(2lf) Quantifying Exosomes, Supermeres and Lipoproteins Subfractions for Early Detection of Cancer and Cardiovascular Disease

Cancer and cardiovascular diseases account for roughly 50% of all global deaths and can often be prevented, cured, or managed with early detection and treatment. The challenge associated with early detection lies in the invasiveness of the detection method, the highly sophisticated required machinery and overall, the high cost of the test. Still, even with the state-of-the-art facility, we cannot detect cancer in its initial stages since most of them rely on looking at methods such as tissue staining, performing biopsies and RNA sequencing, mammography, colonoscopy or the highly inconclusive MRI of the tumor cells – all requiring information on the location of the tumor which is unknown for the undiagnosed patients. This also has long-term side effects – drilling into a healthy patient's brain or spinal cord to check for glioblastoma will lead to a long-term damage. Similar complexity goes for cardiovascular diseases where plaques are difficult to see in their early stages, while cholesterol tests fail to accurately predict cardiovascular risk (~65% true positive and true negative rate). Lipid-protein-nucleic acid assemblies (exosomes, supermeres, lipoprotein, etc.) in the blood can allow us to overcome these issues since their cargo (proteins, miRNA, mRNA, etc.) share the signature of their parent cell and provides us with a window to non-invasively look at tissues from different parts of the body using a routine venous blood draw. For example, exosomes rich with neuronal proteins such as Enolase 2 (ENO2) and L1CAM can allow us to non-invasively investigate the brain for tumors or signs of neurodegeneration since these particles can cross the Blood Brain Barrier (BBB). Looking at proteins such as Carcinoembryonic Antigen (CEA) on these nanovesicles allows us to look at gastrointestinal health and thus colorectal cancers. Additionally, PON1 on HDL is hypothesized to be responsible for significant atheroprotective properties of HDL and may be a much better cardiovascular risk marker.

The bottlenecks associated with the measurement of these particles rely on several factors. Firstly, the gold standard for protein measurement, Enzymatic Immunoassays (ELISA), are incompatible with lipids due to redox interference from lipid peroxides on these particles. At the same time, flow cytometry-based methods cannot be extended to such small vesicles. Secondly, ELISA and other commonly used methods lack the sensitivity (>pM) to assess these particles accurately (~fM-pM concentration), requiring more sensitive methods. Thirdly, there can be significant competition from the soluble versions of these proteins or non-targets, which requires us to have a large dynamic range to operate away from sensor saturation. Lastly, the detection method should be fast (<1 hour) and cheap (~$1) so that everyone can afford these tests – perhaps even as an annual or semi-annual test. This would ensure we can detect the disease early, start the treatment early (>85 percent survival for stage I cancer detection) and use these markers to monitor the efficacy of treatment, remission, and recurrence. We have developed Charge-gated Anion Exchange Membrane (AEM)-based Electrokinetic sensors and Janus particle-based diffusometry sensors that are very sensitive (~fM), cheap (~$1), fast (<1 hour), and have a very large dynamic range and is less susceptible to environmental factors.

We have currently validated our strategy of detecting lipid-protein assemblies as well as our technology by assessing cardiovascular risk, detecting glioblastoma and colorectal cancers non-invasively with ongoing studies that include pancreatic, breast and lung cancers. We were able to achieve an impressive AUC ~ 0.99 with PON1-HDL in distinguishing Coronary Artery Disease patients from healthy controls and an AUC > 0.95 for cancer patients from healthy controls. We compared our platform to the state-of-the-art methods used by healthcare providers for the cardiovascular risk assessment, outperforming Cholesterol tests, Apolipoprotein levels, HDL-P levels (AUC~0.65) that accurately predicted CAD in about 60-70% of the patients compared to our platform that predicted >95% of the CAD patients and controls accurately.

Publications:

1. Kumar, S.; Maniya, N.; Wang, C.; Senapati, S.; Chang, H.-C., Quantifying PON1 on HDL with nanoparticle-gated electrokinetic membrane sensor for accurate cardiovascular risk assessment. Nature Communications 2023, 14 (1), 557.