Break with the old notion that phenotypic variation in cancer cells is often not due to mutations

Most models of how tumors evolve assume that this process is largely based on the genes of cancer cells, and many cancer treatments specifically target disease-associated mutations
.
However, researchers report today (Oct.
26) in the journal Nature that comparing the entire genome sequence data with RNA-seq data from colorectal tumor samples found that the vast majority of gene expression differences between cancer cells could not be explained
genetically.

"So far, much of the work exploring cancer evolution during cancer development has been very much focused on genetics," said Nicholas McGranahan, a computational cancer researcher at the CRUK UCL Centre at the University College London Cancer Institute, who was not involved in the study
.
But according to the new study, "there are so many such alterations that they can't determine a clear (genetic) basis .
This is good research because it highlights some of the limitations of
what we've done before.
”

Computational biologist Andrea Sotttoriva says it's these limitations that prompted him and his colleagues to consider the transcriptome
in cancer evolution.
"Basically, studying genetic evidence doesn't explain everything we see, for example, 'If you just look .
.
.
Among cancer-related genetic mutations, it is not easy to distinguish benign and malignant cancers: in benign cancers, there are just as many
cancer-driven mutations as in malignant cancers.
”

To better understand how cancer cells change at gene expression levels, the team performed whole-transcription RNA-seq and whole-genome sequencing on 27 surgically removed human colorectal tumor samples, eight of which obtained enough data to compare the two types of sequencing
.
Of the 8368 differentially expressed genes included in the analysis, the difference in transcription level could only be traced back to the underlying genes
of 166 genes.
"A large part of the [cancer research] community has always believed that everything is genetically controlled, but these results show that the answer is not strictly genetic
," Sottiva said.

Sottiva says that while some variations may be due to transcriptional noise, as gene expression fluctuates in cells, he suspects that the cell's microenvironment also plays a large role, such as hypoxia or the presence of macrophages affecting transcription programs
.
Cancer cells "can adapt to many different environments.
"

Sottiriva added that this could have an impact
on the emergence of drug resistance.
Typically, he explains, when cancer is ineffective against chemotherapy or targeted therapy, the tumor microenvironment affects transcriptional mechanisms, such as potential autophagy, aging, or quiescence
.
"So that means a lot of evolutionary models need to adapt to this, because a lot of the phenotypic diversity that can lead to drug resistance is not based on genetics
.
"

McGranahan agrees that the significant role of non-genetic variants in cancer cells may affect the effectiveness of treatments, especially if targeted therapies
are selected based on specific mutations found in the cancer.
"Because we see it in genes, we assume it's a driver of tumors
.
" But they are also hinting that in some cases .
.
.
The driver we are talking about is actually a passenger
.
”

McGranahan added that it's important to note that the study involved a batch analysis of tumor samples, meaning some immune cells and other healthy cells were included
.
"Eventually.
.
.
I think single cells (analysis) would be very, very helpful
.
Sottiva agrees that single-cell techniques can provide more insights, but researchers need more samples to get enough data because gene expression changes over time and genome sequences obtained from individual cells with current analysis are "almost always" incomplete
.

Another extension of the study, McGranahan said, is to repeat such studies
in more patients with different types of cancer.
"They're doing well here because they're doing a very deep job, but that necessarily means they're only observing (a few) patients
.
"

Still, he added, the study is an important first step
in considering cancer variants outside the genome.
"Given that evolution acts on phenotypes, not genotypes, we need to start understanding .
.
.
How much does this gene functionally have to do with tumor evolution?"

Sotttoriva
agrees.
"Although genes laid the foundation for the development of cancer, it was not scripted
.
"