OUP user menu

The rate of antigenic variation in fly-transmitted and syringe-passaged infections of Trypanosoma brucei

C.Michael R Turner
DOI: http://dx.doi.org/10.1111/j.1574-6968.1997.tb10486.x 227-231 First published online: 1 August 1997


Rates of antigenic variation were measured in vivo in several populations of cloned lines of Trypanosoma brucei before and after cyclical transmission through tsetse flies. Two cloned lines were adapted for use in laboratory conditions by extensive syringe passaging and rates of antigenic switching/cell/generation were less than 3×10−6 and 1×10−4 in each line. Rates of switching were then determined after fly transmission of the first line and generated per capita rate values of greater than 2×10−3 in three of four populations examined. In the fourth population the switch rate was lower: less than 7×10−5 switches/cell/generation. These data show that rates of antigenic variation are several orders of magnitude lower in syringe-passaged lines, such as those routinely used in the majority of laboratory studies, compared with most recently fly-transmitted lines. They also show that the reduction in switching rate associated with syringe passaging is reversible and is thus unlikely to be caused by mutation.

  • Trypanosoma brucei
  • Antigenic variation
  • Variable antigen type

1 Introduction

Trypanosoma brucei parasites undergo antigenic variation to escape antibody-mediated immune responses and thus prolong the duration of infection and increase the probability of transmission from the mammal to the tsetse fly vector. They do so by switching spontaneously between expression of different variant surface glycoprotein (VSG) genes of which there are several hundred in the genome. More than 107 copies of each VSG molecule enwrap the parasite cell in a surface coat and define its variable antigen type (VAT) [1, 2]. In the field, trypanosome infections are initiated by the bite of an infective tsetse fly but in laboratory studies the most commonly used trypanosome lines have been extensively passaged between mice by syringe inoculation which creates artificial selection pressures to which the parasites respond[2].

The frequency with which trypanosomes undergo antigenic variation requires clarification. Several studies have indicated low rates of switching, similar to mutation rates in other eukaryote systems with per capita switch rate values of approximately 10−7–10−5 switches/cell/generation [35]. These results contrast strongly with those of another study in which much higher rates of switching, >10−3 switches/cell/generation, were observed[6]. Resolving the issue of frequency of antigenic switching is important for two reasons. Firstly, if rates of switching are high, a straightforward host/parasite interaction with rates of antigenic variation and of generation of immunity being similar, leading to a ‘paceD’ stimulation of host responses[7], cannot be correct and other factors, such as competition amongst parasites to reduce growth rates, are required to explain how the characteristic undulating parasitaemias of trypanosome infections arise [7, 8]. Secondly, rates of switching considerably higher than background mutation rates would indicate strongly that the mechanism of antigenic variation is biochemically driven rather than spontaneous. It has been argued that differences between the results of the various studies are due to either syringe passaging (low rate of switching) or cyclical transmission (high rate of switching) of the trypanosomes and the low rates of switching observed in most studies are thus artefacts of laboratory adaptation[6]. Two other possibilities also exist: (1) the switching rates are intrinsic to the trypanosome genotype and the various studies have each been based on different lines of trypanosomes, (2) there were methodological differences between the various studies.

The first objective of this present investigation was to discriminate between these three possibilities by comparing rates of switching in several populations of one cloned line of trypanosomes using the same methodology. The second objective was to investigate whether the low switching rate in syringe-passaged lines has resulted from a heritable change (mutation) or is readily reversible on transmission through tsetse flies.

2 Materials and methods

2.1 Parasites, antibodies and immunofluorescence

Rates of antigenic variation were measured in six populations of T. brucei all of which were antigenically homogeneous (>99%) by immunofluorescence for particular VATs. Two cloned lines, ILTat 1.2 and GUTat 7.2, were adapted to laboratory conditions by extensive rapid syringe passaging – the transfer of trypanosomes between mice by syringe inoculation at 2–3 day intervals for a minimum of 3 months. The first of these was transmitted through tsetse flies and four sub-clones were made in parallel from a mouse infected by the bite of a tsetse fly using previously described methods[6]: ILTat 1.63a, ILTat 1.61c, ILTat 1.64a and ILTat 1.64b. These four sub-clones were grown in mice for 8–11 days between fly bite and use in these experiments. GUTat 7.2 is identical to ILTat 1.63 using both serological and molecular criteria (Barry and Turner, unpublished observations).

VAT-specific rabbit antisera were prepared against GUTat 7.2 and ILTats 1.2, 1.61 and 1.64 using a standard procedure[9]. These antisera were used in antibody-dependent complement-mediated lysis and immunofluorescence assays at pre-determined VAT-specific dilutions. Three previously described VAT-specific monoclonal antibodies were also used in immunofluorescence assays[6]: GUPM 18.7 for ILTat 1.63/GUTat 7.2, GUPM 23.1 for ILTat 1.61 and GUPM 17.1 for ILTat 1.64.

Indirect immunofluorescence on acetone-fixed thin blood films was performed as previously described[9]. The prevalences of specific VATs in a population were based on counts of approximately 200 parasites.

2.2 Experimental design

Switching rates were measured in vivo using an approach that is conceptually based on the classical method of Luria and Delbrück[10] for determining bacterial mutation rates. A key requisite of this method is that a population in which the switching rate is measured is initiated with cells all of which express the same VAT. Infections were therefore initiated (i.p.) from stabilates using inocula that contained, on average, no more than three parasites in mice immunosuppressed by treatment with cyclophosphamide (250 mg/kg) 24 h before inoculation with trypanosomes. Mice were exsanguinated into heparin (10 U/ml) on day 6 post inoculation, the parasite population counted and cells expressing the inoculated VAT killed by antibody-dependent complement-mediated lysis[11]. After incubation for 1 h, dead parasites were observed as immobile ‘ghosts’. In some samples live motile parasites, presumably expressing other VATs, were also occasionally observed.

Groups of 7–20 BALB/c mice were inoculated with a defined number of live trypanosomes plus dead ghosts. These mice were then monitored for patent parasitaemias up to day 14 post inoculation. If no parasites were observed in mice by the end of this period then the inoculum consisted entirely of dead ghosts. This time period over which infections were monitored was chosen because infections initiated by single trypanosomes of these lines become patent on days 6–10 post inoculation[12].

These experiments enabled two parameters to be determined: Po, the proportion of mice in a group that did not develop infections, and N, the number of cells plus ghosts inoculated into each assay mouse. Using these parameters switching rates were estimated from the equation: Embedded Image where ω= rate of switching/cell/generation. This approach to determining the rate of switching is entirely analogous to the method of Luria and Delbrück for measuring mutation rates[10].

Switching rates were determined in all experiments using an N value of 1000 to permit direct comparison of results. In the syringe-passaged populations a second value of N = 5×104 was employed and in fly-transmitted populations a value of N = 50 was used. These values of N were determined empirically to cover a range of switching rates that were likely to be informative [5, 6]. The range of switching rates that could be determined in a particular experiment was dependent not only on N but also the number of mice used in a particular assay and these details are shown in Table 1.

View this table:

Estimates of rates of antigenic variation in syringe-passaged (SP) and fly-transmitted (FT) populations of cloned lines of trypanosomes

Type of lineLineInoculum size, NNo. of mice/groupPercentage of mice uninfected, PoRange of switching rates determinableSwitching rate/cell/generation, ω
SP1.2 (1)103101002×10−3–7×10−5<7×10−5
SP1.2 (2)103201002×10−3–4×10−5<4×10−5
  • In each experiment, two different sizes of inocula, N, were employed. The cloned line ILTat 1.2 was examined twice. The range of switching rates determinable was dependent on N and the number of mice/group.

Three controls were conducted for each experiment. (1) To ensure that trypanosomes were not killed non-specifically by guinea-pig serum (GPS), a portion of each blood sample was incubated in GPS without antiserum and either 50 or 1000 cells plus ghosts (whichever was the smaller number used in any experiment) was inoculated into a single mouse. Microscopical observations showed negligible lysis and mice developed infections in all cases. (2) To establish that antigenic variation had occurred in all mice, even if at a rate lower than the minimum level of detection, 0.5 ml of the infected blood/GPS/antiserum mixture was inoculated into a single mouse. This control also indicated that antisera were not killing trypanosomes non-specifically. Mice developed infections in all cases. (3) To determine that growth of trypanosome populations in assay mice was not due to incomplete or ineffective complement-mediated lysis, blood smears were made from mice that developed infections and screened by immunofluorescence for VAT expression. Combining the data for all experiments, only one of 48 mice screened showed VAT-specific lysis by complement to have been inadequate. That one mouse was excluded from the data analysis.

3 Results

Two separate experiments were conducted using ILTat 1.2 which generated switching rate values of 3×10−6 and <7×10−7 switches/cell/generation. These data demonstrate that the method for determining switching rates in vivo works, as has been shown previously in vitro[5], but that it does not permit fine discrimination of differences in switching rates. The other syringe-passaged population, GUTat 7.2, generated slightly higher values for switching rates – 2×10−4 and 2×10−5 switches/cell/generation (Table 1).

The results from three of the four fly-transmitted populations differed substantially from those of the ILTat 1.2 population from which they were derived (Table 1). In the populations expressing ILTat 1.63a and 1.61c the per capita rates of switching per generation were >4×10−2 and >3×10−2 respectively. Two switching rate values were obtained for ILTat 1.64a – 2×10−2 and 2×10−3 switches/cell/generation, but this latter result is based on only a single mouse failing to develop an infection and so is less reliable. The switch rate values for these three populations were all more than three orders of magnitude higher than those for the parasite line (ILTat 1.2) before transmission through flies. The result for ILTat 1.64b produced, however, a much lower switching rate value of <7×10−5 switches/cell/generation.

Two comparisons indicated that expression of a particular VAT did not influence the rate of switching. ILTats 1.64a and b are separate clones from the same fly-infected mouse, express the same VAT, but had very different switching rates. Also, GUTat 7.2 and ILTat 1.63a are identical by serological and molecular criteria and are both derived from the same field population (Barry and Turner, unpublished results), but with very different switching rates.

4 Discussion

Rates of antigenic variation have been determined in lines of trypanosomes adapted to laboratory conditions by rapid and extensive syringe passaging before transmission through tsetse flies, when a low rate of switching was observed, and after transmission, when a high but variable rate was observed. This study clearly shows that the differences between low and high switching rates found in previous investigations [36] are due to the effects of syringe passaging/fly transmission rather than to differences in methodologies or trypanosome genotypes. These data also show that the reduction in switching rate that accompanies syringe passaging is unlikely to be due to mutation as the trait has been readily reversed on transmission through tsetse flies.

There are two development stage-specific activation mechanisms for VSGs in trypanosomes (bloodstream and metacyclic) and the VATs used in this study partake in them both. Given the time delay between fly infection and use in these studies (8–11 days) it is highly likely that switching rates observed pertain to the bloodstream rather than metacyclic mechanism[13]. The comparison of GUTat 7.2 with ILTat 1.63, together with the results for ILTats 1.64a and b indicate that expression of a particular VAT does not influence the rate of switching. This conclusion from analysis of switching rates in an entire population does not hold true, however, when switching rates are analysed for particular pairs of VATs. At this more detailed level VAT expression does partially influence the rate of switching[6].

The values of switching rates in fly-transmitted infections are the most accurate estimates to date, the single previous study having provided only minimal estimates based on summation of rate values for several specific combinations of VATs. These high rates of switching are consistent with values obtained for all other micro-organisms that undergo antigenic variation in which rates of switching have been determined [6, 14]. Perhaps the most notable feature of T. brucei amongst this group of organisms is that it is the only one studied in any detail where the rate of switching becomes lower on adaptation to the laboratory. A possible cause of this feature may be discerned in the variability of switching rates in fly-transmitted populations where one of four populations gave a low rate value. Although systems of antigenic variation in different organisms are functionally very similar, they differ a great deal in mechanism [14, 15]. The as yet unidentified mechanism causing rapid switching in T. brucei would appear to be unusually labile.


I would like to thank Nasreen Aslam for technical assistance, Dr M. Hope for critical comments on the manuscript, Dr I. Maudlin of the Tsetse Research Laboratory, Bristol for supplying tsetse fly pupae and The Wellcome Trust for financial support. C.M.R. Turner is a Royal Society University Research Fellow.


View Abstract