Characterization of three CpG oligodeoxynucleotide classes with ...

11 downloads 13705 Views 516KB Size Report
class of CpG ODN combines the immune effects of A- and B-Class CpG ODN. C-Class ODN strongly ..... cate that TLR9 is sufficient to at least partially restore.
Eur. J. Immunol. 2004. 34: 251–262

Three distinct immunostimulatory CpG ODN classes

251

Characterization of three CpG oligodeoxynucleotide classes with distinct immunostimulatory activities Jörg Vollmer1, Risini Weeratna2, Paul Payette2, Marion Jurk1, Christian Schetter1, Meike Laucht1, Tanja Wader1, Sibylle Tluk1, Ming Liu2, Heather L. Davis2 and Arthur M. Krieg3 1

Coley Pharmaceutical Group, Langenfeld, Germany Coley Pharmaceutical Group, Ottawa, Canada 3 Coley Pharmaceutical Group, Wellesley, USA 2

Oligodeoxynucleotides (ODN) with unmethylated deoxycytidyl-deoxyguanosine (CpG) dinucleotides (CpG ODN) mimic the immunostimulatory activity of bacterial DNA and are recognized by the Toll-like receptor 9 (TLR9). CpG ODN of the B-Class stimulate strong B cell and NK cell activation and cytokine production. The highest degrees of NK stimulation as well as IFN- § secretion by plasmacytoid DC were found to occur only with A-Class ODN. A third class of CpG ODN combines the immune effects of A- and B-Class CpG ODN. C-Class ODN strongly stimulate B cell or NK cell activation and IFN- § production. In contrast to the AClass, the C-Class is wholly phosphorothioate, has no poly-G stretches, but has palindromic sequences combined with stimulatory CpG motifs. All classes stimulate TLR9-dependent signaling, but with strikingly different dose-response relationships that are quite in contrast to those observed for IFN- § . Effects similar to those on human cells were observed on mouse splenocytes. In contrast, splenocytes from TLR9-deficient mice did not show any response to the three CpG ODN classes. In vivo studies demonstrate that C-Class ODN are very potent Th1 adjuvants. C-Class ODN may represent new therapeutic drugs that combine the effects of A- and B-Class ODN for broad applications in infectious disease or cancer therapy. Key words: CpG / Oligodeoxynucleotides / Immunostimulation / IFN- § / Toll-like receptor 9

1 Introduction

Received Revised Accepted

13/3/03 31/10/03 17/11/03

[DOI 10.1002/eji.200324032]

The bacterial genome, compared to vertebrate DNA, contains a higher frequency of unmethylated deoxycytidyl-deoxyguanosine (CpG) dinucleotides [6, 7]. Small oligodeoxynucleotides (ODN) with unmethylated CpG dinucleotides (CpG ODN) are able to perfectly mimic the immunostimulatory activity of bDNA [7]. Studies using TLR9-deficient (TLR9–/–) mice demonstrated that this TLR subtype is essential for the effects that are mediated by CpG DNA or ODN [1]. In addition, responsiveness to CpG ODN stimulation in vitro was reconstituted by transfection of unresponsive cells with TLR9 [8]. The human TLR9 (hTLR9) is expressed in B cells and plasmacytoid DC (pDC) [8, 9]; mice also express TLR9 (mTLR9) in the myeloid compartment. CpG ODN require internalization and endosomal maturation to activate TLR9 [10], and CpG ODN and TLR9 co-localize intracellularly in lysosomal compartments [11–13].

Abbreviations: bDNA: Bacterial DNA CpG: Deoxycytidyldeoxyguanosine HBsAg: Hepatitis B surface antigen ODN: Oligodeoxynucleotide pDC: Plasmacytoid DC TLR: Tolllike receptor

Certain sequence motifs strongly enhance immunostimulation. We have previously described an optimal human hexamer motif 5’-GTCGTT-3’ [14], and the opti-

Certain molecular structures that are present in pathogens (pathogen-associated molecular patterns or PAMP) are recognized by innate immune cells via pattern recognition receptors (PRR). The cells are activated upon recognition of PAMP and trigger the generation of optimal adaptive immune responses. The Toll-like receptors (TLR) are a large family of PRR consisting of ten different TLR subtypes. Recently, TLR were shown to be essential for the recognition of double-stranded RNA (TLR3), lipopolysaccharide (LPS; TLR4), small anti-viral compounds (TLR7 and TLR8) or bacterial DNA (bDNA; TLR9) [1–5].

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eji.de

252

J. Vollmer et al.

Eur. J. Immunol. 2004. 34: 251–262

mal mouse motif was reported to be 5’-GACGTT-3’ [7]. Different classes of CpG ODN were recently characterized. The B-Class (also known as K type) is a very potent Th1 adjuvant, has anti-tumor activity and stimulates strong B cell and NK cell activation or cytokine secretion [15, 16]. In contrast, the G-rich 3’ and 5’ ends of A-Class (also known as D type) ODN have phosphorothioatemodified ends and a phosphodiester palindromic center portion [17–19]. They are especially potent in activating human pDC to produce high amounts of IFN- § . We herein describe a third class of CpG ODN with intermediate immune effects to the B- and A-Class. These Cclass CpG ODN very strongly stimulate B cells as well as type I IFN secretion in vitro of human peripheral blood mononuclear cells (PBMC). In vivo, they are strong Th1inducing adjuvants. In addition, using TLR9–/– mouse splenocytes we demonstrate that all three CpG ODN classes depend on the presence of TLR9 to induce immunostimulation. In vitro, the different classes all induce TLR9-dependent NF- ‹ B stimulation, although with distinct dose-response relationships.

2 Results 2.1 Sequence requirements of C-Class CpG ODN that activate B cells and induce IFN- > production B-Class ODN induce strong stimulation of B cells but are relatively weak in IFN- § production. A-Class ODN induce high levels of type I IFN. We now describe ODN (e.g. 2395 in Fig. 1) with sequences and structures different from those of the A- or B-Classes that induce high levels of IFN- § as well as strong B cell stimulation. C-Class ODN (e.g. 2395) optimally consist of a stimulatory hexameric CpG motif (5’-TCGTCGTT-3’) linked by a T spacer to GC-rich palindromic sequences (5’-CGGCGCGCGCCG-3’). A wide range of modifications that maintain the palindrome are well tolerated, e.g. 2429 or 5397 (Table 1). In contrast, other palindromes such as CCCCCC (not shown), or destroying the palindrome as in 5709 or 5733, decreased IFN- § secretion (Table 1). Next we examined whether the activity of C-Class ODN was influenced by methylation or CpG inversion. An ODN with the same sequence as 2395 except that the cytosines were methylated lost all IFN- § -inducing activity (5327 in Fig. 1B). IFN- § secretion could be partially restored at a concentration of 1 ? M. No IFN- § secretion could be observed for an ODN with GC modification (5328). Typically, A-Class ODN are poor stimulators of human B cells [17]. In contrast, it turned out that C-Class ODN © 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Fig. 1. C-Class ODN induce strong IFN- § production and B cell stimulation. (A) Human PBMC were incubated for 48 h with B- (2006) or C- (2395) class ODN. SN were harvested and IFN- § determined by ELISA. Shown are means ± SEM for five donors. (B) IFN- § secretion by C-methylated (5327) or CG-inverted (5328) variants of 2395. Shown are means ± SEM for three different donors. (C) For B cell stimulation, human PBMC were cultured with the indicated ODN concentrations for 48 h. Results for expression of CD86 on CD19+ and CD14– B cells were acquired by flow cytometry. Mean ± SEM of the percentage of CD86-expressing B cells for three donors. (D) B cell proliferation induced by ODN 2006 and 2395 was measured by decreasing CFSE content. Shown are means ± SEM of the percentage of proliferating B cells for four donors.

efficiently stimulated B cells. Similar to B-Class ODN, the C-Class ODN induced strong up-regulation of B cell CD86 (Fig. 1C) and other cell surface molecules (not shown). In addition, C-Class ODN also stimulated efficient B cell proliferation (Fig. 1D). In contrast, a typical Aclass ODN, 2216, in similar experiments at these concentrations failed to induce efficient expression of B cell surface molecules and gave no significant B cell proliferation (not shown). In order to determine the separate activity of the two domains of the prototype 2395 C-Class ODN, PBMC were incubated with either 2395, the 2395 5’ sequence (5’-TCGTCGTTTT-3’; 5435), the 2395 3’ GC-rich palindrome (5’-CGGCGCGCGCCG-3’; 5373) or the combination of both the 5’ and 3’ ends. When separated from the 5’ motif, the 3’ palindrome by itself did not induce any IFN- § secretion. Significant stimulation could also not be observed when the two ODN were combined (data not shown), demonstrating that the full immune activity www.eji.de

Eur. J. Immunol. 2004. 34: 251–262

Three distinct immunostimulatory CpG ODN classes

253

Table 1. Effect of palindrome length and structurea)

a)

Human PBMC were cultured with ODN for 48 h. Given are the mean IFN- § amounts ± SEM. Palindromes are underlined.

requires physical linkage between the two domains. Table 2 summarizes the sequences of ODN used in this study.

2.2 Variability of effective palindromes in C-Class ODN Exchange of guanosines or cytosines in C-Class ODN by adenosines or thymidines did not influence their activity if the palindrome was maintained (e.g. 5393 in Table 1). In contrast, exchange of a single guanosine or cytosine that reduced the palindrome length to just six bases (5401 or 5514), decreased the ability to induce IFN- § secretion (Table 1). Within certain limits, the position of the palindrome did not influence IFN- § secretion. An ODN with the palindrome adjacent to the 5’ end (2451) stimulated high levels of IFN- § secretion if the 5’ end was a TCG, but this was dramatically reduced with other 5’ ends (e.g. 5421 in Table 1).

compare the activity to other CpG ODN classes. Fig. 2A shows an IFN- § dose response of typical B-Class (2006), C-Class (2395), A-Class (2216) and control ODN 2243 (2216 GC) or 1982 (non-CpG ODN). The B-Class ODN stimulated low IFN- § levels compared to the C- or AClass ODN. IFN- § induced by the A-Class ODN (2216; or 2336 not shown) exceeded that induced by the C-Class ODN. The control ODN or LPS did induce only background IFN- § levels. In experiments not shown C-Class ODN also stimulated efficient secretion of IFN- + .

2.3 In vitro characterization of C-Class molecules in human PBMC

All CpG ODN induced secretion of IL-6 (Fig. 2B). The maximal IL-6 levels obtained with 2395 exceeded those with 2006 (Fig. 2B). GC modification or methylation of 2395 led to loss of its potent ability to stimulate IL-6 secretion (not shown). IL-6 secretion induced by any CpG ODN did not reach levels observed with LPS. As recently reported for certain non-CpG ODN [20], ODN 1982 induced IL-6 production from B cells at high concentrations. Culture with the A-Class ODN (2216) also led to IL-6 secretion, although significant amounts were only obtained at 0.25 ? M or above, while the GC control was inactive.

We selected ODN 2395 as a representative of this class and performed further immunological assays in order to

IL-6 as well as IL-10 were previously demonstrated to be secreted by CpG-stimulated B cells [14, 21]. Human

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eji.de

254

J. Vollmer et al.

Eur. J. Immunol. 2004. 34: 251–262

Table 2. Sequences of ODN used in this study

*: phosphorothioate linkage; –: phosphodiester linkage; Z: 5’-methylcytosine. Palindromic sequences are underlined.

PBMC were incubated with 0.5 ? M of B-Class (2006), CClass (2429) or control (1982) ODN; and 41.4±11.2% (2006), 52.8±10.8% (2429) or 13.8±2.1% (1982; medium control: 2.9±0.9%) CD19-positive B cells produced IL10 (mean ± SEM of two donors; see also Sect. 4), demonstrating that C-Class ODN induce IL-10 production in human B cells. We further wanted to explore whether the effects on B cells were induced directly. B cells were isolated from human PBMC and cultured with increasing concentrations of B-Class ODN 2006 or C-Class ODN 2395 (Fig. 3A). Both ODN induced IL-6 production by purified B cells. In these experiments a control ODN (1982) gave some IL-6 secretion (not shown). The results demonstrate that human B cells respond directly to stimulation by C-Class ODN.

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

In addition, pDC were isolated from human PBMC, and the pDC-depleted population was compared to whole PBMC for the production of IFN- § upon culture with Bor C-Class ODN. The removal of pDC completely abrogated IFN- § induction (Fig. 3B), demonstrating that pDC are responsible for the IFN- § production. An important activity of CpG ODN, especially of the AClass, is the stimulation of NK cells [18]. We compared NK-mediated cytotoxicity induced by C- and B-Class ODN (2395 and 2006) in several experiments. Table 3 gives an example showing that the C-Class ODN were more efficient in mediating NK-dependent cytotoxicity than B-Class ODN.

www.eji.de

Eur. J. Immunol. 2004. 34: 251–262

Fig. 2. IFN- § and IL-6 secretion induced by the three major classes of CpG ODN. Human PBMC were cultured with BClass ODN (2006), A-Class ODN (2216), C-Class ODN (2395) and control ODN (2243 and 1982) or LPS for 24 h (IL6) or 48 h (IFN- § ). SN were harvested and IFN- § (A) or IL-6 (B) measured by ELISA. Shown are means ± SEM of cytokine secretion for three donors representative for several experiments comparing C-, B- and A-Classes. Not shown are the IL-6 levels obtained with 100 ng/ml LPS for the three donors: (1) 1789±15 pg/ml; (2) 1034±18 pg/ml; (3) 1760±1 pg/ml.

2.4 Decreased ability of C-Class ODN to reconstitute TLR9-mediated NF- k B activation Previous studies demonstrated that B-Class ODN stimulate the TLR9-dependent activation of NF- ‹ B in nonresponder HEK293 cells transfected with hTLR9 or mTLR9 [8, 13]. We wanted to explore whether the different CpG ODN classes all are able to induce TLR9dependent signaling in this experimental system. All CpG ODN gave signaling on the hTLR9 transfectants, although with quite distinct dose-response relationships (Fig. 4A). The B-Class ODN (2006) induced NF- ‹ B activation at very low concentrations and gave strong maximal stimulation. The C-Class ODN (2395) stimulated NF‹ B at higher concentrations and did not reach the same maximal activation. It took even a higher concentration for the A-Class ODN (2216) to activate NF- ‹ B.

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Three distinct immunostimulatory CpG ODN classes

255

Fig. 3. Purified B cells are stimulated by C-Class ODN and pDC are responsible for C-Class IFN- § production. (A) Isolated human B cells were cultured with B-Class ODN (2006), C-Class ODN (2395) or LPS (100 ng/ml) for 24 h, and IL-6 was measured by ELISA. Shown are means ± SEM of five different blood donors. (B) IFN- § production induced by Bor C-Class ODN (black bars) was compared to IFN- § production stimulated after removal of pDC from PBMC (open bars). Cells were cultured for 48 h and IFN- § in the SN measured by ELISA. Shown are means ± SEM of two representative out of four donors of two similar experiments.

Similar results were obtained with mTLR9 transfectants (Fig. 4B). Again, the A-Class ODN needed the highest ODN concentrations to be stimulatory. As previously reported [8], 2006 (optimal human B-Class ODN) acted relatively weak on mTLR9 due to species specificity. All ODN were inferior to the optimal mouse B-Class ODN 1826 (fold induction of luciferase activity at 0.156 ? M: 7.8±0.6 and at 10 ? M: 8.6±0.4). We also found strong inhibition of IFN- § secretion by culturing human PBMC plus C-Class ODN plus chloroquine (C-Class ODN at 1 ? M: 1427±531.32 pg/ml IFN- § ; plus 5 ? g/ml chloroquine: 98.6±4 pg/ml IFN- § ; medium alone: 7±5.6 pg/ml IFN- § ; mean ± SEM for two donors). We further explored the TLR9 dependency of immunostimulation induced by all three CpG ODN classes. TLR9–/– mice were reported to be unresponsive to stimulation by B- or A-Class ODN [1, 2]. We, therefore, compared their potential to stimulate splenocytes from wildtype versus TLR9–/– mice (Fig. 5). All CpG ODN did not stimulate cells from TLR9–/– mice. www.eji.de

256

J. Vollmer et al.

Eur. J. Immunol. 2004. 34: 251–262

Fig. 4. CpG ODN of all classes stimulate TLR9-dependent signaling. HEK293 cells stably transfected with a vector expressing either hTLR9 (A) or mTLR9 (B) and an NF ‹ B-luciferase construct were incubated for 16 h with B-Class ODN 2006 or GC control 2137, A-Class ODN 2216 or GC control 2243, or C-Class ODN 2395 or GC control 5328. NF ‹ B stimulation was measured through luciferase activity. Shown is one representative out of two experiments.

2.5 In vitro effects of C-Class ODN on mouse immune cells We further investigated the effects mediated by C-Class ODN on mouse cells using 1826 or 2006, 2216 (A-Class), 2395 (C-Class) plus controls, 2243 and 2137. The C- and B-Class ODN induced similar IL-6 levels; both were inferior to 1826 (Fig. 6). The A-Class was a very weak IL-6 inducer. Both control ODN appeared inactive. For IL-12 secretion, the A-Class compared well to the B-Class; the C-Class reached similar IL-12 levels obtained with 2006 at the high concentration (Fig. 5B). ODN 1826 was again superior, and the control ODN did not show any activity. The A-Class was the strongest inducer of IFN- + (Fig. 6C). In contrast, 2006 was relatively weak, but the C-Class

was stronger, and was at least comparable to 1826 at the higher concentrations. This suggests that mouse cells respond to C-Class ODN in the same fashion as human cells.

2.6 In vivo effects of C-Class ODN in mice We also carried out early in vivo characterization of the C-Class ODN. BALB/c mice received intramuscular injections at 0 and 4 weeks with hepatitis B surface antigen (HBsAg), alone or in combination with 10 ? g B-Class (2006), C-Class (2429; compare also Table 1) or control ODN 2137. ODN 2006 was used in these experiments because we wanted to perform an in vivo comparison

Table 3. Stimulation of NK cytotoxicity by C-Class ODNa)

a)

Human PBMC were cultured overnight with medium alone or with the indicated ODN or IL-2 concentration. Cells were harvested and incubated for 4 h with K562 target cells. L.U.: lytic units.

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eji.de

Eur. J. Immunol. 2004. 34: 251–262

Three distinct immunostimulatory CpG ODN classes

257

Fig. 5. Splenocytes from TLR9–/– mice do not respond to stimulation through all CpG ODN classes. Naive BALB/c (A, C) or TLR9–/– (B, D) mouse splenocytes were incubated with B-Class (2006), A-Class (2336) or C-Class ODN (2429) for 24 h. SN were harvested, and IL-6 (A, B) or IL-12 (C, D) measured by ELISA. Shown are means ± SD of two ELISA of one representative out of two experiments.

that would not be influenced by species-specific sequence differences (ODN 2006 and 2429 both with human 5’-GTCGTT-3’ motif). Mice were bled at 4 and 6 weeks post-prime and antibody titers were determined. Fig. 7A demonstrates that total anti-HB IgG antibodies were increased for both CpG ODN pre- as well as post-boost. Further tests revealed that 2429 induced a Th1-biased antibody response with increased IgG2a levels comparable to 2006 (Fig. 7B, C). Similar results were obtained 2 weeks post-boost. In contrast, mice injected with ODN 2137 with HBsAg had antibody titers that did not differ in a statistically significant way from those determined in mice immunized with HBsAg alone pre- or post-boost.

3 Discussion Synthetic phosphorothioate CpG ODN are very potent stimulators of immune cells [6, 22]. Two main CpG ODN classes have been described so far. Phosphorothioate B-Class ODN are very potent in stimulating B cells, but also stimulate pDC [14, 17]. In contrast, A-Class ODN induce extremely high amounts of type I IFN from pDC [17, 18] and require palindromic CpG phosphodiester sequences with phosphorothioate G-rich ends [17]. The present study describes ODN sequences that induce very strong IFN- § production, which was known before only for the A-Class [17, 19] (Fig. 1, 2). pDC are specifically required for the strong stimulation of IFN- § © 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Fig. 6. Naive mouse splenocytes respond to stimulation through all CpG ODN classes. BALB/c splenocytes were incubated with mouse optimal B-Class (1826), human optimal B-Class (2006), A-Class (2216), C-Class (2395) and control ODN (2243 and 2137). SN were harvested, and IL-6 (A), IL-12 (B) (both at 24 h) or IFN- + (C) (at 72 h) measured by ELISA. Shown are the results ± SD for one representative out of two experiments.

upon culture with C-Class ODN as demonstrated previously for A-Class ODN (Fig. 3B) [17]. The C-Class ODN lack the need for a chimeric phosphodiesterphosphorothioate backbone or poly-G ends. Such phosphorothioate ODN with a stimulatory CpG plus a palindromic sequence at the 5’ or 3’ end (Table 1) very efficiently induce effects that were previously described to be mediated either by B- or A-Class ODN. In contrast to A-Class ODN, which are relatively poor B cell stimulators [17, 19], C-Class ODN were very efficient in enhancing the expression of costimulatory molecules (Fig. 1C), inducing B cell IL-6 (Fig. 2B) or IL-10 production, as well as stimulating strong B cell proliferation (Fig. 1D). In addition, purified human B cells responded efficiently to stimulation by B- or C-Class ODN (Fig. 3A). This finding confirms the direct stimulatory effects of CClass ODN on TLR9-expressing B cells. www.eji.de

258

J. Vollmer et al.

Eur. J. Immunol. 2004. 34: 251–262

Fig. 7. C-Class ODN have strong Th1 adjuvant activity. BALB/c mice (n=5) were immunized with 1 ? g HBsAg alone (None) or 10 ? g of B-Class (2006), C-Class (2429) or control ODN (2137) by intramuscular injection. Animals were boosted at 4 weeks post-primary immunization. IgG levels were measured at boost or 2 weeks post-boost. (A) Total IgG (log scale; black bars at boost and open bars 2 weeks post-boost); (B) IgG1 titers; (C) IgG2a titers.

We also compared the different CpG classes for their immune effects on mouse splenocytes (Fig. 6). IL-6 secretion was induced very similar to the results observed on human PBMC or isolated human B cells (Fig. 2B, 3A). All CpG ODN (except 1826) stimulated comparable IL-12 production. The C-Class appeared to induce more efficient IFN- + secretion than the B-Class (2006). These data suggest that, comparable to human cells, the C-Class is as efficient if not better than the BClass to induce mouse splenocyte stimulation and Th1 cytokine production. The differences observed between the CpG ODN classes might be explained by previous observations that IL-6 or IL-12 are directly stimulated by CpG ODN in mTLR9-expressing cells whereas IFN- + depends on IFN- § to be produced by NK cells [18]. The high IFN- § production induced by A- and C-Class ODN very efficiently mediated NK lytic activity (Table 3), much more than seen with the B-Class. These results support therapeutic applications of C-Class ODN in the treatment of diseases in which NK activity contributes to host defense. In addition, we also observed that C-Class ODN led to down-regulation of CD14 and up-regulation of CD80 on monocytes (data not shown). This characterizes the transition from monocytes into myeloid DC and indicates that the C-Class promotes the maturation of monocytes into functional DC that will be able to efficiently support antigen-specific cellular responses. As previously reported for the B- and A-Classes, C-Class ODN show specific sequence requirements. The palindrome is needed for strong IFN- § induction, since destroying or shortening it resulted in severely decreased IFN- § production, whereas most other modifications that retained the palindrome were tolerated (Table 1). Optimal C-Class ODN contain a 5’-TCGTCGTT3’ sequence or at least a 5’ TCG, which, when included in a B-Class ODN such as 2006, is very efficient in B cell immunostimulation [16]. GC modification or cytosine methylation strongly decreased IFN- § production (Fig. 1) and B cell stimulation. Our results also suggest that while the palindrome is necessary, it is not sufficient on © 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

its own for strong IFN- § secretion, and a stimulatory CpG motif is essential. Support for this comes from results showing that the stimulatory CpG or palindrome sequence alone or combined in trans were not able to reconstitute stimulation mediated by the full-length CClass ODN where these two elements are found in cis on the same ODN. Surprisingly, the position of the palindrome adjacent to either the 3’ or 5’ end of the ODN had very little influence, if any, on the activity, if the very 5’ end of the ODN had a stimulatory motif (Table 1). There has been much interest in the role of TLR9 in mediating the effects of different ODN classes. Previous studies demonstrated that bDNA and B- or A-Class ODN fail to stimulate cells from TLR9–/– mice [1, 2, 23]. We now extend these observations by demonstrating that all known CpG ODN classes depend on the presence of TLR9, since splenocytes from TLR9–/– mice were nonresponsive to stimulation with any of the three classes (Fig. 5). In addition, stimulation by all classes can be blocked by chloroquine, an agent interfering with endosomal maturation and completely blocking CpGmediated immunostimulation [10]. This indicates that despite the strikingly different immune effects of the three ODN classes, they share at least part of a common signaling pathway that appears to require an early interaction with endosomal TLR9. Although TLR9 is therefore required for mediating the immunostimulatory effects of all three ODN classes, it does not necessarily follow that TLR9 will be sufficient to reconstitute CpG responsiveness in non-expressing cells. Previous work established that TLR9 is sufficient for restoring CpG-mediated signaling by B-Class ODN [8]. A-Class ODN were described to be non-stimulatory on TLR9 transfectants, indicating that TLR9 may not be sufficient to restore responsiveness to the A-Class [24]. We performed a careful comparison between all three ODN classes on TLR9-expressing HEK293 cells. All ODN classes stimulated hTLR9- and mTLR9-dependent NF‹ B signaling, although with significant differences in the dose responses (Fig. 4). B-Class ODN were strongly www.eji.de

Eur. J. Immunol. 2004. 34: 251–262

stimulatory, and we confirmed previously reported differences between optimal human and mouse CpG ODN. Both A- and C-Class ODN induced weaker NF- ‹ B signaling at high ODN concentrations. These results indicate that TLR9 is sufficient to at least partially restore CpG-induced NF- ‹ B signaling and support its role in mediating CpG responses in TLR9-expressing cells. The observed discrepancy between the relatively weak TLR9-dependent NF- ‹ B stimulation and the strong IFN§ secretion induced with A- and C-Class ODN might imply the existence of another mechanism independent of NF- ‹ B that mediates their efficient type I IFN stimulation. One possible explanation could be the existence of yet unidentified TLR adaptor or regulatory molecules, that might be missing in the HEK293 cells used for the TLR9 reconstitution assays, and that would be required for A- and C-Class, but not for B-Class ODN. For example, myeloid differentiation factor 88 (MyD88) is involved in mediating at least some of the signaling events from all known TLR, while Toll-IL-1R domain-containing adaptor protein (TIRAP) and Toll-IL1R domain-containing adaptor inducing IFN- g (TRIF) are involved in certain other TLR signaling pathways, without an apparent role in mediating TLR9-dependent signaling known so far [25–28]. In addition, negative regulators of the TLR-mediated signaling cascade such as suppressors of cytokine signaling (SOCS), extracellular signal-regulated kinases (ERK) or IL-1R-associated kinase (IRAK)-M could in principle modify strength and specific outcome of TLR-mediated cellular activation [29–31]. Clearly, extensive further studies will be required to distinguish these possibilities. Outstanding responses to antigen have been observed in a variety of studies using B-Class ODN as adjuvant [16, 32–34]. Our experiments demonstrate that the CClass ODN, in contrast to the A-Class ODN [6], act as very potent Th1-promoting adjuvants. Mice immunized with recombinant HBsAg plus C-Class ODN had a several times higher antibody response, which was predominantly of the IgG2a isotype and, therefore, Th1dominated (Fig. 7). A- and B-Class ODN might be optimal in the immune therapy of different tumor types but the combination of the immune effects of both classes might have additional effects [35]. Therefore, the C-Class as an intermediate of the A- and B-Classes might have even broader immune therapeutic effects in different tumor types or infectious diseases where a combination of A- and B-Class ODN would be highly desirable. Human clinical trials of CClass ODN are expected to start in 2003.

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Three distinct immunostimulatory CpG ODN classes

259

4 Materials and methods 4.1 Oligodeoxynucleotides All ODN were provided by Coley Pharmaceutical Group (Langenfeld, Germany) and had undetectable endotoxin levels ( X 0.1 EU/ml) measured by the Limulus assay (BioWhittaker, Verviers, Belgium). ODN were suspended in sterile, endotoxin-free Tris-EDTA (Sigma, Deisenhofen, Germany), and stored and handled under aseptic conditions to prevent contamination. All dilutions were carried out using pyrogenfree phosphate-buffered saline (PBS; Life Technologies, Eggenstein, Germany). 4.2 TLR9 assay Stably transfected HEK293 cells expressing hTLR9 or mTLR9 were described before [8] and were a gift from Dr. S. Bauer, University of Munich, Germany. Briefly, HEK293 cells were transfected by electroporation with vectors expressing the hTLR9 or mTLR9 and a 6×NF ‹ B-luciferase reporter plasmid. Stable transfectants (3×104 cells/well) were incubated with ODN for 16 h at 37°C in a humidified incubator. Each data point was performed in triplicate. Cells were lysed and assayed for luciferase gene activity (using the BriteLite kit from Perkin-Elmer, Zaventem, Belgium). Stimulation indices were calculated in reference to reporter gene activity of medium without addition of ODN. 4.3 Cell purification Peripheral blood buffy coat preparations from healthy human donors were obtained from the Blood Bank of the University of Düsseldorf (Germany) and PBMC were purified by centrifugation over Ficoll-Hypaque (Sigma). Cells were cultured in a humidified incubator at 37°C in RPMI 1640 medium supplemented with 5% (v/v) heat-inactivated human AB serum (BioWhittaker) or 10% (v/v) heatinactivated FCS, 2 mM L-glutamine, 100 U/ml penicillin and 100 ? g/ml streptomycin (all from Sigma). 4.4 Cytokine detection PBMC were re-suspended at concentrations of 3×106 to 5×106 cells/ml and added to 48-well flat-bottomed plates (1 ml/well) or 96-well round-bottomed plates (200 ? l/well), and cultured with ODN at different concentrations. In some experiments, LPS (Sigma) was used as control. Culture supernatants (SN) were collected after the indicated time points. If not used immediately, SN were frozen at –20°C until required. Amounts of cytokines in the SN were assessed using commercially available ELISA kits (IL-6, IFN+ or TNF- § ; all from Diaclone, Besancon, ¸ France) or an inhouse ELISA developed using commercially available antibodies (PBL, New Brunswick, NJ; for detection of multiple IFN- § species). www.eji.de

260

J. Vollmer et al.

4.5 Measurement of NK-mediated cytotoxicity For measurement of human NK lytic activity, PBMC were incubated at 5×106 cells/well in 24-well plates with or without ODN; IL-2 (Life Technologies) was added as a control. Cultures were harvested after 24 h, and cells were used as effectors in a standard 4-h 51Cr-release assay against K562 target cells, as previously described [20]. 4.6 Cultures for flow cytometric analysis of PBMC stimulation All monoclonal antibodies (mAb) were purchased from Becton Dickinson (Heidelberg, Germany), except mAb to CD11c from DAKO (Hamburg, Germany) and to blood DC antigen (BDCA)-4 from Miltenyi (Bergisch-Gladbach, Germany). Briefly, PBMC (4×106 cells/ml) were incubated for 48 h with or without ODN. B cells were identified by CD19 expression on CD14– cells. pDC were identified by CD123high, HLADRhigh expression on CD11clow-expressing cells. Flow cytometric data were acquired on a FACSCalibur and were analyzed using the computer program CellQuest (both from Becton Dickinson). For B cell proliferation assays, CD19+ B cells were identified after culturing 5- (and 6-) carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled PBMC with CpG ODN for 5 days by flow cytometry as described before [14]. Decreased CFSE content indicated proliferating B cells. For detection of IL-10-secreting B cells, PBMC were cultured with CpG ODN and incubated with IL-10 catch and detection reagent and mAb to CD19 as described by the manufacturer (Miltenyi), and data acquired by flow cytometry. 4.7 Isolation of human B cells or pDC B cells were isolated from PBMC with the CD19 B cell isolation kit as described by the manufacturer (Miltenyi). To determine purity, cells were stained with mAb to CD20 and CD14 and identified by flow cytometry. In all experiments B cells were more than 95% pure. Purified B cells (2×105 to 5×105 cells/ml) were incubated with increasing concentrations of ODN for 24 h and IL-6 was measured as described above. For the culture of pDC-depleted PBMC, pDC were isolated with the BDCA-4 pDC isolation kit as described by the manufacturer (Miltenyi). pDC isolation was confirmed by staining with mAb to CD123, CD11c and HLA-DR or BDCA-4, and PBMC lacking pDC (4×106 cells/ml) cultured for 48 h with the indicated ODN concentrations. IFN- § in the SN was measured as described above.

Eur. J. Immunol. 2004. 34: 251–262 (Quebec, Canada). TLR9–/– mice were described before [1] and obtained from Dr. Shizuo Akira (Osaka University, Japan). The animals were backcrossed to BALB/c for three generations.

4.9 In vitro mouse assays Naive BALB/c or TLR9–/– mouse splenocytes were used for all in vitro assays. Animals were anesthetized with isofluorane and killed by cervical dislocation. Spleens were removed under aseptic conditions and placed in PBS + 0.2% BSA (Sigma, St. Louis, MO). Spleens were then homogenized and splenocytes were re-suspended in RPMI 1640 (Life Technologies) medium supplemented with 2% normal mouse serum (Cedarlane Laboratories, Canada), 2 mM L-glutamine, penicillin-streptomycin solution (final concentration of 1,000 U/ml and 1 mg/ml, respectively), and 5×10–5 M g -mercaptoethanol (all from Sigma). Splenocytes were plated in 96-well round-bottomed plates (5×106 cells/ ml) and cultured in quadruplicate in a humidified 5% CO2 incubator at 37°C for 24 h (for IL-6 or IL-12) or 72 h (for IFN+ ). SN were harvested and cytokine levels measured (mouse OptEIA kits; PharMingen, Mississauga, Canada).

4.10 Immunization of mice BALB/c mice (n=5/group) were immunized with 1 ? g HBsAg subtype ad (International Enzymes, CA) alone or in combination with 10 ? g CpG ODN in saline. Animals were bled and boosted at 4 weeks post-primary immunization. Two weeks post-boost, animals were again bled to determine antibody titers. Anti-HB antibodies (total IgG, IgG1 and IgG2a) were detected and quantified by end-point dilution ELISA assay, which was performed in triplicate on samples from individual animals. End-point titers were defined as the highest plasma dilution that resulted in an absorbance value (optical density 450 nm) two times greater than that of nonimmune plasma, with a cut-off value of 0.05. These were reported as group mean titers ± SEM.

Acknowledgements: We would like to thank Drs. Eugen Uhlmann and Grayson Lipford for helpful discussions and Andrea Janosch for perfect technical assistance. We also thank Dr. Shizuo Akira for providing TLR9–/– mice, Dr. Stefan Bauer for providing TLR9-transfected HEK293 cells and Dr. Zuhair Ballas (University of Iowa, IA) for providing data on NK-mediated cytotoxicity.

References 4.8 Animals Female BALB/c mice (6–8 weeks of age) were used for all experiments and purchased from Charles River Canada © 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1 Hemmi, H., Takeuchi, O., Kawai, T., Kaisho, T., Sato, S., Sanjo, H., Matsumoto, M., Hoshino, K., Wagner, H., Takeda, K. and Akira, S., A Toll-like receptor recognizes bacterial DNA. Nature 2000. 408: 740–745.

www.eji.de

Eur. J. Immunol. 2004. 34: 251–262

Three distinct immunostimulatory CpG ODN classes

261

2 Hemmi, H., Kaisho, T., Takeuchi, O., Sato, S., Sanjo, H., Hoshino, K., Horiuchi, T., Tomizawa, H., Takeda, K. and Akira, S., Small anti-viral compounds activate immune cells via the TLR7 MyD88-dependent signaling pathway. Nat. Immunol. 2002. 3: 196–200.

17 Krug, A., Rothenfusser, S., Hornung, V., Jahrsdorfer, B., Blackwell, S., Ballas, Z. K., Endres, S., Krieg, A. M. and Hartmann, G., Identification of CpG oligonucleotide sequences with high induction of IFN-alpha/beta in plasmacytoid dendritic cells. Eur. J. Immunol. 2001. 31: 2154–2163.

3 Jurk, M., Heil, F., Vollmer, J., Schetter, C., Krieg, A. M., Wagner, H., Lipford, G. and Bauer, S., Human TLR7 or TLR8 independently confer responsiveness to the antiviral compound R848. Nat. Immunol. 2002. 3: 499.

18 Ballas, Z. K., Rasmussen, W. L. and Krieg, A. M., Induction of NK activity in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial DNA. J. Immunol. 1996. 157: 1840–1845.

4 Alexopoulou, L., Holt, A. C., Medzhitov, R. and Flavell, R. A., Recognition of double-stranded RNA and activation of NFkappaB by Toll-like receptor 3. Nature 2001. 413: 732–738.

19 Verthelyi, D., Ishii, K. J., Gursel, M., Takeshita, F. and Klinman, D. M., Human peripheral blood cells differentially recognize and respond to two distinct CPG motifs. J. Immunol. 2001. 166: 2372–2377.

5 Poltorak, A., He, X., Smirnova, I., Liu, M. Y., Huffel, C. V., Du, X., Birdwell, D., Alejos, E., Silva, M., Galanos, C., Freudenberg, M., Ricciardi-Castagnoli, P., Layton, B. and Beutler, B., Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 1998. 282: 2085–2088. 6 Krieg, A. M., CpG motifs in bacterial DNA and their immune effects. Annu. Rev. Immunol. 2002. 20: 709–760. 7 Krieg, A. M., Yi, A. K., Matson, S., Waldschmidt, T. J., Bishop, G. A., Teasdale, R., Koretzky, G. A. and Klinman, D. M., CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 1995. 374: 546–549. 8 Bauer, S., Kirschning, C. J., Hacker, H., Redecke, V., Hausmann, S., Akira, S., Wagner, H. and Lipford, G. B., Human TLR9 confers responsiveness to bacterial DNA via species-specific CpG motif recognition. Proc. Natl. Acad. Sci. USA 2001. 98: 9237–9242. 9 Hornung, V., Rothenfusser, S., Britsch, S., Krug, A., Jahrsdorfer, B., Giese, T., Endres, S. and Hartmann, G., Quantitative expression of Toll-like receptor 1–10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides. J. Immunol. 2002. 168: 4531–4537. 10 Manzel, L., Strekowski, L., Ismail, F. M., Smith, J. C. and Macfarlane, D. E., Antagonism of immunostimulatory CpGoligodeoxynucleotides by 4-aminoquinolines and other weak bases: mechanistic studies. J. Pharmacol. Exp. Ther. 1999. 291: 1337–1347. 11 Ahmad-Nejad, P., Hacker, H., Rutz, M., Bauer, S., Vabulas, R. M. and Wagner, H., Bacterial CpG-DNA and lipopolysaccharides activate Toll-like receptors at distinct cellular compartments. Eur. J. Immunol. 2002. 32: 1958–1968. 12 Manzel, L. and Macfarlane, D. E., Lack of immune stimulation by immobilized CpG-oligodeoxynucleotide. Antisense Nucleic Acid Drug Dev. 1999. 9: 459–464. 13 Takeshita, F., Leifer, C. A., Gursel, I., Ishii, K. J., Takeshita, S., Gursel, M. and Klinman, D. M., Cutting edge: role of Toll-like receptor 9 in CpG DNA-induced activation of human cells. J. Immunol. 2001. 167: 3555–3558. 14 Hartmann, G. and Krieg, A. M., Mechanism and function of a newly identified CpG DNA motif in human primary B cells. J. Immunol. 2000. 164: 944–953. 15 Gursel, M., Verthelyi, D., Gursel, I., Ishii, K. J. and Klinman, D. M., Differential and competitive activation of human immune cells by distinct classes of CpG oligodeoxynucleotide. J. Leukoc. Biol. 2002. 71: 813–820. 16 Hartmann, G., Weeratna, R. D., Ballas, Z. K., Payette, P., Blackwell, S., Suparto, I., Rasmussen, W. L., Waldschmidt, M., Sajuthi, D., Purcell, R. H., Davis, H. L. and Krieg, A. M., Delineation of a CpG phosphorothioate oligodeoxynucleotide for activating primate immune responses in vitro and in vivo. J. Immunol. 2000. 164: 1617–1624.

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

20 Vollmer, J., Janosch, A., Laucht, M., Ballas, Z. K., Schetter, C. and Krieg, A. M., Highly immunostimulatory CpG-free oligodeoxynucleotides for activation of human leukocytes. Antisense Nucleic Acid Drug Dev. 2002. 12: 165–175. 21 Redford, T. W., Yi, A. K., Ward, C. T. and Krieg, A. M., Cyclosporin A enhances IL-12 production by CpG motifs in bacterial DNA and synthetic oligodeoxynucleotides. J. Immunol. 1998. 161: 3930–3935. 22 Uhlmann, E. and Vollmer, J., Recent advances in the development of immunostimulatory oligonucleotides. Curr. Opin. Drug Discov. Devel. 2003. 6: 204–217. 23 Hemmi, H., Kaisho, T., Takeda, K. and Akira, S., The roles of Toll-like receptor 9, MyD88, and DNA-dependent protein kinase catalytic subunit in the effects of two distinct CpG DNAs on dendritic cell subsets. J. Immunol. 2003. 170: 3059–3064. 24 Klinman, D. M., Takeshita, F., Gursel, I., Leifer, C., Ishii, K. J., Verthelyi, D. and Gursel, M., CpG DNA: recognition by and activation of monocytes. Microbes Infect. 2002. 4: 897–901. 25 Hacker, H., Vabulas, R. M., Takeuchi, O., Hoshino, K., Akira, S. and Wagner, H., Immune cell activation by bacterial CpG-DNA through myeloid differentiation marker 88 and tumor necrosis factor receptor-associated factor (TRAF)6. J. Exp. Med. 2000. 192: 595–600. 26 Schnare, M., Holt, A. C., Takeda, K., Akira, S. and Medzhitov, R., Recognition of CpG DNA is mediated by signaling pathways dependent on the adaptor protein MyD88. Curr. Biol. 2000. 10: 1139–1142. 27 Horng, T., Barton, G. M., Flavell, R. A. and Medzhitov, R., The adaptor molecule TIRAP provides signalling specificity for Tolllike receptors. Nature 2002. 420: 329–333. 28 Yamamoto, M., Sato, S., Mori, K., Hoshino, K., Takeuchi, O., Takeda, K. and Akira, S., Cutting edge: a novel Toll/IL-1 receptor domain-containing adapter that preferentially activates the IFNbeta promoter in the Toll-like receptor signaling. J. Immunol. 2002. 169: 6668–6672. 29 Dalpke, A. H., Opper, S., Zimmermann, S. and Heeg, K., Suppressors of cytokine signaling (SOCS)-1 and SOCS-3 are induced by CpG-DNA and modulate cytokine responses in APCs. J. Immunol. 2001. 166: 7082–7089. 30 Hacker, H., Mischak, H., Hacker, G., Eser, S., Prenzel, N., Ullrich, A. and Wagner, H., Cell type-specific activation of mitogen-activated protein kinases by CpG-DNA controls interleukin-12 release from antigen-presenting cells. EMBO J. 1999. 18: 6973–6982. 31 Kobayashi, K., Hernandez, L. D., Galan, J. E., Janeway, C. A., Jr., Medzhitov, R. and Flavell, R. A., IRAK-M is a negative regulator of Toll-like receptor signaling. Cell 2002. 110: 191–202.

www.eji.de

262

J. Vollmer et al.

32 Davis, H. L., Use of CpG DNA for enhancing specific immune responses. Curr. Top. Microbiol. Immunol. 2000. 247: 171–183. 33 Davis, H. L., Weeratna, R., Waldschmidt, T. J., Tygrett, L., Schorr, J., Krieg, A. M. and Weeranta, R., CpG DNA is a potent enhancer of specific immunity in mice immunized with recombinant hepatitis B surface antigen. J. Immunol. 1998. 160: 870–876. 34 Verthelyi, D., Kenney, R. T., Seder, R. A., Gam, A. A., Friedag, B. and Klinman, D. M., CpG oligodeoxynucleotides as vaccine adjuvants in primates. J. Immunol. 2002. 168: 1659–1663. 35 Blazar, B. R., Krieg, A. M. and Taylor, P. A., Synthetic unmethylated cytosine-phosphate-guanosine oligodeoxynucleotides are potent stimulators of antileukemia responses in naive and bone marrow transplant recipients. Blood 2001. 98: 1217–1225.

Eur. J. Immunol. 2004. 34: 251–262 Note added in proof: Two other reports additionally described CpG ODN sequences with activities similar to the here-characterized C-Class ODN: Hartmann, G., Battiany, J., Poeck, H., Wagner, M., Kerkmann, M., Lubenow, N., Rothenfusser, S. and Endres, S., Rational design of new CpG olignucleotides that combine B cell activation with high IFN-alpha induction in plasmacytoid dendritic cells. Eur. J. Immunol. 2004. 33: 1633–1641 and Marshall, J.D., Fearon, K., Abbate, C., Subramanian, S., Yee, P., Gregorio, J., Coffman, R.L. and Van Nest, G., Identification of a novel CpG DNA class and motif that optimally stimulate B cell and plasmacytoid dendritic cell funtions. J. Leukoc. Biol. 2003. 73: 781–792.

Correspondence: Jörg Vollmer, Coley Pharmaceutical GmbH, Elisabeth-Selbert-Str. 9, D-40764 Langenfeld, Germany Fax: +49-2173-3997-2399 e-mail: jvollmer — coleypharma.com

© 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eji.de