It is a big ask. But it is not an unreasonable ask, or too high a price to pay for scientific credibility. Extraordinary claims demand extraordinary evidence and, in a Cartesian world, what could be more extraordinary than the body being linked to the mind through the immune system?
A stubborn fact
Many of the first immuno-psychiatrists used the same simple, experimental design to get started on their pioneering endeavour. They recruited two groups of volunteers, a group of patients with major depressive disorder, or MDD (called the cases), and a group of healthy people (called the controls). They collected a sample of blood from each volunteer and measured a few inflammatory biomarkers in the blood: either cytokines or C-reactive protein (CRP), which is produced by the liver in response to high levels of cytokines and is therefore useful as an indirect marker of the body’s state of inflammation. Then these biomarker data were analysed to estimate the size of the difference between cases and controls, and to test the probability that the difference was statistically significant, meaning it was unlikely to have occurred by chance.
In the 20-plus years between 1992 and 2014, immunopsychiatrists have reported cytokine measurements on thousands of MDD cases and healthy controls.42 Collectively, these data show that the blood concentrations of CRP and some cytokines are increased in patients with depression. The probability of seeing differences this big by chance is in the order of one in 10,000.
It’s not a massive effect but it’s there:43 on average, people with depression have moderately but significantly increased blood cytokine levels compared to non-depressed people.
These case-control studies, which measure levels of an inflammatory biomarker like CRP in depressed cases compared to healthy controls, start from the idea that it is reasonable to put people into one of two categorically distinct boxes - depressed or healthy. A case-control comparison formalises the common assumption that depression can be conceived in terms of us and them: we are perfectly healthy and they are very different from us because they are depressed. But there is another way of looking at things, which starts from the idea that we are all on a spectrum, we all have some experience of depressive symptoms, mild or severe, and in real life there is not a black-and-white distinction between depressed cases and healthy controls. Taking this more dimensional approach, the question is whether people who are located towards the more severe end of the depressive spectrum tend to have higher blood levels of inflammatory markers; and the answer is yes, they do.
One of the largest studies to date measured CRP and depressive symptoms in 73,131 people recruited from the population of Copenhagen.44 Ordinary members of the Danish general public who frequently experienced low-grade depressive symptoms - such as thinking that they weren’t accomplishing much or wanting to give up - had significantly higher blood CRP levels than those who didn’t. There was something like a dose-response relationship in the data: the higher the dose of inflammation indicated by CRP, the greater the depressive response in terms of negatively biased, self-critical thoughts. The probability that this relationship could have occurred by chance was reckoned to be less than one in a trillion.
This is an impressively solid piece of evidence to add to the cumulative evidence from case-control studies of MDD. We now know that people feeling somewhat depressed in the course of their ordinary lives, as well as the more severely affected patients, are more likely to be inflamed. It is important to be clear that these studies don’t prove that everyone with depression is inflamed; or that everyone who is inflamed will be depressed. But they do give us very robust statistical evidence that depression and inflammation occur together much more often than we would expect as a result of random coincidence or bad luck. And as the studies have accumulated since the pioneering work of the 1990s, the evidence has become progressively more solid. Depression is associated with biomarker evidence of bodily inflammation. It’s the stubborn survival of a disruptive fact. But this result alone does not prove that the relationship between inflammation and depression is causal. We need to get a tighter grip on the slippery but crucial issue of causality to begin to answer the how question.
Causes must come first
We know, by definition, that effects follow causes or that causes precede effects. So, if inflammation caused depression, we would expect that people were inflamed before they became depressed. Many of us will have had personal experiences of feeling low, blue, depressed, subdued or tearful after (not before) an infectious illness or some other inflammatory episode. My experience of post-dental melancholia was certainly consistent with the ordering of inflammation before depression in time: I felt fine before I went to the dentist and suffered a cytokine squall; then I felt lethargic, isolated and pessimistic for the following 24 hours.
One way to study the sequencing of inflammation and depression is by making repeated measurements of cytokine levels and mood states in the same people followed up repeatedly over a period of time. The 2014 study of 15,000 children born in south-west England, which followed them up repeatedly from nine to 18 years old, found that cytokine levels at the age of nine predicted the risk of depression at the age of 18.7 The children who had blood cytokine levels in the top third of the range aged nine were about one-and-a-half times more likely to become depressed aged 18 than the children who had lower levels as nine-year-olds. Importantly, the children with higher cytokine levels were no more depressed than their less inflamed contemporaries when they were first assessed at the age of nine. They only became depressed after they were inflamed.
Similar results emerged from another study, towards the other end of the life cycle, of UK civil servants working in Whitehall.45 About 2,000 people in their sixties were assessed for mood and inflammation on three occasions, in 2004, 2008 and 2012. Low-grade inflammation was quite common in this group of older people. Four hundred of them were chronically inflamed, with high CRP levels in both 2004 and 2008, but were not depressed at the time of these first two assessments. However, their risk of becoming depressed for the first time in 2012 was significantly increased, especially if they were women. A woman with high CRP levels in 2004 and 2008 was about three times more likely to become depressed for the first time in 2012 than a woman with no CRP evidence of inflammation previously.
These long-term follow-up studies, in two very different age groups, demonstrate that increased inflammation can precede depression. But still these results alone are not sufficient to establish a causal mechanism. To put it another way, if inflammation at the age of nine did not increase the risk of depression at the age of 18, a sceptic might reasonably conclude that a causal influence of depression had been ruled out or refuted. However, the opposite result, as was actually observed, is not so decisively affirmative. Showing that cytokine levels in the blood predict scores on a depression questionnaire four or nine years later is permissive of a causal role for inflammation; but not by itself conclusive or compelling. The time interval between inflammatory cause and depressive effect is too long, and the underlying chain of events that presumably sustains the causal process over many years is tenuously understood. However, we can begin to close that explanatory gap by looking at the relationship between inflammation and depression over much shorter periods of time.
One of the many areas of medicine where immunology has delivered therapeutic progress in my lifetime is treatment for hepatitis. This is a viral infection that comes in three forms, A, B and C. Hepatitis B is particularly dangerous because the virus can lurk within liver cells for many years, evading the immune system’s routine efforts to eliminate it completely, causing chronic inflammation and scarring of the liver (also known as cirrhosis), and increasing the risk of cancer of the liver. Among the first treatments to make a difference to this dismal prognosis was interferon, an inflammatory cytokine. The therapeutic rationale is that the hepatitis B virus is immunologically camouflaged; the patient’s immune system doesn’t naturally see it as the mortal threat it is, and so his immune response must be massively boost
ed to help him clear the virus completely.
Since the stakes are so high it has been ethically acceptable for an effective treatment to be extremely unpleasant, which interferon is. All patients immediately react to it as if it was a severe infection - they become feverish, lethargic and anorexic - the same cluster of symptoms, called sickness behaviour, that you’d expect to see in a rat injected with cytokines. This is not a side effect but a sign that the treatment is having its intended, principal effect of stimulating an inflammatory response. Over a few weeks, most patients recover from the acute effects of interferon treatment but about a third become clinically depressed. They are persistently lethargic and anorexic, self-critical, guilty and pessimistic, as well as anhedonic or lacking a sense of pleasure,46 the opposite of pleasure-seeking hedonists.
It is important to be clear that this happens to people who were not depressed immediately before the interferon injection. Their experience provides some of the clearest evidence in humans that an inflammatory stimulus can cause depression. By comparing the patients that become depressed following interferon to those that don’t, we can understand something more about the mechanisms involved. It turns out that patients with a background history of depression are more likely to become depressed again following interferon. This could be because they have a genetic predisposition to respond depressively to inflammation. And indeed there is some evidence that people with a genetic profile that makes them more likely to get inflamed, to produce high levels of cytokines, are more likely to become depressed after interferon treatment.47
Putting it all together, from the long-term epidemiological follow-up studies, to the experiences of patients following interferon treatment, and even including my anecdote about root canal blues, there is evidence that bodily inflammation can precede depression. And if inflammation occurs before depression then it could cause depression. We haven’t yet answered the question how. But we have at least established that it is a question worth asking in more biological detail.
The Berlin Wall in the brain
I was not always a happy medical student in the first half of the 1980s. I did my clinical training at St Bartholomew’s Hospital in the City of London, which was founded in 1123 outside the city walls by Rahere, a monk and a minstrel in the court of Henry I. Henry VIII refounded the hospital in the 16th century; Harvey did his experiments there on the circulation of the blood in the 17th century; Hogarth painted the murals in its Great Hall in the 18th century; and, in 1878, according to Conan Doyle, Dr Watson met Sherlock Holmes for the first time in a chemistry lab at Bart’s, where Holmes, perhaps blazing the trail that Freud would later follow, was investigating the pharmacological properties of a “little pinch of the latest vegetable alkaloid”.48
Bart’s had long been home to some of the most eminent medical minds in London. Percival Pott, to name but one, was an 18th-century surgeon who had the unrivalled honour of two diseases named after him in our textbooks: tuberculosis (TB) of the spine and the chimney-sweep’s disease of cancer of the scrotum. Pott worked out that scrotal cancer was caused by exposure to carcinogens in chimney soot and he led legislation that banned the practice of sending orphan boys as young as five or six climbing up chimneys to clean them. Bart’s proudly claimed him as the first doctor in history to find a cause and a cure for a cancer. The hospital had been there for all of London’s plagues, from the Black Death in the Middle Ages to TB in the Victorian era; it had survived all of London’s disasters, from the Great Fire to the Blitz. It was an ancient, enduring and honourable institution.
However, perhaps for that very reason, the way that Bart’s taught medicine at the time was by a dogmatic and didactic apprenticeship: motto, you can always tell a Bart’s man but you can’t tell him much. We had to learn endless lists of symptoms, signs and pathophysiological catechisms - like the 32 causes of anaemia - and were often quizzed by consultants on teaching rounds to recite these lists in front of other students and staff. It was important under the stress of this ritualised public interrogation not to blurt out wrong answers or, almost as bad, the right answers but in the wrong order.
“A woman comes to see you with a headache, what are the first 10 diagnostic tests you’d do?” If your immediate response was “A brain scan” that deserved some supplementary sarcasm because, as any fule kno,49 that wasn’t the first thing, it was more like the 10th thing, you’d do. “What’s the first thing you’d do?” “Talk to the patient.” “Thank you. And what are the first three questions you’d ask her?”
And so it went on in the time-honoured fashion for three years. We were being drilled as much as trained, over-learning or hard-wiring certain key nuggets of medical knowledge, and learning by example a certain style, and language, and way of working as a doctor. We were generally not encouraged to question too much of the wisdom of senior physicians and surgeons. I was not the only student to resent being taught like this and almost all of us got through it OK. But I think it explains the great pleasure it gives me now to see false facts, which I was taught to learn with black-and-white certainty, on pain of public humiliation, being scientifically usurped by new knowledge (Fig. 9).
Figure 9: What I was taught at medical school, in the 1980s (left) was that depression was caused by reduced levels of serotonin in synapses between nerve cells. It was also common knowledge that the brain was completely separated from the body’s immune system by the blood-brain barrier (BBB), a wall of densely packed endothelial cells, which did not allow circulating macrophages or cytokines to enter the brain.
What we know now (right) is that there are many channels of communication across the BBB. Inflammation of the body can cause activation of the inflammatory microglial cells in the brain, which in turn causes collateral damage to nerve cells in the amygdala, the cingulate and other hubs in the brain’s emotional network. We can also see more clearly that there are many potentially relevant drivers of inflammation. Auto-immune disease, like Mrs P’s arthritis, obesity and physical trauma can all cause bodily inflammation. But so too can social stress, even a brief and relatively mild stress like public speaking. It used to be impossible to imagine that inflammation and depression were mechanistically connected; now we are getting closer to answers to the questions of how and why inflammation could cause depression.
The brain is immune privileged - that is what we were taught at medical school not all that long ago. It sits behind the blood-brain barrier (BBB) and the cells and cytokines of the immune system simply can’t get at it. The BBB rigidly defended the brain from the inflammatory storms of the body. It could only be breached by the immune system in the wake of catastrophic damage to the brain, such as the surprise attack of a stroke or the remorseless growth of a tumour. Under more normal operating conditions, the BBB was assumed to be an impermeable defence of the brain’s unique privilege to work beyond the reach of the immune system. To the extent that this dogma was true, it would obviously be a major obstacle in the mechanistic path from an inflammatory protein in the blood to a state of mind. If peripheral inflammatory signals can’t get across the BBB, they can’t have any effects on the brain; and if they can’t have effects on the brain, then how can they have effects on mood or behaviour? You can see why the Berlin version of the BBB was one of the most concrete formulations of Cartesian dualism. It enforced segregation of the inflamed body and the mind by blocking any communication between them. Happily, it’s mostly wrong.
Even at the time, the analogy wasn’t exact. The Berlin Wall was built from slabs of reinforced concrete whereas the BBB, we were told, was built from millions of cellular bricks, the endothelial cells, that formed the inner lining of the blood vessels in the brain and were very tightly connected to each other. There was literally no space for immune cells, or even large molecules like cytokines, to get between the endothelial cells, as they would have to do if they were to migrate from the bloodstream into the brain tissue on the other side of the barrier. To turn the analogy around, it was as if the Berlin wal
l was supposed to be impenetrable mainly because it was built out of bricks that were held together by an especially dense, adhesive mortar.
Now we know that there are some parts of the brain where this is simply not the case: there are gaps between adjacent endothelial cells - there are chinks in the mortar - that are big enough to allow big molecules, like proteins, to diffuse freely from the blood into the brain. More radically, it turns out that the endothelial bricks in the wall are not inert, like bricks of baked clay; they are double agents in the immune system’s communications network. One side of each cell forms the inner lining of a blood vessel - an artery or vein - and the other side of the cell forms the outer surface of the vessel, in close proximity to nerve cells and microglial cells (the brain’s robocops). The inner surfaces of the endothelial barrier are covered with cytokine receptors, so they can detect inflammatory signals communicated by cytokines circulating in the blood. Endothelial cells can then relay these inflammatory signals into the brain and activate the brain’s resident macrophages, so that the brain becomes inflamed in response to inflammation in the rest of the body.
The “wall” is not only permeable to inflammatory proteins, it is also permeable to the much larger inflammatory cells that are continually circulating through the cardiovascular system. The inner surface of the wall can make itself attractive to circulating white blood cells and actively assist their passage into the brain, squeezing through specially created gaps between endothelial “bricks”. It was even discovered a few years ago that the brain has a system of lymphatic vessels that drain immune cells and proteins from it into nearby lymph nodes, where they can mingle with other cells of the immune system and then return to the blood circulation.50 In direct contradiction to what we were taught as certain knowledge in the 1980s, the brain is not cut off from the immune system in the body. There is free and easy communication by many channels in both directions across the BBB.51
The Inflamed Mind Page 11