Long-term persistence of Ag-experienced CD8 cells, a class of T lymphocytes with cytotoxic function, contributes to immunological memory against intracellular pathogens. After Ag clearance, memory CD8 cells are maintained over time by a slow proliferation, primarily cytokine driven. In this article, we show that the bone marrow (BM) is the crucial organ where such basal division of memory CD8 cells occurs. BM memory CD8 cells contain a higher percentage of proliferating cells than their corresponding cells in either spleen or lymph nodes from C57BL/6 mice. This occurs both in the case of memory-phenotype CD44high CD8 cells and in the case of Ag-specific memory CD8 cells. Importantly, the absolute number of Ag-specific memory CD8 cells dividing in the BM largely exceeds that in spleen, lymph nodes, liver, and lung taken together. In the BM, Ag-specific memory CD8 cells express lower levels of CD127, i.e., the α-chain of IL-7R, than in either spleen or lymph nodes. We interpret these results as indirect evidence that Ag-specific memory CD8 cells receive proliferative signals by IL-7 and/or IL-15 in the BM and propose that the BM acts as a saturable “niche” for the Ag-independent proliferation of memory CD8 cells. Taken together, our novel findings indicate that the BM plays a relevant role in the maintenance of cytotoxic T cell memory, in addition to its previously described involvement in long-term Ab responses.

Immunological memory is the capacity of the immune system to generate a fast and protective response upon re-encounter with the same pathogen. One component of immunological memory against viruses and other intracellular pathogens is provided by CD8 cells, a subset of lymphocytes expressing the αβ TCR for Ag and the CD8 molecule. When encountering their Ag in the context of MHC class I molecule in stimulating conditions, resting “naive” CD8 cells (i.e., cells that have never been stimulated before by their Ag) proliferate and differentiate into effector/memory cells. Effector CD8 cells contribute to pathogen clearance by killing Ag-positive target cells and secreting inflammatory cytokines. After Ag clearance, most of the Ag-specific CD8 cells undergo apoptosis. However, the remaining cells that survive generate a long-lived population of memory CD8 cells that can persist for the life of the immunized individual (memory phase) and rapidly proliferates and assumes effector functions upon re-exposure to Ag (secondary response) (1).

The current view is that memory CD8 cells do not have an intrinsic longevity, but are repeatedly stimulated to proliferate and/or survive during the memory phase (1). Indeed, cycling of Ag-specific memory CD8 cells has been reported to persist at low levels in the absence of the original priming Ag (2). Moreover, a time course of gene expression and functional changes in Ag-specific CD8 cells during an antiviral response suggested that CD8 cells from the spleen acquire full memory cell differentiation several weeks after viral clearance (3). Although it is possible that memory CD8 cells are expanded by persisting Ag, cross-reactive Ags, or self-peptides during the memory phase (4, 5, 6, 7, 8, 9), many findings support the notion that the maintenance of memory CD8 cells over time does not require Ag, MHC class I molecules, CD4 cells, B cells, or Abs (2, 10, 11, 12). Several reports have shown that memory CD8 cells slowly proliferate in response to cytokines, primarily IL-15 and IL-7 (1, 13, 14, 15, 16, 17, 18). It was suggested that this basal proliferation is crucial for long-term memory, because cell proliferation compensates the losses due to cell death, thus maintaining stable over time the frequency of Ag-specific memory cells within the CD8 cells (1).

In this article, we ask whether the basal proliferation maintaining cytotoxic T cell memory occurs predominantly in a specific organ. This question has implications for the control of cytokine-driven stimulation of memory CD8 cells, which is potentially harmful to the host due to the risk of immune-mediated pathology and/or autoimmunity (1, 19, 20). Moreover, within enclosed niches in a specialized organ, excessive proliferation of a memory T cell clone may be limited by other memory T cells competing for limited resources and space (1, 19, 21).

The concept of niche has been applied to memory T cell biology for a few years (12, 19, 21, 22). Still, it is not clear yet whether memory CD8 cells preferentially proliferate in specific lymphoid or extralymphoid organs (23, 24, 25, 26, 27). We focus our study on memory CD8 cells localized in the bone marrow (BM).4 We have shown previously that memory CD8 cells from the BM are in a more activated state than those from lymphoid periphery and proposed that in the BM memory CD8 cells could receive survival/proliferation signals, sustaining the long-term maintenance of these cells (28, 29). Indeed, stroma cells and cells of hemopoietic lineage in the BM produce both IL-7 and IL-15, which are able to stimulate CD8 cell survival and proliferation (18). Moreover, the BM can sustain the long-term survival of plasma cells, i.e., Ag-experienced effector cells of the B lineage (30, 31, 32).

The presence of memory CD8 cells in the BM has been documented previously (25, 28, 29, 33, 34, 35), and it has been shown that anti-lymphocytic choriomeningitis virus memory CD8 cells from the BM are able to mount an effective secondary response upon adoptive transfer to immunodeficient hosts (33). Moreover, previous observations in the adult rat had shown that T cells proliferate in the BM more than in other lymphoid organs, thymus excluded (36).

In this study, we investigate whether memory CD8 cells just sit in the BM or divide in this organ.

C57BL6/J (B6) mice were purchased from Harlan Nossan, and OT-I Ly 5.1 and Ly 5.2 mice were a generous gift from Dr. M. Bellone (San Raffaele Hospital, Milan, Italy). OT-I mice are transgenic for the α- and β-chain of a TCR, recognizing the Kb-restricted OVA257–264 peptide (37). Mice were housed at our institute animal facility, according to the institutional guidelines, and used at 2–4 mo of age. Sentinel mice were screened for seropositivity to Sendai virus, rodent coronavirus, and Mycoplasma pulmonis by Murine ImmunoComb test (Charles River Laboratories) and were found negative.

OVA-derived peptide (OVA257–264, SIINFEKL) was purchased from Primm. Purity was >95%, as analyzed by HPLC and mass spectrometry.

Mice were treated with 0.8 mg/ml BrdU (Sigma-Aldrich) in their drinking water for 3 days. BrdU solution was prepared in sterile water, protected from light exposure, and changed daily. Mice were anesthetized with Avertin and perfused intracardially with 50 ml of PBS before taking their organs out. Spleen, inguinal and brachial (peripheral) lymph nodes (PLN), mesenteric lymph nodes (MLN), tibiae, and femurs were collected; in some experiments, blood samples, liver, and lung were also collected. Single-cell suspensions were prepared from spleen, PLN, and MLN by mechanical disruption and passage through cell strainers, and from tibiae and femurs by syringe insertion into one end of the bone and flushing with PBS. After mechanical disruption and passage through cell strainers, cells from liver were separated on Ficoll gradient (500 × g for 20 min), and cells from lung were separated on 40–80% Percoll gradient (400 × g for 25 min). In both cases, the cells at the gradient interface were collected and washed. In the experiments with memory-phenotype CD8 cells, cells were stained with anti-TCR-CyChrome or -PE, anti-CD8β-biotin or anti-CD8α-CyChrome, anti-CD44-PE Ab, followed by streptavidin-allophycocyanin (BD Pharmingen). Nonspecific staining was blocked with anti-FcγR 2.4G2 mAb. In the experiments with Ag-specific CD8 cells, membrane staining was performed with tetramers and anti-CD8α-CyChrome mAb. In both cases, cells were fixed with 30% methanol and 0.4% paraformaldehyde (PFA) in PBS and left overnight at 4°C in the dark. Cells were permeabilized with 0.01% Tween 20, 1% PFA in PBS, incubated with 500 Kunitz U/ml DNase I, and then stained with either anti-BrdU-FITC (BD Biosciences) or control-FITC mAb (23). Cells were analyzed by flow cytometry, gating either on TCR+ cells and acquiring 10–30,000 events per sample (experiments on memory-phenotype CD8 cells) or on CD8+ cells and acquiring 50–100,000 events per sample (experiments on Ag-specific CD8 cells). The events analyzed in the OVA257–264 peptide-Kb tetramer (OVA-tetr)+ gate were 50–400 in the experiments with B6 mice and 600–6,000 in the experiments with OT-I-B6 mice. Samples from BM, spleen, PLN, and MLN were obtained from individual mice; samples from liver, lung, and blood were pooled from two mice.

Cell cycle was analyzed by the procedure by Carbonari et al. (38), with some modifications. Briefly, after blocking of nonspecific staining with 2.4G2 mAb, cells were stained with anti-TCR-PE and anti-CD8β-FITC mAbs (BD Pharmingen) and fixed with 30% methanol and 0.4% PFA in PBS. Then, cells were resuspended in 10% FCS/PBS containing 20 μg/ml 7-aminoactinomycin D (Sigma-Aldrich) and incubated for 1 h at 37°C, plus overnight at 4°C. Cells were then analyzed by flow cytometry, gating on TCR+CD8+ cells and acquiring at least 50,000 events per sample. The double discrimination module was used to exclude multiple events. The percentage of cells in S/G2/M phase was calculated using the CellQuest software (39). TCR+CD8+ cell FL3-area mean coefficient of variations was always <5. BM total nucleated cells were used as positive control.

Cells were purified from PLN and MLN of OT-I mice and injected i.v. into B6 recipient mice (∼4 × 106 cells/mouse, comprising ∼2.5 × 106 CD8 cells expressing the anti-OVA257–264 transgenic TCR). Normal B6 mice and B6 mice, which had been adoptively transferred 24 h before (OT-I-B6), were injected i.p. with 2.5 mg of OVA and 150 μg of poly(I:C) (40). Two weeks later, mice were boosted i.p. with 250 μg of OVA and 150 μg of poly(I:C).

Cells were purified from spleen of control and immunized mice (responder cells) and tested for anti-OVA257–264 peptide cytotoxic response (41). Briefly, responder cells (4 × 106 cells/well) were stimulated in 24-well plates with irradiated syngeneic spleen cells (stimulator cells, 2 × 106 per well), which had been prepulsed with OVA257–264 peptide at 10 μg/ml, in the presence of IL-2 at 20 U/ml. As a positive control, spleen cells from each mouse were stimulated in parallel cultures with irradiated BALB/c spleen cells bearing the H-2d alloantigen (anti-alloresponse). After 5 days, cells were harvested and tested for cytotoxic activity against [3H]TdR-labeled targets in a 4-h assay. The anti-OVA257–264 peptide responder cells were tested against EL4 target cells (H-2b, syngeneic to B6 mice), which had been either prepulsed or not with OVA257–264 peptide; the anti-alloresponder cells were tested against either P815 (H-2d) or EL4 target cells.

Spleen cells were purified from either control or primed mice (responder cells) and tested in a 40-h ELISPOT assay as described previously (28). Briefly, responder cells were incubated in anti-IFN-γ mAb precoated multiscreen plates with IL-2 at 20 U/ml and 500,000 irradiated syngeneic spleen cells, which had been prepulsed with either the OVA257–264 peptide at 10 μg/ml or the medium alone. Responder spleen cells were either 500,000 per well (B6 responder cells) or 50,000 per well (OT-I-B6 responder cells). As a positive control, responder spleen cells from each mouse (100,000 per well) were incubated with either medium alone or Con A at 10 μg/ml. After 40 h of incubation at 37°C, cells were washed, and the plates were sequentially incubated with anti-IFN-γ-biotinylated mAb, poly-HRP-streptavidin (Pierce), and 3-amino-9-ethylcarbazole substrate (Sigma-Aldrich). IFN-γ-transfected TSA cells and their parental untransfected line were used as controls in each ELISPOT plate, after irradiation. The spots were counted using a Zeiss Axioplan microscope and the KS ELISPOT software.

Cells were incubated with 2.4G2 mAb and streptavidin (Sigma-Aldrich) and stained with either PE-labeled SSYSYSSL Kb tetramer (control peptide-Kb tetramer (ctrl-tetr); ProImmune) (42) or PE-labeled SIINFEKL Kb tetramer (OVA-tetr; ProImmune) in ice for 45 min (43). Background staining with ctrl-tetr was subtracted from each sample. Anti-CD8α-CyChrome plus either control-FITC or anti-CD127-FITC (eBioscience) mAbs were added, and cells were incubated for an additional 15 min. After fixation with 30% methanol and 0.4% PFA in PBS, cells were analyzed by flow cytometry, gating on CD8+ cells, and acquiring from 50,000 to 100,000 events per sample. Annexin V staining was performed with the Annexin V-FITC kit (BD Pharmingen), according to the manufacturer’s instructions (44).

Statistical analysis was performed by Student’s t test. Differences were considered significant when p ≤ 0.05 and highly significant when p ≤ 0.01.

To seek information on CD8 cell basal proliferation occurring in the BM compared with that in lymphoid periphery, we treated normal C57BL/6 (B6) mice for 3 days with BrdU, a thymidine analogue that is incorporated in the DNA of cycling cells, and measured the BrdU-labeled CD8 cells from different lymphoid organs. As shown in Fig. 1,A, the BM TCR+CD8+ cells from BrdU-treated mice contained 3- to 5-fold higher percentages of BrdU+ cells than their corresponding cells in spleen, MLNs, or PLNs. The results were similar for CD8α- and CD8β-positive cells, identified by staining with the corresponding mAb (Fig. 1,A and our unpublished data). To eliminate blood-derived T cells, mice were either perfused intracardially with PBS (Fig. 1,A) or exsanguinated before organ removal, and similar results were obtained in the two sets of experiments (data not shown). Lymphoid organs from untreated control B6 mice had a background staining of ≤1.3% BrdU+ cells within TCR+CD8+ cells (Fig. 1,A). To determine whether the BrdU+ CD8 cells were more abundant in the BM from BrdU-treated mice because of local cell division, or whether CD8 cells had incorporated BrdU in another organ and then migrated to the BM, we measured the ongoing cell division by analyzing the percentage of CD8 cells at different stages of the cell cycle in untreated B6 mice (raw data in Fig. 1,B and summary in C). We found that 0.59% of BM TCR+CD8+ cells were in S/G2/M phases (Fig. 1,C), whereas the percentage was half as high in spleen (0.30%) and even lower in PLNs (0.21%), suggesting that the large majority of BrdU+ cells found in BM proliferated locally. We found a great deal of CD8 cell proliferation in the BM compared with other lymphoid organs, whether measured by BrdU incorporation or by cell cycle analysis (Fig. 1, A–C). However, we cannot exclude the possibility that a few CD8 cells divided in response to an Ag in peripheral lymphoid organs, and then migrated to the BM during the 3 days of BrdU treatment (28), or even initiated a primary response in the BM itself (45).

FIGURE 1.

CD8 cell cycling in lymphoid organs. A and D, BrdU incorporation. BrdU-treated B6 female mice were anesthetized and perfused intracardially with PBS. Single-cell suspensions were prepared from spleen, PLN, MLN, and BM. Cells were stained with anti-TCR-CyChrome, anti-CD8β-biotin-streptavidin-allophycocyanin, and anti-CD44-PE mAbs, plus either control-FITC or anti-BrdU-FITC mAb, and analyzed by flow cytometry. A, BrdU+ cell percentages among CD8 cells. The percentage of BrdU+ cells in the TCR+CD8+ cells from each organ was determined, after subtraction of background staining with control Ab (≤1.5%). BM vs spleen, p ≤ 0.01; BM vs PLN, p ≤ 0.01; and BM vs MLN, p ≤ 0.01. B, Ex vivo cell cycle profiles. Cells from spleen, PLN, and BM of untreated B6 mice were stained with anti-TCR-PE and anti-CD8β-FITC mAb, incubated with 7-aminoactinomycin, and analyzed by flow cytometry. The panels represent the typical histograms of TCR+CD8+ cells from spleen, PLN, and BM. The scales on x-axis are linear, in arbitrary units. The numbers represent the percentages of cells in the indicated region, after gating on TCR+CD8+ cells. BM total nucleated cells were used as positive control. C, Ex vivo cell cycle analysis. The panel shows the percentage of cells in S/G2/M phases among TCR+CD8+ cells from each organ. BM vs spleen, p ≤ 0.01; and BM vs PLN, p ≤ 0.01. D, CD44high and CD44int/low cells among BrdU+ CD8 cells. The histograms represent the average percentages of CD44highBrdU+ and CD44int/lowBrdU+ cells among the TCR+CD8+ cells from each organ. Each panel summarizes the results obtained in three independent experiments, except for B, which shows typical results from one mouse. As a negative control, cell samples from each organ were stained in parallel with control mAbs (and streptavidin-allophycocyanin, where needed). The TCR+CD8+ gate used for cytofluorimetric analysis in A–D contained ≤0.15% positive cells in spleen, PLN, and MLN negative control samples and 0.00% in many of the BM-negative control samples (the rest having a value of 0.03–0.04%).

FIGURE 1.

CD8 cell cycling in lymphoid organs. A and D, BrdU incorporation. BrdU-treated B6 female mice were anesthetized and perfused intracardially with PBS. Single-cell suspensions were prepared from spleen, PLN, MLN, and BM. Cells were stained with anti-TCR-CyChrome, anti-CD8β-biotin-streptavidin-allophycocyanin, and anti-CD44-PE mAbs, plus either control-FITC or anti-BrdU-FITC mAb, and analyzed by flow cytometry. A, BrdU+ cell percentages among CD8 cells. The percentage of BrdU+ cells in the TCR+CD8+ cells from each organ was determined, after subtraction of background staining with control Ab (≤1.5%). BM vs spleen, p ≤ 0.01; BM vs PLN, p ≤ 0.01; and BM vs MLN, p ≤ 0.01. B, Ex vivo cell cycle profiles. Cells from spleen, PLN, and BM of untreated B6 mice were stained with anti-TCR-PE and anti-CD8β-FITC mAb, incubated with 7-aminoactinomycin, and analyzed by flow cytometry. The panels represent the typical histograms of TCR+CD8+ cells from spleen, PLN, and BM. The scales on x-axis are linear, in arbitrary units. The numbers represent the percentages of cells in the indicated region, after gating on TCR+CD8+ cells. BM total nucleated cells were used as positive control. C, Ex vivo cell cycle analysis. The panel shows the percentage of cells in S/G2/M phases among TCR+CD8+ cells from each organ. BM vs spleen, p ≤ 0.01; and BM vs PLN, p ≤ 0.01. D, CD44high and CD44int/low cells among BrdU+ CD8 cells. The histograms represent the average percentages of CD44highBrdU+ and CD44int/lowBrdU+ cells among the TCR+CD8+ cells from each organ. Each panel summarizes the results obtained in three independent experiments, except for B, which shows typical results from one mouse. As a negative control, cell samples from each organ were stained in parallel with control mAbs (and streptavidin-allophycocyanin, where needed). The TCR+CD8+ gate used for cytofluorimetric analysis in A–D contained ≤0.15% positive cells in spleen, PLN, and MLN negative control samples and 0.00% in many of the BM-negative control samples (the rest having a value of 0.03–0.04%).

Close modal

We observed that the large majority of BrdU+ CD8 cells from spleen, lymph nodes, and BM were memory-phenotype, i.e., expressing high levels of CD44, a widely used activation/memory marker for mouse T cells (46) (Fig. 1,D). This is in agreement with previous studies on T cells from peripheral lymphoid organs, showing that memory-phenotype T cells incorporate more BrdU+ than naive-phenotype T cells (23). Because the percentage of CD44high cells within the CD8 cells is higher in the BM than in peripheral lymphoid organs (Fig. 2,A and Ref.28), it was possible that this difference would account for the higher proportion of BrdU+ CD8 cells found in the BM compared with other lymphoid organs. When we analyzed BrdU incorporation within the CD44high CD8 cell subset, we found that even among the CD44high CD8 cells, the proportion of proliferating cells was higher in the BM than in the other lymphoid organs examined (Fig. 2 B). This suggests that the BM is a preferential site for memory-phenotype CD8 cell division.

FIGURE 2.

CD44high CD8 cell BrdU incorporation in lymphoid organs. Single-cell suspensions were prepared from spleen, PLN, MLN, and BM of BrdU-treated, PBS-perfused B6 female mice. After staining with anti-TCR-CyChrome, anti-CD8β-biotin-streptavidin-allophycocyanin, and anti-CD44-PE mAbs, plus either control-FITC or anti-BrdU-FITC mAb, cells were analyzed by flow cytometry (as in Fig. 1, A and C). A, CD44high cell percentages among CD8 cells. The panel shows the percentage of CD44high cells among the TCR+CD8+ cells from each organ. BM vs spleen, p ≤ 0.01; BM vs PLN, p ≤ 0.01; and BM vs MLN, p ≤ 0.01. B, BrdU+ cell percentages among CD44high CD8 cells. The panel shows the percentage of BrdU+ cells among the CD44highTCR+CD8+ cells from each organ. BM vs spleen, p ≤ 0.05; BM vs PLN, p ≤ 0.01; and BM vs MLN, p ≤ 0.01. (Key to symbols in A applies also for B.) Each panel summarizes the results obtained in three independent experiments. C, CD44high CD8 cell numbers. The average total number of CD44highTCR+CD8+ cells from spleen, PLN, MLN, and BM was determined (flow cytometry data from five mice examined in two independent experiments; see Table I for estimated total cell number in each organ). The pie chart represents the contribution of each organ to the CD44highTCR+CD8+ cells contained in the sum of spleen, PLN, MLN, and BM (average numbers of cells from each organ are indicated). D, BrdU+CD44high CD8 cell numbers. The average total number of BrdU+CD44highTCR+CD8+ cells from spleen, PLN, MLN, and BM was determined, and results are represented by a pie chart similarly as in C.

FIGURE 2.

CD44high CD8 cell BrdU incorporation in lymphoid organs. Single-cell suspensions were prepared from spleen, PLN, MLN, and BM of BrdU-treated, PBS-perfused B6 female mice. After staining with anti-TCR-CyChrome, anti-CD8β-biotin-streptavidin-allophycocyanin, and anti-CD44-PE mAbs, plus either control-FITC or anti-BrdU-FITC mAb, cells were analyzed by flow cytometry (as in Fig. 1, A and C). A, CD44high cell percentages among CD8 cells. The panel shows the percentage of CD44high cells among the TCR+CD8+ cells from each organ. BM vs spleen, p ≤ 0.01; BM vs PLN, p ≤ 0.01; and BM vs MLN, p ≤ 0.01. B, BrdU+ cell percentages among CD44high CD8 cells. The panel shows the percentage of BrdU+ cells among the CD44highTCR+CD8+ cells from each organ. BM vs spleen, p ≤ 0.05; BM vs PLN, p ≤ 0.01; and BM vs MLN, p ≤ 0.01. (Key to symbols in A applies also for B.) Each panel summarizes the results obtained in three independent experiments. C, CD44high CD8 cell numbers. The average total number of CD44highTCR+CD8+ cells from spleen, PLN, MLN, and BM was determined (flow cytometry data from five mice examined in two independent experiments; see Table I for estimated total cell number in each organ). The pie chart represents the contribution of each organ to the CD44highTCR+CD8+ cells contained in the sum of spleen, PLN, MLN, and BM (average numbers of cells from each organ are indicated). D, BrdU+CD44high CD8 cell numbers. The average total number of BrdU+CD44highTCR+CD8+ cells from spleen, PLN, MLN, and BM was determined, and results are represented by a pie chart similarly as in C.

Close modal

Because CD8 T cells comprise only ∼1–2% of BM, but 10–12% of spleen, and 20–25% of lymph node cells, it was possible that the BM contribution to memory CD8 cell division was negligible in absolute terms. To address this point, we determined the absolute numbers of BrdU+CD44high CD8 cells present in BM and peripheral lymphoid organs after 3 days of BrdU treatment and found that, on the average, the BrdU+CD44high CD8 cells were ∼2.4 × 105, 1.8 × 105, and 2 × 105 in BM, spleen, and total lymph nodes, respectively (Fig. 2,D and Table I). Overall, the BM contained ∼25% of the total CD44high CD8 cells present in the examined organs (Fig. 2 C) and ∼40% of the total BrdU+ cells among the CD44high CD8 cells (D). Taken together, the results indicate that the highest number of proliferating memory-phenotype CD8 cells is in the BM.

Table I.

Counts and estimates of leukocyte cell numbers from lymphoid and extralymphoid organsa

Examined OrganCell Counts (groups of experiments)Cell Counts (average groups a, b, c)Estimated Cell Recovery from the Examined Organ (%)Total OrganEstimated Percentage of the Total Organ contained in the Examined Organ (%)Estimated Total Cell Number
Spleen a) 81.8 (24.3) 85.9 90 Spleen 100 95.4 
 b) 90.5 (25.0)      
 c) 85.3 (19.7)      
Chain of MLN a) 8.9 (1.9) 8.9 75 Total MLN 70 17.0 
Brachial and inguinal LN a) 8.0 (1.9) 7.8 75 Total PLN 20 52.0 
 b) 6.9 (2.1)      
 c) 8.4 (2.2)      
BM from 1 femur and 1 tibia a) 18.2 (6.9) 16.7 55 Total BM 337.4 
 b) 15.5 (4.0)      
 c) 16.3 (4.4)      
Liver b) 3.8 (1.9) 4.3 20 Liver 100 21.5 
 c) 4.9 (2.7)      
Lung b) 0.9 (0.2) 1.0 20 Lung 100 5.0 
 c) 1.0 (0.4)      
Examined OrganCell Counts (groups of experiments)Cell Counts (average groups a, b, c)Estimated Cell Recovery from the Examined Organ (%)Total OrganEstimated Percentage of the Total Organ contained in the Examined Organ (%)Estimated Total Cell Number
Spleen a) 81.8 (24.3) 85.9 90 Spleen 100 95.4 
 b) 90.5 (25.0)      
 c) 85.3 (19.7)      
Chain of MLN a) 8.9 (1.9) 8.9 75 Total MLN 70 17.0 
Brachial and inguinal LN a) 8.0 (1.9) 7.8 75 Total PLN 20 52.0 
 b) 6.9 (2.1)      
 c) 8.4 (2.2)      
BM from 1 femur and 1 tibia a) 18.2 (6.9) 16.7 55 Total BM 337.4 
 b) 15.5 (4.0)      
 c) 16.3 (4.4)      
Liver b) 3.8 (1.9) 4.3 20 Liver 100 21.5 
 c) 4.9 (2.7)      
Lung b) 0.9 (0.2) 1.0 20 Lung 100 5.0 
 c) 1.0 (0.4)      
a

Organs were collected, and single-cell suspensions were prepared from spleen and LN by mechanical disruption and passage through cell strainers, and from tibiae and femurs by syringe insertion into one end of the bone and flushing with PBS. After mechanical disruption and passage through cell strainers, cells from liver and lung were separated on Ficoll and Percoll gradients, respectively, and cells at the gradient interface were collected. Cells were counted by trypan blue exclusion, after lysis of RBCs. The table summarizes the results obtained in three groups of experiments: 1) four experiments with memory-phenotype CD8 cells with a total of 13 B6 mice; 2) eight experiments with Ag-specific CD8 cells with a total of 23 B6 mice; and 3) eight experiments with Ag-specific CD8 cells with a total of 21 OT-I-B6 mice. Average numbers of cells are indicated in millions (SDs in parentheses). To estimate the fraction of total BM contained in one femur and one tibia, we referred to Benner et al. (6869 ) and Hunt (70 ), considering one femur as ∼6% of total BM, and estimated one tibia as 3%. Our estimates of lymphoid organ total cell counts are in agreement with those of Hunt (70 ) and Benner et al. (53 ). LN, Lymphnode.

One caveat of our experiments is that a few BrdU+CD44high CD8 cells might proliferate in response to unknown Ags in the environment during the 3 days of BrdU treatment. Another potential limitation is that the discrimination between naive and memory CD8 cells based on the expression of activation/memory markers is not entirely reliable, because phenotype shifts may occur independently of Ag-specific responses (47, 48). We thus performed additional studies analyzing Ag-specific memory CD8 cells, long times after priming.

We conducted two sets of experiments. In the first set, we transferred a small number of lymph node cells from anti-OVA257–264 peptide TCR transgenic OT-I mice into wild-type B6 mice. The adoptively transferred mice we obtained (in this article called OT-I-B6 mice) have an artificially increased frequency of naive CD8 cells specific for the OVA257–264 peptide, and, after priming, their lymphoid and extralymphoid organs contain high number of memory CD8 cells that can be easily visualized. In the second set, we used wild-type B6 mice. In both sets of experiments, we immunized mice by giving two i.p. injections of the Ag OVA plus the adjuvant poly(I:C), 2 wk apart. By immunizing with an inert Ag rather than with an infectious agent, we avoided the potential effects of pathogen-host interaction in our analysis. Control mice were either untreated or given injections twice with poly(I:C) alone. One or 2 wk after the second injection, we sacrificed some mice from each group and tested them for anti-OVA257–264 peptide-specific responses. In both sets of experiments, spleen cells from immunized mice displayed Ag-specific IFN-γ production and cytotoxicity, with OT-I-B6 mice showing stronger responses than B6 mice, as expected (Fig. 3, A–C). Control mice had low or negative responses in the experiments with OT-I-B6 mice and negative responses in the experiments with B6 mice (Fig. 3, A–C). We also purified cells from spleen, PLNs, and BM of control and immunized mice of both sets of experiments and measured the OVA257–264 peptide-specific CD8 cell frequency by flow cytometry, after staining with anti-CD8α-CyChrome mAb and either OVA-tetr-PE or ctrl-tetr-PE. Fig. 3,D shows the typical data obtained from spleen, PLNs, and BM samples of one untreated and one OVA plus poly(I:C)-treated B6 mouse. Fig. 4, A and C, summarize the tetramer data from a total of three experiments with OT-I-B6 mice and three with B6 mice, all performed at early times after priming, i.e., 1 or 2 wk after the second Ag injection. We then rested mice of both sets of experiments and measured tetramer-binding CD8 cells from immunized and control groups at late times after priming, i.e., 6–10 wk after the second Ag injection (Fig. 4, B and D). At both early and late time points, the percentage of OVA257–264 peptide-specific cells within the CD8 cells from immunized mice was higher in the BM than in either spleen or PLNs (Fig. 4), and OVA-tetr+ CD8 cells were CD44high in each of the three organs (data not shown). Results were similar in both sets of experiments, apart from the expected higher frequencies of OVA-tetr+ within CD8 cells from OT-I-B6 mice compared with B6 mice (Fig. 4). We detected ≤0.58% OVA-tetr+ within CD8 cells from lymphoid organs of control OT-I-B6 mice, whereas the background staining was ≤0.06% in control B6 mice (Fig. 4). By 6–10 wk after immunization, the frequencies of OVA257–264 peptide-specific cells within the CD8 cells had dropped in all organs (Fig. 4, B and D), as expected considering that most of the primed CD8 cells are short-lived effector cells. However, such decrease was less pronounced in the BM than in the spleen. This pattern was observed in both OT-I-B6 mice and B6 mice, supporting the hypothesis that BM may be a preferential organ for memory CD8 cell homing and persistence.

FIGURE 3.

OT-I-B6 and B6 mice Ag-specific CD8 cell response. OT-I-B6 (A and B) and B6 (A, C, and D) mice were immunized with OVA plus poly(I:C) and tested for OVA257–264-specific IFN-γ production, cytotoxicity, and tetramer staining, at early times after priming, i.e., 1–2 wk after the second Ag injection. A, IFN-γ ELISPOT. Spleen cells from immunized and control mice were tested by a 40-h IFN-γ ELISPOT assay. The numbers of Ag-specific spots are shown on the y-axis, after subtraction of medium background; cells from OT-I-B6 mice were 50,000 per well, and cells from B6 mice were 500,000 per well. As a positive control, spleen cells from each mouse (100,000 per well) were tested for response to Con A. B and C, Cytotoxicity test. Spleen cells from immunized and control mice were tested for cytotoxicity, after in vitro restimulation. The percentages of Ag-specific killing are shown on the y-axis, after subtraction of killing of irrelevant targets. Responder to target (R:T) ratios are shown on the x-axis. Anti-allokilling was used as a positive control. Each panel shows the results of individual mice in one typical immunization experiment of three with OT-I-B6 mice and three with B6 mice. D, Tetramer staining. The panels represent the spleen, PLN, and BM cytograms of one untreated and one OVA plus poly(I:C)-treated B6 mouse. The scales on x- and y-axes are logarithmic, in arbitrary units. The numbers represent the percentages of cells in the indicated region, after gating on CD8+ cells. The panels are representative of the results summarized in Fig. 4.

FIGURE 3.

OT-I-B6 and B6 mice Ag-specific CD8 cell response. OT-I-B6 (A and B) and B6 (A, C, and D) mice were immunized with OVA plus poly(I:C) and tested for OVA257–264-specific IFN-γ production, cytotoxicity, and tetramer staining, at early times after priming, i.e., 1–2 wk after the second Ag injection. A, IFN-γ ELISPOT. Spleen cells from immunized and control mice were tested by a 40-h IFN-γ ELISPOT assay. The numbers of Ag-specific spots are shown on the y-axis, after subtraction of medium background; cells from OT-I-B6 mice were 50,000 per well, and cells from B6 mice were 500,000 per well. As a positive control, spleen cells from each mouse (100,000 per well) were tested for response to Con A. B and C, Cytotoxicity test. Spleen cells from immunized and control mice were tested for cytotoxicity, after in vitro restimulation. The percentages of Ag-specific killing are shown on the y-axis, after subtraction of killing of irrelevant targets. Responder to target (R:T) ratios are shown on the x-axis. Anti-allokilling was used as a positive control. Each panel shows the results of individual mice in one typical immunization experiment of three with OT-I-B6 mice and three with B6 mice. D, Tetramer staining. The panels represent the spleen, PLN, and BM cytograms of one untreated and one OVA plus poly(I:C)-treated B6 mouse. The scales on x- and y-axes are logarithmic, in arbitrary units. The numbers represent the percentages of cells in the indicated region, after gating on CD8+ cells. The panels are representative of the results summarized in Fig. 4.

Close modal
FIGURE 4.

OVA-tetr+ percentages among CD8 cells from lymphoid organs of OT-I-B6 and B6 mice at early and late times after priming. Tetramer staining was performed on spleen, PLN, and BM, from OT-I-B6 (A and B)- and B6 (C and D)-immunized mice at either early (1–2 wk, A and C) or late (6–10 wk, B and D) times after priming with OVA plus poly(I:C). The percentages of OVA-tetr+ cells in the CD8+ cells are shown on the y-axis, after subtraction of background staining obtained with ctrl-tetr (≤0.5% in A and C; and ≤0.2% in B and D). A and C, BM vs PLN, p ≤ 0.05. B and D, BM vs spleen, p ≤ 0.01; and BM vs PLN, p ≤ 0.01. The panels summarize the results of three independent immunization experiments with B6 mice and three with OT-I-B6 mice.

FIGURE 4.

OVA-tetr+ percentages among CD8 cells from lymphoid organs of OT-I-B6 and B6 mice at early and late times after priming. Tetramer staining was performed on spleen, PLN, and BM, from OT-I-B6 (A and B)- and B6 (C and D)-immunized mice at either early (1–2 wk, A and C) or late (6–10 wk, B and D) times after priming with OVA plus poly(I:C). The percentages of OVA-tetr+ cells in the CD8+ cells are shown on the y-axis, after subtraction of background staining obtained with ctrl-tetr (≤0.5% in A and C; and ≤0.2% in B and D). A and C, BM vs PLN, p ≤ 0.05. B and D, BM vs spleen, p ≤ 0.01; and BM vs PLN, p ≤ 0.01. The panels summarize the results of three independent immunization experiments with B6 mice and three with OT-I-B6 mice.

Close modal

In the first 2 wk after immunization, we found no significant difference between the percentages of annexin V-binding cells found in freshly isolated BM (18.49 ± 11.89) and spleen (16.55 ± 9.09) cell samples, after gating on OVA-tetr+CD8+ cells. This suggests that Ag-specific CD8 cells from the BM are not moreprotected from apoptosis as compared with those from the spleen (44).

To investigate where the basal proliferation of Ag-specific memory CD8 cells takes place a long time after priming, we analyzed BrdU incorporation by OVA-tetr+CD8+ cells from different organs of immunized OT-I-B6 and B6 mice at 6–10 wk after the second Ag injection. In these experiments, we also examined cells purified from liver and lung, taking into consideration that Ag-specific CD8 cells can persist in extralymphoid organs for long periods after immunization (25, 27, 49, 50). In both OT-I-B6 and B6 mice, we found that the BM contained the highest percentage of BrdU+ cells within OVA-tetr+CD8+ cells among the organs analyzed, i.e., spleen, PLNs, BM, liver, and lung (raw data in Fig. 5 and summary in Fig. 6, A and D). We also analyzed blood samples from immunized OT-I-B6 mice and found that ∼3.7 ± 1.0% of the blood OVA-tetr+CD8+ cells were BrdU+. Because Ag-specific cells are unlikely to proliferate in the blood, we hypothesize that these cells incorporated BrdU in a solid organ during the 3-day treatment and were in the blood stream at the time of our analysis. These results suggest that the BM environment stimulates Ag-specific memory CD8 cell division.

FIGURE 5.

BrdU incorporation analysis of OVA-tetr+ CD8 cells from OT-I-B6 mice. OT-I-B6 mice were given injections twice with OVA plus poly(I:C) and tested for BrdU incorporation 6–10 wk after the second injection. After 3 days of treatment with BrdU, mice were anesthetized and perfused intracardially with PBS. Cells were purified from spleen, PLN, BM, liver, and lung, stained with anti-CD8α-CyChrome and OVA-tetr-PE plus either control-FITC or anti-BrdU-FITC mAb, and analyzed by flow cytometry. A and B, BrdU staining profiles. The panels represent the cytofluorimetric analysis on either spleen, PLN, BM, or liver cell samples (A) or spleen, PLN, BM, or lung cell samples (B) from representative OVA plus poly(I:C)-treated OT-I-B6 mice from two different experiments. The scales on x- and y-axes are logarithmic, in arbitrary units. The numbers represent the percentage of cells in the upper right quadrant, after gating on OVA-tetr+ cells. The panels are representative of the results summarized in Fig. 6.

FIGURE 5.

BrdU incorporation analysis of OVA-tetr+ CD8 cells from OT-I-B6 mice. OT-I-B6 mice were given injections twice with OVA plus poly(I:C) and tested for BrdU incorporation 6–10 wk after the second injection. After 3 days of treatment with BrdU, mice were anesthetized and perfused intracardially with PBS. Cells were purified from spleen, PLN, BM, liver, and lung, stained with anti-CD8α-CyChrome and OVA-tetr-PE plus either control-FITC or anti-BrdU-FITC mAb, and analyzed by flow cytometry. A and B, BrdU staining profiles. The panels represent the cytofluorimetric analysis on either spleen, PLN, BM, or liver cell samples (A) or spleen, PLN, BM, or lung cell samples (B) from representative OVA plus poly(I:C)-treated OT-I-B6 mice from two different experiments. The scales on x- and y-axes are logarithmic, in arbitrary units. The numbers represent the percentage of cells in the upper right quadrant, after gating on OVA-tetr+ cells. The panels are representative of the results summarized in Fig. 6.

Close modal
FIGURE 6.

Ag-specific CD8 cell BrdU incorporation in lymphoid and extralymphoid organs of OT-I-B6 and B6 mice. At late times (6–10 wk) after priming with OVA plus poly(I:C), OT-I-B6 (A–C) and B6 (D–F) mice were tested for BrdU incorporation. After 3 days of treatment with BrdU, mice were anesthetized and perfused intracardially with PBS. Cells were purified from spleen, PLN, BM, liver, and lung, stained with anti-CD8α-CyChrome and OVA-tetr-PE plus either control-FITC or anti-BrdU-FITC mAb, and analyzed by flow cytometry. A and D, BrdU+ cell percentages among OVA-tetr+ CD8 cells. The percentage of BrdU+ cells in the OVA-tetr+CD8+ cells from each organ was determined, after subtraction of background staining with control Ab (≤1.5%). (Key to symbols in A applies also for D.) Each panel summarizes the results obtained in five independent experiments of BrdU treatment, performed on three groups of immunized mice. A, BM vs spleen, p ≤ 0.01; BM vs PLN, p ≤ 0.01; and BM vs lung, p ≤ 0.01. D, BM vs spleen, p ≤ 0.05; and BM vs liver, p ≤ 0.05. B and E, OVA-tetr+ CD8 cell numbers. The average total number of OVA-tetr+CD8+ cells from spleen, lymph node (LN), BM, liver, and lung was determined (flow cytometry data are from eight to nine mice examined in five independent experiments in either B or E; the flow cytometry data from PLN are used to determine the cell numbers in total LN, considered as sum of PLN and MLN; see Table I for estimated total cell number in each organ). The pie chart represents the contribution of each organ to the OVA-tetr+CD8+ cells contained in the sum of spleen, total LN, BM, liver, and lung (average numbers of cells from each organ are indicated). C and F, BrdU+OVA-tetr+ CD8 cell numbers. The average total number of BrdU+OVA-tetr+CD8+ cells from spleen, LN, BM, liver, and lung was determined, and results are represented by a pie chart similarly as in B and E.

FIGURE 6.

Ag-specific CD8 cell BrdU incorporation in lymphoid and extralymphoid organs of OT-I-B6 and B6 mice. At late times (6–10 wk) after priming with OVA plus poly(I:C), OT-I-B6 (A–C) and B6 (D–F) mice were tested for BrdU incorporation. After 3 days of treatment with BrdU, mice were anesthetized and perfused intracardially with PBS. Cells were purified from spleen, PLN, BM, liver, and lung, stained with anti-CD8α-CyChrome and OVA-tetr-PE plus either control-FITC or anti-BrdU-FITC mAb, and analyzed by flow cytometry. A and D, BrdU+ cell percentages among OVA-tetr+ CD8 cells. The percentage of BrdU+ cells in the OVA-tetr+CD8+ cells from each organ was determined, after subtraction of background staining with control Ab (≤1.5%). (Key to symbols in A applies also for D.) Each panel summarizes the results obtained in five independent experiments of BrdU treatment, performed on three groups of immunized mice. A, BM vs spleen, p ≤ 0.01; BM vs PLN, p ≤ 0.01; and BM vs lung, p ≤ 0.01. D, BM vs spleen, p ≤ 0.05; and BM vs liver, p ≤ 0.05. B and E, OVA-tetr+ CD8 cell numbers. The average total number of OVA-tetr+CD8+ cells from spleen, lymph node (LN), BM, liver, and lung was determined (flow cytometry data are from eight to nine mice examined in five independent experiments in either B or E; the flow cytometry data from PLN are used to determine the cell numbers in total LN, considered as sum of PLN and MLN; see Table I for estimated total cell number in each organ). The pie chart represents the contribution of each organ to the OVA-tetr+CD8+ cells contained in the sum of spleen, total LN, BM, liver, and lung (average numbers of cells from each organ are indicated). C and F, BrdU+OVA-tetr+ CD8 cell numbers. The average total number of BrdU+OVA-tetr+CD8+ cells from spleen, LN, BM, liver, and lung was determined, and results are represented by a pie chart similarly as in B and E.

Close modal

When we calculated the absolute numbers of BrdU+OVA-tetr+CD8+ cells in spleen, PLNs, BM, liver, and lung, we found that the BM contained by far the highest number among the organs examined, accounting for 60–75% of the total number of BrdU+OVA-tetr+CD8+ cells (Fig. 6, C and F, and Table I). The sum of spleen and total lymph nodes accounted for ∼10–30% of the total number of BrdU+OVA-tetr+CD8+ cells, whereas liver and lung together accounted for <15% (Fig. 6, C and F), despite the fact that the liver contained substantial numbers of OVA-tetr+CD8+ cells (Fig. 6, B and E). The results were similar in both OT-I-B6 and B6 mice, although, as expected, the latter group had lower absolute numbers of BrdU+OVA-tetr+CD8+ cells (Fig. 6, C and F). Taken together, our results suggest that, late after Ag encounter, Ag-specific memory CD8 cells predominantly divide in the BM, with rather few of them proliferating in either peripheral lymphoid or extralymphoid organs. After division, some daughter memory CD8 cells recirculate and can be detected in the blood stream.

It has been proposed recently that CD127, the α-chain of IL-7R, is a unique marker for memory CD8 cells, which allows to distinguish between short-lived effector and long-lived memory cells (51, 52). We examined the surface expression of CD127 on Ag-specific CD8 cells from spleen, lymph nodes, and BM at different times after immunization of OT-I-B6 mice with OVA plus poly(I:C). To carefully follow the kinetics of CD127 expression, we analyzed the immunized mice at three different times, i.e., early (1–2 wk), intermediate (4–5 wk), and late (7–8 wk), after the second Ag injection. We found that Ag-specific CD8 cells from either spleen or lymph nodes down-regulated CD127 at early times after immunization, and then up-regulated it at later times, reaching expression levels higher than those of virgin OVA-tetr+CD8+ cells from the corresponding organs of control unimmunized OT-I-B6 mice (raw data in Fig. 7,B, and summary in A). This is in line with previous observations in mice infected with pathogens (51, 52), although in OVA-immunized mice we observe a less pronounced CD127 down-regulation than that reported in infected mice. Interestingly, at early times after immunization, the BM Ag-specific CD8 cells displayed significantly lower levels of CD127 than their corresponding cells in spleen and lymph nodes (raw data in Fig. 7 B, and summary in A). By 4–5 wk after immunization, both the mean fluorescence intensity (MFI) of the BM CD127+OVA-tetr+CD8+ cells and their percentage within OVA-tetr+CD8+ cells were increased. CD127 expression was then stably maintained over late times. At each of the three time points after immunization, the BM memory CD8 cells expressed lower levels of CD127 than either memory CD8 cells from peripheral lymphoid organs or BM virgin CD8 cells. These results confirm that CD127 expression is tightly regulated after immunization and suggest that local signals in the BM lead to a reduced CD127 expression by Ag-specific memory CD8 cells.

FIGURE 7.

CD127 surface expression by Ag-specific CD8 cells from lymphoid organs of OT-I-B6 mice. At early (1–2 wk), intermediate (interm.; 4–5 wk), and late (7–8 wk) times after priming with OVA plus poly(I:C), OT-I-B6 mice were analyzed for CD127 expression. Cells were purified from spleen, PLN, and BM, stained with anti-CD8α-CyChrome and OVA-tetr-PE plus either control-FITC or anti-CD127-FITC mAb, and analyzed by flow cytometry. A, CD127 expression by OVA-tetr+ CD8 cells. The percentage of CD127+ cells in the OVA-tetr+CD8+ cells (top panels) and the MFI of the CD127+OVA-tetr+CD8+ cells (bottom panels) from each organ were determined at the indicated times after immunization. Untreated OT-I-B6 mice were analyzed as controls. The results of four independent immunization experiments with OT-I-B6 mice are summarized. Statistical analysis was performed by Student’s t test. Top panels, Virgin spleen vs early spleen, p ≤ 0.05; early spleen vs interm. spleen, p ≤ 0.05; early spleen vs early BM, p ≤ 0.05; early PLN vs early BM, p ≤ 0.01; and interm. spleen vs interm. BM, p ≤ 0.05. Bottom panels, Virgin spleen vs interm. spleen, p ≤ 0.01; early spleen vs interm. spleen, p ≤ 0.01; and early spleen vs early BM, p ≤ 0.05. B, CD127 staining. The panels represent the cytofluorimetric analysis on spleen and BM samples from OVA plus poly(I:C)-treated OT-I-B6 mice, at different times after priming. The scales on x- and y-axes are logarithmic, in arbitrary units. The numbers represent the percentages of cells in the upper right quadrant, after gating on OVA-tetr+CD8+ cells. MFI of cells in the upper right quadrant are indicated in parentheses. The panels are representative of the results summarized in A.

FIGURE 7.

CD127 surface expression by Ag-specific CD8 cells from lymphoid organs of OT-I-B6 mice. At early (1–2 wk), intermediate (interm.; 4–5 wk), and late (7–8 wk) times after priming with OVA plus poly(I:C), OT-I-B6 mice were analyzed for CD127 expression. Cells were purified from spleen, PLN, and BM, stained with anti-CD8α-CyChrome and OVA-tetr-PE plus either control-FITC or anti-CD127-FITC mAb, and analyzed by flow cytometry. A, CD127 expression by OVA-tetr+ CD8 cells. The percentage of CD127+ cells in the OVA-tetr+CD8+ cells (top panels) and the MFI of the CD127+OVA-tetr+CD8+ cells (bottom panels) from each organ were determined at the indicated times after immunization. Untreated OT-I-B6 mice were analyzed as controls. The results of four independent immunization experiments with OT-I-B6 mice are summarized. Statistical analysis was performed by Student’s t test. Top panels, Virgin spleen vs early spleen, p ≤ 0.05; early spleen vs interm. spleen, p ≤ 0.05; early spleen vs early BM, p ≤ 0.05; early PLN vs early BM, p ≤ 0.01; and interm. spleen vs interm. BM, p ≤ 0.05. Bottom panels, Virgin spleen vs interm. spleen, p ≤ 0.01; early spleen vs interm. spleen, p ≤ 0.01; and early spleen vs early BM, p ≤ 0.05. B, CD127 staining. The panels represent the cytofluorimetric analysis on spleen and BM samples from OVA plus poly(I:C)-treated OT-I-B6 mice, at different times after priming. The scales on x- and y-axes are logarithmic, in arbitrary units. The numbers represent the percentages of cells in the upper right quadrant, after gating on OVA-tetr+CD8+ cells. MFI of cells in the upper right quadrant are indicated in parentheses. The panels are representative of the results summarized in A.

Close modal

Although it has long been known that the migratory capacity of memory T cells is different from that of naive T cells (25, 29, 53, 54), a link had not been established previously between migration to specific anatomical compartments and memory CD8 cell proliferation. Our results show that up to five times more of the memory CD8 cells in the BM proliferate than do their counterparts in spleen or lymph nodes, or in liver and lung. While our manuscript was under revision, another group reported that homeostatic proliferation of CD44high CD8 cells and TCR transgenic memory CD8 cells occurs predominantly in the BM of lymphocytic choriomeningitis virus-immune mice (55). Thus, similar results have been found independently in two different systems, supporting the view that the BM is a crucial organ for CD8 cell proliferation maintaining cytotoxic memory.

One possibility is that memory CD8 cells proliferate in the BM because they gain access to proliferation-inducing molecules, such as IL-7 and IL-15, present in this organ (18). Interestingly, the BM environment is stimulatory also for CD44int/low CD8 cells (28), and, after 3 days of BrdU treatment, a few BrdU+ cells are found in the CD44int/low CD8 cells from the BM, whereas only a tiny proportion of the corresponding cells from peripheral lymphoid organs is BrdU+. An alternative possibility is that the memory CD8 cells that enter the BM are an organ-specific subset with an intrinsic higher proliferative potential (24). We favor the first possibility and suggest that BM memory CD8 cells are part of a recirculating pool, considering that in situ-labeled T cells can traffic from the BM to other lymphoid organs (56) and that, following parabiosis, memory CD8 cells rapidly equilibrate into spleen, lymph nodes, BM, lung, and liver (57). As regards recirculation, memory CD8 cells would differ from plasma cells, which take residence in the BM, whereas their secreted Abs circulate in blood and provide systemic protection (58).

Experiments with blocking Abs and genetically engineered mice have shown previously that the maintenance of memory CD8 cells is impaired severely in the absence of IL-15 and IL-7 (1, 13, 14, 15, 16, 17, 18, 59). IL-15 is also implicated in memory CD8 cell response to poly(I:C) injection, a treatment that mimics viral infections and induces substantial bystander proliferation of memory CD8 cells, peaking at day 2 (60). Poly(I:C)-induced proliferation of memory CD8 cells occurs mostly in the BM and is compromised severely in IL-15−/− mice (55).

As regards constitutive expression, IL-15 has been found in virtually all tissues tested (61), but its function may involve complex cellular interactions, as recently shown in vitro (62). The IL-15R consists of three chains: 1) the α-chain, which is specific for IL-15; 2) the β-chain, which is shared between IL-2R and IL-15R; and 3) the γ-chain, also called common cytokine γ-chain, which is shared by receptors specific for IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21. The β- and γ-chains of IL-15R form an heterodimeric complex, which is sufficient for intracellular signaling upon binding of IL-15. The α-chain of IL-15R can bind IL-15 with high affinity and present IL-15 in trans to neighboring cells, so that the IL-15Rα-expressing cell concentrates IL-15 on its surface and efficiently stimulates the IL-15-responding cell at exceedingly low concentrations of IL-15 (62). Considering such complexity, it is difficult to determine the concentration of biological active IL-15 in different organs. IL-15Rα is expressed by a variety of tissues and cell types, including cells of hemopoietic origin and BM stromal cells (61). In line with the above model, recent reports have shown that the in vivo memory CD8 cell basal proliferation in response to IL-15 does not require the expression of IL-15Rα by the responding CD8 cells, but depends on the presence of BM-derived IL-15Rα-positive cells (40, 63, 64).

In contrast with IL-15, IL-7 has a restricted pattern of in vivo expression and is expressed mostly by BM stromal cells, in addition to thymic and intestinal epithelial cells (15). The IL-7R consists of two chains: 1) the IL-7Rα chain, also called CD127; and 2) the common cytokine γ-chain. According to recent reports, among the Ag-activated CD8 cells expanded at early times after priming, those few that express high levels of CD127 are the committed precursors of long-lived memory CD8 cells (51, 52). In contrast with CD127low short-lived effector cells, CD127high memory cells would selectively persist over time, suggesting that CD127 may be a differentiation marker (51, 52). In agreement with these studies, we found that during the first 2 wk after immunization, the bulk of the Ag-specific CD8 cells in the lymphoid periphery moderately down-regulate CD127, whereas virtually all of the Ag-specific memory CD8 cells persisting at late times after priming are CD127high. CD127 expression by BM Ag-specific memory CD8 cells follows a similar pattern, but BM memory CD8 cells constantly display on their surfaces lower levels of CD127 than their corresponding cells in peripheral lymphoid organs. In contrast with their splenic and lymph node counterparts, at 4–8 wk after immunization, BM memory CD8 cells do not show increased expression of CD127 in comparison with virgin CD8 cells from the same organ. By 4–5 wk after priming, ∼20% of BM memory CD8 cells stably lack CD127 expression. Our results suggest that CD127 surface expression not only depends on the differentiation stage of CD8 cells, but also is regulated by the organ environment. Suppression of CD127 expression has been documented in response to its ligand, IL-7, or other T cell prosurvival cytokines, such as IL-2, IL-4, IL-6, and IL-15 (65). Thus, our findings that BM Ag-specific memory CD8 cells have a decreased expression of CD127 may be an indirect evidence that these cells constantly receive proliferative signals by IL-7 and IL-15 in the BM, and are possibly stimulated also by other cytokines in this organ.

We have shown in this article that the BM accounts for more than half of the total number of proliferating Ag-specific memory CD8 cells present in spleen, lymph nodes, BM, liver, and lung, taken together. Our findings show that, in addition to its involvement in long-term survival of plasma cells (30, 31), the BM plays a dominant role in the proliferation of memory CD8 cells. Thus, both Ab and cytotoxic memory responses are maintained mostly in BM.

We envision the BM microenvironment for memory CD8 T cells as a saturable compartment. We have shown previously that the migration of CD44high T cells to the BM is inhibited in the presence of huge numbers of “rival” CD44high T cells in the lymphoid periphery, suggesting that BM entry is a competitive process (29). Indeed, the “attrition” of the immune response, which has been observed when different virus infections follow each other in the same individual (66), might be due to a displacement of the memory CD8 T cells localized in the BM by new incoming CD8 T cells with different specificities. Thus, competition for lodging in limited niches in the BM might contribute to the maintenance over time of a diverse memory CD8 T cell repertoire.

In conclusion, we propose here that the BM has a previously unrecognized role in the biology of long-term CD8 cell responses, with important implications for the design of successful vaccines against viruses, intracellular parasites, and tumors. In the light of our findings, the presence of antitumor CTLs in the BM of untreated breast cancer patients, which has been associated with the local control of micrometastasis growth (67), might be also due to the preferential maintenance of memory CD8 cells in the BM.

We thank M. Bellone for OT-I mice, M. Carbonari and T. Tedesco for help with cell cycle experiments, S. Zappacosta’s group for help with flow cytometry, M. D’Esposito, U. D’Oro, M. Epstein, M. Iaccarino, and G. Matarese for reading the manuscript, and the Immunology group at the Institute of Genetics and Biophysics, Adriano Buzzati Traverso-Institute of Protein Biochemistry (particularly G. Del Pozzo and P. G. De Berardinis) for discussion. A special thanks to P. Matzinger for her generous intellectual support and for reading the manuscript.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by grants from European Union (Project No. QLK2-CT-2002-0062, EPI-PEP-VAC), Italian Government (Fondi Italiani per la Ricerca di Base 2001, Project No. RBNE01RB9B), and the Italian Association for Cancer Research.

4

Abbreviations used in this paper: BM, bone marrow; PLN, peripheral lymph node; MLN, mesenteric lymph node; PFA, paraformaldehyde; OVA-tetr, OVA257–264 peptide-Kb tetramer; ctrl-tetr, control peptide-Kb tetramer; MFI, mean fluorescence intensity.

1
Sprent, J., C. D. Surh.
2002
. T cell memory.
Annu. Rev. Immunol.
20
:
551
-579.
2
Murali-Krishna, K., L. L. Lau, S. Sambhara, F. Lemonnier, J. Altman, R. Ahmed.
1999
. Persistence of memory CD8 T cells in MHC class I-deficient mice.
Science
286
:
1377
-1381.
3
Kaech, S. M., S. Hemby, E. Kersh, R. Ahmed.
2002
. Molecular and functional profiling of memory CD8 T cell differentiation.
Cell
111
:
837
-851.
4
Beverley, P. C..
1990
. Is T-cell memory maintained by crossreactive stimulation?.
Immunol. Today
11
:
203
-205.
5
Gray, D., P. Matzinger.
1991
. T cell memory is short-lived in the absence of antigen.
J. Exp. Med.
174
:
969
-974.
6
Matzinger, P..
1994
. Immunology: memories are made of this?.
Nature
369
:
605
-606.
7
Selin, L. K., S. R. Nahill, R. M. Welsh.
1994
. Cross-reactivities in memory cytotoxic T lymphocyte recognition of heterologous viruses.
J. Exp. Med.
179
:
1933
-1943.
8
Kundig, T. M., M. F. Bachmann, S. Oehen, U. W. Hoffmann, J. J. Simard, C. P. Kalberer, H. Pircher, P. S. Ohashi, H. Hengartner, R. M. Zinkernagel.
1996
. On the role of antigen in maintaining cytotoxic T-cell memory.
Proc. Natl. Acad. Sci. USA
93
:
9716
-9723.
9
Tanchot, C., F. A. Lemonnier, B. Perarnau, A. A. Freitas, B. Rocha.
1997
. Differential requirements for survival and proliferation of CD8 naive or memory T cells.
Science
276
:
2057
-2062.
10
Lau, L. L., B. D. Jamieson, T. Somasundaram, R. Ahmed.
1994
. Cytotoxic T-cell memory without antigen.
Nature
369
:
648
-652.
11
Hou, S., L. Hyland, K. W. Ryan, A. Portner, P. C. Doherty.
1994
. Virus-specific CD8+ T-cell memory determined by clonal burst size.
Nature
369
:
652
-654.
12
Di Rosa, F., P. Matzinger.
1996
. Long-lasting CD8 T cell memory in the absence of CD4 T cells or B cells.
J. Exp. Med.
183
:
2153
-2163.
13
Kennedy, M. K., M. Glaccum, S. N. Brown, E. A. Butz, J. L. Viney, M. Embers, N. Matsuki, K. Charrier, L. Sedger, C. R. Willis, et al
2000
. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice.
J. Exp. Med.
191
:
771
-780.
14
Ku, C. C., M. Murakami, A. Sakamoto, J. Kappler, P. Marrack.
2000
. Control of homeostasis of CD8+ memory T cells by opposing cytokines.
Science
288
:
675
-678.
15
Schluns, K. S., W. C. Kieper, S. C. Jameson, L. Lefrancois.
2000
. Interleukin-7 mediates the homeostasis of naive and memory CD8 T cells in vivo.
Nat. Immunol.
1
:
426
-432.
16
Becker, T. C., E. J. Wherry, D. Boone, K. Murali-Krishna, R. Antia, A. Ma, R. Ahmed.
2002
. Interleukin 15 is required for proliferative renewal of virus-specific memory CD8 T cells.
J. Exp. Med.
195
:
1541
-1548.
17
Schluns, K. S., K. Williams, A. Ma, X. X. Zheng, L. Lefrancois.
2002
. Cutting edge: requirement for IL-15 in the generation of primary and memory antigen-specific CD8 T cells.
J. Immunol.
168
:
4827
-4831.
18
Schluns, K. S., L. Lefrancois.
2003
. Cytokine control of memory T-cell development and survival.
Nat. Rev. Immunol.
3
:
269
-279.
19
Freitas, A. A., B. Rocha.
2000
. Population biology of lymphocytes: the flight for survival.
Annu. Rev. Immunol.
18
:
83
-111.
20
Berg, R. E., E. Crossley, S. Murray, J. Forman.
2003
. Memory CD8+ T cells provide innate immune protection against Listeria monocytogenes in the absence of cognate antigen.
J. Exp. Med.
198
:
1583
-1593.
21
Butcher, E. C., L. J. Picker.
1996
. Lymphocyte homing and homeostasis.
Science
272
:
60
-66.
22
Stockinger, B., T. Barthlott, G. Kassiotis.
2004
. The concept of space and competition in immune regulation.
Immunology
111
:
241
-247.
23
Tough, D. F., J. Sprent.
1994
. Turnover of naive- and memory-phenotype T cells.
J. Exp. Med.
179
:
1127
-1135.
24
Sallusto, F., D. Lenig, R. Forster, M. Lipp, A. Lanzavecchia.
1999
. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions.
Nature
401
:
708
-712.
25
Masopust, D., V. Vezys, A. L. Marzo, L. Lefrancois.
2001
. Preferential localization of effector memory cells in nonlymphoid tissue.
Science
291
:
2413
-2417.
26
Goldrath, A. W., P. V. Sivakumar, M. Glaccum, M. K. Kennedy, M. J. Bevan, C. Benoist, D. Mathis, E. A. Butz.
2002
. Cytokine requirements for acute and basal homeostatic proliferation of naive and memory CD8+ T cells.
J. Exp. Med.
195
:
1515
-1522.
27
Hogan, R. J., L. S. Cauley, K. H. Ely, T. Cookenham, A. D. Roberts, J. W. Brennan, S. Monard, D. L. Woodland.
2002
. Long-term maintenance of virus-specific effector memory CD8+ T cells in the lung airways depends on proliferation.
J. Immunol.
169
:
4976
-4981.
28
Di Rosa, F., A. Santoni.
2002
. Bone marrow CD8 T cells are in a different activation state than those in lymphoid periphery.
Eur. J. Immunol.
32
:
1873
-1880.
29
Di Rosa, F., A. Santoni.
2003
. Memory T-cell competition for bone marrow seeding.
Immunology
108
:
296
-304.
30
Manz, R. A., A. Thiel, A. Radbruch.
1997
. Lifetime of plasma cells in the bone marrow.
Nature
388
:
133
-134.
31
Slifka, M. K., R. Antia, J. K. Whitmire, R. Ahmed.
1998
. Humoral immunity due to long-lived plasma cells.
Immunity
8
:
363
-372.
32
Cassese, G., S. Arce, A. E. Hauser, K. Lehnert, B. Moewes, M. Mostarac, G. Muehlinghaus, M. Szyska, A. Radbruch, R. A. Manz.
2003
. Plasma cell survival is mediated by synergistic effects of cytokines and adhesion-dependent signals.
J. Immunol.
171
:
1684
-1690.
33
Slifka, M. K., J. K. Whitmire, R. Ahmed.
1997
. Bone marrow contains virus-specific cytotoxic T lymphocytes.
Blood
90
:
2103
-2108.
34
Marshall, D. R., S. J. Turner, G. T. Belz, S. Wingo, S. Andreansky, M. Y. Sangster, J. M. Riberdy, T. Liu, M. Tan, P. C. Doherty.
2001
. Measuring the diaspora for virus-specific CD8+ T cells.
Proc. Natl. Acad. Sci. USA
98
:
6313
-6318.
35
Wherry, E. J., J. N. Blattman, K. Murali-Krishna, R. van der Most, R. Ahmed.
2003
. Viral persistence alters CD8 T-cell immunodominance and tissue distribution and results in distinct stages of functional impairment.
J. Virol.
77
:
4911
-4927.
36
Westermann, J., S. Ronneberg, F. J. Fritz, R. Pabst.
1989
. Proliferation of lymphocyte subset in the adult rat: a comparison of different lymphoid organs.
J. Immunol.
19
:
1087
-1093.
37
Kurts, C., W. R. Heath, F. R. Carbone, J. Allison, J. F. Miller, H. Kosaka.
1996
. Constitutive class I-restricted exogenous presentation of self antigens in vivo.
J. Exp. Med.
184
:
923
-930.
38
Carbonari, M., T. Tedesco, M. Fiorilli.
2001
. A unified procedure for conservative (morphology) and integral (DNA and immunophenotype) cell staining for flow cytometry.
Cytometry
44
:
120
-125.
39
Hedley, D. W., T. V. Shankey, L. L. Wheeless.
1993
. DNA cytometry consensus conference.
Cytometry
14
:
471
-477.
40
Burkett, P. R., R. Koka, M. Chien, S. Chai, F. Chan, A. Ma, D. L. Boone.
2003
. IL-15Rα expression on CD8+ T cells is dispensable for T cell memory.
Proc. Natl. Acad. Sci. USA
100
:
4724
-4729.
41
Matzinger, P..
1991
. The JAM test: a simple assay for DNA fragmentation and cell death.
J. Immunol. Methods
145
:
185
-192.
42
Schott, E., H. L. Ploegh.
2002
. Mouse MHC class I tetramers that are unable to bind to CD8 reveal the need for CD8 engagement in order to activate naive CD8 T cells.
Eur. J. Immunol.
32
:
3425
-3434.
43
Busch, D. H., I. M. Pilip, S. Vijh, E. G. Pamer.
1998
. Coordinate regulation of complex T cell populations responding to bacterial infection.
Immunity
8
:
353
-362.
44
Wang, X. Z., S. E. Stepp, M. A. Brehm, H. D. Chen, L. K. Selin, R. M. Welsh.
2003
. Virus-specific CD8 T cells in peripheral tissues are more resistant to apoptosis than those in lymphoid organs.
Immunity
18
:
631
-642.
45
Feuerer, M., P. Beckhove, N. Garbi, Y. Mahnke, A. Limmer, M. Hommel, G. J. Hammerling, B. Kyewski, A. Hamann, V. Umansky, V. Schirrmacher.
2003
. Bone marrow as a priming site for T-cell responses to blood-borne antigen.
Nat. Med.
9
:
1151
-1157.
46
Budd, R. C., J. C. Cerottini, C. Horvath, C. Bron, T. Pedrazzini, R. C. Howe, H. R. MacDonald.
1987
. Distinction of virgin and memory T lymphocytes: stable acquisition of the Pgp-1 glycoprotein concomitant with antigenic stimulation.
J. Immunol.
138
:
3120
-3129.
47
Bell, E. B., S. M. Sparshott.
1990
. Interconversion of CD45R subsets of CD4 T cells in vivo.
Nature
348
:
163
-166.
48
Murali-Krishna, K., R. Ahmed.
2000
. Cutting edge: naive T cells masquerading as memory cells.
J. Immunol.
165
:
1733
-1737.
49
Hogan, R. J., E. J. Usherwood, W. Zhong, A. D. Roberts, R. W. Dutton, A. G. Harmsen, D. L. Woodland.
2001
. Activated antigen-specific CD8+ T cells persist in the lungs following recovery from respiratory virus infections.
J. Immunol.
166
:
1813
-1822.
50
Masopust, D., V. Vezys, E. J. Usherwood, L. S. Cauley, S. Olson, A. L. Marzo, R. L. Ward, D. L. Woodland, L. Lefrancois.
2004
. Activated primary and memory CD8 T cells migrate to nonlymphoid tissues regardless of site of activation or tissue of origin.
J. Immunol.
172
:
4875
-4882.
51
Kaech, S. M., J. T. Tan, E. J. Wherry, B. T. Konieczny, C. D. Surh, R. Ahmed.
2003
. Selective expression of the interleukin 7 receptor identifies effector CD8 T cells that give rise to long-lived memory cells.
Nat. Immunol.
4
:
1191
-1198.
52
Huster, K. M., V. Busch, M. Schiemann, K. Linkemann, K. M. Kerksiek, H. Wagner, D. H. Busch.
2004
. Selective expression of IL-7 receptor on memory T cells identifies early CD40L-dependent generation of distinct CD8+ memory T cell subsets.
Proc. Natl. Acad. Sci. USA
101
:
5610
-5615.
53
Benner, R., F. Meima, G. M. van der Meulen.
1974
. Antibody formation in mouse bone marrow. II. Evidence for a memory-dependent phenomenon.
Cell. Immunol.
13
:
95
-106.
54
Mackay, C. R., W. L. Marston, L. Dudler.
1990
. Naive and memory T cells show distinct pathways of lymphocyte recirculation.
J. Exp. Med.
171
:
801
-817.
55
Becker, T. C., S. M. Coley, E. J. Wherry, R. Ahmed.
2005
. Bone marrow is a preferred site for homeostatic proliferation of memory CD8 T cells.
J. Immunol.
174
:
1269
-1273.
56
Pabst, R., M. Miyasaka, L. Dudler.
1986
. Numbers and phenotype of lymphocytes emigrating from sheep bone marrow after in situ labelling with fluorescein isothiocyanate.
Immunology
59
:
217
-222.
57
Klonowski, K. D., K. J. Williams, A. L. Marzo, D. A. Blair, E. G. Lingenheld, L. Lefrancois.
2004
. Dynamics of blood-borne CD8 memory T cell migration in vivo.
Immunity
20
:
551
-562.
58
Hauser, A. E., G. F. Debes, S. Arce, G. Cassese, A. Hamann, A. Radbruch, R. A. Manz.
2002
. Chemotactic responsiveness toward ligands for CXCR3 and CXCR4 is regulated on plasma blasts during the time course of a memory immune response.
J. Immunol.
169
:
1277
-1282.
59
Lodolce, J. P., D. L. Boone, S. Chai, R. E. Swain, T. Dassopoulos, S. Trettin, A. Ma.
1998
. IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation.
Immunity
9
:
669
-676.
60
Tough, D. F., P. Borrow, J. Sprent.
1996
. Induction of bystander T cell proliferation by viruses and type I interferon in vivo.
Science
272
:
1947
-1950.
61
Kennedy, M. K., L. S. Park.
1996
. Characterization of interleukin-15 (IL-15) and the IL-15 receptor complex.
J Clin. Immunol.
16
:
134
-143.
62
Dubois, S., J. Mariner, T. A. Waldmann, Y. Tagaya.
2002
. IL-15Rα recycles and presents IL-15 in trans to neighboring cells.
Immunity
17
:
537
-547.
63
Schluns, K. S., E. C. Nowak, A. Cabrera-Hernandez, L. Puddington, L. Lefrancois, H. L. Aguila.
2004
. Distinct cell types control lymphoid subset development by means of IL-15 and IL-15 receptor α expression.
Proc. Natl. Acad. Sci. USA
101
:
5616
-5621.
64
Schluns, K. S., K. D. Klonowski, L. Lefrancois.
2004
. Transregulation of memory CD8 T-cell proliferation by IL-15Rα+ bone marrow-derived cells.
Blood
103
:
988
-994.
65
Park, J. H., Q. Yu, B. Erman, J. S. Appelbaum, D. Montoya-Durango, H. Leighton Grimes, A. Singer.
2004
. Suppression of IL7Rα transcription by IL-7 and other prosurvival cytokines: a novel mechanism for maximizing IL-7-dependent T cell survival.
Immunity
21
:
289
-302.
66
Selin, L. K., M. Y. Lin, K. A. Kraemer, D. M. Pardoll, J. P. Schneck, S. M. Varga, P. A. Santolucito, A. K. Pinto, R. M. Welsh.
1999
. Attrition of T cell memory: selective loss of LCMV epitope-specific memory CD8 T cells following infections with heterologous viruses.
Immunity
11
:
733
-742.
67
Feuerer, M., P. Beckhove, L. Bai, E. F. Solomayer, G. Bastert, I. J. Diel, C. Pedain, M. Oberniedermayr, V. Schirrmacher, V. Umansky.
2001
. Therapy of human tumors in NOD/SCID mice with patient-derived reactivated memory T cells from bone marrow.
Nat. Med.
7
:
452
-458.
68
Opstelten, D., D. G. Osmond.
1983
. Pre-B cells in mouse bone marrow: immunofluorescence stathmokinetic studies of the proliferation of cytoplasmic μ-chain-bearing cells in normal mice.
J. Immunol.
131
:
2635
-2640.
69
Benner, R., A. von Oudenaren, G. Koch.
1981
. Induction of antibody formation in mouse bone marrow. B. Pernis, and I. Lefkovits, eds. In
Immunological Methods
Vol. II
:
247
-267 Academic, New York. .
70
Hunt, S. H..
1987
. Preparation of lymphocytes and accessory cells. G. G. B. Klaus, ed.
Lymphocytes, A Practical Approach
  
1
-34 IRL, Oxford. .