Archive | Nature RSS feed for this section

A multicenter blinded analysis indicates no association between chronic fatigue syndrome/myalgic encephalomyelitis and either xenotropic murine leukemia virus-related virus or polytropic murine leukemia virus

19 Dec

There was much discussion of the possible imprtance of the xenotropic murine leukemia virus-related virus (XMRV) in conditions such as chronic fatigue syndrome or myalgic encephalomyelitis (CFS/ME), prostate cancer and autism. To be clear, the possibility of an autism association was made in the press, not in the research literature. For XMRV in general, there was much discussion in the press, in journals and online as it became clear over time that there were possible problems with the analyses that led to the main papers on the topic. The present study includes work by a multi-site team including the principle author of the original study linking XMRV with CFS/ME.

If one can boil a large, multi-site study result into one line, it would be this:

Here, the original investigators who found XMRV and pMLV (polytropic murine leukemia virus) in blood of subjects with this disorder report that this association is not confirmed in a blinded analysis of samples from rigorously characterized subjects

I.e. there is no link between XMRV and CFS/ME.

Here is the abstract, and the full paper is online as well:

The disabling disorder known as chronic fatigue syndrome or myalgic encephalomyelitis (CFS/ME) has been linked in two independent studies to infection with xenotropic murine leukemia virus-related virus (XMRV) and polytropic murine leukemia virus (pMLV). Although the associations were not confirmed in subsequent studies by other investigators, patients continue to question the consensus of the scientific community in rejecting the validity of the association. Here we report blinded analysis of peripheral blood from a rigorously characterized, geographically diverse population of 147 patients with CFS/ME and 146 healthy subjects by the investigators describing the original association. This analysis reveals no evidence of either XMRV or pMLV infection. IMPORTANCE Chronic fatigue syndrome/myalgic encephalomyelitis has an estimated prevalence of 42/10,000 in the United States, with annual direct medical costs of $7 billion. Here, the original investigators who found XMRV and pMLV (polytropic murine leukemia virus) in blood of subjects with this disorder report that this association is not confirmed in a blinded analysis of samples from rigorously characterized subjects. The increasing frequency with which molecular methods are used for pathogen discovery poses new challenges to public health and support of science. It is imperative that strategies be developed to rapidly and coherently address discoveries so that they can be carried forward for translation to clinical medicine or abandoned to focus resource investment more productively. Our study provides a paradigm for pathogen dediscovery that may be helpful to others working in this field.

There was a lot of hope in the CFS/ME community that this was a breakthrough that could lead to a treatment. Unfortunately, the answers they seek are elsewhere.

As this is an autism-focused site, allow me to bring this back to autism. Unlike CFS/ME, there were no papers claiming an association between autism and XMRV. Instead there were public comments by the researcher involved and inflammatory journalism. In a search for XMRV autism the first article I get is: Is Autism Associated with A Viral Infection?, by David Kirby published at the Huffington Post. Mr. Kirby’s article was probably the first that pushed the (now failed) XMRV/autism hypothesis strongly into the public’s eye. Mr. Kirby was well known for some time previous for his work promoting the idea that vaccines cause autism. In specific, he was a major proponent of the idea that thimerosal in vaccines caused autism, having published a book Evidence of Harm: Mercury in Vaccines and the Autism Epidemic. For his Huffington Post article on XMRV, Mr. Kirby had some rather irresponsbile speculations from XMRV researcher Judy Mikovits and the founder of her reseach institute Annette Whittemore. From those quotes, Mr. Kirby proceeded to present the XMRV news story in his own way, as a series of speculative questions to create an impression built like a house of cards. The impression he left the reader with was that the XMRV story helped to explain a possible link between autism and vaccines. Following a quoted statement by Mikovits, Mr. Kirby wrote

So there you have it – a possible explanation of regressive autism in a significant number of cases associated with immune system deregulation triggered by vaccination.

Of course, much more work is needed to nail down the exact significance of such an association. For example, is the virus implicated in the cause of autism, or do children harbor the virus as a result of autism?

Notice that he doesn’t say, “much more work is needed to show that this is a real association“. No, rather than stress again that the hypothesis was poorly supported, he jumps to assuming the association and asking what significance it has. Classic David Kirby.

To be fair, the comments by Mikovits and the founder of the research center where she worked (Annette Whittemore) fed directly into his story. To say it again, those statements by Mikovits and Whittemore were irresponsible given the early stage this work was in. But even with those statements, Mr. Kirby had no justification to go into this speculative paragraph:

The discovery raises more questions than it answers. What, exactly, is it about immunization that might switch on XMRV viral expression? Could the effect of heavy metals upon cytokine balances be at play? Where did this retrovirus come from, and how did it apparently become so prevalent in children with autism? Did these children inherit the virus from a parent, or was there some other unexplained route of transmission? Why has the NIH said nothing about XMRV in association with autism, and did Dr. Insel know about these findings without sharing them with the IACC

Again, we see the series-of-questions approach that is Mr. Kirby’s style. He isn’t saying immunization switches on XMRV viral expression (whatever he meant by “XMRV viral expression”. It sounds technical though). He’s posing it as a question. Notice how he brought in his mercury hypothesis, but as “heavy metals”. “Could the effect of heavy metals upon cytokine balances be at play?”. This is a great example of a sciency-sounding sentence that has no substance. Whoever was his editor at the Huffington Post should have shot that back with “do you even know what your talking about here?” But if the editor at the Huffington Post was doing his/her job, this article (and many more by Mr. Kirby) wouldn’t have been published there anyway. It is worth noting that by the time this article was written, the evidence was overwhelmingly against the idea that mercury in vaccines raised autism risk, but this was Mr. Kirby’s way of loosely tying his failed hypothesis to his then current speculation.

To pull the last sentence out of Mr. Kirby’s paragraph: “And why had the NIH said nothing about XMRV?”. Perhaps because they were more responsible than Mr. Kirby.

As a point of fact, XMRV is not prevalent in autistics (Lack of infection with XMRV or other MLV-related viruses in blood, post-mortem brains and paternal gametes of autistic individuals and PCR and serology find no association between xenotropic murine leukemia virus-related virus (XMRV) and autism.) In fact, as will be discussed below, it appears to not infect humans. Unfortunately, Mr. Kirby has not seen fit to post corrections. To the XMRV story or others.

The impression Mr. Kirby created with his story was strong. For example, he gathered 298 comments to his article, largely focused on vaccines. Here’s the last one, prominently at the top of the list:

David: As big as this autism story is, it is only one toe of the elephant. Here is another: There are no protections in place to prevent more XMRV from entering the nation’s blood supply. There is as of yet no XMRV screening test for donated blood. And — I just called my local Red Cross – there is as of yet nothing to prevent people diagnosed with CFS from donating blood. We are all at risk.

The elephant: How did our government let this potentially deadly retrovirus spread unchecked for twenty-five years? XMRV has, so far, now has been found to occur in people with autism, lymphoma, a severe form of prostate cancer, atypical MS, ME/CFS, and fibromyalgia. Twenty-three years ago the CDC was first informed of an outbreak of what we now know to be an XMRV-associated local epidemic. Eighteen years ago a study showed a retrovirus was associated with ME/CFS.

The band played on.

Yes, let’s spread fear about the blood supply, based on news reports, speculation and bad science.

Some of the authors of this present XMRV and CFS/ME study were also involved in a separate major multisite study on MMR and autism. I am referring to a study intended to replicate the key findings of some of Andrew Wakefield’s research. That study, by Mady Hornig, W. Ian Lipkin and others, Lack of association between measles virus vaccine and autism with enteropathy: a case-control study been re-interpreted by some as supporting Mr. Wakefield’s work. Some have gone so far as to claim that Mr. Lipkin’s team is signalling support for Mr. Wakefield’s work by citing it in other studies. It’s a stretch, a mind boggling stretch, and it’s wrong.

From the CFS/ME paper:

Sensitive molecular methods for microbial discovery and surveillance have enabled unique insights into biology and medicine. However, increased sensitivity for bona fide signal increases the risk that low-level contaminants may also be amplified. This can lead to spurious findings that pose challenges for public health and require an expensive and complex pathogen dediscovery process. Examples wherein authors of this paper have been engaged in this process include refutation of associations between enterovirus 71 and amyotrophic lateral sclerosis (24) and MMR vaccine and autism (25).

Lipkin and Hornig consider their work to be a “refutation” of the association between MMR and autism. But don’t take that one sentence from the paper as the only proof. Here’s an interveiw with Prof. Lipkin at Nature.

Had we done this when Andrew Wakefield [the former medical researcher who proposed that autism was caused by vaccines] came out with the initial report about the measles, mumps and rubella (MMR) vaccine and autism, and had something this definitive, there are many more children who would have been vaccinated against measles during the ten years it took us to finally complete the MMR–autism work. So I think it’s crucial that we don’t do things in a half-baked fashion, so we can test hypotheses and move on to new ones.

The interviewer even includes the MMR refutation as part of a question: “You have disproved the autism–MMR connection and other controversial disease links.”

In general, what can one say about XMRV? Aside from the drama involved in the story (which I did not discuss in detail in this article), and the questions about CFS/ME, autism, prostate cancer and more, what can we say? Prof. Lipkin says it very clearly in the interview:

We did not find any genetic sequences [of XMRV or related viruses] in the people with CFS or the controls. As far as we know, there is no human being that is infected with XMRV.

But there were papers (some now retracted) claiming some links between XMRV and human disease? What about those? Another quote pulled from the interview:

I think the explanation is that there was contamination. I don’t see any reason to invoke anything beyond that.

For this you have to give Judy Mikovits some credit. She worked with the team that was attempting to replicate her results. Contrast this with, say, Andrew Wakefield. A man whose hospital offered him the opportunity to replicate his own results, and he quit rather than accept that offer. A man who has repeatedly denied the science which has been clearly against his hypothesis. A man who denies the fact that he acted unethically in many ways in conducting his research. Judy Mikovits made some mistakes, both scientific and socially, but she seems to be part of the solution.

But that’s a bit of a sideshow. The main conclusion is that XMRV is not involved with autism. Or, apparently, any human disease.

With apologies for revisiting David Kirby and Andrew Wakefield.


By Matt Carey

Verbal and non-verbal intelligence changes in the teenage brain

24 Oct

A recent Letter in the Journal Nature takes on the question of how stable is intelligence in teenagers. The title is pretty self explanatory: Verbal and non-verbal intelligence changes in the teenage brain.

Consider teenagers. Is their IQ “set in stone”? By the time a kid is, say, 13, isn’t his/her intelligence pretty well demonstrated? What if you heard that IQ changes can change, up or down, by as much as 20 points during the teenage years? Would you be surprised? I was.

The article attracted my attention on its own merits, but also given the nature of the discussion that forms around early intervention and autism. Statements about “windows of opportunity” and times when the brain is “plastic” are common in discussions of autism focused educational therapies. For example, a study entitled “Early intervention and brain plasticity in autism” or “Autism: a “critical period” disorder?“. Certainly early childhood is a period of great learning and growth, but does the “window” of plasticity “close”?

Here is the abstract:

Intelligence quotient (IQ) is a standardized measure of human intellectual capacity that takes into account a wide range of cognitive skills1. IQ is generally considered to be stable across the lifespan, with scores at one time point used to predict educational achievement and employment prospects in later years1. Neuroimaging allows us to test whether unexpected longitudinal fluctuations in measured IQ are related to brain development. Here we show that verbal and non-verbal IQ can rise or fall in the teenage years, with these changes in performance validated by their close correlation with changes in local brain structure. A combination of structural and functional imaging showed that verbal IQ changed with grey matter in a region that was activated by speech, whereas non-verbal IQ changed with grey matter in a region that was activated by finger movements. By using longitudinal assessments of the same individuals, we obviated the many sources of variation in brain structure that confound cross-sectional studies. This allowed us to dissociate neural markers for the two types of IQ and to show that general verbal and non-verbal abilities are closely linked to the sensorimotor skills involved in learning. More generally, our results emphasize the possibility that an individual’s intellectual capacity relative to their peers can decrease or increase in the teenage years. This would be encouraging to those whose intellectual potential may improve, and would be a warning that early achievers may not maintain their potential.

” More generally, our results emphasize the possibility that an individual’s intellectual capacity relative to their peers can decrease or increase in the teenage years.” Those are quite powerful words.

The authors tested 33 subjects at two time periods:

They were first tested in 2004 (‘time 1’) when they were 12–16?yr old (mean, 14.1?yr). Testing was repeated in 2007/2008 (‘time 2’) when the same individuals were 15–20?yr old (mean, 17.7?yr).

And they found that the while the IQ scores were relatively stable, on average, individuals showed large changes:

The wide range of abilities in our sample was confirmed as follows: FSIQ ranged from 77 to 135 at time 1 and from 87 to 143 at time 2, with averages of 112 and 113 at times 1 and 2, respectively, and a tight correlation across testing points (r = 0.79; P<0.001). Our interest was in the considerable variation observed between testing points at the individual level, which ranged from -20 to +23 for VIQ, -18 to +17 for PIQ and -18 to +21 for FSIQ.

Individuals had changes in verbal IQ of -20 to +23 points, with large changes seen in performance IQ (PIQ) and full-scale IQ (FSIQ) as well. Twenty point swings in IQ, up or down? That’s a lot. As noted before, the study is rather small (33 subjects), but what makes this an impressive study is that they have physical data–brain structure data–to correlate with the changes in IQ. The authors performed MRI scans on the subject. These are shown in a Figure from the paper (click to enlarge):

The authors found that grey matter changed in specific areas of the brain and that these changes correlated with the changes in VIQ and PIQ.

The brain is not just “plastic” in terms of IQ scores, it is still able to physically change during teenage years.

Our findings demonstrate considerable effects of brain plasticity in our sample during the teenage years, over and above normal development.

If an early intervention program were to claim that some kids gained 20 IQ points, it would be huge. (Of course, if they had to admit that some kids lost 20 IQ points, it would also be huge, but in a different way)

But if non-autistic kids can see such large swings in IQ during the teenage years, why not autistic kids? Why not kids with intellectual disability and autism? Just as important as the potential for swings up in IQ are the losses in IQ. What if a kid ends up in a placement that is inappropriate to the level that IQ is lost?

I would love to see a study such as this one on autistic kids, especially those with intellectual disabilities, to track IQ and brain structure during age ranges outside of early childhood.

Get ready for the flood of fetal gene screening

21 Feb

In a recent article in Nature, Henry T. Greely, proposes that the time when prenatal genetic screening may be commonplace is closer than, well, I thought.

The world’s news media was buzzing last week after researchers showed that a blood test for mothers could detect Down’s syndrome in their fetuses[1]. Last month, two research groups independently published proof that the fetal genotype — the genetic status at a given locus — can be derived for thousands of sites from samples of fetal DNA with just a 10-millilitre blood draw from a pregnant woman[2, 3].

[1] Chiu, R. W. K. et al. Br. Med. J. doi: 10.1136/bmj.c7401 (2011).

[2] Lo, Y. M. D. et al. Sci. Transl. Med. 2, 61ra91 (2010).

[3] Fan, H. C. & Quake, S. R. Nature Precedings doi:10.1038/npre.2010.5373.1 (2010).

Until now, the main prenatal testing has been for Down Syndrome. It is not common:

Prenatal genetic testing has been clinically available since the late 1960s, but the costs, inconvenience and especially the miscarriage risks have limited its use. Each year, less than 2% of pregnant women in the United States undergo amniocentesis (in which a small amount of amniotic fluid containing fetal cells is taken for analysis) or chorionic villus sampling (CVS — in which fetal tissue is extracted from the placenta). Both procedures increase the risk of miscarriage. Until now, any given sample could be tested for only one or two conditions, typically chromosomal abnormalities such as trisomy 21, the cause of Down’s syndrome.

It is uncommon, but it is offered in high risk situations (older mothers). Since the test has been offered, the prevalence of Down Syndrome has dropped significantly. This in spite of the fact that older mothers are more common now.

Amniocentesis is obviously invasive, resulting in risk to the unborn child. But a blood draw would be a non-invasive prenatal genetic diagnosis (NIPD).

The potential of NIPD goes way beyond Rhesus screening. Two of the leading researchers in cell-free fetal DNA testing — Dennis Lo of the University of Hong Kong and Steve Quake of Stanford University in California — use different methods to analyse fetal cell-free DNA from maternal serum. Each has demonstrated the ability to detect aneuploidies — missing or extra chromosomes, such as in trisomy 21 (refs 5, 6). Last month, both researchers published proof that the fetal genotype could be derived for thousands of sites from cell-free fetal DNA2, 3 — demonstrating the possibility of using maternal blood to test for all fetal genetic traits.

The methods demonstrated are already interesting commercial firms:

Commercial firms are already interested. Sequenom in San Diego, California, is working with Lo; another, Artemis Health of Menlo Park, California, is working with Quake; and still others are also exploring the technology. For-profit development of these methods seems likely within five years, at least for chromosomal abnormalities, such as trisomy 21, and possibly for single-gene traits.

My insurance plan will pay for prenatal genetic testing, but not for genetic testing of a child. I find that thought a bit chilling.

Until now, prenatal genetic testing has been relatively uncommon. The ethics discussions have been largely academic. Important, but academic. The time for academic discussions of the ethics is drawing to a close.

Professional organizations, in medicine and in genetics, need to get involved, both in training their members about these technologies and in beginning to consider guidelines for their use, especially with regard to informed consent. Regulators, companies and consumer advocates need to be talking about pathways for assuring the safety, efficacy and quality of NIPD testing. In the United States, the Food and Drug Administration should start that process immediately. And it is time for ethics commissions, such as the US Presidential Commission for the Study of Bioethical Issues, to report on these issues.

Most importantly, we need to start conversations, between all those concerned, about the limits, if any, to place on this powerful technology. Whether we view NIPD gladly as a way to reduce human suffering, warily as a step towards a eugenic dystopia, or as a mix of both, we should agree that the better we prepare, the more likely we are to avoid the worst misuses of this potentially transformative technology.

What disabilities will eventually have genetic screening possible? How will we as a society take on the ethical challenges? If the drop in Down Syndrome is any indication, I think there is reason to take this discussion very seriously. Now.

Interview with Stephen Scherer

11 Jun

Stephen Scherer is co-author on the recent paper ‘Functional impact of global rare copy number variation in autism spectrum disorders‘. I caught up with him via email to ask a few questions:

1) What is the ‘bottom line’ message readers can take away from your work?

I am always frustrated when I hear at the end of most news stories…’and we don’t know what can cause autism’. Data from the past few years including our new study show alterations in genes can cause autism. We have not found all of the genes yet, and not all autism cases can be accounted for (the genetics can be complex) but genes can cause autism. For the specialists, through our new study we show either de novo or rare inherited copy number variations as one form of genetic alteration involved in autism. The genes affected are often linked together in a connected functional pathway and may of these molecules dictate how brain cells (neurons) develop and communicate. Some of the autism genes we found have already been found to cause intellectual disabilities, which is not entirely surprising since many individuals with autism also have these challenges.

2) How does your paper tie in with other gene/autism studies?

We validate many previous findings, but also find dozens and dozens of new autism risk genes. Our data furthers the hypothesis that rare genetic alterations contribute much more relative risk to developing autism than common genetic variations.

3) What should future researchers use your study for in terms of direction to take their own work?

I think one of the most important impacts of the study is the design itself. Nature really wanted us to include the Figure 1, which outlines the design and analysis. CNV studies are still quite tricky to do and the data has to be of the highest quality to make sense of it. I think our study on autism will set the standard for all other studies going forward, so they should follow it. Moreover, many of the functional pathway studies published before may have been either underpowered, flawed by low resolution arrays or high false discovery rates, or incomplete study designs. We spent alot of time thinking of how to best do this properly so if others are interested they should read the Supplementary Information carefully. Use it as a guide. Finally, for the functional biologists look at the long list of genes we present in the Supplementary Information since they may find their favorite gene to be an autism candidate gene!

4) How difficult was it managing the input from such a very large amount of co-authors?

The Autism Genome Project has some 120 scientists from 11 countries involved (see the authorship list). We selected a ‘writing team’ comprised of genome scientists, statisticians, medical geneticists, psychiatrists and developmental pediatricians. Interesting, everyone saw their own story in the data. I pretty much new in advance what I wanted to see in the final manuscript so much of my job was bring focus to the many other good ideas (note that there are five other papers spinning out of this larger study presenting some of these other data and interpretations). In took about 12 solid months of analysis of the data, three months of writing and editing, alot of cursing, and then submission to Nature (with even more cursing). The review time from submission to publication took ~six months. This was the hardest for me (except other than constantly changing the author list….and affiliations). The reviewers were very very thorough and Nature can (rightfully) be very demanding. In the end we are very proud of the manuscript. Dalila Pinto who is the first author and a post-doc in the lab was the driving force behind the analysis and deserved a lions-share of the credit. Would I do it again? I’ve had two Senior Author papers in Nature this year, which have been very draining. But I would do it again!