NEW YORK – Researchers from the University of Kentucky and their collaborators have used nanopore long-read sequencing to quantify and characterize RNA isoforms in brain samples from Alzheimer's disease patients.
Published in Nature Biotechnology last month, the study, which identified hundreds of new RNA isoforms with high confidence, including those generated by medically relevant genes, offers researchers a window to study their possible association with the etiology of brain-related diseases.
Stemming from the same gene but with different sequences, RNA isoforms enable a gene to diversify its protein-coding capabilities, tapping mechanisms such as alternative splicing and intron retention.
The purpose of this study was to demonstrate the value of long-read sequencing for Alzheimer's disease when it comes to RNA isoform analysis, said Mark Ebbert, a professor at the University of Kentucky and the paper's corresponding author.
Because of the inherent limitations of short-read sequencing, Ebbert explained, the technology practically cannot accurately distinguish the different isoforms from the same genes. In contrast, long-read sequencing has the potential to profile the entire RNA isoforms all at once, offering researchers an avenue to directly study these molecules, he added.
For their study, the University of Kentucky researchers applied deep long-read sequencing on 12 aged, frozen postmortem human frontal cortex brain samples, including six Alzheimer's disease cases and six cognitively unimpaired controls. All samples were collected from individuals of European descent and were split between males and females in both the disease and control groups.
After RNA extraction and sample processing, they underwent sequencing on the Oxford Nanopore Technologies PromethIon device using one flow cell per sample, generating a median of 35.5 million aligned reads per sample. "The main thing is to get ultra-clean RNA in the first place," which required a lot of optimizations, said Jason Brandon, a researcher at Ebbert's lab and a co-first author of the study.
Afterward, the sequencing data were analyzed using previously described software called Bambu for RNA isoform identification and quantification.
Overall, the investigators looked at 5,035 medically relevant genes implicated in brain-related diseases. Of those, they found 1,917 expressed multiple isoforms and 1,018 generated isoforms with varying protein-coding sequences in a single tissue, illustrating the diversity of isoforms within the medically relevant genes. Additionally, of the 1,018 genes with different isoforms, 57 are linked to neurological disorders including major depression, schizophrenia, Parkinson’s disease, and Alzheimer's disease, the authors noted.
The researchers also sought to identify and quantify new RNA isoforms expressed in the human frontal cortex. They reported 428 new "high-confidence" isoforms in the tissue, 53 of which originated from 49 medically relevant genes. "We are trying to be very conservative for reporting [the isoforms]," said Ebbert, noting that the team set a stringent cutoff during analysis to prevent false-positive results.
Additionally, the study discovered five mitochondrially encoded RNA isoforms, demonstrating alternative splicing events in mitochondrial RNA. Of these, four isoforms span the MT-RNR2 transcript, which encodes the mitochondrial 16S rRNA, while the fifth one spans the MT-ND1 and MT-ND2 genes.
In addition to the known genes, the study also investigated RNA isoforms from the so-called new gene bodies — parts of the genome where transcription has not been annotated — and found 267 new high-confidence isoforms from these regions.
"If you don't know if something's there, you really can't measure it and know if it is affecting diseases," said Bernardo Aguzzoli Heberle, a Ph.D. student in Ebbert’s lab and another co-first author of the study. "By revealing these things, we can [start to] tell if they are playing a role in disease."
Furthermore, the researchers performed differential isoform expression analyses on the samples, revealing 99 differentially expressed RNA isoforms between the Alzheimer's disease cases and controls, even though the genes behind these isoforms were not differentially expressed.
While the dataset is "not large enough to draw firm disease-specific conclusions," the study authors noted, the data does cast light on the potential biological function of RNA isoforms for disease etiology and calls for larger studies and further validations.
"This is one of the very few studies that compared patient and control samples to show the difference in the isoforms," said Kin Fai Au, a professor of computational medicine and bioinformatics at the University of Michigan who was not involved in the study. Au noted that while previous studies have explored RNA isoforms in lab samples, such as cell lines or mouse models, this study adds great value to the field by studying real human disease tissues.
Additionally, Au highlighted the team's efforts to sequence the samples at high depth while setting stringent parameters for calling new RNA isoforms to ensure their validity. "While in the basic study, we want a large number of discoveries, in some clinical or biomedical research applications, you want a shortlist of very reliable targets," Au pointed out. "That's another important message sent by this paper to the community."
Ebbert said this most recent proof-of-principle study laid a foundation for a larger study that is currently underway to further validate the results in a larger cohort of 300 samples, splitting between the Alzheimer’s and control groups.
In addition, given that the current study used the older R9.4.1 flow cells from Oxford Nanopore, Ebbert said the group will also incorporate the newer R10 flow cell and Kit 14 chemistry for the new experiments.
Ultimately, Ebbert noted the team aims to harness isoforms as reliable biomarkers for the early diagnosis and treatment of human diseases.
"We want to legitimately identify RNA isoforms that are, at the very least, correlated with [Alzheimer's] disease that can be used as biomarkers, and, hopefully, towards developing a presymptomatic disease diagnostic," he said.