Attention COVID-19 research community: the Mayo Clinic has made a discovery that could help explain why some people who get infected with the SARS-CoV-2 virus suffer badly while others feel only mildly ill. The hypothesis to be tested is whether an increase in protein expression controlled by ACE2 and TMPRESS2 genetic variants might result in elevated COVID-19 susceptibility and severity while a decrease might have a protective effect, according to Richard Weinshilboum, M.D., pharmacologist in the Center for Individualized Medicine at the Mayo Clinic.
The methodology took two years to develop, a task that fell to Lingxin Zhang, Ph.D., a researcher in the center’s Pharmacogenomics Program. The preparatory work involved tailoring cells such that an open reading frame (ORF) could be parked in individual cells and tagged so that changes in amino acids encoded by variants of uncertain significance would be expressed, a complex cellular engineering and bioinformatics exercise accomplished in collaboration with computational and DNA sequencing experts at the Mayo Clinic.
In a study that published recently in Human Molecular Genetics (DOI: 10.1093/hmg/ddac157), Zhang used high throughput sequencing and sorting of cells to determine exactly which variants were damaging and which were not—a technique she had previously applied to common pharmacogenetic genes such as CYP2C9, CYP2C19 and SLC01B1. The deep mutational scanning approach uses a “landing pad cell-based system” designed to analyze hundreds of gene ORF missense variants in a parallel and scalable fashion.
After the COVID-19 pandemic struck, this same platform was used to examine ACE2 and TMPRESS2, key players in SARS-CoV-2 entry into host cells. The information is now being made available to researchers with access to vast quantities of patient outcomes data to explore the clinical implications, says Weinshilboum.
For the past decade, pharmacogenomics has been integrated into the electronic health record (EHR) through Mayo’s Center for Individualized Medicine, which funded the genetics study. Doctors are sent alerts via the EHR when they write a prescription for any of more than 20 drugs whose effects have known genetic influences, Weinshilboum says, prompting them to consider genetic testing.
It has been known for some time that when thiopurine drug therapy is under consideration, for example to treat childhood leukemia, genetic testing is a good idea since the TPMT gene provides instructions for making the enzyme (thiopurine methyltransferase) that breaks down such drugs, says Weinshilboum. Otherwise, people who are genetically TPMT deficient could end up with a 10-fold higher concentration of a cytotoxic agent than intended. Tragically, some hospitalized children didn’t survive in the past because they were treated with a thiopurine drug without first testing for these common genetic variations, he adds.
Among the pharmacogenes studied in recent years by Zhang and Weinshilboum are SLCO1B1, which encodes a protein that can impact uptake of drugs such as statins and certain oral antidiabetic agents (Drug Metabolism and Disposition, DOI: 10.1124/dmd.120.000264). The pair have also used deep mutational scanning to identify severely damaging variants in the CYP2C9 and CYP2C19 genes that could result in decreased drug metabolism (Clinical and Translational Science, DOI: 10.1111/cts.12758).
“When it became clear that the virus responsible for COVID-19 has to attach to a cell and internalize, and that these two proteins [ACE2 and TMPRESS2] play a critical role in that process... Dr. Zhang and I decided to go back and look at the less frequent variants and variants of uncertain significance in the open reading frames of these genes and functionally characterize them,” Weinshilboum says.
Deep mutational scanning was used to study the functional implications of 433 missense variants of the two genes by analyzing abundance of the encoded protein. For ACE2, Zhang reports, the technique identified eight of 127 variants that displaced less than 25% of the wild-type protein expression. For TMPRESS2, 19 of the 157 isoform 1 variants and 13 of 149 isoform 2 variants did so. Four ACE2 variants displayed a 25% or greater elevation of protein expression.
Zhang used color coding to map the relative stability of the protein variants expressed by the engineered cells. Billions of data points were analyzed using a series of technologies, including cell sorting, modern genomics, high throughput DNA sequencing, and a computer algorithm.
The methodology could be implemented to study many additional genes involved in COVID-19 beyond ACE2 and TMPRSS2, says Zhang. Her hope is that “scientists and clinicians across the world... use this information to help their patients.”
An international consortium with sequencing and clinical data on many patients is expected to “very quickly” test the resulting hypothesis that host cell variations are risk factors for COVID-19 infection and disease severity as well as response to therapy, says Weinshilboum. “Common variants are relatively easy to deal with and important whether they inactivate a protein or give you more of it than most people have.”
If it proves valuable to look at the host and not just the sequence of the constantly mutating virus, the Mayo Clinic will be among the first to put that approach into real-world practice, Weinshilboum says. “We went out of our way to be sure that the variants of uncertain significance we were dealing with were not just from European patients,” he also notes.
The study tapped exome sequence data for 70,000 subjects contained in the Genome Aggregation Database (gnomAD), 30,000 of which were from minority populations, says Weinshilboum. These were predominantly African American and Hispanic individuals, as data on large numbers of East Asians is not yet publicly available.
Forthcoming outcomes studies using the newfound information about the ACE2 and TMPRESS2 genetic variants might reveal who is at unique risk for COVID-19 and should therefore take special precautions (e.g., more frequent vaccination) and receive immunomodulating drugs if they become infected, he says. The focus would initially be on adults but could also rapidly shift to the pediatric side.