Cutting Edge: Apoptosis of Superantigen-Activated T Cells Occurs Preferentially After a Discrete Number of Cell Divisions In Vivo¹

Following activation with staphylococcal enterotoxin B (SEB),³ T cells proliferate and eventually undergo apoptosis (1, 2, 3, 4, 5, 6, 7, 8). Surprisingly, it has been shown that the up-regulation of known apoptosis-mediating molecules such as Fas (reviewed in 9) or TNF receptor (reviewed in 10) on the surface of activated T cells does not confer immediate susceptibility to apoptosis (11, 12). One possible explanation of this apparent paradox is that activated T cells are programmed to acquire susceptibility to apoptosis only after a given number of cell divisions. Indeed, a precedent for such “cell division-associated” acquisition of specific function has been established for Ig isotype switch in activated B cells and for activation marker and cytokine expression in activated T cells (13, 14, 15, 16, 17). Linking apoptosis to cell division in activated T cells would allow expansion and effector function of the responding cells, while ensuring that these cells will ultimately undergo apoptosis, thus maintaining homeostasis. To address this possibility, we adapted a recently described cell division monitoring technique (18) that, when used in conjunction with cell surface phenotyping, allows the study of the relationship between the number of cell divisions and activation or death markers. Here, we show that SEB-reactive CD4⁺ T cells divide asynchronously in response to SEB in vivo, and that apoptosis of SEB-reactive T cells occurs preferentially after a discrete number of cell divisions.

We obtained 4- to 8-wk-old BALB/c mice from Harlan Olac (Bicester, U.K.). A total of 5 × 10⁷ syngeneic splenocytes labeled with 5- (and 6-)carboxyfluorescein diacetate succinimidyl ester (CFSE) (Molecular Probes, Eugene, OR) (see below) were transferred i.v. into the tail vein. SEB (10 μg, purchased from Toxin Technology, Saratoga, FL) was injected into each of the hind footpads 1 day following cell transfer.

Single-cell suspensions were made from spleens by homogenization into sterile high glucose DMEM (Life Technologies, Grand Island, NY). Debris was allowed to settle, and splenocytes were subsequently resuspended at 5 × 10⁷/ml in PBS with no protein. CFSE was added to a final concentration of 5 μM, and the suspension was incubated at 37°C for 10 min. At the end of the incubation period, the cells were immediately washed three times in cold DMEM/10% FCS (Irvine Scientific, Santa Ana, CA).

For three-color staining, the following mAbs were used: PE-labeled anti-CD4 (Boehringer Mannheim, Mannheim, Germany) and biotinylated anti-Vβ8 (F23.1) (19) (revealed with streptavidin-tricolor (Caltag, San Fransisco, CA). For four-color staining, anti-Vβ8-PE and anti-CD4-APC were purchased from PharMingen (San Diego, CA) and used in conjunction with annexin V staining (see below). Stained cells were analyzed on a FACScan or FACScalibur (Becton Dickinson, San Jose, CA).

Annexin V (Nexins Research, Maastricht, The Netherlands) was used as described previously (20, 21, 22). Briefly, cells were washed in 100 μl of binding buffer and incubated in a 1/50 dilution of biotin-coupled annexin V solution for 20 min in the dark. Next, cells were washed, revealed with streptavidin tricolor (Caltag), and analyzed on the FACS.

At 3 days after SEB administration, the percentage of cells in CFSE peaks was measured after electronically gating on CD4⁺ Vβ8⁺ cells. To estimate the initial proportion of cells that gave rise to those in the different division peaks, we divided the percentage of cells in division peaks 1, 2, 3, and 4 by 2, 4, 8, and 16, respectively (because every two cells in a given division peak must have arisen from a single cell from the previous division peak). The corrected percentages thus obtained were used in the following formula: Percent undivided cells = U/(U + ΣDc), where U equals the percentage cells in the undivided CFSE peak and ΣDc equals the sum of the corrected percentages of cells in division peaks 1–4.

It should be noted that cell loss/apoptosis, although not accounted for by this analysis, would not significantly alter the outcome; it occurs mainly in division peak 4 (see Fig. 3), which only contributes a minor fraction of ΣDc.

To investigate whether SEB-induced apoptosis of activated T cells is associated with the number of cell divisions in vivo, we used the dye CFSE, which covalently binds to cellular proteins and allows direct measurement by flow cytometry of cell divisions over several cell generations (18). We focused our analysis on the CD4 subset, because SEB-mediated clonal deletion has been shown to be more dramatic in CD4⁺ than in CD8⁺ T cells (8, 23, 24). CFSE-labeled spleen cells were injected i.v. into naive syngeneic mice. Typically, 2–3% CFSE⁺ T cells were detected in the lymphoid organs at 24 h after cell transfer (Fig. 1, inset), and this percentage remained stable over several days (data not shown). To ascertain that CFSE treatment did not compromise the ability of CD4⁺ T cells to respond normally to SEB, we measured the percentage of Vβ8⁺ T cells within the CD4 subset in the draining lymph node (LN) in electronically gated CFSE-positive or -negative cells following injection of SEB. As can be seen in Fig. 1, the response to SEB in donor CFSE⁺ cells was virtually indistinguishable from the typical response of host CFSE⁻ cells, confirming that neither T cell expansion nor deletion was affected by staining with CFSE.

We subsequently confirmed that CFSE can be used to monitor the specific cell division induced by SEB in vivo. In the representative example shown in Fig. 2, five distinct peaks of CFSE fluorescence intensity (each differing by a factor of two from adjoining peaks) were detected among donor-derived CD4⁺ Vβ8⁺ cells in the draining LN 3 days after SEB injection. The brightest peak corresponds to undivided cells, whereas the four remaining peaks identify cells that have undergone 1, 2, 3, and 4 divisions, respectively. As expected, very few CD4⁺ Vβ8⁻ cells divided in response to SEB during this time period, resulting in a dominant CFSE peak that corresponded to undivided cells (Fig. 2). Kinetic analysis of CFSE fluorescence among CD4⁺ Vβ8⁺ cells following SEB injection indicated a progressive increase in the number of cell divisions between 36 and 72 h, when a fourth (and to a lesser extent fifth) division peak was observed (Fig. 3,A). No additional cell division was detected between 72 and 96 h. Quantitation of the proportion of CD4⁺ Vβ8⁺ cells in each CFSE division peak (Fig. 3 B) confirmed that SEB-induced proliferation progressed from 36 to 72 h and subsequently ceased. Several points can be made from these data. First, there is a clearly defined population of undivided CD4⁺ Vβ8⁺ cells that is detectable even at the latest timepoints. Indeed, we calculate (see Materials and Methods for details) that the proportion of undivided cells in the draining LN at 3–4 days after challenge with SEB corresponds to 35–45% (n = 7) of the initial CD4⁺ Vβ8⁺ population, confirming our previous estimates obtained with continuous 5-bromodeoxyuridine labeling (6). The presence of such a substantial fraction of undivided cells expressing a SEB-specific Vβ domain cannot be attributed simply to a failure to contact SEB, because virtually all (>95%) Vβ8⁺ T cells underwent CD62 ligand down-regulation and CD69 up-regulation following SEB injection (Ref. 25 and our unpublished observations). Whatever the explanation, it is likely that these undivided cells represent the so-called “anergic” population that remains 7–10 days after SEB challenge, because our previous studies have demonstrated that virtually all dividing cells eventually die after SEB stimulation (6).

A second important point arising from the CFSE distribution patterns in Figs. 2 and 3 is that CD4⁺ Vβ8⁺ cells are widely distributed among the division peaks at all timepoints after SEB injection. This in turn implies that there is considerable heterogeneity at the single-cell level in the SEB response in vivo. This heterogeneity in cell division among SEB-reactive cells could arise in any of the following ways: 1) asynchrony in contacting SEB and initiating the response, 2) asynchrony in reaching the first mitosis, 3) variations in cell cycle time among dividing cells, 4) variations in cell loss/apoptosis during different division cycles. Because cell cycle parameters are difficult to assess in asynchronous and nonhomogeneous cell populations, we decided to concentrate our attention on apoptosis.

Annexin V is a ligand of phosphatidylserine that, in living cells, is stringently located in the membrane leaflets that face the cytosol. During the early stages of apoptosis, the cell membrane loses asymmetry and phosphatidylserine becomes exposed on the cell surface, where it functions as a tag to signal phagocytes, which then clear the dying cell (20, 21, 22). To directly assess the relationship between apoptosis and cell division in SEB-activated T cells, we evaluated annexin V binding to the surface of dividing CD4⁺ Vβ8⁺ T cells in the draining LN at various times. For these experiments, activated T cells were incubated in vitro (1 h at 37°C) as described previously to allow apoptosis to occur (6, 26). As shown in Fig. 3 C, there was a relatively low frequency of apoptotic cells associated with the first three rounds of division irrespective of the time after SEB injection. In contrast, there was a striking increase in the number of apoptotic cells in CFSE peaks corresponding to CD4⁺ Vβ8⁺ cells that had divided four or five times. Importantly, the association of apoptosis with division peaks 4 and 5 was observed independently of the time after SEB injection, demonstrating that the number of cell divisions (rather than kinetic parameters) was responsible for this phenomenon. It should be noted that the actual number of cell divisions correlating with the induction of apoptosis was constant within a given experiment but varied slightly (usually by no more than one cycle) from one experiment to another.

In conclusion, our data suggest that upon activation by superantigens, T cells are programmed to undergo a fixed number of cell divisions before undergoing apoptosis. The molecular mechanism underlying such a delayed death program in activated T cells remains to be established. One possibility would be that some long-lived inhibitor of apoptosis is induced early after activation and gradually diluted out during subsequent cell divisions, leading eventually to apoptosis. Alternatively, a critical death effector molecule may be induced (or up-regulated) specifically after a fixed number of cell divisions. Finally, it is possible that apoptosis in this model is not causally linked to cell division but rather to some other parameter (such as activation state) that correlates with division status. Whatever the mechanism, it is tempting to speculate that apoptosis is regulated in this way so as to allow T cells to expand and perform their effector functions, while at the same time ensuring homeostasis.