The Inflamed Mind

by Edward Bullmore

  Despite the radioactive fall-out from the implosion of the old business model that still hangs over depression, the still-fresh memories of huge investment that didn’t go according to plan, repurposing could help the industry to start going back there. In the best case, there could be a wave of innovation, based on repurposing some of the hundreds of anti-inflammatory drugs that are already in existence but have not traditionally been regarded as relevant to disorders of the mind. And in the very best case, this could happen relatively quickly, in 5-10 years, say, rather than the 20 years it usually takes to get from a new target, like serotonin circa 1970, to a new medicine, like Prozac circa 1990.

  It is also encouraging that there is a virtually limitless number of potential biomarkers that could be used to predict which patients are more likely to respond to anti-inflammatory drugs. I have talked a lot about CRP, but that doesn’t mean it is the best or only biomarker for inflamed depression. It simply reflects the fact that CRP has been around for a long time in medicine, it was readily available to use in the first immuno-psychiatry studies in the 1990s and it has since served as a leading light in an originally obscure area. But I would expect there to be better biomarkers out there: tests that are capable of showing greater differences between groups of patients, or that are more precisely related to the mechanism of action of a new drug, than blood levels of CRP or cytokines. Modern immunology boasts an extraordinary range of techniques for profiling the peripheral immune system, many of which are only just beginning to find their way into depression research. CRP has already proven to be the first useful immune biomarker for depression; it certainly won’t be the last, or the best for all purposes.

  We can also be encouraged, to some degree, by looking back at old clinical trial data through a new lens. In the last 10 years, following the pioneering lead of anti-TNF (tumour necrosis factor) antibodies for treatment of rheumatoid arthritis, dozens of anti-cytokine antibodies have been tested in trials for many different inflammatory disorders. As you’d expect, all the anti-cytokine antibody trials published to date have been designed according to an experimental plan or protocol that prioritises measurement of the drug’s effect on the patient’s physical health. Most of the studies of new antibodies for rheumatoid arthritis, for example, have used physical examination of swollen joints as the primary endpoint, the key measure of whether the drug works, and whether the trial will be regarded as a success. Mental health has not been entirely neglected in these clinical trials for physical health disorders. It has often been measured, if cursorily, as a secondary endpoint, using questionnaires that simply ask patients, on a scale of 1 to 4: How depressed are you feeling? How much energy do you have? So, it is possible to reanalyse these secondary endpoints, these mental health scores, as the principal outcome of the study, as if the studies had been designed to test the drug effects on depression (rather than joint swelling in arthritis). And the results are apparently impressive. Recent studies that have reanalysed mental health data from dozens of placebo-controlled trials on tens of thousands of patients with various disorders, including rheumatoid arthritis, psoriasis and asthma,86-88 have shown that the anti-inflammatory drugs tested had an anti-depressant effect size of about 0.4, on average. How big a deal is that? 0.4 might not look like a big number but bear in mind that the average effect of SSRIs is only about 0.2 on the same scale. On the face of it, new anti-inflammatory drugs could be twice as effective at treating symptoms of depression as currently standard anti-depressant drugs.

  But there’s a catch. It’s another version of the same old Cartesian catch. In all the clinical trials conducted so far, the drug effects on depression have not been measured for the first time until about two or three months after the start of treatment. By then, many patients will have experienced significant improvements in their physical health. Arthritic patients will have less painful joints and better looking X-rays of their joints; psoriatic patients will have fewer and smaller plaques of inflamed red skin on their faces and elbows. And as any Cartesian will be quick to point out, if you thought you had an incurable disease, then you tried a new treatment that cured it, you would feel a lot less depressed, wouldn’t you? The anti-depressant effects of anti-cytokine treatment are superficially visible but deeply discounted by the same line of reasoning that licenses medical indifference to the mental health symptoms of inflammatory disease.

  It comes back to the question of causality. To show that anti-inflammatory drugs can directly cause an improvement in mental health, rather than a mental reaction to improved physical health, we need to see that the beneficial effects on mental health precede or anticipate any later effects on physical health. The under-investigated Remicade high - the rapid boost in mood that many patients report soon after their first dose of an anti-TNF antibody - indicates that anti-inflammatory drugs could have such rapid anti-depressant effects. And, of course, it would be very useful for doctors and patients to have access to anti-depressant drugs that worked more rapidly than the two to six weeks usually required for SSRIs to work. But we will have to do the studies.

  So far there have been very few placebo-controlled clinical trials deliberately designed to test the anti-depressant effects of anti-inflammatory drugs, and their results are not conclusive. Only one study has been reported of an anti-cytokine antibody.89 Sixty patients with treatment-resistant depression, who had not responded well to conventional anti-depressant drugs, were randomly assigned to treatment with an anti-TNF antibody or a placebo. After eight weeks, the group of patients treated with the antibody reported substantial improvement in the severity of their depressive symptoms; but so too did the patients treated with a placebo. There was no significant difference, on average, between the two groups. In that sense, the trial was negative.

  But when the investigators dug a bit deeper into the data, they found that not all patients responded to treatment in the same way. The patients who had higher levels of CRP at baseline, before the start of the trial, had a stronger antidepressant response to treatment than the patients who had lower levels. In other words, the anti-inflammatory drug was not a panacea: it seemed to work better for depressed patients who were inflamed than it did for depressed patients who were not. In that sense, the trial was positive.

  It points to a future in which trials for an anti-inflammatory drug for depression will routinely use an inflammatory biomarker up front to identify those depressed patients that are most likely to benefit from treatment. Despite the inherent risks in any drug development project, I think there will be significant investment in this new kind of anti-depressant drug trial in the immediate future. And it will certainly be interesting to watch this space over the next few years.

  But what about non-drug treatments for inflamed depression? Are there any other ways, besides drugs, that we can break the vicious cycle that links stress, inflammation and depression?

  We know from the recent discovery of the inflammatory reflex that the vagus nerve controls cytokine release by macrophages in the spleen. We also know that vagal nerve stimulation by an electrical device implanted in the body can dramatically reduce inflammation and improve symptoms in patients with rheumatoid arthritis. What is not widely known is that vagal nerve stimulation has been licensed for treatment of depression since 2005.

  Many depressed patients have had stimulating electrodes implanted close to the vagus nerve as it travels down the neck, with a control device just under the skin so the patient can adjust the timing and duration of vagal stimulation. This procedure was licensed because it is safe and apparently efficacious. It seems to work, although most of the studies have not controlled for a placebo effect, so it has questionable added value. And, if it works, it is still not clear how it works.90 The traditional explanation, which is not strongly supported by experimental data, is that the electrical stimuli from the device travel up the vagus nerve to the brain stem, where they activate the cells that manufacture serotonin and noradrenaline, and this increases serotonin signalling to t
he rest of the brain. In other words, vagal nerve stimulation has been thought to work like an electrical SSRI. But maybe it works more like an electrical anti-cytokine antibody? It may be that the anti-depressant effect depends on the electrical stimulation passing down the vagus to the spleen, not up the vagus to the brain, and is explained by reduced inflammatory cytokines in the body, not by increased serotonin in the brain. We don’t know right now.

  If it became clearer that vagal nerve stimulation worked for depression via its anti-inflammatory effects, this could open the door to using blood biomarkers to predict which patients are most likely to benefit from having an expensive stimulator surgically implanted. And it might foster further research to develop smarter, less invasive ways of implanting stimulators and delivering electrical stimuli to the vagus nerve. The technology of bio-electronics - devices for electrical monitoring and stimulation of biological processes - is moving rapidly, but not yet in the direction of new treatments for depression. It is conceivable that this could change and in the next 10 years or so we might see a new generation of bio-electronic devices appear that can electrically subdue the inflammatory signals that drive depression.91

  We also know, from the recent discoveries about stress-related inflammation, that social and psychological shocks, like public speaking or an abusive relationship, can increase bodily inflammation. That might make you think that a course of psychotherapy or meditation, focused on helping patients to strengthen skills for stress management, could have anti-inflammatory effects. And indeed there is some evidence for this. Mindfulness training reduced loneliness in older adults and also reduced expression of inflammatory genes by white blood cells.92 A recent combined analysis of the results of multiple studies of the immunological effects of mind-body therapies, like meditation or tai chi, found that they significantly reduced the expression of genes that control activation of macrophages in response to infection.93 It seems that the mind can be trained to control the inflammatory response of the body, and this might be one of the mechanisms by which psychological treatments are effective for depression.

  It might not be so obvious how this neuro-immunological explanation could change psychological treatment for depression, since meditation and other stress management techniques are already moderately effective and widely used. But perhaps there is an opportunity for inflammatory markers to be used as a kind of biofeedback, providing detailed information on how bodily inflammation is progressively controlled as people practise meditation and develop stress management skills. You could call it cytokine-guided psychotherapy. As far as I know it hasn’t happened yet. But there is no fundamental reason, in a post-Cartesian world, why the effects of psychological treatment should be restricted to the mind or why it wouldn’t make sense to try measuring the effects of meditation on macrophages.

  Alzheimer’s disease and the yin and yang of microglia

  Although dementia sounds like an ancient word, as old as melancholia or inflammation, it wasn’t invented until the 18th century, when a couple of Latin words were jammed together to create a new one meaning loss of mind. It was only at the end of the 19th century that it began to occur to the first generations of neuroscientists that dementia could be caused by a brain disease rather than by anno domini, the passing of mortal time. The late 19th-century scientist Alois Alzheimer now has name recognition on a par with Freud, and is hugely more famous than his contemporary and mentor Emil Kraepelin, but he wasn’t a towering figure in his own lifetime. His claim to fame rests on a single case, a woman in her fifties called Auguste Deter, one of his patients in an asylum near Frankfurt, who had a rapidly progressive dementia despite being far from senile.94 When she died, at the age of 56, Alzheimer arranged for her brain to be sent to the anatomy lab he had just been invited to set up in Kraepelin’s new Institute for Brain and Mental Health in Munich. Alzheimer looked down the microscope at pieces of her brain and noticed that there were unusual fibres and clumps of stained material in and around the nerve cells. They were what we now call plaques and tangles. Alzheimer described them clearly enough for us to be sure that’s what he saw; but his colleagues weren’t immediately impressed.

  The story goes that at a psychiatric conference in 1907, where Alzheimer first presented his findings, his lecture was immediately followed by a more keenly awaited presentation of a case of compulsive masturbation, and he faced no questions from the audience. I said earlier that it is stressful to be asked questions in public; and it is, but for a scientist it is also humiliating to be asked no questions at all at the end of a lecture. It implies that nothing you have said has been interesting enough even to excite scepticism. Frau Deter’s brain plaques and tangles might have slipped below the surface of history for ever if not for Kraepelin, who remembered them and relaunched her as the world’s first case of Alzheimer’s disease in the eighth edition of his textbook on psychiatry in 1910.

  Kraepelin considered Alzheimer’s disease to be a rare cause of dementia that occurred in a few, younger people like Frau D, but was not the cause of the much commoner senile dementia, which was attributed to a reduced blood supply to the brain. That is roughly what we were still being taught about dementia as medical students at Bart’s in the 1980s: and “crumble” was what we called it under our breath. It is only in the last 25 years or so, since ex-President Ronald Reagan disclosed his diagnosis in 1994, that Alzheimer’s name has become known to us all. We have realised that the majority of cases of dementia in our ageing societies are due to his disease, the accumulation of plaques and tangles in the brain.

  Alzheimer had no idea what the plaques and tangles were. He described them simply as “a peculiar material”. It has since been discovered that they are formed from proteins, abnormally large quantities of abnormally insoluble proteins, called tau and amyloid. As we get older, all of us will form plaques and tangles of these aggregated and mis-folded proteins in our brains, to some extent; but we won’t all “get Alzheimer’s”. We don’t know why plaques and tangles cause a progressive dementia in some people and not in others, but one plausible explanation hinges on the immune system. Tau and amyloid are human proteins, but they are not normal human proteins. From the point of view of the immune system they are antigenic, non-self, alien proteins; and, as you’d expect, they trigger an inflammatory reaction. The microglia, the robocops of the brain, swarm around amyloid plaques, attacking, eating and trying to digest the “peculiarly impregnable” protein they contain. As you might also expect, microglial activation in response to plaques causes collateral damage; nerve cells are damaged or killed by the toxic effects of inflammation in the brain. In fact, it seems that the secondary, inflammatory reaction of the microglia could be a more potent cause of death of nerve cells, and therefore a more powerful driver of progressive loss of memory and other cognitive functions, than the primary problem of plaques and tangles.

  If it is true that dementia is determined as much by the immune response to Alzheimer’s “peculiar material” as by the plaques and tangles themselves, then anti-inflammatory treatments should be effective in slowing or preventing the progression of Alzheimer’s disease. There is some evidence in support of this prediction, but so far only some.95

  Patients like Mrs P, who must take anti-inflammatory drugs on a regular basis to control symptoms of arthritis or other immune disorders of the body, have significantly reduced rates of Alzheimer’s disease.96 Conversely, untreated infection or inflammation in the body is recognised to increase the risk of Alzheimer’s disease and to accelerate the rate of progression of dementia.58 The inflammatory cytokines pumped into the circulation by macrophages dealing with a chronic infection like periodontitis, say, can get across the blood-brain barrier (BBB) and activate microglia, making them more likely to respond aggressively to amyloid plaques and increasing the collateral damage to nerve cells. That’s one reason why I keep going to the dentist, even though it makes me depressed in the short term. I figure that any steps I can reasonably take to calm down inflammat
ion in my gums and teeth are likely to be beneficial to my ageing brain in the long term.

  However, clinical trials of anti-inflammatory drugs for Alzheimer’s have not yet produced a clear winner. As usual with failed trials, there is some disagreement about the reasons why. It is unlikely that all the drugs that have been tested have been given at a high enough dose, or have been able to get across the BBB into the brain. More radically, some scientists have made the important counter-argument that not all microglial activity is bad. The microglial cells, after all, are trying to do the right thing. They are trying to eliminate plaques from the aged brain, and under the microscope you can sometimes see them stuffed full of amyloid protein that they have eaten and are struggling to digest. There is a reasonable case that, therapeutically, we should be trying to support and assist the good work that microglia are doing, not trying to shut them down. That is the rationale for the development of anti-amyloid antibodies that can get into the patient’s brain and bind to plaques, making it easier for microglia to identify and destroy them. It is also the rationale for the development of vaccines against Alzheimer’s, which involve injecting healthy people with fragments of amyloid to stimulate the production of antibodies that could help microglia deal with amyloid plaques when they begin to form in later life. However, to date, none of the new antibodies or vaccines designed to assist “good” microglia have been more effective than any of the anti-inflammatory drugs designed to inhibit “bad” microglia.