Proteomics in action: Opening doorways to new drugs

Proteins are the effector molecules of life and the most common target for new drugs. Proteomics is the analysis of these proteins on a system-wide industrial scale and can be applied to any biological tissue or cell. Examining the right tissue in the right way can open the doorway to drug discovery on this same industrial scale. In the process new protein and peptide drugs can be found to complement drugs from other chemical classes such as small molecules and micro RNA. The potential is vast. From venom there are new bioactive peptides that may become drugs in their own right.Analysing the surface of a pathogen may isolate a novel protein target from which to generate a vaccine. Applying proteomics to human disease yields protein biomarkers that may eventuate in new personalised medicines.

Peptide drugs gain momentum

The role of biomolecular drugs (peptides, protein biosimilars and monoclonal antibodies), also called biologics or new biological entities (NBE's) is increasingly important, representing 32 percent of all new drugs approved between 2003 and 2008. On its own, the peptide market constituted $8.5bn in 2009, and according to the Peptide Therapeutic Foundation, at the beginning of 2010 there were 54 peptides approved for therapeutic use, 113 in early clinical phases, and 18 in phase 3 development and preregistration.

Peptide drugs are a target for biotech and pharma because they exhibit high activity and specificity, and show low accumulation in tissues. Their weaknesses are concurrently stability and oral bioavailability/transport across membranes, however, new delivery technologies such as nanoparticles are overcoming these issues. Sources of peptides are everywhere, and their innate diversity is enormous-consider a peptide composed of only 10 amino acids, with 20 possible amino acids in each position that is approximately 1013 combinations-over 1000 times the number of people on the planet. Given such diversity a mechanism for selecting optimised amino acid sequences is required. This is the time for drug discovery to turn to venom – a collection of molecules that at one dose are deadly toxins, but at another life saving drugs.

Nature's drug factories

Venom is a chemical cocktail of bioactive molecules where peptide sequences have been optimised through millions of years of evolution. Blockbusters drugs already include exenatide for type 2 diabetes (gila monster) and ziconotide, an analgesic for severe pain (marine cone snail). Given the energy required for small invertebrates such as spiders and scorpions to synthesise these molecules it is possible to speculate that each molecule has the potential to become a new drug-pain, inflammation, cell death, metabolism, insecticidal, bactericidal-and proteomics is now being used to open these factory doors.

Targeted discovery - shifting the odds

Traditional high-throughput screening (HTS) hit rates of 0.01- 1 percent are testimony to the wastage in small molecule drug discovery, and applying HTS to venom is further complicated by the quantities required. Much better to take the proteomics approach instead and identify the peptides that are present, recognising that using modern tools each peptide can be fully sequenced. These peptides can then be synthesised as required. The concept therefore becomes one of choosing the source of bioactive peptides, selecting the families of peptides that are likely to be interesting, and building a highly specific candidate library of only tens of peptides with the desired activity. This is possible using in silico or computer based screening, and by taking such an approach subsequent hit rates for in vivo animal models of above 50 percent have been achieved - 100 times more efficient than traditional screening.

The source of new drugs is clearly not restricted to venom. Any organism that makes small proteins (and peptides) is a potential reservoir of new drugs, and emerging evidence suggests that this is all organisms. From bacteria to human plasma the peptidome is ubiquitous and awaiting analysis using proteomics tools. Such data will yield new bioactive molecules and pathways, and in turn, potential new drugs and drug targets. And there is more than the peptidome for drug discovery using proteomics. Natural products from fruit to bacteria are potential sources of protein molecules that have biological activity and potential as new drugs. It is becoming a matter of using proteomics to look for the effector molecules that nature has already given us, some of which have been known for many years but never fully characterised. For example, from Ayurvedic systems of medicine from India which date back 2500 years, Karela - bitter melon (Momordica charantia) - contains Gurmarin, a protein similar to insulin that achieves a sugar regulating effect. Similarly, the paw paw (Carica papaya), which was first reported for its bioactive properties in 1750, has gained recent publicity as treatment in wound healing, and in which a proteomics approach has enabled deeper understandings of the mechanisms responsible.

Drugs from biomarkers

It is not only the direct generation of new drugs that proteomics promises to uncover, but also the targets for drugs of all classes. Biomarker has become a buzz word in an amazingly short period of time and proteomics is part of this, being used to discover protein biomarkers for many diseases. This too will lead to the discovery of new drugs, and here the application of proteomics unfolds in two directions simultaneously – human disease and vaccines against infection. In the former there has been an explosion of research papers on protein biomarkers (3276 in 2008-10) with the initial focus on cancer now expanded into other areas such diabetes and Alzheimer's. The need for such R&D, and the commercial opportunity for the pharmaceutical industry is exemplified by the numbers behind diabetes. Already 250 million people worldwide have diabetes, with the prevalence rate reaching 9.7 percent in China according to the 2010 Chinese National Diabetes and Metabolic Disorders Study. It is forecast that by 2030 the volume of sufferers will increase to 430 million. Along the way the value of the overall diabetes drugs market in China will triple over the next ten years to exceed US$2.1 billion. New treatments are needed.

The proteomics approach relies on comparative analysis of protein samples for healthy versus diseased groups, with the objective of finding which proteins have increased or decreased in concentration in the disease phenotype. At the first level of discovery such a protein may represent a diagnostic biomarker for the disease, finding a role in improving detection in areas such as prostate cancer where the prostate specific antigen (PSA) test is notoriously unreliable. The next step is to move from here rapidly towards prognostic diagnosis, and into companion diagnostics or personalised medicine.

Personalised proteomes

Consider two scenarios. The protein biomarker may be a critical component of the pathway that ultimately manifests itself in disease. It is therefore possible to target this protein, indeed this pathway, with any drug known to have such efficacy, thereby moderating the pathway and hence the outcome of the disease. A simple way to enact this might be to block a ligand-receptor interaction in the pathway, either with a small molecule drug, or perhaps a peptide drug such as exenatide. The second scenario for proteomics drug discovery is to assist in the segregation of the results of clinical trials to establish drug safety and efficacy – personalised medicine. It is acknowledged that many drugs ultimately fall over because they cause adverse effects in a subpopulation that was never detected in the original trial. The consequence is the drug is withdrawn, even if it is beneficial to the majority of the population. Similarly, a drug may fail to achieve significance in its phase 2 or phase 3 clinical trial because the population as a whole is not responsive. Again the drug fails.

Now proteomics is opening the door to personalised drug discovery through the study of these disease specific biomarkers – proteomics provides a therapeutic agent with a companion diagnostic test. The outcome is that for any administered drug the consequences can be followed for the individual recipient, and the movements of the biomarkers studied to indicate if the new drug is effective, or critically, causing unforeseen consequences. The big pharmaceutical companies have been slow to recognise this opportunity but are catching up fast as highlighted by Johnson and Johnson last year establishing its Centre of Excellence in Companion Diagnostics. It is a simple progression of the old adage that 'everyone is different', because in this case 'everyone has a different proteome'. In this context the proteome is the consequence of all levels of genetic transcriptional control (including the regulatory mechanisms of DNA and RNA, and how they interact with each other and with proteins) and that by its study, proteomics will enable the discovery of new drugs targeted to specific groups or individuals.

The increasing sophistication of mass spectrometry continues to offers ever improved levels of sensitivity, and gigabytes of data can be acquired each day. However, the real denominator is the quality of the data itself, that is to say patient cohorts and the design of the study. If the disease phenotypes are not correctly represented in the sample then the answers will not be forthcoming. That is where far sighted epidemiological studies such as the Busselton Health Study of Western Australia (commenced 1966) and the Parsi Zorastrian genome project in India come into their own, providing the opportunity for researchers to ask the specific questions needed to unlock the door.

New routes to traditional drugs

Vaccine discovery then wraps up the set. In the same way that protein biomarkers indicate human disease they can also be used to identify targets for vaccine based drugs. The theory is similar: find a specific marker(s) on the surface of an infectious organism and target that marker. Almost any pathogen can be analysed and its mode(s) of infection investigated, and here genomics and proteomics form a powerful combination allowing the identification and characterisation of new antigens. This time, because the marker is non-human antibodies can be raised to specifically latch onto the target and mark it for destruction via the body's immune system. On this occasion it is not just the doors that are opening, but the floodgates, with proteomics being used to develop new vaccines against numerous bacteria and viruses. Examples range from overcoming loss of livestock production due to the bacterium Brachyspira, the cause of porcine (Swine) dysentery which can be fatal to pigs and its avian strain which damages egg production, to keeping one step ahead of the H5N1 avian flu virus which is just re-emerging in several countries around the Asian region, and represents a possible predecessor of a flu pandemic which according to the World Bank could cost one trillion dollars – a big number, but still ten times less than the number of combinations of a 10 amino acid peptide drug.

Figure 1: Exendin-4 (in blue) binding to the hormone glucogen-like peptide 1 receptor :- the basis for the antidiabetes peptide drug exenatide (derived from the Gila monster)

Figure 2: Accelerating drug discovery from venom

Figure 3: Proteomics is being used in drug discovery to find a trait that is unique to a specific group