Single-cell interaction technology from Scribe Biosciences 1

Promising technology out of UCSF to use droplets as containers for single-cell interaction assays

Way back in 2009 I remember the first time my sales manager at RainDance Technologies presented their core technology to a sophisticated group at the US National Institutes of Health, with an introduction to microfluidics. He began by saying “Microfluidics has gotten a poor reputation in the life sciences…” and that was a saying that has stuck with me.

The promise and reality of microfluidics

Later that same year I struggled to close (in the sales sense of the word) several sizeable target-enrichment customers. I ran into the barrier of fierce competition from Agilent’s then-new SureSelect target enrichment (not requiring any $250K specialized equipment as it was all reagents-based) and soon many other competitors including NimbleGen, Illumina (their now-discontinued Tru-Seq) and Life Technologies (AmpliSeq) joined into that market. The poor reputation for microfluidics has to do with the cost of manufacturing the microfluidic device itself, adding cost and complexity to the overall process.

A first-generation RainDance Technologies RDT1000 microfluidic chip showing the merge of a genomic DNA droplet (bottom left) and a pre-formed oligo pair droplet (upper left) in a 1 to 1 ratio

An introductory slide that I remember so clearly from the NIH presentation was the tantalizing fact that small water-in-oil emulsions can carry single cells, proteins and antibodies, DNA or RNA molecules, illustrated here in a RainDance YouTube video from 2013 (Darren Link was their VP of R&D and a founder of RainDance). Through RainDance’s history they limited themselves to encapsulating genomic DNA or oligo pairs for target enrichment in a 37 pL droplet format, being able to partition genomic DNA into one set of droplets, and then combining them with pre-prepared droplets containing the needed oligo pairs and PCR reagents for emulsion PCR.

Naturally a few years later it was QuantaLife (acquired by BIO-RAD) and RainDance themselves who would develop digital PCR using droplets, again with encapsulating genomic DNA and an oligo pair.

And then another few years later it was Mission Biosciences to encapsulate single cells and then use that single diploid genome as a template for single-cell targeted enrichment. Recently Mission Bio extended their platform to detect proteins with encapsulated antibody and detection reagents.

Another company using a different technology, Berkeley Lights, uses single-cell ‘pens’ to separate single cells, and is able to monitor proliferation on an ongoing basis. I wrote up this technology a few years ago here (complete with short videos of the technology). Their future is clouded, however, due to their inability to scale down their flagship Beacon instrument (at a cool $2M per unit) to something the general research market could use. They were showing data from a less-expensive unit in development but have not brought anything to market yet.

The need to study cell-to-cell interactions with precision at scale

Now Scribe Biosciences wants to look at cell-cell interactions. Understand that for cutting-edge treatments to be developed, scientists have to understand the biological mechanism of particular cellular signals. And for cancer in particular, there is an enormous interest and investment in immunotherapies. It is estimated that half of all oncology therapies are immunotherapies (about a $160B global oncology therapeutics market in total).

For example, with CAR-T (Chimeric Antigen Receptor T-cell) therapy an important milestone was reached: the two first patients that received CAR-T-cell cancer therapies over 10 years ago have been in complete remission, and a recent editorial and research paper in the journal Nature describes 12 years of follow-up with the original two patients with end-stage Chronic Lymphoid Leukemia (CLL), where these CAR-T cells are detected in these patients over ten years from their first administration. These patients are called “cured”, a term that is not used lightly. FDA approval of the first CAR-T cancer treatment occurred in 2017, and subsequently an additional four (a total of five CAR-T treatments for cancer as of this writing).

They are:

• tisagenlecleucel (Kymriah®) for acute lymphocytic leukemia in 2018 by Novartis
• axicabtagene ciloleucel (Yescarta®) for diffuse large B-cell lymphoma in 2017 by Kite, part of Gilead
• brexucabtagene autoleucel (Tecartus®) for mantle cell lymphoma in 2020 by Kite, part of Gilead
• lisocabtagene maraleucel (Breyanzi®) for adult large B-cell lymphoma in 2021 by Juno Therapeutics part of BMS
• idecabtagene vicleucel (Abecma®) for multiple myeloma in 2021 by bluebird bio and BMS

This interest in CAR-T and TCR-T cellular immunotherapies is high because they have demonstrated true breakthroughs in cancer treatment. One study following up Kymriah for ALL determined a response rate of 52% (40% complete response, 12% partial), and after 12 months of the responders 65% had relapse-free survival. 

The challenges of cellular immunotherapy development

These new cellular immunotherapies are contingent upon characterizing T-cell to cancer cell interactions, where the engineered T-cell has a single chain fragment variable domain (called scFv), in-line with a transmembrane domain and co-stimulatory and stimulatory domains. These T-cell-to-target-cancer-cell interactions are currently measured in bulk cell-culture, where assays for different cytokines or cell death or presence of specific cell-surface markers indicate a positive result – the activation of a T-cell in the presence of its target cell, and the death of the target cell.

By doing these experiments in bulk culture, there is room for a single-cell to single-cell approach to improve the accuracy of the measurement as well as enable the scalability of parallelizing the assay. And naturally what can be useful for chimeric antigen receptor T-cell work is applicable to study of other autoimmune disorders, from Multiple Sclerosis to rheumatoid arthritis.

Scribe Biosciences’ approach to study cell-cell interactions inside droplets

Scribe Biosciences is developing a droplet-based microfluidic technology that they call Microenvironment on Demand (MOD). Think of it as an assembly line with four input components: a bucket of effector cells (say the engineered CAR-T cells), a bucket of target cells (say a cancer cell line), a bucket of assay reagents (say a set of labeled anti-interferon-gamma antibodies set to light up in the presence of interferon-gamma and a particular polystyrene bead), and lastly a bucket of special polystyrene beads (designed to work with the prior bucket of anti-interferon-gamma reagents).

The instrument will take each bucket of input cells, beads or reagents and package them into uniformly-dimensioned droplets. And thanks to the ability to dielectrophoretically sort and combine the droplets based upon their contents, one of each type can be arranged, assembly-line style, and merged together into one assay droplet. After incubation for the cells in question to interact with each other (and produce the cytokine(s) searched for in the assay component), a second sorting step can separate out the positive assays from the negative ones, and the positive assays can naturally be collected for single-cell RNA or other analyses.

The proof is in the measurements

The PNAS paper (Madrigal et al, “Characterizing cell interactions at scale with made-to-order droplet ensembles (MODEs)”) has plenty of interesting detail, including the proof-of-concept application to characterize CAR-T activation at scale using this technique, and then go on to demonstrate enrichment for activated CAR-T cells. (By ‘scale’, think tens to hundreds of thousands of cell assemblies per hour, or even higher.) The discussion suggests other readily-applicable cell types (stem cells for hybridoma production, spheroids and organoids). Also the flexibility of the assay components mean not only cytokines can be measured, but also cytotoxicity and antibody production and antibody specificity.

I believe this technology demonstrates great promise in fulfilling the original premise of forming, sorting, and joining microdroplets in a ‘laboratory on a chip’, which has been worked on for many years. Indeed the journal by that same name started publishing in 2001. In a conversation with Russell Cole of Scribe Biosciences he indicated that the instrument they are developing will have four color channels for assay detection, which open up potential for measuring multiple analytes simultaneously.

I also understand that in these early days they are looking for early-access customers / partners for their platform – reach out to Russ directly.