The Compatibility Gene, page 10
Doherty was studying mouse immune responses to Lymphocytic Choriomeningitis Virus (LCMV). This virus is especially interesting to immunologists, because it’s the immune reaction to the virus that causes the problems, rather than the virus itself. Immune cells, activated by the virus, kill cells that are critical for maintaining the blood–brain barrier, something that can lead to acute brain swelling and death. Recall how some parts of the body need to be protected from immune responses: this is an example of why such protection is necessary – and what happens when it fails.
Bonding through their dedication, passion and a mutually appreciated dry humour,25 Zinkernagel and Doherty started working together on LCMV day and night, seven days a week. ‘The experience was truly intense,’ Doherty later recalled. ‘We were totally obsessed.’26 Together they tested first whether immune cells (specifically T cells) taken from the cerebal spinal fluid of infected animals could kill cells that they deliberately infected with the virus.27 They found that the better the immune cells were at killing, the greater was the severity of the disease in mice: as expected for the disease being caused by the immune response, rather than the virus directly. Then, in October 1973, came their big breakthrough experiment.
Inspired by previous studies showing that strains of mice differed in their susceptibility to diseases, they set out to compare the ability of the immune cells from one mouse strain to kill virus-infected cells taken from other strains.28 A series of experiments using cells taken from mice with different genetic backgrounds yielded a sensational discovery: that the immune cells, killer T cells specifically, could not kill just any cell infected with the virus.29 What the pair found was that killer T cells activated by the virus in one mouse strain were only able to detect the virus in other cells that had the same class I compatibility genes. The implication was that genes for transplant compatibility also control immune responses against a virus.
This discovery was dynamite, and Zinkernagel and Doherty knew it. The pair presented their work to Burnet, by now aged seventy-four, but surprisingly he didn’t appreciate the results instantly.30 Undaunted, Zinkernagel and Doherty decided they should get out and tell people about their work: Doherty gave twenty-two talks across the world in six weeks, and Zinkernagel gave many others throughout Europe.31 In April 1974, they published their results in the top journal Nature.32 Even so, many still thought their observation was just some spurious result – perhaps something strange arising from the methods or the particular virus they used.33 The scientific community acknowledged that these genes couldn’t simply exist just to make life difficult for transplant surgeons – but it was hard for scientists to change ingrained patterns of thought and see these genes as being able to control the immune response to a virus. Most scientists at the time thought that an immune cell would recognize a virus infection directly – without any restriction or influence from the type of cell infected. Doherty, in particular, never forgot the names of those who poured cold water on his work.34
It was some two years after Zinkernagel and Doherty’s Nature paper was published – around the time it took other teams to publish their own versions of the experiment – before the scientific community broadly agreed on the importance of this work. Zinkernagel and Doherty coined the term ‘MHC restriction’ to describe the idea that the detection of viruses is restricted to cells with appropriate MHC proteins. (To be clear, MHC refers to compatibility genes in all species, and HLA is the acronym for these genes in humans.) ‘MHC restriction’ remains in the everyday language of immunologists.
In 1975, Zinkernagel and Doherty discussed the implications of their extraordinary discovery in a short but seminal piece for the Lancet.35 In it, they introduced a new idea. Over twenty-five years earlier, in 1949, Burnet had articulated the idea that the immune system works by telling apart self and non-self, distinguishing your own cells and tissues from anything else. Zinkernagel and Doherty’s new idea was to suggest that, in fact, the immune system works through recognition of ‘altered self’. A body’s MHC proteins, they proposed, were ‘altered’ by the presence of a virus – and the body’s immune system could then identify disease as ‘altered self’. This was an important shift in thinking about how the immune system worked.
They went further: to suggest that their data could explain why there is such great diversity in our HLA proteins. At the time, in the mid-1970s, the dominant explanation for this diversity was that transplantation incompatibility could have evolved to prevent tumours spreading among us. Transmissible cancers are not known to occur in humans – but they have been found in a few animals. A transmissible venereal tumour in dogs, for example, is passed on during copulation and may have originated when dogs were first domesticated around 10,000 years ago.36 Zinkernagel and Doherty, however, speculated that diversity in HLA might relate to how a population avoids viral infections.
Perhaps, their argument ran, it would be harder for a virus to evade our immune system if the process of detection varies. Or, put another way, they speculated that we might have evolved diversity in HLA so that we are stronger at fighting off viruses – as a population. The idea proved to be very insightful, especially because it wasn’t clear to anyone how the MHC protein could really be ‘altered’ by the presence of a virus – that wasn’t going to be revealed until Bjorkman, Wiley and Strominger finally worked out what was going on in that elusive 10 per cent of the HLA protein structure. But, at the time, Zinkernagel and Doherty emphasized that such details were not central to their general arguments; like Burnet before them, they were concerned with establishing principles for how the immune system works.
Zinkernagel and Doherty won the Nobel Prize in 1996 – surprisingly a long twenty years after they actually worked together on these ground-breaking experiments and ideas. Did they know it was coming? ‘Well, people told me I’d been nominated for the Nobel Prize,’ Doherty recalled to me in 2011, ‘but I don’t know how they knew as it’s all supposed to be secret.’37
Following Zinkernagel and Doherty, the big question was how MHC proteins and viruses were being recognized together. What exactly – as in what protein or combination of proteins – could immune cells detect? This needed to be answered to understand how detection of a virus is linked to proteins encoded by our compatibility genes. A clear picture of how this works was sorely needed.
In general, cells interact with their surroundings using receptors at their surface – small protein molecules that protrude out from the cell – which bind other molecules either in the surrounding solution or on other cells. For T cells, there were two schools of thought. One was that T cells have a single receptor that could somehow recognize virus protein and MHC proteins together. The other was that these immune cells must have two receptors: one to recognize a virus protein and one to recognize the MHC protein. Both would have to be triggered to cause an immune response.
Different research groups raced to find out the nature of the receptor on T cells. Several made progress,38 but the discovery that made things clear was the identification of the genes that encode the T-cell receptor. Mark Davis at the NIH and then Stanford did this using a technically difficult method which finds genes that T cells use and another cell type, say B cells, that don’t.39 Instrumental to his success was that he was thinking differently from everyone else – and he did a different kind of experiment from everyone else.
He wasn’t thinking about how HLA proteins worked – or about the immune system per se – but was thinking instead about what makes a T cell different from another type of cell.40 All the cells in one person’s body contain the same set of genes, but different genes are ‘switched on’ to make proteins in different cells, giving each cell its unique appearance and role in the body. What Davis particularly wanted to know was which genes were specifically switched on in T cells. That way of thinking led him to find a gene that was variable between different T cells and never found in other cells – so it had to be the main receptor on T cells involved in the recognition of viruses.
He announced his discovery of the T-cell receptor genes in August 1983, in an impromptu talk at the World Immunology Congress in Kyoto, Japan. The journal Science took the highly unusual step of publishing a note on the meeting announcement in September 1983.41 Normally, any scientific result has to be assessed by the scientists – the famous peer-review process – and this usually takes months. Indeed, Davis’s discovery was properly published – after peer review – in March 1984.42 The debate over the nature of the T-cell receptor was settled: there was one receptor that varied from one T cell to the next (allowing each T cell to detect one type of non-self molecule, such as one found on a germ).
But this only led to a new problem: how was this single T-cell receptor able to recognize the presence of a virus in conjunction with MHC protein? The very thing that Zinkernagel and Doherty discovered – that T cells would only detect a virus when it infected a cell with a particular MHC protein – was still mysterious. How this really works – at the level of what the T-cell receptor really detects on another cell – remained crucial to work out, both to fundamentally understand how the immune system worked and also for the aim to medically aid or hinder immune responses.
Again, opinions differed about how the T-cell receptor might work. One view was that the T-cell receptor recognized viral protein stuck or sitting next to HLA proteins. Another view was that a virus modified HLA protein in some way that could be recognized by the T-cell receptor. Nobody knew. And while uncovering the broad concepts of immunology gained much from mere thinking, the molecular details that make it work in practice were very hard to theorize about; simply, more experiments were needed. And another hero.
Alain Townsend was a twenty-three-year-old medic at St Mary’s Hospital in London when he first read the 1974–5 papers by Zinkernagel and Doherty; he was blown away by their importance. Townsend was also fascinated by evidence that susceptibility to some diseases was known to correlate with the types of compatibility genes an individual inherited. So he began a PhD project at the National Institute for Medical Research in Mill Hill, London, to determine how T-cell recognition worked.43 There was one fact that seemed especially puzzling to him – that some viruses could be detected by T cells even when the virus didn’t have any of its own proteins at the cell surface. So how could they be detected?
Townsend and his colleague Andrew McMichael – both men calm, gentle and gifted in thinking clearly – had cultured T cells that could detect one particular ’flu protein. They had looked for this protein at the surface of ’flu-infected cells and were sure it wasn’t there – it was only found inside cells. How could this protein be seen by T cells as a sign of disease on another cell if it isn’t on the surface of the infected cell and is only found inside? Townsend and McMichael would often head to the local pub and thrash about ideas.44
Townsend carried out a pivotal set of experiments which spanned the early 1980s, as part of his PhD project in Mill Hill and continuing when he moved to Oxford University, working in a research institute linked to the John Radcliffe Hospital. In one pioneering experiment he compared how well T cells killed other cells that were treated in three different ways: firstly, cells were infected with the influenza virus; secondly, cells weren’t infected with the whole virus but were engineered to make the one viral protein which the T cells responded to; thirdly, cells were engineered to only make bits of that one viral protein. This third variation was the most important part of the experiment; these cells did not contain influenza virus or even the one protein molecule which the T cells responded to – they just had bits of that protein, called peptides.
Townsend found that T cells could equally well kill cells treated in any of these three ways. He also established that different T cells were activated by different parts of the ’flu protein.45 His follow-up to this is the experiment that is nowadays celebrated the most: instead of using cells engineered to make viral protein, or bits of viral protein, he just directly added to the cells bits of the viral protein – peptides – that had been synthetically made.46 Over twenty-five years later, Townsend still vividly remembers the wondrous moment he looked down the microscope and saw that cells bathed in the right peptide had been destroyed by killer T cells.47
It was obvious that the cells had been killed: they simply came unstuck from the bottom of the dish. He showed McMichael, and they went to the pub to celebrate.48 He knew the outcome of the experiment, but to get a result that was publishable – he couldn’t just write that he saw cells being killed – he had to wait to get a precise answer from his experiment, measured as the amount of radioactivity released into the surrounding liquid as a result of radioactive cells being destroyed. Back from his drink, Townsend got these formal results. It confirmed what he had seen: his experiment apparently showed that killer T cells would recognize and destroy cells that had small viral peptides and a particular HLA protein. This was an astounding result: it showed that the immune system could tell that a cell was diseased if it contained just a small piece of one protein from a virus. That is, small fragments of protein, called peptides, are recognized as signs of disease by the immune system.
But not everyone agreed. Zinkernagel, for one, didn’t like it, probably because there was no simple explanation as to how small parts of the viral protein would be recognized by T cells – this would only be solved when Bjorkman, Wiley and Strominger eventually presented a picture of what the HLA protein looks like, still eighteen months away. Scepticism from Zinkernagel hurt because young Townsend worshipped the ground he stood on.49
To resolve the matter, it was agreed that Townsend would send his cells to Zinkernagel’s lab so that his team could repeat the experiments. Soon after, the pair met at a British Society of Immunology Congress, and Townsend tentatively inquired about the outcome. Zinkernagel said things didn’t work out – he couldn’t repeat the experiments that Townsend had published. Even worse, one of Zinkernagel’s assistants had found out that all the cells sent by Townsend were contaminated with a type of bacteria called mycoplasma, which was known to make cells do strange things. It was a huge blow.
Townsend couldn’t believe it. His laboratory was really careful not to get this type of contamination, and he felt so sure that he hadn’t worked with contaminated cells. So he gently inquired who had done the experiments – and, came the answer, it was one of the junior medics in Zinkernagel’s lab. OK, Townsend replied, I’ll send the reagents and cells again – but this time you have to get somebody experienced in your lab to do it – or do it yourself.
It was a tense situation; there wasn’t anything Townsend could do but wait. Talking to me in 2011, he recalls how he took solace in thinking about Galileo’s 1610 publication, The Starry Messenger, how it caused a similar ruckus. Galileo had built a new ‘spyglass’ – or telescope, as it would later be called. With it, he saw many new stars and discovered four of Jupiter’s moons. And he saw how the moon ‘is not robed in a smooth and polished surface but is in fact rough and uneven, covered everywhere, just like the earth’s surface, with huge prominences, deep valleys, and chasms’.50 He even estimated that some lunar mountains were four miles tall. But, for everybody else, the moon looked smooth, and most just thought Galileo’s observations bunkum. So Galileo made more spyglasses and sent them to some of the sceptics. Only then did it become accepted that the moon was not smooth.
For Townsend, the story had a twofold significance. It showed that the key to having a discovery accepted by others is both that the tools with which the discovery was made have to be shared and that the experiment has to be easy for others to do. If a result relies on extremely rare or new technology, it’s very hard for others to validate it. Townsend’s method to detect cells being destroyed (the so-called radioactive-release assay) was the same type of experiment that Zinkernagel had used in his Nobel-Prize-winning work with Doherty and was easy for any lab to perform.
The second round of experiments was a triumphant vindication of Townsend’s discovery: from then on, Zinkernagel was a staunch supporter of his work. The combination of non-self-peptide with an HLA protein was essentially the ‘altered self’ that Zinkernagel and Doherty had discussed a decade earlier – the idea that a modification to the HLA protein is detected as a signature of disease. When Doherty first read Townsend’s papers, he couldn’t believe he’d missed the discovery for himself, and he thinks that, ‘if they had named a third person on our Nobel, it should have been Alain [Townsend]’.51
But the mystery still remained in picturing clearly how HLA protein and a fragment of virus protein could activate T cells. How did this work? Bjorkman, Strominger and Wiley were on the verge of solving the problem – they had the raw data in hand – but the analysis to reveal the shape of the HLA protein was still difficult. Another postdoctoral researcher had joined Wiley’s group, Mark Saper, and he did a lot of work in analysing Bjorkman’s data.52 Ninety per cent of the protein chain could be traced but there was that fuzzy 10 per cent yet to give up its secret. They didn’t think that it could be a peptide like those that Townsend used, because Bjorkman made her crystals from pure HLA protein taken from cells that weren’t infected with a virus.53
The moment of truth finally came in spring 1987. With all of HLA-A*02 accounted for, the troublesome fuzzy 10 per cent occupied a groove at the top of the HLA protein. It was positioned on top of a flat sheet of protein, cupped between two long helixes of HLA protein. The fuzziness was indeed the size of a peptide. The HLA protein, it transpired, was perfectly formed for clasping and displaying peptides.54
