The Human Protein Atlas at the GoldLab Symposium

By Kevin Davies  

May 31, 2012 | BOULDER, CO—Larry Gold, the founder and CEO of protein biomarker company SomaLogic, hosted a diverse and engaging group of speakers for the third annual GoldLab Symposium* on the future of health care and personalized medicine earlier this month.   

“Technology is only the penultimate step in successful innovation,” Gold said. Putting technological advances into action is more the province of the social sciences. In keeping with that theme, the conference also included a vigorous debate on the definition and value of evidence-based medicine.   

Gold is a champion of the ‘omics revolution but argued that building a better medical infrastructure was essential to reap its rewards. With more longitudinal data on large patient cohorts, emerging technologies (including the biomarker panels being devised by SomaLogic) could facilitate wellness, enabling individuals to take earlier medical action, improve life quality and nutrition.   

“The existence of pre-disease blood is the key to finding treatable disease,” he said. A blood biobank of 1 million people, for example, stored over a decade could reveal dozens of biomarker signatures in a multitude of cancers. “From an N = 1 we learn little; from N = small, we learn a ton!” he said.   

Sweden’s Mathias Uhlen (Royal Institute for Technology, Stockholm), talked about an international consortium designed to identify the building blocks of human beings. Uhlen said the 21^st century would be the century of life sciences, following previous centuries in astronomy, biology, chemistry, and physics. A key resource moving forward would be population-based biobanks, of which there are many in Scandinavia, blessed with excellent clinical records.  

Uhlen, a co-developer of pyrosequencing, said that the Human Genome Project has fueled tremendous increase in our understanding of genome variability and regulation, but has had limited usefulness thus far in revealing new drug targets and signal pathways. The vast majority of pharmaceutical drugs on the market today target only 3% of protein targets. In the United States, an estimated $350 billion worth of prescriptions are ineffective  

“It’s a very good argument for patient stratification for better treatments,” said Uhlen.   

Backed by $150 million from the Wallenberg Foundation, Uhlen is spearheading the Human Protein Atlas (HPA) project, a consortium of some 30 groups in Europe, North America, and Asia. Purified proteins are sent to research centers in China and South Korea for mass spectrometry (MS) analysis and immunoassay studies.   

Uhlen credited a team of pathologists in Mumbai, led by Sanjay Navani, which analyzes some 10,000 immunolocalization images/day in a concerted effort to catalogue the location of every protein within cells and organelles. Some 12 million images are already publicly available, with information on 14,000 gene products expected when version 10.0 of the Atlas is launched this September.  

“If we do ten genes per day [from now on], we’ll have the first draft of the human proteome by 2015,” said Uhlen.  

About 40 commercial companies have provided HPA with more than 14,000 antibodies to characterize the subcellular localization. About 25% of proteins are membranous, while another 16% are secreted. “So 40% of proteins are either exported or sit on surface of cells,” noted Uhlen. About 20% human genes possess isoforms with differential locations.   

The first draft of the subcellular protein atlas is expected by 2015, with refinements continuing for a further five years. “By 2020, we’ll have a quality-assured atlas for all proteins,” said Uhlen.
Interestingly, the latest picture of protein localization suggests that there are very few tissue-specific genes or proteins. “Many proteins are differentially expressed,” said Uhlen. “The kidney has more or less the same proteins as the brain, but the concentrations vary considerably.”   

Uhlen said that the protein atlas will be an invaluable tool for biomarker discovery. “You can search for transcription factors [encoded] by genes on chromosome 14 that are expressed in the skin,” he said. In another example, he retrieved lists of candidate protein biomarkers that show differential expression in breast cancer patients—revealing RBM3 as a potential new biomarker for breast, ovarian, colon cancer, and melanoma.   

According to Uhlen, biomarker discovery in blood has been held back by its reliance on shotgun mass spectrometry. “The sensitivity of this method is a little bit [like an] iceberg. We can see what is sticking out, but not what is under sea level.”  

Celebrating Mass Spec  

Jeremy Nicholson (Imperial College, London), discussed the impact on health care on another key ‘omics field—metabolomics. “Bottom up” systems biology might have been successful in generating journal articles, but not so good for understanding “supraorganisms” such as humans, said Nicholson. “We don’t train doctors to think about connections between genes,” he said. The microbiome consists of an estimated 4 million genes. “We’ve devolved our genomic responsibility to all the microbes within us,” said Nicholson.  

The dozens of genes associated with complex diseases such as Crohn’s disease reveal a fascinating interplay with many other diseases. For example, genes associated with inflammatory bowel disease are not only related to Crohn’s and ulcerative colitis, but type 2 diabetes, multiple sclerosis, lupus, and psoriasis. “Diseases are really complicated. We need to understand those connections while treating diseases,” he said.

Nicholson criticized the incremental value of genome-wide association studies (GWAS). Flashing a newspaper headline—“DOCTORS FIND THE FAT GENE”—Nicholson noted that dozens of GWAS hits explained less than 1.5% variance in body mass index. “The title [of the newspaper story] should have been, ‘It’s not genetic, you fool!’ ” said Nicholson.   

He followed with a number of startling statistics regarding the microbiome, including the fun fact that humans will typically excrete some 25 tons microbes during their lifetime. “The human microbiome probably has more druggable targets than the human genome,” he said.  

Nicholson also showed how state-of-the-art technology is moving into the surgery suite at Imperial. As reported last year in The Lancet, the world’s first solid-state NMR provides near real-time diagnostics data for tissue imaging in certain situations. And a device called the iKnife sends the smoke generated during laser surgery to a mass spec machine to generate lipidomic profiles.  

“You can’t give a mass spectrum to a surgeon—he will stab you!” joked Nicholson. But such technology is having clinical impact. “If you can shorten journeys through the ICU by 10%, you could buy a new 600MHz machine every 3 weeks! Sometimes technology can be very cheap in comparison with the patient journey,” he said.

What Nonsense  

Allan Jacobson, co-founder of PTC Therapeutics, provided an exciting update on the role of nonsense mutations—premature stop codons—in genetic disease and the development of drugs that could provide benefit for a large proportion of some 2,400 monogenic disorders in which some instances are caused by nonsense mutations.  

“This is a new treatment for genetic disorders and about as personalized as you can get,” said Jacobson. “This is a disease modifier, it’s not a cure. It could treat many life-threatening genetic disorders.”

Jacobson, who is on the faculty at the University of Massachusetts Medical School, praised the role of serendipity in taking a basic science project looking at gene expression that morphed into a drug development program to treat fatal inherited disorders. “On average, all genetic disorders are caused by nonsense mutations about 15% of the time. That’s a lot of patients,” he said. And yet he noted, “We started studying this phenomenon in yeast.”  

Founded in 1998, PTC Therapeutics set out to identify small molecules that could allow read through of messenger RNA transcripts containing a premature nonsense mutation. In high-throughput screens, the company tested hundreds of thousands of molecules to see which molecules could promote read through of a premature termination codon in a reporter gene (Luciferase).

The ensuing lead optimization took five years from screening to the synthesis of a development candidate from a pool of 3,500 candidate analogues, ex vivo testing in genes such as dystrophin (for Duchenne muscular dystrophy) and CFTR (for cystic fibrosis), testing in mouse models, and safety screens.

The lead small-molecule candidate—Ataluren (PTC124)—was published in 2007. It provides dose-dependent read through of all three premature stop codons, said Jacobson, but not in the case of normal stop codons.

Ataluren was tested in a Phase 2a trial under a variety of dosing regimens. Immunofluorescence studies show the production of full-length CFTR protein after 45 days in the nasal epithelia. There is also good physiological data that the conductance properties in cystic fibrosis patient nostrils are being restored to normal, as well as statistically significant increases in lung function and decreases in cough frequency.

For Duchenne muscular dystrophy, Ataluren has been associated with improvements in dystrophin expression and patient walking tests compared to placebo. The benefits were sharpest with the lower doses, and improved over several months. Jacobson said the company had been advised to experiment with two different doses. He speculates that the drug acts on different binding sites on the ribosome, and higher doses may tweak the binding and reduce efficacy.  
“There is reason for optimism in certain groups,” said Jacobson, especially younger patients and those suffering with milder disease. “Drug development takes time and a lot of money. It’s an incredibly slow process. We’re now involved in regulatory affairs—what will it take to get our drugs approved?”

If the clinical trials continue to go well, the potential approval of the drug raises the possibility of off-label prescriptions for other diseases in which the same underlying genetic mechanism is at play. Jacobson said the FDA had already advised his company that it would be expecting a few more trials before the agency even thinks about any blanket approval.   

* GoldLab Symposium, May 14-15, 2012; University of Colorado, Boulder