Nucleic acid aptamers

Introduction

Interactions between DNA and RNA and other nucleic acids are exploited by Nature, for example in transcription and DNA replication, and in biotechnology in PCR. The ability of DNA and RNA oligonucleotides to make strong interactions with other types of molecules, including metal ions, small molecules and large macromolecules such as proteins and viruses, has recently been studied: such oligonucleotides are called aptamers. The exceptional ability of aptamers to bind to these non-nucleic acid targets is due to the secondary structures (e.g. hairpins, loops, quadruplexes) that oligonucleotides can adopt, rather than Watson-Crick hydrogen bonding or base stacking.

Comparison with antibodies

Antibodies are large proteins that are used widely in diagnostic tests and in therapeutics. The ability of aptamers to replicate the behaviour of antibodies has led to excitement about the potential of use of aptamers in diagnostic and therapeutic applications. Aptamers offer many significant advantages over antibodies:

• Ease of synthesis: Aptamers can be synthesized using solid phase oligonucleotide synthesis, whereas antibodies have to be obtained using less efficient biochemical or biological methods. This is particularly important if the synthesis is to be scaled up.
• Ease of modification: Because aptamers are made by chemical synthesis, it is easy to extend the synthesis to incorporate modifications.
• Library synthesis: Because of the way solid phase oligonucleotide synthesis works, it is trivial to generate large libraries of similar DNA and RNA aptamers, which are necessary for selection experiments.
• Low immunogenicity: Antibodies often induce an immune response when used as therapeutic agents in the body (they are immunogenic), while DNA and RNA aptamers can be designed to be non-immunogenic
• Thermal stability: Single-stranded oligonucleotides are stable at high temperatures whereas antibodies denature irreversibly at high temperatures.
• Uptake: DNA and RNA aptamers are typically smaller (lower molecular weight) than antibodies, which means that penetration into tissue is faster; however, the small size of aptamers makes them susceptible to renal filtration (this can be overcome by conjugation to peptides or polymers… but this increases the molecular weight).
• Shelf life, storage and distribution: oligonucleotides can be freeze-dried and stored for a long time, and easily transported, while many antibodies have limited shelf life.

Comparison with small molecule drugs

Small molecules still account for most drugs. Although the large-scale synthesis of small molecule drugs can be highly efficient, and combinatorial chemistry enables the synthesis of libraries of small molecules, it is not necessarily easy to design the synthesis of a new drug target from scratch. One advantage aptamers may have over small molecules is that, given an aptamer with moderate affinity for a target (e.g. as the result of a SELEX experiment) it is trivial to synthesize derivatives of that molecule with slightly different sequences, and to incorporate modifications.

Desirable properties of an aptamer

An ideal aptamer has

• high binding affinity for its target
• high specificity for its target (relative to other, similar targets)

Aptamers designed for therapeutic use in the body must also have

• high stability (in aqueous solution, and in the presence of other biological molecules including enzymes)
• high bioavailability (the aptamer must be able to reach its target, which may involve traversing biological membranes).

The dissociation constant

In the reaction between a ligand (aptamer) and receptor (target, e.g. a protein), the association constant, K_a, is the equilibrium constant for the forward reaction, the formation of the complex, while the dissociation constant, K_d, is the equilibrium constant for the reverse reaction, the dissociation of the complex.

L + R -> L:R (complex)

The smaller the dissociation constant, the stronger the complex. Aptamers are often defined by their dissociation constants - a micromolar (10⁻⁶ M) K_d indices a strong complex, a nanomolar K_d (10^-9 M) indicates a very strong complex, and a very good aptamer; a picomolar K_d (10⁻¹² M) is even better! Many low nanomolar and picomolar aptamers have been found for protein targets, which had led to considerable interest in aptamers as diagnostic and therapeutic agents.

Aptamer selection (SELEX)

Much of the success of nucleic acid aptamers is due to SELEX (systematic evolution of ligands by exponential enrichment; Figure 1), an elegant process by which aptamers can be generated for a given target (e.g protein).

In SELEX, a single-stranded DNA library is first generated and exposed to the target. Any olignucleotides within the library that bind to the target are retained, while the non-binding sequences are washed away. Typically, the sequences in the library that bind to the target will be a small fraction of the total library so, after this first step, most of the library will have been washed away, and the overall concentration of DNA will been significantly reduced. The remaining (bound) sequences are then eluted and retained.

In the next step, the remaining (binding) sequences are amplified by PCR. This increases the concentration of DNA in preparation for the next round of SELEX. The library is now biased, or enriched, towards sequences that bind to the target.

The process is repeated multiple times. Often, the binding conditions are made more stringent as SELEX proceeds through each round, increasing the selection pressure. Successive selection and amplification result in a library that, at the end of the SELEX process (typically 8-15 rounds), contains only sequences that bind strongly to the target.

Figure 1

SELEX (Systematic evolution of ligands by exponential enrichment) A single-stranded DNA library is exposed to the target. Non-binding oligonucleotide sequence are washed away; oligonucleotides that bind to the target are retained and amplified by PCR, resulting in an enriched library. Multiple cycles are performed, each starting with the enriched library from the previous step. Eventually, only sequences that bind strongly to the target remain.

After SELEX is complete, the binding sequences need to be identified, which is done by sequencing (link).

SELEX for RNA aptamers

PCR produces DNA, not RNA. So, to use SELEX with RNA, additional reverse transcription (to convert DNA to RNA at the beginning of the SELEX cycle) and transcription (to convert DNA to RNA at the end of the SELEX cycle) and are necessary. It is worth noting that the initial library can be either DNA or RNA.

Improvements to conventional SELEX

While SELEX has proved to be a successful method for generating aptamers in the 25 years since its invention, there is room for improvement. Many of the deficiencies of SELEX stem from its dependence on PCR.

Overcoming PCR bias

While a library may contain equal amounts of different sequences, each of these sequences may not be amplified to the same extent during PCR, and this can result in an unequal distribution of products. This effect, called PCR bias, is exaggerated over multiple rounds of amplification. It is sometimes possible to overcome this effect by carefully optimizing PCR conditions (in particular annealing temperature).

Generating single-stranded DNA

Methods of producing single-stranded DNA using PCR include asymmetric PCR, where one primer is used in excess, which after amplification gives a large excess of the desired strand, which can be separated by electrophoresis. Alternatively, lambda exonuclease is an enzyme that will digest the complementary, unwanted strand after PCR, provided it has a 5’-phosphate label (Figure 2). This can be achieved by using a 5’-phosphate primer.

Figure 2

Generating single-stranded DNA using PCR by digestion

Selectivity

Conventional SELEX will find aptamers that bind strongly to a target, but won't necessarily find aptamers that are highly specific for that target (relative to other, similar targets). The addition of a step that selects against binding to a second target ("counter-SELEX") can help to ensure that any aptamers found are specific to the target of interest.

SELEX variations

The original SELEX publications used nitrocellulose filters to separate unbound sequences from target-bound sequence complexes. This low to medium-throughput binding assay allows the unbound sequences simply pass through the membrane and then the bound sequences can be easily eluted. On the other hand, the efficiency of protein capture depends and the protein and/or the experimental conditions. Aptamers may also bind to the nitrocellulose membrane itself, which could enrich the library with non-binding sequences. Without effective negative selection, these matrix-binding aptamers may prohibit the elucidation of an aptamer for its intended target.

Over the past 25 years, research has focused on developing new SELEX protocols for selecting aptamer candidates. For example, immobilising the target on magnetic beads can be useful as the non-binding sequences can be easily removed from the target and the selection stringency can be finely tuned. In addition, this technique can also be applied to a wide range of targets and no specialized equipment is required. Magnetic beads have been used to isolate aptamers with nanomolar Kd for streptavidin and Botulinum neurotoxin type A. A drawback of the bead-based method is that immobilisation of the target may restrict the surface area that the target can bind to. Furthermore, the density of the target on the support can distort the outcome of the experiment. Ozer (http://nar.oxfordjournals.org/content/41/14/7167.long) showed that non-specific RNA aptamers can concurrently bind up to four H3-C peptide target molecules, increasing the binding affinity by up to two orders of magnitude.

A more recent method for aptamer generation involves microarray technology. A DNA microarray has a surface that contains thousands of single-stranded DNA sequences. In this case, the oligonucleotide is immobilised and the target is incubated with the chip. At present, this method is limited to ssDNA aptamers (due to the scarcity of RNA microarray synthesisers) and the maximum library size is less than 10,000. A new microarray has to be synthesised for each selection round.

In Cell-SELEX, aptamers are incubated to intact cells or cellular fragments. This is primarily used for biomarker discovery and to examine the therapeutic potential of the selected aptamers. For example, it may be used to differentiate between normal cells and tumour cells in their native state. Cell surfaces have many different possible targets, it is important to have a robust negative selection method, since it is likely that an aptamer will be selected for an unintended target.

Common modifications in aptamers

Modifications are incorporated into nucleic acid aptamers for several reasons:

• to increase chemical diversity because the four bases of DNA/RNA are similar in structure, all containing a phosphate backbone, sugar, and heteocyclic base.
• to convey stability to enzymatic (endonuclease) degradation - if an aptamer is to be used as a therapeutic agent, it must be protected against degradation by endonucleases, which are present in cells and digest unmodified DNA and RNA
• to improve cellular uptake
• to increase flexibility/rigidity
• to provide a means of measurement (e.g. incorporation of a fluorophore).

Post-SELEX modifications

Like other technologies that depend on PCR (e.g. DNA sequencing) SELEX cannot generally incorporate modifications other than standard A, G, C and T/U nucleosides. One way to get around this is not to include modifications in the SELEX process. Once an (unmodified) apatmer has been found through SELEX, modified versions of this aptamer are then synthesized.

Incorporating phophorothioate linkages at the 3'-end of DNA oligonucleotides is known to improve the stability of antisense oligonucleotides, and phophorothioate linkages are used in aptamers. LNA bases have been incorporated at the 3'-end to similar effect (Figure 3).

Figure 3

Common modifications Common post-SELEX modifications.

There is a significant chance, however, that introducing modifications will change the nature of the binding between the aptamers and its target, and affect the affinity or specificity of the aptamer. This has led scientists to look for ways of incorporating modifications that are at least partially compatible with SELEX.

Modifications in RNA aptamers

RNA is significantly less stable than DNA, to both chemical and enzymatic degradation (link); so RNA aptamers commonly incorporate 2'-F modifications (the 2'-OH group is replaced with a fluorine; figure X).

Figure 4

Modifications in RNA aptamers 2'-Fluoro and 2'-methoxy RNA, used in RNA aptamers to improve stability.

SOMAmers

SOMAmers, developed by SomaLogic, Inc., have taken an ingenious approach to generating incorporating modifications that are compatible with SELEX. Instead of using the standard deoxyribonucleoside triphosphates (dNTPs), i.e. dATP, dCTP, dGTP and dTTP, the dTTP is replaced by a dUTP modified at the 5-position in these aptamers. The effect of this substitution is that, during PCR, each T is then copied as (modified) T. The structures of some modified bases used in SOMAmer selection are shown in Figure 5.

Figure 5

SOMAmer bases Structures of modified dU bases used in SOMAmer selection.

The SOMAmer approach has a number of caveats. Firstly, the modified dUTP must be compatible with the DNA polymerase enzymes used in PCR. If the modified group is too large, for example, it might distort the structure of the dUTP so that it is not tolerated well by the polymerase enzymes, leading to a PCR bias against this modified base in the library. Secondly, only type of one modification can be incorporated, and every T based in the library will be modified.

Despite these concerns, aptamers that incorporate these modified bases have been found for thousands of targets.

Spiegelmers

Spielgelmers (from German speigel “mirror’) are RNA aptamers built from L-ribose units (i.e. a mirror image of nautral oligonucleotides, which are based on D-ribose sugar units) and therefore are the enantiomers of natural aptamers. Spiegelmers, unlike oligonucleotides are 100% biostable since they cannot be broken down by nucleases.

In the first step, a natural RNA aptamer is selected against the unnatural, enantiomeric form of the target by SELEX. After sequencing, the Spiegelmer is chemically synthesized and binding experiments (with the natural target) are performed. Since enantiomers display identical chemical properties, the binding of a spiegelmer to a natural target should be comparable to that of a natural aptamer to an unnatural target.

Figure 6

Spiegelmers

This approach depends on being able to obtain or produce the enantiomer of the target of interest: while possible for some small molecules, this can be impractical for biological macromolecules.

Polyethylene glycol (PEG) conjugates

The kidneys are very good at removing small molecules (including short oligonucleotides) from the body, which restricts the lifetime of small-molecule drugs in the blood, limiting their potency. One method of overcoming this renal filtration and extending the half-life of an aptamer in the body is to increasing the molecular weight by conjugation to polyethylene glycol (PEG). The effect of PEG conjugation on cellular uptake and binding must be studied if this approach is used.

Figure 7

Structure of an aptamer-PEG conjugate

Aptamer structure

How do aptamers show such high binding affinity AND specificity for such a wide range of targets? - to do with the wide range of three-dimensional structures oligonucleotides (aptamers) can adopt:

• hairpin (stem) loops - occur when two regions of the same strand are complementary to one another and can form Watson-Crick base pairs.
• quadruplexes - occur in guanine-rich sequences, when four guanine bases can associate through hydrogen bonding.
• kissing complexes - formed when the unpaired nucleotides in one hairpin loop base pair with the unpaired nucleotides in another hairpin loop. Usually occurs in RNA.

Figure 8

Aptamer structures

Diagnostic and therapeutic applications

Recently, aptamers have found appliations in both diagnostics and therapeutics, replacing antibodies in some cases.

Pegaptanib

Pegaptanib sodium (brand name Macugen, (Figure 9)), an RNA-based aptamer, was discovered in 2000 and was the first aptamer-based drug to be approved by the US FDA, in 2004. Its target is VEGF, a protein involved in the growth of blood vessels and making them more permeable. Pegaptanib binds to VEGF, reducing the growth of blood vessels in the eye and controls leakage and swelling, and is used in the treatment of neovascular (wet) age-related macular degeneration (AMD).

Figure 9

Structure of Pegaptanib (Macugen), the first nucleic acid aptamer to be approved by the US FDA for therapeutic use Pegaptanib employs a number of different modifications. Almost all of the bases are 2'-fluoro or 2'-methoxy RNA bases and the 5'-end is conjugated to a polyethylene glycol moiety.

The evolution of pegaptanib: from laboratory bench to clinic

As an FDA-approved drug, the first aptamer drug and first RNA drug, the history is pegaptanib is well documented: it is an fascinating case study of the real-world evolution of an aptamer drug. Initial, conventional SELEX experiments identified promising unmodified RNA aptamers with low nanomolar affinity for VEGF. These oligonucleotides were also found to be inhibitors of VEGF (it does not necessarily follow that, if a molecule binds to a protein target, it also inhibits that protein). No attempt was made to confer nuclease stability.

A second study, aimed at producing stable aptamers, included 2'-aminopyrimidine bases in the initial library for SELEX (i.e. standard dA, standard dG, 2'-amino dG and 2'-amino dT), resulting in an aptamer with low nanomolar affinity for VEGF. Following SELEX, 2'-methoxypurine bases were substituted systematically for each unmodified A and G base in the sequence. In the final aptamer, all but four of the purine bases were 2'-methoxy-substituted. Poly-dT regions were introduced at both the 3'- and 5'-ends, with phosphorothioate linkages in between these dT bases. The end result of this study was a molecule with much improved nuclease resistant and also improved affinity (indicating that modifications can improve affinity as well as stability).

A third study replaced the 2'-amino bases with 2'-fluoro bases, to give even higher (low picomolar) affinity, and improved bioavailability by conjugating the aptamer to a 40 kDa polyethylene glycol (PEG) molecule. This aptamer conjugate had significantly increased half-life in the blood, due to a reduction in renal clearance, and increased uptake in organs. Following clinical trials, pegaptanib was approved by the FDA, and many thousands of people have been treated with it. The caveat here is that pegaptanib has been surpassed by antibody drugs more effective for the treatment of AMD. Despite this, pegaptanib has opened the door for other aptamer-based pharmaceuticals.

Aptamers in cancer imaging

Cancer cells overexpress certain proteins, and these cancer cell surface proteins (biomarkers) are targets for aptamers. Highly specific aptamers have been developed for some of these biomarkers, and functionalization the the aptamer with a fluorophore makes imaging of the cancer cells possible, allowing surgeons to distinguish tumours from normal tissue, for example.

Synthesis of aptamer libraries

Aptamer libraries for SELEX are synthesized by conventional solid phase phosphoramidite oligonucleotide synthesis. As SELEX uses PCR, the libraries must contain fixed primer binding regions at the 5'- and 3'-ends. Between these primer binding sites is the variable region, which contains degeneracies (wobbles): mixtures of different DNA bases at a given position in the sequence. For example, the N wobble is an equimolar mixture of A, G, C and T, and is incorporated by mixing the A, G, C and T phosphoramidites in the appropriate ratio (taking into account molecular mass and relative reactivity, which can be determined experimentally; Figure 10).

Figure 10

Library formation for SELEX

For a randomly generated sequence of length n, there are 4ⁿ possible sequences (there are 4 possible bases at each position).