May 15, 2024 | The lab of W. Andy Tao, professor of biochemistry at Purdue University, is pioneering a way to assess drug response based on metabolizing proteins found in circulating extracellular vesicles (EVs), which could serve as a surrogate of their expression level in the liver and become therapeutic targets for some cancers. Their detection technique involves mass spectrometry measuring subtle variations in the protein composition of tested samples as well as a commercially available EV total recovery and purification (EVtrap) technology that Tao co-invented.
The EV field is at once “interesting and messy,” he says, because the particles and their cargos have emerged as promising biomarkers of many diseases but since they are released from almost all types of cells—including those in the brain—their source is difficult to determine. Immature isolation methods have additionally made it hard to know exactly what’s being captured.
Tao and his colleagues are looking to change all that with their quantitative proteomics strategy, which was most recently used to measure changes in metabolic enzyme activities related to drug exposure. In a study published recently in the Proceedings of the National Academy of Sciences Nexus (DOI: 10.1093/pnasnexus/pgae023), they demonstrated that changes in abundances of these EV-housed enzymes paralleled those in hepatic cells isolated from liver tissue and simultaneously analyzed 34 proteins involved in drug absorption, distribution, metabolism, and excretion (ADME) pathways.
Importantly, 16 of the 34 were P450 enzymes known to play pivotal roles in the detoxification of most drugs. If P450 and other metabolic enzymes in EVs are detectable and measurable, they could be used to evaluate immediate drug response through liquid biopsy, Tao says, pointing to the prospect of noninvasive sampling and phenotypic drug ADME monitoring for personalized medicine.
The focus on EVs effectively avoids the technical challenges in measuring ADME proteins in biofluids to evaluate drug efficiency, explains Tao. “Samples usually have a high number of other [irrelevant] proteins and are also potentially unstable.” At his lab, EVs retrieved from samples have been successfully stored in the refrigerator for several years, he notes.
Prior attempts have been made to characterize the variability in drug exposure using EVs, Tao adds, but the techniques couldn’t be easily applied to clinical samples. For example, Pfizer collaborated on a 2019 study (British Journal of Clinical Pharmacology, DOI: 10.1111/bcp.13793) by isolating P450 protein bands on an SDS-PAGE (sodium dodecyl-sulfate polyacrylamide gel electrophoresis) gel to enrich ADME proteins for targeted measurements, which is impractical for clinical applications.
Although the dangers of polypharmacy are well recognized, especially among older adults, the U.S. Food and Drug Administration does not require pharmaceutical companies to conduct drug-drug interaction studies because of the shortage of discernable biomarkers, say Tao. It’s a potentially serious issue currently left to pharmacists to remedy based largely on predicted interactions identified in small-scale clinical studies.
The latest investigation by Tao and his team showed how two drugs—the antibiotic rifampicin and the leukemia chemotherapy drug dasatinib—elevated the level of one P450 enzyme in the plasma EVs of mice. The enzyme corresponds to CYP3A4, the P450 enzyme that metabolizes 30% to 40% of all drugs currently on the market for human consumption.
Taking two or more drugs metabolized by the same enzyme raises the risk of lowering the effectiveness or heightening the toxicity on any one medicine, he continues. Even if companies wanted to look at these unintended consequences, which they may inevitably be forced to do, it could open a Pandora’s box not to mention require a lot of data to address properly.
This all has consequences for patients, including those with cancer, where the therapeutic time window for treatment with the right medicines can be short, says Tao. Drugs also get differently metabolized based on age and genetic factors as well as comorbid conditions being treated, making them significant considerations under a personalized medicine approach. In the future, a simple blood test could quickly say whether an individual’s response to a drug matches the clinical intent.
Last year, Tao’s group teamed up with researchers at Columbia University on a study where EV-encapsulated proteins from urinary EVs were used for the detection of Parkinson’s disease (Nature, DOI: 10.1038/s43856-023-00294-w). The study was part of efforts by Tymora Analytical Operations, a spinoff from Tao’s lab, to identify a biomarker signature for Parkinson’s disease using the company’s EVtrap technology.
The proteins being identified here were leucine-rich repeat kinase 2 (LRRK2) proteins associated with Parkinson’s disease and their altered downstream pathways. The methodology has a long list of potential applications for other neurodegenerative conditions and cancers since EVs carry proteins repeatedly shown to be disease-relevant, says Tao.
When P450 proteins are the target of drug development efforts, companies generally feed the experimental agent to animals then take out their liver and isolate microsomes containing the enzymes to measure efficacy. As Tao’s group has now shown, EVs in the blood could well serve as a surrogate for those in tissue.
A key challenge has been narrowing the search to the small proportion of EVs coming from the liver, he says. In a move toward that end, researchers took blood from mice three and again seven days after treatment with rifampicin or dasatinib to show they could observe and detect P450 enzymes.
EVtrap is quite efficient at isolating specific types of EVs, such as the exosomes involved in intercellular communication and molecular functions linked to cancer, adds Tao. The field is nonetheless stuck in the “wild west” stage because EVs are by nature highly heterogenous with no universal method for isolating the nanoparticles. This helps explain why EVs have acquired at least 10 different names, including ectosomes, microvesicles, microparticles, exosomes, oncosomes, and apoptotic bodies.
One goal of Tymora Analytical Operations is to continuously develop EVtrap so that the protein composition in a biosample might one day be depicted in its totality, Tao says.
The ability to analyze 34 enzymes involved in the ADME process, in parallel, represents a significant development in proteomics, says Tao. Researchers are more typically looking at no more than a handful of proteins at a time via immunoassays.
Forty-four ADME proteins are known to exist, and hundreds of other proteins are thought to play a role in various processes of drug disposition, so there’s plenty of room for improvement, he says. The emerging era of artificial intelligence is expected to aid in the quest for more precise measurement.
Tao and his colleagues have already started to look at the current 34-protein panel in mice given rifampicin, dasatinib, or the two in combination, and over a longer period to monitor their response at different treatment time points. They also plan to move their research to humans by collaborating with clinicians whose patients are already taking the drugs, even as they work on the technical side on isolation of the liver specific EVs.
The study of EVs can be somewhat of a “trap” from a funding standpoint, he jests, because the field still has a lot of question marks despite its promising future. Tao says his big hope is that the pharma industry recognizes the potential of the technology in their drug development efforts and ventures into the realm of drug-drug interactions that can affect the action of individual medicines and, potentially, send people to the emergency room.