Stem cell clinical trial for autism: proceed with caution

Recent headlines have trumpeted an FDA-approved clinical trial of cord blood–derived stem cells for autism, involving 30 children and two i.v. infusions of cells from each child’s own banked cord blood. The stated rationale is a link between inflammation and autism, but I, for one, find that rationale spurious.

The inflammation–autism concept found its footing in part thanks to a study that appearedin 2005 [open access] and described findings linking inflammation in the post-mortem brain and autism. The researchers, based at Johns Hopkins, had examined donated brains from 11 people with autism, six of whom were children, and in particular found evidence of what they called an “inflammatory process” in the cerebellums of brains from autistic people. The autistic group in this study was highly heterogeneous in terms of ages, causes of death, and the presence of epilepsy.  

The Hopkins work and a handful of other papers led some researchers to focus on inflammation, particularly neuroinflammation, as causative in autism. In spite of some faint signposts pointing to an association, however, the Johns Hopkins group that published the 2005 paper warns on its Website that “there is no indication for using anti-inflammatory medications in patients with autism.” Medications like prednisone, they say, would not affect the brain cells they identified as inflamed because the steroids influence a different part of the immune system.

Although research has yet to link autism and neuroinflammation definitively or even sort of definitively, as with many investigative niches in science, a handful of groups dominates the landscape. Among these groups is a Stanford CalTech team led by Paul Patterson, who has written a book about purported links between the immune system and autism. Their most recent findings [paywalled] in a mouse model of autism suggested an association between what they call maternal immune activation (as might be triggered by a viral infection) and autism. In this work, the authors also irradiated affected mice and performed bone marrow transplants that, the team reported, reduced some of the “autism-like” repetitive and “anxiety-like” behaviors.  On his blog, Patterson emphasizes that this work was done in mice, not people, and does not provide an indication for invasive and traumatic bone marrow transplants for autism. Indeed, he notes, “We have yet to establish whether it was the infusion of stem cells or the … transplant procedure itself--complete with irradiation--that corrected the behaviors.”

The point of a bone marrow transplant is in part to replace and replenish cells that participate in immunity. Stem cells form the replacement material, and of course, medical science now has identified many sources of stem cells. One of these sources is umbilical cord blood, which also contains several other cell types. Parents who want to spend the money for peace of mind can arrange to have their infant’s blood taken at birth and stored privately in a cord blood bank. The self-described “largest and most experienced” bank in the United States is the Cord Blood Registry, which also conducts clinical trials with stem cells from the blood samples. I note that the American Academy of Pediatrics discourages private cord blood banking unless an identified need for stem cell therapy already exists.

Stem cells are, obviously, both controversial and potentially big business. Like gene therapy before it, the field of stem cell therapy was expected to be the Next Big Thing, and for some diseases, it might be keeping that promise. The source of the stem cells can be a big factor in ethics, safety, and efficacy, but the safest and most ethically neutral source for an individual receiving a transplant is presumed to be cells that come from that person. Because cord blood stem cells are taken before mutations make cells cancerous, these cells can be effective in transplants to treat, for example, leukemia.

Possibly the most high-profile potential application of stem cells is as treatment for cancers like leukemia and degenerative neurological disorders. Autism is not a degenerative neurological disorder, but because it is neurological and because of some hints at an involvement of the immune system, groups like Patterson’s have homed in on stem cells as a way to replace ostensibly pathological immune system cells with a fresh, presumably healthy batch. Cord blood seems like the perfect source and relatively safe, and the most desirable approach to administration would be a simple intravenous infusion. A Korean group has tried exactly that for a different disorder, cerebral palsy (CP), infusing a specific type of cord blood stem cells called mononuclear cells into a group of 20 children with CP. Five of the patients, or 25%, showed some improvement. Because this type of trial and treatment is in its infancy, no one know what the long-term implications of such treatments are, but the safety profiles thus far are good.

These results and others applying cord blood stem cells for CP have led to interest in trying out a similar treatment for autism, although not enough time has passed for determinations of long-term outcomes, good or bad. Already, some parents seek stem cell transplantation abroad as a treatment for their children’s autism, even as safeguards for the blood products and their contents are lacking and many for-profit clinics overpromise, at best. In fact and not surprisingly, fraud is common. There is likely no more vulnerable population than parents who are desperately seeking to help their children.

In part in an effort to focus on stem cell interventions for autism in a better-regulated environment, the FDA has just approved a clinical trial that will involve intravenous injection of stem cells derived from cord blood into autistic children, with each child receiving cells from his or her own cord blood. News reports have referred to this planned trial as “groundbreaking,” and one outlet went so far as to announce in a headline that it would “cure autism.” Not everyone, however, received the news with unfettered joy.

The trial is the result of a collaboration between the Cord Blood Registry and Sutter Health pediatric neurologist Michael Chez, who also focuses on an immune link with autism and has written a book about the medical management of autism. Given the news saturation the announcement received, I wouldn’t be surprised if phones at the Cord Blood Registry weren’t buzzing all day long as worried parents anticipating a birth dial in, inspired to bank blood as a treatment against their future child’s possible autism. The Cord Blood Registry appears to want to hammer that point home, offering up this graphic on their clinical research site for the study, emphasizing the CDC’s recent report of a 1 in 88 prevalence of autism in the United States and the five-times greater rates among boys. It’s a clear message to expectant parents who know they’re having boys.

But the graphic and accompanying text don’t say much about the research supporting the endeavor. What we have is a few studies suggesting an autism–immunity link, although not all findings support one [paywall], some partial effectiveness of cord blood cell infusion for cerebral palsy, and no data regarding what effect, if any, a cord blood infusion would have in autism. The rationale for the work appears to be the CP trials and Patterson’s work with a mouse model of inflammation.

In a review that Chez co-authored in 2010 [PDF] about the potential for immune therapy in autism, he cites the 2005 paper from the Johns Hopkins group that assessed inflammation in post-mortem autistic brains. In that review, he also notes that each autistic person has a “uniquely different clinical appearance in each individual” and that there are “subsets of autistic patients” with neuroinflammation. One group from Italy has sought to identify these subsets and characterized four groups: one with immune abnormalities, two with little or no immune abnormalities, and a final group with a mix of traits from the other three groups. Of their population of 245 patients, about 18% fell into the “immune abnormality” group.

The cord blood autism trial will include 30 children, all of whom have cord blood banked with the Cord Blood Registry. Fifteen will receive a stem cell infusion and 15 will receive a placebo injection. After 24 weeks, the two groups and their treatments--stem cell or placebo--will be switched. The endpoints of interest are improvements on behavior scales. Pull together the terms “stem cells” and “autism,” and this trial--based on the news reports at least--looks like a hot science winner.

In the midst of all of this celebration and starry-eyed excitement about stem cells, however, the consumer would do well to proceed with caution. The trial suffers from two paradoxical issues. First, there doesn’t seem to be a specific autism-related rationale for treating children with stem cell injections. Cerebral palsy and autism can have some overlap and be co-morbid, but they are not the same thing. Yet, if the trial must go on, the proposed population seems too small. As Chez himself has noted, autism is highly heterogeneous. As the Italian group found, a minority percentage seems to represent a subset that might have immune problems. The inclusion criteria for the trial don’t mention any immune-related endpoints. With a small group and a high heterogeneity, even if cord blood–derived stem cell infusions have an effect, the results might be so individual and sporadic that rather than giving answers and hope, the trial may just end with a soft thud of equivocal results.

There is a real essence of “stem cells are hot and autism is hot so lets throw some stem cells at autism” here. Science and clinical trials involving a pediatric population in particular shouldn’t rely on the “throw it at the wall and see if it sticks” mode of testing. Regardless of how safe the protocol is, administering even intravenous infusions to an autistic child can be traumatic for everyone involved. A trial like this seems hasty and the ballyhoo surrounding it overstated and overpromising. Finally, if they do find an effectiveness of cord-blood–derived stem cells for treating autism--how does that translate into clinical practice? Currently, autism is diagnosed many months or years after birth. Any family hoping to benefit, at least from direct infusion, would have to have had the foresight to privately collect and store their children’s cord blood for any results of this trial to be applicable. And as I’ve noted, that requires no mean outlay for a possibility—autism— that still remains at around 1%.

See also this post from the California Institute for Regenerative Medicine