Scientists at Roche Molecular Systems have been exploring using the 454 Life Sciences GS FLX for high-resolution HLA DNA typing in research, an application that 454 said it believes could become one of the first clinical applications of the platform.
454 says that some of its customers outside the US are close to using the 454 platform for clinical HLA typing. In addition, the company is working with the FDA to find ways to provide the system to US testing labs, some of which doubt that the platform has become robust and cheap enough for such use.
HLA genotyping plays a role in a number of clinical areas — notably in bone-marrow transplants, in which high-resolution HLA typing is used to match transplant donors with recipients.
Specific HLA genotypes are also associated with susceptibility to diseases, notably autoimmune diseases, and have been linked to certain allergic drug responses.
Henry Erlich, vice president of discovery research and director of the department of human genetics at Roche Molecular Systems in Pleasanton, Calif., and his colleagues have explored the GS FLX internally for HLA typing. Erlich presented results from the research at the Advances in Genome Biology and Technology meeting on Marco Island, Fla., last month.
"At the moment, it's a research test and process," he told In Sequence last week. "But we expect it eventually can be used in clinical diagnosis."
He said the 454 platform has the potential to solve a problem with HLA typing called "phase ambiguity," or phasing polymorphisms in HLA genes, which are highly polymorphic.
"If an individual is heterozygous at many positions that are very close to each other, you are not sure how to phase those polymorphisms, that is, which polymorphic sequences go together on one allele and which go together on the other allele," he explained. "That's a very fundamental problem" that occurs with any of the existing HLA-typing methods, including Sanger sequencing, he added.
Resolving these phase ambiguities "takes a lot of work," and is "cumbersome," according to Erlich.
The 454 platform, because it generates sequence information derived from a clonally amplified single molecule, rather than many molecules at once, allows scientists to determine unambiguous HLA sequences, at least within overlapping sequence reads.
"What we are doing is only determin[ing] the unambiguous sequence of each exon," Erlich explained. "But it turns out that the pattern of polymorphisms in HLA is mainly generated by mutation and gene conversion in the exons, so if you have an unambiguous sequence of exons, and you compare it to an HLA sequencing database, you can usually get a unique genotype assignment."
In addition, he said, researchers can use the 454 platform to lower the cost per sample, by sequencing multiple samples in the same run, by partitioning the plate, and by adding indexing tags to each sample.
So far, Erlich and his colleagues have sequenced eight HLA loci in up to 48 individuals in a single run, using 12 different tags. "In principle, you might be able to do over 100 individuals in a single run," Erlich said. "That would bring the cost for each individual sample down considerably."
For the data analysis, the Roche group has teamed up with Conexio Genomics, an Australian software company that sells HLA-genotyping software for Sanger sequencing. Conexio has developed software that analyzes 454 data and compares it to data in HLA-sequence databases to assign a genotype.
454 Life Sciences, for its part, believes that its platform will be used for clinical HLA typing soon. "At least outside the US, we are seeing customers of ours who, if they are not yet past the research stage, are very close to potentially be using this for [laboratory-developed tests]," 454 CEO Chris McLeod told In Sequence last week.
[ pagebreak ]
Last summer, for example, 454 said that the Blood Center Linz, Austria, had purchased a GS FLX for research into rejection reactions in bone marrow and stem cell transplantation. At the time, the center's medical director, Christian Gabriel, told In Sequence that although using the platform is not cheaper than existing ones, it will likely save time and deliver higher-resolution results.
“Our idea is to [develop] this into some sort of method which we can use for patient care,” he said (see In Sequence 9/2/2008)
Gabriel did not respond to a request last week about how far since then his lab has developed HLA typing on the 454 platform.
HLA in the USA
Meantime, in the US, 454 is working with the Food and Drug Administration to determine how the platform could be used by the CLIA-certified high-complexity labs that currently provide most of the HLA typing in the country.
"What we want to do is work with the FDA to define that path to bring [the HLA typing application] to market quickest here," McLeod said. "But until we have an agreement with them on how to do that, we don't want to talk about when that might be available."
But some US HLA-typing labs believe that the technology is not mature enough for clinical use, and does not improve much on Sanger sequence-based typing, or SBT.
According to Nezih Cereb, president of HLA-typing firm HistoGenetics, the sample prep for the 454 platform is not robust enough at the moment, and the cost per run is too high to make the technology interesting for his lab. HistoGenetics, based in Ossining, NY, types up to 360,000 samples per year on 13 Applied Biosystems 3730xl capillary sequencers, making it one of the largest HLA-typing laboratories in the world.
"For our situation … this is still not a practical approach and it's not gong to replace the capillary sequencing anytime soon," Cereb told In Sequence last week.
In particular, he said, the GS FLX's emulsion PCR is not "robust enough to give you consistent results from different kinds of samples," such as DNA from buccal swabs and blood. "We don't have the luxury to quantify the DNA and we don't have the luxury of optimizing emulsion PCR for every sample," Cereb added.
The high cost per 454 run would also require a large number of samples to be run in parallel to make the process cost-effective, he said. Also, his lab is able to resolve phase ambiguities in most cases using existing technology.
Other labs have similar reservations. "The (relative) complexity for sample preparation and the costs, in my opinion, prevent the technology from being employed in both large-throughput laboratories and small-throughput laboratories at this point in time," Daniel Geraghty, a member in the clinical research division of the Seattle-based Fred Hutchinson Cancer Research Center, said in an e-mail message.
His lab has performed sequence-based HLA typing for the HIV Vaccine Trial Network and tested 3,000 samples last year by Sanger sequencing. Geraghty also has "limited" experience with the 454 system through collaborations.
According to Geraghty, currently 454's sample-prep method is complex and difficult to automate, whereas "essentially all of the HLA SBT tests can be pretty much automated, with a lower level of technical expertise required to successfully employ them."
In addition, "the cost of a 454 run requires large quantities of individual samples be run simultaneously to be cost-effective relative to SBT," he added. "Few of the typical HLA-typing laboratories currently have a need that crosses that threshold, although that could change."
[ pagebreak ]
Though Geraghty acknowledged that the 454 sequencer can resolve phase ambiguities, he said that "the short read lengths on 454 runs do impose limits on phasing data from different exons, and HLA [sequence-based typing] has effectively solved this problem, given the use of appropriate, and available, analysis software.
"The limitations of 454 will change with improved read length, with simplified protocols for sample preparation, and of course with decreasing costs for sample runs," he added.
A 454 spokesperson told In Sequence that the company is addressing these complaints. “We are actively pursuing projects to simplify the sample prep workflow and have identified key steps in the process for integration of automation,” including an automation module for the emulsion PCR process, according to the spokesperson.
With regard to cost, she said, “we estimate that a test using 454 sequencing will be five- to six-fold cheaper in consumable cost than the current CE-based test. Considering the fact that the average HLA lab runs 50 samples a week and 200 a month, the use of 454 sequencing is a cost-effective solution.”
Finally, resolving phase within exons “can be completed much faster and cheaper with 454 sequencing than Sanger sequencing,” she said.