Stem cells have been coaxed, coddled, and cajoled to develop into cells of a particular type. Yet many rebuff even the most earnest advances. So stubbornly do they resist, researchers often talk of stem cells as having minds of their own. But it is unclear how individual cells come to possess their different proclivities. Some have argued that they are set in stone: cells are programmed to express different levels of genes, which in turn makes them more or less likely to respond to a growth signal. Over the past five years, a small group of researchers has been offering a different perspective. Borrowing from physics and chemistry, they have argued that a cell’s genes, rather than being expressed stably over time, are subject to random fluctuations, or noise. In this view, an individual stem cell might respond to quite different growth signals over the course of its life.
Sui Huang, Hannah Chang, and colleagues set out to explore these two alternatives. They analyzed an apparently homogeneous population of stem cells and found they were an idiosyncratic bunch, expressing often wildly varying amounts of the same protein, Sca-1. The researchers separated the cells into three groups—those expressing low, medium, and high amounts of the protein. Low–Sca-1 isolates eventually came to include medium– and high–Sca-1–expressing cells. The same was true of the high–Sca-1 cells—they reconstituted the full spectrum of expression patterns. The findings, which appear in the May 22 Nature, suggest that stem cells share a characteristic of even the most resolute of human minds: a tendency to vacillate.
“We talk about stem cells making cell fate decisions,” said Huang, HMS visiting associate professor of surgery at Children’s Hospital Boston. “It turns out, if you talk about humans making decisions, it’s very similar. Your mind wanders around; you scan the space of possibilities. You think, ‘Should I study medicine or go to law school?’ You may have moods where you are more prone to either one. And then you meet a friend who happens to tell you, ‘Oh, medical school is great,’ and you decide to go. So you scan mental states of possibilities and then you have an external trigger. It’s very similar with stem cells.”
On the face of it, the study neatly explains why most stem cells reject researchers’ efforts to get them to differentiate: only a fraction of cells in a population may display the right genetic frame of mind or expression pattern at any given moment.
But the story gets more complicated when one tries to figure out what that ideal expression pattern looks like. In the course of their experiments, Huang and colleagues found that low–Sca-1 cells were much more likely to turn into a red blood cell precursor when exposed to erythropoietin (EPO), a red blood cell–inducing factor; high-Sca-1 cells were more likely to respond to a white blood cell–inducing factor, granulocyte–macrophage colony–stimulating factor (GM-CSF). But the cells’ receptivity was almost certainly the consequence of more than their Sca-1 levels. In fact, Huang, Chang, and colleagues monitored the entire inventory of expressed genes in each of the three groups and found that the low– and high–Sca-1 isolates differed in more than 3,900 genes. “It was stunning, almost scary, that they are so different,” said Huang. Identifying which of those thousands of genes must be turned on for a stem cell to respond to a red or white blood cell–promoting signal is a daunting challenge.
There is another persuasion problem. Biologists are taught that genes are expressed as a consequence of a precise sequence of events. “They are generally taught to think about what the right molecule or the right signaling pathway is that mediates some response or behavior,” said co-author Donald Ingber, the Judah Folkman professor of vascular biology in the Department of Pathology at Children’s. Yet one of the take-home lessons of the current study is that a gene’s expression can wax and wane randomly. At some point, perhaps as the result of feedback interactions with other genes, many of which are also fluctuating randomly, the cell veers toward a more stable gene-expression profile—what has been termed an “attractor” state.
It was an interest in attractors that drew Huang to stem cells in the first place. He had taken the first step toward proving the existence of attractors in 2005 when he took two sets of cells, exposed each to a different -neutrophil--inducing reagent, and monitored the expression of 12,600 genes. Though the cells initially exhibited very different gene expression patterns, after seven days, their patterns were almost congruent (see Focus, April 8, 2005).
Huang, working with Chang, an MD–PhD student at HMS and Children’s, took a different tack for the stem cell project. Biologists have traditionally assumed that clonal populations of stem cells exhibit the same genes, but in fact, they assume a kind of bell curve when it comes to the expression level of any one gene. Using the well-known stem cell marker Sca-1, Huang and Chang decided to look more closely at the outliers to see if they might be pulled toward the mean. If they did, it might argue for the existence of an attractor.
They sorted a clonal population of stem cells into low–, medium–, and high–Sca-1–expressing cells and found that the highest Sca-1–expressing cells produced 1,000 times more protein than the lowest. And the cells were remarkably different in behavior. Low–Sca-1 cells, when exposed to EPO, were seven times more likely to develop into a specific type of red blood cell while the high–Sca-1 cells were more likely to respond to the white blood cell–inducing GM-CSF.
But the real test of the attractor theory came when the researchers observed the isolates over time. If the low– and high–Sca-1 cells were stable variants, they should maintain their gene-expression patterns rather than be pulled toward the mean. In fact, after about nine days, Sca-1 expression in each of the isolates reverted to the original bell-shaped curve, suggesting the majority of cells had been slowly pulled to the mean level of Sca-1 expression by an attractor. What’s more, the entire constellation of expressed genes, the transcriptome, also reverted to the original bell curve distribution.
“This work is not going to have a huge impact on science and medicine until we can connect this physics-based understanding of how things work to which sets of genes or proteins have to change. What’s the combination on the lock to make a fate switch?” asked Ingber. “If you can do that, you’ll revolutionize drug development, diagnosis, and therapeutics—and it will get everybody’s attention.”
Conflict Disclosure: No conflicts of interest were reported for this research.
Funding Sources: The Air Force Office of Scientific Research (AFOSR)
