We describe a technique for the rapid ab-initio discovery of target-tailored tumoricidal DNA oligonucleotides inside an Illumina sequencing chip. By sequencing oligonucleotide pools we generate a physical microfluidic map of hundreds of millions of potential oligo clusters, in which every cluster is mapped to a specific set of spatial coordinates. Tumor cells, pre-loaded with a fluorogenic reporter of apoptosis, are then injected into the chip and monitored over time. Apoptotic tumor cells are identified and analyzed across the entire map, automatically revealing the coordinates of oligos that induced this effect. We demonstrate this method by identifying, within just a few hours, new oligos capable of directly and selectively inducing apoptosis in primary human tumor cells. Such a major capability could lead to a new paradigm of personalized cancer therapy.

No abstract

Our current model of drug discovery is challenged by the relative ineffectiveness of drugs against highly variable and rapidly evolving diseases and their relatively high incidence of adverse effects due to poor selectivity. Here we describe a robust and reproducible platform which could potentially address these limitations. The platform enables rapid, de-novo discovery of DNA aptamers evolved in-vitro to exert specific biological effects on target cells. Unlike conventional aptamers, which are selected by their ligand binding capacity, this platform is driven directly by therapeutic effect and selectivity towards target vs negative target cells. The process could, therefore, operate without any a-priori knowledge (e.g. mutations, biomarker expression, or known drug resistance) of the target. We report the discovery of DNA aptamers with direct and selective cytotoxicity towards several tumor cell lines as well as primary, patient-derived solid and hematological tumors, some with chemotherapy resistance. Aptamers discovered by this platform exhibited favorable biodistribution in animals, persistence in target tumors up to 48 hours after injection, and safety in human blood. These aptamers showed remarkable efficacy in-vivo as well as ex-vivo in freshly obtained, 3D cultured human tumors resistant to multiple chemotherapies. With further improvement, these findings could lead to a drug discovery model which is target-tailored, mechanism-flexible, and nearly on-demand.

The Monty Hall problem is a decision problem with an answer that is surprisingly counter-intuitive yet provably correct. Here we simulate and prove this decision in a highthroughput DNA sequencing machine, using a simple encoding. All possible scenarios are represented by DNA oligonucleotides, and gameplay decisions are implemented by sequencing these oligonucleotides from specific positions, with a single run simulating more than 12,000,000 independent games. This work highlights high-throughput DNA sequencing as a new tool that could extend existing capabilities and enable new encoding schemes for problems in DNA computing.

Drug discovery is challenged by ineffectiveness of drugs against variable and evolving diseases, and adverse effects due to poor selectivity. We describe a robust platform which potentially addresses these limitations. The platform enables rapid discovery of DNA oligonucleotides evolved in vitro for exerting specific and selective biological responses in target cells. The process operates without a priori target knowledge (mutations, biomarkers, etc). We report the discovery of oligonucleotides with direct, selective cytotoxicity towards cell lines, as well as patient-derived solid and hematological tumors. A specific oligonucleotide termed E8, induced selective apoptosis in triple-negative breast cancer (TNBC) cells. Polyethylene glycol-modified E8 exhibited favorable biodistribution in animals, persisting in tumors up to 48-hours after injection. E8 inhibited tumors by 50% within 10 days of treatment in patient-derived xenograft mice, and was effective in ex vivo organ cultures from chemotherapy-resistant TNBC patients. These findings highlight a drug discovery model which is target-tailored and on-demand.