|Home | About | Journals | Submit | Contact Us | Français|
Recent studies have underscored the importance of memory T cells in mediating protective immunity against pathogens and cancer. Pharmacological inhibition of regulators that mediate T cell differentiation promotes the differentiation of activated CD8+ T cells into memory cells. Nonetheless, pharmacological agents have broad targets and can induce undesirable immunosuppressive effects. Here, we tested the hypothesis that aptamer-targeted siRNA inhibition of mTOR complex 1 (mTORC1) function in CD8+ T cells can enhance their differentiation into memory T cells and potentiate antitumor immunity more effectively than the pharmacologic inhibitor rapamycin. To specifically target activated cells, we conjugated an siRNA targeting the mTORC1 component raptor to an aptamer that binds 4-1BB, a costimulatory molecule that is expressed on CD8+ T cells following TCR stimulation. We found that systemic administration of the 4-1BB aptamer-raptor siRNA to mice downregulated mTORC1 activity in the majority of CD8+ T cells, leading to the generation of a potent memory response that exhibited cytotoxic effector functions and enhanced vaccine-induced protective immunity in tumor-bearing mice. In contrast, while treatment with the general mTORC1 inhibitor rapamycin also enhanced antigen-activated CD8+ T cell persistence, the cytotoxic effector functions of the reactivated memory cells were reduced and the alloreactivity of DCs was diminished. Consistent with the immunological findings, mice treated with rapamycin, but not with 4-1BB aptamer-raptor siRNA, failed to reject a subsequent tumor challenge.
Studies in mice have highlighted the importance of persistence of the vaccine-induced immune response (immunological memory) in mediating protective immunity against infectious diseases and cancer (1). A correlation between T cell memory and protective immunity was also seen in nonhuman primates vaccinated against SIV (2–4) and in cancer patients treated with adoptively transferred T cells (5–9). Vaccination protocols must therefore be designed for or complemented with treatments that promote the generation of strong and long-lasting memory responses.
Multiple extrinsic pathways control memory differentiation by regulating the balanced expression of intracellular mediators in activated CD8+ T cells (10–13). For example, elevated levels of mTOR (14), T-bet (15), BLIMP1 (16–18), or GSK3 (19) promote the accumulation of short-lived effectors, whereas products like BCL6 (18), Eomes (20, 21), TRAF6 (22), or TCF1 (21, 23) promote the accumulation of memory cells. Notably, inhibiting the mediators of effector differentiation using genetic means or, whenever available, pharmacological agents, redirected the activated T cells to differentiate along the memory pathway. For example, pharmacological inhibition of mTOR with rapamycin in lymphocytic choriomeningitis virus–infected (LCMV-infected) mice led to enhanced differentiation of the LCMV-specific CD8+ T cells into memory cells (14); activation of AMPK with metformin promoted the development of memory CD8+ T cells (22); and inhibition of GSK3 with TWS119 led to the differentiation of memory precursors with self-renewal capacity (19). The therapeutic potential of promoting memory responses with pharmacological agents like rapamycin, TWS119, or metformin was demonstrated in murine studies using antigen-specific transgenic CD8+ T cells adoptively transferred to tumor-bearing mice (19, 24) or to mice subsequently challenged with tumor or recombinant virus (14, 22, 25). With one exception (19), the transgenic T cells targeted a potent model antigen, chicken OVA, ectopically expressed in the tumor or virus. While providing a proof-of-concept that the promotion of memory responses with pharmacological agents can enhance protective immunity, the question remains whether such strategies will be therapeutically useful in clinical settings.
Pharmacological agents, given the broad distribution of their targets, can also exhibit undesirable immune (suppressive) and nonimmune effects, raising significant, if not insurmountable, challenges in translating those finding to human patients. For example, rapamycin inhibition of mTOR promotes the development of immunosuppressive regulatory Foxp3+ CD4+ T cells (Tregs) (26), polarizes DCs to become tolerogenic APCs (27, 28), and prevents the tissue trafficking of activated T cells (ref. 29 and reviewed in refs. 30, 31). In addition, the development of pharmacological agents designed to modulate the function of intracellular targets that are not accessible to antibodies (“undruggable targets”) is challenging, and their availability, especially for clinical use, is limited. There are currently no pharmacological agents available for the inhibition of intracellular mediators such as BLIMP1 or T-bet, however, their inhibition could provide substantial advantages in promoting immunological memory.
In this study, we describe a versatile, broadly applicable, and clinically feasible approach to promoting the generation of memory T cell responses that addresses the main limitations of pharmacological agents. We used siRNAs to downregulate intracellular mediators of CD8+ T cell effector differentiation. RNAi is broadly applicable to virtually any target, including “nondruggable” intracellular targets such as BLIMP1 or T-bet. To reduce the undesirable effects that could result from the downregulation of the siRNA targets in other cell types, we targeted systemically administered siRNA to CD8+ T cells using oligonucleotide aptamers. Aptamers are high-affinity, single-stranded nucleic acid ligands that can be isolated using a combinatorial chemistry process known as SELEX (systematic evolution of ligands by exponential enrichment) (32). Aptamers exhibit specificity and avidity comparable to or exceeding those of antibodies and can be generated against most targets. Recent publications have illustrated the feasibility and therapeutic potential of aptamers as targeting ligands to eradicate tumors (33), sensitize tumor cells to radiation therapy (34), inhibit HIV replication (35, 36), and potentiate tumor immunity (37, 38). In this study, we show that aptamer-targeted siRNA inhibition of mTOR function in CD8+ T cells potentiates a vaccine-induced memory and an antitumor immunity that are quantitatively comparable, but qualitatively superior, to rapamycin treatment.
To test whether CD8+ T cell memory can be enhanced by targeted inhibition of mTOR, we conjugated an siRNA specific to raptor, a component of the mammalian target of rapamycin complex 1 (mTORC1) (30), to an aptamer that binds to 4-1BB (39), a major costimulatory molecule that is transiently expressed on CD8+ T cells following TCR stimulation (40, 41). Incubation of activated OT-I cells, transgenic CD8+ T cells that recognize a determinant of the chicken OVA product (42), with 4-1BB aptamer-raptor siRNA (4-1BB-raptor) conjugate downregulated raptor mRNA in an aptamer-dependent manner (Figure (Figure1A).1A). Suggestive evidence that the observed reduction in raptor mRNA levels led to the inhibition of mTOR activity was provided by our observations that raptor mRNA downregulation was accompanied by downregulation of Tbet mRNA and upregulation of Eomes mRNA (refs. 25, 43, and Figure Figure1B)1B) and that the 4-1BB-raptor conjugate–treated OT-I cells exhibited prolonged survival and enhanced proliferative capacity comparable with that of rapamycin-treated cells as was previously shown (refs. 14, 25, and Figure Figure1,1, C and D).
Supplemental Figure 1A (supplemental material available online with this article; doi: 10.1172/JCI69856DS1) shows that i.v. administered 32P-labeled 4-1BB, but not scrambled, aptamer-raptor siRNA conjugates become associated with 4-1BB–expressing cells, but not with 4-1BB–negative cells from lymph nodes, bone marrow, and spleen. The systemically administered 4-1BB-raptor conjugate accumulated preferentially in the bone marrow where 4-1BB–expressing memory CD8+ T cells reside, whereas no significant accumulation was noted in the lungs, kidney, or liver, as is often seen upon administration of unformulated siRNA (Supplemental Figure 1B).
To test whether systemic administration of 4-1BB-raptor conjugates is capable of inhibiting mTOR activity in activated CD8+ T cells in vivo, we transferred OT-I cells into C57BL/6 mice, and 2 days later we vaccinated the mice with OVA peptide and treated them with aptamer-siRNA conjugates or with rapamycin. In this experimental system, 4-1BB is upregulated on the OT-I cells as early as 6 hours later, peaks between 12 and 24 hours, and returns to background levels after 48 hours (data not shown). Two days later, we determined conjugate or rapamycin treatment mTORC1 activity in the host cells and donor OT-I cells by measuring the phosphorylation of S6 ribosomal protein, one of the downstream targets of mTORC1. As shown in Figure Figure2A,2A, treatment of the OT-I–bearing mice with rapamycin or with 4-1BB-raptor, but not with 4-1BB-GFP, conjugate downregulated mTORC1 activity in 90% and 59% of the OT-I cells, respectively. The cell and target specificity of raptor siRNA inhibition is indicated by the fact that rapamycin, but not the 4-1BB-raptor conjugate, also inhibited mTOR activity in a proportion of host cells (Figure (Figure2A)2A) and that the 4-1BB-raptor conjugate inhibited the activity of mTORC1, but not mTORC2, whereas rapamycin inhibited both mTOR complexes, as was previously shown (refs. 44, 45, and Figure Figure2B).2B). OT-I cells in mice treated with 4-1BB-GFP conjugate expressed more mTORC1 (64%) compared with mock-treated mice (51%), suggesting that 4-1BB costimulation affected mTORC1 expression. This difference, however, was not always seen. The small decrease of phospho-S6–expressing host (CD45.1–) cells seen in the 4-1BB-raptor conjugate–treated mice may correspond to a small fraction of non-CD8+ T cells that also express 4-1BB (46, 47), reflecting the lack of complete specificity of this targeting protocol. We further evaluated the specificity of aptamer-targeted mTORC1 inhibition by cotransferring mice with OT-I cells together with EL-4 T lymphoma cells that can be readily distinguished by flow cytometry (Figure (Figure2C).2C). Unlike OT-I cells, EL4 cells did not upregulate 4-1BB upon peptide and LPS administration (Figure (Figure2C).2C). As shown in Figure Figure2D,2D, whereas rapamycin treatment inhibited mTORC1 signaling in both OT-I and EL-4 cells, the administration of 4-1BB-raptor conjugate inhibited mTORC1 in 4-1BB–positive OT-I cells, but not in 4-1BB–negative EL4 cells. Figure Figure2E2E shows that on a population level, raptor RNA, but not CD8 or β-actin RNA, is reduced in OT-I cells (upper panel), but not in host CD8+ cells (lower panel) isolated from OT-I–bearing mice treated with 4-1BB-raptor, but not 4-1BB-GFP, conjugate. Given that the efficiency of aptamer-targeted siRNA delivery to circulating OT-I cells was about 60% as measured by functional analysis of mTORC1 (Figure (Figure2A),2A), the extent of raptor downregulation on a per-cell basis was considerably higher. Supplemental Figure 2A shows that 4-1BB-raptor conjugates accumulated preferentially in the transferred OT-I cells compared with host CD4-, CD8-, CD19-, or F4/80-expressing cells, and Supplemental Figure 2B shows that this correlates with inhibition of mTORC1 in OT-I, but not in host CD4-, CD8-, CD19-, or F4/80-expressing cells, whereas treatment with rapamycin led to mTORC1 inhibition in all cell types analyzed. Overall, these observations demonstrate the high efficiency (Figure (Figure2,2, A and E) and specificity (Figure (Figure2,2, B and D, Supplemental Figure 1A, and Supplemental Figure 2) of aptamer-targeted siRNA delivery to circulating 4-1BB–expressing CD8+ T cells in vivo.
We next determined whether the 4-1BB-raptor conjugate–mediated mTOR inhibition leads to the generation of enhanced memory CD8+ T cell responses in mice. Figure Figure3,3, A and B, show that on days 5 or 8, a higher proportion of OT-I cells in mice treated with rapamycin or 4-1BB-raptor, but not 4-1BB-GFP, conjugate exhibited a CD62LhiCD44hi or CD62LhiCD127hi phenotype, respectively, indicative of an enhanced proliferative capacity and the potential to develop into memory cells. At later time points, we found increased numbers of OT-I cells in the blood of both rapamycin- and 4-1BB-raptor–treated mice (Figure (Figure3C)3C) and on day 28 in the spleen (Figure (Figure3D).3D). Thus, both rapamycin and 4-1BB-raptor treatment enhanced the persistence of the antigen-activated OT-I cells that were enriched for cells with a memory precursor phenotype. Of note, in this and subsequent experiments, mice were injected twice daily with low-dose rapamycin because eight or more daily injections as well as higher doses were detrimental (data not shown). This is different from what was seen in poxvirus (24) or LCMV infection (14) models, perhaps reflecting the differences in the conditions prevailing during priming of the CD8+ T cell responses; in LCMV-infected mice or in mice vaccinated with an OVA and costimulatory ligand–expressing poxvirus vector, the CD8+ T cells are primed over an extended period of 7 to 35 days under strong inflammatory conditions, whereas in the OVA peptide plus LPS–vaccinated OT-I–bearing mice, the duration and intensity of inflammatory conditions are limited.
We determined the functional status of the persisting OT-I cells by measuring their proliferative capacity and cytotoxic effector functions following antigenic reexposure. Figure Figure4A4A shows that in the rapamycin and 4-1BB-raptor, but not 4-1BB-GFP, conjugate–treated mice that were revaccinated with OVA peptide 35 days after priming, the OT-I cells accumulated at higher numbers in the spleen. To assess the cytotoxic effector functions of the OT-I cells, we used an in vivo cytotoxicity assay whereby OT-I cells isolated 30 days after vaccination with OVA peptide were mixed with CFSEhi OVA–presenting splenocytes and CFSElo control splenocytes, and equal numbers were injected into the recipient mice. The specific lysis of the OVA targets was determined 17 hours later. As shown in Figure Figure4B,4B, on a per-cell basis, OT-I cells derived from rapamycin-treated mice were less effective than OT-I cells derived from any of the other groups in killing their targets, showing that the rapamycin-generated memory OT-I cells exhibited reduced cytotoxic effector functions.
In separate experiments, we determined the in vitro cytolytic activity of memory OT-I cells isolated from mice treated earlier with either rapamycin or with 4-1BB-raptor or 4-1BB-GFP conjugates using a 51chromium (51Cr) release assay. As shown in Figure Figure4C,4C, when unfractionated splenocytes were used as effectors, the killing activity of memory OT-I cells isolated from rapamycin, but not from 4-1BB-raptor or 4-1BB-GFP conjugate–treated mice, was significantly reduced (4-1BB-raptor versus rapamycin, P = 0.028). However, when purified CD8+ T cells were used as effectors, the killing activity of the rapamycin-generated memory OT-I cells was not compromised (4-1BB-raptor versus rapamycin, P = 0.875). While consistent with the in vivo cytotoxicity experiments shown in Figure Figure4B,4B, the in vitro cytotoxicity experiments suggest that the rapamycin-imprinted cytotoxicity defect is not cell autonomous, but rather is mediated by soluble factors such as IL-10 or TGF-β, and/or by cells such as Foxp3+ Tregs or DCs. This is also consistent with the fact that we did not see differences in the expression of the cytotoxic effector molecules perforin, granzyme B, or CD107a in the OT-I cells derived from the rapamycin-treated mice or from mice in any of the other groups (data not shown). Overall, these observations suggest that while both rapamycin and 4-1BB conjugate treatment during priming can promote the development of memory OT-I cells capable of proliferating in response to antigenic stimulation, rapamycin treatment prevented the full acquisition of cytotoxic effector functions by the memory OT-I cells during subsequent antigenic exposure.
The OT-I CD8+ T cells that recognize a foreign epitope derived from OVA are a poor model for endogenous tumor antigens, most of which correspond to nonmutated products that elicit low-affinity T cells. We therefore sought to determine whether the 4-1BB-raptor conjugate is also capable of expanding memory responses directed at low-affinity epitopes encoded in nonmutated endogenous tumor antigens. To this end, we used transgenic Pmel-1 CD8+ T cells that recognize a determinant in the nonmutated gp100 tumor antigen expressed by B16 melanoma tumor cells (48). Pmel-1 cells bind the cognate epitope with much lower affinity than do OT-I cells (Nicholas Restifo, personal communication). When we revaccinated mice 30 days after vaccination with cognate peptide and LPS, the Pmel-1 cells expanded to higher numbers in mice treated with either 4-1BB-raptor conjugate or rapamycin compared with those observed in untreated mice or mice treated with 4-1BB-GFP conjugate (Figure (Figure5A).5A). Overall, reflecting their lower avidity, memory Pmel-1 cells expanded less effectively than OT-I cells (compare Figure Figure4A4A with Figure Figure55A).
To assess the antitumor effector functions of the memory Pmel-1 cells, 30 days after vaccination mice were challenged s.c. with B16/F10 tumor cells, and tumor growth was monitored. As shown in Figure Figure5B,5B, survival was significantly enhanced in mice treated with 4-1BB-raptor conjugate, but not with rapamycin or with 4-1BB-GFP conjugate, despite the fact that rapamycin treatment also enhanced the generation of Pmel-1 cells, as shown in Figure Figure5A.5A. We found that the memory Pmel-1 cells present 32 days after peptide vaccination and rapamycin treatment also exhibited reduced cytotoxic effector functions, as was seen in OT-I–transferred mice. However, the small number of Pmel-1 cells that accumulated in the spleen of the mouse after 32 days, which was about 100-fold less than that of the OT-I cells, precluded a firm conclusion (data not shown).
Transgenic CD8+ T cells like OT-I and Pmel-1 provide a useful tool to monitor the immunological parameters of memory development and function and have been all but exclusively used to show that pharmacological inhibition of intracellular mediators can enhance immunological memory and protective immunity (14, 19, 22, 24, 25). To assess the therapeutic potential of 4-1BB-raptor conjugates, we tested whether 4-1BB-raptor conjugates can enhance vaccine-induced memory responses against tumor-expressed endogenous antigens that are recognized by the resident, mostly naive T cell repertoire of the mouse. We first used a prophylactic experimental design whereby mice were vaccinated with GM-CSF–expressing irradiated B16 melanoma cells (GVAX) (49) and then challenged s.c. with parental B16 tumor cells. While a number of tumor antigens expressed in B16 tumor cells have been characterized, the antigens dominating the antitumor response induced by GVAX are not known. It is conceivable that the dominant antigens in the GVAX vaccine correspond to tumor-specific mutated products or to oncofetal antigens, and not to products like TRP2 or gp100 expressed in normal melanocytes. As shown in Figure Figure6A,6A, treatment with 4-1BB-raptor, but not 4-1BB-GFP or rapamycin, enhanced GVAX-induced antitumor immunity, underscoring the failure of rapamycin treatment to promote memory responses exhibiting antitumor effector functions (Figure (Figure5B).5B). Our antibody depletion experiments suggest that protective immunity elicited by the 4-1BB-raptor conjugate–generated memory response was mediated in part by CD8+ T cells (Supplemental Figure 3, P = 0.0052), whereas the contribution of CD4+ T cells was not clear, though probable (Supplemental Figure 3, P = 0.064). Given the documented role of CD4+ T cell help in the recall response of memory CD8+ T cells (10, 13, 50), future studies will be required to examine in depth the contribution of both adaptive and innate arms of the immune response in the 4-1BB-raptor conjugate–mediated enhancement of antitumor immunity.
Given the nontargeted nature of rapamycin administration and its documented immunosuppressive effects mediated by tolerizing DCs (27, 28) and promoting Treg differentiation (26), we tested whether the failure to elicit protective immunity shown in Figure Figure5B5B and Figure Figure6A6A correlated with reduced DC function and/or enhanced Treg accumulation in the rapamycin-treated mice. As shown in Figure Figure6B,6B, the allogeneic mixed lymphocyte reaction (alloMLR) activity of splenic DCs isolated from the rapamycin-treated mice was reduced compared with that of untreated mice, especially in the setting of GVAX vaccination. In contrast, treatment with 4-1BB-raptor conjugates with or without GVAX vaccination had no adverse effect on DC function. Why GVAX treatment exacerbated DC dysfunction remains to be determined. On the other hand, we saw no measurable differences in the number and Foxp3 expression of Tregs between the rapamycin-treated mice and mice from any other group shown in Figure Figure6A6A (data not shown).
To further test the therapeutic potential of enhancing vaccine-induced memory responses, we used therapeutic models in which mice were first implanted with tumor and subsequently subjected to vaccination and treatment with 4-1BB-raptor or 4-1BB-GFP conjugates. To simulate scenarios that mimic clinical settings, experimental conditions were chosen in which vaccination or treatment with aptamer-siRNA conjugates had limited or no antitumor effect. As shown in Figure Figure6C,6C, 4-1BB-raptor, but not 4-1BB-GFP, conjugates enhanced GVAX-induced antitumor immunity in mice implanted with B16 tumor cells 5 days prior to vaccination and conjugate administration. In this stringent experimental system, neither GVAX vaccination nor treatment with 4-1BB-raptor conjugates alone affected tumor growth. Similarly, 4-1BB-raptor, but not 4-1BB-GFP, conjugates enhanced vaccine-induced protective immunity in mice implanted with 4T1 breast carcinoma tumor cells 7 days before vaccination and conjugate administration (Figure (Figure6D).6D). Vaccination alone or treatment with 4-1BB-GFP or 4-1BB-raptor conjugates had a small antitumor effect that did not reach statistical significance. Since rapamycin was used in our study for comparison purposes only, it was not included in these experiments (Figure (Figure6,6, C and D) because we could not distinguish between its direct antitumor cytotoxic effects and its enhanced CD8+ T cell memory formation. Indeed, when administered to tumor-bearing mice, rapamycin enhances a protective immunity (51, 52) that in one study, using the OT-I model, was directly linked to the generation of enhanced CD8+ T cell memory responses, notably when used at a high, but not low, dose (24).
In this study, we show that aptamer-targeted siRNA-mediated inhibition of mTORC1 signaling in circulating CD8+ T cells leads to the generation of a potent CD8+ T cell memory response and enhances vaccine-induced protective immunity. Recent studies using transgenic models and/or targeting the model OVA antigen (14, 19, 22, 25) have provided proof of concept that potentiating memory with pharmacological agents can engender protective immunity against infectious diseases and cancer. Here, we demonstrate the therapeutic potential of promoting immunological memory using nontransgenic experimental models in which mice were implanted with poorly immunogenic tumor cells, C57BL/6-derived (H-2b–derived) B16/F10 melanoma tumor cells (53), and BALB/c-derived (H-2d–derived) 4T1 breast carcinoma (54, 55) tumor cells and then vaccinated with syngeneic irradiated tumor cells presenting endogenous tumor antigens to prime cognate T cells present in the repertoire of the mouse. We show that in both prophylactic (Figure (Figure6A)6A) and therapeutic (Figure (Figure6,6, C and D) settings, the 4-1BB aptamer-raptor siRNA conjugate potentiated vaccine-induced protective immunity. These findings underscore the importance of memory responses in protective immunity and show that the approach used in this study could be therapeutically useful.
In this study, we also tested the hypothesis that aptamer-targeted siRNA inhibition of intracellular mediators in vaccine-induced CD8+ T cells will be superior to that of pharmacological agents in terms of enhancing memory CD8+ T cell development and protective antitumor immunity in mice. Cell targeting of RNAi offers important advantages compared with the nontargeted administration of pharmacological agents. Reflecting the broad distribution of their targets, pharmacological agents like rapamycin often exhibit pleiotropic effects that can have undesirable consequences such as promoting the development of immunosuppressive Foxp3+ Tregs and tolerogenic DCs (30, 31). Rapamycin can also inhibit both mTORC1 and mTORC2 (44, 45), as is also shown in Figure Figure2B.2B. Given the roles of mTORC2 in cell survival, cytoskeletal organization, and glucose homeostasis (56, 57), systemic inhibition of mTORC2 could be harmful, especially in predisposed individuals.
The underlying premise of this study was that targeted inhibition of mTORC1 in CD8+ T cells using aptamer-targeted RNAi is superior to across-the-board mTOR inhibition with rapamycin in memory formation and protective antitumor immunity. We show here that while both rapamycin and the 4-1BB aptamer-raptor siRNA conjugate elicited a comparable memory response in terms of numbers (Figure (Figure3,3, B and C) and proliferative capacity (Figure (Figure4A4A and Figure Figure5A),5A), the rapamycin-generated memory CD8+ T cells exhibited reduced cytotoxic effector functions (Figure (Figure4,4, B and C), DCs isolated from the rapamycin-treated mice exhibited reduced MLR activity (Figure (Figure6B),6B), and the rapamycin-treated mice failed to control tumor growth (Figure (Figure5B5B and Figure Figure6A).6A). In contrast, we found that systemic administration of 4-1BB aptamer-raptor siRNA conjugates, which downregulated mTORC1, but not mTORC2, in CD8+ T cells while sparing host cells (Figure (Figure2,2, A and B), did not adversely affect the cytotoxic effector function of the developing memory cells (Figure (Figure4,4, B and C) or the alloMLR activity of resident DCs (Figure (Figure6B)6B) and controlled tumor growth (Figure (Figure5B5B and Figure Figure66A).
How rapamycin imprints a cytotoxic defect in the developing memory CD8+ T cells remains to be determined. Our observation that the memory OT-I cells exhibited reduced cytolytic activity in vitro when total splenocytes, but not purified CD8+ T cells, were used (Figure (Figure4C)4C) suggests that the cytotoxic defect is not cell intrinsic, but rather is mediated in trans by either cellular subsets such as Tregs or DCs and/or soluble factors such as IL-10 or TGF-β. The failure to find significant alterations in the level of cytotoxic effector molecules including CD107a, perforin, or granzyme B in the rapamycin-generated memory OT-I cells (data not shown) and the reduced alloMLR activity of resident DCs isolated from rapamycin-treated mice (Figure (Figure6B)6B) are consistent with this possibility. Though our initial examination has not uncovered differences in the Treg population (data not shown), one possibility is that the rapamycin-compromised DCs drive the differentiation of (tumor) antigen-specific Tregs that compromise the cytolytic functions of the reactivated memory CD8+ T cells. Notwithstanding, given the inherent limitations of in vivo cytotoxicity and alloMLR assays, the biological significance of the differences measured in such experimental systems using arbitrarily set experimental conditions is not clear. Future studies will be required to determined whether the reduced cytotoxicity of the memory cells (Figure (Figure4,4, B and C) and/or the reduced MLR activity of the DCs (Figure (Figure6B)6B) was responsible for the lack of protective immunity in the rapamycin-treated mice (Figure (Figure5B5B and Figure Figure66A).
A second advantage of cell targeting is that it reduces the effective dose of siRNA needed to elicit the desired biological effect, thereby reducing the risk of nonspecific immune activation (58) and enhancing the cost-effectiveness of the therapy. Underscoring the efficiency of aptamer-targeted siRNA delivery, we found that two consecutive injections of 1 nmole of 4-BB-raptor conjugate downregulated mTORC1 activity in over 60% of the circulating antigen-activated CD8+ T cells (Figure (Figure2A),2A), resulting in an enhanced memory development that was not less effective than that seen with rapamycin (Figures (Figures33–5), while three injections of 0.25 nmoles engendered protective antitumor immunity in both prophylactic and therapeutic settings (Figure (Figure6).6). This compares favorably with nontargeted polymer-formulated siRNA delivery requiring 10- to 50-fold higher doses of siRNA to inhibit tumor growth (59, 60).
In this study, the raptor siRNA was targeted to antigen-activated CD8+ T cells, peptide-stimulated transgenic OT-I or Pmel-1 cells, or GVAX vaccine–stimulated tumor-specific T cells using an agonistic dimeric 4-1BB–binding aptamer (39). Thus, the 4-1BB aptamer might have served a dual role: (a) to target the siRNA to 4-1BB–expressing CD8+ T cells, and (b) to promote their survival and potential for differentiating into memory cells. Nonetheless, the agonistic 4-1BB aptamer did not exert a measurable contribution to memory differentiation since 4-1BB aptamer-GFP siRNA conjugates used in every experiment did not affect any of the memory parameters tested. While 4-1BB expression is highly restricted, it is also upregulated on activated conventional CD4+ T cells, Foxp3+ Tregs, activated NK cells, CD40-stimulated mature DCs (reviewed in ref. 46), and on proliferating ECs (47). Given that 4-1BB signaling (46, 47) and mTOR inhibition (30, 31) can exert immune-potentiating or immunosuppressive effects, the positive and negative contributions of 4-1BB–targeted mTOR inhibition in other cells to the protective antitumor immune responses shown in Figure Figure5B5B and Figure Figure6,6, A, C, and D remain to be determined. It is, however, clear that on balance, targeting raptor siRNA to 4-1BB–expressing cells in vivo enhanced CD8+ T cell memory and protective antitumor immunity (Figures (Figures22–6). Arguably, targeting receptors expressed more exclusively on activated CD8+ T cells will further enhance the specificity and therapeutic potential of this approach.
A bivalent murine 4-1BB–binding aptamer (39) was transcribed in vitro using the Durascribe T7 transcription kit (Illumina). The DNA template gggggaattctaatacgactcactataGGGCGGGAGAGAGGAAGAGGGATGGGCGACCGAACGTGCCCTTCAAAGCCGTTCACTAACCAGTGGCATAACCCAGAGGTCGATAGTACTGGATCCCGCCCTCCTGCGGCCGAGAGAGGAAGAGGGATGGGCGACCGAACGTGCCCTTCAAAGCCGTTCACTAACCAGTGGCATAACCCAGAGGTCGATAGTACTGGATCGGCCGCTCCCCTCGGGATCTCTTCCAAAA encoded a T7 promoter (lower case) and two aptamers (bold) separated by a single-stranded linker (italicized) fused with the passenger strand of an siRNA (underlined) against raptor mRNA. Raptor mRNA targeting siRNA was characterized as described previously (61). Briefly, candidate sequences were predicted using HPC Dispatcher (City of Hope, Biomedical Informatics Core, Duarte, California, USA), siRNA scales (Department of Human Genetics, University of Utah, Salt Lake City, Utah, USA), and siDESIGN (Dharmacon, Thermo Fisher Scientific) software. Overlapping predictions featuring a low melting temperature (Tm) were selected and screened for specific activity as 4-1BB aptamer conjugates using siCheck assay in transiently transfected HEK293T cells. The aptamer passenger strand fusion transcript was purified by polyacrylamide gel electrophoresis and annealed to a chemically synthesized guide stand. The siCheck assay was used to confirm that target knockdown by the aptamer-siRNA conjugate was comparable to that of free duplex siRNA.
Splenocytes were isolated from CD45.1+ OT-I mice, and rbcs were lysed with ammonium-chloride-potassium (ACK) solution (Gibco, Life Technologies). Mononuclear cells were washed and incubated overnight with 1 nM of the chicken OVA MHC class I SIINFEKL peptide (AnaSpec) at a density of 106 cells per milliliter in complete RPMI-1640 media supplemented with 10% FBS, essential and nonessential amino acids, Na pyruvate (Gibco, Life Technologies), and 2 ng/ml murine IL-12 (R&D Systems). After overnight incubation, cells were washed extensively to remove excess peptide and used in the subsequent analysis. For quantitative PCR (qPCR), cells were plated in triplicate onto 96-well round-bottom plates (105 cells/well) and incubated with 0.8 μM aptamer-siRNA conjugates three times every 8 hours. Thirty-six hours after the last treatment, cells were lysed with Buffer RLT, and RNA was isolated using an RNeasy kit (both from QIAGEN). RNA was quantified using an Agilent 2100 Bioanalyzer (Agilent Technologies). cDNA synthesis was performed using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). cDNA equivalents of 25 ng of mRNA were used per reaction in a TaqMan qPCR assay using the Step One qPCR machine (Applied Biosystems), with primer sets corresponding to the gene of interest or housekeeping products.
Cells were plated onto 6-well plates (106 cells/5 ml) and incubated with 0.8 μM aptamer-siRNA conjugates three times every 8 hours or with rapamycin at 0.5 μg/5 ml. Seventy-two hours after plating, cells were washed and replated (106 cells/5 ml) and incubated in complete RPMI-1640 media with the addition of 5 ng/ml recombinant IL-7 (R&D Systems). Every 3 to 5 days, cells were collected, and 106 cells were plated per well. On day 25, the cells were harvested, washed, and plated onto 96-well round-bottom plates, with 10 replicates per condition. Five wells for each condition were treated with 0.01 nM SIINFEKL peptide overnight, and five other wells were treated with control SDYEGRLI influenza NP (50–57) peptide. Eighteen hours later, 1 μCi H3-thymidine was added to each well. Six hours later, cells were harvested, and tritium incorporation was measured using a Wallac MicroBeta scintillation counter (PerkinElmer).
Splenocytes from CD45.1 OT-I mice were isolated and enriched for CD8+ T cells using Ly-2 microbeads (Miltenyi Biotec) to greater than 90% purity confirmed by flow cytometry. CD8+ cells (0.5 × 106) were transferred into CD45.2 C57Bl/6 mice by tail-vein injection. Forty-eight hours after transfer, the animals were injected i.v. with 10 μg SIINFEKL peptide, the chicken OVA dominant class I epitope, and 10 μg of E. coli LPS (Sigma-Aldrich) in 100 μl PBS. Starting 6 hours after peptide administration, the animals received an i.v. injection of 0.25 nmoles of aptamer-siRNA conjugate in 100 μl PBS twice daily, or 1.5 μg rapamycin administered i.p. in 200 μl PBS immediately after peptide administration, also twice daily.
To measure mTORC1 activity, 36 hours following the second aptamer-siRNA conjugate or rapamycin injection, splenocytes were isolated, fixed in 4% PFA for 10 minutes in RT, and rbcs were lysed with a 0.1% Triton X-100 PBS solution (37°C for 20 minutes). An equal volume of ice-cold wash buffer was added, and the suspension was centrifuged (300 g for 6 minutes at 4°C). Cells were dehydrated with 80% methanol (prechilled to –20°C) and incubated on ice for 10 minutes. Then, cells were washed with ice-cold PBS, blocked with ice-cold washing buffer (4°C for 30 minutes), and incubated with CD45.1 Pacific Blue (BioLegend) and p-S6-Alexa488 (Cell Signaling Technology) mAbs. To measure mTORC1 and mTORC2 activity in the same preparation, splenocytes were incubated with anti-CD8 and anti-CD45.1 mAbs for 15 minutes at 4°C, washed, and then fixed and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer’s instructions. After permeabilization, cells were stained with anti–phospho-S235/236-S6/Alexa488 together with anti–phospho-S473-AKT/Alexa647 mAbs (Cell Signaling Technology) and analyzed by flow cytometry.
Splenocytes from C57BL/6 mice were incubated with 100 nM of either the OVA SIINFEKL peptide or the SDYEGRLI influenza NP (50–57) peptide MHC class I–restricted (control) peptide (AnaSpec) in serum-free RPMI-1640 media for 1 hour at 37°C, washed and incubated with either 5 μM or 0.5 μM of CFSE solution in PBS and 2% FBS at room temperature for 30 minutes, and then washed extensively in cold PBS. On day 30, splenocytes were isolated from OT-I–transferred mice treated with rapamycin or with aptamer-siRNA conjugates, and the proportion of OT-I cells was determined in each mouse by CD8 and CD45.1 costaining. CFSE-labeled OVA and control peptide–loaded splenocytes (5 × 106) were mixed with splenocytes containing 5 × 104 OT-I cells and injected i.v. into C57Bl/6 mice. Seventeen hours after transfer, splenocytes were isolated, and specific lysis of the OVA peptide–loaded splenocytes was measured as the ratio of remaining SIINFEKL-loaded (CFSEhi)/control peptide–loaded (CFSElo) targets.
OT-I–bearing mice were vaccinated with OVA peptide and LPS and treated with rapamycin or administered either 4-1BB-GFP or 4-1BB-raptor conjugates, as described above. Fourteen days after peptide administration, the mice were injected again with peptide and LPS, sacrificed 72 hours later, and their splenocytes were isolated. EL4 tumor cells (H-2b) were incubated with 1 μM OVA peptide (SIINFEKL), washed and incubated with 0.1 μCi of 51Cr. Cells were plated onto a U-bottom 96-well plate with 105 cells per well. Total splenocytes or purified CD8 cells from OT-I–bearing mice containing 104 OT-I cells were added to wells containing target cells, with 6 wells per condition, and incubated at 37°C. Sixteen hours later, cell supernatant from each well was harvested, and 51Cr release was measured using a gamma counter. Spontaneous release and nonspecific lysis were determined by coincubation of labeled targets with splenocytes from the control mice bearing no OT-I. Maximal release was measured by adding 0.01% Triton-X 100 to completely lyse all targets.
Female C57Bl/6 mice (H-2b) were injected s.c. with 106 irradiated (60 Gy) GM-CSF–secreting B16 cells (GVAX). Twenty-four hours later, GVAX-immunized or untreated mice were injected twice daily with aptamer-siRNA conjugate or rapamycin, and 24 hours after the second injection CD11c+ DCs were purified from the spleen using CD11c MicroBeads (Miltenyi Biotec) and irradiated (20 Gy) prior to the MLR assay. In parallel, splenocytes were isolated from female BALB/c mice (H-2d) mice, and 4 × 105 cells were mixed with 104 CD11c-enriched CD57BL/6-derived splenocytes and plated onto 96-well round-bottom plates in quintuplicate in complete RPMI-1640 media. Four days after plating, 1 μCi 3H-thymidine was added to each well, and tritiated thymidine incorporation was measured 6 hours later using a Wallac MicroBeta scintillation counter (PerkinElmer).
Splenocytes (106 cells) from Pmel-1 transgenic Thy1.1 C57BL/6 mice were transferred to congenic CD90.2 (Ty1.2+) C57BL/6 mice by i.v. injection. Forty-eight hours later, 100 μg of KVPRNQDWL human gp100 (25–33) peptide (AnaSpec) and 10 μg of E. coli LPS were injected i.v., and mice were treated with rapamycin or aptamer-siRNA conjugates as described above. Proliferation and phenotypic analysis was performed as described for OT-I cells, except that cells were costained for CD90.1 (Thy 1.1) instead of CD45.1. To determine the antitumor effector functions of the Pmel-1 cells, 30 days after peptide administration mice were challenged with 105 B16/F10 melanoma tumor cells injected s.c., and tumor growth was monitored daily. Mice were sacrificed when their tumor diameter reached 12 mm.
C57Bl/6 mice were immunized with 106 irradiated (60 Gy) GVAX, and 24 hours later mice were treated with 0.25 nmoles of aptamer-siRNA conjugates administered i.v. or with 1.5 μg of rapamycin administered i.p. three times daily. On day 50, the mice were challenged s.c. with 106 B16/F10 tumor cells, and tumor growth was monitored daily. The mice were sacrificed when their tumor diameter reached 12 mm. Alternatively, mice were injected s.c. with 105 B16/F10 cells and immunized 5 days later with GVAX. Twenty-four hours later, mice were treated with 0.25 nmoles of aptamer-siRNA conjugates administered i.v. or with 1.5 μg of rapamycin administered i.p., and this was repeated on days 9 and 12.
BALB/c mice were injected s.c. with 1.0 × 104 4T1 tumor cells and immunized with a mixture of irradiated B7-1 and MHC class II–expressing 4T1 cells (62) and then treated with aptamer-siRNA conjugates as described above.
Unpaired, 2-tailed Student’s t tests were performed between individual treatment groups using GraphPad Prism, version 5.0 (GraphPad Software). P values less than or equal to 0.05 were considered statistically significant. For survival curves, a log-rank test was performed to determine statistical significance for differences in survival times between treatment groups, with P values below 0.05 being considered statistically significant.
The facilities at the University of Miami Division of Veterinary Resources are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care and the USDA. An OLAW assurance is on file, ensuring that humane animal care and use practices, as outlined in the Guide for the Care and Use of Laboratory Animals (NIH publication no. 85-23. Revised 1985) are followed. All mice were maintained according to the guidelines established by the US Department of Agriculture and the American Association for Accreditation of Laboratory Animal Care (AAALAC). This project was approved by the IACUC of the University of Miami School of Medicine.
We thank Randall Brenneman for his assistance in writing the manuscript. Funding was provided by a bequest from the Dodson Estate and the Sylvester Comprehensive Cancer Center (Miller School of Medicine, University of Miami, Miami, Florida, USA) and by a grant (KG090348) from the Susan G. Komen for the Cure of Breast Cancer Foundation. A. Berezhnoy was responsible for designing and testing the aptamer-siRNA conjugates, in vivo memory studies, and tumor immunotherapy experiments. A. Levay was responsible for generating aptamer-siRNA conjugates and murine immunotherapy studies. I. Castro and T.R. Malek were responsible for designing the OT-I experiments and assisted in the writing of the manuscript. E. Gilboa oversaw the project and was responsible for writing the manuscript.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J Clin Invest. 2014;124(1):188–197. doi:10.1172/JCI69856.