Why we do single-cell RNA sequencing

Going beyond standard of care responsibly

At Valius, our job is to analyze each patient’s tumor biology so that they and their care teams can make the most informed treatment decisions possible.

Since we often assess therapeutic options beyond standard of care, it’s particularly important that we understand every aspect of the therapeutic targets we examine: DNA alterations, RNA expression, protein expression, clinical data from relevant trials, etc.

In practice, this means that while we often identify therapeutic targets for our patients, it is equally, if not more, important to disqualify potential targets. Single-cell RNA sequencing is a critical part of that process.

How single-cell RNA sequencing helps us select the most promising targets  

We start all of our diagnostic workups with whole exome and whole transcriptome sequencing, the latter of which measures average RNA expression across the patient’s sample—that is, how many RNA transcripts of each gene are present in the cells in the sample. Understanding the patient’s RNA expression is critical because, for many novel therapies like antibody-drug conjugates, T cell engagers, CAR-Ts, and radioligands, higher expression of the relevant target in tumor cells is often correlated with a higher likelihood of therapeutic response.

The drawback of whole transcriptome sequencing is that, by calculating average expression levels across the sample, it necessarily includes some data from the non-tumor cells (known as “stromal cells”) in the sample. This means it can confuse high expression in the stromal cells with high expression in the tumor.

By contrast, single-cell RNA sequencing allows us to separate the tumor cells from the stromal cells and analyze them independently. As a result, it can help us understand which targets are specifically expressed in the malignant cells.

Single-cell RNA sequencing in action

Single-cell analysis plays a major role in the hypothesis-generation phase of our patient engagements. Take the example of a myxofibrosarcoma patient we worked with recently.

We started our workup by performing whole exome and whole transcriptome sequencing on two different patient samples collected from two different sites—one from a lung metastasis and one from a metastasis on another tissue—at two different points in time.

In the first (lung) sample, we observed extremely high expression of a gene called NaPi2b. This caught our eye because some investigational therapeutics targeting NaPi2b have shown efficacy in early-stage clinical trials, including one which achieved an impressive objective response rate in ovarian cancer.

However, in the second sample, taken a year later, the NaPi2b expression vanished. Compare the two bars in the chart below: the first bar, corresponding to the first sample, shows a NaPi2b expression of 140 transcripts per million (a common unit of measurement for RNA expression); the second shows expression of just under one.

We often see the expression levels of genes fluctuate over time, but almost never this drastically. Indeed, when we looked at the expression levels of other therapeutic targets (e.g., B7-H3, MET, FAP) across the two samples, we detected only minor changes.

Single-cell sequencing helped us understand what was going on. When we analyzed our sarcoma patient’s first (lung) sample using single-cell sequencing, we found that the NaPi2b expression was largely coming from normal lung tissue (alveolar epithelium) rather than tumor cells. Indeed, compared to the other targetable genes we assessed, NaPi2b was especially differentially expressed in the stromal cells rather than the tumor cells. In the chart below, we've plotted what we call "tumor overexpression" for each gene: the percentage of tumor cells expressing it highly minus the percentage of stromal cells expressing it highly. Note NaPi2b's position in the top-left corner, with one of the most negative values we observed.

This finding clearly explained why NaPi2b didn’t appear to be highly expressed in the second sample: it was predominantly expressed on the patient’s surrounding lung tissue, and there simply weren’t any lung cells in that second sample.

Single-cell sequencing helped us deprioritize NaPi2b as a potential therapeutic target, giving us more time to investigate other plausible targets. Luckily, for this patient, there were several of those. In particular, B7-H3 was highly expressed in both of the patient’s samples and overexpressed on the tumor cells (see again the chart above), and the patient was able to access an investigational B7-H3-targeting therapeutic.

Refining the target list as is important as generating it

When patients go above and beyond the standard of care, their care teams must be extraordinarily careful in evaluating targets and associated therapeutics.

At Valius, we strive to arm our patients and their physicians with only the most promising, biologically-informed therapeutic options. Single-cell sequencing is an indispensable tool in that process.