Are Some Tumors Just ‘Born to Be Bad’? – NIH Director's Blog

Are Some Tumors Just ‘Born to Be Bad’? – NIH Director's Blog

Are Some Tumors Just ‘Born to Be Bad’?

Posted on May 22nd, 2018 by Dr. Francis Collins

Caption: Human colon cancer cells.
Credit: National Cancer Institute, NIH

Thanks to improvements in screening technologies and public health outreach, more cancers are being detected early. While that’s life-saving news for many people, it does raise some important questions about the management of small, early-stage tumors. Do some tumors take a long time to smolder in their original location before they spread, or metastasize, while others track to new, distant, and dangerous sites early in their course? Or, as the authors of a new NIH-funded study put it, are certain tumors just “born to be bad”?

To get some answers, these researchers recently used genomic data from 19 human colorectal tumors (malignant and benign) to model tumor development over time [1]. Their computer simulations showed that malignant tumors displayed distinctive spatial patterns of genetic mutations associated with early cell mobility. Cell mobility is a prerequisite for malignancy, and it indicates an elevated risk of tumors invading the surrounding tissue and spreading to other parts of the body. What’s more, the team’s experimental work uncovered evidence of early abnormal cell movement in more than half of the invasive tumors.

Much more remains to be done to validate these findings and extend them to other types of cancer. But the study suggests that spatial mutation patterns may someday prove useful in helping decide whether to pursue aggressive treatment for early-stage cancer or opt for careful monitoring instead.

A tumor begins when a single cell accumulates mutations that cause it to divide when it shouldn’t. From this founding tumor cell, other abnormal, relentlessly dividing cells arise that pick up additional mutations. Because these initiating mutations drive the development of cancer cells and give them a selective advantage to survive, they will likely persist over time and space as a tumor grows larger.

The NIH-supported study, recently reported in Proceedings of the National Academy of Sciences, was led by Marc D. Ryser, Duke University Medical Center, Durham, NC, and Darryl Shibata, University of Southern California Keck School of Medicine, Los Angeles. The study emerged from their awareness that cancer cells, like people, have unique genetic ancestries. While people assemble their family trees poring over old documents, researchers can do the same for cancer cells by rigorously analyzing the patterns of genetic mutations.

To build a tumor family tree, the researchers developed a computer simulation to model tumor evolution. They began with an initial virtual cancer cell in the colon and tracked the predicted expansion of its progeny until the tumor contained about 10 billion cells. Colorectal cancers were ideal for the study because their glandular structure physically partitions a cancer into distinct neighborhoods of related cells.

The researchers “grew” those virtual tumors under different virtual conditions and then randomly sampled them along the way to capture the predicted spatial patterns of genetic mutations. Those models allowed the researchers to predict the expected distribution of developing cancer cells within the tumors based on their differing degrees of early motility, risk for invasion, and metastasis.

Meanwhile, back in the real world of the lab, the researchers also sequenced whole exomes (the 2 percent of the genome that contains protein-coding genes) from DNA samples that had been pooled from opposite sides of colonic glands, or crypts. Four of the pooled samples were from surgically removed benign colorectal adenomas, while 15 were from excised adenocarcinomas, an invasive form of colorectal cancer.

The researchers then tested their simulated data against the actual sequence data from those 19 tumors. The simulations had predicted that, in benign tumors, cells carrying new mutations are unlikely to move from one side of a growing tumor to the other. As a result, new mutations would tend to cluster in distinct parts of the full tumor, with little evidence that cells from different parts of the early tumor had moved from one side of the gland to the other.

In contrast, the simulations predicted that in more malignant cancers, cells would have greater mobility and could spread more widely, even when a tumor was still quite small. This proposed scattering of cells would cause early-arising mutations to intermix throughout as the tumor grew larger.

And, indeed, that’s exactly what the genomic data showed. None of the benign cancers showed unique mutations on opposite sides of a tumor, indicating that the cells tended to stay put without much intermixing. In contrast, in nine of the 15 more invasive tumors, the spatial patterns of genetic mutations indicated cells within the tumor moved around considerably during the early growth phase and multiplied.

The findings are consistent with Shibata’s earlier proposed “Big Bang model” of colorectal cancer growth [2]. Similar to the theory that the universe started from a single point and exploded outward, Shibata’s model posits that cancers arise from a single, highly capable progenitor cell, which then expands. The new study shows it’s possible to go back in time to infer particular traits associated with the progenitor cell based on the spatial patterns or “footprints” within a tumor observed only later.

If validated in further studies, those footprints might be used to predict the risk of invasion and metastasis in small, developing tumors. To extend the findings, the researchers say they are now pursuing similar studies in breast cancers. This work and other research now underway is helping to point the way to more precise approaches that will ultimately help more people with benign, early-stage tumors to avoid the risks associated with treatment—when it’s safe to do so.

References:

[1] Spatial mutation patterns as markers of early colorectal tumor cell mobility. Ryser MD, Min BH, Siegmund KD, Shibata D. Proc Natl Acad Sci U S A. 2018 May 14.

[2] A Big Bang model of human colorectal tumor growth. Sottoriva A, Kang H, Ma Z, Graham TA, Salomon MP, Zhao J, Marjoram P, Siegmund K, Press MF, Shibata D, Curtis C. Nat Genet. 2015 Mar;47(3):209-16.

Links:

Colorectal Cancer (National Cancer Institute/NIH)

Darryl Shibata (Keck School of Medicine, University of Southern California, Los Angeles)

Marc D. Ryser (Duke University, Durham, NC)

NIH Support: National Cancer Institute