Hype, hope and heresy — or why it is bad to eggsaggerate

Scientists have to be optimistic. How else to survive months or years of frustratingly slow progress in a difficult and often poorly rewarded profession?

But despite the frustrations, scientists retain a responsibility to present their work in a balanced way and not to give unrealistic or false hopes. Unfortunately, several stem cell researchers seem to disregard this all too frequently and the field is particularly prone to hype. It's a risky business — overhyping treatments that turn out not to work out in the clinic leaves patients disappointed and erodes trust in all stem cell research.

Recently, the newspapers were full of headlines about work from the lab of Professor Jonathan Tilly at Harvard. Tilly published a paper (1) reporting the discovery of stem cells in adult human ovaries that could be grown in large numbers in the lab (in vitro) and retain the ability to give rise to eggs (reported in BioNews 646).

For many years, the generally accepted view has been that women have a limited supply of eggs that are generated during embryonic development and are unable to make new ones after birth. Tilly's findings were therefore dramatic and, if true, very important. The cells they claim to have found were offered as a way to study egg cell (oocyte) development in the lab, to allow women to have children after cancer treatment, and as a solution to overcoming premature infertility — all laudable aims. It was also suggested that they could be used to prolong fertility beyond normal menopause — controversial, but an obvious possibility.

I am all for challenging dogma, but to do so requires robust evidence and carefully drawn conclusions. I feel that both were missing in this case.

The view that female mammals have a fixed number of oocytes around birth that cannot be replenished in later life is based on substantial data. For example, we can point to the clear reduction in oocyte numbers as females age, which parallels a decline in fertility. Secondly, scientists had long failed to identify any population of germ cells in the ovary that had not already entered meiosis.

But in 2004, Tilly claimed that there had to be a population of germ line stem cells in the ovaries of adult female mice that could go on to develop into oocytes (2). His reasoning was largely based on work counting oocytes at different postnatal stages, with Tilly concluding that there were insufficient oocytes to account for the normal reproductive lifespan of female mice. But counting oocytes is not as trivial as it sounds, and his analysis was dismissed by others (3).

A subsequent paper (4) suggesting that there were cells in the bloodstream able to colonise the ovary and give rise to oocytes was even more controversial and was disputed by other labs using hard data and robust theory (5).

Tilly changed tack and now claims that factors in the bloodstream can stimulate oocyte development within the ovaries, with the assumption that these new oocytes come from a stem cell population. Such claims were made without clear evidence. What was needed was an independent demonstration of the existence of such cells.

Dr Ji Wu and his team in Shanghai came closest to providing this. They reportedly isolated a subpopulation of cells from mouse ovaries with the ability grow in vitro (6). These cells gave rise to eggs, notably when reintroduced into mouse ovaries. The eggs could be fertilised and some were reported to develop into live-born mice after transfer to surrogate mothers.

However, aspects of this work were hard to understand. In particular, the team's isolation of their ovarian stem cells (OSCs) involved the use of an antibody against a protein (Ddx4) that functions within cells and is not, as far as anyone knows, expressed on the cell surface - despite that being essential for the method to work. The current paper by Tilly and colleagues starts with a more detailed look at the methods used by Wu, notably to see if they can be improved and extended to humans. They claim that they have achieved both these aims. However, there is still no understanding of how the antibody works and whether it really detects Ddx4.

In any case, using this antibody, Tilly identified at best a few hundred OSCs per mouse ovary. Less than one percent of these can be grown in culture, which means that the number of 'germ-line stem cells' per ovary was fewer than four! Their number could not be determined in human ovaries, but it seems they are equally rare.

So, the authors cannot see these cells in the mouse or human ovary, let alone demonstrate that they actually represent stem cells in vivo. Therefore, it is possible that it is the in vitro culture conditions that somehow convert rare ovarian cells into cells with germ-line properties. In other words, their OSCs, which are isolated in a way that is yet to make sense, could also be an in vitro artefact.

Perhaps this doesn't strictly matter for the aims of the research, except subsequent data were also puzzling. For example, a piece of evidence the authors claim proves that they have oocytes forming spontaneously in culture, was the presence of cells with a 1n chromosomal DNA content. This would be typical of male germ cells that have undergone both meiotic divisions to produce haploid (when a cell only contains one set of chromosomes) sperm. However, during oocyte development, the second meiotic division does not occur until after fertilisation — there is no such thing as a haploid oocyte.

Tilly and colleagues claim to show that their human OSCs can also make fully grown oocytes (after being introduced into fragments of human ovary that are then grafted under the skin of mice). For 'ethical reasons' they did not test whether these oocytes could be fertilised or begin normal human embryo development. This means that their normality and potential remain unknown. They did show that the mouse oocytes could be fertilised but why did they not go on to ask (as Wu had done) if the resulting early embryos could give rise to healthy live-born animals?

The human OSCs were isolated from the ovaries of women of reproductive age, but the choice of donor was interesting. These were women who wanted to become men, and had undergone surgery to remove their ovaries as part of the sex-change process. It is usual for such individuals to live as men, usually for at least a year, before surgery can be undertaken, and during this time it is highly likely that they were given androgen (testosterone) hormone replacement therapy (HRT).

This HRT would have had a significant effect on their ovaries. It is hard to know whether this would have helped or hindered Tilly's experiments. Androgens can have powerful effects on cells, but relevant information was not included in the paper. Moreover, it was misleading to refer to them as normal ovaries. If a year of androgen treatment was necessary before OSCs could be derived, I doubt many women would want to proceed with this as a way to overcome infertility.

But even if all my worries about the data are unfounded, how are desperate or vulnerable patients expected to deal with the headlines? I am sure I was not alone in receiving e-mails, and I know there were calls to fertility clinics in London (I expect this was a common occurrence around the world).

How do we answer the patients' concerns? Are the techniques likely to be safe, when will they be available, and what type of patient is likely to be first to have their hopes realised? Where were the qualifying statements, and lists of the scientific issues still to be addressed? It would have been helpful if answers to some of these questions had been provided by the authors — if not in the paper then in accompanying press releases. And how do they expect to achieve societal acceptance of a technology that could have uses that rather stretch current reproductive boundaries? These include not just unlimited numbers of children for women of any age, as spelled out by some of the headlines, but germ line genetic engineering?

So who caused the hype? I suspect the authors of the paper contributed, but their employers, the journal that published it, and a largely unquestioning media did not help. There were some balanced pieces, notably that by Gretchen Vogel writing in Science (7), but too much coverage implied that fertility treatments were just around the corner. They are not. Even assuming the data are valid, the authors are a long way from testing safety and efficiency of OSC-derived oocytes. And in the UK at least, although the research is possible with a licence, it would require a change in primary legislation to permit their use for reproductive purposes, which in itself would take several years.

It is possible to be enthusiastic about a piece of science, but to do so in a qualified manner that respects the hopes and fears of patients. Hype should be avoided for everyone's sake.