One of the great revolutions in modern science rests on the elongated backside of a grotesque, mutant worm. Inexpensive and easy to manipulate in the lab, Caenorhabditis elegans develops from egg to adult in three days and produces a few hundred offspring three days after that. Virtually all of the worms are hermaphrodites, containing both male and female sex organs and capable of making sperm and eggs, so each creature can fertilize itself. And because the worm is transparent and the adult has only 959 cells, development of every stage from egg to adult can be observed under the microscope and documented with near perfect detail while the worm is alive, an achievement accomplished in the 1970s by Sidney Brenner, a University of Cambridge researcher and legend in the field.
C. elegans has been a favorite in biology labs for years due to its transparency, speedy reproductive cycle, and ability to mutate on cue. Just irradiate it or add chemical mutagens to its petri dish, then wait a few days to see what kind of freak worms appear in the progeny. In the late 1970s and 1980s, “worm talks” (as C. elegans lectures were called) inevitably began with a description of development in the normal worm and segued to whatever mutants the lecturer found intriguing.
The “bag of worms” was one such mutant. This version of C. elegans has the singular misfortune of being unable to lay the eggs that it fertilizes. It gets stuck in the earliest stage of wormy development, making the same larval cells repeatedly while failing to form the organs and body parts needed for later life—including the vulva required to get the eggs out of its body. The result is an oversize larva filled with dozens of offspring, hence its nickname. But it does not remain in that state for long. “The eggs hatch inside the worm, and the larvae consume the mother and crawl away,” Victor Ambros, a biologist at the University of Massachusetts, explains.
Ambros first heard about the bag of worms at a 1979 lecture by Robert Horvitz, who had studied C. elegans in Brenner’s Cambridge lab. At the time, Ambros was finishing up a doctoral degree at MIT, studying with Nobel laureate David Baltimore. A few months later he began a postdoctoral fellowship with Horvitz, who suggested he try to identify the defective gene responsible for the grotesquery that was the bag of worms. What followed in fits and starts over a quarter century has evolved into what Baltimore now describes as “a whole new biology.”
It took Ambros 13 years to identify and sequence the defective gene responsible for generating the bag of worms mutant from a normal C. elegans mother; located on chromosome 2 within the worm’s genome, the mutant gene was named lin-4. As it turned out, the gene coded not for a protein—as all genes were then thought to do—but for a tiny snippet of RNA, the simpler molecular cousin of DNA. The RNA molecule was one-hundredth the size of a gene encoding a typical protein, so small it was hard to imagine its having any function whatsoever, let alone producing a mutant as dramatic as the bag of worms.
Ambros’s work on that bizarre mutant provided one of the first signs that RNA might be much more important than anyone had suspected, but not until 2001 did the full story start to unfold. That is when studies finally convinced scientists that the minuscule RNA snippets they had taken to calling “microRNA” were regulating cellular and genetic processes throughout the human body and were critical factors in the determination of health and disease.
The conventional wisdom of what a gene does and how it does it celebrated its 50th birthday just last year. It was in 1958 that Francis Crick of double-helix fame set down the “central dogma of molecular biology,” which could be summed up in six words: DNA makes RNA; RNA makes proteins.
Genes are encoded in the DNA of our chromosomes. They appear as discrete segments of the 3 billion or so pairs of nucleotides, the “letters” of the genetic code that make up the rungs of the double helix. A fertilized human egg begins life with the DNA in its genome, half from the mother and half from the father. From that an entire human being of some 10 trillion cells is programmed. According to the central dogma, this happens as genes are transcribed into RNA, and RNA into proteins. The proteins in turn are the workhorses of biology, spurring chemical reactions inside cells and controlling the expression, transcription, and replication of the genes themselves.
In this picture, RNA—a single-stranded molecule, in contrast to the twin strands of DNA—was perceived as a secondary player, “sort of a slave molecule, copied from DNA in a pretty uninteresting way,” says Philip Sharp, an MIT biologist and Nobel laureate. This belief never wavered, even when geneticists realized that only about 2 percent of the DNA in human cells actually contains genes that make proteins. Most of the remainder was dismissed as “junk DNA,” a term coined in 1972 by the Japanese geneticist Susumu Ohno to capture the notion that most of our DNA is effectively useless, the remains of ancient viruses or now-defunct genes.
With Ambros’s discovery of microRNA came a startling realization: Part of what was considered mysterious junk DNA (and almost 90 percent of all DNA had been so classified) is actually transcribed by the machinery in our cells into bits of RNA that are fundamental controllers of life. Those microRNA molecules have been linked to heart disease and diabetes; to Alzheimer’s, Parkinson’s, and other neurodegenerative diseases; to longevity (at least in worms); and to the entire spectrum of human cancers, including lung, breast, stomach, prostate, colon, pancreatic, and brain.
“People ask me which diseases these RNAs are going to be involved in,” says Thomas Cech, a biochemist at the University of Colorado at Boulder and a Nobel laureate for his work on RNA’s role as an enzyme. “The answer is 100 percent—100 percent of everything that goes on in the body.” The insight demands a fundamental revision in what we thought we knew about how genes work.
Biologists and geneticists now find themselves wondering how they could have missed such a basic aspect of living organisms for so many years. After all, the technology necessary to find microRNA genes has existed since the 1960s. One answer is that they had no particular reason to look. “We bought the paradigm that proteins were the most subtle molecular players in biological systems,” Baltimore says. “We didn’t have a strong feeling that there was something missing, and usually you need a sense that something’s missing to go out and look for it.” Indeed, when Ambros found the first microRNA, he was not looking for it. He had a problem to solve—finding the gene responsible for the lin-4 bag of worms—and the persistence to keep thinking about it when other researchers would have moved on.
Another explanation is that, as with any remarkable scientific discovery, finding microRNA required just the right combination of talent, circumstance, and luck. Ambros found a perfect collaborator in his wife, Candy Lee, who was a lab technician. As Baltimore describes them (having worked with both), they follow the data rather than the scientific fashions; they are both technically adept in the laboratory; and “they have never been ambitious to the point of its getting in the way of reality.” This is not to say that they lack the drive to do good science, but that “they’re not worrying about the trappings of science,” Baltimore says.
Gary Ruvkun, a molecular biologist at Massachusetts General Hospital, has also worked with Ambros and Lee. “Lots of times when you run into people at Harvard and MIT, they’re sort of clenched,” Ruvkun says. “They say, ‘This is my little thing, and I don’t want anybody else to work on it.’ Victor is ‘Hey, let’s crack this thing together!’ Kind of wide-eyed.”
Ambros and Lee met in an organic chemistry class at MIT in 1972, when Ambros was a sophomore and Lee was a freshman. They started dating a year later and were married in 1976, the year Lee graduated. They had both gone to MIT expecting to study physics, but Ambros switched to biology and genetics because he believed they were a better fit for his intellectual talents. There was a narrative about those fields, he says, “and no math that you have to be really good at.” Soon Lee was pursuing biology as well. Because she would receive more pay as a laboratory employee than as a graduate student, she never pursued an advanced degree, instead taking a succession of technician jobs in academic laboratories (including Baltimore’s) and the biotech industry. Eventually she went to work in Ambros’s lab—“the family business,” she says.
When Ambros started working with Horvitz at MIT in 1979, studying a worm like the lin-4 mutant carried a high risk of failure. Geneticists usually prefer to study genes that are easy to mutate so that work can be replicated or varied, but only one lin-4 mutation had ever been observed, in Brenner’s Cambridge lab a few years earlier. The mutants that Ambros and his colleagues worked with were all descendants of that original line.
Ambros spent his first year on the bag of worms trying to generate another mutation in the lin-4 gene. When he failed, he devoted his attention to other genes involved in the development of C. elegans. One of these, with the confusingly similar name lin-14, had the intriguing ability to “revert” lin-4. Here was the first clue about what lin-4 was and how it worked. Whatever lin-4 did to make a bag of worms, a mutated lin-14 could fix by doing the opposite. If a worm began life with the lin-4 mutation and then acquired a second mutation in lin-14, Ambros found, its progeny would appear to be perfectly normal. The ability of lin-14 to reverse the effects of lin-4 implied that the two genes were coupled together in controlling development, and that the protein product of one gene modulated the other.
For two years Ambros worked with Ruvkun, who was then a fellow postdoc in Horvitz’s lab, to clone the lin-14 gene. In 1984 Ambros launched his own lab at Harvard University, taking the lin-4 project with him and leaving the lin-14 gene with Ruvkun. (Ruvkun in turn took the lin-14 project to Mass General when he launched his own lab there just a few months later. )
At Harvard, meanwhile, Ambros and Lee were joined by a new postdoc, Rhonda Feinbaum. Their goal was to use classical genetic techniques to pinpoint the location of the lin-4 gene on the worm’s DNA, progressively narrowing the relevant stretch of nucleotides until all they had left was the gene itself. The process would eventually take four years. Before it could be finished, however, Ambros paid the price for such an unpromising project: He was rejected for tenure at Harvard. His colleagues in the biology department considered his work of insufficient interest to keep him around. “It was a really terrible, dark time,” he says, “when your ego is just pummeled to nothing.”
In 1992 Ambros took a job at Dartmouth in New Hampshire. By that time, Lee and Feinbaum were realizing that the lin-4 gene was astonishingly small. First the two researchers narrowed the possible location of the lin-4 gene to a sequence of DNA 700 nucleotides long, about one-third the size of a typical protein-coding gene. By 1993 they had localized the gene to a stretch of DNA only 70 nucleotides long and established that this little piece of DNA encoded a still smaller piece of RNA, a mere 22 nucleotides long. Until they realized what they had, “we’d been thinking this was a kind of schmutz,” Ambros says, using the Yiddish word for dirt. “We thought nothing meaningful could be this small…. And remember what this gene is doing: You remove it and throughout the animal all these cells are repeating the larval stage over and over, something really pervasive and profound.”
While Lee and Feinbaum were tracking down the lin-4 DNA in Ambros’s lab, Ruvkun and his colleagues at Mass General had identified the product of their lin-14 gene as a typical piece of messenger RNA, the kind that facilitates the construction of proteins. The critical question was how the two genes interacted. Ambros and Ruvkun got on the phone and started reading out sequences on their screens, Ambros providing the 22-nucleotide-long RNA produced by lin-4, and Ruvkun the protein-coding RNA transcribed from lin-14. The two sequences were a perfect match. In the lingo of molecular biology, the tiny piece of RNA from lin-4 could “base pair” with the lin-14 messenger RNA. Ambros and Ruvkun had no doubt that the lin-4 RNA could attach itself to the RNA transcribed from lin-14, and by doing so adjust the amount of protein ultimately created. The significance was profound. Not only did RNA appear to be acting as a gene, but it was a gene that was directly controlled by another piece of RNA. “It was so pretty, it had to be right,” Ruvkun says.
In December 1993 the journal Cell published Ambros’s article on lin-4 back-to-back with an article from Ruvkun’s lab on lin-14. In retrospect, the two papers launched the RNA revolution and changed the face of modern biology. At the time, however, they were effectively ignored.
The big thing then (as now), Ruvkun says, was for researchers to demonstrate that a gene of interest exists in a spectrum of different species—from roundworms and fruit flies to humans. If a gene is important, evolution keeps it around, and the same gene or its homologues will be found again and again in different organisms. But by 1993, researchers had sequenced only a few dozen genes from fruit flies and humans, and neither lin-14 nor lin-4 matched up with any of them. The assumption, then, was that these genes did not matter much.
“I read the papers,” says Phil Sharp. “I knew these guys, and I didn’t do a damn thing about it. We were all saying this is really interesting, but it is just this one silly gene in worms.” Was the gene important? Was it present only in worms or was it widespread? “It would have been a trivial experiment to answer the questions, but I didn’t get puzzled by it.”
Over the next seven years, only a few people in the world, including Ambros and Ruvkun, pressed on to explore whether the regulatory power of the strange RNA seen in the bag of worms was a one-of-a-kind oddity or something more. Finally Ruvkun’s researchers came upon another grotesque mutant of C. elegans, known as let-7, which literally bursts open in the final larval stage. (The “let” in let-7 stands for “lethal.”) When Ruvkun’s team cloned the let-7 gene, they realized that the product was again a short snippet of RNA, this one only 21 nucleotides long. Clearly the lin-4 RNA gene was not alone.
By 1999, when the Ruvkun lab had finished sequencing let-7, a significant portion of the human and fruit fly genomes had been mapped. Now Ruvkun could look for a version of let-7 in humans and flies—and he found it in both. “Then I wrote to people all over the world who were working on weird organisms, and I asked, Is it present in a mollusk? Is it present in a sea anemone? Is it present in a sponge? And we got Fed-Ex packages back from all over the world. And we could do a quick experiment—boom—and there it was.” Finding the normal version of the let-7 gene conserved across the range of organisms indicated that it was functional and serving some useful purpose in all of them. Although researchers are still trying to learn exactly what it does, some have suggested that in humans, at least, it may control the growth of cells. In 2000 Ruvkun published two landmark papers in Nature. The first [pdf] reported that let-7 encoded a 21-nucleotide microRNA. The second [pdf], nine months later, revealed that his lab had detected let-7 RNA “in samples from a wide range of animal species.”
A single short RNA gene (by now dubbed microRNA) could be dismissed as a fluke. With two, including one present throughout the animal kingdom, researchers began to pay attention. “It was one of those times in your life when you spend 10 minutes completely reorganizing your view of the universe,” Ambros says. “We knew instantly that there were many microRNAs, and they had to be in all animals. It’s insane to imagine that let-7 is the only other microRNA after 500 million years of evolution.”
In 2001, drawn by these remarkable findings, three teams were racing to discover what they guessed was an abundance of microRNAs. One group was led by David Bartel of MIT; another was in Germany, led by Tom Tuschl, who had just finished a postdoctoral fellowship with Sharp. The third, of course, consisted of Ambros and Lee, who, despite their head start, had significantly fewer resources than the others. Still at Dartmouth, they recruited the second of their three sons, then a high school freshman, to do their computer programming.
All three labs used a technique that Tuschl had developed to clone small RNA molecules, and their results—103 new microRNA genes in all—were published together in the journal Science in October 2001. Tuschl had identified 14 microRNAs from fruit fly embryos and 19 from a line of human cancer cells. Bartel had identified 55 from C. elegans, and Ambros and Lee, 15. All three teams realized that these were probably just a very small portion of all the microRNAs hidden in these species’ genomes. There could be thousands more.
A year later, Carlo Croce, now director of the Human Cancer Genetics Program at Ohio State University, reported that chronic lymphocytic leukemia (CLL), the most common form of the disease, was caused by deletion of two microRNA genes. That discovery was the result of dogged sleuthing: Croce had spent seven years looking for the genes driving CLL. In about 70 percent of the cases, he found a dislocation in a particular region of a certain chromosome, but at first he could not find any protein-coding gene responsible. Once the new microRNA genes were identified, it turned out that two of them mapped to this region of the chromosome. The realization that mutations in microRNA genes might cause a common form of cancer was “a revelation,” Croce says. “It indicated that a totally new class of gene could play a major role in the disease.”
Soon other researchers began to find more connections between microRNA and cancer, and microRNA research began to spread through medicine and biology—in David Baltimore’s words, like “an infection in the laboratory.” At the Dana-Farber Cancer Institute and the Broad Institute, for instance, Todd Golub, a specialist in the genetics of cancer, had spent a decade learning how to identify malignant tumors by the pattern of protein-coding messenger RNAs expressed in their cells. The technique, which captures a molecular fingerprint of the disease, can be used to customize therapy to best fight a patient’s particular type of cancer. Golub paid little attention to microRNA research until early 2004, when Horvitz phoned to suggest that they collaborate.
Initially Golub was skeptical that microRNA might play a role in cancer, but the evidence soon erased his doubts. “It was amazingly clear that these things were massively, dynamically expressed in different tumors,” he says. Put simply, the pattern of microRNA genes turned on in one kind of tumor was entirely different from the pattern in another kind of tumor, and entirely different again from that seen in healthy cells. “It was hard to find a single microRNA in the human genome that wasn’t differentially turned on or off across different tumor types in different tissues,” Golub says. Indeed, as he and his colleagues went on to report, the 217 microRNAs that had been identified to date could be more effective at classifying tumors than the 20,000 protein-coding RNAs already used for diagnosis.
MicroRNAs turned out to be useful not only for identifying different types of cancer but also for pointing to potential new treatments. Indeed, Golub found that many microRNAs were less active in cancers—regardless of cancer type—than they were in healthy tissue. In a follow-up collaboration with Tyler Jacks of MIT, Golub asked, “What if I could turn off microRNAs in tumor cells? Would the cells care? And the clear observation was that the tumors grow larger.” The converse action—increasing the activity of the microRNAs—could reduce tumors. Once the mechanism is harnessed, a therapy for cancer might emerge.
It now appears that fine-tuning the production of proteins by increasing or decreasing the activity of microRNAs might lead to therapies for a wide range of diseases. In the new model of biology, proteins still do the hard work of catalyzing reactions and switching genes on and off, but the microRNAs regulate the amount of proteins and hence how much of each specific job is done. At a minimum, the result is an incredibly complex web of cellular signals that are constantly working together (or in opposition, as the case may be) to adjust the number and type of proteins expressed in a cell and so, ultimately, the health of the cell itself. The ubiquitous cloud of microRNA regulation amounts to a forward-looking feedback system. Every time a gene is switched on and sets about being transcribed to make a protein, it might also generate one or more microRNAs that would, in turn, trigger a spectrum of simultaneous adjustments or changes throughout the cell.
The implications are vast. Any disease that has a genetic component not yet identified—a long list that includes Alzheimer’s, schizophrenia, bipolar disorder, obesity, heart disease, and diabetes—might be treated, at least in part, by adjustments in genes that code for microRNA. Whenever researchers ask themselves whether microRNA might play a role in a particular disease or health problem, these days the answer is almost invariably yes, because microRNAs appear to be everywhere, part of the underlying health of organs and crucial to biochemical cascades we only thought we understood.
Take myosin, the major protein in heart muscle and “the most studied protein in muscle biology for decades,” according to Eric Olson, a University of Texas molecular biologist who studies the role of microRNA in heart tissue and heart failure. “We thought we knew everything about myosin —how it works, how it makes muscles contract,” he says. “But all this time, myosin had a secret that we didn’t know. It had a microRNA hidden in one of its introns,” the region of DNA that doesn’t code for proteins. “And that microRNA regulates many fundamental properties of muscle. It controls the ability of the heart to respond to injury and to respond to thyroid hormone, a major regulator of heart function. So all this time, we thought myosin was there to make muscles contract. In fact, it had a much deeper and broader role in muscle biology, by encoding this microRNA.”
Olson and his colleagues are now finding distinct patterns of microRNA in different forms of heart disease and tracing the progression of heart disease to progressive changes in microRNA. They have created mice that are protected from heart failure by the removal of a single microRNA gene from their genome; they have also made mice that are particularly susceptible to heart failure by genetically engineering them to overexpress specific microRNAs genes. In light of all this, microRNAs are obvious targets for drug development and may even be useful as drugs themselves. “Now we know that individual microRNAs regulate many facets of heart disease,” Olson says. “So we can develop strategies for modulating or inhibiting those pathological RNAs.”
Where the RNA revolution is going—other than everywhere—is virtually impossible to predict. Revelations are emerging so quickly that researchers describe it as simultaneously exhilarating and intimidating. Ambros and Lee, who recently moved their laboratory to the University of Massachusetts in Worcester, are amazed at the quality and quantity of talent that has flooded their field, once a backwater of science. And they report the surreal sensation of finding themselves at the top of the game.
“You have certain people in science that you admire, like Phil Sharp and David Baltimore,” Ambros says. “My ambition was that someday, later in my career, they would maybe notice what I work on and maybe say, ‘Good job, Victor.’” That Sharp and Baltimore would both be working in a field that Ambros himself pioneered “was just unthought of, and just immensely delightful.”
Micro RNA at Work
A microRNA gene is just one-hundredth the length of a typical gene. The typical gene codes for messenger RNA. The messenger RNA, in turn, directs the assembly of a protein. Recently, scientists learned that microRNA genes can control this essential process by coding for a microRNA strip that binds to the messenger RNA, effectively turning off production of the protein.