By Monica Heger
This story was originally published on March 28.
Next-gen sequencing could make it possible to predict from a blood draw whether a heart transplant recipient will reject a donated organ, according to researchers from Stanford University.
The researchers found that not only can next-gen sequencing detect the DNA from a donated organ in the recipient's blood, but can also detect spikes in the proportion of donor DNA, which is indicative of transplant rejection. These findings, published this week in the Proceedings of the National Academy of Sciences, suggest the technique could serve as an early, minimally-invasive indicator of organ transplant rejection.
The Stanford team developed an approach that uses next-gen sequencing on the Illumina Genome Analyzer to distinguish between donor and recipient DNA by analyzing differences in SNPs. It first requires that the recipient and donor be genotyped using a microarray approach. Then, after sequencing DNA from the recipient's blood, the researchers count the number of reads at SNP positions that differ between donor and recipient in order to determine the proportion of donor DNA.
They first tested their method on HapMap samples, mixing DNA from two samples at varying proportions of "donor" DNA, and found that as the percentage of donor DNA increased, so did the number of sequencing reads corresponding to donor DNA.
Next, they tested the method on transplant patients who had been monitored via biopsy. First, they genotyped both the donors and male recipients of three heart transplants, two of whom had rejection events and one who did not.
They then performed shotgun sequencing of recipient blood DNA on the Illumina GA at each biopsy time point, generating an average of 10 to 12 million uniquely aligning reads per sample.
At each time point, there were about 25,000 reads with SNPs that could be analyzed. Plotting the signal from donor SNPs against the matched recipient SNP background, the researchers found that the percentage of donor DNA tends to hover around 1 percent, but increased to up to 5 percent at the time of organ rejection.
"In patients who are not rejecting, there was a certain low level of donor genome that remained stable and at a low level, but when the patient was rejecting, there was a huge spike in the donor genome that came back down after the patient was treated," Hannah Valantine, a cardiologist at Stanford School of Medicine and an author of the paper, told Clinical Sequencing News. "This has tremendous implications."
Improving the Gold Standard
Currently, heart transplants are monitored by endomyocardial biopsy, which is considered the gold standard, although Valantine said it is hardly golden for patients.
These biopsies, which run conservatively around $5,000 per test, are hugely invasive — doctors surgically extract heart tissue through the patient's jugular vein — and patients have to undergo around 12 of them in the first year after receiving the transplant. "We really want to be able to reduce the numbers of heart biopsies or fully eliminate them," Valantine said.
Additionally, the biopsies typically do not provide an early diagnosis of rejection. Once a biopsy confirms rejection, there is already "intense myocytic damage," she said, and patients have to be put on an extremely high dose of immunosuppressant, which also has serious side effects.
In light of these issues, researchers have been looking for noninvasive ways to diagnose transplant rejection earlier.
In 2008, the US Food and Drug Administration approved the first noninvasive test for this application. The test, developed by XDx and called AlloMap, uses gene expression profiling of the patient's immune response to predict rejection.
Although the gene expression test has been shown to reduce the number of biopsies without resulting in an increase of severe cardiovascular events, it has a low positive predictive power, said Valantine. While it is good at predicting which patients will not experience a rejection, it can't accurately predict which patients are likely to have one.
[ pagebreak ]
Valantine, who also helped developed AlloMap, said that the results of the PNAS study indicate that the sequencing-based method is both sensitive and specific. It could be used in combination with the AlloMap test, she said, because it would be better at predicting which patients will experience a rejection and it may also be able to diagnose those patients earlier.
Eventually, she said, she wants to do a clinical trial comparing the effectiveness of using the sequencing-based test alone and in combination with AlloMap.
Thomas Snyder, the lead author of the paper and a research associate in Stephen Quake's bioengineering lab at Stanford, said that the team is now working on improving the method. While there is a clear spike in the proportion of donor DNA at the time of rejection, the key will be to determine whether these increases are significant enough to make an early diagnosis.
Snyder said the method could also be tweaked by either doing more sequencing or more genotyping. For the current study, the team genotyped about 1 million SNPs and sequenced to only about one-fold coverage. While the low-pass sequencing is still able to detect a sufficient number of SNPs, and keeps costs low, it is possible that deeper sequencing might yield results earlier or be helpful in difficult cases, Snyder said.
The method is similar to approaches developed at Quake's lab and elsewhere for noninvasive prenatal diagnostic tests that use next-gen sequencing to detect fetal DNA in maternal blood.
Dennis Lo, from the Chinese University of Hong Kong, who has been developing such a test for Down syndrome, said in an e-mail that the application of the approach to transplant rejection "is certainly a feasible method" and could be accurate in "differentiating the SNP alleles of the recipient and the donor."
In order to apply it clinically, though, he said that the false positive rate of 16 percent would have to be improved, and also the method would have to be tested in additional patient cohorts and for additional organs.
"It would be important to see if the diagnostic threshold would hold true for another
patient cohort, or indeed for different types of transplant," he wrote.
Snyder said that the group has secured funding from the National Institutes of Health to test the method prospectively in at least 100 patients.
According to the NIH grants database, the National Institute of Allergy and Infectious Disease awarded Quake and Valantine a three-year grant worth $2.5 million in the 2010 fiscal year for a project entitled "Genome Transplant Dynamics: Non-Invasive Sequencing-Based Diagnosis of Rejection."
The grant abstract notes that the prospective cohort study will involve 110 consecutive heart transplant recipients, and the researchers will collect blood samples during and between endomyocardial biopsy procedures to determine whether the donor-recipient DNA ratio "can detect rejection in its early stages and before the development of graft dysfunction."
The team will also apply the approach to lung transplant recipients, according to the abstract.
Diagnosing lung transplant rejection has proven to be even trickier than heart transplants, Snyder said, because tests such as AlloMap base their diagnosis on immune response and cannot distinguish between rejection and infection by a virus or bacterium. "In both cases the immune system will be activated," he said. The sequencing-based test, therefore, could be better since it is looking directly at the DNA of the transplanted organ, he said.
Stanford is also applying for a patent on the method and Snyder said that some potential partners have expressed interest in commercializing the technology. He noted, however, that anyone looking to commercialize the test for clinical use would have to demonstrate that it is cost-effective when compared to the current standard of care.
Both Snyder and Valantine said that while they could not pin an exact price on the test, it would be well below the cost of a biopsy, possibly on the order of 10-fold cheaper, mainly because they sequence to such low coverage.
Snyder said that right now, in a research setting, performing the sequencing probably costs around several hundred dollars.
Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.