shape-zoom2 shape-zoom2 chevron-down edit share2 twitter youtube

Probing Novel Biology

To translate genetic insights into new therapies, Amgen is mounting an expedition into uncharted territories of human biology.



While advances in human genetics are helping to find gene variants strongly linked to disease risk, quite often these newly identified genes code for proteins about which little or nothing is known. To potentially turn these emerging insights into new therapies, Amgen scientists must answer basic questions about the gene's role in human biology.

  • What is the normal biological function of the protein made by the gene?
  • What other molecules does the protein of interest interact with? How exactly does it influence disease risk?
  • Does the risk observed with the variant gene result from a gain or a loss in the protein's function?
  • When is the gene active-primarily during development or in adults as well?

Discovery teams can often find answers to questions like these in scientific publications. "When we consider pursuing a new drug target, we start by reading all the relevant research in the scientific literature," said Paul Kassner, a director in Amgen’s Genome Analysis Unit. "With a well-known target, it may take us three months to get up to speed on the literature." However, with some of Amgen’s new efforts based on genes discovered by its subsidiary, deCODE Genetics, "we might be fully up to speed in two days because there is almost nothing out there," Kassner said.

It’s unusual and quite challenging to focus a drug discovery program on new and unexplored biology. Investigational drugs aim to treat disease through specific effects on targeted proteins-an aim that presumes a hypothesis regarding that protein's influence on disease. In many discovery efforts, this hypothesis forms the basis for launching a program.

In gene-driven drug discovery, you start out with strong genetic clues but may lack any clear mechanistic information about a target. You need to build a solid mechanistic hypothesis by probing the gene and the protein it expresses. All this upfront effort is justified when the human genetics are saying that the biology, once understood, could lead to a major therapeutic advance.

Playing Twenty Questions with Nature

Exploring novel biology is as much an art as a science. There's no checklist or basic strategy teams can use to get the answers they need to go forward. "It's a tougher beast when you start out knowing so little," Kassner said. "You need to be very innovative and take some risks in deciding which experiments to use in trying to figure out the biology."

To support discovery teams with expertise in diseases like diabetes and Alzheimer's, Amgen created elite groups adept at deploying the latest tools of modern biology.the latest tools of modern biology

These tools include:

  • Specialized cell lines
  • RNA interference methods
  • Viral vectors to introduce genes into cells
  • Computational biology
  • Various types of genetically modified mice

In addition, it may be possible to recruit and study people who have the rare variants that Amgen is investigating.

To support discovery teams with expertise in diseases like diabetes and Alzheimer's, Amgen created elite groups adept at deploying the latest tools of modern biology. These tools include specialized cell lines, RNA interference methods, viral vectors to introduce genes into cells, computational biology, and various types of genetically modified mice. In addition, it may be possible to recruit and study people who have the rare variants that Amgen is investigating.

Each tool may help to solve a piece of the puzzle of how a gene functions—or it may not. The exploratory process resembles a game of Twenty Questions: to get the right answers, you need to ask the right questions.

"The questions come down to the details of the experiments you're running," said Peter Coward, a principal scientist in Amgen's Metabolic Disorders Therapeutic Area. "Are you using the right cell type? Have you chosen the right type of assay? We get clues on where to look and what to ask, but the answer to most of these questions is "˜No,' so we need to develop and test another hypothesis. It's difficult, dynamic, highly varied, and iterative work. It can also take a long time, and you have to be okay with that."

Shifting from Exploration to Execution

"Deciding when we have enough data to launch a drug discovery program takes a mix of instinct and experience," said Stefan McDonough, an executive director and head of the Genome Analysis Unit. "If you start screening too soon, you may waste time and resources pursuing the wrong target. But if you expend too much effort making sure that you've got the right target, you may lose the head start that comes from being the first team to explore a new gene variant."

Kassner notes that you don't need to answer every question before beginning the search for a drug. "We will often start a drug discovery campaign without knowing exactly what's going on with the biology," he said. "We need to know the direction of the mutation we're studying"”whether it's a gain of function or a loss of function. And to start screening, we need some activity to screen for, so you need at least some minimal function identified."

Screening efforts and exploration of basic biology can run on parallel tracks, for a time at least. In the long run, the goal is to understand a target's biology in-depth to help choose which type of drug and specific molecule to advance into clinical trials.

The scientists doing this work are well aware of Amgen's history of discovering medicines by working on the leading edge of biology. For example, two Amgen bone therapies can be traced back to the discovery of OPG (osteoprotegerin). Amgen identified this protein through bone density X-rays of transgenic mice engineered to overexpress the OPG gene. A landmark paper announcing these results was published in the journal Cell in 1997.

X-ray of a normal mouse (left) compared to an X-ray of a mouse engineered to overexpress OPG (right). This evidence helped to establish OPG's role in regulating bone resorption. Genetically modified mice are a common tool that scientists use to discern gene function.

"I saw the initial paper in Cell and thought it was so cool that a company was doing this type of research," Kassner recalls. "I wanted to work at Amgen ever since."

An Amgen bone therapy was first tested in humans in June 2001, years after the first genetic clues were uncovered. This timeline is a reminder that finding a gene is a critical but early step in the journey of scientific exploration and drug development. The key difference now is that disease-linked genes can be located more systematically, giving Amgen discovery teams more leads to pursue—and a head start.

"We were way out in front with our bone research because a lot of the biology behind that program was discovered inside Amgen," said Kassner. "In the near term, it's tougher to face a lot of unknowns, but it's the only way to discover our own novel targets and to be out in front in finding entirely new ways to treat disease."


Humans have approximately 20,0001,2 unique genes, and each one can encode multiple distinct proteins. Even a single cell may be home to thousands of different types of proteins, each with its own role to play in the complex interactions that regulate living systems. Discerning the role of newly discovered proteins amidst this complexity is quite challenging. Here are some of the tools and technologies Amgen scientists rely on to meet that challenge.

Gene and protein families

While every protein has its own distinct function, different classes and families of proteins may have structural and functional similarities. These "family resemblances" can be observed in the protein's amino acid sequence, as well as its three-dimensional structure, once that structure has been solved. Structural details can offer broad clues in regard to the protein's biological function—whether it's a kinase (a type of enzyme), a receptor, or a transporter. But more in-depth studies are needed to define function precisely—and sometimes family resemblances prove misleading.

External databases

With a recently discovered gene, there are few if any published studies describing the gene or hinting at its biological role. However, messenger RNA from the gene of interest may still be included in external gene expression databases. Major research programs, such as the Genotype-Tissue Expression project (GTEx) or the Allen Brain Atlas, are cataloguing the tissues and cells in which different genes and their protein products are active. While these data won't tell you how the gene functions, they can tell you which cells to study to gain insights into the function. Once biologists know where the gene operates, they can search databases of cell lines to determine if the cells of interest are commercially available or need to be produced in-house.

Human phenotyping

People with rare variant genes may consent to undergo lab tests, which can be used to study how changes in gene function correlate with changes in physiological function and biochemistry. For example, people with a rare variant that reduces diabetes risk can be tested to look for differences in how their bodies utilize insulin, glucagon, and other hormones involved in regulating blood glucose.

Gene silencing

A gene is expressed when its DNA is used to make RNA, and the RNA is translated into the chain of amino acids that forms the protein. Gene expression can be silenced by RNA interference technology. This approach employs strands of complementary RNA and a naturally occurring protein complex to initiate messenger RNA degradation or inhibit translation, which ultimately prevents protein production. This technology can be used to silence almost any gene in any cell type. Cells in which specific genes have been silenced can then be studied to determine how they differ from cells where gene expression is unencumbered.

A technique called small interfering RNA (siRNA) can provide transient silencing of targeted genes for short-term experiments, while short hairpin RNA (shRNA) can be integrated into the cell's nucleus to provide durable gene silencing for long-term experiments.3

Genetically modified mice

Since a high percentage4 of human genes are also found in one form or another in mice, perturbing a gene of interest in mice may give clues to its function in humans. By modifying mouse embryonic stem cells, scientists can create knock-out mice that are born lacking the gene of interest, as well as knock-in and transgenic mice, which are modified to carry a similar human gene or overexpress the mouse gene. These engineered mice can then be studied in-depth to determine how the inserted or deleted gene affects their overall physiology and the function of particular organs or cells.

Gene editing

Various gene editing techniques are providing powerful and precise ways to modify the DNA inside cells. These techniques combine targeting proteins, engineered to bind to specific sites in the genome, along with nuclease enzymes that can snip out part or all of a targeted gene. These tools can be used to simply remove a gene, alter the gene, or replace it with a different or variant gene. In addition to providing a new way to explore gene function, gene editing techniques have intriguing therapeutic potential.

Antibodies, small molecules, and peptides

In addition to serving as therapeutic agents, small molecules, peptides, and antibodies can be used as research tools. Once a protein of interest has been identified, these drug-like molecules can be used to antagonize (inhibit) or agonize (stimulate) the activity of that protein, enabling experiments to measure the target's activity and map its interactions with other molecules.

Fluorescently tagged antibodies can also be used to determine where the target protein is expressed in the body and how different experimental conditions impact the protein's expression. Developing drug-like molecules as research tools can be time-consuming. On the upside, molecules originally developed as tools can provide a head start for the effort to discover a therapy.

Viral vectors

Viruses are widely used as vectors for introducing genes into cells. The gene that hitches a ride on the virus may code for the normal version of a protein or it may be a variant gene that codes for a protein with lower- or higher-than-normal function. The target destination may be cultured cells grown in vitro or cells inside a live animal, typically a mouse. Viral vectors can be deployed to either increase or decrease the activity of specific genes and their protein products.

Significant expertise may be required to design viruses that can get the desired gene into the targeted cell type and ensure that gene is adequately expressed. Amgen is an industry leader in the use of sophisticated viral vectors.


This information is current as of December 8, 2014. Amgen's product pipeline will change over time as molecules move through the drug development process, including progressing to market or failing in clinical trials, due to the nature of the development process. This description contains forward-looking statements that involve significant risks and uncertainties, including those discussed in Amgen's most recent Form 10-K and in Amgen's periodic reports on Form 10-Q and Form 8-K, and actual results may vary materially. Amgen is providing this information as of the date above and does not undertake any obligation to update any forward-looking statements contained in this as a result of new information, future events or otherwise.



  1. Ezkurdia I, Juan D, Rodriguez JM, Frankish A, Diekhans M, Harrow J, et al. Multiple evidence strands suggest that there may be as few as 19 000 human protein-coding genes. Hum Mol Genet. 2014 Nov 15;23(22):5866-78.
  2. Clamp M, Fry B, Kamal M, Xie X, Cuff J, Lin MF, et al. Distinguishing protein-coding and noncoding genes in the human genome. PNAS. 2007 Oct 3;104(49):19428"“19433
  3. Rao DD, Vorhies JS, Senzer N, Nemunaitis J. siRNA vs. shRNA: Similarities and differences. Adv Drug Deliv Rev. 2009 Jul 25;61(9):746-59
  4. Why Mouse Matters. Genome. 2010 Jul 23. Available from: