Our new paper due out in Dec 5 CELL

This is the paper showing probiotic therapy works in the mouse model of an autism risk factor.

We have just published our first paper on extending the rodent model of maternal immune activation to monkeys. As repeat readers of this blog, and the accompanying book, well know, an important risk factor for both autism and schizophrenia is infection during pregnancy. We modeled this in mice by simply activating the mother’s immune system at mid-pregnancy. The resulting offspring display behaviors consistent with both of these disorders as well as neuropathology seen in these disorders. Since monkey brains and behavior more closely resemble humans than do mice, it was important to test whether the same findings would hold true when activating the immune system during pregnancy in monkeys.

Spearheading this work at UC Davis and the California Regional Primate Center, Melissa Bauman and David Amaral found that, indeed, the cardinal symptoms of autism (stereotyped/repetitive behavior, deficits in verbal communication, and highly abnormal social behaviors) are found in the offspring of mothers immune activated during either the first or second trimester, although more pronounced effects were seen in the former offspring. In further work yet to be published, we also found that there are eye tracking abnormalities (lack of eye contact with others) in these offspring, as is seen in human autism. These findings not only establish a new model in non-human primates, but they also validate prior work with rodents: the latter findings are not restricted just to mice and rats.

Summary of autism treatments

Our collaborator at Arizona State University, James Adams, has put together a useful summary and analysis (with references) of the incredibly diverse array of autism treatments out there – including many pharmaceuticals, diets, dietary supplements, chelation, hyperbaric oxygen, etc. This is intended for the lay public.

Another fascinating feature of this PDF is a table showing the results of ratings of all of these treatments obtained from surveys of >27,000 parents of autistic kids. Such results must be taken with a grain of salt because these are not double blind clinical trials but rather the opinions of parents on the benefits and drawbacks of treatments they have tried. Thus, there is the risk of a placebo effect. Nonetheless, there are so many negative ratings on this list that the placebo effect may not be such a big problem – altho it could still be relevant to particular treatments. I reprint this summary table below. Altho it may be too small read all the entries, below the table I’ve picked out some of the points that are of particular interest to me.

I’m struck by the nearly total failure of the plethora of pharmaceuticals; we knew this before, but seeing so many in one table is impressive. One bright spot is the apparent success of two anti-fungals, Diflucan and Nystatin. These were used selectively, where indicated by evidence of infection. In my opinion, this points to the gut microbiota. In this same context, it is surprising to find that antibiotics failed, as I know of two double blind studies that showed positive effects of antibiotics on ASD symptoms, altho the symptoms returned after the antibiotic was discontinued. Maybe the latter finding is the reason for the negative rating. A really impressive finding in this survey is the marked success of a variety of special diets. 8 of these were rated very highly. Similarly, many dietary supplements were rated highly, the best being fatty acids. On the other hand, I was quite surprised to see that both chelation and hyperbaric oxygen treatments were rated highly. The section on the latter treatment summarizes situations in which it can be quite dangerous, and the lack of strong support from double blind studies. The same caveats  apply to chelation.

Stem cell therapy for autism – first clinical trial

A paper just out from the Shenzen Beike Cell Engineering Research Institute in Guangdong, China presents the first results of a proof-of-concept, safety trial of stem cell (SC) injections in 37 autistic children. The justification for this intervention is primarily that animal studies of various inflammatory conditions can be ameliorated by SC therapies, and there is good evidence of immune dysregulation in autism (as well as Rett, OCD, PANDAS and Tourette). It is not suggested that the SCs will populate the brain and participate in the circuitry.

Two types of cells were used: human cord blood mononuclear cells and umbilical cord-derived mesenchymal SCs. The first group of patients initially received the cord blood cells by intravenous infusion and then, at 5-7 day intervals, they received 3 more injections by the intrathecal (spinal) route. They also received behavioral therapy. The second group of patients two cord blood intravenous and intrathecal injections each, followed by two umbilical SC intrathecal injections. This group is termed the combination group. They also received behavioral therapy. The third group of patients received only behavioral therapy.

The drawbacks in this study are that it was non-randomized, open-label (everyone knew who received which treatment), conducted at a single center, primarily by people who were employees at the company that produces these SCs (Beike Cell Engineering Research Inst). Thus, we can’t take the autism-related results too seriously, but we can probably rely on the results related to safety – they report that none of the 23 subjects that received multiple injections of these (foreign) cells displayed signs of toxicity or changes in blood biochemistry over the 6 months of the study. This bodes well for future studies of this kind.

Regarding the behavioral data, they did not use the gold standard ADOS method of scoring autism symptoms. Rather, they used the descriptive CARS, ABC Chinese Version, and the CGI (global) scales. In all of these 3 measures, the combination treatment group showed improvement over the control group. The single, cord blood treatment group was somewhat intermediate in efficacy. The large grain of salt is, as mentioned above, the folks doing the evaluations were aware of the treatment used on each patient. Plus, the number of subjects in each group was small (but not minuscule!). No doubt, these results will push this group and other SC labs and companies to move forward with further testing, hopefully double-blind!

Stress sends immune cells into the brain

As discussed in my book, stress promotes inflammation and immune dysfunction, as well as anxiety. A new and very complete paper from John Sheridan and colleagues at Ohio State University provides evidence that stress also induces an influx of immune cells into the brain, and that this causes anxiety symptoms. They stressed mice by putting an aggressive intruder into their cage for 2 hours each night for 1, 3 or 6 nights. This intruder subdues the residents and makes them subordinate, but does not injure them. Increasing the days of this stress causes and increase in anxiety in the resident mice, as measured in separate assays (not in the presence of an intruder)  by their reluctance to enter the center of an open field and their reluctance to go from a dark box into the light. These are some of the standard ways of measuring anxiety in mice. Increasing the days of stress also causes an increase in immune cells called monocytes in the blood. Using a genetic labeling technique, the investigators went on to show that some of the monocytes then entered the brain, where they became macrophages. Moreover, the parts of the brain where these invading cells are found are the areas known to control anxiety behavior.

This entry into the brain is caused by chemokine signals eminating from the brain itself, and if these signals are blocked, the immune cells do not enter the brain and, importantly, the mice do not display anxiety. That is, if mice who cannot do this chemokine signaling are placed in this highly stressful situation, they do not become anxious and their brains do not display the invading immune cells. The suggestion is then that the stressed brain recruits immune cells to do something that leads to anxiety. This ‘something’ could certainly involve cytokines, as stress increases cytokines in the brain, and cytokines can induce anxiety. But why recruiting macrophages is necessary when the brain’s own cells can produce those cytokines is a mystery. Perhaps the recruited immune cells are also carrying out another function that is helpful (or harmful?) in dealing with social stress. What this might be remains mysterious. It is also not clear how long these invaders remain the brain after the stress is over. Creepy?

Important advance for major depressive disorder

Major depressive disorder (MDD) is quite common and can be very disabling if not treated properly. However, fewer than 40% of patients achieve remission with their initial treatment. In the case of ineffective therapy, those patients often then experience months of trials using various different selective serotonin reuptake inhibitors (SSRIs), or they move on to other psychotherapists, or they drop out of treatment attempts altogether, which can be disastrous. Therefore, it would be of great importance if an objective, quantitative test were available to guide the patient towards the optimal treatment direction at the outset. Helen Mayberg and colleagues at Emory University in Atlanta have recently published an important step forward in this direction. They measured brain glucose metabolism by PET (positron emission tomography) imaging in a series of MDD patients before any form of treatment was started. They then randomly assigned the patients to either of two  groups: treatment with a standard anti-depression medication (SSRI) or treatment with evidence-based psychotherapy. Of the 38 patients with clear outcomes, 12 responded well to cognitive behavioral therapy, 11 to SSRI treatment, 9 did not respond to psychotherapy, and 6 did not respond to the SSRI. They then asked whether glucose metabolism in various parts of the brain clearly corresponded to these 4 patient groups. In fact, measurement of the metabolism in the right anterior insula was able to discriminate among the groups: Lower metabolism was associated with remission in response to psychotherapy and no remission in response to SSRI treatment, while higher metabolism was associated with remission in response to SSRI treatment and no response to psychotherapy. If this finding can be reproduced in a large, prospective trial, it would mean, first, that new MDD patients could be assigned to the treatment path that is appropriate for them and with the best chance  to succeed., and second, it would argue that these two groups of  patients have a distinct neurophysiological basis for their MDD.

Incidentally, Mayberg has also spearheaded the use of deep brain stimulation for MDD that is resistant to any type of standard treatment. While this approach is deeply invasive, it appears to be quite promising for those patients who have run out of other treatment options.

Pictures from Africa

Some of the classes I taught in Uganda in June –

Very bright students in Kampala prep school.Chalk talk

Class of ~90 students

Singing Uganda national anthem after my lecture

Maternal antibodies and ASD risk for the offspring

ASDresearchinitiative asked me to comment on the new papers from UC Davis, so now that I’m back from a grueling trip to E Africa (for lectures and touring), I should get back to work…

In one paper, Melissa Bauman, David Amaral and colleagues report a follow up to their prior work, further showing that when pregnant monkeys are injected with antibodies taken from human mothers of autistic children, the monkey offspring display a couple of abnormal behaviors, including unusual social interactions. Such behaviors are not observed in the offspring of pregnant monkeys  given antibodies taken from human mothers of neurotypical kids. While the behavioral differences between the two sets of monkey offspring are convincing, they do not include some of the cardinal signs of human autism such as stereotyped/repetitive behavior, which was reported in their prior paper on this model. However, the type of antibodies injected in this new experiment is different from the type used in the prior work. The antibodies used in the new work are known to be a subset of the various auto-immune antibodies that recognize antigens in fetal monkey brains.  Overall, these results support the hypothesis that some mothers of autistic children generate antibodies against the fetal brain and that these antibodies cross the placenta and alter fetal brain development. In fact, the investigators showed that the brains of the animals with altered behavior do display some differences from the controls when analyzed by MRI.

In an independent set of experiments that will be published very soon, Melissa, David and I found that activating the pregnant monkey immune system, using the same technique that we used in mice, results in offspring with a series of autism-like behaviors, including repetitive behaviors, a deficit in verbalizations and a highly abnormal type of social interaction. A subsequent paper will also show an autism-like abnormality in eye tracking – not looking at faces as much as controls. Could these two models be related – maternal antibodies and maternal immune activation (MIA)? It is indeed possible that MIA stimulates the production of antibodies. However, I favor a mechanism in which MIA induces the cytokine IL-6, which we find alters placental endocrine function and also directly activates subsets of fetal brain cells.

A second, and related new paper from UC Davis by Judy van de Water and colleagues follows up their prior work on characterizing the antibodies found in the blood of mothers of autistic children. The new paper identifies the proteins to which these antibodies bind in extracts of monkey fetal brain. Interestingly, many of these protein antigens are known to have functions in brain development. However, most of these proteins are thought to reside in the cytoplasm, so it remains to be shown how the maternal antibodies would access them in living cells in the fetal brain. Further animal studies will presumably get at this question. It will also be interesting to find out when these antibodies are generated during pregnancy, and whether they are present in subsequent pregnancies in the same mother that involve neurotypical or autistic offspring. The latter point is important because van de Water et al. propose that the presence of combinations of several of these antibodies could be used as a diagnostic biomarker for an autistic outcome of a pregnancy. In fact, they have formed a company to develop such a diagnostic.

Positive sides to autism

In LA Times article, Temple Grandin sees the advantages that she and others with autism bring to society.

Important message from the Director of the National Institute of Mental Health

Thanks to ADResearchInitiative blog for the tip on this message.

On another note, our family is off to E Africa for 3 weeks to give lectures (me, viruses and mental health; wife Carolyn, special education) in Uganda, Kenya and Tanzania. We are also looking forward to touring national parks, including tracking mountain gorillas in Biwindi. So I’ll be out of touch till July, but may come back with some interesting stories.