PEGylation of adenovirus with retention of infectivity and protection from neutralizing antibody in vitro and in vivo.

Replication-defective recombinant adenovirus (Ad) vectors are under development for a wide variety of gene therapy indications. A potential limiting factor associated with virus gene therapy requiring repeated treatment is the development of a humoral immune response to the vector by the host. In animal models, there is a dose-dependent rise in neutralizing antibodies after primary vector administration, which can preclude effective repeat administration. The strategy we have developed to circumvent the neutralization of adenovirus vectors by antibodies is to mask their surface by covalent attachment of the polymer polyethylene glycol (PEG). Covalent attachment of PEG to the surface of the adenovirus was achieved primarily by using activated PEG tresylmonomethoxypolyethylene glycol (TMPEG), which reacts preferentially with the epsilon-amino terminal of lysine residues. We show that the components of the capsid that elicit a neutralizing immune response, i.e., hexon, fiber, and penton base, are also the main targets for PEGylation. Several protocols for PEGylation of an adenovirus vector were evaluated with respect to retention of virus infectivity and masking from antibody neutralization. We show that covalent attachment of polymer to the surface of the adenovirus can be achieved with retention of infectivity. We show further that PEG-modified adenovirus can be protected from antibody neutralization in the lungs of mice with high antibody titers to adenovirus, suggesting that PEGylation will improve the ability to administer Ad vectors on a repeated basis.

PEGylation of cytokines and other therapeutic proteins and peptides: the importance of biological optimisation of coupling techniques.

Polyethylene glycol (PEG) modification, PEGylation, is a well established technique which has the capacity to solve or ameliorate many of the problems of protein and peptide pharmaceuticals. It is one of the most important of the molecule altering structural chemistry (MASC) techniques and in many settings is enabling technology. The use of PEG as a linker molecule is also beginning to make a contribution to the production of exciting new products. We have previously reviewed the marked differences between methods of PEGylation and the surprising and dramatic impact of different coupling techniques (using different activated PEGs) on factors such as retention of bioactivity, stability and immunogenicity of the resulting PEGylated proteins and peptides. Numerous factors play a part in this variation: the presence or absence of linkers between the PEG and the target molecule; the nature and stability of the bond(s) between the PEG, linker and target; the impact of PEG attachment on surface charge; the coupling conditions; and the relative toxicity of the activated polymer and/or coproduct(s). These are not, however, the only sources of qualitative differences in PEGylated products. Our own experience whilst developing a linkerless PEGylation technique (i.e. one attaching only PEG to the target molecule), which we devised to overcome all the major problems of pre-existing PEGylation techniques, was that considerable modification of the prototype method and a process of 'biological optimisation' was required to achieve good results in terms of conservation of bioactivity. Biological optimisation has not, as far as we are aware, been systematically applied by other groups working in PEGylation. It is the term we use to describe an iterative process for examining and refining all the steps in the PEGylation process, including manufacturing the activated polymer, in order to achieve the best possible conservation of bioactivity and other beneficial features of the method. The application of this biologically optimised PEGylation technique, using tresyl monomethoxy PEG (TMPEG), to a variety of target proteins reveals, as outlined in this review, an exceptional ability to conserve biological activity of the target. This, and the benefit of adding nothing other than PEG itself (which has an excellent safety record), to the protein, as well as other manufacturing and practical advantages, makes the method ideal for the modification of cytokines and other therapeutic proteins.

DNA-protein interaction sites in differentiating cells. II. A subset of alphoid repetitive sequences with retinoic acid induced protein attachment and an unusual purine-pyrimidine 'signature'.

Two independent methods were used to recover the tightly or covalently attached DNA-protein complexes which form during the differentiation induction of HL60 cells by all-trans retinoic acid. One method employed nitrocellulose filter binding (described in the accompanying paper) to recover these protein-DNA complexes, following by representational difference analysis (RDA). RDA is an extremely powerful new technique for cloning the difference between complex DNA samples, exploiting a combination of kinetic enrichment during PCR amplification, and subtractive hybridization. In this case, we used filter-bound DNA from undifferentiated cells for subtraction of equivalent DNA from differentiating cells. A second approach used a combination of sodium dodecyl sulfate (SDS)/KCI precipitation and affinity phase partitioning for purification of DNA bound to proteins, followed by selection of clones showing differentiation-specific attachment of proteins (by differential hybridization to protein-attached DNA from differentiating and undifferentiated HL60 cells). Both procedures yielded a high proportion of alphoid repetitive sequences, although slightly higher in the RDA sample than that of the other method (approximately 50% and approximately 30%, respectively). Plots of purine-pyrimidine composition showed that the alphoids recovered by both techniques were unusual and remarkably similar. Although the clones were related to each other, they differed more from the alphoid consensus than examples of alphoid sequences selected from data banks. There was also a contrast between the purine-pyrimidine composition plots of archetypal alphoid monomers and this subset. The principal difference observed related to two adjacent homopyrimidine tracts present in the archetypal monomers but altered in the subset. The possible significance of these differences is discussed with respect to the established roles of alphoid sequences and known/putative protein binding sites including that for the centromeric binding protein (CENB-P box) and response elements for retinoic acid receptors. Following induction of differentiation with retinoic acid, protein attachment (monitored by filter binding) was rapid, maximal at 40 minutes, and still elevated at 165 minutes. This was specific to the alphoid subset, and induced protein binding at a nonrepetitive site in DNA had different kinetics. Changing protein attachment at this subset of alphoid repetitive sequences contributes to differentiation-associated chromatin structural changes.

DNA-protein interaction sites in differentiating cells. I. Two-dimensional mapping of modulated sites.

We have previously shown that DNA-protein attachment sites form during the induction of hematopoietic cell differentiation. Affinity phase-partitioning studies of DNA/protein complexes demonstrated that the DNA involved is not randomly distributed throughout the genome. The object of this study was to use filter binding followed by two-dimensional (2D) polyacrylamide gel electrophoresis (using a neutral 6% gel in the first dimension and a denaturing gradient gel in the second dimension) to gain insight into changes in DNA-protein interactions during induced granulocytic and monocytic differentiation of HL60 cells. Nitrocellulose filter-binding enriched samples for protein-associated DNA sufficiently to change the pattern of DNA spots on 2D gels. The patterns of spots obtained was reasonably reproducible between experiments and highly reproducible within experiments. Gels obtained from cells induced to differentiate by either phorbol ester or all-trans retinoic acid (RA) showed identical patterns for the majority of spots but changes in a small proportion of spots with respect to uninduced controls. Both intensification and reduction/disappearance of spots was observed, demonstrating the existence of both invariant and variant DNA/protein attachment sites during the early stages of hematopoietic cell differentiation. Previous studies have implicated DNA topoisomerase II in chromatin structural changes that are necessary for induction of granulocytic differentiation. We therefore examined the filter-binding DNA preparation by 5'-exonuclease digestion (since topoisomerase II is known to bind covalently to the 5'termini on either side of its cleavage sites). The filter-associated DNA exhibited increased 5' exonuclease protection (with respect to filter flow-through DNA), and the degree of protection increased significantly with exposure to phorbol ester and less markedly with retinoic acid. However, since not all filter DNA was 5' protected, it remains unresolved whether the specific differentiation-associated DNA-protein interactions revealed here involve DNA topoisomerase II or some other protein.

Enhanced tumour specificity of an anti-carcinoembrionic antigen Fab' fragment by poly(ethylene glycol) (PEG) modification.

Polyethylene glycol (PEG) modification of a chimeric Fab' fragment (F9) of A5B7 (alpha-CEA), using an improved coupling method, increases its specificity for subcutaneous LS174T tumours. PEGylation increased the area under the concentration-time curve (AUC0-144) in all tissues but there were significant differences (variance ratio test, F = 27.95, P < 0.001) between the proportional increases in AUC0-144, with the tumour showing the greatest increase. The increase in AUCtumour from F9 to PEG-F9 was similar to the reported increase from Fab' to F(ab')2 while the increase in AUCblood by PEGylation of F9 was only 21% of the reported increase from Fab' to whole IgG. A two sample t-test showed no significant differences between maximal tumour/tissue ratios for PEG-F9 and F9 while the tumour/tissue ratios for PEG-F9 remained high over a longer period, with tumour levels at least double those for F9. PEG-F9 emerges as a new generation antibody with potential advantages for both radioimmunotherapy and tumour imaging. Since there was a reduction in antigen binding, optimisation of PEGylation might further improve tumour specificity. The latter resulted from complex effects on both the entry into and exit rates from tumour and normal tissues in a tissue-specific fashion.

Polyethylene glycol modification: relevance of improved methodology to tumour targeting.

Of all the polymers applied to molecule altering structural chemistry, polyethylene glycol (PEG) modification has numerous benefits and relatively few drawbacks. PEG is now increasingly being applied to the problems of tumour targeting, both in the context of the passive targeting of PEG-liposomes and in active targeting strategies using PEGylated anti-tumour antibodies. PEG can also serve as a useful linker molecule between targeting moieties and other agents, including cytotoxic or imaging agents and targeted liposomes. Despite these demonstrated benefits and the level of attention which PEGylation has received, relatively little consideration has been given to two key areas: first, the extent to which the coupling method has an impact on both the functionality of the PEG-adduct and the acquisition of beneficial properties; second, that the impact of PEGylation on biodistribution is complex, thus any attempt to optimise a PEG-peptide or PEG-liposome for a particular task must involve an examination of all the individual facets of the effects of PEGylation. Studies investigating the underlying principles of tumour targeting suggest that current views concerning the optimisation of PEGylated vehicles for tumour localisation need to be re-examined.

Retinoic acid and phorbol ester induced hyperphosphorylation of topoisomerase II-alpha is an early event in HL-60 human leukaemia cell differentiation: effect on topoisomerase activity and etoposide sensitivity.

Treatment of HL-60 with phorbol myristate acetate (PMA) for 30 min, or all-trans retinoic acid (RA) for 60 min, results in hyperphosphorylation (3-5x) of topoisomerase II (p170, topo II) in vivo. RA and PMA activate a coprecipitating kinase, respectively inducing 1.6 and 2.7-fold increases in phosphorylation of topo II in immunoprecipitates. The activity of the co-precipitating kinase is inhibited by heparin and unlabelled GTP suggesting that casein kinase II (CKII) is, at least in part, responsible for the topo II hyperphosphorylation in response to differentiation signals. Although following dephosphorylation of the enzyme with alkaline phosphatase there was virtual abrogation of activity, the differentiation associated hyperphosphorylation had little impact on the decatenation activity of topo II in nuclear extracts. There were, however detectable changes in topo II function in vivo which affected the formation of the etoposide stabilised cleavable complex, but only after PMA treatment. PMA resulted in a rapid reduction in etoposide induced cleavage, 30 min treatment with PMA reducing cleavage by 20%. However, treatment with RA for 1 or 2 h when hyperphosphorylation was maximal did not affect cleavage. Immunoband depletion assays suggested that differentiation associated changes in chromatin structure rather than alterations in the enzyme per se are responsible for the reduction in cleavable complex formation following PMA treatment. Etoposide cytotoxicity was significantly reduced following just 30 min PMA treatment, but not reduced and even possibly enhanced by retinoic acid treatment. These findings are relevant not only to the dissection of the role of topo II in differentiation but also to its exploitation as a therapeutic target.

Quantitative analysis of polyethylene glycol (PEG) in PEG-modified proteins/cytokines by aqueous two-phase systems.

Covalent attachment of poly(ethylene glycol) (PEG) to proteins produces conjugates with altered/improved physicochemical and biological properties which depend upon the number of PEG chains linked. Quantification of the attached PEG is however not a trivial issue. The partition coefficient, K, of the PEG-protein conjugate in PEG/dextran two-phase systems provides a quantitative measure for the degree of modification. A linear relationship between log K and the number of PEG chains was observed in fractionated PEG-modified-granulocyte-macrophage colony stimulating factor conjugates having 1 to 3 substitutions. Furthermore, in mixtures of PEG-bovine-serum-albumin conjugates with increasing degrees of modification, a linear relationship was found between log K and n, the average substitution. The increment in log K per PEG chain added is protein specific and this suggests that the interactions between the PEG-protein conjugate and the polymers in the phase system are more complex than just a simple affinity of the PEG for the PEG-rich top phase. Increasing the polymer concentration in the phase system produces larger increments in log K per PEG molecule attached and the proportionality between log K and number of PEG molecules is only compromised for conjugates with high degree of substitution when partitioned in biphasic systems of high concentration of polymers.

Theoretical and practical considerations in screening strategies for differentiation agents.

This paper reviews strategies for identification of new compounds with differentiation inducing activity. The primary objective is to discriminate 'lead' compounds with novel mechanisms of action. Complementary to this is the need to generate a taxonomy of known compounds, identifying those with similar mechanisms, preferably in a way that provides clues as to the nature of those mechanisms. Experimental data suggest that the methods to do this are already to hand and that some existing search strategies are readily adapted to these objectives. The principal property required of the screen is that distinct mechanisms of induction produce different outcomes. Response patterns can thus be used to group drugs with similar mechanisms and hence identify novel activities. To be successful, such patterns have to be relatively insensitive to potency (so that agents with the same mechanism but different potency are classed together). Two strategies to achieve mechanism-sensitive response patterns are outlined, one based on the varied response capacities of multiple target cell types and another exploiting drug interaction patterns. With the former, the principle exploited is that the differences in the panel cells' capacities to respond depend on the components and assembly of their signal transduction and effector mechanisms for differentiation. With the latter, it is the varied and complex ways in which the components of the intracellular mechanism are assembled which provides distinctive interaction patterns, depending upon which components are the molecular targets of a particular pair of drugs. Pattern generation, analysis and optimisation of efficacy and cost/benefit are also discussed. Differentiation end-points are difficult to assess and cell number/mass may be more appropriate to high throughput designs. The problems of discriminating growth arrest due to differentiation, from other antiproliferative or simply cytotoxic effects are therefore considered and suggestions made for feasibility studies of these approaches.

Effects of DNA topoisomerase II inhibitors on human bone marrow progenitor cells.

Topoisomerase II (topo II) is a target for many cytotoxic agents. Two observations, however, warrant caution in their therapeutic use: first, these agents can inhibit differentiation and second, perturbations in function render the enzyme error-prone. Illegitimate recombination events occurring at sites where topo II acts in differentiation could be particularly important in the development of secondary malignancies (relatively frequent after therapy with agents that target topo II). Topo II inhibitors are heterogeneous in mechanisms of action; in site-specificity of cleavable complex 'entrapment' (where present) and in the relative potency against the two topo II isoforms, all potentially influencing the site of maximum DNA damage. The object of this study was to examine the effect of topo II inhibitors on human haemopoietic precursor cells, to determine which have most impact on differentiation. We selected two which act via cleavable complex entrapment, but with different site preferences (m-AMSA and VP-16), and two acting via other mechanisms (merbarone and fostriecin). VP-16 and m-AMSA showed similar patterns with low dose stimulation of granulocyte-macrophage colony formation and high dose inhibition of all colony types. The stimulation was accompanied by an increase in colony size and blast content, consistent with a low dose inhibition of differentiation. Forstriecin, in contrast, stimulated predominantly mixed and erythroid colonies. Merbarone failed to increase colony formation. Neither produced substantial inhibition of colony formation. The effects on granulocyte-macrophage progenitors were confirmed using 7-day suspension cultures, using nitroblue tetrazolium (NBT) reduction and 3-4,5,dimethylthiazol 2,5-diphenyl tetrazolium bromide (MTT) assays for differentiated cells and total cell mass, respectively. These results demonstrate that the effects of topo II inhibitors on haemopoietic cell proliferation and differentiation are agent-specific and can involve lineage-restricted partial inhibition of differentiation.

Polyethylene glycol (PEG) modification of granulocyte-macrophage colony stimulating factor (GM-CSF) enhances neutrophil priming activity but not colony stimulating activity.

PEG-modified proteins have numerous advantages over their unmodified counterparts (increased half life, reduced antigenicity, improved solubility), but almost without exception, they show a modest to marked reduction in biological or enzymatic activity. However, while investigating a new protocol for the preparation of PEG-proteins, we compared PEG-modified and unmodified GM-CSF with respect to their polymorphonuclear neutrophil granulocyte (PMN) priming activities. PEG-GM-CSF was unexpectedly more active than GM-CSF in its ability to prime neutrophils to respond to the synthetic peptide n-formyl-methionyl-leucyl-phenylalanine (FMLP) with an oxidative burst (assessed both by nitroblue tetrazolium reduction and ferricytochrome c reduction). These results were in contrast to the findings for colony stimulating activity and with GM-CSF induced thymidine uptake, where the biological activity was unchanged or reduced. The enhanced neutrophil priming activity of PEG-GM-CSF was confirmed using FPLC fractionated PEG-modified GM-CSF. This showed changes in the bioactivity profile consistent with both the shift in protein elution profile and enhanced activity of the PEG-modified material (reflected in the increased area under the bioactivity curve). We also excluded a neutrophil priming action for PEG-modified fetal calf serum proteins, carrier proteins and 'irrelevant' cytokine, erythropoietin. The dissociation of the two bioactivities was confirmed using individual FPLC fractions. These results suggest the presence of differences in either binding, receptor/ligand processing or signal transduction for neutrophils versus progenitors, that are differentially affected by PEG-modification of GM-CSF. The demonstration that PEG-modification can partially dissociate two biological activities suggests the feasibility of using PEG-modification to produce proteins with subtly altered spectra of biological activity and hence new ranges of clinical applications.

Polyethylene glycol (PEG)-modified granulocyte-macrophage colony-stimulating factor (GM-CSF) with conserved biological activity.

Polyethylene glycol (PEG) modification improves the pharmacological properties of proteins, usually extending plasma half-life and concomitantly increasing in vivo bioactivity, reducing both antigenicity and immunogenicity, and increasing solubility and resistance to proteolysis. Despite these established benefits, few PEG proteins are in use. Current coupling methods are either traumatic for the protein or involve lengthy and difficult procedures to activate monomethoxyPEG (MPEG). We have applied a new coupling method that allows coupling of MPEG directly to proteins under physiological conditions. Using this method with recombinant human (rh)granulocyte-macrophage colony-stimulating factor (GM-CSF) we were able to construct biologically active PEG-GM-CSF. Fast protein liquid chromatography (FPLC) and phase-partitioning confirmed the presence of PEG modification, and the former was used to fractionate modified and unmodified material. Bioactivity was measured in colony assays of normal human bone marrow cells and by tritiated thymidine uptake (of chronic myeloid leukemia cells and TF-1 cells). With both uptake and colony assays, using unfractionated material, we observed only a modest reduction in biological activity. Assays of FPLC-fractionated material confirmed that much of the bioactivity of the PEG-GM-CSF preparations was due to the modified species and any residual unmodified GM-CSF. Species uncontaminated by tresylmonomethoxyPEG (TMPEG; which was somewhat inhibitory in the thymidine uptake assay and eluted over a broad region of the FPLC profile) had no significant reduction in activity, but we cannot rule out the possibility that PEG-GM-CSF species eluting elsewhere in the profile had modest reduction of activity. Subcutaneous injection into mice confirmed the anticipated improved half-life in vivo and demonstrated a longer uptake from the injection site. This is, as far as we are aware, the first successful construction of PEG-GM-CSF with conserved biological activity.

The uses and properties of PEG-linked proteins.

Poly(ethylene glycol) (PEG) is a water soluble polymer that when covalently linked to proteins, alters their properties in ways that extend their potential uses. PEG-modified conjugates are being exploited in many different fields. The improved pharmacological performance of PEG-proteins when compared with their unmodified counterparts prompted the development of this type of conjugate as a therapeutic agent. Enzyme deficiencies for which therapy with the native enzyme was inefficient (due to rapid clearance and/or immunological reactions) can now be treated with equivalent PEG-enzymes. PEG-adenosine deaminase has already obtained FDA approval. PEG-modified cytokines have been constructed and, interestingly, one of the conjugates, PEG-modified granulocyte-macrophage colony-stimulating factor, showed dissociation of two biological properties. This novel observation may open new horizons to the application of PEGylation technology. The biotechnology industry has also found PEG-proteins very useful because PEG-enzymes can act as catalysts in organic solvents, thereby opening the possibility of producing desired stereoisomers, as opposed to the racemic mixture usually obtained in classical organic synthesis. Covalent attachment of PEG to proteins requires activation of the hydroxyl terminal group of the polymer with a suitable leaving group that can be displaced by nucleophilic attack of the epsilon-amino terminal of lysine residues (other nucleophilic groups can also interact). Several chemical groups have been exploited to activate PEG, thereby giving rise to a variety of PEG-proteins. Some of these varieties retain part of the activating group as a coupling moiety between PEG and protein and others provide a direct linkage. For each particular application, different coupling methods provide distinct advantages.(ABSTRACT TRUNCATED AT 250 WORDS)

Separation of cell mixtures by immunoaffinity cell partitioning: strategies for low abundance cells.

The partitioning of cells in aqueous two-phase systems formed by poly(ethylene glycol) (PEG) and dextran can be changed by incubating the cells with a PEG-modified antibody directed specifically against its surface. We have developed a new approach for immunoaffinity cell partitioning (IACP) in which the antibodies are first reacted with tresylated monomethoxy PEG (TMPEG) in sodium phosphate buffer, pH 7.5, the excess TMPEG is quenched by reaction with bovine serum albumin, and the resulting preparation is used directly for incubation with the cells without any isolation of the monomethoxyPEG (MPEG)-antibody conjugates. We have demonstrated the specificity of this IACP method by showing that MPEG-modified anti-human red blood cell antibody increases the partition of human erythrocytes from the interface to the PEG-rich top phase (up to 100%) but not the partitioning of either neutrophils or HL60 cells. Irrelevant antibodies do not affect the partitioning of red blood cells. The partitioning behaviors of erythrocytes and HL60 cells in mixtures varying from 75 to 10% red blood cells subjected to IACP are similar to those of the pure cell population, i.e., erythrocytes ca. 100% and HL60 cells 3% in top phase. Thus, the population of erythrocytes can be almost completely extracted into the top phase in a single step. The contaminant cells represent only a small percentage (less than 5% in most of the cases) of the cell mixture recovered in top phase. Both cell populations can be completely separated by countercurrent distribution (CCD).(ABSTRACT TRUNCATED AT 250 WORDS)