Drug Discovery

Overview

There are two major methods we use to develop small molecule therapeutics in my lab. A target-based approach requires reliable structural information on the protein target. However, we often use a structural biology approach to predict the target structure based on homology modeling. Once we identified the potential drug binding cavities, we usually dock 3-5 million compounds for potential ligands. The top score-ranking molecules will be selected and evaluated using in vitro experiments. In the scenario in which there is little to no structural data on the protein target, we will employ a ligand-based approach using HTS. We have several compound libraries (e.g. FDA-approved compound library, 100K diversity compound library, etc) that we routinely use. We deploy various phenotypic screens to identify compounds that disrupt E3 ligase/substrate interactions. If the E3 ligase/substrate pathways are well characterized, we sometimes use fluorescence resonance energy transfer (GFP-E3 ligase, YFP-Substrate) as the HTS assay. Other approaches such as FP assay and lantha screen are also used frequently.

Hit identification

This figure shows a routine work flow on hit compound identification and validation studies using both approaches. The most critical step is to identify robust primary and secondary screen assays.  See Chemical screening section for further info.

Hit to lead

In both approaches, once the “hit” molecules have been identified and tested, we will start the “hit to lead” and “lead optimization” campaign. The candidate “hit” compound will be chemically modified by changing the size, hydrophobic/ hydrophilic properties of its side chain, backbone, and other characteristics. One important aspect of developing the “lead” molecules is that we must thoroughly evaluate their physiochemical properties, not just their in vitro efficacy. We have developed a sophisticated testing algorithm to fully evaluate these compounds focusing on their 1) target engaging and in vitro efficacy, 2) compound off-targeting effect, 3) in vitro and in vivo ADME. This figure shows a routine work flow on this step:

IND-enabling

In both approaches, once the “hit” molecules have been identified and tested, we will start the “hit to lead” and “lead optimization” campaign. The candidate “hit” compound will be chemically modified by changing the size, hydrophobic/ hydrophilic properties of its side chain, backbone, and other characteristics. One important aspect of developing the “lead” molecules is that we must thoroughly evaluate their physiochemical properties, not just their in vitro efficacy. We have developed a sophisticated testing algorithm to fully evaluate these compounds focusing on their 1) target engaging and in vitro efficacy, 2) compound off-targeting effect, 3) in vitro and in vivo ADME. This figure shows a routine work flow on this step:

Clinical Trials

One of our lead compound BC-1261 was approved for IND in Feb 2019. We are on target for starting a Phase I trial in March 2019.

Insilico Screen

We routinely carry out protein modeling work and use several docking algorithm for in silico compound screen. We use several compound libraries containing 3-10million compounds. Depending on the size of potential drug binding cavity, the docking studies on a typical 3 million compound library range from 18h to 3 weeks using a 48 threads supermicro server. Recently we also installed the docking program on a dell Bladecenter which hosts 32 CPUs, 320 cores, 640 threads and 1TB ram. This new system drastically improved the docking time. We also use this system for QSAR studies and HCS imaging analysis.

Library

Our lab curates and annotates a small molecule library containing 101,106, natural compound, low molecular weight and diverse screening compounds, pharmacological active compounds, clinically used compounds and compound fragments.

HTS Chemical Libraries

FDA 1106
Beyond Flatland 7360
Ion Channels 7200
Anti viral 6480
PPI 4720
Anti cancer 3280
NPB 2720
Anti inflammatory 2320
CNS 2080
Soluble Diversity 2000
Macrocycles 2000
Current FL 1760
3D-Flat land 1600
Singletons 12640
Diversity 43840

RNAi Libraries
We have a collection of MISSION human esiRNA libraries composed of 16,744 esiRNAs. The libraries are kept in 384-well plates.

Plasmid/cDNA Libraries
We have a collection of 13,000 DNA plasmids representing individual proteins in the human genome (MGC premier Human Lentiviral ORF Library). The plasmids are in carried in E. coli glycerol stocks, and can be used for protein expression along with Lentiviral packaging.

Chemical screening

We have the capability to execute HTS assays using absorbance, fluorescent, FP, TRF, TRFRET, AlphaScreen, LanthaScreen, bioluminescence and cellular fluorescence imaging. Assay targets can include soluble proteins, channels, receptors, trafficking proteins, protein/protein interaction, etc. For various liquid handling and detecting equipment, see HTS.

Medicinal Chemistry

Our facility has variety of medicinal and synthetic chemistry equipment including Biotage microwave synthesizer, Teledyne flash chromatography (Biotage and Teledyne.jpg), Dionex HPLC (Dionex.jpg) and thermos TSQ Vantage mass spectrometry (TSQ.jpg). If the chemistry is straight forward with only 3-5 steps, we usually perform it in house so we can quickly test the potential hit molecules in assays. For more complex SAR studies, we routinely have 8-12 FTEs with CRO to carry out these works. For confirming compound target engagement and measuring binding kd, we are also equipped with a SPR (SPR.jpg) and an ITC(ITC.jpg) machine.

Data Storage and analysis

Since we deploy many high-content based screens, we have setup several servers to handle data storage, archival and analysis. A routine HCS run with 50K compounds will require ~0.5-1TB storage space. Currently we have >200TB storage space in our lab; and it can be easily expanded by adding NAS drives as shown here (100TB in a small package!). We also use discovery studio and ScienceCloud to analysis the HTS run, generate SAR table and design new molecules.