Name: Mao Mao
Title: Research fellow at Pfizer Oncology Research
President, Asian Cancer Research Group
Experience and Education: Director of molecular profiling and pharmacogenomics and China project liaison, Merck Research Labs, 2007-2009
Associate director of oncology clinical research, Merck Research Labs, 2005-2007
Research fellow, research biologist, research scientist, Rosetta Inpharmatics, 1999-2004
Assistant director, National Human Genome Center at Shanghai, 1998
PhD, MD, Shanghai Second Medical University, 1996
Masters of Medical Genetics, Sun Yat-sen University of Medical Sciences, 1991
Bachelor of Medicine, Suzhou Medical College, 1986
As next-generation sequencing becomes increasingly used in disease research, pharmaceutical companies are also looking to use the technology for drug development, to identify patients for clinical trials, and to find disease biomarkers.
Recently, Mao Mao, a research fellow at Pfizer Oncology Research who is working on several different cancer sequencing projects, spoke to In Sequence about the company's use of sequencing in drug development and the hurdles that still remain in implementing the technology in pharmaceutical research.
The following is an edited version of the conversation.
How are you using next-generation sequencing to develop drugs, and how is that approach different from traditional drug development?
We are using sequencing mainly for two purposes. One is for target discovery. So we want to look at those key driver mutations in cancer, and then develop drugs that target those mutations. The other purpose is to identify patients that will respond to the drug.
In traditional drug development, you first develop drugs, and then you do a large clinical trial to try to figure out if that drug is effective in the whole population, but many drugs have failed in that way. Those drugs all have a well-defined target, but not all patients have that target. That's why the efficacy will be diluted in the whole population, because the drug only works in a specific population.
Do you think that using sequencing in drug development will become more common?
Yes, I think so. If you look at all successful targeted therapies — drugs like Herceptin and Gleevec — they all have well-defined targets. So we need to identify those targets. Next-gen sequencing can help identify the single nucleotide mutations, gene fusions, or gene amplifications within those targets. We certainly believe next-gen sequencing can help us identify the Achilles' heel in cancer — the key oncogenic drivers. Therefore, it will help us design better targeted therapies and then deliver [them] to the targeted population.
Can you describe the different collaborations under which you have using next-generation sequencing?
We are working with the Asian Cancer Research Group, the University of Hong Kong, the Ontario Institute for Cancer Research, and the British Columbia Cancer Agency.
We are working on hundreds of tumor samples from hepatocellular carcinoma, gastric, breast, and colon cancer patients for next-generation sequencing, including whole genome, whole-exome, and transcriptome sequencing. The goal is to use next-gen sequencing to characterize tumors in the studies and identify cancer mutations and structural changes for target discovery and cancer molecular classification.
We're not just doing the discovery — not just trying to identify mutations. [We're] also trying to characterize those tumors. Eventually, we can implant the tumors in mice and do drug testing in vivo. And we can use these studies to characterize patients. If we do the sequencing, we know the mutational spectrum of those patients. And if we have some kind of drug we can match to the mutation, we can eventually enroll those patients in clinical trials because we've already used next-gen sequencing to characterize those patients' tumors.
[With the BC Cancer Agency] the main focus is on breast cancer. They have a large collection of breast cancer samples. We have started with gene expression profiling of approximately 800 tumors, and are then sequencing a subset of those tumors. We are trying different types of sequencing — whole-genome, whole-exome, and transcriptome — and looking at up to 100 samples.
The collaboration with the Asian Cancer Research Group is going to be a comprehensive genomic study. The initial stage will be 1,000 lung adenocarcinoma and 1,000 gastric tumors. And we will do gene expression studies and genotyping first. Then we will select a subset for next-gen sequencing. This group is a not-for-profit organization, co-funded by Pfizer, Lilly, and Merck. We will focus on the Asian-prevalent cancer types. So we'll have lung and gastric and also think we'll include HCC.
The Ontario Institute for Cancer Research is focused on colon cancer. Basically we want to use whole-genome SNP array and next-gen sequencing to characterize those tumors. For this one, we have a large cohort, but we are focusing on a subset of tumors we can characterize and do drug testing.
Where are you doing the sequencing, and on what platforms?
All the sequencing is being done at the academic centers and with the next-gen sequencing vendors. We are mainly using the Illumina sequencing platform.
Additionally, Pfizer’s Oncology Research Unit is looking at accessing next-gen sequencing for molecular diagnostics purpose. We're looking at all the popular platforms and also some platforms that may be more suitable for diagnostic purposes.
What are some of the hurdles that remain in using sequencing in drug development?
One is the cost. It's still very expensive to do next-gen sequencing. Also, we need to do it on a large scale, because when we talk about target discovery, we're not talking about sequencing one genome. We're talking about sequencing hundreds of genomes or more because we need to know what is in the population. So that's one big issue, cost.
If we're sequencing hundreds of genomes, or even thousands of genomes, that's going to be very expensive.
There are ways to reduce the cost. Maybe we can start with one well-annotated cohort, sequence just one hundred, and then validate those mutations in a larger cohort, instead of sequencing thousands of samples.
The other hurdle is we will have thousands of mutations from the sequencing, so the question is how to identify those key driver mutations. Again, we'll need a larger cohort to look at the mutational spectrum and to look at the frequency of those mutations. And we'll have to do a functional assay to understand whether the mutation is critical for cancer or not.
Some sequencing studies have found that a lot of potential drug targets for certain cancers are in a very small subset of the population. For instance, you have said that the EML4-ALK gene fusion, while a potential drug target, is only present in about 4 percent of non-small cell lung cancer patients. Why would a pharmaceutical company be interested in developing a drug that is effective in such a small population?
I think for us, the most important thing is still efficacy. If a drug is very effective, but only in a small population, the benefit is still going to be big. For a drug like Gleevec, it doesn't treat a big population, but it is a well-defined population, so the drug is effective, and is successful in treating those patients.