Kenneth Chien, Massachusetts General Hospital

While most therapies aim to stall the advance of debilitating disease, regenerative medicine strives to reverse it. Kenneth Chien of the Harvard Stem Cell Institute and the Cardiovascular Research Center at Massachusetts General Hospital describes how stem-cell biology will aid this goal and the lessons that can be gleaned from other fields of medicine.

(Note: This is part of a series of interviews conducted to accompany Nature Insight Regenerative Medicine .)

Is there a difference between regenerative medicine and stem cell biology?

Regenerative medicine as a goal lies beyond stem cell biology. Stem cell biology is necessary, but not sufficient.

What we have learned from the clinical studies is that there's still a lot of work to be done. You can view the cup as not half full — perhaps an eighth full, which is better than nothing. You can view this as relatively early days, similar to gene therapy. We know what we'd like to do, but the technology just hasn't caught up yet.

How long will that take?

Sometimes in looking forward it's good to look back. In cardiac regenerative medicine, probably the only clear success to date is heart transplantation. From the initial grant that Norman Shumway received in 1958 [to study the possibility of heart transplantation] it took more than two decades before the procedure became routine.

Shumway was a careful, thoughtful man. He not only didn't do the first heart transplant; he didn't do the second. He was slowed down in the United States because of the regulatory barriers and ethical concerns. Christiaan Barnard, on the other hand, went back to South Africa and decided to just go for it. Sounds familiar?

We realized very quickly that this was not working, that the science was not there. In 1968, a year after his first attempt, Barnard gave up on the procedure and considered it a failure. Everyone gave up, except Shumway. He went back to the lab and spent the next ten years figuring it out. He realized that the issue was rejection.

Then he showed that he could get successful transplantation in large animals. That was key. It wasn't one of these 'let's just try it out in patients' kind of things.

So it's going to be a long time before stem cells are used for regenerative medicine?

I believe that in two years we will see very good models of human disease based on human embryonic stem (ES) cells1. In the traditional mode, you work your way up through larger and larger animals and end up eventually in man, but there could be a paradigm shift. Humans may be the model organism.

For example, if you can get reprogrammed fibroblasts from a patient that you know, you might find a biomarker or a pathway that really tips you off as to whether the patient has a higher risk [of disease progression] . You can then call them back and examine them. My prediction is that this will lead to a resurgence in human physiology. That's going to lead to the first phase of regenerative medicine. Cell-based therapy will be on the heels of this.

Is science ready for that paradigm shift?

While basic and clinical science have been converging, science and medicine have been diverging from a physical and practical and infrastructure standpoint. The stem-cell institutes are beginning to reverse that, but if they don't have a tangible and continuous interaction with the clinic, they probably will not be the leaders in regenerative medicine.

This type of work is a bit like a decathlon. You have to be very skilled in a couple of suits; you can't make it with just one. In moving stem cells to regenerative medicine, having physicians with PhDs or with research experience similar to a PhD is going to be important. The funding for generating these MD PhDs is getting better and so is the funding for training them, but the funding to support them as they become junior faculty has never been worse.

What are the barriers to using stem cells to study disease?

One is their inherent variability. If you take any ES cell line and differentiate it, it will form embryoid bodies, and some of them will be beating . So there is variability even within a single cell line.

But the highest level of variability is actually between the different parental ES cell lines. Doug Melton and I wrote a paper2 showing that cell lines that happened to be very good at making one germ layer, say mesoderm, were not very good at making endoderm. It's an inherent variability that is much larger than we initially thought. We are talking about an order of magnitude.

Why do different ES cell lines behave so differently?

We suspect there are a number of variables, one of which is a developmental difference. Different ES cell lines may have been derived at different timepoints or from different regions of the blastocyst. In addition, lines come from individuals whose genetic backgrounds are quite different, and there are also stochastic and epigenetic differences. Probably the answer is a combination.

How will this variability affect research?

Given that there's an order of magnitude difference between lines in generating a cell of one of the easier lineages, one can only imagine what the variability would be in getting stem cells to differentiate into lineages for extremely rare cell types.

I really believe it is imperative for us to generate new ES cell lines for specific interests and to screen them for their ability to differentiate into rare cell types.

I really believe it is imperative for us to generate new ES cell lines for specific interests and to screen them for their ability to differentiate into rare cell types. Take a human ES cell line that goes readily into your favourite cell type. If you know what gene causes the disease you want to study, you can get that gene in the ES cell, Then pretty soon you can study the gene of interest in diseased cells differentiated from the ES cells and compare them with the parental ES cell line.

Comparing lines from different patients could be very difficult. There could be more differences between the cells independent of any effects of disease mutations.

For induced pluripotent stem (iPS) cells we have all the same problems but an additional layer of variability. There are going to be significant differences even between cell lines from the same patient, because the transforming genes will have inserted in different loci, they will have different levels of expression, and you're selecting for cells at different states of development. Thousands of genes are differentially expressed between iPS cells and ES cells. So it's going to be important to compare iPS and ES cells and make sure that they are the same.