High-performance affinity capture-removal of bacterial pyrogen from solutions

Synthetic peptide S3Δ has high affinity for bacterial endotoxin or lipopolysaccharide (LPS). Under tested conditions of pH 5–9 and 0–0.4 M NaCl, the affinity constant, K_D ranged from 2·10⁻⁶ to 2·10⁻⁹ M⁻¹. A novel affinity matrix based on peptide S3Δ was developed for removal of LPS from solutions such as: water; buffers with a wide range of ionic strength and pH; medium for cell culture; and protein solutions under optimized conditions. At a starting LPS of ≈100 EU/ml, a post-purification level below 0.005 EU/ml was achieved.     

亲和介质及溶液条件对蛋白质溶液中内毒素去除的影响

生物制品中内毒素的去除是一项十分重要的工作。为了更好地去除各种生物制品中的内毒素,采用合成的多粘菌素B琼脂糖亲和介质,通过静态吸附的方法去除蛋白质溶液中的内毒素。重点考察了介质的间臂长度、配基密度以及各种溶液条件(pH值、盐种类和浓度、蛋白质种类和浓度、内毒素浓度、添加剂等)对内毒素去除率及蛋白质回收率的影响。分别采用动态浊度法和Lowry法检测内毒素含量和蛋白质浓度。结果表明该介质具有载量高、去除速度快、去除率高、可重复使用的特点。此外,配基密度、pH值、盐浓度和蛋白质特性(等电点和疏水性)对内毒素去除效果均有重要影响。在优化的条件下,血红蛋白、人血清白蛋白和溶菌酶的回收率分别达到87.2%、73.4%和97.3%,相应的内毒素去除率分别达到99.8%、97.9%和99.7%。阐明了各种因素对内毒素去除率和蛋白质回收率的影响规律,为生物制品中内毒素的高效去除提供了参考。     

Factors affecting endotoxin removal from recombinant therapeutic proteins by anion exchange chromatography

Removal of endotoxins from recombinant proteins is a critical and challenging step in the preparation of injectable therapeutics, as endotoxin is a natural component of the bacterial expression systems widely used to manufacture therapeutic proteins. In this study we investigated various parameters affecting anion exchange chromatography to selectively remove endotoxins from therapeutic proteins. NY-ESO-1, Melan-A, and SSX-2 are different recombinant proteins used in this study, all of them are cancer antigens currently developed as potential immunotherapeutic agents. We found that by using a commercially available Q XL resin in a flow-through mode, endotoxin could be effectively removed from these proteins while maintaining very acceptable protein yields. The ratio of resin volume to endotoxin load was analyzed to determine the endotoxin binding capacity of the resin. In our hands at least 900,000 endotoxin units (EU) could be loaded per ml of Q XL resin. Solution conductivity could be increased to 20 mS/cm to minimize protein loss by weakening protein-resin attraction, and pH could be increased to enhance endotoxin removal by weakening endotoxin-protein attraction. Endotoxin levels were ultimately decreased to below 0.5 EU per microg of protein, an over 2000-fold reduction in this single step. A successful scale-up of these processes in which column volume was increased 100-fold was performed under cGMP conditions with over 80% protein recovery.     

Removal of Endotoxin during Purification of Poly(3-Hydroxybutyrate) from Gram-Negative Bacteria

Poly(3-hydroxybutyrate) (PHB) was produced by cultivating several gram-negative bacteria, including Ralstonia eutropha, Alcaligenes latus, and recombinant Escherichia coli. PHB was recovered from these bacteria by two different methods, and the endotoxin levels were determined. When PHB was recovered by the chloroform extraction method, the endotoxin level was less than 10 endotoxin units (EU) per g of PHB irrespective of the bacterial strains employed and the PHB content in the cell. The NaOH digestion method, which was particularly effective for the recovery of PHB from recombinant E. coli, was also examined for endotoxin removal. The endotoxin level present in PHB recovered by 0.2 N NaOH digestion for 1 h at 30°C was higher than 10⁴ EU/g of PHB. Increasing the digestion time or NaOH concentration reduced the endotoxin level to less than 1 EU/g of PHB. It was concluded that PHB with a low endotoxin level, which can be used for various biomedical applications, could be produced by chloroform extraction. Furthermore, PHB with a much lower endotoxin level could be produced from recombinant E. coli by simple NaOH digestion.     

Endotoxin removing method based on lipopolysaccharide binding protein and polyhydroxyalkanoate binding protein PhaP

Polyhydroxyalkanoates (PHAs) granule associated protein PhaP has a strong affinity to PHA and other hydrophobic polymers. Human lipopolysaccharide binding protein (hLBP) is a natural endotoxin receptor in plasma. In this study, genes encoding hLBP fused with PhaP were expressed in Pichia pastoris GS115 for production of the fusion protein. The purified rhLBP-PhaP fusion protein was immobilized on particles of polyhydroxybutyrate (PHB), which is a member of microbial polyhydroxyalkanoates (PHA). The rhLBP-PhaP-coated PHB particles were added to endotoxin containing water and protein solutions to study their endotoxin removal and protein recovery efficiencies. The influences of ionic strengths and pH on endotoxin removal and protein recovery in different protein solutions were also studied using acidic proteins including bovine serum albumin (BSA), ovalbumin, and basic protein α-chymotrypsinogen as model proteins. The results showed that rhLBP-PhaP particles could remove endotoxin with an efficiency of over 90%. All endotoxin removal and protein recovery efficiencies were only slightly affected by ionic strengths but were drastically affected by pH changes. Our results demonstrated that rhLBP-PhaP particles with their high efficiency, ease of preparation, and nontoxicity will be a suitable system for endotoxin removal in the protein purification industry.     

Production and quality control assessment of a GLP-grade immunotoxin,D2C7-(scdsFv)-PE38KDEL,for a phase I/II clinical trial

D2C7-(scdsFv)-PE38KDEL (D2C7-IT) is a novel recombinant Pseudomonas exotoxin A-based immunotoxin (IT), targeting both wild-type epidermal growth factor receptor (EGFRwt) and mutant EGFR variant III (EGFRvIII) proteins overexpressed in glioblastomas. Initial pre-clinical testing demonstrated the anti-tumor efficacy of D2C7-IT against orthotopic glioblastoma xenograft models expressing EGFRwt, EGFRvIII, or both EGFRwt and EGFRvIII. A good laboratory practice (GLP) manufacturing process was developed to produce sufficient material for a phase I/II clinical trial. D2C7-IT was expressed under the control of the T7 promoter in Escherichia coli BLR (λ DE3). D2C7-IT was produced by a 10-L batch fermentation process and was then purified from inclusion bodies using anion exchange, size exclusion, and an endotoxin removal process that achieved a yield of over 300 mg of purified protein. The final vialed batch of D2C7-IT for clinical testing was at a concentration of 0.12 ± 0.1 mg/mL, the pH was at 7.4 ± 0.4, and endotoxin levels were below the detection limit of 10 EU/mL (1.26 EU/mL). The stability of the vialed D2C7-IT has been monitored over a period of 42 months through protein concentration, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), isoelectric focusing, size exclusion chromatography, cytotoxicity, sterility, and pH measurements. The vialed D2C7-IT is currently being tested in a phase I/II clinical trial by intratumoral convection-enhanced delivery for 72 h in patients with recurrent glioblastoma (NCT02303678, D2C7 for Adult Patients with Recurrent Malignant Glioma; clinicaltrials.gov).

     

Membrane cartridges for endotoxin removal from interferon preparations

The suitability of membrane cartridges for the removal of endotoxin from both distilled water and interferon preparations was examined. The endotoxin concentrations were reduced to 4.0 and 7.3 EU/ml, respectively, when about 4000 ml of distilled water with 20 and 28 EU/ml were passed through the deoxycholate and chitosan immobilized membrane cartridges. When 200 ml of interferon preparation with endotoxin concentration more than 80 EU/ml and pH 3.9 were applied to a deoxycholate immobilized membrane cartridge at a flow-rate of 9 ml/min, the endotoxin concentration was reduced to less than 10 EU/ml. However, if an interferon preparation of 450 ml, with more than 80 EU/ml of endotoxin and pH 3.9 was applied to the chitosan immobilized membrane cartridge at a flow-rate of 18 ml/min, the endotoxin concentration was reduced to less than 10 EU/ml.     

Membrane adsorbers for selective removal of bacterial endotoxin

Surface-modified flat-sheet microfiltration membranes were functionalised with poly-l-lysine, polymyxin B, poly(ethyleneimine), l-histidine, histamine, α-amylase and DEAE as well as deoxycholate. Their suitability to remove endotoxin from both buffers and protein solutions was examined using bovine serum albumin, murine IgG₁ and lysozyme as model proteins. In protein-free solutions reduction from 6000 EU/ml to <0.1 EU/ml was achieved with all applied ligands; only α-amylase as well as l-histidine and histamine, when immobilized via the non-ionic spacer bisoxirane, exhibited low clearance factors at neutral pH. The adsorption of endotoxin is mainly ruled by electrostatic interaction forces. Thus in multi-component systems, such as endotoxin-contaminated protein solutions, competing interactions take place: acidic proteins compete with endotoxin for binding sites at the membrane adsorbers, basic proteins compete with the ligands for endotoxin and act as endotoxin carriers. With properly chosen conditions the membrane adsorbers presented here show exceptional effectiveness also in the presence of proteins. They are generally superior to functionalised Sepharose chromatographic sorbents and allow fast processing. They may contribute to reduce the risks in the application of parenterals and diagnostics.     

Characterization of variant diphtheria toxin–interleukin-3 fusion protein,DTIL3K116W,for phase I clinical trials

We have produced clinical grade of DTIL3K116W, a variant diphtheria toxin–interleukin-3 fusion protein, for treatment of acute myeloid leukemia. The product was filter sterilized, aseptically vialed, and stored at ?80 °C. It was characterized by Coomassie-stained SDS-PAGE, endotoxin assay, cytotoxicity assay, sterility, mass spectroscopy, receptor binding affinity, ADP-ribosylation, inhibition of normal human CFU-GM, disulfide bond analysis, immunoblots, stability, size exclusion chromatography–HPLC, sequencing, and immunohistochemistry. Vialed product was sterile in 0.25 M NaCl/5 mM Tris, pH 7.9, and had a protein concentration of 1.08 mg/ml. Purity by SDS-PAGE was >99%. Aggregates by HPLC were <1%. Endotoxin levels were 0.296 EU/mg. Peptide mapping and mass spectroscopy confirmed its composition and molecular weight. The vialed drug kept reactivity with anti-IL3 and DT antibodies. Potency study revealed a 48-h EC₅₀ of 0.5 pM on TF1/H-ras cell. Its binding properties were confirmed by competitive experiments showing IC₅₀ of 1.4 nM. ADP-ribosylation activity was equivalent to DTGM-CSF. Drug did not react with tested frozen human tissue sections by immunohistochemistry. There was no evidence of loss of solubility, proteolysis aggregation, or loss of potency over 6 months at ?80 °C. Further, the drug was stable at 4 and 25 °C in the plastic syringe and administration tubing for 48 h.     

Strategies for the depyrogenation of contaminated immunoglobulin G solutions by histidine-immobilized hollow fiber membrane

The depyrogenation of different IgG solutions using the histidine-linked hollow fiber membrane developed in our laboratory is presented here. Three strategies for endotoxin (ET) removal were investigated according to the immobilized histidine's ability to bind different immunoglobulins: (1) ET removal from 1 mg/ml non histidine-binding mouse monoclonal IgG₁ (MabCD4) solution was achieved in the presence of acetate buffer (pH 5.0) without any protein loss. (2) For contaminated human IgG, combined adsorption of ET and IgG in the presence of MOPS or Tris buffer was tested, followed by differential elution using increasing salt concentrations. This attempt was not successful since ET were quantitatively found in the IgG elution fraction. (3) Alternatively, it was proposed to adsorb selectively ET in the presence of acetate buffer (pH 5.0) under non binding conditions for human IgG. Human IgG could then be purified if necessary with the same membrane in the presence of MOPS buffer (pH 6.5). With a 1 m² histidine-PEVA module under these operating conditions, it is estimated that the depyrogenation of 3 1 of 1 mg/ml IgG (human or murine) solution containing 80 EU/ml of ET should be possible.     

Removal of endotoxin during purification of poly(3-hydroxybutyrate) from gram-negative bacteria.

Poly(3-hydroxybutyrate) (PHB) was produced by cultivating several gram-negative bacteria, including Ralstonia eutropha, Alcaligenes latus, and recombinant Escherichia coli. PHB was recovered from these bacteria by two different methods, and the endotoxin levels were determined. When PHB was recovered by the chloroform extraction method, the endotoxin level was less than 10 endotoxin units (EU) per g of PHB irrespective of the bacterial strains employed and the PHB content in the cell. The NaOH digestion method, which was particularly effective for the recovery of PHB from recombinant E. coli, was also examined for endotoxin removal. The endotoxin level present in PHB recovered by 0.2 N NaOH digestion for 1 h at 30 degrees C was higher than 10(4) EU/g of PHB. Increasing the digestion time or NaOH concentration reduced the endotoxin level to less than 1 EU/g of PHB. It was concluded that PHB with a low endotoxin level, which can be used for various biomedical applications, could be produced by chloroform extraction. Furthermore, PHB with a much lower endotoxin level could be produced from recombinant E. coli by simple NaOH digestion.     

Binding domains of colicins E1, E2 and E3 in the receptor protein btub ofEscherichia coli

Eight reagents specifically modifying amino acids were applied to cells of a standardEscherichia coli colicin indicator strain to followin vivo changes of its binding capacity for colicins E1–E3 and hence the binding domains (epitopes) for them in the outer membrane receptor protein BtuB. The effect of these reagents was also investigated in a mutant strain carrying an extensive BtuB deletion. The following differences of the binding epitopes could be ascertained.Colicin E1: Blockage of OH-groups, just as N-substitution of His and modification of Arg and Trp enhance binding of colicin E1. In the deleted receptor, also abolition of carboxylic anion bonds enhances its affinity for colicin E1. It follows that colicin E1 is bound, most of all, to the hydrophobic domain A (loops 1+2) of BtuB.Colicins E2 and E3: both exert rather analogous binding parameters. In contrast to E1, O-substitution of Ser and Thr dramatically decreases the E2 and E3 binding, similarly to modification of Lys. There is also a clear difference in the binding affinity of the domain for E2 and/or E3 and for E1 following modifications of their Arg and His. Colicins E2 and E3 are rather bound to the hydrophilic domain B (loops 5–7) of the receptor. In this respect, interactions of colicins E2 and E3 with deeper parts of A and B domains (Trp, several Arg, Lys and His residues) exhibited subtle differences. Acidic pH (4.5–6.0) shows a positive, while pH 7.0–8.5 a rather negative impact on the receptor-binding function for the colicins. It was clearly demonstrated that there is just a partial difference between the binding behavior of colicins E1, E2 and/or E3.     

Removal of tightly bound endotoxin from biological products

The method for endotoxin removal described in this paper is useful for separation of tightly bound endotoxin from biological products, particularly those produced in Escherichia coli in the form of inclusion bodies for which a denaturation step is required to solubilise the product. We employed guanidine hydrochloride and ammonium sulphate in combination with hydrophobic interaction chromatography (HIC). These conditions enable binding of the endotoxin to the matrix, giving unbound product in the column flow-through. This makes the method generally applicable to biological products. An endotoxin reduction of about 3.7 logs was achieved; from as much as 1,100,000 EU mg(-1) in the solubilised material to about 200 EU mg(-1) in the product purified by this method. The method was developed for a cervical dysplasia vaccine, a fusion protein comprising L2, E7 and E6 from Human Papilloma Virus type 16, because both conventional and commercially available methods of endotoxin removal were ineffective in removing the tightly bound endotoxin from this product.     

Studies on protein kinase activity and the binding of adenosine 3'5-monophosphate by membranes of hereditary spherocytosis erythrocytes.

Evidence has been presented recently of a deficiency of an endogenous membrane-associated protein kinase in erythrocytes of patients with hereditary spherocytosis (HS). We have measured endogenous protein kinase activity in erythrocyte membranes of 4 HS subjects using different membrane isolation and reaction conditions and find that the phosphorylation of the spectrin component (mean ± S.E. 17.1 ± 1.2 pmoles/10 mins per mg protein) is not significantly different to that of 4 normal controls (mean ± S.E. 20.7 ± 1.1 pmoles/10 mins per mg protein). Phosphorylation of exogenous proteins such as casein and protamine is also not deficient in HS erythrocyte membranes. Adenosine 3′5-monophosphate (cyclic AMP) binding to normal and HS erythrocyte membranes was also studied using a Millipore filtration assay. The affinity of cyclic AMP for erythrocyte membranes as determined by Hill plots of binding data from 4 HS subjects (K_D mean ± S.E. = 2.2 ± 0.2 nM) was not significantly different to 4 normal controls (K_D mean ± S.E. = 2.8 ± 0.6 nM). The rate of dissociation of bound cyclic AMP from HS membranes was also similar to control membranes. We thus cannot confirm the prediction by others that an abnormality of cyclic AMP interaction with the erythrocyte membrane underlies HS.     

Single step protocol to purify recombinant proteins with low endotoxin contents

Endotoxin is an unwanted by product of recombinant proteins purified from Escherichia coli. The inherent toxicity of endotoxins makes their removal an important step for the proteins' application in several biological assays and for safe parenteral administration. The method described in this paper is a one-step protocol which is effective at removing tightly bound endotoxin from over-expressed tagged proteins in E. coli. We combined affinity chromatography with a non-ionic detergent washing step, to remove most of the endotoxin contaminants from the end product. An endotoxin reduction of less than 4 to 0.2 EU mg(-1) was achieved with protein recovery close to a yield 100%. As this new protocol requires only one step to simultaneously purify tagged proteins and eliminate endotoxins, it represents a substantial advantage in time, effort, and expense.     

The separation, properties and possible subunit composition of adenosine 3',5'-monophosphate-dependent protein kinases in brown adipose tissue.

Two 8.5-S protein kinases (ATP : protein phosphotransferase EC 2.7.1.37) and one 6.6-S protein kinase were purified 500--1000-fold from the acid-soluble fraction of brown adipose tissue. The catalytic properties of the kinases were similar. Each kinase was activated by cyclic AMP and had two components of cyclic AMP binding. In the presence of 200 nM cyclic AMP, undissociated kinase activity sedimented at 7.7 or 5.5 S. Free catalytic activity (3.2 S) could be detected but was unstable. Free regulatory units could not be detected. The 8.5-S protein kinase was dissociated by freezing and thawing to a 7.7-S variety with loss of the higher affinity component of binding. The 7.7-S kinase was sedimented through linear gradients of sucrose containing different concentrations of cyclic AMP. At each concentration, kinase activity lost from the holoenzyme peak (% of original) was identical with the amount of cyclic AMP bound at equilibrium (% oof maximum). Similar experiments on the 8.5-S kinase showed that the binding component with higher affinity was not associated with the release of catalytic activity. The results were consistent with the propostal that the kinases isolated contained one more cyclic AMP binding subunit than catalytic subunit (3 : 2 for 8.5 S and 2 : 1 for 6.6 S) and that this extra subunit was released to give an equal number of subunits of each type before catalytic activity was liberated.     

Endotoxin removal by anion-exchange polymeric matrix.

The effects of anion-exchange polymeric matrices on endotoxin removal from albumin and gamma-globulin solutions are evaluated. The positively charged cellulose acrylic media carrying DEAE or QAE functional groups remove significant amounts of endotoxin from tap water, but are less effective in protein solutions. With properly controlled pH levels and salt concentrations, the endotoxin level in a protein solution can be reduced; however, low endotoxin concentrations, less than 100 pg/ml, are more difficult to remove. The endotoxin removal capacity depends on the number of functional groups existing in the matrix, expressed as the number of milliequivalents (meq), and on the pH operable range, which is directly related to the pK alpha value of the matrix. The effects of pH and salt on endotoxin removal from albumin and gamma-globulin solutions by an anion-exchange polymeric matrix were evaluated statically in test tubes. In addition, a dynamic flow was performed under statically defined conditions on a 250-ml DEAE cartridge for the removal of endotoxin from albumin at a flow rate of 40 ml/min. A greater than 75% reduction in the endotoxin can be achieved, with protein loss occurring only in the early stage of removal. Such processes are useful for the reduction of endotoxin from biological solutions produced by natural sources or recombinant DNA technology.     

An adenosine 3':5'-monophosphate-adenosine binding protein from mouse liver. Factors affecting the activation of the binding protein by adenosine 5'-triphosphate.

A cyclic AMP-adenosine binding protein, whose binding sites are activated by preincubation in the presence of Mg⁺-ATP, has been purified to apparent homogeneity from mouse liver (P.M. Ueland and S.O. Døskeland, 1977, J. Biol. Chem.,252, 677–686). The degree of activation of both the cyclic AMP binding site and a high-affinity site for adenosine depends on the concentration of ATP during the preincubation. The velocity and the degree of activation are dependent on the temperature and the presence of Mg²⁺ and K⁺. The NH₄⁺ ion can be substituted for K⁺, whereas Na⁺ is inefficient. Low pH promotes the conversion from the inactive to the active form. The apparent affinity for adenosine to the high-affinity site for this adenine derivative and the affinity for cyclic AMP to the site specific for this nucleotide are independent of the degree of activation as judged from the slope of Scatchard plots. The activation of the cyclic AMP binding site by ATP (6 mm) was determined at pH 7 in the presence of 10 μm cyclic AMP, AMP, ADP, or adenosine. Adenosine specifically inhibits the activation and does not promote the inactivation of the binding protein. The possibility that the apparent inhibition of activation was effected by interference with cyclic AMP binding by adenosine was ruled out.     

Endotoxin removal with poly(ethyleneimine)-immobilized adsorbers: Sepharose 4B versus flat sheet and hollow fibre membranes

Poly(ethyleneimine) was immobilized on poly(vinyl alcohol)-coated nylon flat sheet membranes, poly(vinyl alcohol) and poly(ethylenevinyl alcohol) hollow fibre membranes as well as Sepharose 4B. The resulting poly(ethyleneimine)-immobilized adsorbers were used for removal of E. coli derived endotoxin from buffers and bovine serum albumin solutions. The efficiency of poly(ethyleneimine) proved to be constant over a wide pH range, including phosphate buffered saline. The performance depended upon the matrix type employed: endotoxin clearance factors varied from 100 to 120 000 in protein-free solutions and 40 to 33 000 in solutions of bovine serum albumin using 6000 EU/ml as feed concentration. The best adsorber was the flat sheet membrane-immobilized poly(ethyleneimine), followed by the hollow fibre-immobilized poly(ethyleneimine) and poly(ethyleneimine)-Sepharose. The factors influencing endotoxin clearance were the mass transport (convective systems were superior to the diffusive system), the chemical composition and the surface structure of the underlying matrix.