By Kevin Davies
March 8, 2012 | SAN DIEGO—Two leaders in the clinical application of whole genome sequencing offered further signs of progress in a pair of keynote talks at the 2012 XGen Congress*.
Rick Wilson, director of The Genome Institute (TGI) at Washington University, St Louis, and Elizabeth Worthey, bioinformatician at the Medical College of Wisconsin (MCW), shared new findings on the sequencing of cancer and pediatric cases. Although encouraging—and in one unpublished case potentially life-saving—Worthey in particular expressed some strong cautionary notes that are hampering current efforts in ending diagnostic odysseys.
Wilson described the sequencing of a cancer patient conducted in St Louis last August that resulted in a potentially life-saving outcome. The male patient was initially diagnosed with a form of acute leukemia (B-ALL) at age 25 and received standard therapy. But after a relapse five years later, he received a bone marrow transplant from his younger brother.
Last July, the patient experienced another severe relapse, including involvement of his central nervous system. “He was not expected to survive past Christmas,” said Wilson.
Last August, Wilson’s team performed whole-genome, exome and transcriptome sequencing. A couple of deletions were noted, although nothing appeared particularly actionable at first. While no mutations in the FLT3 gene were observed, the gene turned out to be massively over-expressed (about 800-fold). This made the patient a prime candidate for treatment with the oral drug Sunitinib (Sutent), a drug used in cases of over-active FLT3 in acute myeloid leukemia (AML). Wilson’s colleagues achieved complete clinical remission in just 12 days, enabling a stem cell transplant.
“When do we stop Sunitinib?” Wilson asked rhetorically. That’s hard to say, because this was likely the first B-ALL patient ever treated with this off-label drug. The drug would now be considered for all B-ALL patients treated at Wash U.
Liz Worthey opened her presentation by reviewing the well publicized case of Nicholas Volker, a young patient with a severe inflammatory intestinal disorder (see, “Seize the New: Laying the Clinical Genomics Pipeline in Wisconsin”). Exome sequencing led to a positive diagnosis and a bone marrow transplant. Nicholas is back at school and there has been no recurrence of the disease. Interestingly, many previous children with the same original diagnosis as Nicholas (XLP2, susceptibility to viral infections) turn out to have severe colitis. “Associations will come along after the fact,” said Worthey.
To date, 13 pediatric patients and their families have gone through the MCW sequencing process. Only one family elected not to receive any results. Given the choice, 10 of the 12 families wanted to learn about the risk of non-actionable childhood disorders; the same number requested any results on actionable adult disorders as well.
Worthey and colleagues have built a host of tools to manage clinical genome data, beginning with Carpe Novo, a clinical variant analysis platform, which has been approved by CAP (College of American Pathologists). “The goal is we don’t want bioinformaticians to do the analysis. It must be clinician friendly,” said Worthey.
The informatics suite includes Valcrie—a variant repository; GapMine—to identify genes with insufficient depth of coverage that may harbor unrecognized mutations (“You can’t identify the mutation if you didn’t sequence that region!”); GATool for genome annotation (a “one-stop shop” for functional information) with a report that can go to the clinical geneticist; DiseasePortal and TextMining.
As Worthey stressed, there are numerous hurdles to reaching a genome-based diagnosis. Nomenclature typos between major databases such as HGMD and NCBI; widely varying allele frequencies from different sources; confounding pseudogenes present in more than 1,200 disease-associated genes in OMIM; and many variants that appear damaging are in fact benign polymorphisms. “In whole-genome sequencing, no-one can know all the data,” said Worthey.
Of the 13 clinical genomes sequenced at MCW, Worthey and colleagues have made three positive diagnoses. The other ten cases remain under analysis, she said. One positive diagnosis revealed mutations in a known gene (TWINKLE) that indicated a costly liver transplant would have been futile. The child died at age 6 months, but spent its remaining time with the mother in comfort.
Another case revealed a missense mutation in a mitochondrial gene, which has allowed other family members to be tested. In a third case, a 4-year-old female presented with repeated infections, poor growth and diarrhea. Worthey said her team had found several variants of unknown significance, any of which could be the disease-causing mutation.
In the case of a 10-year-old female with leukodystrophy, ataxia, and cerebellar atrophy, Worthey’s analysis presented two putatively pathogenic variants in a pair of genes, but neither one is associated with the primary phenotype.
“We haven’t and won’t always find the mutation,” said Worthey. It was vital that patients and families be made aware of that possibility. Indeed, Worthey said the anecdotal success rate based on findings from several institutes was only about 1 in 20 so far.
On a brighter note, she said, “insurance companies are agreeing to pay for whole-genome sequencing.” Four of her group’s sequencing cases had been covered so far. “This will likely be [the insurance companies’] initial guidance based on cost,” she said, predicting that wouldn’t take very long before implementation.
“We are fundamentally changing the practice of medicine,” Worthey concluded, but there is a major need for genomics education for all types of medical providers. Work was also needed to clean up genomics data for clinical use as well as provide new methodologies for the appropriate use of reference genomes.
*CHI’s XGen Congress 2012: San Diego, March 5-8, 2012.