She is mostly speaking about embryonic stem cells, but some of her points are the same as what Dr. Centeno has been saying all along. The FDA's GMP guidelines, for instance, do not really apply to stem cell therapy and yet that is what they are requiring of doctors wanting to do adult stem cell therapy in this country. Makes no sense. Time to get that regulatory agency moved out of the way so that we can all benefit from the therapies that are already taking place all over the world with adult stem cells. ICMS already has guidelines in place for safe adult stem cell treatment. Join if you haven't www.safestemcells.org
Nature Reports Stem Cells
Published online: 27 August 2009 | doi:10.1038/stemcells.2009.113
Melissa Carpenter: making stem cells for many, safely
Monya Baker
Barriers must be overcome to move cell therapies from the laboratory to the clinic
Melissa Carpenter runs a cell-therapy consulting business and is a senior scientific advisor to Proteus Venture Partners, an investment group focusing on regenerative medicine. She began her professional career studying neural stem cells at CytoTherapeutics Inc. and then became director of stem cell biology at Geron Corp. just weeks after the first publication about human embryonic stem cells. She worked there for five years before moving first to the Robarts Research Institute in Ontario, Canada, and then to Novocell as vice president of research and development. Monya Baker spoke with her about her work and her recent Nature Biotechnology review on developing safe therapies from human pluripotent stem cell lines1.
What is the biggest challenge facing the stem cell field right now?
Translation of discoveries to therapies.
What needs to happen to meet that challenge?
Generating enough cells to treat lots of patients, and that's where it gets difficult.
If you wanted to make enough cells to treat 5 patients or even 20 patients, you can imagine doing that in a way that's difficult and inconvenient and requires a type of brute-force production. But if you want to make enough cells to treat 100 patients, or 1,000 patients, then you need manufacturing in a very real sense.
That transition is going to require fundamental biology. Cells in a 96-well plate are very different than cells in a bioreactor. And if you're going to treat 1,000 patients, it's going to require bioengineering and all kinds of new technologies. There is going to be discovery involved in developing manufacturing on a large scale.
What we do in the lab right now is so artful, and you have to minimize the art to deliver cells to thousands of patients. That's going to take money, and that's going to take time.
Whose money?
This is where it gets kind of ugly. In many cases, there isn't money to do all the routine optimization that needs to happen to deliver cells to patients. I can give you an example of how difficult this is. At Robarts, I wanted to work on media formulations, ways to grow ES [embryonic stem] cells that were more defined and would give you more consistent cultures. I thought I'd write a grant for this, and I called up NIH [National Institutes of Health, which does fund some research in Canada] and they told me, "Melissa, it's incredibly important, but the NIH funds hypothesis-driven research, not this type of research".
What about start-ups and companies?
Right now with the economy collapsing and investors being risk intolerant, it's very difficult to raise money. What the investors would like to see is proof of concept in human trials as opposed to proof of concept in a mouse. Likely a lot of groups will not be successful in raising money, which is really unfortunate.
What advice do you have for regulatory agencies about getting to the clinic?
The FDA [Food and Drug Administration] has done a remarkable job working with diverse groups trying to understand how to make their processes applicable to patients. It's really a bidirectional process. The investigators are educating the FDA about the science and the product, and the FDA is educating the investigators about how to achieve safety and how to get a product configuration that is suitable for clinical application.
When new guidelines or recommendations are being put forth, it would be good to keep in mind not just the research process but the translation process. I remember looking at the GTP/GMP [good tissue practices/good manufacturing practices] guidelines and thinking some of them weren't practical for the generation of hESC [human embryonic stem cell] therapies.
Can you give me an example?
Say you're going to start with an embryo. The embryo will have been made in an IVF [in vitro fertilization] clinic. The first problem is that most IVF clinics don't operate under strict GMP. Also, most of the embryos used for the derivation of hESCs have been frozen for many years, so acquiring patient histories and blood samples from the donors may not be possible or relevant at the time of derivation. So from the beginning, you're not compliant with GTP/GMP.
So what do you do?
You go back and you look at the spirit of the guidelines, and the spirit of the guidelines is around safety and testing and document control and testing for adventitious viruses and all of the things that ultimately have to do with patient safety. So you look at the process that you're going through and you try to build in safety, and you do it in ways that make sense for your process and you work with FDA to make sure that you are fulfilling the spirit of the guidelines.
In your review, you say the FDA can't just come up with a minimum allowable number of unwanted, undifferentiated cells in therapeutics. Why not?
What we know about ES cells is that different cell lines are different; they have different stabilities and abilities to differentiate. I've handled a lot of the lines, and you can tell.
And what that translates to is that it's very likely that different human ES cell lines will have different probabilities of making teratomas [a characteristic tumour formed by pluripotent cells]. We also know if you completely dissociate them as single cells you won't get as robust teratoma formation. If you inject them in clumps, then you have a higher likelihood of getting a teratoma. So we know that there are handling issues.
We also know if you put them into the subcutaneous space it's going to take more cells than in the kidney capsule, which has more growth factors and is highly vascularized [and so has a plentiful supply of blood]. So, your next question is where do I put [the cells] to test tumour formation? Do I put them into a permissive site [of the body], multiple sites, the site where I'll deliver my cell product? If I am delivering cells to the heart, many of those cells are going to go into the bloodstream, and those cells are going to go all over the body.
The cell product is not just going to contain my target cell, the cell that carries the functional benefit, but other cells too. And those other cells are going to be impacting that product. So it's not just the tissue environment, it's the microenvironment provided by the heterogeneous cell product.
What do you mean by heterogeneous cell product?
hESCs [human embryonic stem cells] don't differentiate absolutely homogeneously. So, your cell product will contain your cell that is your active ingredient, which is going to confer the functional benefit. And then you have other cells: you have cells that might be accessory cells that are providing support for your target cell; there might be inappropriate cells, undifferentiated ES cells. And there might be 'bystander' cells ? we don't know what they're doing, but they are there and you can measure them. So when you do your quality analysis for your cell product, you are going to want to measure those components and look for consistency.
And all that is in vitro testing. You say some of the most important properties, tumorigenicity and immunogenicity, are also the hardest to test in animals.
Tumorigenicity can only be tested for the life span of the animal. Say you are going to deliver 100 million cells to the patient. What you want to know is whether any contamination with, say, undifferentiated cells, will result in a tumour. Let's say your in vitro assays can detect 0.1% of your population as undifferentiated ES cells. That's the limit of the assay. Then you have to assume that, for your 100 million cells, 0.1% or 100,000 cells may be undifferentiated hESCs. What you want to know is, if you inject 100,000 cells into an animal, will you get a tumour? Let's say you put your cell product into an immunocompromised rodent and you wait a year ? that's about as long as you can go in an immunocompromised animal. If the animal doesn't have a tumour at the end of the year, what does that result mean? Does that mean your cell product is safe? Or that it simply takes two years to make a tumour? Or five years?
If you go into a large animal [with human cells], it's going to be immunosuppressed. Immunosuppression is a pretty nasty thing and still incomplete. So if you see a negative result, is that because [the cells] were immunorejected or because your immunosuppressive drugs adversely affected your graft? What does a negative result mean?
Making a mouse ES cell and putting it into a mouse is probably not going to answer the question of putting human cells into a human. If you start thinking about making [monkey ES cells to test in monkeys] you can very quickly spend all your money making a really brilliant therapy for primates.
The only place you're going to be able to do the definitive experiment is in patients. Therefore it is critical to balance the risk to the patients with the possible benefit.
Are companies reinventing the wheel as they try to answer these questions? What can be done to try to prevent redundant work?
I don't think it's realistic to expect competing companies to share proprietary data about their products. The companies are struggling with similar problems, and they really are competing to get to the answers first. So I don't know that they are going to share that data. That information is costly.
You say hESC products are going to need to be evaluated throughout many steps. Can you describe what those steps are?
The way most people envision hESC therapies is you start with ES cells and you differentiate them, you get some population, you put that population into a patient. So, you're going to need to assess your ES cells, and then you're going to need to assess your cell population throughout the differentiation process, and hopefully your assessment will be predictive.
Let's say you have a 20-day differentiation process, but you know 5 days after the differentiation process begins that you wouldn't get a cell population that would function in an animal unless they had a particular marker. That would be great, because then after 5 days you would know that it wasn't going to work, and then you would stop, and that would save you a whole bunch of money.
Assessing your hESCs for stability will be extremely important as well, and there's another component of concern. Say you make neurons, and you put them into somebody's brain: Would those neurons be stable? How would we know if they aren't stable? Would an experiment that lasts for a year be long enough? We just don't know; we're still accumulating data.
What about induced pluripotent stem (iPS) cells?
hESCs are setting the groundwork; a lot of the assessment is getting sorted out, like the ways in which we can measure tumorigenicity. What we're figuring out is the framework of how you make a cell therapy from pluripotent stem cells, while [at the same time] we're figuring out the fundamentals of iPS cells, how to make them consistently, how to make them without using genes.
What's going to happen is that iPS cells will be accelerated by this foundation of work on hESCs, but there are many things to be sorted out with iPS cells before we can consider making them into a therapeutic entity.
You were director of stem cell biology at Geron but went to the Robarts Research Institute in Canada as an academic when Geron turned toward product development. Why?
I wanted to stay more on a research track and less of a management track, and I had this desire to be a professor. When I was at Robarts, I was recruited to Novocell. I ended up spending a year working at both places. It seemed like a really good idea when I signed for that, but it became impractical.
At that point you decided you wanted to be on the management side rather than research?
Yeah. It's really cool to be in the lab. I really like working the data and handling the cells myself, but I could have a greater impact if I took on a manager's role.
How did you make the transition to consulting?
I wish I could say that it was a great strategic plan, but it just sort of started. Consulting is really interesting; you get to see how different companies work. How it works well, what happens when it doesn't work well, how to help them get around that, how to help them produce, and I'm exposed to lots of interesting technologies.
What advice do you have for young scientists?
My advice is to always look at opportunities even if they don't fit in your plan because they could take you in some very interesting directions that you couldn't have predicted.
And do things that are really fun. You should be asking questions that you find very exciting. You should be enjoying how you get through your day. Science is hard. There are lots of failures, a lot of experiments that don't give you the answers you want. If you don't enjoy the actual process, you're not going to like the job.
Related articles
1. Carpenter, M. K., Frey-Vasconcells, J. & Rao, M. S. Developing safe therapies from human pluripotent stem cells Nature Biotechnol. 27, 606?613 (2009). | Article
Nature Reports Stem Cells
Published online: 27 August 2009 | doi:10.1038/stemcells.2009.113
Melissa Carpenter: making stem cells for many, safely
Monya Baker
Barriers must be overcome to move cell therapies from the laboratory to the clinic
Melissa Carpenter runs a cell-therapy consulting business and is a senior scientific advisor to Proteus Venture Partners, an investment group focusing on regenerative medicine. She began her professional career studying neural stem cells at CytoTherapeutics Inc. and then became director of stem cell biology at Geron Corp. just weeks after the first publication about human embryonic stem cells. She worked there for five years before moving first to the Robarts Research Institute in Ontario, Canada, and then to Novocell as vice president of research and development. Monya Baker spoke with her about her work and her recent Nature Biotechnology review on developing safe therapies from human pluripotent stem cell lines1.
What is the biggest challenge facing the stem cell field right now?
Translation of discoveries to therapies.
What needs to happen to meet that challenge?
Generating enough cells to treat lots of patients, and that's where it gets difficult.
If you wanted to make enough cells to treat 5 patients or even 20 patients, you can imagine doing that in a way that's difficult and inconvenient and requires a type of brute-force production. But if you want to make enough cells to treat 100 patients, or 1,000 patients, then you need manufacturing in a very real sense.
That transition is going to require fundamental biology. Cells in a 96-well plate are very different than cells in a bioreactor. And if you're going to treat 1,000 patients, it's going to require bioengineering and all kinds of new technologies. There is going to be discovery involved in developing manufacturing on a large scale.
What we do in the lab right now is so artful, and you have to minimize the art to deliver cells to thousands of patients. That's going to take money, and that's going to take time.
Whose money?
This is where it gets kind of ugly. In many cases, there isn't money to do all the routine optimization that needs to happen to deliver cells to patients. I can give you an example of how difficult this is. At Robarts, I wanted to work on media formulations, ways to grow ES [embryonic stem] cells that were more defined and would give you more consistent cultures. I thought I'd write a grant for this, and I called up NIH [National Institutes of Health, which does fund some research in Canada] and they told me, "Melissa, it's incredibly important, but the NIH funds hypothesis-driven research, not this type of research".
What about start-ups and companies?
Right now with the economy collapsing and investors being risk intolerant, it's very difficult to raise money. What the investors would like to see is proof of concept in human trials as opposed to proof of concept in a mouse. Likely a lot of groups will not be successful in raising money, which is really unfortunate.
What advice do you have for regulatory agencies about getting to the clinic?
The FDA [Food and Drug Administration] has done a remarkable job working with diverse groups trying to understand how to make their processes applicable to patients. It's really a bidirectional process. The investigators are educating the FDA about the science and the product, and the FDA is educating the investigators about how to achieve safety and how to get a product configuration that is suitable for clinical application.
When new guidelines or recommendations are being put forth, it would be good to keep in mind not just the research process but the translation process. I remember looking at the GTP/GMP [good tissue practices/good manufacturing practices] guidelines and thinking some of them weren't practical for the generation of hESC [human embryonic stem cell] therapies.
Can you give me an example?
Say you're going to start with an embryo. The embryo will have been made in an IVF [in vitro fertilization] clinic. The first problem is that most IVF clinics don't operate under strict GMP. Also, most of the embryos used for the derivation of hESCs have been frozen for many years, so acquiring patient histories and blood samples from the donors may not be possible or relevant at the time of derivation. So from the beginning, you're not compliant with GTP/GMP.
So what do you do?
You go back and you look at the spirit of the guidelines, and the spirit of the guidelines is around safety and testing and document control and testing for adventitious viruses and all of the things that ultimately have to do with patient safety. So you look at the process that you're going through and you try to build in safety, and you do it in ways that make sense for your process and you work with FDA to make sure that you are fulfilling the spirit of the guidelines.
In your review, you say the FDA can't just come up with a minimum allowable number of unwanted, undifferentiated cells in therapeutics. Why not?
What we know about ES cells is that different cell lines are different; they have different stabilities and abilities to differentiate. I've handled a lot of the lines, and you can tell.
And what that translates to is that it's very likely that different human ES cell lines will have different probabilities of making teratomas [a characteristic tumour formed by pluripotent cells]. We also know if you completely dissociate them as single cells you won't get as robust teratoma formation. If you inject them in clumps, then you have a higher likelihood of getting a teratoma. So we know that there are handling issues.
We also know if you put them into the subcutaneous space it's going to take more cells than in the kidney capsule, which has more growth factors and is highly vascularized [and so has a plentiful supply of blood]. So, your next question is where do I put [the cells] to test tumour formation? Do I put them into a permissive site [of the body], multiple sites, the site where I'll deliver my cell product? If I am delivering cells to the heart, many of those cells are going to go into the bloodstream, and those cells are going to go all over the body.
The cell product is not just going to contain my target cell, the cell that carries the functional benefit, but other cells too. And those other cells are going to be impacting that product. So it's not just the tissue environment, it's the microenvironment provided by the heterogeneous cell product.
What do you mean by heterogeneous cell product?
hESCs [human embryonic stem cells] don't differentiate absolutely homogeneously. So, your cell product will contain your cell that is your active ingredient, which is going to confer the functional benefit. And then you have other cells: you have cells that might be accessory cells that are providing support for your target cell; there might be inappropriate cells, undifferentiated ES cells. And there might be 'bystander' cells ? we don't know what they're doing, but they are there and you can measure them. So when you do your quality analysis for your cell product, you are going to want to measure those components and look for consistency.
And all that is in vitro testing. You say some of the most important properties, tumorigenicity and immunogenicity, are also the hardest to test in animals.
Tumorigenicity can only be tested for the life span of the animal. Say you are going to deliver 100 million cells to the patient. What you want to know is whether any contamination with, say, undifferentiated cells, will result in a tumour. Let's say your in vitro assays can detect 0.1% of your population as undifferentiated ES cells. That's the limit of the assay. Then you have to assume that, for your 100 million cells, 0.1% or 100,000 cells may be undifferentiated hESCs. What you want to know is, if you inject 100,000 cells into an animal, will you get a tumour? Let's say you put your cell product into an immunocompromised rodent and you wait a year ? that's about as long as you can go in an immunocompromised animal. If the animal doesn't have a tumour at the end of the year, what does that result mean? Does that mean your cell product is safe? Or that it simply takes two years to make a tumour? Or five years?
If you go into a large animal [with human cells], it's going to be immunosuppressed. Immunosuppression is a pretty nasty thing and still incomplete. So if you see a negative result, is that because [the cells] were immunorejected or because your immunosuppressive drugs adversely affected your graft? What does a negative result mean?
Making a mouse ES cell and putting it into a mouse is probably not going to answer the question of putting human cells into a human. If you start thinking about making [monkey ES cells to test in monkeys] you can very quickly spend all your money making a really brilliant therapy for primates.
The only place you're going to be able to do the definitive experiment is in patients. Therefore it is critical to balance the risk to the patients with the possible benefit.
Are companies reinventing the wheel as they try to answer these questions? What can be done to try to prevent redundant work?
I don't think it's realistic to expect competing companies to share proprietary data about their products. The companies are struggling with similar problems, and they really are competing to get to the answers first. So I don't know that they are going to share that data. That information is costly.
You say hESC products are going to need to be evaluated throughout many steps. Can you describe what those steps are?
The way most people envision hESC therapies is you start with ES cells and you differentiate them, you get some population, you put that population into a patient. So, you're going to need to assess your ES cells, and then you're going to need to assess your cell population throughout the differentiation process, and hopefully your assessment will be predictive.
Let's say you have a 20-day differentiation process, but you know 5 days after the differentiation process begins that you wouldn't get a cell population that would function in an animal unless they had a particular marker. That would be great, because then after 5 days you would know that it wasn't going to work, and then you would stop, and that would save you a whole bunch of money.
Assessing your hESCs for stability will be extremely important as well, and there's another component of concern. Say you make neurons, and you put them into somebody's brain: Would those neurons be stable? How would we know if they aren't stable? Would an experiment that lasts for a year be long enough? We just don't know; we're still accumulating data.
What about induced pluripotent stem (iPS) cells?
hESCs are setting the groundwork; a lot of the assessment is getting sorted out, like the ways in which we can measure tumorigenicity. What we're figuring out is the framework of how you make a cell therapy from pluripotent stem cells, while [at the same time] we're figuring out the fundamentals of iPS cells, how to make them consistently, how to make them without using genes.
What's going to happen is that iPS cells will be accelerated by this foundation of work on hESCs, but there are many things to be sorted out with iPS cells before we can consider making them into a therapeutic entity.
You were director of stem cell biology at Geron but went to the Robarts Research Institute in Canada as an academic when Geron turned toward product development. Why?
I wanted to stay more on a research track and less of a management track, and I had this desire to be a professor. When I was at Robarts, I was recruited to Novocell. I ended up spending a year working at both places. It seemed like a really good idea when I signed for that, but it became impractical.
At that point you decided you wanted to be on the management side rather than research?
Yeah. It's really cool to be in the lab. I really like working the data and handling the cells myself, but I could have a greater impact if I took on a manager's role.
How did you make the transition to consulting?
I wish I could say that it was a great strategic plan, but it just sort of started. Consulting is really interesting; you get to see how different companies work. How it works well, what happens when it doesn't work well, how to help them get around that, how to help them produce, and I'm exposed to lots of interesting technologies.
What advice do you have for young scientists?
My advice is to always look at opportunities even if they don't fit in your plan because they could take you in some very interesting directions that you couldn't have predicted.
And do things that are really fun. You should be asking questions that you find very exciting. You should be enjoying how you get through your day. Science is hard. There are lots of failures, a lot of experiments that don't give you the answers you want. If you don't enjoy the actual process, you're not going to like the job.
Related articles
1. Carpenter, M. K., Frey-Vasconcells, J. & Rao, M. S. Developing safe therapies from human pluripotent stem cells Nature Biotechnol. 27, 606?613 (2009). | Article