WO1999036092A2

WO1999036092A2 - Hiv inhibition using ligand for gp60 receptor or albumin receptor on target cells

Info

Publication number: WO1999036092A2
Application number: PCT/GB1999/000174
Authority: WO
Inventors: Paul Stuart O'shea
Original assignee: Quadrant Healthcare (Uk) Limited
Priority date: 1998-01-19
Filing date: 1999-01-19
Publication date: 1999-07-22
Also published as: WO1999036092A3; GB9801070D0; AU2175299A

Abstract

HIV infection may be prevented by using a material that does not bind lipids but binds preferentially to a receptor selected from the gp60 receptor and albumin receptors, on target cells. Such materials include albumin fragments having no lipid-binding domain.

Description

INHIBITION USING LIGAND FOR GP60 RECEPTOR OR ALBUMIN RECEPTOR ON TARGET CELLS

Field of the Invention

This invention relates to methods and products that may be used to inhibit HIV infection. Background of the Invention

Various proposals have been made, for inhibiting infection by HIV, generally regarded as the causative agent of AIDS. However, substances that may inhibit HIV in vitro may not be useful therapeutic agents. Human immunodeficiency virus type I (HIV-1) entry in target cells is mediated by the viral envelope glycoproteins designated gpl20 and gp41, derived by proteolytic cleavage of the gpl60 precursor. gpl20 is the outer surface glycoprotein and contains the site necessary for viral binding to specific target cells. gp41 appears to possess one transmembrane domain and is thought to exhibit a multi-role function, being involved in the anchoring of the envelope glycoprotein complex to the viral membrane, the oligomerisation of the envelope glycoprotein and the interaction between viral and cell membranes. Similar processes occurs in other species, with related viruses (e.g. SIV in monkeys) .

The viral membrane-interaction sequences of gp41 appear to possess a relatively high content of Ala and Gly residues and to be characterised by a short amino-terminal hydrophobic segment which is likely to interact with plasma membranes. There is also a high degree of conservation of Phe-Leu-Gly, i.e. the so-called FLG motif, which is 7-8 residues from the N-terminus. Substitution of these residues with charged residues in a membrane-interaction sequence results in abolition of syncytium formation, underlining the importance of these sequences in infection. The late membrane interaction, therefore, is one of the key events during viral infection. Baba et al , PNAS USA 85:6132-6 (1988), discloses that sulphated polysaccharides (polyanions, e.g. dextran sulphate) inhibit HIV infection of target cells. The site of binding is unclear.

Taka i et al , Biochim. Biophys. Acta. 1180:180-6 (1992) , confirms this effect. This article reports that a modified polyanionic protein (maleylated human serum albumin, or Mal-HSA) mimics polyanionic species such as dextran sulphate. Mal-HSA inhibits HIV infection, via two binding proteins, of 155 kDa and 220 kDa molecular weight, respectively. The effective minimum concentration of Mal- HSA required to show this effect in vitro is 1.25 μg/ml (1.25 mg/L) .

Takami et al reports: "We performed a preliminary experiment to investigate the relationship [between the number of Mal-HSA binding sites and the affinity for Mal- HSA] using Molt-4 clone 8 cells and Molt-4 clone 3 cells. These cells express the CD4 molecule equally, but Molt-4 clone 8 cells and clone 3 cells are highly infectable and low infectable cells respectively. However these cells possessed almost the same affinity and number of Mal-HSA binding sites. Furthermore there was no significant difference between 200 kDa and 155 kDa proteins". These data strongly suggest that the use of polyanionic molecules would not affect the potential infectivity of cells by HIV.

Human serum albumin (HSA) is present in the blood at a level of 40 g/L (i.e. a concentration of 32,000 times that reported by Takami et al ) without any noticeable effect. For example, new immature T-cells in the blood are readily infected in the HIV-positive individual, despite the presence of HSA. Further, any increase in HSA blood concentration by artificial means would significantly disturb the pH and oncotic balance of the blood, with deleterious results to the subject. Therefore, HSA is apparently unsuitable for use in the therapy of HIV infection. The transport of ligands across cellular barriers, known as transcytosis, involves mechanisms of selectivity to facilitate specific ligand transport. These processes appear to involve the recently discovered plasma membrane albumin receptor known as gp60 or albondin. For example, WO-A-9710850 discloses that transcytosis of a drug such as insulin may be enhanced by formulation with albumin, fragments thereof, gp60 peptide fragments or anti-gp60 antibody.

Curtain et al , AIDS Res. Human Retroviruses 9:1145-56 (1993) discloses that the amino terminal peptide of HIV gp41 interacts with HSA. Mitsuya et al , Science 226:172-4 (1984), discloses that suramin, a polysulphonated naphthylurea that binds to lymphocyte membranes, protects T cells in vitro against infectivity and cytoplasmic effect of HTLV III.

Moriuchi et al , J. Virol. 12:9664-71 (1997), suggests that a factor as yet undiscovered is necessary for infection by HIV. Summary of the Invention

This invention is based on the discovery that HIV gp41 interacts with the gp60 receptor. Experimental evidence presented below also shows that a receptor for a gp41 fragment exists on the surfaces of T-cells but not on B-cells. Therefore, pharmaceutically-acceptable agents that bind preferentially (which term may also mean mere selectivity, with respect to albumin) to such a receptor, which in use may be any receptor having the binding characteristics of gp60, or another albumin receptor, can be used to inhibit HIV infection. Improved selectivity may comprise reduced, substantially zero or zero membrane- binding or lipid-binding. This may exclude charged, e.g. polyanionic, species.

Brief Description of the Drawings

The accompanying drawings are graphs of results that illustrate the basis of this invention, and in particular show fluorescence plotted against time. Fluorescence is recorded at 518 nm after excitation at 490 nm.

Fig. 1 shows the interactions of B and T lymphocytes labelled with FPE and albumin. Fig. 2 shows the interactions of B and T lymphocytes labelled with FPE, preincubated with albumin and challenged with HIV peptide.

Figs. 3 and 4 show the interactions of gp60 liposomes labelled with FPE and challenged with HIV peptide and albumin, respectively.

Fig. 5 shows the binding of HSA and fragments thereof to gp60 liposomes, over a first period of 0.02 sec and a second period of up to 100 sec. Description of the Invention

While naturally-occurring human serum albumin may interact with the gp60 receptor at least, the present invention utilises a different agent. HSA exists in a ligand-bound state. While lipids such as oleate can be removed, and it has been found that defatted albumin interacts with the gp60 receptor, it rapidly binds lipids in vivo and is thus not a suitable active agent. A preferred agent for use in this invention is albumin that has no lipid-binding capacity, e.g. by having its lipid- binding domain (s) deleted or inactivated. This can be done while retaining the function of interaction with the gp60 receptor. For example, suitable albumin fragments include those consisting of or including amino-acids 45-65, 45-585, 123-298 or 298-585 of albumin, or a fragment thereof. Certain such fragments may be novel per se, or at least for therapeutic use.

Other agents that may be used in the invention are anti-gp60 antibody, gp60 receptor protein, gp41 receptor- binding domain, i.e. gp41 fusion protein without its fusion capability, and fragments thereof that bind to the gp60 receptor. Certain fragments of different parent molecules may show homology. Combinations of any such agents, or combinations of fragments of the same parent molecule, may also be used. The invention also envisages the use of materials that have the additional capability of targeting receptor, e.g. CD4, on B/T cells; materials having this function will be well known to those of ordinary skill in the art.

The gp60 receptor may have two functions, i.e. recognition in addition to transcytosis. An agent for use in the present invention may effectively interact with one function, and not necessarily with both.

The appropriate binding by the active agent of the target receptor may be a function of conformation, or of a structural constraint preventing any change in conformation. Alternatively or in addition, an active agent may interact with either, or both, of low and high affinity binding sites.

Although other materials may have the desired characteristics, including small molecules that may mimic peptideε, agents currently preferred for use in the invention are peptides. Depending on their ability to retain the desired functional characteristics, such peptides may comprise at least 4, 5, 6, 8, 10, 15 or 20 amino-acids, or they may be longer, e.g. up to 20, 50 or 100 amino-acids.

Peptides are generally substantially insoluble. Formulations for administration of such materials are well known, and can be used in this invention. The preferred route of administration is by inhalation, e.g. using a metered dose inhaler, buccal, transdermal, etc. The amount of the agent to be administered depends on factors routinely considered by those skilled in the art, and can be determined by routine procedures.

Since gp60 apparently has a valuable function in transcytosis, it may be desirable to avoid blocking it completely. The level of agent that is used in the invention may have to be controlled accordingly. Alternatively, the effect of the agent may enhance (by blocking receptors) or be enhanced by the co-administration of a compound that is known to inhibit HIV infection, e.g. an anti-replicase such as AZT. The following Experiments serve to illustrate the basis of the present invention. The following abbreviations apply:

FPE fluorescein phosphatidyl ethanolamine P V phospholipid vesicle

PC phosphatidyl choline PS phosphatidyl serine Although the nature of the binding of many such ligands to the albumin molecule is becoming better understood, the mechanism of selectivity by the receptor for albumin carrying particular ligands is unknown. To shed light on this problem, gp60 was isolated from human endothelial cells and reconstituted into model membranes (see experimental protocol, below) . This well-defined experimental system was studied using a technique that reveals the interactions of proteins with membranes; see Golding et al , Biochem. 35:1093-7 (1996). By employing stopped-flow rapid mixing, the complete time-evolution of the interaction of albumin with gp60 was observed. This represents the first demonstration of the time-course of a specific ligand-receptor interaction in which the receptor is membrane-located.

Membrane systems with well-defined lipid compositions were used in an attempt to characterise the sequence of interactions of the membrane interactions of the viral fusion peptides. Synthetic peptides corresponding to the 12 residue N-terminal region of SIV-gp32 (NH₂-Gly-Val-Phe-Val- Leu-Gly-Phe-Leu-Gly-Phe-Leu-Ala-CONH₂; see also SEQ ID NO.l) and the 16 residue sequence corresponding to the N- terminus extremity of gp41 of HIV-1 (NH2-Ala-Val-Gly-Ile- Gly-Ala-Leu-Phe-Leu-Gly-Phe-Leu-Gly-Ala-Ala-Gly-CONH₂; see also SEQ ID NO.2) were studied using fluorescence techniques; see Wall et al , Mol. Memb. Biol. 12:181-190 (1995), and also J. Cell Sci. 108:2673-82 (1995). The time evolution and overall extent of the interactions of many types of molecule with membranes, including proteins and peptides, have been shown to be accessible using this technique. Measurements rely on the fact that most proteins that interact with membranes possess a net charge and once bound, result in changes of the electric field present on the membrane surface. Such changes influence the fluorescence yield of FPE previously added in small amounts (lmol FPE: 800-1000 mol lipid) with the fluorophore precisely located on the membrane surface. The underlying theory and application provide the additional virtue that the technique may be applied to living cells as well as to model membrane systems.

Interactions between albumin, gp41 and gp32 with purified and reconstituted gp60 from bovine lung micro- vascalature or human umbilical vein endothelia and with lymphocyte membranes were studied, therefore, as outlined in Wall et al , supra , and Golding et al , supra . Fluorescence changes associated with the binding of albumin to the cell surface of B and T lymphocytes are illustrated in Fig. 1. The fluorescence was observed to decrease consistent with the addition of negative charge to the membrane surface as the albumin molecule overall is known to be negatively charged. These signals may be plotted culmulatively, as in Fig. 1, to estimate the affinity of the plasma membrane for albumin. Both cell types exhibit saturation of binding with similar dissociation constants (ca. 4-6μM) , but the relative abundance of the putative albumin receptor is of the order of 4-5 times greater in the T cells.

Similar studies performed with the HIV peptides are shown in Fig. 2, and indicate that the relative capacity of each type of lymphocyte for the HIV peptides is similar to that for the albumin molecule. Similarly, a number of cell types such as endothelia, fibroblasts and icroglia that have also been shown to bind albumin have also been studied in an analogous manner and yielded results similar for those of the lymphocytes.

A more compelling demonstration of the ligand- dependent affinity for albumin and the HIV proteins by the receptor known as gp60 involved complementary studies. Thus the interactions of HIV and albumin with purified gp60 reconstituted into artificial membranes in the absence of other membrane proteins were investigated. This was achieved by first isolating gp60 from human umbilical vein endothelial cells (HUVECs) ; purity was gauged as a single band on a SDS-PAGE electrophoretogram followed by autoradiographic ligand blots between I^"125-Albumin and gp60 that indicated selective binding. gp60 was then reconstituted into membranes made up with a well-defined phospholipids mixture (85% PC: 15% PS w/w) followed by labelling with FPE (0.1-0.3M%). Gel filtration (Pharmacia- PD10) of the reconstituted vesicles facilitated selection of the reconstituted proteoliposome vesicles on the basis of their diameter (100 nm) . Studies were then performed as with each of the cell types described in Fig. 1. The results of some of these studies are shown in Figs. 3 and 4. In the case of albumin, shown in Fig. 4 and as with cells, the net fluorescence was ultimately diminished following the addition of albumin to gp60. During the course of the first encounters of albumin with the receptor over the millisecond time-domain, however, as shown in Fig. 4, the fluorescence increases rapidly with a rate approaching the resolution of the stopped-flow mixing technique (>1000 s^-1) . Such rates are so rapid that they are unlikely to represent processes other than binding. These early encounters leading to the binding of albumin to gp60, therefore, must first involve the addition of positive charges to the membrane. Albumin is known to possess a number of specific domains or regions on its surface which are heavily positively or negatively charged; the gp60- albumin-binding domain would seem to interact initially with one of these regions.

Following the initial increment over the milliseconds time domain, it is clear from Fig. 4 that the fluorescence signals then decrease more slowly over a period of seconds. This behaviour has been observed with other membrane- protein interactions and indicates that conformational changes takes place that reveal the more dominant negatively charged groups and/or part of the gp60/albumin complex has become more deeply embedded within the membrane bilayer.

Control experiments were performed with the same membranes in the absence of gp60. Although there are interactions between the albumin or the HIV peptide and the simple membrane they are quite different from those observed in the presence of gp60.

Similar studies of the interactions of the isolated gp60 reconstituted into model membranes labelled with FPE with the gp41 and gp32 peptides indicate that they possess the ability to interact with gp60. Studies of the interactions shown in Fig. 3 also indicate that there is an initial fast binding reaction, followed by a slower conformational change.

These results indicate clearly that there are interactions between albumin, gp60 and HIV peptides. Agents that interfere with gp60 can prevent HIV from entering a target cell.

The following experimental protocols illustrate particular procedures that have been used. Production of Crude Cell Extract from ECV304 Endothelial Cells

ECV304 endothelial cells were grown to confluence in 225 cm² flasks then harvested and stored in their culture medium at -18°C. Approx 1 billion cells were collected washed twice with 30ml PBS (lO M sodium and potassium phosphate buffer containing 2.7mM potassium chloride, 137mM sodium chloride, pH 7.4) by precipitation at 4500g for 20 minutes at 4°C. Cells (approx. 4ml) were re-suspended in 6ml PBS containing 1% (v/v) Triton X-100 and then gently stirred on ice for ca. 60 mins. Centrifugation of this mixture at approx. 3000g for 5 minutes produced a pellet of 3-4ml volume and 6ml of supernatant. The supernatant was rapidly added to 45ml absolute ethanol at -18°C. The precipitate that formed was allowed to settle out overnight at -18°C. The protein precipitate in 88% cold ethanol and was precipitated at 3000g for 5 minutes. The ethanol supernatant was removed and 16ml 70% ethanol at -18°C was added to the pellet, gently resuspended and precipitated by centrifugation at 3000g for 5 minutes. The pellet was then gently stirred with 8ml PBS on ice for ca. 60 mins. Further centrifugation at approx. 3000g for 5 minutes produced a pellet of l-2ml. The supernatant (10ml) was adjusted to 1% Triton X-100 and 0.2% sodium dodecylsulphate (SDS) . Extraction of the pellet with PBS was repeated and a second aliquot of 10ml of the supernatant adjusted to 1% Triton X- 100 and 0.2% SDS.

An affinity column was prepared as follows: wheat germ agglutinin from Triticum vulgaris cross-linked to 4% beaded agarose (1ml bed volume) was regenerated by washing with PBS containing 0.5M NaCl, ImM MgCl₂, ImM MnCl₂, ImM ZnCl₂, and ImM CaCl₂. This solution was removed using 15ml PBS containing 1% Triton X-100 and 0.2% SDS. About 20ml of the cell extract was passed down the WGA-affinity column at a flow rate of approx. 0.2ml. min^"1, then washed with 15ml PBS containing 1% Triton X-100 and 0.2% SDS. The GP60 was released from the affinity column by eluting it with PBS buffer containing 1% Triton X-100, 0.2% SDS and 0.5M N- acetylglucosamine. Six 1ml aliquots were collected from the column and the N-acetylglucosamine-containing buffer replaced by PBS buffer containing 0.1% Triton X-100, 0.02% SDS. This was carried out using Centricon-10 ultrafiltration units that were centrifuged at 4500g. Fraction 2 was found to contain significant amounts of gp60 as indicated by ligand blot analysis using I¹²⁵-HSA . Preparation of FPE-labelled GP60 Vesicles and Control Liposomes

A sample of GP60 which had been stored in liquid nitrogen was thawed and dialysed against approx. 300 volumes of PBS containing 2.0% sodium cholate pH 7.4 overnight. A dispersion of PC and PS was prepared by drying a mixture of 17mg PC and 2.8mg PS under argon, washing with 0.5ml ethanol, and then redrying under argon. The phospholipid mixture was then dispersed in 1ml PBS containing 2% sodium cholate using a whirlimix (5-10 mins) . The sample was subjected to sonication using a microtip probe at power setting 4, 30% duty cycle for 10 mins. This procedure was carried out at 4°C. After sonication the sample was centrifuged at approx. 2000g for 5 mins. Two samples of dispersed phospholipids were prepared in this manner so that one could be used in the preparation of the control liposome sample.

0.9ml of the dialysed GP60 sample was mixed with 1ml of the PC/PS dispersion and transferred to benzoylated visking tubing (1000 Da cut off) . The samples were then dialysed against 250ml, 250mM sucrose, lOmM HEPES(Na+) , 0.68% sodium cholate pH 7.4 for 4 hours, then 250ml, 250mM sucrose, lOmM HEPES(Na-t-), 0.3% sodium cholate for 4 hours, and finally 250ml, 280mM sucrose, lOmM Tris/HCl pH 7.5 overnight (at least 16 hours) . This procedure was repeated but in the absence of GP60 in order to prepare a sample of control liposomes.

GP60 vesicles and control liposomes prepared as described above were labelled on their outer leaflet with FPE at a concentration of approx. 0.2% as described by Wall et al , supra . The integrity of the membrane vesicles was tested by observing fluorescence changes according to procedures described by Wall et al , supra .

Rapid kinetic stopped-flow studies were carried out on a SX-17 MV Applied Photophysics spectrofluorimeter apparatus equipped with a 150W xenon arc lamp light source. A small aliquot of protein sample was rapidly mixed with an equal volume of gp60 vesicles (or control liposomes) , typically approx. 150μl, and the sample illuminated with light of 490nm wavelength. All optical fluorescence emitted by the sample above approx. 512nm was monitored over time. The dead time was estimated to be between 1.2 and 1.5 ms. Consequently, analysis of the time courses was carried out no earlier than 1.8ms after the start of initiating each time course. All experiments were carried out in 280mM sucrose, lOmM Tris/HCl pH 7.5 buffer. The HSA was obtained from Armour Ltd. and treated to remove any residual bound fatty acids. HSA/oleate was prepared by equilibrating a six-fold excess of oleic acid with the HSA sample over approx. 5 hours at room temperature with gentle stirring. Binding of HSA Fragments

Different HSA fragments were tested, for their binding to gp60 liposomes. The results are shown in Fig. 5; FragA, FragB, FragC and FragD are respectively amino-acids 1-123, 1-298, 298-585 and 123-298. FragA shows little binding.

Claims

1. Use of a material for the manufacture of a medicament for preventing HIV infection, wherein the material does not bind lipids but binds preferentially to a receptor selected from the gp60 receptor and albumin receptors, on target cells.

2. Use of a material for the manufacture of a medicament for preventing HIV infection, wherein the material is selected from anti-gp60 antibody, gp60 receptor protein, gp41 receptor-binding domain, non-fatty acid-binding albumin, and fragments thereof that bind to the gp60 receptor .

3. Use according to claim 1 or claim 2 , wherein the material is an albumin fragment.

4. Use according to claim 3, wherein the material consists of or includes amino-acids 45-65, 45-585, 123-298 or 298-585 of albumin, or a fragment thereof.

5. Use according to any preceding claim, wherein the material is a peptide and the medicament is adapted to be administered via a route suitable for an insoluble material.

6. Use according to any preceding claim, which additionally comprises using an anti-replicase, e.g. AZT.

7. An albumin fragment as defined in claim 4, for therapeutic use.