- Email: [email protected]
Aptamers as therapeutic and diagnostic agents
Reviews in Molecular Biotechnology 74 Ž2000. 5]13
Review article
Aptamers as therapeutic and diagnostic agents U
Edward N. Brody ,1, Larry Gold1 Department of Molecular, Cellular and De¨ elopmental Biology, Uni¨ ersity of Colorado, Boulder, CO 80309-0347, USA Received 29 November 1999; accepted 2 December 1999
Abstract Aptamers are oligonucleotides derived from an in vitro evolution process called SELEX. Aptamers have been evolved to bind proteins which are associated with a number of disease states. Using this method, many powerful antagonists of such proteins have been found. In order for these antagonists to work in animal models of disease and in humans, it is necessary to modify the aptamers. First of all, sugar modifications of nucleoside triphosphates are necessary to render the resulting aptamers resistant to nucleases found in serum. Changing the 29OH groups of ribose to 29F or 29NH 2 groups yields aptamers which are long lived in blood. The relatively low molecular weight of aptamers Ž8000]12 000. leads to rapid clearance from the blood. Aptamers can be kept in the circulation from hours to days by conjugating them to higher molecular weight vehicles. When modified, conjugated aptamers are injected into animals, they inhibit physiological functions known to be associated with their target proteins. A new approach to diagnostics is also described. Aptamer arrays on solid surfaces will become available rapidly because the SELEX protocol has been successfully automated. The use of photo-cross-linkable aptamers will allow the covalent attachment of aptamers to their cognate proteins, with very low backgrounds from other proteins in body fluids. Finally, protein staining with any reagent which distinguishes functional groups of amino acids from those of nucleic acids Žand the solid support. will give a direct readout of proteins on the solid support. Q 2000 Published by Elsevier Science B.V. All rights reserved. Keywords: SELEX; Photo-SELEX; Universal protein stain; Clearance; Oligonucleotide; Pharmaceuticals
U
Corresponding author. Tel.: q1-303-499-6846; fax: q1-303-499-8327. E-mail address: [email protected] ŽE.N. Brody. 1 Present address: SomaLogic, Inc.; 1033 5 th Street; Boulder, CO 80302, USA. 1389-0352r00r$ - see front matter Q 2000 Published by Elsevier Science B.V. All rights reserved. PII: S 1 3 8 9 - 0 3 5 2 Ž 9 9 . 0 0 0 0 4 - 5
6
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
1. Introduction Aptamers are single-stranded nucleic acids evolved in vitro to perform a specific function ŽTuerk and Gold, 1990; Ellington and Szostak, 1990.. This in vitro procedure has been called SELEX ŽSystematic Evolution of Ligands by EXponential Enrichment . ŽTuerk and Gold, 1990.. SELEX begins with a large population of singlestranded nucleic acid molecules which is challenged with a specific task. Usually, the task is to bind to a purified protein, but many other tasks have been successfully selected. These include cell binding, peptide binding, small molecule binding, nucleic acid binding and catalysis of a variety of chemical reactions ŽGold et al., 1995; Osborne and Ellington, 1997.. A requirement for a successful SELEX experiment, is to determine a method for partitioning those nucleic acid molecules that have performed the task from those that have not. For example, nucleic acids that bind to proteins can be separated from free nucleic acids by nitro-cellulose filter binding, affinity chromatography, size fractionation on columns, or electrophoresis on native polyacrylamide gels. The selected nucleic acid molecules are subsequently amplified as DNA. If the single-stranded library is RNA or RNA analogs, it is first copied as complementary DNA by reverse transcriptase. This DNA is then amplified by PCR. Single-stranded DNA molecules are amplified directly by PCR. The DNA population, now pared down by one round of selection, is subsequently used to generate a new library of single-stranded molecules Žtranscription for RNA; strand separation for DNA. which is again subjected to the selection procedure. This process is repeated until a population of single-stranded molecules is selected which performs the task adequately. The starting population generally has fixed ends for PCR-primer annealing and, when appropriate, to allow copying at one end by T7 RNA polymerase. Starting libraries often contain either 30- or 40nucleotide long random sequences in between the fixed ends. This means that there are potentially 4 30 Ž10 18 . or 4 40 Ž10 24 . individual sequences available for partitioning. For reasons of commodity, the starting round contains no more than
10 14 ]10 15 individual sequences. This is, obviously, a very large number. In the decade or so since the SELEX procedure was first introduced, an abundant literature ŽGold et al., 1995; Osborne and Ellington, 1997; Brody et al., 1999., has reinforced the initial insight that single-stranded nucleic acids are a source of a vast number of three-dimensional shapes. Because of this, the SELEX procedure almost always finds, in the initial library, a family of sequences which will bind strongly to any protein. Of the first 100 SELEX experiments done at NeXstarrGilead Sciences, approximately 80% had K d values under 1 nM ŽBrody et al., 1999.. This binding equilibrium compares very favorably with the binding strength of most antigen]antibody complexes ŽGriffiths et al., 1994.. Moreover, the comparison made is usually between a monovalent aptamer and a divalent antibody. Using two different multimeric proteins, dimerization of the cognate aptamer led to a large increase in binding strength. In one of these cases, this increase was due, as expected, to a decreased dissociation rate constant for the dimerized aptamer from the protein ŽRingquist and Parma, 1998.. Because aptamers can be evolved to bind tightly and specifically to almost any protein, they have been regarded as potential pharmaceutical and diagnostic agents ŽGold, 1995.. This review will focus on the use of aptamers in these applications. It shall address issues of specificity of binding, the stability and pharmacokinetics of aptamers, cost of goods of these compounds and covalent cross-linking of aptamers to proteins for diagnostics.
2. Therapeutics In order for aptamers to be useful therapeutic reagents, they must bind tightly to proteins, inhibit a specified function of that protein Žif an antagonist is desired, as is usually the case., and have no harmful side-effects. Thus, the same criteria that apply to therapeutic antibodies apply to therapeutic aptamers. Basic fibroblast growth factor ŽbFGF. has been implicated in neovascularization of tumors as well as in a number of other
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
disease processes ŽBasilico and Moscatelli, 1992; Jellinek et al., 1995.. This protein belongs to a family of FGF proteins, which in addition to bFGF, includes at least eight other members ŽBasilico and Moscatelli, 1992; Tanaka et al., 1992; Miyamoto et al., 1993.. An aptamer evolved to bind bFGF did not bind significantly to any of the other members of the FGF family ŽJellinek et al., 1995.. This aptamer binds to the heparin-binding site of bFGF. Since aptamers, like heparin, are polyanions, one would have imagined that aptamers would show little specificity for heparinbinding sites on proteins. Four heparin-binding proteins ŽVEGF, PDGF AB, anti-thrombin III and thrombin. were tested with the bFGF aptamer; however, none bind significantly. This illustrates a general finding; although aptamers bind frequently to epitopes of proteins that are polyanion binding sites, each polyanion binding site]aptamer pair shows specificity. In fact, aptamers evolved to the same protein from different mammalian species show specificity unless there is a very high sequence of conservation in the protein. Another example of specificity comes from NX 1838, an aptamer evolved to bind to the 165 amino acid isoform of the vascular endothelial growth factor ŽVEGF.. VEGF165 includes exon 7 sequences, which code for a heparin-binding site. This aptamer does not react at all with VEGF121 , an isoform lacking exon 7 sequences ŽRuckman et al., 1998.. Furthermore, this aptamer as a monomer, is specific for the VEGF165 dimer, as all binding is lost when the protein dimer is reduced by a sulfydryl reagent ŽTable 1.. Although NX 1838 does not bind to the dimer of the related Ž53% homology. placenta growth factor ŽPIGF., a hetero-dimer of VEGFrPIGF does bind NX 1838 approximately 10-fold less efficiently than it binds VEGF165 ŽTable 1.. We should emphasize that this specificity has been seen in simple SELEX procedures. For therapeutic aptamers, we have not yet needed to exploit a parameter which, in principle, can add exquisite specificity: this parameter is counterSELEX. If, in the search for a therapeutic aptamer, one wants to guarantee that the aptamer will distinguish between a target and a known
7
Table 1 This table shows the K d of aptamer NX 1838 which was evolved to bind to human vascular endothelial growth factor 165 ŽVegF165 . a Protein
Kd ŽpM.
Human VEGF165 Mouse VEGF164 Human VEGF165 Žreduced. Human VEGF121 Human PIGF129 Human VEGF165 rPIGF129 PDGF-AB TGFb1 bFGF
49 " 6 90 " 22 ) 1 000 000 ) 1 000 000 ) 1 000 000 432 " 154 800 000 ) 1 000 000 220 000
a The aptamer binds to this 165 amino acid form of the protein with a K d of 49 pM. It binds almost equally strongly to the homologous mouse protein. When the homodimer VegF is reduced to monomers with sulfhydryl reagents, binding is destroyed. Binding is to the heparin-binding domain of VegF. The VegF121 isoform of this protein, which does not contain this heparin binding site, does not bind to the aptamer. A related heparin binding protein, PIGF129 , does not bind the aptamer, nor do three other heparin-binding proteins listed at the bottom of the table. A VegF165 rPIGF129 heterodimer binds NX 1838 approximately one-tenth as well as does the VegF165 homodimer ŽRuckman et al., 1998..
closely related protein Žor proteins. then one can employ alternate rounds of positive selection for the target Žtaking only molecules that bind. with negative selection for the related protein Žtaking only molecules that do not bind.. This has worked successfully in aptamer evolution to small molecules, permitting an aptamer to bind 10 000fold better to theophylline than to caffeine, even though the only difference between these two molecules is one methyl group on the N7 position of a purine ring ŽJenison et al., 1994.. Specificity is only one of the requirements for turning aptamers into therapeutics. Another is stability. Clearly, unmodified RNA cannot be used as a therapeutic agent since blood is rich in ribonucleases. However, appropriate modification of single-stranded RNA and DNA can produce molecules which are stable in blood. The most active RNAase in blood has the same specificity as pancreatic RNAase, which cuts specifically after pyrimidine ribonucleotides. Thus, changing the 29OH group of ribose to 29NH 2 or 29F-modified sugars on pyrimidines is sufficient to give resis-
8
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
Fig. 1. This cartoon is based on the crystallographic structure of a protein target for SELEX Žin dark blue. and the NMR structure of an aptamer to a closely related protein Žin white.. In frame 1, the yellow residue on the protein and the maroon and green residues on the aptamer represent chemical groups that will be hidden by the complex protein]aptamer. In frames 1]4, the red groups represent functional groups on the aptamer and on the protein which, after interaction lead to functional groups that can be cross-linked chemically once aptamer formation has been optimized, and the protein]aptamer complex has been exposed to chemical cross-linking protocols. In a photo-SELEX protocol these residues do not cross-link, although they are in close proximity. The blue residue, by contrast Žframes 3 and 4., is in a photo-cross-linkable orientation, and will be cross-linked to an appreciable degree after applying the photo-SELEX protocol. In frame 4, the yellow residues represent lysine residues reacting specifically with a universal protein stain which distinguishes lysine residues on proteins from the N7 residues of guanine, its closest competing nucleophile.
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
tance to nuclease breakdown in serum for over 2 days. Because these modifications on nucleoside triphosphates are compatible with their use by T7 RNA polymerase and the resultant modified oligonucleotides are faithful substrates for reverse transcriptase, such modifications are compatible with the SELEX protocol ŽEaton and Pieken, 1995; Jellinek et al., 1995.. All aptamers evolved for therapeutic use have either 29F or 29NH 2 groups within each pyrimidine nucleotide. Further modifications are used in preparing aptamers for therapeutic purposes. Once a winning sequence is selected, it is synthesized chemically using machine synthesis with the appropriate phosphoramidites. Ordinarily, this synthesis is started Žsince synthesis is 39 ª 59. with a dT 39]39 dT cap Žfor either DNA or RNA single-strand synthesis . which renders the oligonucleotides resistant to 39]59 exonucleases which are present in blood and in other tissues. Finally, the 29OH groups on purine nucleosides which could be substrates for other Žminor. RNAases in blood are methylated to discover which ones are essential for binding. Usually, from one-half Žusual. to all Žvery rarely. of the 29OH groups on purine nucleosides can be converted to 29OCH 3 groups without substantially changing the binding properties of the aptamer. For example, NX 1838, the aptamer currently used in clinical trials for agerelated macular degeneration, is a 29F pyrimidine-containing RNA analogue in which all but two purine nucleosides are modified by 29OCH 3 groups ŽRuckman et al., 1998.. Sugar modifications may ensure stability but they do not guarantee adequate pharmacokinetics for aptamers to be therapeutically active. Aptamers are in the range of M W 8000]12 000; they are cleared from plasma within minutes of IV injection, probably through renal excretion. Keeping intact aptamers in the blood from hours to days after injection has been accomplished by conjugating them to larger macromolecules. Polyethyleneglycol ŽPEG. can be linked to a primary amine on the 59 end of an RNA analogue ŽRuckman et al., 1998.; PEG of 20 000 and 40 000 M W have been tested. The PEG of 40 000 M W has lower plasma clearance and, as shall be discussed later, is adequate to show in vivo inhibition of
9
targeted proteins. Decreasing plasma clearance of aptamers has also been achieved by embedding them in liposomes ŽWillis et al., 1998.. A stabilized aptamer to VEGF was derivatized at its 59 end with a dialkylglycerol ŽDAG . moiety. The lipophilic DAG inserts into the lipid bilayer of liposomes Žsome with the aptamer pointing into the liposome and some with the aptamer pointing into solution.. The liposomal VEGF aptamers were inhibitors of both vascular permeability and angiogenesis in vivo ŽWillis et al., 1998.. It should be emphasized here, that for multimeric proteins and monovalent aptamers, concentrating aptamers on liposomes could lead to a decrease in K off of the aptamer and thus a large gain in efficacy. This has yet to be approached experimentally. Thus, aptamers can be stabilized chemically and modified for pharmacokinetics. Do they then work as therapeutics? The response is that aptamers work in vivo in animal models quite consistently; one aptamer, to VEGF } NX 1838 } is currently in clinical trials for age-related macular degeneration, the leading cause of blindness in the elderly. Aptamers to a number of extra-cellular targets, including L-selectin ŽWatson et al., 1999., VEGF ŽOstendorf et al., 1999., P-selectin ŽJenison et al., 1998., and PDGF ŽFloege et al., 1999; Leppanen et al., 1999. have each shown in vivo efficacy in animal models. NX 1838, the 29F pyrimidine aptamer to VEGF is, when conjugated to 40 kDa PEG, active in a variety of animal models. This aptamer inhibited VEGF-induced vascular permeability by more than 80% ŽRuckman et al., 1998.. In these experiments, VEGF Ž20 nM. was pre-mixed with the aptamer Ž0.1 mM. before injection. More recently, this aptamer has been shown to be as effective, injected daily at 10 mgrkg body wt., in slowing human rhabdomyosarcoma xenografts in mice, as is twice weekly injection of an anti-VEGF monoclonal antibody ŽBlake Tomkinson, personal communication.. Also, in rats with mesangioproliferative nephritis, use of NX 1838 showed that VEGF165 mediates endothelial proliferation in damaged glomerular capillaries ŽOstendorf et al., 1999.. No effect of NX 1838 was seen on normal glomeruli. This, in fact, suggests that endothelial cells in
10
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
glomeruli, like endothelial cells in tumors, are VEGF-dependent during growth and remodeling, but lose this VEGF dependence once associated with pericytes Žor podocytes. ŽBenjamin et al., 1999; Darland and D’Amore, 1999.. Finally, NX 1838 was shown to be non-inflammatory and non-toxic when injected into the vitreous humor of rabbit eyes, and to be therapeutically active Žafter IV injection. in combating O 2-induced retinopathy in newborn mice ŽHenninger and Bill, personal communication.. Such data, among others, led to an Investigational New Drug acceptance for NX 1838 for human age-related macular degeneration; phase 1 trials for direct intravitreal injection of NX 1838 are currently underway. Perhaps the most thorough in vivo study of an aptamer was recently completed by David Parma and his collaborators ŽWatson et al., 1999.. A 29F pyrimidine-containing aptamer evolved against human L-selectin was studied in severe immunodeficient ŽSCID. mice. The aptamer binds to Lselectin on human lymphocytes and prevents leucocyte rolling, a prerequisite for attachment to capillary walls and extravasation to peripheral lymph nodes. A complete pharmacokinetic analysis of the 40-kDa PEG-modified aptamer showed that plasma clearance was the most important determinant of physiological activity. Clearance, presumably through renal excretion, is inversely related to the size of the aptamer conjugate. This was true for the L-selectin aptamer as the molecular weight of the conjugate increased to 50 kDa, beyond which there was no further decrease in plasma clearance. Over a 30-fold range of in vitro K d values, in vivo inhibition was not affected by the K d . Another result of this study is that nonspecific binding to leucocytes is not a problem at doses of aptamer which show a 50% reduction of leucocyte trafficking. Thus, for this cell surface target, efficient inhibition by the aptamer could be measured even when human leucocytes were injected 10 h after the IV bolus of aptamer had been injected. Because the in vivo activity of aptamers depends on their being conjugated to a higher molecular weight carrier, much research remains to be done on in vivo efficacy. Up to now, only
liposomes and PEG have been tested as carriers, both successfully. It is possible that conjugation of aptamers to other carriers, including human proteins that naturally circulate for weeks in the blood, will substantially decrease the clearance rate from plasma. Furthermore, liposomal presentation of aptamers has been carried out with only one liposomal formulation. Here too, further progress can be expected. The successful use of aptamers as therapeutic agents depends not only on their efficacy and specificity, but also on economics } that is, what are their costs of manufacture? The phosphoramidites for 29F pyrimidine and 29OCH3 purine monomers are readily available as is the 40 kDa molecular weight PEG used for conjugation. A number of companies now have Good Manufacturing Practice ŽGMP. facilities for the large scale synthesis of oligonucleotides, including RNA analogs. If we extrapolate from the most successful animal experiments in which target concentration is not limiting for efficacy, an aptamer dose of 1]2 mgrkg body wt. is usually an effective dose Žthis is derived from experiments on aptamers inhibiting VEGF, PDGF, L-selectin, and P-selectin .. For a 70-kg adult, this means that each injected dose would be 70]140 mg. Current quotes for aptamers approximately 35 nucleotides long containing modified ribonucleotides are approximately $2000rg Žwhen orders are for, minimally, 100 g.. This means that the manufacturing cost is $140]$280 per dose. For acute indications, such as organ transplant, myocardial infarcts, toxic or septic shock, angioplasty, or pulmonary embolism treatment every 3 days for 15 days would involve $700]$1400 in cost of goods. This is not out of line with many treatments, such as therapeutic antibodies, already in clinical use. Clearly, for chronic indications, the cost of the goods is an issue. There are a number of possible ways to reduce these costs. Firstly, a more efficient chemistry for aptamer synthesis, Product Anchored Sequential Synthesis ŽPASS., has been
1
Pieken and Gold Ž1999. Method for Solution Phase Synthesis of Oligonucleotides and Peptides. US Patent No. 5 874 532.
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
described.1 The PASS technology, although not yet optimized, could reduce the cost of aptamer synthesis by as much as a factor of 10 Žto approx. $200rg.. Also, as mentioned earlier, the studies on the pharmacokinetics of aptamers have only just begun. Any increase in circulatory half-life should reduce the cost of goods. This could come from new conjugates, new formulation of liposomes, or new delivery vehicles to slow down release from depots. Finally, it is worthwhile focusing on the relatively poor therapeutic results in cancer with standard chemotherapy. Although some cancers Že.g. leukemia and lymphomas. are curable by chemotherapy, many other frequent cancer types are almost never cured by standard cytotoxic reagents Žlung, brain, pancreas, colon, and breast cancer } once it becomes estrogen-independent.. The average life expectancy of a patient diagnosed with pancreatic cancer is only 8 months; if aptamer-based therapy against specific targets extended life to 5 or more years, even the present cost of goods would be found acceptable.
3. Diagnostics If one pictures aptamers as equivalent to antibodies, keeping in mind that their in vitro mode of production has the advantages discussed earlier, then it becomes obvious that aptamers can be ideal tools for the diagnostic evaluation of proteins. The standard diagnostic protocol for detecting proteins in blood, urine, saliva, and other body fluids, is the two-antibody sandwich assay. In this assay, a capture monoclonal antibody is used Žusually on a solid surface . to bind the protein in question. The signal-to-noise ratio of such a single capture step in a mixture of proteins as complex as blood, is never sufficient to provide a useful test. Because of this, a second antibody, either monoclonal to a different epitope, or polyclonal aimed at a different epitope, is then used to detect the bound protein. The second antibody is then detected by fluorescence or a colorimetric assay. In order for such an antibody-based test to be useful in diagnostics, it is essential that these two binding steps be indepen-
11
dent so that the background is reduced by the multiplicative specifity of the two steps. It is feasible to use the in vitro-derived aptamers in such a way that one aptamer can substitute for two antibodies. We have detailed the various steps involved in a recent review ŽBrody et al., 1999.. Here we shall outline the essential argument. Diagnostics is on the brink of a revolutionary change, which will be the consequence of sequencing the human genome during the next 2]3 years. It is expected that novel reagents Žwhich allow detection of pertinent human proteins in body fluids. will usher in this dramatic change in diagnostic capacity. There are three elements in the use of aptamers that should make them the reagents of choice to keep pace with the availability of new proteins as they are produced subsequently to sequencing the human genome. The first is the rapidity with which aptamers can be produced. This is a consequence of the construction of robots to perform the SELEX process ŽCox et al., 1998. ŽZichi et al., personal communication.. Since SELEX is an in vitro iterative protocol involving only enzymatic steps and separation of bound from unbound nucleic acids, it has been amenable to automation. It is now possible to adapt SELEX to a robotic procedure which can do approximately three rounds on 96 samples a day. Since the average successful SELEX procedure takes less than 10 rounds to complete Žespecially in the robotics format., one robotics machine can identify aptamers aimed at 3 = 96 = 365r10s; 10 000 protein reagents per year. The second element is the introduction of photo-SELEX ŽWillis et al., 1994; Jensen et al., 1995; Meisenheimer et al., 1996; Golden et al., 1999.. Photo-SELEX is a derivative of SELEX in which substrates for either DNA or RNA polymerase include a 5-iodo or 5-bromo-substituted dUTP or UTP ŽWillis et al., 1994; Meisenheimer et al., 1996.. A series of publications show that such substitutions allow for normal evolution of tight binding aptamers, but also allows that photoexcitation by long wavelength UV light leads to cross-linking of the aptamer to its targeted protein. Such SELEX experiments, in which the protocol demands evolution of aptamers which can be specifically photo-cross-linked to their cognate
12
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
proteins, have been successfully carried out ŽBrody et al., 1999; Golden et al., 1999.. The importance of this for diagnostics is in the independence of the cross-linkability parameter from the binding parameter. This independence is great enough to give multiplicative background reduction analogous to the independent binding of two different antibodies to the same protein in an ELISA format. This leads to the following diagnostic scenario. With arrays of aptamers on a solid surface, blood or any other body fluid can be incubated with this aptamer ‘chip’. Proteins will sort themselves out in a manner analogous to affinity chromatography ŽRomig et al., 1999.. As long as the dissociation rate constant for binding of each aptamer to its cognate protein is sufficiently low, a gentle wash step should increase the signal-to-noise ratio at each specific aptamer cluster. Then, laser-induced photo-cross-linking will introduce covalent bonds only with those aptamer]protein pairs whose geometry has been evolved to position a } Br or } I sufficiently close to an electron-rich amino acid. The geometric, thus chemical, requirements for this happening are very constrained in this protocol ŽWillis et al., 1994; Meisenheimer et al., 1996; Ruckman et al., 1998; Golden and Willis, personal communication.. Other bound proteins, specifically non-cognate proteins, will not be covalently cross-linked. Next, a harsh detergent wash should keep only the cross-linked proteins on the aptamer ‘chip’ ŽBrody et al., 1999. ŽFig. 1.. The third element of this diagnostic protocol results directly from the use of aptamers and photo-SELEX. The detection step with antibodies involves necessarily indirect measurements; this is because antibodies are proteins, like the analytes they are supposed to be measuring in blood. Such indirect measurements add expense and a further source of error into antibody-based detection systems. When one uses photo-cross-linkable aptamers, the only proteins on the ‘chips’ after the wash cycles are the analytes themselves. Thus, any stain which distinguishes proteins from nucleic acids Žany universal protein stain. can be used directly to measure protein levels in a body fluid. Very simple compounds are available. For example, any lysine-specific reagent could func-
tion well as such a stain ŽBrody et al., 1999. ŽFig. 1.. No nucleophile in nucleic acids is as strong as lysine. Other amino acid-specific reagents have been under investigation for more than 50 years ŽWong, 1991.. Aptamers have been investigated since 1990 ŽEllington and Szostak, 1990; Tuerk and Gold, 1990.. The SELEX protocol is simple and automatable. When one considers that both liposomes and monoclonal antibodies were discovered approximately 20 years before the first therapeutic applications were approved for clinical use, it seems that the medical applications of aptamers are on a very fast historical track. Acknowledgements The authors would like to thank all of their colleagues at NeXstar Pharmaceuticals, now a wholly owned subsidiary of Gilead Sciences, Foster City, CA, USA. Nebojsa Janjic and his colleagues contributed greatly to the therapeutic aptamer studies; David Parma and co-workers were instrumental in the L-selectin pharmacokinetic and efficacy studies; Dwight Henninger, Jerry Bill, Blake Tomkinson and Wendy Gillette, in Ray Bendele’s group allowed citation of some of their unpublished work. The Diagnostic and Robotic effort was directed by Dom Zichi, Sumedha Jayasena, Mike Willis, and Drew Smith. Our thanks also, to Kris Cordova for preparation of the manuscript. References Bsasilico, C., Moscatelli, D., 1992. The FGF family of growth factors and oncogenes. Adv. Cancer Res. 59, 115]165. Benjamin, L.E., Golijanin, D., Itin, A., Pode, D., Keshet, E., 1999. Selective ablation of immature blood vessels in established human tumors follows vascular endothelial growth factor withdrawal Žsee comments.. J. Clin. Invest. 103, 159]165. Brody, E.N., Willis, M.C., Smith, J.D., Jayasena, S., Zichi, D., Gold, L., 1999. The Use of Aptamers in Large Arrays for Molecular Diagnostics. Mol. Diagn., 4, 381]388. Cox, J.C., Rudolph, P., Ellington, A.D., 1998. Automated RNA selection. Biotechnol. Prog. 14, 845]850. Darland, D.C., D’Amore, P.A., 1999. Blood vessel maturation: vascular development comes of age comment. J. Clin. Invest. 103, 157]158.
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13 Eaton, B.E., Pieken, W.A., 1995. Ribonucleosides and RNA. Annu. Rev. Biochem. 64, 837]863. Ellington, A.D., Szostak, J.W., 1990. In vitro selection of RNA molecules that bind specific ligands. Nature 346, 818]822. Floege, J., Ostendorf, T., Janssen, U. et al., 1999. Novel approach to specific growth factor inhibition in vivo: antagonism of platelet-derived growth factor in glomerulonephritis by aptamers. Am. J. Pathol. 154, 169]179. Gold, L., 1995. Oligonucleotides as research, diagnostic, and therapeutic agents. J. Biol. Chem. 270, 13 581]13 584. Gold, L., Polisky, B., Uhlenbeck, O., Yarus, M., 1995. Diversity of oligonucleotide functions. Annu. Rev. Biochem. 64, 763]797. Golden, M., Collins, B.D., Willis, M.C., Koch, T.H., 1999. Diagnostic potential of photo SELEX-evolved ss aptamers. J. Biotechnol., submitted for publication. Griffiths, A.D., Williams, S.C., Hartley, O. et al., 1994. Isolation of high affinity human antibodies directly from large synthetic repertoires. EMBO J. 13, 3245]3260. Jellinek, D., Green, L.S., Bell, C. et al., 1995. Potent 29-amino29-deoxypyrimidine RNA inhibitors of basic fibroblast growth factor. Biochemistry 34, 11 363]11 372. Jenison, R.D., Gill, S.C., Pardi, A., Polisky, B., 1994. High-resolution molecular discrimination by RNA. Science 263, 5152. Jenison, R.D., Jennings, S.D., Walker, D.W., Bargatze, R.F., Parma, D., 1998. Oligonucleotide inhibitors of P-selectindependent neutrophil]platelet adhesion. Antisense Nucleic Acid Drug Dev. 8, 265]279. Jensen, K.B., Atkinson, B.L., Willis, M.C., Koch, T.H., Gold, L., 1995. Using in vitro selection to direct the covalent attachment of human immunodeficiency virus type 1 Rev protein to high-affinity RNA ligands. Proc. Natl. Acad. Sci. USA 92, 12 220]12 224. Leppanen, O., Janjic, N., Carlsson, M.A. et al., 1999. PDGF B-chain Aptamer Reduces Pseudointimal Hyperplasia in a Rat Model of Arterial Restenosis, submitted for publication . Meisenheimer, K.M., Meisenheimer, P.L., Willis, M.C., Koch, T.H., 1996. High yield photocrosslinking of a 5-iodocytidine ŽIC. substituted RNA to its associated protein. Nucleic Acids Res. 24, 981]982. Miyamoto, M., Naruo, K., Seko, C., Matsumoto, S., Kondo, T., Kurokawa, T., 1993. Molecular cloning of a novel cytokine cDNA encoding the ninth member of the fibroblast growth
13
factor family, which has a unique secretion property. Mol. Cell Biol. 13, 4251]4259. Osborne, S.E., Ellington, A.D., 1997. Nucleic acid selection and the challenge of combinatorial chemistry. Chem. Rev. 97, 349]370. Ostendorf, T., Kunter, U., Eitner, F. et al., 1999. VEGF165 mediates glomerular endothelial repair. J. Clin. Invest., in press . Ringquist, S., Parma, D., 1998. Anti-L-selectin oligonucleotide ligands recognize CD62L-positive leukocytes: binding affinity and specificity of univalent and bivalent ligands. Cytometry 33, 394]405. Romig, T., Bell, C., Drolet, D., 1999. Aptamer affinity chromatography; combinatorial chemistry applied to protein purification. J. Chromatogr., in press . Ruckman, J., Green, L.S., Beeson, J. et al., 1998. 29-Fluoropyrimidine RNA-based aptamers to the 165-amino acid form of vascular endothelial growth factor ŽVEGF165.. Inhibition of receptor binding and VEGF-induced vascular permeability through interactions requiring the exon 7-encoded domain. J. Biol. Chem. 273, 20 556]20 567. Tanaka, A., Miyamoto, K., Minamino, N. et al., 1992. Cloning and characterization of an androgen-induced growth factor essential for the androgen-dependent growth of mouse mammary carcinoma cells. Proc. Natl. Acad. Sci. USA 89, 8928]8932. Tuerk, C., Gold, L., 1990. Systematic Evolution of Ligands by Exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase. Science 249, 505]510. Watson, S.R., Chang, Y.F., O’Connell, D., Weigand, L., Parma, D., 1999. Anti-L-selectin aptamers: binding characteristics, pharmacokinetic parameters and activity against an intravascular target in vivo. Antisense Nucl. Acid Drug Dev., in press . Willis, M.C., Collins, B.D., Zhang, T. et al., 1998. Liposomeanchored vascular endothelial growth factor aptamers. Published erratum appears in Bioconjug Chem 1998 Sep]Oct; 9Ž5.:633. Bioconjug. Chem. 9, 573]582. Willis, M.C., LeCuyer, K.A., Meisenheimer, K.M., Uhlenbeck, O.C., Koch, T.H., 1994. An RNA-protein contact determined by 5-bromouridine substitution, photocrosslinking and sequencing. Nucleic Acids Res. 22, 4947]4952. Wong, S., 1991. Chemistry of Protein Crosslinking . CRC Press.
Review article
Aptamers as therapeutic and diagnostic agents U
Edward N. Brody ,1, Larry Gold1 Department of Molecular, Cellular and De¨ elopmental Biology, Uni¨ ersity of Colorado, Boulder, CO 80309-0347, USA Received 29 November 1999; accepted 2 December 1999
Abstract Aptamers are oligonucleotides derived from an in vitro evolution process called SELEX. Aptamers have been evolved to bind proteins which are associated with a number of disease states. Using this method, many powerful antagonists of such proteins have been found. In order for these antagonists to work in animal models of disease and in humans, it is necessary to modify the aptamers. First of all, sugar modifications of nucleoside triphosphates are necessary to render the resulting aptamers resistant to nucleases found in serum. Changing the 29OH groups of ribose to 29F or 29NH 2 groups yields aptamers which are long lived in blood. The relatively low molecular weight of aptamers Ž8000]12 000. leads to rapid clearance from the blood. Aptamers can be kept in the circulation from hours to days by conjugating them to higher molecular weight vehicles. When modified, conjugated aptamers are injected into animals, they inhibit physiological functions known to be associated with their target proteins. A new approach to diagnostics is also described. Aptamer arrays on solid surfaces will become available rapidly because the SELEX protocol has been successfully automated. The use of photo-cross-linkable aptamers will allow the covalent attachment of aptamers to their cognate proteins, with very low backgrounds from other proteins in body fluids. Finally, protein staining with any reagent which distinguishes functional groups of amino acids from those of nucleic acids Žand the solid support. will give a direct readout of proteins on the solid support. Q 2000 Published by Elsevier Science B.V. All rights reserved. Keywords: SELEX; Photo-SELEX; Universal protein stain; Clearance; Oligonucleotide; Pharmaceuticals
U
Corresponding author. Tel.: q1-303-499-6846; fax: q1-303-499-8327. E-mail address: [email protected] ŽE.N. Brody. 1 Present address: SomaLogic, Inc.; 1033 5 th Street; Boulder, CO 80302, USA. 1389-0352r00r$ - see front matter Q 2000 Published by Elsevier Science B.V. All rights reserved. PII: S 1 3 8 9 - 0 3 5 2 Ž 9 9 . 0 0 0 0 4 - 5
6
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
1. Introduction Aptamers are single-stranded nucleic acids evolved in vitro to perform a specific function ŽTuerk and Gold, 1990; Ellington and Szostak, 1990.. This in vitro procedure has been called SELEX ŽSystematic Evolution of Ligands by EXponential Enrichment . ŽTuerk and Gold, 1990.. SELEX begins with a large population of singlestranded nucleic acid molecules which is challenged with a specific task. Usually, the task is to bind to a purified protein, but many other tasks have been successfully selected. These include cell binding, peptide binding, small molecule binding, nucleic acid binding and catalysis of a variety of chemical reactions ŽGold et al., 1995; Osborne and Ellington, 1997.. A requirement for a successful SELEX experiment, is to determine a method for partitioning those nucleic acid molecules that have performed the task from those that have not. For example, nucleic acids that bind to proteins can be separated from free nucleic acids by nitro-cellulose filter binding, affinity chromatography, size fractionation on columns, or electrophoresis on native polyacrylamide gels. The selected nucleic acid molecules are subsequently amplified as DNA. If the single-stranded library is RNA or RNA analogs, it is first copied as complementary DNA by reverse transcriptase. This DNA is then amplified by PCR. Single-stranded DNA molecules are amplified directly by PCR. The DNA population, now pared down by one round of selection, is subsequently used to generate a new library of single-stranded molecules Žtranscription for RNA; strand separation for DNA. which is again subjected to the selection procedure. This process is repeated until a population of single-stranded molecules is selected which performs the task adequately. The starting population generally has fixed ends for PCR-primer annealing and, when appropriate, to allow copying at one end by T7 RNA polymerase. Starting libraries often contain either 30- or 40nucleotide long random sequences in between the fixed ends. This means that there are potentially 4 30 Ž10 18 . or 4 40 Ž10 24 . individual sequences available for partitioning. For reasons of commodity, the starting round contains no more than
10 14 ]10 15 individual sequences. This is, obviously, a very large number. In the decade or so since the SELEX procedure was first introduced, an abundant literature ŽGold et al., 1995; Osborne and Ellington, 1997; Brody et al., 1999., has reinforced the initial insight that single-stranded nucleic acids are a source of a vast number of three-dimensional shapes. Because of this, the SELEX procedure almost always finds, in the initial library, a family of sequences which will bind strongly to any protein. Of the first 100 SELEX experiments done at NeXstarrGilead Sciences, approximately 80% had K d values under 1 nM ŽBrody et al., 1999.. This binding equilibrium compares very favorably with the binding strength of most antigen]antibody complexes ŽGriffiths et al., 1994.. Moreover, the comparison made is usually between a monovalent aptamer and a divalent antibody. Using two different multimeric proteins, dimerization of the cognate aptamer led to a large increase in binding strength. In one of these cases, this increase was due, as expected, to a decreased dissociation rate constant for the dimerized aptamer from the protein ŽRingquist and Parma, 1998.. Because aptamers can be evolved to bind tightly and specifically to almost any protein, they have been regarded as potential pharmaceutical and diagnostic agents ŽGold, 1995.. This review will focus on the use of aptamers in these applications. It shall address issues of specificity of binding, the stability and pharmacokinetics of aptamers, cost of goods of these compounds and covalent cross-linking of aptamers to proteins for diagnostics.
2. Therapeutics In order for aptamers to be useful therapeutic reagents, they must bind tightly to proteins, inhibit a specified function of that protein Žif an antagonist is desired, as is usually the case., and have no harmful side-effects. Thus, the same criteria that apply to therapeutic antibodies apply to therapeutic aptamers. Basic fibroblast growth factor ŽbFGF. has been implicated in neovascularization of tumors as well as in a number of other
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
disease processes ŽBasilico and Moscatelli, 1992; Jellinek et al., 1995.. This protein belongs to a family of FGF proteins, which in addition to bFGF, includes at least eight other members ŽBasilico and Moscatelli, 1992; Tanaka et al., 1992; Miyamoto et al., 1993.. An aptamer evolved to bind bFGF did not bind significantly to any of the other members of the FGF family ŽJellinek et al., 1995.. This aptamer binds to the heparin-binding site of bFGF. Since aptamers, like heparin, are polyanions, one would have imagined that aptamers would show little specificity for heparinbinding sites on proteins. Four heparin-binding proteins ŽVEGF, PDGF AB, anti-thrombin III and thrombin. were tested with the bFGF aptamer; however, none bind significantly. This illustrates a general finding; although aptamers bind frequently to epitopes of proteins that are polyanion binding sites, each polyanion binding site]aptamer pair shows specificity. In fact, aptamers evolved to the same protein from different mammalian species show specificity unless there is a very high sequence of conservation in the protein. Another example of specificity comes from NX 1838, an aptamer evolved to bind to the 165 amino acid isoform of the vascular endothelial growth factor ŽVEGF.. VEGF165 includes exon 7 sequences, which code for a heparin-binding site. This aptamer does not react at all with VEGF121 , an isoform lacking exon 7 sequences ŽRuckman et al., 1998.. Furthermore, this aptamer as a monomer, is specific for the VEGF165 dimer, as all binding is lost when the protein dimer is reduced by a sulfydryl reagent ŽTable 1.. Although NX 1838 does not bind to the dimer of the related Ž53% homology. placenta growth factor ŽPIGF., a hetero-dimer of VEGFrPIGF does bind NX 1838 approximately 10-fold less efficiently than it binds VEGF165 ŽTable 1.. We should emphasize that this specificity has been seen in simple SELEX procedures. For therapeutic aptamers, we have not yet needed to exploit a parameter which, in principle, can add exquisite specificity: this parameter is counterSELEX. If, in the search for a therapeutic aptamer, one wants to guarantee that the aptamer will distinguish between a target and a known
7
Table 1 This table shows the K d of aptamer NX 1838 which was evolved to bind to human vascular endothelial growth factor 165 ŽVegF165 . a Protein
Kd ŽpM.
Human VEGF165 Mouse VEGF164 Human VEGF165 Žreduced. Human VEGF121 Human PIGF129 Human VEGF165 rPIGF129 PDGF-AB TGFb1 bFGF
49 " 6 90 " 22 ) 1 000 000 ) 1 000 000 ) 1 000 000 432 " 154 800 000 ) 1 000 000 220 000
a The aptamer binds to this 165 amino acid form of the protein with a K d of 49 pM. It binds almost equally strongly to the homologous mouse protein. When the homodimer VegF is reduced to monomers with sulfhydryl reagents, binding is destroyed. Binding is to the heparin-binding domain of VegF. The VegF121 isoform of this protein, which does not contain this heparin binding site, does not bind to the aptamer. A related heparin binding protein, PIGF129 , does not bind the aptamer, nor do three other heparin-binding proteins listed at the bottom of the table. A VegF165 rPIGF129 heterodimer binds NX 1838 approximately one-tenth as well as does the VegF165 homodimer ŽRuckman et al., 1998..
closely related protein Žor proteins. then one can employ alternate rounds of positive selection for the target Žtaking only molecules that bind. with negative selection for the related protein Žtaking only molecules that do not bind.. This has worked successfully in aptamer evolution to small molecules, permitting an aptamer to bind 10 000fold better to theophylline than to caffeine, even though the only difference between these two molecules is one methyl group on the N7 position of a purine ring ŽJenison et al., 1994.. Specificity is only one of the requirements for turning aptamers into therapeutics. Another is stability. Clearly, unmodified RNA cannot be used as a therapeutic agent since blood is rich in ribonucleases. However, appropriate modification of single-stranded RNA and DNA can produce molecules which are stable in blood. The most active RNAase in blood has the same specificity as pancreatic RNAase, which cuts specifically after pyrimidine ribonucleotides. Thus, changing the 29OH group of ribose to 29NH 2 or 29F-modified sugars on pyrimidines is sufficient to give resis-
8
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
Fig. 1. This cartoon is based on the crystallographic structure of a protein target for SELEX Žin dark blue. and the NMR structure of an aptamer to a closely related protein Žin white.. In frame 1, the yellow residue on the protein and the maroon and green residues on the aptamer represent chemical groups that will be hidden by the complex protein]aptamer. In frames 1]4, the red groups represent functional groups on the aptamer and on the protein which, after interaction lead to functional groups that can be cross-linked chemically once aptamer formation has been optimized, and the protein]aptamer complex has been exposed to chemical cross-linking protocols. In a photo-SELEX protocol these residues do not cross-link, although they are in close proximity. The blue residue, by contrast Žframes 3 and 4., is in a photo-cross-linkable orientation, and will be cross-linked to an appreciable degree after applying the photo-SELEX protocol. In frame 4, the yellow residues represent lysine residues reacting specifically with a universal protein stain which distinguishes lysine residues on proteins from the N7 residues of guanine, its closest competing nucleophile.
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
tance to nuclease breakdown in serum for over 2 days. Because these modifications on nucleoside triphosphates are compatible with their use by T7 RNA polymerase and the resultant modified oligonucleotides are faithful substrates for reverse transcriptase, such modifications are compatible with the SELEX protocol ŽEaton and Pieken, 1995; Jellinek et al., 1995.. All aptamers evolved for therapeutic use have either 29F or 29NH 2 groups within each pyrimidine nucleotide. Further modifications are used in preparing aptamers for therapeutic purposes. Once a winning sequence is selected, it is synthesized chemically using machine synthesis with the appropriate phosphoramidites. Ordinarily, this synthesis is started Žsince synthesis is 39 ª 59. with a dT 39]39 dT cap Žfor either DNA or RNA single-strand synthesis . which renders the oligonucleotides resistant to 39]59 exonucleases which are present in blood and in other tissues. Finally, the 29OH groups on purine nucleosides which could be substrates for other Žminor. RNAases in blood are methylated to discover which ones are essential for binding. Usually, from one-half Žusual. to all Žvery rarely. of the 29OH groups on purine nucleosides can be converted to 29OCH 3 groups without substantially changing the binding properties of the aptamer. For example, NX 1838, the aptamer currently used in clinical trials for agerelated macular degeneration, is a 29F pyrimidine-containing RNA analogue in which all but two purine nucleosides are modified by 29OCH 3 groups ŽRuckman et al., 1998.. Sugar modifications may ensure stability but they do not guarantee adequate pharmacokinetics for aptamers to be therapeutically active. Aptamers are in the range of M W 8000]12 000; they are cleared from plasma within minutes of IV injection, probably through renal excretion. Keeping intact aptamers in the blood from hours to days after injection has been accomplished by conjugating them to larger macromolecules. Polyethyleneglycol ŽPEG. can be linked to a primary amine on the 59 end of an RNA analogue ŽRuckman et al., 1998.; PEG of 20 000 and 40 000 M W have been tested. The PEG of 40 000 M W has lower plasma clearance and, as shall be discussed later, is adequate to show in vivo inhibition of
9
targeted proteins. Decreasing plasma clearance of aptamers has also been achieved by embedding them in liposomes ŽWillis et al., 1998.. A stabilized aptamer to VEGF was derivatized at its 59 end with a dialkylglycerol ŽDAG . moiety. The lipophilic DAG inserts into the lipid bilayer of liposomes Žsome with the aptamer pointing into the liposome and some with the aptamer pointing into solution.. The liposomal VEGF aptamers were inhibitors of both vascular permeability and angiogenesis in vivo ŽWillis et al., 1998.. It should be emphasized here, that for multimeric proteins and monovalent aptamers, concentrating aptamers on liposomes could lead to a decrease in K off of the aptamer and thus a large gain in efficacy. This has yet to be approached experimentally. Thus, aptamers can be stabilized chemically and modified for pharmacokinetics. Do they then work as therapeutics? The response is that aptamers work in vivo in animal models quite consistently; one aptamer, to VEGF } NX 1838 } is currently in clinical trials for age-related macular degeneration, the leading cause of blindness in the elderly. Aptamers to a number of extra-cellular targets, including L-selectin ŽWatson et al., 1999., VEGF ŽOstendorf et al., 1999., P-selectin ŽJenison et al., 1998., and PDGF ŽFloege et al., 1999; Leppanen et al., 1999. have each shown in vivo efficacy in animal models. NX 1838, the 29F pyrimidine aptamer to VEGF is, when conjugated to 40 kDa PEG, active in a variety of animal models. This aptamer inhibited VEGF-induced vascular permeability by more than 80% ŽRuckman et al., 1998.. In these experiments, VEGF Ž20 nM. was pre-mixed with the aptamer Ž0.1 mM. before injection. More recently, this aptamer has been shown to be as effective, injected daily at 10 mgrkg body wt., in slowing human rhabdomyosarcoma xenografts in mice, as is twice weekly injection of an anti-VEGF monoclonal antibody ŽBlake Tomkinson, personal communication.. Also, in rats with mesangioproliferative nephritis, use of NX 1838 showed that VEGF165 mediates endothelial proliferation in damaged glomerular capillaries ŽOstendorf et al., 1999.. No effect of NX 1838 was seen on normal glomeruli. This, in fact, suggests that endothelial cells in
10
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
glomeruli, like endothelial cells in tumors, are VEGF-dependent during growth and remodeling, but lose this VEGF dependence once associated with pericytes Žor podocytes. ŽBenjamin et al., 1999; Darland and D’Amore, 1999.. Finally, NX 1838 was shown to be non-inflammatory and non-toxic when injected into the vitreous humor of rabbit eyes, and to be therapeutically active Žafter IV injection. in combating O 2-induced retinopathy in newborn mice ŽHenninger and Bill, personal communication.. Such data, among others, led to an Investigational New Drug acceptance for NX 1838 for human age-related macular degeneration; phase 1 trials for direct intravitreal injection of NX 1838 are currently underway. Perhaps the most thorough in vivo study of an aptamer was recently completed by David Parma and his collaborators ŽWatson et al., 1999.. A 29F pyrimidine-containing aptamer evolved against human L-selectin was studied in severe immunodeficient ŽSCID. mice. The aptamer binds to Lselectin on human lymphocytes and prevents leucocyte rolling, a prerequisite for attachment to capillary walls and extravasation to peripheral lymph nodes. A complete pharmacokinetic analysis of the 40-kDa PEG-modified aptamer showed that plasma clearance was the most important determinant of physiological activity. Clearance, presumably through renal excretion, is inversely related to the size of the aptamer conjugate. This was true for the L-selectin aptamer as the molecular weight of the conjugate increased to 50 kDa, beyond which there was no further decrease in plasma clearance. Over a 30-fold range of in vitro K d values, in vivo inhibition was not affected by the K d . Another result of this study is that nonspecific binding to leucocytes is not a problem at doses of aptamer which show a 50% reduction of leucocyte trafficking. Thus, for this cell surface target, efficient inhibition by the aptamer could be measured even when human leucocytes were injected 10 h after the IV bolus of aptamer had been injected. Because the in vivo activity of aptamers depends on their being conjugated to a higher molecular weight carrier, much research remains to be done on in vivo efficacy. Up to now, only
liposomes and PEG have been tested as carriers, both successfully. It is possible that conjugation of aptamers to other carriers, including human proteins that naturally circulate for weeks in the blood, will substantially decrease the clearance rate from plasma. Furthermore, liposomal presentation of aptamers has been carried out with only one liposomal formulation. Here too, further progress can be expected. The successful use of aptamers as therapeutic agents depends not only on their efficacy and specificity, but also on economics } that is, what are their costs of manufacture? The phosphoramidites for 29F pyrimidine and 29OCH3 purine monomers are readily available as is the 40 kDa molecular weight PEG used for conjugation. A number of companies now have Good Manufacturing Practice ŽGMP. facilities for the large scale synthesis of oligonucleotides, including RNA analogs. If we extrapolate from the most successful animal experiments in which target concentration is not limiting for efficacy, an aptamer dose of 1]2 mgrkg body wt. is usually an effective dose Žthis is derived from experiments on aptamers inhibiting VEGF, PDGF, L-selectin, and P-selectin .. For a 70-kg adult, this means that each injected dose would be 70]140 mg. Current quotes for aptamers approximately 35 nucleotides long containing modified ribonucleotides are approximately $2000rg Žwhen orders are for, minimally, 100 g.. This means that the manufacturing cost is $140]$280 per dose. For acute indications, such as organ transplant, myocardial infarcts, toxic or septic shock, angioplasty, or pulmonary embolism treatment every 3 days for 15 days would involve $700]$1400 in cost of goods. This is not out of line with many treatments, such as therapeutic antibodies, already in clinical use. Clearly, for chronic indications, the cost of the goods is an issue. There are a number of possible ways to reduce these costs. Firstly, a more efficient chemistry for aptamer synthesis, Product Anchored Sequential Synthesis ŽPASS., has been
1
Pieken and Gold Ž1999. Method for Solution Phase Synthesis of Oligonucleotides and Peptides. US Patent No. 5 874 532.
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
described.1 The PASS technology, although not yet optimized, could reduce the cost of aptamer synthesis by as much as a factor of 10 Žto approx. $200rg.. Also, as mentioned earlier, the studies on the pharmacokinetics of aptamers have only just begun. Any increase in circulatory half-life should reduce the cost of goods. This could come from new conjugates, new formulation of liposomes, or new delivery vehicles to slow down release from depots. Finally, it is worthwhile focusing on the relatively poor therapeutic results in cancer with standard chemotherapy. Although some cancers Že.g. leukemia and lymphomas. are curable by chemotherapy, many other frequent cancer types are almost never cured by standard cytotoxic reagents Žlung, brain, pancreas, colon, and breast cancer } once it becomes estrogen-independent.. The average life expectancy of a patient diagnosed with pancreatic cancer is only 8 months; if aptamer-based therapy against specific targets extended life to 5 or more years, even the present cost of goods would be found acceptable.
3. Diagnostics If one pictures aptamers as equivalent to antibodies, keeping in mind that their in vitro mode of production has the advantages discussed earlier, then it becomes obvious that aptamers can be ideal tools for the diagnostic evaluation of proteins. The standard diagnostic protocol for detecting proteins in blood, urine, saliva, and other body fluids, is the two-antibody sandwich assay. In this assay, a capture monoclonal antibody is used Žusually on a solid surface . to bind the protein in question. The signal-to-noise ratio of such a single capture step in a mixture of proteins as complex as blood, is never sufficient to provide a useful test. Because of this, a second antibody, either monoclonal to a different epitope, or polyclonal aimed at a different epitope, is then used to detect the bound protein. The second antibody is then detected by fluorescence or a colorimetric assay. In order for such an antibody-based test to be useful in diagnostics, it is essential that these two binding steps be indepen-
11
dent so that the background is reduced by the multiplicative specifity of the two steps. It is feasible to use the in vitro-derived aptamers in such a way that one aptamer can substitute for two antibodies. We have detailed the various steps involved in a recent review ŽBrody et al., 1999.. Here we shall outline the essential argument. Diagnostics is on the brink of a revolutionary change, which will be the consequence of sequencing the human genome during the next 2]3 years. It is expected that novel reagents Žwhich allow detection of pertinent human proteins in body fluids. will usher in this dramatic change in diagnostic capacity. There are three elements in the use of aptamers that should make them the reagents of choice to keep pace with the availability of new proteins as they are produced subsequently to sequencing the human genome. The first is the rapidity with which aptamers can be produced. This is a consequence of the construction of robots to perform the SELEX process ŽCox et al., 1998. ŽZichi et al., personal communication.. Since SELEX is an in vitro iterative protocol involving only enzymatic steps and separation of bound from unbound nucleic acids, it has been amenable to automation. It is now possible to adapt SELEX to a robotic procedure which can do approximately three rounds on 96 samples a day. Since the average successful SELEX procedure takes less than 10 rounds to complete Žespecially in the robotics format., one robotics machine can identify aptamers aimed at 3 = 96 = 365r10s; 10 000 protein reagents per year. The second element is the introduction of photo-SELEX ŽWillis et al., 1994; Jensen et al., 1995; Meisenheimer et al., 1996; Golden et al., 1999.. Photo-SELEX is a derivative of SELEX in which substrates for either DNA or RNA polymerase include a 5-iodo or 5-bromo-substituted dUTP or UTP ŽWillis et al., 1994; Meisenheimer et al., 1996.. A series of publications show that such substitutions allow for normal evolution of tight binding aptamers, but also allows that photoexcitation by long wavelength UV light leads to cross-linking of the aptamer to its targeted protein. Such SELEX experiments, in which the protocol demands evolution of aptamers which can be specifically photo-cross-linked to their cognate
12
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13
proteins, have been successfully carried out ŽBrody et al., 1999; Golden et al., 1999.. The importance of this for diagnostics is in the independence of the cross-linkability parameter from the binding parameter. This independence is great enough to give multiplicative background reduction analogous to the independent binding of two different antibodies to the same protein in an ELISA format. This leads to the following diagnostic scenario. With arrays of aptamers on a solid surface, blood or any other body fluid can be incubated with this aptamer ‘chip’. Proteins will sort themselves out in a manner analogous to affinity chromatography ŽRomig et al., 1999.. As long as the dissociation rate constant for binding of each aptamer to its cognate protein is sufficiently low, a gentle wash step should increase the signal-to-noise ratio at each specific aptamer cluster. Then, laser-induced photo-cross-linking will introduce covalent bonds only with those aptamer]protein pairs whose geometry has been evolved to position a } Br or } I sufficiently close to an electron-rich amino acid. The geometric, thus chemical, requirements for this happening are very constrained in this protocol ŽWillis et al., 1994; Meisenheimer et al., 1996; Ruckman et al., 1998; Golden and Willis, personal communication.. Other bound proteins, specifically non-cognate proteins, will not be covalently cross-linked. Next, a harsh detergent wash should keep only the cross-linked proteins on the aptamer ‘chip’ ŽBrody et al., 1999. ŽFig. 1.. The third element of this diagnostic protocol results directly from the use of aptamers and photo-SELEX. The detection step with antibodies involves necessarily indirect measurements; this is because antibodies are proteins, like the analytes they are supposed to be measuring in blood. Such indirect measurements add expense and a further source of error into antibody-based detection systems. When one uses photo-cross-linkable aptamers, the only proteins on the ‘chips’ after the wash cycles are the analytes themselves. Thus, any stain which distinguishes proteins from nucleic acids Žany universal protein stain. can be used directly to measure protein levels in a body fluid. Very simple compounds are available. For example, any lysine-specific reagent could func-
tion well as such a stain ŽBrody et al., 1999. ŽFig. 1.. No nucleophile in nucleic acids is as strong as lysine. Other amino acid-specific reagents have been under investigation for more than 50 years ŽWong, 1991.. Aptamers have been investigated since 1990 ŽEllington and Szostak, 1990; Tuerk and Gold, 1990.. The SELEX protocol is simple and automatable. When one considers that both liposomes and monoclonal antibodies were discovered approximately 20 years before the first therapeutic applications were approved for clinical use, it seems that the medical applications of aptamers are on a very fast historical track. Acknowledgements The authors would like to thank all of their colleagues at NeXstar Pharmaceuticals, now a wholly owned subsidiary of Gilead Sciences, Foster City, CA, USA. Nebojsa Janjic and his colleagues contributed greatly to the therapeutic aptamer studies; David Parma and co-workers were instrumental in the L-selectin pharmacokinetic and efficacy studies; Dwight Henninger, Jerry Bill, Blake Tomkinson and Wendy Gillette, in Ray Bendele’s group allowed citation of some of their unpublished work. The Diagnostic and Robotic effort was directed by Dom Zichi, Sumedha Jayasena, Mike Willis, and Drew Smith. Our thanks also, to Kris Cordova for preparation of the manuscript. References Bsasilico, C., Moscatelli, D., 1992. The FGF family of growth factors and oncogenes. Adv. Cancer Res. 59, 115]165. Benjamin, L.E., Golijanin, D., Itin, A., Pode, D., Keshet, E., 1999. Selective ablation of immature blood vessels in established human tumors follows vascular endothelial growth factor withdrawal Žsee comments.. J. Clin. Invest. 103, 159]165. Brody, E.N., Willis, M.C., Smith, J.D., Jayasena, S., Zichi, D., Gold, L., 1999. The Use of Aptamers in Large Arrays for Molecular Diagnostics. Mol. Diagn., 4, 381]388. Cox, J.C., Rudolph, P., Ellington, A.D., 1998. Automated RNA selection. Biotechnol. Prog. 14, 845]850. Darland, D.C., D’Amore, P.A., 1999. Blood vessel maturation: vascular development comes of age comment. J. Clin. Invest. 103, 157]158.
E.N. Brody, L. Gold r Re¨ iews in Molecular Biotechnology 74 (2000) 5]13 Eaton, B.E., Pieken, W.A., 1995. Ribonucleosides and RNA. Annu. Rev. Biochem. 64, 837]863. Ellington, A.D., Szostak, J.W., 1990. In vitro selection of RNA molecules that bind specific ligands. Nature 346, 818]822. Floege, J., Ostendorf, T., Janssen, U. et al., 1999. Novel approach to specific growth factor inhibition in vivo: antagonism of platelet-derived growth factor in glomerulonephritis by aptamers. Am. J. Pathol. 154, 169]179. Gold, L., 1995. Oligonucleotides as research, diagnostic, and therapeutic agents. J. Biol. Chem. 270, 13 581]13 584. Gold, L., Polisky, B., Uhlenbeck, O., Yarus, M., 1995. Diversity of oligonucleotide functions. Annu. Rev. Biochem. 64, 763]797. Golden, M., Collins, B.D., Willis, M.C., Koch, T.H., 1999. Diagnostic potential of photo SELEX-evolved ss aptamers. J. Biotechnol., submitted for publication. Griffiths, A.D., Williams, S.C., Hartley, O. et al., 1994. Isolation of high affinity human antibodies directly from large synthetic repertoires. EMBO J. 13, 3245]3260. Jellinek, D., Green, L.S., Bell, C. et al., 1995. Potent 29-amino29-deoxypyrimidine RNA inhibitors of basic fibroblast growth factor. Biochemistry 34, 11 363]11 372. Jenison, R.D., Gill, S.C., Pardi, A., Polisky, B., 1994. High-resolution molecular discrimination by RNA. Science 263, 5152. Jenison, R.D., Jennings, S.D., Walker, D.W., Bargatze, R.F., Parma, D., 1998. Oligonucleotide inhibitors of P-selectindependent neutrophil]platelet adhesion. Antisense Nucleic Acid Drug Dev. 8, 265]279. Jensen, K.B., Atkinson, B.L., Willis, M.C., Koch, T.H., Gold, L., 1995. Using in vitro selection to direct the covalent attachment of human immunodeficiency virus type 1 Rev protein to high-affinity RNA ligands. Proc. Natl. Acad. Sci. USA 92, 12 220]12 224. Leppanen, O., Janjic, N., Carlsson, M.A. et al., 1999. PDGF B-chain Aptamer Reduces Pseudointimal Hyperplasia in a Rat Model of Arterial Restenosis, submitted for publication . Meisenheimer, K.M., Meisenheimer, P.L., Willis, M.C., Koch, T.H., 1996. High yield photocrosslinking of a 5-iodocytidine ŽIC. substituted RNA to its associated protein. Nucleic Acids Res. 24, 981]982. Miyamoto, M., Naruo, K., Seko, C., Matsumoto, S., Kondo, T., Kurokawa, T., 1993. Molecular cloning of a novel cytokine cDNA encoding the ninth member of the fibroblast growth
13
factor family, which has a unique secretion property. Mol. Cell Biol. 13, 4251]4259. Osborne, S.E., Ellington, A.D., 1997. Nucleic acid selection and the challenge of combinatorial chemistry. Chem. Rev. 97, 349]370. Ostendorf, T., Kunter, U., Eitner, F. et al., 1999. VEGF165 mediates glomerular endothelial repair. J. Clin. Invest., in press . Ringquist, S., Parma, D., 1998. Anti-L-selectin oligonucleotide ligands recognize CD62L-positive leukocytes: binding affinity and specificity of univalent and bivalent ligands. Cytometry 33, 394]405. Romig, T., Bell, C., Drolet, D., 1999. Aptamer affinity chromatography; combinatorial chemistry applied to protein purification. J. Chromatogr., in press . Ruckman, J., Green, L.S., Beeson, J. et al., 1998. 29-Fluoropyrimidine RNA-based aptamers to the 165-amino acid form of vascular endothelial growth factor ŽVEGF165.. Inhibition of receptor binding and VEGF-induced vascular permeability through interactions requiring the exon 7-encoded domain. J. Biol. Chem. 273, 20 556]20 567. Tanaka, A., Miyamoto, K., Minamino, N. et al., 1992. Cloning and characterization of an androgen-induced growth factor essential for the androgen-dependent growth of mouse mammary carcinoma cells. Proc. Natl. Acad. Sci. USA 89, 8928]8932. Tuerk, C., Gold, L., 1990. Systematic Evolution of Ligands by Exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase. Science 249, 505]510. Watson, S.R., Chang, Y.F., O’Connell, D., Weigand, L., Parma, D., 1999. Anti-L-selectin aptamers: binding characteristics, pharmacokinetic parameters and activity against an intravascular target in vivo. Antisense Nucl. Acid Drug Dev., in press . Willis, M.C., Collins, B.D., Zhang, T. et al., 1998. Liposomeanchored vascular endothelial growth factor aptamers. Published erratum appears in Bioconjug Chem 1998 Sep]Oct; 9Ž5.:633. Bioconjug. Chem. 9, 573]582. Willis, M.C., LeCuyer, K.A., Meisenheimer, K.M., Uhlenbeck, O.C., Koch, T.H., 1994. An RNA-protein contact determined by 5-bromouridine substitution, photocrosslinking and sequencing. Nucleic Acids Res. 22, 4947]4952. Wong, S., 1991. Chemistry of Protein Crosslinking . CRC Press.