Murine myeloblastic NFS-60 cells were purchased from Cell Line Service (Eppelheim, Germany). RPMI 1640 medium, fetal bovine serum (FBS), fetal calf serum (FCS), penicillin and streptomycin were provided by Gibco Life Technologies (Monza, Italy). Mouse interleukine 3 (IL-3) was from Sigma-Aldrich (St. Louis, MO, USA) while WST-1 reagent was from Roche (Milan, Italy). Materials for gel electrophoresis were provided by Bio-Rad Laboratories Inc. (Hercules, CA, USA). If not otherwise reported, all other reagents and chemicals were of analytical grade and provided by Sigma Aldrich-Fluka, Merck or Carlo Erba (Milan, Italy).
Recombinant Met-G-CSF (BK0023) was produced using a Bioker’s manufacturing technology at Eurogenetec S.A., Belgium, in GMP conditions by high biomass E.coli fermentation according to a previously described protein fusion technology starting from a gene coding for filgrastim bearing an N-terminal enhancer peptidic tag . Briefly, LacZ8-PNP20-Kex-1-Met-G-CSF hybrid fusion protein (where LacZ8 represents the first 8 aminoacids of the LacZ protein fragment [UniProt/SwissProt accession Q37953], PNP20 is a sequence coding for the first twenty amino acids of E.coli purine nucleoside phosphorylase  and the N-terminal methionine residue of filgrastim is immediately preceded by a sequence coding for the short flexible peptide -Glu-Ser-Ser-Met-Ser-Gly-Leu-Phe-Lys-Arg- ending with a Lys-Arg basic dipeptide which in vitro is specifically cleaved at C-terminal site of arginine by the endoprotease ss-Kex-1-C611 to leave the mature sequence of filgrastim) was expressed in E.coli as cytoplasmic inclusion bodies. After isolation of inclusion bodies, the fusion protein was dissolved in 7 M guanidine, renaturated by dilution, dialyzed and cleaved by treatment with the patented recombinant endoprotease ss-Kex1-C611 and the released filgrastim was purified to homogeneity by column chromatography to obtain a clinical grade preparation coded BK0023. Bulk BK0023 in pH 4.5 buffered solution, was sterile filtrated, formulated in glass vials containing for each milliliter 300 μg filgrastim, 50 mg sorbitol, 0.04 mg polysorbate 80, 0.59 g acetate ions 0.83 mg sodium ions and stored at 2-8°C. The reference filgrastim was Neupogen® 300 μg/ml, (Dompé Biotec, Italy) purchased from a pharmacy.
Sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and Western Blotting
SDS-PAGE was carried out in reducing and non-reducing conditions on polyacrylamide gels (4% stacking gel, 12% running gel) prepared according to Laemmli . Samples were diluted 1:1 v/v with 62.5 mM Tris–HCl-pH 6.8-25% glycerol-2% SDS-0.01% Bromophenol Blue ± 5% 2-mercaptoethanol buffer and heated for 5 min at 95°C. Ten microliters samples (3–5 μg of protein) were loaded in the gels which, after the electrophoretic migration, were stained with Coomassie Brilliant Blue R 250 and washed with 10% v/v acetic acid-15% v/v ethanol solution.
Western Blotting was carried out, after SDS-PAGE separation and electrophoretic transfer, using anti-G-CSF rabbit polyclonal antibody (PeproTech, Hamburg, Germany) as primary antibody and anti-rabbit goat antibody HRP conjugate (Santa Cruz Biotechnology, Heidelberg, Germany) as secondary antibody.
Reversed Phase-High Pressure Liquid Chromatography (RP-HPLC)
BK0023 purity and contamination with its oxidized forms were detected by RP-HPLC of 3–5 μg/10 μl samples performed on a Symmetry C18 300 Å column, 2.1 × 150 mm, 3.5 μm (Waters Corp., Milford Massachusetts, USA) with a HP 1100 liquid chromatography system (Agilent Technologies, Waldbronn, Germany) equipped with automatic sample injection and UV detection at 215 nm. The column was equilibrated in 46.3% buffer A (HPLC grade H2O containing 0.1% trifluoroacetic acid) and 53.7% buffer B (acetonitrile containing 0.08% trifluoroacetic acid). Gradient elution was carried out at a flow rate of 0.285 ml/min and 53°C as follows: from 53.7% to 58.3% buffer B in 8 min; from 58.3% to 63,3.6% buffer B in 8 min; from 63.3% to 72.7% buffer B in 6 min.
Size-Exclusion High-Pressure Liquid Chromatography (SE-HPLC)
BK0023 contamination with its aggregated forms was detected by SE-HPLC of 5 μl samples performed on a TSKgel Super SW2000, 4 μm, 4.6 × 30 cm column (Tosho Bioscience, Stuttgart, Germany) with a HP 1100 liquid chromatography system and UV detection at 214 nm. The column was equilibrated and isocratically eluted with 63 mM phospate-pH 7- 3% isopropanol mobile phase at a flow-rate of 250 μl/min and 25°C.
Amino acid sequence analysis and peptide mapping
Amino acid sequences of BK0023 and Neupogen® were compared by NH2-terminal amino acid sequence analysis of the first 15 residues by standard Edman degradation procedure on an automated Procise 610A Protein Sequencer (Applied Biosystems, Foster City CA, USA) and by peptide mapping after proteolysis in non-reducing and reducing conditions with endoprotease Glu-C from S aureus followed by RP-HPLC separation of resulting peptides and MS/MS identification. Mass spectrometry was performed in positive ion mode on a MALDI-TOF Reflex III instrument (Bruker, Bremen, Germany).
Assay of process derived contaminants
Endotoxin contamination of formulated BK0023 was assayed by Lymulus Amoebocyte Lysate (LAL) test using Pyrochrome LAL kit (Cape Code Inc, Falmouth, MA, USA), while residual E.coli proteins were detected by immunoenzymatic method assay using an E.coli HCP Elisa kit (Cygnus Technologies. Wrentham, MA, USA). Both tests were performed according to the manufacturers’ instructions.
Quantization of host/vector DNA contamination has been performed under Good Laboratory Practice conditions by an external contractor (NewLab Bioquality AG, Erkrath, Germany).
A three months stability study of formulated BK0023 was performed maintaining aliquots of product solutions at 5 ± 3°C and 25 ± 2°C. At time intervals, samples were assayed by RP- and SE-HPLC to detect the appearance of new minor degradation peaks.
In vitro cell proliferation
Cell proliferation effect of BK0023 in comparison to Neupogen® was measured on a murine myeloblastic NFS-60 cell line according to a described colorimetric method . Briefly, NFS-60 cells were cultured at a confluence of 5 × 105 cells/ml in RPMI 1640 medium containing 200 mM L-glutamine, 1 mM Na-pyruvate, 10% FBS and 33 IU/ml of mouse IL 3 at 37°C, under a 5% CO2 humidified atmosphere. After a starvation period of 17 hours without serum, 1 × 104 cells/well were seeded in triplicate in microtiter plates and incubated for 48 hours with samples in concentration ranging from 0.0015 to 5 ng/ml. At the end of incubation, 20 μl of WST-1 cell proliferation reagent were added and, after 4 hours of incubation, the formation of soluble formazan dye was estimated by reading the absorbance at 450 nm with an ELISA plate reader. EC50 of BK0023 and Neupogen were calculated from the sigmoidal concentration-response curve using the GraphPad Prism program (GraphPad Software, San Diego, CA, USA). Raw data for the parallel line method were processed according to the European Pharmacopoeia  and the results were expressed as rate of BK0023 towards Neupogen®’s potency.
Radioligand binding assay
The potencies of BK0023 and Neupogen® in displacing 125I-radiolabelled h-G-CSF (Amersham Biosciences, Amersham, UK) bound to the G-CSF receptor were determined using NFS-60 cell line as described . The experiments were carried out in duplicate by incubating 2.5 × 106 cells for 1 hour at 37°C in 150 μl of RPMI 1640 medium containing 10% FCS, 20 mM Hepes-NaOH (pH 7.3), 500 pM [125I]G-CSF and increasing concentrations of either BK0023 or Neupogen®.
After the incubation, the tubes were chilled on ice and the mixtures were layered over 400 μl of phtalate oil (dibutylphtalate: dioctylphtalate 3:2) in 1.5 ml Eppendorf tubes and centrifuged at 6,000 g for 10 min at 4°C. Following aspiration of the aqueous and organic supernatants, the pellets were solubilized and mixed with Formula-989 liquid scintillation cocktail (Perkin-Elmer, Monza, Italy). The radioactivity was determined by liquid scintillation counting in a TRI-CARB 1600TR liquid scintillation counter (Canberra-Packard, Cassina de’ Pecchi, Italy) with an efficiency of 70%. Nonspecific binding was determined in the presence of 500 nM unlabeled r-G-CSF. Assays were performed in duplicate.
BK0023 and Neupogen® concentrations were measured in rat plasma samples by an ELISA method using the commercial kit G-CSF Instant ELISA (Bender Med. System, Vienna, Austria) according to the the manifacture’s instructions.
Care and handling of animals used for preclinical studies were in accordance with the provisions of the European Economic Community Council Directive 86/209 recognized and adopted by the Italian Government with the Ministerial Decree No. 230/95-B and the NIH publication No. 85–23, revised in 1985.
Seven-nine week old Cr:CD Sprague Dawley rats and 13–18 week old New Zealand White female rabbits were purchased from Charles River Srl (Calco, Italy) and stabulated for acclimatization for about 4 and 2 weeks, respectively, under thorough observation by a veterinarian.
Local tolerability of BK0023 in rabbits
Local tolerability after a single subcutaneous, intravenous (into ear vein) or intramuscular injection of 300 μg/ml/rabbit of formulated BK0023 was studied using 6 female rabbits for each administration route.
The reaction to treatment was assessed by visual inspections on the day of treatment and on the day after in the first 3 animals of each group and for 14 days in the remaining animals. For microscopic evaluation, three animals from each group were killed 48 hours after treatment, while the remaining animals from each group were killed 14 days after dosing.
Repeated dose toxicity in rats
Rats of both sexes (10 males and 10 females for each group) were treated for 4 weeks by daily subcutaneous injection of formulated BK0023 or Neupogen® at dosage of 20, 100 and 500 μg/kg/day. Upon treatment completion, the animals were observed for a further 4 week recovery period.
All the animals were submitted to routine clinical observations (local tolerance, clinical signs, body weight, food and water consumption and ophthalmoscopic examination), standard laboratory analyses (blood chemistry, haematology and urinalysis) and post-mortem examinations (autopsy, organ weight and histology).
The pharmacodynamic parameters (white blood cells and neutrophil counting) were determined by withdrawing blood samples into tubes containing EDTA as anticoagulant before the start of treatment, at days 14 and 28 and upon completion of the recovery period. Blood samples were stored at 4°C and used for counting white blood cells (WBC) and neutrophils (ANC) on a Animal Blood Counter ABC (ABX Diagnostic, Montpellier, France) according to manufacturer’s instructions.
The toxicokinetic parameters were determined by collecting serum samples at 1, 2, 4, 8 and 24 hours post-dose on day 1 and on day 14 from all the treated animals. The serum h-G-CSF concentration was measured using an ELISA technique. The formation of anti-G-CSF antibodies was also evaluated by competitive ELISA assay to examine whether a relationship existed between antibody response and toxicokinetics.
Pharmacodynamics and pharmacokinetics in neutropenic and non-neutropenic rats
Male Crl:CD rats were allocated in 12 groups of 10 animals, while one group of 6 animals was used as a control. 6 groups of 10 animals were treated by intraperitoneal administration at day -1 with a single 50 mg/kg dose of cyclophosphamide to induce a marked increase of neutropenia, lasting for several days and of comparable severity to WHO grade III neutropenia in humans . All animals were then treated for 4 consecutive days by subcutaneous injections of 10, 30 and 100 μg/kg of BK0023 and Neupogen® while control group was administered the vehicle alone.
The pharmacodynamic parameters (WBC and ANC) were determined on blood samples withdrawn from the tail vein of the first 6 rats taken from each group at the following time points: pre-dose, 1, 2, 3, 4, 5, 6, 8, 10 and 12 days after the first dose.
In addition, pharmacokinetics was investigated by collecting serum samples at 6 time points (pre-dose, 1, 2, 4, 8 and 24 hours after dosing) on day 1 of administration from the remaining 4 animals of each treatment group. The sera were stored frozen pending the determination of G-CSF concentration with a commercial ELISA kit.
Phase I clinical study
Evaluation of pharmacodynamic equivalence and pharmacokinetics bioequivalence of BK0023 to Neupogen® were assessed by an external Clinical Research Organization in healthy male volunteers after single and multiple-dose subcutaneous administration according to an experimental protocol authorized by Canton Ticino Ethical Committee and by the Swiss Central Health Authority.
Healthy men were treated with formulated BK0023 or Neupogen® by subcutaneous injections at the following doses according to a cross-over design in 2 consecutive study periods, separated by a wash-out period of at least 28 days.
Group 1 (16 subjects): 2.5 µg/kg/day for 7 consecutive days;
Group 2 (16 subjects): 5 µg/kg/day for 7 consecutive days;
Group 3 (16 subjects): 10 µg/kg/day for 5 consecutive days.
The primary clinical phase I study endpoints were as follows:
Equivalence in terms of pharmacodynamic effects of BK0023 versus Neupogen® assessed on baseline adjusted increase of ANC.
Bioequivalence of pharmacokinetic profile of BK0023 and Neupogen® assessed as peak concentration of G-CSF (Cmax) and area under the curve (AUC0-24 hours) for BK0023 and Neupogen at both day 1 and at steady-state (last treatment day).
The secondary clinical phase I study endpoints were as follows:
Equivalence in terms of induction of CD34+ cells of BK0023 versus Neupogen® assessed on CD34+ peak concentration and AUCCD34+ versus time from day 1 to 10 (2.5 and 5 μg/kg/day dose groups) or from day 1 to 8 (10 μg/kg/day dose group)
Bioequivalence in terms of G-CSF peak concentration of BK0023 and Neupogen® at both day 1 and at steady-state (last treatment day)
Bioequivalence of pharmacokinetic parameters of BK0023 and Neupogen® at both day 1 and at steady-state (last treatment day) assessed as Tmax (time to peak concentration), t1/2 (half life) and clearance.
Safety and tolerability of BK0023 as compared to Neupogen®
The study was designed as a group sequential clinical trial in order to recalculate the correct sample size necessary to obtain statistically acceptable pharmacodynamic and pharmacokinetic results for each group, based on the intra subject variability, to achieve the primary study objectives.
Consequently, the study was conducted in two parts:
Part 1: 16 subjects per dose group were enrolled and treated with both drug products in cross-over. The pharmacodynamic and pharmacokinetic primary parameters, that is the AUC of ANC, the maximal attained ANC and the serum filgrastim AUC on day 1 and at steady state were calculated after the end of the second period of each dose group. A pharmacokineticist, independent from the clinical study management and from the sponsor, temporarily broke the blind with the aim of evaluating intra-subject variability. The biostatistician re-calculated the sample size with the actual study ad interim results in order to verify whether additional subjects were needed to obtain statistically valid results.
Part 2: On the basis of sample size re-calculation, an additional number of subjects for each dose group was enrolled: 16 subjects in the Group 1 (2.5 μg/kg/day), 20 subjects in the Group 2 (5 μg/kg/day) and 6 subjects in the Group 3 (10 μg/kg/day).
The pharmacodynamic effects of both BK0023 and Neupogen® were determined by ANC performed on whole blood samples with a fully automated laser ADVIA 120 Haematology System (Bayer) which incorporates flow cytometric principles and cytochemistry and by evaluating the count of peripheral blood CD34 positive (CD34+) hematopoietic stem cells by flow cytometric analysis as described . In fact, beside regulating granulopoiesis, G-CSF is also able to mobilize CD34+ cells from the marrow to the blood enabling the use of stem cell enriched peripheral blood in hematopoietic transplantation in lieu of bone marrow cells .
The pharmacokinetic profile of filgrastim was determined in serum samples withdrawn after the 1st and the last dose and stored frozen until the determination of G-CSF concentration with a commercial ELISA kit
Statistical analysis of clinical and preclinical data
Results were analyzed using SAS® version 9.1.2 for Windows software. The AUC of ANC and CD34+ cells and their maximal attained counts were compared based on analysis of variance for a cross-over design: the 95% CI for the ratio of the averages (population geometric means, if parameters are log-transformed) of the values for the test and reference and the two onesided t-tests of Schuirmann at the level of significance of 2.5%. Criteria for equivalence on ANC parameters were that the 95% CI of the ratio between treatment means fell within 85%-115% (or 85%-118% when parameters are log-transformed). AUC and peak concentration of serum filgrastim were compared based on Latin-square analysis of variance for a cross-over design: the 90% CI for the ratio of the averages (population geometric means) of the values for the test and reference and the two one-sided t-tests of Schuirmann at the level of significance of 5%. Criteria for bioequivalence were that the 90% CI of the treatment means ratio fell within the range of 80 to 125%. Pharmacokinetic data were calculated using Kinetica™ Version 4.4.1 software (Thermo Electron Corporation, USA); the area under the blood concentration-time curve (AUC) was calculated using the trapezoidal rule.
Experimental data of in vitro inhibition of radioactive G-CSF binding by BK0023 and Neupogen to cell expressing G-CSF receptor were analyzed by using the GraphPad Prism Program (San Diego, USA).
Information was given in both oral and written form to each volunteer. Before being admitted to the clinical study subjects expressed their consent to participate singing copy of the written informed consent form.