NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Frank SA. Immunology and Evolution of Infectious Disease. Princeton (NJ): Princeton University Press; 2002.

Cover of Immunology and Evolution of Infectious Disease

Immunology and Evolution of Infectious Disease.

Show details

Chapter 14Experimental Evolution: CTL Escape

CTL pressure favors amino acid substitutions in pathogen epitopes that escape recognition. Escape substitutions may avoid peptide processing and transport, reduce binding to MHC molecules, or lower affinity for the T cell receptor (Borrow and Shaw 1998). In this chapter, I discuss experimental evolution studies of CTL escape. I also discuss nonevolutionary studies that provide background or suggest promising experimental systems.

The first section reviews mechanisms of escape during peptide cleavage and transport. Two studies of murine leukemia virus describe single amino acid substitutions that changed patterns of peptide cleavage in cellular proteasomes. One substitution added a cleavage site within an epitope. Before the substitution, a significant amount of that epitope was transported to the endoplasmic reticulum, bound to MHC molecules, and presented to CTLs at the cell surface. The new cleavage site greatly reduced the abundance of the epitope available for MHC binding. A different substitution abrogated cleavage at the carboxyl terminus of an epitope, preventing transport of the peptide from the proteasome to the endoplasmic reticulum.

The second section describes escape from binding to MHC molecules during experimental evolution. Studies of influenza and lymphocytic choriomeningitis virus used a transgenic mouse with almost all of its CTLs specific for a single epitope, creating intense selective pressure for escape. Structural analyses of the peptide-MHC complex illuminate the biochemical mechanisms by which escape variants reduce binding to MHC. Experiments with simian immunodeficiency virus compared the spread of amino acid substitutions in several epitopes when in different host MHC genotypes. If the host had an MHC molecule that could present a particular epitope, that epitope was much more likely to evolve escape substitutions during infection.

The third section summarizes escape from binding to the T cell receptor (TCR). The experimental evolution studies of influenza and lymphocytic choriomeningitis with transgenic, monoclonal T cells yielded some TCR escape substitutions. Structural studies of peptide-MHC complexes and binding affinity studies of the complexes with TCRs clarified the biochemical mechanisms by which escape occurs.

The fourth section considers how the function of pathogen proteins may be altered by CTL escape substitutions. The Tax protein of human T cell leukemia virus type 1 provides a major target for CTL attack. Intense immune pressure selects for escape substitutions in naturally occurring infections. Tax plays a key role in many viral and cellular processes that affect viral fitness. Functional studies of Tax mutants suggest that substitutions reduce Tax performance. In HIV, amino acid substitutions in response to drugs sometimes increase binding to MHC molecules. It may be that the wild-type sequence reflects a balance between protein function and avoidance of MHC binding. Drugs or other experimental perturbations may upset that balance, exposing the mechanisms that mediate balancing selection.

The fifth section lists kinetic processes that determine the success or failure of escape variants. Kinetic processes connect the biochemical mechanisms of molecular interaction to the ultimate fitness consequences that shape observed patterns of antigenic variation. No experimental evolutionary studies have focused on these kinetic processes.

The final section highlights some topics for future research.

14.1. Cleavage and Transport of Peptides

Cellular proteasomes continuously chop up proteins into smaller peptides. Proteasomal cleavage patterns of proteins determine which peptides will be transported into the endoplasmic reticulum by transporter (TAP) molecules and made available for binding and presentation by the MHC system.

Single amino acid substitutions can affect proteasomal cleavage patterns (references in Beekman et al. 2000). Ossendorp et al. (1996) compared an eight-residue CTL epitope in two different murine leukemia viruses. Those epitopes differ by a K→R substitution in the first position. The R residue adds an additional, strong cleavage site within the epitope, causing a large reduction in the abundance of the R-containing epitope available for MHC presentation.

Cleavage does not occur instantaneously for all proteins. Instead, varying sites affect rates of cleavage and consequently relative abundances of different peptides. Ossendorp et al. (1996) present kinetic data for the accumulation of the K- and R-containing epitopes, and show that a large difference in abundance arises early in digestion, on the order of one hour.

Beekman et al. (2000) studied a different epitope in the same pair of murine leukemia viruses. In this case, an amino acid substitution at the residue flanking the C-terminus of the epitope affected both cleavage and transport. An aspartic acid residue at this position prevented cleavage precisely at the C-terminal site of the epitope and prevented transport by TAP. Thus, the epitope remained intact, but the peptide was not carried to the site of MHC binding.

These studies demonstrate that cleavage and transport can affect CTL response. But no data show how commonly amino acid substitutions abrogate effcient cleavage and transport. Experimental evolution studies could manipulate immunodominance and kinetic aspects of within-host infections to measure the frequency of the escape mechanism under different conditions.

14.2. MHC Binding

Amino acid substitutions can reduce affnity of epitopes for binding to MHC molecules. This may prevent MHC from transporting bound epitopes to the cell surface. Alternatively, peptide-MHC complexes may be presented on the cell surface, but higher o3-rates of the peptide reduce the opportunity for interaction with T cell receptors (TCRs). Several experimental evolution studies report mutations that reduce peptide-MHC binding.


Mice infected with lymphocytic choriomeningitis virus (LCMV) form the best-studied experimental system (Pircher et al. 1990; Weidt et al. 1995; Moskophidis and Zinkernagel 1995). LCMV is a noncytopathic RNA virus that naturally infects mice. The infection can be controlled or eliminated by a strong CTL response of the host. In H2-b mice, the MHC molecule Db presents the viral glycoprotein epitope GP33–43 and the MHC molecule Kb presents the overlapping GP34–43 epitope (Puglielli et al. 2001).

Genetically modified (transgenic) mouse lines have been developed that express a TCR specific for GP33–43 presented by MHC Db. Most (75–90%) of the CTLs in these mice express the TCR for GP33–43 (Pircher et al. 1990). These CTLs clear low doses of LCMV. High doses produce an initial viremia and subsequent decline under CTL pressure, followed by the appearance of CTL escape variants (Pircher et al. 1990). The rather extreme immunodominance of this experimental system provides a good model for studying molecular details of escape variants.

Puglielli et al. (2001) used this system to study amino acid substitutions in response to CTL pressure against Db-restricted GP33–43. They infected transgenic mice with high doses of LCMV virus. After the initial viremia and subsequent decline in titers in response to CTL pressure, viral titers increased. They isolated viruses from this later period to determine if escape variants had evolved and, if so, by what mechanism. These late viruses had a V→A substitution at the third position (site 35) of GP33–43 that nearly abolished binding to MHC Db.

Binding affnity of a peptide to MHC class I molecules typically depends on a small number of anchor residues in the peptide (Janeway et al. 1999). For example, an MHC molecule may have two anchor positions such that the fifth and ninth amino acids from the amino terminus (lower-numbered end) of the peptide determine the main portion of binding affnity. Structurally, such anchors may be pockets in the MHC molecule into which side chains from amino acids can fit. An amino acid with a side chain that fits well into the MHC pocket will bind with high affnity. A substitution in the peptide at an anchor position to a different amino acid with significantly altered shape or charge often diminishes or abolishes effective binding of the peptide-MHC complex. Substitutions at nonanchor residues usually have much smaller effects on binding affinity.

The third binding position of MHC Db is neither the primary nor auxiliary anchor residue according to previous studies (Rammensee et al. 1995). However, Tissot et al. (2000) solved the structure of Db bound to the LCMV wild-type epitope at GP33-43 (fig. 14.1). They found that the peptide residue at position three had its side chain buried in the Db binding cleft and, apparently, certain substitutions such as V→A at this location can disrupt binding in the manner of an anchor position (Puglielli et al. 2001).

Figure 14.1. Binding of LCMV epitope GP33-41 to the H-2 mouse MHC molecule Db.

Figure 14.1

Binding of LCMV epitope GP33-41 to the H-2 mouse MHC molecule Db. The nine amino acids of the epitope in positions 33–41 of the protein are labeled as P1–P9. (a) The epitope fits into the MHC binding groove. The bound epitope exposes P4 (more...)

Moskophidis and Zinkernagel (1995) studied the same system with H2-b mice and LCMV virus. Evolution within experimentally infected mice produced substitutions in immunodominant CTL epitopes. The substitutions N→D and N→S at position 280 of epitope GP275–289 abolished presentation by MHC molecule Db. This position forms an anchor residue for binding to Db.

Moskophidis and Zinkernagel (1995) also studied the V→L substitution at position 35 of GP32–42 originally obtained in the experimental evolution studies of Pircher et al. (1990). Both Db and Kb can present this epitope. The V→L change, which occurred in a nonanchor residue for both MHC molecules, did not significantly reduce binding to either MHC molecule. The substitution did significantly reduce CTLs directed against this epitope when presented by Db.

From these studies, two different substitutions at GP35 were analyzed (position P3 in fig. 14.1). Puglielli et al. (2001) found that a V→A substitution at GP35 abrogated binding to Db, whereas Moskophidis and Zinkernagel (1995) found that a V→L change did not affect binding to Db. Although this site was previously considered as a nonanchor residue, the L and A substitutions had significantly different effects on MHC binding. Interestingly, the V→L substitution, which did not affect binding affinity to the MHC molecule, did reduce the affinity of the peptide-MHC complex for a TCR (see section 14.3 below, TCR Binding).


Several authors suggest that HIV escape from CTLs plays an important role in the dynamics and persistence of infection within individual hosts (e.g., McMichael and Phillips 1997; Nowak and May 2000). Many lines of evidence from human hosts support this argument. But human infections cannot be controlled or manipulated experimentally. Several experimental evolution studies of simian immunodeficiency virus (SIV) infections in rhesus macaques have recently been published.

Allen et al. (2000) infected eighteen rhesus macaques with cloned SIV. All ten hosts expressing the MHC class I molecule Mamu-A*01 made CTLs to Tat28–35, the SL8 epitope in the Tat protein. Up to 10% of circulating CTLs recognized this epitope during the acute phase of viremia three to four weeks after infection. However, the frequency of Tat-specific CTLs dropped sharply after the acute phase, suggesting escape from recognition.

Sequencing at eight weeks after infection showed that five of ten Mamu-A*01 positive animals had mutations in the SL8 epitope, with little variation outside of this epitope. By contrast, only one of eight Mamu-A*01 negative hosts had mutations in the SL8 epitope. Four of the amino acid substitutions in SL8 effectively destroyed binding to Mamu-A*01. Three of these substitutions occurred at position 8, the primary anchor site, and one substitution occurred at position 2, the secondary anchor site. Two other substitutions reduced binding by less than two orders of magnitude: a substitution at position 1 reduced binding by 67%, and a substitution at position 5 reduced binding by 85%.

Z. W. Chen et al. (2000) observed CTL escape in SIV infections of rhesus macaques mediated by a single T→A substitution in an epitope of the Gag protein. The Mamu-A*01 MHC molecule presents this Gag181–189 epitope on the cell surface. The substitution did not affect binding affinity of Mamu-A*01 for Gag181–189. However, the peptide-MHC complexes could stimulate CTL response only when in vitro target cells where experimentally pulsed with high concentrations of the mutated epitope. Z. W. Chen et al. (2000) conclude that the substitution increases the off-rate of binding, causing a high dissociation rate of peptide-MHC complexes on the cell surface. High experimental concentrations of the epitope in vitro may overcome the high dissociation rates and provide enough peptide-MHC complexes on the cell surface to stimulate CTLs.

Many studies of CTL escape either use nearly monoclonal T cells or follow only one viral epitope. This leaves open the problem of whether polyclonal CTL responses to multiple epitopes favor escape in the same way as CTL responses to a single epitope. Evans et al. (1999) addressed this issue by following the within-host evolution of five distinct CTL epitopes of the Env and Nef proteins in five rhesus macaques experimentally infected by SIV. All hosts had the same MHC class II genotype and thus similar presentation of epitopes to helper T cells.

Hosts B and B′ had the same MHC class I genotype and progressed rapidly to disease without making a strong CTL response. The other three hosts made strong initial CTL responses. Hosts A and D progressed slowly to disease, whereas host C progressed at an intermediate rate. The intermediate progressor, host C, differed from the fast progressors by having MHC class I molecules Mamu-A*11 and Mamu-B*17. These MHC molecules presented two epitopes, Env497–504 and Nef165–173. One slow progressor, host A, differed from the fast progressors by having MHC class I molecules Mamu-B*03 and Mamu-B*04, which present three epitopes, Env575–583, Nef136–146, and Nef62–70. The other slow progressor, host D, had all four class I molecules listed for hosts A and C, and presented all five epitopes.

Evans et al. (1999) sequenced env and nef genes at various times during the course of infection. For the five epitopes listed above, each host had viral populations dominated by escape mutants only for those epitopes that they could present by their class I MHC molecules. For example, host C viruses were dominated by escape mutants in Env497–504 and Nef165–173 but not in the other three epitopes. This demonstrates selective pressure on multiple epitopes, defined by MHC class I presentation. Some of the escape variants abolished MHC binding to the epitopes, whereas others apparently reduced TCR recognition.


Price et al. (2000) infected mice with a human isolate of influenza. These transgenic mice expressed a monoclonal TCR specific for the influenza nucleoprotein epitope NP366–374, leading to a narrow and strong CTL response directed against this epitope. This intense selective pressure favored escape variants in this nine-residue epitope at positions 5, 6, 7, and 9, each escape variant having only one altered position.

Young et al. (1994) described the structure of the MHC Db molecule bound to this NP epitope. This structural information allowed Price et al. (2000) to interpret the substitutions they observed in response to CTL pressure. Positions 5 and 9 form anchor sites buried in the MHC binding groove. Three different amino acid replacements at position 5 greatly reduced binding affinity for Db and consequently abrogated CTL stimulation.

A substitution at position 9 reduced affinity for Db less than 10-fold. In spite of the relatively small change in binding affinity for MHC, this substitution also abolished CTL response. Price et al. (2000) cite data to suggest that peptide processing and transport do not play a role, so the mechanism of escape by this substitution remains unclear.

14.3. TCR Binding

I continue with the influenza study by Price et al. (2000), which ended the previous section. I then return to the LCMV experimental system, which provides the first combined information on structure, binding affinity, and escape mutations with respect to peptide-MHC interactions with the TCR.


The structural study of Young et al. (1994) demonstrated that binding between MHC Db and the NP366–374 epitope exposes positions 6 and 7. These exposed positions could potentially interact with the TCR. Haanen et al. (1999) showed a dominant role for position 7 by tetramer staining of TCR bound to peptide-MHC complexes. These previous structural and binding studies did not implicate position 6 as important for TCR affinity.

Price et al. (2000) observed three substitutions at positions 6 and one at position 7 that bound to Db with the same affinity as the wild type. These escape variants avoided binding by the transgenic, monoclonal CTL.

Price et al. (2000) compared the ability of the wild type and an M→I substitution at position 6 to stimulate a CTL response in immunocompetent mice with a polyclonal repertoire. The M→I substitution attracted 3–10-fold fewer CTLs than did the wild type. Thus, this substitution at position 6 that escaped the transgenic monoclonal CTLs did not abolish polyclonal CTL stimulation. These observations suggest that the total TCR repertoire includes a set of overlapping specificities with varying affinity, allowing recognition of the altered ligand.


In the section above on MHC binding, I described the V→L substitution of LCMV at GP35 (P3) that provided escape from transgenic CTLs in experimental infections of mice (Moskophidis and Zinkernagel 1995). This substitution did not significantly reduce binding affinity of the GP33–41 epitope for the MHC molecule Db. Moskophidis and Zinkernagel (1995) concluded that this substitution interfered with stimulation of CTLs by the peptide-MHC complex.

Tissot et al. (2000) analyzed the structure and affinity of the peptide-MHC complex bound to the same TCR used by Moskophidis and Zinkernagel (1995). The V→L substitution at P3 reduced binding affinity to the TCR by a factor of 50, even though P3 is buried in the peptide-MHC binding groove (fig. 14.1). Tissot et al. (2000) suggest that the relatively bulky leucine reside caused shifting of the structure and a slight movement in residues P2 and P4. The residue at P4 is exposed and had the strongest effect on binding affinity to the TCR, so movement of P4 could be responsible for the change in affinity. Escape variants with a Y→F substitution at P4 obtained during experimental evolution in vivo cause a 100-fold decline in affinity for the TCR.

14.4. Functional Consequences of Escape

Escape substitutions change amino acids in pathogen proteins. Those changed proteins may have altered performance, affecting pathogen fitness in ways other than CTL escape. Ideally, experimental studies of escape would provide information about changed functional characteristics of pathogen proteins and the associated fitness consequences. I am not aware of any such analyses for CTL escape in experimentally controlled systems. Two uncontrolled studies provide some clues.

First, Niewiesk et al. (1995) analyzed CTL escape variants of human T cell leukemia virus (HTLV-1) in naturally infected human hosts. They focused on the Tax protein, a major target of CTLs. Individuals with MHC type HLA-A2 simultaneously recognize at least five epitopes of Tax (Parker et al. 1994). Niewiesk et al. (1995) found that in HLA-A2 subjects, 24 of 179 isolates had substitutions in epitopes presented by HLA-A2. By contrast, in subjects without HLA-A2, only one of 116 of these epitopes had a substitution. CTLs appear to be imposing strong selective pressure that favors escape. Nine different substitutions occurred across the five Tax epitopes. Each substitution abolished CTL attack of the associated epitope.

HTLV-1 is a retrovirus that integrates itself into the host genome. The Tax protein is a trans-acting transcriptional regulator that modulates expression of several viral and cellular genes (Yoshida 2001). Because HTLV-1 typically occurs as an integrated provirus in host cells, viral replication occurs by transmission within the lineages of host cells and by transmission between cells. Tax appears to affect several aspects of the cell cycle, potentially enhancing cell division and reducing cell death.

Niewiesk et al. (1995) tested the nine observed Tax substitutions for potency as activators of one viral and two cellular promoters of transcription. Potencies were compared with activation by a consensus sequence. Three substitutions had lowered ability to activate the viral promoter, and all nine substitutions caused lowered or no activation of two cellular promoters. The fitness consequences of these substitutions could not be measured directly. In vitro studies introduced mutations into the Tax protein and demonstrated that most mutations abolished Tax function (Smith and Greene 1990; Semmes and Jeang 1992). Thus, Tax appears to be highly constrained, suggesting that substitutions accumulate only under very strong CTL pressure.

A second study analyzed the selective pressure imposed by a drug (Samri et al. 2000). This study of human patients with HIV compared the viral reverse transcriptase (RT) protein before and after application of nucleoside inhibitors of RT. Substitutions in RT that escape drug pressure also reduce viral fitness (Coffin 1995; Back et al. 1996; Harrigan et al. 1998; Sharma and Crumpacker 1999). Samri et al. (2000) showed that, on average, drug escape mutants increased CTL recognition, most likely by enhanced binding to common MHC molecules.

Samri et al.'s (2000) preliminary study raises some interesting problems. Amino acid sequences of viral proteins may be shaped by two opposing pressures: contribution to viral function and escape from immune recognition. Thus, amino acid substitutions in response to a third force, such as a drug, may be likely to reduce protein performance or enhance recognition by the host immune system. In the case of RT, both reduced performance and enhanced MHC recognition may have occurred.

A particular viral sequence reflects the balance between functional performance and avoidance of CTL recognition via MHC presentation. Experimentally applied selective pressures such as drugs may provide information about the functional and immune selective pressures that shaped the wild-type sequence.

14.5. Kinetics of Escape

Experimental evolutionary studies have not focused on the kinetics by which escape variants arise and spread within hosts or within populations. I briefly list six issues.


The first experimentally controlled studies of escape from CTLs used extreme immunodominance (Pircher et al. 1990). In that system, genetically constructed mice produced the identical TCR on 75–90% of circulating T cells. That extreme, monoclonal TCR distribution creates powerful selection favoring escape mutants for epitopes recognized by the dominant TCR.

More realistic polyclonal distributions of TCRs may not favor escape so easily (Borrow and Shaw 1998; Haanen et al. 1999). A single viral mutation can abrogate recognition of a particular epitope, but the virus carrying the mutant will likely express other epitopes recognized by different CTLs. By this argument, partial escape means partial recognition and death.

The degree of immunodominance plays an important role. For some pathogens and hosts, a typical response may primarily target a single epitope, with fewer CTLs focusing on subdominant epitopes. In this case, the pressure on the lead epitope favors escape. Other infections may have a broader and more even CTL response against several epitopes. Escape at one epitope does not alleviate recognition at several other epitopes. However, escape at multiple epitopes may be observed within individual hosts (Evans et al. 1999). The role of immunodominance in escape depends on the rate of killing by CTLs relative to the rate of viral transmission between cells (McMichael and Phillips 1997; Nowak and May 2000).

Rate of Killing versus Rate of Transmission

Consider a cytopathic virus—one that bursts its host cell when liberating progeny virions. A CTL escape mutant gains if it enhances the probability of cellular burst before CTL-mediated death. This probability depends on the race between the CTLs to kill infected cells and the viruses to liberate progeny. Escape at a dominant epitope provides benefit if the aggregate rate of killing via subdominant epitopes allows a higher probability of burst before death.

Noncytopathic viruses leak progeny virions from intact host cells. Here the race occurs between, on the one hand, CTL-induced death and, on the other hand, the time before the first viral progeny release and then the subsequent rate of progeny production. CTL escape has no benefit if pressure on other epitopes still kills before initial progeny production. If some infected cells survive to produce new virions, the benefit of escape at one epitope depends on the expected increase in cellular longevity during the productive phase of virion release and the probability that released viruses transmit to new host cells.

The roles of these different rate processes could be combined into a mathematical model by extending the approach of Nowak and May (2000). Experimental control over TCR diversity, CTL intensity, and comparison of cytopathic and noncytopathic viruses could provide tests of the mechanistic processes contained in the mathematical formulation.

Multiplicity of Infection

Higher multiplicity of infection may reduce the rate at which escape mutants spread (McMichael and Phillips 1997). Suppose two viruses infect a host cell. One virus expresses an escape mutant that avoids recognition by the dominant CTLs, whereas the other virus expresses the common epitope recognized by the dominant CTLs. This dually infected cell presents the common epitope on its surface, making it susceptible to recognition by CTLs. The escape mutant benefits only to the extent that fewer recognized peptides occur on the cell surface—lower density may reduce the rate of killing, and that reduction may in turn allow more of the escape variant's progeny to be transmitted.

Dosage and Population Size within Hosts

In experimental studies, escape mutants arise more often as infecting dose increases (Pircher et al. 1990). At low doses, the host clears infection before escape mutants spread. Higher dose most likely produces larger population size during the initial viremia, increasing the time and the number of pathogens available to make a particular mutant.

Experimental manipulations could test the contributions of dosage, pathogen population size within the host, and time to clearance. Waiting time for an escape mutant also depends on the mutation rate, which could perhaps be varied by comparing genotypes that differed in mutation rate.

Rate of Clearance and Transmission between Hosts

Rapidly cleared infections provide little opportunity for the transmission of escape variants. Such variants spread within hosts only after intense CTL pressure begins. If the infection clears rapidly, then the potential escape variants do not increase sufficiently within the host to contribute significantly to transmission to other hosts.

MHC Polymorphism and Transmission between Hosts

CTL escape may depend on reduced binding to an MHC molecule or on changed affinity between TCRs and the peptide-MHC complex. Hosts vary in the highly polymorphic MHC alleles. Escape from one host's MHC binding does not provide protection if most other hosts carry different MHC alleles. Long-term spread of an MHC escape variant depends on the frequency at which viruses encounter the particular MHC molecule and the intensity of forces against the escape variant. Such countering forces include creation of new sequences that bind well to different MHC molecules and functional attributes that affect the performance of the viral protein.

Further insight may be gained by multiple passage experiments. Viruses could be passed through a sequence of hosts with either the same or varying MHC genotypes. The changing frequencies of amino acid substitutions could be tracked under different regimes of fluctuating selection.

14.6. Problems for Future Research

1. Immunodominance and timing of epitope expression

Infected cells express different pathogen proteins at different times. Proteins expressed relatively early may be more likely to attract a dominant CTL response because they occur on the cell surface for longer periods of time. The role of timing could be studied in the following experimental evolution design.

Create a host with a biclonal TCR repertoire focused on two different pathogen epitopes, one expressed early and the other expressed late. Also create two different hosts with monoclonal TCRs, one focused on the early epitope and the other focused on the late epitope. Finally, create a host immunodeficient in CTL response. Infect the four different host types with cloned genotypes of a pathogen. If early expressed epitopes attract stronger CTL responses, then the biclonal host should induce fewer escape substitutions in the late expressed epitope than the monoclonal host focused on the late epitope. The relative escape rates in the monoclonal hosts focused on early and late epitopes calibrate escape rates in the absence of competition between epitopes.

A similar design may be accomplished by using different host MHC genotypes to turn on or off the pressure on epitopes expressed at different times.

2. Immunodominance and probability of escape

Pressure from a single kind of MAb often gives rise to antibody escape mutants. By contrast, simultaneous pressure from two or more MAbs can prevent spread of escape substitutions because a pathogen needs to escape simultaneously from multiple killing agents. A similar experiment can be developed for CTLs by using different MHC host genotypes. In experimental evolution studies, hosts that can effectively present a broader variety of epitopes should restrict the spread of escape substitutions relative to hosts with narrower presentation.

3. Performance versus avoiding MHC recognition

Substitutions that avoid MHC recognition may reduce other components of fitness. This hypothesis can be tested by evolving pathogens in different regimes of MHC-mediated selection and then competing the evolved pathogens to determine relative fitness. In this design, one starts with a cloned pathogen genotype. Some lineages can be passaged in a sequence of hosts with the same MHC genotype, others in hosts with varying MHC genotypes, and controls in hosts immunodeficient for CTL response. Competition between various pairs of evolved pathogens can be used to estimate the fitness costs of substitutions that avoid MHC recognition.

Some sites may have the potential to abrogate MHC recognition but fail to acquire escape substitutions during experimental evolution. Non-varying sites may identify key functional residues. This can be tested by site-directed mutagenesis.

4. Spread of TCR escape substitutions between hosts

Klenerman and Zinkernagel (1998) demonstrated original antigenic sin of TCR escape mutants in LCMV. To simplify a bit, suppose epitope B is a TCR escape variant derived from epitope A. If a host is first exposed to epitope A, subsequent exposure to epitope B tends to reinforce the response against epitope A. Similarly, hosts initially exposed to B, then challenged with A, enhance their response to the first epitope, B. Thus, memory tends to recall TCRs to previous, cross-reacting epitopes.

This memory effect may influence the spread of TCR escape substitutions within a population. To study the evolutionary consequences, an experimental evolution design could follow the fate of particular substitutions through serial passage in hosts with various histories of prior exposure. For example, what sort of evolutionary response would occur in a series of hosts each previously exposed to epitope A? Each previously exposed to epitope B? What about a sequence of hosts with varying exposure histories?

5. Altered peptide ligands

The affinity and kinetics of TCR binding to peptide-MHC ligands influences regulatory control of CTL clones (Davis et al. 1998; Germain and Štefanová 1999). Variant epitopes present altered peptide ligands (APLs) to the TCR. It may be that APLs can interfere with CTL attack by preventing expansion of CTLs with matching TCR specificities. Experimental evolution may favor APLs that interfere with CTL attack. APL variants could be measured for binding properties to TCRs (Tissot et al. 2000) and tendency to spread under various experimental conditions.

6. Dependence of TCR escape on TCR germline genes

Germline genes recombine to make diverse TCRs. The role of the germline genotype in the TCR repertoire could be studied by selecting for TCR escape mutants in hosts with different TCR germline genotypes.

7. Laboratory adaptation

Experimental evolution creates adaptations to the particular in vitro or in vivo laboratory conditions. These conditions only partially reflect selection in the wild. Laboratory studies provide an opportunity to relate biochemical mechanism to kinetics, and kinetics to fitness. Mathematical models aid the controlled, experimental dissection of these relations (Nowak and May 2000). Controlled analysis must be complemented by study of variation and adaptation in natural isolates. The next chapter discusses aspects of natural variation.

Copyright © 2002, Steven A Frank.
Bookshelf ID: NBK2382


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...