Why is Spatial Transcriptomics Critical to Studying Sarcomas? Q&A with Dr. Troy McEachron, Ph.D
“Whereas the EWS specimens tended to have very simple genomes, the osteosarcoma specimens looked like a hand grenade had gone off in the genome. These were the most fragmented and complex genomes I had ever seen”. Troy McEachron, Ph.D
The first word that came to our minds when we started interviewing Troy A. McEachron, Ph.D., was “enthusiasm.” This brilliant Assistant Professor exudes passion, a depth of knowledge, and an eagerness to share his love for research and hope that one day children and young adults with the devastating diagnosis of cancer will have more than a treatment: they will have a cure.
Dr. McEachron received his doctorate from the University of North Carolina at Chapel Hill School of Medicine. He did his first post-doc at St. Jude Children’s Research Hospital and a subsequent post-doc fellowship at the Translational Genomics Research Institute (TGen). Dr. McEachron is now an Assistant Professor in the Department of Translational Genomics at the Keck School of Medicine at the University of Southern California.
In our interview, Dr. McEachron talked to us about new approaches to studying sarcomas using NanoString technology.
NS: Dr. McEachron, what led to your research focus on using Next Generation Sequencing (NGS) and GeoMx® Digital Spatial Profiling (DSP) to profile refractory cancer in children, particularly in sarcomas (cancers of mesenchymal cells)?
TM: My background is quite varied and includes immunology, cancer biology, precision medicine, and translational genomics. My interest in sarcomas started during my post-doc at TGen while working on a precision medicine project. We were using both whole genome sequencing and RNA sequencing to profile recurrent and refractory pediatric cancers for use in clinical decision-making. I remember seeing the profile of a patient with Ewing sarcoma (EWS). I was very confused by the genomic profile because it was very quiet except for the defining EWSR1-FLI1 translocation. I found it very intriguing, and it sparked my interest in sarcomas. Shortly thereafter, we profiled several osteosarcomas and saw that they were the complete opposite of the EWS sample. Whereas the EWS specimens tended to have very simple genomes, the osteosarcoma specimens looked like a hand grenade had gone off in the genome. These were the most fragmented and complex genomes I had ever seen. The osteosarcoma genome is so incredibly complex that it makes you wonder how a tumor cell with this amount of DNA damage can actually survive. My focus is to use NGS and other technologies to profile osteosarcomas because the vast majority of tissues we get are indeed from patients.
NS: You have been studying copy number variations and alterations in targetable immune checkpoints in osteosarcomas, opening doors for new therapeutic interventions. Could you briefly walk us through what led you to this discovery?
TM: The genome of osteosarcoma is incredibly complex and has relatively few point mutations, but it does have a lot of copy number alterations. That was interesting to me, and so was the fact that patients who are treated with a variety of immune checkpoint inhibitors don’t respond to treatment very well. I started to wonder if there was something hidden in these copy number variations Were there alterations or amplifications of inhibitory immune checkpoints that could potentially confer resistance to checkpoint blockade? After looking at whole genome sequencing data and some older SNP 6.0 array data, we decided to look at targetable immune checkpoint inhibitors such as PD-L1, PD-L2, IDO1, and B7-H3 as these would be expressed on the tumor cells. What we discovered is that there were a handful of osteosarcomas that demonstrated somatic copy number gains at the PD-L1/PD-L2 locus, the IDO1 locus, and the B7-H3 locus. What was really interesting was that these were largely non-overlapping subgroups of specimens. The frequency of these alterations at each locus was about 8-9%, so cumulatively you have about 25% of osteosarcoma specimens that have copy number alterations for these checkpoints.
NS: Why did you choose NanoString technology to study the microenvironment of these tumors?
TM: There were two specific reasons for our choice. The first was that the vast majority of samples we have access to are archived FFPE specimens. The most common way to prep a bone tumor for sectioning is to decalcify the specimen using a very strong acid, which severely compromises the integrity of nucleic acids. Using traditional RNA sequencing methods would have generated a low complexity library with a lot of noise. I was drawn to the NanoString hybridize-capture approach. Rather than try to do any kind of amplifications, we would have two different probes that hybridize directly to the transcripts. That allowed for a much more sensitive quantification of the genes of interest. This leads me to my second point that traditional RNA sequencing would not have been sensitive enough to capture our transcripts of interest. When you have tumors that don’t have robust lymphocyte infiltration, these transcripts are so low in abundance that they usually get filtered out by the assay noise. By using NanoString’s direct hybridization approach, we saw an increase in the signal-to-noise ratio, and we accurately and reproducibly quantified the expression of scarce transcripts coming from the diffuse lymphocyte population. The NanoString assay just seemed the best choice considering the complexities of working with bone tumors.
NS: What did you learn about the microenvironment of these tumors using this technology?
TM: We discovered a few things when we studied high-risk patients who have failed the standard-of-care chemotherapy regimen. The microenvironment of the primary tumor versus the metastatic tumors was very different. In samples of lung metastases, we noticed that there was a considerable lack of infiltrating T cells. The T cells were confined to the periphery of the metastatic lesions and did not infiltrate the core of the lesions. That explained a lot about why these patients did not respond to checkpoint therapy– there were no target T cells present for the therapy to act upon. As a result, we’re now focusing on how to get T cells to infiltrate these metastatic tumors.
NS: Your research showed that the 18-gene Tumor Inflammation Signature (TIS) embedded within the PanCancer IO 360™ Panel did not distinguish between metastatic and non-metastatic cancer despite the fact that non-metastatic cancer is characterized by lymphocyte infiltration. How do you explain this?
TM: I wish I had an explanation for that—it would be an entirely new paper! I do think about a few things, though. The first is that the TIS was derived from studies in carcinomas; sarcomas are entirely different in origin and location. The fact that sarcomas are a very different tumor type may be the underlying reason why using a TIS score developed by studying epithelial cell tumors may not be relevant to a sarcoma. The second thing is that I’m not certain that an 18-gene signature is enough to capture the complexity of osteosarcoma. These primary tumors grow in the bone and often invade the medullary cavity of the bone, where we have lots of immature immune cells. This is a unique environment that won’t be found in any other epithelial-derived primary tumor. That leads to a very complex microenvironment, and it’s likely that this setting is not captured by or reflected in the 18-gene TIS. We’re very interested in comparing and revising this epithelial-derived signature to see how we might make it more useful for different cancer types like sarcomas or for cancers that might metastasize and traffic to a bone, such as breast cancer or prostate cancer. Is there something about the bone environment that might result in a different and better predictive signature? That’s what we are looking to find out.
NS: What other gene signatures in the IO 360 Panel caught your attention when you reviewed the data from your study?
TM: The observation about leukocyte trafficking was very important because it was the first thing we started to think about how few T cells were able to penetrate the metastatic microenvironment. After that, we wanted to focus specifically on the vasculature to see what we could learn about its role using the IO 360 Panel. The panel did the job incredibly well, and we were able to supplement the panel with additional downstream profiling. That was extremely useful for us, and it is something we plan on utilizing as we move on to DSP with the GeoMx system. We’re excited to try and generate some of our own gene signatures and explore new regions of interest. The IO 360 Panel enabled us to generate a lot of data, and we are well-positioned for these next lines of investigation. The GeoMx platform will be very powerful! We’ve been collaborating with the teams at NanoString for about two years now, looking at various cancer types. With osteosarcoma specifically, we are interested in applying our home-brewed gene signatures to be able to generate data and push the limits of the technology. We want to see how much multiplexing we can actually do. We’re trying to maximize what we can do because these spatial approaches are the future. The amount of data that we can obtain just by being creative is going to be unparalleled. Combine nCounter and GeoMx DSP with NGS-type approaches, and we will have a highly integrated and very powerful technology pipeline to answer a lot of questions that we could only dream about before. That will be the fun part, and we’re looking forward to working closely with NanoString to make it happen.
FOR RESEARCH USE ONLY. Not for use in diagnostic procedures.