H2-M3-Restricted Memory T Cells: Persistence and Activation Without Expansion

Mice, bacteria, and inoculations

Female CB6/F₁ and BALB/c mice were purchased from The Jackson Laboratory (Bar Harbor, ME). C57BL/6 K^d-transgenic mice were generated by injection of fertilized oocytes with a 20-kDa fragment of the genomic H2-K^d gene, which was kindly provided by Dr. F. Lemonnier (Pasteur Institute, Paris, France). L. monocytogenes strain 10403s was obtained from D. Portnoy (University of California, Berkeley, CA) and grown in brain heart infusion broth at 37°C, shaking, until growth was exponential (OD₆₀₀ = 0.05–0.1). Titers were calculated based on the following ratio: OD₆₀₀ of 0.1 = 2 × 10⁸ L. monocytogenes. All infections were performed i.v. by tail vein injection in a volume of 200 μl of PBS. Doses of 2 × 10³ or 1 × 10⁵ L. monocytogenes were used for inoculation of mice for primary and recall infections, respectively; any variations in dose are noted in the figures.

For peptide immunization, syngeneic (CB6/F₁) splenocytes were harvested as described below, irradiated (1000 rad), incubated with 2 × 10⁻⁶ M LLO_91–99 and fMIGWII peptides for 2 h at 37°C, and washed twice. Approximately 3 × 10⁷ peptide-coated splenocytes and 100 μg of anti-CD40 mAb (clone FGK45) were injected i.v. in a volume of 200 μl of PBS. For concurrent live infection and anti-CD40 mAb treatment, mice received 100 μg of Ab 48 and 24 h before infection, and a third dose at the time of infection.

Generation of tetramers

MHC class I tetramers were generated following the approach described previously (22, 23, 24). Modification of the murine H2-K^d cDNA to introduce a specific biotinylation site and the amplification of K^d and β₂-microglobulin were described in Ref.22 . Modification of H2-M3 cDNA was described in Ref.23 . H2-M3 and K^d H chain and β₂-microglobulin were expressed as recombinant proteins in Escherichia coli, purified from inclusion bodies, and refolded in the presence of ∼60 μg/ml LLO_91–99 peptide (for H2-K^d) or fMIGWII(A) peptide (for H2-M3). H2-M3 tetramers generated with both the hexamer (fMIGWII) and the heptamer (fMIGWIIA) forms of the listerial LemA epitope specifically stain cells from Listeria-infected mice (25); in the experiments shown in this study, tetramers made with the two peptides were used interchangeably. Peptides were synthesized by Research Genetics (Huntsville, AL).

Enrichment, staining, and FACS analysis of CD8⁺ T cells

On the indicated day after infection, spleens were harvested, dissociated through wire mesh screens, and incubated for 5 min in ammonium chloride-Tris to lyse erythrocytes. Splenocytes were filtered through mesh before MACS sorting or FACS staining. Staining of splenocytes for FACS analysis was performed on either CD8⁺-enriched cells or whole splenocytes, as specified in the figures for specific experiments.

MACS (Miltenyi Biotec, Bergisch Gladbach, Germany) was used to enrich splenocytes for CD8⁺ T cells by depletion. Cells were first incubated with Abs against CD4 (GK1.5), MHC class II (TIB120), and MAC-1 (TIB128) for 20 min on ice in separation buffer (PBS, 0.5% BSA, and 2 mM EDTA (pH 7.45)). After extensive washing, spleen cells were incubated with goat anti-rat IgG magnetic microbeads (Miltenyi Biotec) for 20 min at 4°C. Splenocytes were washed again, applied to a type LS column (Miltenyi Biotec), and separated using the MidiMACS (Miltenyi Biotec) following the manufacturer’s instructions. Before FACS staining, ∼3 × 10⁵ enriched cells (per well in a 96-well plate) were blocked with unconjugated streptavidin (0.5 mg/ml; Molecular Probes, Eugene, OR) and Fc-block (BD PharMingen, San Diego, CA) in staining buffer (SB; PBS, 0.5% BSA, and 0.02% sodium azide (pH 7.45)) for 20 min on ice. Splenocytes were then stained for 1 h on ice, covered, with anti-CD62L-FITC (clone MEL-14; BD PharMingen); anti-CD8α-CyChrome (clone 53-6.7; BD PharMingen); and PE-conjugated H2-M3/fMIGWII(A) or H2-K^d/LLO_91–99 tetrameric complexes (0.25–0.5 mg/ml). After three washes in SB, cells were resuspended in SB (100 μl) and transferred to FACS tubes containing an equal volume of 2% paraformaldehyde/PBS (pH 7.45) (final 1% paraformaldehyde).

Staining of whole splenocytes (1 × 10⁶/well) was performed as above, but different Abs were used. Following the blocking step, cells were stained on ice for 1 h with PE-conjugated tetramers, anti-CD8α-allophycocyanin (clone 53-6.7; BD PharMingen), and one of the following FITC-conjugated Abs specific for murine surface molecules (all from BD PharMingen): CD62L (clone MEL-14), CD44 (clone IM7), CD45RB (clone 16A), CD69 (clone H1.2F3), CD25 (IL-2Rα, clone 7D4), CD122 (IL-2Rβ, clone TM-β1), and Ly-6C (clone AL-21). Either propidium iodide (PI) or ethidium monoazide bromide (EMA; Molecular Probes) was added to stain dead cells (specified in the figures). When PI was used, splenocytes were resuspended in SB after the final wash, transferred to chilled FACS tubes, and kept on ice at 4°C (unfixed) until acquisition. Just before acquisition, an equal volume of 8 μg/ml PI (in SB, final 4 μg/ml) was added to the cells. When EMA was used, a concentration of 1.25 μg/ml (in SB; stock 2.5 mg/ml in dimethylformamide) was added to the buffer during the staining of the cells. For the last 10 min of the (1-h) incubation, the cells were exposed to light to cross-link DNA-bound EMA; unbound EMA was washed away in the three subsequent washes. After the final wash, the splenocytes were fixed as above.

All data were acquired using a FACSCalibur flow cytometer and analyzed using CellQuest software (BD Biosciences, Mountain View, CA).

Activation and memory markers on H2-M3- and H2-K^d-restricted T cells are similar after primary Listeria infection

Immunological memory is generated after infection with most pathogens, resulting in faster and more effective immune responses when the same pathogen is re-encountered (1). In mice reinfected with L. monocytogenes, H2-K^d-restricted, Listeria-specific T cells respond more quickly and reach much higher frequencies than seen during primary infection (6, 22, 23, 24). In contrast, H2-M3-restricted T cell populations undergo only minimal expansion following a second exposure to Listeria (23, 26). To investigate the differences between MHC class Ia (H2-K^d)- and class Ib (H2-M3)-restricted memory responses, we performed phenotypic analyses of these populations following primary and secondary L. monocytogenes infection.

H2-M3/fMIGWII and H2-K^d/LLO_91–99 tetramer-positive cells expressed similar activation and memory markers during primary and secondary L. monocytogenes infection. Seven days following primary infection, M3/fMIGWII and K^d/LLO_91–99 tetramers stained ∼4% and 2.3% of CD8⁺ cells, respectively (Fig. 1⇓, left columns). Both populations were CD62L^low, CD44^high, CD45^high, CD69^low, and CD25^low 7 days postinfection and throughout the primary response (days 5, 7, 9, 11, 14, 17, and 21 postinfection). Differences in activation marker expression between the H2-K^d- and H2-M3-restricted populations were detected only for CD122 (IL-2Rβ) and Ly-6C. These differences most likely reflect the distinct kinetics of the T cell populations; on day 5 postinfection (not shown), H2-M3-restricted T cells resembled H2-K^d-restricted T cells that are present 7 days postinfection (Fig. 1⇓), with higher levels of CD122 and more heterogeneous expression of Ly-6C. By day 9 postinfection, LLO_91–99-specific cells were CD122^low and more uniformly high for Ly-6C (not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Similar expression of T cell activation and memory markers by H2-M3-restricted and H2-K^d-restricted T cells following L. monocytogenes infection. CB6/F₁ mice were infected with 2000 L. monocytogenes, and splenocytes were stained with mAb specific for CD8α and activation/memory markers, H2-M3/fMIGWII and H2-K^d/LLO_91–99 tetramers, and EMA to exclude dead cells 7 (left columns) or 28 (right columns) days following infection. Activation/memory marker staining (specified to the right) is shown on the x-axis; tetramer staining (tetramer indicated at the top) is on the y-axis. Dot plots are gated on live CD8⁺ cells. Percentages in the upper right corners represent the percentage of activation/memory marker-positive cells among tetramer-staining cells. Three mice were analyzed each day, and data are shown for one representative animal.

Four weeks following primary L. monocytogenes infection, fMIGWII- and LLO_91–99-specific T cell populations constituted ∼0.65% and 0.40% of CD8⁺ cells, respectively. Both populations had become more heterogeneous in expression of CD62L (Fig. 1⇑, right columns), confirming our previous finding that some LLO_91–99-specific T cells re-express CD62L over time (22). Some heterogeneity in the expression of CD44, and particularly CD45RB, also developed 4 wk postinfection, although the majority of epitope-specific cells retained high levels of these markers. Some studies have indicated that memory cells are CD45RB^low; in our system, only a very small percentage of cells would fit that criterion. Finally, in contrast to the memory-like cells that undergo homeostatic expansion in “empty” mice and maintain uniformly high levels of CD122 (17, 20, 21), neither fMIGWII- nor LLO_91–99-specific T cells expressed high levels of CD122 in the weeks following primary infection. No clear differences between the H2-M3- and H2-K^d-restricted T cell populations were detected, supporting the notion that H2-M3-restricted memory cells are generated and maintained following primary infection with L. monocytogenes.

Recall infection with L. monocytogenes activates H2-M3-restricted T cells

To further characterize the disparity between H2-M3- and H2-K^d-restricted memory T cells, we measured cell size and intensity of CD8α staining, as an indication of T cell activation, at early time points following reinfection. Changes in both parameters were very similar in the two populations (Fig. 2⇓). Both M3/fMIGWIIA and K^d/LLO_91–99 tetramer-positive (and CD62L^low) T cell populations increase in size on day 3 postchallenge and return to a small size (equivalent to the tetramer-negative, CD62L^high population) by day 7. H2-M3-restricted CTL also appear to be activated on the basis of CD8 expression levels; the kinetics of CD8 up-regulation and subsequent down-regulation following secondary Listeria infection are nearly identical with those detected for the LLO_91–99 peptide-specific population (Fig. 2⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Similar changes in FSC and cell surface CD8α levels are detected for H2-M3- and H2-K^d-restricted T cell populations following L. monocytogenes challenge. CB6/F₁ mice were immunized with 2000 L. monocytogenes and then, 48 days later, challenged with 1 × 10⁵ bacteria; day 0 represents the memory population after 7 wk (not reinfected). Splenocytes were harvested on the indicated days (left), enriched for CD8⁺ cells, and stained with mAb specific for CD8α and CD62L and with H2-M3/fMIGWIIA or H2-K^d/LLO_91–99 tetramers. After gating on CD8⁺ cells, M3/fMIGWIIA and K^d/LLO91–99 tetramer-positive and tetramer-negative cells (gates as indicated in dot plot at the top) from the same mice were analyzed by flow cytometry for FSC (left panels) and CD8α (right panels) levels. Each histogram is from one representative mouse (of three; two on day 0) per day.

M3/fMIGWII tetramer-positive T cells also express early activation markers following challenge with L. monocytogenes (Fig. 3⇓). One day following reinfection, fMIGWII- and LLO_91–99-specific T cell populations constitute ∼0.5 and 0.3% of total CD8⁺ T cells, respectively, and significant percentages of both populations express CD69 (>60%) and, particularly for M3/fMIGWII-positive cells, CD25 (20–45%). During primary infection, populations of epitope-specific CTL expressing CD69 and CD25 are not detectable (data not shown) because of the low frequencies at early time points. Three days following reinfection, levels of CD122 are dramatically up-regulated, with 65% of fMIGWII- and nearly 80% of LLO_91–99-specific CTL expressing the marker. However, CD122 levels decrease by day 5 postinfection and remain low at subsequent time points (data not shown). Expression of these activation markers is not detected on all tetramer-positive cells, which may reflect heterogeneity within the populations or, given the low frequency of memory T cells, the inadvertent inclusion of nonspecifically stained cells. Nevertheless, by size parameters, modulation of CD8 levels, and expression of early activation markers, the activation states of H2-M3- and H2-K^d-restricted CTL are indistinguishable at early time points following reinfection.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Up-regulation of early activation markers on H2-M3-restricted T cells following L. monocytogenes reinfection. Four weeks following primary L. monocytogenes infection, CB6/F₁ mice were reinfected with 1 × 10⁵ bacteria. Splenocytes were harvested 1 (left columns) and 3 (right columns) days postchallenge and stained as described in Fig. 1⇑. Dot plots are gated on live CD8⁺ T cells, with staining for activation/memory markers (listed to the right) on the x-axis and tetramer staining on the y-axis. The tetramer used (H2-M3/fMIGWII or H2-K^d/LLO_91–99) is specified above each column. The percentage of tetramer-staining cells positive for each activation marker is shown. Three mice were analyzed per day, and representative data for one individual are shown.

Extended exposure to L. monocytogenes during secondary infection does not increase the expansion of H2-M3-restricted T cells

We next considered the possibility that the amount of N-formyl peptide present during secondary infection, although sufficient to induce activation, is not enough to result in dramatic proliferation of H2-M3-restricted T cells. Although immune mice are typically challenged with doses of L. monocytogenes 50-fold larger than the primary inoculum, bacteria are cleared much more quickly (27). To extend secondary infection and increase the duration of in vivo Ag presentation, immune mice were repeatedly infected with high doses of L. monocytogenes (Fig. 4⇓). Infection with 1 × 10⁶ L. monocytogenes for 2 (Fig. 4⇓, middle column) or 3 (right column) consecutive days did not increase expansion of H2-M3/fMIGWII or H2-K^d/LLO_91–99 tetramer-positive cells; in fact, frequencies of epitope-specific cells were remarkably similar to expansion detected following a typical challenge with 1 × 10⁵ L. monocytogenes (Refs.22 and23 , and data not shown). The size of the spleens in mice reinfected with one, two, or three doses of L. monocytogenes was similar (2.8, 2.7, and 2.5 × 10⁸ splenocytes, respectively), and viable bacteria were detected at the time of spleen harvest in most repeatedly infected mice (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Repeated exposure to L. monocytogenes does not enhance expansion of H2-M3/fMIGWII tetramer-positive memory cells during recall infection. Female CB6/F₁ mice were infected with 2000 L. monocytogenes and, 15 wk later, reinfected with 1 × 10⁶ Listeria one, two, or three times at daily intervals (left to right, as indicated above the dot plots). Splenocytes harvested 5 days after the first reinoculation were MACS enriched for CD8⁺ cells and stained with anti-CD8α and anti-CD62L mAb as well as MHC class I tetramers (listed to the left). Dot plots are gated on CD8⁺ T cells, with CD62L on the x-axis and tetramer staining on the y-axis. Percentages of activated, tetramer-positive cells among CD8⁺ cells are indicated in the upper left quadrants.

To rule out the possibility that H2-M3-restricted T cells fail to expand in response to challenge because of overstimulation and subsequent deletion of epitope-specific cells, immune mice were also reinfected with lower doses of L. monocytogenes. Following challenge with 1 × 10⁴ L. monocytogenes, LLO_91–99-specific T cells expanded to somewhat lower frequencies than those detected after challenge with higher doses. However, no expansion of H2-M3-restricted T cells was detected (data not shown).

H2-M3-restricted T cells do not undergo enhanced expansion if re-exposed to L. monocytogenes during the primary immune response

H2-M3-restricted T cell populations in Listeria-immune mice become activated, but do not expand dramatically if mice are reinfected 4–7 wk after the primary infection (Ref.23 ; Figs. 2⇑ and 3⇑). To determine whether H2-M3-restricted CTL expand if re-exposed to L. monocytogenes earlier following initial inoculation, immune mice were boosted with 1 × 10⁶ L. monocytogenes 6 days after primary infection. LLO_91–99-specific T cells underwent further expansion under these conditions; at the peak of expansion, the LLO_91–99-specific CTL population was ∼5-fold larger than the peak response in control mice (Ref.28 ; Fig. 5⇓A). The kinetics of the additional expansion, with a peak 5 days after the boost, suggest that LLO_91–99 peptide-specific cells with the potential to undergo memory expansion are present within the primary effector population (22, 23). In contrast to the enhanced expansion of H2-K^d-restricted T cells, there was little additional expansion of fMIGWII-specific T cells following boosting (Fig. 5⇓A). The slight increase in the frequency of fMIGWII-specific T cells is not a consistent finding and was not detected in a similar, independent experiment (D. H. Busch and E. G. Pamer, unpublished results).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Minimal expansion of H2-M3-restricted T cells following rechallenge with L. monocytogenes during the course of primary infection. BALB/c (A) or CB6/F₁ (B) mice were infected with 2000 L. monocytogenes (day 0). Six (A) or 4 (B) days after the initial inoculations, one group of mice was reinfected with 1 × 10⁶ L. monocytogenes (▪ and •). The other group was allowed to proceed through the normal course of infection (□ and ○). Splenocytes were isolated and stained on the indicated days. A, CD8⁺-enriched BALB/c splenocytes were stained with mAb specific for CD8α and CD62L as well as H2-K^d/LLO_91–99 or H2-M3/fMIGWIIA PE-conjugated tetramers. The values of two mice per group at each time point (one mouse for fMIGWIIA on day 11) are averaged. B, Unenriched CB6 splenocytes were stained with anti-CD8α and anti-CD62L mAb, H2-K^d/LLO_91–99 or H2-M3/fMIGWII tetramers, and PI to exclude dead cells. The averages of three mice per group are shown, with the exception of day 8 (four mice reinfected, two mice not reinfected). The time course of each experiment is shown on the x-axis. Absolute numbers of CD8⁺, CD62L^low (activated), and tetramer-positive cells were calculated and plotted on the y-axis. □ and ▪, H2-K^d/LLO_91–99 tetramer-positive cells; ○ and •, H2-M3/fMIGWII tetramer-positive cells. SD are indicated.

We were concerned that the differences between H2-M3- and H2-K^d-restricted CTL expansion in response to the day-6 boost might reflect the distinct expansion kinetics of these T cell populations following primary L. monocytogenes infection; on day 6 after primary infection, H2-M3-restricted CTL are beginning the contraction phase, while H2-K^d-restricted T cells are in the midst of vigorous expansion (23). To address this issue, mice were boosted 4 days after primary infection (Fig. 5⇑B) because, on day 4 after primary infection, both H2-M3- and H2-K^d-restricted T cell populations are expanding. The results of this experiment also show that LLO_91–99-specific T cells undergo additional expansion while fMIGWII-specific T cells are unaffected. Thus, the different capacities of H2-M3- and H2-K^d-restricted T cells to respond to an Ag re-exposure are already apparent during the primary expansion phase of the T cell populations.

Peptide immunization primes fMIGWII and LLO_91–99 peptide-specific T cells that can respond to subsequent infection with L. monocytogenes

The challenge inoculum of L. monocytogenes, whether administered 4 days or 7 wk after the primary infection, is cleared very rapidly (data not shown). To determine whether Ag-experienced, H2-M3-restricted CTL are capable of expansion under primary infection conditions, we primed epitope-specific T cells by immunization of mice with peptide-coated splenocytes and anti-CD40 mAb. Immunization with fMIGWII- and LLO_91–99-coated splenocytes primed peptide-specific T cell responses restricted by H2-M3 and H2-K^d, respectively (Fig. 6⇓). Subsequent L. monocytogenes infection of peptide-immunized mice resulted in impressive expansion of M3/fMIGWII and K^d/LLO_91–99 tetramer-positive T cells (Fig. 6⇓). The frequency of LLO_91–99-specific T cells 5 days after reinfection was slightly lower in mice that were peptide-immunized than in mice previously immunized with L. monocytogenes (Fig. 6⇓B, left panels). However, the absolute numbers of LLO_91–99-specific T cells that expanded under these two conditions were similar (Fig. 6⇓A). The recall frequencies in this experiment are low compared with previous studies (23) because the challenge dose was only 5000 bacteria, because peptide-immunized mice were not expected to have high-level protective immunity. Remarkably, the fMIGWII-specific T cell population expanded to a similar extent as LLO_91–99-specific T cells in peptide-primed mice (Fig. 6⇓, B, lower right panel, and A). As expected, the H2-M3 memory response in mice previously immunized with live L. monocytogenes was small (Fig. 6⇓A) and stained with lower intensity with the tetramers (data not shown). It is interesting that LLO_91–99-specific T cells did not expand dramatically in L. monocytogenes-primed mice reimmunized with peptide-coated splenocytes. It is possible that LLO-coated APCs are rapidly eliminated by memory T cells in these immune mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Memory-like response of H2-M3-restricted T cells to L. monocytogenes following primary exposure to Listeria-derived epitopes. Female CB6/F₁ mice were immunized with 2000 L. monocytogenes or peptide-coated splenocytes (fMIGWII and LLO_91–99 plus anti-CD40 mAb) at 8 wk of age (1°, Primary exposure). Four weeks later, Listeria- and peptide-immunized groups were exposed again (2°) to either Listeria Ag in the form of 5000 L. monocytogenes or peptide-coated splenocytes. Naive mice from the same cohort were infected with 5000 Listeria (1° Lm) or immunized with peptide-coated splenocytes at the same time (left columns in A). Five days after the 2° exposure, splenocytes were stained with anti-CD8α and anti-CD62L mAb, H2-K^d/LLO_91–99 or H2-M3/fMIGWII tetramers, and PI. A, Total CD8⁺, CD62L^low, and tetramer-positive cells per spleen were averaged (y-axis) for three mice from each immunization protocol (x-axis). SD are indicated. B, Frequencies and FSC data are shown for representative mice from the 1° Listeria→2° Listeria (top panels) and 1° peptide→2° Listeria groups. Dot plots are gated on live CD8⁺ T cells; CD62L is on the x-axis, with tetramer staining (H2-K^d/LLO_91–99 (left) or H2-M3/fMIGWII (right)) on the y-axis. Percentages of activated tetramer-positive cells (among CD8⁺ cells) are indicated in the upper left quadrants. Histograms of FSC (x-axis) are gated on CD8⁺ cells and compare activated tetramer-positive cells (bold line) to tetramer-negative, CD62L^low cells. C, C57BL/6-H2-K^d-transgenic mice were either treated with FGK45 anti-CD40, as described in Materials and Methods, or left untreated. Mice were infected with 2,000 L. monocytogenes and, 3 wk later, reinfected with 100,000 bacteria. Five days following reinfection, splenocytes were stained with H2-M3/fMIGWII tetramers or H2-K^d/LLO_91–99 tetramers, as indicated.

To determine whether anti-CD40 stimulation during primary infection with live bacteria alters the generation of subsequent H2-M3-restricted memory responses, we treated mice with three doses of 100 μg of FGK45 anti-CD40 Ab, 48 and 24 h before and at the time of primary L. monocytogenes infection. Fig. 6⇑C demonstrates that anti-CD40 treatment in the context of a live infection did not enhance the H2-M3-restricted memory T cell response.

The increased frequencies of fMIGWII- and LLO_91–99-specific T cells in peptide-immunized mice (compared with naive mice) indicates that the T cells are Ag experienced. However, the T cell memory responses in mice primed with peptide and live L. monocytogenes are not identical. For example, LLO_91–99- and fMIGWII-specific memory T cells in mice immunized with live L. monocytogenes are large, as measured by forward scatter (FSC), 3 days postchallenge and become small by the fifth day (Figs. 2⇑, and 6⇑B, top histograms). In contrast, memory T cells in peptide-immunized mice remain large on the fifth day after live bacterial challenge, suggesting that these T cells are still actively proliferating (Fig. 6⇑B, bottom histograms). Although we interpret these differences in T cell size as a reflection of different kinetics of bacterial clearance and inflammation in the two experimental groups, it is also possible that they result from disparities in the priming of CD8⁺ T cells by live infection vs peptide immunization.