• Open Access

Antitumor activity of chimeric immunoreceptor gene-modified Tc1 and Th1 cells against autologous carcinoembryonic antigen-expressing colon cancer cells


To whom correspondence should be addressed. E-mail: tak24@igm.hokudai.ac.jp


To generate tumor-specific and interferon (IFN)-γ-producing Tc1 and Th1 cells applicable for many cancer patients, we previously developed a protocol for generating carcinoembryonic antigen (CEA)-specific Tc1 and Th1 cells from healthy human T cells by transduction with a lentivirus containing a chimeric immunoglobulin T-cell receptor (cIgTCR) gene composed of single-chain variable fragments from an anti-CEA-specific monoclonal antibody fused to an intracellular signaling domain of CD28 and CD3ζ. These cells, designated Tc1-T and Th1-T bodies, respectively, showed strong antitumor activity against CEA-expressing tumor cells in RAG2–/– mice when both of them were transferred. However, it remains unclear whether it is possible to generate Tc1-T and Th1-T bodies from cancer patients with defective T-cell function because of significant immunosuppression. Here, we prepared Tc1-T and Th1-T bodies from T cells of a colon cancer patient, and asked whether these T bodies can exert effective T-cell function against autologous tumor cells. These T bodies showed high cytotoxicity and produced IFN-γ in response to CEA-expressing autologous tumor cells, even in the presence of soluble CEA. It was also demonstrated that Th1-T bodies supported the survival of Tc1-T bodies through cell-to-cell interactions. Furthermore, our protocol utilized retrovirus for cIgTCR transduction to achieve better induction efficiency compared to lentivirus-mediated transduction. Taken together, our findings here indicate that retrovirally transduced Tc1-T and Th1-T bodies will become a promising strategy for adoptive immunotherapy of human cancer. (Cancer Sci 2006; 97: 920–927)

L  ymphocytes have been shown to play an important role in antitumor immune responses.(1–3) Among them, CD4+ T helper cells are known to play an essential role in the effective response of CD8+ cytotoxic T cells.(4–10) CD4+ T cells are reported to be required for the priming(11–13) and memory generation(14–17) of CD8+ T cells. It was shown that CD4+ T cells exert their help by activating antigen-presenting cells such as dendritic cells that are mediated by the CD40–CD154 interaction(18–20) or by direct CD4+–CD8+ T-cell communication via lymphokines or cell-to-cell interaction.(12,14,21)

It is well accepted that Th1-dominant immunity is critical for the successful induction of antitumor immunity in tumor-bearing hosts.(22–24) Moreover, establishment of a Th1/Tc1 circuit in a tumor-bearing host was shown to be crucial for complete tumor eradication.(25–27) Therefore, we established a method to prepare tumor-specific Tc1-T and Th1-T cells with the same antigenic specificity that is independent of major histocompatibility complex (MHC) restriction, based on the strategies of chimeric immune receptors, which combine antibody specificity with the trafficking properties and effector activity of T lymphocytes,(28,29) as an efficient strategy to induce type I antitumor immunity as a novel adoptive tumor immunotherapy.(30)

We have reported previously that Tc1 and Th1 cells generated from non-specifically activated CD8+ and CD4+ T cells of healthy human donors were transduced successfully with a chimeric immunoglobulin T-cell receptor (cIgTCR) composed of single-chain variable fragments (scFv) derived from a carcinoembryonic antigen (CEA)-specific monoclonal antibody and an intracellular signaling domain derived from the cytoplasmic portions of membrane-bound CD28 and CD3ζ.(30) Such artificially generated antigen-specific T cells, which were termed Tc1-T and Th1-T bodies, thereby possess the advantage of binding with antigen in an MHC-independent manner, while maintaining the capacity to be activated and exert Tc1 or Th1 cell function in an antigen-specific manner. These T bodies showed both cytotoxicity and interferon (IFN)-γ production in response to CEA-expressing tumor cells and exhibited a synergistic antitumor effect in a collaborative manner.

However, Tc1-T and Th1-T bodies generated from healthy donors are allogeneic responder cells against tumor cell lines, even though they have the antigen-specific receptor that was artificially transduced. As these T bodies were generated by activating them with non-specific stimulation such as anti-CD3 monoclonal antibody, they were not naive T cells. Therefore, it is arguable that Tc1-T and Th1-T bodies generated from non-specifically activated T cells of healthy donors potentially react to tumor cell lines in an allogeneic manner, to some degree. Therefore, it is crucial to address whether the CEA-specific reaction of Tc1-T and Th1-T bodies is fully efficient without the potentiation provided from allogeneic stimulation.

In addition, it is well known that tumor-bearing hosts often exhibit significant negative regulatory mechanisms that compromise their antitumor immunity.(31,32) These negative regulatory mechanisms include the induction of tumor-specific T-cell tolerance mediated by regulatory T cells,(22,33) the production of immunosuppressive cytokines such as vascular endothelial growth factor (VEGF), interleukin (IL)-10 and transforming growth factor (TGF)-β,(34–37) and downregulation of molecules important for antigenicity of tumor cells.(38–40) Particularly, tumor-bearing hosts are known to exhibit impaired T-cell function, especially in T-cell receptor signaling.(37,41–43) Therefore, it is critical to address whether the Tc1-T and Th1-T bodies generated from T cells derived from cancer patients can exert efficient T-cell function on the recognition and destruction of autologous tumor cells.

The mechanisms of CD4+ T cell assistance of CD8+ T cells have not been elucidated for T bodies against tumor cells. As several different mechanisms and pathways for CD4+ T cell assistance of CD8+ T cells have been proposed,(11–21) it is essential to address the cell-to-cell interaction requirement of Tc1-T and Th1-T bodies, which can not be substituted by soluble factors, to test the importance of Th1-T bodies in adoptive cell therapy.

In the present work, we aimed to test whether our established protocol to prepare Tc1-T and Th1-T bodies represents a promising strategy for effective adoptive immunotherapy of cancer patients. For this purpose, we generated CEA-specific Tc1-T and Th1-T bodies from peripheral blood mononuclear cells (PBMC) of a colon cancer patient utilizing retrovirus-containing cIgTCR genes. We asked whether these T bodies exhibited cytotoxicity and IFN-γ production in response to autologous tumor cells, even in the presence of blocking soluble CEA. Moreover, we addressed how CD4+ Th1-T bodies cooperated with CD8+ Tc1-T bodies while responding to CEA-expressing autologous tumor cells.

Materials and Methods

Cell lines

A CEA-positive human colon cancer cell line, SC1, was established from the liver metastatic lesion of colon carcinoma from a 54-year-old female patient that was resected at Hokkaido University Hospital in 2004 after obtaining written informed consent, approved by the medical ethics committees of Hokkaido University Graduate School of Medicine. An Epstein–Barr (EB) virus-immortalized B-cell line, SLCL1, was also established from the PBMC of the same patient. SC1, SLCL1, human lung cancer cell (HLC)-1, human pancreas cancer (PCI)-10, human gastric cancer (SH10), Daudi (human lymphoma) and Jurkat (human lymphoma) were all cultured in RPMI-1640 (Sigma, St Louis, MO, USA) containing 10% heat-inactivated fetal bovine serum (Life Technologies/Invitrogen, Carlsbad, CA, USA) with 2 mM l-glutamine, 0.05 mM 2-mercaptoethanol (Sigma), 10 mM HEPES, 100 U/mL penicillin and 100 µg/mL streptomycin sulfate (RPMI-10S). Packaging cell lines PG13 and 293 gp were maintained in Dulbecco's modified Eagle's medium (Sigma) containing 10% fetal bovine serum supplemented with 2 mM l-glutamine, 10 mM HEPES, 100 U/mL penicillin and 100 mg/mL streptomycin (DMEM-10S). Human leukocyte antigen (HLA)-typing of CEA-positive tumor cell lines was determined using the polymerase chain reaction (PCR)-reverse Sequence Specific Oligonucleotide method(44,45) (SRL, Tokyo, Japan). SC1 appeared to have HLA-A24, A26, B61 and B62. HLC-1 had HLA-A24, A26, B46 and B56. PCI-10 appeared to be A24 homozygous and B52 homozygous.

cIgTCR gene and retroviral vector

The cIgTCR gene, F39scFV/CIR-2, comprising scFv derived from human CEA-specific monoclonal antibody (F11-39), the CD8 hinge lesion, the transmembrane and cytoplasmic lesion of CD28, and the intracellular signaling domain of the CD3z chain, has been described.(46,47) As shown in Fig. 1, F39scFV/CIR-2 cDNA was inserted into the NotI and BamHI restriction enzyme sites within the multiple cloning site of the pGCDΔNsamIRESGFP vector, a derivative of the pGC-based retroviral vector pGCDNsamIRESGFP.(48)

Figure 1.

Schematic representation of the chimeric immunoglobulin T-cell receptor cDNA. Cyto, cytoplasmic domain; GFP, green fluorescent protein; IRES, internal ribosomal entry site; L, leader peptide gene; TM, transmembrane domain; Vκ, light chain variable region gene of F11-39 anticarcinoembryonic antigen monoclonal antibody; VH, heavy chain variable region of F11-39 monoclonal antibody.

Preparation of retrovirus containing cIgTCR

The 293 gp cells (5 × 106) were plated on 10-cm dishes (Nunc, Roskilde, Denmark) precoated with 0.002% poly l-lysine (Sigma). After 24 h, cells were transfected with pGCDΔNsamIRESGFP containing F39scFV/CIR-2 by calcium phosphate transfection.(49) At 12 h later, the medium was aspirated and 8 mL of fresh DMEM-10S was added. After a further 24 h, virus-containing supernatant was collected and passed through a 0.45-µm filter. The virus-containing supernatant and polybrene (8 µg/mL) were then added to the PG13 cells that were plated on the six-well plate (Nunc) with concentrations of 1 × 104 cells per well, 36 h before virus transduction. Virus-containing supernatant was again added 24 h later. Virus-exposed PG13 cells were cloned using the limiting dilution method, and selected for high titer production measured by the transduction efficiency into the Jurkat T lymphoma cell line. The clone with the highest virus titer was used as a producer of retrovirus-containing supernatant.

cIgTCR transduction into Tc1 and Th1 cells

Peripheral blood mononuclear cells were isolated using Ficoll-Paque Plus (Amersham Biosciences, Björkgatar, Sweden) from healthy human donors and cancer patients after obtaining informed consent. Cells were then sorted into CD4+ and CD8+ T cells with purity of 99% or higher using FACS Vantage (BD Biosciences, San Jose, CA, USA) with anti-CD4 or anti-CD8 monoclonal antibodies (Pharmingen, San Diego, CA, USA). The sorted T cells (1.5 × 106 cells) were cultured with mitomycin C (MMC)-treated autologous PBMC (1.5 × 106 cells) in the RMPI-10S medium containing phytohemagglutinin (PHA) (20 µg/mL) and Th1 cytokines (20 ng/mL IFN-γ, 100 IU/mL of IL-2, and 50 IU/mL of IL-12) for 72 h in 12-well plates. The resulting polyclonally activated Tc1 and Th1 cells (1 × 105 cells) were transduced with cIgTCR in 96-well flat-bottomed plates coated with RetroNectin (25 mg/mL; Takara Bio, Otsu, Japan) and anti-CD3 antibody (5 µg/mL; Pharmingen) by administration of retrovirus-containing supernatant twice, separated by a 24-h interval. The infected T cells were expanded in RPMI-10S with Th1 cytokines (20 ng/mL IFN-γ, 100 IU/mL IL-2 and 50 IU/mL of IL-12) for 7 days, plus an additional 2–5 days in RPMI-10S with IL-2 (20 IU/mL) alone.

Cytokine production and detection

Effector cells (1 × 105 cells) and tumor cell lines (5 × 104 cells) were cocultured in 96-well round-bottomed plates in 200 µL RPMI-10S medium. After a 24-h incubation, supernatants were collected and the IFN-γ level was measured by enzyme-linked immunosorbent assay using OptEIA Human IFN-γ set (Pharmingen). In the assay for the blocking effect of soluble CEA, the patient's plasma or RPMI-10S with recombinant CEA protein (2 µg/mL) was used as the culture medium.

Cytotoxicity assay

The cytotoxicity mediated by effector cells was measured using standard 12-h 51Cr-release assays, and the percentage cytotoxicity was calculated as described previously.(50) In the assay for the blocking effect of soluble CEA, the patient's plasma or RPMI-10S with recombinant CEA protein (2 µg/mL) was used as culture medium.

Determination of the effect of Th1-T bodies on the survival of Tc1-T bodies

The cIgTCR-transduced Tc1 cells (1 × 106 cells) were cultured alone, or were cocultured with MMC-treated SC1 cells (1 × 105 cells) and/or cIgTCR-transduced Th1 cells (1 × 106 cells) in 12-well plates in RPMI-10S for 72 h. The number of surviving CD8+ T cells was calculated from the total viable cell number, and the CD8+ T cell ratio was determined by flow cytometry. The Transwell tissue culture system (cell culture insert; Falcon, Franklin Lakes, NJ, USA) was used to determine the requirements of cell-to-cell contact.


Retroviral transduction improves transduction efficacy of the cIgTCR gene into Th1 and Tc1 cells

We first examined the efficiency of viral transduction of the cIgTCR gene by retrovirus vector. Using the PBMC of three healthy donors and a patient with colon cancer, retroviral transduction of cIgTCR gene into CD4+ Th1 or CD8+ Tc1 cells was carried out, and the transduction efficiency was determined by measuring the percentage of green fluorescent protein (GFP)+ cells using flow cytometry. The efficiency of cIgTCR transduction into the T cells derived from a patient with colon cancer was 71% for Th1 cells and 66% for Tc1 cells, which was almost comparable to those using T cells from healthy donors. The mean efficiency of transduction into T cells from five different healthy donors was 72% for Th1 cells and 64% for Tc1 cells (data not shown). This efficiency was significantly higher than the efficiency with lentivirus gene transduction that we reported previously, where 42% of Th1 cells and 23% of Tc1 cells were transduced successfully with the cIgTCR gene.(30)

Establishment of a CEA-expressing colon cancer cell line

A colon cancer cell line, SC1, was established from resected cancer tissue of a female patient who also provided PBMC for this study (Fig. 2A). The EB virus-immortalized B-cell line SLCL1 was established from the same colon cancer patient. CEA expression of SC1, SLCL1 and four other human tumor cell lines was examined by flow cytometry using the CEA-specific monoclonal antibody F11-39. SC1 as well as HLC-1 (lung cancer) appeared to be positive for cell surface CEA expression. PCI-10 (pancreatic cancer) was also positive for cell surface CEA; however, the level was significantly lower compared to that of SC1 or HLC-1. SLCL1, SH-10 (gastric cancer) and Daudi (lymphoma) did not express CEA (Fig. 2A). Cell surface MHC expression of the CEA-positive cell lines was determined using monoclonal antibody to MHC-class I or MHC-class II. As shown in Fig. 2B, all of the three cell lines were MHC-class I positive but MHC-class II negative.

Figure 2.

Carcinoembryonic antigen (CEA) and major histocompatibility complex (MHC) expression of the cell lines established from a colon cancer patient. (A) CEA expression of tumor cell lines. Tumor cell lines were stained with F11-39 anti-CEA monoclonal antibody (solid lines) or control mouse IgG (dotted lines). After extensive washing, cells were stained with fluorescein-isothiocyanate (FITC)-conjugated rabbit antimouse IgG antibody. CEA expression of (a) SC1, (b) HLC-1, (c) PCI-10, (d) Daudi, (e) SH10 and (f) SLCL1, which is an EB-transformed B-cell line established from the same colon cancer patient of SC1, are shown in histograms. (B) MHC expression of tumor cell lines. CEA-positive cell lines (a,d) SC1, (b,e) HLC-1 and (c,f) PCI-10 were stained with FITC-conjugated (a) anti-human leukocyte antigen (HLA)-A, (b) anti-HLA-B, or (c) HLA-C monoclonal antibodies, (d) anti-HLA-DR, (e) anti-HLA-DP or (f) anti-HLA-DQ monoclonal antibodies (solid lines), or FITC-conjugated control mouse IgG (dotted lines).

Cytokine production by cIgTCR gene-transduced Tc1 and Th1 cells in response to CEA+ or CEA autologous tumor cell lines

The IFN-γ production of cIgTCR-transduced effector cells in response to CEA+ or CEA cells was investigated by measuring IFN-γ in the supernatant of cocultured tumor cells and T bodies. Tc1 cells or Th1 cells of cancer patient from whom SC1 and SLCL1 were established were transduced with retrovirus containing cIgTCR, and were used as effector T bodies. Both CD8+ Tc1-T bodies and CD4+ Th1-T bodies produced high levels of IFN-γ when they were cultured with CEA+ autologous SC1 cells as well as CEA+ allogeneic HLC-1 cells (Fig. 3). These T bodies also produced significant but lower level of IFN-γ in response to PCI-10, which expresses a lower level of CEA compared to SC1 or HLC-1. These T bodies did not produce IFN-γ in response to CEA autologous SLCL1 or CEA allogeneic tumor cells. Moreover, non-specifically activated control T cells did not respond to any of the tumor cells tested, indicating that IFN-γ production from T bodies was induced in an antigen-specific manner (Fig. 3). Although both Tc1-T and Th1-T bodies produced significant amounts of IFN-γ, Th1-T bodies always exhibited higher production than Tc1-T bodies. These data indicate that T bodies generated from a cancer patient were highly efficient in their antigen-specific response to produce IFN-γ. The level of IFN-γ was comparable to those from T bodies generated from healthy donor T cells (data not shown).

Figure 3.

Antigen-specific interferon (IFN)-γ production from T bodies after 24-h coincubation with tumor cells. T bodies or control T cells established from a colon cancer patient (1 × 105 cells) and target tumor cells (5 × 104 cells) were cocultured in 96-well round-bottomed plates, and the IFN-γ levels in the supernatants were measured by enzyme-linked immunosorbent assay. ND, not detectable. Similar results were obtained in three separate experiments. Statistical analysis was carried out using two-tailed Student's t-test. The Th1-T body group was significantly different from the other groups (*P < 0.01).

Cytotoxicity mediated by T bodies against autologous CEA+ or CEA tumor cell lines

The cytotoxicity of these effector cells against CEA+ or CEA tumor cells was determined using the 12-h 51Cr-release assay (Fig. 4). Both Tc1 and Th1 cells transduced with the CEA-specific cIgTCR gene showed high cytotoxicity against CEA+ autologous SC1 cells and allogeneic HLC-1 cells with equivalent efficiency. They also showed clear but lower cytotoxicity against PCI-10 cells that expressed lower levels of CEA compared to SC1 or HLC-1 cells. These effector cells did not lyse CEA tumor cells. Control Tc1 and Th1 cells did not show cytotoxicity to any of the tumor cells tested. Although Th1-T bodies showed cytotoxicity against CEA+ tumor cells, Tc1-T bodies always tended to lyse target cells more efficiently compared to Th1-T bodies. The cytotoxic effect of Tc1-T and Th1-T bodies on CEA-expressing autologous SC1 cells was determined visually by microscopic observation (Fig. 5). Five hours after coculture of SC1 cells and Tc1-T or Th1-T bodies, rosette formation of T bodies surrounding tumor cells were observed. Morphological changes that suggested cellular activation were found in the T bodies of rosette formation. After 12 h, adherent SC1 cells were not found in the culture with Tc1 or Th1 cells transduced with the CEA-specific cIgTCR gene.

Figure 4.

Antigen-specific cytotoxic activities of Th1-T and Tc1-T bodies. Cytotoxic activity of Th1-T or Tc1-T bodies against tumor cells including autologous tumor cells were determined using the 12-h 51Cr-release assay. (A) SC1; (B) HLC-1; (C) PCI-10; (D) Daudi; (E) SH10; and (F) SLCL1. Similar results were obtained in three separate experiments. Statistical analysis was carried out using the two-tailed Student's t-test. (A,B) Th1-T bodies and Tc1-T bodies were significantly different from control Th1 and control Tc1, respectively, with *P < 0.01 in every effector/target (E/T) ratio. (C) Similar differences were observed in the E/T ratio of 60/1 and 30/1 with **P < 0.05.

Figure 5.

Efficient eradication of carcinoembryonic antigen (CEA)-expressing autologous tumor cells by Th1-T or Tc1-T bodies. Autologous effector cells were plated in a 48-well plate in which SC1 cells were attached in advance and microscopic images (×100) were recorded (A–D) 5 h after coculture, or (E,F) 12 h after coculture with washing by phosphate-buffered saline. (A,E) Th1-T bodies; (B,F) Tc1-T bodies; (C,G) control Th1 cells; and (D,H) control Tc1 cells. Similar results were obtained in three separate experiments.

Essential role of cell-to-cell contact between Th1-T and Tc1-T bodies for the survival of Tc1 cells

By coculture with SC1 cells, autologous Th1-T bodies could expand easily, and survived for several months (data not shown). In contrast, SC1-stimulated Tc1 cells could only survived for 2 or 3 days (Fig. 6). The supporting effect of Th1-T bodies on the survival of Tc1-T bodies cultured with SC1 cells was determined by measuring the cell number of viable CD8+ Tc1-T bodies after stimulation with SC1 cells in the presence or absence of CD4+ Th1-T bodies (Fig. 6). The requirement of cell-to-cell contact on the effect of Th1-T bodies was evaluated using the Transwell tissue culture system, which separated Th1-T bodies from Tc1-T bodies through a semipermeable membrane (0.4-µm pore size), allowing the diffusion of cytokines. As shown in Fig. 6, Tc1-T bodies survived significantly longer in the presence of Th1-T bodies and SC1 cells (13-fold increase in the number of cells surviving). However, prolonged survival of Tc-1 T bodies was abolished when they were cultured separately with Th-1 T bodies, suggesting that cell-to-cell contact is an important event for supporting the survival of Tc1-T bodies by Th1-T bodies. The addition of IL-2 failed to support the survival of Tc1-T bodies stimulated with SC1 cells (data not shown).

Figure 6.

The effect of Th1-T bodies on the survival of Tc1-T bodies. Chimeric immunoglobulin T-cell receptor (cIgTCR)-transduced Tc1 cells (1 × 106 cells) were cultured alone or with mitomycin C (MMC)-treated SC1 cells (1 × 105 cells) and/or cIgTCR-transduced Th1 cells (1 × 106 cells) in 12-well plates for 72 h. The number CD8+ T cells that survived was calculated. In the groups with inline image (group 5 and 6), the Transwell tissue culture system was used to separate Th1-T bodies from Tc1-T bodies. Similar results were obtained in three separate experiments. Statistical analysis was carried out using the two-tailed Student's t-test. Group 3 was significantly different to the other groups (*P < 0.01).

Soluble CEA did not inhibit the function of CEA-specific T bodies

Increased CEA levels in the plasma of patients with CEA-expressing tumors is well characterized, and soluble CEA level has been applied clinically as a marker for the screening of disease and the assessment of treatment. We evaluated the effect of soluble CEA on the activation of T bodies by measuring the cytotoxicity and IFN-γ production against SC1 cells in the presence or absence of soluble CEA. The plasma CEA concentration of the patient from whom SC1 was established appeared to be 753 ng/mL. Using the plasma of this patient and recombinant CEA protein (2000 ng/mL), it was shown that soluble CEA did not block the cytotoxicity of T bodies (Fig. 7). IFN-γ production from Th1-T bodies decreased marginally with patient plasma. However, recombinant CEA protein at higher concentrations did not show any inhibition, suggesting the effect of conditions other than CEA in patient plasma. Moreover, soluble CEA alone did not stimulate T bodies to induce IFN-γ production.

Figure 7.

Evaluation of the effect of soluble carcinoembryonic antigen (CEA) on the antigen recognition of CEA-specific chimeric immunoglobulin T-cell receptor (cIgTCR)-transduced T bodies. The colon cancer patient's plasma, which contained high levels of CEA (753 ng/mL) or recombinant CEA protein (2 µg/mL), was added in the (A) 12-h 51Cr-release cytotoxicity assay (E/T ratio was 60) or (B) interferon (IFN)-γ production assay with Th1-T or Tc1-T bodies and autologous S1 tumor cells. ND, not detectable. Similar results were obtained in two separate experiments.


The strategy to induce chimeric immune receptors to T cells is considered to have several advantages for tumor immunotherapy. First, T bodies can be induced easily from non-specifically activated polyclonal T cells. Therefore, it can overcome the difficulties in inducing and expanding natural tumor-specific CD8+ and CD4+ T cells.(51–53) Second, T bodies can recognize the target cells to exert their T-cell function in an MHC-independent manner. Therefore, T-bodies are applicable to cancer patients without the restriction of HLA types. Furthermore, T bodies can exhibit their function against tumor cells that do not express MHC and costimulatory molecules, which is often observed in a variety of tumor tissues.(38–40,54)

However, the majority of studies on T bodies have utilized T cells from healthy donors and allogeneic tumor cells. For developing clinical applications of T body-mediated technology, it is critical to test the effective recognition and eradication of autologous tumor cells by T bodies derived from advanced cancer patients who are frequently reported to exhibit impaired T-cell function.(37,41–43) It is also reported that cancer patients show difficulties in inducing and expanding tumor-specific T cells in a vaccination setting, or ex vivo expansion from PBMC or tumor-infiltrating T cells of cancer patients.(51–53)

The present report is in line with limited previous studies using T cells from cancer patients and their autologous tumor cells.(55–57) Hombach et al.(55) and Sheen et al.(56) both demonstrated the induction of CEA-specific T bodies from cancer patients that could efficiently recognize autologous tumor cells to produce cytokines. Sheen et al. described low cytotoxicity of T bodies against autologous tumor cells. Hombach et al. also showed marginal cytotoxic activity of T bodies against autologous tumor cells. In the present work, fully functional Tc1-T and Th1-T bodies that can kill autologous tumor cells and allogeneic tumor cells equivalently were generated efficiently from a colon cancer patient (Fig. 4), suggesting an advantage of this method in overcoming the tolerance of peripheral blood T cells from colorectal cancer patients and producing fully functional T cells. It is possible that this advantage was achieved because T cells from a cancer patient were once exposed to type-I cytokine conditions in vitro, and also because the signal from transduced chimeric receptor was sufficient to activate T cells without the requirement of other signals such as the signals from costimulatory molecules. Induction of similar specific activities in other cancer patients is an important next step of this study. As shown in Figs 3 and 4, Tc1-T and Th1-T bodies generated from a colon cancer patient could exert specific T-cell activity against two different allogeneic CEA-positive tumor cell lines. These results suggest that similar specific activities may be inducible in other cancer patients in a non-autologous setting.

Because the cIgTCR molecule does not have any conformational character to bind to MHC molecules, the T bodies are considered to react with tumor cells in an MHC-independent manner. By flow cytometry analysis, all of the three CEA-positive cell lines used in the present study were MHC-class II-negative (Fig. 2B), indicating that the Th1-T bodies were reacting with tumor cells in an MHC-independent manner. All three of the cell lines differed in HLA-B typing (as described in Materials and Methods), indicating that HLC-1 and PCI-10 were allogeneic to the T bodies used in this study. All three of the CEA-positive tumor cell lines shared HLA-A24. However, the control Th1 or Tc1 cells with endogenous T-cell receptor (TCR) did not react to CEA-positive cells (Figs 3, 4), suggesting that even the Tc1-T bodies reacted to CEA-positive cells in an MHC-independent manner. Moreover, we demonstrated that the cytotoxicity as well as cytokine production capacity of Tc1-T and Th1-T bodies was blocked strongly by anti-CEA monoclonal antibody,(30) but not by anti-HLA-class I monoclonal antibody or anti-HLA-A24 monoclonal antibody (data not shown).

The advantage of coinjection of Th1-T bodies with Tc1-T bodies in overcoming immunosuppression in tumor-bearing hosts was shown in our previous report.(30) In a model of tumor adoptive immunotherapy, we found that a combination of Tc1-T and Th1-T bodies could induce complete rejection of tumor cells, although Th1-T or Tc1-T bodies alone were unable to reject the tumor completely. The mechanism underlining the advantage of coinjection was also addressed here. As shown in Fig. 6, the survival of CD8+ Tc1-T bodies was greatly supported by direct coculture with CD4+ Th1-T bodies in the presence of autologous CEA-expressing tumor cells, and therefore was suggested as one of the important mechanisms of the improved antitumor effect of coinjection. This finding is consistent with previous reports that emphasize the effect of CD4+ T cells on the maintenance of CD8+ T cells.(14–17) Utilizing the Transwell tissue culture system, cell-to-cell contact was found to be essential for Th1 assistance in Tc1 survival, suggesting that administration of cytokines may not be sufficient to substitute the advantage of T helper cells. Giuntoli et al. also reported the requirement of direct costimulation mediated by cell surface molecules for the CD4+ T cell help on the proliferation of a tumor–reactive cytotoxic T lymphocytes (CTL) clone.(21) Both the Tc1 cells and Th1 cells generated in our protocol expressed CD28, CD152, CD86, CD70, MHC class I, and MHC class II, but not CD40. Tc1 cells additionally expressed CD80 (data not shown). It is possible that these molecules play some important role in the cell-to-cell interaction between these cells by providing costimulatory signals resulting in the prolonged survival of Tc1 cells. Giuntoli et al. reported that CD27 (CD70L), CD134 and MHC-class II on CTL are capable of interacting directly with the corresponding ligands on Th cells, resulting in enhanced survival and proliferation of the CTL.(21) It is possible, however, that other mechanisms also play important roles for the collaboration of Tc1-T and Th1-T bodies in vivo. Support of CD4+ T cells on the survival of CD8+ T cells may also be provided through activation of antigen-presenting cells or cytokines in vivo.(12,18–20) The existence of CD4+ T cells may also contribute to the priming of newly arising tumor-specific T cells, or the trafficking of CD8+ T cells into tumor sites or draining lymph nodes.(11–13)

Soluble antigen, which can be detected frequently in cancer patients in high serum concentrations, is considered to block the receptor of grafted effector cells and thus prevent the recognition of target cells and effector function.(58,59) In particular, patients with CEA expression are frequently found to have high serum concentrations of CEA. Therefore, adverse effects of serum CEA should be tested for the application of T bodies specific to CEA. We did not found any blocking effect or non-specific activating effect of soluble CEA on the Th1-T and Tc1-T bodies specific to CEA (Fig. 7), consistent with a previous report utilizing a murine T-cell clone.(60) The exact mechanism for the absence of adverse effects was unclear, although preferential triggering of chimeric receptors by the membrane-bound and immobilized CEA was speculated. Preferential triggering may be achieved by specificity of the monoclonal antibody F11-39 used in this study. However, it is also possible that soluble CEA lacks the blocking activity of grafted cIgTCR. In this case, soluble CEA may differ from immobilized CEA in the conformation critical for the blocking activity of the chimeric receptor.

In the present paper, we established a highly efficient protocol to generate T bodies by retroviral transduction with defined Tc1 and Th1 phenotypes. This protocol was applicable for the PBMC from a cancer patient without the difficulties of inducing efficient T-cell functions against autologous tumor cells. The cell-to-cell interaction of CD4+ Th1-T bodies and CD8+ Tc1-T bodies was found to be essential for CD4+ T-cell assistance of CD8+ T-cell survival, and therefore was suggested as a mechanism of for the collaborative effect of Tc1-T and Th1-T bodies on tumor eradication. In vivo studies utilizing mice with disrupted RAG2 genes and common γ chain or SCID-hu model are planned to further evaluate the application of this protocol for the clinical setting. Taken together, these findings suggest that Tc1-T and Th1-T bodies represent a novel and promising strategy for inducing tumor-specific type-I immunity in cancer patients as an efficient adoptive cancer immunotherapy.