GEN sat down with Joe Beechem to learn more about his perspective on the explosive growth of spatial biology technologies.
This interview has been edited for length and clarity.
GEN: I want to start by asking you where you think the state of spatial biology is today?
Beechem: I had hopes that this field was going to grow like this. Now it is growing like I had hoped. A lot of what we talked about, when we spoke a few years ago, has played out—probably even bigger than we anticipated.
Who would’ve thought we’d have this many people wanting to get into spatially resolved work? And there is a reason for that. We have impactful papers that are coming out almost every week that are discovering new areas of biology. New places where spatial context is essential for showing how tissues respond and how drugs work.
You just can’t ignore that kind of raw discovery rate. I’m super happy and excited about it.
GEN: There is a lot of excitement about spatial, for sure. Some may call it hype. Do you think it’s hype?
Beechem: It’s not hype at all. We’ll look back someday and wonder, “How did we do modern research for a hundred years and not do spatial?” Every living organism on this planet lives in three dimensions of X, Y, and Z. You must ask yourself, “Why didn’t we study it in X, Y, and Z for the last hundred years?” The reason was the tools were no good.
It’s almost the exact opposite. It’s not hype. It’s like the air you breathe. Can you think of a case on this planet where a living organism exists in zero dimensions? It doesn’t happen. Everything exists in space. We’ve just been ignoring it because we didn’t have any tools.
GEN: Do you think spatial is going to take over single-cell RNA-seq?
Beechem: No. There is going to be a place for single-cell analysis. There will always be cells in solution (like blood) that you will want to work with.
But you have to scratch your head pretty hard to think about cases where spatial isn’t important. Instead of the bulk technologies being the focal point of a study, it’s going to be the other way around. People are going to interrogate these systems spatially, and single-cell analysis will provide the ancillary pieces of information.
That’s the way tools work in biotechnology. They don’t just disappear. Take something like qPCR. We used to wonder if it was going to disappear. But it continues to fulfill a key role in a particular area. That’s where most of these other techniques will go.
GEN: How many people can use spatial biology right now? And how are we going to get it to the point where every single-cell lab can run spatial experiments?
Beechem: Well, that’s where we’re going. Single-cell experiments aren’t cheap; spatially resolved technologies will be commensurate—if not less. I don’t have to throw down $20,000 every time I do a spatial experiment. Spatial will become just another piece of instrumentation people have in their lab.
We’re trying to eliminate some of the specialized parts that were out of the range of the average scientist. It’s not really the instrumentation; for $300,000 you can get a state-of-the-art spatial instrument. And then, for a couple of thousand dollars a sample, you can get a whole transcriptome. So, as far as cost structure goes, it is probably a little bit less than single-cell RNA-seq.
Also, the number of cells in spatial makes single-cell RNA-seq look embarrassing. I’m doing 4 million cells every time I turn the instrument on. How long does it take someone to do 4 million single cells? It’s a bunch of work to do it with single-cell RNASeq.
I have not met a person who conducts single-cell experiments who isn’t seriously planning on putting an X, Y, and Z coordinate on every cell that they’ve got. I haven’t met one that says they don’t want to have spatial.
GEN: Is wanting to do spatial experiments different from having the budget for spatial?
Beechem: How can you not have it? Imagine you’re going to do 10 to 20 single-cell experiments at scale in a year. That’s all the funding you need for a spatial platform buildout.
GEN: With spatial, there are the big whole transcriptome instruments. And there are platforms that have 20 probes, or maybe a hundred probes, for more targeted imaging. I’ve heard pros and cons to both. Why is it that NanoString thinks that big data is the way to go for spatial?
Beechem: It’s not big data for big data’s sake. It is high-dimensional data that has a lot of information content in it. We believe the plex (that is, the number of transcripts) is incredibly important, and we’ll continue to push that technology forward.
When you generate this high-information-content data, the beauty is that it can just sit there and you can continue to discover and learn more from it for years. It’s a batch of information that never goes stale.
One of the other reasons for plex is that we want people to be able to share the data. So the data collected at high plex from a pancreatic cancer study, can be compared to data from a study on rheumatoid arthritis (RA) or breast or ovarian cancer. If studies work with the same or nearly identical large-scale panels, they can start cross-fertilizing each other.
And that will lead to an understanding of how ovarian cancer is linked to breast cancer, and of how the immune system and RA interact. What do we learn from that?
One of the things I learned on my last tour through Europe is that we’re making biomarker discoveries in space. And the same biomarkers are addressing autoimmune organ transplantation and cancer on both sides of the same coin. These groups are starting to share information. I learned as much from the organ transplantation group about cancer on this last trip that I learn from going to a cancer meeting. We’re starting to see these fields cross-fertilize each other.
GEN: Do you think that cross-fertilization will still happen if one group is on a NanoString platform and another group is using a different platform? Or is there going to have to be a set of standards?
Beechem: What we’re going to find is that high-quality digital data just works. That was the beauty of digital technologies interacting with next-generation sequencing because high-quality digital data just fits together well. The other thing is that spatial is in situ data. All the in situ technologies will be able to share data, and we’ll be able to make sense out of it together.
GEN: Is moving into the clinic the long-term vision for NanoString?
Beechem: We’re a big believer in translational research. That’s why we made all our platforms work on real-world, archived, FFPE samples. We don’t tack it on at the end like a Band-Aid. Every chemistry project I do, every platform I build, is built for that sample type. A lot of us got into life sciences to change the world for our kids and our parents. So, translational is huge for us. Now, it’s just the time scale. But we don’t control that, our customers do. And I’m just amazed at what they’re doing with the platforms.
Spatial is moving into the clinic two times faster, or three times faster, than I had ever expected. Knight Diagnostic Laboratories, Oregon Health & Science University, is a great example of that. They have validated over 85 of our barcoded antibodies now and are pushing them in the diagnostics lab.
There are certain questions that you can answer at the protein level that you cannot do with sequencing. All the phosphorylation states of these core needle biopsies that come from breast cancer samples can be quantitated beautifully using our antibodies. And it’s exciting to see that. The Mayo Clinic is also doing this type of work. It’s the translational people that want to change the world, and this is a path for them to do it.
GEN: I know that you worked in next-generation sequencing (NGS) for a long time. How did this spatial work all begin for you?
Beechem: It was listening to the unmet needs that are out there. I spend a lot of time in the outside world, and that’s what I’m always listening for. Sometimes the investigators don’t know how to ask; you just have to see what they’re trying to do. The field of immuno-oncology led me to realize that we had to do this. Checkpoint inhibitors re-architecture the spatial immune system in tissues. But how are we going to develop good ones if we can’t measure that in space?
So, I literally invented the GeoMX waiting in the lobby of the Marriott Hotel when I was traveling to give a talk. I asked, “How can we take the power of NGS and not lose all the spatial coordinate information?”
And that was when I realized that all we had to do was put a sequencing barcode on a photocleavable linker, and we’d be there. I scribbled it down on a piece of paper in January of 2015. And then we we’re off to the races in spatial biology.
Our first paper, using the prototype instruments in my lab, helped us discover, together with Jen Wargo and Jim Allison, tertiary lymphoid structures in tumors. Now, there’s an entire field that does nothing but work on tertiary lymphoid structures (TLS) and tumors because they’re
GEN: Were you at NanoString in January 2015?
Beechem: Absolutely. We weren’t a spatial company back then, though. When I came back with this idea, everybody was looking at me like I had three eyes growing out of my head. They asked, “Why is Beechem all of a sudden talking about the importance of spatial?”
Our investigators put it together on their own. People came in at night to make little microscope slides with pieces of tape on them, to pipette liquids over tissue slices, and to measure signals in space. And it was working. Slowly but surely, we were convinced. Then, I had to convince the inside of the company that this was a worthwhile thing to do, and we did that. And the rest is history. It’s been a fun seven years.