Goodbye thymus, hello CSF

The great watershed year was 1965, wasn’t it? Something quite dramatic changed it for all of you.

Yes. It arose from the phenomenon that individual cells in a culture of bone marrow cells growing in semi-solid medium, agar, could generate enormous colonies. Now, that technique was discovered by accident by Ray Bradley, a scientist working in the University of Melbourne with whom I had collaborated over the years. Two things became pretty obvious. For the first time in history, people could grow blood-forming cells as colonies. It turned out that (as had seemed likely) they were clones, each one coming from a single cell – and they made a colony of daughter cells during a week of incubation. But unless you added something to the medium in the culture, colonies would not grow. That something we called colony stimulating factor, CSF.

The point about the cultures was that they gave you a technique for measuring CSF concentrations, because the number of colonies that develop reflects the concentration of CSF. So we had a way of doing three things: working in tissue culture, which I knew we needed; detecting some factor that, hopefully, was a regulator of the sort we had been seeking for a decade; and measuring it. So yes, almost overnight all work on the thymus stopped.

It wasn’t that we immediately rushed over to Ray Bradley and taught ourselves how to culture colonies. We worked for the next year as a team, in which I continued to do the formal haematology and general cell biology, but eventually we did teach ourselves how to do the technique and take the next logical steps. Every so often there is an accidental occurrence like that, when you would have to be blind not to realise that here is something astonishing that warranted a few decades’ work – and so it proved.

Early steps toward purifying CSF

So what to do about that watershed in the mid-’60s? It’s no good simply believing that you have a technique for discovering your favourite unknown hormone-regulating blood cells. You’re working with cells in a culture dish, artefacts abound, maybe colony formation was all just an artefact. To get further forward, several things were needed. The first was to be able to show that CSF was detectable in the serum and hopefully in the urine. Why? Because it would make sense if it’s a regulator that detectable levels of CSF should be present in the serum and urine. It would be nice also if you found that there were CSFs to be detected in tissues. It would make sense if you had an infection and needed to make extra protective white cells (granulocytes and macrophages) that CSF levels should go up, otherwise it would not be a good candidate for a regulator.

We spent about three years surveying patients with infections, looking at CSF levels in their urine and serum and looking at different tissues to see which had the greatest content of CSF – assaying all the time by the culture method, which was the only one available to us. And by, perhaps, late 1968 it was obvious that there was enough indirect evidence to support the notion, ‘Yes, CSF is a good candidate for a regulator. Let’s spend some time purifying it and putting a biochemical basis to it.’

I put a poor unfortunate PhD student, Richard Stanley, onto this 'simple' job of purifying CSF. (He is now a distinguished professor in New York; his photograph was on last month’s issue of Cancer Research.) We started with human urine because it was a good, cheap starting material. We had buckets for collection of urine in the Institute. First you had to take the cigarette butts out of it – these were the days when you could smoke in a research institute – and then you had to dialyse it in great evil-smelling tanks in 50 litre batches. Great stuff! It took nine years to purify CSF from human urine. Richard did not complete the job until he was in Toronto working as a post-doc.

So many CSFs!

This must have been getting into the early 1970s, was it?

Yes. Meanwhile, the situation was becoming a little bit murky and uncomfortable. The CSF from urine did not stimulate colony formation all that well. In particular, mostly we got only small macrophage colonies, not the large beautiful granulocytic macrophage colonies seen with the original Bradley technique. Clearly things were a bit more complicated than we had thought. There must be more than one type of CSF. When we began to analyse what type of CSF was being made by different tissues, it became appallingly obvious that lung tissue was making a CSF that had no chemical relationship whatsoever with urine CSF (which was now being called M-CSF because it pretty much only stimulated macrophage colony formation). Lung CSF was a much smaller molecule and it stimulated the formation of beautiful granulocyte macrophage colonies, so we called it GM-CSF.

It also became obvious that if you took lymphocytes and stimulated them with mitogens they produced another type of CSF with some remarkable properties. While all this had been going on, we and others had developed culture techniques that would grow colonies of other types of blood cell. (There are eight major families of blood cells.) CSF made by activated T lymphocytes could stimulate the formation of red cell or megakaryocyte colonies. Urine CSF or lung CSF could not do this. So there appeared to be yet another CSF.

It took us quite a while to realise there was yet another, fourth CSF. This turned out to be the most famous CSF of all – G-CSF. For two years I had missed the fact that there were miserable little colonies developing in certain culture dishes. I thought they were merely dead colonies! But the CSF causing the formation of these small granulocytic colonies came to be known as G-CSF, and it’s the one that is making mega-millions for drug companies.

So everything was happening simultaneously. You might say we were very slow to purify the CSFs, but the project had become four times more complicated. This is partly why the project took fifteen years to complete. Other sorts of assays were being developed all the time, we had to figure out all the novel biology behind why one type of colony was being made and why another, and we ended up with a project that needed four different purifications (for four different CSFs) and a much broader range of assays to be done.

Focus questions

• Metcalf says that successfully growing blood cells as colonies in agar gave him a way of doing three things. What were they?

• What did Metcalf discover when he and his colleagues attempted to purify CSF?