NEW YORK – A new image-based spatial transcriptomics method may be able to better predict heart transplant outcomes by providing more mechanistic insights into acute allograft rejection types.
The method, presented on Monday at the annual Precision Medicine Tri-Con meeting in San Diego and described online in a preprint in BioRxiv, demonstrated distinct cell-specific gene expression patterns in response to different therapies and associated with specific short- and long-term rejection-related outcomes in heart transplant recipients.
Researchers affiliated with Vanderbilt University and the Translational Genomics Research Institute (TGen) and led by Nicholas Banovich, director of the division of bio-innovation and genome sciences at TGen, used the method to examine longitudinal heart biopsies from 13 pediatric and 49 adult transplant recipients.
Patients in this cohort experienced multiple rejection subtypes, comprising acute cellular rejection (ACR), antibody-mediated rejection (AMR), and mixed rejection.
The team observed that distinct patterns of cellular organization accompanied different types of rejection. In both ACR and AMR samples, for instance, T-cell subsets tended to cluster together, although in ACR, this clustering also tended to involve plasmacytoid dendritic cells and myeloid dendritic cells, whereas AMR samples featured more colocalization of macrophages and fibroblasts.
Using the Xenium high-resolution single-cell spatial platform from 10x Genomics with a panel of 477 genes, the investigators examined patient biopsies across the grading range of zero, for no rejection, to three, for severe rejection. They found that biopsies from grades zero, one, and two tended to be very transcriptionally similar to one another and distinct from grade three biopsies.
"We were kind of hoping for the opposite," Banovich said in his conference presentation, "which would be that even a grade-zero biopsy from a grade-three patient would allow us to molecularly characterize that patient as being grade three. But that wasn't the case."
He and his colleagues next examined whether pretreatment biopsies contained molecular information that could be used to predict treatment responses and outcomes. They observed that patients with more activated immune cells pretreatment responded better to immuno-modulating drugs, suggesting that the effects they were seeing might have diagnostic and clinical utility.
"The differences in the pretreatment [molecular] states suggested that molecular information exists to tell us if a patient will respond to a given treatment or not," Banovich said.
Among 32 individuals who experienced ACR, for example, 15 responded to anti-rejection therapies. From these individuals, the team identified 16 unique differentially expressed genes across seven cell types, none of which overlapped with genes differentially expressed in non-responders. From these data, the team was able to identify differences in pretreatment gene expression profiles that might mediate therapy response.
In addition to acute rejection subtypes, the researchers also examined cardiac allograft vasculopathy (CAV) samples. CAV is characterized by the vasculature of the heart shrinking and thickening, and acute rejection of any type is a major risk factor for its development. Few therapeutic options exist to manage this outcome.
As with the other rejection types, CAV samples showed cell-type specific gene expression patterns characteristic of different aspects of CAV pathogenesis, such as angiogenesis and vascular inflammation.
Organ transplant rejection risk is currently assessed via direct examination of tissue biopsies by a pathologist, which is considered the gold standard, and to some degree by liquid biopsies, which are becoming more established in clinical settings. While a key advantage of liquid biopsies is their minimal invasiveness and relative ease of use, Banovich noted in his talk that the main advantage of this spatial transcriptomic method is the ability to more precisely match biomarkers to their respective pathologies.
While liquid biopsies continue to improve and expand, Banovich said that they still comprise a "pseudo-bulk" approach that might miss certain features that are specific to the actual tissue remodeling observed in different types of acute rejection.
Banovich and his team are now preparing studies in larger cohorts. For the time being, he plans to continue focusing on the heart, although he mentioned that conceptually, this methodology can be broadly applied to other organs as well.
Spatial transcriptomics is a growing field with numerous companies developing tools to apply this technology across the research and translational landscape. Illumina, for instance, recently unveiled its own spatial platform and is working with the Broad Institute to demonstrate its potential for producing large-scale spatial transcriptome datasets.
Other players in this space include Bruker, Singular Genomics, and BGI subsidiary STOmics, some of whom have recently announced new spatial biology offerings.
Although Banovich acknowledged the commercial potential of this method, he said that such considerations remain a long way off.
"We are starting to think about [how] we could use these types of data to predict outcomes and [whether] there's additional immunomodulation that you could do on the front end before they ultimately end up with this phenotype," Banovich said.
The TGen/Vanderbilt study was funded by the International Society for Heart and Lung Transplantation and Enduring Hearts.