Octant's drug screening platform (we call it 'the Platform'), is the core of Octant. We not only use the Platform to screen chemical ligands for drug discovery, but also as a continuous development and testing infrastructure to build our next set of technologies. This blog post gives a high level overview of the Platform, how it all started, and how we use it to rationalize drug discovery.
At Octant, we engineer biology to rationalize drug discovery. To do that, we're mapping thousands of chemicals across hundreds of cellular receptors and signaling pathways to create "chemical profiles" that can be used to treat complex diseases (Figure 1). That's why we invented Octant's drug screening Platform, which tests many cellular receptors in a single assay for their response to a drug or other molecule - contrasted with traditional assays that test each receptor individually in a single well. This enables us to use a single drugging condition to inform on the activity of many receptors simultaneously, a technology we call "multiplexing." To put this in perspective, let's say you want to test 100 cellular receptors against 100 different drugging conditions in triplicate. 100 receptors x 100 conditions x 3 repeats would require 30,000 single-well assays, or ~80 384-well plates. At Octant, we accomplish the same assay using a single 384-well plate.
Figure 1. Making a "chemical profile". GPCRs are expressed across the cellular surface, with each cell expressing several types of GPCRs (example receptors shown in pink, purple, and green). Each GPCR signals through one, and sometimes more, downstream signaling pathways to elicit a cellular response (signaling pathways shown as Gi, Gs, and Gq). The chemical profile (graphs on right) shows how each GPCR responds to the ligand (shown in blue circles) through each signaling pathway.
WHAT IS MULTIPLEXING?
Multiplexing to increase data throughput isn't new. It dates back to the 1800s when it was developed to improve telecommunications and electronics. In this context, multiplexing means to send and receive many independent signals through the same channel, making our current communication pathways like phone, email, and internet, possible. It has also been used to measure multiple biological outputs simultaneously. For example, multiplexing is used in fluorescence-based assays such as qPCR, flow cytometry, and microscopy, by coupling multiple biological outputs with proteins that emit light wavelengths in different color ranges. Although fluorescence-based multiplexing has facilitated a whole host of new technologies and discoveries, it is limited by a small handful of discrete color ranges.
At Octant, instead of colors, we use short strings of nucleotide sequences (we call them "barcodes"), to tag and identify cellular receptors and their corresponding pathways. Our ability to inexpensively design, build, map, and engineer these barcodes into the desired biological pathways is made possible by recent advancements in DNA sequencing, synthesis, and editing. Now, instead of pairing a biological output to a small number of luminescent colors, we engineer biological pathways to produce these genetic barcodes and later "read", or count, them using high-throughput sequencing technologies. Using nucleotide barcodes massively increases the multiplexable outputs of an assay. For example, a string of 15 nucleotides has 4^15, or 1,073,741,824, unique combinations readable by a sequencer. Increase the string length, and the possible unique identifiers increases exponentially!
MULTIPLEXING ENABLES LARGE SCALE CHEMICAL-RECEPTOR MAPPING
Humans have ~800 G-protein coupled receptors (GPCRs), which mediate olfaction, taste, and numerous physiological processes throughout the body. Approximately 400 of the GPCRs mediate physiological functions, including significant roles in diseases like schizophrenia, heart disease, diabetes, and others. In fact, roughly 30% of all FDA approved drugs target GPCRs. Interestingly, although humans have a large number of GPCRs, GPCRs signal through a small number (5-ish) of common downstream signaling pathways. The combinatorial possibilities of these GPCRs and their pathways, while finite, is very complex. Most receptors signal through more than one of the pathways, and molecules can cause different cell activity by preferentially activating one downstream pathway over another (sometimes referred to as "biased" activation or agonism). While the task in front us is daunting, by gaining a more thorough understanding of this complexity, we can more rationally pursue better treatments for complex diseases.
In order to understand GPCR signaling complexity, we needed a way to measure multiple receptors and signaling pathways at scale. This is where multiplexing comes in! Although advanced DNA sequencing technology is still relatively new, it has undergone rapid development and expanded beyond genome sequencing. Back at UCLA, where Octant began, Jones et al. used genetic barcodes to multiplex GPCR signaling and showed that we can decode a portion of receptor activity in olfactory GPCRs. By doing a broad screen consisting of tens of olfactory receptors across several chemicals, they identified 79 new interactions in the sense of smell. With prototype in hand, Octant was born, and began tackling the pharmacologically relevant GPCRs.
To achieve our goal of mapping interactions between pharmacologically relevant sets of GPCRs and chemicals, we need to do five things: 1) Engineer cell lines to express a specific GPCR of interest; 2) Design biological circuits that report on one of the ~5 signaling pathways activated by that GPCR - we call these circuits "reporters;" 3) Link a unique molecular barcode to that reporter and GPCR → now we have a DNA plasmid that contains our GPCR of interest, a reporter, and a barcode; 4) Engineer the cell such that when our GPCR of interest is activated along the reporter pathway, the barcode we assigned to it is expressed; 5) Repeat all of the above for all GPCRs and all their reporter pathways, each combination tagged with a unique molecular barcode. Although there are finite number of GPCRs, the number of reporter types seems infinite, with each GPCR requiring a specific cellular background, signaling pathway modification, and more. Given this, we have A LOT of unique GPCR - reporter - barcode combinations to build and screen.
After making each of these uniquely barcoded individual cell lines, we mix them together forming a cell library that can be seeded into each well of a high-throughout assay plate. We can now interrogate the entire library in a single well by introducing a molecule we'd like to screen into that well. Any particular GPCR pathways that are activated by that molecule will express its associated unique barcode via the engineered reporter pathway. Basically, more activated GPCR = more barcode. By counting all barcodes in the well, we can tell which pathways were activated by that drugging condition, and how intensely activated relative to any other activated pathways. By treating the cell library with a variety of drugging conditions in different wells, we can determine the biological output of every single individual cell line to each of those conditions in a multiplexed, high-throughput fashion. This enables us to map interactions along three axes: 1) thousands of chemicals, 2) hundreds of receptors, and 3) the signaling pathways they activate.
Figure 2. Following a unique molecular barcode through Platform Development. First, GPCR - reporter - barcode constructs with unique barcodes linked to individual receptors are made using HT library cloning. These are individually stably integrated into mammalian cells and then mixed together to make a 'Cell Library'. Cell libraries are tested on the Platform where each well on the plate tests a different drugging condition. Finally, GPCR response to a condition can be measured by sequencing the barcodes and counting the number of times a barcode is transcribed in a well.
OCTANT'S PLATFORM DEVELOPMENT TEAM
As you can imagine, generating the thousands of cell lines required to design and improve on the above process is an incredible amount of molecular biology, tissue culture work, and analysis. It requires iterating on plasmids for every single GPCR - reporter- barcode combination, individual cell lines that express these plasmids, and blending them into cell libraries. Over the past few years we've developed technologies to streamline this process and can go from reporter design to complex, multiplexed sequencing data in <6 weeks. For perspective, this amount of data might take the same number of scientists using traditional methods well over a year. This process goes a little something like this...
High-Throughput Library Cloning (~2 days): We generate hundreds of plasmids containing each GPCR - reporter - barcode combination. Using pooled library cloning methods, we generate dozens of individual constructs in a single reaction.
Whole-Plasmid Sequencing (~1 week): We use our in-house plasmid sequencing platform, OCTOPUS, to sequence the plasmids containing each GPCR - reporter - barcode combination. By sequencing the entire plasmid, OCTOPUS automatically identifies and "maps" a receptor - reporter pair with a unique molecular barcode that can later be identified on the drug screening platform.
Cell Library Generation (~2 weeks): We stably integrate each individual plasmid into individual cell lines in parallel. These are later blended into cell libraries to test on the Platform.
Testing (~1 week): A Platform run by a single member of our team can test 1000s of conditions at a time with. After a number of optimizations and scale-ups, we've significantly reduced the amount of hands on time and screening large compound libraries can feel almost trivial (even though its really, really not).
Analysis (~1 day): Automated computational pipelines analyze the sequenced barcodes to call statistically significant "hits", or interactions, fast.
Add in a handful of robots, some custom software, and a few hard working humans, and voila! you have Octant's Platform Development Team.
Although our drug screening platform was designed to screen chemicals for drug design, we've found that we can utilize it to develop its own successor improvements. We regularly use the Platform's multiplex superpowers to accelerate development of our reporter designs, promoter optimizations, cellular background manipulation, and more. "Use the platform to improve the platform" is a common phrase here at Octant, and our team of ante-disciplinary scientists use it and their creativity to bring more ideas to life. At this time they've stretched the Platform to its limits, forcing us to come up with new strategies to increase our scale, find new analytical methods, and be creative with how we sequence barcodes from various reporters. This push has only made us improve our technologies across the board, and we're excited to share a few of those developments with you over the next several months.