The long-term survival of Ag-specific memory CD4 T cells provides enhanced protection against previously encountered pathogens (40
). As a site rich in survival cytokines, the bone marrow is predicted to be a critical ‘survial niche’ for memory T cells (3
). However, it is unknown if the entry of memory CD4 T cells into the bone marrow is required for their long-term survival. In this study, we use a peptide:MHC class II tetramer enrichment technique to examine the role of β1 integrin in the maintenance of endogenous, polyclonal Ag-specific CD4 T cells in the bone marrow following bacterial infection. We demonstrate that CD4 T cells lacking β1 integrin enter the bone marrow early following an infection but are not maintained there long-term. This lack of Ag-specific CD4 T cell maintenance in the bone marrow does not result in a systemic loss of CD4 T cell memory or a reduced ability to mount a secondary response. These findings support the idea that the localization of CD4 T cells to the bone marrow is not required for the long-term maintenance of CD4 T cell memory.
We observed comparable numbers of wt and β1−/− Ag-specific CD4 T cells in the bone marrow at early time points after a primary Listeria
infection. Co-homing assays of day 5 post-infection cells into day 5 post-infection hosts confirmed that localization of recently activated CD4 T cells to the bone marrow was independent of β1 integrin expression. Similar findings were also obtained when day 5 cells were transferred into uninfected hosts. This revealed that potential infection-induced alterations in the bone marrow microenvironment or the presence of antigen in the bone marrow is not required for β1 integrin-independent localization of recently activated CD4 T cells to the bone marrow. Our results are consistent with the previously reported promiscuous distribution of recently activated CD4 T cells to normal and inflamed tissue without the presence of antigen (1
). Thus, the initial entry of β1−/− T cells into the bone marrow is mediated by cell intrinsic factors rather than systemic changes that occur during acute infection.
Our observation that β1 integrin expression is not required for the entry of recently activated CD4 T cells into the bone marrow is surprising, as previous studies have reported that the binding of the α4β1 integrin to VCAM-1 promotes firm arrest and subsequent entry of T cells into the bone marrow (21
). However, T cell migration into the bone marrow also involves the function of other adhesion molecules, such as PSGL-1, which promotes T cell rolling (15
). In addition, LFA-1 integrin plays a minor role in bone marrow entry that becomes more prominent when the α4β1/VCAM-1 interaction is compromised (21
). Of note, the vast majority of systemic Ag-specific CD4 T cells responding to Listeria
infection express high amounts of LFA-1 and functional PSGL-1. Furthermore, the CD4 T cells recovered from the bone marrow are enriched for this population. Thus, the high level expression of functional PSGL-1 and LFA-1 on Ag-specific CD4 T cells early after infection may allow for efficient bone marrow entry in the absence of β1 integrin expression. Alternatively, the enhanced expression of α4β7 on β1−/− CD4 T cells might compensate for the absence of β1 integrin (29
). α4β7 integrin-mediated compensation could occur either through direct binding to VCAM-1 (41
) or binding to the α4β7 ligand MAdCAM-1, which can be unregulated on the bone marrow vasculature during inflammation (42
). Our data argues against either of these possibilities, as β1−/− CD4 T cells additionally lacking β7 integrin (β1β7−/−) enter the bone marrow in similar numbers as wt cells at day 5.
Previous work has demonstrated that T cells emigrate from and recirculate back to the bone marrow (34
). Parabiosis studies also predict that T cells recirculate through the bone marrow (36
) rather than take up long-term residence, as can be seen in the intestine (37
). Steady-state maintenance of CD4 T cells in the bone marrow is thus highly related to the rates of cell entry and exit. In our system, by day 12 post-infection, the number of 2W1S-specific CD4 T cells in the bone marrow of β1−/− mice is ~60% reduced compared to wt controls. This decreased number of Ag-specific CD4 T cells in the bone marrow of β1−/− mice continues to decline and becomes ~90% reduced compared to wt by day 120. In an attempt to determine the relative contributions of entry and retention to the gradual loss of β1−/− CD4 T cells from the bone marrow, we co-transferred wt and β1−/− CD4 T cells 20 days following an infection into day 20 hosts. There was a modest reduction in the localization of β1−/− CD4 T cells in the bone marrow at 2 hours post-transfer, but this finding did not reach statistical significance (p>0.05). Although these cells express lower levels of LFA-1 and functional PSGL-1 than at day 5 post-infection, high level expression of these molecules is still observed and may provide some functional redundancy. These findings predict that 20 days after an infection, CD4 T cells lacking β1−/− would only be minimally impaired in their ability to enter into the bone marrow. However, by 18 hours post-transfer there is a significant reduction in the maintenance of β1−/− CD4 T cells in the bone marrow. This is also observed when day 20 cells are transferred into day 5 hosts. Thus, this effect is likely cell intrinsic as it is not overcome by potential alterations in the bone marrow vasculature or stroma that may occur with an acute infection. Together, our results suggest that the rapid decline of β1−/− CD4 T cells in the bone marrow following bacterial infection is due predominantly to impaired retention. However, as transferred T cells are continuously entering and exiting the bone marrow, the modest reduction in the entry of β1−/− CD4 T cells observed at 2 hours may, over time, also contribute to the loss of T cell accumulation in the bone marrow.
In addition to a reduction in retention of cells, another possible mechanism for the reduced maintenance of β1−/− CD4 T cells in the bone marrow is increased cell death. Signals from β1 integrins have been proposed to promote the survival of memory T cells in non-lymphoid tissue (24
). In addition, β1 integrin may be important for the localization of CD4 T cells to VCAM-1+ stromal cells producing IL-7 in the bone marrow (16
). Staining for effector caspases 3 & 7 revealed no differences between wt and β1−/− CD4 T cells in either the bone marrow or the spleen. In both wt and β1−/− mice, 2W1S-specific CD4 T cells did demonstrate significantly reduced caspase activity in bone marrow compared to the spleen. This reduced caspase activity is consistent with the idea of the bone marrow as a T cell ‘survival niche’. As both wt and β1−/− CD4 T cells demonstrate higher viability in the bone marrow, this suggests that access to survival factors in the bone marrow is independent of β1 integrin expression. This is a surprising result, as the IL-7 producing stromal cells express the β1 integrin ligand VCAM-1 (28
). Another possibility is that the most viable population of 2W1S-specific CD4 T cells from the spleen are the cells that migrate to the bone marrow. Our experimental results do not allow us to distinguish between these two possibilities. Overall, our results suggest that cell death is not a major factor reducing the abundance of β1−/− CD4 T cells in the bone marrow.
In many systems, β1 integrins function as T cell retention molecules in peripheral tissue (38
). Maintenance of memory T cells in peripheral tissue is predicted to provide rapid protection against pathogen reinfection. The collagen binding integrin α1β1 is critical for T cell retention in the lung following influenza infection (24
). In this system, lung resident α1β1+ effector-memory CD4 cells are the major IFN-γ-secreting, rapid responders during a secondary infection (26
). In humans, the collagen binding integrin α1β1 is a marker of similar TH
1 phenotype CCR7low
effector-memory CD4 T cells (43
). Additionally, skin resident effector-memory CD4 T cells specifically express α1β1, which is involved in their retention in the epidermis (44
). In our system, the majority of the CCR7low
effector-memory like CD4 T cells are maintained in the spleen and lymphoid organs (20
). The CD4 T cells we recover from the bone marrow are enriched for this CCR7low
effector-memory like population compared to the spleen. This CCR7low
effector-memory like population expresses high levels of both β1 and LFA-1, consistent with the characterization of effector-memory T cells in humans (45
). By day 20, CD4 T cells expressing a CCR7low
effector-memory like phenotype remain the predominant population in the bone marrow (data not shown), suggesting that this population is particularly efficient at localizing to this site. This also corresponds to the population of CD4 T cells reported to localize to the bone marrow in other systems (5
). Bone marrow resident CD4 T cells express high levels of the collagen binding integrin α2β1 (17
), but a role for α1β1 as a retention molecule in the bone marrow has not been excluded.
For naïve CD4 T cell homeostasis, stromal components of the lymph nodes and spleen are thought to be the major source of the survival cytokine IL-7 (46
). In contrast, the bone marrow has recently been proposed to be the major site for memory CD4 T cell maintenance and IL-7 survival signaling (17
). The bone marrow has unique properties that situate it somewhere between peripheral tissue and lymphoid tissue (3
). Unlike peripheral tissue, where the presence of antigen is thought to drive retention (47
), localization to the bone marrow can occur with or without local antigen (3
). Thus, local immunosurveillance may not completely explain the presence of CD4 T cells in the bone marrow and its function as a survival reservoir has been gaining support (3
). If the bone marrow is the major source of IL-7 for memory CD4 T cells, we expected to observe a global loss of CD4 T cell memory over-time in the β1−/− mice. Surprisingly, the decreased ability of β1−/− CD4 T cells to be maintained in the bone marrow does not result in a systemic decrease in CD4 T cell memory or loss of a rapid proliferative response to antigen re-challenge. At nearly one year post-infection, when we are unable to detect β1−/− memory CD4 T cells in the bone marrow using our peptide:MHC class II tetramer enrichment technique, the secondary response in the spleen and bone marrow is comparable to wt mice. In our system, the spleen remains the predominant site of memory CD4 T cell localization, although the total percentage of CD4 T cells maintained in the bone marrow compared to the spleen in wt mice does increase over time. These results are consistent with other recently published work utilizing this system (20
). The ability to maintain long-term CD4 T cell memory in the presence of decreased bone marrow localization is seemingly in contrast to other work using adoptive transfer systems (17
). The reasons for this discrepancy are unclear but may relate to the analysis of polyclonal versus TCR transgenic T cells, or the route/method of antigen challenge.
In summary, we have demonstrated an important function for β1 integrin in the long-term maintenance of memory CD4 T cells in the bone marrow. Our findings reveal that β1 integrin expression is critical for steady-state maintenance of Ag-specific CD4 T cells in the bone marrow during the memory phase. Although we demonstrate significantly reduced numbers of CD4 T cells maintained in the bone marrow of β1−/− mice during the memory phase, this does not result in a global decrease in CD4 T cell memory survival or proliferative response upon re-challenge. Our results suggest that survival signals produced by sources other than the bone marrow, such as the stromal compartment of the spleen and lymph nodes (46
), may be sufficient for long-term memory CD4 T cell survival.