T-cell activation, proliferation and apoptosis in primary Listeria monocytogenes infection

Abstract

Listeria monocytogenes infection of mice leads to a rapid expansion of activated T cells, followed by a decline in specific cells once the bacteria are eliminated. In order to define the relationship between T-cell proliferation and activation, and to investigate the role of apoptosis in limiting the expansion, the expression of activation markers, uptake of 5-bromo-2′-deoxyuridine (BrdU) in vivo and the incidence of apoptosis was investigated. Increased numbers of T cells expressing the activated phenotype CD25⁺, CD44^hi and CD62L^lo were detected 4 days after infection. Expression of CD25 (IL-2Rα chain) on CD4⁺ and CD8⁺ T cells peaked at this time and returned to normal by day 7. In contrast, CD44^hi and CD62L^lo persisted, with the maximum proportion occurring at 7 days after infection. This was accompanied by a burst of in vivo proliferation of CD4⁺ and CD8⁺ T cells occurring between day 5 and 7. Apoptosis, which is presumably needed to control this expansion of T cells, also peaked at 7 days after infection. Apoptosis occurred preferentially amongst T cells which had proliferated. Most but not all proliferating T cells had down-regulated their CD62L marker. While most apoptotic T cells were CD62L^lo, again not all had down-regulated this marker. Hence, CD25 expression peaked early, but expression of other activation markers, in vivo proliferation and apoptosis coincided after Listeria infection. T cells that had proliferated were over-represented in the apoptotic population.

Introduction

Listeria monocytogenes is a Gram-positive, facultative intracellular bacterium that has been a powerful model for the study of the cellular immune response against intracellular bacteria.¹ Intravenous injection of viable bacteria leads to an acute infection, which is cleared within 7–10 days after infection depending on mouse strain.² CD8⁺ cytotoxic T cells play a prominent role in protection against listeria, while CD4⁺ T cells produce interferon-γ (IFN-γ).³ Protection is dependent upon IFN-γ, interleukin (IL)-12, tumour necrosis factor-α (TNF-α), IL-6 and other cytokines.⁴

Listeria infection provides a very strong stimulus for T-cell activation and proliferation. Numbers of T cells in the spleen more than double in the first few days after intravenous infection of mice.^5–7 It is not possible to tell whether this is due entirely to proliferation or also involves recruitment to the spleen. It is doubtful that all the cells are specific for the organism, although even among specific T cells there is a marked expansion, depending on the technique used to identify them. The protective activity of transferred T cells increases about 100-fold between days 3 and 6 of infection of the donor.² Limiting dilution assays suggested 1/1000 T cells were specific for Listeria 5–7 days after infection⁸ while 0·5–2% T cells stained for intracytoplasmic IFN-γ at the peak of infection.⁹ Enzyme-linked immunospot (ELISpot) assay of Listeria-specific CD4⁺ T cells producing IFN-γ showed an increase from <10/10⁶ in uninfected mice to 500/10⁶ 7 days postinfection (Smith and Cheers, unpublished information). Epitope-major histocompatibility complex (MHC) class I tetramer staining showed that 7 days post infection, 1·75% of CD8⁺ T cells in the mouse spleen were specific for just four Listeria epitopes.¹⁰ Specific or not, the overwhelming response of both CD4⁺ and CD8⁺ T cells calls for equal control measures to return T-cell numbers to a resting state.

Antigen-driven T-cell proliferation typically sensitises T cells to apoptosis. Apoptosis has emerged as an important mechanism for maintaining T-cell homeostasis following the dramatic increases in cell numbers such as those described above.¹¹ Infection of mice with lymphocytic choriomeningitis virus causes a dramatic expansion of T cells, closely followed by a wave of T-cell apoptosis.¹² The injection of the staphylococcal superantigen, enterotoxin B, into mice also leads to a burst of proliferation, followed by apoptosis.¹³ Recently, it has been shown apoptosis occurs predominantly in those cells that have divided four or more times.¹⁴

Apoptosis amongst T cells has been identified just 48 h after infection of mice with Listeria¹⁵ but this relates to an anomalous early depletion of T cells from T-dependent areas of lymphoid tissues^5,6 and may precede induction of specific immunity. The effect was transient, and could not be attributed to the action of Fas/FasL or any of a number of likely cytokines.¹⁵ Its induction required infection with live Listeria organisms expressing the secreted toxin listeriolysin. In relation to feedback control of the T-cell population after the induction of specific T-cell mediated immunity, it has been shown that when lpr mice, defective in Fas, were infected intraperitoneally with Listeria, activated (CD44^hi) cells persisted in the peritoneal cavity well after the clearance of infection at 10 days and for some days longer than seen in control mice.⁷ This suggests a role for Fas/FasL interaction and apoptosis in down-regulation of the activated T cells.

The aims of this work therefore were to provide a direct measure of in vivo T-cell proliferation and relate this to activation and apoptosis of CD4⁺ and CD8⁺ T cells during primary Listeria infection, to test whether the events were occurring as a sequence in the same cells. The T-cell activation markers chosen for study were the early activation markers, CD25 (α chain of the IL-2 receptor, IL-2Rα) and CD62L (l-selectin), and CD44 (pgp-1), whose expression has been associated by some with immunological memory.¹⁶ The expression of IL-2R is clearly a key to IL-2-driven T-cell proliferation. CD62L is down regulated upon activation of T cells, while CD44 expression is increased. Both molecules regulate cell circulation. CD44 allows binding of cells to hyaluronidate allowing their entry into tissues, while down regulation of CD62L releases cells from continual circulation between lymph nodes.¹⁷ As a marker of early apoptosis we used annexin V. Annexin V binds to phosphatidylserine, which is translocated from the inner to the outer side of the membrane at the early stages of apoptosis. Fluorochrome-conjugated annexin V has been used to detect cells in the early stages of apoptosis by flow cytometry.¹⁸ This method has the advantage that it allows the expression of other cell surface markers to be determined at the same time as early apoptosis.

Our experiments comparing both CD4⁺ and CD8⁺ T cells showed that a strong proliferation of T cells during Listeria infection was matched by early expression of CD25, followed by other activation markers and apoptosis. Apoptosis occurred preferentially amongst CD62L^lo cells which had previously divided.

Materials and Methods

Bacteria and infection

L. monocytogenes strain EGD obtained from R. V. Blanden (Australian National University, ACT, Australia) were grown overnight on horse blood agar (HBA) plates, dilutions of bacteria were prepared in saline and mice were infected with 5×10³ colony-forming units (CFU) into the lateral tail vein. The dose of bacteria was confirmed retrospectively for all experiments. Bacterial loads and infectious doses were determined by plating serial 10-fold dilutions of bacterial suspensions or tissue homogenates on HBA plates.²

Mice

Male C57B/10 mice, naturally resistant to Listeria infection (intravenous LD₅₀ 2×10⁵)² were used in all experiments along with age- and sex-matched, uninfected controls. All experiments were conducted according to Animal Ethics Experimentation Committee standards in the animal facility Department of Microbiology and Immunology, University of Melbourne. Groups of three mice, analysed individually were used in all experiments. Results are representative of at least two independent experiments.

Activation marker staining

Mice were killed by CO₂ overdose at the times after infection indicated. Spleens were removed and disrupted by passage through a sieve and red blood cells were lysed by NH₄Cl solution.¹⁹ Two-colour flow cytometry was used to analyse the expression of activation markers on T cells during infection. One to two×10⁶ cells were stained with anti-CD44 (clone IM7, Pharmingen, San Diego, CA), anti-CD62L (clone mel-14, a gift from Dr W. Heath, Walter and Eliza Hall Institute, Victoria, Australia) and biotinylated anti-CD25 (clone 7D4, Pharmingen), in phosphate buffered saline (PBS) with 0·1% bovine serum albumin (BSA) and 0·01% NaN₃. CD44 and CD62L staining was detected by fluorescein isothiocyanate (FITC) anti-rat immunoglobulin (IgG; Silenius, Melbourne, Australia), whereas CD25 staining was detected by streptavidin Cy-Chrome (Pharmingen). After blocking with 1% rat serum in PBS, cells were stained with CD4–phycoerythrin (PE) or CD8–PE (Pharmingen). Cells were either fixed in 1% paraformaldehyde in PBS and analysed the next day, or resuspended in PBS and analysed immediately. Five to 10 thousand CD4⁺ or CD8⁺ events were collected, gating on live cells by forward and side scatter. Negative staining controls were antirat FITC alone or streptavidin Cy-Chrome alone. CD4 and CD8 staining was controlled by a PE-labelled isotype control (Pharmingen).

Detection of apoptosis

Spleen cell suspensions were centrifuged through Ficoll-Histopaque (Sigma, Castle Hill, NSW, Australia) to remove both dead cells and red blood cells, resulting in preparations which were >95% viable. Apoptosis was detected and analysed by fluorescence-activated cell sorting (FACS), using annexin V and 7-amino actinomycin (7-AAD), as described previously.²⁰ For CD62L staining, cells were stained with CD4–FITC or CD8–FITC and biotinylated Mel-14, followed by streptavidin Cy-Chrome. Cells were then stained with annexin V–PE and analysed by flow cytometry.

In vivo proliferation

Mice were fed drinking water containing 1 mg/ml 5-bromo-2′-deoxyuridine (BrdU, Sigma, Castle Hill, NSW, Australia) and 1% glucose, for 2 days prior to analysis. Spleens were removed and single-cell suspensions were prepared and cells were stained with CD4–PE or CD8–PE as above. Staining for BrdU was performed as described by Tough et al.¹⁶ Briefly, cells were dehydrated in an alcohol solution, fixed and permeabilized in 1% paraformaldehyde/0·01% Tween-20, treated with 50 U/ml DNase and then stained with 10 µl of FITC conjugated anti-BrdU (Becton Dickinson, San Jose, CA). For CD62L staining cells were stained with CD4–PE or CD8–PE, then biotinylated Mel-14 (prepared in our lab), followed by streptavidin Cy-Chrome then BrdU staining as above. Negative controls included mice that did not receive BrdU and test samples that were not treated with DNase. For annexin V and BrdU staining, cells were stained for CD4 or CD8 and annexin V in annexin V binding buffer.¹⁸ BrdU staining was performed as above except all washes were done in annexin V binding buffer. Annexin V staining in PBS was used as a control for non-specific binding of annexin V.

Statistical analysis

Student's t-test was used to determine statistical significance. P<0·05 was considered significant.

Results

Infection and activation of T cells

Initial experiments aimed to define the changes in the expression of the activation markers, CD25, CD44 and CD62L, on T cells after infection with Listeria. Following intravenous infection with 5×10³ CFU, the bacterial burden in the spleen and liver increased to 5 log 2 days after infection, then fell to 2 log, the limit of detection, by 7–10 days. No bacteria were detected at later time points (Fig. 1a). Flow cytometry showed that CD4⁺ and CD8⁺ T cells as a percentage of total cells in the spleen had declined significantly (P<0·001) 2 days after infection, compared to uninfected mice (Fig. 1b), as previously reported for total T cells.^5,6 By 7 days the percentage of CD4⁺ and CD8⁺ T cells had recovered to normal levels. However, when the absolute number of splenic CD4⁺ and CD8⁺ T cells were calculated there was a significant 3–4-fold increase in the number of CD4⁺ (P<0·01) and CD8⁺ (P<0·01) spleen cells after 7 days compared to uninfected mice (Fig. 1c,d). By 15 days after infection the number of spleen cells had returned to near preinfection levels.

Figure 1.

Bacterial load and changes in T-cell populations during Listeria infection. (a) Time course of infection, showing log₁₀ number of bacteria in the spleen and liver of mice infected i.v. with 5×10³ L. monocytogenes. (b) Changes in the percentage of CD4⁺ and CD8⁺ cells from spleens of mice at different times after infection as determined by FACS analysis. (c) Total spleen cells recovered from individual mice. (d) The absolute number of CD4⁺ and CD8⁺ cells were calculated by multiplying the total spleen cells recovered from individual mice by the percentage CD4⁺ or CD8⁺ in the same mice, as determined by FACS analysis. Each point represents the arithmetic mean±standard deviation of groups of three mice. Where error bars are not visible, they are hidden by the symbols. Each experiment was repeated 4 times with similar results.

Flow cytometry was used to assess the effect of infection on the expression of the activation markers CD44, CD62L and CD25 on CD4⁺ and CD8⁺ T cells. Figure 2 shows representative histograms of the staining pattern of CD4⁺ T cells from individual mice at the peak of expression of each of the markers. For CD25 this peak was achieved at day 4, when the numbers of CD25⁺ cells were approximately double those in uninfected mice. At 7 days post infection, CD44^hi cells formed a markedly higher proportion of cells from infected mice compared with naive (60% compared with 21% in the example shown), but CD44 was expressed to some degree on all CD4⁺ T cells. In contrast, infection caused CD62L on T cells to be down-regulated, reversing the proportions of CD62L^lo CD62L^hi cells (39/61 for naïve mice, 73/27 for infected mice).

Figure 2.

Activation marker expression on T lymphocytes from naive and infected mice. The FACS histograms of spleen cells from individual mice that were either naive (uninfected, left) or infected seven days earlier (right) are shown for three activation markers: (a), CD25; (b), CD44; (c), CD62L. Histograms show analysis of 5000 CD4⁺ cells. The dotted line represents the negative control stained with anti-rat–FITC or streptavidin Cy-Chrome alone as appropriate. The percentage activated cells is indicated by horizontal bars for each histogram.

Figure 3 shows the time course of expression of all three markers on CD4⁺ and CD8⁺ T cells. It was notable that expression of CD25 (IL-2Rα) on both CD4⁺ and CD8⁺ T cells reached an early peak at 4 days, both in terms of percentage of the cell population and absolute numbers (Fig. 3). Expression declined rapidly, returning to background levels by day 7. In contrast, numbers of CD44^hi cells and CD62L^lo cells peaked at the later time of 7 days and was still elevated at 15 days, well past the time of resolution of the infection. Despite the prominence of CD8⁺ T cells in protection against this infection³ cells bearing these activation markers made up a lower percentage of CD8⁺ cells in the spleen than of CD4⁺ cells. This was compounded by the lower total numbers of CD8⁺ cells, so that even at the peak they represented less than 1/3 of activated T cells.

Figure 3.

Changes in the percentage and number of activated CD4⁺ and CD8⁺ T cells during Listeria infection: Groups of three mice were killed 4, 7, or 15 days after i.v. infection with 5×10³ L. monocytogenes and analysed individually. Spleen cells were gated on CD4 (a and b) or CD8 (c and d) and activation markers, CD25, CD44 and CD62L were assessed. The percentage (a and c), or the number (b and d) of CD44^hi, CD62L^lo or CD25⁺ cells were determined. One representative of two independent experiments is shown. Each point represents mean±standard deviation for three mice.

Changes in the in vivo proliferation of T cells during infection

To determine whether the increased numbers of T cells during L. monocytogenes infection reflected increased proliferation, BrdU was used to measure the rate of T-cell proliferation in vivo following infection. Mice were fed BrdU in their drinking water for 2 days immediately before they were killed and their spleen cells were stained with a BrdU-specific monoclonal antibody. Thus the numbers of BrdU-positive cells represent those proliferating on the previous 2 days. The results (Fig. 4a) showed a dramatic increase in rate of proliferation of both CD4⁺ and CD8⁺ cells, compared with uninfected (day 0). There was no increase above background (7% CD4⁺ and 6% for CD8⁺) during the first 2 days of infection. However by 7 days after infection, a significant increase in the rate of proliferation occurred, 54% of CD4⁺ (P<0·02) and 43% of CD8⁺ (P<0·05) T cells having proliferated in the proceeding 2 days. By 12 days after infection the rate of proliferation had dropped to 14% of CD4⁺ and 12% of CD8⁺ cells, which was not significantly greater than uninfected mice.

Figure 4.

T lymphocyte proliferation and activation during Listeria infection. Groups of three mice were fed drinking water containing BrdU for two days immediately before to sacrifice 2, 7, or 12 days after iv infection with 5×10³ L. monocytogenes. (a). Proliferation amongst CD4⁺ and CD8⁺ subpopulations. Spleen cells from individual mice were stained with anti-CD4 or anti-CD8 and BrdU. One representative of four independent experiments is shown. (b, c) Proliferation amongst activated subpopulations. Spleen cells were stained for CD4 or CD8 and CD62L then for BrdU incorporation. The percentage of BrdU⁺ CD4⁺ and CD8⁺ is shown when gated on CD62L^lo (b) or CD62L^hi cells (c). One representative of two independent experiments is shown. Each point represents mean±standard deviation for three mice.

Next, three-colour FACS analysis was used to measure the proliferation of CD62L^lo T cells at different times after infection. Most of the proliferation occurred amongst the CD62L^lo population (Fig. 4b versus Fig. 4c). The time course of proliferation of the CD62L^lo T-cell population was similar to the whole CD4⁺ and CD8⁺ T cells. For the first 2 days after infection, proliferation was not greater than the uninfected controls. The BrdU incorporation between 5 and 7 days after infection was significantly greater than the naive controls for both CD4⁺ (P<0·01) and CD8⁺ (P<0·02) T cells, indicating that a burst of proliferation occurred within the CD62L^lo population (Fig. 4b). By 10–12 days after infection the proportion of CD4⁺ CD62L^lo T cells that had proliferated declined to 16% of CD4⁺, whereas 26% of CD8⁺ CD62L^lo T cells were still proliferating. Amongst the CD62L^hi cells, the proportion of cells that had proliferated was lower at all times compared with CD62L^lo cells (Fig. 4c). There was a small increase in the proliferation of both CD4⁺ and CD8⁺ CD62L^hi T cells 5–7 days after infection. This was statistically significant for CD4⁺ CD62L^hi cells (P<0·05 at 7 days compared with cells from uninfected mice) but not for the CD8⁺ population.

Apoptosis during Listeria infection

Given the acute activation and proliferation of T cells in response to infection it seemed probable that a high rate of apoptosis was needed to return the T-cell pool to its normal composition once the infection was cleared. To measure this directly, annexin V and 7-AAD staining were used to detect the early apoptotic cells amongst the T cells during infection. The results (Fig. 5a) show that the rate of apoptosis was more sustained among CD4⁺ compared with CD8⁺ T cells. The peak of apoptosis occurred 4–7 days after infection. By 15 days after infection the rate of apoptosis had returned to background levels.

Figure 5.

Apoptosis of T lymphocytes during Listeria infection. Groups of three mice were killed 2, 5, 7 and 15 days after infection with 5×10³ L. monocytogenes. Spleen cells were stained for CD4 or CD8 plus activation and apoptosis markers. (a) The percentage of apoptotic CD4⁺ and CD8⁺ cells. Spleen cells from individual mice were stained with anti-CD4 or anti-CD8, plus annexin V and 7-AAD. (b) Activation of CD4⁺ and CD8⁺ cells. Spleen cells from individual mice were stained with anti-CD4 or anti-CD8 plus anti-CD62L. (c, d) Apoptosis amongst activated populations. Spleen cells were stained for CD4 or CD8 and CD62L then for annexin V. (c) shows the percentage of annexin V⁺ CD4⁺ when gated on CD62L^lo or CD62L^hi cells, while (d) shows the equivalent data for CD8⁺ cells. Each point represents mean±standard deviation for three mice. One representative of three independent experiments is shown.

To test whether apoptosis was related to prior activation of the cells, the proportion of CD62L^lo T cells versus CD62L^hi cells that were undergoing apoptosis during infection was determined. The time course of changes in activation markers was similar for CD4⁺ and CD8⁺ T cells but, as before, more CD4⁺ cells down-regulated their CD62L than did CD8⁺ (Fig. 5b). The rate of apoptosis amongst CD4⁺ T cells was similar for the CD62L^lo and CD62L^hi populations (Fig. 5c), representing about 25% of each population at its peak. Amongst the CD8⁺ T cells, apoptosis occurred preferentially amongst the CD62L^lo population, being about 24% of CD62L^lo and 15% of CD62L^hi (Fig. 5d). In terms of absolute numbers, the majority of apoptotic CD4⁺ and CD8⁺ T cells were CD62L^lo, because these were the majority of T cells.

Expanded T cells are deleted by apoptosis

To determine the relationship between proliferation and apoptosis, spleen cells were stained for CD4 or CD8, annexin V and BrdU. Representative staining of cells from individual mice, either uninfected or infected 7 days previously, and gated on CD4⁺ T cells is shown in Fig. 6. As noted in Fig. 5, even cells from uninfected mice showed a background of apoptosis, but in infected mice apoptosis was increased amongst proliferating (BrdU⁺) cells. Similar staining was seen when gated on CD8⁺ cells (data not shown). When the complete time course was investigated, there was a steady increase in the proportion of annexin V⁺ T cells, peaking 7 days after infection (Fig. 7a). Apoptosis occurred preferentially amongst cells that had proliferated, both CD4⁺ and CD8⁺ (Fig. 7b), with a gradual, but not significant, increase over the course of infection in the percentage of each population staining with annexin V. Comparing the BrdU incorporation by T cells that were annexin V⁺ or annexin V⁻ (Fig. 7c), it was found that 39% of annexin V⁺ CD4⁺ T cells had proliferated in the previous 2 days, while 32% of annexin V⁺ CD8⁺ T cells had proliferated during the same interval. In contrast significantly fewer (P<0·001) annexin V⁻ CD4⁺ or CD8⁺ T cells had proliferated during the same time. Hence, T cells that proliferated were being deleted by apoptosis as the infection was resolved.

Figure 6.

Annexin V staining of proliferating cells from infected or naive mice. Scatter plots of BrdU versus annexin V staining of spleen cells from individual mice that were either naive (uninfected controls) or infected 7 days earlier. Cells were stained as described in Materials and Methods. Controls are cells from the same animals, which were treated in the same manner except that annexin V staining was done in PBS and the cells were not treated with DNAse prior to staining with anti BrdU. The number represent the percentage of cells in each quadrant, when gated on CD4⁺ cells.

Figure 7.

T lymphocyte proliferation and apoptosis during Listeria infection. Groups of three mice 4, 7 and 10 days after i.v. infection with 5×10³ L. monocytogenes. Spleen cells were stained with CD4 or CD8, plus annexin V and BrdU. (a) The percentage of annexin V⁺ CD4⁺ and annexin V⁺ CD8⁺ cells. (b) apoptosis amongst dividing and nondividing cells. Cells which had or had not incorporated BrdU were analysed separately for annexin V staining. (c) Cell division amongst apoptotic cells. Annexin V⁺ and annexin V⁻ populations were analysed separately for BrdU incorporation. All points represent mean±standard deviation for three mice. One representative of three independent experiments is shown.

Discussion

Here we show that experimental L. monocytogenes infection leads to the rapid activation, proliferation and apoptosis of CD4⁺ and CD8⁺ T cells. The peak in apoptosis coincided with the peak in T cell activation judged by CD44^hi CD62L^lo cells and in vivo proliferation measured by BrdU uptake. Interestingly, the peak in CD25 expression preceded other activation markers. Apoptosis occurred preferentially amongst proliferating lymphocytes that had down-regulated CD62L.

It is notable that protective T cells, judged by adoptive transfer, first become apparent at 4 days after Listeria infection and reach a peak at 7–10 days.² CD25 expression changed rapidly in the present experiments, peaking 4 days after infection, and then was rapidly down regulated by 7 days after infection. CD25, the alpha chain of the IL-2R, is a component of the high affinity IL-2R.²¹ It was interesting that proliferation continued after the loss of the IL-2R expression. However, while IL-2 is well known to drive the proliferation of T cells in vitro, in vivo IL-15 may be critical. Indeed, following allogeneic stimulation in vivo, it was shown that IL-15, not IL-2, promoted T-cell proliferation.²² IL-2 was not produced until several rounds of T-cell division, and was then responsible for T-cell apoptosis by down-regulating the γ-c chain shared by IL-2 and IL-15 receptors and decreasing Bcl2 expression. It is thus interesting that the loss of expression of IL-2Rα preceded the peak of apoptosis in the present experiments. It would clearly be of interest to examine the expression of IL-15R during infection.

In contrast with CD25, the expression of CD44^hi and CD62L^lo phenotypes followed a different, but parallel kinetic, both peaking 7 days after infection and remaining above normal levels thereafter, in agreement with previous reports.^7,10 CD44 and CD62L have, broadly speaking, different roles from CD25. CD44 allows cells to bind to hyaluronate, while CD62L binds to the high endothelial venules which keeps the naïve T cells circulating between the lymph nodes¹⁷ so that its down-regulation allows the cells to enter other tissues. The CD25⁺ population may represent the acutely activated population at four days after infection, which gives rise to the rapidly proliferating, CD44^hi CD62L^lo population by 7 days after infection.

The use of BrdU in drinking water allowed the direct identification of T cells that had proliferated in vivo at different time intervals after infection. The results support data from other investigators who used tetramer staining¹⁰ or adoptive transfer systems²³ to analyse the expansion of T cells after primary and secondary Listeria infection. All methods show a vigorous expansion of T cells at 5–7 days after infection, which the BrdU staining here has shown to include a considerable component of proliferation. Recent data suggests that the expansion of CD8⁺ T cells during Listeria infection is programmed by initial antigenic stimulus and is independent of persisting infection.²⁴ The fact that expansion of CD4⁺ T cells here parallels the expansion of CD8⁺ T cells suggests the same may apply to this subpopulation. Interestingly, it has recently been shown that proliferation of T cells is not necessarily required for their activation, despite the long held assumption that the two are mechanistically linked.²⁵ These results show that the two at least run in parallel. The CD62L^lo cells showed very similar kinetics of proliferation to the whole T-cell population. The CD62L^hi population contained fewer proliferating cells, although a small peak in proliferation of these cells mirrored the peak in proliferation in the CD62L^lo population. This could be due to the expression of CD62L by activated cells as previously described for CD8⁺ T cells in other systems.²⁶ Specific CD62L^hi CD8⁺ T cells have been identified by tetramer staining in secondary Listeria infection.²⁷ Alternatively, they could be cells that have recently emigrated from the thymus. A comparison of thymectomized and intact mice, infected with Listeria would be required to rigorously test this hypothesis. However, Tough and Sprent have shown that recent emigrants which had proliferated in the thymus showed lower intensity of staining with BrdU compared with activated cells that had proliferated in the periphery¹⁶ and we did not see any lesser intensity when comparing BrdU staining of CD62L^lo and CD62L^hi populations.

Apoptosis of T cells became evident 2 days after infection. This early apoptosis was observed by Merrick et al.¹⁵ and used to explain the depletion of T cells observed early in infection.^5,6 Whether this time-point represents a distinct phase of apoptosis particularly related to that early depletion, or whether it is the precursor of the later apoptosis observed here is not clear. Like activation marker expression and proliferation, apoptosis peaked 7 days after infection. It is not surprising that apoptosis occurred preferentially, although not entirely, amongst the population of T cells that had previously proliferated, since a similar relationship has been found in a number of systems^12,13,28,29. In particular, Renno et al.¹⁴ showed that apoptosis in a superantigen-driven system occurred mainly amongst T cells which had divided four or more times. Our method of identifying apoptotic, dividing cells involved staining with anti-CD4 or anti-CD8 plus annexin V and anti-BrdU. Annexin V identifies both early apoptotic and dead cells, the latter usually being ruled out by 7-AAD or similar vital staining.¹⁸ However, with already three colours, there was no opportunity to do this. Instead, dead cells were removed at the beginning of the assay by centrifugation through Ficoll-Histopaque. Using this approach, we find that dead cells make only a minor contribution to the populations staining with annexin V. After purification of control cell preparations in this manner only 5–8% of annexin V⁺ cells stain with 7-AAD, a fraction which is undetectable in practice.

The relationship of apoptosis and cells bearing an activation marker was less clear cut. For the CD4⁺ T cells, apoptosis during Listeria infection occurred more or less equally amongst the CD62L^lo and CD62L^hi populations, representing approximately 25% of each. For CD8⁺ T cells, apoptosis occurred preferentially amongst the CD62L⁻ population, but even at its peak represented only 24% of CD62L^lo and 15% of CD62L^hi. Given the clear-cut relationship of apoptosis with proliferation and proliferation with CD62L regulation, this at first sight appears anomalous. However, it may well be that the subpopulation of CD62L^hi cells which had proliferated (25% of all T cells) accounted for the excess numbers of CD62L^hi cells undergoing apoptosis. In terms of absolute numbers, the majority of apoptotic CD4⁺ and CD8⁺ T cells were CD62L^lo, as these represented the majority of T cells.

The mechanism of apoptosis during the course of infection is not clearly understood. Merrick et al.¹⁵ were unable to attribute the early apoptosis they observed 2 days after Listeria infection to any of the known mechanisms. Fas/FasL interactions do appear to be one major pathway of T-cell apoptosis in later stages of the infection⁷ while perforin and IFN-γ feedback have also been implicated.³⁰ It is known that NO, produced to high concentrations during listeria infection³¹ induces apoptosis amongst CD4⁺ Th1 cells.³² The use of inducible nitric oxide synthase deficient mice, unable to up-regulate NO production, would clarify its role in apoptosis of CD8⁺ T cells.

As in other systems, it remains unclear how the cells that will be deleted by apoptosis are separated from those that will persist to become memory cells. Our data allow us to identify the population from which the long lived memory cells will derive; i.e. those that did proliferate in response to infection but did not enter the apoptotic pathway. Analysis of the cells that proliferated after infection, but were not deleted shortly afterwards, may allow insights into how the decision to become a memory cell is made. The development of methods to genetically tag activated T cells may allow this problem to be addressed.³³

In summary, Listeria infection causes a dramatic and acute activation and proliferation of T cells, with concomitant apoptosis. This means of controlling expansion of the T-cell population presumably works well for the host that has cleared the infection, as in listeriosis, and only requires a small pool of memory T cells in the event of re-infection. However the process may be less to the host's advantage if the infectious organism is not rapidly eliminated, as with mycobacterial infection.^34–37 Recent work from our lab has shown that chronic Mycobacterium avium infection is associated with an increase in T-cell apoptosis and elevated, but sustained levels of in vivo proliferation (20 and Mannering and Cheers, in press). The relationship between activation, proliferation and apoptosis is under study in that model.

Abbreviations

7-AAD

7-amino actinomycin

BrDU

5-bromo-2′-deoxyuridine

CFU

colony-forming units

FACS

fluorescence-activated cell sorter

FITC

fluorescein isothiocyanate

HBA

horse blood agar

LD₅₀

lethal for 50%

PBS

phosphate-buffered saline

PE

phycoerythrin