NEW YORK (GenomeWeb) – A new study out of the University of Texas Southwestern Medical Center has cast doubt on the widely accepted role of a pair of microRNAs as cell-autonomous tumor suppressors in colon cancer and other epithelial cell tumor types, providing evidence that the anti-oncogenic effects observed in past studies were actually the result of sampling bias.
The research instead points to an indirect role for the miRNAs in epithelial wound healing in the intestine and adds to a growing body of literature demonstrating that many of the small, non-coding RNAs are only active under stress conditions.
The first miRNAs to be identified were found during genetic screening experiments in C. elegans and revealed themselves when their deletion triggered developmental defects in the worms. This helped contribute to the viewpoint that miRNA functioning predominantly occurs during development, according to Joshua Mendell, UT Southwestern researcher and senior author of this week's study.
"What has become clear in the last five years, as more and more microRNAs have been knocked out, especially in mice, is that that is not the case," he said.
Although there are key exceptions, most miRNAs now appear to be part of stress-activated pathways, with the effects of their deletion only exposed upon injuring or taxing an organ system. In keeping with this, miRNAs have in recent years become promising biomarkers and therapeutic targets in human disease, most notably in cancer.
Two of these, miR-143 and miR-145, have been widely studied after early reports of their downregulation in colorectal adenocarcinoma — a discovery that was first reported in 2003 by a team from Australia's Flinders Medical Center and has proven to be highly reproducible by other groups.
Additionally, the two miRNAs have "very potent anti-tumorigenic activity when ectopically expressed," Mendell told Gene Silencing News. "Many labs including ours have expressed them in cancer cells in tissue culture and seen very potent anti-tumorigenic effects" such as reduced cellular proliferation and apoptosis.
In light of these observations, Mendell and his team began a series of in vivo studies to confirm that the loss of miR-143 and miR-145 promoted tumorigenesis, or at least reveal the miRNAs' natural function to understand its relationship to cancer.
As detailed in a paper appearing last week in Cell, the researchers generated conditional knockout mice in which the two miRNAs could be deleted either in the entire body of the animals or in specific tissue compartments.
When the miRNAs were knocked out of the mice entirely, the scientist saw no obvious phenotype, Mendell noted. "The mice develop normally, they live a normal lifespan, and the colon and intestine function completely normally as far as we can tell … [which is] consistent with the emerging view … that many [miRNAs] don't seem to play important roles in development."
Still, miR-143 and miR-145 are highly conserved among all vertebrates, suggesting that they do have a critical function. So Mendell and his colleagues then looked at the effects of the miRNAs' deletion under a stress condition.
Because of the miRNAs' link to colorectal cancer, they focused their experiments on the intestine and provided the mice dextran sulfate sodium (DSS) in their drinking water. DSS ingestion causes an intestinal epithelial injury regeneration sequence that is commonly used to model ulcerative colitis.
In wild-type mice, when the intestine is injured by DSS, the uninjured epithelium — the layer of cells covering the intestine — enters a hyperproliferative state that heals the wounded tissue and the animals recover rapidly. However, when the knockout mice were subjected to DSS, most died quickly.
"That was a big surprise to us," Mendell said, noting that they had expected better wound healing since this process is characterized by enhanced cellular proliferation — a hallmark of cancer that would be expected to increase in the absence of tumor suppressors.
Upon closer inspection, the investigators discovered that, while the mice lacking miR-143 and miR-145 were ostensibly normal before injury, once treated with DSS, their intestinal epithelia were unable to initiate the hyperproliferative state needed for repair.
"It was as if the mice couldn't sense the injury at all and they didn't heal," Mendell said.
To figure out what was happening, Mendell's team began looking at the expression of miR-143 and miR-145 in other intestinal cell types.
Although colorectal cancers arise from epithelial cells, intestinal tissue is composed of a number of other supporting cell types collectively known as the mesenchymal compartment. And while it was assumed that the miRNAs must be expressed in the epithelium if they were acting as tumor suppressors, Mendell's team studied their expression in the entire mesenchymal compartment both before DSS injury and after.
Unexpectedly, they found that the miRNAs are not expressed at any detectable levels in epithelial cells, but rather are highly and exclusively expressed in supporting mesenchymal cells.
"That was a big surprise because it essentially told us that the microRNAs cannot function as cell-autonomous tumor suppressors because that implies their loss of function in the epithelial cells promotes cancer," he explained. "But they are not in the epithelial cells to begin with."
In collaboration with surgeons at UT Southwestern, Mendell and his team obtained human colorectal cancer samples and paired normal tissue for further examination. When miRNA levels were measured in whole tissue, they were found to be lower in tumors. But when epithelial cells were purified from both the cancerous and normal tissue, the miRNAs were absent in both cases.
This finding suggests that previous observations of abnormal miR-143 and miR-145 expression in cancer were the result of sampling bias during tissue isolation, according to the Cell report.
"Tumors arise from epithelial cells, so tumor tissue itself is highly enriched for the epithelial cells and relatively depleted for the mesenchymal cells as compared to normal tissue," Mendell said. "So microRNA levels are going to reflect that cell composition."
This finding also led the UT Southwestern team to another surprising conclusion — that the miRNAs are helping control epithelial wound healing, but from the supporting mesenchymal cells.
Taking advantage of the conditionality of the miRNA knockouts in the mice, the scientists found that deletion of the miRNAs in mesenchymal cells triggered the intestinal regeneration defect observed in their previous experiments. When miRNA loss of function was induced only in epithelial cells, the wound-healing process was unaffected.
This observation suggested that the epithelial injury response is guided by signaling between the epithelium and the supporting mesenchymal cells, and that the signal is somehow controlled by the miRNAs.
The researchers ultimately identified a protein called insulin-like growth factor binding protein 5 (Igfbp5) as at least one part of the intracellular signal. Insulin-like growth factor (IGF) is known to be involved in intestinal growth and repair, and IGF binding proteins such as Igfbp5 typically function to competitively inhibit IGF signaling.
"We think when these microRNAs are knocked out, [Igfbp5] is derepressed because it is no longer controlled by the microRNAs," Mendell said. "So it is excessively secreted in these wounds and prevents the normal signaling through the IGF pathway," thereby interfering with the normal IGF healing process.
Mendell said that he and his team don't think that Igfbp5 alone is responsible for the phenotype observed with miR-143/miR-145 deletion, and that they are currently exploring other targets of the miRNAs that might be involved.
"But we feel this is at least one component of the signaling that's abnormal in the absence of the microRNAs," he said.