It's a reminder of the path that the Hempels have walked for eight years, of a world that few parents can imagine, until it stares them in the face.
The Hempels are on a break from the world of fluorescent lights, hard white floors and the antiseptic smells of hospital corridors. It's the world they landed in when their twin daughters were diagnosed with a rare genetic disease, a disease they were told is invariably fatal.
The Hempels met while working at Netscape, when the company was just getting off the ground. Hugh was in marketing and Chris helped run the public relations department. They left in 1999, cashing in stock options to settle down, work for themselves and raise a family.
In January 2004, they welcomed twin girls into the world. They called them Addison and Cassidy -- Addi and Cassi, for short.
The story took a dark turn when the girls were nearly 2 years old. There were unexplained fevers. A doctor found the girls had enlarged spleens. More than one doctor told the Hempels not to worry, but over the next several months, it got worse. Stumbles. Lost words. Soon, the girls were staggering.
Increasingly frantic, their parents made the rounds of specialists; eventually, tests confirmed what could truly be described as a worst-case scenario: The girls were victims of Niemann-Pick Type C, or NPC, a genetic disorder that prevents the body from properly processing cholesterol.
Most people think of cholesterol as bad stuff in food, but the reality is complicated. The body makes its own cholesterol in each and every cell; it's used to build cell membranes and to produce key proteins. We need it to live.
Cholesterol, like many other substances in the body, is processed in a compartment of each cell known as the lysosome.
"You can think of it as a kind of recycling center," says Dr. Steven Walkley, a specialist in what are known as lysosomal storage diseases, who teaches at the Albert Einstein College of Medicine in New York.
In patients with Niemann-Pick, the recycling machinery is broken. Cholesterol gradually accumulates, until it reaches toxic levels. Neurons -- nerve cells -- are particularly sensitive, although no one really knows why, says Walkley.
In patients with the mutation, the nervous system is destroyed, slowly but surely. Patients struggle with movement, gradually lose the ability to speak and eventually the ability to think. Some doctors refer to it as "childhood Alzheimer's," since patients with the most severe genetic mutation -- like Addi and Cassi -- typically die as teenagers, if not sooner.
At the time of Addi and Cassi's diagnosis, there was no approved treatment for Niemann-Pick, although one drug -- miglustat, sold as Zavesca -- was in clinical trials.
The Hempels weren't scientists, but they poured themselves into research -- parsing the language of obscure scientific papers they found on the Internet.
In December 2007, Chris spotted one in the Journal of Lipid Research. Scientists at the University of Texas Southwestern Medical School kept a colony of mice with a version of Niemann-Pick disease. When they were treated with a compound called cyclodextrin, the animals lived nearly twice as long.
Cylclodextrin is a form of sugar. It's used in consumer products like detergent and air freshener because it can bind to odor-causing molecules. In the pharmaceutical industry, it is sometimes used as an excipient: a delivery vehicle to mix with an active ingredient. No one thought of it as medicine.
The experiment at UT-Southwestern was a lucky accident, according to Dr. Benny Liu, who wrote the paper that Chris saw. A few years earlier, other scientists had noted that mice with Niemann-Pick did better if they were given a type of steroid, which happened to be mixed with cyclodextrin.
When Liu's lab tried to replicate the results, they found the mice who did best weren't the mice getting a steroid. They were the animals receiving pure cyclodextrin, what was supposed to be the inactive ingredient.
"Initially we thought we contaminated our samples, or mixed up the two drugs. And so we repeated these experiments over and over again, and got the same exact results," says Liu. "It was a blind stumble onto this discovery."
The Hempels tracked down a distributor of cyclodextrin, CTD Holdings in Alachua, Florida, and placed a bulk order. They began giving it to the twins, mixing it into drinks in their sippy cups. At times the girls seemed to improve, but it was hard to tell, since they were also taking an array of supplements and medicines, including Zavesca.
But a year after learning about cyclodextrin, Chris came across the work of Charles Vite, a veterinary researcher at the University of Pennsylvania. He was studying cats with a feline version of NPC.
Most striking to Chris were the videos Vite had recorded. Untreated cats were twitching and staggering, a terrifying reminder of what might lie in Addi and Cassi's future. But Vite had given some of the cats infusions of cyclodextrin. In the videos, those cats looked normal.
Infusions were the key. It was becoming clear that eating or drinking cyclodextrin might be ineffective, because the drug wouldn't cross the blood-brain barrier to reach the most-affected cells. After consulting with their physician -- Dr. Caroline Hastings, at Children's Hospital in Oakland -- the Hempels decided they wanted to inject cyclodextrin by infusion, into the spinal fluid.
It wouldn't be simple. Cyclodextrin wasn't a medicine; the Hempels had to order it from a bulk supplier in Florida and have it mixed and sterilized at a compounding pharmacy. What's more, you can't just go into a hospital and inject something new into a patient. The Food and Drug Administration wanted proof it was safe.
That might have been an insurmountable hurdle. Safety studies, even in animals, typically take years and cost millions of dollars. But this case was different. Cyclodextrin was already used in the anti-fungal drug Sporanox, made by Janssen Pharmaceuticals, part of Johnson and Johnson. That meant the safety studies were done, part of what's known as the drug master file.
"We captured lightning in a bottle," Chris says now.
But first, more roadblocks. Chris' initial calls to Janssen went nowhere, and she wrote an angry blog post: "Dear Johnson and Johnson, Do Kids Really Matter to You?" The very next morning, she got a call from Dr. Steven Silber, a top executive at Janssen. Silber recalls it well. "I said, 'How can I help?'"
Less than a week later, the drug master file was in the hands of the FDA. Just a few weeks after that, the twins were receiving their first infusions. This time, the results were more dramatic.
"A couple of hours after they woke up, they were more bright-eyed. I mean it was just obvious," says Hugh.
The brightness didn't last more than a day or two, but as the months went on, there were other improvements. The twins were steadier on their feet. They'd been nearly deaf, and new tests showed their hearing was better. "That just doesn't happen with a progressive neurological disease," says Chris.
Hastings agreed. "I think if the girls were not receiving cylodextrin, they would be in a much worse predicament. It's even possible they might not be with us."
Still, it wasn't good enough. The Hempels and Hastings wanted to implant a tiny pump, called an Omayya, to deliver cyclodextrin directly to the brain. The thinking went: a steady supply might be better than infusions every two weeks. Aside from that, no one wanted to face a lifetime of spinal taps twice a month.
But the FDA had more questions. The pumps were used to safely deliver other drugs, but with cyclodextrin there were unknowns. Would it degrade the plastic and cause contamination? Did the disease itself mean a higher risk of infection?
The debate went on and on. The Hempels waited.
By now, more parents were pursuing the same path as they were, getting individual approvals known as INDs, case-by-case from the FDA. Johnson and Johnson had agreed to produce the drug in its facilities, and was providing it free. At the same time, the National Institutes of Health had announced a clinical trial of cyclodextrin. No one liked to talk about it, but there was a subtle sense of competition.
It wasn't about ego. Some NIH scientists and their supporters said it was critical to support the clinical trial. Otherwise, how would anyone know if the treatment worked?
"We need to get an answer," says Dr. Marc Patterson, a leading NPC expert at the Mayo Clinic in Minnesota. "If you do it at NIH, it's controlled, and it's likely the data will help show whether or not it works. If you do it in your own child, it's hard unless you have a treatment that really reverses the symptoms."
Dr. Forbes Porter, the scientist running the NIH study, is blunt. "The feasibility of a clinical trial decreases if patients are not available or willing to participate in a formal clinical trial due to availability of INDs."
By this time, the Hempels had ruffled other feathers, including some of the other families coping with the same disease.
"Lots of parents sort of said, 'Oh my God, I would never do that,'" Hugh recalls. "Or, 'They're monsters, they're experimenting on their children.' My answer is, to each their own. If it's not a risk you want to take on for your kids, I respect that. In return, I hope you'll respect that we feel it's a worthy risk."
In early 2013, three years after they proposed the idea of implanting an Omayya pump, the FDA gave the green light and both twins underwent surgery at Children's Hospital of Oakland.
It seemed to go well -- the operations were uneventful -- but a few days later, back home, Cassi started vomiting. It turned out there was bleeding at the site of the operation, and the blood was putting pressure on her brain. She underwent an emergency operation to remove the pump, and spent six weeks in the hospital, fighting for her life.
She has largely recovered, but she's still paralyzed on her left side.
The NIH trial ran into unexpected complications, too. The very first patient came down with an infection and had to have the pump removed. The second and third patients also ran into complications.
It was somewhat mysterious. The surgery itself wasn't supposed to be difficult. "To be honest, I think it was just really bad luck," says Patterson.
After the setbacks, the NIH switched its protocol: Patients now receive cyclodextrin through spinal taps, much as the Hempels did for years. The safety looks good enough that the NIH is now planning for a larger study, and Porter says they've been able to gather useful data about NPC.
Meanwhile, about a dozen patients in the United States are receiving cyclodextrin infusions outside the study, through individual, FDA-approved treatment plans. There are also several children receiving treatment overseas, in Brazil, Spain and Japan.
Cassi is back to getting infusions every other week. Addi, who still has the pump in her brain, has done better. For the Hempels, the story is bittersweet. It's clear that much of the damage is permanent.
On the other hand, all the hard work and struggle almost certainly paid off. Aside from improvements in hearing and alertness, Addi and Cassi have avoided the lung problems and other complications that many Niemann-Pick children suffer.
Without cyclodextrin, says Chris, "There's no question in my mind they wouldn't be here now."
The Hempels are testing new computer technology to help the twins communicate better with the outside world. They're experimenting with medical marijuana to better control the girls' seizures. They still scour the scientific literature.
As the family sits together, watching the fall sunlight dissolve into shadows over the garden, it's just a rest. The story continues.