CD11c^high Dendritic Cell Ablation Impairs Lymphopenia-Driven Proliferation of Naive and Memory CD8⁺ T Cells¹

Regulated production, expansion, and death keep the overall numbers of T lymphocytes relatively constant in the course of the mammalian life (1, 2). In addition, T cells grafted into lymphopenic hosts undergo, in the absence of exogenous antigenic stimuli, clonal expansion until their number plateaus (3, 4, 5). This behavior was initially named homeostatic proliferation. The observation that expanding naive T cells acquire an effector/memory phenotype (6, 7), however, resulted in renaming the phenomenon lymphopenia-driven proliferation (LDP).⁴

More recently, LDP of naive CD4⁺ and CD8⁺ T cells has also been demonstrated in a physiological lymphopenic environment, i.e., the developing neonatal immune system (8, 9). Furthermore, infrequent Ag-independent proliferation plays a crucial role in memory T cell maintenance (10, 11, 12). LDP is also of major clinical relevance, both in a beneficial context, as a source of T cell regeneration in therapy-related lymphopenia (13), and in a harmful setting, as an obstacle to transplantation tolerance (14).

The molecular requirements for LDP of naive and memory CD8⁺ T cells have been defined in some detail. The cytokine IL-7- and TCR-mediated recognition of low affinity self-peptide/MHC complexes are required for the maintenance of naive CD8⁺ T cells (15). In contrast, the survival and proliferation of memory CD8⁺ T cells are generally independent of MHC recognition (16), but require the cytokines IL-7 and IL-15, respectively (17, 18, 19, 20). Interestingly, recent studies suggest that IL-15 must be presented to CD8⁺ memory T cells by neighboring bone marrow (BM)-derived cytokine-presenting cells on a high affinity receptor (IL-15Rα) (20, 21). This indicates that cell-cell contact and a limiting cellular surface are important in lymphopenia-driven/homeostatic proliferation.

Although the molecular interface is relatively well defined, the identity of the BM-derived cell type that provides the surface screened by T cells remains under debate. LDP of naive T cells requires entry of the grafted cells into the dendritic cell (DC)-rich T cell areas of lymphoid organs (22). This together with the observation that adoptively transferred DC can boost the proliferation of naive CD8⁺ T cells in lymphopenic recipient mice (23) have led to the suggestion that DC play a major role in lymphopenia-driven proliferation. More recently, Gruber and Brocker (24) reported that transgenic DC-restricted MHC class expression is sufficient to promote homeostatic CD8⁺ T cell proliferation. However, although these reports suggest that DC can support LDP, none has addressed whether this is a unique DC feature or whether other BM-derived cells carrying self-peptide/MHC class I complexes and expressing the relevant cytokines are sufficient to sustain LDP of naive and memory CD8⁺ T cells.

In this study we report impairment of lymphopenia-driven/homeostatic proliferation of naive and memory CD8⁺ T cells in mice depleted of CD11c^high DC after diphtheria toxin-induced conditional cell ablation. Our results suggest that CD11c^high DC define the limiting cellular surface that determines the size of the peripheral naive and memory CD8⁺ T cell pool.

This study involved the use of CD11c-diptheria toxin (DT) receptor (DTR) transgenic mice (line DTR57; B6.FVB-Tg(Itgax-DTR/GFP)57Lan/J) on the following genetic backgrounds: C57BL/6, CB6F₁ (C57BL/6 × BALB/C):Rag1^−/−, C57BL/6:TCRα^−/−, and CB6F₁. All DTR transgenic mice used in this study were heterozygote transgenic. CD11c-DTR mice were screened by PCR using the following primers: DTR1, 5′-gcc acc atg aag ctg ctg ccg-3′; and DTR2, 5′-tca gtg gga att agt cat gcc-3′. For systemic DC depletion, CD11c-DTR transgenic mice were injected i.p. twice with diphtheria toxin (in PBS; D-2918; Sigma-Aldrich) at the doses of 4 and 2 ng/g body weight 8 h before and 2 days after T cell transfer, respectively. All mice were maintained under specific pathogen-free conditions and handled under protocols approved by the Weizmann Institute animal care committee according to international guidelines.

T cell grafts were isolated from spleens and lymph nodes of the respective donor mice and enriched by magnetic positive selection of CD8⁺ T cells (Miltenyi Biotec). CD8⁺ T cells were labeled with the intracellular fluorescent dye CFSE (Molecular Probes; C-1157) by incubating them at 10⁷ cells/ml in 5 μM CFSE in the absence of serum for 8 min at room temperature. CFSE loading was terminated by addition of an equal volume of cold FCS. Cells were washed twice in complete RPMI 1640 medium. One to 2 × 10⁶ CD8⁺ T cells in 200 μl of PBS were injected into the tail veins of the recipient mice. OVA-specific CD8⁺ T cells carrying a Vα2/Vβ5 TCR specific for the SIINFEKL peptide presented in the context of MHC class I K^b were isolated from OT-I TCR transgenic C57BL/6 or B6.SJL (CD45.1) mice (25). T3.70 clonotype-positive CD8⁺ T cells directed against the male SMCY gene-encoded Ag peptide (KCSRNRQYL) presented in the context of MHC (H2-D^b) and T3.70 clonotype-negative CD8⁺ T cells were isolated from HY TCR transgenic B6.SJL (CD45.1) mice (26). L^d-specific CD8⁺ T cells were isolated from 2C TCR transgenic C57BL/6 donors (27) and detected with the 1B2 Ab, both gifts from Y. Reisner (Department of Immunology, Weizmann Institute of Science, Rehovot, Israel).

Fluorochrome-labeled mAbs used in this study included the PE-coupled Abs anti-Vα2, anti-CD45.1, anti-I-A/I-E (MHC class II), and anti-H-Y; the allophycocyanin-coupled Abs anti-CD11c, anti-CD45.1, and anti-CD8 as well as streptavidin; the PerCP-coupled Ab anti-CD8; the FITC-coupled Abs anti-CD45.1, anti-CD8, and anti-CD44; and the biotinylated Ab anti-CD45.1. The reagents were obtained from BD Pharmingen or eBioscience. At the indicated time points, recipients were killed, spleens and lymph nodes (inguinal, mesenteric) were collected, and single-cell suspensions were prepared in FACS buffer (PBS, 1% FBS, 2 mM EDTA, and 0.05% sodium azide). Mononuclear cells were analyzed by multicolor flow cytometry on a FACSCalibur cytometer (BD Biosciences) using CellQuest Pro software (BD Biosciences). A live lymphocyte gate was set according to forward and side scatter parameters.

To generate CD45.1⁺ memory CD8 T cells, C57BL/6 mice (CD45.2) were grafted with 0.5 × 10⁶ OVA-specific T cells isolated from OT-I mice (CD45.1) and infected with a recombinant Listeria monocytogenes strain expressing OVA (gift from L. Lefrancois, Division of Immunology, University of Connecticut Health Center, Farmington, CT (28)) that was grown in brain-heart infusion broth supplemented with 5 μg/ml erythromycin to an OD₆₀₀ of ∼0.1, diluted in PBS, and injected i.v. (0.2 ml/animal). Mice received 2000 live Listeria in the primary infection and 10,000 live bacteria in the secondary infection. CD44^high memory CD8⁺ T cells, constituting 20–40% of the peripheral CD8⁺ T cell compartment of the infected mice, were isolated by magnetic positive selection using biotinylated anti-CD45.1 Abs (eBioscience) and streptavidin-conjugated beads (Miltenyi Biotec).

C57BL/6 wild-type (wt) or CD11c-DTR transgenic recipient mice were exposed to a single lethal dose of 950 rad total body irradiation. The following day the mice i.v. received 2–5 × 10⁶ of C57BL/6 wt or CD11c-DTR transgenic BM as indicated. The mice were allowed to rest for 8 wk before use.

To investigate the role of DC in LDP, we took advantage of a recently established mouse model that allows for conditional ablation of CD11c^high DC (29). Diphtheria toxin (DT)-induced transient (4-day) DC depletion was instrumental in defining essential in vivo functions of CD11c^high DC, e.g., their requirement for the initiation of cytotoxic T cell responses to infection with intracellular bacteria, protozoa, and viruses (29, 30). Lymphopenia-driven T cell expansion is readily studied in murine experimental systems using adoptive T cell transfer into hosts rendered lymphopenic through genetic manipulation. We therefore crossed the CD11c-DTR transgene onto a TCRα-deficient background that lacks αβ T cells (31). CD8⁺ T cell grafts isolated from wt C57BL/6 were labeled with CFSE before their adoptive transfer. Dilution of this intracellular fluorescent dye can serve as an indicator of in vivo cell proliferation. Nontransgenic and CD11c DTR transgenic TCRα^−/− recipient mice were treated 8 h before and 2 days after T cell transfer with DT (4 and 2 ng/g body weight, respectively). Four days after T cell transfer, recipient mice were killed and analyzed for DC depletion and CD8⁺ T cell proliferation. As shown in Fig. 1,A, DT treatment of CD11c DTR-transgenic mice resulted in the depletion of DC that express the DTR-GFP fusion protein. Furthermore, the absence of DC impaired LDP of the naive T cell graft in the DT-treated CD11c DTR-transgenic C57BL/6.TCRα^−/− mice. To substantiate this finding, we resorted to an alternative system involving the transfer of OT-I TCR transgenic CD8⁺ T cells into nontransgenic and CD11c DTR-transgenic CB6F₁.RAG1^−/− recipients. Again, DT treatment and, hence, DC depletion impaired the LDP of the naive CD8⁺ T cell graft observed in untreated mice (Fig. 1 B).

It has been noted that the outcome and requirements of homeostatic proliferation in lymphopenic mice generated through gene ablation might differ from those observed in irradiated mice, because irradiation results in the induction of inflammatory cytokines. We therefore next examined the role of CD11c^high DC in irradiated CD11c-DTR-transgenic mice. Nontransgenic and CD11c-DTR-transgenic C57BL/6 mice were irradiated (450 rad) and treated with DT 8 h before and 2 days after receiving a CFSE-labeled OT-I TCR transgenic CD8⁺ T cell graft. Four days after transfer, spleens and lymph nodes of lymphopenic recipients were isolated and analyzed by flow cytometry for DC depletion and proliferation of the grafted CD45.1⁺ CD8⁺ T cells. As shown in Fig. 1 C, CD11c^high DC ablation impaired LDP of the naive T cell graft in sublethally irradiated C57BL/6 mice.

In adult mice, thymic output vastly exceeds the cell number required to maintain peripheral T cell pools. LDP is, therefore, unlikely to contribute to the maintenance of the naive repertoire under normal conditions. In contrast, long-term survival of memory T cells in the absence of exogenous Ag relies on their spontaneous slow division (11, 12). This homeostatic self-renewal of CD8⁺ memory T cells is independent of MHC class I recognition (16), but requires the cytokines IL-7 and IL-15 (17, 20), as does the more rapid acute LDP of adoptively transferred CD8⁺ memory T cells (18). To determine whether LDP of memory CD8⁺ T cells, like that of naive CD8⁺ T cells, depends on CD11c^high DC, we transferred naive and memory CD8⁺ T cells into untreated and DC-depleted lymphopenic recipients. Pure CD8⁺ memory T cells were isolated from C57BL/6 mice (CD45.2⁺) that had received an adoptive OT-I TCR transgenic CD8⁺ T cell graft (CD45.1⁺) before sequential infection with an OVA-expressing recombinant L. monocytogenes strain (rLM-OVA; provided by L. Lefrancois (28)). Two months after the last infection, CD45.1⁺CD44^high memory OT-I CD8⁺ T cells were isolated by magnetic purification from spleens and lymph nodes, labeled with CFSE, and adoptively transferred into sublethally irradiated, DT-treated, nontransgenic and CD11c-DTR-transgenic recipient mice (Fig. 2,B). A control group received CD44^low naive OT-I CD8⁺ T cells (Fig. 2 A). Recipient mice were analyzed 4 days after T cell transfer. We found that DT-induced DC ablation in CD11c-DTR transgenic recipient mice impaired lymphopenia-driven expansion of both naive and memory CD8⁺ T cell grafts.

Lymphopenia-driven lymphocyte proliferation is a phenomenon of major clinical importance. For example, the expansion of donor memory T cells that are refractory to immunomodulation, such as anti-CD154 or CTLA-4 Ig treatment, is believed to be a major barrier to transplantation tolerance (32). In contrast, the expansion of mature T cells that contaminate allogeneic BM grafts is believed to promote beneficial host immunoreconstitution and to potentially contribute to the eradication of residual host tumor cells (graft-vs-leukemia effect) in the treatment of hematologic malignancies. Mature donor CD8⁺ T cells are also the main cause of graft-vs-host disease (GVHD) due to their activation by major and minor histocompatibility Ag mismatches (33). Interestingly, a number of recent reports have suggested that host APC (34) and, more specifically, host DC (35) may be responsible for the priming of alloreactive CD8⁺ T cells, and that inactivation of host DC might, hence, be of therapeutic value. Because our results suggest that such a strategy would interfere with host immune reconstitution, we studied the potential benefit of DC ablation in GVHD prevention.

To compare the impact of DC ablation on homeostatic vs alloreactive T cell proliferation, we exploited the observation that anti-HY-TCR-transgenic CD8⁺ T cells directed against the minor histocompatibility Ag HY (26) fail to undergo homeostatic proliferation (36). Expansion of CD8⁺ T cells isolated from HY-TCR-transgenic mice in lymphopenic females, therefore, is restricted to the T3.70 clonotype-negative population (36) (Fig. 3,A). When grafted into irradiated male mice, T3.70 clonotype-negative and T3.70 clonotype-positive populations undergo lymphopenia and alloantigen-driven proliferation, respectively. We transferred CFSE-labeled CD8⁺ T cells isolated from HY-TCR-transgenic mice into nonirradiated and irradiated CD11c-DTR and control mice of both genders and analyzed T cell expansion in the presence and the absence of DC. As expected, DC ablation abrogated LDP of HY clonotype-negative CD8⁺ T cells in DTR-transgenic females and males. Surprisingly, however, DC ablation did not impair the allogeneic response of the HY clonotype-positive T cells to the male Ag (Fig. 3 A).

To extend this finding to a GVH reaction elicited by a major histocompatibility mismatch, we examined the response of L^d-specific CD8⁺ T cells isolated from C57BL/6 2C transgenic donors (27) against irradiated CB6F₁ recipients (H2^b/d). 2C TCR-transgenic cells were isolated from spleens and lymph nodes of donor mice, labeled with CFSE, and transferred into irradiated (950 rad), DT-treated, nontransgenic and DTR-transgenic mice. The C57BL/6 (H2^b) and CB6F₁ (H2^b/d) recipient mice were analyzed 4 days later for the effect of DC ablation on lymphopenia-driven and alloantigen-driven proliferation of the CD8⁺ T cell graft, respectively. DC depletion completely abrogated proliferation of the grafted CD8⁺ T cells in C57BL/6 (H2^b) mice, but the L^d-alloantigen-driven response in CB6F₁ (H2^b/d) remained unaffected (Fig. 3 B). Taken together, these results show that minor and major alloantigen-driven proliferation of naive CD8⁺ T cells can persist in the absence of CD11c^high DC. This argues that alloreactive T cells are stimulated by APC other than CD11c^high DC. Alternatively, the number of residual CD11c^high DC surviving the DT treatment of CD11c-DTR-transgenic mice may be sufficient to promote the GvH response. Future experiments involving mixed BM chimeras are required to resolve this issue.

CD11c-DTR-transgenic mice do not tolerate repeated DT injections (29), and the time window of their DT-induced DC depletion is therefore limited. In contrast, syngeneic BM chimeras generated through reconstitution of lethally irradiated recipient mice with CD11c-DTR BM can be treated with DT for prolonged periods of time without adverse side effects (Fig. 4,A) (37). The observed lethality in CD11c-DTR-transgenic mice thus seems to be due to DTR expression and, hence, DT sensitivity of a radioresistant, potentially nonhemopoietic cell type. This idea is also supported by the fact that lethally irradiated DTR-transgenic mice reconstituted with wt BM succumb after DT treatment to weight loss and death as do CD11c-DTR transgenic mice (Fig. 4,A). Toxin treatment of CD11c-DTR BM chimeras results in DC depletion and abrogation of CTL responses to challenge with intracellular pathogens, such as Listeria (Fig. 4,B), confirming previous results obtained from CD11c-DTR-transgenic mice (29). To investigate whether cells other than CD11c^high DC can compensate for LDP at later time points after T cell transfer, we generated lymphopenic CD11c-DTR BM chimeras and engrafted them with naive (wt) and memory (OT-I) CD8⁺ T cells. As shown in Fig. 4 C, in contrast with CD11c-DTR transgenic recipients in CD11c-DTR BM chimeras, both types of grafts underwent LDP in the absence of CD11c^high DC.

Through transient in vivo depletion of CD11c^high DC in lymphopenic hosts, we have been able to show their requirement for homeostatic expansion of grafted naive and memory CD8⁺ T cells. Our results suggest that grafted CD8⁺ T cells fail to sense lymphopenia when the number of CD11c^high DC in recipient mice is reduced. Cotransfer of bystander cells efficiently blocks LDP of T cell grafts, suggesting that T cells compete for specific microcompartments or cytokines (1, 22). Naive and memory CD8⁺ T cells differ in their requirements for survival and division and are thus segregated into independent ecological niches (38). CD11c^high DC seem to define these limiting niches for both peripheral T cell repertoires. This idea is supported by the recent finding that DC are sufficient to support LDP (24) and the observation that DC in vivo expansion by Flt3 ligand promotes LDP (39).

LDP of naive CD8⁺ T cell depends on the presence of MHC class I and the cytokine IL-7. Although it has been shown that MHC has to be expressed on a BM-derived cell, IL-7 could be provided from an irradiation-resistant cellular source (40). This indicates that although DC can synthesize IL-7 (41), their crucial contribution in driving homeostatic expansion of naive CD8⁺ T cells might be to provide self-peptide/MHC complexes.

In contrast to naive CD8⁺ T cells, memory CD8⁺ T cells are independent of MHC expression for their maintenance and homeostatic expansion, but, for the latter, require the cytokines IL-7 and IL-15, respectively (18, 19, 20). Furthermore, a number of recent reports point toward a unique role of the cytokine IL-15, presented on its high affinity receptor (IL-15Rα) by BM-derived cells, in this process. In vitro studies have shown that IL-15Rα can be recycled and presents exogenous IL-15 in trans (42). More recent in vivo studies using mixed BM chimeras indicate that memory CD8⁺ T maintenance requires BM-derived cells capable of producing both receptor and ligand (21). Our results suggest that these BM-derived, cytokine-presenting cells might be CD11c^high DC. Finally, although our results indicate a critical role of CD11c^high DC in acute lymphopenia-driven memory CD8⁺ T cell expansion, it remains to be shown whether CD11c^high DC ablation also interferes with long-term memory T cell maintenance.

We found that in DTR-transgenic, but otherwise syngeneic, BM chimeras, cells other than CD11c^high DC are able to promote LDP of naive and memory CD8⁺ T cells. Syngeneic BM chimeras were previously shown to develop after cyclosporin treatment syngeneic GVHD and, hence, are believed to lack regulatory mechanisms that control autoreactive T cells (43). Adoptive transfer of CD4⁺CD25⁺ regulatory T cells into DTR transgenic BM chimeras did not restore the need for CD11c^high for LPD of grafted CD8⁺ T cells (data not shown). Therefore, the reason for the discrepancy between the DTR-transgenic mice and BM chimeras remains unclear.

We recently noticed that DT treatment of CD11c-DTR-transgenic mice results in rapid depletion of marginal zone (MZ) macrophages in addition to CD11c^high DC (S. Jung, unpublished observation) (44). We currently do not know whether depletion of the MZ macrophages is due to CD11c/DTR expression or to secondary effects. In this respect, it is notable that MZ macrophages are generally sensitive to manipulation and are absent in multiple strains of mutant mice, including CCL19/CCL21-deficient plt/plt mice (45). Homeostatic proliferation of CD8⁺ T cells was reported to proceed normally in plt/plt mice (46), consistent with the impairment of homeostatic proliferation reported in this study resulting from the depletion of CD11c^high DC rather than MZ macrophages.

Lymphopenia-driven proliferation of mature grafted T cells contributes to the effective reconstitution of the recipient’s immune system after allogeneic BM transplantation (47). Our results are therefore of clinical relevance, particularly in the context of recent suggestions that the use of peritransplant ablation of host APC/DC could ameliorate GVHD (34, 35). Surprisingly, we found that the ablation of conventional CD11c^high host DC had no impact on the graft-vs-host response in two established experimental systems of minor and major alloreactivity. Interestingly, recent data suggest a crucial role of CD11c^neg-low epidermal Langerhans cells (LC) in the priming of GVHD (48). However, CD11c expression is up-regulated in LC upon their migration from the epidermis to lymph nodes, and LC thus become sensitive to DT-induced ablation in CD11c-DTR mice (A. Sapoznikov, unpublished observation). Future experiments will be needed to determine whether the observed expansion of alloreactive CD8⁺ T cells is due to residual BM-derived host APC.

Cumulatively, the results presented in this study establish a critical role for a subset of mononuclear phagocytes, the CD11c^high DC, in the control of lymphopenia-driven expansion of both naive and memory CD8⁺ T cells. Previous studies defined distinct membrane-associated molecular requirements for the expansion of naive and memory CD8⁺ T cells, i.e., self-peptide/MHC and IL-15Rα-bound IL-15, respectively. Our findings therefore support the idea of a novel dual role of CD11c^high DC as unique APC and cytokine-presenting cells.

We thank Drs. Alexandra Mahler and Yair Reisner for critical review of the manuscript and helpful discussion, and J. Chermesh and O. Amram for animal husbandry.

The authors have no financial conflict of interest.