a long culture period. Unfortunately, the granuloma and the inflammatory response disappeared after chemotherapy, and, hence, the conditions of this model did not resemble those found in humans. Actually, it might have just been further evidence of the tolerance of mice to destroy M. tuberculosis, which was demonstrated by the triggering of a weak response only based on the control of a relevant concentration of growing bacilli 45.
Concerning the development of granulomas in a murine model of tuberculosis, the current authors' have observed how M. tuberculosis "escapes" from the granulomas (fig. 1). Infact, granulomas in mice are generated by an initial accumulation of macrophages in the infectious focus, which is surrounded by lymphocytes triggered by specific immunity. Subsequently, another cellular ring of foamy macrophages starts to surround these granulomas 46. This is a consequence of the migration of macrophages, filled with tissue debris and bacilli, to the alveolar spaces 47. Interestingly, most macrophages are activated either directly by specific lymphocytes or by ingesting cell wall components from M.tuberculosis, as evidenced by the expression of nitric oxide synthase (NOS) 48. It is difficult to explain why some of these cells have one or two bacilli inside. The answer may be that these bacilli withstand the bactericidal mechanisms by triggering a starvation response 16, thus stopping their growth. Taking into account that the immunity against M. tuberculosis seems to be directed against peptides synthesised by growing bacilli 49, 50, it may be hypothesised that these bacilli may not be recognised by the immune system. Another explanation may be related to the activated nature of these foamy macrophages, since, by synthesising NOS, these macrophages induce immunosuppression in any effector lymphocytes that lie in their vicinity 51. Therefore, these foamy cells may be a kind of a "sanctuary" for the stressed bacilli. Finally, some of them start to grow inside the foamy macrophages, until they are destroyed 45, 46, 48. This growth inside the alveolar space and outside the granulomas is especially harmful, as dissemination takes place very easily from this area. This might be the reason why all mice died due to an almost total occupation of the lung parenchyma 52, 53.
|
55, support the theory of a constant bacillary growth and destruction. Even the progressive occupation of the lung parenchyma during chronic infection 48, 52, 53 supports this idea. The findings in the murine model of TB must be considered with caution, since the human response against M. tuberculosis is not that tolerant, but the murine model might give us an idea about what happens at the beginning of infection when there is nowell-established fibrosis in granulomas and the chances fortrafficking material from these to the alveolar spaces is feasible. In addition, the fact that phospholipids may resuscitate dormant bacilli is interesting because these molecules are found throughout host tissues, rendering possible the reactivation of latent bacilli. This phenomenon would be very difficult to explain if only Rfp was able to resuscitate these cells, as it would require the previous presence of growing bacilli to synthesise Rfp. It may be hypothesised that a nontoxic environment with phospholipids, such as the one found inside foamy macrophages, is a suitable scenario to start reactivation. Once growth has started, the synthesis of Rfp would enhance the process.
Special attention must be given to studies using strains of M. tuberculosis with one specific gene deleted. The infection ofmice with those strains has provided a great amount of information on the role of each gene, regarding the capacity to survive inside a host. Out of these genes, the repression of isocitrate lyase (icl1 mutant) or Hsp70 heat shock protein (hspR mutant) affects bacillary persistence in host tissues 27, 28, showing that both genes are important for the generation of latency. ICL1 enhances the role of isocitrate lyase and theglyoxylate shunt to obtain carbon and energy for the metabolism of M. tuberculosis from the fatty acids of the host 29, instead of being a way to survive in an environment with a low oxygen pressure 14, whereas HSPR confirms the importance of counterbalancing the stress conditions generated against the bacillus.
How do latent bacilli deal with the dynamic nature of the pulmonary parenchyma?
Some authors have recently described the presence of M. tuberculosis DNA in human lung parenchyma inside endothelial and epithelial cells, or fibrocytes, and mainly outside granulomas 56, and they have stated that the bacilli would be responsible for the maintenance of LTBI. Without considering the limited significance of detecting DNA by in situ hibridisation in tissue 57–59, there is a paramount issue: the turnover of those pulmonary cells ranges 28–125 days 60. If we accept that latent bacilli are reclused within this niche, latency is limited as some energy would be required to periodically invade younger parenchymal cells. Therefore, latent bacilli must deal with this dynamic nature of the pulmonary parenchyma. Careful analysis of published results led the current authors to consider two possibilities: bacteria constantly disseminate and reactivate, as supported by the murine model; or the bacteria are kept dormant inside the necrotic material of a fibrotic granuloma, where the movement of macrophages would be limited for a long time until being finally reabsorbed, that is, if they are not calcified orbecome a scar. Obviously, this resuscitation should be veryfast, before the bacilli are drained out by the host. Experimental data from the in vitro model (i.e. from bacilli submitted to a stationary nonstressed culture) has shown that they require up to 4–5 months in the most ideal conditions 22 to reactivate them. At this point, bacilli would be drained out of the lungs. As a consequence, the idea of latent bacilli waiting for immunosuppression should be changed by a constant trend of bacilli to disseminate in order to reach an adequate environment for reactivation. If this reactivation takes place in an area where bacilli may grow quickly, such as the apex ofthe lungs, and where there is a lack of immunity, then acavitary lesion (and, thus, pulmonary TB, which is the evidence of LTBI) may develop.
The histopathological characteristics of human TB seem to suggest that intragranulomatous necrosis is induced at the beginning of granuloma formation 45 and, thus, adds extra stress for bacilli. This would explain why humans develop a significant population of extracellular latent bacilli (which is hardly seen in the murine model of experimental TB) that would be phagocytosed by the new macrophages. In this case, the lack of growth may be beneficial for the survival of bacilli, since they would not activate the new macrophages, and then they would be easily removed from the granuloma once the infected cell had become a foamy macrophage.
Many questions arise when the pathologies observed in mice and humans are compared, but probably the most important is whether the initial lesions are "cleaned" by macrophages and then surrounded by foamy macrophages, such as in other chronic inflammatory responses in the lung 61. The first results obtained from the current authors' studies carried out using material from autopsies of patients with TB seem to support this idea (data not published). Another concern is related to the time needed to become fibrotic and effectively close the granuloma in humans. This is not seen in mice, and that is why they die with extensive lung dissemination. The time taken to effectively turn a lesion fibrotic may help to establish the "risk period" to develop active TB.
Undoubtedly, the time for bronchogenic dissemination is limited, and it may be thought to be even more limited in the case of second-generation granulomas, since they would have developed under an immunological response, and the chances to grow would be lower than in the primary foci. Therefore, it is believed that chronicity also has a time limit. We must stress that, in order to be reactivated, latent bacilli must "escape" from the granuloma. Apparently, a low oxygen pressure and certain toxic materials from destroyed macrophages constitute a "nonresuscitating" medium. Moreover, once the granuloma is fibrotic or even calcified, the chances for these latent bacilli (probably dormant at this point) to escape are almost nonexistent. In this regard, nobody has demonstrated the ability of these bacilli to survive for years in an adverse environment. Conversely, classical studies showed that once intragranulomatous necrosis has been induced, the survival chances of bacilli decrease and are negligible after fibrosis andcalcification. Up to 50% of necrotic lesions and 85% of calcified or fibrotic lesions are sterile 62. Interestingly, Opie and Aronson 63 found that homogenates from fibrocaseous lesions in the upper areas of the lungs used to cause TB in guinea pigs, whereas homogenates from caseous encapsulated or calcified lesions rarely caused the disease. This study confirmed the difficulty in sterilising an infectious focus in the apical zone of the lung, and, consequently, both bacillary growth and inflammatory response were more significant, thus supporting the theory of a chronic active infection in this zone, rather than an induction of latency. Amazingly, in the same study, almost half of the samples from superficially normal lung tissue were infectious in guinea pigs. These data support a constant dissemination through the alveolar spaces.
What are the chances of reactivation after a single infection?
The diagnosis of LTBI is based on the tuberculin test. The existence of live bacilli is not necessary to retain a strong immune memory, since these cells live for long periods of time 64 and many people with LTBI have already killed the bacilli; in this case, many LTBI bacilli will never reactivate. Infection with M. tuberculosis triggers protective immunity. Ithas been estimated that BCG vaccination can induce immunity for 15–20 yrs 3, therefore suggesting a similar period of protection after M. tuberculosis infection. Considering the current authors' hypothesis that the constant "escape" of bacilli from granulomas before fibrosis is the primary source of bacteria, reactivation would never occur after a specific time period, unless the host suffered an immunosuppressive episode. Another question is whether the immune system would be able to stop bacillary growth in theupper zones of the lungs due to high oxygen pressure 65. The answer may be found in the classical literature. Since calcified primary lesions have also been detected in the upper zones of the lung 7, it seems clear that the immune system would be able to stop bacillary growth at this point.
The theory concerning the chances of developing the disease during life after a single infection needs to be reconsidered (table 2). The epidemiological data suggest that the risk of developing TB is higher immediately after infection with M. tuberculosis. Historical data from chemoprophylaxis trials using untreated TB-infected household contacts demonstrated that the disease occurred at a rate of 0.74%·yr–1 during years 1 and 2, 0.31%·yr–1 during years 3–5, and 0.16%·yr–1 during years 6 and 7 4. These data may be significantly adjusted to a decreasing linear regression, where the chances of developing TB would be nonexistent from 8 yrs after infection.
|
On the other hand, data from molecular fingerprinting of TB cases seem to give a renewed role to reinfection 68 compared with the almost impossibility suggested in historical studies 69. These positions need to be balanced definitively by looking at the epidemiological evidence. In populations with a high risk of infection, reinfection may be the major contributor to the rate of TB in adults. In populations with alow risk, the overall cases may probably be a result of reactivation 31.
In conclusion, latent Mycobacterium tuberculosis is a complex mixture of both slow metabolism and dormant bacilli (probably depending on the severity of the environmental stress suffered). In both cases, the fate of latent bacilli is determined by the dynamic physiology of the tissue where they remain. Thus, it seems feasible to suggest that there areonly two possible mechanisms to establish latency and promote disease late in life due to reactivation: constant reactivation once the stressful conditions have disappeared (e.g. when bacilli leave the granuloma inside foamy macrophages); or keeping dormant inside the necrotic tissues, waiting for a late drainage, and then resuscitate in a short period of time before being definitively removed from the host. In both cases, it seems important that reactivation takes place without any specific immunity against the bacteria and after reaching a privileged zone, where they would be able togrow as much as possible and to generate a strong inflammatory response that, in turn, would induce liquefaction and form a cavitary lesion. Since there is no in vivo nor invitro evidence for quick resuscitation of dormant bacilli, the current authors strongly favour the possibility that latent tuberculosis infection can be maintained for no longer than 10 yrs, which is, nowadays, a time period very close to that considered for "primary" tuberculosis.
Acknowledgements
The authors would like to thank M. Correia-Neves and R. Appelberg for their suggestions, and V. Ausina and S. Gordillo for their careful reading.
Footnotes
For editorial comments see page 895. 
References
This article has been cited by other articles:
|
|
 |
|
  J.A. Caminero and A. Torres Controversial topics in tuberculosis Eur. Respir. J., December 1, 2004; 24(6): 895 - 896. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
