Autoimmune Encephalitis
and Related Disorders of the Nervous System

by J. Dalmau and F. Graus
Cambridge, 2022, 662 pages
reviewed by T. Nelson

or the past two months, the press has been arguing about whether Vladimir Putin has Parkinson's or terminal cancer. Some videos show him grimacing as if in pain. Others show him with an unsteady gait and maybe a tremor. What the heck, they ask, could be wrong with this guy?

I have no idea, but as this book shows, it is now well established that cancer is a major cause of neurological disorders. So it's quite possible that someone with cancer can exhibit dyskinesia, involuntary movements, postural tremor, gait instability, and psychiatric disturbances that may appear to a layman to be purely neurological.

How does that work? For reasons not fully understood, cancer breaks self-immune tolerance, which is supposed to prevent the body from making antibodies against itself. When that happens, the body makes antibodies against proteins on the surface or interior of cancer cells—which is not supposed to happen. These antibodies may be helpful at suppressing the cancer, but if they cross-react with proteins in the brain they will cause neurological problems such as psychosis, ataxia, seizures, memory loss, insomnia, and encephalopathy (irritability, confusion, and lethargy progressing to coma), depending on the antibody. This is called paraneoplastic encephalitis.

The first example, discovered in 2005, was anti-NMDA receptor encephalitis. It is an autoimmune reaction to a teratoma and occurs mainly in women. Patients or the people around them first notice psychiatric problems: they become combative, paranoid, agitated, and hyperemotional. They become insomniac. As the disease progresses, they experience severe psychosis along with seizures and rapid cognitive decline and make involuntary wiggling hand move­ments. Some have catatonia-like episodes in which their eyes are tightly shut and their jaw remains open. Eventually some become hypoventilated and must be placed on a ventilator. It is not a fun thing to have.

These aren't new diseases: the authors discuss a case from 1830 where a patient had psychiatric symptoms before vomiting up an abdominal teratoma, which cured her. We now know the cause is IgG1 antibodies that bind to the GluN1 subunit of the NMDA receptor in the brain, which is essential for neurotransmission.

At least 18 similar antibody diseases are known, including SOX1 antibody, which can cause para­cere­bellar degeneration or Lambert-Eaton myasthenic syndrome (LEMS), a form of partial paralysis. LEMS is often a sign of a serious cancer called SCLC or small cell lung cancer. Myasthenia gravis, a generalized weakness caused by antibodies against either the acetylcholine receptor or the esterase AChE, is typically a sign of a tumor in the thymus gland, called a thymoma. Antibodies to MOG, or myelin oligo­dendrocyte protein, cause seizures, encephalopathy, and a demyelinating syndrome similar to MS. In children it is now thought to be identical to ADEM (acute dissem­inated encephalo­myelitis), a dangerous autoimmune response to measles infection, whereas in teens and adults it more often causes optic neuritis, where the patient experiences loss of visual function and feels pain when trying to move their eyes.

Many of these disorders are early signs of cancer, most often SCLC, testicular cancer, or thymoma. But they can also be produced by medical treatment. A new type of anti-cancer drug known as an immune checkpoint inhibitor turned out not to be as benevolent as once thought. Viruses can also cause them: the authors say that 27% of patients with herpes simplex encephalitis also get anti-NMDA receptor encephalitis. And levamisole, a drug formerly given to cancer patients as an immune-modu­lating agent and now found in 80% of the illegal cocaine in the US, is now known to cause leuko­enceph­alo­pathy (a disorder of the white matter of the brain) and agranulo­cytosis (loss of granular leukocytes, a type of white blood cell).

Cancer patients may see CAR-T cells as a miracle cure, but 12–31% of those who receive them get severe neurotoxicity, which shows up as encephalo­pathy (a decreased level of consciousness, confusion, and seizures) due to CRS (cytokine release syndrome) or HLH (hemo­phago­cytic lympho­histio­cytosis, an often fatal disorder in which the red blood cells are destroyed).

We often also think the cerebellum is immune from neurodegeneration, but it's not so. In cerebellar ataxia, which can be caused by antibodies to gliadin, DNER, GAD, CASPR2, Yo, or carbonic anhydrase-related protein VII, patients have extensive loss of Purkinje cells, giving them an irreversible syndrome charac­terized by agitation, negativism, unintelligible speech, and unsteady gait. It often starts after cancer surgery.

Classical autoimmune and autoinflammatory disorders are not discussed much. The treatment is purely medical with almost no molecular biology. The authors include many case histories. Videos and an online copy of the book are available to owners of the hardcopy version. Sections 1 and 2 summarize the disorders. In Section 3 the disorders are presented again in individual chapters.

The treatment seems somewhat disorganized and repetitive. For example, on page 299, they repeat the same facts about MOG antibodies in ADEM several times. The diseases are presented in roughly decreasing order of the authors' confidence in their importance (and, in some cases, their existence). They emphasize over and over the importance of measuring antibodies in cerebro­spinal fluid (CSF) instead of blood plasma and complain a lot about the quality of commercial antibodies and results derived from them.

Finally, in Section 4, we find out why. The authors are making a case that many cases previously diagnosed as schizophrenia are actually anti-NMDAR encephalitis:

[A]ll the symptoms of schizophrenia can also occur in autoimmune encephalitis, particularly anti-NMDAR encephalitis. The difference between these diseases does not reside in the psychiatric symptoms, but in the timing of the presentation, evidence of prodromal psychiatric features, accompanying neurological symptoms, abnormal diagnostic tests, . . . and detection of NMDAR IgG antibodies in the CSF, which are absent in schizophrenia. [p.505]

They say they re-analyzed CSF samples from some patients treated for psychiatric problems and found evidence that they were actually suffering from anti-NMDAR encephalitis.

There are a few kinks to work out on this theory, namely that the sex ratios don't match up and the fact that schizophrenia starts at the wrong age for cancer, so there must be something else that triggers it. But if they're right, we could be in for a minor revolution in the treatment of psychiatric patients.

jun 26 2022

The Neuropathology of Schizophrenia

by Matthew Williams, ed.
Springer, 2021, 241 pages
reviewed by T. Nelson

his multi-author book discusses changes in brain anatomy that occur in schizophrenia (SZ). Each chapter focuses on a different brain region.

A big challenge in studying SZ arises from the fact that as soon as a patient is diagnosed, they're adminis­tered powerful antipsychotic drugs whose effects on brain physiology tend to overwhelm any physio­logical differences. That is, of course, their purpose, but it means that many findings described so far are artifacts of treatment. Even so, there are some consistent findings:

• Overall brain symmetry looks more normal in schizophrenic brains than in controls. That is, the centerline of the brain from front to back is straight in schizo­phre­nia brains but bent in normal brains. This is called brain torque.

• At the same time, brains of schizophrenia patients have “occipital bending”, which is a develop­mental condition in which the occipital lobe in the back is distorted. This is only seen in 35% of patients and its significance, if any, is unknown.

• There is a generalized loss of synapses and gray matter in many regions, particularly the temporal and frontal cortex, the amygdala, which is involved in emotional valence, and the nucleus accumbens, which is involved in addiction (though some studies on the latter two find no difference, and there are studies that find that gray matter volume is increased). Loss of gray matter automat­ically means ventric­ular enlargement, which is also found. Some regions, like the hippocampus, are hardly changed. There are many conflicting results, mostly because of differences in what is being measured, but also because of patient variability, drug treatment, and other factors.

• Genome-wide association studies showed 8300 SNPs (single nucleotide polymorphisms, i.e. mutations) for schizophrenia, while proteomic studies found 84 proteins and 74 miRNAs. Thus, findings implicating one specific protein or neuro­trans­mitter are automat­ically suspect due to the vast number of ways a protein can be affected by other proteins and by general loss of neurons and/or synapses.

• Dopamine synthesis is hugely increased in the substan­tia nigra, apparently due to elevated tyrosine hydrox­ylase, which is the rate-limiting enzyme in its synthesis. You would think this would protect them from Parkinson's disease, but a new finding (not mentioned in this book) shows they have an elevated risk. Antipsychotic drugs affect the striatum (which is where the SN projects to) and cause tarditive dyskinesia, a different disorder also caused by L-DOPA, which is metabolized to dopamine in the brain.

The overall style is descriptive: in many cases authors simply say that something was ‘increased’ or ‘decreased’ without bothering to mention by how much; in almost no cases are error terms mentioned. Researchers seem mainly concerned with determining whether it's more productive to study SZ at the molecular, cellular, or anatomical level. The earlier hypothesis that SZ is caused by too much dopamine is still clinging to life, but other neuro­trans­mitters like GABA and glutamate are catching up fast.

There are strong hints that SZ may be a neuro­develop­mental disorder. One author mentions, just in passing, that 22q11.2 deletion syndrome, which is a partial loss of chromosome 22 that happens during gamete formation or in early devel­op­ment, produces a high risk of SZ. (Typical of this book, he doesn't say the amount of risk, but other sources say it's about 30%.) The lost section of DNA codes for several proteins including COMT, an enzyme that degrad­es dopamine and other catecholamines. Mice with an orthologous deletion (16qA13) get SZ-like abnormal behaviors.

Oddly enough, this syndrome, which is also called DiGeorge syndrome, can also induce autoimmune rheumatoid arthritis, which epidemiological studies show seems to protect against schizophrenia. This is what we in science call a Big Clue, but the book is too focused on reporting hard-to-reproduce neuro­ana­tomical observations to mention it.

So, this book isn't the best way to learn about schizophrenia, but it's a pretty good way to learn neuroanatomy, as there are many color diagrams for each brain region. There is no index and no table of abbreviations.

An alternative text on neuroanatomy is John H. Martin's Neuroanatomy Text and Atlas, 5e, which has photos and 3D colored diagrams. If you just want to find things, a good one is the spiral-bound 1989 Structure of the Human Brain: A Photographic Atlas by DeArmond, Fusco, and Dewey. (It doesn't matter that this one is in grayscale; everything in the brain is some shade of beige.) If you're not too squeamish, get Nolte's The Human Brain in Photo­graphs and Diagrams. All of 'em even have an index. Woo-hoo!

dec 29 2021