The Compatibility Gene, page 20
For Shatz’s work, the main source of contention was the lack of anyone having a detailed understanding about how the MHC proteins could influence organization of neuronal synapses in the brain. For any phenomenon that strays from what’s expected, biologists like a good explanation of how it works before taking the new idea as fact; as the mantra goes: show me how it works before I believe it’s true. Although sceptical, most scientists were intrigued, and if she was right then a burning issue was whether or not this reflected a trivial recycling of molecules to be used for a different task or whether, instead, this cracked open a fundamental link between our nervous and immune systems.
Given that compatibility genes – and the proteins they encode – vary so much, it might seem unlikely for the brain to co-opt such a complex system unless their variability was somehow important for their role in the brain. But this argument is flawed. Because it requires that we – and other living things – have evolved to end up working in an efficient or sensible way. In fact, there are many examples in nature where things work fine without being slick like an iPad. The way that we have been formed – through a process of evolution – means that everything must be built upon what went before, not constructed elegantly from scratch.
Take, for example, the path of a particular thin tube in men called the vas deferens, which transports sperm from the testicles to the urethra (it’s the connection that’s severed in a vasectomy). Instead of this tube taking a straight path, it follows a route far longer than it needs to by looping over another bit of tubing in the area, the ureter. There doesn’t seem to be any reason why it couldn’t follow a direct path; it just happened to evolve to go the long way round. This has been explained by suggesting that the position of the testicles changed as we evolved from our ancestors and as that move occurred, the vas deferens tube got caught over the ureter rather than going under it. So, things don’t evolve to a perfect design, and that’s why it can’t be taken for granted that the brain uses the same molecules as the immune system for any fundamental reason; it may have just happened to evolve that way. To claim that the connection between our immune and nervous systems is intimate – and not just a chance recycling of parts – there needs to be specific evidence.
We’ve discussed in detail how compatibility genes are linked to susceptibility or resistance to various types of infectious diseases, but also our variation in these genes has been linked to many neurological disorders, such as schizophrenia or bipolar disorder.22 This is consistent with an intimate connection between compatibility genes and our nervous system. But researchers studying schizophrenia, and other neurological diseases, differ in their view of how important these genes are. Although many tens of studies link compatibility genes to schizophrenia, dispute remains because each comes to a different conclusion about which versions of these genes are risk factors for the illness.23 This might reflect different diagnostic criteria used in studies: mental illnesses are notoriously difficult to categorize, and it is possible that different compatibility genes are important in particular versions of these multi-faceted diseases. There are at least three ways in which compatibility genes could, in principle, influence neurological illnesses.
An infection may underlie some mental illnesses; and then one way in which compatibility genes would be linked to neurological diseases is through their normal role in immune defence. In this scenario, the situation is identical to how these genes affect our susceptibility or resistance to other infectious diseases, such as AIDS. A second possibility is typified by the rare sleeping disorder narcolepsy, which has one of the strongest links of any disease to our HLA genes. About 1 in 2,000 people are affected so that their brain is unable to regulate the normal cycle of being awake and asleep.24 Sufferers can fall asleep at inappropriate times during the day and sleep poorly at night. The vast majority of people with narcolepsy have particular versions of class II compatibility genes.25 These versions of compatibility genes are found in almost all people with narcolepsy, but they are also common in those without the disease – so they are not sufficient to cause the illness and instead play some role in how it starts.
For some sufferers, there is evidence that narcolepsy is an autoimmune disease. That is, it arises because the immune system mistakenly attacks the body’s own healthy tissue. Symptoms could be caused by an immune reaction against the neurons that are important for regulating sleep and wakefulness. In support of this, a protein made by neurons triggers an immune reaction in narcolepsy patients.26 So HLA proteins could influence our susceptibility to narcolepsy according to the extent that they aid an immune attack on healthy neurons.
The third way in which these proteins could influence neurological disorders directly relates to Shatz’s discovery: that these proteins can influence the way in which neuronal connections or synapses are configured in the brain. One possibility of how this might relate to disease stems from the fact that, during any kind of immune reaction in the body, immune cells secrete proteins – called cytokines. Cytokines do many things to help the immune reaction, one of which is to increase the production of MHC proteins. So if cytokines were secreted from immune cells at a time when these MHC proteins are important in sculpting the developing brain, it is feasible that the change in their production could lead to abnormalities in the structure of the brain.27 But this is only an idea – nothing more than a feasible way in which compatibility genes link with neurological illness. Sadly, the causes of most mental illnesses remain baffling.
Aside from disease, Shatz’s discovery also begs the question of whether or not compatibility genes affect normal brain functions – perhaps in some way that’s more subtle than influencing the course of illness. Since her observations indicated that MHC proteins were important in how the brain changed in response to external stimuli – both in the visual cortex and then in the hippocampus – Shatz and her team decided to next investigate, in 2008, whether or not MHC proteins affect learning, a process that must involve changes to the brain.
It’s exceptionally difficult – perhaps impossible – to test whether or not immune system genes influence how well humans learn, but Shatz could more easily explore the idea in mice. First, Shatz clarified which types of compatibility genes were being used in the mouse brain. The genes that vary enormously among us – HLA-A, -B, -C and so on – are called classical compatibility genes. But there are also a number of other genes – in humans and animals alike – that encode similarly shaped proteins which don’t vary much between us. These are named the non-classical compatibility genes. Shatz’s earlier work didn’t distinguish between these different types of compatibility genes, but for their investigation into learning her team focused on the genes in mice that are equivalent to our HLA-A and -B genes – the classical compatibility genes that vary hugely in individual mice or people.
Mice genetically altered to lack these variable genes did indeed have defects in the cerebellum at the base of the brain. Specifically, synapses were not weakened in the right way, reminiscent of what she had discovered in other parts of the brain. The cerebellum is important for motor learning – how we learn skills by practising like riding a bike. Although we have little idea about how it works in any detail, it’s generally accepted that motor learning will somehow require a reconfiguration of the connections between neurons – some being strengthened while others weaken. So if compatibility genes can influence which synapses weaken in the cerebellum, Shatz reasoned, could they affect the ability of mice to actually learn a skill?
To test this directly, Shatz’s team used a simple apparatus called a Rotarod – a horizontal cylinder that rotates, on which mice can learn to balance so that they don’t fall off. It’s not harmful for the mice because the bar isn’t so high that it hurts when they fall. But timing how long an individual mouse can stay balanced on the bar is a way of measuring their ability. After some practice, normal mice were able to stay on the rotating rod for around a minute, but mice lacking the equivalent genes to our HLA-A and –B genes could learn to balance for nearly double that time.28 A lack of these genes isn’t something that occurs naturally, but, by deliberately testing what does happen if they are missing, this experiment revealed that compatibility genes can influence the ability of mice to learn.
Mice lacking these genes also remembered their skill for much longer because, after a four-month break from practising, they were still much better at balancing on the rod. How would our hero from Chapter 1, Peter Medawar, have reacted if a know-it-all alien had visited and whispered: Peter, the genes you’re studying – those that control the transplantation reaction – also work to fight infections, help organize the brain and influence learning? Still, stunning as these results are, there are at least two caveats – both of which Shatz acknowledges herself. First, it remains unknown whether or not humans (not just mice) use MHC proteins in this way in the brain; the problem is that it’s very hard to think up an ethically sound experiment to test these ideas on people. Second, Shatz’s experiments only show that the brain is affected in mice whose compatibility genes are completely incapacitated – an artificial situation just to test what happens if these proteins are removed. This doesn’t really tell us whether or not the brain could be influenced by the natural diversity in these genes. If there is an effect caused by differences in these genes it is surely going to be subtle, compared to when these genes are incapacitated.
Our understanding of compatibility genes in the brain has simply not yet benefited from the decades it took for everyone to agree, for example, the importance of Medawar and Burnet’s work. It is simply too soon for there to be unanimous agreement over the importance of Shatz’s observations (or pretty much anybody else’s discovery made within the last decade or so). Almost every statement in any scientific textbook is the outcome of many years’ hard work and debate; and MHC proteins working in the brain is beyond current textbook-level science. While many scientists would argue that a popular-level book like this one should also stick to established decades-old ideas, my view is that nothing can be more exciting than what’s happening at the edge of knowledge. And most questions about the brain are at the edge of our knowledge, but the field is ripe for breakthroughs in the twenty-first century.
More than 32,000 people attended a neuroscience congress held in Washington, DC, in 2011, for example, while the same meeting forty years earlier attracted just over 1,000 people.29 No other scientific discipline enjoys such stadium-sized lectures; Shatz says it feels like everyone’s becoming a neuroscientist. When she first presented her data indicating that compatibility genes are active in the brain, in 1998, it was as a lone mention of the subject, but the neuroscience congress in 2011 was chock-full of presentations about how immune-system genes influence the brain and nervous system.30 In fact, many components of our immune system are now known to influence our nervous system.
Receptors that immune cells use to detect bacteria, for example, can affect the extent of brain damage that occurs in a stroke.31 This, with several related discoveries, has established that stroke and many other neurological problems can be triggered or exacerbated by immune responses. The medical implication is that targeting immune-system components might help. Drugs that block secretions from immune cells (cytokines), for example, could alleviate the symptoms caused by neuronal injury in stroke or brain trauma.32 That Shatz helped establish connections between immune and nervous systems that soon might be medically useful must be gratifying, but when I put this to her in November 2011, she replied that, ‘like all discoveries, first everyone says it’s wrong, then everyone finds it out for themselves, and eventually everybody forgets that you made the discovery in the first place’.33
With hindsight, it maybe shouldn’t have been so surprising that molecules are shared between our immune and nervous systems. These systems must be intimately connected; we’ve all experienced feeling sad or sleepy when ill. Indeed, immune responses are connected to all kinds of physiological processes.34 In times of stress, for example, steroid hormones are released to alter energy use around the body and increase blood-sugar levels. These same hormones also dampen immune responses. They switch off excessive inflammation, which can damage tissues unnecessarily – this underlies the use of steroid hormones in preventer inhalers for asthma, to dampen immune responses in airways. A release of adrenaline triggered after an acute injury has the opposite effect; it stimulates or primes immune cells for action.
The interaction between our nervous and immune systems goes in both directions, as secretions from immune cells also affect the brain and central nervous system (and are the likely cause of us feeling sad and sleepy when ill). In fact, there is a vast web of neuro-immune circuitry that is critical to our well-being. Exercise, for example, affects the levels of various hormones and other proteins that circulate through blood, including adrenaline and cortisol levels, which in turn influence immune responses.35 Regular exercise can have an anti-inflammatory effect, for example, which can protect against diseases in which a chronic immune response is part of the problem – such as in type 2 diabetes.
More research in this area is important but in deciding what next to do – where to target funding – it’s useful to remember that Shatz’s discovery came from exploration into how vision works, not a specific attempt to study the intersection between the immune and nervous systems, nor a direct effort to tackle any specific disease. Sparks might emerge from fostering greater interaction between scientists working in silos for understanding either immune cells or neurons. Much can be learned by thinking about the differences and similarities between these two types of cell.
In 1994, immunologists Bill Paul and Bob Seder at the National Institute of Health in Bethesda wrote a speculative but hugely influential article suggesting that neurons and immune cells have some similarity in how they work.36 They reached this view because of an experiment that others performed in 1988 which showed that immune cells could secrete molecules in a specific direction, something that neurons had long been known to do.37 The importance of this was in showing that immune cells and neurons can affect another cell in contact, rather than all the cells in the vicinity. Soon after Paul and Seder’s article was published, Abraham ‘Avi’ Kupfer, working alongside his wife Hannah at the National Jewish Medical and Research Centre in Denver, performed experiments which directly showed a striking similarity in the way in which immune cells and neurons work.38
Kupfer’s discovery came through watching immune cells in action with a high-powered microscope. In 1995, he stood before an unsuspecting crowd of a few hundred immunologists gathered for one of the prestigious Keystone symposia – named after a US ski resort where the meetings are often held. He showed images of immune cells interacting with other cells which revealed that the contacts between them involved aggregates of proteins organized into bull’s-eye patterns.39 His images showed two cells in contact – like two balls squashed together – and across the flattened connection between the two cells a patch of one protein coloured red could be seen surrounded by a ring of another protein labelled green. Before that moment, nobody thought these proteins would arrange themselves into a pattern at the contact between cells – but it was reminiscent of the organization of molecules at neuronal synapses.
This led to the common use of the term ‘immune synapse’ to describe the contacts that immune cells make with other cells. Both types of synapse involve rings of proteins to promote adhesion between cells and patches of other proteins particular to the discussion between the cells. To the Keystone audience, Kupfer’s pictures were instantly accessible, and the immediate implication was that our thoughts and the detection of a virus both work through a complex choreography of molecules at the contacts between cells. One immunologist in the audience, Anton van der Merwe from Oxford University, remembers the event well:
I recall us looking at these beautiful images for the first time in stunned silence. Although his talk overshot the allotted time no one showed any sign of leaving. After he had finished there was prolonged applause followed by many questions. When the chairperson ended the question session many of us crowded around Avi to continue the discussion.40
Independently, Mike Dustin, then at Washington University School of Medicine, St Louis, and his collaborators were also imaging immune cells but with an interesting twist. Instead of imaging two cells interacting together, they replaced one of the cells with a surrogate membrane composed of the lipids or fat molecules from a real cell but laid out flat on a glass slide. As immune cells landed on this glass-slide-supported mimic of a cell surface, they could also see dramatic movements of proteins labelled with different-coloured dyes. This artificial system was easier to image because the microscope could rapidly capture pictures of the flat synapse laid out over the glass slide. Their approach revealed that the immune synapse is dynamic so that arrangements of proteins change when, for example, a T cell responds to the presence of non-self (peptide).41
My contribution to the story of compatibility genes comes here. While working with Jack Strominger at Harvard University, independently from Kupfer and Dustin, I also discovered a structured immune synapse – but this time, formed by human Natural Killer cells.42 I vividly recall looking at a computer screen, which relayed what was being detected down the microscope, to see patterns of differently coloured proteins at the contact between cells. Unsure of my ability to make an important discovery, I had to ask my girlfriend at the time (now my wife), who isn’t a biologist, to come into the lab and use the microscope herself – to be sure I wasn’t doing anything wrong. My research showed that synapses were important for different types of immune cell – and showed that different organizations of the synapse can switch immune cells on or off. The new science opened up by my research – together with Kupfer and Dustin – is that changing arrangements of molecules control immune-cell interactions, turning them on and off when needed, analogous to what happens at neuronal synapses.
