PASylation: a biological alternative to PEGylation for extending the plasma half-life of pharmaceutically active proteins

Abstract

A major limitation of biopharmaceutical proteins is their fast clearance from circulation via kidney filtration, which strongly hampers efficacy both in animal studies and in human therapy. We have developed conformationally disordered polypeptide chains with expanded hydrodynamic volume comprising the small residues Pro, Ala and Ser (PAS). PAS sequences are hydrophilic, uncharged biological polymers with biophysical properties very similar to poly-ethylene glycol (PEG), whose chemical conjugation to drugs is an established method for plasma half-life extension. In contrast, PAS polypeptides offer fusion to a therapeutic protein on the genetic level, permitting Escherichia coli production of fully active proteins and obviating in vitro coupling or modification steps. Furthermore, they are biodegradable, thus avoiding organ accumulation, while showing stability in serum and lacking toxicity or immunogenicity in mice. We demonstrate that PASylation bestows typical biologics, such as interferon, growth hormone or Fab fragments, with considerably prolonged circulation and boosts bioactivity in vivo.

Keywords: biologic; dosing; kidney filtration; pharmacokinetics; therapeutic protein.

Fig. 1.

Concept of PASylation. (A) Modelled structure of the PASylated Fab fragment of an antibody. Both Ig chains are colored red whereas the antigen-binding site is shown in black. Twenty-four arbitrarily selected random conformations of the PAS polypeptide tag at the C-terminus of the light chain are presented as snapshots and shown in different shades of grey in order to illustrate the fluctuating random coil-like, space-filling behavior. (B) Comparison of the chemical composition between the synthetic linear polymer PEG (potentially carrying a reactive group R′ at the left end) with its hydrophilic ether oxygen bridges and an unfolded biological polypeptide with its polar peptide groups (side chains abbreviated with R). (C) Nucleotide and encoded amino acid sequences of the building blocks for PAS#1, PAS#1P2 and PAS#5 gene cassettes obtained by hybridization of two complementary oligodeoxynucleotides, giving rise to two non-palindromic sticky ends that can be mutually ligated or cloned via a SapI restriction site. (D) Schematic representation of expression cassettes on the plasmids used in this study with coding elements and singular restriction sites indicated. All genes are under control of the chemically inducible tet^p/o while the vector backbone corresponds to pASK75 (Skerra, 1994).

Fig. 2.

Biochemical and biophysical analysis of PAS fusion proteins. (A) Analysis of the recombinant 4D5 Fab fragment and its PASylated versions, all purified from the periplasmic cell fraction of E.coli, by Coomassie-stained 12% SDS–PAGE. Lane 1, the original Fab fragment; lanes 2–4, its fusions with PAS#1(200), PAS#1(400) and PAS#1(600), respectively; lane 5, the Fab fragment carrying a PAS#1(200) polymer at both of its light and heavy chains. Samples on the right are the same but were not reduced with 2-mercaptoethanol. (B) Analysis of recombinant IFN and its PASylated versions as in (A). Lane 1, original IFN; lanes 2–4, its fusions with PAS#1(200), PAS#1(400) and PAS#1(600), respectively. (C) Isoelectric focussing (IEF) of IFN and PAS#1(600)-IFN. Lane M, IEF marker proteins; lane 1, recombinant IFN; lane 2, PAS#1(600)-IFN. Both proteins show single bands at almost the same position corresponding to pH ∼6.6, in line with the calculated pI of 6.51 for IFN. (D) Analysis of hydrodynamic volume for the purified recombinant Fab fragment and its PASylated versions by SEC (peaks normalized). Two hundred and fifty microliters (250 µl) of each protein at a concentration of 0.25 mg/ml was applied to a Superdex S200 10/300 GL column equilibrated with PBS. The arrow indicates the void volume of the column (7.8 ml). (E) Analytical SEC of IFN and its PASylated versions as in (D). (F) CD spectra of IFN and its PASylated versions. Spectra were recorded at room temperature in 50 mM K₂SO₄, 20 mM K-P_i pH 7.5 using a 0.01 cm quartz cuvette and normalized to the molar ellipticity (Θ_M) for each protein. (G) Molar difference CD spectra for the PASylated IFN versions with 200, 400 and 600 residues obtained by subtracting the spectrum of the original IFN from (F).

Fig. 3.

Biophysical characterization of PAS#1(600)-hGH. (A) Mass spectrometric analysis of PAS#1(600)-hGH by ESI-MS, confirming the calculated molecular mass of 73121.65 Da as well as precisely monodisperse composition, without any indication of prematurely terminated gene products. (B) Hydrophobicity analysis of PASylated hGH by reverse-phase (RP) HPLC. HIC of the original recombinant hGH and its PASylated version was performed on a Resource RPC column using an elution gradient from 2% v/v acetonitrile, 0.065% v/v TFA to 80% v/v acetonitrile, 0.05% v/v TFA. Both profiles show a single homogeneous peak with slightly earlier elution of PAS#1(600)-hGH (51.8% acetonitrile) in comparison with the corresponding unfused protein (52.3% acetonitrile). (C) Thermal denaturation of hGH and PAS#1(600)-hGH as determined by far UV CD measurement. Thermal unfolding of a 8.2 µM protein solution in 50 mM K₂SO₄, 20 mM K-P_i pH 7.5 was followed at a wavelength of 208 nm where maximal spectral change upon unfolding was observed and the normalized unfolded fraction, f(u), was plotted as a function of temperature. The melting temperatures (T_m) of hGH and PAS#1(600)-hGH were 85.4°C and 88.6°C, respectively.

Fig. 4.

Effect of PASylation on protein PK. (A) PK of recombinant IFN and its PASylated versions in the blood of female BALB/c mice up to 48 h post i.v. injection at a dose of 5 mg/kg body weight (b.w.). The protein concentration in plasma was quantified in a sandwich ELISA using appropriate calibration curves and plotted against the time of sampling. Data were fitted with WinNonlin ver. 6.1 assuming a bi-exponential decay. Resulting parameters are listed in Table I. (B) PK of the 4D5 Fab fragment and its PASylated forms—as well as a PEGylated version and an ABD fusion (see text)—measured as in (A). Plasma concentration values were plotted in a semi-logarithmic fashion. (C) Plot of the AUC values (normalized to the unmodified protein, left) determined from the PK analysis in (B) and of the apparent M_W from the SEC analysis in Fig. 2D (right) against the length of the PAS sequence for the PASylated Fab fragments. For comparison, the dotted line indicates the true protein mass.

Fig. 5.

In vitro/in vivo activity and plasma half-life of PASylated hGH. (A) Real-time kinetic analysis of PAS#1(600)-hGH binding to the hGH receptor as Fc chimera immobilized on a Xantec CMPD sensorchip (ΔRU ≈ 200) measured on a BIAcore 2000 instrument. The signals for various hGH concentrations are depicted as red lines while curve fits according to a 1 : 1 Langmuir model are shown in black. The resulting kinetic and affinity parameters are listed in Table I. (B) PK of hGH and its PAS#1(600) fusion in the blood of C57BL6/J mice up to 24 h post i.v. versus s.c. injection. In the case of i.v. application, WinNonlin data analysis revealed a terminal half-life of 4.42 h for PAS#1(600)-hGH, contrasting with 0.047 h for hGH. The PK profiles for s.c. administration show distinct resorption and elimination phases with a terminal half-life of 3.72 h for PAS#1(600)-hGH. (C) PD study in growth-retarded homozygous (lit/lit) mice of strain C57BL/6J (age: ca. 8 weeks, weight: 10–12 g). Four groups (N = 7 or 8) were injected s.c. starting on Day 0 with either PBS vehicle or a fixed dose of 43 nmol/kg (b.w.) of hGH or of PAS#1(600)-hGH each day or of PAS#1(600)-hGH every second day (see arrows). Body weights were measured and normalized according to the mean b.w. for each mouse during eight days before the first injection (i.e. set to 100%). These normalized values (depicted with error bars representing the standard deviation) were averaged for each group on a daily basis and linearly fitted. The slope of the straight line represents the daily increase in b.w.: PBS: 0.86 ± 0.14%/day; hGH: 1.48 ± 0.08%/day; PAS#1(600)-hGH: 3.01 ± 0.09%/day; PAS#1(600)-hGH injected every 2nd day: 2.84 ± 0.12%/day. (D) Comparison of IGF-1 biomarker concentration in the plasma of mice from (C) at the end of the experiment as measured in a sandwich ELISA.