Previous studies have shown that the action of bacterial or synthetic oligodeoxynucleotide (oligo-DNA) on mouse NK cells to produce IFN-γ is mediated mostly by monocytes/macrophages activated by olig-DNA. However, its action on human IFN-γ-producing cells has not been well investigated. In the present study, we examined the effect of oligo-DNAs on highly purified human NK and T cells. Bacillus Calmette-Guérin-derived or synthetic oligo-DNAs induced NK cells to produce IFN-γ with an increased CD69 expression, and the autocrine IFN-γ enhanced their cytotoxicity. The response of NK cells to oligo-DNAs was enhanced when the cells were activated with IL-2, IL-12, or anti-CD16 Ab. T cells did not produce IFN-γ in response to oligo-DNAs but did respond independently of IL-2 when they were stimulated with anti-CD3 Ab. In the action of oligo-DNAs, the palindrome sequence containing unmethylated 5′-CpG-3′ motif(s) appeared to play an important role in the IFN-γ-producing ability of NK cells. The changes of base composition inside or outside the palindrome sequence altered its activity: The homooligo-G-flanked GACGATCGTC was the most potent IFN-γ inducer for NK cells. The CG palindrome was also important for activated NK and T cells in their IFN-γ production, although certain nonpalindromes acted on them. Among the sequences tested, cell activation- or cell lineage-specific sequences were likely; i.e., palindrome ACCGGT and nonpalindrome AACGAT were favored by activated NK cells but not by unactivated NK cells or activated T cells. These results indicate that oligo-DNAs containing CG palindrome act directly on human NK cells and activated T cells to induce IFN-γ production.

In the 1970s, the successful treatment of cancer with Mycobacterium bovis bacillus Calmette-Guérin (BCG)3 in experimental animals and humans (1, 2, 3, 4) prompted a number of investigators to isolate the components that exhibit the antitumor activity from BCG or from other bacteria (3, 4). Tokunaga et al. (5, 6, 7, 8, 9, 10, 11) showed that the DNA-rich fraction of BCG, MY-1, exhibits an antitumor effect by activating the host innate immune response. We examined the biological activities of this fraction and found that the single-stranded oligo-DNAs with certain sequences of hexamer palindromes containing 5′-CpG-3′ (CG) motif(s) were active for both mouse spleen cells and human PBMC (12, 13, 14, 15, 16, 17). These sequences are widely observed in DNAs from other types of bacteria, viruses, and an invertebrate animal as well and are rarely present in vertebrate DNAs (10). Therefore, the palindrome sequences with CG motif(s) are foreign DNAs for mammalian immunocompetent cells, and this may be one of the reasons why BCG-DNA exhibits immunogenicity in mice and humans.

The immunogenicity of oligo-DNAs have also been confirmed by other investigators with findings that DNA extracted from various strains of bacteria (other than BCG) and their synthetic counterparts or plasmid DNA can induce mouse and human immunocompetent cells to produce IFN-γ (18, 19, 20, 21, 22, 23, 24), IL-6 (18, 20, 21, 25, 26), IL-12 (18, 19, 21, 22, 23, 28), IL-1β (29), TNF-α (20, 26, 29, 30), macrophage inflammatory protein-2 (26), type 1 IFN (23, 28), and IL-18 (23). These DNAs also enhance NK activity (27) and stimulate B cells for their growth and immunoglobulin production (18, 25, 31, 32, 33, 34, 35). This cumulative evidence supports the current concept that bacterial DNA promotes both cellular and humoral responses in protective and/or defensive immunity in mice and humans (21, 23, 24, 28, 36, 37, 38, 39, 40, 41, 42).

It has been reported that bacterial DNA promotes NK cell function both directly (18) and indirectly in mice. In the indirect mode, bacterial DNA-stimulated monocytes/macrophages (Mos/Mφs) produce IL-12, TNF-α, and type 1 IFN, and these cytokines induce IFN-γ production by NK cells and their enhanced cytotoxicity (12, 14, 18, 19, 22, 27). Bacterial DNA appears also to induce T cell activation in Ag-mediated responses in vivo (28, 36, 39, 42) and in vitro in mice (41). However, to our knowledge, there have been no corresponding studies regarding human NK and T cells, although Roman et al. (23) showed that human T cells do not respond to oligo-DNA at the resting state. If oligo-DNAs directly target human NK and T cells, immunotherapy with the oligo-DNAs would be more efficient because the direct activation of NK cells leads to an enhancement of the MHC-nonrestricted cytotoxicity (43), and IFN-γ produced by activated NK and/or T cells induces the generation of Th1 cells (44). Therefore, we examined whether the oligo-DNAs act directly on human NK or activated T cells.

Recent studies on the DNA structures that determine the immunogenicity have revealed that there are some differences in the immunogenic sequences of bacterial DNA between those identified by us and those of other investigators; the active sequences determined by us are hexamer palindromes containing the CG motif(s) (12, 13, 14, 15, 16, 17), but theirs are the sequences containing CG in a particular sequence context, with less importance of palindrome sequence (18, 25, 27, 33, 34). In the present study, we also tested whether the CG palindromes and other CG-oligo-DNAs are truly immunogenic in human NK and T cells.

RPMI 1640 (Nissui, Tokyo, Japan) supplemented with 10% heat-inactivated FCS (Equitech-Bio, Ingram, TX; endotoxin, <0.05 ng/ml), 100 U/ml penicillin G potassium (Banyu Pharmaceutical, Tokyo, Japan), and 100 μg/ml streptomycin sulfate (Meiji Seika, Tokyo, Japan) was used as a complete medium for cell culture. Human rIL-2, rTNF-α, and IFN-α were generously provided by Shionogi Pharmaceutical (Osaka, Japan), Dainippon Pharmaceutical (Osaka, Japan), and Hayashibara Biochemical Laboratories (Okayama, Japan), respectively. Human rIL-12 was purchased from R&D Systems (Minneapolis, MN), and human rIL-15 and rIL-18 were commercially obtained from BioSource International (Camarillo, CA). mAbs against human IFN-γ (IgG2a), TNF-α (IgG1), IL-12 (IgG1, clone C8.6), and IL-15 (IgG1) were purchased from Genzyme (Boston, MA). Anti-IL-18 (IgG2a) and anti-IFN-α mAbs were purchased from R&D Systems and Pestka Biomedical Laboratories (New Brunswick, NJ), respectively. Polyclonal rabbit anti-IL-2 Ab was commercially obtained from Collaborative Research (Bedford, MA). Based on our preliminary experiments, 1 μg of the anti-IFN-γ mAb neutralizes 4 ng of human rIFN-γ, 100 ng of anti-TNF-α mAb neutralizes 200 pg of human rTNF-α, and 1 μg/ml anti-IL-12 mAb neutralizes IFN-γ secretion by human NK cells induced by 625 pg/ml human rIL-12. Anti-IL-18 (2 μg/ml), 2 μg/ml anti-IL-15, 5 μg/ml anti-IFN-α, and 1 μg/ml anti-IL-2 neutralized 50 ng/ml rIL-18, 5 ng/ml rIL-15, 1000 U/ml rIFN-α, and 100 U/ml IL-2, respectively. Purified mouse myeloma IgG1 and IgG2a proteins and rabbit serum purchased from ICN Pharmaceuticals (Costa Mesa, CA) were used as an isotype-matched control Ig for the mAbs and as a control serum for the IL-2 Ab, respectively, and were shown not to alter the IFN-γ production or cytotoxicity of NK cells in our experiments. The following reagents were commercially obtained: polymyxin B (Sigma Chemical, St. Louis, MO); Dynabeads M-450 CD3, CD14, CD19, and anti-mouse IgG (Dynal, Oslo, Norway); mouse anti-human mAbs directed CD3, CD14, CD16, CD19, CD25, CD30, CD38, CD56, CD69, CD71, CD94, CD97, CD134, CDw137, HLA-DR, and HLA-ABC (PharMingen Becton Dickinson, San Diego, CA, and/or DAKO, Glostrup, Denmark); FITC- or PE-labeled anti-CD3, anti-CD14, anti-CD16, anti-CD19, and anti-CD56 (PharMingen); and goat anti-mouse Ig (DAKO, Becton Dickinson Immunocytometry Systems, San Jose, CA, or Caltag, San Francisco, CA).

A single-stranded oligo-DNA-rich fraction designated MY-1 was extracted from BCG as described previously (5). MY-1 does not contain any detectable cell wall components. We purchased the oligo-DNAs from Nisshinbo (Tokyo, Japan), who prepared them using an Expedite Model 8909 Nucleic Acid Synthesis System (PerSeptive Biosystems, Framingham, MA). The endotoxin level in the synthetic oligo-DNAs was less than 50 pg/100 μM when measured by the Limulus test (Seikagaku, Tokyo, Japan), which specifically detects endotoxin. The sequences of the oligo-DNAs are presented in Figs. 4 and 9 and in Table II.

FIGURE 4.

Responsiveness of NK cells to synthetic oligoDNAs. NK cells (2 × 106/ml) were cultured in triplicate for 40 h with medium, MY-1 (50 μg/ml), or various kinds of 30-mer synthetic oligo-DNAs (5 μM, almost equivalent to 50 μg/ml for each oligoDNA). Data are means ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the control culture with medium alone. Similar results were obtained in two other separate experiments.

FIGURE 4.

Responsiveness of NK cells to synthetic oligoDNAs. NK cells (2 × 106/ml) were cultured in triplicate for 40 h with medium, MY-1 (50 μg/ml), or various kinds of 30-mer synthetic oligo-DNAs (5 μM, almost equivalent to 50 μg/ml for each oligoDNA). Data are means ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the control culture with medium alone. Similar results were obtained in two other separate experiments.

Close modal
FIGURE 9.

Effect of different sequences on IFN-γ production by NK cells cultured with IL-2. NK cells (2 × 106/ml) were cultured in triplicate for 22 h with culture medium, 50 μg/ml MY-1, or 5 μM concentrations of various kinds of oligo-DNAs containing palindromes or nonpalindromes in the presence of 10 U/ml rIL-2. The capital letters in the name of the oligo-DNAs listed in the figure represent the sequences that were introduced to the position of Ns of 5′-accgatNNNNNNgccggtgacggcaccacg-3′. Data are the means ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the respective controls with medium alone. The results are representative of six similar experiments, and cells obtained from different donors were used in each experiment.

FIGURE 9.

Effect of different sequences on IFN-γ production by NK cells cultured with IL-2. NK cells (2 × 106/ml) were cultured in triplicate for 22 h with culture medium, 50 μg/ml MY-1, or 5 μM concentrations of various kinds of oligo-DNAs containing palindromes or nonpalindromes in the presence of 10 U/ml rIL-2. The capital letters in the name of the oligo-DNAs listed in the figure represent the sequences that were introduced to the position of Ns of 5′-accgatNNNNNNgccggtgacggcaccacg-3′. Data are the means ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the respective controls with medium alone. The results are representative of six similar experiments, and cells obtained from different donors were used in each experiment.

Close modal
Table II.

Effect of oligo-DNAs containing CG, GC, or methylated CG on the IFN-γ production by activated human T cellsa

Oligo-DNAIFN-γ Amount in the Culture Supernatant (pg/ml)
Sequences (5′-3′)Name
Medium  56.2 ± 7.8 
tccatgacgttcctgatgct 1668 45.0 ± 2.6 
tctcccagcgtgcgccat 1758 58.4 ± 1.7 
ttgcttccatcttcctcgtc 2105 58.4 ± 5.0 
gagaacgctcgaccttcgat 1643 59.5 ± 8.5 
   
accgataacgctgccggtgacggcaccacg AACGCT-30 127.6 ± 15.8b 
accgataagcctgccggtgacggcaccacg AAGCCT-30 62.6 ± 6.0 
ggggggggggggaacgctgggggggggggg g12AACGCT 52.8 ± 3.4 
accgataacggtgccggtgacggcaccacg AACGGT-30 67.2 ± 10.8 
accgataacgatgccggtgacggcaccacg AACGAT-30 52.0 ± 3.7 
accgataacgtcgccggtgacggcaccacg AACGTC-30 131.9 ± 19.4b 
accgataagctcgccggtgacggcaccacg AAGCTC-30 65.9 ± 5.0 
ggggggggggggaacgtcgggggggggggg g12AACGTC 66.2 ± 7.7 
accgataacgtagccggtgacggcaccacg AACGTA-30 119.7 ± 15.4b 
accgataagctagccggtgacggcaccacg AAGCTA-30 57.3 ± 8.3 
ggggggggggggaacgtagggggggggggg g12AACGTA 58.3 ± 7.8 
accgataacgtggccggtgacggcaccacg AACGTG-30 66.5 ± 11.3 
accgataacgttgccggtgacggcaccacg AACGTT-30 118.0 ± 10.0b 
accgataagcttgccggtgacggcaccacg AAGCTT-30 48.0 ± 6.9 
ggggggggggggaacgttgggggggggggg g12AAC 284.7 ± 14.0b 
ggggggggggggaacgttgtgggggggggg Methylated g12AAC 53.0 ± 6.0 
   
BCG-DNA fraction MY-1 114.0 ± 9.6b 
Oligo-DNAIFN-γ Amount in the Culture Supernatant (pg/ml)
Sequences (5′-3′)Name
Medium  56.2 ± 7.8 
tccatgacgttcctgatgct 1668 45.0 ± 2.6 
tctcccagcgtgcgccat 1758 58.4 ± 1.7 
ttgcttccatcttcctcgtc 2105 58.4 ± 5.0 
gagaacgctcgaccttcgat 1643 59.5 ± 8.5 
   
accgataacgctgccggtgacggcaccacg AACGCT-30 127.6 ± 15.8b 
accgataagcctgccggtgacggcaccacg AAGCCT-30 62.6 ± 6.0 
ggggggggggggaacgctgggggggggggg g12AACGCT 52.8 ± 3.4 
accgataacggtgccggtgacggcaccacg AACGGT-30 67.2 ± 10.8 
accgataacgatgccggtgacggcaccacg AACGAT-30 52.0 ± 3.7 
accgataacgtcgccggtgacggcaccacg AACGTC-30 131.9 ± 19.4b 
accgataagctcgccggtgacggcaccacg AAGCTC-30 65.9 ± 5.0 
ggggggggggggaacgtcgggggggggggg g12AACGTC 66.2 ± 7.7 
accgataacgtagccggtgacggcaccacg AACGTA-30 119.7 ± 15.4b 
accgataagctagccggtgacggcaccacg AAGCTA-30 57.3 ± 8.3 
ggggggggggggaacgtagggggggggggg g12AACGTA 58.3 ± 7.8 
accgataacgtggccggtgacggcaccacg AACGTG-30 66.5 ± 11.3 
accgataacgttgccggtgacggcaccacg AACGTT-30 118.0 ± 10.0b 
accgataagcttgccggtgacggcaccacg AAGCTT-30 48.0 ± 6.9 
ggggggggggggaacgttgggggggggggg g12AAC 284.7 ± 14.0b 
ggggggggggggaacgttgtgggggggggg Methylated g12AAC 53.0 ± 6.0 
   
BCG-DNA fraction MY-1 114.0 ± 9.6b 
a

T cells were cultured in triplicate at 2 × 106/ml for 21 h with medium, 50 μg/ml MY-1, 5 μM oligo-DNAs containing certain sequences (underlined) with CG, GC, or methylated CG (bold c), in the presence of 2 × 106 particles/ml M-450 CD3. The results are representative of four experiments with cells obtained from different donors with similar results. The IFN-γ concentrations in culture supernatants are expressed as means ± SD.

b

p < 0.01 compared with the control value with medium alone.

Isolation of PBMC.

PBMC were isolated from the venous blood of healthy volunteers by 60% osmolarity-adjusted Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden) density gradient centrifugation. Platelets were removed from the PBMC suspension or the density-fractionated cells by centrifugation on Nyco Prep 1.063 (Daiichi Pure Chemicals, Tokyo, Japan), throughout the experiments, because they inhibit MY-1-induced IFN-γ production by NK or T cells (our unpublished observation).

Preparation of nonadherent cells (NAC).

PBMC suspended in complete medium were incubated in plastic dishes for 2 h at 37°C in 5% CO2 in a humidified atmosphere. After the plastic adherence was repeated at least twice, the cells that floated up by gentle agitation of the dishes were recovered, loaded on a nylon wool fiber (Polysciences, Warrington, PA) column, and incubated for 2 h at 37°C in 5% CO2 in humidified air. The cells that passed through the nylon wool column were collected as NAC depleted of Mos/Mφs and B cells.

Purification of NK and T cells.

NAC were layered over a discontinuous density gradient composed of 42.9% (F1), 46.2% (F2), 50.0% (F3), 54.5% (F4), and 60% (F5) or F1, F2, and 52.6% (F3/F4) of osmolarity-adjusted Percoll and centrifuged for 30 min at 1500 or 1250 rpm, respectively, at room temperature. The cell layers on F3–F4 or F3/F4 were collected as the large granular lymphocyte (LGL)-rich (morphologically 60–90% of LGL), and those on F5 were collected as the T cell fraction (>99% CD3+ determined by flow cytometry (Fig. 10,A)). When T cell purity was not sufficient as determined by flow cytometry, B cells and Mos/Mφs were removed with the use of M-450 CD19 and M-450 CD14, respectively, or by cell sorting (Epics Elite, Beckman Coulter, Fullerton, CA) with FITC-conjugated CD14, CD16, and CD19 mAbs (PharMingen). NK cells were purified from the LGL-rich fraction by negative or positive selection. In the negative selection, the LGL-rich population was depleted of Mos/Mφs and T/B cells by the serial use of M-450 CD14, M-450 CD3, and M-450 CD19 magnetic beads or by an indirect method using anti-CD14, anti-CD3, and anti-CD19 mAbs as the first Abs, then with M-450 goat anti-mouse IgG or M-450 sheep anti-mouse IgG magnetic beads as the second Abs. The immunomagnetic depletion was repeated at least twice in each method. In the positive selection, indirect immunomagnetic separation was performed with a combination of anti-CD56 mAb and M-450 IgG after the repeated depletion of Mos/Mφs, using M-450 CD14 to avoid the trapping of Mos/Mφs in the NK cell population, which may be caused by phagocytosis of the immunobeads or nonspecific cell aggregation. Cells obtained by these methods contained >97% CD56-positive cells as determined by flow cytometry (Fig. 1 A) and <1% Mos/Mφs as evaluated with nonspecific esterase (Muto Pure Chemical, Tokyo, Japan) or flow cytometric analysis of CD14 expression. In some experiments, NK cells were isolated with FITC-conjugated CD56 mAb (PharMingen) by cell sorting (Epics Elite). NK cells purified by CD56-positive selection in the immunomagnetic separation method were used after 6 h of incubation at 37°C in 5% CO2 to detach the beads but were used without removing the beads when cells were stimulated with anti-CD16 mAb. Both procedures for the NK cell purification did not alter the responsiveness of the NK cells to oligo-DNAs.

FIGURE 10.

A, Purity of T cell preparation. T cells isolated and used in the present study contained >99% of CD3+ cells and <1% of CD14+ and/or CD19+ cells. The cells shown in A expressed 99.8% of CD3+ cells. B, Effect of MY-1 and g10GACGA on IFN-γ production of unstimulated and M-450 CD3-stimulated T cells. T cells were cultured in triplicate at 2 × 106/ml for 24 h with culture medium, 50 μg/ml MY-1, or 10 μg/ml g10GACGA in the presence or absence of 2 × 106 particles/ml M-450 CD3. The culture with 100 U/ml IL-2 was performed as the positive control for IFN-γ production. Data are means ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the control with medium alone in M-450 CD3-stimulated T cell culture. C, Effect of anti(α)-IL-2 Ab on the MY-1-induced IFN-γ production. M-450 CD3-activated T cells (a mixture of 2 × 106/ml T cells and 2 × 106/ml M-450 CD3) were cultured for 24 h with medium, 50 μg/ml MY-1, or 100 U/ml IL-2 in the presence or absence of anti-IL-2 Ab. Data are means ± SD. ∗∗, p < 0.01 compared with the control value with IL-2 alone. NS, not significantly different compared with the corresponding controls with medium or MY-1 alone.

FIGURE 10.

A, Purity of T cell preparation. T cells isolated and used in the present study contained >99% of CD3+ cells and <1% of CD14+ and/or CD19+ cells. The cells shown in A expressed 99.8% of CD3+ cells. B, Effect of MY-1 and g10GACGA on IFN-γ production of unstimulated and M-450 CD3-stimulated T cells. T cells were cultured in triplicate at 2 × 106/ml for 24 h with culture medium, 50 μg/ml MY-1, or 10 μg/ml g10GACGA in the presence or absence of 2 × 106 particles/ml M-450 CD3. The culture with 100 U/ml IL-2 was performed as the positive control for IFN-γ production. Data are means ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the control with medium alone in M-450 CD3-stimulated T cell culture. C, Effect of anti(α)-IL-2 Ab on the MY-1-induced IFN-γ production. M-450 CD3-activated T cells (a mixture of 2 × 106/ml T cells and 2 × 106/ml M-450 CD3) were cultured for 24 h with medium, 50 μg/ml MY-1, or 100 U/ml IL-2 in the presence or absence of anti-IL-2 Ab. Data are means ± SD. ∗∗, p < 0.01 compared with the control value with IL-2 alone. NS, not significantly different compared with the corresponding controls with medium or MY-1 alone.

Close modal
FIGURE 1.

A, Purity of NK cells. NK cells isolated and used in the present study contained >97% CD56+ cells and <1% of CD14+ and/or CD19+ cells. The cells shown in A contained 98.9% of CD56+ cells and 0.4% of CD14/CD19+ cells. B, Dose-response effects of MY-1 and synthetic oligo-DNA, g10GACGA, on the IFN-γ production by NK cells. NK cells were cultured in triplicate at 3 × 106/ml for 24 h with various concentrations of MY-1 or g10GACGA. IFN-γ concentrations in the culture supernatants are presented as the mean ± SD. ∗∗, p < 0.01 compared with the control culture with medium alone. Similar results were obtained in three other separate experiments.

FIGURE 1.

A, Purity of NK cells. NK cells isolated and used in the present study contained >97% CD56+ cells and <1% of CD14+ and/or CD19+ cells. The cells shown in A contained 98.9% of CD56+ cells and 0.4% of CD14/CD19+ cells. B, Dose-response effects of MY-1 and synthetic oligo-DNA, g10GACGA, on the IFN-γ production by NK cells. NK cells were cultured in triplicate at 3 × 106/ml for 24 h with various concentrations of MY-1 or g10GACGA. IFN-γ concentrations in the culture supernatants are presented as the mean ± SD. ∗∗, p < 0.01 compared with the control culture with medium alone. Similar results were obtained in three other separate experiments.

Close modal

NAC, NK, or T cells were placed in 96-well plates (round-bottom plates for NAC and NK cells and flat-bottom plates for T cells (Corning Glass Works, Corning, NY) and cultured in complete medium at 37°C in a humidified atmosphere with 5% CO2 under the conditions described in Results. IFN-γ secreted in the culture supernatants was measured by an ELISA kit (Cytoscreen Immunoassay Kit) (Biosource International). The lower limit of detection for human IFN-γ was 4 pg/ml.

NK cells were cultured with K562 cells at the indicated E:T ratios in triplicate, for 4 h at 37°C in 5% CO2. The activity of lactate dehydrogenase released from damaged cells into the culture medium were measured by a cytotoxicity detection kit (lactate dehydrogenase) (Boehringer Mannheim, Mannheim, Germany), and the cytotoxicity is expressed here as a percentage of target cell lysis.

Flow cytometry analysis was performed on a EPICS XL (Beckman Coulter). Cells were incubated with mAbs, followed by washing and labeling with FITC or PE-conjugated goat anti-mouse Ig. Data were obtained in a logarithmic scale.

Data were analyzed with the Wilcoxon signed rank test, ANOVA, or Student’s t test. Differences in the results were considered significant at p < 5%.

To determine the cell type(s) that is responsive to BCG-DNA, MY-1, to produce IFN-γ, we separated LGL and T cells from NAC. When these two cell fractions were cultured for 24 h at a concentration of 2 × 106 cells/ml, only the LGL fraction produced IFN-γ in the presence of MY-1 (Table I, experiments 1 and 2). NAC which contain 20–30% of NK and 70–80% of T cells, produced IFN-γ in response to MY-1 when the cell density was increased to 4 × 106 cells/ml (experiment 2), whereas the T cell fraction did not produce IFN-γ even when cultured at 1 × 107 cells/ml (experiment 3) or for longer periods (data not shown). Polymyxin B, a LPS inhibitor, did not affect the MY-1-induced IFN-γ production, and DNase treatment of MY-1 abolished the IFN-γ-inducible activity (data not shown).

Table I.

Effect of BCG-derived DNA, MY-1, on IFN-γ production by NK and T cellsa

Expt.Fractionated CellsCell DensitiesIFN-γ Produced (pg/ml)
Without MY-1With MY-1
LGL 2 × 106/ml 10.5± 3.5 47.0± 6.4* 
 T cells 2 × 106/ml <4 <4 
 NAC 2 × 106/ml <4 <4 
LGL 2 × 106/ml <4 46.3± 3.1* 
 T cells 2 × 106/ml <4 <4 
  4 × 106/ml <4 <4 
 NAC 2 × 106/ml <4 <4 
  4 × 106/ml <4 17.1± 2.0* 
CD56 cells in LGL 2 × 106/ml <4 <4 
 CD56+ cells in LGL 2 × 106/ml <4 58.7± 5.6* 
 T cells 2 × 106/ml <4 <4 
  1 × 107/ml <4 <4 
Expt.Fractionated CellsCell DensitiesIFN-γ Produced (pg/ml)
Without MY-1With MY-1
LGL 2 × 106/ml 10.5± 3.5 47.0± 6.4* 
 T cells 2 × 106/ml <4 <4 
 NAC 2 × 106/ml <4 <4 
LGL 2 × 106/ml <4 46.3± 3.1* 
 T cells 2 × 106/ml <4 <4 
  4 × 106/ml <4 <4 
 NAC 2 × 106/ml <4 <4 
  4 × 106/ml <4 17.1± 2.0* 
CD56 cells in LGL 2 × 106/ml <4 <4 
 CD56+ cells in LGL 2 × 106/ml <4 58.7± 5.6* 
 T cells 2 × 106/ml <4 <4 
  1 × 107/ml <4 <4 
a

Data are representative of eight experiments performed with similar results. In each experiment, cells were fractionated from PBMC obtained from different donors and cultured at indicated cell densities for 24 h with or without 50 μg/ml MY-1. Data are the means ± SD (n = 3). ∗, p < 0.01 compared with the respective controls without MY-1.

We then purified CD56+ cells from the LGL fraction (Fig. 1,A) and tested their responsiveness to MY-1. As shown in Table I (experiment 3), NK cells produced IFN-γ in response to MY-1. The doses of MY-1 necessary to induce the maximum amount of IFN-γ were between 12.5 and 50 μg/ml in the culture of NK cells (Fig. 1,B). IFN-γ production in the culture with MY-1 was first observed at 18 h and increased thereafter (Fig. 2). The amounts of IFN-γ produced without MY-1 at 24-h culture were mostly below 4 pg/ml and did not exceed 13 pg/ml in any NK cell sources examined. These results show that NK cells are responsive to MY-1 in terms of IFN-γ production.

FIGURE 2.

Time course studies of MY-1-induced IFN-γ production by NK cells. NK cells were cultured in triplicate at 3 × 106/ml for 12, 18, 24, and 48 h in the presence or absence of 50 μg/ml MY-1. Data are means ± SD. This analysis was repeated three times with very similar results. ∗∗, p < 0.01 compared to the respective controls without MY-1.

FIGURE 2.

Time course studies of MY-1-induced IFN-γ production by NK cells. NK cells were cultured in triplicate at 3 × 106/ml for 12, 18, 24, and 48 h in the presence or absence of 50 μg/ml MY-1. Data are means ± SD. This analysis was repeated three times with very similar results. ∗∗, p < 0.01 compared to the respective controls without MY-1.

Close modal

To prove that the MY-1-induced IFN-γ production is caused by a direct action on NK cells, we added neutralizing concentrations of mAbs against IL-12 or TNF-α to the culture of NK cells in the presence or absence of oligo-DNA. In this experiment, instead of MY-1, g10GACGA (synthetic oligo-DNA) was used because of its potent ability to induce IFN-γ (see below). As shown in Fig. 3, neither anti-IL-12 nor anti-TNF-α mAb influenced the IFN-γ production by NK cells cultured with or without g10GACGA. The combined addition of mAbs against IL-12 and TNF-α also did not inhibit the production of IFN-γ. No inhibitory effect of these Abs on the IFN-γ production was observed when 10 μg/ml g10GACGA was applied for NK cell stimulation. In addition, 1–10 μg/ml mAbs to IL-18, IL-15, or IFN-α did not alter the level of IFN-γ production induced by the oligo-DNA (data not shown).

FIGURE 3.

Effect of mAbs against IL-12 or TNF-α on g10GACGA-induced IFN-γ production. NK cells (2 × 106/ml) were cultured in triplicate for 24 h with medium, 50 μg/ml g10GACGA, 500 pg/ml rIL-12, or 200 pg/ml rTNF-α in the presence of the mAbs against IL-12 or TNF-α, mouse IgG1, or medium alone. Anti-IL-12 and anti-TNF-α mAbs and mouse IgG1 were added at concentrations of 1, 0.1, and 1 μg/ml, respectively. Data are means ± SD. NS, not significant compared with the corresponding controls with medium, g10GACGA, rIL-12, or rTNF-α alone. §, p < 0.01 compared with the value with medium alone. ∗∗, p<0.01 compared with the value with rIL-12 alone. The results are representative of three independent experiments performed with cells obtained from different individuals. α, anti.

FIGURE 3.

Effect of mAbs against IL-12 or TNF-α on g10GACGA-induced IFN-γ production. NK cells (2 × 106/ml) were cultured in triplicate for 24 h with medium, 50 μg/ml g10GACGA, 500 pg/ml rIL-12, or 200 pg/ml rTNF-α in the presence of the mAbs against IL-12 or TNF-α, mouse IgG1, or medium alone. Anti-IL-12 and anti-TNF-α mAbs and mouse IgG1 were added at concentrations of 1, 0.1, and 1 μg/ml, respectively. Data are means ± SD. NS, not significant compared with the corresponding controls with medium, g10GACGA, rIL-12, or rTNF-α alone. §, p < 0.01 compared with the value with medium alone. ∗∗, p<0.01 compared with the value with rIL-12 alone. The results are representative of three independent experiments performed with cells obtained from different individuals. α, anti.

Close modal

To determine the effective sequence(s) in MY-1 for human NK cells, 30-mer synthetic oligo-DNAs possessing various palindrome sequences were first tested. These oligo-DNAs are analogues of BCG-4a (12), the sequence of which was chosen from the cDNA encoding 64-kDa heat shock protein AgA of BCG. GACGTC in BCG-4a was replaced with different types of hexamer palindromes. These included nine active palindromes which enhanced NK cytotoxicity more strongly than GACGTC in mice and one inactive form (17). The ability of each 30-mer oligo-DNAs to induce IFN-γ production was compared at a concentration of 5 μM (almost equivalent to 50 μg/ml for each oligo-DNA) with 50 μg/ml MY-1. Among the oligo-DNAs examined in three separate experiments, the sequences containing ATCGAT, TCGCGA, CGTACG, CGGCCG, and GACGTC (named ATCGAT-30, TCGCGA-30, CGTACG-30, CGGCCG-30, and GACGTC-30; BCG-4a, respectively) were shown to induce IFN-γ production by NK cells as potently as did MY-1. The oligo-DNAs containing the CGATCG and AACGTT sequences (CGATCG-30 and AACGTT-30) also exhibited IFN-γ-inducing activity, but their activities were less potent than that of MY-1. The oligo-DNAs with AGCGCT (AGCGCT-30) and GCGCGC (GCGCGC-30) had a slight ability to induce IFN-γ. ACCGGT, which we reported as an exceptionally inactive palindrome sequence (14, 15, 16, 17), showed no IFN-γ-inducing effect on NK cells. The representative data are shown in Fig. 4.

The oligo-G introduced at the position of the extrapalindrome sequence potentiates the ability of the palindrome sequence to enhance NK activity in mice (17). To examine whether the IFN-γ induction abilities of the weakly active palindromes, CGATCG and AACGTT, are augmented by inducing the homooligo-G extrapalindrome, we flanked these palindromes with 12-mer oligo-G at both 5′- and 3′-ends (g12CGA and g12AAC, respectively). As shown in Fig. 4, the ability of CGATCG, but not AACGTT, was augmented 2-fold by oligo-G flanking of the original sequence. An elongated palindrome (GACGATCGTC) that flanked with 10-mer oligo-G (g10GACGA) exhibited the most potent ability to induce IFN-γ production, but ACAACGTTGT (g10ACAAC) did not. The doses of g10GACGA capable of inducing the plateau level of IFN-γ production by NK cells were from 6.3 to 50 μg/ml (10 μg/ml is almost equal to 1 μM) (Fig. 1). The level of IFN-γ production at 24-h culture with 10 μg/ml of g10GACGA was comparable with those induced by 10 U/ml IL-2 or 10 pg/ml IL-12 (see Fig. 8). The amount of IFN-γ detected in the culture with 30-mer oligo-G (g30) appeared slightly higher than that in the control culture, but the increase was not significant.

FIGURE 8.

Effect of g10GACGA on the production of IFN-γ by NK cells in the presence of rIL-2, rIL-12, or anti-CD16 mAb. NK cells were isolated by positive selection and cultured in triplicate at 2 × 106/ml for 20 h with or without 10 μg/ml g10GACGA in the presence of culture medium, 10 U/ml rIL-2, 10 pg/ml rIL-12, or 1 μg/ml anti (α)-CD16 mAb. Data are means ± SD. ∗∗, p < 0.01 compared with the value without g10GACGA in each culture condition. These experiments were independently repeated three times with similar results.

FIGURE 8.

Effect of g10GACGA on the production of IFN-γ by NK cells in the presence of rIL-2, rIL-12, or anti-CD16 mAb. NK cells were isolated by positive selection and cultured in triplicate at 2 × 106/ml for 20 h with or without 10 μg/ml g10GACGA in the presence of culture medium, 10 U/ml rIL-2, 10 pg/ml rIL-12, or 1 μg/ml anti (α)-CD16 mAb. Data are means ± SD. ∗∗, p < 0.01 compared with the value without g10GACGA in each culture condition. These experiments were independently repeated three times with similar results.

Close modal

The sequences of immunostimulatory DNAs that have been extensively studied by other investigators do not have the hexamer palindrome. In mice, gagaacgctcgaccttcgat (1643) is mitogenic to B cells (33), and tccatgacgttcctgatgct (1668) induces not only B cell activation (33) but also the production of inflammatory cytokines by lymphocytes (18, 20). Furthermore, tctccagcgtgcgccat (1758, antisense hBcl-2) activates NK cells (37). In humans, ttgcttccatcttcctcgtc (2105) directly activates B cells (34). We tested whether these sequences are effective for human NK cells to induce IFN-γ production. Unlike the active sequences mentioned above, they did not induce IFN-γ production of human NK cells; these sequences were tested at concentrations ranging from 0.2 to 20 μM, by extending the culture periods to 5 days. The representative data (IFN-γ amount, pg/ml) from three separate experiments with NK cells, which were purified by cell sorting and cultured at 3 × 106/ml for 3 days in the presence of 1643, 1668, 1758, 2105, AACGTT-30, g10GACGA, or medium alone, were 20.0 ± 3.6 (mean ± SD, n = 3), 25.4 ± 4.8, 26.2 ± 5.1, 18.4 ± 1.8, 75.4 ± 15.4, 163.7 ± 25.2, and 18.8 ± 3.6, respectively. Further, replacement of the AACGTT motif of AACGTT-30 with the immunostimulatory core sequence, PuPuCGPyPy, i.e., AACGCT, AACGTC, or AACGCC, and with other sequences containing one thymine at the 3′-side of AACG, such as AACGAT, AACGGT, AACGTA, and AACGTG, did not induce IFN-γ production (data not shown).

We tested whether oligo-DNAs can affect the cytotoxicity of purified NK cells. As shown in Fig. 5, when NK cells were cultured with MY-1 or g10GACGA, the ability to lyse K562 cells was enhanced and the enhancement was prominent in the culture with g10GACGA. The 30-mer homooligo-G, g30, used as a control DNA, did not alter the cytotoxic activity. These results indicate that the enhanced cytotoxicity is directly elicited by purified NK cells cultured with synthetic oligo-DNAs. To examine the role of IFN-γ induced by oligo-DNAs in the augmentation of NK activity, we added a neutralizing mAb against IFN-γ to the NK cell culture in the presence or absence of g10GACGA. The ability of g10GACGA to enhance NK activity was diminished by the addition of anti-IFN-γ mAb (Fig. 6). Similarly, MY-1 enhanced NK cells and the enhancement was inhibited in the presence of the anti-IFN-γ mAb (data not shown).

FIGURE 5.

Effect of synthetic oligo-DNAs on NK activity. NK cells were cultured in triplicate for 24 h with 50 μg/ml MY-1, synthetic oligo-DNAs, g10GACGA and g30, or medium alone. NK activities are expressed as percent lysis (mean ± SD) at the indicated E:T ratios. ∗∗, p < 0.01 compared with the value with medium alone at the respective E:T ratios. The oligo-DNA-enhanced NK activity was similarly reproduced in experiments that were repeated three times using cells from different individuals.

FIGURE 5.

Effect of synthetic oligo-DNAs on NK activity. NK cells were cultured in triplicate for 24 h with 50 μg/ml MY-1, synthetic oligo-DNAs, g10GACGA and g30, or medium alone. NK activities are expressed as percent lysis (mean ± SD) at the indicated E:T ratios. ∗∗, p < 0.01 compared with the value with medium alone at the respective E:T ratios. The oligo-DNA-enhanced NK activity was similarly reproduced in experiments that were repeated three times using cells from different individuals.

Close modal
FIGURE 6.

Effect of anti (α)-IFN-γ mAb on g10GACGA-enhanced NK activity. NK cells were cultured in triplicate for 44 h with medium, 50 μg/ml g10GACGA, or 400 pg/ml rIFN-γ in the presence of 1 μg/ml mouse IgG2a or anti-IFN-γ mAb. NK activities are presented as percent lysis at E:T = 12, 6, and 3 (mean ± SD). ∗∗, p < 0.01 compared with the value with medium alone. §, p < 0.01 compared with the respective control values without anti-IFN-γ. No significant difference was observed between the values with anti-IFN-γ plus g10GACGA or rIFN-γ and the value with anti-IFN-γ plus medium. The results shown are representative of three experiments with similar results.

FIGURE 6.

Effect of anti (α)-IFN-γ mAb on g10GACGA-enhanced NK activity. NK cells were cultured in triplicate for 44 h with medium, 50 μg/ml g10GACGA, or 400 pg/ml rIFN-γ in the presence of 1 μg/ml mouse IgG2a or anti-IFN-γ mAb. NK activities are presented as percent lysis at E:T = 12, 6, and 3 (mean ± SD). ∗∗, p < 0.01 compared with the value with medium alone. §, p < 0.01 compared with the respective control values without anti-IFN-γ. No significant difference was observed between the values with anti-IFN-γ plus g10GACGA or rIFN-γ and the value with anti-IFN-γ plus medium. The results shown are representative of three experiments with similar results.

Close modal

To identify the activation-associated molecule(s) which would be induced by oligo-DNA, we examined the expression of CD25, CD69, and CD94 molecules and HLA-ABC on cultured NK cells. As shown in Fig. 7, NK cells strongly expressed the CD69 molecule when cultured with g10GACGA with 2.44 ± 0.56 (mean ± SD, n = 3) times more in the percent positive cells than in those cultured with medium alone. Expression of other molecules such as CD25 and CD94 was unchanged by the culture with g10GACGA. In these experiments, the fluorescence of NK cells stained with HLA-ABC was always intensified by the culture with oligo-DNA (e.g., medium: 502 ± 17, g10GACGA: 590 ± 19, and IL-2 as a positive control: 705 ± 17 as expressed by the mean intensity ± SD), indicating the autocrine stimulation of NK cells by IFN-γ induced by g10GACGA.

FIGURE 7.

Oligo-DNA enhances the expression of CD69 in NK cells. NK cells were cultured for 36 h with and without 1 μM g10GACGA, washed, and stained for CD69 surface expression.

FIGURE 7.

Oligo-DNA enhances the expression of CD69 in NK cells. NK cells were cultured for 36 h with and without 1 μM g10GACGA, washed, and stained for CD69 surface expression.

Close modal

We then tested the ability of NK cells to produce IFN-γ in response to oligo-DNAs in the presence of IL-2, IL-12, or anti-CD16 mAb to examine the influence of the activation status of NK cells on their responsiveness to oligo-DNAs. g10GACGA was used in this experiment because of its potent activity. As shown in Fig. 8, g10GACGA could induce IFN-γ production by NK cells in the absence of the stimuli. The addition of IL-2, IL-12, or anti-CD16 mAb to this culture significantly enhanced the IFN-γ production. The increase was synergistic in the culture with IL-2, whereas in the culture with IL-12 or anti-CD16 mAb, the increases were additive. Therefore, the activated NK cells appear to be more susceptible to oligo-DNAs in terms of IFN-γ production, especially with IL-2 stimulation.

With IL-2, MY-1 also enhanced IFN-γ production by NK cells (Fig. 9). We then examined the effect of different palindrome sequences, which often occur in MY-1 (17), on the IFN-γ production by NK cells in the presence of IL-2, in a manner similar to that used for unactivated NK cells. The synthetic oligo-DNAs that induced IFN-γ production by unactivated NK cells, i.e., ATCGAT-30, GACGTC-30, TCGCGA-30, CGTACG-30, and CGGCCG-30, and those that showed weak or modest abilities to induce IFN-γ production by the unactivated NK cells, i.e., AGCGCT-30, CGATCG-30, GCGCGC-30, and AACGTT-30 all enhanced the IFN-γ production by NK cells in the presence of IL-2. When the IFN-γ-inducing activity of these palindromes was expressed as a percentage of the control in six separate experiments, the order of potency was as follows: AACGTT (432 ± 95 pg/ml, mean ± SE), ACCGGT (408 ± 48), CGTACG (376 ± 42), AGCGCT (376 ± 72), GCGCGC (320 ± 32), CGATCG (259 ± 37), TCGCGA (256 ± 35), ATCGAT (249 ± 37), CGGCCG (246 ± 22), and GACGTC (238 ± 40). That of MY-1 was 397 ± 34. These values were not statistically different. Unlike those in the culture of unactivated NK cells, however, AACGTT was the most potent and GACGTC the weakest palindrome in the culture of IL-2-activated NK cells. In these results, there was a striking difference in the sequence pattern of the induction of IFN-γ production from those observed in the unactivated NK cells. That is, an oligo-DNA with the ACCGGT palindrome (which was inactive in unactivated NK cells) was able to induce IFN-γ in the presence of IL-2. One of these data is shown in Fig. 9 (experiment 1) as the amount of IFN-γ produced in the cell culture supernatant. A palindrome that contains GC instead of CG, AAGCTT, showed no effect on the IFN-γ production (experiment 2).

It has been reported by Chace et al. (22) that oligo-DNAs act on mouse NK cells in the presence but not in the absence of IL-12. Activated NK cells may be more susceptible for oligo-DNA stimulation, regardless of the particular contexts such as a palindrome with internal CG or the PuPuCGPyPy sequence. Then two bases of AACGTT were replaced at the 3′-side with theoretically possible dinucleotides containing one thymine, to test for the IFN-γ-inducing ability using IL-2-activated NK cells. As shown in Fig. 9 (experiment 3), activated NK cells responded to the sequences containing CG irrespective of particular contexts such as the palindrome or PuPuCGPyPy. Among them, the sequences with TT, CT, AT, TC, and TA at the 3′-side of AACG were more potent.

Resting T cells did not respond to MY-1 (Table I). However, the synergism between oligo-DNAs and IL-2 observed in NK cells prompted us to examine the responsiveness of activated T cells to MY-1 and synthetic oligo-DNAs. We stimulated purified T cells (Fig. 10,A) with Dynabeads M-450 CD3, which is able to activate T cells (manufacturer’s information), and evaluated their IFN-γ production in response to MY-1 and g10GACGA. As shown in Fig. 10,B, T cells produced IFN-γ in the presence of M-450 CD3, and this production was significantly enhanced by the addition of MY-1 or g10GACGA. A 10-μg/ml concentration of g10GACGA was almost equipotent to 100 U/ml IL-2 for the induction of IFN-γ in M-450 CD3-stimulated T cells. The effect of g10GACGA on the IFN-γ production was not influenced by the addition of anti-IL-2 Ab to these cultures (Fig. 10 C), indicating that oligo-DNA-induced IFN-γ production is independent of IL-2 production by activated T cells.

The expression of CD25, CD69, CD94, HLA-ABC, CD30, CD38, CD71, CD94, CD97, CD134, CDw137, and HLA-DR was also tested as to whether specific activation marker(s) are induced by oligo-DNA. Anti-CD3 stimulation of T cells expressed higher levels of these molecules, and further enhancement was not observed when examined at 24 and 48 h of cultures with MY-1 or g10GACGA (data not shown).

To seek out the effective sequences for activated T cells, the sequences involved in the induction of IFN-γ production were examined in a manner similar to that used for NK cells. All oligo-DNAs that contained hexamer palindromes with CG motif(s), except for ACCGGT, induced IFN-γ production by M-450 CD3-stimulated T cells. When the activity was presented as a percentage of the control in six independent experiments, the order of potency among the active palindromes was as follows: CGGCCG (273 ± 57 pg/ml, mean ± SE); TCGCGA (223 ± 45); AGCGCT (220 ± 38); AACGTT (218 ± 13); ATCGAT (173 ± 19); CGATCG (166 ± 19); GCGCGC (163 ± 6); CGTACG (160 ± 21); and GACGTT (132 ± 6). That of MY-1 was 233 ± 12. These values were not statistically different, but one of the weakly active sequences for IL-2-activated NK cells, CGGCCG, was distinctly active for anti-CD3-stimulated T cells. A representative result of these experiments is shown as the amount of IFN-γ in the culture supernatant in Fig. 11. A non-CG palindrome, CG of which was replaced with GC, i.e., AAGCTT-30, showed no effect on IFN-γ production (Fig. 11 and Table II).

FIGURE 11.

Effect of different sequences of palindromes on IFN-γ production by M-450 CD3-stimulated T cells. T cells (2 × 106/ml) were cultured in triplicate for 24 h with medium, 50 μg/ml MY-1, or 5 μM oligo-DNAs that contain different sequences of palindrome or nonpalindrome, in the presence of 2 × 106 particles/ml M-450 CD3. The sequences of oligo-DNAs are listed in Figs. 3 and 8. The results shown are representative of six independent experiments with cells obtained from different donors. The IFN-γ concentrations in culture supernatants are expressed as the mean ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the control value with medium alone.

FIGURE 11.

Effect of different sequences of palindromes on IFN-γ production by M-450 CD3-stimulated T cells. T cells (2 × 106/ml) were cultured in triplicate for 24 h with medium, 50 μg/ml MY-1, or 5 μM oligo-DNAs that contain different sequences of palindrome or nonpalindrome, in the presence of 2 × 106 particles/ml M-450 CD3. The sequences of oligo-DNAs are listed in Figs. 3 and 8. The results shown are representative of six independent experiments with cells obtained from different donors. The IFN-γ concentrations in culture supernatants are expressed as the mean ± SD. ∗ and ∗∗, p < 0.05 and p < 0.01, respectively, compared with the control value with medium alone.

Close modal

We then examined the activities of nonpalindrome oligo-DNAs. Although replacement of AACGTT of AACGTT-30 by nonhexamer palindromes such as AACGCT, AACGTC, or AACGTA exhibited IFN-γ inducing activity, when replaced by AACGGT or AACGTG, the activity was very weak or completely undetectable. AACGAT remained inactive (Table II). The substitution of GC for CG in the active motifs abolished their activities (Table II). However, the oligo-DNA, 1643, which contains AACGCT, did not induce IFN-γ production. Then, AACGTT, AACGCT, AACGTC, and AACGTA were flanked by oligo-G to determine whether a backbone sequence changes the activity of these sequences. In the oligo-G-flanked sequences, only AACGTT (g12AAC) showed potent activity (Table II), and this seemed most potent for activated T cells among the sequences tested. To examine whether the IFN-γ-inducing activity of g12AAC is modified by the methylation of CG, we synthesized methylated g12AAC in which CG inside the palindrome was methylated. As shown in Table II, the methylated g12AAC did not induce the IFN-γ production. Other well-investigated immunostimulatory oligo-DNAs, 1668, 1758, and 2105, which activate mouse spleen cells or human B cells, had little activity for anti-CD3-stimulated T cells (Table II).

In this study, we demonstrated that both BCG-derived DNA (MY-1) and synthetic oligo-DNAs directly induce human NK cells and activated T cells to produce IFN-γ and that the autocrine IFN-γ enhances NK activity. Our present study reveals that the oligo-DNA responsiveness of human IFN-γ-producing cells differ with that of mice in terms of the sequence requirement and that the effective sequences are different according to the cell types and/or the activation status. Furthermore, the IFN-γ-inducing activity of the oligo-DNA was somehow interdependent on the presence of CG, the context of the core motif of CG, and its outer flanking sequences.

To exclude an indirect action caused by contaminated Mos/Mφs, we purified the NK and T cell fractions to more than 97% CD56+ and 99% CD3+ cells, respectively, and less than 1% Mos/Mφs. We also added mAbs against IL-12, TNF-α, IL-15, IL-18, or IFN-α to these cell cultures with oligo-DNAs to determine whether these cytokines produced by contaminated Mos/Mφs (<1%) were involved in the IFN-γ production. However, the levels of IFN-γ production remained unchanged. Indeed, none of these cytokines was detected in the culture supernatants of the NK or T cell population (data not shown). For instance, the concentration of IL-12 measured by ELISA was less than the detectable dose (1 pg/ml) which was not sufficient to induce IFN-γ production by these cells in our preliminary experiments. Furthermore, g10GACGA induced TNF-α production in the same culture conditions; however, the level was too low (2.86 ± 1.70 pg/ml in 40-h culture, mean ± SD) to induce IFN-γ production (Ref. 45 and our observation). Our unpublished data suggest that TNF-α detected in the NK cell culture with g10GACGA may be produced by NK cells rather than by Mos/Mφs. These facts thus indicate that the effect of Mos/Mφs contaminating the NK or T cell fraction at <1% was negligible in the IFN-γ production.

It has been reported in mice that bacterial DNA or oligo-DNA does not induce IFN-γ production or NK enhancement when purified NK cells or Mo/Mφ-depleted nonadherent cells are used without additional stimuli as the responders (19, 22, 27). In humans, these DNAs directly activated NK cells (present study). This implies that mouse and human NK cells behave differently in response to the DNA stimulation. It is unclear as to what caused the difference in the responsiveness of NK cells to oligo-DNAs between our present study and the others. In B cells, a differential requirement regarding the oligo-DNA sequence between humans and mice has been observed (33, 34). The sensitivity of NK cells to oligo-DNA sequences may also, therefore, be different between mice and humans. The PuPuCGPyPy sequences that were immunostimulatory for mouse spleen cells (33), such as AACGCT or AACGTC, did not directly activate mouse NK cells (27). In the present study, these sequences did not act directly on human NK cells either. Although the hexamer palindromes with CG dinucleotide(s) also did not activate mouse NK cells (9, 12, 14, 19, 27), they were active for human NK cells in our study. Human NK cells thus appear sensitive to oligo-DNA, especially when the particular sequences such as the CG palindromes are present. For Mos/Mφs, hexamer palindromes behave actively regardless of the species (12, 14, 15, 19, 23, 28, 46). In this study, these sequences were shown to be effective also for activated human T cells to enhance IFN-γ production. The palindrome sequences containing the CG motif may therefore be some of the most potent sequences for immunocompetent cells involved in the induction of IFN-γ production in humans.

Among the CG-oligo-DNAs tested in this study, the favorable sequences for the IFN-γ induction differed with the cell lineage and/or its activation status. For example, in palindromes, ATCGAT was more effective in unactivated NK cells than in activated NK cells, and vice versa in AACGTT. Further, the weakly active CGGCCG in activated NK cells was distinctly active for activated T cells. More importantly, ACCGGT, one of the palindrome sequences that was inactive for unactivated NK cells turned out to be active for IL-2-activated NK cells. Because in humans ACCGGT is inactive for other types of cells such as Mos/Mφs (15), B cells (34), and even M-450 CD3-activated T cells (Fig. 11), this sequence seems to be specific for NK cells in relation to their activation status. In addition to the palindromes, certain nonpalindrome sequences acted effectively when NK or T cells were activated (Figs. 9 and 11 and Table II). Also in these sequences, however, the action differed with the targeted cells as well, and AACGAT seemed favorable for activated NK cells but not for activated T cells. From these facts, cell activation- and/or cell lineage-specific sequences are likely in terms of the IFN-γ-inducing activity. Oligo-DNA-binding molecules which would be expressed differently by the cell lineage and/or the activation status may cause the different responsiveness of IFN-γ-producing cells to the sequences.

The CG palindrome oligo-DNAs used as the IFN-γ inducer in this study are analogues of BCG-4a (GACGTC-30 in this study) which was randomly selected from the cDNA-encoding 64-KDa protein (Ag A) of M. bovis BCG (12). These analogues contain 4 CG dinucleotides besides the CG(s) inside palindrome. In this study, the change of CG to GC or to methylated CG in palindrome diminished the IFN-γ-inducing activity (Figs. 9 and 11 and Table II). Therefore, unmethylated CG palindrome in these oligo-DNAs appears critical for their activity. However, it should be also noted that the flanking bases outside the CG palindrome seemed to cooperate with the active core motif to implement the activity. This indicates that the immunostimulatory sequences, PuPuCGPyPy, such as AACGCT, exhibited the IFN-γ-inducing activity when inserted into BCG-4a instead of the palindrome motif (Table II). This sequence was inactive when used as the 1643 oligo-DNA itself or by being flanked with oligo-G (Table II). These facts signify that the activity of the CG-oligo-DNA is somehow interdependent on the presence of CG, the context of the core motif of CG, and its outer flanking sequences.

The regulatory role of the flanking bases could be further suggested by the following results. For instance, when the palindrome was flanked with oligo-G, CGATCG acquired more ability to induce IFN-γ production by human NK cells, but AACGTT failed to do so (Fig. 4). Oligo-G has an increased affinity to cellular membranes (46), and it has been hypothesized that the G quartet formed by four contiguous guanosine residues reduces the degree of rotational freedom of oligo-DNA (47). In addition, as for the biological action, oligo-G itself is not only mitogenic for murine B lymphocytes (32) but also inhibitory for NK enhancement (46) and IFN-γ production in the phosphorothioate form (48). Therefore, oligo-G may act as a regulator of palindromes by promoting the cell entry and binding to targeted intracellular molecules; thus, the efficacy of palindromes to modulate the targeted molecules may be augmented by being flanked with the oligo-G. This could result in the enhancement of IFN-γ production with some types of palindrome such as CGATCG (g10GACGA) or could conversely cause the suppression of IFN-γ production by perturbing the interacting signals toward the IFN-γ production in other types of palindrome such as AACGTT in NK cells. This may also be one of the reasons why oligo-G-flanked AACGTT (19) or TCAACGTTGA (27) did not activate the murine NK cells. Changes of the IFN-γ-inducing activity by oligo-G flanking was further observed when activated T cells were targeted: the activity of the palindrome AACGTT was greatly enhanced by the oligo-G flanking; and that of the nonpalindrome AACGCT, AACGTC, or AACGTA was decreased to negligible levels. The flanking sequence thus appears to influence the oligo-DNA activity in different ways according to the sequence of the core motif and also to the lineage and/or activation status of the target cells, in terms of IFN-γ production.

We previously reported that the biological activity of palindromes is triggered after their entry into the cells (46, 49). As shown in other types of cells (50, 51), oligo-DNA containing palindrome with CG may be located in endosomes and in the nucleus once it is taken up by NK or activated T cells. However, the mechanisms by which oligo-DNA induces IFN-γ production in these cells remain to be identified. Yi et al. (25) showed that oligo-DNA directly increases the transcriptional activity of IL-6 promoter, suggesting an interaction of oligo-DNA with responsive elements. Stacey et al. (29) and Sparwasser et al. (30) demonstrated that oligo-DNA modulates the activity of transcription factors. As an alternative mechanism, a certain structure of oligo-DNA may interact, as a charged structure, with second messenger-delivered signals which are involved in IFN-γ production, because the calf thymus-DNA structure activates p68 kinase, which has specific ATP-binding sites (52). This mechanism should be examined in the oligo-DNA-induced IFN-γ production in human NK or activated T cells, because the expression of CD69 Ag, which was reported to be involved in signal transduction (53), was enhanced by oligo-DNA in NK cells.

A clinical trial with MY-1 has been performed in Japan to assess its efficacy as an immunotherapeutic agent for malignant diseases, with positive results (54). Our present results imply that not only MY-1 but also the immunogenic synthetic oligo-DNAs induce multiple immune responses in vivo including NK cell activation and the possible induction of cytotoxic T cells, both of which are major components of the immune defense system against neoplasms. If NK or T cells are activated under some circumstances, the in vivo action of these oligo-DNAs may be augmented under those circumstances. Further studies are required to identify physiological potentiators of oligo-DNAs for the better therapeutic efficacy of these agents.

We thank Professor C. Kumakura, Kinjo Gakuin University, for research assistance and helpful advice; Professor Y. Kimura, Fukui Medical University, for the use of his laboratory equipment while performing this work; Chino Kobayashi for help preparing the manuscript; and Junko Yamamoto, Central Research Laboratories, Fukui Medical School, for her helpful discussion regarding flow cytometric analysis.

1

This work was supported in part by grants from the Japan Health Science Foundation and Smoking Research Foundation.

3

Abbreviations used in this paper: BCG, bacillus Calmette-Gúerin; CG, CpG; oligo-DNA, oligodeoxynucleotide; Mo, monocyte; Mφ, macrophage; LGL, large granular lymphocyte; NAC, nonadherent cells.

1
Bast, R. C., Jr, B. Simpson Bast.
1976
. Critical review of previously reported animal studies of tumor immunotherapy with non-specific immunostimulants.
Ann. NY Acad. Sci.
277
:
60
2
Mastrangelo, M. J., D. Berd, R. E. Bellet.
1976
. Critical review of previously reported clinical trials of cancer immunotherapy with non-specific immunostimulants.
Ann. NY. Acad. Sci.
277
:
94
3
Lamoureux, G., R. Turcotte, and V. Portelance. 1976. BCG in Cancer Immunotherapy. Grune & Stratton, New York.
4
Terry, W. D., Y. Yamamura.
1979
. In
Immunobiology and Immunotherapy of Cancer: Developments in Immunology
Vol. 6
: Elsevier/North-Holland, New York.
5
Tokunaga, T., H. Yamamoto, S. Shimada, H. Abe, T. Fukuda, Y. Fujisawa, Y. Furutani, O. Yano, T. Kataoka, T. Sudo, N. Makiguchi, T. Suganuma.
1984
. Antitumor activity of deoxyribonucleic acid fraction from Mycobacterium bovis BCG. I. Isolation, physicochemical characterization, and antitumor activity.
J. Natl. Cancer Inst.
72
:
955
6
Shimada, S., O. Yano, H. Inoue, E. Kuramoto, T. Fukuda, H. Yamamoto, T. Kataoka, T. Tokunaga.
1985
. Antitumor activity of the DNA fraction from Mycobacterium bovis BCG. II. Effects on various syngeneic mouse tumors.
J. Natl. Cancer Inst.
74
:
681
7
Shimada, S., O. Yano, T. Tokunaga.
1986
. In vivo augmentation of natural killer cell activity with a deoxyribonucleic acid fraction of BCG.
Gann
77
:
808
8
Mashiba, H., K. Matsunaga, H. Tomoda, M. Furusawa, S. Jimi, T. Tokunaga.
1988
. In vitro augmentation of natural killer activity of peripheral blood cells from cancer patients by a DNA fraction from Mycobacterium bovis BCG.
Jpn. J. Med. Sci. Biol.
41
:
197
9
Yamamoto, S., E. Kuramoto, S. Shimada, T. Tokunaga.
1988
. In vitro augmentation of natural killer cell activity and production of interferon-α/β and -γ with deoxyribonucleic acid fraction from Mycobacterium bovis BCG.
Jpn. J. Cancer Res.
79
:
866
10
Yamamoto, S., T. Yamamoto, S. Shimada, E. Kuramoto, O. Yano, T. Kataoka, T. Tokunaga.
1992
. DNA from bacteria, but not from vertebrates, induces interferons, activates natural killer cells and inhibits tumor growth.
Microbiol. Immunol.
36
:
983
11
Kuramoto, E., S. Toizumi, S. Shimada, T. Tokunaga.
1989
. In situ infiltration of natural killer-like cells induced by intradermal injection of the nucleic acid fraction from BCG.
Microbiol. Immunol.
33
:
929
12
Tokunaga, T., O. Yano, E. Kuramoto, Y. Kimura, T. Yamamoto, T. Kataoka, S. Yamamoto.
1992
. Synthetic oligonucleotides with particular base sequences from the cDNA encoding proteins of Mycobacterium bovis BCG induce interferons and activate natural killer cells.
Microbiol. Immunol.
36
:
55
13
Kataoka, T., S. Yamamoto, T. Yamamoto, E. Kuramoto, Y. Kimura, O. Yano, T. Tokunaga.
1992
. Antitumor activity of synthetic oligonucleotides with sequences from cDNA encoding proteins of Mycobacterium bovis BCG.
Jpn. J. Cancer Res.
83
:
244
14
Yamamoto, S., T. Yamamoto, T. Kataoka, E. Kuramoto, O. Yano, T. Tokunaga.
1992
. Unique palindromic sequences in synthetic oligonucleotides are required to induce INF and augment INF-mediated natural killer activity.
J. Immunol.
148
:
4072
15
Yamamoto, T., S. Yamamoto, T. Kataoka, K. Komuro, M. Kohase, T. T. Tokunaga.
1994
. Synthetic oligonucleotides with certain palindromes stimulate interferon production of human peripheral blood lymphocytes in vitro.
Jpn. J. Cancer Res.
85
:
775
16
Yamamoto, T., S. Yamamoto, T. Kataoka, T. Tokunaga.
1994
. Ability of oligonucleotides with certain palindromes to induce interferon production and augment natural killer cell activity is associated with their base length.
Antisense Res. Dev.
4
:
119
17
Kuramato, E., O. Yano, Y. Kimura, M. Baba, T. Makino, S. Yamamoto, T. Yamamoto, T. Kataoka, T. Tokunaga.
1992
. Oligonucleotide sequences required for natural killer cell activation.
Jpn. J. Cancer Res.
83
:
1128
18
Klinman, D. M., A.-K. Yi, S. L. Beaucage, J. Conover, A. M. Krieg.
1996
. CpG motifs present in bacterial DNA rapidly induce lymphocytes to secrete interleukin 6, interleukin 12, and interferon γ.
Proc. Natl. Acad. Sci. USA
93
:
2879
19
Halpern, M. D., R. J. Kurlander, D. S. Pisetsky.
1996
. Bacterial DNA induces murine interferon-γ production by stimulation of interleukin-12 and tumor necrosis factor-α.
Cell. Immunol.
167
:
72
20
Cowdery, J. S., J. H. Chace, A.-K. Yi, A. M. Krieg.
1996
. Bacterial DNA induces NK cells to produce IFN-γ in vivo and increases the toxicity of lipopolysaccharides.
J. Immunol.
156
:
4570
21
Klinman, D. M., G. Yamshchikov, Y. Ishigatsubo.
1997
. Contribution of CpG motifs to the immunogenicity of DNA vaccines.
J. Immunol.
158
:
3635
22
Chace, J. H., N. A. Hooker, K. L. Mildenstein, A. M. Krieg, J. S. Cowdery.
1997
. Bacterial DNA-induced NK cell IFN-γ production is dependent on macrophage secretion of IL-12.
Clin. Immunol. Immunopathol.
84
:
185
23
Roman, M., E. Martin-Orozco, J. S. Goodman, M. D. Nguyen, Y. Sato, A. Ronaghy, R. S. Kornbluth, D. D. Richman, D. A. Carson, E. Raz.
1997
. Immunostimulatory DNA sequences function as T helper-1-promoting adjuvants.
Nat. Med.
3
:
849
24
Chu, R. S., O. S. Targoni, A. M. Krieg, P. V. Lehmann, C. V. Harding.
1997
. CpG oligodeoxynucleotides act as adjuvants that switch on T helper 1 (Th1) immunity.
J. Exp. Med.
186
:
1623
25
Yi, A.-K., D. M. Klinman, T. L. Martin, S. Matson, A. M. Krieg.
1996
. Rapid immune activation by CpG motifs in bacterial DNA.
J. Immunol.
157
:
5394
26
Schwartz, D. A., T. J. Quinn, P. S. Thorne, S. Sayeed, A.-K. Yi, A. M. Krieg.
1997
. CpG motifs in bacterial DNA cause inflammation in the lower respiratory tract.
J. Clin. Invest.
100
:
68
27
Ballas, Z. K., W. L. Rasmussen, A. M. Krieg.
1996
. Induction of NK activity in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial DNA.
J. Immunol.
157
:
1840
28
Sato, Y., M. Roman, H. Tighe, D. Lee, M. Corr, M.-D. Nguyen, G. J. Silverman, M. Lotz, D. A. Carson, E. Raz.
1996
. Immunostimulatory DNA sequences necessary for effective intradermal gene immunization.
Science
273
:
352
29
Stacey, K. J., M. J. Sweet, D. A. Hume.
1996
. Macrophages ingest and are activated by bacterial DNA.
J. Immunol.
157
:
2116
30
Sparwasser, T., T. Miethke, G. Lipford, A. Erdmann, H. Hacker, K. Heeg, H. Wagner.
1997
. Macrophages sense pathogens via DNA motifs: induction of tumor necrosis factor-α-mediated shock.
Eur. J. Immunol.
27
:
1671
31
Messina, J. P., G. S. Gilkeson, D. S. Pisetsky.
1991
. Stimulation of in vitro murine lymphocyte proliferation by bacterial DNA.
J. Immunol.
147
:
1759
32
Pisetsky, D. S., C. Reich.
1993
. Stimulation of in vitro proliferation of murine lymphocytes by synthetic oligodeoxynucleotides.
Mol. Biol. Rep.
18
:
217
33
Krieg, A. M., A.-K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale, G. A. Koretzky, D. M. Klinman.
1995
. CpG motifs in bacterial DNA trigger direct B-cell activation.
Nature
374
:
546
34
Liang, H., Y. Nishioka, C. F. Reich, D. S. Pisetsky, P. E. Lipsky.
1996
. Activation of human B cells by phosphorothioate oligodeoxynucleotides.
J. Clin. Invest.
98
:
1119
35
Sun, S., C. Beard, J. Jaenisch, P. Jones, J. Sprent.
1997
. Mitogenicity of DNA from different organisms for murine B Cells.
J. Immunol.
159
:
3119
36
Lipford, G. B., M. Bauer, C. Blank, R. Reiter, H. Wagner, K. Heeg.
1997
. CpG-containing synthetic oligonucleotides promote B and cytotoxic T cell responses to protein antigen: a new class of vaccine adjuvants.
Eur. J. Immunol.
27
:
2340
37
Wooldridge, J. E., Z. Ballas, A. M. Krieg, G. J. Weiner.
1997
. Immunostimulatory oligodeoxynucleotides containing CpG motifs enhance the efficacy of monoclonal antibody therapy of lymphoma.
Blood
89
:
2994
38
Weiner, G. J., H.-M. Liu, J. E. Wooldridge, C. E. Dahle, A. M. Krieg.
1997
. Immunostimulatory oligodeoxynucleotides containing the CpG motif are effective as immune adjuvants in tumor antigen immunization.
Proc. Natl. Acad. Sci. USA
94
:
10833
39
Davis, H. L., R. Weeranta, T. J. Waldschmidt, L. Tygrett, J. Schorr, A. M. Krieg.
1998
. CpG DNA is a potent enhancer of specific immunity in mice immunized with recombinant hepatitis B surface antigen.
J. Immunol.
160
:
870
40
Kline, J. N., T. J. Waldschmidt, T. R. Businga, J. E. Lemish, J. V. Weinstock, P. S. Thorne, A. M. Krieg.
1998
. Modulation of airway inflammation by CpG oligodeoxynucleotides in a murine model of asthma.
J. Immunol.
160
:
2555
41
Zimmermann, S., O. Egeter, S. Hausmann, G. B. Lipford, M. Rocken, H. Wagner, K. Heeg.
1998
. CpG oligodeoxynucleotides trigger protective and curative Th1 responses in lethal murine leishmaniasis.
J. Immunol.
160
:
3627
42
Sun, S., H. Kishimoto, J. Sprent.
1998
. DNA as an adjuvant: capacity of insect DNA and synthetic oligodeoxynucleotides to augment T cell responses to specific antigen.
J. Exp. Med.
187
:
1145
43
Lewis, C. E., and J. O’D. McGee. 1992. The Natural Immune System: the Natural Killer Cell. IRS Press at Oxford University Press, New York.
44
Seder, R. A., W. E. Paul.
1994
. Acquisition of lymphokine-producing phenotype by CD4+ T cells.
Annu. Rev. Immunol.
12
:
635
45
Wherry, J. C., R. D. Schreiber, E. R. Unanue.
1991
. Regulation of γ interferon production by natural killer cell in scid mice: roles of tumor necrosis factor and bacterial stimuli.
Infect. Immun.
59
:
1709
46
Kimura, Y., K. Sonehara, E. Kuramoto, T. Makino, S. Yamamoto, T. Yamamoto, T. Kataoka, T. Tokunaga.
1994
. Binding of oligoguanylate to scavenger receptors is required for oligonucleotides to augment NK cell activity and induce IFN.
J. Biochem.
116
:
991
47
Stein, C. A..
1995
. Does antisense exist?.
Nat. Med.
1
:
1119
48
Halpern, M. D., D. S. Pisetsky.
1995
. In vitro inhibition of murine IFNγ production by phosphorothioate deoxyguanosine oligomers.
Immunopharmacology
29
:
47
49
Yamamoto, T., S. Yamamoto, T. Kataoka, T. Tokunaga.
1994
. Lipofection of synthetic oligodeoxyribonucleotide having a palindromic sequence of AACGTT to murine splenocytes enhances interferon production and natural killer activity.
Microbiol. Immunol.
38
:
831
50
Zhao, Q., T. Waldschmidt, E. Fisher, C. J. Herrera, A. M. Krieg.
1994
. Stage-specific oligonucleotide uptake in murine bone marrow B-cell precursors.
Blood
84
:
3660
51
Beltinger, C., H. U. Saragovi, R. M. Smith, L. LeSauteur, N. Shah, L. DeDionisio, L. Christensen, A. Raible, L. Jarette, A. M. Gewirtz.
1995
. Binding, uptake, and intracellular trafficking of phosphorothioate-modified oligodeoxynucleotides.
J. Clin. Invest.
95
:
1814
52
Hovanessian, A. G., J. Galabru.
1987
. The double-stranded RNA-dependent protein kinase is also activated by heparine.
Eur. J. Biochem.
167
:
467
53
Lopez-Cabrera, M., A. G. Santis, E. Fernandez-Ruiz, R. Blacher, F. Esch, P. Sanchez-Mateos, F. Sanchez-Madrid.
1993
. Molecular cloning, expression, and chromosomal localization of the human earliest lymphocyte activation antigen AIM/CD69, a new member of the C-type animal lectin superfamily of signal-transmitting receptors.
J. Exp. Med.
178
:
537
54
Ochiai, T., T. Suzuki, K. Nakajima, T. Asano, M. Mukai, T. Kouzu, K. Okuyama, K. Isono, H. Nishijima, M. Ohkawa.
1986
. Studies on lymphocyte subsets of regional lymph nodes after endoscopic injection of biological response modifiers in gastric cancer patients.
Int. J. Immunother.
2
:
259