|Home | About | Journals | Submit | Contact Us | Français|
We previously established a model to study CD8+ T cell (TCD8)-based adoptive immunotherapy of cancer using line SV11 mice that develop choroid plexus tumors in the brain due to transgenic expression of Simian Virus 40 large T antigen (Tag). These mice are tolerant to the three dominant TCD8-recognized Tag epitopes I, II/III and IV. However, adoptive transfer of spleen cells from naïve C57BL/6 (B6) mice prolongs SV11 survival following TCD8 priming against the endogenous Tag epitope IV. In addition, survival of SV11 mice is dramatically increased following transfer of lymphocytes from Tag-immune B6 mice. In the current study, we compared the kinetics and magnitude of Tag-specific TCD8 accumulation at the tumor site following adoptive transfer with a high dose of either Tag-immune or naïve donor cells or decreasing doses of Tag-immune lymphocytes. Following adoptive transfer of Tag-immune cells, epitope I and IV-specific TCD8 accumulated to high levels in the brain of SV11 mice, peaking at 5 to 7 days, while epitope IV-specific TCD8 derived from naïve donors required three weeks to achieve peak levels. A similar delay in the peak of epitope IV-specific TCD8 accumulation was observed when ten-fold fewer Tag-immune donor cells were administered, reducing control of tumor progression. These results suggest that efficient and prolonged control of established autochthonous tumors is associated with high-level early accumulation of adoptively transferred T cells. We also provide evidence that although multiple specificities are represented in the Tag immune donor lymphocytes, epitope IV-specific donor TCD8 play a predominant role in control of tumor growth.
Tumor immunotherapy utilizes the exquisite specificity of CD8+ T lymphocytes (TCD8) to eliminate the growth of tumors. TCD8 recognize either cross-presented or directly-presented tumor antigen derived from both mutated [1, 2] and over-expressed [3, 4] self-proteins or viral oncoproteins [5–7] presented in the context of MHC class I molecules. Following antigen stimulation, naïve TCD8 follow a program of differentiation and proliferation into effector cells and long-lived antigen-specific memory cells . Since TCD8 present at each stage of differentiation are qualitatively different, current challenges for tumor immunotherapy are to identify which characteristics of the T cell response are associated with long-term control of tumor progression .
Current clinical protocols for adoptive cell transfer therapy involve the ex vivo stimulation and expansion of autologous tumor-specific T cells followed by re-infusion into the patient . Most often this is accomplished via repeated in vitro stimulation of clonotypic T cells to achieve large numbers of highly-activated T cells. Although adoptive cell transfer was initially met with limited success , the combination of recipient lymphodepletion and adoptive cell transfer has resulted in tumor regression in both murine studies [12–17] and clinical trials . However, recent studies have suggested that the stage of effector T cell differentiation may be important for achieving successful adoptive immunotherapy due to constraints on T cell survival and proliferative capacity [19, 20].
It is well established that TCD8 can control tumor growth in vivo . However, questions remain regarding the requirements for effective tumor control following adoptive cell transfer. In the current study, we investigated the basis for the successful control of advanced stage autochthonous tumors using the SV11 line of SV40 Tag transgenic mice, which develop spontaneous choroid plexus tumors due to expression of full-length Tag from the SV40 enhancer/promoter . Choroid plexus papillomas are first evident in SV11 mice by ~36 days of age, and tumors grow progressively until mice succumb to tumor burden at a mean age of 104 days . Tag is also expressed in the thymus, resulting in central tolerance toward the immunodominant Tag epitopes . Therefore, potentially tumor-reactive TCD8 are deleted from the repertoire.
In non-transgenic C57BL/6 (B6) mice, Tag is normally the target of a strong TCD8-mediated immune response against four epitopes (I, II/III, IV and V) . Epitope IV, an H-2Kb-restricted epitope, is the most immunodominant. Epitopes I, II/III and V are H-2Db-restricted, with epitopes I and II/III being subdominant to epitope IV. Epitope V is classified as immunorecessive since TCD8 specific for this epitope are only induced following immunization with Tag variants lacking the dominant epitopes . Due to the lack of endogenous TCD8 that can respond to the immunodominant Tag epitopes in SV11 mice, we previously established a model of adoptive cell transfer [13, 23]. In this model, the T antigen is expressed as a transforming nonstructural self antigen in SV11 mice, but represents a non-self antigen to the donor T cells derived from normal B6 mice. However, donor T cells remain subject to the potential effects of peripheral tolerance and the immunosuppressive tumor environment following adoptive transfer into SV11 mice [23, 25]. We previously found that reconstitution of irradiated 80 day-old SV11 mice bearing advanced tumors with Tag-immune B6 splenocytes resulted in a dramatic increase in their lifespan (mean of 277 days) and led to the accumulation of epitope I, II/III and IV-specific TCD8 in the brain . In contrast, only epitope IV-specific TCD8 are initially primed by the endogenous Tag following adoptive transfer of naïve B6 donor splenocytes, resulting in a more modest (mean of 170 days), yet significant increase in lifespan. These findings demonstrated that activated Tag-specific TCD8 are more effective at controlling tumor progression in the SV11 mouse model than naïve donor lymphocytes, despite the presence of Tag epitope IV-specific TCD8 at the tumor site in each case.
The goal of the present study was to elucidate whether the kinetics of TCD8 accumulation and persistence at the tumor site is associated with prolonged tumor control in SV11 mice induced by transfer of Tag-immune lymphocytes. The results demonstrate that Tag-specific TCD8 from immune donors rapidly accumulated at the tumor site and exhibited effector function earlier than following transfer of naïve donor cells. This effect was dose dependent and may be enhanced by the presence of Tag-specific TCD8 targeting multiple epitopes within Tag. Lastly, Tag-specific TCD8 persisted in the tumorigenic environment for an extended period of time, consistent with long-term surveillance against tumor growth in SV11 mice.
C57BL/6 (B6) (H-2b) and B6.SJL-H2 (H-2b) mice were purchased from The Jackson Laboratory and maintained at the animal facility of the Milton S. Hershey Medical Center. SV11 mice (H-2b) express full-length Tag from the SV40 promoter/enhancer  and are maintained by breeding Tag-positive males with B6 females. SV11 positive mice are identified by PCR amplification of the transgene as previously described . Transgene positive (SV11) and transgene negative (B6) littermates were used at 80 days of age unless otherwise specified. All experimental procedures were performed in accordance with guidelines established by the institutional animal care and use committee of the Pennsylvania State University College of Medicine and complied with federal guidelines.
B6/WT-19 cells are derived from SV40 transformed primary mouse embryonic fibroblasts and express wild type Tag . B6/15Bb cells (B6/Tag-IV) are Tag transformed mouse embryonic fibroblasts that express a mutant form of Tag in which epitope I (amino acids 206–215), II/III (amino acids 223–231) and V (amino acids 489–497) are deleted . All cell lines were maintained in DMEM supplemented with 100 U/mL penicillin, 100 mg/mL streptomycin, 100 mg/mL kanamycin, 2 mM L-glutamine, 10 mM HEPES, 0.075% (w/v) NaHCO3 and 5–10% FBS. Synthetic peptides corresponding to the SV40 Tag epitope I (SAINNYAQKL; peptide I [7, 29], a variant of epitope IV in which the carboxy-terminal C is replaced with L (VVYDFLKL; peptide IV) [7, 28] and the HSV gB epitope gB498–505 (SSIEFARL; peptide gB)  were synthesized in the Macromolecular Core Facility of the Milton S. Hershey Medical Center by F-moc chemistry using an automated peptide synthesizer (9050 MiliGen PepSynthesizer; Milipore). Peptides were solubilized in DMSO and diluted to 1 mM in RPMI 1640 medium.
To generate Tag-immune donor lymphocytes, naïve B6 mice were immunized i.p. with 3–5×107 of the indicated cell line in 0.5 mL of HBSS. On day 10 post-immunization, spleens and LN (cervical, brachial and mesenteric) were processed to single-cell suspensions, and splenocytes were depleted of RBCs using Tris NH4Cl as described previously . RBC depleted lymphocytes derived from the spleens of either naïve or immunized mice were analyzed by flow cytometry prior to transfer to determine the percentage of Tag epitope-specific TCD8. Lymphocytes from gender-matched mice were resuspended at the indicated dose (typically 5×107 cells) in HBSS, filtered and injected i.v. into the tail vein in 0.2 mL HBSS. One day prior to adoptive transfer, SV11 mice and B6 littermates received 400 rads of whole body gamma irradiation from a 60Co source (Gammacell 220; MDS Nordion, Ottawa, Canada). At varying times post-adoptive transfer SV11 mice were anesthetized by i.p. injection of ~100 mg/kg sodium pentobarbital (Nembutal®) diluted in 10% ethanol and euthanized by exsanguination. Spleen and brain were removed and placed in cold RPMI 1640 medium, and were processed as described previously .
MHC class I tetramers corresponding to the H-2Db/Tag epitope I (Db/I), H-2Kb/Tag epitope IV (Kb/IV) and H-2Kb/HSV epitope gB (Kb/gB) were prepared as previously described . Purified anti-CD16/32 (Fc receptor block) was purchased from BD Pharmingen. The following antibodies were purchased from eBioscience: Pe-Cy5-labeled anti-mouse CD8a (clone 53-6-7), FITC-labeled anti-mouse CD44 (clone IM7), FITC-labeled anti-mouse L-selection (clone MEL-14), and PE-labeled anti-mouse CD45.1 (clone A20). For ex vivo characterization of TCD8, lymphocytes isolated from the spleen and brain were resuspended at 2×107/mL in FACS buffer (PBS with 2%FBS/0.01%NaN3) and incubated in the presence of anti-CD16/CD32 to block Fc receptors and unconjugated streptavidin for 30 min on ice. Following a wash in FACS buffer, cells were resuspended in a mixture of the indicated fluorochrome-conjugated mAb (diluted 1:100) and PE-conjugated tetramers (diluted 1:200) for 1 hr in the dark on ice. Cells were washed 3 times to remove unbound antibody and fixed with 2% paraformaldehyde/PBS and analyzed using a FACScan or FACSCalibur flow cytometer (BD Biosciences). 100,000 events were typically collected and analyzed using FlowJo software (Tree Star). The percentage of TCD8 that stained specifically for Db/I and Kb/IV was determined by subtracting the percentage of TCD8 which stained positively for the control tetramer.
For staining of intracellular IFN-γ, lymphocytes isolated from the spleen and brain were processed as described and incubated in 0.2 mL of RPMI 1640/10% FBS in U-bottom 96-well plates with 1 µM of the indicated peptides plus 1 µg/mL Brefeldin A (Sigma-Aldrich) for 5–6 hr at 37°C/5%CO2. TCD8 were stained for IFN-γ using the Cytofix/Cytoperm kit (BD Pharmingen) in accordance with the manufacturer’s instructions as described  and analyzed by flow cytometry. The percentage of cells that stained specifically for IFN-γ following stimulation with peptide I or IV was determined by subtracting the percentage of TCD8 which stained for IFN-γ in the presence of the control gB peptide.
In vivo cytotoxicity assay was performed as described previously . Briefly, target cells were prepared from gender-matched B6.SJL (CD45.1+) splenocytes. RBC-depleted splenocytes were incubated with the indicated peptide (1 µM) in RPMI 1640 medium/10% FBS at 37° C for 60 min. and then washed 3 times to remove excess peptide. Targets were then differentially labeled with CFSE (5 µM for peptide I; 0.5 µM for peptide IV; 0.25µM for peptide gB) for 10 min. at 37° C in PBS/0.1% BSA and washed 3 times. A total of 6×106 (2×106 of each) targets were injected i.v. into the tail vein in 0.2 mL HBSS. The elimination of CD45.1+ targets was monitored the following day by staining splenocytes for CD45.1 and analysis by flow cytometry. The formula used to determine the percentage of specific killing was: % lysis = [1-(ratio unprimed/ratio primed)×100], where ratio = (% of CFSE low cells/percent of CFSE high cells).
SV11 mice were monitored for the development of lethargy and hydrocephalus which are indicative of end-stage choroid plexus tumor burden. Mice were euthanized in a timely manner following development of these symptoms. Survival curves were generated by the Kaplan-Meier method using GraphPad Prism software (GraphPad Software). Significance was determined by single-factor ANOVA and validated using the log-rank test. p values < 0.05 are considered significant.
Prior to attempting to elucidate the mechanism responsible for increased control of tumor burden by immune cells, it was necessary to fully characterize the Tag-immune donor lymphocytes. B6 mice were immunized with B6/WT-19 cells expressing wild type Tag. Ten days post-immunization, the frequency of epitope I and epitope IV-specific TCD8 was determined by MHC tetramer analysis. Directly ex vivo, epitope I and IV-specific TCD8 routinely represented ~2.6% and ~8.0% of total TCD8, respectively (Fig. 1a). A portion of these Tag-specific TCD8 harvested at the peak of the immune response  produced IFN-γ following ex vivo stimulation with peptide (Fig. 1b). Both the epitope I and IV-specific TCD8 exhibited an effector phenotype as evidenced by high levels of CD44 and low levels of CD62L (Fig. 1c), although a subset of each remained CD62Lhi. These findings demonstrate that following exogenous immunization of B6 mice with B6/WT-19 cells, the Tag-immune donor population consists of activated and functional TCD8 including those specific for epitopes I and IV.
The Tag-immune donor cell population provides an increased frequency and a more differentiated phenotype of Tag-specific TCD8 as compared to the naïve lymphocyte repertoire. We therefore hypothesized that increased control of tumor burden in SV11 mice may be associated with more rapid accumulation of Tag-specific TCD8 at the tumor site, since initial T cell activation is not required. To investigate this possibility, irradiated 80 day-old SV11 mice received 5×107 Tag-immune lymphocytes, and the frequency of epitope I- and epitope IV-specific TCD8 recovered at various time points was assessed by MHC tetramer analysis. A small population of both epitope I and epitope IV-specific TCD8 was first detected in the spleen at 4 days post-adoptive transfer (Fig. 2a). Few total TCD8 were present in the brain at this time point, although a proportion were specific for epitope I (3.5%) or epitope IV (8.9%). On day 4 post-adoptive transfer a subset of the Tag-specific TCD8 produced IFN-γ upon ex vivo stimulation with peptide. An in vivo cytotoxicity assay demonstrated that epitope I and IV peptide-pulsed CFSE-labeled target cells were eliminated from mice on day 4 post-adoptive transfer (Fig. 2b). In vivo cytolytic activity was not detected in the spleen on day 4 post-adoptive transfer of naïve donor cells (data not shown). Thus, Tag-specific TCD8 are rapidly detected in SV11 mice following transfer of Tag immune cells.
To further characterize the kinetic accumulation of the Tag-specific TCD8 from Tag-immune and naïve donor cells, the frequency of epitope I and epitope IV-specific TCD8 was monitored over 3 weeks in SV11 mice. Following adoptive transfer of Tag-immune donor cells, the frequency of epitope I and IV-specific TCD8 increased in both the spleen (Fig. 3a) and brain (Fig. 3b) over the first week following transfer. Both epitope I and IV-specific TCD8 peaked in the spleen on day 7 post-adoptive transfer at a frequency of ~13% of the TCD8 population. The epitope I-specific TCD8 frequency began to decline by 2 weeks post-transfer while the epitope IV-specific TCD8 frequency was maintained through 2 weeks and declined at 3 weeks post-transfer. In contrast, a high percentage of epitope IV-specific TCD8 accumulated early in the brain (~25%) and persisted at this high level over 5 weeks post-adoptive transfer. Epitope I-specific TCD8 were also detected in the brain, but were delayed in their accumulation and never reached the high levels observed for epitope IV-specific T cells (only ~12%), despite reaching similar levels in the spleen. At each day analyzed the frequency of epitope I-specific TCD8 in the spleen and brain was comparable and began to decline after 10 days post-adoptive transfer.
Following adoptive transfer of naïve donor cells, a small population of epitope IV-specific TCD8 were first detected in the spleen (Fig. 3a) and brain (Fig. 3b) at 7 days post-adoptive transfer. Previous studies demonstrated that these Tag-specific TCD8 produce IFN-γ . Only epitope IV-specific TCD8 are detected following adoptive transfer of naïve donor cells, suggesting that epitope I-specific TCD8 are insufficiently primed by the endogenous Tag expressed in SV11 mice. The frequency of splenic epitope IV-specific TCD8 derived from naïve donors remained below that achieved following transfer of immune donor cells, peaking at day 14 (~5%). In the brain, epitope IV-specific TCD8 frequency gradually increased over the 3 week period, reaching levels similar to that derived from Tag-immune donors only at 3 weeks post-adoptive transfer. We note that in contrast to the previous study , a lower frequency of epitope IV-specific TCD8 was observed in the spleen and brain at day 10 post adoptive transfer, although the frequency detected in that study was variable (15–56% of CD8+ cells in the brain; Fig. 3B of ). In the current study, multiple experiments using a larger number of mice revealed a reduced initial frequency of epitope IV-specific TCD8 following transfer of naïve B6 cells. Despite the initial differences in epitope IV-specific TCD8 numbers at early timepoints, the frequency of epitope IV-specific TCD8 derived from either Tag-immune donors or naïve donors was similar in the spleen (4.6% vs. 3.4%) and brain (25.5% vs. 22.6%) at the later time point of 40 days post-adoptive transfer (120 days of age). These kinetic data demonstrate that Tag-specific TCD8 derived from the Tag-immune donor repertoire traffic and accumulate to a high magnitude with increased kinetics in both the spleen and tumor site compared with Tag-specific TCD8 derived from the naïve donor repertoire. In fact, greater than 3 weeks was required for epitope IV-specific TCD8 from naïve donor mice to achieve similar frequencies to that observed with Tag-immune donors. In addition, epitope I-specific TCD8 were detected at the tumor site only when Tag-immune donor cells were transferred.
Human and murine studies have indicated that the persistence of tumor-specific TCD8 is associated with long-term control of tumor burden [34–37]. We found previously that although SV11 mice show a significant increase in their survival following adoptive cell therapy with Tag immune donor cells (mean of 277 days), all mice eventually succumbed to tumor burden . To determine whether Tag-specific TCD8 persisted long term in SV11 mice and whether the cells retained functionality in the tumor-prone environment, the frequency and functionality of epitope I and IV-specific TCD8 was assessed in SV11 mice at 90, 120, 150, 190 and 257 days of age following adoptive transfer of Tag-immune cells at 80 days of age (Table 1). At 90, 120 and 150 days of age, mice exhibited no symptoms of tumor burden and no tumor was observed upon gross examination. At 190 days of age, one of two mice examined had a detectable tumor at the time of analysis, while two of four mice examined at 257 days of age had tumors. Tetramer analysis demonstrated that the frequency of both epitope I and epitope IV-specific TCD8 decreased in the spleen over the lifespan of the mice. However, a small but stable population of epitope I-specific and a significant population of epitope IV-specific TCD8 remained in the brain even at 257 days, the latest time point examined. We note that one mouse with detectable tumor at 190 days of age had reduced frequencies of epitope I and IV-specific TCD8 compared to a mouse with no tumor at this time point. However, at 257 days of age, no difference in the response was found between mice with and without obvious tumors. Importantly, a proportion of the Tag-specific TCD8 in the brain at this late time point exhibited effector function, as detected by IFN-γ production. These data indicate that Tag-specific TCD8 specific for both epitope I and epitope IV persist at the tumor site in SV11 mice and maintain function at late time points following adoptive transfer of Tag immune cells.
So far, the results of this study indicate that adoptive transfer of Tag-immune donor lymphocytes into irradiated hosts leads to rapid accumulation and persistence of Tag-specific TCD8 within the tumor site. We postulated that the increased control of tumor progression by Tag-immune versus naïve donor cells could be explained by either initial differences in the frequency of Tag-specific TCD8 or phenotypic differences between effector cells derived from Tag-immune or naïve donors. We asked whether reducing the number of immune donor cells would hinder rapid accumulation in the brain and/or control of tumor progression. Irradiated 80 day-old SV11 mice received either 5×107, 5×106 or 5×105 Tag-immune donor lymphocytes. Flow cytometric analysis prior to adoptive transfer demonstrated that the frequency of epitope I and IV-specific TCD8 was ~1.5×105 and ~1×106, respectively at the 5×107 dose. The frequency of epitope I and IV-specific TCD8 recovered from the brain at 90 days of age correlated with the number of cells transferred (epitope IV; ~42% vs. ~10% vs. ~1%) (Fig. 4a). Similar results were observed in the spleen (data not shown). In contrast, at 120 days of age (Fig. 4b), the frequency of epitope IV-specific TCD8 at the tumor site was ~25% of TCD8 whether 5×107 or 5×106 Tag-immune donor cells were transferred. No TCD8 specific for either epitope was detected following transfer of the lowest dose of cells. These data indicate that reducing the initial dose of Tag-immune donor cells increases the amount of time required to achieve high level accumulation of TCD8 in the brain.
It was possible that following adoptive transfer of the lower doses of lymphocytes, 5×106 and 5×105, the cells may seed poorly into the recipient SV11 mice. To address this issue, the varying doses of Tag-immune donor cells were combined with naïve lymphocytes obtained from line T350gB mice, which are tolerant to the Tag epitopes, in order to maintain the total number of cells transferred at 5×107. Mice of the line T350gB express a full length SV40 Tag variant containing the HSV gB498–505 epitope inserted at position 350 of Tag (Blaney, Yorty, Bonneau, Schell and Tevethia, manuscript in preparation). Tag expression in the thymus results in tolerance toward all four H-2b-restricted epitopes. Maintaining the total number of cells transferred at 5×107 did not restore the frequency of Tag-specific TCD8 in SV11 mice that received 5×106 or 5×105 Tag-immune cells (data not shown; result comparable to Fig. 4a). Therefore, the observed decrease in the frequency of Tag-specific TCD8 in the brain was not simply due to poor seeding of the donor cells when the lower doses of 5×106 and 5×105 cells were transferred.
Since fewer Tag-specific TCD8 were present at early time-points in SV11 mice receiving the lower dose of immune cells, it was important to determine whether there was an effect on tumor burden. SV11 mice were euthanized at 120 days of age and tumor burden was assessed by H&E staining. At the highest dose of 5×107 donor cells, complete elimination of tumors was observed (Fig. 4c), as these mice would have had significant tumor burden at the time of adoptive transfer. As the frequency of Tag-specific cells transferred was decreased, the tumor burden in the SV11 mice increased. Indeed, several of the mice treated with the 5×105 dose were euthanized prior to 120 days of age because they exhibited symptoms of advanced-stage tumors. Interestingly, the 5×106 dose of donor cells exhibited results similar to adoptive transfer of naïve donor cells . These data indicate that efficient control of advanced tumors requires adoptive transfer of a relatively high frequency of activated TCD8 against the target antigen in order to achieve high level accumulation of tumor-specific TCD8 early after transfer. Hence, changing the kinetics of TCD8 accumulation alters the effect on tumor burden.
Both epitope I and IV-specific TCD8 persist at late time points in SV11 mice following adoptive transfer of immune donor cells. However, epitope I-specific TCD8 exhibit a decreased ability to accumulate at the tumor site and persist at relatively low levels. Therefore, we sought to determine whether donor cells primed against only epitope IV could control tumor progression. To elicit an epitope IV-specific response, B6 mice were immunized with the epitope IV-expressing cell line, B6/Tag-IV (Fig. 5a). Since epitope I-specific TCD8 responses are rarely primed by the endogenous Tag in SV11 mice, this approach allowed us to monitor the contribution of activated epitope IV-specific donor TCD8 in isolation.
Irradiated 80 day-old SV11 mice received 5×107 donor cells from mice immunized specifically against epitope IV. Reconstitution of SV11 mice with epitope IV immune cells resulted in a significant increase in the lifespan of SV11 mice (Fig. 5b, mean age of 226 days; p < 0.001), suggesting that the immunodominant epitope IV-specific TCD8 from immune donors play a significant role in the observed extension of the SV11 lifespan. At the time that SV11 mice exhibited symptoms of late-stage tumor burden, few functional epitope IV-specific TCD8 were present in either the spleen or brain (data not shown). We note that the lifespan of SV11 mice given epitope IV immune donor cells is slightly lower than that reported previously when mice received wild type Tag-immune cells containing epitope I-, epitope II/III- and epitope IV-specific TCD8 (mean age of 277 days) . This result suggests that epitope IV-specific TCD8 are largely responsible for control of choroid plexus tumors, but that TCD8 targeting additional epitopes might further prolong the already significant affect.
The studies described here address the basis for efficient control of spontaneous choroid plexus tumor progression in the SV11 mouse model following adoptive immunotherapy with Tag-immune donor lymphocytes . The results demonstrate that Tag-immune donor cells are detected at the tumor site earlier and accumulate to peak levels more quickly than TCD8 derived from naïve donors. Both functional epitope I and epitope IV-specific TCD8 persisted long-term in SV11 mice, although only epitope IV-specific TCD8 persisted at high levels in the brain. Decreasing the dose of transferred Tag-immune cells delayed the peak of TCD8 accumulation within the brain and the ability to control tumor burden, further supporting the idea that T cell frequency rather than phenotype is associated with control of tumor progression. Combined, our data suggest that rapid, high level accumulation of Tag-specific TCD8 within the brain optimizes the immunotherapeutic effect against advanced-stage choroid plexus tumors.
One factor contributing to successful control of tumor progression in this model of adoptive immunotherapy may be the use of freshly isolated immune donor cells. Results from clinical adoptive immunotherapy studies have revealed that extensive in vitro expansion of tumor-specific T cells may diminish their effectiveness in vivo. While T cells expanded in vitro were found to be highly cytotoxic and produce large amounts of IFN-γ, these donor cells were undetectable within the peripheral blood one week post-infusion . Supplemental treatment with IL-2 was able to prolong T cell survival for only a limited time period . Failure of in vitro cultured T cells to persist long-term post-infusion may be associated with decreased expression of homing receptors on activated cells and subsequent failure of these cells to traffic to lymph nodes and interact with professional antigen presenting cells . Indeed, the stimulation of T cells in vitro does not abrogate the requirement for recognition of professional antigen presenting cells in the host . In the current study, the polyclonal Tag-specific TCD8 derived from Tag-immune donors exhibited an effector phenotype directly ex vivo as evidenced by downregulation of CD62L and antigen-specific production of IFN-γ following ex vivo stimulation with peptide. However, a small portion of the Tag-specific TCD8 remained CD62L high. This suggests that at least a subset of the Tag-specific cells transferred may be in a relatively early effector stage which may aid in the long-term persistence of the cells in SV11 mice .
The role of lymphodepletion in enhancement of adoptive immunotherapy approaches is now well established [13, 15–17, 40, 41]. Following non-myeloablative chemotherapy, the survival, differentiation, and persistence of multiple tumor-specific T cells has been associated with clinical regression of melanoma [18, 34, 35]. Current research suggests that lymphodepletion strategies aid in adoptive immunotherapy (reviewed in ) through mechanisms such as elimination of competition for homeostatic cytokines  and the depletion of regulatory T cells . In previous studies with SV11 mice, irradiation prior to adoptive transfer allowed for enhanced priming of epitope IV-specific naïve donor cells by the endogenous Tag . Thus, irradiation may also facilitate cross-priming of tumor antigen . Indeed, a recent study indicates that localized irradiation results in increased cross-presentation by tumor stromal cells .
Accumulation of adoptively-transferred tumor-specific TCD8 at the tumor site is critical for inhibiting growth and for ultimate destruction of the tumor [25, 45, 46]. Regarding the kinetics of T cell accumulation, we found that Tag-immune donor TCD8 arrive only a few days earlier within the spleen and brain of tumor-bearing SV11 mice than TCD8 derived from naïve donor cells. Although there is a relatively small difference in arrival time, the Tag-immune TCD8 achieved maximum frequency within the brain very quickly (~5 days), while transfer of naïve donor cells took more than 3 weeks to achieve a similar frequency. This rapid accumulation is likely the result of an increased precursor frequency as well as a decreased amount of time required for the T cells to traffic to the tumor since effector cells express the appropriate homing molecules for migration and extravasation into peripheral tissues. A similar lag in accumulation was observed when lower doses of immune T cells were transferred into SV11 mice, resulting in less efficient control of tumors. This latter result suggests that T cell numbers, rather than phenotypic differences in effector T cell populations contributes strongly toward inducing tumor regression.
Although our results demonstrate that TCD8 from Tag immune donors exhibited an early advantage in T cell accumulation within the brain over TCD8 from naïve donors, the numeric advantage was eventually normalized at later time points. Yet, the SV11 mice that received the Tag-immune donor cells exhibited an ~100 day increase in lifespan . This suggests that the initial attack of tumor-specific TCD8 is critical for extended control of tumor burden and that long-term T cell persistence, although likely important, is not sufficient for extended control of tumor progression. This suggestion is supported by the findings of Hwang, et al.  who recently demonstrated that the race between arrival of the T cells and tumor growth dictates the outcome of tumor immunotherapy. In that study the authors found that depletion of tumor-specific TCR transgenic T cells at four, but not ten days post-adoptive transfer into mice with established B16 melanomas abrogated control of tumor progression, indicating that early accumulation following T cell transfer is critical for control of tumor progression.
In SV11 mice, tumor progression is rapid after 80 days of age, at which time the more aggressive grade III and IV tumors appear within the choroid plexus . Hence, the period around 80 days of age may represent a critical window within which control of tumor burden can be achieved if enough tumor-specific cells are available at the tumor site. Alternatively, irradiation may provide a brief period during which the tumor is more sensitive to effector T cells due to changes such as transient upregulation of antigen presentation [17, 44] or death receptors  on the tumor cells, or elimination of immunosuppressive cells . In support of this idea, Zhang et al.  demonstrated that the optimal time for T cell transfer was within two days following localized irradiation, due to a transient increase in tumor antigen presentation by the stromal cells. In SV11 mice, rapidly arriving Tag immune cells may encounter a radiation-induced increase in Tag epitope presentation at early times after irradiation, leading to high level expansion in the brain. In contrast, T cells derived from naïve donors may only encounter reduced levels of antigen presentation due to delayed arrival. Taken together, these results suggest that approaches which facilitate the rapid influx of adoptively transferred T cells into the tumor will be most effective.
The suggestion that maximal accumulation of TCD8 within the tumor must occur rapidly in order to significantly reduce tumor progression is further supported by the results of the dosage data in which decreasing the initial number of Tag-immune donor lymphocytes from 5×107 to 5×106 cells resulted in decreased tumor protection, resembling that achieved following adoptive transfer of the naïve donor cells. Previous studies have also documented decreased tumor protection following adoptive immunotherapy with reduced numbers of effector cells [19, 41, 46]. Our data is especially intriguing given that a ten-fold reduction in donor Tag-immune cells still provides a much higher starting frequency of epitope I and epitope IV-specific T cells (~1.5×104 and ~1×105) as compared to the naïve donor pool. Despite the higher starting frequency, the 5×106 dose of Tag immune donor cells resulted in life span and T cell accumulation kinetics similar to that observed when naïve donor cells were used, indicating that the size of the donor T cell pool must be large to promote maximum T cell accumulation within the tumor in time to facilitate regression.
In addition to the difference in kinetics, the Tag-immune donor lymphocytes may have an advantage over naïve donor cells since multiple epitopes within Tag are targeted. Epitope I- and II/III-specific TCD8 from the naïve repertoire are inefficiently primed by the endogenous Tag, even following prior irradiation of SV11 mice. Hence, only epitope IV-specific TCD8 are recruited . Following transfer of Tag-immune cells, both epitope I and IV-specific TCD8 rapidly trafficked to the tumor site. However, only epitope IV-specific TCD8 were maintained at high levels within the brain of SV11 mice. It is unknown whether this distinction is due to differences in T cell proliferation or retention within the tumor site or whether these T cell subsets may be variably susceptible to negative signals at the tumor site. The enhanced survival of the SV11 mice following adoptive transfer of Tag-immune donor cells appears to be highly dependent upon the immunodominant epitope IV-specific TCD8, as transfer of lymphocytes from Tag IV-only immunized donors dramatically prolonged the survival of SV11 mice. The possibility that epitope I- and II/III-specific TCD8 play a supporting role in control of tumor progression cannot be ruled out. Indeed, clinical studies have documented that the persistence of T cells specific for multiple tumor epitopes is associated with tumor regression .
In the present study, donor T cells were derived from mice not exposed to the tumor antigen during T cell development and thus represent high quality/avidity tumor-reactive T cells. This scenario is not easily duplicated in human cancer patients, as the self-reactive T cell repertoire has been purged of the most strongly self-reactive T cells that could be used to target the autochthonous tumor. Several potential strategies have emerged which could provide higher avidity tumor-reactive T cells including the targeting of subdominant tumor epitopes where T cell tolerance may be incomplete [49, 50], targeting of mutated antigens for which central tolerance has not eliminated strongly reactive T cells [35, 51] and the use of redirected T cells engineered to express high avidity tumor antigen-specific TCRs . Regardless of the donor cell source, understanding the mechanisms that facilitate effective adoptive immunotherapy will be important for ensuring donor cell survival, efficient trafficking and appropriate activity following transfer into the environment of the tumor-bearing host.
We thank Melanie Epler and Jeremy Haley for excellent technical support and Nate Sheaffer from the Flow Cytometric Core Facility of the M.S. Hershey Medical Center for assistance with flow cytometry. This work was supported by CA25000 from the National Cancer Institute/National Institutes of Health.