RESEARCH ARTICLE – Pharmaceutical Biotechnology

Antibody Responses in Mice to Particles Formed from Adsorption of a Murine Monoclonal Antibody onto Glass Microparticles MALIHEH SHOMALI,1 ANGELIKA FREITAG,2 JULIA ENGERT,2 MICHAEL SIEDLER,3 ZEHRA KAYMAKCALAN,3 GERHARD WINTER,2 JOHN F. CARPENTER,4 THEODORE W. RANDOLPH1 1

Department of Chemical and Biological Engineering, Center for Pharmaceutical Biotechnology, University of Colorado, Boulder, Colorado 80303 2 Department of Pharmacy, Pharmaceutical Technology and Biopharmaceutics, Ludwig-Maximilians-University of Munich, Munich D-81377, Germany 3 AbbVie Bioresearch Center, 100 Research Drive, Worcester, Massachusetts 01605 4 Department of Pharmaceutical Sciences, Center for Pharmaceutical Biotechnology, University of Colorado Anschutz Medical Campus, Aurora, Colorado 80045 Received 13 August 2013; revised 2 October 2013; accepted 14 October 2013 Published online 13 November 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23772

ABSTRACT: Immunogenicity of therapeutic monoclonal antibodies (mAbs) is a concern because of the effects of anti-drug antibodies (ADAs) on therapeutic efficacy. Particulate matter has been suggested as a potential contributing factor to immunogenicity. In this study, we investigated ADA levels in mice in response to administration of a murine immunoglobulin G (IgG)2c/␬ mAb (mAb1) that was generated in C57BL/6J mice. Particles of mAb1 were formed by adsorbing the protein to glass microparticles. Formulations containing microparticles were administered subcutaneously to mice of either the syngeneic strain, C57Bl/6J, or the allogeneic strain, BALB/c. ADA levels were measured using an isotype-specific enzyme-linked immunosorbent assay method. Whereas BALB/c mice showed strong IgG1 and IgG2b responses against both the particulate and native mAb1 samples, adsorption of mAb1 to particles rendered it slightly more immunogenic than its native, soluble form. In BALB/c mice, immunoglobulin M (IgM) was produced after the first week of injections and then faded gradually. In contrast, C57BL/6J mice showed moderate IgM, IgG1, IgG2b, and IgG3 responses to injections of glass particle-adsorbed mAb1. ADA responses were higher in the allogeneic BALB/c mice, which do not produce mAbs of the IgG2c/␬ isotype. Thus, the presence C 2013 Wiley Periodicals, Inc. and the American of both foreign epitopes and particles may be important in inducing ADA responses.  Pharmacists Association J Pharm Sci 103:78–89, 2014 Keywords: protein; aggregates; immunogenicity; microparticles; adsorption; adjuvant effect; anti-drug antibody (ADA); monoclonal antibody (mAb); ELISA assay; infrared spectroscopy

INTRODUCTION Therapeutic monoclonal antibodies (mAbs) are used for treatment of numerous diseases and conditions. A major concern with these types of drugs is that they may stimulate immune responses in patients, leading to the formation of anti-drug antibodies (ADAs).1,2 Despite the development of fully humanized mAbs, there can be a high frequency of ADAs in patients.3–5 ADAs may block a drug’s activity and decrease its efficacy and/or lead to rapid clearance from the blood.1,6–11 The causes for immunogenicity of mAbs and other therapeutic proteins are not well understood, but it has been suggested that protein aggregates or other particulate contaminants may play a critical role.1,12–15 For decades, the immunogenic potential of aggregated proteins has been studied in human clinical studies and in animal experiments. For example, patients receiving antibody therapies from non-human sources typically show an immune response that leads to rapid clearance of the therapeutic. Abbreviations used: mAb, monoclonal antibody; ADA, anti-drug antibody; ELISA, enzyme-linked immunosorbent assay; BGG, bovine gamma-globulin; PBS, phosphate-buffered saline; Trp, tryptophan; TD, T-dependent. Correspondence to: Theodore W. Randolph (Telephone: +303-492-4776; Fax: +303-492-4341; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 103, 78–89 (2014)

 C 2013 Wiley Periodicals, Inc. and the American Pharmacists Association

78

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

However, Weksler et al.16 demonstrated that patients could be tolerized to horse anti-human lymphocyte immunoglobulin G (IgG) if particles and aggregates were removed prior to administration by preparative ultracentrifugation. Similarly, Ring et al.6 showed that removal of aggregates reduced immune response in humans. They showed that heat-aggregated human serum albumin could induce anaphylactic reaction in dogs, whereas ultracentrifuged solutions were well-tolerated.6 They also showed that ultracentrifuged horse anti-human lymphocytes globulin were well-tolerated in human patients, whereas the aggregate-containing solutions were not.6 They also observed the same result when patients were administered solutions containing aggregated human serum albumin.6 Also, numerous cross-species studies in animals have documented that immunogenicity to IgG’s could be modulated by increasing or decreasing the aggregate and particle contents of samples. Dresser17 succeeded in inducing immunological tolerance in adult CBA mice by injection of bovine gamma-globulin (BGG) that had been centrifuged to remove particulate matter. Biro and Garcia18 observed the same result when they tested aggregated and aggregate-free (obtained by preparative ultracentrifugation) human gamma-globulin (HGG) in rabbits. They reported that heat-aggregated HGG is an excellent antigen for rabbits, whereas aggregate-free HGG is not capable of inducing the primary response. Gamble19 and Sassen et al.20 studied

RESEARCH ARTICLE – Pharmaceutical Biotechnology

the immunogenicity of HGG in mice for the initiation of primary and secondary antibody responses. They reported that heat-treated, but not untreated, HGG has potential to initiate a primary antibody response in mice. In the work of Sassen et al.,20 no difference was reported between heat-treated and untreated HGG for their capacity to induce a secondary response; however, at limited antigen dose levels, heat-treated HGG was observed to induce larger secondary responses than untreated HGG.20 One mechanism by which particles of therapeutic proteins may form is by adsorption to microparticulate contaminants; these may derive from several sources. For example, Tyagi et al.21 showed that stainless steel nanoparticles shed from filling pumps can nucleate particles of mAbs. Also, glass nanoparticles and microparticles have been shown to induce protein particles from mAbs and other therapeutic proteins.22 Glass particles shed from containers (vials, syringes, and cartridges) have been responsible for numerous recent product recalls because of the visible particulate matter in drug products, including therapeutic proteins.23,24 Moreover, one recent study showed that glass particles can be potent adjuvants. Fradkin et al.25 demonstrated that murine growth hormone adsorbed to microparticles obtained from ground glass syringes (nonsiliconized) stimulated a robust immune response in mice, and that particle-free samples were not immunogenic. Van Beers et al.26 found that human interferon-beta adsorbed to the same glass particles slightly enhanced immunogenicity of the protein in transgenic mice. Considering the immunogenicity potential of therapeutic antibodies and the potential role of pharmaceutically relevant glass particles as adjuvants, our first goal in this study was to test the hypothesis that immunogenicity of a monoclonal antibody would be enhanced by adsorption to glass microparticles. For comparison, the mAb1 was adsorbed onto aluminum hydroxide, an adjuvant commonly used in vaccine formulations. To conduct this investigation, we chose a murine monoclonal IgG2c/6 (mAb1) generated in C57BL/6J mice, which would be expected to be minimally immunogenic in its soluble native form when administered to mice of this same strain. The second goal of this study was to investigate the immunogenicity of mAb1 in BALB/c mice, which do not produce IgG2c/6 antibodies. It has been shown to be very challenging to make BALB/c mice unresponsive to ultracentrifuged human or BGG compared with C57BL/6J mice.27–29 It was reported by Golub and Weigle29 that when even an extremely small amount of ultracentrifuged HGG (as low as 50 :g) was introduced to C57BL/6J mice, it could induce immunological tolerance, whereas to make BALB/c mice tolerant, a dose as high as 10 mg was needed. They postulated BALB/c mice may process the trace amount of aggregates remaining in samples efficiently, whereas C57BL/6J mice may not.29 However, when the trace amount of aggregates was removed by salt fractionation, mice from both strains became unresponsive to small doses of HGG. Similar results were observed by Das and Leskowitz28 when single doses of ultracentrifuged BGG were administered to BALB/c and DBA/2 mice. In that study, they showed that DBA/2 mice became tolerant at doses of 0.2 mg, whereas BALB/c mice required more than 20 mg to induce tolerance. Also, BALB/c mice do not have a gene to produce the IgG2c immunoglobulin isotype.30,31 There is a 16% difference in amino acid sequences between IgG2c and IgG2a.30 Therefore, we should be able to

DOI 10.1002/jps.23772

79

avoid interference between injected IgG2c and the endogenous immunoglobulin in the bioanalytical assay. Samples of adsorbed mAb1 on glass microparticles were administered by subcutaneous injections in the two mouse strains. Sera were collected and evaluated for levels of immunoglobulin M (IgM) and G (IgG1, IgG2b, and IgG3) ADAs against mAb1. An enzyme-linked immunosorbent assay (ELISA) coupled with an acid-dissociation step22 was used for ADA detection.

MATERIALS AND METHODS Materials All chemicals used in this work were of reagent grade or higher quality. Sterile water for injection was used entirely and all materials used for injection were of USP grade. mAb1, a mouse monoclonal antibody (IgG2c/6, 145 kDa) against tumor necrosis factor alpha (TNF-a), was provided by AbbVie Bioresearch Center (Worcester, Boston, Massachusetts). HRP-goat antimouse IgM, IgG1, and IgG2b were purchased from Jackson ImmunoResearch Laboratories Inc. (West Grove, Pennsylvania) and HRP-rabbit anti-mouse IgG3 was purchased from Fitzgerald Industries International Inc. (Acton, Massachusetts). Alum adjuvant (Alhydrogel ) was purchased from Brenntag Biosector (Frederikssund, Denmark), L-Histidine was obtained from RPI (Prospect, Illinois), sucrose was obtained from Sigma– Aldrich (St. Louis, Missouri), citric acid, tri-sodium salt dehydrate was obtained from ACROS ORGANICS (Fair Lawn, New Jersey), and sulfuric acid was obtained from Mallinckrodt (Hazelwood, Missouri). The other chemical reagents were purchased from Fisher: acrylamide from Fisher BioReagents (Pittsburgh, Pennsylvania), urea from Fisher Scientific (Fair Lawn, New Jersey), sodium chloride, polysorbate20, and phosphate-buffered saline (PBS; 10× solution, DNase–RNaseand protease-free, 1.37 M sodium chloride, 0.027 potassium chloride, and 0.119 phosphate buffer) from Fisher Scientific (Pittsburgh, Pennsylvania), and TMB from Thermo Scientific (Rockford, Illinois). R

METHODS Sample Preparation and Characterization

Model mAb Monoclonal IgG2c/k antibody (mAb1) was received as a frozen solution at a concentration of 24 mg/mL in 15 mM histidine buffer (pH 6.0). Size-exclusion chromatographic analysis of the thawed solution showed that the mAb1 preparation contained 97.3% monomer, 2.1% low-molecular-weight species, and 0.6% high-molecular-weight species. The original thawed sample was aliquotted into 1-mL tubes under aseptic conditions and stored in a freezer at −80◦ C until further use.

Preparation and Characterization of Ground Glass Microparticles Glass shards from vials (5cc, Type 1 Glass, USP/PhEur, nontreated; Schott Inc., Syracuse, New York) were ball-milled with zirconia (ZrO2 ) medium, and the resulting powders were sieved through a 45-:m screen. The sieved particles were washed with 1 M ammonium sulfate and thoroughly rinsed with purified water to dissolve and wash away exposed oxides.32 Finally, the particles were dried under vacuum at 100◦ C for 1 h, and the

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

80

RESEARCH ARTICLE – Pharmaceutical Biotechnology

dried cake was gently broken with a mortar and pestle. A suspension of glass microparticles (100 mg/mL) in 20 mM histidine buffer pH 5.7 was used for determination of particles size and adsorption studies. Particle size distributions of glass microparticles suspended in 20 mM histidine buffer (pH 5.7) were measured in triplicate using a Beckman Coulter LS230 laser diffraction instrument (Beckman Coulter Inc., Miami, Florida). Particle-specific surface area was determined for the dried glass powder using nitrogen adsorption and Brunauer– Emmett–Teller (BET) isotherm analysis using Micromeritics Gemini V (Micromeritics Instrument Corporation, Norcross, Georgia).

Preparation of Adsorbed mAb1 Suspensions of 500 :L total volume were prepared in 20-mM histidine buffer pH 5.7 containing 7.5 mg glass microparticles and various concentrations of mAb1 (ranging from 0.08 to 0.4 mg/mL). The samples were gently rotated at 8 rpm for 30 min at room temperature to allow the mAb1 to adsorb onto glass microparticles. Then, samples were centrifuged for 30 min at 15,000g. The mAb1 concentration in the supernatant was determined by UV absorbance, and a mass balance was applied, with the fraction of mAb1 remaining in solution subtracted from original concentration to obtain the amount of mAb1 adsorbed on the glass particles. An adsorption isotherm curve was constructed by plotting the amount of mAb1 adsorbed on particles (mg/m2 ) versus the concentration of mAb1 remaining in the solution (mg/mL).33

Reversibility of mAb1 Adsorption In Vitro To determine the extent to which adsorbed mAb1 might desorb from microparticles in vivo, an in vitro incubation was used to mimic in vivo conditions. A sample of mAb1 (0.25 mg/mL) was incubated with a sufficient amount of glass microparticles to be fully adsorbed. The suspension with adsorbed mAb1 was centrifuged (30 min at 15,000g), and the supernatant was removed. Then, to mimic the pH and salt content found in vivo, the pellet was resuspended in PBS and diluted in three times the initial sample volume . Samples were incubated for another 30 min with rotation at 8 rpm and then centrifuged again. The mAb1 concentration in the supernatant was measured by UV absorption at 280 nm, and by applying a mass balance, the reversibility of mAb1 adsorption on glass particles in PBS was calculated.

Preparation of Samples Containing Adjuvant To adsorb mAb1 onto aluminum hydroxide adjuvant, a mass ratio of 2:3 (g/g) for mAb1–adjuvant was used. Samples were incubated for 30 min by rotating gently at 8 rpm to allow sufficient time for mAb1 to adsorb. To ensure that 30-min incubation is sufficient for mAb1 to adsorb, a solution of adsorbed mAb1 onto Alhydrogel was centrifuged for 30 min at 15,000g and the UV absorbance at 280 nm of the supernatant was measured to quantify any remaining antibody. The UV analysis confirmed that essentially all mAb1 was adsorbed on Alhydrogel . R

R

Infrared Spectroscopy: Characterization of Secondary Structure of Adsorbed mAb1 A sufficient amount of glass particles (98.6:1 glass–mAb1, w/w, ratio) was added to a 1 mg/mL solution of mAb1 to adsorb esShomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

sentially all of the protein. After gently rotating the sample for 30 min at 8 rpm, the particles were pelleted by centrifugation for 30 min at 15,000g, and the mAb1 content of the supernatant was measured by UV absorption at 280 nm, which showed no detectable mAb1 remaining in the supernatant. Then, the mAb1 particle pellets were resuspended gently using a pipette tip in 20–30 :L of 20 mM histidine buffer (pH 5.7). Secondary structure of the adsorbed protein was analyzed using infrared spectroscopy of the amide I band.34–36 A BioCell sample cell with CaF2 windows and a 6.5 :m path length was used. Spectra were obtained on a Bomem MB-series spectrometer (Quebec, Canada) from 4000 to 400 cm−1 , with 256 single-beam scans at 4 cm−1 resolution. A spectrum for 20 mM histidine buffer was subtracted from each protein sample spectrum, and water vapor noise was subtracted.34 Then, the second derivative of the spectrum for each sample was calculated as described in Bee et al.32 Triplicate samples were analyzed, and the mean value of second derivative of the spectra was reported. Because the cell had a path length of 6.5 :m, we were limited to measuring protein adsorbed onto finer particles. Therefore, samples of mAb1 adsorbed on glass were prepared using the fine fraction of glass microparticles obtained after allowing larger particles to settle under gravity from a suspension for 5 min. To unfold mAb1 for infrared spectroscopy analysis, 1 mL of a solution containing 1 mg mAb1 was loaded in a centrifuge tube and heated in a water bath at 80◦ C for 10 min.

Front-Face Acrylamide Fluorescence Quenching: Characterization of Tertiary Structure of Adsorbed mAb1 Front-face fluorescence quenching was used to monitor the tertiary structure of mAb1 after adsorption onto microparticles.37–39 A sample of mAb1 at 0.1 mg/mL was adsorbed onto glass microparticles by gently mixing for 30 min at 8 rpm using a rotary mixer. The sample was loaded in a cuvette and front-face fluorescence quenching was monitored in an SLM-Aminco Spectrofluorometer (SLM-Aminco, Urbana, Illinois) as described by Bee et al.32 Fluorescence of a protein relies mostly on emission from tryptophan (Trp) residues. Trp residues adsorb light at 280 nm or a longer wavelength at 295 nm and emit light at higher wavelength usually around 300–350 nm. The emission peak for Trp residues that are buried in the hydrophobic core of protein is shifted by 10–20 nm from the peak for Trp that exposed to the solvent. Therefore, a wavelength of 295 nm was used to excite the Trp fluorescence, which was monitored at 328 nm for both native and adsorbed mAb1, and at 348 nm for mAb1 unfolded in 10 M urea. Acrylamide was used to quench the intrinsic Trp fluorescence of mAb1, both in its soluble, native state and when fully bound to particles. To unfold mAb1 for fluorescence analysis, a sample containing mAb1 at 0.1 mg/mL was prepared in 10 M urea solution and incubated at room temperature overnight. The quenching data were averaged (two to four replicates) then analyzed using the Stern–Volmer equation: F0 = 1 + K SV [Q] F here, F0 and F are the fluorescence intensities in the absence and presence of quencher, respectively, and [Q] is the quencher concentration (M). The Stern–Volmer constant (KSV ) indicates the accessibility of the quencher to Trp. It is obtained from the initial slope of the line (first seven points) when F0 /F is DOI 10.1002/jps.23772

RESEARCH ARTICLE – Pharmaceutical Biotechnology

81

All mice were in the range of 17.5–21 g and were weighed prior to each bleeding.

followed by another three cycles of washing. A 4-:L serum sample was preincubated in 155-:L acetic acid (300 mM) for 1 h at room temperature.45 After pH adjustment to 7.4 with 77.5 :L 1 M Tris pH 9.5, 100 :L of the solution was pipetted into the first row of the plate, and 100 :L of the same solution mixed with 100 :L dilution buffer (14.4 mM citric acid tri-sodium salt dehydrate, 1.1 M sodium chloride, and 0.1% polysorbate20, pH 7.4) in the second row of the plate. The serum was serially diluted in dilution buffer from the second row down the plate and was incubated for another 1–2 h at room temperature. Next, HRP-conjugated goat anti-mouse antibodies of subclasses IgM, IgG1, IgG2b, and rabbit anti-mouse IgG3 antibody were applied. The plate wells were washed five times. Then, goat or rabbit anti-mouse antibodies, which had been diluted 1:15,000 in a solution of 90% dilution buffer + 10% blocking reagent, were allowed to incubate in the wells with shaking for 1–1.5 h at 600 rpm. After washing the plate wells for five cycles, 1step Ultra TMB was added as a substrate. After incubation for 10–15 min, the reaction was quenched with 0.5 M sulfuric acid. Absorbance at 450 nm, which is proportional to the antimAb1 antibody concentration, was measured using a MAXline microplate Reader (Molecular Devices Corporation, Sunnyvale, California).

Preparation of Samples for Injection into Mice

Data Analysis

For the first week of injections, all samples were prepared at the beginning of the week. For the next 3 weeks, all samples were prepared freshly the day before injection. mAb1 was diluted from the original sample in a filtered 20 mM histidine plus 8% (w/v) sucrose pH 5.7 to achieve a protein concentration of 0.25 mg/mL. All samples were tested for endotoxin levels using LAL Endotoxin assay according to the protocol in the kit (LONZA, Walkersville, Maryland). Samples injected had an endotoxin level lower than 0.1 EU/mL.40

The ELISA data were analyzed and reported as antibody titers for the different immunoglobulin isotypes (IgM, IgG1, IgG2b, and IgG3). The titer is the greatest dilution above the cutoff point at which ADAs can be still detected. Cutoff values were calculated as the mean value of the absorbance at 450 nm for preimmunized serum at day 1 plus two times the standard deviation of the absorbance values. For graphical presentation, absorbance values equal to or lower than the cutoff value were reported as titer 10 on the log scale.

plotted versus quencher concentration [Q]. KSV values for mAb1 adsorbed on particle surfaces were compared with KSV values for native and unfolded mAb1 in solution. Animal Study

Mouse Strains Immunogenicity of native and glass adsorbed mAb1 was tested in two mouse strains, BALB/c and C57BL/6J.

Animal Protocol Animal experiments were approved by the University of Colorado IACUC Institutional Animal Care and Use Committee under protocol number 09-05-RAN-02. Female BALB/c or C57BL/6J mice at the age of 8 weeks were purchased from Charles River Laboratories (Wilmington, Massachusetts). Four or five mice per cage were kept with available food and water ad libitum. After 1 week of acclimation, groups of five or eight mice were treated according to the schedule below.

Mouse Weight

Statistical Analysis

Injection and Bleeding Schedule Mice were injected subcutaneously in the scruff of the neck with 200 :L of sample containing 50 :g mAb1 four times in the first week (on days 1, 2, 4, and 6) followed by one injection per week for the next 3 weeks (on days 13, 20, and 27).13,41,42 Blood was collected from retro-orbital venous sinus before each injection on days 1, 6, 13, 20, 27, and on day 71 after 6 weeks of recovery. Blood samples were collected in centrifuge tubes and kept on ice during the procedure. Later, blood samples were centrifuged for 10 min at 3000g (4◦ C), and the serum samples obtained were divided into smaller aliquots and stored at −80◦ C until further analysis. A group of mice was injected with 20 mM histidine buffer plus 8% (w/v) sucrose (pH 5.7) as a placebo. Also, mAb1 adsorbed on the surface of Alhydrogel adjuvant microparticles was injected to test the effect of administering mAb1 in combination with a well-known vaccine adjuvant.43,44 R

Indirect ELISA Assay The production of antibodies with specificity to mAb1 was measured by indirect ELISA assay. In this technique, a plate was coated with 150 :L/well mAb1 (10 :g/mL) and incubated overnight (20–22 h) at 4◦ C. After washing three times with washing solution (0.05% polysorbate 20 in PBS), the plate was treated with 250 :L/well blocking reagent [1% bovine serum albumin (BSA) in PBS] and kept at room temperature for 2 h DOI 10.1002/jps.23772

All statistical analyses were performed using SigmaPlot software, version 12.0 (Systat Software, Inc. San Jose, CA). Nonparametric Mann–Whitney analysis was used to compare the isotypic anti-mAb1 antibody titer responses between each group that received the various mAb1 formulations. Titers were compared on day 71, except for the IgM response in BALB/c mice, for which titers were compared on day 6. Nonresponders (serum samples with undetectable anti-mAb1 titers) were assigned a value of zero for statistical significance determination. The antibody titers resulting from administration of mAb1 adsorbed on glass or on Alhydrogel were compared with titers from mice that had been administered with native mAb1 as well as with each other, and the p values are presented in Table 1. Differences in immune responses were considered to be significant when p values of less than 0.05 were obtained. R

RESULTS Sample Preparation and Characterization The diameter-based size distribution measured by laser diffraction ranged from 1 to 2 :m for glass particles and 0.09 :m for Alhydrogel particles. The surface area of the glass particles measured by BET analysis was 4.06 ± 0.03 (m2 /g). An adsorption isotherm was created by plotting the amount of mAb1 adsorbed on glass particles (mg/m2 ) versus the R

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

82

RESEARCH ARTICLE – Pharmaceutical Biotechnology

2.5

2D derivative (AU)

Adsorbed protein (mg/m 2 )

3

2 1.5 1 0.5 0 0

0.05

0.1 0.15 0.2 Soluble protein(mg/mL)

0.25

0.3

1720

1700

1680

1660

1640

1620

1600

1580

Wavenumber (cm−1 )

Figure 1. Adsorption isotherm curve of mAb1 on glass microparticles.

concentration of mAb1 remaining (mg/mL) in the solution after adsorption (Fig. 1). The data were fitted using a Langmuir adsorption isotherm that assumes that adsorption is a reversible process and that the adsorbed layer forms a monolayer. The monolayer surface coverage value was determined to be 2.75 mg/m2 , which is in the range of the predicted monolayer surface coverage value (0.91–3.7 mg/m2 ).32 However, in order to have essentially no free protein in the solution after adsorption, sufficient glass to provide a protein mass to glass microparticles surface area of 2.5 mg/m2 (∼10% less than 2.75 mg/m2 ) was used when adsorbing mAb1 onto glass microparticles. In addition, the sharp initial slope of the isotherm indicated that a high affinity exists between mAb1 and glass microparticle surface.33 This is likely because of the interaction between positively charged mAb1 (at pH 5.7) and the negatively charged glass particle surface. The > potential of the glass microparticles in 20 mM histidine buffer was reported previously32 to be −41.7 ± 1.76 mV. Also, the adsorption of mAb1 on glass microparticles was observed to be partially reversible. When collected pellets of adsorbed mAb1 (as above) were resuspended in three times the initial volume of solution using PBS, 70 ± 0.008% of mAb1 desorbed into the media. It was assumed that the shift in pH from 5.7 (20 mM histidine buffer) to 7.4 (PBS) weakened the binding of mAb1 to the glass microparticles.

Figure 2. Second-derivative transmission infrared spectra of native mAb1 (solid line), mAb1 adsorbed on glass microparticle (dotted line), mAb1 adsorbed on AlhydrogelR (dashed line), and boiled mAb1 (dashdotted line).

Analysis of Tertiary Structure of the Adsorbed mAb1 Using Trp Quenching Fluorescence Figure 3 shows the Stern–Volmer plots for native mAb1, mAb1 adsorbed on glass, mAb1 adsorbed on Alhydrogel , and mAb1 unfolded in 10 M urea. The initial slope of the Stern–Volmer plot at low concentrations of acrylamide was used to obtain the KSV .37,38 This value reflects the degree of exposure of the Trp residues of mAb1 to solvent and thereby reveals whether mAb1 unfolded after adsorption onto microparticles. The Stern–Volmer plot for unfolded mAb1 has a much steeper slope (KSV = 8.83±0.28) that of native mAb1 (KSV = 4.65 ± 0.19), reflecting greater solvent exposure of Trp residues in the unfolded protein. The KSV values for adsorbed mAb1, either on glass (KSV = 4.47 ± 1.39) or on Alhydrogel (KSV = 5.32 ± 0.50), are close to the value for native mAb1, suggesting that the tertiary structure of mAb1 did not change significantly upon adsorption. These results were consistent with those reported by Bee et al.32 and Hoehne et al.,46 who reported minimal structural changes for mAbs adsorbed to surfaces of glass microparticles. R

R

Analysis of Secondary Structure of Adsorbed mAb1 Using Infrared Spectroscopy Figure 2 shows the second-derivative infrared spectra for native mAb1, mAb1 adsorbed on glass microparticles, mAb1 adsorbed on Alhydrogel , and mAb1 unfolded by heating for 10 min at 80◦ C in a water bath. The band at 1640 cm−1 is characteristic of the native $-sheet secondary structure of IgG antibodies.35,36 Comparing the spectrum of mAb1 adsorbed on glass with the spectra of native and heated samples, we infer that the model mAb1 retains near-native secondary structure when adsorbed on glass microparticles. Also, the secondary structure of mAb1 was not changed after adsorption to Alhydrogel . In contrast, the spectrum of heated mAb1 showed a strong loss of $-sheet structure (1640 cm−1 ) and the concomitant appearance of a new band at 1625 cm−1 , which is attributed to intermolecular $-sheet structures.36 In addition, the small new band that was formed at 1630 cm−1 for mAb1 after adsorption on glass microparticles can be assigned to intermolecular $-sheet structure.35

7 6 5

R

F0 /F

4 3 2 1

R

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

0 0

0.1

0.2

0.3

0.4

0.5

Acrylamide (M)

Figure 3. Stern–Volmer plots for the acrylamide quenching of native mAb1 (diamonds), mAb1 adsorbed on glass microparticles (circles), mAb1 adsorbed on AlhydrogelR (squares), and unfolded mAb1 in 10 M urea (triangles). DOI 10.1002/jps.23772

RESEARCH ARTICLE – Pharmaceutical Biotechnology

83

Figure 4. Titers of IgM in BALB/c mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse.

Animal Study All mice were inspected visually during the study. Nothing unusual was observed. Over the course of the study, the mice gained weight and appeared healthy. By day 71, BALB/c mice gained up weight to 30% and C57BL/6J gained up to 60%. Administration of 20 mM histidine buffer did not induce significant immune response in either mouse strain. Some weak responses detected in the group of mice injected with 20 mM histidine buffer are obviously nonspecific responses as they were not exposed to mAb1 (Figs. 4a, 5a, 6a, 7a, 8a, 9a, 10a, and 11a).

Immunogenicity of mAb1 Formulations in BALB/c Mice In BALB/c mice, the levels of IgM increased slightly after the first week of injections of samples containing mAb1, and then

faded away after a week, presumably as a result of class switching to IgGs47 (Fig. 4). In the group injected with native mAb1, 50% of mice responded (Fig. 4b). The number of responders was highest (100%) in the group that received adsorbed mAb1 on Alhydrogel (Fig. 4d). Significant differences in the IgM response (p = 0.002) were detected between the group that was administered mAb1 adsorbed on Alhydrogel (Fig. 4c) and the group that had received injections of native mAb1. However, the differences between IgM levels in the group of mice that received mAb1 adsorbed on glass microparticles and the group that received native mAb1 were not significant (p = 0.105). Immunoglobulin IgG1 was produced in a few of the BALB/c mice that received native or adsorbed mAb1 samples shortly after the first week of injections. Then, IgG1 developed in all eight mice in those groups and persisted until day 71 (Figs. 5b, 5c, and 5d). The ADA level detected for IgG1 seemed R

R

Figure 5. Titers of IgG1 in BALB/c mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse. DOI 10.1002/jps.23772

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

84

RESEARCH ARTICLE – Pharmaceutical Biotechnology

Figure 6. Titers of IgG2b in BALB/c mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse.

saturated, and comparison of the immune responses of the groups that received native mAb1 and the adsorbed form was difficult. Therefore, available sera were tested for IgG1 levels at higher dilution ratios. For the group of mice that received native mAb1, the IgG1 titer was 25,000. In contrast, the IgG1 titer in the groups that received adsorbed mAb1 on glass or Alhydrogel was even higher, up to 50,000. This result indicated that, in BALB/c mice, the adsorbed form of mAb1 stimulated a slightly stronger immune response than mAb1 in its native state. The IgG2b response in BALB/c mice was not as strong as the IgG1 response. However, the number of mice that responded to mAb1 adsorbed on glass microparticles or on Alhydrogel was higher than the number of mice responding to the native, soluble mAb1 formulation (Fig. 6). In terms of the observed IgG2b titers, the response observed on day 71 for animals that received mAb1 adsorbed on glass microparticles was not significantly different (p = 0.234) from that of the group that received R

R

native mAb1. In contrast, IgG2b titers were significantly higher (p = 0.001) in the group that had been administered mAb1 adsorbed on Alhydrogel . At day 71, the production of IgG3 in BALB/c mice was not significant in any of the BALB/c groups (Fig. 77). Over the course of the 71 day period, transient IgG3 responses were detected in 25% of the mice injected with native mAb1. In contrast, three of eight and eight of eight of the BALB/c mice that had been administered mAb1 adsorbed on glass and Alhydrogel , respectively, showed transient IgG3 responses. R

R

Immunogenicity of mAb1 Formulations in C57BL/6 Mice Because mAb1 was derived from C57BL/6J mice, no ADAs were expected in C57BL/6J mice when native mAb1 was injected. However, an increase was detected in the form of IgM, IgG1, IgG2b, and IgG3 subclass antibodies, mostly in mice that received adsorbed forms of mAb1 (Figs. 8–11).

Figure 7. Titers of IgG3 in BALB/c mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse. Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

DOI 10.1002/jps.23772

RESEARCH ARTICLE – Pharmaceutical Biotechnology

Table 1.

85

Calculated p Values for Anti-mAb1 Antibody Titers Between Experimental Groups

Mouse Strain

Isotype (Day of Bleed)

Formulation

C57BL/6J

IgG1 (Day 71)

Native mAb1 Adsorbed glass Native mAb1 Adsorbed glass Native mAb1 Adsorbed glass Native mAb1 Adsorbed glass Native mAb1 Adsorbed glass Native mAb1 Adsorbed glass Native mAb1 Adsorbed glass Native mAb1 Adsorbed glass

IgG2b (Day 71) IgG3 (Day 71) IgM (Day 71) BALB/c

IgG1 (Day 71) IgG2b (Day 71) IgG3 (Day 71) IgM (Day 6)

Adsorbed Glass

Adsorbed Al

0.69

0.31 0.548 0.008 0.056 0.008 0.151 0.010 0.222 1.0 1.0 0.001 0.083 1.0 1.0 0.234 0.002

0.032 0.008 0.095 1.0 0.234 1.0 0.105

Figure 8. Titers of IgM in C57BL/6J mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse.

Figure 9. Titers of IgG1 in C57BL/6J mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse. DOI 10.1002/jps.23772

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

86

RESEARCH ARTICLE – Pharmaceutical Biotechnology

Figure 10. Titers of IgG2b in C57BL/6J mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse.

Figure 11. Titers of IgG3 in C57BL/6J mice after administration of 20 mM histidine buffer (a), mAb1 in its native state (b), mAb1 adsorbed on glass particles (c), and mAb1 adsorbed on AlhydrogelR (d), on days D1 (preimmunization), D6 (after three injections), D13, D20, and D71 (end day; after 6-week recovery period). Each bar represents the titer for an individual mouse.

After the second week of injections, an increase in the IgM level was observed in a few of the mice that received adsorbed mAb1, and was detected in more mice by day 71 (Fig. 8). On day 71, insignificant differences were detected between IgM titers for the group of mice that had received native mAb1 and the group that had received mAb1 adsorbed to glass particles (p = 0.095). Titers were significantly higher in mice that received mAb1 adsorbed to Alhydrogel (p = 0.01). Only one mouse out of five showed a response to native mAb1, whereas the numbers of responders were four of five and five of five in the groups receiving mAb1 adsorbed on glass and adsorbed on Alhydrogel , respectively. For IgG1, IgG2b, and IgG3 anti-mAb1 antibody production, a similar tendency was observed (Figs. 9–11). Few mice responded to injection of adsorbed mAb1 on day 13, but later, more mice developed ADAs, which were detectable by 71 days. R

R

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

Two to three mouse in each group produced detectable levels of IgG1 in response to injections of mAb1 adsorbed on glass or Alhydrogel , but none responded in the group treated with the native form. The titers in responding mice were small, and the differences in IgG1 titers were not significant between the group of mice that received native mAb1 and the groups that received mAb1 adsorbed on glass (p = 0.69) or on Alhydrogel (p = 0.31). After administration of mAb1 adsorbed on either glass or Alhydrogel , all C57BL/6 mice showed an IgG2b response, but only 20% of mice responded to receiving the native form of mAb1. The p values between titers for the group that received native mAb1 and those that received mAb1 adsorbed on particles were 0.032 and 0.08, for mAb1 absorbed to glass and Alhydrogel , respectively. R

R

R

R

DOI 10.1002/jps.23772

RESEARCH ARTICLE – Pharmaceutical Biotechnology

No C57BL/6 mice were observed to produce IgG3 antibodies following injections of native mAb1, but significantly increased titers were observed for mice that were administered mAb1 adsorbed on glass (p = 0.008) or on Alhydrogel (p = 0.008). By day 71, all five (out of five) mice in each of the groups that were administered mAb1 adsorbed to glass or to Alhydrogel showed IgG3 responses. R

R

DISCUSSION The role of particulate adjuvants in stimulating immune responses has been studied widely. Dresser48 and Dresser and Gowland49 showed the effect of adjuvants on stimulating immune response by injecting BSA in mice or rabbits in the presence of an adjuvant. They reported that BSA by itself was not immunogenic in mice or rabbits, but an immune response was stimulated if Freund’s adjuvant was introduced along with BSA. Alum, a mixture of aluminum hydroxides, was first used as an adjuvant in 1926 and until recently, it was the only adjuvant approved in the USA for use in humans.43 Alum is known to be effective in provoking humoral immunity by inducing an antibody (Th2) response in the form of IgG1.43,44,50 Proteins can adsorb to alum and produce particulates roughly 3–4.5 :m in size.43 The mechanisms by which alum enhances the immune response is not entirely clear, but several mechanisms are cited in literature: the formation of an antigen depot in the tissue and the immunostimulation of uptake by dendritic and macrophage cells.44,51 In addition to recruiting monocytes and initiating the release of pro-inflammatory cytokines, alum can provide the second signal to activate the Nalp3 inflammasome and causes the activation of adaptive immune system in a Nalp3-dependent way.52,53 We hypothesized that other pharmaceutically relevant microparticles such as glass microparticle may be capable of acting as adjuvants in a similar fashion. It was published recently that adsorbed murine growth hormone on glass microparticles has higher potential than the native form to stimulate an immune response.25 Also, recombinant human interferon-$ was shown to be more immunogenic in wild-type mice when it was adsorbed on glass microparticles, stainless steel microparticle, or polystyrene nanoparticles.26 Classic work by Dresser54 showed that subcutaneous administration of powdered silica or sterile alum did not enhance the immune response in mice to an intraperitoneal injection of BGG administered the following day. In contrast, when BGG was adsorbed to alum and coadministered intraperitoneally to mice, a strong immune response was seen. Protein aggregates are known to have a higher potential to stimulate immune responses than their monomeric counterparts. In this study, we investigated this potential by adsorbing a murine monoclonal antibody onto microparticles and testing its effect in two mouse strains. From the in vitro study of the reversibility of adsorption of mAb1 to glass microparticles in PBS buffer, it was anticipated that at least 30% of the mAb1 injected would remain adsorbed on glass particles once administered in vivo. Furthermore, based on the results of IR spectroscopy and Trp fluorescence quenching, the secondary and tertiary structure of mAb1 after adsorption on particles was not changed significantly from that of native, soluble mAb1. mAb1 adsorbed in this fashion presumably presents multiple immunogenic epitopes similar to those of near-native structure, making the microparticles highly efficient in eliciting an immune response.1 DOI 10.1002/jps.23772

87

BALB/c mice developed anti-mAb1 antibodies after receiving any of the samples that contained mAb1, either in its native soluble state or as adsorbed on particles. The BALB/c strain is very immunologically resistant to induction of tolerance compared with other strains, and BALB/c mice do not produce IgG2c.30 Thus, even the minimal nonnative epitopes presented by administration of the IgG2c antibody mAb1 could not be tolerated in BALB/c. The anti-mAb1 responses in BALB/c mice were slightly higher when mAb1 was administered adsorbed onto glass or Alhydrogel particles than when administered in the soluble native form, presumably because the particles contributed an adjuvant effect. It is also interesting to highlight that BALB/c mice exhibited a strong IgG1 response. The increase in IgG1 level is consistent with a T-cell-dependent immune response,55–57 although confirmation of the detailed mechanism is beyond the scope of this study. As mAb1 was generated in C57BL/6J mice, it might be expected that administration of mAb1 in this strain would be tolerated and that no immune responses would occur. Indeed, native mAb1 was tolerated by majority of C57BL/6J mice. However, the adsorbed forms of mAb1 elicited immune responses even in this syngeneic strain of mouse. The response was not as strong as that observed in BALB/c (especially for IgG1 isotype), but a moderate level of ADAs was detected. C57BL/6J mice responded to adsorbed mAb1 mostly by secreting IgG3. IgG3 secretion has been reported to be the characteristic of Tindependent B cell responses,58,59 in which the levels of IgM and IgG3 are higher. When adsorbed mAb1 was injected to C57BL/6J mice, an increase in IgM level was observed and persisted to day 71. IgM responses to soluble antigens would be expected to diminish in approximately 1 week after injection as result of class switching.47 Instead, we observed a more persistent IgM response. We speculate that the microparticulate nature of the adsorbed mAb1 provided a depot effect that resulted in a persistent IgM level. It was recently shown by Filipe et al.60 that the biodistribution of aggregates differs from that of monomers when a humanized IgG1 is administered subcutaneously into mice. Using fluorescence imaging, they showed that monomeric IgG1 is distributed rapidly in the body, whereas aggregated forms remain at the site of injection for more than a month. Thus, we speculate that subcutaneous injection of mAb1 adsorbed to glass microparticles acted as an antigen depot, providing the antigen continuously to the newly arriving immune system cells up to 10 weeks after the initial wave of ADA production. This mechanism was not observed in the BALB/c strain, most likely because a T-dependent (TD) B cell activation. In TD pathway, high-affinity antibodies with specificity to antigen are produced. Also, as the result of TD response, some B memory cells are formed that can survive for longer periods, and then specific antigens are cleared faster in further exposure.61 On the basis of this, ADAs that were released after the first wave of activation in BALB/c mice might circulate for longer periods of time, enabling antigen to be more rapidly cleared. This assumption was not studied in further detail. In summary, both strains showed stronger response to the administration of mAb1 when adsorbed to microparticle surfaces than to mAb1 in its soluble native form. The response in BALB/c mice was stronger than in C57BL/6J mice, particularly in IgG1 isotype secretion. All BALB/c mice produced antibodies, even against native mAb1, most likely because of the “foreign” characteristic of the sample. In addition, the response was slightly higher to adsorbed forms of mAb1. R

Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

88

RESEARCH ARTICLE – Pharmaceutical Biotechnology

The mechanism(s) of immune response in C57BL/6J mice, the host for mAb1 generation, are not entirely clear, but adsorbed mAb1 may present a new arrangement of epitopes mimicking TI type II antigens that have higher potential than mAb1 in its soluble native form for eliciting immune responses. This assumption is in agreement with the antigen model that was proposed by Dintzis and Vogelstein.62 They proposed several criteria that an immunon capable of stimulating B cell activation in a T-independent manner would meet, including a MW exceeding 100 kDa and arrangement of epitopes in a repeating fashion.62 Although the precise orientation of mAb1 on the surface of the microparticles that we tested is unknown, we expect that the monolayer surface coverage of mAb1 in a near-native conformation that we observe would provide a density and spacing of epitopes sufficient to confer immunon character to the microparticles. Our result is also in agreement with the study of Joubert et al.63 In that study, the effect of aggregates of therapeutic antibodies on enhancing the innate immune response of a population of na¨ıve human peripheral blood mononuclear cells was investigated. Aggregates were generated by stirring or pumping through a siliconized syringe of human IgG antibodies, and adsorption of IgG2 on polymeric microspheres. The adsorbed IgG2, which maintained a high degree of properly folded secondary structure, induced a much higher response than any of the other aggregate types. Furthermore, adsorbed mAb1 might induce responses in the same fashion as Alhydrogel by enhancing the uptake by macrophages, or they might induce immunostimulatory cytokines leading to faster maturation of dendritic cells.44 R

CONCLUSIONS The loss of efficacy because of the formation of antibodies against mAbs is reported frequently in the treatment of patients, even when mAbs are fully humanized. In this study, ADA response was detected in mice in response to subcutaneous administration of a murine antibody adsorbed on glass and aluminum hydroxide microparticles. Glass microparticles were shown to have potential to adsorb protein and stimulate immune responses by acting as an adjuvant, even in the host mouse strain from which the mAb was derived. Although the absolute level of any risk to human patients cannot be defined in this study, the microparticle-induced enhancement of immune response to a murine antibody that we observed in mice suggests that it is prudent to minimize levels of microparticulate contaminants in formulations of therapeutic proteins.

ACKNOWLEDGMENTS We would like to thank AbbVie GmbH & Company KG, Ludwigshafen, Germany for supporting this project financially. Also, thanks to Professor Raul Torres (National Jewish Health, Denver, Colorado) for scientific discussions and suggestions in immunogenicity study. Zehra Kaymakcalan and Michael Siedler are employees of AbbVie and are Abbvie stockholders. Maliheh Shomali and Theodore W. Randolph are employees at the Department of Chemical and Biological Engineering, Center for Pharmaceutical Biotechnology, University of Colorado. Gerhard Winter, Angelika Freitag, and Julia Engert are employees at the Department of Pharmacy, Pharmaceutical Technology and Biopharmaceutics, Ludwig-Maximilians-University Munich. John Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014

F. Carpenter is an employee at the Department of Pharmaceutical Sciences, Center for Pharmaceutical Biotechnology, University of Colorado Anschutz Medical Campus. The University of Colorado and the Ludwig-Maximilians-University Munich received research funds from AbbVie (former Abbott Laboratories) to conduct the study. AbbVie (former Abbott Laboratories) provided financial support, provided the murine antibody used in this study, as well as resources to support the in vivo studies and the bioanalytical characterization. Furthermore, Abbott authors were involved in study design, research, analysis, data collection, interpretation of data, reviewing, and approving the publication.

REFERENCES 1. Rosenberg AS. 2006. Effects of protein aggregates: An immunologic perspective. AAPS J 8:E501–E507. 2. Randolph TW, Carpenter JF. 2007. Engineering challenge of protein formulations. AICHE J 53:1902–1907. 3. Aarden L, Ruuls SR, Wolbink G. 2008. Immunogenicity of anti-tumor necrosis factor antibodies—Toward improved methods of anti-antibody measurement. Curr Opin Immunol 20:431–435. 4. Anderson PJ. 2005. Tumor necrosis factor inhibitors: Clinical implications of their different immunogenicity profiles. Semin Arthritis Rheum 34:19–22. 5. Getts DR, Getts MT, McCarthy DP, Chastain EM, Miller SD. 2010. Have we overestimated the benefit of human(ized) antibodies?. MAbs 2:682–694. 6. Ring J, Seifert J, Jesch F, Brendel W. 1977. Anaphylactoid reactions due to non-immune complex serum protein aggregates. Monogr Allergy 12:27–35. 7. Pedotti R, Mitchell D, Wedemeyer J, Karpuj M, Chabas D, Hattab EM, Tsai M, Galli SJ, Steinman L. 2001. An unexpected version of horror autotoxicus: Anaphylactic shock to a self-peptide. Nat Immunol 2:216–222. 8. West RL, Zelinkova Z, Wolbink GJ, Kuipers EJ, Stokkers PCF, Van der Woude CJ. 2008. Immunogenicity negatively influences the outcome of adalimumab treatment in Crohn’s disease. Aliment Pharmacol Ther 28:1122–1126. 9. Baert F, Noman M, Vermiere S, Van Assche G, D’ Haens G, Carbonez A, Rutgeerts P. 2003. Influence of immunogenicity on the long-term efficacy of infliximab in Crohn’s disease. N Engl J Med 348:601–608. ˜ L, Bonilla G, 10. Pascual-Salcedo D, Plasencia C, Ramiro S, Nuno Nagore D, Ruiz Del Agua A, Mart´ınez A, Aarden L, Mart´ın-Mola E, Balsa A. 2011. Influence of immunogenicity on the efficacy of long-term treatment with infliximab in rheumatoid arthritis. Rheum 50:1445– 1452. 11. Plasencia C. 2012. Influence of immunogenicity on the efficacy of long-term treatment spondyloarthritis with infliximab. Ann Rheum Dis 71:1955–1960. 12. Braun A, Kwee L, Labow MA, Alsenz J. 1997. Protein aggregates seem to play a key role among the parameters influencing the antigenicity of interferon alpha in normal and transgenic mice. Pharm Res 14:1472–1478. 13. Hermeling S, Aranha L, Damen JM, Slijper M, Schellekens H, Crommelin DJ, Jiskoot W. 2005. Structural characterization and immunogenicity in wild-type and immune tolerant mice of degraded recombinant human interferon alpha2b. Pharm Res 22:1997–2006. 14. Fradkin AH, Carpenter JF, Randolph TW. 2009. Immunogenicity of aggregates of recombinant human growth hormone in mouse models. J Pharm Sci 98:3247–3264. 15. Van Beers MMC, Sauerborn M, Gilli F, Brinks V, Schellekens H, Jiskoot W. 2010. Aggregated recombinant human interferon beta induces antibodies but no memory in immune-tolerant transgenic mice. Pharm Res 27:1812–1824. 16. Weksler ME, Bull G, Schwawrz GH, Stenzel KH, Rubin AL. 1970. Immunologic responses of graft recipients. J Clin Invest 49:1589–1595. DOI 10.1002/jps.23772

RESEARCH ARTICLE – Pharmaceutical Biotechnology

17. Dresser D. 1962. Specific inhibition of antibody production. II. Paralysis induced in adult mice by small quantities of protein antigen. Immunol 5:378–388. 18. Biro CE, Garcia G. 1965. The antigenicity of aggregated and aggregate-free human gamma-globulin for rabbits. Immunol 8:411– 419. 19. Gamble C. 1966. The role of aggregates in the primary immune response of mice to human gamma globulin. Int Arch Allergy Appl Immunol 30:446–455. 20. Sassen A, Perkins EH, Brown RA. 1968. Immunogenic potency of human gamma-globulin in mice. Immunology 14:247–256. 21. Tyagi AK, Randolph TW, Dong A, Maloney KM, Hitscherich CH Jr, Carpenter JF. 2009. IgG particle formation during filling pump operation: A case study of heterogeneous nucleation on stainless steel nanoparticles. J Pharm Sci 98:94–104. 22. Freitag A, Shomali M. 2013. Detection of anti-drug antibodies in murine serum samples: The need for immune complex dissociation. Unpublished Results. 23. Wen Z, Torraca G, Masatani P, Sloey C, Phillips J. 2012. Nondestructive detection of glass vial inner surface. J Pharm Sci 101:1378– 1384. 24. US FDA. 2011. Advisory to drug manufacturers: Formation of glass lamellae in certain injectable drug. Accessed, at: http://www.fda.gov/drugs/drugsafety/ucm248490.htm on 11/03/2013. 25. Fradkin AH, Carpenter JF, Randolph TW. 2011. Glass particles as an adjuvant: A model for adverse immunogenicity of therapeutic proteins. J Pharm Sci 100:4953–4964. 26. Van Beers MMC, Gilli F, Schellekens H, Randolph TW, Jiskoot WIM. 2012. Immunogenicity of recombinant human interferon beta interacting with particles of glass, metal, and polystyrene. J Pharm Sci 101:187–199. 27. Fujiwara M. 1977. Cellular mechanisms of the resistance to the induction of immunological tolerance. Immunol 32:783–791. 28. Das S, Leskowitz S. 1974. The cellular basis for tolerance or immunity to bovine-gamma-globulin in mice. J Immunol 112:107–114. 29. Golub ES, Weigle WO. 1969. Studies on the induction of immunologic unresponsiveness: III. Antigen form and mouse strain variation. J Immunol 102:389–396. 30. Martin RM, Brady JL, Lew AM. 1998. The need for IgG2c specific antiserum when isotyping antibodies from C57BL/6 and NOD mice. J Immunol Methods 212:187–192. ¨ a¨ I, Sarvas H, P´eterfy F, Makel ¨ 31. Seppal a¨ O. 1981. The four subclasses of IgG can be isolated from mouse serum by using protein A-sepharose. Scand J Immunol 14:335–342. 32. Bee JS, Chiu D, Sawicki S, Stevenson JL, Chatterjee K, Freund E, Carpenter JF, Randolph TW. 2009. Monoclonal antibody interactions with micro- and nanoparticles: Adsorption, aggregation, and accelerated stress studies. J Pharm Sci 98:3218–3238. 33. Haynes CA, Norde W. 1994. Globular proteins at solid/liquid interfaces. Colloids Surfaces B: Biointerfaces 2:517–566. 34. Dong A, Huang P, Caughey WS. 1990. Protein secondary structures in water from second-derivative amide I infrared spectra. Biochem 29:3303–3308. 35. Buijs J, Norde W, Lichtenbelt JWT. 1996. Changes in the secondary structure of adsorbed IgG and F(ab‘) 2 studied by FTIR spectroscopy. Langmuir 12:1605–1613. 36. Giacomelli C, Bremer M, Norde W. 1999. ATR-FTIR study of IgG adsorbed on different silica surfaces. J Colloid Interface Sci 220:13–23. 37. Eftink MR, Ghiron CA. 1981. Fluorescence quenching studies with proteins. Biochem 114:199–227. 38. Eftink M, Ghiron C. 1977. Exposure of tryptophanyl residues and protein dynamics. Biochem 16:5546–5551. 39. Lakowicz JR. 1999. Principles of fluorescence spectroscopy. 2nd ed. New York: Springer. 40. Malyala P, Singh M. 2008. Endotoxin limits in formulations for preclinical research. J Pharm Sci 97:2041–2044. 41. Maas C, Hermeling S, Bouma B, Jiskoot W, Gebbink MF. 2007. A role for protein misfolding in immunoDOI 10.1002/jps.23772

89

genicity of biopharmaceuticals. J Biol Chem 282:2229– 2236. 42. Hermeling S, Jiskoot W, Crommelin D, Bornaes C, Schellekens H. 2005. Development of a transgenic mouse model immune tolerant for human interferon Beta. Pharm Res 22:847–851. 43. Wilson-welder JH, Torres MP, Kipper MJ, Mallapragada SK, Wannemuehler MJ, Narasimhan B. 2009. Vaccine adjuvants: Current challenges and future approaches. J Pharm Sci 98:1278–1316. 44. HogenEsch H. 2002. Mechanisms of stimulation of the immune response by aluminum adjuvants. Vaccine 20:S34–S39. 45. Patton A, Mullenix MC, Swanson SJ, Koren E. 2005. An acid dissociation bridging ELISA for detection of antibodies directed against therapeutic proteins in the presence of antigen. J Immunol Methods 304:189–195. 46. Hoehne M, Samuel F, Dong A, Wurth C, Mahler H, Carpenter JF, Randolph TW. 2011. Adsorption of monoclonal antibodies to glass microparticles. J Pharm Sci 100:123–132. 47. Murphy K, Travers P, Walport M. 2008. Janeway’s immunobiology. 7th ed, Garland Science, New York and London. 48. Dresser D. 1961. Acquired immunological tolerance to a fraction of bovine gamma globulin. Immunology 4:13–23. 49. Dresser D, Gowland G. 1964. Immunological paralysis induced in adult rabbits by small amount of a protein antigen. Nature 203:733– 736. 50. Petrovsky N, Aguilar JC. 2004. Vaccine adjuvants: Current state and future trends. Immunol Cell Bio 82:488–496. 51. McKee AS, Munks MW, MacLeod MKL, Fleenor CJ, Van Rooijen N, Kappler JW, Marrack P. 2009. Alum induces innate immune responses through macrophage and mast cell sensors, but these sensors are not required for alum to act as an adjuvant for specific immunity. J Immunol 183:4403–4414. 52. Eisenbarth SC, Colegio OR, O’Connor W, Sutterwala FS, Flavell RA. 2008. Crucial role for the Nalp3 inflammasome in the immunostimulatory properties of aluminium adjuvants. Nature 453:1122–1126. 53. Kool M, P´etrilli V, De Smedt T, Rolaz A, Hammad H, Van Nimwegen M, Bergen IM, Castillo R, Lambrecht BN, Tschopp J. 2008. Cutting edge: Alum adjuvant stimulates inflammatory dendritic cells through activation of the NALP3 inflammasome. J immunol 181:3755–3759. 54. Dresser DW. 1968. An assay for adjuvanticity. Clin Exp immunol 3:877–888. 55. Rosenberg YJ. 1981. The ability of nonspecific T-cell stimulators to induce helper-cell-dependent increases in either polyclonal or isotyperestricted l g production in vivo. Cell Immunol 61:416–424. 56. Mond JJ, Vos Q, Lees A, Snapper CM. 1995. T cell independent antigens. Curr Opin Immunol 7:349–354. 57. Swanson CL, Wilson TJ, Strauch P, Colonna M, Pelanda R, Torres RM. 2010. Type I IFN enhances follicular B cell contribution to the T cell-independent antibody response. J Exp Med 207:1485–1500. 58. Perlmutter RM, Hansburg D, Briles DE, Robert A, David JM. 1978. Subclass restriction of murine anti-carbohydrate antibodies. J Immunol 121:566–572. 59. Snapper CM, McIntyre TM, Mandler R PL. 1992. Induction of IgG3 secretion by interferon gamma: A model for T-cell-independent class switching in response to T cell-independent type 2 antigens. J Exp Med 175:1367–1371. 60. Filipe V. July 19-21, 2011. IgG monoclonal antibody aggregates: In vitro characterization in human serum and in vivo imaging in a mouse model. Poster presented at the Colorado Protein Stability Conference, Breckenridge, Colorado. 61. Tangye SG, Tarlinton DM. 2009. Memory B cells: Effectors of longlived immune responses. Eur J Immunol 39:2065–2075. 62. Dintzis HM, Vogelstein B. 1976. Molecular determinants of immunogenicity: The immune model of immune response. Proc Natl Acad Sci USA 73:3671–3675. 63. Joubert MK, Hokom M, Eakin C, Zhou L, Deshpande M, Baker MP, Goletz TJ, Kerwin BA, Chirmule N, Narhi LO, Jawa V. 2012. Highly aggregated antibody therapeutics can enhance the in vitro innate and late-stage T-cell immune responses. J Biol Chem 287:25266–25279. Shomali et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:78–89, 2014