We aimed to use cell-based carriers to direct vector production to target sites for systemic therapy. We used T cells engineered to express a chimeric T cell receptor that can specifically recognize target cells expressing the tumor-associated carcinoembryonic antigen (CEA). These T cells were modified to produce a retrovirus under tight pharmacological control using the rapamycin-inducible transcriptional regulatory system. The retroviral vectors produced were transcriptionally targeted to CEA-expressing target cells. We found that vector production and transgene expression from these T cells in vitro was dependent on pharmacological induction and expression of CEA in target cells, respectively. Mice bearing metastatic tumors that received cell carriers delivering the HSVtk gene demonstrated a significant increase in survival, but only in response to pharmacological induction of vector production. Interestingly, the therapeutic effect required the presence of the tumor-specific chimeric receptor on T cells. Further studies demonstrated that systemic delivery of tumor-specific T cells to mice bearing metastatic tumors caused recruitment of nonspecific T cells to the tumor site. We hypothesize that this enhanced targeting to tumor sites is responsible for the efficiency of T cell-mediated retroviral gene transfer and that this principle can be used to enhance systemic therapies using immune-cell carriers.

T cells have the desirable properties of systemic trafficking and the ability to recognize target cells (1). These properties represent the currently unattained goal of successful gene therapy vectors (2, 3). We sought to use these features of T cells to enhance systemic delivery of gene therapy vectors and to enhance the efficacy of adoptive T cell therapy in vivo.

We have recently reported on the successful use of T cells to deliver retrovirus to metastatic tumor deposits in the lung and liver (4). As a model system we used Jurkat T cells expressing a CIR7 recognizing the tumor-associated carcinoembryonic antigen CEA or a control CIR recognizing the hapten NIP (5). In those studies, we used a novel retroviral vector in which the U3 region of the 5′-LTR was replaced with a promoter that is activated on CIR ligation. Additionally, the U3 region of the 3′-LTR was replaced with a CEA promoter. This ensured that proviral transgene expression would only occur in transduced cells with CEA promoter activity. Using this system we showed in vitro that retrovirus production occurs only upon CIR ligation and proviral gene expression only occurs in CEA+ve infected cells. Furthermore this system was used to significantly reduce the metastatic deposits in the lung and liver of nude mice bearing CEA+ve tumors (4).

This approach is therefore very promising for the systemic delivery of retrovirus by T cell based carriers. To attain exogenous control over the production of retrovirus from engineered T cells, we set out to place vector production under pharmacological control. Such an approach should allow us to distinguish the functional and regulatory contribution of the effects of TCR engagement from those of retroviral production. Furthermore, this strategy would allow us to investigate the possibility of uncoupling virus production from T-cell activation thereby opening the way for the use of other cell carriers in addition to tumor-specific T cells. Finally, in a clinical setting, pharmacological control would significantly enhance safety, allowing virus production to be terminated by drug withdrawal should it become unnecessary or harmful.

Several systems for the small molecule control of gene transcription have been described. Here we have investigated the use of the dimerizer-regulated gene expression system based on rapamycin, a small molecule natural product that mediates heterodimer formation between the immunophilin FKBP and the lipid kinase FRAP (6, 7). By fusing FKBP domains to a DNA-binding domain, and a portion of FRAP to a transcriptional activation domain, assembly of an active transcription factor and expression of a target gene can be made absolutely dependent on the presence of rapamycin or analogues of rapamycin (“rapalogs”) that lack the antiproliferative and immunosuppressive activity of rapamycin (7). The rapamycin-inducible transcription switch has several advantages as a pharmacological control for retrovirus production including tightly regulated expression that should allow for strict temporal control of retroviral production (7).

In this report, we demonstrate that human T cells can be engineered to produce retroviral vectors using rapamycin induction and that tight transcriptional specificity for target cells can be conferred upon the retroviral particles. Additionally, we show that retrovirus production linked to pharmacological control is effective at enhancing survival in a metastatic model of colon cancer, but that this therapy still requires a TCR specific for the tumor cell. We provide evidence that suggests the requirement for a tumor-specific T cell is, at least in part, attributable to the increased recruitment of other T cells to the tumor site, which is likely a result of the elaboration of pro-inflammatory mediators from activated TCR signaling.

Cell Culture.

The CEA+ve human colorectal tumor cell line HCT116, the CEA−ve human melanoma line Mel624 and Jurkat T lymphocytes derived from an acute T-cell leukemia were obtained from the American Type Culture Collection. Jurkat.CEA and Jurkat.NIP are lines derived from Jurkat cells that were stably transfected with two different chimeric TCR constructs, both of which contain a functional CD3ζ intracellular domain. The Jurkat.NIP line expresses a B1.8-derived scFv with specificity against the hapten NIP as a ligand-binding domain; Jurkat.CEA contains the MFE23 scFv with specificity against human CEA (4, 5).

Adherent cell lines were grown in DMEM medium with 10% FCS. Jurkat lines were maintained in RPMI 1640 medium with 10% FCS. Jurkat.CEA and Jurkat.NIP were maintained in RPMI, 10% FCS with hygromycin at 400 μg/ml. For in vitro cell killing, the medium was supplemented with ganciclovir to a final concentration of 5 μg/ml. For induction of gene expression, cells were grown in normal medium supplemented with 10 nm rapamycin or the rapamycin analogue (details of this analogue are available from ARIAD8). Cells were transfected using the Effectene reagent (Qiagen, Valencia, CA) according to the manufacturer’s instructions for both adherent and nonadherent cells. To obtain a population of Jurkat cells enriched for transfectants with rapamycin-inducible vectors, we transfected cells with the appropriate vector DNA (pNANA or pSB-STN-Rapa), diluted them to approximately 10–100 cells per well, infected them with the AdC4 virus (multiplicity of infection of 100), incubated them in medium containing rapamycin, and examined them for GFP expression by fluorescence microscopy. Wells containing the highest level of fluorescence were expanded and used for future experiments on the basis that these populations contained significant proportions of Jurkat cells stably transfected with the appropriate vector. Populations of Jurkat cells stably transfected with pBabePuro were obtained by selection in puromycin after transfection. Populations of Jurkat.CEA and Jurkat.NIP cells transfected with vectors, such as pSB-STN (Rapa)-tk, were obtained by cotransfection of the retroviral vector (2 μg) with pSV2-puro (0.1 μg) followed by selection in puromycin for up to 2 weeks. These populations were, therefore, enriched for cells containing the retroviral vector, although they were not single-cell cloned.

Plasmids.

The C-type retroviral vectors used in this study were derived from the pBabePuro vector (8). Modifications were made by standard molecular techniques and are described in Fig. 1 legend for pNANA, pSB-STN-Rapa, and pSB-STN-tk. The rapamycin-inducible transcriptional regulatory system involves coexpression of two fusion proteins: (a) Z1F3, a fusion between the DNA binding domain ZFHD1 and three FKBP domains; and (b) R(H1)S, a fusion between a modified rapamycin-binding portion of FRAP and part of the p65 subunit of human NFκB (6). These two proteins were expressed constitutively on a bicistronic message from the plasmid pC4EN-R(H1)S-Z1F3 (6). The Z12 dimerizer-regulated promoter, comprising 12 binding sites for the ZFHD1 DNA-binding domain followed by the minimal interleukin-2 promoter (6), is designated as “Rapa” in Fig. 1 A. The HSVtk gene was cloned into the vectors from the pCR3.1-HSV-tk as described previously (9, 10). The 3′-LTR replacement in pSB-STN-Rapa was carried out by cloning the 365-bp CEA promoter from the plasmid pSW121B4 (a gift from Dr. P. Searle, University of Birmingham CRC Institute for Cancer Studies, Medical School, Edgbaston, Birmingham, United Kingdom) into an isolated copy of the Mo-MLV LTR in the shuttle plasmid pSKLTR [a gift of Dr. Rhys Jagger, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Headington, Oxford, United Kingdom, and (11)] followed by reassembly into the pBabePuro plasmid as described previously (11, 12). Full details of the cloning and plasmid maps are available on request.

Retroviral Infections.

Cell-free supernatants were harvested from cultures engineered to produce retroviral vectors and centrifuged twice to pellet any cells. Retroviral infections were performed by exposing exponentially growing target cells to different dilutions of viral supernatants in serum free medium for 4 h in 24-well plates in the presence of Polybrene (Sigma) at 4 μg/ml. The infectious media was then removed, and cells were washed three times before being allowed to grow in normal media for 48 h. At this stage, the infected cells were FACS analyzed for the expression of the eGFP transgene.

Adenoviral Vectors.

The Ad-gag/pol and Ad-GALV adenoviral vectors have been described previously (13). To generate the adenovirus, Ad-C4, which encodes the dimerizer-regulated transcription factors, the coding regions of the DNA-binding and activation-domain fusion proteins were excised from the C4 plasmid and cloned into the shuttle vector pAdBN. The resulting construct was transfected into 293 cells, and plaques representing candidate recombinant virus clones were chosen. Virus was expanded from individual plaques, and the identity of the recombinant virus was confirmed by a functional assay. After two rounds of plaque purification, a positive recombinant was expanded on 293 cells, and titer was determined by limiting dilution analysis on 293 cells.

In Vivo Studies.

The Institutional Animal Care and Use Committee at the Mayo Foundation approved all procedures. To establish liver metastasis, athymic mice received i.v. injections of 106 HCT116 tumor cells or 2 × 106 Mel624 tumor cells. On days 7 and 8, 4 × 106 Jurkat.CEA or Jurkat.NIP cells pre-engineered to produce pSB-STN-(Rapa)-tk virus as described previously were injected i.v. The mice were then given 5 consecutive days of ganciclovir treatment at 50 mg/kg with or without the rapamycin analogue AP21967 delivered i.p. at 1 mg/kg in PBS. Mice were examined daily and sacrificed when they developed evidence of prominent metastatic disease.

To track viral distribution within mice receiving systemic T-cell therapy, athymic mice were given 1 × 106 HCT116 cells i.v. On days 7 and day 8, 4 × 106 Jurkat.CEA engineered to produce pSB-STN-(Rapa)-GFP virus as described previously were injected i.v. On day 12, lung, liver, spleen, brain, ovary, small intestine, blood and kidney were harvested and half of each organ was frozen for genomic PCR isolation and half was FACS analyzed for GFP and binding of PE conjugated HLA-ABC (BD PharMingen, San Diego, CA).

To study the recruitment of nonspecific Jurkat cells to HCT116 tumors by specific Jurkat.CEA cells, 1 × 106 HCT116 were injected either s.c. or i.v. Seven days following tumor inoculation, 4 × 106 Jurkat.CEA plus 4 × 106 Jurkat cells or 4 × 106 Jurkat.NIP, plus 4 × 106 Jurkat cells were injected i.v. Jurkat cells were labeled with Cell Tracker Orange (Molecular Probes, Eugene, OR). Control mice received PBS, Jurkat alone, or Jurkat.CEA alone. On day 9, livers were removed and analyzed by confocal microscopy for orange cell infiltrate. On day 14, s.c. tumors were removed and analyzed by confocal microscopy.

Genomic PCR Analysis.

Genomic DNA was isolated from frozen lung, liver, spleen, brain, ovary, small intestine, blood, and kidney using the DNAeasy kit (Qiagen, Valencia, CA) according to manufacturer’s instructions. Ten nanograms of DNA were amplified by PCR using primers specific for the target genes. The reaction mixtures were analyzed by agarose gel electrophoresis. The target genes were specific for human cells, mouse cells, virally infected cells, or the T-cell carriers.

Construction of a Rapalog-inducible Retroviral Vector That Generates Virus That Is Transcriptionally Specific for Colorectal CEA+ve Tumor Cells.

We constructed several different retroviral vectors based on the C-type retroviral vector pBabe Puro (8). pNANA contains the eGFP cDNA cloned into pBabe Puro in place of the SV40-Puro cassette, with the 5′- and 3′-LTRs retained as the wild-type LTRs (Fig. 1,A). In pSB-STN-Rapa, the 5′-LTR of pNANA is replaced with the rapamycin-responsive Z12 promoter (see Refs. 6, 7; Fig. 1,A). In addition, using a previously described strategy to generate transcriptionally targeted retroviral vectors (11, 12), we replaced the viral enhancer and promoter in the 3′-LTR of pSB-STN-Rapa with a 365-bp element of the human CEA promoter, which has been shown to confer colorectal specific gene expression (see Ref. 14; Fig. 1,A). Thus, the rapamycin-inducible promoter would drive expression of the viral genome in the producer cell. However, in the provirus resulting from infection of a target cell with pSB-STN-Rapa, the CEA promoter would direct expression of the transgene GFP (Fig. 1 A). In pSB-STN-(Rapa)-tk, the eGFP cDNA was replaced with the cytotoxic gene HSV-tk.

We next investigated whether human Jurkat T cells could be converted into retroviral-producer cells using vector-modified/selected Jurkat populations transfected with the pSB-STN-Rapa vector. To deliver rapalog dimerized transcription factor, we used an adenovirus encoding the C4 DNA binding and activation domains of the rapamycin system (Ad-C4; see “Materials and Methods”). To provide the remaining required retroviral genes we used the Ad-gag/pol and Ad-GALV viruses, which encode the Mo-MLV GAG and POL proteins and the GALV envelope protein, respectively (13). In the absence of Ad-C4 and/or rapalog, the supernatants from pSB-STN-Rapa-transfected Jurkat cells infected with Ad-gag/pol and Ad-GALV failed to transfer GFP expression to any cell lines (Fig. 1,B). However, in the presence of both Ad-C4 and the rapamycin analogue AP21967, the supernatants from pSB-STN-Rapa-transfected Jurkat cells infected with Ad-gag/pol and Ad-GALV generated retrovirus that passed GFP expression into CEA+ve, but not CEA−ve cells (Fig. 1 B). Rapalog-stimulated Jurkat cells were able to transfer GFP expression to a range of other CEA+ve cell lines including SW620, LoVo, and SW116 (data not shown).

Pharmacologically Regulated Retrovirus Production from Systemically Delivered T Cells Results in Efficient Transduction of Metastatic Tumors.

Our previous data demonstrated that systemic therapy with vector-packaging T cells resulted in decreased metastatic burden, where vector production was dependent on chimeric TCR (CIR) ligation (4). To study vector distribution where virus production was under the control of a pharmacologically regulated promoter system rather then dependent upon the presence of CEA antigen, we established metastases of HCT116 in nude mice by tail vein injection. Seven days later, we administered J.CEA cells transfected with pSB-STN-Rapa and infected with Ad-gag/pol, Ad GALV, and Ad-C4. For the next 5 days, the rapalog AP21967 was administered i.p. We used a rapalog rather than rapamycin, in these in vivo studies because the rapalog is devoid of antiproliferative activity in vivo and yet retains the ability to activate transcription (7). This prevented the natural antitumor activity of rapamycin from complicating the interpretation of our studies. We isolated genomic DNA from the liver, lungs, spleen, ovaries, brain, kidney, blood, and small intestine for PCR analysis. Separate primer pairs were designed to detect vector DNA (Fig. 1,A), which should be present only in Jurkat cells; human DNA, present in the HCT116 metastases and circulating Jurkat cells, or integrated proviral DNA (Fig. 1,A), which should be present in only infected cells. We were only able to detect vector DNA in the spleen, suggesting that the number of Jurkat cells in other organs was below the detection level (Fig. 2,A). Integrated proviral DNA was detected in both the liver and lung, and this correlated well with the presence of detectable human genomic DNA (Fig. 2 A). This suggests that viral infection was prominent in areas of tumor localization.

To examine gene expression from integrated proviral DNA, we examined GFP expression in these same organs by FACS. To determine whether GFP was expressed in human or mouse cells, we costained the cell types with a PE conjugated human HLA-ABC antibody (Fig. 2,B). High numbers of human cells were detected in the liver and lung, correlating well with the PCR data (Fig. 2 A). These human cells are likely to be HCT116 cells rather than virus producer Jurkat.CEA cells given that we were unable to detect any viral producer cells in the liver and lung by PCR. In addition, our previous work has shown that J.CEA cells are killed after encountering CEA+ve tumor cells (4). In the liver, ∼25% of the HLA-ABC cells expressed GFP, and in the lung ∼60% of the tumor cells were positive for GFP expression. This represents an impressive level of systemic gene transfer achieved using T cells and pharmacologically induced retrovirus production.

Pharmacologically Regulated Retrovirus Production Requires T-Cell Recognition of Tumor Cells for Systemic Therapy.

The PCR and FACS data showed that effective systemic transgene expression was achieved in metastatic tumors in vivo after pharmacological induction of retrovirus from tumor specific T cells. To determine whether this could translate into a therapeutic benefit, we replaced the eGFP cDNA with an HSV-tk gene to create the vector pSB-STN-Rapa-tk (Fig. 1,A). Nude mice bearing HCT116 or Mel624 metastatic tumors received injects of J.CEA or J.NIP cells transfected with pSB-STN-Rapa-tk and infected with Ad-gag/pol, Ad-GALV, and Ad-C4. Administration of AP21967 and GCV resulted in a significant increase in survival (P < 0.005; Fig. 3,A and B). A therapeutic effect was only achieved in mice bearing the CEA+ve tumor, HCT116 (Fig. 3,A), and not in mice bearing CEA−ve Mel624 tumors (Fig. 3,B), showing dependence on the CEA promoter. Importantly, a therapeutic effect was also achieved only when rapalog was given to the mice to induce retroviral production from the T cells (Fig. 3 A). Thus, vector production from these cells was essential for therapy and could be tightly controlled by rapalog administration.

In this study, it was possible that adoptively transferred T cells would circulate and nonspecifically localize to tumors in the liver. In this scenario, it would be expected that J.NIP-pSB-STN-Rapa-tk cells should be as therapeutic as J.CEA-pSB-STN-Rapa-tk cells after administration of rapalog in vivo. However, interestingly, this was not the case and the therapeutic effect still required T-cell recognition of tumor cells for systemic therapy. Thus, T cells specific for CEA on the surface of tumors (Jurkat.CEA) conferred a therapeutic effect when rapalog was provided to induce retrovirus production; however, T cells specific for the hapten NIP (Jurkat.NIP) showed no therapy even in the presence of pharmacological induction of retrovirus from these cells (Fig. 3 A). Therefore, even if retrovirus production is uncoupled from T-cell specificity and, instead, controlled in a temporal manner by pharmacological induction, CIR specificity is still required to achieve therapy.

Tumor-specific Jurkat.CEA Cells Increase the Recruitment of Nonspecific Jurkat Cells to Tumor-bearing Livers and s.c. Tumors.

There are several ways that CIR specific interaction could have enhanced the observed therapeutic effect. One attractive possibility is that CIR-ligand interaction resulted in the elaboration of factors such as cytokines and chemokines. According to this hypothesis a significant accumulation of Jurkat cells to tumor sites would be achieved in mice receiving tumor specific Jurkat.CEA cells. To test this hypothesis we performed a cell chasing experiment in vivo using labeled Jurkat cells (Fig. 4 and Table 1). In this study, nude mice received i.v. or s.c. HCT116 tumor cells followed 7 days later by injections of Jurkat.CEA or Jurkat.NIP cells, along with Cell Tracker Orange labeled parental Jurkat cells. Two days or 7 days later, livers or s.c. tumors, respectively, were removed and examined by confocal microscopy for evidence of localization of orange cells. Only in mice receiving both Jurkat.CEA and orange Jurkat cells were orange cells detectable within either the liver or the s.c. tumor (Fig. 4 and Table 1). Typically for livers, two to three clusters of orange cells were seen in sections measuring approximately 0.5 cm × 0.5 cm × 0.5 mm. Jurkat.NIP cells did not recruit orange cells to the liver or s.c. tumor, nor were there any detectable orange cells in the s.c. tumor or liver of mice receiving only orange-labeled Jurkat cells. Our data suggest that signals elaborated in vivo as a result of tumor/Jurkat cell interaction lead to the recruitment of other Jurkat cells, including nonspecific cells, to the tumor site.

In Vitro Jurkat Cell/Tumor Cell Interaction Releases Factors That Stimulate NFκB Promoter-driven Retroviral Production.

We hypothesized that this recruitment occurred through the release of inflammatory mediators from T cells, which enhance T-cell entry into sites of ligand-receptor interaction. To test this hypothesis, we used a previously described retrovirus construct, pSB-STN(NFκB)3, in which virus production is induced by NFκB activation in producer cells (4). In this experiment Jurkat.CEA or Jurkat.NIP were cocultured for 24 h with HCT116 tumor cells. Cell-free supernatants from the cocultures were transferred onto Jurkat.CEA or Jurkat.NIP cells transfected with pSB-STN (NFκB)3 and infected with Ad-gag/pol and Ad-GALV. Retrovirus production will only occur in cells exposed to supernatants that lead to activation of the (NFκB)3. Supernatants from these cells were then transferred onto HCT116 cells, and 24 h later GFP expression was examined by FACS (Fig. 5). GFP expression was seen only from T cells stimulated with supernatants from Jurkat.CEA/HCT116 cocultures. Both Jurkat.CEA and Jurkat.NIP cells transfected with pSB-STN (NFκB)3 were able to produce retrovirus when they received supernatants from Jurkat.CEA/HCT116 cocultures, but neither produced retrovirus if it received supernatants from other test or control cocultures. This provides further in vitro evidence that TCR interaction with cognate ligand results in the elaboration of inflammatory signals that, in this case, are sufficient to activate NFκB promoter activity. Such inflammatory signals are consistent with the recruitment of nonspecific Jurkat T cells to tumor sites by an initial antigen specific localization of Jurkat.CEA cells (Fig. 4) and would greatly potentiate the in vivo therapeutic effects observed (Fig. 3).

In this report, we describe that T cells can be modified to produce retrovirus after pharmacological induction with a rapalog dimerizer (Fig. 1). Furthermore, these T cells were able to deliver systemically a reporter transgene into a very significant fraction of metastatic tumor cells in the lung and liver (Fig. 2, A and B). When the retrovirus was modified to encode a therapeutic transgene, HSV-tk, we were able to show a significant increase in survival in mice receiving both rapalog to induce viral production and GCV as the cytotoxic prodrug in comparison with animals in which either or both drugs were withheld (Fig. 3,A). Interestingly, for achieving a therapeutic effect, the retroviral carrier Jurkat cells had to be tumor-specific. This raises the possibility that CIR signaling results in a benefit over and above that which is produced by nonspecific cells trafficking to tumors and releasing virus. We showed in vivo that tumor specific Jurkat.CEA cells recruit parental Jurkat cells to the tumor site, whereas nonspecific Jurkat.NIP cannot recruit parental Jurkat cells (Fig. 4). Finally, in vitro we have shown that coculture of untransfected Jurkat.CEA with HCT116 results in the elaboration of secreted signals that are sufficient to activate NFκB promoter activity in other Jurkat cells (Fig. 5). These data suggest that one mechanism whereby specific T cells are enhancing the therapeutic effects of pharmacologically induced retroviral gene therapy may be through the recruitment of other T-cell carriers to the tumor site for local release of their virus. In addition, these effects offer the opportunity to enhance the trafficking of endogenous, or adoptively transferred, T cells to tumor sites, whether or not these cells are modified to produce vector. Preliminary experiments in an immunocompetent murine model have confirmed that such effects lead to enhanced accumulation of adoptively transferred T cells into s.c. tumors. Experiments are under way to exploit these effects to create more effective protocols for both adoptive transfer of antigen-specific T cells and for direct, T cell-released, retroviral tumor cell killing.

Pharmacological induction of the retroviral genome allowed for tight on/off regulation of retrovirus production. However, after initiating these experiments, it was less clear what the requirement for tumor-specific T cells would be once retroviral vector production was under pharmacological control. We envisioned several models of how the Jurkat T cells might be trafficking in vivo. The model that best fits the data described here is that T cells circulate and a few initial T cells arrive at the tumor site. After TCR-antigen engagement, a series of cytokines and chemokines are released resulting in a secondary recruitment of further circulating activated T cells that may be either specific or nonspecific. This model raises the interesting possibility that a limited low level of specific T-cell trafficking is required to initiate an inflammatory response to tumor associated antigen, which then allows the recruitment of additional nonspecific T cells to the tumor. This is characteristic of reported data in adoptive transfer experiments with viral specific CD8+ T cells (15, 16, 17). After acute lung infection in mice carrying adoptively transferred antigen-specific CD8 T cells, between 10 and 70% of the CD8+ T cells recovered from the respiratory tract were nonspecific (17). The authors suggest that whereas antigen is required to activate T cells, localization of T cells to inflammation is a property of recently activated and memory T cells (17). Adoptive transfer of tumor-antigen-specific T cells into tumor-bearing patients can result in high numbers of activated, antigen-specific cells at the site of antigen, in proportions many times greater than found in peripheral blood (18). This nonrandom association (19) is the basis for the use of activated specific T cells as vector delivery agents for cytotoxic gene therapy (4).

The cytokines and chemokines released from the various subsets of T cells after TCR/ligand interaction have been well defined (20, 21). Cytokines released from activated T cells include IFN-γ and tumor necrosis factor-α (20). These cytokines have the potential to increase immune-cell recruitment by activating resident cells (22) and tissue macrophages to secrete chemokines (23). These cytokines also directly activate the endothelium (24) and increase adherence of circulating T cells with the appropriate phenotype (25, 26, 27). Additionally, activated T cells themselves secrete high levels of chemokines including Mip-1α, Mip-1β, and RANTES (28). This ability of T cells to initiate an inflammatory cascade is not unique. Many studies point to similar abilities of NK cells and the γδ-T cell population as key cells in initiating the early phases of local immune responses in situations where antigen-specific cells do not exist (29, 30). Our model system, in which the Jurkat cells are engineered to recognize a tumor-associated antigen molecule, directly bypasses the conventional TCR/MHC/tumor-associated antigen dependence and suggests a potential targeting strategy in immunotherapy/gene therapy by generating local antigen-specific inflammation. It is important to point out that our model system used human tumor cell lines seeded in nude mice with the adoptive transfer of human T cells. Clearly, this model system does not fully reflect the situation in a human patient where there will also be large numbers of additional T cells present. We are now pursuing experiments with adoptive transfer of T cells in fully syngeneic murine models in which these issues, as well as possible confounding effects such as compatibility of chemokines, cytokines, and tumor cell/endothelial interactions, are no longer relevant.

Our model of established metastatic tumor is likely to consist of multiple, disparate noninflammatory sites. Therefore, it is expected that initially T cells will travel into these tumors at a relatively low rate and be found in much larger numbers in sites such as the spleen (31, 32). Although accumulation of vectors at nontarget sites is an undesirable feature of most vector-delivery approaches, our system has the initial advantage of pharmacological inducibility, such that vector production can be turned off at signs of toxicity. Importantly, it has the second advantage of tight transcriptional specificity, such that expression of the therapeutic gene will occur only in target cells. For this reason, the presence of many vector-carrier cells in sites such as the spleen will not negate the highly efficient local delivery of transgene to the tumor site. In addition to systemic trafficking, our data suggest that antigen-specific T cells set off an inflammatory cascade after signaling through their TCR. The current goal of tumor therapy involving adoptive transfer of T cells is specific tumor cell killing via perforin/granzyme and Fas-mediated mechanisms. This requires that high numbers of tumor-specific cytotoxic T cells must be isolated from patients and activated and expanded in vitro(32, 33, 34). On the basis of our data, a complimentary approach could involve a limited number of tumor specific T cells that could act as “scouts” to detect sites of antigen expression. We hypothesize that these T cells would circulate and traffic into various peripheral tissues and, in response to TCR ligation in target sites these cells, would secrete inflammatory cytokines and chemokines. This inflammatory response within tumors would, in turn, result in endothelial activation and a chemokine gradient to attract viral carrier T cells to the tumor site. This second T cell would need to express the integrins and chemokine receptors that are characteristic of inflammatory homing T cells (25, 26, 27), but need not be antigen specific. The appeal of this approach is that the multiple manipulations required to generate a vector-producing cell would not need to occur in the more precious patient-derived tumor-specific T cells.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was funded by the Mayo Foundation.

7

The abbreviations used are: CIR, chimeric T cell immune receptor; CEA, carcinoembryonic antigen, LTR, long terminal repeat; TCR, T-cell receptor; NFκB, nuclear factor κB; FACS, fluorescence-activated cell sorting; GALV, Gibbon ape leukemia virus; NIP, nitrophenyl; FKBP, FK506-binding protein; FP-AP, FKBP-papamycin-associated proton.

8

Internet address: www.ariad.com/regulationkits.

Fig. 1.

A, retroviral Vectors. pNANA contains the eGFP cDNA cloned into pBabe Puro in place of the SV40-Puro cassette with the 5′- and 3′-LTRs retained as the wild-type LTRs. In pSB-STN-Rapa, the 5′-LTR of pNANA is replaced with the Z12 promoter activated by the dimerizer-regulated transcription system (6, 7). The transcription factor DNA-binding and activation-domain fusion proteins that are activated by rapamycin or analogues (“rapalogs”) are encoded on a separate plasmid, pC4EN-R(H1)S-ZIF3 (pC4; data not shown). In a cell expressing C4, and in the presence of rapamycin or rapalog, transcription is initiated from the Z12 promoter in pSB-STN-Rapa. In addition, the 3′-LTR of pSB-STN-Rapa contains a 365-bp fragment of the human CEA promoter (14) in place of the Mo-MLV viral enhancer and promoter to generate a hybrid LTR as described previously (11, 12). When pSB-STN-Rapa is packaged as free virus and infects a target cell, the enhancer/promoter region of the 3′-LTR forms the template for reverse transcription of the integrated proviral genome such that expression of the eGFP gene is directed by the hybrid LTR (11, 12), that is, by the CEA enhancer/promoter. Location for primers used for PCR analysis in Fig. 2 A are labeled. B, transcriptionally targeted retroviral vectors can be induced by rapamycin from human Jurkat T cells. Jurkat cells previously transfected with the plasmids pSB-STN-Rapa or pNANA and partially selected for high proportions of GFP expressing cells (as described in Materials and Methods.), or Jurkat cells transfected with pBabe Puro and selected in puromycin, were infected with Ad-gag/pol and Ad-GALV adenoviruses at an multiplicity of infection of 100. In addition, some samples were also infected with Ad-C4, an adenovirus encoding the C4 transcription factors, also at a multiplicity of infection of 100. Four hours after infection, cells were washed three times in PBS and replated in normal growth medium. Forty-eight hours later, supernatants were used to infect 5 × 105 CEA+ve HCT116 cells or CEA−ve HT1080 or Mel624 cells, followed 48 h later by FACS analysis for GFP expression (Y axis). Treatments shown below each graph indicate the plasmid initially used to transfect the Jurkat cell populations; +RAPA indicates that the transfections were carried out in the presence of 10 nm rapalog AP21967; Ad-C4 indicates a co-infection with the Ad-C4 adenovirus, as well as the Ad-gag/pol and Ad-GALV adenoviruses with which all samples were treated.

Fig. 1.

A, retroviral Vectors. pNANA contains the eGFP cDNA cloned into pBabe Puro in place of the SV40-Puro cassette with the 5′- and 3′-LTRs retained as the wild-type LTRs. In pSB-STN-Rapa, the 5′-LTR of pNANA is replaced with the Z12 promoter activated by the dimerizer-regulated transcription system (6, 7). The transcription factor DNA-binding and activation-domain fusion proteins that are activated by rapamycin or analogues (“rapalogs”) are encoded on a separate plasmid, pC4EN-R(H1)S-ZIF3 (pC4; data not shown). In a cell expressing C4, and in the presence of rapamycin or rapalog, transcription is initiated from the Z12 promoter in pSB-STN-Rapa. In addition, the 3′-LTR of pSB-STN-Rapa contains a 365-bp fragment of the human CEA promoter (14) in place of the Mo-MLV viral enhancer and promoter to generate a hybrid LTR as described previously (11, 12). When pSB-STN-Rapa is packaged as free virus and infects a target cell, the enhancer/promoter region of the 3′-LTR forms the template for reverse transcription of the integrated proviral genome such that expression of the eGFP gene is directed by the hybrid LTR (11, 12), that is, by the CEA enhancer/promoter. Location for primers used for PCR analysis in Fig. 2 A are labeled. B, transcriptionally targeted retroviral vectors can be induced by rapamycin from human Jurkat T cells. Jurkat cells previously transfected with the plasmids pSB-STN-Rapa or pNANA and partially selected for high proportions of GFP expressing cells (as described in Materials and Methods.), or Jurkat cells transfected with pBabe Puro and selected in puromycin, were infected with Ad-gag/pol and Ad-GALV adenoviruses at an multiplicity of infection of 100. In addition, some samples were also infected with Ad-C4, an adenovirus encoding the C4 transcription factors, also at a multiplicity of infection of 100. Four hours after infection, cells were washed three times in PBS and replated in normal growth medium. Forty-eight hours later, supernatants were used to infect 5 × 105 CEA+ve HCT116 cells or CEA−ve HT1080 or Mel624 cells, followed 48 h later by FACS analysis for GFP expression (Y axis). Treatments shown below each graph indicate the plasmid initially used to transfect the Jurkat cell populations; +RAPA indicates that the transfections were carried out in the presence of 10 nm rapalog AP21967; Ad-C4 indicates a co-infection with the Ad-C4 adenovirus, as well as the Ad-gag/pol and Ad-GALV adenoviruses with which all samples were treated.

Close modal
Fig. 2.

Systemically administered T cells can deliver retrovirus into tumor cells in vivo. HCT116 cells (106) were delivered i.v. into nude mice. On days 7 and day 8, 4 × 106 Jurkat.CEA pSB-STN-Rapa cells infected with Ad-C4, Ad-gag/pol, and Ad GALV were administered i.v., and mice received rapalog i.p. for 5 days. On day 12, organs were removed. Half of each organ was frozen for subsequent genomic DNA isolation and PCR analysis (A), and half was used to prepare a single-cell suspension for FACS analysis for GFP and human class I expression (B). PCR analysis for vector used a 5′ primer located within the rapamycin promoter (primer 1) and a 3′ primer located within the pBabePuro backbone 5′ of the transgene (primer 2; Fig. 1,A). PCR analysis for provirus used a 5′ primer located within the CEA promoter (primer 3) and primer 2 (Fig. 1 A). Primer sequences can be provided on request. A, Lane 1, brain; Lane 2, kidney; Lane 3, small intestine; Lane 4, blood; Lane 5, liver; Lane 6, Lung; Lane 7, spleen. B, GFP+ non-tumor: gated GFP+ HLA-ABC; GFP− tumor: gated GFP HLA-ABC+; GFP+ tumor: gated GFP+ HLA-ABC+; Total tumor: gated HLA-ABC+.

Fig. 2.

Systemically administered T cells can deliver retrovirus into tumor cells in vivo. HCT116 cells (106) were delivered i.v. into nude mice. On days 7 and day 8, 4 × 106 Jurkat.CEA pSB-STN-Rapa cells infected with Ad-C4, Ad-gag/pol, and Ad GALV were administered i.v., and mice received rapalog i.p. for 5 days. On day 12, organs were removed. Half of each organ was frozen for subsequent genomic DNA isolation and PCR analysis (A), and half was used to prepare a single-cell suspension for FACS analysis for GFP and human class I expression (B). PCR analysis for vector used a 5′ primer located within the rapamycin promoter (primer 1) and a 3′ primer located within the pBabePuro backbone 5′ of the transgene (primer 2; Fig. 1,A). PCR analysis for provirus used a 5′ primer located within the CEA promoter (primer 3) and primer 2 (Fig. 1 A). Primer sequences can be provided on request. A, Lane 1, brain; Lane 2, kidney; Lane 3, small intestine; Lane 4, blood; Lane 5, liver; Lane 6, Lung; Lane 7, spleen. B, GFP+ non-tumor: gated GFP+ HLA-ABC; GFP− tumor: gated GFP HLA-ABC+; GFP+ tumor: gated GFP+ HLA-ABC+; Total tumor: gated HLA-ABC+.

Close modal
Fig. 3.

Systemically administered Jurkat cells can be used to deliver retrovirus expressing a transcriptionally targeted cytotoxic gene for the treatment of metastatic disease. HCT116 (106) or Mel624 (2 × 106) cells were delivered i.v. into nude mice. On days 7 and day 8, 4 × 106 Jurkat.CEA or Jurkat.NIP cells pre-engineered to produce pSB-STN-(Rapa)-tk virus after induction with rapalog were injected i.v. The mice were then given ganciclovir treatment for 5 consecutive days at 50 mg/kg with or without rapalog i.p. Mice were examined daily and sacrificed when they developed evidence of prominent metastatic disease. Survival data are shown for HCT116 (A) and Mel624 (B) tumors. ▪, J.CEA pSB-STN-Rapa-tk + rapalog; ▴, J.NIP pSB-STN-Rapa-tk + rapalog; ▾, J.CEA pSB-STN-Rapa-tk no rapalog; ♦, J.NIP pSB-STN-Rap-tk no rapalog. In HCT116 treated with J.CEA pSB-STN-Rapa-tk + rapalog P < 0.005 compared with J.NIP pSB-STN-Rapa-tk + rapalog.

Fig. 3.

Systemically administered Jurkat cells can be used to deliver retrovirus expressing a transcriptionally targeted cytotoxic gene for the treatment of metastatic disease. HCT116 (106) or Mel624 (2 × 106) cells were delivered i.v. into nude mice. On days 7 and day 8, 4 × 106 Jurkat.CEA or Jurkat.NIP cells pre-engineered to produce pSB-STN-(Rapa)-tk virus after induction with rapalog were injected i.v. The mice were then given ganciclovir treatment for 5 consecutive days at 50 mg/kg with or without rapalog i.p. Mice were examined daily and sacrificed when they developed evidence of prominent metastatic disease. Survival data are shown for HCT116 (A) and Mel624 (B) tumors. ▪, J.CEA pSB-STN-Rapa-tk + rapalog; ▴, J.NIP pSB-STN-Rapa-tk + rapalog; ▾, J.CEA pSB-STN-Rapa-tk no rapalog; ♦, J.NIP pSB-STN-Rap-tk no rapalog. In HCT116 treated with J.CEA pSB-STN-Rapa-tk + rapalog P < 0.005 compared with J.NIP pSB-STN-Rapa-tk + rapalog.

Close modal
Fig. 4.

Systemically administered Jurkat.CEA cells recruit Jurkat cells to liver and s.c. tumors (see Table 1 also). HCT116 cells (106) were delivered i.v. or s.c. into nude mice. On days 7 and day 8, 4 × 106 Jurkat.CEA or Jurkat.NIP were co-administered into mice with 4 × 106 Cell Tracker Orange labeled Jurkat cells. Two days later livers were removed from mice receiving i.v. tumors and examined by confocal for evidence of orange cells. Livers were scored as + if they contained evidence of multiple areas of orange cells and − if they contained no such areas (Table 1). s.c. tumors were removed 7 days after T-cell injection and examined by confocal microscopy for evidence of orange cells. Representative confocal fluorescence images of tumors from mice receiving Jurkat.CEA or Jurkat.NIP cells are shown. Arrows indicate foci of labeled cells.

Fig. 4.

Systemically administered Jurkat.CEA cells recruit Jurkat cells to liver and s.c. tumors (see Table 1 also). HCT116 cells (106) were delivered i.v. or s.c. into nude mice. On days 7 and day 8, 4 × 106 Jurkat.CEA or Jurkat.NIP were co-administered into mice with 4 × 106 Cell Tracker Orange labeled Jurkat cells. Two days later livers were removed from mice receiving i.v. tumors and examined by confocal for evidence of orange cells. Livers were scored as + if they contained evidence of multiple areas of orange cells and − if they contained no such areas (Table 1). s.c. tumors were removed 7 days after T-cell injection and examined by confocal microscopy for evidence of orange cells. Representative confocal fluorescence images of tumors from mice receiving Jurkat.CEA or Jurkat.NIP cells are shown. Arrows indicate foci of labeled cells.

Close modal
Fig. 5.

Factors secreted by Jurkat.CEA cells cocultured with HCT116 tumor cells activate viral production from a NFκB regulated retrovirus. HCT116 cells (105) were cocultured with 106 Jurkat.CEA or Jurkat.NIP cells. Twenty-four hours later, cell-free supernatants were transferred onto Jurkat.CEA pSB-STN-GFP-(NFκB)3 or Jurkat.NIP pSB-STN-GFP-(NFκB)3 both infected with Ad-GALV and Ad-gag/pol. Twenty-four hours later, cell-free supernatants were collected and cultured with HCT116 cells. Forty-eight hours later, HCT116 cells were FACS analyzed for GFP expression (Y axis). Lower plots show HCT116 cells cocultured with Jurkat.CEA retrovirus-producer clones or Jurkat.NIP retrovirus-producer clones for 48 h.

Fig. 5.

Factors secreted by Jurkat.CEA cells cocultured with HCT116 tumor cells activate viral production from a NFκB regulated retrovirus. HCT116 cells (105) were cocultured with 106 Jurkat.CEA or Jurkat.NIP cells. Twenty-four hours later, cell-free supernatants were transferred onto Jurkat.CEA pSB-STN-GFP-(NFκB)3 or Jurkat.NIP pSB-STN-GFP-(NFκB)3 both infected with Ad-GALV and Ad-gag/pol. Twenty-four hours later, cell-free supernatants were collected and cultured with HCT116 cells. Forty-eight hours later, HCT116 cells were FACS analyzed for GFP expression (Y axis). Lower plots show HCT116 cells cocultured with Jurkat.CEA retrovirus-producer clones or Jurkat.NIP retrovirus-producer clones for 48 h.

Close modal
Table 1

Systemically administered Jurkat.CEA cells recruit Jurkat cells to liver and s.c. tumors

Orange (number of positive mice/total mice)
Jurkat.CEA + Jurkat (orange) 3/3 
Jurkat.NIP + Jurkat (orange) 0/3 
PBS 0/1 
Jurkat.CEA 0/2 
Jurkat (orange) 0/3 
Orange (number of positive mice/total mice)
Jurkat.CEA + Jurkat (orange) 3/3 
Jurkat.NIP + Jurkat (orange) 0/3 
PBS 0/1 
Jurkat.CEA 0/2 
Jurkat (orange) 0/3 

We thank Toni Higgins for her help in preparing the manuscript.

1
Harrington K., Alvarez-Vallina L., Crittenden M., Gough M., Chong H., Diaz R. M., Vassaux G., Lemoine N., Vile R. Cells as vehicles for cancer gene therapy: the missing link between targeted vectors and systemic delivery?.
Hum. Gene Ther.
,
13
:
1263
-1280,  
2002
.
2
Verma I., Somia N. Gene therapy–promises, problems and prospects.
Nature
,
389
:
239
-242,  
1997
.
3
Vile R. G., Russell S., Lemoine N. Cancer gene therapy: hard lessons and new courses.
Gene Ther.
,
7
:
2
-8,  
2000
.
4
Chester J., Ruchatz A., Gough M., Crittenden M., Chong H., Loic-Cosset F., Diaz R. M., Harrington K., Alvarez-Vallina L., Vile R. Tumor antigen-specific induction of transcriptionally targeted retroviral vectors from chimeric immune receptor-modified T cells.
Nat. Biotechnol.
,
20
:
256
-263,  
2002
.
5
Alvarez-Vallina L., Yanez R., Blanco B., Gil M., Russell S. J. Pharmacologic suppression of target cell recognition by engineered T cells expressing chimeric T-cell receptors.
Cancer Gene Ther.
,
7
:
526
-529,  
2000
.
6
Rivera V. M., Clackson T., Natesan S., Pollock R., Amara J. F., Keenan T., Magari S. R., Phillips T., Courage N. L., Cerasoli F., Jr., Holt D. A., Gilman M. A humanized system for pharmacologic control of gene expression.
Nat. Med.
,
2
:
1028
-1032,  
1996
.
7
Pollock R., Issner R., Zoller K., Natesan S., Rivera V. M., Clackson T. Delivery of a stringent dimerizer-regulated gene expression system in a single retroviral vector.
Proc. Natl. Acad. Sci. USA
,
97
:
13221
-13226,  
2000
.
8
Morgenstern J. P., Land H. Advanced mammalian gene transfer: high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line.
Nucl. Acids Res.
,
18
:
3587
-3596,  
1990
.
9
Wagner M. J., Sharp J. A., Summers W. C. Nucleotide sequence of the thymidine kinase gene of herpes simplex virus type 1.
Proc. Natl. Acad. Sci. USA
,
78
:
1441
-1445,  
1981
.
10
Vile R. G., Miller N., Chernajovsky Y., Hart I. R. A comparison of the properties of different retroviral vectors containing the murine tyrosinase promoter to achieve transcriptionally targeted expression of the HSVtk or IL-2 genes.
Gene Ther.
,
1
:
307
-316,  
1994
.
11
Diaz R. M., Eisen T., Hart I. R., Vile R. G. Exchange of viral promoter/enhancer elements with heterologous regulatory sequences generates targeted hybrid long terminal repeat vectors for gene therapy of melanoma.
J. Virol.
,
72
:
789
-795,  
1998
.
12
Emiliusen L., Gough M., Bateman A., Ahmed A., Voellmy R., Chester J., Diaz R. M., Harrington K., Vile R. A transcriptional feedback loop for tissue-specific expression of highly cytotoxic genes which incorporates an immunostimulatory component.
Gene Ther.
,
8
:
987
-988,  
2001
.
13
Duisit G., Salvetti A., Moullier P., Cosset F. L. Functional characterization of adenoviral/retroviral chimeric vectors and their use for efficient screening of retroviral producer cell lines.
Hum. Gene Ther.
,
10
:
189
-200,  
1999
.
14
Richards C. A., Austin E. A., Huber B. E. Transcriptional regulatory sequences of carcinoembryonic antigen: identification and use with cytosine deaminase for tumor-specific gene therapy.
Hum. Gene Ther.
,
6
:
881
-893,  
1995
.
15
Murali-Krishna K., Altman J. D., Suresh M., Sourdive D., Zajac A., Ahmed R. In vivo dynamics of anti-viral CD8 T cell responses to different epitopes. An evaluation of bystander activation in primary and secondary responses to viral infection.
Adv. Exp. Med. Biol.
,
452
:
123
-142,  
1998
.
16
Flynn K. J., Belz G. T., Altman J. D., Ahmed R., Woodland D. L., Doherty P. C. Virus-specific CD8+ T cells in primary and secondary influenza pneumonia.
Immunity
,
8
:
683
-691,  
1998
.
17
Topham D. J., Castrucci M. R., Wingo F. S., Belz G. T., Doherty P. C. The role of antigen in the localization of naive, acutely activated, and memory CD8(+) T cells to the lung during influenza pneumonia.
J. Immunol.
,
167
:
6983
-6990,  
2001
.
18
Yee C., Thompson J. A., Roche P., Byrd D. R., Lee P. P., Piepkorn M., Kenyon K., Davis M. M., Riddell S. R., Greenberg P. D. Melanocyte destruction after antigen-specific immunotherapy of melanoma: direct evidence of t cell-mediated vitiligo.
J. Exp. Med.
,
192
:
1637
-1644,  
2000
.
19
Pockaj B. A., Sherry R. M., Wei J. P., Yannelli J. R., Carter C. S., Leitman S. F., Carasquillo J. A., Steinberg S. M., Rosenberg S. A., Yang J. C. Localization of 111indium-labeled tumor infiltrating lymphocytes to tumor in patients receiving adoptive immunotherapy. Augmentation with cyclophosphamide and correlation with response.
Cancer
,
73
:
1731
-1737,  
1994
.
20
Guidotti L. G., Chisari F. V. Noncytolytic control of viral infections by the innate and adaptive immune response.
Annu. Rev. Immunol.
,
19
:
65
-91,  
2001
.
21
Guidotti L. G., Ando K., Hobbs M. V., Ishikawa T., Runkel L., Schreiber R. D., Chisari F. V. Cytotoxic T lymphocytes inhibit hepatitis B virus gene expression by a noncytolytic mechanism in transgenic mice.
Proc. Natl. Acad. Sci. USA
,
91
:
3764
-3768,  
1994
.
22
Fujiie S., Hieshima K., Izawa D., Nakayama T., Fujisawa R., Ohyanagi H., Yoshie O. Proinflammatory cytokines induce liver and activation-regulated chemokine/macrophage inflammatory protein-3alpha/CCL20 in mucosal epithelial cells through NF-kappaB [correction of NK-kappaB].
Int. Immunol.
,
13
:
1255
-1263,  
2001
.
23
Vaday G. G., Franitza S., Schor H., Hecht I., Brill A., Cahalon L., Hershkoviz R., Lider O. Combinatorial signals by inflammatory cytokines and chemokines mediate leukocyte interactions with extracellular matrix.
J. Leukoc. Biol.
,
69
:
885
-892,  
2001
.
24
Boehm U., Klamp T., Groot M., Howard J. C. Cellular responses to interferon-gamma.
Annu. Rev. Immunol.
,
15
:
749
-795,  
1997
.
25
Cerwenka A., Morgan T. M., Harmsen A. G., Dutton R. W. Migration kinetics and final destination of type 1 and type 2 CD8 effector cells predict protection against pulmonary virus infection.
J. Exp. Med.
,
189
:
423
-434,  
1999
.
26
Austrup F., Vestweber D., Borges E., Lohning M., Brauer R., Herz U., Renz H., Hallmann R., Scheffold A., Radbruch A., Hamann A. P- and E-selectin mediate recruitment of T-helper-1 but not T-helper-2 cells into inflamed tissues.
Nature
,
385
:
81
-83,  
1997
.
27
Sallusto F., Lenig D., Mackay C. R., Lanzavecchia A. Flexible programs of chemokine receptor expression on human polarized T helper 1 and 2 lymphocytes.
J. Exp. Med.
,
187
:
875
-883,  
1998
.
28
Schrum S., Probst P., Fleischer B., Zipfel P. F. Synthesis of the CC-chemokines MIP-1alpha, MIP-1beta, and RANTES is associated with a type 1 immune response.
J. Immunol.
,
157
:
3598
-3604,  
1996
.
29
Biron C. A., Nguyen K. B., Pien G. C., Cousens L. P., Salazar-Mather T. P. Natural killer cells in antiviral defense: function and regulation by innate cytokines.
Annu. Rev. Immunol.
,
17
:
189
-220,  
1999
.
30
Allison J. P., Havran W. L. The immunobiology of T cells with invariant gamma delta antigen receptors.
Annu. Rev. Immunol.
,
9
:
679
-705,  
1991
.
31
Economou J. S., Belldegrun A. S., Glaspy J., Toloza E. M., Figlin R., Hobbs J., Meldon N., Kaboo R., Tso C. L., Miller A., Lau R., McBride W., Moen R. C. In vivo trafficking of adoptively transferred interleukin-2 expanded tumor-infiltrating lymphocytes and peripheral blood lymphocytes. Results of a double gene marking trial.
J. Clin. Investig.
,
97
:
515
-521,  
1996
.
32
Yee C., Riddell S. R., Greenberg P. D. In vivo tracking of tumor-specific T cells.
Curr. Opin. Immunol.
,
13
:
141
-146,  
2001
.
33
Yee C., Gilbert M. J., Riddell S. R., Brichard V. G., Fefer A., Thompson J. A., Boon T., Greenberg P. D. Isolation of tyrosinase-specific CD8+ and CD4+ T cell clones from the peripheral blood of melanoma patients following in vitro stimulation with recombinant vaccinia virus.
J. Immunol.
,
157
:
4079
-4086,  
1996
.
34
Yannelli J. R., Hyatt C., McConnell S., Hines K., Jacknin L., Parker L., Sanders M., Rosenberg S. A. Growth of tumor-infiltrating lymphocytes from human solid cancers: summary of a 5-year experience.
Int. J. Cancer
,
65
:
413
-421,  
1996
.