Viewpoint Expert opinions on trends and controversies in autism research.
Illustration by Julia Yellow
Opinion / Viewpoint

Same DNA deletion paves paths to autism, schizophrenia

by  /  18 October 2016
The Expert:

Carrie Bearden

Professor, University of California, Los Angeles

From their earliest descriptions, schizophrenia and autism have been inextricably linked.

In the early 1900s, Swiss psychiatrist Eugen Bleuler coined the term ‘schizophrenia’ to mean the splitting of psychic functions. People with the condition tend to have false beliefs, called delusions, and sometimes experience hallucinations, seeing or hearing things that aren’t there.

Within his description of schizophrenia, Bleuler used the term ‘autism’ to mean a withdrawal from reality. He considered autism to be a feature of schizophrenia, rather than a separate condition.

Similarly, after Leo Kanner first described autism in 1943, some clinicians called the condition ‘childhood-onset schizophrenia.’ But in the 1970s, evidence began to mount that autism and schizophrenia are distinct conditions that differ in when their features appear and what those features are. This view has gone largely unchallenged, until recently.

Studies from the past several years show that the same genetic glitch — deletion of a stretch of DNA on chromosome 22 — raises the risk for both conditions. Up to 30 percent of individuals missing this region, called 22q11.2, develop a psychotic disorder — most commonly schizophrenia. And up to 50 percent are diagnosed with autism.

It’s possible that despite the same genetic cause, these seemingly disparate conditions manifest differently because their features show up at distinct times: autism during early childhood and schizophrenia during adolescence.

In fact, prominent researchers, such as Maria Karayiorgou, professor of psychology at Columbia University, have proposed that the social dysfunction in children with 22q11.2 deletions is an early manifestation of schizophrenia that is misinterpreted as autism. They suggest that the dominant cause of autism is a duplication, rather than a deletion, in the 22q11.2 region.

But we believe that the deletion is also an important cause of autism. The biology as well as characteristics of autism stemming from the deletion are quite distinct from those of schizophrenia arising from the same genetic glitch — a phenomenon known as pleiotropy.

Spectrum of signs:

First, if all deletions of 22q11.2 were truly associated with schizophrenia (and not autism), as some contend, we would expect that a diagnosis of autism in childhood in these cases would be associated with an increased risk of schizophrenia later in life.

But there is no such association. In 2013, psychiatrist Jacob Vorstman and his colleagues at the Utrecht University Medical Center in the Netherlands reported that adults with 22q11.2 deletions who had shown signs of autism during childhood did not show an increased risk for schizophrenia in adulthood1. The researchers have since followed children with 22q11.2 deletions over five years. Again, they found that those diagnosed with autism are no more likely to develop schizophrenia than those without autism.

Findings from my laboratory provide additional evidence that two distinct conditions can arise from a 22q11.2 deletion2. Using a network analysis, we identified a group of genes that show atypical expression patterns in individuals with a 22q11.2 deletion and psychosis. These genes play a role in regulating other genes.

We identified a separate group of genes that are differentially expressed in people with a 22q11.2 deletion and autism. These genes are involved in the body’s immune response. And, interestingly, they show significant overlap with genes previously linked to ‘idiopathic autism,’ in which a person’s autism involves no clear genetic cause.

What’s in a name:

Another important question is whether autism in children with the 22q11.2 deletion is the same as idiopathic autism. It is most likely that children with the 22q11.2 deletion share some features with children who have idiopathic autism, but that they also have unique behavioral characteristics. For instance, children with a 22q11.2 deletion who have an autism diagnosis tend to repeat the same question, which is not typical of autism.

This unique profile is consistent with findings from previous studies suggesting that ‘syndromic’ forms of autism associated with a particular genetic variant have characteristic features.

In 2014, Hilgo Bruining and his colleagues at the Utrecht University Medical Center showed they could predict which of six autism-related syndromes a person has by his or her profile on a diagnostic test, the Autism Diagnostic Interview-Revised (ADI-R). One of the six was 22q11.2 deletion syndrome. This finding points to ‘signature’ behavioral profiles that track with underlying genetic risk factors.

Ultimately, what we call a particular condition matters less than understanding how its underlying cause triggers certain behaviors. There is increasing interest in conceptualizing what we now know as ‘autism’ and ‘schizophrenia’ from a dimensional rather than categorical perspective. That is, researchers are viewing features of conditions along a continuum that is rooted in biology rather than simply grouping features together under a name.

Although we believe that autism associated with 22q11.2 deletion is separate from schizophrenia linked to the genetic anomaly, seeking a better understanding of the similarities and differences between the two conditions is important. That knowledge can bring us closer to the goal of developing better treatments for features that cross diagnostic boundaries.

  1. Vorstman J.A. et al. Schizophr. Res. 143, 55-59 (2013) PubMed
  2. Jalbrzikowski M. et al. PLOS One 10, e0132542 (2015) PubMed

12 responses to “Same DNA deletion paves paths to autism, schizophrenia”

  1. MaryTormey says:

    DNA is not lost it just changes to suit the environment. The mutagen in sodium fluoride can affect the neurological timing. People with different nerves can see and hear things other people can’t. We need to put an end to the sexist idea that if a woman sees or hears something she is consider “crazy” if a man can’t see or hear it. The fact is that just being a different sex, can alter the way people see and hear the world. Perhaps the nernst equation could give you a better idea of what is going on. Schizophrenia is a result of people repeatedly being forced to make decisions faster than they can. It is often a side effect of the medical industries rape culture. I notice light moves faster for schizophrenics, they see my eyes as blue, others often see them as green or brown. They just have blue shift and need a stable non-abusive reality.

  2. Julian Baker says:

    As far as I know 22q11.2 deletions are not significantly enriched in genome-wide searches for autism susceptibility CNVs unless they are lumped up together with 22q11.2 duplications, which is quite strange. If that is true, then it does not support the argument that 22q11.2 deletions are a cause of autism. Enrichment in genome-wide CNV searches was how all the prominent autism associated CNVs were discovered.

    • Alexis Ren says:

      This is probably right. Also 22q11.2 deletions are conistently found highly enriched in genome-wide scans for schizophrenia CNVs. So the statements “22q11.2 deletion paves path to schizophrenia” and “22q11.2 deletion paves path to autism” are not equivalent as it was presented in this piece.

  3. Carrie Bearden says:

    Thank you for these important points. Sanders et al (Neuron 2011) find over-representation of 22q11.2 deletions in the Simons Simplex Cohort, but you are correct that this finding only reaches significance threshold when considered together with duplications (which is indeed interesting). Samples of 22q11.2 patients in which elevated rates of autism have found tend to be clinically ascertained samples, which I agree is a limitation, although these samples are now quite large, multi-site (Schneider et al AJP 2014). We wish to make 2 primary points: 1) despite the unequivocal association of 22q11.2 deletion with schizophrenia, an autism diagnosis in the context of this deletion is not an indicator of premorbid schizophrenia, as it is not predictive ; and 2) autism in the context of 22q11.2 deletion (or any other syndromic form of autism) may be better understood as a cluster of dimensional traits. I would agree with Alexis that the paths (to schizophrenia vs autism) are not equivalent in terms of specificity of association. thank you for raising this.

    • Alexis Ren says:

      Dr. Bearden, thanks for your response. I maybe missing something but
      first, I am not sure why Sanders et al (Neuron 2011) combined 22q11 deletions and duplications in their analysis. As far as I know deletions and duplications have often opposite phenotypes so combining them does not make much biological sense
      second, I am not sure that the diagnosis of autism should be a moving target. By the diagnostic criteria used to establish the Simons Simplex Cohort -which is sort of a gold standard- 22q11 deletions are not a cause of autism (at most, a very minor one).

      • Carrie Bearden says:

        hi Alexis- i won’t quibble with “a minor cause’, but It is also important to think about the inclusion criteria for the SSC- nonverbal mental age below 18 months, neurological deficits, birth trauma, perinatal complications- all somewhat common for 22q11.2 deletion patients. There are of course good reasons for these exclusion criteria when recruiting on the basis of ASD diagnosis, but meta-analysis estimates 22q11.2 deletion syndrome is associated with 8·3 fold increase in ASD risk (95% CI 1·04–65·73) (

        • Alexis Ren says:

          there is a huge discrepancy between the 8-fold (CI 1·04–65·73) increase in ASD risk in the Lancet Psychiatry paper and the lack of statistically significant enrichment in the Simons Simplex Collection. If the Lancet paper is correct the implication is that there are serious problems with the diagnostic criteria used to generate SSC which is very alarming for SSC and reflects bad on the reliability of ASD genetics.

          • Carrie Bearden says:

            I would not make that interpretation. Rather, I think ascertainment and inclusion criteria have a major impact on the study population. UCLA was one of the SSC recruitment sites, and our 22q11 patients were phenotyped right alongside the SSC participants by the same interviewers – a substantial proportion received ASD diagnoses, but would have been ineligible for SSC for various reasons.

          • Alexis Ren says:

            Why not? that is the simplest interpretation. Although given on one hand the inconsistency and unreliability of autism diagnosis you describe (…”various reasons”) and on the other hand the spectacular success that SSC had with identifying autism risk loci, I would not consider a CNV an autism-risk CNV unless it is found highly significantly enriched in SSC. Just keeping things simple and reproducible across studies to reduce “noise”. By this criteria, it seems to me that 22q11 deletions are unlikely to be an autism-risk locus.

          • Carrie Bearden says:

            Dear Alexis, I think you are misunderstanding my point. I think the ASD diagnoses in the SSC were done extremely reliably and indeed should be considered a gold standard. The issue is who is included in the first place. the SSC excluded those with perinatal complications, low nonverbal mental age, etc – for good reason, but this would of course affect who is included in the sample.

  4. kenfine says:

    I have Asperger’s, and have reviewed the research into psychiatric disorders for 30 years. Common factors include immune system upregulation and receptor abnormalities. In particular, the transcription factor, nuclear factor kappa beta (NF-kB), is seen to be activated.

    NF-kB has multiple functions, but a principal function is the control of intracellular pathogens which establish latency within cells. NF-kB can become chronically and aberrantly activated, resulting in disease. Many genetic and environmental factors contribute to this aberrant activation. Genetic factors include immune system, receptor and enzyme polymorphisms. Environmental factors include infections, lack of immune challenge in infancy, exposure to environmental chemicals and metals, lack of vitamin D, pro-inflammatory ‘western’ diet, gut microbiota, chronic psychological stress etc.

    The review concludes that activation of NF-kB occurs via two principal mechanisms. These are:

    1) Chronic, excessive production of intracellular reactive oxygen species (ROS), usually by metalloenzymes.

    2) Chronic decreased or increased activation of cell membrane receptors.

    1) was already known. 2) appears to be a novel finding of the review.

    To control cellular infection, the NF-kB driven immune response causes ROS to be produced inside cells. These ROS are destructive to microbes. It also downregulates cell membrane receptors which pathogens use to enter cells, in order to prevent infection. This downregulation can take three forms – reduced expression, desensitization, and receptor blockade. This causes cells to become hypofunctional. To compensate for this, receptor activation on other cells may become hyperfunctional.

    If chronic excessive ROS production, or receptor hypoactivation or hyperactivation occurs in the absence of infection, the NF-kB driven immune response may become aberrantly activated. The review concludes that this gives rise to the bimodal character of psychiatric disorders – schizophrenia positive and negative syndromes, mania and depression in bipolar disorder, depending on whether receptor downregulation or compensation (upregulation) is predominant.

    In the case of autism, the evidence strongly suggests that the same bimodal mechanism is at work, with ADHD being the positive syndrome. The usual age of onset is similar in both, gender disparity is similar, many children exhibit symptoms of both conditions, and associations with prenatal exposure to the same environmental chemicals, such as pthalates and organophosphates have been found in both.

    In schizophrenia, the disease often begins soon after late developmental changes in the brain in adolescence and early adulthood, involving increased expression of dopamine receptors and resulting increased dopaminergic activity. In genetically predisposed individuals, this may result in chronic activation of NF-kB, by mimicking the compensation mechanism.

    In autism, the GABA(A) receptor is profoundly involved in early development of the brain, and the same increase in gabaergic transmission may give rise to autism in genetically predisposed individuals.

    I’ve written a blog with more than 900 links to PubMed papers, which goes into much greater detail:

Leave a Reply

Your email address will not be published. Required fields are marked *