The science of secretin: will studies of rodent brains give new life to a controversial autism therapy?

In 1998 in two television appearances, a boy named Parker Beck helped make the protein secretin the latest in a long line of would-be cures of the puzzling neurological disease known as autism. The autistic child was plagued by frequent diarrhea and vomiting--it's unclear why, but many children with autism suffer from gastrointestinal problems in addition to the typical problems of social withdrawal.

As part of a routine diagnostic test, Parker's doctors at the University of Maryland in Baltimore injected him with a single dose of secretin. This hormone, naturally produced by the small intestine, stimulates the pancreas to secrete digestive enzymes and the liver, to produce bile.

Over the next several months, the boy's social behavior, sleep patterns, and use of language improved considerably, according to his parents. They became convinced that secretin was responsible for reducing the symptoms of autism.

After two other autistic children seemed to benefit from secretin injections, Parker's parents and the Maryland physicians began to champion the hormone in public, including on two national television shows. A mad rush for the drug broke out, overwhelming the limited supply. At the time, secretin was derived only from pig intestines. "People were literally mortgaging their house to get infusions of secretin" for their children, recalls Marie M. Bristol-Power of the National Institute for Child Health and Human Development in Bethesda, Md.

That concerned many veteran autism researchers. "It pays to be somewhat skeptical because every 6 months on average, there's some new cure touted," says Fred R. Volkmar of the Child Study Center at Yale University.

Secretin particularly mystified scientists because there was no obvious explanation for how a gut hormone could influence the course of a brain disorder. Their wariness seemed justified when a handful of small studies over the next 2 years produced little evidence that secretin helps autistic children more than a placebo injection does.

Now, however, scientists have for the first time obtained compelling evidence that secretin has a role inside the brain. In fact, the hormone appears to work in brain regions previously implicated in autism.

Relying on such work and other, unpublished data, the company that manufactures a synthetic form of secretin plans to launch two new trials intended to persuade the Food and Drug Administration to approve the drug for use in autism. The latest evidence has also made some researchers more open-minded about secretin.

"I've switched sides on a number of occasions. Now I think there's something to [secretin]," says Paul Shattock, director of the autism research unit at the University of Sunderland in England. "It's not the wonder drug that was promised 3 years ago, but there are certainly some people who may benefit from it." Bristol-Power adds that even if secretin doesn't become part of a therapy, it could provide insight into some cases of autism.

First described in 1943, autism affects 5 to 15 out of every 10,000 children. Striking boys far more often than girls, the disorder remains a mystery. The most common explanation, albeit a broad one, is that abnormalities in brain development produce the condition.

Autism is not usually considered a form of mental retardation. While autistic people can have trouble picking up the fundamentals of language, some display significant intelligence. Still, most autistic individuals don't connect emotionally and socially with other people. They often appear to live in their own world and frequently display highly focused, ritualistic, and bizarre behavior, sometimes even repeatedly hitting their heads on walls.

Physicians have little to offer parents of autistic children. "The most commonly used drug for autism in the United States is [the stimulant] Ritalin, and there's no evidence whatsoever of its efficacy," says Shattock.

The absence of any proven therapy is one reason that the story of secretin spread like wildfire among families with autistic children. After Parker's story made headlines, several research groups rushed to test the safety and effectiveness of secretin injections on dozens of children and adults with autism and related conditions.

The National Institutes of Health in Bethesda, Md., helped fund several trials. Half the participants in these tests received a placebo injection instead of secretin, and neither the researchers nor the participants knew who got the drug until the end of the trial.

At the end of 1999, the bad news started to roll in. Secretin emerged as safe, but none of the trials found that it helps patients, at least not in a statistically significant way. For example, the first trial to be reported started with 56 children age 3 to 14, two-thirds with true autism and one-third with pervasive developmental disorder, a condition marked by autistic symptoms.

Among 27 children injected with a single dose of secretin, 9 were judged "much improved" or "very much improved" in their autistic symptoms, Adrian D. Sandier of Thomas Rehabilitation Hospital in Asheville, N.C., and his colleagues reported in the NEW ENGLAND JOURNAL OF MEDICINE. Yet 7 out of 25 children in the placebo group showed similar improvements. Other researchers soon reported equally equivocal results in other journals.

"The body of work on secretin as a therapeutic agent has unfortunately been disappointing," concludes Volkmar.

Walter C. Herlihy, president of synthetic-secretin manufacturer Repligen in Needham, Mass., disputes that. He contends that these initial studies confirmed the safety of the drug but were too limited to gauge the value of secretin. The trials involved few participants, with broad ranges in age and symptoms, so he says he's not surprised that no clearly beneficial effect emerged.

Researchers are still struggling to design trials that gauge whether a therapy helps an autistic child, Herlihy points out. "Clinical trials are really something new to the field of autism. There isn't a lot of precedence about how you should go about them," he says. "How do you measure the symptoms of autism? That's a fundamental question."

At the moment, investigators depend on a battery of subjective evaluations by parents and psychologists. Yet, says Shattock, many of these appraisals were designed to diagnose autism, not discern changes in a behavioral condition whose symptoms can naturally wax and wane.

Skeptical from the start, many autism researchers have considered the negative clinical trials the last word on the secretin frenzy. The evidence that secretin plays a role in the brain may force a review of that judgment, however.

In the Sept. 15 JOURNAL OF NEUROSCIENCE, a research team led by Billy K.C. Chow of the University of Hong Kong reports that the hormone and its receptor--the cell-surface protein that recognizes secretin--are both present in the cerebellum of rats. There were previous reports that secretin appears in the mammalian brain, but most scientists dismissed them as products of experimental artifacts, notes Andrew Leiter of the New England Medical Center in Boston.

Chow's group uses a technique called in situ hybridization histochemistry, which is much more accurate and sensitive than previous detection methods. With it, the researchers show that the genes for secretin and its receptor are active in the rat cerebellum. Moreover, the investigators pinpoint specific brain cells. The first, called Purkinje cells, make the two proteins. The secretin receptor is also made by a second group of neighboring brain cells, ones that a neurotransmitter called GABA activates. Secretin released by Purkinje cells may regulate the responsiveness of the nearby cells to GABA, the researchers suggest.

These results are provocative because the cerebellum and its Purkinje cells have long been of interest to autism researchers. In the 1980s, investigators conducted the first systematic autopsy studies of the brains of autistic individuals and found that some displayed a loss of cerebellar Purkinje cells (SN: 9/6/86, p. 155). This offered a possible explanation for certain symptoms of autism, such as insensitivity to pain and oversensitivity to sounds, because the cerebellum plays a crucial role in regulating sensory information.

"We are extremely excited to see that our work supports strongly the role of secretin in the cerebellum and a connection between secretin and autism," says Chow. He notes that since this initial study, his group has found other brain areas that make both secretin and its receptor. The investigators are creating mice with mutations in the genes for each protein and will examine the animals' behavior and how their brains develop.

At this week's Society for Neuroscience meeting in San Diego, Repligen plans to present its own evidence for secretin's influence on the brain. In one study, company researchers and their academic colleagues created a radioactively labeled version of secretin, injected it into the bloodstream of mice, and documented that small quantities of the protein make it into the brains of the rodents.

This represents the first evidence that secretin can cross the blood-brain barrier that keeps so many molecules and therapeutic drugs away from the brain. It therefore supports the possibility that secretin injected into an autistic person can influence the central nervous system.

Secretin "is directly accessing the brain," says Herlihy.

In a second study, Repligen investigators injected secretin into mice and over the next 6 hours looked in the brain for signs of a protein called c-fos. "It's one of the first proteins expressed in the bucket brigade of proteins that gets turned on when a neuron is activated," says Herlihy. "We hoped to find out if there was an effect [of secretin] on any specific brain structure."

The one region of the brain that did activate neurons when secretin was injected turned out to be the amygdala. "There's a huge literature about the amygdala and autism," says Herlihy. "It did not escape our attention that this was an interesting finding."

For example, the autopsy studies that highlighted the cerebellum also found abnormalities in the amygdalas of autistic brains. This almond-shape brain structure integrates social and emotional stimuli, including such specific abilities as discerning the feelings of others and focusing attention on what others are interested in. This so-called joint attention is often absent in autistic children, which makes learning difficult.

Do the new studies in rodents confirm secretin's effectiveness as an autism therapy? Bristol-Power cautions that even if abnormalities in the brain's use of secretin underlie some cases of autism, secretin injections won't necessarily correct those problems. The damage to the brain may have already been done.

"I am not convinced myself that secretin can be used as a therapy for every autistic patient. I believe it may have effects on a very limited group of autistic patients, presumably those who at the same time have GI [gastrointestinal] disorders," adds Chow.

Repligen argues that a controversial new study, which it financed, supports the notion that there is a subset of individuals with autism that could benefit from the hormone. The largest secretin-autism test to date, it was conducted by five autism centers across the country. The trial evaluated 126 children 3 to 6 years of age who were given at 3-week intervals three injections of either secretin or a placebo.

The results are under debate. At the beginning of the study, the company had designated the Childhood Autism Rating Scale (CARS) as its primary tool for gauging autism. In a CARS evaluation, a psychologist interviews a study participant for about 90 minutes. Based on CARS scores before and after treatments, the finding was negative: Children treated with secretin didn't improve significantly more than those receiving a placebo did.

Yet a Repligen press release this summer called the trial's results positive. Instead of focusing on the CARS data, the company highlighted the favorable results achieved with other kinds of evaluations, including one in which parents evaluated their children's behavioral changes using a scale of 1 ("very much improved") to 7 ("very much worse").

Herlihy defends that decision, noting that part of the study's intent was to learn what types of evaluation best document changes in autistic behavior. He contends that parents are much more sensitive to changes in their children than are outsiders who see the kids only briefly.

The company also asserts that secretin was even more clearly beneficial to a subset of children who had normal concentrations of two gastrointestinal proteins, calprotectin and chymotrypsin, in stool samples taken at the beginning of the trial. Children with abnormal concentrations of the proteins may have greater gastrointestinal problems, causing greater variation in their autism symptoms and frustrating any attempt to evaluate secretin's effects, Herlihy says.

Autism researchers unconnected to the company worry that Repligen is doing too much interpretation of its data to show that secretin is beneficial. "If look hard at any data set, you can usually find something that correlates with something," says Geraldine Dawson of the University of Washington in Seattle.

Nevertheless, the company on Nov. 1 announced plans to conduct two more secretin trials with about 300 children. Some researchers say that they would like first to see the smaller study's data published in a scientific journal. Repligen has so far discussed those results only at research meetings, in a press release, and during a conference phone call for investors,

"Going by the published data, I could not support" a much larger trial of secretin, says Volkmar.

The secretin roller coaster ride will probably continue. Dawson and her colleagues are preparing to release the results of yet another placebo-controlled study--this one involving about 85 children.

"It will not change the story," says Dawson. "Our findings are consistent with previous studies failing to show effects."

Undeterred, Repligen is forging ahead with other secretin experiments. The company just announced that it's joining with a brain-imaging group in Boston to investigate whether secretin injections activate brain regions, such as the amygdala, in people.

"Eventually, the truth about secretin will come, whether good, bad, or indifferent," says Herlihy.