Dear Erkki,
Congratulations for receiving the 2022 Lasker Basic Medical Research Award for your discovery of integrins, jointly with Richard Hynes and Timothy Springer. Can you briefly tell us what integrins are and how they were discovered? We will provide references for those that want more details.
Integrins are receptors that bind cells to the extracellular matrix or to other cells. The integrins that mediate cell-matrix interactions are critical to multicellular organisms. They direct cell migration during development and maintain tissue integrity by keeping cells in their appropriate places in adult tissues. Normal cells need matrix attachment; they undergo apoptosis without it. Cancer cells survive without matrix attachment, which allows them to metastasize. Platelets have an integrin that is central to blood clotting. Integrins that mediate cell-cell interactions are associated with the immune system.
My work on the matrix-binding integrins started with the discovery, together with Antti Vaheri, of a cell surface protein we later named fibronectin. (It became clear later that fibronectin is a matrix protein, it just was closely associated with cell surfaces in cultured cells). Two other laboratories published at that time, in 1973, on proteins that turned out to be that same protein, fibronectin. Once it was shown a few years later, that cells attached to and spread on fibronectin-coated surfaces, it was clear that there had to be a cellular receptor for this protein. It took 10 years to find the fibronectin receptor (one of many as it turned out).
My laboratory did painstaking biochemistry to get at the receptor. We first located the sites in the 250K fibronectin polypeptide for the known binding activities of this protein, including the site responsible for the cell attachment activity. We then isolated the smallest proteolytic fragment that reproduced the attachment activity and determined its amino acid sequence by protein sequencing (we also cloned the cDNA for the site, but the sequence from the protein fragment came first). The fragment had 108 amino acids, and we were able to synthesize it in 4 slightly overlapping peptides. One of them was active, so we made smaller peptides based on its sequence and ultimately reduced the site to the now well-known RGD (arginine-glycine-aspartic acid) sequence. Peptides containing this sequence mediated cell attachment when coated onto a surface and inhibited it when used in solution as a competitor. RGD was the first Eureka moment.
We then attached an RGD peptide to an affinity column and run detergent extracts of cells through it. Elution with the same peptide brought out a protein fraction that separated into two bands. A result that initially caused us a lot of consternation was that when we incorporated this presumed fibronectin receptor from the column to liposomes, they did not bind to fibronectin. Instead, they bound to another adhesive protein we worked on, vitronectin. We decided to go back and use a large protein fragment in the affinity column, hoping that that would give us the receptor we were after. That was successful. RGD peptide elution brought out a protein that did make liposomes attach to fibronectin. The puzzling affinity of the first receptor for vitronectin was explained when we cloned vitronectin and saw that it also had an RGD sequence. At this point we knew that we had two RGD-directed receptors that we also knew were real receptors because we could reproduce the activity in liposomes with purified proteins. One could not necessarily conclude that from antibody inhibition of cell attachment because antibodies that do not bind to fibronectin can inhibit attachment to fibronectin.
We now had the beginnings of a beautiful recognition system: two receptors with a similar polypeptide composition and with related RGD-directed) but distinct specificities. I wrote in the discussion of our paper on the vitronectin receptor: “(We) propose the existence of a family of receptors each recognizing the Arg-Gly-Asp sequence within the context of individual proteins”. This was the second Eureka moment.
We subsequently cloned our two receptors, which turned out to be closely related in sequence. Richard Hynes, who also was one of the discoverers of fibronectin, cloned at the same time the chicken fibronectin receptor using a monoclonal antibody to identify the clones. Tim Springer separately identified and cloned three related lymphocyte cell-cell receptors. The sequences showed that all these receptors were members of the same family of proteins, now known as integrins.
Could the strategy you used for the discovery of integrins be extended to other components of cell-cell communications? We have many additional technologies available now, some of them recently discovered.
One would probably would not want to do as much biochemistry as we did 40 years ago. The antibody approach we and the other two groups used to clone the receptors is still valid, although functional genetics approaches are likely to be quicker. One thing in my approach that I think will remain valuable despite current “omics”, is that I had a hypothesis. As a postdoc at Caltech, I became aware of experiments, some of them done in sponges, that suggested the existence of cell surface recognition proteins guiding cell movements. It became the so-called “area code” or “zip code” hypothesis; different areas of the body would have specific molecular codes that would guide cells to their appropriate places. I decided to try to identify these hypothetical recognition molecules when I started my own laboratory in 1970. What we found was a recognition system like the one I had set out to find. The point is that a hypothesis keeps one focused on the goal.
I stopped working on integrins soon after the work I have described here was completed but continued working on zip codes. We have been using in vivo phage display screening of phage libraries to identify molecules that are specific for the vasculature of different normal tissues and diseases. There was no getting away from RGD and integrins; tumor screens yield predominantly RGD peptides because alphav integrins are expressed in tumor vessels and not in the vessels of normal tissues.
REFERENCES
- Ruoslahti E. Molecular ZIP codes in targeted drug delivery. Proceedings of the National Academy of Sciences. 2022 Jul 12;119(28):e2200183119.
- Rothman JE. The gripping story of integrins. Cell. 2022 Oct 13;185(21):3844-8.