Phosphorothioate oligodeoxynucleotides with CpG motifs (CpG-ODNs) activate various immune cell subsets and induce production of numerous cytokines. To evaluate whether CpG-ODNs can induce rejection of established malignant tumor, A/J mice were challenged by the s.c. implantation of a syngenic neuroblastoma cell line (neuro2a) and subsequently injected with CpG-ODNs in the vicinity of the tumor. Daily injections of 10 μg CpG-ODNs for 15 days seemed to be the most potent regimen, leading to the eradication of 5-mm-diameter tumors in one-half of the animals and a significant tumor growth inhibition when compared with controls (88% reduction volume; P < 0.001). CpG-ODN-cured animals were further protected against a new tumor challenge. The antitumoral effect of CpG-ODNs was dependent on CpG motifs, and natural killer cells seemed to play a critical role in tumor rejection. We conclude that immunostimulatory CpG-ODNs may induce the rejection of established tumors and warrant further evaluation as a potential immunotherapeutic agent.

Increasing the ability of the immune system to mediate tumor regression in vivo is a promising area in the management of cancer. Studies in animal models have shown that immunization against a selected tumor antigen can delay tumor growth and occasionally cure established tumors (1, 2). In neuroblastomas, both Hu and gangliosides GD2 have been described as potential antigens for immunization with partial protection against a tumor challenge (3, 4). Direct activation of the immune system and stimulation of tumor infiltrating lymphocytes (TILs) by various immunological agents represent another approach in cancer immunotherapy, which overcomes the need for the selection and purification of an immunologically relevant cancer antigen. IL-23 for example, has been used to stimulate NK cells and has shown antitumoral effect in murine neuroblastomas (5), but clinical applications are limited by moderate efficacy and IL-2 toxicity (6).

Some synthetic ODNs containing an unmethylated CpG dinucleotide have recently been shown to stimulate immune responses independently of any known antisense effect (7, 8). When injected into mice, they activate various immune cell subsets, including macrophages and NK cells (8, 9), and induce production of a wide variety of T helper-1-promoting cytokines (9, 10, 11). CpG motifs flanked by two 5′ purines and two 3′ pyrimidines seemed to be the most potent immunostimulatory sequences (8), and ODNs containing such immunostimulatory CpG motifs are promising new immune adjuvants (11, 12). To evaluate whether a CpG-ODN could stimulate the immune system and lead to tumor rejections, A/J mice were inoculated s.c. with a syngenic neuroblastoma cell line and subsequently treated with injections of CpG-ODN. We here show that CpG-ODN treatment displays a potent antitumoral effect, sometimes leading to complete eradication of tumors, and that this effect is critically dependent on NK cells.

Oligonucleotides.

Purified single-stranded phosphorothioate ODNs were purchased from Genset (Paris, France). The sequences used in this study were CpG-ODN, 5′-TGACTGTGAACGTTCGAGATGA-, which contains two CpG dinucleotides, one being part of an immunostimulatory sequence (5′-AACGTT) and IMM-ODN 5′-TGACTGTGAAGGGTAGAGATGA-, in which both of the CpG motifs have been mutated, as described in Roman et al.(11). Lipopolysaccharide levels in ODNs, assessed by the Limulus assay, were less than 1 ng/mg (Biowhittaker, Emerainville, France).

Neuroblastoma Cell Line and in Vitro Toxicity Assays.

The neuro-2a cell line, a subclone of the C1300 murine neuroblastoma that was developed in A/J mice (CCL-131, American Type Culture Collection) was maintained in MEM with 0.1 mm nonessential amino acids and 10% fetal bovine serum (Life Technologies, Grand Island, NY). For toxicity assays, 1500 cells were plated into 1.5-cm2 culture flasks, and CpG-ODN was added at various concentrations (5.10−6, 5.10−7, or 5.10−8m) at day 1 in culture. Cells were harvested on day 3 by trypsinization and counted on a Malassez hematocytometer. Cell viability was checked by trypan blue exclusion. The experiments were performed in triplicate, and results were expressed as the mean ± SE.

Tumor Implantation and CpG-ODN Treatment.

Six-week-old male A/J mice (Charles River, St. Aubin les Elbeufs, France) and Nude mice (Centre d’Elevage Regional Janvier, Lyon, France) were given injections s.c. of 106 neuro2a cells into the right flank. Depending on protocols, mice were treated either 2 days after tumor inoculation or when the tumor diameter had reached 5 mm (approximatively 10 days after tumor inoculation). Mice were injected with 50 μl of isotonic sodium chloride (controls) or ODNs (see below for schedule) dissolved in 50 μl of saline, either i.p. or s.c. in the vicinity of the tumor. Tumor volumes were assessed with a caliper every 4 days by using the formula: φ/6 × length × width2(4).

Immunohistochemistry.

For histological analysis, mice bearing 5-mm-diameter tumors were given injections daily with 50 μg of CpG-ODNs or 50 μl saline for 2 days, and killed on the third day. Tumors were surgically removed, snap-frozen, and stored at −80°C. Frozen sections (10 μm thick) were fixed for 10 min in 10%-buffered formalin, and sequentially incubated for 1 h with 10% normal goat serum (Jackson Immunoresearch, West Grove, PA), and for 2 h with a rat antimouse-CD8a (1:30; clone KT15, Serotec, Oxford, United Kingdom). The sections were further incubated for 1 h with fluoresceine isothiocyanate-conjugated goat antimouse IgG1 (1:50; Clinisciences, Montrouge, France) and examined under a Leica fluorescent microscope. Quantitative analysis of labeled cells were performed using three different sections for each sample by an investigator who was blinded to the animal’s history.

In Vivo Depletion Studies.

Mice were depleted of NK+ cells by i.p. injections of 200 μg monoclonal anti-NK1.1 antibody (clone PK136, Serotec, Oxford, United Kingdom), which have been reported to deplete NK cells in mice for at least 7 days (13). Mice in the control groups were injected with 200 μg of normal mouse IgG (Sigma, Saint-Quentin, France).

Statistics.

Statistical analysis of differences in tumor size among the various groups was determined by the ANOVA repeated-measures test.

CpG-ODNs Inhibited Tumor Growth and the Antitumoral Effect Was Dependent on CpG Motifs.

In a first set of experiments, mice that had been implanted s.c. with 106 neuro2a cells into the right flank were injected 2 days later at the same location with a single injection of either 50 μl of sodium chloride or 50 μg of CpG-ODNs. By day 22 after the tumor implantation, animals that had received 50 μg CpG-ODNs (n = 8) had tumor volumes that were 67% smaller than controls (n = 8; mean tumor volume ± S.E., 474 ± 259 versus 1434 ± 409 mm3; P < 0.01). If CpG-ODNs (50 μg) were not injected s.c. but i.p. (n = 8), a similar but less marked effect was seen, with tumor volumes 37% smaller than controls (910 ± 313 versus 1434 ± 409 mm3; P = 0.38, not significant). The antitumoral effect of CpG-ODNs was dependent on the CpG motifs, inasmuch as no tumor growth inhibition was seen when mice were treated s.c. with 50 μg of IMM-ODNs [n = 8; mean tumor volume + SE, 1603 + 420 mm3 (Fig. 1)].

Peritumoral Injections of CpG-ODNs Induced Rejection of Established Tumors Measuring 5 mm.

We then investigated whether or not CpG-ODN treatment could induce regression of large established tumors. A/J mice bearing neuro2a tumors of 5 mm in diameter were given injections s.c. around the tumor with saline (50 μl daily for 15 days; n = 6), or CpG-ODNs (10 μg daily for 15 days; n = 6), or CpG-ODNs (50 μg daily for 3 days; n = 6). All of the control animals treated with saline developed progressive tumors and died. Treatment with CpG-ODNs induced tumor disappearance in three of six mice in the group treated with 15 injections, and in two of six mice in the group treated with 3 injections. In these animals, the tumors disappeared progressively within 4–15 days, without recurrence over a 3-month period of observation. In mice which did not reject the tumors, CpG-ODN treatment resulted in a dramatic reduction in tumor growth rate. By day 22 after the initiation of treatment, when compared with controls, the mean tumor volume was only 12% for the group treated with 15 injections of 10 μg (mean tumor volume ± S.E., 910 ± 728 versus 7289 ± 1142 mm3; P < 0.001) and 36% for the group treated with 3 injections of 50 μg (2656 ± 851 versus 7289 ± 851 mm3; P < 0.004). The difference between both CpG-ODN regimens was statistically significant (P < 0.05; Fig. 2).

Mice Cured by CpG-ODNs Rejected a New Inoculation with Neuro2a Cells.

Some of the surviving animals (n = 4) which had rejected tumors of 5 mm in diameter were subjected to a second tumor challenge with 106 neuro2a cells, 12 weeks after completion of the initial treatment with CpG-ODNs. All of them remained tumor-free and survived without any treatment.

CpG-ODN Was Not Toxic on Neuro2a Cells in Vitro.

Neuro2a cells were cultured for 48 h with various concentrations (5.10−6, 5.10−7, and 5.10−8m) of CpG-ODNs. No inhibition of proliferation was seen with CpG-ODNs when compared with control [(mean ± S.E., 5750 ± 287; 5667 ± 333; 6000 ± 577; and 5667 ± 905 cells for control (5.10−6, 5.10−7, and 5.10−8m CpG-ODNs, respectively)], which suggests that the action mechanism of CpG-ODN is not mediated by a direct toxicity on tumor cells.

The Antitumoral Effect of CpG-ODNs Was Abrogated by NK Cell Depletion.

A/J mice bearing neuro2a tumors of 5 mm in diameter were randomly assigned to experimental groups of n = 4 animals. Mice were either injected with normal IgG (groups 1 and 2) or depleted of NK cells by injections of 200 μg monoclonal anti-NK1.1 antibody (group 3). Twenty-four h after these initial injections, mice were treated daily for 7 days with 50 μl saline (group 1) or 10 μg CpG-ODNs (groups 2 and 3) s.c. around the tumor. The tumor volumes were measured on day 7, to assess the effect of NK cell depletion on tumor growth inhibition. Whereas CpG-ODNs treatment in nondepleted mice resulted in a 85% reduction of tumor volumes when compared with controls (mean tumor volume + S.E., 204 ± 113 versus 1000 ± 221 mm3; P < 0.05), the antitumor effect was dramaticaly reduced by NK cell depletion (mean tumor volume ± S.E., 725 ± 175 mm3). (Fig. 3).

CpG-ODNs Displayed an Antitumoral Effect in Nude Mice.

Nude mice bearing neuro2a tumors of 5 mm diameter were divided in 2 groups (n = 5 in each) and treated s.c. around the tumor with 50 μl of saline or 10 μg of CpG-ODNs daily for 7 days. On day 7, nude mice treated with CpG-ODNs showed a 49% reduction of their mean tumor volume when compared with controls (mean tumor volume ± S.E., 1193 ± 248 versus 2362 ± 532 mm3; P < 0.05; Fig. 4). Despite this antitumoral effect of CpG-ODN treatment, all mice ultimately died from progressive tumor growth within 7 weeks.

CpG-ODNs Did Not Promote Recruitment of CD8-positive Lymphocytes into the Tumors.

A subset of A/J mice bearing 5 mm diameter tumors was treated with saline (n = 4) or 50 μg CpG-ODNs (n = 4) s.c. around the tumor for 2 days and they were killed on day 3. The tumor were removed surgically and examined for CD8+ lymphocytic infiltrations. None of the tumor sections showed significant CD8+ cells, which were always below 5 per 0.135 mm2 field.

These data show that treatment with phosphorothioate ODNs containing a CpG immunostimulatory sequence can cure established neuroblastoma in syngenic mice. Whereas none of the control animals survived the tumor challenge, 50% of the A/J mice treated with daily peritumoral injections of CpG-ODNs for 15 days were cured, and all of the animals treated with CpG-ODNs had reduced tumor growth. i.p. injections of CpG-ODNs were less efficient than peritumoral injections, suggesting that CpG-ODNs exert their effects locally and must reach a sufficient concentration at the tumor site. A direct cytotoxic effect of CpG-ODNs is unlikely, for no cytotoxicity of CpG-ODNs on neuro2a cells was detected in vitro. The mechanism is probably immunogenic inasmuch as the immunostimulatory CpG motif included in our ODNs did appear critical in vivo. This was confirmed by NK-cell depletion experiments, which almost completely abrogated the antitumoral effect of CpG-ODNs in A/J mice. On the contrary, the effect of CpG-ODNs was observed in nude mice who lack CD4 and CD8 lymphocytes but not NK cells, although this antitumoral effect was less pronounced than in the A/J mice. This critical role of NK cells in our model is in accordance with the fact that CpG-ODNs are well known to enhance NK activity both in vitro and in vivo(14, 15), and that neuroblastoma cells are sensitive to natural killing in vitro(5, 16). The role of Ag receptor-negative host defense cells such as NK cells in tumor immunity has long been suspected because some tumors grow faster in NK-depleted animals than in controls (13, 17). In a murine neuroblastoma model, IL-2 can decrease the tumor growth rate through activation of NK cells (5). Similarly, IL-2-transfected fibrosarcoma cells elicited predominantly NK cells at the site of injection and did not seem to require T cells for rejection in unprimed animals (18). We here provide evidence that NK cells display a potent antitumoral effect in nontransfected neuroblastomas when stimulated with CpG-ODNs. Interestingly, a long-term immunity was also primed in our model, inasmuch as cured A/J mice were able to reject additional tumor challenges, which suggests that CD4+ T-cells, at least, were also involved in the immune reaction triggered by CpG-ODN injections.

ODNs with immunostimulatory motifs are under investigation as potential adjuvants in immunization against selected antigens. Promising results have been reported in infectious diseases such as hepatitis B (12). Similarly, such ODNs have been used as adjuvant for immunization against a tumor antigen in a murine lymphoma model (19, 20). In contrast, our data show that direct injections of CpG-ODNs alone around tumors represent a very simple means of achieving therapeutic effects in neuroblastomas without the need for selection and purification of tumoral antigens. This strategy widens the potential therapeutic application for CpG-ODNs and may be a promising new approach in cancer immunotherapy.

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

      
1

This work was supported by Assistance Publique-Hopitaux de Paris and Université Paris VI.

            
3

The abbreviations used are: IL, interleukin; NK, natural killer; ODN, oligodeoxynucleotide; CpG-ODN, ODN containing an immunostimulatory CpG motif.

Fig. 1.

Neuro2a tumor volumes (mean + SE) in A/J mice that received injections 2 days after tumor implantation either into the tumor bed with sodium chloride (♦; n = 8), 50 μg of CpG-ODN (▪; n = 8), 50 μg IMM-ODN (□; n = 8), or i.p. with 50 μg CpG-ODN (▴; n = 8). Days, the number of days after tumor implantation.

Fig. 1.

Neuro2a tumor volumes (mean + SE) in A/J mice that received injections 2 days after tumor implantation either into the tumor bed with sodium chloride (♦; n = 8), 50 μg of CpG-ODN (▪; n = 8), 50 μg IMM-ODN (□; n = 8), or i.p. with 50 μg CpG-ODN (▴; n = 8). Days, the number of days after tumor implantation.

Close modal
Fig. 2.

Neuro2a tumor volumes (mean + SE) in A/J mice bearing neuro2a tumors of 5 mm diameter and injected s.c. around the tumor with 50 μl sodium chloride daily for 15 days (♦; n = 6), or 50 μg of CpG-ODNs daily for 3 days (▪; n = 6), or 10 μg CpG-ODNs daily for 15 days (□; n = 6). Days, the number of days after the initiation of treatment.

Fig. 2.

Neuro2a tumor volumes (mean + SE) in A/J mice bearing neuro2a tumors of 5 mm diameter and injected s.c. around the tumor with 50 μl sodium chloride daily for 15 days (♦; n = 6), or 50 μg of CpG-ODNs daily for 3 days (▪; n = 6), or 10 μg CpG-ODNs daily for 15 days (□; n = 6). Days, the number of days after the initiation of treatment.

Close modal
Fig. 3.

A/J mice bearing neuro2a tumors of 5 mm diameter were injected with normal mouse IgG (group 1, n = 4, and group 2, n = 4) or anti-NK1.1 antibody (group 3, n = 4) and subsequently treated daily for 7 days with saline (group 1) or 10 μg CpG-ODNs (groups 2 and 3) s.c. around the tumor. Tumor volumes (mean + SE) were measured on day 7 to assess the effect of NK cell depletion on tumor growth inhibition.

Fig. 3.

A/J mice bearing neuro2a tumors of 5 mm diameter were injected with normal mouse IgG (group 1, n = 4, and group 2, n = 4) or anti-NK1.1 antibody (group 3, n = 4) and subsequently treated daily for 7 days with saline (group 1) or 10 μg CpG-ODNs (groups 2 and 3) s.c. around the tumor. Tumor volumes (mean + SE) were measured on day 7 to assess the effect of NK cell depletion on tumor growth inhibition.

Close modal
Fig. 4.

Nude mice bearing neuro2a tumors of 5 mm diameter were treated daily for 7 days with saline (group 1, n = 5) or 10 μg CpG-ODN (group 2, n = 5) s.c. around the tumor. The tumor volumes (mean + SE) were measured on day 7.

Fig. 4.

Nude mice bearing neuro2a tumors of 5 mm diameter were treated daily for 7 days with saline (group 1, n = 5) or 10 μg CpG-ODN (group 2, n = 5) s.c. around the tumor. The tumor volumes (mean + SE) were measured on day 7.

Close modal

We thank J. G. Guillet, B. Zalc, and P. F. Pradat for their helpful comments in the redaction of the manuscript.

1
Bueler H., Mulligan R. C. Induction of antigen-specific tumor immunity by genetic and cellular vaccines against MAGE: enhanced tumor protection by coexpression of granulocyte-macrophage colony-stimulating factor and B7-1.
Mol. Med.
,
2
:
545
-555,  
1996
.
2
Syrengelas A. D., Chen T. T., Levy R. DNA immunization induces protective immunity against B-cell lymphoma.
Nat. Med.
,
2
:
1038
-1041,  
1996
.
3
Bergman I., Arbit E., Rosenblum M., Larson S. M., Heller G., Cheung N. K. Treatment of spinal epidural neuroblastoma xenografts in rats using anti-GD2 monoclonal antibody 3F8.
J. Neurooncol.
,
15
:
235
-242,  
1993
.
4
Carpentier A. F., Rosenfeld M. R., Delattre J. Y., Whalen R. G., Posner J. B., Dalmau J. DNA vaccination with HuD inhibits growth of a neuroblastoma in mice.
Clin. Cancer Res.
,
4
:
2819
-2824,  
1998
.
5
Reynolds J. V., Shou J., Choi H., Sigal R., Ziegler M. M., Daly J. M. The influence of natural killer cells in neuroblastoma.
Arch. Surg.
,
124
:
235
-239,  
1989
.
6
Pession A., Prete A., Locatelli F., Pierinelli S., Pession A. L., Maccario R., Magrini E., De Bernardi B., Paolucci P., Paolucci G. Immunotherapy with low-dose recombinant interleukin 2 after high-dose chemotherapy and autologous stem cell transplantation in neuroblastoma.
Br. J. Cancer
,
78
:
528
-533,  
1998
.
7
Yamamoto S., Yamamoto T., Kataoka T., Kuramoto E., Yano O., Tokunaga T. Unique palindromic sequences in synthetic oligonucleotides are required to induce IFN and augment IFN-mediated natural killer activity.
J. Immunol.
,
148
:
4072
-4076,  
1992
.
8
Krieg A. M., Yi A. K., Matson S., Waldschmidt T. J., Bishop G. A., Teasdale R., Koretzky G. A., Klinman D. M. CpG motifs in bacterial DNA trigger direct B-cell activation.
Nature (Lond.)
,
374
:
546
-549,  
1995
.
9
Chace J. H., Hooker N. A., Mildenstein K. L., Krieg A. M., Cowdery J. S. Bacterial DNA-induced NK cell IFN-γ production is dependent on macrophage secretion of IL-12.
Clin. Immunol. Immunopathol.
,
84
:
185
-193,  
1997
.
10
Chu R. S., Targoni O. S., Krieg A. M., Lehmann P. V., Harding C. V. CpG oligodeoxynucleotides act as adjuvants that switch on T helper 1 (Th1) immunity.
J. Exp. Med.
,
186
:
1623
-1631,  
1997
.
11
Roman M., Martin-Orozco E., Goodman J. S., Nguyen M. D., Sato Y., Ronaghy A., Kornbluth R. S., Richman D. D., Carson D. A., Raz E. Immunostimulatory DNA sequences function as T helper-1-promoting adjuvants.
Nat. Med.
,
3
:
849
-854,  
1997
.
12
Davis H. L., Weeranta R., Waldschmidt T. J., Tygrett L., Schorr J., Krieg A. M. CpG DNA is a potent enhancer of specific immunity in mice immunized with recombinant hepatitis B surface antigen.
J. Immunol.
,
160
:
870
-876,  
1998
.
13
Seaman W. E., Sleisenger M., Eriksson E., Koo G. C. Depletion of natural killer cells in mice by monoclonal antibody to NK-1. 1. Reduction in host defense against malignancy without loss of cellular or humoral immunity.
J. Immunol.
,
138
:
4539
-4544,  
1987
.
14
Ballas Z. K., Rasmussen W. L., Krieg A. M. Induction of NK activity in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial DNA.
J. Immunol.
,
157
:
1840
-1845,  
1996
.
15
Boggs R. T., McGraw K., Condon T., Flournoy S., Villiet P., Bennett C. F., Monia B. P. Characterization and modulation of immune stimulation by modified oligonucleotides.
Antisense Nucleic Acid Drug Dev.
,
7
:
461
-471,  
1997
.
16
Main E. K., Lampson L. A., Hart M. K., Kornbluth J., Wilson D. B. Human neuroblastoma cell lines are susceptible to lysis by natural killer cells but not by cytotoxic T lymphocytes.
J. Immunol.
,
135
:
242
-246,  
1985
.
17
Frey A. B. Role of host antigen receptor-bearing and antigen receptor-negative cells in immune response to rat adenocarcinoma 13762.
J. Immunol.
,
156
:
3841
-3849,  
1996
.
18
Bannerji R., Arroyo C. D., Cordon-Cardo C., Gilboa E. The role of IL-2 secreted from genetically modified tumor cells in the establishment of antitumor immunity.
J. Immunol.
,
152
:
2324
-2332,  
1994
.
19
Weiner G. J., Liu H. M., Wooldridge J. E., Dahle C. E., Krieg A. M. Immunostimulatory oligodeoxynucleotides containing the CpG motif are effective as immune adjuvants in tumor antigen immunization.
Proc. Natl. Acad. Sci. USA
,
94
:
10833
-10837,  
1997
.
20
Wooldridge J. E., Ballas Z., Krieg A. M., Weiner G. J. Immunostimulatory oligodeoxynucleotides containing CpG motifs enhance the efficacy of monoclonal antibody therapy of lymphoma.
Blood
,
89
:
2994
-2998,  
1997
.