Crystalline monoclonal antibodies for subcutaneous delivery

Abstract

Therapeutic applications for mAbs have increased dramatically in recent years, but the large quantities required for clinical efficacy have limited the options that might be used for administration and thus have placed certain limitations on the use of these agents. We present an approach that allows for s.c. delivery of a small volume of a highly concentrated form of mAbs. Batch crystallization of three Ab-based therapeutics, rituximab, trastuzumab, and infliximab, provided products in high yield, with no detectable alteration to these proteins and with full retention of their biological activity in vitro. Administration s.c. of a crystalline preparation resulted in a remarkably long pharmacokinetic serum profile and a dose-dependent inhibition of tumor growth in nude mice bearing BT-474 xenografts (human breast cancer cells) in vivo. Overall, this approach of generating high-concentration, low-viscosity crystalline preparations of therapeutic Abs should lead to improved ease of administration and patient compliance, thus providing new opportunities for the biotechnology industry.

Fig. 1.

Crystals of rituximab (A and B), trastuzumab (C and D), and infliximab (E and F). Crystallization protocols are described elsewhere (15). Crystals were produced in 400-μl batches with the following precipitants: 12% polyethylene glycol (PEG) 400, 1.17 M Na₂SO₄ in 100 mM Hepes buffer, pH 7.7 (A); 35% PEG 400, 0.2 M CaCl₂ in 0.1 M Hepes buffer, pH 7.5 (B); 25% PEG 400, 5% PEG 8000, 10% propylene glycol, 0.1% Tween 80 in Tris buffer, pH 8.5 (C); 20% PEG 400, 10% PEG 8000, 10% glycerol in 0.1 M Tris buffer, pH 7.0 (D); 20% PEG 300, 5% PEG 8000, 10% glycerol in 0.1 M Tris buffer, pH 7.0 (E); and 0.2 M NaCl, 10% isopropanol in 0.1 M Hepes buffer, pH 7.6 (F). Crystals were examined under an Olympus BX60 microscope equipped with a DXC-970MD 3CCD color video camera with camera adapter (CMA D2) and analyzed with image-pro plus software (Media Cybernetics, Silver Spring, MD). Samples of protein crystals were covered with a glass coverslip, mounted, and examined under ×10 magnification by using an Olympus microscope with an Olympus UPLAN F1 objective lens ×10/0.30 PH1 (phase contrast).

Fig. 2.

In vitro bioactivity of rituximab. Cultured RAJI lymphoma cells were detached, diluted to 0.5 × 10⁵ cells per ml, and added to 96-well plates (100 μl per well). (A) Direct cytotoxicity studies were performed by incubating RAJI lymphoma cells for 3 days in the presence of various concentrations of rituximab or dissolved crystals of rituximab (added in 25-μl volumes). (B) Complement-dependent cytotoxicity was determined by exposing a mixture of RAJI lymphoma cells and rituximab preparations (as in A), followed by the addition of various concentrations of human serum. (C) In vitro bioactivity of infliximab. Cultured L-929 mouse fibroblast cells were detached, diluted to 2 × 10⁵ cells per ml, and added to 96-well plates (100 μl per well). TNF-α neutralization assays were performed by incubating mouse fibroblast cells overnight in the presence of various concentrations of infliximab or dissolved crystals of infliximab. The number of viable cells was determined by using a CellTiter 96 AQ_ueous One Solution Cell Proliferation Assay kit (Promega).

Fig. 3.

Viscosity measurements of infliximab. The viscosity of soluble and crystalline suspensions of infliximab was measured by using the Cannon-Fenske (State College, PA) viscometer according to the manufacturer's instructions. Commercially available infliximab [100 mg per vial containing 500 mg of sucrose, 0.5 mg of polysorbate 80, 2.2 mg of monobasic sodium phosphate (monohydrate), and 6.1 mg of dibasic sodium phosphate (dihydrate)] was reconstituted in water to concentrations of 10, 25, 50, 100, 125, and 150 mg/ml and compared with various concentrations of crystalline suspensions. The excipients in the commercial lyophilized formulation did not contribute significantly to the viscosity of the reconstituted infliximab. Even at 150 mg/ml, the viscosity of the vehicle was only 26 centipoise (cps; 1 cps = 10^–3 Pa·sec), compared with 275 cps for soluble infliximab. For crystalline suspensions of infliximab, a 200 mg/ml solution (which was in formulation buffer containing 10% ethanol, 10% PEG 3350, 0.1% Tween 80, and 50 mM trehalose in 50 mM sodium phosphate buffer, pH 7.0) was tested for viscosity in addition to the concentrations mentioned above for soluble infliximab.

Fig. 4.

PK studies of infliximab. Approximately 100 μlofinfliximab (20 mg/ml) was administered to BBDR/Wor rats having a body weight of ≈250 g to provide a dose of 8 mg/kg. The mAb was administered in a soluble form either i.v. (▴) or s.c. (□) or as a crystalline suspension s.c. (⋄). The maximum concentration (C_max) for soluble infliximab was 257 μg/ml for the i.v. sample and 95 and 112 μg/ml for the s.c. soluble and crystalline samples, respectively. The half-life of soluble infliximab was 270 h for the i.v. sample; the half-life of the s.c. soluble and crystalline infliximab was 390 and 779 h, respectively. The total area under the curve (AUC_0-t) for soluble infliximab was 29.3 mg·h·ml^–1 for the i.v. sample and 37.5 and 42.8 mg·h·ml^–1 for the soluble and crystalline s.c. samples, respectively.

Fig. 5.

s.c. injection site analysis for trastuzumab. Injection sites were isolated at necropsy 12 h after the injection of 0.1 ml of 200 mg/ml trastuzumab crystals in a formulation buffer, which is different from the mother liquor, containing 10% ethanol and 10% PEG 3350 (800 mg/kg of body weight; A), 0.1 ml of a 7.5 mg/ml soluble commercial formulation of trastuzumab (30 mg/kg; B), or lipopolysaccharides (LPS; positive control; C). A portion of the skin at the site of injection was fixed in neutral buffered formalin oriented with the internal surface on a cassette-size square of paper, embedded in paraffin, and sectioned for histology. Paraffin sections were processed through hematoxylin/eosin stain. (All micrographs shown are ×40.)

Fig. 6.

In vivo bioactivity/efficacy of trastuzumab in nude mice. Human breast cancer BT-474 cells were used to establish tumor nodules that could be monitored by measuring their dimensions every 3–4 days with vernier calipers. Animals were placed randomly into treatment groups (n = 6). Tumor volume was calculated by the formula π/6 × [larger diameter × (smaller diameter)²]. Ab treatments were initiated when tumors became >20–30 mm³ in size in one set of animals. Crystalline trastuzumab suspensions and control solutions (containing a nonspecific IgG also at 30 mg/kg) were administered s.c. in 10 doses over 5 weeks (2 doses per week). Mice then were killed for pathological and histopathological examination and compared with the untreated animals (not injected with BT-474 cells). The mean of each experimental group is shown. Statistical significance of data comparing vehicle to anti-human epidermal growth factor receptor 2 (HER2) was determined by using one-way ANOVA software (Microsoft excel).