This article has been updated to correct Stephen Quake's title at the Chan Zuckerberg Biohub Network.
BALTIMORE – A new blood-based cell-free DNA methylation sequencing assay developed by researchers at Cornell University and their collaborators can detect or predict major complications of allogeneic hematopoietic cell transplantation (HCT) early on, promising to improve the care of stem cell transplant patients.
The developers have filed patents related to the technology, which Cornell is in the process of licensing to an undisclosed company.
In a proof-of-concept study published last week in the Proceedings of the National Academy of Sciences, Iwijn De Vlaminck, associate professor of biomedical engineering at Cornell, and his team described an assay that uses bisulfite sequencing of cfDNA from blood to simultaneously scan for common complications after allogeneic HCT: graft-versus-host disease (GVHD), infection, graft failure, and disease relapse.
Previously, De Vlaminck was involved in studies demonstrating cfDNA as an indicator of heart, lung, and kidney transplant rejection. The underlying principle for these studies is that when rejection happens, the cells and tissues of the transplanted organ are injured, shedding DNA into the bloodstream. Thus, by quantifying the amount of donor DNA, researchers can infer transplant rejection.
"That is basically what inspired the current study," said De Vlaminck. "We've been thinking about ways in which we could develop a blood test for rejection [in allogeneic HCT]."
Unlike in solid organ transplantation, rejection after stem cell transplantation, or GVHD, can happen anywhere in the recipient’s body, potentially affecting any tissue type, De Vlaminck said. He added that allogeneic HCT patients are typically "deeply immunosuppressed," and therefore often deal with numerous other complications, such as infection or graft failure.
To detect GVHD, the team performed methylation profiling of cfDNA from blood using shallow bisulfite whole-genome sequencing. "This allowed us to obtain information about what tissue types are contributing to the [cfDNA] because different tissues have different methylation profiles," De Vlaminck noted.
Simultaneously, with metagenomic cfDNA profiling, analysis of tumor-specific chromosomal aberrations, and quantifying the donor and recipient-specific cfDNA proportion, the scientists were able to infer infection, cancer relapse, and graft failure, respectively, using the same sequencing data.
For their study, they tested the assay on 170 blood samples collected from 27 allogeneic HCT recipients at Dana-Farber Cancer Institute at multiple time points before and after transplantation. They found that cfDNA derived from solid tissue could help predict the future onset of GVHD as early as one month after HCT. They also discovered the frequent presence of cfDNA from anelloviruses, cytomegalovirus, herpesvirus 6, Epstein-Barr virus, and polyomavirus in the samples. Moreover, the assay was effective in predicting graft failure and disease relapse post transplantation.
The study sequenced the cfDNA samples on Illumina sequencers, but De Vlaminck said the method "would be compatible with any sequencing platform."
The paper is "terrific," said Stephen Quake, professor of bioengineering at Stanford University and president of the Chan Zuckerberg Biohub Network. "It's one thing to cut out your heart and lung and put another one in," he said. "It's another thing to try to bleach your immune system and give you a new one." This paper, he noted, helps demonstrate that "liquid biopsy has an enormous amount to offer to people who are having stem cell transplants."
Before joining Cornell, De Vlaminck was a postdoc in Quake’s lab at Stanford, where he spearheaded cfDNA assays to predict solid organ transplant rejection.
In 2014, CareDx licensed a cfDNA method invented by Quake's lab to detect heart transplant rejection. The same year, CareDx acquired Quake’s transplant diagnostics startup ImmuMetrix.
Last December, CareDx published two studies, one on the clinical validation for its AlloMap Kidney gene expression test for kidney transplant rejection and another on the long-term performance results for AlloSure Kidney, its donor-derived cfDNA assay for post-transplant rejection.
In addition to CareDx, Natera has also been offering cfDNA-based assays, called Prospera, to detect rejection in kidney, heart, and lung transplants. Another company, Transplant Genomics, part of Eurofins Scientific, has been offering cfDNA testing kits for kidney and liver transplant rejections, as well.
Nelson Chao, chief of hematologic malignancies and the cellular therapy division at Duke University School of Medicine, said the concept behind this study is "quite remarkable." In current clinical practice, he said, diagnosis of GVHD relies solely on clinical features. For instance, to examine GVHD of the gut, an endoscopy along with an upper GI or lower GI biopsy could be required. "Obviously, that's quite intrusive," Chao said. To confirm infection, current clinical methods use PCR or culturing. While PCR is relatively fast, culturing for most bacteria takes 48 to 72 hours, said Chao. Also, in most cases, the way to find out relapse "is just time."
However, the results of the study need to be confirmed in a larger cohort, he said. "The key question is what happens with large datasets."
De Vlaminck said his team has recruited "many more patients" since the pilot, adding that "one of the things we plan to do is to look at more samples and do more to analyze the robustness of the approach."
Another limitation of the assay could be a lack of reference methylation profiles. The cfDNA methylation profiles in the study were compared against a reference set of methylation profiles from specific cell and tissue types. "At the tissue level, I’d say [the reference methylation profiles] have been pretty well saturated," said Stanford’s Quake. "Whereas at the cell type level, not quite as much."
De Vlaminck concurred, adding that "the better the references, the better the analysis we can put forward."
Additionally, since the assay uses DNA-based sequencing, it is also insensitive to RNA viruses, as well as to microbes that lack a reference genome.
The assay currently costs "a few hundred dollars" per sample, De Vlaminck said, much of which is related to the cost of DNA sequencing.
Moving forward, he said his team is planning to further validate the assay so it can quantify the degree of involvement of different tissues when it comes to GVHD, "for which we only have circumstantial evidence at this stage." Additionally, he hopes to continue optimizing the method by "developing further bioinformatics" and by further eliminating environmental DNA contamination during the workflow.
"There's a lot of work between here and there," Quake said, alluding to the potential commercialization of the assay. But the study "convinces people that … we should do a really large clinical trial to validate [the assay] and apply it as a real diagnostic for people."