We've all heard the analogies. Chopping up cells and analyzing them in bulk is like trying to pick up the taste of a single bad blueberry in a slurpy-sized fruit smoothy. Single-cell analysis is important and essential for studying rare cell populations that play important roles in cancer and immunology. But it's also expensive and technically challenging.

We've also heard the phrase send a thief to catch a thief, a spy to catch a spy. Well, what if we could send a cell to catch a cell? That's exactly what the emerging biotech, 

Sampling Human, is doing. They've figured out how to produce bioparticles from engineered yeast cells to detect rare cell populations with unprecedented sensitivity and resolution. 

‍

The Immune System Is the Most Advanced Diagnostic Tool That We All Get for Free

Think about that for a moment. Right now, there are billions of immune cells circulating through your veins. You probably imagine them all looking pretty similar, but now consider that just among your B and T cells, there are likely 1,000,000,000 variations at any given time from 100,000,000,000,000,000,000,000,000 different possible configurations.

That's billions (with a B, plural) of micro-diagnostics poised to recognize an attack. We could never possibly hope to invent a test that is that sensitive and that comprehensive, not by many orders of magnitude. That's why Daniel Georgiev, Co-founder and CEO of Sampling Human, calls the immune system "the most advanced diagnostic tool that we all get for free."

And he's not alone. Jian Han, renowned immunologist from the Hudson Alpha Institute and founder of immune sequencing company iRepertoire, calls the immune system "nature's best doctor" and notes that it's "more sensitive than any test we could invent." 

Companies like iRepertoire and Adaptive Biotechnologies are hyper-focused on characterizing and harnessing the immune landscape to diagnose and treat disease. And they've made major strides, but picking out extremely rare cell types remains a challenge.

The problem goes back to those huge numbers. In a single milliliter of blood (about the volume of your pinky), there are 5 million blood cells. Only about 100 of those are CD34+ hemopoietic stem cells – the kind that can each self-renew and differentiate into the very, very, very varied immune cells in your body. That's just 0.002% of the sample.

We can't even measure 100 in 5 million, let alone pick out the 1 in 1 billion that will be able to identify and neutralize a particular threat. But one cell is all it takes. A single immune cell can recognize a grain of pollen and initiate a response you can see with the naked eye. 

‍

How Do We Measure With That Much Sensitivity?

Obviously, we're talking about single cells, but standard single-cell technologies aren't sensitive enough. 

Most single-cell sequencing technologies start by partitioning individual cells using wells, beads, or droplets. But if you're looking for a very rare cell population, that means screening through a lot of irrelevant cells. The approach doesn't scale well and can require specialized instrumentation. Additionally, the cell sorting process disrupts the cellular environment, eliminating critical information about cell-to-cell interactions or cellular health. 

Sampling Human's approach is different. They use tunable bioparticles to capture cells of interest based on biological markers like cellular surface receptors or apoptotic (cell death) signals. The bioparticles encapsulate the cells and any cells they are interacting closely with. “Once encapsulated, they capture the target cell RNA and let off a bioluminescent signal that can be read by a simple plate reader device already present in many labs,” Georgiev explained. 

Importantly, this elegant approach means that we can now study single cells without first sorting them—eliminating the need for expensive FACS instruments that require specially trained, expert users.

And what are these bioparticles? The cell walls of engineered yeast. 

" We call them bioparticles because they work like smart beads, and for most applications, they don’t need to be alive. When someone hears yeast cells they immediately think that you have to grow or culture them, and that’s not the case,” explains Georgiev. “The yeast cell wall is an excellent chassis that provides lots of functionality, is easily programmable, and can be manufactured in tons.”

‍

What Can Yeast-Derived Bioparticles Tell Us About CAR-T Therapy?

CAR-T therapy is a blockbuster technology that uses the patient's own T cells engineered to recognize cancer. But it's incredibly expensive and a weird numbers game.

The reason our immune system is able to turn a nanoscale threat into a 6-foot, 250-pound response is clonal expansion. When that 1 in a billion immune cell recognizes a target, it makes an army of clones to neutralize it. 

If you're in the business of CAR-T therapy, you pump in the cells at a rate of about half a million dollars, cross your fingers, and wait. 

“Once it's infused, the frequency of those CAR-T cells may be one in 100,000 or one in 1,000,000, which is below the detection limit," explained Georgiev.

This is an uncomfortable prospect, especially when the waiting period means precious time for a critically ill patient. 

With Sampling Human's smart encapsulation technology, the bioparticles can be engineered to recognize the CAR-T cells so they can be tracked over time despite their low concentration. 

Back to the numbers—flow cytometry, the standard approach for cellular detection—has a sensitivity of about 1 in 10,000 cells. Cutting-edge single-cell technologies offered by industry leaders like 10x Genomics, Parse Biosciences, and Scale Biosciences enable researchers to study ~40-100K single cells simultaneously. That means, if you're really lucky, you might be able to catch a single infused CAR-T cell. Sampling Human's bioparticles can detect cells at a concentration of as little as 1 in 1 million, increasing our likelihood of measuring infused CAR-T cells at least 10-fold. 

‍

What About the Cost?

Assay sensitivity challenges are one thing, but even for more abundant cell populations, the cost is a challenge. 

Myeloid-derived suppressive cells (MDSCs) are immune cells with the counter-intuitive function of suppressing the immune system. In healthy individuals, they moderate an inappropriate immune response, but in cancer patients, they can mistakenly protect tumor cells from immune attack. 

MDSCs are less rare than hemopoietic stem cells or infused CAR-T cells at about 1 MDSC cell/1,000 cells in a typical blood draw. That means that if you sequence 100,000 whole blood cells, you'll see up to roughly 100 MDSC cells. That's not bad until you consider that it will cost about $10,000. 

Dr. Xiaoqiang Wang, a scientist at City of Hope National Cancer Center, is currently recruiting 136 patients for a clinical trial on MDSC suppression. By other, still cutting-edge single-cell approaches, the cost of sequencing alone would be well over $1.3M ($10,000 X 136). With Sampling Human's technology, they expect to cut that cost by at least half. 

"Half the cost, more clear information, and a more targeted cell signal. That's the benefit," said Dr. Wang about the proposed work with Sampling Human. Dr. Wang will be using Sampling Human's technology to not only identify and measure the rare MDSCs but also to sequence them. 

‍

How Does Rare Cell Detection Translate To Sequencing?

The first step in Sampling Human's workflow is what Georgiev calls "smart encapsulation." Smart, because the bioparticles are engineered to detect very particular cellular sub-populations, and encapsulation, because the bioparticles organize to form a tight capsule around their target. 

All that's required to detect and measure the cells—a process Sampling Human calls biocytometry—is encapsulation. But once the bioparticles encapsulate the cell of interest, we can do all sorts of other interesting things, from studying cell-to-cell interactions, adding additional signals for markers of cell health or activity, and even sequencing. 

Once the cells are encapsulated, they can be broken apart (lysed). When this happens, the bioparticles create a concentrated pocket of the RNA that was present in those cells, while the RNA from all other cells dissipates into the sample. 

These concentrated pockets of RNA have an interesting feature, they increase the likelihood of reverse transcription (copying RNA into DNA)—the first step of RNA sequencing. Also, the proximity of the bioparticles—along with some additional steps to concentrate them—allows for a kind of barcoding that enables the specific identification of signals from the cells of interest after sequencing.

"This technology would not have been possible three years back," explained Georgiev. That's because even just three years ago, we did not have the machine learning capabilities to deconvolute the sequencing data. 

‍

Where Do We Go From Here?

Assuming this all works as prescribed, we can now sequence single cells with orders of magnitude more sensitivity and at half the cost – according to Dr. Wang’s estimate. The actual costs, particularly for very rare sample types where you have to sequence a lot of samples, are likely to be significantly lower. “The cost per patient using our technology will be in the hundreds of dollars,” Georgiev noted, which could signal a fundamental shift in healthcare.

We all know sequencing costs have decreased faster than Moore's law. To date, single-cell sequencing has not seen the same logarithmic drop. A paradigm change in methods could be the shift that kicks off that drop. 

Parse Bioscience and Scale Bioscience made big strides by eliminating the need for specialized instruments. Factorial Bio is flipping the script by enabling library preparation within cells. Enter Sampling Human with bioparticles from engineered yeast. 

The Human Genome Project promised to translate the genetic code into digital data. The impact has been astronomical, but we've also learned that bulk genomic DNA is just the beginning of the story. If single-cell sequencing technologies can really see dramatic price drops concurrent with increases in sensitivity, we stand to create a digital fingerprint of the immune system. 

With that fingerprint, we could read the output of the most advanced diagnostic tool instead of trying to reinvent it. 

‍