is gene therapy going down (or do we just play dumb?)

Maybe I am just having a bad day and you'll think otherwise. What is that disappoints me? The idea that another chance was missed, or the idea that these people chose a gene for their gene therapy trials that had a high chance to send everything to hell? Maybe it's more because of the second thing I've said. More below the fold..

I have just found this through Science Daily, and I decided to follow the thread. This is the story:

Journal of Clinical Investigation (2008, March 23). Gene Therapy Can Cause Leukemia In Large Animals. ScienceDaily.

ScienceDaily (Mar. 23, 2008) — Individuals with a number of life-threatening genetic diseases of the immune system have been successfully treated by gene therapy -- that is, they were infused with early precursors of immune cells that had the correct form of the defective gene delivered into them by agents known as retroviral vectors.

The corresponding author, Hans-Peter Kiem, works at the Fred Hutchinson Cancer Research Center in Seattle, while another of the senior scientists behind the paper, Keith Humphries, works at the BC Cancer Research Centre in Vancouver, BC. The reason why I am so pissed off?

Well, the paper the Science Daily article is referring to talks about the effects of gene therapy to try and expand hematopoietic stem cell populations. To sum it up very briefly, trying to expand blood-forming stem cells using gene therapy, and introducing a gene called HOXB4 into these cells using a retrovirus...well, this ends up producing a leukemia in long-lived mammals receiving these stem cells as part of a therapy. About 2 years after being transplanted with cells overexpressing HOXB4, 2 out of 2 dogs and 1 out of 2 macaques contracted leukemia.

Of course, this therapy was previously tried ex vivo first, and in mice then. Through ex vivo studies, it was found that overexpression of HOXB4 grants hematopoietic stem cells the capacity to expand - to multiply. In vivo studies in mice had shown that cells that had received a retrovirus allowing them to overexpress this same gene were in fact able to expand within the mice, and that these mice were not contracting leukemia even after 12 weeks.

The 12-week time point, as well as the nature of the gene being used for potential gene therapy are crucial here, and they are the main two reasons why I am pissed off.

First of all, I owe you an explanation of the time point, and why it matters. Even now, it is still not known how long blood-forming (hematopoietic) stem cells actually live in an adult human. We do know that there seem to be different populations of these cells with different functions: some of them have only limited regeneration potential, and are most probably lineage-restricted (which means, they can only give rise to some cells of the blood, but not all of them). These are also cells that have a more limited "life span" - in the sense that they cannot reproduce for an unlimited time, nor can they repopulate their own population for an indefinite time.

Bona fide hematopoietic stem cells, though, are multipotent and can give rise to all blood cells in an organism. And it is very probable that these cells are able to repopulate their specific cellular compartment (which means, produce more multipotent stem cells) for the lifetime of the individual. Does this mean that they reproduce a lot, while still having a high turnaround, or that they are long-lived cells that, just like mature memory B cells do, end up homing in the bone marrow and living (as well as repopulating their compartment slowly) throughout our life? Tough question - I think we do not know. But evidence points to the possibility that these cells are very, very long-lived.

Now, these long-lived cells are present usually in a very small number. And throughout your life, they might never accumulate those kinds of mutations that eventually lead to the development of leukemia (and mind you, here I am assuming that leukemic stem cells are basically mutated hematopoietic stem cells - which might not always be the case). But sometimes, these cells accumulate a range of mutations. And some of these are able to turn them into leukemic cells - cells that replicate indiscriminately, and often generate other cells that never fully differentiate, and that end up flooding your bone marrow, and impairing hematopoietic function.

The Humphries group, as well as other research groups around the world, studies the leukemogenic potential of specific gene fusions involving certain HOX genes (among which is HOXB4) and some nuclear proteins - I am talking about the HOX-NUP fusions. These fusions are responsible for a small number of cases of acute myeloid leukemia. HOX genes are truly amazing players, especially during early development. Interestingly, we know that HOX fusions can induce expansion of hematopoietic stem cells, and that HOXB4 can induce this even without being in the context of a fusion, but simply while being overexpressed in these cells.

But I am diverging. The main thing you need to get home is that it might take you a long time to accumulate the mutations that lead to leukemia in your stem cell compartment - and in fact, you might never manage to accumulate them in the course of an entire lifetime.

How exactly do you find out whether a group of cells will cause leukemia or not? This is a crucial question for one to answer if one wants to be able to study the small number of cells that are responsible for the generation of leukemic blasts in a patient. The way people normally do this is by testing this cell population in immuno-compromised mice. If these cells are able to reproduce the leukemia, they probably are (or at least some of them are) leukemic stem cells. You would usually realize whether these cells engrafted or not (getting used to their new host and starting making leukemic blasts) by checking the blood and/or bone marrow of these mice at different time points. And you would get the final result by doing a blood/marrow aspiration when these mice have lived about 12 weeks after the initial injection.

But we are talking about finding out about cells that already cause leukemia! Full-blown leukemic cells injected in mice need, depending on their potency, from few weeks to about 3 months to show their nasty potential. But what if the cells still needed an additional "hit", another mutation (or set of mutation) to become fully leukemic? And what if you had already provided them with a competitive advantage over all the other cells?

Of course, HOXB4 was chosen for a reason. While other related genes, such as HOXB9, can actually induce AML if overexpressed, HOXB4 seemed not to be able to do that - at least up to 12 weeks after the injection of mice.

Unfortunately, that was not the case after two years. I know that it was probably worth trying. But I wonder whether this was more about getting exposure and a few cool papers rather than trying to find a viable way to get hematopoietic cells to expand to be possibly used in transplants. Because especially if you know (and study!) HOX genes and their involvement in leukemia, trying to overexpress one of them to make stem cells is the last thing you want to do. Maybe trying to find a way to switch them on and off, so that they can only be switched on during the initial expansion and not throughout the life of the host cells, that might have been a better idea - and there are technologies that actually allow you to do this by using conditional knock-in's. But simply overexpressing, and ectopically express HOX genes in a stem cell population?

Keith & co., you have got to be kidding me...

Here is the citation of the paper, and the commentary in the same issue of the Journal of Clinical Investigation.

Zhang, X., Beard, B.C., Trobridge, G.D., Wood, B.L., Sale, G.E., Sud, R., Humphries, R.K., Kiem, H. (2008). High incidence of leukemia in large animals after stem cell gene therapy with a HOXB4-expressing retroviral vector. Journal of Clinical Investigation DOI: 10.1172/JCI34371

Larochelle, A., Dunbar, C.E. (2008). HOXB4 and retroviral vectors: adding fuel to the fire. Journal of Clinical Investigation DOI: 10.1172/JCI35326


View blog reactions