DOESN’T Philippe Collas get the message? If he wants to use engineered cells to help treat diseases such as Parkinson’s, diabetes and Alzheimer’s, he should be working with stem cells, right? After all, these multitalented cells are the hottest thing in biology. With the potential to form all sorts of replacement tissues from scratch, their promise seems boundless, their popularity unrivalled.
But Collas has his own agenda. It’s as if he can see the stem cell bandwagon pass him by, but doesn’t really care. As he enthusiastically describes his research over the phone from his lab at the University of Oslo, you can almost picture the anticipation of a eureka moment carving its way into a fleeting half-smile.
That’s because Collas believes he has found another way to create cells for transplants and therapy – and what’s more, cells that are compatible with the patient’s tissue type and immune system. All he needs, he says, is a skin biopsy, and he can teach the cells to become the type of cell that patient requires.
Advertisement
This is an astonishing claim because it challenges one of biology’s most established dogmas. Once an adult cell has developed into its specialised cell type, the theory goes, you can’t force it to switch to become a different type. For this reason, many scientists are highly sceptical about Collas’s ideas. But the evidence that it really could work is starting to trickle in from his lab, and others too.
Although it will be years before anyone knows for sure whether Collas’s idea will form the basis of a feasible therapy, he’s prepared to bet on it and has started a company to develop the method. And because the technique bypasses the need for stem cells of any kind – as well as the controversy surrounding the use of human embryos, eggs and therapeutic cloning – it’s already generating a buzz of excitement among biologists.
The concept is surprisingly simple, and in fact researchers have toyed with it for decades before Collas’s discovery. The fancy name for it is transdifferentiation. It means transforming an already specialised, or differentiated type of adult cell, such as a skin cell, into a completely different type, such as a nerve cell. Collas’s ambition is to turn skin cells into brain cells, immune cells and insulin-producing pancreas cells, in the hope of treating conditions such as Parkinson’s, cancer and diabetes.
Not surprisingly, his vision is highly controversial. Cell differentiation has always been thought of as a one-way street, taking a less-specialised cell towards a specific fate (see Diagram). It all starts with a single blank-slate – the fertilised egg – which contains all the instructions needed to make an organism such as a human. As the cells in the embryo start to divide, they pick up chemical and environmental cues about what sort of cell to turn into. These signals tell the developing cell to switch on the key genes and proteins that will govern its physical shape and biological characteristics. As the embryo becomes a fetus, and the fetus a baby, its cells become more and more specialised. After that, the general idea goes, there is no way back. Once a differentiated cell, always a differentiated cell.
This is why many researchers have turned to embryonic stem cells in their quest to engineer replacement adult cells. If you catch these cells early enough, they can still turn into many different cell types. But the development of treatments based on such cells has been dogged by obvious ethical concerns.
Then there are the famous adult stem cells. The adult body has a reservoir of these cells. They are part of nature’s repair kit, a subset of cells that have already acquired some general clues about their destiny, but can still give rise to more than one final cell type. But even they are fixed in their fates to a certain extent.
The great thing about transdifferentiation, at least in principle, is that it sidesteps the limitations of stem cells, and shatters the long-held idea of differentiation being irreversible. It suggests that you could take a cell of one supposedly fixed identity and turn it into a completely different type of cell. It would be like suddenly opening a bridge between parallel highways that normally never meet.
The first hint that transdifferentiation could indeed happen came from a handful of situations where it occurs naturally. In the eyes of salamanders and chickens, for example, if the lens is removed from the eye, iris cells will begin to divide, lose their pigment and form a brand new lens, even though the two tissues are very different. żěè¶ĚĘÓƵs also found that if rats are fed a particular diet, some of their pancreatic cells can change into liver cells.
But a series of experiments done in the 1980s hinted for the first time that transdifferentiation was more than a mere curiosity and that a number of different cell types could be made to do it. Helen Blau and her team at Stanford University in California were studying how cells become specialised. They devised a way of fusing different kinds of cells together, to see what would happen if the components of cell A got mixed with the components of cell B. They chose to fuse pairs of cells that had originated from different parts of the embryo and whose genetic routes to specialisation were entirely unrelated.
What they saw came as a complete shock. When they fused muscle cells with unrelated cell types such as liver cells, the unrelated cell type would invariably switch on muscle-specific genes. It was as if the fused cell were trying to change its identity to muscle. Blau hypothesised that something in the cytoplasm of the muscle cell, which had mingled with that of the second cell, was able to flip the switch. “It surprised us,” says Blau. “We thought cells closely related to muscle would turn on muscle genes, but not cells that were not related.”
Yet as tantalising as those results were, and even though others were able to reproduce the work, there was a lack of good experimental tools to dig deeper, so the enthusiasm slowly died away. “None of us quite knew how to take it further,” Blau says. As a result, little else was discovered about how cells are able to reprogram their identity. Transdifferentiation research ground to a halt.
Two decades later, many scientists are still unconvinced that transdifferentiation plays any significant role in biology. Many think it is just an aberrant behaviour brought on by culturing cells in artificial conditions. “I think it is still at the very brink of scientific legitimacy,” says Tanja Dominko, a scientist at Advanced Cell Technology (ACT) in Massachusetts. “The general view is that you are unlikely to change a highly specialised cell into another specialised cell,” agrees John Gurdon, a developmental biologist at the University of Cambridge.
But the new work by Collas is challenging these doubts. Partly thanks to his experiments, scientists are beginning to bring up the “T” word again, Dominko says. “We first heard about the Collas stuff about a year ago,” adds David Ayares of PPL Therapeutics in Blacksburg, Virginia. He says scientists from the company came back from a conference very excited about the results. “As we saw the data, it looked feasible.”
Collas’s team published their results earlier this year, describing the first intriguing evidence in decades that changing “fixed” cellular identities may be easier than previously thought. They took human fetal connective tissue cells called fibroblasts and treated their outer membranes with an enzyme to make the cells permeable. Then they mashed up some immune system cells called T cells to release their contents and make a cellular soup, and then added it to the skin cells. An hour later, the team removed the fibroblasts and put them back into cell culture solution.
Identity switch
To their surprise, Collas and his colleagues saw that after just one hour’s exposure to the T cell extract, the fibroblasts began to change their characteristics. They turned off genes normally active in fibroblasts, and activated genes present in T cells. These included genes for cell surface proteins that are typical of T cells. Intrigued, the team used a “DNA array” technique that allowed them to compare the activities of all the genes in treated and non-treated skin cells. They found that the activity of over 120 genes had changed. But exactly how these genes were altering the cells’ behaviour remained a mystery.
When the team repeated the experiment, this time using extract from nerve cells instead, the fibroblasts not only changed their patterns of gene expression, but they also changed shape, extending outgrowths that resembled the fine, branching axons of nerve cells. “It looks quite impressive,” says Gurdon. “It’s remarkable that such a small treatment makes such a big change.”
Even more impressive, these changes in the cells’ behaviour and shape have lasted for months, says Collas. He thinks the results are so promising that he has gone on to co-found a biotechnology firm called Nucleotech in Westport, Connecticut, to capitalise on the technology. The company’s optimistic, albeit speculative, ambition for the technology is this: a patient with diabetes, say, gives a skin biopsy and those cells are grown in culture. The skin cells are then treated with an extract from insulin-producing pancreatic islet cells. The characteristics of the skin cells change, and they start producing insulin. Those cells are then transplanted back in the patient, hopefully providing a lifelong supply of the hormone. And here’s the best part: there’s no possibility of immune rejection because the cells came from the patient in the first place.
Of course, a full therapy will be many years away if it ever works, but Collas claims Nucleotech has already had promising results turning skin cells into insulin-producing cells. And he’s not the only one hoping that the technology will be a success. “We are trying to induce skin cells to become neural cells,” says Dominko, who recently won a small business grant from the National Institutes of Standards and Technology to pursue a very similar transdifferentiation project. Although she won’t be drawn on the exact details of her experiments, Dominko hopes to publish her preliminary results by the end of the year.
And why would a company like ACT, better known for its work with cloning and stem cells, want to pursue transdifferentiation? Clearly because if it really works, transdifferentiation offers an enticing alternative to stem cells. “I think it’s the ultimate answer,” says Dominko. “It’s ethically acceptable to everybody, and you can really customise cells for every patient.”
Not only that, but it may also represent a significant shortcut on the long route to obtaining cells for therapy. Most attempts to develop cells as therapeutic agents focus on stem cells of some sort. The most coveted goal is to find a way of creating stocks of stem cells for each individual – sources of unspecialised cells from which any kind of replacement tissue could be made to order. Although embryonic stem cells are the most versatile cells for the job, they are far from ideal. They have to be harvested from human embryos, and no one has yet figured out how to reliably coax them to form specific cell types.
To get around this problem, other groups are trying instead to reverse the differentiation of adult cells. One idea is to take adult cells and somehow erase their memory by exposing them to cytoplasm from egg cells. PPL Therapeutics and ACT, for example, have ongoing projects of this kind. Another idea is to take adult stem cells, which have some degree of flexibility, and develop them in culture until they become almost like embryonic stem cells. Either way, the hope is that if a cell can be taken back in time, it can be reprogrammed with the identity of choice.
But the common idea in these approaches is that a cell must somehow be regressed to embryonic-like status before it can be reprogrammed as the desired adult cell type. That’s like telling a lawyer who wants to become a doctor that he has to go back to elementary school, relearn his way into college and then apply for medical school. In contrast, transdifferentiation could allow you to take cell A and turn it into cell B with far fewer intermediate steps. It’s like having the lawyer enrol in medical school right away.
Right now, Collas’s main challenge is to convince others that what he sees is real. One of the main sticking points is whether his results really represent the beginnings of a complete reprogramming from one cell type to another, says Azim Surani of the University of Cambridge. He points out that although Collas’s cells switched on their T cell genes, they didn’t switch off all of their original fibroblast genes. This means that what Collas and colleagues observed was only a partial shift to the new cell type. Gurdon agrees: “They may not be seeing a change in cell type. It could be that [fibroblasts] are taking on characteristics of [T cells] but it’s not a complete switch,” he says.
Collas agrees with the cautionary notes, but points out that they aren’t necessarily a barrier to making the technique work. He concedes it could be impossible to completely switch cell types in the test tube. But he speculates that the “soup” treatment may be all that’s needed to jump-start the process. Once the cells are put back in a patient, he says, the environment within the body may prompt the cells to complete the switch. And even if your skin cells don’t make perfect pancreatic islet cells, does it really matter, as long as they make insulin?
Breaking the rules
Besides, Collas adds, this is not the first time that the scientific community has been sceptical about a notion that challenges the accepted dogma. Until Dolly the sheep was cloned, most people didn’t think an adult nucleus could be completely reprogrammed. Now it’s widely accepted that there must be special substances in egg cells or embryonic cells that can erase the nucleus’s memory and reset the cell. So it’s not that much of a leap to suppose that the cytoplasm from an adult cell could do the same under the right circumstances, Collas says.
“Transdifferentiation is in a grey area, but it’s the same grey area that cloning was five years ago,” says Ayares. He adds that Collas’s work “is going to turn some attention” to the field. But however intriguing Collas’s preliminary results may be, the future of transdifferentiation as a viable therapy hangs on proving that it works in the body as well as in the cell culture dish. With this in mind, Collas is now gearing up to test the technique on animals. “Within a year we will be doing experiments with mice,” he says. The researchers plan to implant diabetic mice with rat skin cells that have been transdifferentiated into insulin-producing cells. These experiments will help answer crucial questions about how safe and effective the technology is. For example, it remains to be seen whether the cells will heed their new instructions when transplanted into the body. “Are these cells actually functional, could they integrate into tissues, stay alive, and more importantly, would they not make tumours?” asks Dominko.
The worst case scenario for Collas would be to find out in a few years that the cells don’t work when transplanted back into live tissue. But even then, he will still have developed a new way to study how cells can begin to be reprogrammed. Such a framework is badly needed to understand some of the most basic questions in cell biology. Nobody fully understands what goes on inside the nucleus of a donor cell during cloning, or inside an embryonic stem cell when researchers try to make it form a particular cell type.
For Blau, almost two decades after her first set of fusion experiments, it just feels great to see a field that had almost been forgotten be brought back to the forefront of scientific research. “I’m really excited that people are starting to think about cell reprogramming again,” she says. And for Collas, the idea that something as complex as differentiation could be moulded at will with something as simple as cell soup is nothing short of amazing. “It’s almost too simple to be true,” he says.