Engineered CD8+ cytotoxic T cells with fiber-modified adenovirus-mediated TNF-α gene transfection counteract immunosuppressive interleukin-10-secreting lung metastasis and solid tumors

Abstract

T-cell suppression derived from tumor-secreted immunosuppressive interleukin (IL)-10 becomes a major barrier to CD8+ T-cell immunotherapy of tumors. Tumor necrosis factor-alpha (TNF-α) is a multifunctional cytokine capable of activating T and dendritic cells (DCs) and counteracting IL-10-mediated DC inhibition and regulatory T-cell-mediated immune suppression. In this study, we constructed a recombinant adenovirus MFAdVTNF with fiber-gene modified by RGD insertion into the viral knob's H1 loop and a melanoma cell line B16OVA/IL−10 engineered to express ovalbumin (OVA) and to secrete IL-10 (2.2 ng/ml/106 cells/24 h). We transfected OVA-specific CD8+ T cells with MFAdVTNF, and found a fivefold increase in transgene human TNF-α secretion (4.3 ng/ml/106 cells/24 h) by the engineered CD8+ TTNF cells transfected with MFAdVTNF, compared to that (0.8 ng/ml/106 cells/24 h) by CD8+ T cells transfected with the original AdVTNF without viral fiber modification. The engineered CD8+ TTNF cells exhibited enhanced cytotoxicity and elongated survival in vivo after adoptive transfer. TNF-α derived from both the donor CD8+ T cells and the host cells plays an important role in donor CD8+ T-cell survival in vivo after adoptive transfer. We also demonstrated that the transfected B16OVA/IL−10 tumor cells secreting IL-10 are more resistant to in vivo CD8+ T-cell therapy than the original B16OVA tumor cells without IL-10 expression. Interestingly, the engineered CD8+ TTNF cells secreting transgene-coded TNF-α, but not the control CD8+ Tcontrol cells without any transgene expression eradicated IL-10-secreting 12-day lung micrometastasis in all 10/10 mice and IL-10-secreting solid tumors (5 mm in diameter) in 6/10 mice. Transfer of the engineered CD8+ TTNF cells further induced both donor- and host-derived memory CD8+ T cells, leading to a stronger long-term antitumor immunity against the IL-10-secreting B16OVA/IL−10 tumor cell challenges. Therefore, CD8+ T cells engineered to secrete TNF-α may be useful when designing strategies for adoptive T-cell therapy of solid tumors.

Introduction

Cytotoxic T lymphocytes (CTLs) play a crucial role in host immune responses to cancer. Effective adoptive cancer immunotherapy with tumor-sensitized CTLs has been well documented in animal models, where transfer of such tumor-specific T cells into mice bearing tumors resulted in tumor eradication,1, 2, 3 but not autoimmunity.4 Recently, it has further been reported that adoptive transfer of cytotoxic T type 1 cells can cure 70% palpable tumors.5 More recently, Becker et al.6 have shown that Tc cells enriched through an interferon (IFN)-γ-capture method displayed enhanced tumor-specific cytotoxicity and the adoptive therapy with these T cells cured 3/5 mice bearing 3-day tumors. In most of these animal models, T-cell adoptive therapy is all limited in treatment of early-stage tumors (e.g., 3-day or palpable tumors or 3–10-day early lung metastases), but not large tumors. However, it is the well-established tumors in animal models that mimic the clinical patients with existing tumor burdens. Therefore, new strategies with improved therapeutic efficiency are expected.

IL-10 has been identified as a key immunomodulatory cytokine capable of mediating immunosuppressive effects in immune systems. It inhibits cytokine production and proliferation of CD4+ T cells7 and reduces expression of major histocompetibility complex (MHC) class II, intracellular adhesion molecule (ICAM)-1 and CD80 on dendritic cells (DCs), resulting in induction of T-cell anergy8, 9 or regulatory T (Tr) cells that suppress antigen-specific T-cell responses10, 11 IL-10 was found in many human tumors12, 13 It has been demonstrated that tumor-induced IL-10 (i) blocks tumor-specific helper T type 1 (Th1) responses,14 (ii) inhibits tumor-specific CD8+ CTL cytotoxicity15 and (iii) supports the tumor growth.16 in animal tumor models. Therefore, T-cell suppression derived from tumor-secreted IL-10 becomes one of the major barriers to T-cell immunotherapy of tumors. Clinically, the detectible serum IL-10 in cancer patients becomes a negative indicator for clinical outcome17, 18

TNF-α is a multifunctional cytokine with a broad spectrum of activities. It activates IL-2R/IFN-γ expression and cytotoxicity of T cells,19 and stimulates T-cell proliferation and activation20 through activation of nuclear factor-kappa B (NF-κB) involving multiple cytokine activation.21, 22 Recently it has been shown that TNF-α can provide the costimulatory survival signals for CD8+ T cells in the absence of B7-CD28 and IL-2 costimulatory pathways.23 In addition to T-cell activation, TNF-α can also stimulate the DC maturation24 and reduce/counteract IL-10-mediated DC inhibition.25

We hypothesized that TNF-α-gene engineered CD8+ T cells may directly counteract IL-10-induced T-cell suppression and improve CD8+ T-cell therapy of solid tumors. In this study, we constructed an expression vector pcDNAIL-10 expressing immunosuppressive IL-10 and a recombinant fiber-modified adenovirus (MFAdVTNF) expressing the human TNF-α. By using the above reagents, we then generated the OVA-specific TNF-α gene-engineered CD8+ T (TTNF) cells with MFAdVTNF and the IL-10 gene-transfected OVA-expressing melanoma cell line B16OVA/IL−10 with pcDNAIL-10. We further investigated the potential therapeutic effect of engineered CD8+ TTNF cells in treatment of IL-10-secreting and OVA-expressing B16OVA/IL−10 solid tumors and lung micrometastasis.

Materials and methods

Antibodies, cytokines, adenoviruses, cell lines and animals

Biotin-conjugated anti-mouse MHC class I (H-2Kb), CD8, CD25, CD45.1, CD45.2, CD69, Vα2Vβ5+ T-cell receptor (TCR) and Fas ligand (FasL) antibodies (Abs), the R-phycoerythrin (PE)-conjugated anti-mouse IL-4, IL-10, IFN-γ and perforin Abs and anti-human TNF-α (hTNF-α) Ab were all obtained from Pharmingen Inc. (Mississauga, Ontario, Canada). Fluorescein isothiocyanate (FITC) or energy-coupled dye (ECD)-conjugated avidin or anti-mouse immunoglobulin G (IgG) Ab was obtained from Bio/Can Scientific (Mississauga, Ontario, Canada). PE-labeled H-2Kb/OVA257–264 tetramer and FITC-labeled rat anti-mouse CD8 Ab were obtained from Beckman Coulter (San Diego, CA). The recombinant mouse granulocyte monocyte-colony stimulating factor (GM-CSF), IL-2 and IL-4 were purchased from R&D Systems (Minneapolis, MN). The chicken egg ovalbumin (OVA) protein, concanamycin A (CMA) and emetin were obtained from SIGMA (St Louis, MO). The mouse anti-OVA Ab was purified by using Protein G affinity column from antisera of the mice immunized with OVA emulsified in adjuvant. The recombinant adenovirus (AdV) AdVTNF expressing the human TNF-α was previously constructed using the adenoviral vector pLpATNF in our laboratory.26 The highly lung metastatic melanoma cell line BL6-10 (B16) and its derivative OVA-transfected cell line BL6-10OVA (B16OVA) were generated in our laboratory.27 The mouse thymoma cell line EL4 and its derivative OVA-transfected cell line EG7 were obtained from American Type Culture Collection (ATCC, Rockville, MD). Both B16OVA/IL−10 and EG7 tumor cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS) and G418 (0.5 mg/ml). The human embryonic kidney cell line 293 containing adenovirus (AdV) 5 E1 region was obtained from ATCC. The OVA-specific TCR transgenic OT I mice having a transgenic Vα2Vβ5 TCR specific for OVA257–264 epitope in context of H-2Kb,28 C57BL/6(TNF-α−/−) mice with TNF-α gene knockout (KO), C57BL/6 (B6, CD45.2+) and C57BL/6.1 (B6.1, CD45.1+) mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Homozygous OT I (TNF-α−/−) mice were generated by backcrossing female C57BL/6(TNF-α−/−) mice onto the male OT I background for two generations. Animal care was in accord with the guidelines of the University of Saskatchewan.

Construction of pcDNA/IL-10 expression vector

Total RNA was extracted from mouse splenocytes using RNeasy Mini Kit (Qiagen, Mississauga, Ontario, Canada). A 1 kb cDNA fragment coding for the full open-reading frame of mouse IL-10 gene was cloned by reverse transcription-polymerase chain reaction (RT-PCR) from a cDNA library derived from the splenocyte RNA using PFU polymerase. Two primers specific for the mouse IL-10 gene were used, namely the sense primer (5′-IndexTermatgag acttc tcctc ctg-3′) and the anti-sense primer (5′-IndexTermttacc cagtc agggt tactg-3′). The cloned cDNA fragment was ligated into the pCR2.1 (Invitrogene, Carlsbad, CA) to form the pCR2.1IL-10 vector. The IL-10 sequence was verified by dideoxy nucleotide sequencing method. The cDNA fragment of IL-10 (XbaI/HindIII) from the pCR2.1IL-10 vector was further ligated into pcDNA vector (Invitrogene) containing the hygromycin-resistant gene to form the expression vector pcDNAIL-10.

Engineering B16OVA/IL−10 tumor cells to secrete IL-10

Twenty million B16OVA tumor cells were resuspended in 0.7 ml phosphate-buffered saline (PBS) and mixed with another 0.3 ml PBS containing 10 μg pcDNAIL-10 or the control pcDNA DNA. Tumor cells were transfected with the above vector DNA using a Bio-Rad gene pulser (Bio-Rad Laboratories, Mississauga, Ontario, Canada) with parameters of 250 V and 125 μFD capacitance. The transfected B16OVA/IL−10 or the control B16OVA/control tumor cells by using pcDNAIL-10 and the control pcDNA vectors were selected for growth in the medium containing G418 (0.5 mg/ml)/hygromycin (4 mg/ml), respectively, and then maintained in the medium containing G418 (0.5 mg/ml)/hygromycin (0.5 mg/ml).

Construction of fibre-modified FMAdVTNF expressing TNF-α

The recombinant fiber-modified adenovirus MFAdVTNF expressing the transgene human TNF-α26 was constructed using the AdEasy system (Stratagene, La Jolla, CA) as previously described.29 The fiber gene was modified by insertion of RGD motif into the viral H1 loop similarly to the method described by Liu et al.30 Briefly, the pAdEasy-1 vector was digested with BamHI, resulting in 21.7 and 11.7-kb DNA fragments. The 11.7-kb fragment was gel purified and inserted into pCRBlunt (Invitrogen, Carlsbad, CA) to form a pCR11.7 vector. The pCR11.7 vector was digested with EcoRI, resulting in 9.3 and 5.6-kb DNA fragments. The 9.3-kb fragment containing the pCRBlunt backbone underwent self-ligation, forming a pCR9.3 vector. PCR amplification was used to replace the DNA sequence within the StuI–AflII sites of the fiber knob domain in pCR9.3 vector. In PCR, the sense primer 1 (5′-IndexTermcaaca aaggc cttta cttgt ttaca gcttc a-3′) and the antisense primer 2 (5′-IndexTermtgaca tagag tactg gttta gtttt gtctc cgttt aag-3′) were used to amplify a 680-bp P1+P2 fragment (nucleotides 31 950–32 630 of the AdV5 genome), whereas the sense primer 3 (5′-IndexTermactaa accag tactc tatgt cattt tcatg ggact ggt-3′) and the antisense primer 4 (5′-IndexTermtggac agcga catga acttt aagtg agctg-3′) were used to amplify a 435-bp P3+P4 fragment (nucleotides 32 690–33 125 of the AdV5 genome). The P1+P2 and P3+P4 fragments were gel purified, mixed, and joined by PCR using P1 and P4, resulting in a 1.1-kb fragment. This fragment contains part of the knob sequence of AdV5 with AflII (nucleotide 31 950) and StuI (nucleotide 33 125) sites. A deletion was created from 32 631 to 32 689 by removing amino-acid residue VTLTI TLNGT QETGD TTPSA, and incorporating a single mutation from T to A to create a ScaI site in the H1 loop. Plasmid pCR9.3(AS) was constructed by inserting the 1.1-kb PCR fragment into the pCR9.3 vector. A duplex was formed by annealing two complimentary oligonucleotides (5′-IndexTermaacac taacc attac actaa acggt acaca ggaaa cagga gacac aactt gcgac tgtag aggag actgc ttttg tccaa gtgca t-3′ and 5′-IndexTermatgca cttgg acaaa agcag tctcc tctac agtcg caagt tgtgt ctcct gtttc ctgtg taccg tttag tgtaa tggtt agtgt t-3′). This 86-bp duplex was cloned into ScaI-digested pCR9.3(AS), forming the pCR9.3(RGD) plasmid which contains the AdV5 complete sequence (nucleotides 31 950–33 125) and an additional RGD-4C sequence (CDCRGDCFC) in the HI loop (between nucleotides 32 679 and 32 680). The previously generated 5.6-kb fragment was inserted into EcoRI-digested pCR9.3(RGD) to form pCR11.7(RGD). The resulting 11.7-kb BamHI band from pCR11.7(RGD) was ligated into the previous 21.7-kb BamHI fragment of pAdEasy-1, resulting in pAdEasy(RGD). This 33.4-kb plasmid contains the pAdEasy-1 sequence plus an additional RGD-4C sequence in the H1 loop. All of the insert orientations within the vectors were determined by dideoxy nucleotide sequencing method and restriction enzyme-digestion analysis. The shuttle vectors pLpATNF with the human TNF-α gene and the control pLpA without any transgene were used.26 The PmeI-digested shuttle vectors were then cotransformed into BJ5183 Escherichia coli cells already containing the backbone vector pAdEasy(RGD) for increased efficiency of homologous recombination31, 32 to form the recombinant vectors pAd(RGD)TNF (Figure 1) and pAd(RGD), respectively. The recombinant pAd(RGD)TNF and pAd(RGD) vectors were linealized by PacI digestion, and then transfected into 293 cells using Lipofectamine (Gibco/BRL, Burlington, Ontario, Canada) to generate the fiber-modified adenoviruses MFAdVTNF with the transgene human TNF-α expression (Figure 1) and the fiber-modified control MFAdV without any transgene expression. The fiber-modified MFAdVGFP expressing the marker gene GFP was obtained from Dr Zheng, NIH, Bethesda, MD. All recombinant adenoviruses were amplified in 293 cells, purified by a series of cesium chloride ultracentrifugation gradients, and stored at −80°C until use.

Figure 1
figure1

Schematic representation of recombinant adenoviruses. The recombinant adenoviruses AdVTNF and MFAdVTNF were constructed using the adenoviral vectors pLpATNF26 and pAd(RGD)TNF, respectively. The RGD (bold) motif was incorporated into the H1 loop of the fiber. For detail, see text. CMV, cytomegalovirus promoter; poly A, bovine growth hormone polyadenylation signal; LITR, left inverted terminal repeat; RITR, right inverted terminal repeat.

Preparation of bone marrow-derived DCs

The preparation of bone marrow (BM)-derived DCs was previously described.27 Briefly, BM cells prepared from femora and tibiae of normal C57BL/6 (H-2Kb) mice were depleted of red blood cells with 0.84% ammonium chloride and plated in DC culture medium (DMEM plus 10% FCS, GM-CSF (20 ng/ml) and IL-4 (20 ng/ml)). On day 3, the nonadherent granulocytes, and T and B cells were gently removed, and fresh media were added. After 2 days, the loosely adherent proliferating DC aggregates were dislodged and replated. On day 6, the nonadherent cells were harvested. The DCs generated in this manner displayed (i) typical morphologic features of dendritic cells (i.e., numerous dendritic processes) and (ii) expression of MHC class I (H-2Kb) and II (Iab) antigens, co-stimulatory molecules (CD40, CD80 and CD86) and adhesion molecules (ICAM-1, CD11b and CD11c) (data not shown).

Preparation of OVA-specific CD8+ T cells

Naïve OVA-specific CD8+ T cells were isolated from the spleens and lymph nodes of OT I or OT I(TNF-α−/−) TCR-transgenic mice, enriched by passage through nylon wool columns (C&A Scientific, Manassa, VA), and then fractionated by negative selection using anti-mouse CD4 (L3T4) paramagnetic beads (DYNAL Inc., Lake Success, NY) according to the manufacturer’s protocols. BM-derived DCs were loaded with OVA protein (0.3 mg/ml) at 37°C for overnight, and referred to as DCOVA. The above purified naïve CD8+ T cells from OT I mice (3 × 105 cells/ml) were stimulated with irradiated (4000 rad) DCOVA (2 × 105 cells/ml) in presence of IL-2 (20 U/ml) for 3 days, and purified using Ficoll-Paque (Sigma, St Louis, MO) density gradient centrifugation and then CD8 microbeads (Miltenyi Biotec, Auburn, CA).33 DCOVA-activated CD8+ T cells derived from OT I and OT I(TNF-α−/−) mice were referred to as CD8+ T and T(TNF-α−/−) cells, respectively. The phenotypes and cytokine profiles of CD8+ T(TNF-α−/−) cells are similar to CD8+ T cells, except for the TNF-α deficiency (data not shown).

Engineering CD8+ TTNF cells with FMAdVTNF

To assess the optimal multiplicity of infection (MOI) for maximal transgene expression, CD8+ T cells were transfected with AdVGFP at various MOIs for 1 day and examined by fluorescence microscopy. The number of transfected CD8+ T cells expressing GFP at various MOIs was calculated. The above DCOVA-activated CD8+ T cells were further incubated with the fiber-modified FMAdVTNF-α and the control FMAdV at MOI of 100.34 Following the viral adsorption for 1 h at 37°, the T-cell culture medium was replaced with DMEM/10% FCS/IL-2 (10 U/ml), and the T cells were incubated for another 24 h at 37°. CD8+ T cells transfected with FMAdVTNF and the control FMAdV were referred to as CD8+ TTNF and CD8+ Tcontrol cells, respectively.

RT-PCR analysis of cytokine expression

Total RNA was obtained from the transfected B16OVA/IL−10 or the control B16OVA cells, as well as the DCOVA-activated CD8+ T cells, the engineered CD8+ TTNF cells and the control CD8+ Tcontrol cells. The first-strand cDNA synthesis for RT-PCR was performed with 5 μg of RNA using a commercial kit (Stratagene, La Jolla, CA), following the manufacturer's instructions. The PCR primers included the sense primer (5′ IndexTermatgag acttc tcctc ctg-3′) and the antisense primer (5′-IndexTermttacc cagtc agggt tactg-3′) for the mouse IL-10, and the sense primer (5′-IndexTermatgag acttc tcctc ctg-3′) and the antisense primer (5′-IndexTermttacc cagtc agggt tactg-3′) for the human TNF-α. In addition, another one set of primers was also used, including the sense primer (5′-IndexTermcaggt tgtct cctgc gactt-3′) and the antisense primer (5′-IndexTermcttgc tcagt gtcct tgctg-3′) for the control gene glyceraldehyde phosphate dehydrogenase (GAPDH). The PCR conditions comprised one cycle at 94°C (5 min), 54°C (1 min) and 72°C (1 min), and 25 cycles at 94°C (1 min), 55°C (1 min) and 72°C (1 min). All PCR reaction products were resolved on 1% agarose gels with ethidium bromide staining.33

Phenotypic characterization of engineered B16OVA/IL-10 tumor cells and TTNF cells

For phenotypic analysis, the engineered B16OVA/IL−10 or the original B16OVA cells as well as the engineered CD8+ TTNF and the control DCOVA-activated CD8+ T (Tcontrol) cells or DCOVA-activated CD8+ T cells, which were further cultured in medium containing recombinant IL-2 (10 U/ml) and recombinant human TNF-α (10 ng/ml) for 1 day and referred to as ‘T+rTNF’ cells, were stained with a panel of biotin-conjugated Abs (2 μg/ml) for 1 h on ice, washed with PBS, and then incubated for 1 h on ice with FITC-conjugated avidin. After another three washes with PBS, the above cells were analyzed by flow cytometry. To examine the intracellular expression of cytokines, the above cells were processed using a commercial kit (Cytofix/CytoPerm Plus with GolgiPlug; Pharmingen, Inc.), and stained with PE-conjugated anti-IL-4, -IL-10, -IFN-γ, -hTNF-α or -perforin Abs, according to the manufacturers’ protocols.33 Cytokine secretion by tumor or T cells in the supernatants from their 24 h cell culture in vitro was assessed using the commercial enzyme-linked immunosorbent assay (ELISA) kits (Endogen, Woburn, MA). To examine IL-10 secretion by B16OVA/IL−10 tumors in vivo, the protein extracts were prepared by homogenization of tumor tissues in 125 mM Tris, 0.05% sodium dodecyl sulfate (SDS), 10% β-mercaptoethanol, followed by centrifugation at 1000 g for 5 min. Cytokine secretion by tumors in vivo in the supernatants were also assessed using enzyme-linked immunosorbent assay (ELISA) kits. The results were normalized to the recombinant cytokine standard curves.

Tumor cell growth in vitro and in vivo

To assess tumor cell growth in vitro, the engineered B16OVA/IL−10 and the original B16OVA tumor cells (0.2 × 106 cells) were plated in each of the small flasks containing 5 ml of culture medium, and then counted viability of the cells using trypin blue exclusion method daily. To assess tumor cell growth in vivo, the above tumor cells (0.5 × 106 cells per mouse) were s.c. inoculated into the right thighs of C57BL/6 mice, and tumor growth was monitored daily using a caliper.

T-cell cytotoxicity assay

The engineered CD8+ TTNF and the control CD8+ Tcontrol cells were used as effector (E) cells, whereas the OVA-expressing B16OVA/IL−10 and the original B16 tumor cells without OVA expression were used as target (T) cells. The target cells were radio-labeled by culturing these cells for 1 h in the culture medium in presence of 50 μl of sodium[51Cr]-chromate (36 mCi/ml; Amersham, Arlington Heights, IL), and then washed twice with DMEM. Approximately 1 × 104 labeled target cells per triplicate well were mixed with effector cells at various E:T cell ratios, and then incubated for 6 h. For testing the killing mechanism, the effector cells were pre-incubated with CMA (1 μ M) and emetin (5 μ M) for 2 h to prevent perforin (CMA)- or Fas/FasL interaction (emetin)-mediated cytotoxicity.34 The percentage of specific lysis was calculated as: 100 × ((experimental count per minute (cpm)−spontaneous cpm)/(maximal cpm−spontaneous cpm)). Spontaneous cpm release in the absence of E cells was <10% of specific lysis. The maximal cpm release was determined by lysis of T cells with 0.25% Triton X-100.

Adoptive CD8+ T-cell immunotherapy models

For treatment of lung micrometastasis, C57BL/6 mice (10 per group) were i.v. injected with B16 and B16OVA/IL−10 tumor cells (0.5 × 106 cells per mouse), respectively. At 12 days after tumor cell injection, mice had lung micrometastasis.5 These mice were then treated with i.v. injection of the engineered CD8+ TTNF and the control CD8+ Tcontrol or ‘T+rTNF’ cells (2 × 106 cells per mouse). The mice were killed 4 weeks after tumor cell injection, and the lung metastatic tumor colonies were counted in a blind fashion.27 Metastases on freshly isolated lungs appeared as discrete black-pigmented foci that were easily distinguishable from normal lung tissues and confirmed by histological examination. Metastatic foci too numerous to count were assigned an arbitrary value of >100. For treatment of established solid tumors, C57BL/6 mice (10 per group) were s.c. inoculated with 0.5 × 106 B16OVA/IL−10 tumor cells in their thighs. At 10 days after tumor cell inoculation, tumor grew to 5 mm in diameter. The tumor-bearing mice were then i.v. injected with the engineered CD8+ TTNF and the control CD8+ Tcontrol cells (5 × 106 cells per mouse), respectively. The tumor growth was then monitored daily using a caliper. To assess long-term antitumor immunity, C57BL/6 mice (12 per group) were i.v. immunized with the engineered CD8+ TTNF and the control Tcontrol cells (5 × 106 cells per mouse). The immunized mice were then challenged with s.c. injection of 0.3 × 106 or 2 × 106 B16OVA/IL−10 tumor cells 3 months after immunization. The tumor growth was then monitored daily by measuring the diameter of the tumor mass using a caliper. For humanitarian reasons, all mice with tumors that achieved a size of 1.5 cm in diameter were killed. Log rank test and Graphpad Prism software were used to compare the mouse survival data.35

T-cell survival

Naive C57BL/6.1. (CD45.1+, B6.1) mice were i.v. injected with the engineered CD8+(45.2+) TTNF and the control CD8+(45.2+) Tcontrol cells (5 × 106 cells per mouse) derived from CD45.2+ OT I mice. A H-2Kb/OVA257–264 tetramer staining assay was performed to examine the presence of OVA-specific CD8+ T cells in the mouse peripheral blood at 5 days and once a week for 3 months after adoptive transfer. The tail blood samples were incubated with PE-labeled H-2Kb/OVA257–264 tetramer, FITC-labeled rat anti-mouse CD8 Ab as previously described.36 At 3 months after adoptive transfer, mice were boosted by i.v. injection of irradiated (4000 rad) DCOVA (0.5 × 106 cells per mouse). At 4 days after the boost, the tail blood samples from these mice were stained with PE-labeled H-2Kb/OVA257–264 tetramer, FITC-labeled anti-mouse CD8 and ECD-labeled anti-CD45.1 or CD45.2 Abs, and analyzed by flow cytometry. To assess the role of TNF-α in CD8+ T-cell survival, wild-type C57BL/6 mice and C57BL/6(TNF-α−/−) mice with TNF-α gene KO were i.v. injected with DCOVA-activated CD8+ T cells (5 × 106 cells per mouse) derived from OT I mice. In another set of experiments, naive C57BL/6 mice were i.v. injected with DCOVA-activated CD8+ T(TNF-α−/−) cells (5 × 106 cells per mouse) derived from OT I(TNF-α−/−) mice. A tetramer staining assay was performed to examine the presence of OVA-specific CD8+ T cells in the mouse peripheral blood at 5 days and once a week for 3 months after the adoptive transfer as described above.

Results

Transfected B16OVA/IL−10 cells express the transgene IL-10

We first cloned the mouse IL-10 gene and constructed an expression vector pcDNAIL-10 expressing the mouse IL-10. We then transfected the mouse melanoma cell line B16OVA expressing OVA with the expression vector pcDNAIL-10 and the control vector pcDNA to generate the transfected B16OVA/IL−10 and the control B16OVA/control tumor cell lines. To examine the IL-10 transgene expression, RNA extracted from the transfected B16OVA/IL−10, the control B16OVA/control and the original B16OVA tumor cells was subjected to RT-PCR analysis. As shown in Figure 2a, IL-10 expression was only found in the engineered B16OVA/IL−10 tumor cells, but not in the control B16OVA/control and the original B16OVA tumor cells by RT-PCR analysis. To quantify the IL-10 secretion, we performed ELISA analysis. IL-10 secretion of the engineered B16OVA/IL−10 tumor cells was estimated to be 2.2 ng/ml/106 cells/24 h.

Figure 2
figure2

Characterization of transfected B16OVA/IL−10 tumor cells. (a) Expression of IL-10. RNA was extracted from tumor cells including B16OVA, B16OVA/IL−10 and B16OVA/control. The first-strand cDNA was synthesized from RNA using the reverse transcriptase, and the PCRs were conducted using two sets of primers for IL-10 and the control ‘house-keeping’ gene GAPDH, respectively. (b) Phenotypic characterization of B16OVA/IL−10 tumor cells by flow cytometry. Tumor cells were stained with a panel of antibodies (solid lines) or isotype-matched irrelevant antibodies (dotted lines), and then analyzed by flow cytometry. (c) In vitro and in vivo tumor cell growth. In in vitro tumor cell growth assay, viable tumor cells in triplicate cultures of each cell line were counted daily. The curves are plotted from a representative assay. Each point represents the mean of triplicates. In in vivo tumor growth assay, tumor cells (0.5 × 106 cells per mouse) were s.c. injected into the right thigh of C57BL/6 mice and tumor growth was monitored daily by measuring tumor diameter using a caliper. The mean size of tumors in each group of five mice is presented. One representative experiment of three is shown.

Transfected B16OVA/IL−10 tumor cells show a similar phenotype and growth rate compared to B16OVA tumor cells

For phenotypic analysis, the transfected B16OVA/IL−10 and the control B16OVA/control tumor cells were analyzed by flow cytometry. The transfected B16OVA/IL−10 tumor cells expressed a similar amount of cell surface MHC class I and OVA as the control B16OVA/control and the original B16OVA tumor cells (Figure 2b). We then performed in vitro and in vivo growth rate assays. The transfected B16OVA/IL−10 tumor cells displayed a similar in vitro and in vivo growth rate compared to the original B16OVA tumor cells (Figure 2c). The IL-10 secretion by B16OVA/IL−10 tumors in vivo was estimated to be 16 μg/g of tumor tissues as determined by ELISA.

Engineered CD8+ TTNF cells express the transgene human TNF-α

The OVA-specific CD8+ T cells were generated by incubation of naïve CD8+ T cells from OT I transgenic mice with OVA-pulsed DCOVA. To assess the optimal MOI for maximal transgene expression, CD8+ T cells were transfected with AdVGFP at various MOIs and examined by fluorescence microscopy. The number of transfected CD8+ T cells expressing GFP increased with increased MOI. The maximal transfection with 85% GFP-positive CD8+ T cells was obtained at MOIs of 100. Therefore, MOI of 100 was selected for transfection of CD8+ T cells in this study. These CD8+ T cells were further transfected with the fiber-modified MFAdVTNF and the control MFAdV at MOI of 100, and termed as the engineered CD8+ TTNF and the control CD8+ Tcontrol cells. As shown in Figure 3a, these CD8+ T cells all expressed IL-2 receptor (CD25), active T-cell marker (CD69) and Vα2Vβ5+ TCR, indicating that they are active OVA-specific CD8+ T cells. These active CD8+ T cells also expressed intracellular mouse IFN-γ, but not IL-4, indicating that they are type 1 CD8+ T cells. In addition, CD8+ Tcontrol cells did not express intracellular human TNF-α, whereas CD8+ TTNF cells did express the transgene-coded human TNF-α. To quantify the cytokine secretion, we performed ELISA analysis. These CD8+ T cells secreted the mouse IFN-γ (0.6 ng/ml/106 cells/24 h) and the mouse TNF-α (0.4 ng/ml/106 cells/24 h), but did not secrete any mouse IL-4 in their supernatants. In addition, they also displayed cell surface FasL and intracellular perforin. The phenotypes of CD8+ ‘T+rTNF’ cells cultured in medium containing recombinant human TNF-α were similar to the control CD8+ Tcontrol cells cultured in medium without recombinant TNF-α (data not shown). The phenotypes of the engineered CD8+ TTNF cells were also similar to those of the control CD8+ Tcontrol cells, except for an increased expression of perforin in engineered CD8+ TTNF cells. To assess the transgene expression, we performed RT-PCR analysis. As shown in Figure 3b, the human TNF-α expression was only found in the engineered CD8+ TTNF cells, but not in the control CD8+ Tcontrol and the untransfected CD8+ T cells. To quantify transgene (human TNF-α) secretion of the engineered TTNF cells, we also performed ELISA assay. Human TNF-α secretion of the engineered TTNF cells was estimated to be 4.3 ng/ml per 106 cells per24 h, which is more than 10-fold than the secreted mouse TNF-α. Interestingly, the human TNF-α secretion of the engineered T cells generated by using the original AdVTNF without fiber modification26 was estimated to be only 0.8 ng/ml/106 per cells per 24 h, more than fivefold less than that (4.3 ng/ml/106 cells/24 h) of the engineered CD8+ TTNF cells generated by using fiber-modified MFAdVTNF, indicating that the adenoviral fiber modification with RGD motif insertion into the viral knob's H1 loop greatly enhances AdV infection to CD8+ T cells. The cytokine profiles of the engineered CD8+ TTNF cells are also similar to those of the control CD8+ Tcontrol cells and the untransfected CD8+ T cells (data not shown), except that the later two did not express the human TNF-α.

Figure 3
figure3

Characterization of engineered CD8+ TTNF cells. (a) Phenotypic characterization of engineered CD8+ TTNF cells. CD8+ T cells were stained with a panel of antibodies (solid lines) and isotype-matched irrelevant antibodies (dotted lines), and then analyzed by flow cytometry. (b) Expression of the transgene human TNF-α. RNA was extracted from T cells. The first-strand cDNA was synthesized from RNA using the reverse transcriptase, and the PCRs were conducted using a set of primers for the human TNF-α and the control ‘house-keeping’ gene GAPDH, respectively. One representative experiment of two is displayed.

Engineered CD8+ TTNF cells display stronger cytotoxicity to OVA-expressing B16OVA tumor cells in vitro

To assess the T cell cytotoxicity, we performed chromium release assay. As shown in Figure 4a, the engineered CD8+ TTNF cells exhibited stronger cytotoxicity (75% specific killing at an E:T cell ratio of 50) for OVA-expressing B16OVA/IL−10 tumor cells than the control CD8+ Tcontrol cells (48% specific killing at an E:T cell ratio of 50) (P<0.05). However, the above killing activities were not for the parental B16 tumor cells without OVA expression, indicating that the killing activity is OVA specific. To study the killing mechanism, concanamycin A (CMA) and emetin were used to inhibit the perforin- and Fas/FasL interaction-mediated cytotoxicity, respectively. Our data showed that both CMA and emetin demonstrated a dose-dependent inhibition of the engineered CD8+ TTNF (Figure 4b) and the control CD8+ Tcontrol (data not shown)-cell cytotoxicity. The treatment of CMA at 1 μ M induced almost a complete inhibition (94%) of TTNF-cell cytotoxicity, whereas the treatment of emetin even at 5 μ M resulted in a maximal inhibition (12%), indicating that the perforin-mediated pathway plays a major role in CTL cytotoxicity.

Figure 4
figure4

Cytotoxicity assay. (a) The engineered CD8+ TTNF and the control CD8+ Tcontrol cells were used as effector (E) cells, whereas the OVA-expressing B16OVA/IL−10 and the original B16 tumor cells were used as target (T) cells. The data are presented as the percent-specific lysis of the target cells in a 6 h 51Cr-release assay. Each point represents the mean of triplicate cultures. *P<0.05 versus cohorts of the control CD8+ Tcontrol cells (Student’s t-test). (b) In inhibition assay, the engineered CD8+ TTNF cells were pre-incubated with different concentrations of CMA and emetin for 2 h and then used as effector cells in cytotoxicity assay. The data are presented as the percent inhibition of CD8+ TTNF-cell-mediated B16OVA/IL−10 cytolysis at an E:T ratio of 50:1 in the presence of different concentrations of CMA and emetin. One representative experiment of three is depicted.

IL-10-secreting tumors are more resistant to CD8+ T-cell therapy

To treat lung metastasis, the control CD8+ Tcontrol cells were i.v. transferred into the mice bearing 12-day lung micrometastasis. As shown in Experiment I of Table 1, the control CD8+ Tcontrol cells eradicated B16OVA lung metastasis without IL-10 secreting in all 10/10 mice, but eliminated IL-10-secreting B16OVA/IL−10 lung metastasis in only 2/10 mice, indicating that immunosuppressive IL-10-secreting tumors are more resistant to CD8+ T-cell therapy. To assess the therapeutic effect on solid tumors, the control CD8+ Tcontrol cells were i.v. transferred into the mice bearing 10-day solid tumors (5 mm in diameter). All mice died within 16 days subsequent to tumor cell challenge. In mice bearing B16OVA tumors, 40% (4/10) of the mice treated with the control CD8+ Tcontrol cells were free of tumors (Figure 5a). The original tumors (5 mm in diameter) in these mice with tumor regression gradually disappeared starting at day 2 after CD8+ T-cell transfer. In mice bearing B16OVA/IL−10 tumors, however, all (10/10) mice with treatment of the control CD8+ Tcontrol cells died within 24 days after tumor cell challenge (Figure 5b), confirming that IL-10-secreting tumors are more resistant to CD8+ T-cell therapy.

Table 1 Transfer of engineered CD8+ TTNF cells eradicate IL-10-secreting B16OVA/IL−10 lung micrometastasis
Figure 5
figure5

Impact on mortality rates in adoptive engineered TTNF-cell therapy. Mice (10 per group) bearing 10-day (a) B16OVA and (b) B16OVA/IL−10 tumors (5 mm in diameter) were given i.v. injections of 5 × 106 CD8+ TTNF or Tcontrol cells, respectively. Tumor growth was monitored and the tumor size (diameter) measured daily using a caliper. The evolution of the tumors in individual mice is depicted, as are the fractions of mice in each treatment group that were tumor free at 60 days post-treatment. The survival of CD8+ TTNF-cell-treated mouse group* is significantly longer than that of the Tcontrol-cell-treated group (Log rank test, P<0.05). One representative experiment of two is shown.

Engineered CD8+ TTNF cells have enhanced therapeutic effect on IL-10-secreting lung metastasis and solid tumors

To assess the therapeutic efficacy on lung metastasis, the engineered CD8+ TTNF cells were i.v. transferred into the mice bearing 12-day lung micrometastasis. As shown in Table 1, DCOVA-activated CD8+ T (Tcontrol) and DCOVA-activated CD8+ T (T+rTNF) cells with further culturing in the medium containing recombinant human TNF-α did not efficiently cure IL-10-secreting B16OVA/IL−10 lung metastasis because 80–90% of mice in these two groups of mice still died after the treatment. However, the engineered CD8+ TTNF cells not only eradicated the original B16OVA, but also the IL-10-secreting B16OVA/IL−10 lung metastasis in all 10/10 mice, indicating that the engineered CD8+ TTNF cells have enhanced therapeutic effect on the IL-10-secreting B16OVA/IL−10 lung metastasis. To assess the therapeutic effect on solid tumors, the engineered CD8+ TTNF cells were i.v. transferred into the mice bearing 10-day solid tumors (5 mm in diameter). In mice bearing the original B16OVA tumors, 90% (9/10) of the mice treated with the engineered CD8+ TTNF cells were free of tumors (Figure 5a). In mice bearing the IL-10-secreting B16OVA/IL−10 tumors, 60% (6/10) of the mice with treatment of the engineered CD8+ TTNF cells were still free of tumors, confirming that the engineered CD8+ TTNF cells have stronger therapeutic effect on both B16OVA and IL-10-secreting B16OVA/IL−10 solid tumors than the control CD8+ Tcontrol cells. The tumor growth rate in the five mice still bearing tumors in CD8+ TTNF-cell-treated group was also significantly delayed compared to the group of mice treated with the control CD8+ Tcontrol cells (P<0.05).

Engineered CD8+ TTNF cells have prolonged survival in vivo

The successful cancer T-cell therapy is critically dependent upon duration of in vivo persistence of the transferred CD8+ T cells.37 To assess in vivo survival of the engineered CD8+ TTNF cells, we performed a kinetic study. In the kinetic study, we measured the amount of OVA-specific CD8+ T cells in the peripheral blood of B6.1 mice (CD45.1+) with i.v. transfer of CD8+ T cells derived from OT I mice (CD45.2+) using PE-labeled H-2Kb/OVA257–264 tetramer staining at different time points after adoptive transfer. As shown in Figure 6a and b, the number of totally detected OVA-specific CD8+ T cells in peripheral blood of the mice with the engineered CD8+ TTNF and the control CD8+ Tcontrol cell transfer accounted for 18 and 12% of the total CD8+ T-cell population, respectively, at day 5 after adoptive transfer. The numbers then gradually dropped to 14 and 9.3% at day 12, and 1.8 and 1.2% at day 30 subsequent to adoptive transfer, and then stably maintained for at least 3 months, indicating that the engineered CD8+ TTNF cells have prolonged survival in vivo, compared with the control CD8+ Tcontrol cells. Since these long-term survival CD8+ T cells should be considered to be memory T (Tm) cells with the capacity of expansion upon antigen stimulation,38 we boosted these mice with OVA-pulsed DCOVA 3 months after adoptive transfer, and then measured the number of OVA-specific CD8+ T cells in the peripheral blood 4 days after the boost. As shown in Figure 6c, the expanded OVA-specific CD8+ T cells accounted for 22 and 34% of the total CD8+ T-cell population in the mice with previous transfer of CD8+ Tcontrol and TTNF cells, respectively. To examine the origin of these expanded OVA-specific CD8+ T cells, we performed ECD-CD45.1 or ECD-CD45.2 staining. The majority (80%) of the expanded OVA-specific CD8+ T cells were CD45.2 positive, indicating that they are derived from the originally transferred donor CD8+ T cells. Interestingly, we also found that around 20% of the expanded OVA-specific CD8+ T cells were CD45.1-positive, indicating that they are derived from the host.

Figure 6
figure6

CD8+ T-cell survival in vivo after adoptive transfer. (a, b) The engineered CD8+45.2+ TTNF and the control CD8+45.2+ Tcontrol cells derived from OT I mice (CD45.2+) were i.v. injected into C57BL/6.1 (CD45.1+) mice (10 per group). Mouse tail blood cells were stained with PE-H-2Kb/OVAI tetramer (PE-tetramer), FITC-anti-CD8 Ab (FITC-CD8) and ECD-anti-CD45.1 (ECD-CD45.1) or -anti-CD45.2 (ECD-45.2) Ab, and analyzed by flow cytometry at the indicated time points after adoptive transfer. The value in each panel represents the percentage of PE-tetramer-positive CD8+ T cells versus the total peripheral CD8+ T-cell population. The value in parenthesis represents the standard deviation. *P<0.05 versus cohorts of the control CD8+ Tcontrol cells (Student’s t-test). (c) The PE-tetramer-positive CD8+ T cells in the circle are also gated for further analysis of either ECD-45.1 or ECD-45.2 expression by flow cytometry. (d) At 3 months after adoptive transfer, the mice (12 per group) were s.c. challenged with 0.3 × 106 or 2 × 106 B16OVA/IL−10 tumor cells. Mouse survival was monitored daily. The survival of TTNF-cell-transferred mouse group* against 2 × 106 B16OVA/IL−10 tumor cell challenge is significantly longer than that of the Tcontrol-cell-transferred group (Log rank test, P<0.05). The results presented are representative of three in the above different experiments is shown.

Adoptive transfer of engineered CD8+ TTNF cells induces long-term antitumor immunity

We next investigated the potential long-term immunity derived from adoptive transfer of CD8+ TTNF cells. The immunized mice with CD8+ TTNF and Tcontrol-cell transfer were challenged with s.c. injection of B16OVA/IL−10 tumor cells 3 months after adoptive transfer. As shown in Figure 6d, transfer of both the engineered CD8+ TTNF and the control CD8+ Tcontrol cells completely protected 12/12 (100%) mice from a challenge of 0.3 × 106 B16OVA/IL−10 tumor cells. However, when challenging the mice with 2 × 106 B16OVA/IL−10 tumor cells, only 3/12 (25%) mice immunized with the control CD8+ Tcontrol cells were protected from tumor growth. In the mice immunized with the engineered CD8+ TTNF cells, 10/12 (83% mice) were protected from tumor growth (P<0.05), indicating that the long-term antitumor immunity induced by adoptive transfer of engineered CD8+ TTNF cells is significantly stronger than that of the control CD8+ Tcontrol cells. In addition, the tumor-bearing mice with complete solid tumor regression after transfer of the engineered CD8+ TTNF cells also showed a similar long-term antitumor protection as the wild-type C57BL/6 mice with the engineered CD8+ TTNF-cell transfer (data not shown).

Transgene TNF-α expression sustains engineered CD8+ T-cell survival in vivo

To investigate the role of TNF-α derived from the CD8+ T cells and the host in transferred CD8+ T-cell survival in vivo, we conducted another kinetic study by transferring CD8+ T cells derived from OT I or OT I(TNF-α−/−) mice to C57BL/6 mice or C57BL/6 mice with TNF-α gene KO. As shown in Figure 7a and b, the numbers of detected OVA-specific CD8+ T cells in peripheral blood of C57BL/6 mice with DCOVA-activated CD8+ T- and T(TNF-α−/−)-cell transfer accounted for around 12% of the total CD8+ T-cell population at day 5 subsequent to adoptive transfer. The numbers of detected OVA-specific CD8+ T cells became 8.2 and 1.2% in the peripheral blood of C57BL/6 mice with transfer of CD8+ T cells at 12 and 30 or even 90 days subsequent to adoptive transfer (Figure 7a). However, the numbers dramatically dropped to 0.5%, and then completely disappeared in the peripheral blood of C57BL/6 mice with transfer of CD8+ T(TNF-α−/−) cells at 12 and 30 or even 90 days subsequent to adoptive transfer (Figure 7a), indicating that TNF-α derived from the donor CD8+ T cells plays an important role in transferred donor CD8+ T-cell survival in vivo. In addition, the number of detected OVA-specific CD8+ T cells in peripheral blood of C57BL/6 mice with TNF-α gene KO were 10% of the total CD8+ T-cell population at day 5 subsequent to adoptive transfer of DCOVA-activated CD8+ T cells. However, the numbers also dramatically dropped to 0.2%, and then completely disappeared in the peripheral blood at 12 and 30 or even 90 days subsequent to adoptive T-cell transfer (Figure 7a), indicating that TNF-α derived from the host also plays an important role in transferred donor CD8+ T-cell survival in vivo.

Figure 7
figure7

The in vivo survival of CD8+ T cells with TNF-α gene knockout after adoptive transfer. (a, b) The DCOVA-activated CD8+ T cells derived from OT I mice and the DCOVA-activated CD8+ T(TNF−/−) cells derived from OT I mice with TNF-α gene KO were i.v. injected to C56BL/6 (B6) mice for kinetic study, and termed T/B6 and T(TNF−/−)/B6, respectively. The DCOVA-activated CD8+ T cells derived from OT I mice were i.v. injected to C57BL/6 mice with TNF-α gene KO with and termed T/B6(TNF−/−). Mouse tail blood cells were stained with PE-H-2Kb/OVAI tetramer (PE-tetramer) and FITC-anti-CD8 Ab, and analyzed by flow cytometry at the indicated time points after T-cell transfer. The value in each panel represents the percentage of PE-tetramer-positive CD8+ T cells versus the total peripheral CD8+ T-cell population. The value in parenthesis represents the standard deviation. *P<0.05 versus cohorts of CD8+ T(TNF−/−) cells in B6 mice or CD8+ T cells in B6 mice with TNF-α KO (Student’s t-test).One representative experiment of two in the above different experiments is shown.

Discussion

Retrovirus and adenovirus are commonly used for delivery of transgenes to cells or tissues.39 Retroviruses integrate their genome into the host DNA of dividing cells, and this provides the possibility of long-term transgene expression. However, its random integration may cause insertional mutagenesis in the host cells. Recently it has been shown that retroviral vectors can cause leukemia formation.40 Efficient infection of target cells by AdV requires two predominant viral-host cell receptor interactions. Primary adsorption of AdV relies on the interaction of the viral knob domain with a cellular surface receptor designated coxsackievirus and adenovirus receptor (CAR).41 Subsequently, the interaction of RGD motifs on the viral penton base with the cellular αVβ3 and αVβ5 integrins facilitates AdV entry via clathrin-mediated endocytosis.42, 43 Although AdV can infect a wide range of cell types, lymphocytes and DCs are not generally susceptible to AdV infection, in part because of the absence of CAR expression44, 45 Fiber-gene modification by insertion of RGD sequence in the viral fiber knob's H1 loop has been shown to enhance AdV infection of DC through binding of RGD motif to DC integrins leading to enhanced DC-mediated antitumor immunity.46, 47 To enhance T-cell infection, we constructed a recombinant MFAdVTNF with fiber-gene modification and transfected CD8+ T cells with MFAdVTNF. We found a fivefold increase of transgene human TNF-α secretion by CD8+ TTNF cells transfected with the fiber gene-modified MFAdVTNF, compared to CD8+ T cells transfected with the original AdVTNF without viral fiber modification. CTLs are able to lyse target cells by two mechanistically distinct, but functionally similar mechanisms:48 an antigen-specific Ca2+-dependent perforin/granzyme mechanism and an antigen-nonspecific Ca2+-independent Fas/FasL mechanism. We showed that the engineered CD8+ TTNF cells exhibited an upregulation of perforin and displayed stronger in vitro cytotoxicity to tumor cells than the control Tcontrol cells. We also demonstrated that the in vitro TTNF cell cytotoxicity is mainly mediated by perforin secretion, which is consistent with our previous reports.33, 34

In this study, we also analyzed the therapeutic effect of engineered OVA-specific CD8+ TTNF cells in mice bearing either 12-day lung micrometastasis or solid tumors (5 mm in diameter) secreting IL-10. We showed that the engineered CD8+ TTNF cells secreting the transgene-coded human TNF-α eradicated the IL-10-secreting B16OVA/IL−10 lung metastasis in all 10/10 mice, whereas the control DCOVA-activated CD8+ T (Tcontrol) and DCOVA-activated CD8+ T (T+rTNF) cells with further culturing in the medium containing the recombinant human TNF-α eliminated lung metastasis in only 2/10 and 1/10 mice, respectively, indicating that the engineered CD8+ TTNF cells have enhanced in vivo therapeutic effect on the IL-10-secreting B16OVA/IL−10 lung metastasis. In mice bearing solid B16OVA/IL−10 tumors, all 10/10 mice with treatment of the control CD8+ Tcontrol cells died within 24 days after tumor cell challenge, whereas 6/10 mice with treatment of the engineered CD8+ TTNF cells were free of tumors, confirming that the engineered TTNF cells have stronger in vivo therapeutic effect on the IL-10-secreting B16OVA/IL−10 solid tumors than the control CD8+ Tcontrol cells. To improve further the therapeutic effect, a combinational immunotherapy using the engineered CD8+ TTNF cells and CD4+ Th cells and/or the chemokine gene (lymphotactin and IP-10) therapy for enhancement of CD8+ T-cell tumor infiltration33, 49, 50 is underway in our laboratory to eradicate completely IL-10-secreting B16OVA/IL−10 tumors (5 mm in diameter and even larger in size).

It has been reported that the mechanism of in vivo CD8+ T-cell-mediated tumor regression is distinct from that of in vitro CD8+ T-cell-mediated cytotoxicity, and is independent of perforin and FasL by using CD8+ T cells with perforin gene knockout or FasL mutant51, 52 Recently, Hollenbaugh et al.53 demonstrated that CD8+ T cell-secreted, but not the host-derived IFN-γ play an important role in eradication of solid tumors by regulating CD8+ T-cell expansion and migration and inducing tumor cell apoptosis. Blohm et al.54 visualized adoptive CD8+ T-cell-mediated tumor destruction and found that tumors ‘melt’ from the inside were derived from destruction of tumor vasculature and accumulation of granulocytes. The enhancement of in vivo therapeutic effect of the engineered TTNF cells observed in this study may not be due to its upregulated expression of perforin, but may be associated with its enhanced secretion of the transgene-coded human TNF-α as well as its elongated survival in vivo. TNF-α plays an essential role in CD8+ T-cell-mediated elimination of tumors in vivo52, 55 possibly by its direct cytotoxicity to tumor cells leading apoptosis through engagement of the p55 TNF-α receptor I.56 The inflammatory cytokine TNF-α also activates macrophages, natural killer cells and T cells20, 57 by its binding to the p75 TNF-α receptor II (TNFRII).55 In addition, it deactivate αVβ3 integrin on angiogenic endothelial cells leading to lack of adhesion and apoptosis58 and promoting intravascular thrombosis.59 Recently, it has further been reported that TNF-α inhibits the suppressive function of regulatory T (Tr) cells by upregulated expression of TNFRII on Tr cells leading to downregulation of FoxP3 expression.60 Therefore, in addition to the IFN-γ secreted by engineered CD8+ T cells, enhanced secretion of TNF-α by engineered CD8+ T cells may (i) induce destruction of tumor vasculatures, (ii) activate immune cells, (iii) suppress Tr cells and (iv) directly induce tumor cell apoptosis, leading to enhancement of in vivo therapeutic effect of the engineered TTNF cells in this study.

The critical importance for a successful T-cell therapy is duration of in vivo persistence of the transferred CD8+ T cells.37 Unfortunately, in vitro cultured antigen-specific CD8+ T cells, particularly CD8+ T-cell clones, often die only a few hours after adoptive transfer and generally do not survive more than a matter of days, which greatly limits treatment efficacy.61, 62 This is because the vast majority of activated CD8+ T cells die by activation-induced cell death via apoptosis, and only a small portion of these T cells survives for extended periods.63 In this study, we found that the number of detected OVA-specific CD8+ T cells in peripheral blood of the mice with CD8+ TTNF and CD8+ Tcontrol−cell transfer accounted for 18 and 12% of the total CD8+ T-cell population at day 5 after adoptive transfer. However, the numbers gradually dropped to 14 and 9.3% at day 12, and 1.8 and 1.2% at day 30 subsequent to adoptive transfer, but stably maintained for at least 3 months thereafter, indicating that the engineered CD8+ TTNF cells have prolonged survival in vivo, compared with the control CD8+ Tcontrol cells. The importance of TNF-α in CD8+ T-cell survival in vivo has been elucidated by using C57BL/6 and OT I mice with TNF-α gene KO. Our data showed that the number of detected OVA-specific CD8+ T cells in peripheral blood of C57BL/6 mice with CD8+ T(TNF-α−/−)-cell transfer or in peripheral blood of C57BL/6(TNF-α−/−) mice with CD8+ T-cell transfer all dramatically dropped to 0.5 and 0.2% of the total CD8+ T-cell population and both completely disappeared at days 12 and 30 after transfer, respectively, indicating that both the donor’s cellular and the host TNF-α are important in CD8+ T-cell survival in vivo. Recently, it has been reported that membrane TNF-α can delay T-cell activation-induced cell death by activation of AKT and NF-κB,64 which leads to elongated T-cell survival. Our results are thus consistent with a previous report by Dobrazanski, et al.,65 demonstrating that CD8+ T-cell therapeutic effect was significantly less efficient in TNF-α−/− KO mice. These long-lived memory T cells possess the unique ability to respond rapidly and specifically to the antigen.38 Our data showed that, after the boost, the expanded OVA-specific CD8+ T cells accounted for 22 and 36% of the total CD8+ T-cell population in the mice with previous transfer of the control CD8+ Tcontrol and the engineered CD8+ TTNF cells, respectively. Interestingly, our data also demonstrated that transfer of the engineered CD8+ TTNF cells further leads to a stronger long-term antitumor immunity against the IL-10-secreting B16OVA/IL−10 tumor cell challenge.

Dobrazanski et al.65 previously reported the induction of the host antigen-specific CD8+ Tm cells in lung tissues after adoptive transfer of the donor antigen-specific CD8+ effector T cells. However, the mechanism involved in this interesting finding was not elucidated. In this study, we also examined the origin of expanded OVA-specific CD8+ Tm cells in C57BL/6.1 (CD45.1+) mice with transfer of OVA-specific CD8+45.2+ TTNF cells using PE-tetramer, FITC-CD8 and ECD-CD45.1 or ECD-CD45.2 stainings. Our data demonstrated that a majority (80%) of the expanded OVA-specific CD8+ T cells were still CD45.2 positive, indicating that they are derived from the originally transferred donor CD8+ T cells. Interestingly, we also found that around 20% of the expanded OVA-specific CD8+T cells were CD45.1 positive, indicating that they are derived from the CD45.1-positive host. We have recently demonstrated that DC-activated CD8+ T cells acquired peptide/MHC I (pMHC I) complexes and costimulatory molecules from DC and acted as T-antigen presenting cell (T-APC) capable of stimulating the host central memory CD8+ T-cell responses via acquired pMHC I and CD80 costimulation, and IL-2 secretion.36 Therefore, the expanded CD45.1-positive CD8+ T cells observed in this study should be derived from the donor’s CD8+ TTNF-cell-stimulated host central memory CD8+ T-cell responses.

Taken together, our data show that the recombinant fiber-modified adenovirus MFAdVTNF has enhanced capacity to mediate T-cell transfection. The engineered CD8+ TTNF cells with MFAdVTNF-mediated gene transfer secrete significant amount of transgene-coded TNF-α, exhibit enhanced cytotoxicity and elongated survival in vivo, eradicate immunosuppressive IL-10-secreting lung metastasis and solid tumors, and induce enhanced long-term antitumor immunity after adoptive transfer. Therefore, CD8+ T cells engineered to secrete TNF-α may be useful when designing strategies for adoptive T-cell therapy of solid tumors.

References

  1. 1

    Plautz GE, Touhalisky JE, Shu S . Treatment of murine gliomas by adoptive transfer of ex vivo activated tumor-draining lymph node cells. Cell Immunol 1997; 178: 101–107.

  2. 2

    Peng L, Shu S, Krauss JC . Treatment of subcutaneous tumor with adoptively transferred T cells. Cell Immunol 1997; 178: 24–32.

  3. 3

    Tanaka H, Yoshizawa H, Yamaguchi Y, Ito K, Kagamu H, Suzuki E et al. Successful adoptive immunotherapy of murine poorly immunogenic tumor with specific effector cells generated from gene-modified tumor-primed lymph node cells. J Immunol 1999; 162: 3574–3582.

  4. 4

    Eck SC, Turka LA . Adoptive transfer enables tumor rejection targeted against a self-antigen without the induction of autoimmunity. Cancer Res 2001; 61: 3077–3083.

  5. 5

    Dobrzanski MJ, Reome JB, Dutton RW . Type 1 and type 2 CD8+ effector T cell subpopulations promote long-term tumor immunity and protection to progressively growing tumor. J Immunol 2000; 164: 916–925.

  6. 6

    Becker C, Pohla H, Frankenberger B, Schuler T, Assenmacher M, Schendel DJ et al. Adoptive tumor therapy with T lymphocytes enriched through an IFN-gamma capture assay. Nat Med 2001; 7: 1159–1162.

  7. 7

    Gastl GA, Abrams JS, Nanus DM, Oosterkamp R, Silver J, Liu F et al. Interleukin-10 production by human carcinoma cell lines and its relationship to interleukin-6 expression. Int J Cancer 1993; 55: 96–101.

  8. 8

    Smith DR, Kunkel SL, Burdick MD, Wilke CA, Orringer MB, Whyte RI et al. Production of interleukin-10 by human bronchogenic carcinoma. Am J Pathol 1994; 145: 18–25.

  9. 9

    Taga K, Mostowski H, Tosato G . Human interleukin-10 can directly inhibit T-cell growth. Blood 1993; 81: 2964–2971.

  10. 10

    Steinbrink K, Graulich E, Kubsch S, Knop J, Enk AH . CD4(+) and CD8(+) anergic T cells induced by interleukin-10-treated human dendritic cells display antigen-specific suppressor activity. Blood 2002; 99: 2468–2476.

  11. 11

    Groux H, Bigler M, de Vries JE, Roncarolo MG . Interleukin-10 induces a long-term antigen-specific anergic state in human CD4+ T cells. J Exp Med 1996; 184: 19–29.

  12. 12

    Hsieh CL, Chen DS, Hwang LH . Tumor-induced immunosuppression: a barrier to immunotherapy of large tumors by cytokine-secreting tumor vaccine. Hum Gene Ther 2000; 11: 681–692.

  13. 13

    Hara M, Kingsley CI, Niimi M, Read S, Turvey SE, Bushell AR et al. IL-10 is required for regulatory T cells to mediate tolerance to alloantigens in vivo. J Immunol 2001; 166: 3789–3796.

  14. 14

    Yang AS, Lattime EC . Tumor-induced interleukin 10 suppresses the ability of splenic dendritic cells to stimulate CD4 and CD8 T-cell responses. Cancer Res 2003; 63: 2150–2157.

  15. 15

    Kawamura K, Bahar R, Natsume W, Sakiyama S, Tagawa M . Secretion of interleukin-10 from murine colon carcinoma cells suppresses systemic antitumor immunity and impairs protective immunity induced against the tumors. Cancer Gene Ther 2002; 9: 109–115.

  16. 16

    Garcia-Hernandez ML, Hernandez-Pando R, Gariglio P, Berumen J . Interleukin-10 promotes B16-melanoma growth by inhibition of macrophage functions and induction of tumour and vascular cell proliferation. Immunology 2002; 105: 231–243.

  17. 17

    Chau GY, Wu CW, Lui WY, Chang TJ, Kao HL, Wu LH et al. Serum interleukin-10 but not interleukin-6 is related to clinical outcome in patients with resectable hepatocellular carcinoma. Ann Surg 2000; 231: 552–558.

  18. 18

    De Vita F, Orditura M, Galizia G, Romano C, Lieto E, Iodice P et al. Serum interleukin-10 is an independent prognostic factor in advanced solid tumors. Oncol Rep 2000; 7: 357–361.

  19. 19

    Robinet E, Branellec D, Termijtelen AM, Blay JY, Gay F, Chouaib S et al. Evidence for tumor necrosis factor-alpha involvement in the optimal induction of class I allospecific cytotoxic T cells. J Immunol 1990; 144: 4555–4561.

  20. 20

    Lahn M, Kalataradi H, Mittelstadt P, Pflum E, Vollmer M, Cady C et al. Early preferential stimulation of gamma delta T cells by TNF-alpha. J Immunol 1998; 160: 5221–5230.

  21. 21

    Zhang W, Chen Z, Li F, Kamencic H, Juurlink B, Gordon JR et al. Tumour necrosis factor-alpha (TNF-alpha) transgene-expressing dendritic cells (DCs) undergo augmented cellular maturation and induce more robust T-cell activation and anti-tumour immunity than DCs generated in recombinant TNF-alpha. Immunology 2003; 108: 177–188.

  22. 22

    Lyakh LA, Koski GK, Telford W, Gress RE, Cohen PA, Rice NR et al. Bacterial lipopolysaccharide, TNF-alpha, and calcium ionophore under serum-free conditions promote rapid dendritic cell-like differentiation in CD14+ monocytes through distinct pathways that activate NK-kappa B. J Immunol 2000; 165: 3647–3655.

  23. 23

    Sepulveda H, Cerwenka A, Morgan T, Dutton RW . CD28, IL-2-independent costimulatory pathways for CD8 T lymphocyte activation. J Immunol 1999; 163: 1133–1142.

  24. 24

    Brunner C, Seiderer J, Schlamp A, Bidlingmaier M, Eigler A, Haimerl W et al. Enhanced dendritic cell maturation by TNF-alpha or cytidine-phosphate-guanosine DNA drives T cell activation in vitro and therapeutic anti-tumor immune responses in vivo. J Immunol 2000; 165: 6278–6286.

  25. 25

    Brossart P, Zobywalski A, Grunebach F, Behnke L, Stuhler G, Reichardt VL et al. Tumor necrosis factor alpha and CD40 ligand antagonize the inhibitory effects of interleukin 10 on T-cell stimulatory capacity of dendritic cells. Cancer Res 2000; 60: 4485–4492.

  26. 26

    Wright P, Braun R, Babiuk L, Littel-van den Hurk SD, Moyana T, Zheng C et al. Adenovirus-mediated TNF-alpha gene transfer induces significant tumor regression in mice. Cancer Biother Radiopharm 1999; 14: 49–57.

  27. 27

    Xiang J, Huang H, Liu Y . A new dynamic model of CD8+ T effector cell responses via CD4+ T helper-antigen-presenting cells. J Immunol 2005; 174: 7497–7505.

  28. 28

    Li M, Davey GM, Sutherland RM, Kurts C, Lew AM, Hirst C et al. Cell-associated ovalbumin is cross-presented much more efficiently than soluble ovalbumin in vivo. J Immunol 2001; 166: 6099–6103.

  29. 29

    He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, Vogelstein B et al. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci USA 1998; 95: 2509–2514.

  30. 30

    Liu Y, Ye T, Sun D, Maynard J, Deisseroth A . Conditionally replication-competent adenoviral vectors with enhanced infectivity for use in gene therapy of melanoma. Hum Gene Ther 2004; 15: 637–647.

  31. 31

    Chan T, Sami A, El-Gayed A, Guo X, Xiang J . HER-2/neu-gene engineered dendritic cell vaccine stimulates stronger HER-2/neu-specific immune responses compared to DNA vaccination. Gene ther 2006; 13: 1391–1402.

  32. 32

    Zeng M, Smith SK, Siegel F, Shi Z, Van Kampen KR, Elmets CA et al. AdEasy system made easier by selecting the viral backbone plasmid preceding homologous recombination. Biotechniques 2001; 31: 260–262.

  33. 33

    Huang H, Liu Y, Xiang J . Synergistic effect of adoptive T-cell therapy and intratumoral interferon gamma-inducible protein-10 transgene expression in treatment of established tumors. Cell Immunol 2002; 217: 12–22.

  34. 34

    Chen Z, Huang H, Chang T, Carsen S, Saxena A, Marr R et al. Enhanced HER-2/neu-specific antitumor immunity by cotransduction of mouse dendritic cells with two genes encoding HER-2/neu and alpha tumor necrosis factor. Cancer Gene Ther 2002; 9: 778–786.

  35. 35

    Liu Y, Huang H, Chen Z, Zong L, Xiang J . Dendritic cells engineered to express the Flt3 ligand stimulate type I immune response, and induce enhanced cytoxic T and natural killer cell cytotoxicities and antitumor immunity. J Gene Med 2003; 5: 668–680.

  36. 36

    Xia D, Hao S, Xiang J . CD8+ cytotoxic T-APC stimulate central memory CD8+ T cell responses via acquired peptide-MHC class I complexes and CD80 costimulation and IL-2 secretion. J Immunol 2006; 177: 2976–2984.

  37. 37

    Speiser DE, Romero P . Toward improved immunocompetence of adoptively transferred CD8+ T cells. J Clin Invest 2005; 115: 1467–1469.

  38. 38

    Zammit DJ, Cauley LS, Pham QM, Lefrancois L . Dendritic cells maximize the memory CD8 T cell response to infection. Immunity 2005; 22: 561–570.

  39. 39

    Dyer MR, Herrling PL . Progress and potential for gene-based medicines. Mol Ther 2000; 1: 213–224.

  40. 40

    Marshall E . Gene therapy. Second child in French trial is found to have leukemia. Science 2003; 299: 320.

  41. 41

    Bergelson JM, Cunningham JA, Droguett G, Kurt-Jones EA, Krithivas A, Hong JS et al. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 1997; 275: 1320–1323.

  42. 42

    Wickham TJ, Mathias P, Cheresh DA, Nemerow GR . Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 1993; 73: 309–319.

  43. 43

    Nemerow GR . Cell receptors involved in adenovirus entry. Virology 2000; 274: 1–4.

  44. 44

    Huang S, Endo RI, Nemerow GR . Upregulation of integrins alpha v beta 3 and alpha v beta 5 on human monocytes and T lymphocytes facilitates adenovirus-mediated gene delivery. J Virol 1995; 69: 2257–2263.

  45. 45

    Huang S, Kamata T, Takada Y, Ruggeri ZM, Nemerow GR . Adenovirus interaction with distinct integrins mediates separate events in cell entry and gene delivery to hematopoietic cells. J Virol 1996; 70: 4502–4508.

  46. 46

    Okada N, Saito T, Masunaga Y, Tsukada Y, Nakagama S, Mizuguchi H et al. Efficient antigen gene transduction using Arg-Gly-Asp fiber-mutant adenovirus vectors can potentiate antitumor vaccine efficacy and maturation of murine dendritic cells. Cancer Res 2001; 61: 7913–7919.

  47. 47

    Okada N, Masunaga Y, Okada Y, Iiyama S, Mori N, Tsuda T et al. Gene transduction efficiency and maturation status in mouse bone marrow-derived dendritic cells infected with conventional or RGD fiber-mutant adenovirus vectors. Cancer Gene Ther 2003; 10: 421–431.

  48. 48

    Kagi D, Vignaux F, Ledermann B, Burki K, Depraetere V, Nagata S et al. Fas and perforin pathways as major mechanisms of T cell-mediated cytotoxicity. Science 1994; 265: 528–530.

  49. 49

    Huang H, Li F, Gordon JR, Xiang J . Synergistic enhancement of antitumor immunity with adoptively transferred tumor-specific CD4+ and CD8+ T cells and intratumoral lymphotactin transgene expression. Cancer Res 2002; 62: 2043–2051.

  50. 50

    Huang H, Bi XG, Yuan JY, Xu SL, Guo XL, Xiang J . Combined CD4+ Th1 effect and lymphotactin transgene expression enhance CD8+ Tc1 tumor localization and therapy. Gene Ther 2005; 12: 999–1010.

  51. 51

    Winter H, Hu HM, Urba WJ, Fox BA . Tumor regression after adoptive transfer of effector T cells is independent of perforin or Fas ligand (APO-1L/CD95L). J Immunol 1999; 163: 4462–4472.

  52. 52

    Poehlein CH, Hu HM, Yamada J, Assmann I, Alvord WG, Urba WJ et al. TNF plays an essential role in tumor regression after adoptive transfer of perforin/IFN-gamma double knockout effector T cells. J Immunol 2003; 170: 2004–2013.

  53. 53

    Hollenbaugh JA, Dutton RW . IFN-gamma regulates donor CD8 T cell expansion, migration, and leads to apoptosis of cells of a solid tumor. J Immunol 2006; 177: 3004–3011.

  54. 54

    Blohm U, Potthoff D, van der Kogel AJ, Pircher H . Solid tumors ‘melt’ from the inside after successful CD8 T cell attack. Eur J Immunol 2006; 36: 468–477.

  55. 55

    Prevost-Blondel A, Roth E, Rosenthal FM, Pircher H . Crucial role of TNF-alpha in CD8 T cell-mediated elimination of 3LL-A9 Lewis lung carcinoma cells in vivo. J Immunol 2000; 164: 3645–3651.

  56. 56

    Sarraf C . Tumor necrosis factor and cell death in tumors. Int J Oncol 1994; 5: 1333–1339.

  57. 57

    Tracey KJ, Cerami A . Tumor necrosis factor: a pleiotropic cytokine and therapeutic target. Annu Rev Med 1994; 45: 491–503.

  58. 58

    Ruegg C, Yilmaz A, Bieler G, Bamat J, Chaubert P, Lejeune FJ et al. Evidence for the involvement of endothelial cell integrin alpha V beta 3 in the disruption of the tumor vasculature induced by TNF and IFN-gamma. Nat Med 1998; 4: 408–414.

  59. 59

    Denekamp J . Review article: angiogenesis, neovascular proliferation and vascular pathophysiology as targets for cancer therapy. Br J Radiol 1993; 66: 181–196.

  60. 60

    Valencia X, Stephens G, Goldbach-Mansky R, Wilson M, Shevach EM, Lipsky PE . TNF downmodulates the function of human CD4+CD25hi T-regulatory cells. Blood 2006; 108: 253–261.

  61. 61

    Walter EA, Greenberg PD, Gilbert MJ, Finch RJ, Watanabe KS, Thomas ED et al. Reconstitution of cellular immunity against cytomegalovirus in recipients of allogeneic bone marrow by transfer of T-cell clones from the donor. N Engl J Med 1995; 333: 1038–1044.

  62. 62

    Dudley ME, Wunderlich J, Nishimura MI, Yu D, Yang JC, Topalian SL et al. Adoptive transfer of cloned melanoma-reactive T lymphocytes for the treatment of patients with metastatic melanoma. J Immunother 2001; 24: 363–373.

  63. 63

    Sprent J, Tough DF . T cell death and memory. Science 2001; 293: 245–248.

  64. 64

    Zhang HG, Liu C, Su K, Yu S, Zhang L, Zhang S et al. A membrane form of TNF-alpha presented by exosomes delays T cell activation-induced cell death. J Immunol 2006; 176: 7385–7393.

  65. 65

    Dobrzanski MJ, Reome JB, Hollenbaugh JA, Dutton RW . Tc1 and Tc2 effector cell therapy elicit long-term tumor immunity by contrasting mechanisms that result in complementary endogenous type 1 antitumor responses. J Immunol 2004; 172: 1380–1390.

Download references

Acknowledgements

This study was supported by research grants (MOP 67230/81228) of the Canadian Institutes of Health Research to JX. We thank Mark Boyd for his excellent technical support of flow cytometric analysis.

Author information

Correspondence to J Xiang.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ye, Z., Shi, M., Chan, T. et al. Engineered CD8+ cytotoxic T cells with fiber-modified adenovirus-mediated TNF-α gene transfection counteract immunosuppressive interleukin-10-secreting lung metastasis and solid tumors. Cancer Gene Ther 14, 661–675 (2007). https://doi.org/10.1038/sj.cgt.7701039

Download citation

Keywords

  • TNF-α
  • IL-10
  • engineering
  • cytotoxicity
  • survival
  • T-cell therapy

Further reading