The Compatibility Gene, page 16
Herberman’s and Kiessling’s discovery of a new immune cell was not immediately hailed as a breakthrough because, at the time, technology for discriminating between different types of cells was relatively primitive. And it took another four years for the equivalent cells to be identified in humans. Human NK cells looked very different to T cells when viewed under a microscope: much more speckled, or, to use the technical term, more granular.15 Their speckled appearance easily stood out – and this was formative in everyone agreeing that NK cells really were a new type of immune cell.16
The discovery of NK cells solved the transplantation mystery; these immune cells were responsible for the rejection of bone-marrow transplants (in F1 hybrids). The killer had been identified – but what was their motive? As one enigma unravelled another opened up: what caused the NK cells to attack transplanted bone marrow? Did they use a similar strategy to T cells to detect disease – self/non-self – or did they have an entirely different battle-plan? NK cells could not have evolved to detect transplanted bone marrow because that would never happen naturally. But there had to be a clue in the transplantation mystery as to how NK cells really worked – because the way that NK cells detect diseased cells must cause them to attack transplanted bone marrow.
Another clue to how NK cells worked was that some tumour cells, readily killed by these immune cells, actually lacked proteins encoded by compatibility genes – the HLA proteins. Most people took this to mean that NK cells must recognize some signature of disease which had nothing to do with HLA. As a result, research programmes around the world focused on finding something other than HLA proteins on transplanted cells that would be detected by NK cells. But the young Swedish PhD student Kärre came up with a different idea.
For the final sentence of his PhD thesis, Kärre wanted to write down the strategy used by NK cells to detect signs of disease in other cells.17 Focusing on overarching principles about how the immune system works is reminiscent of Burnet’s ambition to discover the grand unified theory of immunology, but Kärre’s approach was more pragmatic. Kärre collated all the available data – his own as well as everything he could read about – and then set about trying to summarize them. This was different from the way Burnet thought in more abstract terms about how an immune system could work: Kärre simply listed all the situations in which NK cells were good at killing diseased cells to see if he could fathom a common denominator.18
He had begun his journey to this point because of one crucial phone call that went unanswered. In 1975, at age twenty-one, while in his second year of studies to become a medical doctor, he read that George Klein, the head of the Tumour Biology Research, was looking for someone to help in their research. At his interview, Kärre began to tell Klein all about himself, but Klein interrupted to tell him that none of that was important: ‘What I need to know is when can you start?’19 Kärre said he could start the next day, which got him the job, and immediately Klein tried to phone the researcher he had in mind to work with Kärre, but he wasn’t around. Flustered, Klein just called someone else and arranged for them instead to work with Kärre. A tiny moment on which great things hinged; this second researcher was Kiessling – and that’s how Kärre got to work on the newly discovered NK cells.20
After years of studying these new cells, it was during a stay at his soon-to-be parents-in-law – trying to finish his PhD thesis – that Kärre had an epiphany. It occurred to him that cells readily killed by NK cells were often resistant to killing by T cells, and vice-versa. Thinking about the problem in this way, he realized that T cells only killed cells that have proteins encoded by compatibility genes, while NK cells were best at killing cells in which these proteins were absent. His stroke of genius was to recognize that the very fact these proteins were not there might be the signal NK cells used to detect that something was wrong.
The unlikely source of inspiration for Kärre’s revelation was a defence strategy used by the Swedish navy.21 The navy command had become worried about the presence of foreign submarines entering their waters and decided that a low-cost option for surveillance would be to educate local fishermen to keep an eye out for them. Their initial plan was to give fishermen leaflets showing silhouettes of the submarines they should look out for. Anybody who saw a submarine that looked like one in the leaflet should alert the navy. But there were all kinds of possible submarines, and anybody seeing one would have to go through several pages to check for a match. Realizing that this was cumbersome, the navy came up with a second plan of issuing fishermen with one single page showing three submarines – the Swedish ones – along with the instruction that if they saw a submarine that was not one of those pictured then they should alert the navy. Rather than directly looking for something foreign, the best strategy was to check for what was expected and, if it wasn’t that, then it should be considered a problem.
Kärre realized the navy strategy applied to NK cells. Rather than directly looking for unexpected molecules as a signature of disease, NK cells could check for proteins that are normally on cells as a sign of health. Compatibility genes are at the centre of this. It is the proteins that they encode for – the HLA proteins – that is precisely what is checked for. Nearly all cells in your body have HLA proteins at their surface – to allow them to be checked by T cells – so any cell that lacks these proteins at the surface must be abnormal in some way. Indeed, many cancer cells contain mutations that cause them to lack HLA proteins at their surface – presumably helping the cancerous cell avoid detection by T cells. And HIV, for example, has its own way of stopping HLA proteins doing their job properly.22 NK cells defend against this – by noticing whether HLA proteins are missing. It was the epiphany that gave Kärre the grand ending he wanted for his PhD thesis: the immune system searches for things that should not be in your body and checks for the presence of normal proteins that should be there.
Back again to the transplantation mystery. Recall that we got to the point where (F1 hybrid) NK cells were discovered to attack bone-marrow transplants from their parents, but the new problem was in understanding why. The situation was mysterious, because an F1 hybrid inherits all of Mum and Dad’s compatibility genes, so there is nothing that could be detected as non-self to trigger an immune reaction against the transplant. Kärre’s epiphany provided the solution: when bone-marrow cells from one parent – say Mum’s cells – are transplanted, they only have half the set of compatibility genes that the F1 hybrid has – Mum’s but not Dad’s. So some proteins are missing in the transplanted cells; NK cells, abundant in bone marrow, detect this loss of protein – and attack.
No longer thinking this should just be one sentence in his PhD thesis, Kärre thought to announce his idea in his thesis title. Kiessling, his supervisor, considered it unwise, because this was still merely an unproven hypothesis. Kiessling and Kärre were similar in age and openly criticized each other’s opinions. This was important to their success: heated, back-and-forth banter is often necessary for thoughts to crystallize. Kärre was awarded his PhD in 1981, but his big idea was indeed left out of his thesis title. It didn’t really matter, because, despite being the culmination of many years’ hard work, a PhD thesis is rarely read by anyone anyway.
Eventually, Kärre’s hypothesis got the wider attention it deserved when, aged thirty, he was one of the youngest presenters at the second-ever NK Cell Workshop, held near Detroit in 1984. Although the audience thought his new idea interesting, it was far from being instantly accepted. Most delegates were sceptical – and when scientists politely say in public that they are sceptical, they’ll be clearer in the bar and say they think it’s bullshit.
Kärre thought to himself that, if it turned out that he was wrong, he would leave scientific research and embrace a clinical career instead.23 Resistance in the community wasn’t simply inertia against anything new; scientists were sceptical because there was some data that didn’t seem consistent with Kärre’s idea. For example, some cells could be killed by NK cells even when they had MHC proteins present, so it wasn’t clear-cut that only those without them were killed. At least initially, other views about NK cells seemed equally plausible: maybe NK cells were able to recognize some other protein – one that remained to be discovered – as a sign of disease.
But the biggest problem with acceptance of Kärre’s idea was that it seemed to suggest something completely opposite to what was already known about these proteins. The proteins encoded by compatibility genes were established as being important for activating immune cells – capable of switching on T cells to attack. Now, they were being proposed as also being able to turn off immune cells – albeit different immune cells; NK cells. One scientist likened the idea to somebody saying you could play the Beatles backwards and still get meaningful lyrics.24
In 1986, two years after Kärre first presented his idea at the Detroit meeting, he found that mouse tumour cells deliberately chosen to lack MHC proteins were indeed efficiently killed by NK cells.25 With this result, Kärre was sure that his idea couldn’t be entirely wrong. A few years later, in 1990, he published a hugely influential article with his first doctoral student, Hans-Gustaf Ljunggren, elaborating on the idea and naming it the ‘missing self’ hypothesis.26 Good names are important – the Big Bang is another example. Art also helped; the cover to the journal featuring the ‘missing self’ hypothesis showed a woman looking into a large mirror with her reflection absent.
As Kärre’s idea gained traction, attention shifted to the problem of how immune cells could know when a normal molecule was missing on another cell. One way this could work would be if NK cells had receptor proteins on their surface that stop NK cells from killing whenever they bind to MHC protein on another cell. In this way, when an NK cell touches another cell that has MHC proteins, the NK cell receives a signal to not kill. But if MHC proteins are missing on any cell, the NK cell receptor proteins have nothing to interact with – the brake signal is lost and the NK cell delivers its kiss of death. This seemed the simplest way in which NK cells could survey other cells for the presence of a normal self protein – but the question was whether or not such a receptor really existed; the search was on for the missing ‘missing self’ receptor.27
Once again, the problem was solved out of left field. Without thinking about the ‘missing self’ hypothesis at all, Wayne Yokoyama – born in Hawaii – had been working at the Washington University School of Medicine, St Louis, to identify new receptor proteins on immune cells and in 1989 he found one that he called A1.28 It was later renamed Ly49 and Wayne’s wife, Lynn Yokoyama, truly believes he named it after her. In truth, it’s named as the forty-ninth member in the Ly family of proteins that share certain chemical features.29 Yokoyama found out that Ly49 came in many different forms. He hadn’t any clue that Ly49 would turn out to be central to Kärre’s ‘missing self’ hypothesis but he was confident that he was on to something important – because he thought that the protein’s variability must mean something interesting.
Yokoyama – who has a zen-like quality of always seeming calm and relaxed – had initially been turned on to medical research through a high school project he did.30 Yokoyama’s father had died when he was fourteen, and his school teacher mentored and supported him through the difficult times that followed, later helping him obtain a summer placement at the hospital where Hawaii’s first kidney transplant between twins had just happened.31 Aged seventeen, Yokoyama worked with doctors who were beginning to optimize transplantation success using tests that had recently been established by the early pioneers of HLA – those we met in Chapter 3. The population of Hawaii is exceptionally diverse, and Yokoyama could see patterns in blood reactivity when he ordered his results according to each individual’s ethnicity. He could see that this could have a major impact in medical transplantation and decided he should become a medical doctor.32 Later in life, when it became clear that part of the job as a medic is to ask patients intimate details about issues such as their sex life, he decided that his place was in the lab.
After working at the NIH, Yokoyama earned a faculty position in University of California San Francisco (UCSF) and decided to focus his new lab on the protein he had recently discovered, Ly49. At that point, the protein could have had any role at all in the body. But he found out that the Ly49 gene was positioned close to another gene already known to be active in NK cells, and that gave him a clue that he should look at those cells. By separating NK cells according to whether or not they had Ly49, he found out that NK cells which lacked Ly49 killed other cells especially well. This indicated that the presence of the Ly49 protein on NK cells worked to prevent them killing.
By comparing how well NK cells could kill different types of tumour cells, it was eventually worked out that the Ly49 receptor turned off the NK cell killing machinery when it bound to MHC proteins on another cell in contact. If MHC proteins were missing on a cell in contact with an NK cell, the signal from Ly49 to ‘switch-off’ gets lost and the NK cell delivers its lethal hit.33 Yokoyama had identified the molecular process that lets the immune system search for ‘missing self’ – through a receptor protein that checks for presence of MHC proteins on other cells. Yokoyama’s discovery vindicated Kärre’s original idea about NK cells, and both scientists became well known as a result. In 2009, Kärre gained the prestigious position of chairing the committee that decides who wins the Nobel Prize for Medicine and Physiology.34 He doesn’t like to talk about it; the committee conducts its work in high secrecy, for reasons that are obvious, given the gossip we’re about to come across.
Yokoyama’s discovery was in mice. It was important to find out whether or not NK cells worked the same way in humans; the next race was to find the equivalent of Ly49 in us. Research teams searched for a protein that looked like Ly49 on human NK cells – or a gene that encoded for one. Many years passed, and the human equivalent simply couldn’t be found. The question arose: could a defence strategy as elegant as the search for ‘missing self’ only be used by rodents?
Then – as we’ve seen happen before – two teams reported the solution at the same time.35 But gossip circulated soon after to suggest that one of the team’s leaders had listened to a lecture by the other – and from that had gained a vital clue that they had then used to identify the receptor. It could well be a rumour without merit, but I mention it, not to take anything away from the achievement of the scientists involved, but to show that gossip circulates in science probably no differently to the banter that goes on in art, music, literature or any other area of human achievement. Tales like this exist for many big discoveries – and even small ones where establishing priority has no impact anyway.
The clue that one scientist may or may not have gained from the other was the protein’s type (and its chemistry at one end). It was crucial information, because the human version of Ly49 was not much like mouse Ly49; for years, everybody had been looking for something that would be similar to mouse Ly49, but the human version turned out to be altogether different. The human counterpart to mouse Ly49 has very different chemical features – and is part of the so-called immunoglobulin family of proteins, which gives them their cumbersome name of Killer Immunoglobulin-like Receptors, or KIR.
Important processes for life usually evolved in a common ancestor to mice and humans. That is why genes and proteins are often very similar in both species. But here, different kinds of protein operate an identical aspect of the immune system. The implication is that the strategy of looking for missing self as a signature of disease is useful to many species – but the process to make it work has evolved relatively recently so that it is different in mice and men.
This difference – of mice and men – is important because mice are commonly used for testing and developing new medicines; they’re relatively easy to keep and breed fast, and a lot of technology has been developed around mice so that, for example, it’s easy to genetically manipulate them. I write this as fact – not as an advocate for animal experiments, which is a complex issue that needs a separate discussion. In support of using mice for medical research – as opposed to another animal – it’s often quoted that the genomes of all mammals are remarkably similar. Precise details depend on what’s compared; roughly 99 per cent of mouse genes have a human counterpart, each being around 85 per cent similar on average. However, there’s no mouse version of AIDS or many other human diseases. And the different kinds of NK cell receptors used to search for ‘missing self’ are a good example of where mouse and human immune systems differ. A drug designed to block these receptors on human NK cells, for example, cannot be tested in mice because they simply don’t have that protein for the drug to act on. This is not an academic point; it’s a major problem for testing new medicine.
Such a drug – to block the NK cell’s ‘switch-off’ receptors – is, for example, in development at the French company InnatePharma (and licensed to the pharma giant Bristol-Myers Squibb). The idea is to unleash NK cells so they may be able to kill cancer cells, or other diseased cells, more effectively than normal. To test the drug in mice, receptors from human NK cells have to be genetically added to the animals. Technically it’s not that difficult to do this – scientists have been shuffling genes between species for a long time now – but it’s easy to see how this approach soon gets fiddly. Because, when adding the human NK cell receptors to mice, the human MHC proteins they bind to should also be genetically added – and then how much more besides?36
