Schizophrenia, Hubris and Science

By Neuroskeptic | February 1, 2016 3:14 pm

 

sekar

A press-release from the Harvard-MIT Broad Institute reaches astonishing heights of hyperbole in announcing a new schizophrenia study (Sekar et al. 2016). Here’s the release:

Genetic study provides first-ever insight into biological origin of schizophrenia

Landmark analysis reveals excessive “pruning” of connections between neurons in brain predisposes to schizophrenia

A landmark study, based on genetic analysis of nearly 65,000 people, has revealed that a person’s risk of schizophrenia is increased if they inherit specific variants in a gene related to “synaptic pruning” – the elimination of connections between neurons. The findings represent the first time that the origin of this devastating psychiatric disease has been causally linked to specific gene variants and a biological process…

And so it goes on.

When I first heard about this on Twitter, I couldn’t believe it. Surely Broad hadn’t actually claimed the “first-ever insight” into the biological origin of schizophrenia? Yet, absurdly, they did. They seem to be saying that no-one has ever found out anything biological or genetic about schizophrenia until now.

Further down the press release, Broad Institute head Eric Lander chips in, saying that “For the first time, the origin of schizophrenia is no longer a complete black box”. For Lander, it seems, everyone was completely in the dark until Broad stepped, Prometheus-like, onto the scene!

lander_prometheus

Then again, Lander is known for his Broadocentric view of science. He caused outrage recently in the field of molecular biology with a review paper called The Heroes of CRISPR. Critics say that Lander’s paper, a history of the discovery of gene-editing technique CRISPR/Cas9, is an attempt to assign most of the credit to researchers at the Broad Institute. (A patent dispute over who invented the technique first is ongoing.)

Is this press release an attempt to frame a similar “Heroes of Schizophrenia” narrative? Perhaps so. Certainly, the language in the release goes well beyond the accepted limits of hype. Hype would be to say ‘this research is very important’. Every press release does this to some extent. But Broad are saying ‘no-one else has ever done research like this’, which is beyond hype: it’s hubris.

In my view, Sekar et al.’s paper is a very interesting and impressive contribution to science, but it’s not unique. It’s one more contribution among others.

Sekar et al. show that genetic variants in the gene encoding complement factor C4, a protein involved in immune responses, are associated with the risk of schizophrenia (somewhat). They further show that these risk variants increase expression of a form of C4 called C4A, and that C4A is expressed in the brain where, at least in mice, it helps control synapse “pruning” by marking synapses for destruction by cells called microglia. Therefore, Sekar et al. trace a pathway from genetics, via protein expression, to synapse dysfunction… and thence, presumably, to the symptoms of schizophrenia.

This is cool stuff, but there have been plenty of cool studies before in this field (not all of which have turned out to be correct.) C4A is not the first schizophrenia risk gene to have been discovered, nor is it the first whose neurobiological function has been established using a gene-protein-brain approach (e.g. one of many previous examples.) It’s not even the first schizophrenia risk gene whose role in synaptic pruning has been elucidated!

As for the “excessive synaptic pruning” model, it’s very far from new: it was proposed in 1983 and has since been discussed in 3580 papers. So it’s simply untrue to imply that Sekar et al. are the first to shine light into a biological “black box”. Sekar et al.’s results are important, but not Promethean. As for Broad, I hope they remember that while giants may have broad shoulders, even Broad scientists stand on the shoulders of giants.

ResearchBlogging.orgSekar A, Bialas AR, de Rivera H, Davis A, Hammond TR, Kamitaki N, Tooley K, Presumey J, Baum M, Van Doren V, Genovese G, Rose SA, Handsaker RE, Schizophrenia Working Group of the Psychiatric Genomics Consortium, Daly MJ, Carroll MC, Stevens B, & McCarroll SA (2016). Schizophrenia risk from complex variation of complement component 4. Nature PMID: 26814963

ADVERTISEMENT
  • http://jayarava.blogspot.com Jayarava

    Thanks for the reality check.

  • Bernard Carroll

    Your title reminds me of the joke about the difference between hubris and chutzpah. Hubris, from the Greek, is the term that applies when mere mortals presume to give advice to the Gods. Chutzpah, from the Yiddish, is the term for when they follow up by sending the Gods a bill for the advice.

  • http://www.mazepath.com/uncleal/qz4.htm Uncle Al

    Perhaps they are off-put because the Voices decline to speak to them. If the voices in your head don’t stop for breath, that’s the Toni.

  • http://twitter.com/donotgogently Chris Patil

    Neuroskeptic’s “shoulders of giants” metaphor betrays an astonishing ignorance of the state of despair around translational psychiatry, and the blogger completely misses the boat on the importance of these findings. If the purpose of this blog post was to argue for a more reserved tone in scientific press releases, and support that view by challenging the significance of a study described in such a release, the blogger made an absolutely awful decision in their choice of which paper to make an example of.

    Drugs used to treat schizophrenia today have the same molecular mechanism as compounds discovered serendipitously 60 years ago. In the intervening time we’ve seen no innovative drugs, and no new biomarkers — a sad lack of progress at a time when we’ve made enormous progress in treatment of cancer and cardiovascular disease. At the same time, pharma companies have exited neuropsychiatry in droves – despite enormous unmet need – citing lack of scientific foundations.

    Because, see, drug developers don’t just just count Google Scholar hit results; they need to have a real understanding of mechanism with a solid experimental basis. Neuroskeptic’s abstract-counting seems to imply that every paper in PubMed reflects a major advance, but this confuses quantity with quality. Others feel that the view from the “shoulders of giants” is not especially clear. While it’s true that thousands of papers about schizophrenia have been published, many of them followed false leads, described failed animal “models”, or told genetic ‘just-so’ stories.

    The Sekar et al. discovery in Nature connected the human genome’s largest influence on risk of schizophrenia with a molecular mechanism that would explain the age of onset and major pathological observations. That’s a major advance. Many others (including NIMH Director Cuthbert, and well-known GWAS skeptic David Goldstein) also called it a major advance (e.g., see Goldstein’s News and Views in Nature.)

    I agree with them. Given that this paper not only points the way toward a novel risk gene but provides convincing evidence of molecular mechanism, it should be considered an unprecedented advance. (Unless the blogger feels like well-studied, potentially druggable pathways have been implicated by recent work on this disease; if so, they were not mentioned in the original post).

    The press release doesn’t say that Sekar et al identified the “first schizophrenia risk gene”, but rather that “the first time that the origin of this devastating psychiatric disease has been causally linked to specific gene variants *AND* a biological process” [emphasis mine]. This isn’t just another lab reporting a locus identified in GWAS. Instead, this is a situation in which a team of (very) clever geneticists analyzed a (very hard to analyze) set of genetic variants and then showed that the protein encoded by the risk allele has a molecular activity pertinent to an established (but heretofore unexplained) disease phenotype. This paper is an interdisciplinary tour de force, and it is philistine to challenge the magnitude of the achievement to defend a quibble about the tone of a press release. In any case, the basic claim of that release is true: Literally no one else has published research like this.

    Beyond that, I’m puzzled by the implication that major advances in an extensively studied field are less important because they confirm (or deny) widely discussed models. In any event, hypotheses are different from experimental confirmation. The idea that excessive synaptic pruning is involved in schizophrenia has indeed been discussed in many papers (although most of those 3000+ mentions are one-liners and asides that cite a handful of originating papers).

    Pardon my caps, but THAT IS EXACTLY WHY THIS IS IMPORTANT. The idea that SCZ could involve excessive synaptic pruning is so provocative and exciting that thousands of authors have decided to mention it in their papers, but before Sekar et al, we had no idea about the underlying molecular mechanism. Now we have not just a candidate mechanism, but one that unifies a great many disparate findings and points the way toward both further genetic research and rational design of therapeutics.

    The Internet makes it easy for us to nip at the ankles of giants, especially from behind the cover of a pseudonym, and lately it’s been fashionable to pile on Eric Lander for doing basically what any chair is supposed to do: popularize the work of scientists affiliated with his institution. Also, granted, press releases could be toned down. But this paper is the wrong place to plant that flag; this is the wrong hill to die on. This paper is gigantic, and it’s easy for me to see how it could go down in history as a turning point in what was once a hopeless battle.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Thanks for the detailed comments. I will reply to different points in different replies.

      “The human genome’s largest influence on risk of schizophrenia”

      C4 may be the largest common genetic influence on schizophrenia, but it is far from having the largest effect size when rare variants are considered too. 22q11.2 deletion syndrome, for instance, is a rare disorder associated with an absolute risk of schizophrenia of about 25%, compared to a population prevalence of 1%.

    • http://adamgradzki.com Adam Gradzki

      Thank you for your detailed reply

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      “Given that this paper not only points the way toward a novel risk gene but provides convincing evidence of molecular mechanism, it should be considered an unprecedented advance. (Unless the blogger feels like well-studied, potentially druggable pathways have been implicated by recent work on this disease; if so, they were not mentioned in the original post).”

      I agree that Sekar et al. represents a major advance because it implicates a druggable pathway. However, I disagree that this is unprecedented. For instance, glutamate signalling has lately emerged as a major pathway of interest, and in the post I cited as an example a paper by Weickert et al. (2011) who linked schizophrenia risk variants in NMDA receptor genes to reduced expression of sub-units of the receptor in schizophrenia brains post mortem, and to cognition in vivo.

      See also Kevin Mitchell’s comment, below.

    • Costa Vakalopoulos

      ‘Unifies a great many disparate findings” is somewhat exaggerated. it potentially explains synaptic pruning which is not even specific to schizophrenia

  • http://twitter.com/donotgogently Chris Patil

    Also…

    >”It’s not even the first schizophrenia risk gene whose role in synaptic pruning has been elucidated!”

  • http://twitter.com/donotgogently Chris Patil

    > It’s not even the first schizophrenia risk gene whose role in synaptic pruning has been elucidated!

    This misleading statement bends the word “elucidated” past the breaking point. Neuroskeptic might feel that hyperbole in the service of fighting hyperbole is no vice, but I find this line disingenuous (and not a little ironic).

    The cited paper describes a microRNA that is differentially regulated in schizophrenia, not a risk gene, and the authors themselves admit that the connections they propose are speculative (I excerpted some text below, and the entire paper is free to read, although that is clearly no guarantee that everyone who links to it will actually do so.).

    “Given the suggested role of miR-132 in potentiation of NMDAR signaling, abnormalities in miR-132 expression during this critical period may disrupt the molecular maturation of GABA interneurons. A core molecular characteristic of schizophrenia is reduced expression of the GABA-synthesizing enzyme GAD67, encoded by the GAD1 gene, and hypermethylation of schizophrenia risk genes, including GAD1 and REELIN, has been proposed as a cause of reduced GAD67 expression in schizophrenia (41). Notably, one of the miR-132 targets most strongly repressed during the period of peak miR-132 expression in adolescent mice is Dnmt3a, a DNA methyltransferase and confirmed miR-132 target. Reduced miR-132 levels in adolescents at risk for schizophrenia would be predicted to result in derepression of DNMT3A and subsequent hypermethylation of the genes targeted by DNMT3A. Our data do not prove a direct interaction between miR-132 and DNMT3A. However, we demonstrate that miR-132 regulates—directly or indirectly—RNA and protein levels of this important DNA methyltransferase. Furthermore, inhibition of NMDA signaling results in abnormal up-regulation of Dnmt3a expression in adulthood. The ultimate effect is still speculative, as the functions of the DNMT family members often overlap and may compensate for dysregulation of one family member.”

  • Sarah St Jean

    Bruce Cuthbert (NIMH Director) to NPR: “a crucial turning point in the fight against mental illness.”

    David Goldstein (leading brain geneticist) to NY Times: “This paper gives us a foothold, something we can work on, and that’s what we’ve been looking for now, for a long, long time.”

    There are many other similar comments from independent experts, in many news outlets. The press release’s view that this is not your typical paper seems widely shared by independent scientists.

  • Elizabeth Nicholson

    Like other posters, I find Neuroskeptic’s blog entry misleading.

    I found that press release online (http://www.eurekalert.org/pub_releases/2016-01/biom-gsp012316.php)

    Neuroskeptic appears to have stopped quoting the press release immediately before it cites (in the third sentence): “decades-old observations: synaptic pruning is particularly active during adolescence, which is the typical period of onset for schizophrenia symptoms, and brains of schizophrenic patients tend to show fewer connections between neurons.”

    Then later in his piece Neuroskeptic suggests that he had to find out about those earlier observations from his own Google Scholar search. Those observations are the third sentence of the release, and in almost all news coverage of the Nature paper. It is misleading of the Neuroskeptic blog to hide this.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Thanks for the comment. I’m unsure why you had to “find” the press release since I linked to it in the first line of my post. As for myself, I linked to Google Scholar by way of a citation for my point that there has been lots of prior work on this topic.

      It’s true that the press release mentions earlier work on the synaptic pruning hypothesis, but my point is that it downplays all the earlier work as essentially unenlightening. Maybe this is a matter of opinion.

      • http://jayarava.blogspot.com Jayarava

        Exactly. The press release reads as though the discovery of the link between synaptic over-pruning and schizophrenia has no precedent. But it does. As the scholar search shows. The PR is hyperbolic. I see this in my field also – highly placed researchers, with links to media companies (National Geographic in my example), that hype genuine discoveries into earth shattering breakthroughs because, I presume, when it comes to applying for more grants they are widely seen as the people that make earth shattering breakthroughs by a panel who may have read the press release but are unlikely to have read the paper.

        In these cases the press release deserves to be demolished with vigour because the researchers are just gaming the system. I don’t think we should encourage that kind of behaviour.

        Here’s what I did in response to a parallel in my field: http://jayarava.blogspot.co.uk/2013/11/the-earliest-buddhist-shrine.html (Behind the scenes, I received many plaudits from people in the field for writing this critique, but few in the open)

        • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

          Thanks. I agree with all that – although I should stress that I don’t think “the researchers are gaming the system”. Rather I’m criticizing whoever wrote the press release (probably not the researchers) and Eric Lander, who is a researcher but wasn’t an author on this paper. The authors themselves may be embarrassed at how their work was hyped.

  • http://www.wiringthebrain.com/ Kevin Mitchell

    The paper by Sekar et al is a beautiful dissection of a GWAS
    signal, identification of common variants that are most likely the functional ones conveying risk and demonstration of effects on expression in human brains. They also link to functional studies of this gene in mice showing a role in synaptic pruning (we sort of knew that last bit already). So, a really nice advance, a huge amount of work, some very careful and rigorous methodology and a nice potential link to a cellular function.

    But the press release was off the hook! By suggesting this was the first ever biological insight into schizophrenia from genetics it completely ignored a vast literature on rare mutations that confer moderate to very high risk of the disorder (and of the many other conditions that show overlapping etiology).

    This review (of mine) covers the genetics of neurodevelopmental disorders in general (which is appropriate because they’re not etiologically distinct), lists dozens of CNVs and individual genes and discusses some important pathways and insights into biology http://www.ncbi.nlm.nih.gov/pubmed/20832285

    This one from MIke Owen’s group also highlights synaptic pathologies: http://www.ncbi.nlm.nih.gov/pubmed/25152434

    This one’s good on 22q11: http://www.ncbi.nlm.nih.gov/pubmed/20920576

    This one’s decent on CNVs in general: http://www.ncbi.nlm.nih.gov/pubmed/22424231

    And this one is a very nice study of effects of NRXN1 mutations on synaptic fnxn in human-derived neurons http://www.ncbi.nlm.nih.gov/pubmed/26279266

    So, there are loads of examples of biological insights, many implicating synaptic development and plasticity (including pruning). I could go on (and on and on).

    These get dismissed because each of them is rare, but this ignores the fact that there are so many of them that they can collectively account for many cases of neurodevelopmental disorders, with that number growing all the time (http://www.ncbi.nlm.nih.gov/pubmed/26817790; http://www.ncbi.nlm.nih.gov/pubmed/24866043).

    But even if they are rare, the fact that they carry high risk allows one to translate from large biological effects to large pathogenic effects.

    By contrast, the common variants – even the ones affecting C4 copy number – have tiny effects on risk – on the order of 1.2 or less. Tying the variants to effects on expression is great, but that still leaves open the question of how they influence disease risk, as discussed here: http://www.wiringthebrain.com/2015/11/what-do-gwas-signals-mean.html

    Moreover, showing that null mutations of the gene have an effect in mice does not necessarily say that a slight increase in expression in humans will do much of anything. I’m not trying to say it’s not a nice discovery – just that it’s a small step, not the breakthrough being trumpeted.

    More generally, if you believe that SZ arises due to the cumulative effects of >8000 common SNPs, as some people (including from the authors’ groups) maintain, then how important could this one SNP possibly be? Yes, it points to a cellular mechanism, but is that process also supposed to be affected by all those other loci? That seems unlikely, to say the least. So, we’re left not really knowing how central a role it plays in pathogenesis in general (if there is such a thing for such a heterogeneous disorder).

    So, all in all, a really nice study, sadly tarnished by an absurd level of hype and deep inaccuracies in the press release and associated hyperbolic quotes.

    Kevin Mitchell

  • Chris Roche

    I have scitzophrenia and take classes 3 times a week. Your looking in the oceans for tigers with looking at genes. It’s simply put a thought disorder. I’ve knoticed that there are two groups that have this disease, one with a developmental disability and ones that don’t. Stress caused my symptoms. Others just perceive the world wrong.

  • Dwight Dickinson

    This was elegant and exciting work, to be sure. I hope it will be the tip of the iceberg and that more actionable findings will follow. But important to keep in perspective. Two points. First, the paper says that the highest risk form of the C4 gene (which they found in about 10% of the samples they examined) increased odds of developing illness from 1:100 to 1.27:100. This is a small difference in odds — maybe something like the difference between having an IQ of 100 and having an IQ of 99.75. Certainly, it’s nothing you would base a drug development effort on — even if it were possible to intervene safely in the delicate and prolonged process of synaptic pruning. Second, the paper is much stronger in decoding a cryptic association between C4 and schizophrenia risk than it is in providing evidence for a causal connection to synapse pruning. The mouse model is very cool, but it concerns a different neural system (visual vs association cortex), a different period of life (early puphood vs human adolescence), fundamentally different biology (mice don’t even have differentiated C4A and C4B genes), and it addresses the effects of a deficiency of C4 protein in relation to pruning, not an excess of C4 protein, which is the authors’ central hypothesis here. I think this is wonderful work and hope it leads to more fascinating findings — but it is a very small step in the end.

  • Pingback: Virtuality Bites – Enlaces interesantes de la semana |()

  • Pingback: Virtuality Bytes – Intersting links of the week 6 February, 2016 – QUEROLUS.ORG – A DIGITAL LIFE EXPOSED()

  • Pingback: OYM94: C4 Blows the Lid off Schizophrenia>On Your Mind Podcast()

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Neuroskeptic

No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.

ADVERTISEMENT

See More

@Neuro_Skeptic on Twitter

ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar
+