Colorectal cancer is on the rise, particularly among younger adults. Thus, the recommended age at first screening has shifted downward from 50 to 45 for some patients.
Additional advances, however, have been stymied by a continuing lack of detail concerning the disease’s basic mechanisms, including how it progresses on a cellular level.
The Human Tumor Atlas Network aims to change that. This National Cancer Institutes-funded Cancer Moonshot effort involves premier research centers from across the nation, each focused on creating a detailed guide, or “atlas,” of a particular cancer.
These guides will chart the cellular, morphological and molecular features at each step of the disease – from precancerous lesions to advanced. They will also be able to capture potentially contributing features of the cell’s surrounding microenvironment, according to Vanderbilt University Medical Center researchers.
“The Human Tumor Atlas Network aims to create the ‘Google Maps’ of tumors,” said Ken S. Lau, Ph.D., director of the Center for Computational Systems Biology at Vanderbilt.
“The Human Tumor Atlas Network aims to create the ‘Google Maps’ of tumors.”
“We’re trying to use next-generation, cutting-edge molecular profiling approaches to look at all the cells in a tumor, how they are organized, how they are communicating with each other, and where they’re located.” Lau said. “We are looking carefully at defining features and the transition steps between the different stages of the progression toward cancer, including what exactly is changing and what might be fueling those changes.”
Vanderbilt University and VUMC are together assigned the job analyzing colorectal cancer through multimodal molecular profiling, such as spatial transcriptomics and multiplex imaging.
Now in its second phase, the project’s goals are so ambitious that the team expects to have to invent some of the technologies necessary to capture the targeted information. Artificial intelligence and big-data mining are filling in the gaps where existing technologies don’t exist yet.
“All of it is with the goal of identifying targets for earlier intervention and developing a better understanding of risk to inform refined precision screening strategies,” Lau said.
Paths from Polyp to Cancer
During Phase I of Vanderbilt’s foray into the Human Tumor Atlas Network project, researchers zeroed in on two primary pathways to colorectal cancer: adenoma polyps and serrated polyps.
While adenomas have been recognized for several decades as precursors to colorectal cancer and their pathway well-documented, serrated lesions were not well understood.
“Our work demonstrated that conventional adenomas followed the well-established knowledge of pathways towards cancer, whereas the serrated pathway was quite different and actually appeared to have a process of metaplasia that’s common in gastric cancer,” said Martha Shrubsole, Ph.D., co-lead of the Cancer Epidemiology Research Program at Vanderbilt-Ingram Cancer Center and whose findings were published in the journal Cell in 2021.
“This is a paradigm-shifting observation that has changed the field. We now see these non-stem cells as the origin of these particular types of lesions.”
“This had not been previously fully characterized or recognized as a contributor to colorectal carcinogenesis.
“We saw that the origin for conventional adenomas does indeed appear to be the stem cell compartment at the base of a colon crypt. For a serrated polyp, it appears to be triggered, most likely from damage at the surface of the crypt, which is quite different.
“This is a paradigm-shifting observation that has changed the field. We now see these non-stem cells as the origin of these particular types of lesions.”
Targets for Immunotherapy
Human Tumor Atlas Network Phase I is complete, and the research team has applied the sequential, stage-focused, microenvironment-inclusive atlas to other aspects of colorectal cancer.
In one follow-up study, the research team looked at later stages of advanced colorectal cancer. They noticed that some microsatellite-stable tumors, traditionally thought of as “immune cold,” are in fact flooded with immune cells. Retrospective analysis uncovered a mechanism that potentially converts “immune cold” tumors to “immune hot” states. This finding laid the groundwork for their proposal that immunotherapy may be an effective therapy for these cancers, which make up about 90 percent of colorectal cancers.
Additionally, the group’s recent publication in the journal Nature built on the Human Tumor Atlas Network effort by taking a similar atlas-like approach and applying it to normal, non-cancerous cells to analyze the transition from normal to precancerous and on to cancerous.
“That is a very difficult problem, because normal cells accumulate mutations at a very slow rate,” Lau said. “So, we used a mutational barcoding approach, introducing synthetic mutations into normal cells to accelerate the timeline, allowing us to do the same sort of temporal and phylogenetic analysis on normal cells that we did with precancerous and cancerous cells in the Human Tumor Atlas Network.”
In doing so, Lau and the research team have gathered evidence that could completely change the way we envision the moment a cell turns cancerous.
“The paradigm in the field is that cancer starts from a single cell, usually from a single stem cell,” Lau said. “People came to this conclusion by looking at cancers and not looking at precancers. What we actually found is these precancerous lesions actually have a polyclonal initiation mechanism, meaning that multiple normal cells come together and cooperate with each other to form the tumor.”
Lau believes this has significant implications for early interventions targeting the prevention of polyclonal tumor development.
Targeting Early-Onset Colorectal Cancer
Human Tumor Atlas Network Phase II began in late 2024. It will take a multimodal 3D-atlas approach, this time with a focus on distinguishing differences between early-onset and late-onset cases of colorectal cancer.
“The highest rate of increase we see in colorectal cancer of any demographic group is among those under the age of 50, which is considered early-onset colorectal cancer,” Shrubsole said. “We see this really concerning increase that’s been a trend for several years, and we’re unsure why it’s happening.”
The Human Tumor Atlas Network aims to understand these trends and inform more precise screening guidelines.
“We have been gathering contextual data – demographics, health behaviors, health histories – from the beginning,” Shrubsole said. “It could potentially help us learn how to prevent some of these cancers or provide people with more informed decision-making context for screening by better defining who is ‘high risk.’”
Much of Phase I, while it pushed the envelope in terms of methods of analysis, took place using 2D images. One of the aims in Phase II is to construct the atlas in 3D, which is where AI technology and big data computational approaches can help the team.
Lau and Shrubsole attribute these advances, as well as the success of the entire Human Tumor Atlas Network work thus far, to the collaborative spirit at Vanderbilt and VUMC.
Robert Coffey, M.D., Ingram Professor of Cancer Research, was a principal investigator on Phase I. He will continue in Phase II alongside Lau and Shrubsole. They will be joined by Jeffrey Spraggins, Ph.D., of the Mass Spectrometry Research Center.
“This is very much a transdisciplinary project where we have investigators and leaders from multiple scientific disciplines, from population-based research, epidemiology to computational biology, bioinformatics, molecular biology, surgeons involved in research, gastroenterology,” Shrubsole said. “That team and the collaborative spirit and possibilities at Vanderbilt enables this type of work, as well as a patient population that is willing to participate in research.”
