Conjugation of immunostimulatory DNA to the short ragweed allergen Amb a 1 enhances its immunogenicity and reduces its allergenicity

Conjugat ion of im m unost im ulat ory D N A t o t he short ragw eed allergen A m b a 1 enhances it s im m unogenicit y and reduces it s allergenicit y Helen Tighe, PhD,a Kenji Takabayashi, PhD,a David Schwartz, PhD,c Gary Van Nest, PhD,c Stephen Tuck, PhD,c Joseph J. Eiden, M D, PhD,c Anne Kagey-Sobotka, M D,d Peter S. Creticos, M D,d Law rence M . Lichtenstein, M D, PhD,d Hans L. Spiegelberg, M D,b and Eyal Raz, M Da La Jolla and Berkeley, Calif, and Baltimore, Md Background: Allergen immunotherapy is inconvenient and associated with the risk of anaphylaxis. Efforts to improve the safety of immunotherapy by means of chemical modification of allergens have not been successful because it greatly reduced their antigenicity. Recently, immunostimulatory DNA sequences (ISS or CpG motifs) have been shown to act as strong TH1 response–inducing adjuvants. Objective: We sought to determine whether conjugation of ISS to the major short ragweed allergen Amb a 1 results in enhanced immunotherapeutic potential in mice and decreased allergenicity in human subjects. Methods: A 22-mer ISS oligodeoxynucleotide (ISS-ODN) was coupled to Amb a 1 and used for immunization of mice, rabbits, and monkeys. Results: In mice the Amb a 1-ISS conjugate induced a TH1 response (IFN-γ secretion), whereas Amb a 1 induced a TH2 response (IL-5 secretion). The TH1 response was not observed with an Amb a 1-non-ISS conjugate. Coinjection of Amb a 1 with ISS-ODN was much less effective in inducing a TH1 response. In mice primed for a TH2 response, injection with Amb a 1-ISS conjugate induced a de novo TH1 response and suppressed IgE antibody formation after challenge with Amb a 1. Amb a 1-ISS conjugate induced high-titer anti-Amb a 1 IgG antibodies in rabbits and cynomolgus monkeys, whereas Amb a 1 alone or Amb a 1 coinjected with ISS-ODN did not induce a detectable response. Amb a 1-ISS conjugate was less allergenic than Amb a 1 alone, as shown by a 30-fold lower histamine release from human basophils of patients with ragweed allergy, whereas mixing ISS-ODN with Amb a 1 did not reduce histamine release. Conclusion: Amb a 1-ISS conjugate has an enhanced TH1biased immunogenicity and reduced allergenicity. It may offer From athe Department of Medicine and The Sam and Rose Stein Institute for Research on Aging, and bthe Department of Pediatrics, University of California San Diego School of Medicine, La Jolla; cDynavax Technologies Corporation, Berkeley; and dthe Department of Medicine, Johns Hopkins University Asthma and Allergy Center, Baltimore. Supported in part by grant AI40682 from the National Institutes of Health. Received for publication Feb 4, 2000; revised Apr 7, 2000; accepted for publication Apr 13, 2000. Reprint requests: Eyal Raz, MD, Associate Professor, Department of Medicine, University of California San Diego School of Medicine, 9500 Gilman Dr, La Jolla, CA 92093-0663. Copyright © 2000 by Mosby, Inc. 0091-6749/2000 $12.00 + 0 1/1/107927 doi:10.1067/mai.2000.107927 124 a more effective and safer approach for allergen immunotherapy than currently available methods. (J Allergy Clin Immunol 2000;106:124-34.) Key words: Allergen-ISS conjugate, immunotherapy, immunostimulatory DNA sequence, CpG motifs, basophil histamine release, ragweed allergy Although the efficacy of allergen immunotherapy has been proven in placebo-controlled clinical trials for allergic rhinitis and asthma, the risk of anaphylaxis, the inconvenience and discomfort of frequent dosing, and the duration of several years of therapy are all factors that limit efficacy.1-3 Attempts at improving immunotherapy by reducing allergenicity have been paralleled by reduced immunogenicity4,5 and did not result in an effective form of immunotherapy. Recent understanding of the immunologic basis of allergic diseases and the mechanisms of immunotherapy revealed that immunotherapy changes the TH2 response to allergens6-9 to a TH1 response,10-13 and this is likely one of the mechanisms responsible for the beneficial and long-lasting effects. In contrast, the commonly used symptomatic treatment of allergic disorders does not have such prolonged immunomodulatory effects. The advantage of immunotherapy compared with symptomatic allergy treatments would be further enhanced by development of an immunotherapy method that could more rapidly and more strongly induce an antiallergen TH1 response and reduce its allergenicity. The ability to bias an immune response toward TH1 has recently been demonstrated with bacterial DNA or immunostimulatory oligodeoxynucleotides (ISS-ODN), both of which contain CpG motifs.14-17 ISSs cause macrophages and other antigen-presenting cells (APCs) to secrete type 1 cytokines, such as IFNs, IL-12, and IL-18, which cause naive T cells to differentiate into TH1 cells.15 Coimmunization with a mixture of antigens and ISSs elicits a TH1 response15-17 and also has been shown to be effective in animal models relevant to human allergic diseases.18-22 Furthermore, antigen-ISS conjugates formed by an avidin-biotin bridge or chemical conjugation were reported to be very immunogenic.23,24 Based on the crucial role of APCs in the above-mentioned system, we hypothesized that by chemical conjugation of ISS-ODN Tighe et al 125 J ALLERGY CLIN IM M UNOL VOLUM E 106, NUM BER 1, PART 1 Abbreviations used APC: Antigen-presenting cell ISS: Immunostimulatory DNA sequence ISS-ODN: Immunostimulatory oligodeoxynucleotide mODN: Mutated (non-ISS) oligodeoxynucleotide RW: Short ragweed to an allergen, we would deliver the allergen and ISSODN to the same APC and thus minimize the dose of ISSODN and amplify the subsequent TH1 response. To test this approach, ISS-ODN was chemically conjugated to the major short ragweed (RW) allergen Amb a 1 (Amb a 1ISS conjugate) and evaluated for induction of a primary antiallergen response and effects on preexisting TH2 responses in an animal model. The Amb a 1-ISS conjugates were also tested for reduction of allergenicity by reaction with human IgE antibodies in an ELISA test and by reduction of histamine release from basophils of patients with RW allergy. M ETHODS Purification of Amb a 1 Defatted RW pollen (Ambrosia artemisiifolia) was purchased from Greer Laboratories, Inc (Lenoir, NC). The Amb a 1 was isolated by using a modified method, as described by King et al.24,25 After DEAE sepharose fast-flow anion exchange chromatography and Butyl sepharose fast-flow hydrophobic interaction chromatography, the purified Amb a 1 was concentrated by means of diafiltration to 5 mg/mL. Endotoxin levels were tested with a Limulus amoebocyte lysate analysis kit (BioWhittaker, Walkersville, Md) and found to be less than 20 ng/mL. Oligodeoxynucleotides The ISS 5´-TGACTGTGAACGTTCGAGATGA phosphorothioate ODN and the mutated (non-ISS) 5´-TGACTGTGAACCTTAGAGATGA phosphorothioate ODN (mODN) were purchased from either Trilink BioTechnologies (San Diego, Calif) or Hybridon Specialty Products (Milford, Mass). Endotoxin levels were less than 20 ng/mL. Preparation of Amb a 1-ISS conjugate The purified Amb a 1 was treated with N-ethylmaleimide and activated with sulfosuccinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate. Residual reagents were removed by chromatography on a G-25 desalting column. 5´-Disulfide ISS-ODN was reduced with tris(2-carboxyethyl)phosphine, residual reagents were removed with chromatography on a G-25 desalting column, and the resulting 5´-Thio ISS-ODN was mixed with the activated Amb a 1 to produce the Amb a 1-ISS conjugate. The conjugate was purified by Superdex HR 200 Gel Filtration chromatography. Nonreducing SDS-PAGE, combined with Coomassie blue and DNA-specific silver staining, was used as a qualitative measure of composition of the conjugate. The method used a Novex 8% to 16% TRIS-glycine minigel at a constant current of 25 mA per gel. Conjugate bands were visualized with either Coomassie brilliant blue G-250 stain or DNA silver stain (Pharmacia DNA Silver Staining Kit). The average number of ODNs per Amb a 1 in the conjugate was estimated by the ratio of the molar ODN content divided by the molar protein content. The ISS-ODN content was determined by using A260, and the protein content was determined by using the BCA assay. The extinction coefficient at 260 nm for 5´ ODNs was determined to be 25.6 mg·mL–1cm–1. Endotoxin levels were less than 20 ng/mL, resulting in less than 1 ng in the 50 µL injected per mouse. Immunization Groups of 4 to 10, 4- to 6-week-old female BALB/c mice were rested a minimum of 3 days after receipt (Jackson Laboratories, Bar Harbor, Me) and were injected intradermally at the base of the tail with 50 µL of saline containing either (1) 10 µg of Amb a 1 alone, (2) a mixture of 10 µg of Amb a 1 and 7.9 µg of ISS-ODN, or (3) an Amb a 1-ISS conjugate (containing 10 µg of Amb a 1 and 7.9 µg of ISSODN). In some experiments Amb a 1 was also injected intraperitoneally mixed with 3 mg of alum in 500 µL of saline solution. Mice were bled from the retro-orbital plexus after achievement of anesthesia, and serum was separated by means of centrifugation and stored at –20°C before being analyzed. Three days before being killed, the mice were injected intravenously with 5 µg of Amb a 1 to increase the frequency of antigen-specific T cells for subsequent in vitro assays.15 Groups of 5 New Zealand White rabbits were injected subcutaneously with either (1) 10 µg of Amb a 1, (2) 10 µg of Amb a 1 mixed with 50 µg of ISS-ODN, (3) 10 µg of Amb a 1 mixed with 500 µg of ISS-ODN, or (4) Amb a 1-ISS conjugate (containing 10 µg of Amb a 1 and 7.9 µg of ISS-ODN) each in saline solution. Rabbit sera were obtained before immunization and 2 weeks after each immunization and stored frozen at –20°C. Groups of 4 cynomolgus monkeys were injected subcutaneously in the thigh at weeks 0, 4, and 12 with 50 µg of Amb a 1 or with Amb a 1-ISS conjugate (containing 39 µg of ISS-ODN conjugated to 50 µg of Amb a 1) each in saline solution. Sera were obtained before immunization and 2 weeks after each immunization and stored frozen at –20°C. All animal experiments were approved by the appropriate committees on animal experimentation at the University of California San Diego, the Sidney Kimmel Cancer Center, or Sierra Biomedical, Inc (Sparks, Nev). Approval for rabbit experiments was given by Babco (Berkeley, Calif). Antibody analyses Mouse sera were analyzed for IgG1 and IgG2a anti-Amb a 1 antibodies by using an ELISA, as previously described.26 The values of the test sera were calculated from a standard serum pool that was previously titrated. This standard serum pool was assigned a value in units per milliliter that was equal to the reciprocal of the highest dilution that gave an OD reading double that of the background. IgE anti-Amb a 1 antibody titers of the mice were determined by using a modified ELISA, as previously described.26 The modification consisted of first absorbing the serum with protein G to remove IgG anti-Amb a 1 antibodies that compete with IgE antibodies for antigen in the ELISA test. Aliquots of individual sera were added to a 50% slurry of protein G Sepharose beads (Pharmacia, Piscataway, NJ) in borate-buffered saline (pH 8.5) at a final 1:10 dilution and rotated overnight at 4°C. ELISA plates were coated with 5 µg/mL Amb a 1 in carbonate buffer (pH 9.0). Each plate contained a titration of a protein G–absorbed standard serum pool that was assigned a value in units per milliliter that was equal to the reciprocal of the highest dilution that gave an OD reading double that of the background. Units per milliliter of the test sera were calculated relative to this standard serum. The ability of Amb a 1-ISS conjugate to bind human Amb a 1–specific IgE was measured by using a competition ELISA assay. ELISA plates were coated overnight at 4°C with Amb a 1 at 1 µg/mL. A 1:600 dilution of a high-titer human IgE anti-Amb a 1 sera was incubated with 500 ng/mL or serial 3-fold dilutions of Amb a 1 or Amb a 1-ISS conjugate for 1 hour at room temperature 126 Tighe et al J ALLERGY CLIN IM M UNOL JULY 2000 at 37°C in a 5% CO2 and water-saturated atmosphere. Supernatants were harvested after 3 days and frozen at –20°C before being analyzed. The levels of IFN-γ and IL-5 were quantitated by ELISA, as previously described,26 by using paired antibodies obtained from Pharmingen (San Diego, Calif). Histamine release from human basophils A Venous blood was collected from volunteers with RW allergy and mixed with a solution of dextrose, EDTA, and dextran to sediment the erythrocytes, as previously described.27 The leukocyte supernatant was used as the source of basophils. Cells were mixed with Amb a 1, Amb a 1-ISS conjugate, Amb a 1 mixed with ISS-ODN, or ISS-ODN alone and incubated for 45 minutes at 37°C with mixing every 15 minutes. The cells were centrifuged, and the histamine content of the supernatant was determined with an autoanalyzer. Control tubes for determining the background histamine release contained cells and buffer only. Background histamine release was usually not more than 2% of the total cellular histamine level, which was determined by using lysis of the cells with 2% HClO4. Statistical analyses The differences between groups are presented as the mean ± SE. Statistical significance of differences between groups (P < .05) was determined by using ANOVA. The P values of the IgG responses of rabbits and monkeys were calculated by using nonparametric analysis. RESULTS Characterization of Amb a 1-ISS conjugate B FIG 1. Nonreduced SDS-PAGE analysis of short RW extract, purified Am b a 1, and Am b a 1-ISS conjugate having an average Am b a 1/ISS-ODN m olar ratio of 1:4 (gel: 4%-20%, TRIS-glycine-SDS buffer). A, Coom assie stain; B, DNA silver stain. Lanes 1, M olecular w eight standards (M ark 12, Novex); lanes 2, blank; lanes 3, RW extract; lanes 4, purified Am b a 1; lanes 5, purified Am b a 1ISS conjugate. before being tested for binding to Amb a 1 in the ELISA assay by using a goat anti-human ε-chain–specific biotin conjugate (Biosource International, Camarillo, Calif) diluted 1:50. Rabbit antisera were similarly analyzed for Amb a 1–specific IgG responses by means of ELISA. A goat anti-rabbit IgG horseradish peroxidase–conjugated antibody (Boehringer Mannheim, Indianapolis, Ind) was used to detect IgG anti-Amb a 1 antibodies. Monkey sera were also similarly analyzed for Amb a 1–specific IgG responses by means of ELISA with a goat anti-monkey IgG horseradish peroxidase–conjugated antibody (ICN/Cappel, Costa Mesa, Calif). Cytokine profile of splenocytes Mice were killed by means of cervical dislocation, and their spleens were removed. Spleens were teased into single-cell suspensions, and 5 × 105 splenocytes were added to triplicate wells in a total volume of 200 µL and stimulated with either 20 µg/mL Amb a 1 protein or medium as a negative control. Cultures were incubated Analysis of the conjugates revealed a population of 10 Amb a 1-ISS or Amb a 1-mODN conjugate bands migrating between the apparent molecular weights of 50 and 120 kd, and all were positive as determined by DNAspecific silver staining, indicating that DNA was bound to Amb a 1 (Fig 1, A and B). The Amb a 1-mODN conjugate (non-ISS) had an appearance similar to that of the Amb a 1-ISS conjugate (not shown). The average number of ODN molecules conjugated to a single Amb a 1 molecule was estimated from the ratio of the molar ODN content to the molar Amb a 1 content, which revealed an average number of ODNs per Amb a 1 molecule of 4. Primary immune response in mice to Amb a 1 or Amb a 1-ISS conjugate The experimental protocol for the primary, secondary, and tertiary immunization experiments is shown in Fig 2. In the primary immunization experiment BALB/c mice were injected intradermally at the base of the tail with either Amb a 1-ISS conjugate or Amb a 1 in saline solution. Control groups received Amb a 1-mODN conjugate or a mixture of Amb a 1 and ISS-ODN (equivalent in amount to that in the conjugate). A control group of mice was also injected intraperitoneally with Amb a 1 in alum to induce a strong TH2 response and IgE antibodies. The IgG2a, IgG1, and IgE antibody formation in these mice is shown in Fig 3. Mice immunized with Amb a 1 alone, Amb a 1-mODN conjugate, or Amb a 1 in alum formed IgG1 antibodies against Amb a 1 but no IgG2a antibodies. In contrast, mice injected with Amb a 1-ISS conju- J ALLERGY CLIN IM M UNOL VOLUM E 106, NUM BER 1, PART 1 Tighe et al 127 FIG 2. Experim ental protocols for prim ary im m unization, boosting w ith Am b a 1 in alum , and im m unotherapy after prim ary im m unization w ith Am b a 1 in alum and subsequent challenge w ith Am b a 1 in saline solution that w ere used for the Am b a 1–based vaccines in m ice. SAC, Sacrifice. FIG 3. Antibody profile in BALB/c m ice injected intraderm ally w ith 10 µg each of either Am b a 1, Am b a 1ISS conjugate, Am b a 1 m ixed w ith 7.9 µg of ISS-ODN (equivalent of ISS-ODN in Am b a 1-ISS conjugate), or Am b a 1-m ODN conjugate. Am b a 1 in alum w as injected in another group to induce a strong TH2 response. M ean ± SE of data from 4 m ice and representative data from one of 4 experim ents. 128 Tighe et al J ALLERGY CLIN IM M UNOL JULY 2000 TABLE I. Cytokine profile of Am b a 1–activated spleen cells from m ice im m unized as indicated Antigen used for immunization IFN-γ (pg/ mL) IL-5 (pg/ mL) Amb a 1 Amb a 1-ISS conjugate Amb a 1-mODN conjugate Amb a 1/ISS-ODN mix Amb a 1/alum Saline (naive) <10 8340 ± 2170* <10 170 ± 110 490 ± 240 <10 630 ± 150 <10* 500 ± 150 610 ± 280 4410 ± 880 <10 Values are means ± SE of data from 8 mice per group (killed at week 16). See Fig 2 for experimental design. *P < .05 for the Amb a 1-ISS conjugate compared with the other groups. gate responded predominantly with IgG2a antibody directed against Amb a 1. Mice injected with a mixture of Amb a 1 and ISS-ODN also generated an IgG2a antibody response; however, the titer was 3-fold (P < .05) lower than that observed with the conjugate. The IgG1 response to Amb a 1 was 20-fold higher than that of mice that received the conjugate. Amb a 1 injected in alum induced significantly (P < .05) higher titers of IgG1 and IgE antibodies than the other groups, which persisted for 16 weeks. Cytokine profile of activated splenocytes after the primary immune response After the last injection, mice were killed, and their spleen cells were stimulated in vitro with Amb a 1. The levels of the IFN-γ and IL-5 secreted in vitro by Amb a 1-stimulated spleen cells are shown in Table I. Spleen cells from mice immunized with Amb a 1-ISS conjugate secreted almost 50-fold more IFN-γ than cells from the mice immunized with the mixture of Amb a 1 and the ISS-ODN. IFN-γ was not detected in supernatants obtained after stimulation of splenocytes from mice injected with either Amb a 1 alone or with Amb a 1-mODN conjugate. Splenocytes from Amb a 1 in alum-injected mice secreted significantly more IL-5 than cells from the other groups (6- to 9-fold more). No IL-5 was detected from splenocytes from mice injected with Amb a 1-ISS conjugate. Control splenocytes without antigen did not have detectable levels of either IFN-γ or IL-5. The IgG2a antibody titer and IFN-γ production after primary immunization indicated that Amb a 1-ISS conjugate induced the strongest TH1 response to Amb a 1, and therefore this modality was selected for the subsequent immunotherapy protocols. Prevention of a TH2 response by priming w ith Amb a 1-ISS conjugate To determine whether the primary TH1 response to Amb a 1-ISS conjugate persisted even after boosting with Amb a 1 in alum, mice were first primed with the Amb a 1-ISS conjugate and then administered a booster injection with Amb a 1 in alum (Fig 2). Control mice were primed with either (1) Amb a 1 alone, (2) Amb a 1mODN conjugate, (3) Amb a 1 mixed with ISS-ODN, or (4) Amb a 1 in alum. As shown in Fig 4, boosting with Amb a 1 in alum caused an increase of the IgG2a titer in mice previously primed with Amb a 1-ISS conjugate, whereas it did not induce IgG2a antibodies in the Amb a 1–primed control mice. Amb a 1 in alum boosted the IgG1 response in all groups, even those injected with Amb a 1-ISS conjugate. The IgE titer rose in all mice boosted with Amb a 1 in alum; however, the response was significantly less (P < .05) in the mice primed with Amb a 1-ISS conjugate or primed with Amb a 1 mixed with ISS-ODN. Cytokine profile of activated splenocytes after priming w ith Amb a 1-ISS conjugate and boosting w ith Amb a 1 The in vitro lymphokine profile in response to Amb a 1 was measured in activated spleen cells obtained 7 weeks after the booster injection of Amb a 1 in alum (Table II). Splenocytes from mice primed with Amb a 1ISS conjugate secreted up to 5-fold higher levels of IFNγ than did spleen cells obtained from mice primed with a mixture of Amb a 1 and ISS-ODN, Amb a 1-mODN conjugate, Amb a 1 in saline solution, or Amb a 1 in alum. The lowest IL-5 levels were in the Amb a 1-ISS conjugate–primed group, down to 15% of that obtained from the control groups. Control splenocytes cultured without antigen had neither detectable levels of IFN-γ nor IL-5. These data indicate that boosting with Amb a 1 in alum reenforced the preexisting TH1 response in mice primed with Amb a 1-ISS conjugate and only induced a weak de novo TH2 response in the same animals. Effect of Amb a 1-ISS conjugate immunization on TH2-primed mice and subsequent response to challenge w ith Amb a 1 Although a mouse model cannot precisely mimic the clinical situation of an allergic patient receiving allergen immunotherapy, the immune response of a TH2-primed animal is useful in evaluating a potential therapeutic approach. For this purpose, mice were primed with Amb a 1 in alum to induce an Amb a 1–specific TH2 response and then were injected 3 times with Amb a 1-ISS conjugate (Fig 2). Control mice received secondary injections 3 times with Amb a 1 in saline solution (analogous to immunotherapy used in human subjects), and another control group did not receive any secondary immunization. To evaluate the effect of immunotherapy in this model on subsequent allergen exposure, mice were subsequently challenged (tertiary immunization) with Amb a 1 in saline solution 15 weeks after the last injection, and the antibody response was determined. As shown in Fig 5, the IgG2a titer increased after the secondary immunization only in the group of mice injected with Amb a 1ISS conjugate. The IgG1 titer increased in all groups after the secondary immunization, although this increase Tighe et al 129 J ALLERGY CLIN IM M UNOL VOLUM E 106, NUM BER 1, PART 1 FIG 4. Antibody profile in BALB/c m ice boosted intraperitoneally w ith 10 µg of Am b a 1 in alum after intraderm al prim ing w ith 10 µg each of either Am b a 1, Am b a 1-ISS conjugate, and Am b a 1-m ODN conjugate in saline solution or intraperitoneal prim ing w ith Am b a 1 in alum . M ean ± SE of 4 m ice and representative data from one of tw o experim ents. TABLE II. Cytokine profile of Am b a 1–activated spleen cells from m ice as indicated and boosted w ith Am b a 1/alum at w eek 13 Antigen used for priming Secondary immunization IFN-γ (pg/ mL) IL-5 (pg/ mL) Amb a 1 Amb a 1-ISS conjugate Amb a 1-mODN conjugate Amb a 1/ISS-ODN mix Amb a 1/alum Saline (naive) Amb a 1/alum Amb a 1/alum Amb a 1/alum Amb a 1/alum Amb a 1/alum Saline 659 ± 197 4595 ± 1527* 397 ± 34 788 ± 164 472 ± 127 <10 840 ± 373 144 ± 32* 613 ± 121 708 ± 93 1361 ± 331 <10 Values are means ± SE of spleens of 4 mice per group (killed at week 20). See Fig 2 for experimental design. *P < .05 for the Amb a 1-ISS conjugate compared with the other groups. was small and not significantly different from the prechallenge level. Secondary immunization with Amb a 1 in saline solution but not with Amb a 1-ISS conjugate increased the IgE titer. After the tertiary immunization with Amb a 1 in saline solution (rechallenge after immunotherapy), the IgG2a titer increased 4-fold in the Amb a 1-ISS conjugate–treated group, whereas it increased only slightly in the other groups. The IgE titer rose over 10-fold in the mice that did not receive secondary immunization. The small rise in the IgE titer in mice that were injected with Amb a 1-ISS conjugate during secondary immunization was significantly less (P < .05) than that of the untreated or Amb a 1–treated mice. Mice that were injected with Amb a 1 in saline solution showed a 1.5-fold rise in IgE; however, the rise was significantly less than that of the untreated mice. Cytokine profile of activated splenocytes after priming w ith Amb a 1 in alum, immunotherapy w ith Amb a 1-ISS conjugate, and challenge w ith Amb a 1 The lymphokine profile of splenocytes from mice killed 8 weeks after the tertiary injection with Amb a 1 in saline solution is shown in Table III. Mice that received Amb a 1-ISS conjugate during the secondary immunization secreted the highest levels of IFN-γ approximately 4 times higher than that of the mice that received Amb a 1 in saline solution or alum during the secondary immunization. The IL-5 secretion was 2.5-fold lower in the mice that received a secondary immunization, either with Amb a 1-ISS conjugate or Amb a 1 compared with that of untreated mice. IFN-γ and IL-5 were not detected in 130 Tighe et al J ALLERGY CLIN IM M UNOL JULY 2000 FIG 5. Effect of secondary im m unization w ith Am b a 1 or Am b a 1-ISS conjugate on tertiary IgG2a, IgG1, or IgE antibody form ation in response to challenge w ith 10 µg of Am b a 1 in saline solution of BALB/c m ice previously prim ed w ith Am b a 1 in alum . M ean ± SE of 4 m ice and representative data from one of tw o experim ents. TABLE III. Cytokine profile of Am b a 1–activated spleen cells of m ice prim ed w ith Am b a 1 in alum and then im m unized w ith Am b a 1-ISS conjugate or Am b a 1 alone and challenged by intraperitoneal injection of Am b a 1 in saline solution Priming Amb a 1/alum Amb a 1/alum Amb a 1/alum Saline Secondary immunization Tertiary immunization IFN-γ (pg/ mL) IL-5 (pg/ mL) Amb a 1 Amb a 1 Amb a 1 Saline 602 ± 202 2318 ± 101* 1033 ± 133 <10 2380 ± 587† 2584 ± 506† 6276 ± 1332 <10 Amb a 1 Amb a 1-ISS conjugate Saline Saline Values are means ± SE of spleens of 4 mice per group (killed at week 32). See Fig 2 for experimental design. *P < .05 for the group primed with Amb a 1/alum and treated with Amb a 1-ISS conjugate compared with the other groups. †P < .05 of the Amb a 1 and the Amb a 1-ISS conjugate–treated groups compared with the untreated control group. the control splenocytes that did not have antigen. The antibody response and the lymphokine profile in this set of experiments demonstrate that a secondary immunization with either Amb a 1-ISS conjugate or Amb a 1 prevents the large increase of the IgE titers and significantly reduces IL-5 secretion after re-exposure to Amb a 1. Although the absolute amount of secreted IL-5 was still substantial, treatment with Amb a 1-ISS conjugate induced very high levels of IFN-γ that provided an IFNγ/IL-5 ratio that is likely to cause a decrease of the effects of the residual IL-5 formation. IgG antibody response to Amb a 1, Amb a 1 mixed w ith ISS, or Amb a 1-ISS conjugate in rabbits and cynomolgus monkeys Rabbits were immunized with Amb a 1, Amb a 1 mixed with 50 or 500 µg of ISS-ODN, or Amb a 1-ISS conjugate (Fig 6, A). All groups showed low antibody titers in preimmunization bleeds, ranging from 263 to 495 U/mL. No increase in anti-Amb a 1 IgG titers was detected after two immunizations with Amb a 1 mixed with 50 or 500 µg of ISS-ODN. In contrast, two immunizations with Amb a 1-ISS conjugate increased the mean anti-Amb a 1 IgG titers from 495 to 106,000 U/mL (P < .05). Cynomolgus monkeys were immunized with Amb a 1 or Amb a 1-ISS conjugate (Fig 6, B). All the animals had some background anti-Amb a 1 IgG titers (prebleed samples), ranging from 350 to 450 U/mL. No increases in the anti-Amb a 1 IgG titers were detected after one, two, or three immunizations with Amb a 1. In contrast, two immunizations with Amb a 1-ISS conjugate enhanced the mean anti-Amb a 1 IgG antibody titer 18-fold (449 vs 8040 U/mL, P < .05). No IgE anti-Amb a 1 antibodies were detectable in any of the monkey sera before or after immunization (data not shown). Tighe et al 131 J ALLERGY CLIN IM M UNOL VOLUM E 106, NUM BER 1, PART 1 A B FIG 6. A, IgG antibody titers of rabbits injected w ith either 10 µg of Am b a 1, 10 µg of Am b a 1-ISS conjugate, or 10 µg of Am b a 1 m ixed w ith either 50 or 500 µg of ISS-ODN. B, IgG antibody titers of cynom olgus m onkeys injected w ith 50 µg of Am b a 1 or 50 µg of Am b a 1-ISS conjugate. Circles represent the titer of an individual anim al, and the bars represent the m ean values. Reaction w ith human IgE antibodies and histamine release by human basophils from patients allergic to RW induced by Amb a 1 or Amb a 1-ISS conjugate To determine whether the chemical conjugation of ISSODN to Amb a 1 diminished the ability of the Amb a 1 to react with human IgE anti-Amb a 1 antibodies, the conjugate was tested in an ELISA assay for its inhibition of the reactivity of an IgE anti-Amb a 1–containing sera with Amb a 1. As shown in Fig 7, the Amb a 1-ISS conjugate reacted less well with human IgE anti-Amb a 1 antibodies than did Amb a 1. An approximate 3-fold higher concentration of Amb a 1-ISS conjugate was necessary to achieve 50% inhibition compared with native Amb a 1. To determine whether the conjugation of ISS-ODN to Amb a 1 changed its allergenic properties, basophils were obtained from 4 patients allergic to RW, and histamine release was assessed in vitro after stimulation with Amb a 1 or Amb a 1-ISS conjugate. As shown in Fig 8, on aver- age a 30-fold higher concentration of Amb a 1-ISS conjugate was required to induce 50% of maximum histamine release in comparison with the histamine release induced by Amb a 1 alone. The histamine release induced by Amb a 1-mODN conjugate was also about 30-fold lower than that induced by Amb a 1 (data not shown). ISS-ODN alone did not induce histamine release nor did it decrease histamine release induced by native Amb a 1. DISCUSSION This study demonstrates that chemical conjugation of an ISS-ODN to the major RW allergen Amb a 1 enhances its immunogenicity and reduces its allergenicity. Mice immunized with Amb a 1-ISS conjugate generated high levels of IgG2a antibodies and IFN-γ–secreting cells, whereas native Amb a 1 induced IgG1 and IgE antibodies and IL-5–secreting T cells. Mice injected with Amb a 1-ISS conjugate also formed IgG1 antibodies; however, IgG1 formation is not always restricted to TH2 responses. The observed secreted 132 Tighe et al J ALLERGY CLIN IM M UNOL JULY 2000 FIG 7. Inhibition of hum an IgE binding of a high IgE titer serum from a patient allergic to RW to Am b a 1 by Am b a 1-ISS conjugates in an ELISA inhibition test. Data show n are from one of tw o experim ents that gave sim ilar results. Each data point represents the m ean of duplicate analyses. The variability of duplicate w ells w as less than 20%, and the differences betw een the m eans of the first three inhibition concentrations w ere statistically significant (P < .05). FIG 8. Histam ine release from hum an basophils induced by either Am b a 1, Am b a 1-ISS conjugate, Am b a 1 m ixed w ith ISS-ODN at a dose equivalent to that conjugated to 10 µg of Am b a 1, or ISS-ODN alone. M ean ± SE histam ine release from basophils of 4 different donors w ith RW allergy is show n. lymphokines were typical for a TH1 and TH2 immune response, respectively.28 Amb a 1 conjugated to a non-ISS control ODN (mODN) in which the two CG dinucleotides were mutated to CC and AG did not have this effect. The increased immunogenicity of the Amb a 1-ISS conjugate is different from the modified allergens (allergoids) that were used in the past4,5 because allergoids lost most of their immunogenicity after chemical modification. Moreover, J ALLERGY CLIN IM M UNOL VOLUM E 106, NUM BER 1, PART 1 and clinically most importantly, a preexisting Amb a 1–specific TH2 response did not prevent induction of a de novo allergen-specific TH1 response to immunization with Amb a 1-ISS conjugate and suppressed IgE antibody formation to a tertiary injection of Amb a 1. Furthermore, as judged by in vitro histamine release, the conjugate was less allergenic: about a 30-fold higher concentration of conjugate was necessary to induce the same amount of histamine release as that induced by Amb a 1. These properties of allergen-ISS conjugates indicate that they offer a novel type of modified allergen that may provide an effective and safer approach to immunotherapy in human subjects than that presently available. The effect of immunotherapy on downregulation of an allergic response in an animal model has to our knowledge not been thoroughly investigated in the past, presumably because it was used as an established therapy for allergic disorders in human subjects long before the immunologic mechanisms underlying allergic disorders were discovered. Because we prepared the Amb a 1-ISS conjugate for a more effective and safer immunotherapy, we compared the effect of Amb a 1-ISS conjugate immunotherapy with that of Amb a 1 in mice primed with Amb a 1 in alum to induce a preexisting TH2 response. Perhaps not too surprisingly, both the Amb a 1ISS conjugate and Amb a 1 injections prevented a large increase of the IgE antibody formation on a tertiary challenge with Amb a 1, and both caused a significant decrease of the IL-5 secretion by splenic T cells compared with untreated mice. These data parallel the clinical observations of the blunting of the IgE response of immunotherapy-treated patients during the ragweed pollen season and data showing that immunotherapy could change allergen-specific T cells from a TH2- to a TH1-like type.10-13 However, in mice the Amb a 1-ISS conjugate therapy had a qualitatively much stronger beneficial effect than the Amb a 1 injections. Amb a 1-ISS conjugate treatment induced much higher titers of IgG2a antibodies than Amb a 1 injections, suggesting a stronger TH1 response. Although the T cells from the Amb a 1-ISS conjugate– and Amb a 1–treated mice still secreted substantial amounts of IL-5, the high IFN-γ secretion by the conjugate-treated mice resulted in an IFN-γ/IL-5 ratio that is likely to decrease any residual IL-5 effects. Second, Amb a 1-ISS conjugate injections suppressed the IgE antibody formation more strongly than the Amb a 1 injections. Whether Amb a 1-ISS conjugate immunotherapy is more efficient than classical immunotherapy in allergic patients remains to be investigated. Coinjection of Amb a 1 with ISS-ODN also induced a TH1 response in mice; however, it was weaker than that induced with the corresponding conjugate. Coinjection of mice with a mixture of Amb a 1 and ISS-ODN induced a lower IgG2a titer, and T cells from these mice secreted 10-fold less IFN-γ than T cells from mice immunized with a conjugate that contained an equivalent amount of Amb a 1 and ISS-ODN. The enhanced immunogenicity of Amb a 1-ISS conjugates was even more dramatic in rabbits and monkeys. These species formed high titers of Tighe et al 133 IgG antibodies to the conjugate but failed to respond to the mixture or antigen alone. These findings indicate that conjugation of ISS-ODN to antigens may be crucial for successful induction of immune responses to weak antigens or very low doses of antigen, as is used in allergen immunotherapy. As noted, the increased immunogenicity of Amb a 1ISS conjugate was not limited to murine species. Both rabbits and cynomolgus monkeys generated a significantly enhanced IgG antibody response after injection with the conjugate. This response was not observed after injection with either Amb a 1 alone (rabbits and monkeys) or after injection with a mixture of Amb a 1 and ISS-ODN (rabbits). These results are the first to report in vivo immunostimulatory properties in mice, rabbits, and nonhuman primates by the same ISS-ODN. The mechanisms by which allergen-ISS conjugate induces a TH1 response are not yet fully explored. ISSODNs were shown to activate components of the innate immune system that result in a subsequent TH1 response.14-17 In particular, ISS-ODN induces the secretion of type 1 cytokines15-17,29 and induces a distinct costimulatory molecule profile on APCs.30 Allergen-ISS conjugate allows for the codelivery of antigen and ISS to the same cell, such as an APC, and thereby optimizes the immunomodulatory effects of ISS with antigen processing. In contrast, the coinjection of Amb a 1 with ISSODN may result in the in vivo dissociation of the ISS from the coinjected antigen, resulting in a less efficient ISS effect on antigen processing and a weaker TH1 response. Subcutaneous injection of allergen, as it is performed in current standard immunotherapy, carries the risk of inducing anaphylactic reactions.13 These adverse reactions severely limit the tolerated doses and are one of the major reasons why immunotherapy is used less today than it has been in the past. Amb a 1-ISS conjugates were less reactive with human IgE anti-Amb a 1 antibodies and required 30-fold more conjugate on a molar basis than Amb a 1 to release the same amount of histamine from basophils of patients allergic to RW. The reasons for the decreased histamine release by the conjugates are not yet fully understood. Steric hindrance or chemical alteration of the allergenic epitopes by the ISS-ODN coupled to the Amb a 1 may be one of the reasons. Should this property of allergen-ISS conjugate also result in a lower anaphylactic potential in vivo, patients might be safely injected with higher doses of allergen-ISS conjugate compared with unmodified allergen alone. In summary, these studies demonstrate that chemical conjugation of ISS-ODN to allergen confers on the allergen two new properties that make it an excellent candidate for human immunotherapy. First, allergen-ISS conjugate induces strong TH1 and high IgG responses, even in the face of an ongoing TH2 response. The induction of a rapid and strong allergen-specific IFN-γ release (ie, a TH1 response), as well as the IgG antibody formation to injections of low doses of allergen, are both desired properties that would enhance the beneficial response to an 134 Tighe et al immunotherapeutic agent. Second, the markedly lower histamine release from basophils than allergen or allergen mixed with ISS should portend a much lower risk of anaphylaxis or serious adverse events to immunotherapy. This combination of enhanced immunogenicity and reduced allergenicity is encouraging for the development of allergen-ISS conjugates as a novel and efficacious mode of allergen immunotherapy. We thank Mary Malo, Cindy Beck, John Zhu, Robin Marsden, De Shon Hall, Dawn Dawson, Debbie Higgins, Paula Traquina, Bob Milley, Roberto Rodriquez, and Christina Abbate for technical assistance; Jane Uhle for secretarial assistance; and Drs R. Schleimer and R. Zeiger for critical review of the manuscript. J ALLERGY CLIN IM M UNOL JULY 2000 13. 14. 15. 16. 17. 18. REFERENCES 1. Malling H-J, Weeke B, eds. Immunotherapy. Allergy 1993;48(Suppl):35. 2. Creticos PS. Immunotherapy with allergens. JAMA 1992;268:2834-9. 3. Bousquet J, Lockey RF, Malling H-J, eds. Allergen immunotherapy: therapeutic vaccines for allergic diseases. Allergy 1998;53(Suppl):44. 4. Marsh DG, Lichtenstein LM, Campbell DN. Studies on allergoids prepared from naturally occurring allergens. I. Assay of allergenicity and antigenicity of formalized rye group I components. Immunology 1970;18:705-14. 5. Lee WY, Sehon AH. Abrogation of reaginic antibodies with modified proteins. Nature 1977;267:618-20. 6. Wierenga EA, Snoek M, de Groot C, et al. Evidence for compartmentalization of functional subsets of CD4+ T lymphocytes in atopic patients. J Immunol 1990;144:4651-6. 7. Parronchi P, Macchia D, Piccini M-P, et al. Allergen-and bacterial antigen-specific T-cell clones established from atopic donors show a different profile of cytokine production. Proc Natl Acad Sci USA 1991;88:4538-42. 8. Pene J, Rousset F, Briere F, et al. IgE production by normal human lymphocytes is induced by interleukin 4 and suppressed by interferons gamma and alpha and prostaglandin E2. Proc Natl Acad Sci USA 1988;85:6880-4. 9. Lopez AF, Sanderson CJ, Gamble JR, Campbell HD, Youg IG, Vadas MA. Recombinant human interleukin 5 is a selective activator of human eosinophil function. J Exp Med 1988;167:219-24. 10. Secrist H, Chelen CJ, Wen Y, Marshall JD, Umetsu DT. Allergen immunotherapy decreases interleukin 4 production in CD4+ T cells from allergic individuals. JExp Med 1993;178:2123-30. 11. Jutel M, Pichler WJ, Skrebic D, Urwyler A, Dahinden C, Muller UR. Bee venom immunotherapy results in decrease of IL-4 and IL-5 and increase of IFN-gamma secretion in specific allergen-stimulated T cell cultures. J Immunol 1995;154:4187-94. 12. McHugh SM, Deighton J, Stewart AG, Lachmann PJ, Ewan PW. Bee venom immunotherapy induces a shift in cytokine responses from a TH- 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 2 to a TH-1 dominant pattern: comparison of rush and conventional immunotherapy. Clin Exp Allergy 1995;25:828-38. Durham SR, Till SJ. Immunologic changes associated with allergen immunotherapy. J Allergy Clin Immunol 1998;102:157-64. Sato Y, Roman M, Tighe H, et al. Immunostimulatory DNA sequences necessary for effective intradermal gene immunization. Science 1996;273:352-4. Roman M, Martin-Orozco E, Goodman J, et al. Immunostimulatory DNA sequences function as Th1 inducing adjuvants. Nat Med 1997;3:849-54. Chu RS, Targoni OS, Krieg AM, Lehman PV, Harding CV. GpC oligodeoxynucleotides act as adjuvant that switch on T helper 1 (Th1) immunity. J Exp Med 1997;186:1623-31. Sun S, Kishimoto H, Sprent J. DNA as an adjuvant: capacity of insect DNA and synthetic oligonucleotides to augment T cell responses to specific antigen. J Exp Med 1998;187:1145-50 Jahn-Schmid B, Wiedermann U, Bohle B, Repa A, Kraft D, Ebner C. Oligodeoxynucleotides containing CpG motifs modulate the allergic TH2 response to BALB/c mice to Bet v 1, the major birch pollen allergen. J Allergy Clin Immunol 1999;104:1015-23. Sur S, Wild JS, Choudhury BK, Sur N, Alam R, Klinman DM. Long term prevention of allergic lung inflammation in a mouse model of asthma by CpG oligodeoxynucleotides. J Immunol 1999;162:6284-93. Broide D, Schwarze J, Tighe H, et al. Immunostimulatory DNA sequences inhibit IL-5, eosinophilic inflammation, and airway hyperresponsiveness in mice. J Immunol 1998;161:7054-62. Kline JN, Waldschmidt TJ, Businga TR, et al. Modulation of airway inflammation by CpG oligodeoxynucleotides in a murine model of asthma. J Immunol 1998;160:2555-9. Van Uden J, Raz E. Immunostimulatory DNA and applications to allergic disease. J Allergy Clin Immunol 1999;104:902-10. Klinman DM, Barnhart KM, Conover J. CPG motifs as immune adjuvants. Vaccine 1999;17:19-25. Tighe H, Takabayashi K, Schwartz D, et al. Coupling CpG motif immunostimulatory DNA to ragweed allergen Amb a 1 induces a Th1 response to allergen. J Allergy Clin Immunol 1999;103:S48. King TP, Norman PS, Lichtenstein LM. Isolation and characterization of allergens from ragweed pollen. IV. Biochemistry 1967;6:1992-2000. Raz E, Tighe H, Sato Y, et al. Preferential induction of a Th1 immune response and inhibition of specific IgE antibody formation by plasmid DNA immunization. Proc Natl Acad Sci USA 1996;93:5141-5. Lichtenstein LM, Norman PS, Winkenwerder WL. A single year of immunotherapy for ragweed hay fever. Immunological and clinical studies. Ann Intern Med 1971;75:663-71. Mosmann TR, Coffman RL. The Th1 and Th2 cells: different pattern of lymphokine secretion lead to different functional properties. Ann Rev Immunol 1989;7:145-73. Klinman DM, Verthelyi D, Takeshita F, Ishii KJ. Immune recognition of foreign DNA: a cure for bioterrorism? Immunity 1999;11:123-9. Martin-Orozco E, Kobayashi H, Van Uden J, Nguyen M-D, Kornbluth RS, Raz E. Enhancement of antigen-presenting cell surface molecules involved in cognate interactions by immunostimulatory DNA sequences. Int Immunol 1999;11:1111-8.