Characterization of Aluminum Hydroxide Gel and Oil-in-Water Emulsion Formulations Containing CpG ODNs

October 2, 2010

Adjuvant activity can be greatly improved by appropriate formulation of cytosine-phosphorothioate-guanine oligodeoxynucleotides (CpG ODNs).


Oligonucleotide sequences known as CpG ODNs are potent TLR9 agonists. The literature shows that formulation of CpG ODNs elicits more potent immune responses than the formulation or the adjuvant molecule alone. Given the importance of formulation on adjuvant biological activity and toxicity, formulation parameters should be well characterized and rationally designed. An understanding of the biophysical phenomena related to adjuvant interactions with formulation components can be obtained through complementary analytical techniques. Microeletrophoresis, UV spectrophotometry, and HPLC are used to characterize the association of CpG ODNs with aluminum hydroxide and oil-in-water emulsion formulations.

Synthetic oligodeoxynucleotides containing cytosine-phosphorothioate-guanine sequences (CpG ODNs) are designed to mimic bacterial genomes that contain a higher proportion of unmethylated CpG motifs compared to vertebrate genomes. CpG ODNs have demonstrated potent vaccine adjuvant activity as TLR9 agonists, inducing higher antibody titers and an array of cellular immune responses.1


Several studies have demonstrated the importance of appropriate formulation of CpG ODNs to maximize adjuvant response, including a recent special issue dedicated to this topic.1 For example, CpG ODN mixed with aluminum hydroxide or MF59 elicited higher hemagglutination inhibition antibody titers compared to CpG or either formulation alone, and a more Th1-biased cytokine response compared to either formulation alone in a mouse influenza model.2 Similarly, adding CpG ODN to MF59 induced higher total antibody levels compared to either component alone and increased IgG2a/IgG1 ratios in mice compared to MF59 alone on immunization with HIV p55 gag antigen.3

It was also shown that CpG ODN formulated with aluminum hydroxide or an oil-in-water emulsion induced significantly higher antibody titers than either component alone in a rabbit model with various antigens or a piglet model with swine streptococcic septicemia killed vaccine.4,5 In addition, total antibody levels and CTL activity induced in mice by a hepatitis B surface antigen formulated with CpG ODN and aluminum hydroxide were significantly higher than CpG or alum alone.6 Finally, encapsulation of CpG ODNs in polymer microparticles has been shown to increase antibody titers to a meningitis antigen in mice compared to non-encapsulated CpG.7

Given that the combination of adjuvant molecules such as CpG ODNs with particulate formulations translates into significantly more potent immune responses or reduced toxicity, it is important to optimize the formulation and presentation of the adjuvant to the immune system. Improving adjuvant formulation design requires an understanding of the basic biophysical principles underlying adjuvant association with the particulate formulation. This topic will be discussed using several specific examples of the formulation of CpG ODNs with particulate formulations commonly employed as vaccine adjuvants, namely aluminum hydroxide gel and oil-in-water emulsion. A particular emphasis has been placed on practical aspects of material properties and characterization methodology.


CpG ODNs. CpG oligodeoxynucleotides (ODNs) were obtained from Avecia Biotechnology and Coley Pharmaceutical Group. Several different sequences are available and efficacy varies between animal models depending on the sequence. The ODNs described here are a modified CpG ODN similar to 10103 (the order of two sequential base pairs were switched from CpG ODN 10103): TCG TCG TTT TTC GGT GCT TTT, CpG 2395: TCG TCG TTT TCG GCG CGC GCC G, and CpG ODN 1826: TCC ATG ACG TTC CTG ACG TT.8,9 CpG ODN (similar to 10103) from Avecia is supplied as a dry powder and stored at –20 °C. CpG 1826 was purified with an ethanol precipitation step and reconstituted in 10 mM pH7 Tris with 1 mM EDTA. CpG ODN absorbance values were measured using a Beckman Coulter DU 640i or a Hitachi 3900H UV-Vis spectrophotometer.

Aluminum. Alhydrogel "85" was manufactured by Brenntag Biosector. Alhydrogel "85" consists of aluminum hydroxide gel containing small aggregates (~1 to 10 mm) of high surface area and positive charge.10 X-ray diffraction and infrared spectroscopy data indicate that the structure of aluminum hydroxide adjuvant resembles boehmite with the chemical formula AlO(OH), instead of the implied formula of Al(OH)3.10,11 The aluminum content in Alhydrogel "85" is 1% or 10 mg/mL.12 Alhydrogel is stored at room temperature and settles with time, so vigorous shaking is required before use. Aluminum hydroxide zeta potential measurements were measured using a Zetasizer Nano-ZS and an MPT-2 multi-purpose titrator from Malvern Instruments.

Oil-in-water emulsion. An oil-in-water stable emulsion (SE) was manufactured at the Infectious Disease Research Institute. Squalene (Sigma-Aldrich), purified from shark liver, was used for the oil source and was present at 10% (v/v) in the final product. High purity (99%) phosphatidylcholine (Avanti Polar Lipids) from chicken egg yolks, was used as a natural emulsifier and was present at 1.92% (w/v) in the final product. Poloxamer 188 (Pluronic F68 from BASF), a synthetic triblock copolymer nonionic surfactant, was used as a coemulsifier at a final concentration of 0.09% (w/v). To create an isotonic formulation, glycerol (Sigma-Aldrich) was included in the SE at 1.8% (v/v). Alpha-tocopherol (Spectrum Chemical) was used as an antioxidant and present at 0.05% (w/v). A 25 mM ammonium phosphate buffer was used to maintain the aqueous phase at pH 5.1 ± 0.05.

The oil phase, consisting of the squalene, phosphatidylcholine (PC), and alpha-tocopherol, was prepared by sonication in a heated (~50 °C) water bath until the PC was fully dissolved (typically 1–2 h). The aqueous phase was prepared by combining deionized water with glycerol, Pluronic F68, and ammonium phosphate buffer. The buffered aqueous phase was added at 90% (v/v) to the oil phase and then mixed with a Silverson heavy-duty laboratory mixer emulsifier (3/4 inch tubular square hole high shear screen attachment) at ~8,000–10,000 rpm for several minutes. The mixture was subjected to high-pressure homogenization using the Microfluidics M-110EH-30 for 12 passes at ~207 MPa (~30,000 psi).

Particle diameter was ~100–110 nm as measured by dynamic light scattering using the Malvern Instruments Zetasizer Nano-S. Visually, the SE had the appearance of homogenized milk. The SE formulation was stored in rubber-stoppered glass vials at 2–8 °C and demonstrated good stability (minimal change in particle size and no visual phase separation) for at least 1 year. The aqueous phase separation procedure described in the text was carried out using Ficoll PM400 obtained from GE Healthcare. In addition to the manufacture of SE, an MF59-like emulsion was manufactured using similar procedures to produce an oil-in-water formulation containing 5% v/v squalene and 0.5% v/v each of Tween 80 and Span 85 as surfactants; the water in this case was unbuffered.

Results and Discussion

CpG-aluminum hydroxide adsorption isotherm. Because of the negatively charged phosphate groups that make up the CpG ODN backbone, these molecules are electrostatically attracted to the positively charged Alhydrogel. Besides electrostatic considerations, other mechanisms, including phosphate ligand exchange and hydrophobic interactions, also can attract adjuvants or antigens to the aluminum adjuvant.10

The following illustrates two complementary procedures for determining the extent of CpG ODN adsorption to Alhydrogel "85". These techniques also are applicable to antigen adsorption studies.10 Aliquots of 100 mL Alhydrogel "85" (10 mg/mL aluminum) were pipetted into each of 10 microcentrifuge tubes. One hundred mg of CpG ODN (1 mg/mL) were added to the first microcentrifuge tube, 200 mg to the second tube, etc., with the last tube containing 1 mg of CpG ODN. Each tube contained the same total solution volume (1.1 mL). After briefly vortexing, the samples were incubated for 1 h at room temperature to allow for binding of CpG ODN to Alhydrogel "85". The samples were then centrifuged for 9 min at 16,000g to separate Alhydrogel "85" from aqueous solution (Alhydrogel "85" settles). The absorbance of supernatant at 260 nm was measured in each sample. The non-adsorbed concentration of CpG ODN using a standard curve was determined and subtracted from the original CpG ODN concentration to calculate adsorbed CpG ODN. Adsorbed CpG ODN versus original CpG ODN concentration in solution is plotted in Figure 1.

Figure 1. Adsorption isotherm of a modified CpG ODN (similar to CpG 10103) on aluminum hydroxide (Alhydrogel "85") as measured by centrifugal separation and absorbance at 260 nm.

A CpG ODN-aluminum hydroxide adsorption isotherm also can be constructed by measuring the zeta potential of aluminum hydroxide particles during CpG ODN titration. Increased negative charge on the particles reflected in lower zeta potential represents more bound CpG ODN, whereas a plateau in zeta potential represents maximum CpG ODN adsorption. To measure the zeta potential of CpG ODN molecules adsorbed to Alhydrogel "85", a stock solution of aluminum hydroxide was prepared at 1 mg/mL aluminum and CpG ODN at 4 mg/mL. The titrator was configured to inject 0.05 mg/mL aliquots of CpG ODN and collect duplicate measurements of zeta potential after each injection. Of course, if a titrator is not available, the zeta potential measurements can be carried out by preparing separate solutions, as was done for the UV absorption experiment above. The titration was continued until the final CpG concentration was 1.25 mg/mL, corresponding to 1.8 mg CpG ODN/mg aluminum. Zeta potential versus CpG ODN/aluminum w/w is plotted in Figure 2. Notice that the agreement between both adsorption isotherms is quite good (Figures 1 and 2).

Figure 2. Adsorption isotherm of a modified CpG ODN (similar to CpG 10103) on aluminum hydroxide as measured by zeta potential titration. Error bars represent duplicate measurements at each concentration in three separate experiments.

Adsorption Effects of Excipients, Other Adjuvants, or Antigen

The excipients present in some antigen or adjuvant stock solutions can alter the binding behavior of the CpG-Alhydrogel complex. One must be especially cautious of ligand exchange with phosphate-containing solutions. Before the effects of the antigen on the adsorption of CpG and vice versa can be monitored, the effects of the excipients present in the antigen preparation must be evaluated. To determine excipient effects on CpG-Alhydrogel adsorption behavior, the binding capacity protocol from above was repeated using a CpG ODN solution containing the excipients from a lyophilized antigen preparation instead of water. Thus, for a 100 mL in vivo injection consisting of 10 mg CpG and 100 mg aluminum, a sample phosphate buffer containing EDTA and sucrose was mixed with the adjuvant and aluminum. The sample was centrifuged as described above and the absorbance of supernatant was measured at 260 nm. For this example, it was found that the adsorption of CpG ODN to Alhydrogel was not diminished in the presence of the excipient mixture (Table 1). It should be noted that order of addition of each component and time to bind also may affect adsorption behavior. The effects on CpG-Alhydrogel adsorption of several different buffers and excipients has been discussed elsewhere.13

Table 1. Adsorbtion of CpG ODN to aluminum hydroxide gel

A similar procedure can be followed to determine the effects of other adjuvants or antigens on CpG-Alhydrogel adsorption. However, detection using spectrophotometry becomes more complicated since other adjuvants and especially antigens can interfere with the 260 nm absorbance signal. In these cases, other detection methods may be desirable. Supernatant CpG ODN can be detected by SDS-PAGE/laser densitometry, in addition to spectrophotometry.13 Size exclusion HPLC can detect CpG ODN present in the supernatant after centrifugation.

The mobile phase for size exclusion HPLC methods often is the same as the sample buffer, making these methods simple to perform. Other HPLC methods for ODNs or DNA, including reversed phase and ion exchange, also have been demonstrated.14,15 In addition, a dissociation step may be necessary before HPLC analysis to separate CpG ODN bound to protein or other materials.14 In this case, because the CpG ODN was stored in TE buffer, a mobile phase was prepared consisting of 10 mM Tris buffer at pH 8. The flow rate was set to 0.5 mL/min (TSKgel G3000PWXL column), with an injection volume of 20 μL, and the UV-Vis absorbance detector set to 260 nm. To develop the detection range for this method, samples of CpG 2395 were prepared at 1, 0.1, and 0.01 mg/mL, with a blank of TE buffer. Detection of CpG ODN was possible at a concentration of 0.01 mg/mL (Figure 3).

Figure 3. HPLC detection of CpG 2395 in TE buffer

Association of CpG ODN with Emulsion Formulations

The association of CpG with the oil (squalene) phase is thought to be weak because of the unfavorable energetics of hydrophobic interactions. The partitioning of oligonucleotides is dominated by strong solvation in the aqueous phase, though surface or interfacial effects may be occurring through mediation of surfactants. However, even weak association with particles may have an effect on biological activity. Because of the small difference in density of the two phases, it is necessary to increase the density gradient by adding a high MW sucrose polymer such as Ficoll PM400 (GE Healthcare). Ficoll can be dissolved in water to create a solution that has a significantly higher density than water because of its high molecular weight and excellent water solubility. This allows separation of the phases in an emulsion using a standard microcentrifuge. Alternatively, an ultracentrifuge may be used to separate the phases without adding the sucrose polymer.3

To determine the extent of CpG ODN associated with the oil particulates in the SE formulation, a 30% (w/v) solution of Ficoll in water was mixed at a 1:1 ratio with the SE formulation (10% oil v/v and 1.25 mg/mL CpG ODN) in a microcentrifuge tube and vortexed briefly. The sample was centrifuged for 30 min at 16,000g. After centrifugation, the sample was carefully removed and the tube bottom carefully punctured with a 27g half-inch needle to extract an aliquot of the lower aqueous phase. The absorbance at 260 nm was then measured and compared to a CpG ODN standard. The difference in absorption represents the amount of CpG ODN associated with the oil phase (Table 2). In this example, all of the CpG ODN was in the aqueous phase, as was the case for a CpG-MF59 formulation that underwent ultracentrifugation.3 This probably is because these emulsions use nonionic surfactants or neutral phospholipids as emulsifiers. Emulsions made with charged emulsifiers show higher CpG ODN association with the oil phase and somewhat better immune activity; however, strong oil association is not necessary for synergistic adjuvant effects demonstrated with neutral emulsion-CpG ODN formulations.3

Table 2. Association of CpG ODN with oil-in-water emulsion

Because CpG ODNs are efficiently ionized at physiological pH, a change in electrophoretic mobility (zeta potential) of the emulsion can be observed if the CpG ODN molecules associate preferentially into the interface between the oil and water phases. Differences in the composition and local chemical potentials at the interface can adjust to balance the electrostatic–hydrophobic interactions that shift as the activity of CpG grows with increasing solution concentration. To measure the change in emulsion interfacial charge caused by CpG ODN interactions, a titration of zeta potential versus CpG ODN concentration was performed. The emulsion used here was the MF59-like emulsion described above. The emulsion was diluted 10-fold in water to permit light beam transmission of the scattering solution, as well as lowering the ionic strength to decrease the conductivity and increase the Debye length, permitting a more accurate measurement.

The diluted emulsion was titrated against 4.0 mg/mL CpG ODN and changes in emulsion surface charge and curvature were measured by measuring both zeta potential and particle size as described above. In this example, the effects were subtle, with a gradual downward slope in zeta potential from 0 to ~0.1 mg/mL CpG ODN, a gradual upward slope from ~0.1 to 0.25 mg/mL (Figure 4), and an abrupt shift of ~2 mV at ~0.25 mg/mL CpG ODN, possibly caused by a reorganization of the molecules in the interfacial Stern layer to accommodate the changes in the aqueous phase as solvation of the added CpG becomes more extensive. At higher CpG ODN concentrations up to 0.9 mg/mL, the zeta potential is relatively constant.

Figure 4. Zeta potential measurement of the titration of CpG 1826 in an MF59-like emulsion. Error bars represent triplicate measurements from a single experiment.


Appropriate formulation of CpG ODN can significantly improve vaccine adjuvant activity. Rational design of CpG ODN formulations is enabled by thorough characterization of fundamental formulation parameters. In addition, changes in vaccine excipient composition should be monitored to determine effects on CpG ODN formulation. In this work, simple but informative physicochemical analytical techniques have been effectively used to determine the extent of CpG ODN association with particulate formulations, namely aluminum hydroxide and oil-in-water emulsion formulations.


This work was supported in part by Grant 42387 from the Bill & Melinda Gates Foundation. The authors would like to thank Martin Freide, PhD, for helpful discussions regarding method development.

CHRISTOPHER B. FOX is the lead formulation engineer, TIMOTHY S. DUTILL is the process development specialist II, JAMES CHESKO is a senior scientist, STEVE G. REED is the head of research and development, and THOMAS S. VEDVICK is the director of process sciences, all at the Infections Disease Research Institute, Seattle, WA, 206.330.2527, cfox@idri.orgRYAN C. ANDERSON, formerly of the Infections Disease Research Institute, currently is the account consultant—northern California, at Chemical Abstracts Service, Columbus, OH.


1. Salem AK. CpG oligonucleotides as immunotherapeutic adjuvants: innovative applications and delivery strategies. Adv Drug Del Rev. 2009;61:193–4.

2. Wack A, Baudner BC, Hilbert AK, Manini I, Nuti S, Tavarini, S, et al. Combination adjuvants for the induction of potent, long-lasting antibody and T-cell responses to influenza vaccine in mice. Vaccine. 2008;26:552–61.

3. Ott G, Singh M, Kazzaz J, Briones M, Soenawan E, Ugozzoli M, et al. A cationic sub-micron emulsion (MF59/DOTAP) is an effective delivery system for DNA vaccines. J Control Rel. 2002;79:1–5.

4. Ioannou XP, Gomis SM, Hecker R, Babiuk LA, van Drunen Littel-van den Hurk S. Safety and efficacy of CpG-containing oligodeoxynucleotides as immunological adjuvants in rabbits. Vaccine. 2003;21:4368–72.

5. Linghua Z, Xingshan T, Fengzhen Z. The efficacy of CpG oligodinucleotides, in combination with conventional adjuvants, as immunological adjuvants to swine streptococcic septicemia vaccine in piglets in vivo. Int Immunopharm. 2006;6:1267–76.

6. Brazolot Millan CL, Weeratna R, Krieg AM, Siegrist CA, Davis HL. CpG DNA can induce strong Th1 humoral and cell-mediated immune responses against hepatitis B surface antigen in young mice. Proc Natl Acad Sci USA. 1998;95:15553–8.

7. Singh M, Chesko J, Kazzaz J, Ugozolli M, Malyala P, O'Hagan D. Surface-charged poly(lactide-co-glycolide) microparticles as novel antigen delivery systems. In: Singh M, editor. Vaccine adjuvants and delivery systems. Hoboken, NJ: John Wiley & Sons; 2007. p. 223–48.

8. Krieg AM. WO/2004/005476, WIPO, Nucleic acid compositions for stimulating immune responses, Coley Pharmaceutical Group, Inc., 2004.

9. CpG ODN Information Sheet, Coley Pharmaceutical Group, 2006.

10. Hem SL, HogenEsch H. Aluminum-containing adjuvants: properties, formulation, and use. In: Singh M, editor. Vaccine adjuvants and delivery systems. Hoboken, NJ: John Wiley & Sons; 2007. p. 81–114.

11. Lindblad EB. Aluminium adjuvants—in retrospect and prospect. Vaccine. 2004;22:3658–68.

12. Alhydrogel "85" 2% Certificate of Analysis. Brenntag Biosector. 2007.

13. Aebig JA, Mullen GED, Dobrescu G, Rausch K, Lamber L, Ajose-Popoola O, et al. Formulation of vaccines containing CpG oligonucleotides and alum. J Immunol Methods 2007;323:139–46.

14. Kritsch CE, Berger A, Heinrich-Cseh C, Bugajska-Schretter A, Zauner W. Separation and quantification of a novel two-component vaccine adjuvant. J Chromatogr B. 2005;822:263–70.

15. Dupuis M, Murphy TJ, Higgins D, Ugozzoli M, van Nest G, Ott G, et al. Dendritic cells internalize vaccine adjuvant after intramuscular injection. Cell Immunol. 1998;186:18–27.