Stability of Virus-Specific CD4⁺ T Cell Frequencies from Acute Infection into Long Term Memory¹

Although the mechanism of T cell activation and proliferation is becoming increasingly understood, the specificities of T cells responding to infection and the factors contributing to the conversion of an acute T cell response into a memory response remain poorly defined. Most examinations of these processes in viral infections have focused on CD8⁺ T cells, which vigorously respond to many viral infections, often resulting in T cell hyperplasia and a conversion of the CD4 to CD8 ratio from 2:1 to 1:2. Rather surprising recent data have shown that when apoptotic events reduce the T cell number after the viral clearance, the virus-specific CTL precursors (CTLp)³ per CD8⁺ T cell remain remarkably high in the memory state, within 12 to 50% of the levels seen at the peak of the acute infection (1). In addition, this virus-specific memory CD8⁺ T cell response is extremely long-lasting and, in the case of memory CD8⁺ T cell precursors to lymphocytic choriomeningitis virus (LCMV), is maintained at steady state levels for the lifetime of the mouse (1, 2, 3). These results may indicate that long after a viral infection has cleared, the memory CD8⁺ T cell pool remains highly biased with virus-specific T cells, closely reflecting the population present at its peak during infection but partially diluted with naive cells.

Although usually not as vigorously stimulated as CD8⁺ T cells, CD4⁺ T cells also become activated and respond to viral infections (4). In some but not all systems, their response is required for the induction of CD8⁺ CTL, and even in systems such as LCMV, where their response is not required for CTL induction (5, 6, 7), CD4⁺ T cells may be required for the long term maintenance of CD8⁺ T cell memory (8). Surprisingly little quantitative information, however, is known about the acute CD4⁺ T cell responses to experimental viral infections and about the frequencies and stability of virus-specific CD4⁺ T cells in the memory state. How memory CD4⁺ T cells may contribute to the stability of memory CD8⁺ T cells is also unclear. To begin to address these issues, we have undertaken a quantitative analysis of the mouse CD4⁺ T cell response to LCMV, a virus for which the CD8⁺ T cell response has been very well analyzed. We show here that, like the CD8⁺ T cell frequencies, the CD4⁺ T cell frequencies remain elevated after the acute infection subsides and stay elevated throughout long term immunity, albeit in a less active state.

C57BL/6 (H-2^b) mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and used at 2 to 14 mo of age for most experiments. The LCMV Armstrong strain was propagated in BHK21 baby hamster kidney cells (9). Mice were inoculated i.p. with 7 × 10⁴ plaque-forming units of LCMV diluted 1:100 in PBS in 0.1 ml vol/mouse. This dilution in PBS was done to prevent CD4⁺ T cell activity against cell debris or FCS Ags present in the virus stock.

Mice were killed by cervical dislocation, and spleens were removed aseptically. Splenic leukocytes were obtained by preparing single-cell suspensions from spleens and treating them with 0.84% NH₄Cl to lyse Es, as described previously (10). Cell culture medium used was RPMI medium (Sigma, St. Louis, MO) containing 10% FCS (Sigma), 100 U/ml penicillin G, 100 mg/ml streptomycin sulfate, 2 mM l-glutamine, 5 × 10⁻⁵ M 2-ME, 0.1 mM sodium pyruvate (Life Technologies, Grand Island, NY), 0.1 mM nonessential amino acids (Life Technologies), and 10 mM HEPES. Splenic leukocytes from individual animals were analyzed in two-color mode on a FACScan cytometer (Becton Dickinson, Mountain View, CA). The mAb used for staining were phycoerythrin (PE) conjugated (H129.19 anti-CD4 and PK136 anti-NK1.1) or FITC conjugated (53-6.7 anti-CD8, IM7 anti-CD44, RA3-6B2 anti-CD45R/B220, MEL-14 anti-CD62L, and H1.2F3 anti-CD69). All mAb were purchased from PharMingen (San Diego, CA). Approximately 1 × 10⁶ cells were stained, and 5,000 to 10,000 events were acquired from each preparation. The data were analyzed using either PC-Lysis or Cell Quest software (Becton Dickinson).

ELISPOT assays for cytokine-secreting cells were performed based on slight modifications of established protocols (11, 12). Briefly, 96-well nitrocellulose-based microtiter plates (Millititer HA; Millipore, Bedford, MA) were coated overnight at 4°C with 50 μl/well of anti-cytokine mAb diluted in PBS. After the plates were washed with PBS, all wells were blocked with 200 μl of RPMI 1640 containing 10% FCS for 2 h at 37°C. After a washing with PBS, the lymphocyte populations (either splenic leukocytes or FACS-separated T cell populations) were then added to the wells (1 × 10⁵ to 1.3 × 10⁴ or from 1 × 10⁵ to 7.8 × 10² cells/well for responders with 3 × 10⁴ LCMV-infected irradiated (2000 rads) peritoneal exudate cells (PEC) as stimulators) in RPMI 1640, 10% FCS (200 μl/well total volume) and incubated for 20 h at 37°C. After the wells were washed in PBS-Tween, biotinylated anti-cytokine mAb were added, diluted in PBS containing 10% FCS (100 μl/well), and incubated overnight at 4°C. Plates were washed in PBS-Tween, and 100 μl of an anti-biotin mAb conjugated with peroxidase (1/250 dilution in PBS, 10% FCS; Jackson Immunoresearch Laboratories, West Grove, PA) per well were added, followed by another overnight incubation at 4°C. Spots representing individual cytokine-secreting cells were visualized by developing with the substrate 3-amino-9-ethylcarbazole and counted using an Olympus SZ-STS Stereozoom microscope (Lake Success, NY). All assays were performed in triplicate. Mean numbers of cytokine-secreting cells were calculated from the triplicate assays. The results shown are means ± SD for two to three separate experiments. The following pairs of mAb were used in the ELISPOT assays: anti-IL-4, BVD4-1D11, and biotinylated BVD6-24G2; anti-IFN-γ, R4-6A2, and biotinylated XMG1.2. The mAb were used at a concentration of 2 μg/ml, with the exception of R4-6A2, which was used at 10 μg/ml. All of the mAb were obtained from PharMingen.

Unsorted splenic leukocytes or FACS-separated T cells were resuspended at a final density of 1 × 10⁶ cells per ml in RPMI 1640 supplemented as above. Responders (100 μl) were added to 3 × 10⁴ LCMV-infected irradiated (2000 rads) stimulator PEC in RPMI 1640 (200 μl/well total volume) and incubated at 37°C. Supernatants were harvested at 24, 48, 72, and 96 h and stored at −80°C before being assayed for cytokines by ELISA. Dynatech Immulon 4 plates (Dynatech, Chantilly, VA) were coated by incubation overnight at 4°C with anti-cytokine mAb diluted in PBS (50 μl per well). The plates were then washed with PBS-0.05% Tween 20 after each of the following steps. Plates were blocked with PBS-10% FCS (200 μl/well) for 2 h at room temperature. Samples were added at 100 μl/well, and a standard curve was constructed for each plate by using eight twofold dilutions of recombinant cytokine, and the plates were again incubated overnight at 4°C before the addition of biotinylated anti-cytokine mAb at 100 μl/well. After a 1-h incubation at room temperature, 100 μl of avidin-peroxidase (1/400 dilution in PBS, 10% FCS; obtained from Sigma) per well were added, and the plates were incubated at room temperature for 30 min before detection with 100 μl of the peroxidase substrate, 3,3′,5,5′-tetramethylbenzidine dihydrochloride (Sigma) dissolved in 0.05 M phosphate-citrate buffer, pH 5.0. The reaction was stopped with 25 μl of 2 N H₂SO₄. Plates were read at 450 nm using a THERMO_MAX plate reader and analyzed using SoftMax 2.3 (both from Molecular Devices, Menlo Park, CA). The pairs of anti-cytokine mAb listed above for the ELISPOT assays were used in the ELISA, all at a concentration of 2 μg/ml. Recombinant murine IFN-γ and IL-4 used as standards were obtained from PharMingen.

CD4 limiting dilution assay (LDA) were performed based on modifications of established protocols (13, 14, 15). Splenic leukocytes from a pool of three to four mice were stained with PE-conjugated anti-CD4 and FITC-conjugated anti-CD8. They were then sorted in two-color mode on a FACStar^Plus flow cytometer from Becton Dickinson. Sorted CD4⁺ and CD8⁺ cells were consistently >94% pure. Cultures were set up in 96-well U-bottom plates (Falcon, Lincoln Park, NJ) in a final volume of 200 μl in RPMI medium. Dilutions of sorted CD4⁺ T cells in 100 μl were plated as replicates of 16 or 24 microcultures with 3 × 10⁴ uninfected (mock-infected with BHK21 cell supernatant) or LCMV-infected irradiated (2000 rads) PEC added in an additional 100 μl. The cultures were incubated at 37°C in 5% CO₂. At various times after the initiation of the cultures, the plates were centrifuged at 1000 rpm for 5 min, 50-μl aliquots of the supernatants were transferred to new plates, and IL-2 activity was measured using the CTLL-2 biologic assay. Briefly, the CTLL-2 cells were grown in 25-cm² flasks in RPMI 1640 medium supplemented with 1 U/ml of recombinant IL-2 (Cellular Products, Buffalo, NY). The CTLL-2 cells were washed three times with RPMI 1640, and 5 × 10³ CTLL-2 cells in 50 μl were added to 50 μl of the cell culture supernatants in 96-well plates and incubated for 18 h at 37°C in 5% CO₂. Wells were then pulsed with 1 μCi of [³H]TdR in 50 μl of RPMI for 6 h. The plates were harvested onto glass fiber filters, and [³H]TdR incorporation was assayed in a Wallac Betaplate scintillation counter (Wallac Gaithersburg, MD). In two experiments, IL-4 activity was determined by the proliferation of the IL-4-dependent cell line, CT.4S (provided by Dr. William Paul, National Institutes of Health, Bethesda, MD). Briefly, 50-μl aliquots of the supernatants were incubated for 24 h with 5 × 10³ CT.4S cells in 50 μl and then pulsed overnight with 1 μCi of [³H]TdR. Microcultures giving values in the CTLL-2 or CT.4S assay of >3 SDs above the mean values obtained for APC alone were scored as positive. The frequencies were corrected for the purity of the sorted CD4⁺ populations. Frequencies were calculated using χ² analysis according to the method of Taswell (16) on a computer program provided by Dr. Richard Miller (University of Michigan, Ann Arbor, MI). The frequencies reported against uninfected APC are estimated, given that the curves generated from the responder dilution series often yielded linear regression lines that never crossed the 37% negative well threshold. Thus, these frequencies may reflect slight overestimates of the background response to uninfected APC. However, the response against uninfected APC was always significantly lower than that against virus-infected APC in all of the LCMV-infected mice.

Figure 1 shows the percentage of CD4⁺ and CD8⁺ T cells in the spleen of C57BL/6 mice acutely infected with LCMV. Also shown in Figure 1 is the twofold increase in the total leukocyte number in the spleen that occurs during the acute LCMV infection. The increase in the percentage of CD8⁺ T cells combined with the increase in the total number of cells results in a massive 5- to 20-fold expansion in the number of CD8⁺ T cells in the spleen during LCMV infection (10). In contrast, there is only a modest increase in the total number of CD4⁺ T cells in the spleen, because the percentage of CD4⁺ T cells has decreased at the peak in the leukocyte number (Fig. 1). Although the total CD4⁺ T cell number remains stable, further analyses revealed evidence of CD4⁺ T cell activation as shown by blastogenesis and activation Ag expression. The mean percentages of blast-sized CD4⁺ T cells from eight individual mice per group were as follows: day 0, 8; day 3, 9; day 5, 16; day 7, 25; day 9, 21; day 11, 19; day 15, 11. Figure 2 shows the distribution of CD4⁺ T cells following an acute infection with LCMV with regard to the Ags CD44 (Pgp-1), a marker the cell surface expression of which increases on activated T cells, and CD62L (MEL-14), a molecule with a cell surface expression that decreases on activated T cells (17). Table I shows that there is an increase in the mean fluorescence intensity (MFI) of CD44 on gated CD4⁺ T cells, and as can be seen in Figure 2, this increase in the MFI is mainly due to an increase in the relative proportion of CD4⁺ cells expressing a high level of CD44 rather than all of the CD4⁺ T cells staining more highly with CD44. Likewise, Table I shows that there is a decrease in the MFI of CD62L on gated CD4⁺ T cells, and as can be seen in Figure 2, this decrease in the MFI seems to be due to an increase in the relative proportion of CD4⁺ cells expressing a low level of CD62L. We also performed a more detailed time course of the cell surface expression of CD44 and CD62L at days 0, 3, 5, 7, 9, 11, and 15 postinfection (p.i.). The peak in cell surface expression of the activation marker CD44 was at day 11 p.i. and correlated with the lowest cell surface expression of CD62L (data not shown).

In addition, we examined the cell surface expression of several other activation markers and found increases in the MFI of CZ-1 (a sialated form of CD45RB), CD25 (IL-2R), and CD69 (very early activation Ag) as well as of the adhesion molecules CD49d (VLA-4) and LFA-1 (CD11a) on gated CD4⁺ T cells after LCMV infection (Table I). In contrast to CD44 and CD62L discussed above, CD4⁺ T cell surface expression of the very early activation Ag CD69 peaked much earlier, at 3 days p.i., and declined thereafter (data not shown). CD45RB cell surface expression has been shown to increase on effector CD4⁺ T cells (17). In agreement with this, we observed an increase in the cell surface expression of CD45RB on CD4⁺ T cells at day 10 p.i. with LCMV. Interestingly, while CD45RB expression increases on activated cells and then decreases on memory CD4⁺ T cells, a sialated form of this protein (CZ-1), previously shown to be bright on activated cells, remains high on memory CD4⁺ T cells, and this is reflected in the higher expression of CZ-1 vs CD45RB on CD4⁺ immune cell populations (day 60, Table I) (18). Thus, LCMV infection induces an activated phenotype on CD4⁺ T cells that peaks by day 11 p.i.

The above experiments show that CD4⁺ T cells expressing an activated cell surface phenotype are induced following an acute LCMV infection. Since one of the primary effector functions of CD4⁺ T cells is the secretion of various cytokines, we utilized the ELISPOT assay to assess IFN-γ and IL-4 production at the single-cell level in C57BL/6 mice acutely infected with LCMV. Uninfected mice (day 0) had no detectable IFN-γ- or IL-4-secreting cells (hereafter referred to as IFN-γ⁺ and IL-4⁺ cells) in the spleen. Following an infection with LCMV, the frequency of IFN-γ⁺ cells per spleen leukocyte started to increase by day 5 p.i. and reached its peak by day 9 p.i. The mean numbers of IFN-γ-secreting cells per 10⁵ splenic leukocytes from eight individual mice per group were as follows: day 0, 0; day 3, 0; day 5, 6; day 7, 177; day 9, 346; day 11, 178; day 15, 124. On a per spleen basis, the peak in the total number of IFN-γ⁺ cells was even more pronounced at days 7 to 11 p.i., because the number of cells in the spleen had nearly doubled (Fig. 1). By day 15 p.i., the frequency of IFN-γ⁺ cells had started to decline. The decline in the frequency of IFN-γ⁺ cells after days 7 to 11 p.i. coincided with the clearance of the virus and the decline of the CD8⁺ CTL response. In contrast to IFN-γ, few IL-4⁺ cells could be detected in the spleen, agreeing with previous work suggesting that LCMV induces primarily a Th1 response (19).

CD4⁺ and CD8⁺ T cells from the spleens of mice that had been acutely infected with LCMV were purified by cell sorting to determine whether virus-specific CD4⁺ T cells contributed to the secretion of IFN-γ in the ELISPOT assays described above. Table II, which presents the results from two to three separate experiments per time point, shows that CD8⁺ T cells made up the majority of IFN-γ-secreting cells during the acute LCMV infection. However, there were detectable frequencies of LCMV-specific CD4⁺ T cells secreting IFN-γ. Furthermore, we noticed in these assays that, while on average 11-fold more CD8⁺ than CD4⁺ T cells secreted IFN-γ, the CD4⁺ “spots” in the assays were consistently larger. This suggested that the CD4⁺ T cells may be making more IFN-γ on a per cell basis than the CD8⁺ T cells. Previous work with in vitro systems has shown that CD4⁺ T cells can secrete higher levels of cytokines than similarly stimulated CD8⁺ T cells (20). To examine whether this were the case during LCMV infection, we FACS-purified CD4⁺ and CD8⁺ T cells at various times p.i., restimulated them in vitro with virus, and performed ELISA assays on the culture supernatants. Table II shows that CD4⁺ T cells secreted as much IFN-γ protein into the cell culture supernatant as CD8⁺ T cells, even though there were 11-fold more CD8⁺ T cells secreting IFN-γ than CD4⁺ T cells, as detected with ELISPOT assays. This indicates that the CD4⁺ T cells make more IFN-γ on a per cell basis than do the CD8⁺ T cells.

When splenocytes from LCMV-immune mice (day 180 p.i.) were used in the ELISPOT assay, the frequency of unsorted and sorted CD8⁺ T cells secreting IFN-γ in the 20-h assay was greatly reduced compared with cells isolated during the acute LCMV infection. We were unable to detect any CD4⁺ T cells secreting IFN-γ from the memory mice. This reduced frequency likely reflects a decrease in the activation status of these cells rather than a reduction in the frequency of cells capable of secreting IFN-γ, as discussed below. The relatively short assay time (20 h) of the ELISPOT assay is probably not long enough to fully stimulate the memory T cells as opposed to the already activated cells taken from the acute LCMV infection.

To quantitate the virus-specific CD4⁺ T cell response during LCMV infection, an assay for LCMV-specific Thp was adapted from the method used by recent studies in the influenza and Sendai virus systems (14, 15, 21). Figure 3 shows a typical regression line obtained from C57BL/6 mice during the acute infection with LCMV (day 7) and into memory (day 60). We performed kinetic studies (Fig. 3) to confirm that we were measuring the peak of the IL-2 production in the LDA cultures. These kinetic studies revealed that high frequencies of Thp during the acute LCMV infection and into memory were found after at least 48 h following the initiation of the culture. These results and interpretations agree with the initial studies performed using Sendai virus (14).

Analysis of Thp frequencies during the acute LCMV infection and on into memory showed that the Thp frequencies remain quite stable for at least 1 year p.i. (Table III). The LCMV-specific Thp frequency rose from <1/100,000 in a naive animal (day 0) to ∼1/600 by day 10 p.i. with LCMV. This peak in the Thp frequency by day 10 p.i. corresponds with that of the activation markers CD44 and CD62L discussed above. Interestingly, the Thp frequency per CD4⁺ T cell dropped only twofold from the peak of the response (days 9–11) on into long term memory. This relatively small reduction in the Thp frequency mirrors the twofold drop in frequency that we have previously reported for the CD8⁺ T cell response during LCMV infection (1). In fact, on a per spleen basis, the decline in the total number of CD4⁺ memory T cells is even less than the decline in the total number of CD8⁺ memory T cells, because the percentage of CD4⁺ T cells increases between the peak of the acute infection and memory, whereas the percentage of CD8⁺ T cells decreases (see Fig. 1).

The CD4⁺ Thp frequencies obtained against both uninfected and LCMV-infected APC (Tables III and IV) are presented because there is consistently a low level response against uninfected APC. Two additional controls were performed to ensure that we were measuring LCMV-specific CD4⁺ Thp producing IL-2 in these assays. Even though we used FACS-purified CD4⁺ T cells in these LDA, we could not rule out that a very small contamination (<1%) by CD8⁺ T cells might be present in our cultures and release IL-2 in response to the virus-infected APC. CD8⁺ T cells, when present, make enough IL-2 to be detected in this assay (data not shown). As shown in Table IV, treatment with anti-CD4 mAb blocked all of the virus-specific IL-2 production, returning the precursor frequency to a level comparable to that obtained against uninfected APC. Second, because there was the possibility that the CTLL-2 cells used to monitor IL-2 production in these LDA could be detecting IL-4, we added anti-IL-4 mAb to the assays. Table IV shows that there was no effect (within the 95% confidence limits) in three of the four experiments when anti-IL-4 mAb was added to the LDA cultures. In one experiment, using LCMV-immune animals, the presence of the anti-IL-4 mAb only slightly (within the 95% confidence limits) reduced the resulting precursor frequency. In two additional experiments, we used the very IL-4-sensitive CT.4S cells line in the LDA. No IL-4 could be detected using this indicator cell line at day 15 or 240 p.i. in the LDA. LCMV infection is thought to stimulate primarily a Th1 cytokine profile in the responding T cell populations (19). The anti-IL-4 blocking experiments and use of the CT.4S cells support this, suggesting that the LCMV-specific CD4⁺ T cells are primarily of a Th1 phenotype.

Here we report that the frequency of LCMV-specific CD4⁺ T cells within the CD4⁺ T cell pool remains remarkably constant from its peak in the acute viral infection and throughout long term memory after clearance of the virus. Although this memory CD4⁺ T cell population is in a lower activation state, as shown by activation markers and the inability to produce IFN-γ in short-term assays, the Thp frequencies 1 year after infection are only twofold lower than their peak at day 10 p.i. Virus-specific CD8⁺ CTL memory has also been shown to be long-lived (2, 22), and we recently reported that the frequency of LCMV-specific precursor CTL per CD8⁺ T cell drops only twofold from its peak, as the acute LCMV infection converts into a memory state (1). This 2-fold drop per CD8⁺ T cell occurs during the more global 5- to 10-fold drop in the total CD8⁺ T cell number per spleen, as the immune response silences and the CD8/CD4 ratio converts from 2:1 to 1:2, as shown in Figure 1. A similar observation has been made in an LCMV-specific CD8⁺ TCR transgenic model, in which the frequency of adoptively transferred transgenic T cells reached 70% of the CD8⁺ T cells during the peak of the acute infection and remained at 27% in the memory state (23). Thus, in the wake of a substantial reduction in the number of CD8⁺ T cells, the proportion of virus-specific cells within this compartment is, like our present observation with CD4⁺ T cells, only moderately modulated, in a manner consistent with a 1:1 dilution with naive cells. These results indicate that after an acute viral infection resolves, both the CD4⁺ and CD8⁺ T cell pools remain heavily biased with virus-specific memory T cells.

Little quantitative information has been generated on the stability of CD4⁺ T cells in viral infections. Some quantitation of CD4⁺ T cell precursors has been performed in the Sendai and influenza virus systems (14, 15, 21). In each of these systems, the observed peak in Thp frequency is ∼1/100, higher than that we have observed here for LCMV-specific Thp. However, the Thp frequency in the memory state in each of these systems is near 1/1400, very close to the frequency of 1/1200 we have shown here for LCMV-specific Thp (Table III). In addition, Topham et al. (15) show that the influenza-specific Thp frequency remains relatively stable in the spleen for 6 mo but drops off during this time in both the mediastinal and cervical lymph nodes. Thus, in both the Sendai and influenza virus systems, there seems to be a larger, 10-fold drop in the Thp frequency from the peak of the response into long term memory due to a higher peak in the Thp frequency (4, 15). We have not seen such a drop in the LCMV system for either CD4⁺ or CD8⁺ T cells (Table III and 1 . Whether this reflects differences in the immune response to these viruses or differences in the assay systems is unclear. T cells in the resting memory state may have different activation requirements than those during the acute infection, and subtle differences in the LDA might affect that outcome. Earlier work on LCMV has suggested more extreme drops on CTP precursor (CTLp) frequencies between the acute infection and the memory state (2, 8, 24), but analysis of our own data and of recent data from other laboratories indicates that there is very little drop when the assays are optimized and the data are expressed as CTLp per CD8⁺ T cell (1, 2, 23).

Factors contributing to the maintenance and preservation of memory T cells are poorly understood. Recent work has indicated that the presence of class II and class I MHC molecules is required for the respective survival of CD4⁺ and CD8⁺ T cells in vivo (25, 26). This argues that the TCR on memory cells need to be triggered for their survival, but this is likely to be a lower affinity event than that which initially activated the cells, because memory cells can persist in vivo in the absence of the Ag to which they were initially generated (2, 22). The potent CD8⁺ CTL response during the LCMV infection does not require CD4⁺ T cells for its induction (5, 6, 7), but in the absence of CD4⁺ T cells LCMV-specific CD8⁺ CTL memory rapidly wanes (8). This suggests that the CD8⁺ T cells receive growth or survival factors from the CD4⁺ T cells to perpetuate in this environment devoid of the strong Ag stimulus that elicited the CD8⁺ T cell response in the first place. At least some memory CD8⁺ T cells appear to be in a higher activation state than memory CD4⁺ T cells. At any given time, a higher percentage of CD8⁺ T cells is undergoing blastogenesis than CD4⁺ T cells (27), and Table II shows that in 20-h ELISPOT assays, some IFN-γ production can be detected in the CD8⁺ T cell population but not in the CD4⁺ T cell population. We have recently reported that enriched blast-sized CD8⁺ T cells >1 year after the infection has cleared can still mediate direct cytotoxic activity against sensitive targets (28). We are left with the irony that the memory CD8⁺ T cells are more active than the memory CD4⁺ T cells and yet require the CD4⁺ T cells to maintain their activity. Whether autologous stimulation is sufficient for the maintenance of the more dormant CD4⁺ T cell population is not known.

In this study, we utilized a panel of cell surface markers to define activated CD4⁺ T cells and showed that by day 10 p.i. with LCMV, there was an increase in the cell surface expression of activation markers such as CD44, CD49d, LFA-1, and CZ-1 on gated CD4⁺ T cells (Fig. 2 and Table I). While there was little increase in the absolute numbers of CD4⁺ cells in the spleen expressing these activation markers, there was an increase in the relative proportion of CD4⁺ cells expressing these activation markers (Fig. 2). This observation agrees well with a recent study using the Traub strain of LCMV in BALB/c mice (29). Our previous work has shown that many of these CD4⁺ T cells are blast-sized at day 6 p.i., in agreement with an activated state (18). Here we report that 25% of the CD4⁺ T cells in the spleen were blast-sized at day 7 p.i. These data contrast with a report by Christensen et al. using the Traub strain of LCMV in BALB/c mice, in which they show that only 2% of CD4⁺ T cells exhibit increased DNA content during the peak of the infection (30). The discrepancy between these findings and ours may be due to the differences in the strains of mice or viruses that were used in these studies. Even though 25% of the CD4⁺ T cells were blast-sized, there was little or no increase in the total number of CD4⁺ T cells in the spleen during an acute LCMV infection (Fig. 1). This observation highlights a difference between the LCMV virus system and other protein-Ag systems in which peptide-specific CD4⁺ T cells have been shown to expand in number (31, 32). Interestingly, the peak of expression of the very early activation Ag CD69 occurred at 3 days p.i., well before the peak of the other markers discussed above (Table I). This expression pattern parallels the type I IFN response in these mice (33). To our knowledge, type I IFN has not been previously reported to up-regulate CD69 expression on CD4⁺ T cells, but it has been shown to be induced by IFN-α on NK cells (34).

The frequency of IFN-γ-secreting cells in the spleen peaked by day 9 p.i. in general agreement with Gessner et al. (Table II) (35, 36). Our frequency of IFN-γ producing splenocytes at day 7 p.i. was 10-fold higher than that reported by Cousens et al. (37) using the same strain of mice and virus used here. The reason for this discrepancy may be due to us adding LCMV-infected PEC to our ELISPOT assays to enhance the virus-specific stimulation during these assays. We found consistently more CD8⁺ T cells secreting IFN-γ than CD4⁺ T cells during the acute LCMV infection, as detected at the single-cell level utilizing the ELISPOT assay. This ratio was not surprising, given that there were ∼10-fold more CTLp than Thp at these time points during the acute LCMV infection (1). Interestingly, in four independent experiments, the CD4⁺ and CD8⁺ T cells secreted comparable amounts of IFN-γ protein into the supernatant, even though there were many more IFN-γ-secreting CD8⁺ T cells than CD4⁺ T cells as detected by ELISPOT assays (Table II). This observed difference may be explained by fact that LCMV infection induces CD8⁺ CTL with high cytotoxic activities that may be capable of destroying virus-infected APC and thus reducing the amount of IFN-γ that is produced by the CD8⁺ CTL (38). However, this seems unlikely given that the number of LCMV-infected APC used in these assays is in great excess over the frequency of responding CD8⁺ T cells. Of interest is that recent studies have found that CTL clones with high cytotoxic activity produced low amounts of cytokines (39).

Examination of the CD4⁺ Thp frequency during an acute LCMV infection and on into the memory state revealed that within 10 days p.i. with LCMV, the Thp frequency rose from <1/100,000 to ∼1/600. The Thp frequency remained elevated in immune animals, dropping only twofold to ∼1/1200, and remained very stable into long term memory. While the CD4⁺ T cell response to LCMV resembles that of the CD8⁺ T cells, there are significantly fewer virus-specific CD4⁺ than CD8⁺ T cells. It is not clear whether the difference in precursor frequencies between the CD4⁺ and CD8⁺ T cells reflects real differences or is due to the inherent limitations of the CD4⁺ Thp LDA. The efficiency of these assays is understood not to be 100%, and studies have provided indirect evidence that most of the activated and proliferating CD8⁺ T cells during the acute LCMV infection are LCMV specific (40), even though only 3 to 4% can be detected by LDA. It is also possible that some virus-specific Thp secrete cytokines other than IL-2 (such as IFN-γ or TNF-α) and would not be detected by the CTLL-2 assay used here.

In the present study, we have shown that there is little change in the frequency of virus-specific CD4⁺ T cells from the peak of the response into long term memory. Thus, the host T cell repertoire remains biased with a high frequency of LCMV-specific memory CD4⁺ T cells after clearance of the acute virus infection.

We thank Carey O’Donnell for her technical assistance, Liisa K. Selin for helpful discussions, and Tammy Krumpoch and Barbara Fournier for help with the FACS analysis.