Sulforaphane and autism continued: metabolomics wades in...

"We identified 77 urinary metabolites that were correlated with changes in symptoms, and they clustered into pathways of oxidative stress, amino acid/gut microbiome, neurotransmitters, hormones, and sphingomyelin metabolism."

So said the findings reported by Stephen Bent and colleagues [1] continuing a theme in autism research circles examining the use of a compound called sulforaphane - "a supplement with indirect antioxidant effects that are derived from broccoli sprouts and seeds" - in the context of [some] autism (see here). Once again, I'm sure that there may be people out there with brows furrowing when it comes to talk of a 'broccoli chemical' potentially impacting on the presentation of autism. But peer-reviewed science (placebo-controlled) is peer-reviewed science [2] and not just to be 'put to one side' because it doesn't follow the trends or [research] fashions of the day.

This latest work from Bent et al represents a not-so-methodologically strong attempt (i.e. not placebo-controlled) to bring the science of metabolomics into research proceedings to "examine changes in physiological markers that may underlie beneficial treatment effects from sulforaphane by analyzing changes in urinary metabolites." Metabolomics by the way, as well as being music to my research ears, is something that percolates through quite a lot of autism research these days (see here and see here for examples) as small molecules in urine, blood and other biofluids are separated, detected and elucidated all in the name of science.

A small group of children participated in the Bent study; all had a "formal diagnosis of autism", all were reasonably happy to swallow a tablet containing sulforaphane ("weight-based dosing of sulforaphane") and all were able to provide urine samples before the study started and at the conclusion of the 12 week research period. Parents of participants were also willing and able to complete a couple of behavioural schedules: "the Aberrant Behavior Checklist (ABC) and... the Social Responsiveness Scale (SRS)" at "baseline, 4 weeks, and 12 weeks using an online and secure platform" about their children too.

Results: alongside looking at pre- and post-intervention behavioural scores, the authors also "examined the number of participants who had a clinical response." This is a particularly important detail in the context of autism and the continuing discussions about how the spectrum is well and truly heterogeneous (the autisms?) and so one shouldn't expect every single person diagnosed to somehow have the same genetics and/or biochemistry; also affecting response to any particular intervention. With that in mind, authors also reported that (group) scores on one of their primary outcome measures - the SRS - were significant, indicative of some positive change noted to behaviour over the course of the intervention period. I say this bearing in mind that this was an open-trial, where everyone took sulforaphane and everyone knew that they were taking sulforaphane (including the parents who did the scoring). As for those potential 'best-responders' to sulforaphane use, we are told that: "Eight participants had a clinical response compared to seven who were classified as non-responders."

Then to those metabolomic results, and from a total of nearly 700 compounds identified in urine, approaching 80 of them seemed to show some correlation with the behavioural symptom changes noted. They were put into various categories depending on their biochemical form and/or action, and correlations with behavioural scores (and significance) were presented. Some of the best correlations that I could see were with regards to sphingomyelin metabolism. The authors did seem a bit surprised by results in this area, but added: "It is not clear how sulforaphane might alter sphingomyelin metabolism or availability and whether this is related to clinical benefits, but if this association is confirmed, it has important clinical and treatment implications." I daresay that examination of sphingomyelin metabolism in other contexts (see here for example) might be revealing, particularly in the context of other behavioural/psychiatric labels [3] that could (and do) overlap with autism.

There is a further scheme of work to be followed when it comes to sulforaphane and autism on the basis of these and other scientific results [4]. I'd also suggest that the continued incorporation of the science of metabolomics is a good thing, and adds a further tier of investigation when it comes to studying intervention more generally in the context of autism. It's also an important step in (eventually) coming up with a 'test' for who might be a best-responder to the use of sulforaphane and perhaps associated compounds...

----------

[1] Bent S. et al. Identification of urinary metabolites that correlate with clinical improvements in children with autism treated with sulforaphane from broccoli. Molecular Autism. 2018; 9: 35.

[2] Singh K. et al. Sulforaphane treatment of autism spectrum disorder (ASD). Proc Natl Acad Sci U S A. 2014 Oct 28;111(43):15550-5.

[3] Castillo RI. et al. From Molecules to the Clinic: Linking Schizophrenia and Metabolic Syndrome through Sphingolipids Metabolism. Frontiers in Neuroscience. 2016;10:488.

[4] Sedlak TW. et al. Sulforaphane Augments Glutathione and Influences Brain Metabolites in Human Subjects: A Clinical Pilot Study. Mol Neuropsychiatry. 2018 May;3(4):214-222.

----------

Another double-blind, placebo-controlled trial of vitamin D in autism. But...

The findings reported by Conor Kerley and colleagues [1] continue a theme in autism research circles of 'one study reporting one thing, and another study reporting something different'.

The name of the research game this time around was vitamin D supplementation in relation to autism spectrum disorder (ASD) and yet another gold-standard double-blind, placebo-controlled trial of the sunshine vitamin/hormone to look-see what happens to autistic signs and symptoms following supplementation. This continues on from a previous trial - again with gold-standard metholodgy - suggesting that there might be some benefit when it comes to supplementation (see here).

The Kerley study had been previously listed on ClinicalTrials.gov (see here). In short, supplementation with "2000 IU [international units] vitamin D3 supplementation or placebo daily for 20 weeks" were the intervention options for some 40 children diagnosed with ASD. The Aberrant Behaviour Checklist (ABC) was the primary outcome but other, more autism-specific measures (e.g. Social Responsiveness Scale) were also included in the scientific mix. Importantly in the context that vitamin D overdose is not unheard of in the context of autism (see here), researchers also probed vitamin D status in study participants.

Results: well, vitamin D supplementation did affect measures of vitamin D status as would be expected: "Following vitamin D3 supplementation, there was a significant increase in 25(OH)D to 83.8 nmol/L (p=0.0016)." When however it came to the behaviour side of things, scores on the primary outcome measure did not show any significant changes. The authors did find that "an improvement in self-care" in the supplemented group based on another scheduled used over the course of the study but that was about it.

Despite the lack of any significant effect, the authors seem to be still up-beat about the need for further study on vitamin D in relation to autism. They note: "Considering the other promising data as well as the relative safety and cheapness of vitamin D supplementation, further trials are warranted."

I would tend to agree that one study does not a scientific conclusion make. I can't say that there was a specific issue with the amount of vitamin D supplemented in the Kerley study compared to that used by Saad and colleagues [2] nor that the study timescale was too short. I might however draw your attention to the schedules used by the different studies and how perhaps the Saad study - using CARS and ATEC as well as the ABC and SRS - was perhaps better suited to detecting specific changes related to presenting autism symptoms than the Kerley study. But that is by no means an excuse for the Kerley results. There may also be other factors to consider [3].

Vitamin D deficiency/insufficiency is seemingly over-represented in relation to autism (see here for example) as it is in quite a few other behavioural/psychiatric labels (see here and see here). I do think this is an area that requires a lot more study; not least when one considers that (a) there may be groups on the autism spectrum who are genetically predisposed to lower vitamin D levels (see here) and (b) there may be potential 'best-responders' and 'non-responders' to this type of intervention (as there seems to be with many different intervention approaches with autism in mind). All this in the context that the label of autism is a great descriptor of symptoms but rather less splendid when it comes to research validity (see here).

So, just before anyone goes and touts vitamin D as 'not relevant to autism', think on...

----------

[1] Kerley CP. et al. Lack of effect of vitamin D3 supplementation in autism: a 20-week, placebo-controlled RCT. Arch Dis Child. 2017 Jun 16. pii: archdischild-2017-312783.

[2] Saad K. et al. Randomized controlled trial of vitamin D supplementation in children with autism spectrum disorder. J Child Psychol Psychiatry. 2016 Nov 21.

[3] Scragg R. Limitations of vitamin D supplementation trials: why observational studies will continue to help determine the role of vitamin D in health. J Steroid Biochem Mol Biol. 2017 Jun 13. pii: S0960-0760(17)30157-7.

----------

Autism and the GFCF diet: ScanBrit episode 2

"These preliminary observations on potential best responder characteristics to a gluten- and casein-free diet for children with autism require independent replication".

That sentence, taken from a recent (pre-print) publication I was very peripherally involved in writing, is probably the most important thing to take from the paper by Lennart Pedersen and colleagues* and certainly is a message that I would be very keen to promote.

Skandinavism @ Wikipedia 

ScanBrit, was/is a meeting of minds between the group I work with and other research groups based in Denmark and Norway, who wanted to experimentally examine the question of whether a diet devoid of foods containing gluten and casein might be able to impact on the presentation of autism in children.

Combined with a small but growing bank of research, the net outcome was that science should perhaps be looking at a possible dietary effect at least in relation to some cases of autism. One of the main issues being that the evidence base (including our own results**) was still limited in terms of the methodological quality of the investigations*** being undertaken (see here for a review). Oh and mechanism-wise, (i.e. how and why did diet 'work'?) we are still feeling around the edges, despite a few avenues of potential further inquiry (see here and here) including revisiting some older ideas.

In the most recent paper, (we) the ScanBrit collaboration were looking to further analyse the wealth of data which was generated over the course of the 2 years of experimental study and in particular, try and ascertain whether there were certain characteristics which might be markers for dietary response or not. As any person following the GFCF diet or their parents know, this is not an easy diet to follow and might, under certain circumstances, place a person at risk of nutritional deficiencies (see here) and/or other more socially-mediated issues (see here). Identifying markers of who might be a responder or non-responder would therefore save quite a lot of time, effort and expense.

Interestingly, as I've indicated in a previous post, some of the most significant and consistent behavioural changes which were noted in our original trial as correlating with the installation of a gluten- and casein-free (GFCF) diet were actually outside of the core symptoms of autism; instead potentially reflecting changes to attention and hyperactivity, more traditionally linked to the symptoms of attention-deficit (hyperactivity) disorder (ADHD). In other words, the effect of a GFCF diet might be more peripheral than anything else in cases of autism, in terms of targeting issues associated with ADHD which might also be present alongside the core autism symptoms (see my ESSENCE post and the very interesting paper by Sprenger and colleagues****).

Chronological age was identified as potentially being a factor influencing response to the GFCF diet, and in particular, those children aged between 7 and 9 years old as being potential best responders. I have to say that I am still mystified by this finding. Traditionally, I'd always assumed that there was a 'younger, the better' sentiment when applied to the possible effectiveness of dietary intervention on cases of autism, as per the rising tide of literature on other early interventions. Outside of the relatively small participant numbers we included for study, this could mean that there is some tie-in with the ADHD symptoms which seem to be targeted, bearing in mind my relative inexperience when talking about the manifestation of ADHD throughout childhood. In light of the recent case study from Dr Martha Herbert on a GFCF diet morphing into a ketogenic diet (see here), I do wonder if there might be other comorbidities also being affected which might also have some tie-in too. Who knows.

Again, without making too much of this variable at the current time, there was also a possibility for some involvement of one of the urinary compounds***** we focused on in the original ScanBrit trial as potentially being involved in response. I know biomarkers and autism is still a very complicated subject (see here) bearing in mind the heterogeneity of presentation and the potential influence of all that heightened risk of comorbidity. That all being said, trans-indolylacryloylglycine (IAG) whilst not being a biomarker for autism****** continues to garner interest with its possible connection to gastrointestinal (GI) comorbidity*******. More investigation needed here methinks.

I don't think we have made any startling discoveries with the publication of this paper. We were limited to the variables that we examined over the course of the study and so were unable to talk about whether other, more diet-relevant factors such as coeliac disease (CD) or related pathology - non-coeliac gluten sensitivity for example - might also be involved in response to diet. This was a major shortcoming of our original paper I'll freely admit.

That we found that inattention and hyperactivity might be key factors in dietary response however, is to my mind, potentially very important. That targeting such behavioural issues might also have a knock-on effect to more core autism presentation (as per our ADOS results on communication) is also an area requiring further investigation and perhaps reflects how autism (or should that be the autisms) is so very much more than just the formal clinical descriptions we give it. RDoC anyone?

Blogging about ones own peer-reviewed research is something that I've always been a little bit hesitant about doing. Aside from the possible charge of self-publicity and the natural tendency to think that yours is obviously an important paper (who wouldn't think that about their own work?), one might argue that important features such as the objectivity required for good science blogging, might to some degree be compromised. Sort of like marking your own homework as per one of my colleagues oft-cited phrases. I very much hope that I've not over-stepped the mark on good science blogging in this post on Lennart's paper. Indeed I'll end by reiterating part of the opening sentence to this post on the requirement for independent replication before anyone gets too excited about these results, as indeed the dietary research base as it presently stands is in need of more inquiry********.

And since we're on the topic of diet and autism, would that be one hump or two*********?

----------

* Pedersen L. et al. Data mining the ScanBrit study of a gluten- and casein-free dietary intervention for children with autism spectrum disorders: behavioural and psychometric measures of dietary response. Nutr Neurosci. 2013 (pre-print)

** Whiteley P. et al. The ScanBrit randomised, controlled, single-blind study of a gluten- and casein-free dietary intervention for children with autism spectrum disorders. Nutr Neurosci. 2010 Apr;13(2):87-100.

*** Millward C. et al. Gluten- and casein-free diets for autistic spectrum disorder. Cochrane Database Syst Rev. 2008 Apr 16;(2):CD003498.

**** Sprenger L. et al. Impact of ADHD symptoms on autism spectrum disorder symptom severity. Res Dev Disabil. 2013 Aug 21;34(10):3545-3552.

***** Anderson RJ. et al. Identification of indolyl-3-acryloylglycine in the urine of people with autism. J Pharm Pharmacol. 2002 Feb;54(2):295-8.

****** Wright B. et al. Is the presence of urinary indolyl-3-acryloylglycine associated with autism spectrum disorder? Dev Med Child Neurol. 2005 Mar;47(3):190-2.

******* Wang L. et al. Is urinary indolyl-3-acryloylglycine a biomarker for autism with gastrointestinal symptoms? Biomarkers. 2009 Dec;14(8):596-603.

******** Winburn E. et al. Parents' and Child Health Professionals' Attitudes Towards Dietary Interventions for Children with Autism Spectrum Disorders. J Autism Dev Disord. 2013 Sep 1.

********* Al-Ayadhi LY. & Elyass Elamin N. Camel Milk as a Potential Therapy as an Antioxidant in Autism Spectrum Disorder (ASD). Evidence-Based Complementary and Alternative Medicine, 2013, Article ID 602834, 8 pages, 2013. doi:10.1155/2013/602834

----------

Lennart Pedersen, Sarah Parlar, Kajsa Kvist, Paul Whiteley, & Paul Shattock (2013). Data mining the ScanBrit study of a gluten- and casein-free dietary intervention for children with autism spectrum disorders: Behavioural and psychometric measures of dietary response Nutritional Neuroscience