Synergistic Enhancement of Antitumor Immunity with Adoptively Transferred Tumor-specific CD4⁺ and CD8⁺ T Cells and Intratumoral Lymphotactin Transgene Expression

INTRODUCTION

CTLs play a crucial role in the host immune response to cancer. Effective adoptive cancer immunotherapy with tumor-sensitized CTLs has been well documented in animal models (1, 2, 3, 4) , where transfer of such tumor-specific cells into mice bearing established tumors has resulted in tumor eradication. However, even in such model systems, this therapeutic approach is limited to early-stage tumors (e.g., 3- or 10-day lung metastases) or established s.c. tumors that have been previously irradiated (to facilitate T-cell infiltration; Refs. 3, 4, 5, 6 ). In clinical settings, only limited numbers of patients have responded to CTL therapy (7, 8, 9) , with an objective response rate of ∼30% (10) . In general, it is assumed that the antitumor efficacy of the transferred T cells is, to a large extent, determined by their ability to leave the vasculature and infiltrate the tumor (5 , 6) , but the overall fraction of transferred T cells that accumulate in tumors is rather small (11 , 12) . The resistance of tumors to T-cell infiltration may be attributable in part to limited expression on the neovasculature within these growths of the adhesion molecules that are essential for T-cell adherence and transendothelial migration, e.g., vascular cell adhesion molecule-1 and intracellular adhesion molecule-1 (13) .

The trafficking of lymphocytes from the systemic circulation into tissues is a dynamic, multistep process. It requires selectin-mediated rolling and tethering, lymphokine-induced activation of integrins, firm adhesion of the lymphocytes to endothelial cells and their diapedesis through the endothelium, as well as migration within the connective tissues along established chemoattractant gradients (14) . Although adhesion to endothelial cells and various extracellular matrix proteins is critical for successful T-cell recruitment into inflammatory sites, the signals that regulate these processes have not been fully elucidated. It is believed that essential steps are mediated by chemokines produced at the sites of inflammation (15) .

Chemokines are a superfamily of cytokines that attract and activate leukocytes (16) . They are produced by multiple cell types (e.g., leukocytes, endothelial cells, fibroblasts, and tumor cells) in response to viruses, bacteria, lipopolysaccharide, and proinflammatory cytokines (e.g., IL-1 3 and TNF-α; Ref. 17 ). The superfamily’s four major branches are defined by the spacing of the first two cysteines in a conserved four-cysteine motif. The two cysteines of the α subfamily (C-X-C) are separated by another residue and those of the β subfamily (C-C) are adjacent, whereas the γ subfamily (C) has only one cysteine at its NH₂ terminus (18) . The newly identified C-X₃-C subfamily has two cysteines separated by three other residues (19) . Generally, C-X-C chemokines are potent activators and chemoattractants for neutrophils, whereas the C-C chemokines have the potential to chemoattract monocytes and T lymphocytes (17) . The C chemokine Lptn was originally reported to induce T and NK cell, but not monocyte, migration in vitro (20) through its interactions with the G-protein-coupled, seven-transmembrane domain receptor XCR-1 (21) . Recently, it has been reported that Lptn also chemoattracts neutrophils and B cells expressing XCR-1 (22) and acts as an innate mucosal adjuvant (23) .

More recently, we engineered a mouse myeloma cell line, SP2/0, to express a Lptn transgene. We found that such engineered SP2/0-Lptn tumor cells invariably regressed when implanted into syngeneic BALB/c mice, after Lptn-dependent infiltration of the tumors by CD4⁺ and CD8⁺ T cells and neutrophils (24) . Furthermore, the wild-type SP2/0 cells induced very little in the way of effector immune responses after tumor inoculation, such that neither IFN-γ nor significant tumor-specific CTL responses could be discerned in these animals. Because Lptn can induce T-cell migration in vitro (20) and in vivo (24 , 25) , we reasoned that Lptn expression within tumors, induced by adenovirus-mediated Lptn gene transfer, should enhance T-cell infiltration and thus the therapeutic efficiency of adoptively transferred tumor-specific T cells. In the present study we tested this hypothesis. We phenotypically characterized the SP2/0-specific T lymphocytes, studied the chemotactic and functional effects of Lptn on these cells, and then studied the efficacy of adoptive T-cell therapy using these activated T cells in combination with intratumoral adenovirus-mediated Lptn gene transfer.

MATERIALS AND METHODS

Antibodies, Cell Lines, and Animals.

Monoclonal antimouse MHC class I (H-2K^d) and II (Ia^d) antigens; anti-CD3, -CD4, -CD8, -CD25, -l-selectin, and -FasL antibodies, and the PE-conjugated antimouse IL-4, IFN-γ, perforin, CD11c, and TNF-α antibodies were all obtained from Pharmingen, Inc. (Mississauga, Ontario, Canada). FITC-conjugated goat antirat and antimouse IgG antibodies were obtained from Bio/Can Scientific (Mississauga, Ontario, Canada). Recombinant mouse Lptn, IL-2, and IFN-γ, as well as the polyclonal goat anti-Lptn and MIP-1β antibodies, were purchased from R&D Systems (Minneapolis, MN).

The wild-type myeloma cell line SP2/0, B-cell lymphoma cell line A20 (of BALB/c mouse origin), and the mouse lymphoma cell line Yac-1 were obtained from the American Type Culture Collection (Rockville, MD) and maintained in DMEM containing 10% FCS and 50 μg/ml gentamicin. The wild-type SP2/0 tumor cells displayed cell surface expression of H-2K^d, but not Ia^d. To stimulate Ia^d expression, SP2/0 cells were cultured in DMEM containing 10% FCS and 50 ng/ml IFN-γ (SP2/0/IFN-γ cells) for 3 days and then harvested for flow cytometric analysis and cytotoxicity assay. For some experiments, SP2/0 in vivo tumor cell suspensions were freshly prepared from SP2/0 tumors grown in BALB/c mice by mincing the tumor masses into small pieces and pressing them through a fine mesh. Splenic B cells were prepared from the plastic nonadherent splenocytes of BALB/c mice, using antimouse pan-B cell Dynabeads (DYNAL Inc., Lake Success, NY) according to the supplier’s protocol (22) . B-Cell blasts expressing Ia^d (data not shown) were prepared by culturing these B cells with lipopolysaccharide (0.2 μg/ml; Sigma-Aldrich Canada Ltd, Oakville, Ontario, Canada;) Ref. (26) . The engineered tumor cell line SP2/0-Lptn, which secretes the C chemokine Lptn, was generated in our laboratory by transfecting SP2/0 cells with the plasmid pCIneo-Lptn (24) . The IL-4-secreting engineered tumor cell line J558-IL-4 (27) was provided by Dr. Hook, Harvard Medical School (Boston, MA). Female BALB/c mice (4–6 weeks of age) were obtained from the Animal Resource Center, University of Saskatchewan. Female athymic nude mice (4–6 week of age) were obtained from Charles River Laboratories (St. Constant, Quebec, Canada). All mice were maintained in the animal facility at the Saskatoon Cancer Center.

Construction of Recombinant Adenoviral Vector AdVLptn.

The 1-kb cDNA fragment of Lptn, obtained from DNAX (Palo Alto, CA; Ref. (24) , was ligated into the pLpA vector (28) to form pLpA-Lptn. The recombinant adenovirus AdVLptn was generated by cotransfecting 293 cells with the pLpA-Lptn and pJM17 vectors (28) . Homologous recombinants of the two plasmids were ultimately packaged into AdV capsids released into the medium. Clones of AdVLptn were obtained by limiting-dilution analysis of the cytopathic effects of this conditioned medium on fresh 293 cells. Working stocks of the virus were obtained by harvesting 293 cells at 36 h after infection with the appropriate seed stock; at this time, the transfected 293 cells exhibited cytopathic effects, but had not yet undergone lysis. The cells were lysed by three freeze-thaw cycles, and the adenovirus was then purified by two rounds of cesium chloride density gradient ultracentrifugation and subsequently stored at −80°C until used. AdVLacZ, which contains the Escherichia coli β-gal marker gene under the transcriptional control of the human cytomegalovirus promoter, and a control adenovirus AdVpLpA with no transgene insert were constructed in our laboratory as reported previously (28) . Each of these recombinant viruses is an E1-deleted replication-deficient human type-V adenovirus.

Adenoviral Infection of SP2/0 Tumors in Vivo.

The ability of AdVLacZ to infect SP2/0 tumors in vivo was investigated. Briefly, 0.5 × 10⁶ SP2/0 cells were s.c. injected into each athymic nude mouse. Twelve days later, when the SP2/0 tumors were ∼8 mm in diameter, 50-μl bolus aliquots of AdVLacZ virus (2 × 10⁹ pfu or 5-fold dilutions thereof) were injected into the SP2/0 tumors. One day after viral injection, each SP2/0 tumor was removed and cut into three approximately equal sections. These tumor tissues were mounted in tissue-freezing medium (Triangle Biomedical Sciences, Durham, NC) and snap-frozen by immersion in 2-methylbutane (J. T. Baker, Phillipsburg, NJ), that had been chilled over liquid nitrogen. Frozen 6-μm sections were obtained for the analysis of β-galactosidase expression (22) . To accomplish this, transfected SP2/0 tumor tissue sections were fixed in PBS containing 37% formaldehyde and 25% glutaraldehyde, stained with X-gal, and then counterstained with nuclear fast red (Poly Scientific, Bay Shore, NY). The mean proportion of blue-staining SP2/0 cells from three tumor sections was taken as the percentage of transfection. We observed a dose-dependent response to the adenovirus, with maximum staining (∼45% of cells) at 2 × 10⁹ pfu AdVLacZ. Thereafter, 2 × 10⁹ pfu was selected as the optimal dose for delivery of AdVLptn to tumors in vivo.

Northern and Western Blot Analysis.

To examine Lptn expression after in vivo AdVLptn transfection, 12-day SP2/0 tumor masses from nude mice were injected with 2 × 10⁹ pfu of AdVLptn, and 1 day later these tumors, or AdVpLpA-infected SP2/0 control tumors, were removed for extraction of proteins (below) or RNA, using a commercial kit (Qiagen Inc., Mississauga, Ontario, Canada). The RNA was denatured in formaldehyde and separated on agarose-formaldehyde gels by electrophoresis, visualized by UV illumination to assess its integrity, and transferred to Nytran membranes (Schleicher & Schuell, Keane, NH). The EcoRI Lptn cDNA fragment, labeled with [α-³²P]dCTP (Amersham, Arlington Heights, IL) by random priming, was used as a probe. The membranes were hybridized overnight at 42°C in 6× SSC-5× Denhardt’s solution containing 0.5% SDS and 20 μg/ml salmon sperm DNA. The filters were washed sequentially at 65°C in 3× SSC-0.1% SDS, 1× SSC-0.1% SDS, 0.3× SSC-0.1% SDS, and 0.1× SSC-0.1% SDS, and then exposed to Kodak X-ray film until the desired signal densities were obtained.

The protein extracts were prepared by homogenization of the tumor tissues in 125 mm tris, 0.05% SDS, 10% β-mercaptoethanol, followed by centrifugation at 1000 × g for 5 min. The supernatants containing the protein samples were electrophoresed through 10% polyacrylamide gels and electrophoretically transferred onto nitrocellulose papers. The blots were blocked with 10% BSA in PBS and incubated with the goat anti-Lptn antibody, followed by peroxidase-conjugated rabbit antigoat IgG, and finally with an enhanced chemiluminescence reagent (New England Nuclear Life Science Products, Boston, MA), according to the manufacturer’s protocol.

Preparation of Activated SP2/0 Tumor-specific T Lymphocytes.

Spleens were removed from immunized mice that had previously rejected SP2/0-Lptn tumors (24) , and single-cell suspensions were prepared by pressing the spleens through fine nylon mesh. The red blood cells were lysed using 0.84% ammonium chloride. The splenic lymphocytes (5 × 10⁶) were cocultured in 24-well plates (Costar Corp, Cambridge, MA) with 1 × 10⁵ γ-irradiated (6000 rad) SP2/0 cells in 2 ml of DMEM supplemented with 10% FCS and IL-2 (10 U/ml; DMEM-FCS-IL-2 medium). After 4 days, the lymphocytes were purified from the cultures using Ficoll-Paque density gradient centrifugation, and for some experiments, the CD4⁺ and CD8⁺ T cells were fractionated by negative selection using antimouse CD8 (Ly2) and CD4 (L3T4) paramagnetic beads, respectively (DYNAL Inc.), according to the manufacturer’s protocols. Hereafter, these populations of cells are referred to as CD4⁺ and CD8⁺ T cells, respectively. The unfractionated T cells, as well as the purified CD4⁺ or CD8⁺ cells, were then subject to (a) phenotypic analysis by flow cytometry and cytokine ELISA, (b) functional analyses using chemotaxis and cytotoxicity assays, and (c) animal studies.

Purification and Phenotypic Characterization of Activated CD4⁺ and CD8⁺ Lymphocytes and SP2/0 Tumor Cells.

In the phenotypic analyses, splenic T lymphocytes from naive BALB/c mice were used as negative-control cells. Red cells were lysed using 0.84% ammonium chloride, and the T cells were purified on nylon wool columns as described previously (26) . Naive or activated T lymphocytes were incubated for 1 h on ice with rat antimouse antibodies against CD3, CD4, CD8, CD25, L-selectin, or FasL (each, 2 μg/ml), washed with PBS, and then incubated for 1 h on ice with FITC-conjugated antirat IgG antibody (5 μg/ml). For phenotypic analysis of SP2/0 cells, the wild-type SP2/0, IFN-γ-cultivated SP2/0/IFN-γ, and SP2/0/in vivo tumor cells were incubated for 1 h on ice with monoclonal antimouse antibodies against H-2K^d and Ia^d (each, 2 μg/ml), washed with PBS, and then incubated for 1 h on ice with FITC-conjugated antimouse IgG antibody (5 μg/ml). After another three washes with PBS, the above T cells and SP2/0 tumor cells were analyzed by flow cytometry. Isotype-matched irrelevant specificity monoclonal antibodies were used as controls. CD4⁺ and CD8⁺ T-cell subpopulations were assessed for purity by flow cytometry using FITC-anti-CD4 and -CD8, and PE-anti-CD11c antibodies. These CD4⁺ and CD8⁺ populations each contained 2–3% CD11c⁺ cells (i.e., APCs).

Cytokine Expression and Secretion.

To examine the intracellular expression of cytokines by the purified CD4⁺ and CD8⁺ T cells, the cells were processed using a commercial kit (Cytofix/CytoPerm Plus with GolgiPlug; Pharmingen, Inc), and stained with PE-conjugated anti-IL-4, -IFN-γ, -TNF-α, or -perforin antibodies, according to the manufacturers’ protocols. Cytokine secretion by the cells was assessed by ELISA, using commercial kits (Endogen, Woburn, MA). The cells were cultured at 0.5 × 10⁶ cell/well in DMEM-FCS-IL-2 medium in flat-bottom 96-well plates (Costar Corp), and the culture supernatants were harvested at 1 and 3 days for IFN-γ and IL-4 analyses. The results were normalized to the recombinant cytokine standard curves (29) .

Chemotaxis Assay.

The chemotactic responses of activated T lymphocytes to Lptn were examined using modified Boyden microchemotaxis chambers (Neuroprobe, Gaithersburg, MD) and polyvinyl pyrrolidone-free 5 μm pore size polycarbonate membranes, essentially as described previously (24) . Monocytes, collected by peritoneal lavage of normal BALB/c mice receiving i.p. injections of 1.5 ml of thioglycollate broth (Becton-Dickonson, San Jose, CA; Ref. (30) 4 days previously, were used as a negative control in the chemotaxis assays. To purify the monocytes, the peritoneal lavage cells were washed with PBS and plated into Petri dishes for 1 h in serum-free DMEM; the nonadherent cells were then removed and the adherent cells harvested. More than 98% of these cells were macrophages, as determined by morphological and phagocytic criteria. Recombinant Lptn, diluted in DMEM-0.1% BSA to 0.1–1000 ng/ml, was added to triplicate lower chambers of the wells, and 10⁵ activated T lymphocytes or monocytes in DMEM-BSA were added to the upper chambers. After incubation for 2 h at 37°C, the cells that had not migrated into the membranes were wiped from the upper surfaces of the membranes, which were then fixed in 70% methanol and stained using a Diff-Quik kit (American Scientific Products, McGraw Hill, IL). The lymphocytes or monocytes that were associated with the membranes were enumerated by direct counting in a blinded fashion of at least nine × 40 objective fields per well. The results are expressed as the mean number of cells/× 40 field (± SE).

RT-PCR.

Total RNA was obtained from activated T lymphocytes and monocytes as noted above. The first-strand cDNA synthesis was performed with 5 μg of total RNA using a RT-PCR kit (Stratagene, La Jolla, CA), according to the manufacturer’s instructions. Two sets of PCR primers were used, including the sense (5′-ctcct gtcta ctgcc tgtgt tg-3′) and antisense (5′-tgact gttcg gtgtc tctgt ct-3′) primers for the Lptn receptor gene XCR-1 (21) and the sense (5′-caggt tgtct cctgc gactt-3′) and antisense (5′-cttgc tcagt gtcct tgctg-3′) primers for the control gene GAPDH. The protocol for amplification of both mRNA species included 1 cycle of 94°C for 5 min, 54°C for min, and 72°C for 1 min; and 40 cycles of 94°C for 1 min, 54°C for 1 min, and 72°C for 1 min. All PCR reaction products were resolved using ethidium bromide-stained 1% agarose gels.

T-Cell Proliferation Assay.

Purified in vitro-activated CD4⁺ and CD8⁺ T cells were incubated at 0.5 × 10⁶ cells/well in DMEM-FCS-IL-2 medium in flat-bottom 96-well plates, alone or with Lptn (0.1–1000 ng/ml). To confirm the specificity of the Lptn effects on the T cells, rabbit anti-Lptn antibody (15 μg/ml) was added to one set of wells. After 48 h, all wells were pulsed for 18 h with 1 μCi of [³H]thymidine (Amersham) and then harvested onto glass fiber filters. Thymidine incorporation was determined by liquid scintillation counting.

T-Cell Cytotoxicity Assay.

In the T-cell cytotoxicity assays, the freshly activated T lymphocytes or their purified CD4⁺ or CD8⁺ T-cell subsets were used as effector cells, whereas ⁵¹Cr-labeled SP2/0 or irrelevant control A20 tumor cells or B-cell blasts were used as target cells. The target cells were radiolabeled by culture for 1 h in the presence of 50 μl of sodium [⁵¹Cr]chromate (36 mCi/ml; Amersham), then washed twice with DMEM. Approximately 10⁵ labeled target cells per triplicate well were mixed with effector cells at various E:T cell ratios, either alone or with exogenous Lptn (1 μg/ml) or anti-H-2K^d (15 μg/ml) or anti-Ia^d antibody (15 μg/ml) added, and were incubated for 6 h. The percentage of specific lysis was calculated as: 100 × [(experimental cpm − spontaneous cpm)/(maximal cpm − spontaneous cpm)]. Spontaneous cpm release in the absence of effector cells was <10% of specific lysis. The maximal cpm release was determined by lysis of the target cells with 1% Triton X-100.

Assessment of T-Cell Infiltration of the SP2/0 Tumors in Vivo.

To track T-cell migration into the tumors, on experimental day 0 we injected 2 × 10⁹ pfu of AdVLptn and AdVpLpA into well-established (i.e., 10-day) SP2/0 tumors in the right and left thighs, respectively, of the mice. The next day, 5 × 10^{6 51}Cr-labeled activated T lymphocytes (labeled as described above) were injected into the tail vein of each mouse, and 2 or 24 h later the mice were sacrificed. Each tumor and a panel of organs or tissues were removed and weighed; the associated radioactivity was then determined using a gamma counter. The biodistribution of labeled T cells was expressed as the localization index (cpm/g of tissue). To unequivocally confirm the infiltration of the tumors by intact T cells, in a second set of experiments, the T cells were labeled by incubation with TRITC dye (2 μg/ml:DMEM; Sigma Chemical Co; Ref. (31) for 45 min at 37°C. The TRITC-labeled T cells were washed three times with PBS and injected i.v. into the mice (1 × 10⁷ cells/mouse). One day later, the mice were sacrificed, and their tumors were removed and processed to 5-μm frozen sections. The TRITC-labeled T-cell infiltration of the tumors was assessed by fluorescence microscopy, whereas histologic analyses were performed using frozen sections that had been stained with H&E. The numbers of TRITC-positive cells were determined by counting 15 randomly chosen low power microscope fields (× 10 objective) per tumor mass.

Effects of Adoptive Immunotherapy and Intratumoral Lptn Transgene Expression on Animal Mortality and Tumor Growth.

Mice (eight per group) received s.c. injections of 0.5 × 10⁶ SP2/0 tumor cells in each thigh. At 10–12 days postinoculation, when the tumors were ∼5 mm in diameter, each mouse received an injection of 2 × 10⁹ pfu of AdVLptn or AdVpLpA. One and 2 days after virus injection, each mouse received an i.v. injection of 1 × 10⁷ activated T lymphocyte, or purified CD4⁺ or CD8⁺ T-cell subsets. Animal mortality and tumor growth were monitored daily for up to 10 weeks; for humanitarian reasons, all mice with tumors that achieved a size of 1.5 cm in diameter were sacrificed. In experiments to examine the antitumor immune memory, mice cured of their tumors by these means were again challenged with 0.5 × 10⁶ SP2/0 cells 8 weeks subsequent to tumor regression. The tumor growth was then monitored as above.

RESULTS

Activated T Lymphocytes Generated from SP2/0-Lptn-immunized Mice Comprise Th1 and Tc1 Cells.

Activated T lymphocytes were prepared by coculturing splenocytes of SP2/0-Lptn-immunized mice with irradiated SP2/0 tumor cells for 4 days and then were subjected to phenotypic characterization by flow cytometry. Essentially all of these cells were CD3⁺, with 66% CD4⁺ and 33% CD8⁺. They uniformly displayed high-level expression of CD25 (IL-2R), indicating that they were indeed activated, and exhibited low- and high-level L-selectin and FasL expression, respectively, all relative to T cells from naive mice (data not shown).

The CD4⁺ and CD8⁺ cells from this activated T-cell pool were then fractionated by negative selection using the anti-CD8 and -CD4 paramagnetic beads, respectively, and their expression of IFN-γ and IL-4 was assessed by flow cytometry (intracellular cytokines; Fig. 1A ⇓ ) and ELISA (secreted cytokine; Fig. 1B ⇓ ). After in vitro SP2/0 cell challenge, both populations stained positively for intracellular IFN-γ and secreted abundant IFN-γ protein. Neither population stained strongly for intracellular IL-4, although the CD4⁺ cells did secrete low levels of IL-4 (relative to their IFN-γ responses) at 3 days. This pattern of cytokine expression is consistent with the type 1 immune responses usually associated with antitumor immunity (29) and indicates that the CD4⁺ and CD8⁺ populations comprised Th1 and Tc1 phenotype cells, respectively. High proportions of both populations also expressed significant amounts of intracellular TNF-α and perforin (Fig. 1A) ⇓ .

Activated CD8⁺ Tc1 and CD4⁺ Th1 T Cells Display in Vitro Cytotoxicity to H-2K^d- and Ia^d-expressing SP2/0 Cells, Respectively.

Because the phenotypic markers associated with the activated T cells were consistent with CTLs, we next tested their cytotoxic activities against ⁵¹Cr-labeled SP2/0 cells. In these experiments, the in vitro-activated T cells from the SP2/0-Lptn-immunized mice did indeed display significant cytotoxicity for wild-type SP2/0 cells (57% specific killing at an E:T cell ratio of 100), but not for the irrelevant A20 or NK cell-sensitive Yac-1 cells (Fig. 2A) ⇓ . As noted previously (24) , the CTL activity of mice inoculated with SP2/0-Lptn tumors was 7–8-fold greater than that of mice bearing SP2/0 tumors. These latter mice also did not mount discernible antitumor responses in vivo, nor did their lymphocytes secrete IFN-γ after in vitro challenge with irradiated SP2/0 cells.

Because these activated T lymphocytes comprised CD4⁺ Th1 and CD8⁺ Tc1 cells and both subpopulations expressed TNF-α and perforin, we also assessed their respective abilities to act as effector cells against SP2/0 targets. As shown in Fig. 2B ⇓ , the CD8⁺ Tc1 cells displayed significant cytotoxicity for wild-type SP2/0 tumor cells (H-2K^d+/Ia^d−, 84% specific killing at an E:T cell ratio of 100), whereas CD4⁺ Th1 cells were ineffective, indicating that the CD8⁺ Tc1 cells had mediated the SP2/0 tumor-specific cytotoxicity observed above. Nearly 80% of CD8⁺ Tc1-mediated killing could be blocked by anti-H-2K^d, but not anti-Ia^d antibodies (data not shown).

To determine whether the CD4⁺ cells, which would be expected to be MHC II-restricted, could display cytotoxic effector functions, we induced Ia^d expression on SP2/0 cells by culture in IFN-γ (32) and then used these as targets for the CD4 cells. These IFN-γ-stimulated SP2/0 cells displayed significant amounts of cell surface Ia^d (data not shown) and, as shown in Fig. 2C ⇓ , were susceptible to CD4⁺ Th1 cell cytotoxicity (38% specific killing at an E:T cell ratio of 100); Ia^d-expressing B-cell blasts were not susceptible. This CD4 T cell-dependent tumor-specific killing was almost completely blocked by the anti-Ia^d (Fig. 2C) ⇓ , but not by the anti-H-2K^d antibody (data not shown). The potential in vivo relevance of this CD4 T-cell response was assessed by testing the cytotoxic effector activities of these cells against SP2/0 cells purified from freshly excised tumor masses (SP2/0/in vivo tumor cells). These SP2/0/in vivo tumor cells also displayed both MHC class I and II antigens and susceptibility to activated CD4⁺ Th1 cell cytotoxicity at levels equivalent to those for the IFN-γ-stimulated SP2/0/IFN-γ cells (data not shown). This clearly suggests that the activated CD4⁺ Th1 cells could also play a significant role in directly mediating cytotoxicity against SP2/0 tumors in vivo.

Lptn Chemoattracts and Stimulates Proliferation of CD8⁺ Tc1 Cells in Vitro.

These data, when taken together with our previous report (24) , suggest that Lptn could be an important stimulatory molecule in the induction of SP2/0-specific immunity. The data from our previous assessment of Lptn expression in SP2/0-Lptn tumors indicated that, as might be expected, these tumors express exceptionally high levels of their transgene product (i.e., ∼25 μg/g of tumor tissue), as determined in bioassays (25) . Thus, we wished to assess the specific role(s) of Lptn in these responses. We first evaluated the abilities of Lptn to chemoattract and stimulate proliferative responses among T cells from SP2/0-Lptn-immunized mice. As shown in Fig. 3 ⇓ , Lptn was able to chemoattract these T cells in a dose-dependent manner in vitro, but had no such effect on the control monocyte populations. To verify the molecular mechanisms mediating these chemotactic effects, we used RT-PCR to determine whether these cells expressed the XCR1 Lptn receptor (21) . The T cells, but not control monocytes, did express XCR-1 mRNA (Fig. 3 ⇓ , inset).

When we addressed the impact of Lptn signaling on the proliferative responses on our T cells, we used CD4⁺ and CD8⁺ cells purified by negative selection magnetic cell sorting from splenocyte cultures that had been pulsed for 4 days with irradiated SP2/0 cells. In addition, both of these populations were “contaminated” with 2–3% CD11c⁺ cells (above), likely dendritic cells or macrophages, which would be thus loaded with antigen for continued antigen-specific stimulation of the T cells within the proliferation assays themselves. We found that Lptn had dose-dependent inhibitory effects on these activated CD4⁺ T cells from the SP2/0-Lptn-immunized mice (Fig. 4A) ⇓ , whereas it induced a marked dose-dependent augmentation of the proliferative responses of the antigen-activated CD8⁺ T cells from these animals (Fig. 4B) ⇓ . That Lptn mediated these effects was confirmed by the demonstration that the anti-Lptn, but not irrelevant anti-macrophage inflammatory protein-1β, antibodies blocked 90% of the inhibitory effects of exogenous Lptn on the CD4⁺ T-cell cultures (P < 0.01), and blocked 99% of its CD8⁺ cell stimulatory effects (P < 0.01).

To determine whether this proliferative effect of Lptn on the CD8⁺ cells also translated into an augmented cytotoxic potential, we tested its impact on the cell’s abilities to lyse ⁵¹Cr-labeled SP2/0 cells. At an E:T ratio of 100, the CD8⁺ cells effected equivalent levels of specific killing (i.e., 81–84%) in the presence or absence of exogenous Lptn (1 μg/ml), and similarly matched responses were observed at all other E:T ratios examined (i.e., 6:1–50:1; data not shown). Thus, whereas Lptn would increase the size of the CD8⁺ T cell pool through its proliferative effects, it did not augment the cytotoxic activities of constant numbers of these CD8⁺ cells.

Expression of Transduced Lptn by Established Tumors Is Associated with Tumor Infiltation by T Cells.

Because Lptn has T-cell chemotactic effects in vitro, we next assessed the potential in vivo relevance of this observation. We first confirmed that tumors that had been infected with the AdVLptn virus detectably expressed authentic Lptn. We injected 2 × 10⁹ pfu of AdVLptn or control adenovirus AdVpLpA directly into SP2/0 tumors growing in mice; 24 h later, we used Northern and Western blot analysis to assess the steady-state levels of Lptn mRNA and protein within the tumors. As expected, very strong signals were detected for both Lptn RNA and proteins from the AdVLptn-treated tumors, whereas only background levels of signal were observed from the AdVpLpA-treated tumors (Fig. 5) ⇓ . We then assessed the infiltration of T cells into SP2/0 tumors that had been transfected with the experimental or control virus. One day after injection of the virus into tumors on the right (AdVLptn) and left (AdVpLpA) thighs of each mouse, the animals received i.v. injections of ⁵¹Cr-labeled activated T cells from SP2/0-immune mice. At 2 or 24 h post-treatment, the distribution of the radiolabeled T cells within the organ systems of the mice was determined. At 2 h the radiolabel had largely accumulated in the capillary-rich organs, such as the spleen, lungs, liver, and kidneys (Table 1) ⇓ . The levels of radioactivity detected in the AdVLptn- and AdVpLpA-treated tumors were much lower at this time (Table 1) ⇓ , and there were no significant differences between the experimental and control tumors (P > 0.05). As expected, the amount of radioactivity in the lungs decreased substantially over 24 h, when most of the residual activity was in the spleens and livers. Of note is the fact that at this time the level of radioactivity in the AdVLptn-treated tumors was more than three times than that of the AdVpLpA-treated tumors (P < 0.05; Table 1 ⇓ ), which clearly suggested that the ⁵¹Cr-labeled T cells were preferentially localizing to the tumors that expressed Lptn. We confirmed the validity of these results by direct tracking of TRITC-labeled T cells infiltrating Lptn-expressing and control virus-treated tumor tissues. TRITC-labeled T cells were 4.2-fold more abundant (P < 0.01) in the AdVLptn-treated tumors (15.5 ± 2.9 cells/field) than in the control adenovirus-treated tumors (3.7 ± 1.5 cells/field). These data confirm unequivocally that induced local expression of Lptn within SP2/0 tumors in mice can enhance tumor infiltration by adaptively transferred tumor-specific T cells and likely also the trafficking of other XCR-1-positive T cells through the tumors as well.

Synergistic Enhancement of Antitumor Immunity with Adoptively Transferred Tumor-specific CD4⁺ and CD8⁺ T Cells and Intratumoral Lptn Transgene Expression.

In clinical practice, most candidates for cancer therapy are patients with existing tumor burdens. Thus, although we and others have demonstrated efficacious prophylactic treatment of experimental mice (24 , 29 , 33) , if we are to accurately model the clinical case we must direct our questions of therapeutic efficacy to the elimination of established tumors in animals. To this end, we next tested the effects on 10-day established tumors of combining induced intratumoral expression of Lptn with the adoptive transfer of tumor-specific T cells. In our hands, 10-day SP2/0 tumors are ∼5 mm in diameter and have a well-developed vasculature (28) . We directly injected 2 × 10⁹ pfu of AdVLptn or AdVpLpA into the tumors; after 24 and 48 h, we injected 10⁷ activated T lymphocytes i.v. (total, 2 × 10⁷ cells/mouse). Control mice were given intratumoral injections of AdVLptn or PBS (day 0), but not the T cells. Tumor growth within the mice and their mortality rates were then monitored daily for up to 10 weeks. As expected, the tumors within the PBS-treated mice grew very aggressively (Fig. 6) ⇓ , and all mice in this group died within 21 days of tumor inoculation (Fig. 7A) ⇓ . On the other hand, mice with tumors that had been treated with the AdVLptn virus alone or with the control AdVpLpA virus in combination with an adoptive T-cell transfer experienced somewhat slower tumor growth than the PBS-treated animals (Fig. 6) ⇓ , although both groups of mice died within 24–40 days of tumor inoculation (Fig. 7A) ⇓ . Of great significance here was the observation that six of eight (75%) mice that had been treated with the AdVLptn virus and had received the adoptive T-cell transfers were entirely tumor free at 10 weeks postinoculation. The growth rates of the tumors in the remaining two (i.e., tumor-bearing) mice in this group were very much slower than those in the other groups of mice (Fig. 6) ⇓ ; these two mice died at 40–42 days postinoculation (Fig. 7A) ⇓ . These results demonstrate compellingly that a combined adoptive T-cell therapy and adenovirus-mediated Lptn gene transfer can synergistically enhance antitumor immunity in mice with preexisting SP2/0 tumors of substantial mass. Furthermore, we found that six of six mice that had rejected such tumors had developed solid tumor-specific immunologic memory, such that they also rejected tumor challenges given 8 weeks later.

Our final experiments addressed the subset of T cells that mediated this antitumor immunity in vivo. Thus, we repeated these experiments using purified CD4⁺ or CD8⁺ T cells for the passive transfers. Although either CD4⁺ or CD8⁺ cell transfer reduced the mean growth rates of AdvLptn-treated tumors and prolonged the life spans of these mice (Figs. 6 ⇓ and 7B ⇓ ), all of the CD4⁺ cell recipients were dead by 55 days, and six of eight CD8⁺ cell recipients were dead at 40 days after tumor inoculation (Fig. 7B) ⇓ . The remaining two CD8⁺ cell recipients in this group were tumor free at 10 weeks postinoculation. These results clearly indicate that both Th1 and Tc1 cells may contribute importantly to the enhancement of antitumor immunity in mice undergoing Lptn gene therapy, but more importantly, that both CD4⁺ and CD8⁺ cells are required for optimal combined immunotherapy and gene therapy.

DISCUSSION

We demonstrated previously that inoculation of mice with SP2/0 myeloma tumor cells engineered to express the C chemokine Lptn resulted in a failure of otherwise inevitable tumor establishment as a consequence of a strongly induced SP2/0 tumor-specific T-cell response and neutrophil infiltration (24) . In this study, we further characterized the activated SP2/0 tumor-specific T cells from such mice. Our data show that, after coculture with irradiated SP2/0 cells, the activated CD4⁺ and CD8⁺ T cells from these mice expressed IFN-γ, perforin, and TNF-α, but little IL-4, indicating that they were Th1 and Tc1 phenotype cells, respectively. Their in vitro cytotoxicity for tumor cells was mediated mainly by the CD8⁺ subpopulation and was specific for SP2/0 tumor cells, but not for the irrelevant A20 tumor cell line.

L-Selectin is an adhesion molecule that is highly expressed on naive T cells (34) . On the other hand, it has recently been reported that tumor-specific L-selectin^low T cells play a major role in tumor cell killing in vitro, as well as in the eradication of lung metastasis or s.c. tumors in vivo after adoptive T-cell transfers (4 , 5 , 34) . However, in those studies the target tumors were either very small (i.e., 3- or 10-day metastases; Refs. 4 , 5 ) or were well-established ones for which prior tumor irradiation was a prerequisite of the desired immunotherapeutic outcome (5 , 35) . In our hands, the in vitro-stimulated SP2/0 tumor-specific T cells derived from SP2/0-Lptn tumor-immunized mice were all L-selectin^low. Nevertheless, these cells alone were insufficient to achieve curative effects, very likely because the tumors were so well established in our system. This observation is consistent with those of Narvaiza et al. (36) , who also used otherwise untreated well-established tumors.

The importance to successful T-cell immunotherapy of T-cell infiltration into target tumor tissues is increasingly being recognized (37 , 38) , as is the direct relevance to this process of chemokines expressed in the tissues and the receptors for these molecules on the T cells (e.g., Ref. (39) . In the present study, our working hypothesis was that transgene-derived intratumoral Lptn expression would enhance the infiltration of adoptively transferred (XCR-1⁺) T cells into the tumors and thus improve the therapeutic efficiency of the T-cell therapy. Consistent with a previous report (40) , we found that adenovirus-mediated Lptn gene transfer alone did not cure any of the mice that were bearing well-established tumors. Similarly, a combination of the adoptive T-cell transfers and the control adenovirus AdVpLpA treatments did not effect tumor elimination, although for both of these groups (i.e., AdVLptn and T cell+AdVpLpA), tumor growth was significantly dampened. In contrast, combinational immunotherapy with adenovirus-mediated Lptn gene transfer and adoptive T-cell therapy not only very substantially inhibited tumor growth, but also eradicated 10-day SP2/0 tumors in six of eight mice. These results clearly highlight the synergistic therapeutic effects of local expression of Lptn within tumors in the context of adoptive T-cell therapy.

The synergistic effect of Lptn in adoptive T-cell therapy is likely derived from the combined chemotactic and stimulatory effects of this chemokine for activated T cells. We demonstrated that these activated T lymphocytes express XCR-1 mRNA and that they chemotactically respond to Lptn in in vitro chemotaxis assays. More importantly, we also showed that intratumoral injection of AdVLptn significantly enhanced T-cell infiltration of tumors in vivo. In addition to these chemotactic effects, our data also confirmed that Lptn stimulates CD8⁺ T-cell proliferation and inhibits CD4⁺ T-cell growth in vitro, as noted previously in another system (41) . However, we also demonstrated that, at defined E:T cell ratios, there were no differences in the cytotoxicity of Lptn-treated and untreated CD8⁺ cells from SP2/0-immune mice. Nevertheless, it is tempting to speculate that through its ability to expand CD8⁺ T-cell populations, Lptn expression by the tumor tissues could augment the Tc1 responses even without increasing the CTL activity on a “per cell” basis. It might be expected that effective tumor rejection by adoptively transferred T-cell populations could potentially be dependent on both direct and indirect CD8⁺ cell-mediated mechanisms. For example, we know that CD8⁺ Tc1 cells can directly eradicate tumor cells through cognate interactions that may involve either perforin- or Fas-mediated lytic mechanisms (42) . In addition, it has been shown that the release of Tc1 cytokines, such as IFN-γ, can directly inhibit tumor cell growth (43) and stimulate APCs (44) to further foster these responses.

CD4⁺ T cells can play important roles in facilitating antitumor immune responses (45 , 46) . In general, CD4⁺ T cells could provide IL-2 or other help to the CD8⁺ T cells because it has been shown that IL-2 administration, for example, can facilitate the eradication of some tumors (47) . In addition to the helper effect of CD4⁺ T cells for CD8⁺ CTLs, CD4⁺ T cells can also induce direct killing of tumor cells (26 , 48) . In this study, we demonstrated that (a) SP2/0 tumor cells cultivated with IFN-γ in vitro or freshly prepared from in vivo SP2/0 tumors displayed the cell surface Ia^d molecule, and (b) that CD4⁺ Th1 cells could also be directly cytotoxic to Ia^d+, but not Ia^d−, SP2/0 tumor cells. Our data suggest that SP2/0 tumor-specific CD4⁺ Th1 cells may mediate a direct killing of SP2/0 tumors in vivo. Recently, the cooperative role of CD4⁺ T cells has been reported for CD8⁺ CTLs in tumor eradication (49) , and some new insights into potential roles for CD4⁺ cells in this response may be gained from other studies. Specifically, it has been shown that CD4⁺ T-cell-stimulated APCs can better present tumor antigens to CD8⁺ T cells, which in turn augments the cytotoxic effector function of these cells (50, 51, 52) . Thus, passive transfer of tumor-specific CD4⁺ Th1 clones can enhance the ability of host APCs to initiate endogenous CD8⁺ T-cell responses and thereby also enhance tumor immunity (53) . Tumor-specific CD4⁺ T cells can also play a major “postlicensing” role in CTL-mediated antitumor immunity through maintenance of the CTL pool and its function, thereby allowing for augmented CD8⁺ T-cell infiltration of the tumor (54) .

Taken together, our data provide clear evidence of a potent synergy between adoptive CD4⁺ and CD8⁺ T-cell immunotherapy and adenovirus-mediated Lptn gene transfer into tumor tissues. This combinational immunotherapeutic strategy resulted in the cure of well-established tumors and thus becomes a tool of considerable conceptual interest in the implementation of future clinical objectives.