In the early 1980s, I did a post-doc in Dundee, Scotland, with Philip Cohen—now Sir Philip. When I was in Cohen’s lab, I was given the job of purifying a novel protein kinase, which was quite difficult in those days. Had I known there were 500 of them, I wouldn’t have attacked it, but I had the naiveté of youth. I purified this protein kinase, called GSK3, which was thought to be regulated by insulin. Cohen spent a long time working on this. But when I left Dundee, I didn’t want to have anything to do with insulin signaling. I wanted to establish my own identity and I started to get into the world of molecular biology and cloning of protein kinases. At the beginning, there were only around 10 of them, and then after the orgy of PCR, as one Nobel laureate called it, we ended up identifying 518. During this period I discovered what turned out to be a novel kinase, which is now known as PKB or AKT. There’s a European name and an American name.

Well, we discovered it first in Basel and, in parallel, a man called Steve Staal, working at NIH, was working on a retrovirus isolated from a mouse tumor and this was called v-akt. In 1991, we published our cloning of PKB as a novel kinase related to PKA and PKC, and within a year, the v-akt sequence was published. It was found to be an oncogenic form of PKB, so the two worlds collided, and that’s why it has two names. We were screening for PKA cDNAs and we picked up one that looked like PKA, and that turned out to be a distinct gene, so we carried on working on it.

“...this PIK3/PKB signaling pathway is a major pathway in human cancers.”

The point is nothing really happened for a while. There were only two labs in the world working on this, my lab and Phil Tsichlis’s. In truth, we actually identified the PKB cDNA around 1986 and we were so impressed by it that we left it in the fridge for a couple of years. We didn’t know how to handle a new kinase. In 1995, there were many seminal papers produced. One was by Franke and Tsichlis, another from Burgering and Coffer, and this was when the whole world came together; it was realized that PKB was downstream of another major player called PI3kinase—PI3K. This was a lipid kinase, attached to tyrosine kinase receptors. Previous to the work in 1995, there was no known downstream target for PI3K, yet it was one of the first proteins that was activated by receptor tyrosine kinases which sit at the cell membrane.

There’s one other piece to this that also came together at this point. This is the pleckstrin homology domain. Somebody worked out that maybe the pleckstrin homology domain binds PIP3, and this turned out to be the crucial ingredient in this pathway. When you activate receptor tyrosine kinase, you elevate PIP3 in the membrane, and this leads to PKB being recruited to the membrane and activated. This filled a big hole in the signal transduction pathway when it was reported in 1995.

At that time I went to Dundee to interview a Ph.D. student and Philip Cohen was there and for old times’ sake he showed me new data on GSK3 and how it was regulated by insulin. And he said that he was purifying the upstream kinase. This is when I told him that we had a kinase regulated by insulin and I suggested to him that he check for this to save himself the next year or two purifying the unknown kinase. We had a very good antibody to it, and when I got back we sent it to him and he tested it in the purification and there it was: PKB was the upstream kinase of GSK3. So we were able to connect the signaling pathway from insulin receptor through PI3kinase through PKB and then to GSK3. This was quite a breakthrough at the time.

We started to collaborate on this on June 15^th and we worked very intensely and actually published the paper on December 24^th of the same year. So we conceived the experiments, did them and wrote the paper and published it in Nature all within six months.

This was a major step forward for the insulin-signaling field. And it was first and came out of nowhere. That’s why it’s so highly cited. That was the first substrate for PKB and it led to probably 50 to 70 other proteins being identified as substrates for PKB and established PKB as a major signaling molecule. So it went from being an orphan kinase—not having a substrate or a signaling pathway—to being center stage, and an oncogene. So the whole field exploded in 1995. It went from about 10 papers in ‘95 to maybe 10,000 papers now, nearly all of which list PKB/AKT as an important component.

We had done some early experiments in the lab. We were searching for a pathway for this orphan kinase, and one of the classical things you throw on cells is insulin, growth factors, serum, all sorts of things. And we had a very good antibody; we could immuno-precipitate it and measure its activity.

Not really. It was a bit of a whirlwind. It was only in the next two years that we realized how mega it was. It had all the components and then suddenly there it was; the components had all come together and we had a major story and a major impact on signal transduction for the next 10 years.

Yes. From 1990 to 1995, this was a minor activity in my lab. There was just one student, one post-doc, working on it. Then in 1995, when we found out the PIK3-insulin connection, I rapidly shifted more of my people in the lab to work on it so we could stay current. We had to keep up with all the other people who suddenly joined in. One minute you’re alone and the next minute you’re surrounded by people. Now about two-thirds of the work in the lab is on PKB/AKT and this pathway.

We’ve been trying to establish the molecular mechanism for activating the kinase. We worked some of this out with Philip Cohen and Dario Alessi. All three of us have benefited from this protein. We worked out that PKB was regulated by multi-site phosphorylation. There were some competing mechanisms proposed, but our mechanism is the one that stood the test of time and it made many antibody companies quite rich, because of the phospho-specific antibodies they made to the sites we identified that control the activity. Then it was a race to find the upstream kinases in between PIK3 and PKB, which were eventually found.

It’s still not nailed down completely, but we have contributed to this by identifying PDK1 as a master regulatory kinase that regulates PKB and also several other kinases. So it’s a kinase that regulates a kinase that regulates a kinase. It’s a bit like a loop. Meanwhile, the second phosphorylation of PKB has been a controversial issue, but we think that maybe there is some order in this field now. We found that members of the PIKK family of kinases regulate PKB, as well. So it has to get two signals: one from PDK1 and one from the PIKK family and they phosphorylate different sites on PKB. 

We also had a collaboration with David Barford in London, and David was able to work out the crystal structure of PKB in its inactive and active state. And we showed that how this multi-site phosphorylation led from a partially-disordered structure to an ordered structure and approximately a 1000-fold increase in activity, which is the pivotal event in this signaling pathway.

Now we’ve moved from this cell culture work and this signaling pathway, which has grown to gigantic proportions, into the area of studying the kinase in the animal. We have started to do mouse genetics, and we knocked out the three different isoforms of PKB to establish the true in vivo function. In the cell culture system you can make predictions of how it works and when it works, but to really show that the signal pathway works in a whole organism you have to knock out or ablate the genes. We’ve done that in all three isoforms of PKB and the result is three different phenotypes.

If you knock out PKB alpha, you end up with about 30% of the mice dead. The ones that survive are small and stay small. So this looks like a condition of intrauterine growth retardation. It’s probably due to the fact that PKB plays major a role in the placenta and in angiogenesis of the placenta. So the embryos are a little bit starved, which is not a good thing. We’re currently looking to see what happens to these mice in adulthood. They stay small but they have difficulties breeding. So this starvation of the womb is imprinted for the rest of their life.

In Philadelphia, Morris Birnbaum knocked out PKB-beta and this mouse is normal size, but it develops a kind of diabetic phenotype when it ages; it becomes insulin resistant. So this brings us back to insulin signaling. Then we also knocked out the third isoform, PKB gamma. These mice are normal size, but they actually have problems with post-natal brain development. They end up with a brain about 25% smaller than the normal mouse brain. They may have additional problems: basically their brain has fewer cells in it and the cells are smaller.

Now we’ve gone onto the second generation and made double knockouts and we’re attempting triple knockouts so we can work out all the in vivo roles of PKB. And these mice all have additional phenotypes, as well.

The other major thing is that this PIK3/PKB signaling pathway is a major pathway in human cancers. So many of the mutations lead to activation of this pathway, which helps a cell become cancerous. Now that we have described much of the normal physiology, we also need to describe the pathophysiology and to find a therapy to down-regulate this cancer-forming pathway, which is prevalent in about 10 or 15 different cancers, probably more. We need a drug, basically. This is what the big aim is: to find a drug to the PI3K pathway and turn it off in tumor cells.

Yes. What we want to do for therapy would be to turn down the pathway and not interfere with normal insulin signaling. In effect, we’re now doing what they call systems biology, where we look at the whole organism and try to understand the problem that way, rather than the reductionist approach, where we look at small components. Now that we’ve mapped many of the signaling pathways, we can stand back and look to see how they’re affected in our organisms and interpret how they’re working in the pathophysiological states—in Type 2 diabetes, for instance, where there are very high circulating insulin levels, and so insulin can change from a hormone that regulates your glucose uptake to actually function as growth factor. And this is what makes this work so important now from a clinical point of view.