Rapamycin is a commonly used immunosuppressive drug in transplant recipients and specifically inhibits the intracellular kinase mTOR 5
. Several recent studies have shown that rapamycin has various effects on the immune system such as inhibiting type I interferon production by plasmacytoid dendritic cells 6
, modulating T cell trafficking 7
, and regulating Foxp3 expression in regulatory T cells 8, 9
. However, the role of the mTOR pathway in regulating CD8 T cell responses is not known. To address this issue we treated B6 mice with rapamycin during the course of an acute LCMV infection and monitored the virus specific CD8 T cell response (). We made the surprising observation that rapamycin enhanced the LCMV specific CD8 T cell response. Increased numbers of antigen specific CD8 T cells were seen in both lymphoid and non-lymphoid tissues ( and Supplementary Fig. 1). The striking thing about this result was the decreased contraction of the T cell response in the rapamycin treated group. Similar frequencies of virus specific effector CD8 T cells were observed in both groups of mice at the peak of the T cell response on day 8 post-infection but there was minimal contraction of the T cells in the rapamycin treated group (). To determine whether the decreased T cell contraction seen between days 8 ~ 30 post-infection in the rapamycin treated group was due to increased cell proliferation and/or reduced cell death, mice were infected with LCMV in the presence or absence of rapamycin and then given BrdU during the T cell contraction phase from days 10 ~ 22 (Supplementary Fig. 2). We found that there was minimal incorporation of BrdU by antigen specific CD8 T cells in either group of mice showing that the decreased contraction of T cells in the presence of rapamycin was not due to increased cell proliferation. Thus, it appears that the major effect of rapamycin is to enhance the survival of antigen specific CD8 T cells.
Rapamycin enhances the number and quality of virus specific memory CD8 T cells
We next examined the phenotype of the memory CD8 T cells present in the two groups of mice at day 36 post-infection (). To investigate this, we performed phenotypic analysis of virus-specific memory CD8 T cells using four markers that are useful in defining memory CD8 T cells: CD127; IL-7 receptor α and essential for memory T cell maintenance 10-13
, CD62L; lymph node homing receptor and associated with high proliferative capacity 14
, KLRG-1; inversely-correlated with long lived memory cells 15, 16
, and Bcl-2; anti-apoptotic and expressed at high levels in memory T cells 12, 17
. Memory CD8 T cells generated in the presence of rapamycin expressed higher levels of CD127, CD62L, and Bcl-2, and had a higher frequency of KLRG-1Low
cells compared to control mice ( and Supplementary Fig. 3). These data strongly suggested that inhibition of mTOR pathway using rapamycin not only increased the magnitude of the virus specific CD8 T cell response () but also improved the functional qualities of the memory CD8 T cells since memory cells with the phenotype (CD127High
) are associated with long-lived protective immunity 14-16
. To directly test this we examined the ability of these memory CD8 T cells to undergo homeostatic proliferation, a property essential for long-term memory maintenance, and to make re-call responses upon re-exposure to antigen. As shown in , virus specific memory CD8 T cells generated in mice treated with rapamycin were superior to memory cells generated in untreated mice in both of these hallmark memory properties.
In the experiment shown in mice were continuously treated with rapamycin during the entire course of the T cell response (day -1~35 post-infection). We next examined how rapamycin would effect the CD8 T cell response if it was only given during the T cell expansion phase (days -1~8 post-infection). These results () were strikingly similar to what we had seen earlier (); even if the rapamycin treatment was discontinued during the contraction phase (days 8~30) there was only minimal death of the effector CD8 T cells generated in the presence of the drug. Previous studies have shown that the day 8 effector CD8 T cell population consists of two subsets; the terminal effector T cells (CD127Low
) that mostly die over the ensuing 2~4 weeks and the memory precursor cells (CD127High
) that mostly survive and further differentiate to give rise to the pool of long-lived memory cells 10, 15, 16
. Our results suggested that rapamycin was enhancing the formation of these memory precursor cells. This was indeed the case and day 8 virus specific effector CD8 T cells generated in rapamycin treated mice contained a higher proportion of CD127High
cells and these cells also expressed higher levels of Bcl-2 (). However, we noticed that the phenotype of memory CD8 T cells at day 36 post-infection was similar in the drug treated and control mice (). This was different from the results obtained upon continuous rapamycin treatment (compare versus ). Taken together these results clearly show that rapamycin enhances the formation of memory precursors during the naïve to effector T cell differentiation phase but these results suggest that, in addition, rapamycin may also regulate the effector to memory transition phase.
Rapamycin treatment during T cell expansion phase increases the number of memory precursors
To test this hypothesis we treated mice with rapamycin only during the T cell contraction phase (days 8~35) following acute LCMV infection (). We found that the number of memory cells generated were not affected by the drug () but the phenotype of these memory CD8 T cells was strikingly different (). Thus, rapamycin treatment during the effector to memory transition phase enhanced the memory differentiation program resulting in a significantly higher number of virus specific CD8 T cells with the phenotype characteristic of highly functional memory cells (p value; <0.0001~0.0022) (). It was important to determine if this represented cell proliferation and outgrowth of a subset of effector CD8 T cells already expressing these memory markers or if rapamycin truly increased the expression of these markers in the surviving effector T cells during this effector to memory differentiation phase. To address this issue, we transferred a highly purified (>99.7%) and CFSE labeled population of day 8 CD62LLow antigen specific effector CD8 T cells into naïve mice and monitored both cell division and memory differentiation of these transferred effector cells in the presence or absence of rapamycin (). We found that there was no cell division during this effector to memory transition phase (days 1~25 post-transfer) but that the memory T cells that differentiated in the presence of rapamycin re-expressed CD62L much faster (). More importantly, these memory CD8 T cells were functionally superior and exhibited better re-call responses and protective immunity (viral control) following challenge with vaccinia virus expressing the LCMV GP33 epitope (). Thus, inhibiting mTOR during the effector to memory transition phase improved the functional qualities of memory T cells.
Rapamycin treatment during effector to memory transition phase accelerates memory differentiation
The results so far have shown that rapamycin can enhance both the magnitude and quality of the CD8 T cell response following a primary viral infection. We next examined whether similar effects would be seen during a secondary response. As shown in Supplementary Fig. 4 rapamycin also enhanced re-call responses when the drug treatment was only done during secondary LCMV infection. Thus, rapamycin regulates both primary and secondary T cell responses and this would have important implications in designing strategies for improving memory T cell qualities during prime-boost vaccine regimens.
To determine if our findings from the mouse model of LCMV infection could be generalized to other systems we examined the effect of rapamycin following immunization of mice with a non-replicating vaccine. In these experiments mice were vaccinated with VLPs (virus-like particles) derived from hepatitis B core antigen genetically fused to the LCMV GP33 epitope 18
. Rapamycin again enhanced both the magnitude and the quality of the VLP-induced memory CD8 T cells (Supplementary Fig. 5). It should be noted that the effects of rapamycin treatment were very long-lasting; memory T cell numbers remained 10-fold higher even 165 days after stopping the drug treatment (Supplementary Fig. 5a). We also tested the applicability of this approach in a non-human primate model. Rhesus macaques previously immunized with vaccinia virus were boosted with MVA in the presence or absence of rapamycin and antigen-specific CD8 T cell responses were analyzed by intracellular IFN-γ staining. We found clear differences in frequencies of antigen specific CD8 T cells between rapamycin treated and untreated monkeys. In the presence of rapamycin, maintenance of higher number of memory CD8 T cells was observed (Supplementary Fig. 6a, b) and slower T cell contraction was evident compared to control animals (Supplementary Fig. 6c). These results demonstrate that rapamycin enhances T cell immunity in both mice and non-human primates following vaccination with either live or inactivated vaccines.
Our results clearly establish that mTOR is a major regulator of memory CD8 T cell differentiation. However, a critical question that needs to be answered is whether mTOR is acting intrinsically in antigen specific CD8 T cells to regulate memory differentiation or if the observed effects of rapamycin on memory formation are mediated by some other cells of the immune system. It is important to resolve this issue since mTOR is ubiquitously expressed by many cells and several recent studies have shown that rapamycin can modulate the functional properties of several other cells of the immune system 6, 8, 9, 19, 20
. To address this question, we used a retrovirus based RNA interference (RNAi) system to specifically knock-down various genes of the mTOR pathway (mTOR, raptor, S6K1, eIF4E, and FKBP12) in antigen specific CD8 T cells. Retroviruses marked by GFP and expressing RNAi for a particular gene or a control retrovirus were used to infect LCMV specific transgenic CD8 T cells (P14 cells) and these transduced cells were then adoptively transferred into naïve mice, followed by LCMV infection (Supplementary Fig. 7a). This system allows us to compare the phenotypic changes that occur during memory T cell differentiation in GFP positive retrovirus transduced versus GFP negative non-transduced antigen specific CD8 T cells in the same environment (ie., the same mouse) (Supplementary Fig. 7b). Thus, any differences in memory differentiation that are seen between these two cell populations can be ascribed to the intrinsic effects of that particular gene in antigen specific T cells.
First, we knocked down mTOR itself in antigen specific CD8 T cells. We found that mTOR RNAi retrovirus-transduced GFP+
P14 cells showed significantly higher expression of the canonical memory T cell markers (CD127, CD62L, Bcl-2, CD27) and lower expression of KLRG-1 compared to non-transduced or control vector transduced P14 cells (). These data show that mTOR acts intrinsically in antigen specific CD8 T cells to regulate memory differentiation. However, since mTOR forms two distinct complexes, the rapamycin-sensitive mTOR complex 1 (mTORC1) and the rapamycin - insensitive mTORC2 5
, mTOR knockdown does not completely mimic rapamycin treatment. To distinguish between these two pathways, we knocked down the gene, raptor, that is an essential component of the mTORC1 complex 21, 22
. As shown in , inhibition of raptor in antigen specific T cells gave results identical to what was seen upon knock down of mTOR identifying the mTORC1 complex as the regulator of memory differentiation. To gain more insight into mechanisms by which mTOR regulates memory formation we examined the role of S6K1 and eIF4E 5
. We found that knockdown of these mTORC1 downstream effectors significantly enhanced memory CD8 T cell differentiation (Supplementary Fig. 8). Thus, our results show that mTOR is exerting its effect through these two downstream molecules.
mTOR acts intrinsically in antigen specific CD8 T cells through the mTORC1 pathway to regulate memory T cell differentiation
To further explore the role of mTOR in T cell intrinsic versus external effects on memory differentiation we made rapamycin-insensitive antigen specific CD8 T cells by knockdown of the FKBP12 protein. This intracellular protein binds rapamycin and it is this FKBP12 - rapamycin complex that inhibits the mTORC1 pathway. Thus, by knocking down FKBP12 in P14 CD8 cells we made these cells insensitive to any intrinsic effects of rapamycin but the drug could still act effectively on all the other cells in the mouse. This system allows us to examine if inhibition of mTOR in other cells can effect memory CD8 T cell differentiation. As shown in inhibiting mTOR in other cells when the antigen specific cells themselves were rapamycin insensitive did not effect memory differentiation. The effect of rapamycin on memory differentiation almost disappeared upon knock down of FKBP12 from the P14 cells and these cells did not show increased expression of the characteristic memory markers (CD127, CD62L, Bcl2, etc.) (see last column of the figures in ). Thus, taken together the results shown in establish that mTOR not only acts intrinsically in antigen specific CD8 T cells but that inhibiting mTOR in other cells has minimal to no effect on memory T cell differentiation.
During the past few years considerable progress has been made in understanding the lineage relationships between naïve, effector and memory T cells and in defining the phenotypic and functional changes that underlie memory CD8 T cell differentiation 1, 23
. However, much less is known about the intracellular molecules and pathways that regulate the generation of memory T cells. In this study we now identify a molecular pathway that regulates memory T cell differentiation and also provide a strategy for modulating the formation of memory cells. In particular, the ability to increase the functional qualities of memory T cells provides a new approach for enhancing the efficacy of vaccines against infectious diseases and cancer.