- Email: [email protected]
Does programmed CTL proliferation optimize virus control?
Opinion
TRENDS in Immunology
Vol.26 No.6 June 2005
Does programmed CTL proliferation optimize virus control? Dominik Wodarz1 and Allan Randrup Thomsen2 1
Department of Ecology and Evolutionary Biology, 321 Steinhaus Hall, University of California, Irvine, CA 92697, USA Institute of Medical Microbiology and Immunology, University of Copenhagen, The Panum Institute, Build. 22.5.16, 3C Blegdamsvej, DK-2200 Denmark 2
CD8 T-cell or cytotoxic T-lymphocyte responses develop through an antigen-independent proliferation and differentiation program. This is in contrast to the previous thinking, which was that continuous antigenic stimulation was required. This Opinion discusses why nature has chosen the proliferation program and how it compares to continuous stimulation. Although the two mechanisms should not lead to significantly different dynamics during chronic infection, they do make a difference in acute infection. We argue that programmed proliferation is better at clearance, whereas continuous stimulation is better at limiting acute symptoms. The 7–10 programmed cell divisions observed in vivo might be an optimization of this trade-off. We also discuss the conditions under which the program does or does not require CD4 T-cell help for clearance. Introduction In recent years, immunological research has provided important new insights into the mechanisms by which cytotoxic T lymphocytes (CTLs or CD8 T cells) respond to infectious agents [1–5]. According to these data, a single encounter with antigen triggers a program of CTL expansion and differentiation, which is independent from further antigenic stimulation events. This is referred to as ‘programmed proliferation’ [6] (Figure 1). It is thought that the CTLs undergo w7–10 cell divisions, which result in the generation of effector cells and subsequently in their differentiation into memory cells [7,8]. If this does not result in the clearance of the pathogen, the memory CTLs are re-activated and further expansion occurs. These insights changed our view of how CTLs respond to antigenic stimulation. Previously, it was believed that continuous antigenic stimulation was required for CTL division and differentiation (Figure 1). We refer to this as the ‘continuous stimulation model’. It should be noted that this hypothesis of CTL expansion was not based on rigorous experimental observations but on common sense. These concepts have been characterized well at the cellular level [1–5]. However, questions remain regarding the ‘design principles’ of immune responses. In other words, why do we observe programmed CTL proliferation and what are the advantages of this over continuous Corresponding author: Wodarz, D. ([email protected]). Available online 22 April 2005
stimulation? Does programmed CTL proliferation optimize the ability of CTLs to control infections? In this Opinion, we discuss these questions and provide some hypotheses and speculations. These are based on previous work, in which we (and others) have combined mathematical modeling of immune responses and the analysis of in vivo CTL dynamics in mice. Although this approach must obviously involve a measure of oversimplification, it presents a powerful and unique tool to address such overall design questions. We have chosen to focus on viral infections because CTLs are essential for the clearance of most viral infections and viral pathogens are likely to represent a major evolutionary reason for the development of this cell type. We argue that there exists a trade-off between limiting the initial virus growth and the ability of the CTLs to drive the virus population to extinction. Continuous stimulation is the better strategy to limit peak virus load (which is reached during primary infection) and acute symptoms, whereas programmed proliferation is more efficient at clearing the virus. The larger the number of programmed divisions, the more efficient the response is at virus clearance, although the less efficient it is at catching up with a growing virus population during the acute phase. We speculate that the w7–10 cell divisions observed in vivo represent an optimum, which makes sure that infections can be cleared while the severity of acute symptoms is limited. Next, we consider the dynamics of persistent infections in the context of the two mechanisms. We argue that, in persistent infection, continuous stimulation and programmed proliferation have similar properties and similar effects on infection dynamics. Therefore, many of the conclusions reached in earlier studies, which assumed continuous stimulation of CTLs, remain valid. Finally, we discuss factors that determine whether or not the programmed response can clear an infection. This is done in the context of CD4 T-cell help. Comparing experimental data with mathematical modeling studies suggests those circumstances under which help is required for CTL-mediated resolution of a primary infection and those circumstances in which help is not required. Acute symptoms versus clearance The main point of our discussion is that there is a trade-off between the ability of CTLs to limit acute symptoms and
www.sciencedirect.com 1471-4906/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.it.2005.04.007
Opinion
306
TRENDS in Immunology
Vol.26 No.6 June 2005
(a)
Antigen Program
Program
Program
Program
+
•••
Naïve CTLp
Activated CTLp
Effector CTLs
Memory CTLs
Constant rate of division, independent from antigenic stimulation
(b)
Antigen
Antigen
+
Naïve CTLp
+
Activated CTLp
Antigen +
Antigen
Antigen
+
+
Rate of CTL expansion depends on antigenic stimulation
•••
Effector CTLs
Memory CTLs TRENDS in Immunology
Figure 1. Representations of (a) the programmed proliferation and (b) the continuous stimulation concepts. With programmed proliferation, a single encounter with antigen induces a program of cell division and differentiation, which results in the generation of effector and memory CTLs. With the assumption of continuous stimulation, cell division and differentiation requires the presence of antigen on a continuous basis. If antigen is removed, CTL proliferation and differentiation stops.
their ability to clear the infection. Continuous stimulation is better at limiting acute symptoms but worse at clearing the infection. Programmed proliferation is better at clearing the infection but worse at limiting acute symptoms. Here, we present this argument in detail and discuss it in the light of mathematical modeling studies. Let us consider acute symptoms first. We assume that the severity of acute symptoms correlates with the peak virus load, which is reached during primary infection (note that acute symptoms could also be brought about by immunopathology but this would correlate with relatively weak immunity and thus also be associated with relatively high virus loads [9]). During the first stages of infection, when the pathogen enters the host, the immunological reactions will follow similar principles both with continuous stimulation and with programmed proliferation [1,10–12]. That is, naı¨ve CTL precursors (CTLp) become activated on antigenic stimulation. How many of them become activated depends on a variety of factors. These include the route of infection [13,14] (e.g. peripheral versus intravenous), however, the most important factor could be the initial virus load [15]. Higher virus loads trigger more CTLp. After this activation event, however, programmed proliferation and continuous stimulation have different properties [1–5]. With continuous stimulation, the rate of CTL expansion increases as virus load increases. This www.sciencedirect.com
enables the response to catch up more efficiently with a growing virus population and limit peak loads (Figure 2). With programmed proliferation, the rate of CTL expansion is independent from antigenic stimulation and remains constant as virus load increases. This renders the response less efficient at catching up with a growing virus population and at limiting peak virus loads (Figure 2). Next, consider the post-acute phase of the infection. Again, there is an important difference between continuous stimulation and programmed proliferation. Consider continuous stimulation first. As the virus load declines, the reduced antigenic stimulation weakens the CTL response because CTL division depends on the amount of antigen [16,17]. This renders virus clearance more difficult and the virus population is likely to persist at low levels (Figure 2). This is nicely underlined by results from mathematical models, which assume continuous stimulation of CTLs. If the equilibrium virus load predicted by the model is low and indicates extinction (i.e. the virus dynamics converge on average to steady state levels, which are less than a single virus particle), the virus population settles at a significantly higher level in numerical simulations [16,17]. It declines at a slow rate (Figure 2) and only reaches extinction after a long period of time (this behavior is termed ‘quasi-equilibrium’ [16,17]). By contrast, with programmed proliferation, the CTL
Opinion
TRENDS in Immunology
Number of CTL5
Programmed proliferation Continuous stimulation
Virus load
Time
Quasi-equilibrium Time TRENDS in Immunology
Figure 2. Representation of acute infection dynamics assuming programmed proliferation and continuous stimulation. Continuous stimulation can catch up more efficiently with a growing virus population and can thus better limit acute virus load. However, it is less efficient at virus clearance because antigenic stimulation is lost at low virus loads. Mathematical work suggests that virus load is reduced to a ‘quasi-equilibrium’. That is, instead of being cleared, the virus population decays only with a slow rate and thus persists in the long term. Programmed proliferation is more efficient at virus clearance, however, although less efficient at catching up with a growing virus population and at limiting acute loads. This is because the rate of cell division in the context of the program is not influenced by the level of antigenic stimulation.
response does not weaken as virus load declines because the execution of the program is independent from antigenic stimulation. This greatly promotes virus clearance (Figure 2). Persistent infection dynamics Now, assume that the execution of the proliferation program does not result in the clearance of infection because the CTL response is weaker. Assume that persistent infection develops and that memory CTLs become re-activated by the persisting antigen and proliferate. If the virus persists in the long term, the CTLs will experience repeated antigen-dependent phases. This means that the CTL dynamics, in the context of the program and continuous stimulation, become similar: the response is continuously boosted by antigen. The only difference is that, with the program, the boosting events are separated by several divisions. In principle, this should not, however, change the chronic infection dynamics significantly. In fact, this scenario can be approximated with mathematical models that assume continuous antigenic stimulation [18]. Whether each cell division requires an antigenic stimulus or whether every nth cell division requires an antigenic stimulus does not make a significant difference in the context of a relatively weak CTL response, which fails to clear the infection. In both cases, the dynamics are expected to converge to the same steady states in the chronic phase of infection. Thus, we suggest that the properties of persistent infection do not depend significantly on the difference between the two mechanisms considered. www.sciencedirect.com
Vol.26 No.6 June 2005
307
It is also possible that chronic infection dynamics are slightly different than described here. Instead of repeatedly executing a full program, which results in the generation of central memory cells, the CTLs might get trapped in a cycle between effector cells and so called ‘effector memory’ cells, which can be triggered quickly by exposure to antigen [11]. In this case, the properties of programmed proliferation and continuous stimulation become even closer during chronic infection. What is the relevance of this comparison? Previous mathematical work on the dynamics of persistent infections assumed continuous stimulation [19–22]. Based on the discussion here, we argue that the insights gained from these studies are not obsolete but remain robust in the context of programmed CTL proliferation. We would like to illustrate this point with a specific example, which has been heavily debated in the literature: the role of effector molecules, such as interferon-g and perforin, in determining CTL homeostasis. In mice deficient in perforin or interferon-g (IFN-g), elevated levels of CTLs are observed [23–27]. Continuous stimulation models suggest that higher antigenic drive, brought about by reduced effector activity, accounts for the elevation of CTL levels [26,27]. However, it has been argued that this explanation cannot work in the context of programmed proliferation because the program is not influenced by virus load. Instead, it was proposed that the effector molecules somehow directly modulate the CTL response [23–25]. Although we cannot refute or validate this hypothesis with our current state of knowledge, our arguments suggest that programmed proliferation is not at odds with the hypothesis that higher levels of antigenic stimulation, owing to reduced effector activity, can explain the observation. This requires that memory cells become re-stimulated at least once before the infection is cleared. In this case, the absence of CTL effector molecules results in higher antigenic drive when the CTLs are re-stimulated and this leads to higher levels of CTLs, even in the context of programmed proliferation. In this context, it is interesting to consider recent results by Christensen et al. [28]. They studied CTL dynamics in vesicular stomatitis virus (VSV)-infected mice deficient in perforin and/or IFN-g. In the context of VSV, these molecules do not contribute substantially to effector activity, virus load or clearance. The experiments showed that perforin and IFN-g do not significantly regulate the level of virus-specific CTLs in this case. This result supports the notion that differences in the amount of antigenic stimulation, influenced by the antiviral effector activity of perforin and IFN-g, might account for the influence of these molecules on CTL homeostasis. CD4 T-cell help Earlier sections discussed virus clearance during acute infection, as well as the dynamics during persistent infection. The question then arises of under which circumstances each outcome is observed. Related to this is the topic of CD4 T-cell help. Although some infections (such as faster replicating LCMV strains) strictly require help for virus clearance, other infections (such as influenza virus infection in mice) can be resolved in the
308
Opinion
TRENDS in Immunology
Vol.26 No.6 June 2005
absence of help [29–33]. What is the role of help for viral clearance and under which conditions is help crucial? Over the last few years, significant advances have been made regarding the effect of CD4 T-cell help on CTLs. Experiments comparing infection dynamics during primary and secondary responses suggest that initial CTL expansion and differentiation in the acute phase do not depend on help [33–38]. The maintenance of memory and/or the re-activation of the memory cells, however, crucially depend on help. The mechanism underlying this observation is unclear [33,37–39]. It has been suggested that the helper cells somehow ‘program’ the CTLs during acute infection and that the subsequent availability of help has no influence [37,38]. However, a more recent study indicates that the presence of help might be continuously required to maintain memory in the long term [39]. With either mechanism, clearance can occur in the absence of help if the virus is eradicated during the first round of programmed proliferation and no re-activation of memory cells is needed (this might be the case in murine influenza virus infection [31]). If memory cells need to be reactivated, however, help is required for clearance. Now, we discuss the conditions in which these two outcomes are observed (note, we only discuss the basic principles of these dynamics here; viral evasion strategies, such as latency or escape mechanisms, will obviously promote viral persistence and the need to re-activate memory CTLs [40]).
Consider host parameters first. No doubt, a higher rate of CTL-mediated activity promotes clearance after one round of expansion. Similarly, it seems intuitive that a higher rate of CTL activation and proliferation promotes virus clearance. However, our previous discussion suggests that it can also decrease the chances to clear the virus (Figure 3a). The reason is that for high CTL activation and proliferation rates, the program is executed fast and completed before the virus is cleared. Thus, helper-dependent memory cells would have to be reactivated to complete clearance. This means that there is an optimal rate of CTL activation and proliferation, at which the chances of helper-independent clearance are maximized (Figure 3a). Although this is an interesting observation, it is unlikely that variation in the CTL activation and proliferation rate determines whether an infection can be resolved in the absence of help or not. It is more likely that this will depend on viral parameters, which are considered next. The main viral parameter that can influence whether an infection is cleared after one round of programmed proliferation is the replication rate of the virus. According to intuition, we would expect that faster viral replication correlates with virus persistence. This is, however, not the whole truth (Figure 3b): (i) at lower rates of viral replication, an increase in the replication kinetics might promote virus clearance in a helper-independent manner because it leads to higher acute virus loads. Consequently,
(a)
Slow activation and proliferation rate results in a response that is too weak to clear
Optimum
Fast activation and proliferation rate completes the program too fast, before clearance is achieved
Rate of CTL activation and proliferation
(b)
Very slow replication rate enables the virus to ‘sneak past’ the CTL response
Faster replication rate provides more antigenic stimulation and promotes clearance
Even faster replication rate does not increase acute virus load further; without increased stimulation, it only counters CTL-mediated activity
Viral replication rate TRENDS in Immunology
Figure 3. How can host and viral parameters influence the ability of a helper-independent CTL program to clear an infection? (a) Effect of the CTL activation and proliferation rate. An increase in the rate of CTL activation and proliferation can promote virus clearance because the response becomes more efficient. If the activation and proliferation rate is too high, however, the program is completed before the virus is cleared. Thus, clearance would require the re-activation of memory cells and CD4 T-cell help. Therefore, there is an optimal rate of CTL activation and proliferation, at which the chances of helper-independent virus clearance are maximized. (b) Effect of the viral replication rate. At low replication kinetics, an increase in the rate of viral replication can reduce the ability of a helper-independent program to clear an infection. This is because faster replication provides a higher antigenic stimulus, which triggers more CTLs. Slower replication triggers fewer CTLs and enables the virus to ‘sneak past’ the response. However, if the viral replication rate is higher, an increase in the replication kinetics counters the ability of the program to clear the infection without help. This is because an increase in the replication rate no longer results in higher acute virus loads and higher levels of antigenic stimulation because of target cell saturation. Thus, faster replication only counters the activity of the CTLs. www.sciencedirect.com
Opinion
TRENDS in Immunology
more CTLp are triggered. Reduced replication rates enable the virus to ‘sneak past’ the response by triggering fewer CTLp. It has been suggested that in hepatitis B and C virus infection, slower strains might ‘underwhelm’ the immune response and sneak past it during acute infection, increasing the chance of viral persistence [41]. (ii) By contrast, at high viral replication kinetics, an increase in the replication rate reduces the chances of clearing the virus (Figure 3b). This is because acute virus load has already reached maximum levels and cannot be increased any further by faster spread (due to target cell limitation). In this case, faster virus spread only counters CTL-mediated activity because a faster viral replication rate opposes the negative effect of CTL-mediated killing on the virus population. In this respect, it is interesting to consider LCMV infection. Although it is generally a fast replicating virus, LCMV exists as a collection of strains, which differ in their exact replication kinetics in vivo [42]. Because all strains spread relatively fast and result in high acute virus loads, our arguments suggest that slower strains should be cleared or controlled, whereas faster strains should establish persistent infection (Figure 4). In the context of a fast replicating strain and no CD4 T-cell help, our arguments further suggest that, initially, the virus population should be reduced to low levels before resurging to higher loads (Figure 4). This is because the absence of help only reduces the proliferation of memory CTLs but does not influence the first round of programmed proliferation. This is indeed observed [43,44]. The slower LCMV strains (e.g. Armstrong) are controlled in the absence of help [30]. Faster replicating strains, such as LCMV-Traub, are initially controlled but re-grow and (a) Slowest replicating strain: does not resurge (b) Intermediate replication rate of strain: resergence observed (c) Fast replicating strain: resurgence occures earlier
(c)
(b)
Virus load
(c) (b) (a)
Time TRENDS in Immunology
Figure 4. Representation of infection dynamics in CD4 Th-deficient hosts, assuming a collection of relatively fast replicating viruses. The virus population is initially reduced to low numbers because, in the acute phase, the CTL program is executed independent from help. The subsequent dynamics depend on the exact replication rate of the virus. (a) If the viral replication rate is lower and lies below a threshold, the virus is cleared or controlled in the long term. If the viral replication rate is faster, the virus population grows back. (b,c) The faster the replication kinetics, the faster the resurgence of the virus population. These dynamics are observed because the virus is not cleared after one round of programmed CTL proliferation and the absence of CD4 T-cell help prevents an efficient re-activation and expansion of the memory cells. www.sciencedirect.com
Vol.26 No.6 June 2005
309
persist [33]. The time it takes the virus to resurge correlates with its replicating rate [26] (Figure 4). Moreover, for slow LCMV strains, a deficiency in IFN-g shifts the dynamics in helper-deficient hosts away from virus control and towards persistent infection [43]. This is because the absence of IFN-g essentially enables the virus to replicate at a faster rate. Conclusion In this Opinion, we argue that there is a trade-off between virus clearance and the ability to contain acute infection. Programmed proliferation should be more efficient at virus clearance but less efficient at containing acute virus growth, whereas, continuous stimulation should be more efficient at containing acute virus growth but less efficient at virus clearance. We propose that the w7–10 programmed divisions observed in vivo might be the optimal solution to this trade-off. That is, there are enough programmed divisions to ensure that viruses can, in principle, be cleared, although there are not too many programmed divisions so that acute pathology is too high. Although acute pathology might be caused by the CTLs themselves as well as the virus (immunopathology), the occurrence of pathology tends to correlate with high virus loads in both cases [9,44]. Another solution to this trade-off could be to have the best of both worlds: that is, the existence of two types of CTLs, which respond by programmed division and continuous stimulation, respectively. There is, however, no evidence for such a separation among CTLs. In the context of persistent infections, we argue that the dynamics are similar with both the proliferation program and with continuous stimulation. Finally, we argue that CTLs can only resolve infections in the absence of help if clearance occurs after a single round of programmed proliferation, without the re-activation of memory CTLs. If memory CTLs need to be re-activated to achieve clearance, CD4 T-cell help is strictly required for the resolution of infection. In this case, absence of help would result in a temporary reduction of the virus load to a low level, followed by resurgence to high levels. The replication rate of the virus and the CTL activation and proliferation rates should be the most important parameters that determine whether help is required for CTL-mediated clearance. Further work should investigate in more detail the mechanisms by which memory CTLs become re-activated during chronic infection as well as on secondary challenge. In this context, it is of great interest that new results suggest that, although memory CD8 T cells following antigen clearance are sustained through cytokine-driven homeostatic proliferation, virus-specific CD8 T cells in persistently infected mice seem to be consistently activated, and this prevents them from acquiring the cardinal memory T-cell property of long-term antigen-independent persistence [45]. References 1 Kaech, S.M. and Ahmed, R. (2001) Memory CD8C T cell differentiation: initial antigen encounter triggers a developmental program in naı¨ve cells. Nat. Immunol. 2, 415–422
310
Opinion
TRENDS in Immunology
2 van Stipdonk, M.J. et al. (2001) Naı¨ve CTLs require a single brief period of antigenic stimulation for clonal expansion and differentiation. Nat. Immunol. 2, 423–429 3 van Stipdonk, M.J. et al. (2003) Dynamic programming of CD8C T lymphocyte responses. Nat. Immunol. 4, 361–365 4 Wong, P. and Pamer, E.G. (2001) Cutting edge: antigen-independent CD8 T cell proliferation. J. Immunol. 166, 5864–5868 5 Badovinac, V.P. et al. (2002) Programmed contraction of CD8C T cells after infection. Nat. Immunol. 3, 619–626 6 Mercado, R. et al. (2000) Early programming of T cell populations responding to bacterial infection. J. Immunol. 165, 6833–6839 7 Murali-Krishna, K. et al. (1998) Counting antigen-specific CD8 T cells: a reevaluation of bystander activation during viral infection. Immunity 8, 177–187 8 Butz, E.A. and Bevan, M.J. (1998) Massive expansion of antigenspecific CD8C T cells during an acute virus infection. Immunity 8, 167–175 9 Wodarz, D. and Krakauer, D.C. (2000) Defining CTL-induced pathology: implications for HIV. Virology 274, 94–104 10 Opferman, J.T. et al. (1999) Linear differentiation of cytotoxic effectors into memory T lymphocytes. Science 283, 1745–1748 11 Wherry, E.J. et al. (2003) Lineage relationship and protective immunity of memory CD8 T cell subsets. Nat. Immunol. 4, 225–234 12 Veiga-Fernandes, H. et al. (2000) Response of naı¨ve and memory CD8C T cells to antigen stimulation in vivo. Nat. Immunol. 1, 47–53 13 Kundig, T.M. et al. (1996) On the role of antigen in maintaining cytotoxic T-cell memory. Proc. Natl. Acad. Sci. U. S. A. 93, 9716–9723 14 Kundig, T.M. et al. (1996) On T cell memory: arguments for antigen dependence. Immunol. Rev. 150, 63–90 15 Badovinac, V.P. and Harty, J.T. (2002) CD8C T-cell homeostasis after infection: setting the ‘curve’. Microbes Infect. 4, 441–447 16 Wodarz, D. et al. (2000) The role of antigen-independent persistence of memory CTL. Int. Immunol. 12, 467–477 17 Wodarz, D. et al. (2000) A new theory of cytotoxic T-lymphocyte memory: implications for HIV treatment. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355, 329–343 18 Jansen, V.A.A. et al. (2005) Contrasting B cell and T cell based protective vaccines. J. Theor. Biol. 234, 39–48 19 De Boer, R.J. and Perelson, A.S. (1995) Towards a general function describing T cell proliferation. J. Theor. Biol. 175, 567–576 20 De Boer, R.J. and Perelson, A.S. (1998) Target cell limited and immune control models of HIV infection: a comparison. J. Theor. Biol. 190, 201–214 21 Perelson, A.S. (2002) Modelling viral and immune system dynamics. Nat. Rev. Immunol. 2, 28–36 22 Nowak, M.A. and Bangham, C.R. (1996) Population dynamics of immune responses to persistent viruses. Science 272, 74–79 23 Stepp, S.E. et al. (2000) Perforin: more than just an effector molecule. Immunol. Today 21, 254–256 24 Matloubian, M. et al. (1999) A role for perforin in downregulating T-cell responses during chronic viral infection. J. Virol. 73, 2527–2536 25 Badovinac, V.P. et al. (2000) Regulation of antigen-specific CD8C T cell homeostasis by perforin and interferon-g. Science 290, 1354–1358 26 Wodarz, D. (2001) Mechanisms underlying antigen-specific CD8C T cell homeostasis. Science 292, 595 27 Bartholdy, C. et al. (2000) Persistent virus infection despite chronic
Vol.26 No.6 June 2005
28
29
30
31
32 33
34 35
36
37
38 39
40
41
42
43
44 45
cytotoxic T-lymphocyte activation in g interferon-deficient mice infected with lymphocytic choriomeningitis virus. J. Virol. 74, 10304–10311 Christensen, J.E. et al. (2004) Perforin and IFN g do not significantly regulate the virus-specific CD8C T cell response in the absence of antiviral effector activity. Eur. J. Immunol. 34, 1389–1394 Eichelberger, M. et al. (1991) Clearance of influenza virus respiratory infection in mice lacking class I major histocompatibility complexrestricted CD8C T cells. J. Exp. Med. 174, 875–880 Ahmed, R. et al. (1988) T4C T helper cell function in vivo: differential requirement for induction of antiviral cytotoxic T-cell and antibody responses. J. Virol. 62, 2102–2106 Belz, G.T. et al. (2002) Compromised influenza virus-specific CD8C-T-cell memory in CD4C-T-cell-deficient mice. J. Virol. 76, 12388–12393 Zajac, A.J. et al. (1998) Viral immune evasion due to persistence of activated T cells without effector function. J. Exp. Med. 188, 2205–2213 Thomsen, A.R. et al. (1996) Exhaustion of CTL memory and recrudescence of viremia in lymphocytic choriomeningitis virusinfected MHC class II-deficient mice and B cell-deficient mice. J. Immunol. 157, 3074–3080 Janssen, E.M. et al. (2003) CD4C T cells are required for secondary expansion and memory in CD8C T lymphocytes. Nature 421, 852–856 Thomsen, A.R. et al. (1998) CD40 ligand is pivotal to efficient control of virus replication in mice infected with lymphocytic choriomeningitis virus. J. Immunol. 161, 4583–4590 Borrow, P. et al. (1998) CD40 ligand-mediated interactions are involved in the generation of memory CD8C cytotoxic T lymphocytes (CTL) but are not required for the maintenance of CTL memory following virus infection. J. Virol. 72, 7440–7449 Sun, J.C. and Bevan, M.J. (2003) Defective CD8 T cell memory following acute infection without CD4 T cell help. Science 300, 339–342 Shedlock, D.J. and Shen, H. (2003) Requirement for CD4 T cell help in generating functional CD8 T cell memory. Science 300, 337–339 Sun, J.C. et al. (2004) CD4C T cells are required for the maintenance, not programming, of memory CD8C T cells after acute infection. Nat. Immunol. 5, 927–933 Sarawar, S.R. et al. (2001) Stimulation via CD40 can substitute for CD4 T cell function in preventing reactivation of a latent herpesvirus. Proc. Natl. Acad. Sci. U. S. A. 98, 6325–6329 Bocharov, G. et al. (2004) Underwhelming the immune response: effect of slow virus growth on CD8C-T-lymphocyte responses. J. Virol. 78, 2247–2254 Moskophidis, D. et al. (1993) Virus persistence in acutely infected immunocompetent mice by exhaustion of antiviral cytotoxic effector T cells. Nature 362, 758–761 Christensen, J.P. et al. (2001) Depletion of CD4C T cells precipitates immunopathology in immunodeficient mice infected with a noncytocidal virus. J. Immunol. 166, 3384–3391 Thomsen, A.R. et al. (2000) Host factors influencing viral persistence. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355, 1031–1041 Wherry, E.J. et al. (2004) Antigen-independent memory CD8 T cells do not develop during chronic viral infection. Proc. Natl. Acad. Sci. U. S. A. 101, 16004–16009
Free journals for developing countries The WHO and six medical journal publishers have launched the Access to Research Initiative, which enables nearly 70 of the world’s poorest countries to gain free access to biomedical literature through the Internet. The science publishers, Blackwell, Elsevier, the Harcourt Worldwide STM group, Wolters Kluwer International Health and Science, Springer-Verlag and John Wiley, were approached by the WHO and the British Medical Journal in 2001. Initially, more than 1000 journals will be available for free or at significantly reduced prices to universities, medical schools, research and public institutions in developing countries. The second stage involves extending this initiative to institutions in other countries. Gro Harlem Brundtland, director-general for the WHO, said that this initiative was ’perhaps the biggest step ever taken towards reducing the health information gap between rich and poor countries’. See http://www.healthinternetwork.net for more information. www.sciencedirect.com
TRENDS in Immunology
Vol.26 No.6 June 2005
Does programmed CTL proliferation optimize virus control? Dominik Wodarz1 and Allan Randrup Thomsen2 1
Department of Ecology and Evolutionary Biology, 321 Steinhaus Hall, University of California, Irvine, CA 92697, USA Institute of Medical Microbiology and Immunology, University of Copenhagen, The Panum Institute, Build. 22.5.16, 3C Blegdamsvej, DK-2200 Denmark 2
CD8 T-cell or cytotoxic T-lymphocyte responses develop through an antigen-independent proliferation and differentiation program. This is in contrast to the previous thinking, which was that continuous antigenic stimulation was required. This Opinion discusses why nature has chosen the proliferation program and how it compares to continuous stimulation. Although the two mechanisms should not lead to significantly different dynamics during chronic infection, they do make a difference in acute infection. We argue that programmed proliferation is better at clearance, whereas continuous stimulation is better at limiting acute symptoms. The 7–10 programmed cell divisions observed in vivo might be an optimization of this trade-off. We also discuss the conditions under which the program does or does not require CD4 T-cell help for clearance. Introduction In recent years, immunological research has provided important new insights into the mechanisms by which cytotoxic T lymphocytes (CTLs or CD8 T cells) respond to infectious agents [1–5]. According to these data, a single encounter with antigen triggers a program of CTL expansion and differentiation, which is independent from further antigenic stimulation events. This is referred to as ‘programmed proliferation’ [6] (Figure 1). It is thought that the CTLs undergo w7–10 cell divisions, which result in the generation of effector cells and subsequently in their differentiation into memory cells [7,8]. If this does not result in the clearance of the pathogen, the memory CTLs are re-activated and further expansion occurs. These insights changed our view of how CTLs respond to antigenic stimulation. Previously, it was believed that continuous antigenic stimulation was required for CTL division and differentiation (Figure 1). We refer to this as the ‘continuous stimulation model’. It should be noted that this hypothesis of CTL expansion was not based on rigorous experimental observations but on common sense. These concepts have been characterized well at the cellular level [1–5]. However, questions remain regarding the ‘design principles’ of immune responses. In other words, why do we observe programmed CTL proliferation and what are the advantages of this over continuous Corresponding author: Wodarz, D. ([email protected]). Available online 22 April 2005
stimulation? Does programmed CTL proliferation optimize the ability of CTLs to control infections? In this Opinion, we discuss these questions and provide some hypotheses and speculations. These are based on previous work, in which we (and others) have combined mathematical modeling of immune responses and the analysis of in vivo CTL dynamics in mice. Although this approach must obviously involve a measure of oversimplification, it presents a powerful and unique tool to address such overall design questions. We have chosen to focus on viral infections because CTLs are essential for the clearance of most viral infections and viral pathogens are likely to represent a major evolutionary reason for the development of this cell type. We argue that there exists a trade-off between limiting the initial virus growth and the ability of the CTLs to drive the virus population to extinction. Continuous stimulation is the better strategy to limit peak virus load (which is reached during primary infection) and acute symptoms, whereas programmed proliferation is more efficient at clearing the virus. The larger the number of programmed divisions, the more efficient the response is at virus clearance, although the less efficient it is at catching up with a growing virus population during the acute phase. We speculate that the w7–10 cell divisions observed in vivo represent an optimum, which makes sure that infections can be cleared while the severity of acute symptoms is limited. Next, we consider the dynamics of persistent infections in the context of the two mechanisms. We argue that, in persistent infection, continuous stimulation and programmed proliferation have similar properties and similar effects on infection dynamics. Therefore, many of the conclusions reached in earlier studies, which assumed continuous stimulation of CTLs, remain valid. Finally, we discuss factors that determine whether or not the programmed response can clear an infection. This is done in the context of CD4 T-cell help. Comparing experimental data with mathematical modeling studies suggests those circumstances under which help is required for CTL-mediated resolution of a primary infection and those circumstances in which help is not required. Acute symptoms versus clearance The main point of our discussion is that there is a trade-off between the ability of CTLs to limit acute symptoms and
www.sciencedirect.com 1471-4906/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.it.2005.04.007
Opinion
306
TRENDS in Immunology
Vol.26 No.6 June 2005
(a)
Antigen Program
Program
Program
Program
+
•••
Naïve CTLp
Activated CTLp
Effector CTLs
Memory CTLs
Constant rate of division, independent from antigenic stimulation
(b)
Antigen
Antigen
+
Naïve CTLp
+
Activated CTLp
Antigen +
Antigen
Antigen
+
+
Rate of CTL expansion depends on antigenic stimulation
•••
Effector CTLs
Memory CTLs TRENDS in Immunology
Figure 1. Representations of (a) the programmed proliferation and (b) the continuous stimulation concepts. With programmed proliferation, a single encounter with antigen induces a program of cell division and differentiation, which results in the generation of effector and memory CTLs. With the assumption of continuous stimulation, cell division and differentiation requires the presence of antigen on a continuous basis. If antigen is removed, CTL proliferation and differentiation stops.
their ability to clear the infection. Continuous stimulation is better at limiting acute symptoms but worse at clearing the infection. Programmed proliferation is better at clearing the infection but worse at limiting acute symptoms. Here, we present this argument in detail and discuss it in the light of mathematical modeling studies. Let us consider acute symptoms first. We assume that the severity of acute symptoms correlates with the peak virus load, which is reached during primary infection (note that acute symptoms could also be brought about by immunopathology but this would correlate with relatively weak immunity and thus also be associated with relatively high virus loads [9]). During the first stages of infection, when the pathogen enters the host, the immunological reactions will follow similar principles both with continuous stimulation and with programmed proliferation [1,10–12]. That is, naı¨ve CTL precursors (CTLp) become activated on antigenic stimulation. How many of them become activated depends on a variety of factors. These include the route of infection [13,14] (e.g. peripheral versus intravenous), however, the most important factor could be the initial virus load [15]. Higher virus loads trigger more CTLp. After this activation event, however, programmed proliferation and continuous stimulation have different properties [1–5]. With continuous stimulation, the rate of CTL expansion increases as virus load increases. This www.sciencedirect.com
enables the response to catch up more efficiently with a growing virus population and limit peak loads (Figure 2). With programmed proliferation, the rate of CTL expansion is independent from antigenic stimulation and remains constant as virus load increases. This renders the response less efficient at catching up with a growing virus population and at limiting peak virus loads (Figure 2). Next, consider the post-acute phase of the infection. Again, there is an important difference between continuous stimulation and programmed proliferation. Consider continuous stimulation first. As the virus load declines, the reduced antigenic stimulation weakens the CTL response because CTL division depends on the amount of antigen [16,17]. This renders virus clearance more difficult and the virus population is likely to persist at low levels (Figure 2). This is nicely underlined by results from mathematical models, which assume continuous stimulation of CTLs. If the equilibrium virus load predicted by the model is low and indicates extinction (i.e. the virus dynamics converge on average to steady state levels, which are less than a single virus particle), the virus population settles at a significantly higher level in numerical simulations [16,17]. It declines at a slow rate (Figure 2) and only reaches extinction after a long period of time (this behavior is termed ‘quasi-equilibrium’ [16,17]). By contrast, with programmed proliferation, the CTL
Opinion
TRENDS in Immunology
Number of CTL5
Programmed proliferation Continuous stimulation
Virus load
Time
Quasi-equilibrium Time TRENDS in Immunology
Figure 2. Representation of acute infection dynamics assuming programmed proliferation and continuous stimulation. Continuous stimulation can catch up more efficiently with a growing virus population and can thus better limit acute virus load. However, it is less efficient at virus clearance because antigenic stimulation is lost at low virus loads. Mathematical work suggests that virus load is reduced to a ‘quasi-equilibrium’. That is, instead of being cleared, the virus population decays only with a slow rate and thus persists in the long term. Programmed proliferation is more efficient at virus clearance, however, although less efficient at catching up with a growing virus population and at limiting acute loads. This is because the rate of cell division in the context of the program is not influenced by the level of antigenic stimulation.
response does not weaken as virus load declines because the execution of the program is independent from antigenic stimulation. This greatly promotes virus clearance (Figure 2). Persistent infection dynamics Now, assume that the execution of the proliferation program does not result in the clearance of infection because the CTL response is weaker. Assume that persistent infection develops and that memory CTLs become re-activated by the persisting antigen and proliferate. If the virus persists in the long term, the CTLs will experience repeated antigen-dependent phases. This means that the CTL dynamics, in the context of the program and continuous stimulation, become similar: the response is continuously boosted by antigen. The only difference is that, with the program, the boosting events are separated by several divisions. In principle, this should not, however, change the chronic infection dynamics significantly. In fact, this scenario can be approximated with mathematical models that assume continuous antigenic stimulation [18]. Whether each cell division requires an antigenic stimulus or whether every nth cell division requires an antigenic stimulus does not make a significant difference in the context of a relatively weak CTL response, which fails to clear the infection. In both cases, the dynamics are expected to converge to the same steady states in the chronic phase of infection. Thus, we suggest that the properties of persistent infection do not depend significantly on the difference between the two mechanisms considered. www.sciencedirect.com
Vol.26 No.6 June 2005
307
It is also possible that chronic infection dynamics are slightly different than described here. Instead of repeatedly executing a full program, which results in the generation of central memory cells, the CTLs might get trapped in a cycle between effector cells and so called ‘effector memory’ cells, which can be triggered quickly by exposure to antigen [11]. In this case, the properties of programmed proliferation and continuous stimulation become even closer during chronic infection. What is the relevance of this comparison? Previous mathematical work on the dynamics of persistent infections assumed continuous stimulation [19–22]. Based on the discussion here, we argue that the insights gained from these studies are not obsolete but remain robust in the context of programmed CTL proliferation. We would like to illustrate this point with a specific example, which has been heavily debated in the literature: the role of effector molecules, such as interferon-g and perforin, in determining CTL homeostasis. In mice deficient in perforin or interferon-g (IFN-g), elevated levels of CTLs are observed [23–27]. Continuous stimulation models suggest that higher antigenic drive, brought about by reduced effector activity, accounts for the elevation of CTL levels [26,27]. However, it has been argued that this explanation cannot work in the context of programmed proliferation because the program is not influenced by virus load. Instead, it was proposed that the effector molecules somehow directly modulate the CTL response [23–25]. Although we cannot refute or validate this hypothesis with our current state of knowledge, our arguments suggest that programmed proliferation is not at odds with the hypothesis that higher levels of antigenic stimulation, owing to reduced effector activity, can explain the observation. This requires that memory cells become re-stimulated at least once before the infection is cleared. In this case, the absence of CTL effector molecules results in higher antigenic drive when the CTLs are re-stimulated and this leads to higher levels of CTLs, even in the context of programmed proliferation. In this context, it is interesting to consider recent results by Christensen et al. [28]. They studied CTL dynamics in vesicular stomatitis virus (VSV)-infected mice deficient in perforin and/or IFN-g. In the context of VSV, these molecules do not contribute substantially to effector activity, virus load or clearance. The experiments showed that perforin and IFN-g do not significantly regulate the level of virus-specific CTLs in this case. This result supports the notion that differences in the amount of antigenic stimulation, influenced by the antiviral effector activity of perforin and IFN-g, might account for the influence of these molecules on CTL homeostasis. CD4 T-cell help Earlier sections discussed virus clearance during acute infection, as well as the dynamics during persistent infection. The question then arises of under which circumstances each outcome is observed. Related to this is the topic of CD4 T-cell help. Although some infections (such as faster replicating LCMV strains) strictly require help for virus clearance, other infections (such as influenza virus infection in mice) can be resolved in the
308
Opinion
TRENDS in Immunology
Vol.26 No.6 June 2005
absence of help [29–33]. What is the role of help for viral clearance and under which conditions is help crucial? Over the last few years, significant advances have been made regarding the effect of CD4 T-cell help on CTLs. Experiments comparing infection dynamics during primary and secondary responses suggest that initial CTL expansion and differentiation in the acute phase do not depend on help [33–38]. The maintenance of memory and/or the re-activation of the memory cells, however, crucially depend on help. The mechanism underlying this observation is unclear [33,37–39]. It has been suggested that the helper cells somehow ‘program’ the CTLs during acute infection and that the subsequent availability of help has no influence [37,38]. However, a more recent study indicates that the presence of help might be continuously required to maintain memory in the long term [39]. With either mechanism, clearance can occur in the absence of help if the virus is eradicated during the first round of programmed proliferation and no re-activation of memory cells is needed (this might be the case in murine influenza virus infection [31]). If memory cells need to be reactivated, however, help is required for clearance. Now, we discuss the conditions in which these two outcomes are observed (note, we only discuss the basic principles of these dynamics here; viral evasion strategies, such as latency or escape mechanisms, will obviously promote viral persistence and the need to re-activate memory CTLs [40]).
Consider host parameters first. No doubt, a higher rate of CTL-mediated activity promotes clearance after one round of expansion. Similarly, it seems intuitive that a higher rate of CTL activation and proliferation promotes virus clearance. However, our previous discussion suggests that it can also decrease the chances to clear the virus (Figure 3a). The reason is that for high CTL activation and proliferation rates, the program is executed fast and completed before the virus is cleared. Thus, helper-dependent memory cells would have to be reactivated to complete clearance. This means that there is an optimal rate of CTL activation and proliferation, at which the chances of helper-independent clearance are maximized (Figure 3a). Although this is an interesting observation, it is unlikely that variation in the CTL activation and proliferation rate determines whether an infection can be resolved in the absence of help or not. It is more likely that this will depend on viral parameters, which are considered next. The main viral parameter that can influence whether an infection is cleared after one round of programmed proliferation is the replication rate of the virus. According to intuition, we would expect that faster viral replication correlates with virus persistence. This is, however, not the whole truth (Figure 3b): (i) at lower rates of viral replication, an increase in the replication kinetics might promote virus clearance in a helper-independent manner because it leads to higher acute virus loads. Consequently,
(a)
Slow activation and proliferation rate results in a response that is too weak to clear
Optimum
Fast activation and proliferation rate completes the program too fast, before clearance is achieved
Rate of CTL activation and proliferation
(b)
Very slow replication rate enables the virus to ‘sneak past’ the CTL response
Faster replication rate provides more antigenic stimulation and promotes clearance
Even faster replication rate does not increase acute virus load further; without increased stimulation, it only counters CTL-mediated activity
Viral replication rate TRENDS in Immunology
Figure 3. How can host and viral parameters influence the ability of a helper-independent CTL program to clear an infection? (a) Effect of the CTL activation and proliferation rate. An increase in the rate of CTL activation and proliferation can promote virus clearance because the response becomes more efficient. If the activation and proliferation rate is too high, however, the program is completed before the virus is cleared. Thus, clearance would require the re-activation of memory cells and CD4 T-cell help. Therefore, there is an optimal rate of CTL activation and proliferation, at which the chances of helper-independent virus clearance are maximized. (b) Effect of the viral replication rate. At low replication kinetics, an increase in the rate of viral replication can reduce the ability of a helper-independent program to clear an infection. This is because faster replication provides a higher antigenic stimulus, which triggers more CTLs. Slower replication triggers fewer CTLs and enables the virus to ‘sneak past’ the response. However, if the viral replication rate is higher, an increase in the replication kinetics counters the ability of the program to clear the infection without help. This is because an increase in the replication rate no longer results in higher acute virus loads and higher levels of antigenic stimulation because of target cell saturation. Thus, faster replication only counters the activity of the CTLs. www.sciencedirect.com
Opinion
TRENDS in Immunology
more CTLp are triggered. Reduced replication rates enable the virus to ‘sneak past’ the response by triggering fewer CTLp. It has been suggested that in hepatitis B and C virus infection, slower strains might ‘underwhelm’ the immune response and sneak past it during acute infection, increasing the chance of viral persistence [41]. (ii) By contrast, at high viral replication kinetics, an increase in the replication rate reduces the chances of clearing the virus (Figure 3b). This is because acute virus load has already reached maximum levels and cannot be increased any further by faster spread (due to target cell limitation). In this case, faster virus spread only counters CTL-mediated activity because a faster viral replication rate opposes the negative effect of CTL-mediated killing on the virus population. In this respect, it is interesting to consider LCMV infection. Although it is generally a fast replicating virus, LCMV exists as a collection of strains, which differ in their exact replication kinetics in vivo [42]. Because all strains spread relatively fast and result in high acute virus loads, our arguments suggest that slower strains should be cleared or controlled, whereas faster strains should establish persistent infection (Figure 4). In the context of a fast replicating strain and no CD4 T-cell help, our arguments further suggest that, initially, the virus population should be reduced to low levels before resurging to higher loads (Figure 4). This is because the absence of help only reduces the proliferation of memory CTLs but does not influence the first round of programmed proliferation. This is indeed observed [43,44]. The slower LCMV strains (e.g. Armstrong) are controlled in the absence of help [30]. Faster replicating strains, such as LCMV-Traub, are initially controlled but re-grow and (a) Slowest replicating strain: does not resurge (b) Intermediate replication rate of strain: resergence observed (c) Fast replicating strain: resurgence occures earlier
(c)
(b)
Virus load
(c) (b) (a)
Time TRENDS in Immunology
Figure 4. Representation of infection dynamics in CD4 Th-deficient hosts, assuming a collection of relatively fast replicating viruses. The virus population is initially reduced to low numbers because, in the acute phase, the CTL program is executed independent from help. The subsequent dynamics depend on the exact replication rate of the virus. (a) If the viral replication rate is lower and lies below a threshold, the virus is cleared or controlled in the long term. If the viral replication rate is faster, the virus population grows back. (b,c) The faster the replication kinetics, the faster the resurgence of the virus population. These dynamics are observed because the virus is not cleared after one round of programmed CTL proliferation and the absence of CD4 T-cell help prevents an efficient re-activation and expansion of the memory cells. www.sciencedirect.com
Vol.26 No.6 June 2005
309
persist [33]. The time it takes the virus to resurge correlates with its replicating rate [26] (Figure 4). Moreover, for slow LCMV strains, a deficiency in IFN-g shifts the dynamics in helper-deficient hosts away from virus control and towards persistent infection [43]. This is because the absence of IFN-g essentially enables the virus to replicate at a faster rate. Conclusion In this Opinion, we argue that there is a trade-off between virus clearance and the ability to contain acute infection. Programmed proliferation should be more efficient at virus clearance but less efficient at containing acute virus growth, whereas, continuous stimulation should be more efficient at containing acute virus growth but less efficient at virus clearance. We propose that the w7–10 programmed divisions observed in vivo might be the optimal solution to this trade-off. That is, there are enough programmed divisions to ensure that viruses can, in principle, be cleared, although there are not too many programmed divisions so that acute pathology is too high. Although acute pathology might be caused by the CTLs themselves as well as the virus (immunopathology), the occurrence of pathology tends to correlate with high virus loads in both cases [9,44]. Another solution to this trade-off could be to have the best of both worlds: that is, the existence of two types of CTLs, which respond by programmed division and continuous stimulation, respectively. There is, however, no evidence for such a separation among CTLs. In the context of persistent infections, we argue that the dynamics are similar with both the proliferation program and with continuous stimulation. Finally, we argue that CTLs can only resolve infections in the absence of help if clearance occurs after a single round of programmed proliferation, without the re-activation of memory CTLs. If memory CTLs need to be re-activated to achieve clearance, CD4 T-cell help is strictly required for the resolution of infection. In this case, absence of help would result in a temporary reduction of the virus load to a low level, followed by resurgence to high levels. The replication rate of the virus and the CTL activation and proliferation rates should be the most important parameters that determine whether help is required for CTL-mediated clearance. Further work should investigate in more detail the mechanisms by which memory CTLs become re-activated during chronic infection as well as on secondary challenge. In this context, it is of great interest that new results suggest that, although memory CD8 T cells following antigen clearance are sustained through cytokine-driven homeostatic proliferation, virus-specific CD8 T cells in persistently infected mice seem to be consistently activated, and this prevents them from acquiring the cardinal memory T-cell property of long-term antigen-independent persistence [45]. References 1 Kaech, S.M. and Ahmed, R. (2001) Memory CD8C T cell differentiation: initial antigen encounter triggers a developmental program in naı¨ve cells. Nat. Immunol. 2, 415–422
310
Opinion
TRENDS in Immunology
2 van Stipdonk, M.J. et al. (2001) Naı¨ve CTLs require a single brief period of antigenic stimulation for clonal expansion and differentiation. Nat. Immunol. 2, 423–429 3 van Stipdonk, M.J. et al. (2003) Dynamic programming of CD8C T lymphocyte responses. Nat. Immunol. 4, 361–365 4 Wong, P. and Pamer, E.G. (2001) Cutting edge: antigen-independent CD8 T cell proliferation. J. Immunol. 166, 5864–5868 5 Badovinac, V.P. et al. (2002) Programmed contraction of CD8C T cells after infection. Nat. Immunol. 3, 619–626 6 Mercado, R. et al. (2000) Early programming of T cell populations responding to bacterial infection. J. Immunol. 165, 6833–6839 7 Murali-Krishna, K. et al. (1998) Counting antigen-specific CD8 T cells: a reevaluation of bystander activation during viral infection. Immunity 8, 177–187 8 Butz, E.A. and Bevan, M.J. (1998) Massive expansion of antigenspecific CD8C T cells during an acute virus infection. Immunity 8, 167–175 9 Wodarz, D. and Krakauer, D.C. (2000) Defining CTL-induced pathology: implications for HIV. Virology 274, 94–104 10 Opferman, J.T. et al. (1999) Linear differentiation of cytotoxic effectors into memory T lymphocytes. Science 283, 1745–1748 11 Wherry, E.J. et al. (2003) Lineage relationship and protective immunity of memory CD8 T cell subsets. Nat. Immunol. 4, 225–234 12 Veiga-Fernandes, H. et al. (2000) Response of naı¨ve and memory CD8C T cells to antigen stimulation in vivo. Nat. Immunol. 1, 47–53 13 Kundig, T.M. et al. (1996) On the role of antigen in maintaining cytotoxic T-cell memory. Proc. Natl. Acad. Sci. U. S. A. 93, 9716–9723 14 Kundig, T.M. et al. (1996) On T cell memory: arguments for antigen dependence. Immunol. Rev. 150, 63–90 15 Badovinac, V.P. and Harty, J.T. (2002) CD8C T-cell homeostasis after infection: setting the ‘curve’. Microbes Infect. 4, 441–447 16 Wodarz, D. et al. (2000) The role of antigen-independent persistence of memory CTL. Int. Immunol. 12, 467–477 17 Wodarz, D. et al. (2000) A new theory of cytotoxic T-lymphocyte memory: implications for HIV treatment. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355, 329–343 18 Jansen, V.A.A. et al. (2005) Contrasting B cell and T cell based protective vaccines. J. Theor. Biol. 234, 39–48 19 De Boer, R.J. and Perelson, A.S. (1995) Towards a general function describing T cell proliferation. J. Theor. Biol. 175, 567–576 20 De Boer, R.J. and Perelson, A.S. (1998) Target cell limited and immune control models of HIV infection: a comparison. J. Theor. Biol. 190, 201–214 21 Perelson, A.S. (2002) Modelling viral and immune system dynamics. Nat. Rev. Immunol. 2, 28–36 22 Nowak, M.A. and Bangham, C.R. (1996) Population dynamics of immune responses to persistent viruses. Science 272, 74–79 23 Stepp, S.E. et al. (2000) Perforin: more than just an effector molecule. Immunol. Today 21, 254–256 24 Matloubian, M. et al. (1999) A role for perforin in downregulating T-cell responses during chronic viral infection. J. Virol. 73, 2527–2536 25 Badovinac, V.P. et al. (2000) Regulation of antigen-specific CD8C T cell homeostasis by perforin and interferon-g. Science 290, 1354–1358 26 Wodarz, D. (2001) Mechanisms underlying antigen-specific CD8C T cell homeostasis. Science 292, 595 27 Bartholdy, C. et al. (2000) Persistent virus infection despite chronic
Vol.26 No.6 June 2005
28
29
30
31
32 33
34 35
36
37
38 39
40
41
42
43
44 45
cytotoxic T-lymphocyte activation in g interferon-deficient mice infected with lymphocytic choriomeningitis virus. J. Virol. 74, 10304–10311 Christensen, J.E. et al. (2004) Perforin and IFN g do not significantly regulate the virus-specific CD8C T cell response in the absence of antiviral effector activity. Eur. J. Immunol. 34, 1389–1394 Eichelberger, M. et al. (1991) Clearance of influenza virus respiratory infection in mice lacking class I major histocompatibility complexrestricted CD8C T cells. J. Exp. Med. 174, 875–880 Ahmed, R. et al. (1988) T4C T helper cell function in vivo: differential requirement for induction of antiviral cytotoxic T-cell and antibody responses. J. Virol. 62, 2102–2106 Belz, G.T. et al. (2002) Compromised influenza virus-specific CD8C-T-cell memory in CD4C-T-cell-deficient mice. J. Virol. 76, 12388–12393 Zajac, A.J. et al. (1998) Viral immune evasion due to persistence of activated T cells without effector function. J. Exp. Med. 188, 2205–2213 Thomsen, A.R. et al. (1996) Exhaustion of CTL memory and recrudescence of viremia in lymphocytic choriomeningitis virusinfected MHC class II-deficient mice and B cell-deficient mice. J. Immunol. 157, 3074–3080 Janssen, E.M. et al. (2003) CD4C T cells are required for secondary expansion and memory in CD8C T lymphocytes. Nature 421, 852–856 Thomsen, A.R. et al. (1998) CD40 ligand is pivotal to efficient control of virus replication in mice infected with lymphocytic choriomeningitis virus. J. Immunol. 161, 4583–4590 Borrow, P. et al. (1998) CD40 ligand-mediated interactions are involved in the generation of memory CD8C cytotoxic T lymphocytes (CTL) but are not required for the maintenance of CTL memory following virus infection. J. Virol. 72, 7440–7449 Sun, J.C. and Bevan, M.J. (2003) Defective CD8 T cell memory following acute infection without CD4 T cell help. Science 300, 339–342 Shedlock, D.J. and Shen, H. (2003) Requirement for CD4 T cell help in generating functional CD8 T cell memory. Science 300, 337–339 Sun, J.C. et al. (2004) CD4C T cells are required for the maintenance, not programming, of memory CD8C T cells after acute infection. Nat. Immunol. 5, 927–933 Sarawar, S.R. et al. (2001) Stimulation via CD40 can substitute for CD4 T cell function in preventing reactivation of a latent herpesvirus. Proc. Natl. Acad. Sci. U. S. A. 98, 6325–6329 Bocharov, G. et al. (2004) Underwhelming the immune response: effect of slow virus growth on CD8C-T-lymphocyte responses. J. Virol. 78, 2247–2254 Moskophidis, D. et al. (1993) Virus persistence in acutely infected immunocompetent mice by exhaustion of antiviral cytotoxic effector T cells. Nature 362, 758–761 Christensen, J.P. et al. (2001) Depletion of CD4C T cells precipitates immunopathology in immunodeficient mice infected with a noncytocidal virus. J. Immunol. 166, 3384–3391 Thomsen, A.R. et al. (2000) Host factors influencing viral persistence. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355, 1031–1041 Wherry, E.J. et al. (2004) Antigen-independent memory CD8 T cells do not develop during chronic viral infection. Proc. Natl. Acad. Sci. U. S. A. 101, 16004–16009
Free journals for developing countries The WHO and six medical journal publishers have launched the Access to Research Initiative, which enables nearly 70 of the world’s poorest countries to gain free access to biomedical literature through the Internet. The science publishers, Blackwell, Elsevier, the Harcourt Worldwide STM group, Wolters Kluwer International Health and Science, Springer-Verlag and John Wiley, were approached by the WHO and the British Medical Journal in 2001. Initially, more than 1000 journals will be available for free or at significantly reduced prices to universities, medical schools, research and public institutions in developing countries. The second stage involves extending this initiative to institutions in other countries. Gro Harlem Brundtland, director-general for the WHO, said that this initiative was ’perhaps the biggest step ever taken towards reducing the health information gap between rich and poor countries’. See http://www.healthinternetwork.net for more information. www.sciencedirect.com