Emerging trends at the interface of Chemistry and Biology: Applications to the design of human therapeutics

SANTANU BHATTACHARYA1 and RAGHAVAN VARADARAJAN2

Department of Organic Chemistry, Indian Institute of Science, Bangalore 560 012, India. e-mail: [email protected] 2 Molecular Biophysics Unit, Indian Institute of Science, Bangalore 560 012, India. e-mail: [email protected]

This article describes recent developments in the design and implementation of various strategies towards the development of novel therapeutics using rst principles from biology and chemistry. Strategies for multi-target therapeutics and network analysis with a focus on cancer and HIV are discussed. Methods for gene and siRNA delivery are presented along with challenges and opportunities for siRNA therapeutics. Advances in protein design methodology and screening are described, with a focus on their application to the design of antibody based therapeutics. Future advances in this area relevant to vaccine design are also mentioned.

1. Introduction Over the past few decades, there have been signicant advances in the elds of chemistry, molecular, structural and cell biology, and genetics, which have led to an increased understanding of the various life processes in general. However, much remains to be explored. Rapid increases in human population, overcrowding, environmental degradation and rapid transport have all enhanced the spread of existing infectious diseases as well as the threat of newly emerging infections. At the same time increased mechanization, sedentary lifestyles, dietary changes and an increase in life expectancy have lead to increased rates of a variety of non-infectious diseases. Advances in diverse areas of chemistry and biology are being used to discover therapeutics for both infectious and non-infectious diseases. Novel strategies have been discovered to design and make new active principles. This article discusses a few of them, which appear to be promising. The article is divided into two parts; the rst part describes progress in the area of small molecule therapeutics, gene delivery and siRNA technology. The next part

describes recent progress in the area of protein therapeutics. Organic chemistry is an enabling principle for the development of molecules that can selectively perturb or probe various biological systems. There is a growing understanding of the identication of molecules or systems that engage in distinct interaction with one or several components of complex and dynamic biological milieu. Herein we present instances where chemical biology principles have been extended for dierent therapeutic leads. 2. Multi-target therapeutics and network analysis The modern drug discovery has shifted its focus from the concept of one gene, one drug, and one disease, i.e. single-target strategy to multitarget strategy, which is based on network biology (gure 1) [1,2]. Former suers from lack of ecacy, toxicity, drug resistance and huge investment. Network biology is a tool to overcome this, which emphasizes on tackling the elements that disturb the diseased network rather than targeting the specic gene causing the disease.

Keywords. Multi-target therapeutics; network biology; gene delivery; protein therapeutics; protein design.

10

SANTANU BHATTACHARYA AND RAGHAVAN VARADARAJAN

Figure 1.

Drug development strategies.

This is also used for managing pain caused by nerve damage in patients suering from diabetic neuropathy. Now to make polypharmacological drugs, one has to identify a lead compound which has activity against multiple targets without becoming nonselective and then this should be transformed into a drug with a good pharmaceutical prole. More challenging is to optimize the potency at both the targets. To predict a multi-target compound requires high level computational methods to explore a vast number of possible drug and target combinations which are compatible with each other. 4. Multi-target therapeutics and network analysis for cancer treatment Most of the current anticancer drugs have low therapeutic index because their action is based on inhibiting the essential functions that are present in both normal and cancerous cells. Their eectiveness is because of rapidly dividing nature of the cancer cells compared to the normal cells. It would be ideal if anticancer agents target proteins or interactions that are essential in cancer cells but non-essential in normal cells. The genetic changes in a cancer cell change the relative dependence on specic proteins relative to normal cells. Thus the new methodology takes advantage of this difference in topology of the biological network to produce drugs with better therapeutic index. 4.1 Drug design based on network analysis Eect of a drug is studied in the context of relevant interactions (link) present in protein, regulatory, metabolic or signaling networks where the respective target is called element. Analysis based on genomic, proteomic, metabolism, etc. are the methods to identify the systemic analysis of multitarget drug action. A simple topological model is created by suppressing an element or a link and then its eect on the network is studied. It is found that multiple but partial attacks on various elements or links is more ecient than complete attack on a single element or link because the former aect more number of network links [4]. Thus, multi-target drugs are better than single target drugs because they aect more sites of network. The eect is more if they are distributed over the entire network. 4.2 Mutations in phosphatidylinositol 3-kinase a cause for cancer Phosphatidylinositol 3-kinases/Phosphoinositide 3-kinases (PI 3-kinases or PI3Ks) are enzymes

3. Strategies of multi-target therapeutics (A) Multiple drugs are given in combination, as in antiretroviral therapy for AIDS. But patient compliance and adverse drug-drug interactions are the main drawbacks of this method. (B) Multicomponent therapeutics where two or more active ingredients are eectively combined into a single delivery system and administered as a drug. It is easy to nd the optimal formulation with dierent agents that achieve maximum potency. If required, the target can be subjected to dierential exposure of the individual agents by a programmed release of the constituting agents. However, dierences in pharmacokinetics, bioavailability and metabolism of the individual ingredients cause many problems. Examples include, drugs e.g. Atripla (HIV drug), Advair (a drug for chronic obstructive pulmonary diseases), Caduet (used for treatment of high blood pressure and high cholesterol), etc. For either of the above two-drug combination strategies, not only should the individual drug be safe but should be so also in combination. This limits the availability of safe drug combinations. (C) Individual multi-targeted single agents, which use the principle of polypharmacology, where a single compound specically binds to two or more molecular targets [3]. Such a drug circumvents the problems regarding equitable pharmacokinetics and biodistribution. The formulation of an individual active agent is easier compared with that of a mixture. Examples of multi target drugs are: Symbalta (cymbalta), Sutent (an oral multikinase inhibitor, an anti-cancer drug for multiple cancer), Nexavar (an oral multi-kinase inhibitor, an anti-cancer drug for liver and kidney), etc. Cymbalta is an antidepressant drug belonging to serotonin and norepinephrin reuptake inhibitors, which works by restoring their balance in the brain. It is used to treat depression and anxiety disorder.

NOVEL THERAPEUTIC STRATEGIES

11

which can phosphorylate the position-3 hydroxyl group of inositol ring of phosphatidylinositol. Discovery of cancer specic mutations in the catalytic subunit p110 of class I PI3K has made it a promising cancer target [5]. The p110 mutations show a preference to three mutational hot spots within the coding region of PIK3CA. Many of the mutations in p110 due to cancer are missense mutations where an amino acid is substituted with the change in a single nucleotide. The p110 hot spot mutants can grow independently, resistant to apoptosis and can spread the tumour elsewhere [6]. Many other rare mutations accompany p110 mutations with some gain of function like activating growth factor-independent phosphorylation of Akt and p70s6K and enhancing the lipid kinase activity. Nonetheless rare types of mutants are less ecient than the hot-spot mutants with respect to initiation or replication of the cancerous cells. Based on molecular modeling, function of mutations present in various functional domains is proposed. Homology model proposes that the gain of function in p110 might involve cellular localization, interaction with other proteins and modication of the catalytic activity. 4.3 Mutants in p110 as cancer targets The drug should be isoform specic because isoform is the main regulator of the cell growth [7]. Small molecule inhibitors are good candidates because they can be cancer specic and PI3K being an enzyme can be targeted easily by them. The presence of mutational hot spots can be taken into advantage to achieve mutant specicity. It is found that substitution of H1047R creates most robust cancerous p110 with highest frequency. Now a compound that can dierentiate between the mutant and wild-type protein, thus binding with two adjacent sites on the target protein, can be a promising candidate. A deeper understanding of the structure of wildtype and mutant p110 would greatly enhance the identication of mutant-specic small molecule inhibitors and throw light on how mutant-mediated gain of function takes place. 5. Challenges in the development of HIV medicine Most dreaded disease of the modern day world is AIDS (Acquired Immuno Deciency Syndrome), which is caused by HIV-1 virus. A vaccine for HIV-1 seems to be still elusive for the scientists because the virus is highly diverse in nature, and escapes the adaptive immune responses, dicult to

induce broad reactive antibody responses and its viral reservoirs are set up very early [8]. Though there are drugs which partially prevent the disease by controlling the viral replication to a certain extent, there is no vaccine which completely blocks the infection and provides full immunity. Life cycle of HIV. The viral RNA and other contents enter into the cytoplasm of the cell after the virus attaches itself to specic receptors in lymphocyte membrane. The enzyme reverse transcriptase transcribes the viral RNA sequence into a complementary DNA sequence and the enzyme integrase coalesces with host cell DNA. Replication of viral RNA is triggered by activation of lymphocyte which is followed by translation to produce various structural proteins like reverse transcriptase, protease, integrase, etc. Proliferation occurs when infected proteins bind to healthy immune cells [9].

6. Viral entry inhibitors A. Enfuvirtide (DP-178, T-20), a peptide inhibits HIV fusion by preventing the fusion of a glycoprotein (gp41) of the viral envelope with the cell membrane. This is taken along with other antiretrovirals but its low genetic barrier to resistance undermines its usefulness. B. Small molecule antagonists like mozobil, maraviroc, vicriviroc and aplaviroc hinder the viral entry by targeting human chemokine receptors CCR5 and CXCR4. 6.1 HIV protease inhibitors HIV protease belongs to an aspartic protease family which plays an indispensable role in the life cycle of HIV. Protease inhibitors block protease activity by mimicking the structure of its regular substrate and subsequent competitive binding with the virus. Some examples are atazanavir, tipranavir and darunavir, etc. 6.2 HIV reverse transcriptase inhibitors These are of two types nucleoside reverse transcriptase inhibitors (NRTIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs). NRTIs are rst converted into their active metabolites and then they act as competitive inhibitors or alternative substrates with respect to the normal substrates (dATP, dGTP, dCTP or dTTP) and lead to the termination of chain elongation. Some examples are zidovudine, didanosine,

12

SANTANU BHATTACHARYA AND RAGHAVAN VARADARAJAN

zalcitabine, etc. Nucleotide reverse transcriptase inhibitors (NtRTIs) also have similar mode of action but they dier in the number of steps required for their activation. 6.3 HIV non-nucleoside reverse transcriptase inhibitors

transduction occurs. The size of a lipid raft can be anywhere between 10 nm to 1000 nm and they always restructure themselves by continuous forming and breaking. So it is quite dicult to study their evolution. 8. Gene delivery

They block the enzymes activity by interacting with an allosteric binding site on HIV-1 reverse transcriptase. Some examples of this category include nevirapine, delavirdine, etc. 7. Chemical interface in lipid rafts Stier and Sackmann postulated the concept of membrane microdomains in 1970 [10] which was later supported and further studied in detail by various other groups. Gerrit van Meer coined the term lipid rafts for these microdomains which are enriched with lipids, cholesterol, glycolipids and sphingolipids present in cell membranes [11]. Although this concept was invoked to explain the transport of cholesterol from the trans golgi network to plasma membrane, recent research [12] has shown that these lipid rafts have roles in diverse processes like signal transduction, endocytosis apart from cholesterol tracking. Composition of lipid rafts can be a well-ordered, cholesterol- and glycosphingolipid-enriched domain or a diverse collection of lipids, showing variable enrichment in cholesterol and sphingolipids depending on the method of preparation. Lipid rafts are related to the immiscibility of ordered and disordered liquid microdomains in cells which cause decrease in free energy between two phases. Membrane proteins are categorized based on their location whether they are present in rafts or in liquid-disordered phase or in the intermediate state. Thus lipid rafts constitute proteins and lipids that not only oat freely within the liquiddisordered bilayer of cellular membranes but can also aggregate to form clusters. Removal of cholesterol which partitions between raft and non-raft phase makes the proteins non-functional. Lipid rafts are also implicated in many viral and bacterial infections [13]. For example, inuenza virus and HIV-1 virus enter into the cell by binding at the rafts followed by altering the cell functions. In diseases like HIV and leishmaniasis, pathogens interfere with the function of rafts of the host cell and use them to gain entry into the cell interior [14]. Apart from this the size of lipid raft is very important in raft-borne signaling proteins. They are in o state when rafts are small and turned on when rafts aggregate to form clusters where functionally related proteins interact and signal

According to Miller, gene therapy is dened either as the use of genes as medicines to treat disease or the delivery of nucleic acid (with a vector) to patients for some therapeutic purpose [15]. In other words, it involves the delivery of the DNA to the targeted cells, thus curing the disease at its causative level. For this to be successful, an ecient delivery system should be available which transports the gene expression system into the cell that encodes a therapeutic protein that is either not produced in the patient or produced in a nonfunctional form. The fact that naked DNA is not suitable for the in vivo transport of genetic materials into selected cell body types due to its degradation by serum nucleases, search for ecient carrier system has intensied over the past few years. Vectors can be divided into two classes: viral or non-viral. Some of the viral vectors are adenovirus, retrovirus, adenoassociated virus, etc. However immunogenicity, toxicity, insertional mutagenesis and vector inactivation, etc. have led the scientists to develop nonviral vectors. Among non-viral vectors, DEAE-dextran mediated transfection and calcium-phosphate methods are classical and popular methods. Now the research is focused on the design and synthesis of, cationic lipids [16], cationic polymers [17] and peptides which are highly ecient gene transfection vectors. Cationic liposomes or polymers bind to negatively charged DNA due to electrostatic interactions and form a complex. This binds to the cell surface and endocytosis takes over. But after entering lysosome, there is no clear understanding as to how the release of DNA takes place before the occurrence of lysosomal degradation although the addition of buering agents can reduce this to some extent. Cationic lipids are converted into liposomes which contain inner aqueous pools capable of encapsulating drugs or DNA [16]. Lipoplexes thus formed gain entry into the cell by slow endocytosis. The nucleic acids which escape from endosomal compartments enter into the cell cytoplasm. Nucleic acids may face further diculty in entering the nucleus. Once it enters the nucleus, expression of delivered genes occurs. Typically a cationic lipid consists of a hydrophobic domain (usually two long aliphatic chains or

NOVEL THERAPEUTIC STRATEGIES

13

cholesterol-type group), a cationic head group and a linker which connects these two parts. Many of the cationic lipids have glycerol backbone on which the hydrophobic tails are present [18]. Even the head group is attached through a hydrophilic domain like polyethylene glycol to impart special properties to the lipid. The systematic modication of each part and understanding of the structureactivity relationship is an active area of research by many groups [19,20]. There are many advantages of gene delivery by cationic lipids. Lipoplexes are easy to produce, they are non-immunogenic, can deliver large polynucleotides into the cell. Ease of synthesis allows one to incorporate various functionalities to ne tune the eciency. However, poor cell specic binding, low cellular uptake of plasmid DNA, serum incompatibility and toxicity are some of the issues that need to be addressed carefully in order to obtain an ecient cationic lipid for gene transfection. Small interfering RNA (siRNA), sometimes known as short interfering RNA or silencing RNA, is a class of 2025 nucleotide-long double stranded molecules that play a variety of roles in biology [21]. Most notably, siRNA is involved in the RNA interference (RNAi) pathway, where it interferes with the expression of a specic gene. It was observed that when double-stranded RNAs (ds-RNA) are injected into the worm Caenorhabditis elegans, genes, whose sequences were complementary to those of the introduced ds-RNAs were silenced [22]. Now it is known that ds-RNAs are processed into short interfering RNAs (siRNAs) by the RNase enzyme Dicer. These siRNAs are then incorporated into a silencing complex called RISC (RNA-induced silencing complex), which identies and silences complementary messenger RNAs [23]. Delivery of siRNA directly in cells can be achieved by using microinjection, electroporation or by the use of a transfection reagent. In addition to their role in the RNAi pathway, siRNAs also act in RNAi-related pathways, e.g., as an antiviral mechanism or in shaping the chromatin structure of a genome. Also it has been shown that ds-RNAs targeting gene promoters induce potent transcriptional activation of associated genes and these are termed small activating RNAs (saRNAs). The fact that RNAi via siRNAs can knock down any gene of interest, it has a great potential to contribute to modern medicine [24]. Because disease processes also depend on the activity of multiple genes, it is expected that in some situations turning o the activity of a gene with a siRNA could produce a therapeutic benet. There is also potential for using RNAi for the treatment of viral diseases such as those caused by the

hepatitis C virus (HCV) and the human immunodeciency virus (HIV). Synthetic siRNAs and expressed shRNAs have been used to target several early and late HIV-encoded RNAs in cell lines and in primary haematopoietic cells. When a hepatitis B virus (HBV) was co-delivered with an expression unit encoding an anti-HBV shRNA in mice, a signicant knockdown (99%) of the HBV core antigen in liver hepatocytes could be achieved. Many studies have used siRNAs as an experimental tool to dissect the cellular pathways that lead to uncontrolled cell proliferation and to cancer. Thus RNAi has been proposed as a potential treatment for cancer.

9. Challenges using siRNA technology Sometimes non-specic eects are triggered by the experimental introduction of a siRNA. A mammalian cell may mistake siRNA as a viral by-product and mount an immune response. O-targeting can also occur where the genes with incomplete complementarity (microRNAs) are inadvertently downregulated by the siRNA and cause misinterpretation of data and toxicity. Introduction of too much siRNA can result in non-specic events due to activation of intrinsic immune responses. Thus intracellular delivery at therapeutically eective concentrations is yet to be achieved. To derive maximum benet from this emerging technology, strategies that maintain a high level of siRNA eciency while minimizing the non-specic o-target eects associated with siRNA expression should be explored in detail.

10. Protein therapeutics 10.1 Folding and inverse folding problems Proteins are biopolymers that are typically composed of 20 dierent kinds of monomer units (amino acids) linked by peptide bonds. Most proteins fold into a well dened three dimensional structure. The structure of a protein is intimately linked to its function. For many proteins, the information specifying the structure is encoded in the amino acid sequence. This can be shown by denaturing puried protein in aqueous solution and allowing the protein to refold by removal of the denaturing agent. Although it is widely accepted that sequence species structure, it is still not possible to predict the structure of a protein from rst principles given its amino acid sequence. This problem is known as the protein folding problem

14

SANTANU BHATTACHARYA AND RAGHAVAN VARADARAJAN

Figure 2. Schematic description of protein folding and inverse folding problems. (A) The protein folding problem deals with prediction of a protein three-dimensional structure from its amino acid sequence. (B) The inverse folding problem is to predict all possible sequences compatible with a given protein three-dimensional structure.

and it remains one of the major unsolved problems in biology. A related problem, known as the inverse folding problem has to do with enumeration of all possible sequences consistent with a given protein structure (gure 2). Solutions to this problem have important implications for protein design. Below we briey describe progress in this area and applications to the design of biological therapeutic agents. In proteins, bond lengths and bond angles between dierent types of atoms are highly conserved and depend only on the type of atom(s) involved, not the specic amino acid or protein. Dierences in three dimensional structures between proteins arise primarily from dierences in dihedral angles. This insight was made use of by Ramachandran and colleagues [25] who showed that protein main chain congurations are described by two dihedral angles ( and ) and that substantial regions of conformational space are disallowed because of steric overlap between non-bonded atoms. While Ramachandran and colleagues focused on steric restrictions imposed primarily by the main chain, protein structures are also constrained by side chain interactions. A major advance in addressing the inverse folding

problem came from the work of Ponder & Richards [26]. They demonstrated that approximating side chain conformations by discrete rotamers from a rotamer library, combined with a few additional constraints (avoidance of steric overlap, buried charge and large cavities) greatly restricted the number of sequences energetically compatible with a given main chain conguration/fold. One of the major challenges in inverse folding or protein design is addressing the combinational nature of the problem. For example, for a 100 residue protein, there are 20 possible amino acids at each position. If we assume that each amino acid on an average can be modeled by three rotameric congurations, there are still (20 3)100 i.e. 10177 possible congurations that need to be sampled. A number of elegant approaches and approximations have been used [27,28] to reduce the search space and provide approximate solutions to the problem. A more tractable and commonly encountered problem is the structure based prediction of a limited number of mutations to alter specic properties of a protein. For example, mutations to change protein stability or to modulate protein: protein or protein: substrate interactions. Major limitations in protein design arise from errors in

NOVEL THERAPEUTIC STRATEGIES

15

Figure 3. Antibody structure. (A) Schematic diagram of an immunoglobulin G type of Ab. Variable and constant domains are prexed with V and C, respectively. The VH and VL variable regions are involved in antigen binding. The Fc region involved in eector function consists of the CH2 and CH3 domains. (B) Structure of mAb b12 (adapted from the pdbid 1 HZH) using the program RASMOL (http://www.rasmol.org). Colours of the corresponding regions of (A) and (B) are identical.

the potential function(s) employed and the use of various approximations and cutos to compute energies of specic sequence-structure combinations. It is therefore essential to be able to experimentally construct and test a large number of protein sequences and then experimentally select the optimal ones from a large pool. There are a variety of methodologies for doing this, which include phage display, yeast display and ribosome display [29]. 11. Antibodies as therapeutics The vast majority of therapeutics currently used to treat disease are small molecules with molecular weights of < 500 Da [30]. In recent years there have been increasing numbers of biological therapeutics developed by the biotechnology industry. Monoclonal antibodies (mAbs) and related molecules constitute the major fraction of protein therapeutics with 2006 US sales of $ 14.5 billion [31] and an annual growth rate > 35%. However, the high cost of these treatments (of the order of Rs. 100,000/- per dose in some cases), have put them out of reach of the vast majority of people, especially in the developing world. It is therefore imperative to develop better and cheaper ways of developing and producing these molecules Antibodies (Abs) typically produced by the body in response to infection by a foreign agent are a complex mixture of molecular species, termed polyclonal antibodies. However, in 1975 Milstein and

Kohler [32] showed that large amounts of homogenous antibody can be produced by fusion of an Ab producing cell with a myeloma cell. Since the Ab producing cell is typically from a mouse that has been immunized with the target protein (antigen), the resulting Abs need to be modied (humanized) before they can be used for therapeutic purposes in humans, as mouse derived Abs will provoke an immune response in humans. However, mouse derived Abs can be used without modication in diagnostic applications. Other drawbacks of conventional hybridoma technology are that Abs cannot be raised against self molecules, toxins and some small molecules because of toxicity and/or lack of immunogenicity. An alternative is the use of in vitro technology such as phage display or yeast surface display [29].

12. Antibody structure A schematic description of an immunoglobulin G (lgG) type Ab molecules is shown in gure 3. These constitute the major type of Ab in serum. The molecule consists of a total of four polypeptide chains, two identical pairs of light chains and two identical pairs of heavy chains. Each light chain interacts with a single heavy chain through disulde bonds and noncovalent interactions. The N terminal 110 amino acids of each chain constitute the variable regions. Within the variable regions, maximal sequence diversity occurs within the so called complementarity determining regions

16

SANTANU BHATTACHARYA AND RAGHAVAN VARADARAJAN

(CDRs) which are directly involved in antigen binding. The CH2 region contains a structurally and functionally important conserved glycosylation site. In addition there may be other nonconserved glycosylation sites in other regions of the molecule. The constant region (Fc) of an Ab is involved in eector functions which involve binding to receptors (Fc receptors) on immune cells, such as natural killer (NK) cells or macrophages. Binding of multiple Abs to an antigen in vitro results in clustering of Fc regions and binding to Fc receptors, which in turns targets cells or particles displaying the antigen for destruction. In the following sections, we describe approaches that have been used to design and select therapeutic antibodies (Abs) for use in humans and the possible use of similar technology for vaccine development. The rst mAbs to be tested in humans were rodent derived Abs developed by conventional hybridoma technology. However these suered from several problems, such as low half-lives in humans, poor eector function and high immunogenicity (as they were recognized as foreign objects by the human immune system). More recently, these problems have been substantially addressed by the use of chimeric as well as humanized mAbs [33]. Chimeric mAbs are ones in which the VH and VL chains are murine (mouse) derived, while the constant regions are of human origin. Such mAbs contain approximately 33% of murine derived amino acid sequence. Humanized mAbs contain only about 10% of murine sequence primarily in the hypervariable loops. Recent methodologies such as phage display of Ab libraries or transgenic mice with human immunoglobulin genes [34,35] have made it possible to produce fully human mAbs (containing only human derived sequence) with desired specicities. It still remains to be shown whether fully human mAbs provide signicant therapeutic benet over chimeric or humanized mAbs [36]. 13. Designer antibodies Protein design and selection methodology has been extensively used to modulate antigen binding, eector function and the in vitro half lives of mAbs and to design Ab fragments that retain one or more functions of the intact molecule. The procedure for the in vitro selection of Abs from large libraries is summarized in gure 4. For mAb libraries generated in vitro, the diversity of the library can be generated either from synthetic variable genes or using variable regions cloned from human B cells. During in vitro selections, the phenotype of the mAb displayed on a particle is linked to the genotype (DNA sequence encoding the mAb/fragment) which is packaged inside the particle (gure 4).

Figure 4. Scheme for in vitro selection of Ab/Ab fragments that bind to a given antigen.

The library consists of a large number of such particles (bacteriophages, viruses or cells). In one round of selection, the library is bound to immobilized antigen and unbound library members are washed o. Bound members are eluted, allowed to replicate in an appropriate host, repuried and subjected to further rounds of selection. At each round, the fraction of the input pool that binds is measured. Typically after a few rounds of selection, a signicant fraction of the input pool shows binding. At this point individual members are sequenced, expressed, puried and characterized for binding anity to the desired target. The diversity of libraries that can be screened in vitro is typically limited by the transformation eciency of the host cell which is typically less than 109 transformants/g of DNA. Since there are 20 possible amino acids at each position, even a library where only 8 positions are randomized will have a diversity (208 = 256 108 ) in excess of what can be conveniently transformed. It is therefore important to use computational analyses to guide the choice of regions to be mutated as well as to introduce specic types of diversity designed to simulate the types of amino acids that are abundant in natural binding sites [37]. Recent studies have attempted to dene the minimum diversity in amino acid sequence required for ecient recognition of antigens by mAbs. Eleven synthetic libraries were constructed with CDR diversity restricted to only four amino acids (including at least one small amino acid to provide conformational exibility) and screened for binding to four dierent protein antigens. Two libraries (with amino acids Y/A/D/S and H/R/D/G) generated Abs against 3 of the 4 antigens [38]. In another study [39] it was shown that functional Abs could be obtained with a diversity in CDRs generated from only two amino acids (Y and S). High anity (Kd = 60 nM) and high specity Abs against human vascular

NOVEL THERAPEUTIC STRATEGIES

17

endothelial growth factor were obtained even with such a binary recognition code. In addition to improving antigen binding anity, protein design and selection strategies have also been used to improve Ab eector function. In a recent study [40], automated protein design methodology combined with high throughput screening was used to engineer mutations in the Fc region of an anti-CD52 mAb. The best of the designed variants showed an increase in eector function of over 100 fold because of increased binding to the Fc RIIIa receptor. In another study [41] the in vivo half of Abs and Ab fragments was modulated by mutations in the Fc region to alter interactions with the protective neonatal Fc receptor FcRn. This has both therapeutic and imaging applications. One major barrier to the use of mAbs as therapeutics is high cost. In part this is because they need to be expressed using expensive mammalian cell culture systems to ensure correct glycosylation. Recently it has been possible to engineer yeast and other microbes to produce glycosylation patterns similar to those found in human cells [4245]. These advances have the potential to dramatically lower costs of mAbs by expression in microbial systems. Alternative methods that can potentially lower costs of treatment are the use of genetically engineered cells for the continuous product of Abs in vivo [46]. 14. Vaccine design As can be seen from the above discussion, the development of Ab therapeutics has been facilitated by a combination of developments in computational methodology for protein design, library construction, screening via phage/yeast display, development of appropriate transgenic mice and advances in understanding of factors modulating Ab eciency in vitro. All the tools used in Ab design can also be applied to design better vaccines. In the developing world, infectious diseases are the leading cause of death. For three of the major infections diseases (TB, malaria and HIV) no vaccine exists. For other diseases such as inuenza, a satisfactory vaccine currently exists. However, the advent of highly pathogenic avian inuenza (bird u) and pandemic H1N1 is a reminder that new strategies to counter such newly emerging infections are urgently required. New approaches to vaccine design involve identication of the target antigen, followed by determination of its molecular structure as well as mapping regions of the structure (neutralization epitopes) where Ab binding will reduce or abrogate infectivity/pathogenicity. Following this, protein design methodology can

be used to construct fragments/domains of the protein containing the desired neutralization epitopes. The designed proteins can be screened for binding to known neutralizing mAbs. Selected molecules (immunogens) can be injected in appropriate animal models to see if they can induce desired Ab responses and, where feasible, protect the animals from pathogenic challenge. Immunogens which show encouraging responses in small animals can be further tested in non-human primates and eventually in humans. This approach, known as reverse vaccinology or retrovaccinology [47] has the potential to accelerate the development of vaccines for emerging infections. It can also be used to reduce costs of existing vaccines, by redesigning immunogens so that they can be expressed in microbial systems instead of currently used more expensive cell culture systems. However, further work needs to be done to assess the benets of this methodology over conventional methods of using attenuated or killed versions of the infectious organism as vaccines.

15. Conclusions Multitarget therapeutics as well as network analysis derived multicomponent therapeutics have the potential to treat diseases that have proved resistant to conventional single target therapeutics. Gene therapy and siRNA are alternatives to conventional drug treatment. Advances in gene delivery via cationic lipids are important for the enormous potential benets of gene therapy to be realized. Small molecule drugs are by far, the most common allopathic therapeutics. However, the market for biological therapeutic proteins, especially antibodies and antibody derivatives is rapidly growing. Advances in computational biology, structural biology and immunology have led to the introduction of numerous mAb therapeutics in recent years. Improvements in protein design and screening methodologies have made it possible to modulate the stability, in vivo half-life and function of mAbs and similar molecules in a semi rational fashion. Similar methodologies can potentially be applied to design immunogens and vaccines for newly emerging global health threats.

Acknowledgments S B thanks V Guru Raja for his help with the manuscript organization. R V thanks Bharat Adkar and Piyali Saha for providing gures and Bhagyashree K G for help with manuscript preparation.

18

SANTANU BHATTACHARYA AND RAGHAVAN VARADARAJAN

References
[1] Russell R B and Aloy P 2008 Targeting and tinkering with interaction networks; Nat. Chem. Biol. 4 666673. [2] Zimmermann G R, Lehar J and Keith C T 2007 Multitarget therapeutics: When the whole is greater than the sum of the parts; Drug Discov. Today 12 3442. [3] Hopkins A L 2008 Network pharmacology: The next paradigm in drug discovery; Nat. Chem. Biol. 4 682690. [4] Whitehurst A W, Bodemann B O, Cardenas J, Ferguson D, Girard L, Peyton M, Minna J D, Michno C, Hao W, Roth M G, et al 2007 Synthetic lethal screen identication of chemosensitizer loci in cancer cells; Nature 446 815819. [5] Engelman J A, Luo J and Cantley L C 2006 The evolution of phosphatidylinositol 3-kinases as regulators of growth and metabolism; Nat. Rev. Genet. 7 606619. [6] Hennessy B T, Smith D L, Ram P T, Lu Y and Mills G B 2005 Exploiting the PI3K/AKT pathway for cancer drug discovery; Nat. Rev. Drug. Discov. 4 9881004. [7] Knight Z A, Gonzalez B, Feldman M E, Zunder E R, Goldenberg D D, Williams O, Loewith R, Stokoe D, Balla A, Toth B, et al 2006 A pharmacological map of the PI3-K family denes a role for p110alpha in insulin signaling; Cell 125 733747. [8] Barouch D H 2008 Challenges in the development of an HIV-1 vaccine; Nature 455 613619. [9] De Clercq E 2007 The design of drugs for HIV and HCV; Nat. Rev. Drug Discov. 6 10011018. [10] Stier A and Sackmann E 1973 Spin labels as enzyme substrates. Heterogeneous lipid distribution in liver microsomal membranes; Biochim. Biophys. Acta. 311 400408. [11] Simons K and van Meer G 1988 Lipid sorting in epithelial cells; Biochemistry 27 61976202. [12] Simons K and Toomre D 2000 Lipid rafts and signal transduction; Nat. Rev. Mol. Cell. Biol. 1 3139. [13] Carter G C, Bernstone L, Sangani D, Bee J W, Harder T and James W 2009 HIV entry in macrophages is dependent on intact lipid rafts; Virology 386 192202. [14] Garg R, Lodge R, Descoteaux A and Tremblay M J 2008 Leishmania infantum promastigotes reduce entry of HIV-1 into macrophages through a lipophosphoglycan-mediated disruption of lipid rafts; J. Infect. Dis. 197 17011708. [15] Miller A D 1998 Cationic liposomes for gene therapy; Angew. Chem. Int. Ed. 37 17691785. [16] Bhattacharya S and Bajaj A 2005 Recent advances in lipid molecular design; Curr. Opin. Chem. Biol. 9 647655. [17] Bajaj A, Kondaiah P and Bhattacharya S 2008 Synthesis and gene transfection ecacies of PEIcholesterol-based lipopolymers; Bioconjug. Chem. 19 16401651. [18] Bhattacharya S and De S 1999 Synthesis and vesicle formation from dimeric pseudoglyceryl lipids with (CH2)(m) spacers: Pronounced m-value dependence of thermal properties, vesicle fusion, and cholesterol complexation; Chemistry-A European Journal 5 23352347. [19] Bajaj A, Kondaiah P and Bhattacharya S 2007 Synthesis and gene transfer activities of novel serum [20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31] [32] [33]

[34]

[35]

[36]

compatible cholesterol-based gemini lipids possessing oxyethylene-type spacers; Bioconjug. Chem. 18 15371546. Bajaj A, Kondiah P and Bhattacharya S 2007 Design, synthesis, and in vitro gene delivery ecacies of novel cholesterol-based gemini cationic lipids and their serum compatibility: A structure-activity investigation; J. Med. Chem. 50 24322442. Hannon G J and Rossi J J 2004 Unlocking the potential of the human genome with RNA interference; Nature 431 371378. Fire A, Xu S, Montgomery M K, Kostas S A, Driver S E and Mello C C 1998 Potent and specic genetic interference by double-stranded RNA in Caenorhabditis elegans; Nature 391 806811. Preall J B, He Z, Gorra J M and Sontheimer E J 2006 Short interfering RNA strand selection is independent of dsRNA processing polarity during RNAi in Drosophila. Curr. Biol. 16 530535. Gregory R I, Chendrimada T P, Cooch N and Shiekhattar R 2005 Human RISC couples microRNA biogenesis and posttranscriptional gene silencing; Cell 123 631640. Ramachandran G N, Ramakrishnan C and Sasisekharan V 1963 Stereochemistry of polypeptide chain congurations; J. Mol. Biol. 7 9599. Ponder J W and Richards F M 1987 Tertiary templates for proteins. Use of packing criteria in the enumeration of allowed sequences for dierent structural classes; J. Mol. Biol. 193 775791. Lasters I, De Maeyer M and Desmet J 1995 Enhanced dead-end elimination in the search for the global minimum energy conformation of a collection of protein side chains; Protein Eng. 8 815822. Kuhlman B and Baker D 2000 Native protein sequences are close to optimal for their structures; Proc. Natl. Acad. Sci. U S A 97 1038310388. Hoogenboom H R 2005 Selecting and screening recombinant antibody libraries; Nat. Biotechnol. 23 11051116. Lipinski C A, Lombardo F, Dominy B W and Feeney P J 2001 Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings; Adv. Drug. Deliv. Rev. 46 326. Aggarwal S 2007 Whats fueling the biotech engine? Nat. Biotechnol. 25 10971104. Milstein C 1980 Monoclonal antibodies; Sci. Am. 243 6674. Lonberg N 2008 Fully human antibodies from transgenic mouse and phage display platforms; Curr. Opin. Immunol. 20 450459. Green L L, Hardy M C, Maynard-Currie C E, Tsuda H, Louie D M, Mendez M J, Abderrahim H, Noguchi M, Smith D H, Zeng Y, et al 1994 Antigen-specic human monoclonal antibodies from mice engineered with human Ig heavy and light chain YACs; Nat. Genet. 7 1321. Lonberg N, Taylor L D, Harding F A, Trounstine M, Higgins K M, Schramm S R, Kuo C C, Mashayekh R, Wymore K, McCabe J G, et al 1994 Antigenspecic human antibodies from mice comprising four distinct genetic modications; Nature 368 856859. Gibson T B, Ranganathan A and Grothey A 2006 Randomized phase III trial results of panitumumab, a fully human anti-epidermal growth factor receptor

NOVEL THERAPEUTIC STRATEGIES

19

[37]

[38]

[39]

[40]

[41]

monoclonal antibody, in metastatic colorectal cancer; Clin. Colorectal Cancer 6 2931. Sidhu S S, Li B, Chen Y, Fellouse F A, Eigenbrot C and Fuh G 2004 Phage-displayed antibody libraries of synthetic heavy chain complementarity determining regions; J. Mol. Biol. 338 299310. Fellouse F A, Wiesmann C and Sidhu S S 2004 Synthetic antibodies from a four-amino-acid code: A dominant role for tyrosine in antigen recognition; Proc. Natl. Acad. Sci. U S A 101 1246712472. Fellouse F A, Li B, Compaan D M, Peden A A, Hymowitz S G and Sidhu S S 2005 Molecular recognition by a binary code; J. Mol. Biol. 348 11531162. Lazar G A, Dang W, Karki S, Vafa O, Peng J S, Hyun L, Chan C, Chung H S, Eivazi A, Yoder S C, et al 2006 Engineered antibody Fc variants with enhanced eector function; Proc. Natl. Acad. Sci. USA 103 40054010. Olafsen T, Kenanova V E and Wu A M 2006 Tunable pharmacokinetics: Modifying the in vivo half-life of antibodies by directed mutagenesis of the Fc fragment; Nat. Protoc. 1 20482060.

[42] Chiba Y and Jigami Y 2007 Production of humanized glycoproteins in bacteria and yeasts; Curr. Opin. Chem. Biol. 11 670676. [43] Hamilton S R, Davidson R C, Sethuraman N, Nett J H, Jiang Y, Rios S, Bobrowicz P, Stadheim T A, Li H, Choi B K, et al 2006 Humanization of yeast to produce complex terminally sialylated glycoproteins; Science 313 14411443. [44] Wildt S and Gerngross T U 2005 The humanization of N-glycosylation pathways in yeast; Nat. Rev. Microbiol. 3 119128. [45] Kowarik M, Numao S, Feldman M F, Schulz B L, Callewaert N, Kiermaier E, Catrein I and Aebi M 2006 N-linked glycosylation of folded proteins by the bacterial oligosaccharyltransferase; Science 314 11481150. [46] Luo X M, Maarschalk E, OConnell R M, Wang P, Yang L and Baltimore D 2009 Engineering human hematopoietic stem/progenitor cells to produce a broadly neutralizing anti-HIV antibody after in vitro maturation to human B lymphocytes; Blood 113 14221431. [47] Burton D R 2002 Antibodies, viruses and vaccines; Nat. Rev. Immunol. 2 706713.