At CSHL Conference, Researchers Highlight Importance of RNA-seq Data to Guide Cancer Treatment

October 05, 2011

By Monica Heger

Targeted sequencing of select gene panels may miss druggable mutations in cancer patients when compared to a more comprehensive sequencing strategy that includes RNA-seq, according to several case studies presented at last weekend's Personal Genomes meeting at Cold Spring Harbor Laboratory.

Both Elaine Mardis, co-director of the Genome Institute at Washington University, and Karin Kassahn, a researcher in Sean Grimmond's lab at the University of Queensland in Brisbane, presented case studies of patients for whom druggable mutations were identified via a comprehensive sequencing strategy.

The findings add to the debate about whether targeted sequencing of cancer genes or a more in-depth sequencing approach that includes whole-genome and transcriptome sequencing makes the most sense for clinical sequencing in cancer.

RNA-seq to ID Drug Targets

The Wash U team follows a comprehensive cancer sequencing strategy that includes whole-genome sequencing of tumor and matched normal to 60-fold and 30-fold coverage, respectively, as well as exome sequencing for tumor and normal, tumor transcriptome sequencing, and deep digital sequencing of the tumor sample to verify mutations.

The higher coverage for tumor whole-genome sequencing helps "give us a lot of certainty that what we're seeing somatically is real," Mardis said in a presentation at last weekend's meeting.

Additionally, she added, transcriptome sequencing in particular has helped the team verify mutations, identify targets that they wouldn't have otherwise found, and rule out potentially druggable mutations that turned out not to be expressed in the tumor.
"If we're going to target a mutation, we better make darn sure it's expressed in the tumor genome," she said.

One example Mardis presented was of a patient in her mid-50s with metastatic breast cancer. After an initial biopsy showed that her tumor was ER-negative and HER2-positive, she was treated with paclitaxel and Genentech's Herceptin (trastuzumab). In 2010, however, the disease progressed, metastasizing to her brain, and surgery was performed.

The Wash U team then sequenced tissue from the surgical sample. An initial analysis of the RNA-seq results revealed that only about 40 percent of the mutations in the tumor genome were expressed.

Additionally, the team used the RNA-seq data to identify an amplification of HDAC2, a druggable target. "This is not evidence of a mutation, but evidence of over-expression," she said. The gene is "amplified to almost the same degree as HER2 … and this is a targetable amplification."

Two HDAC inhibitors — Merck's Zolinza (vorinostat) and Celgene's Istodax (romidepsin) — have been approved for use in cutaneous T-cell lymphoma, and several are in clinical trials for various other cancers.

While the patient is currently on a combination drug regimen that has stabilized her disease, the identification of the HDAC2 amplification will most likely guide future treatment, Mardis added.

The Wash U team found that including transcriptome sequencing also helped guide treatment for a patient with acute lymphocytic leukemia. The patient, whose leukemia developed when he was in his 20s, initially received a bone marrow transplant from a sibling, but then relapsed.

Following the relapse, he was given induction therapy and appeared to achieve remission. The Wash U team, which had sequenced the patient's tumor genome on his initial diagnosis, sequenced it again to see if there was any residual disease that couldn't be identified with other methods. The team identified the two populations of tumor subclones that were characteristic of his initial disease and likely led to his relapse. Those subclones indicated that even if he received a second transplant, "he would still be at risk for disease," Mardis said.

Looking at the whole-genome sequencing data, the team identified 91 somatic tier 1 SNVs. Transcriptome sequencing confirmed that 42 of them were expressed in the tumor tissue. However, "none were obvious targetable mutations, so we didn't know what to do."

A close look at the RNA-seq data revealed "unusual FLT3 expression, at extraordinarily high levels," Mardis said.

Two weeks ago, the patient was prescribed Pfizer's Sutent (sunitinib), which targets FLT3 among several other kinases. Three days following the initial Sutent treatment, a blood test indicated marked improvement.

The plan now, said Mardis, is to keep the patient on Sutent for three weeks and then re-evaluate his remission status. If remission is complete, a bone marrow donor has already been identified and he will receive a transplant.

Both cases illustrate the need for comprehensive sequencing, said Mardis.

Identifying Pathways

A team from the University of Queensland in Brisbane, Australia, has also found that a comprehensive sequencing approach helps to identify druggable mutations. Researchers in Sean Grimmond's lab at the Queensland Center for Medical Genomics, including Kassahn, have been sequencing the exomes of pancreatic cancer patients as part of its contribution to the International Cancer Genome Consortium.

Additionally, as part of a collaboration with Andrew Biankin's group at the Garvan Institute, the team is doing more comprehensive sequencing on select samples to try and determine how clinical sequencing could be employed on cancer patients. Biankin is also spearheading a clinical trial to use sequencing to guide treatment — the Individualized Molecular Pancreatic Cancer Therapy, or IMPACT, clinical trial — which will use next-gen sequencing to determine second-line treatment for patients.

While cohort genomics — sequencing the exomes or whole genomes of many patients —has the "power to detect recurrence" for projects like the ICGC, a more comprehensive strategy is needed for personal genomics, where the goal is to suggest therapy, Kassahn said.

At the Cold Spring Harbor meeting she presented an example of a patient her team sequenced as part of a pilot to determine the feasibility of sequencing cancer patients.

The woman had been diagnosed with pancreatic cancer at the age of 84. She had surgery and was treated with Eli Lilly's Gemzar (gemcitabine), but became resistant, and passed away 15 months after her initial diagnosis.

Kassahn said the team used an integrated sequencing strategy, performing whole-genome and exome sequencing, RNA-seq, and miRNA-seq on Life Technologies' SOLiD, and methylation array analysis. Additionally, they used the Ion Torrent PGM for amplicon sequencing to verify somatic mutations.

They identified 29 somatic missense and nonsense mutations, only 11 of which were expressed in the tumor. Additionally, they identified 22 regions of homozygosity loss, two indels, 14 interchromosomal events including one fusion transcript, 10 intrachromosomal events, three inversions, and 167 complex rearrangements that they are still verifying.

They identified mutated genes typical of pancreatic cancer — such as TP53, PTEN, and KRAS — as well as some private point mutations.

Making sense of the complex data to try and identify an alternative therapy would not have been possible if the team had just done whole-genome or whole-exome data, Kassahn said. By incorporating the transcriptome sequencing data and methylation data, the team was able to piece together the networks and pathways that were most heavily involved in the cancer, rather than focusing on specific genes.

The network analysis pointed to the DNA replication pathway as the top network. "Three mutations in that network stood out to us," Kassahn said. Two genes are known to be sensitive to mitomycin-C. One of them is part of the BRCA network, while the other is involved in cell survival following DNA damage and has also been seen in other cancers. The third gene also interacts with BRCA, is involved in DNA replications, and is chemosensitive to genotoxic stress.

Based on their analysis, the researchers tested mitocycin-C first in xenograft models of the tumor. The tumor stopped growing and in some replicates, there was a slight reduction, Kassahn said.

"That was exciting and pointed that we identified a useful target," she said.

In this case, however, the patient had already passed away, but Kassahn said the team will apply its lessons learned as part of the IMPACT clinical trial.

In particular, Kassahn said the integrated analysis, which included "overlaying the copy number changes, somatic mutations, expression data, and methylation status, … made us confident that we were on the right track."

The case also illustrated the difference between sequencing cohorts to identify common mutations in cancer and sequencing individual patients to identify possible therapies.

"A specific patient doesn't want to know what's common," Kassahn said. And in fact, the targetable mutations identified in this patient have not been seen in the 70 exomes the team has sequenced as part of the ICGC project, she added.

Targeted vs. Comprehensive Sequencing

While other groups, such as the University of Michigan's Center for Translational Pathology, are employing a similar comprehensive sequencing strategy for individual patients (CSN 8/31/2011), targeted sequencing has recently been gaining ground for cancer diagnostics (CSN 8/17/2011).

A number of organizations and companies, including a team at Dana Farber Cancer Institute, the Ontario Institute of Cancer Research, and Foundation Medicine are developing targeted cancer gene panels that will evaluate the mutational status of between about 50 and 200 known cancer genes.

These targeted panels could serve as a lower-cost and quicker method for helping to guide treatment for cancer patients, but Mardis said that they might miss genes that are over-expressed or deleted.

While the number of mutations these panels miss will depend in part on the tumor, she said that her team has estimated that between "20 to 30 percent of what needs to be known to inform treatment won't be captured by a targeted approach."

The debate is still unsettled as to which technique will ultimately prove more effective, and ultimately both models could prove necessary depending on the specific case.

Note: Taken from http://www.genomeweb.com/sequencing/cshl-conference-researchers-highlight-importance-rna-seq-data-guide-cancer-treat

Next-generation Sequencing (Illumina)

Illumina's $600 million acquisition of Solexa in November 2006 gave the company a head start in the next generation sequencing market.

This post is a brief overview of Solexa's sequencing-by-synthesis chemistry. The sample prep methods used differ slightly from that used in ABI's SOLiD system, but the basic goals are the same: generate large numbers of unique "polonies" (polymerase generated colonies) that can be simultaneously sequenced. These parallel reactions occur on the surface of a "flow cell" (basically a water-tight microscope slide) which provides a large surface area for many thousands of parallel chemical reactions.

Step 1: Sample Preparation

The DNA sample of interest is sheared to appropriate size (average ~800bp) using a compressed air device known as a nebulizer. The ends of the DNA are polished, and two unique adapters are ligated to the fragments. Ligated fragments of the size range of 150-200bp are isolated via gel extraction and amplified using limited cycles of PCR.

Complete detailed protocols for DNA and small RNA library preparation can be found in the documents provided in the attachments to this post. ("dna_libe_prep.pdf" and "rna_libe_small_prep.pdf", respectively). This process is a fairly straightforward multi-step molecular biology process, however there are many pitfalls that can result in skewed results downstream.

Steps 2-6: Cluster Generation by Bridge Amplification

In contrast to the 454 and ABI methods which use a bead-based emulsion PCR to generate "polonies", Illumina utilizes a unique "bridged" amplification reaction that occurs on the surface of the flow cell.

The flow cell surface is coated with single stranded oligonucleotides that correspond to the sequences of the adapters ligated during the sample preparation stage. Single-stranded, adapter-ligated fragments are bound to the surface of the flow cell exposed to reagents for polyermase-based extension. Priming occurs as the free/distal end of a ligated fragment "bridges" to a complementary oligo on the surface.

Repeated denaturation and extension results in localized amplification of single molecules in millions of unique locations across the flow cell surface. This process occurs in what is referred to as Illumina's "cluster station", an automated flow cell processor.




Steps 7-12: Sequencing by Synthesis

A flow cell containing millions of unique clusters is now loaded into the 1G sequencer for automated cycles of extension and imaging.

The first cycle of sequencing consists first of the incorporation of a single fluorescent nucleotide, followed by high resolution imaging of the entire flow cell. These images represent the data collected for the first base. Any signal above background identifies the physical location of a cluster (or polony), and the fluorescent emission identifies which of the four bases was incorporated at that position.

This cycle is repeated, one base at a time, generating a series of images each representing a single base extension at a specific cluster. Base calls are derived with an algorithm that identifies the emission color over time. At this time reports of useful Illumina reads range from 26-50 bases.




The use of physical location to identify unique reads is a critical concept for all next generation sequencing systems. The density of the reads and the ability to image them without interfering noise is vital to the throughput of a given instrument. Each platform has its own unique issues that determine this number, 454 is limited by the number of wells in their PicoTiterPlate, Illumina is limited by fragment length that can effectively "bridge", and all providers are limited by flow cell real estate.


Source from http://seqanswers.com/forums/showthread.php?t=21... credit to "ECO".